From 61fb778b1cd69053c54460b0c46da32241f044ac Mon Sep 17 00:00:00 2001 From: vr8hub Date: Sun, 27 Aug 2023 16:27:12 -0600 Subject: [PATCH 1/9] Move tests to functions by type (syntax, typography, etc.) --- se/se_epub_lint.py | 4438 +++++++++++++++++++++++--------------------- 1 file changed, 2311 insertions(+), 2127 deletions(-) diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py index e97987d5..ec70de60 100644 --- a/se/se_epub_lint.py +++ b/se/se_epub_lint.py @@ -528,6 +528,68 @@ def __init__(self, section_id: str, depth: int, has_header: bool): # Wealth of Nations by Adam Smith self.depth = depth if depth <= 6 else 6 +def _build_section_tree(self) -> List[EbookSection]: + """ + Helper function used in self.lint() + Walk the files in spine order to build a tree of sections and heading levels. + + INPUTS + self + + OUTPUTS + section_tree, A list of EbookSections + """ + + # section_tree is a list of EbookSections. + # For example, The Woman in White by Wilkie Collins would have a section_tree that looks like this: + # titlepage (1) + # imprint (2) + # dedication (2) + # preface (2) + # halftitlepage (2) + # part-1 (2) + # division-1-1 (3) + # chapter-1-1-1 (4) + # chapter-1-1-2 (4) + # chapter-1-1-3 (4) + # chapter-1-1-4 (4) + section_tree: List[EbookSection] = [] + + for filename in self.spine_file_paths: + dom = self.get_dom(filename) + for dom_section in dom.xpath("/html/body//*[re:test(name(), '^(section|article|nav)$')][@id]"): + # Start at h2 by default except for the titlepage, which is always h1 + starting_depth = 1 if regex.findall(r"\btitlepage\b", dom_section.get_attr("epub:type") or "") else 2 + + section = _find_ebook_section(dom_section.get_attr("id"), section_tree) + + section_parent_id = dom_section.get_attr("data-parent") + + has_header = bool(dom_section.xpath("./*[re:test(name(), '^h[1-6]$')]")) + if not has_header: + has_header = bool(dom_section.xpath(".//*[not(re:test(name(), '^(section|article|nav)$'))]//*[re:test(name(), '^h[1-6]$')]")) + + if not section_parent_id: + # We don't have a data-parent, but do we have a direct parent section in the same file as this section? + direct_parent = dom_section.xpath("./ancestor::*[re:test(name(), '^(section|article|nav)$')][@id][1]") + + if direct_parent: + section_parent_id = direct_parent[0].get_attr("id") + + # If we don't have the section in our list yet, and it has a parent, append it to the parent + if not section and section_parent_id: + parent = _find_ebook_section(section_parent_id, section_tree) + if parent: + # Only increment the depth if the parent has a header. If it doesn't then we don't want to increment, + # because we will end up skipping an h# level (like h2 -> h4 instead of h2 -> h3) + parent.children.append(EbookSection(dom_section.get_attr("id"), parent.depth + 1 if parent.has_header else parent.depth, has_header)) + + # If we don't have it in our list yet, and it has no parent, append it to the top + elif not section: + section_tree.append(EbookSection(dom_section.get_attr("id"), starting_depth, has_header)) + + return section_tree + def _find_ebook_section(section_id: str, sections: List[EbookSection]) -> Union[EbookSection, None]: """ Find an ebook section in a tree of sections. @@ -690,7 +752,7 @@ def files_not_in_spine(self) -> set: INPUTS None - OUTPITS + OUTPUTS Set of files not in the spine (typically an empty set) """ @@ -698,2532 +760,2640 @@ def files_not_in_spine(self) -> set: spine_files = set(self.spine_file_paths + [self.toc_path]) return xhtml_files.difference(spine_files) -def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = None) -> list: +def _update_missing_styles(filename: str, dom: se.easy_xml.EasyXmlTree, local_css: dict) -> list: """ - Check this ebook for some common SE style errors. + Identify any missing CSS styles in the file being checked INPUTS - None + filename: the name of the file being checked + dom: The dom of the file being checked + local_css: Dictionary containing several flags concerning the local CSS OUTPUTS - A list of LintMessage objects. + Set of files not in the spine (typically an empty set) """ - local_css_path = self.content_path / "css/local.css" - messages: List[LintMessage] = [] - has_halftitle = False - has_frontmatter = False - has_cover_source = False - cover_svg_title = "" - titlepage_svg_title = "" - xhtml_css_classes: Dict[str, int] = {} - headings: List[tuple] = [] - double_spaced_files: List[Path] = [] - unused_selectors: List[str] = [] - id_attrs: List[str] = [] - missing_metadata_elements = [] - abbr_elements_requiring_css: List[se.easy_xml.EasyXmlElement] = [] - metadata_xml = self.metadata_dom.to_string() - has_glossary_search_key_map = False - glossary_usage = [] - has_images = False + missing_styles: List[str] = [] - # This is a dict with where keys are the path and values are a list of code dicts. - # Each code dict has a key "code" which is the actual code, and a key "used" which is a - # bool indicating whether or not the code has actually been caught in the linting run. - ignored_codes: Dict[str, List[Dict]] = {} + if not local_css["has_elision_style"]: + missing_styles += [node.to_tag_string() for node in dom.xpath("/html/body//span[contains(@class, 'elision')]")] - # Cache the browser default stylesheet for later use - with importlib_resources.open_text("se.data", "browser.css", encoding="utf-8") as css: - self._file_cache["default"] = css.read() # pylint: disable=protected-access + # Check to see if we included poetry or verse without the appropriate styling + if filename.name not in IGNORED_FILENAMES: + nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')][./p/span]") + for node in nodes: + if "z3998:poem" in node.get_attr("epub:type") and not local_css["has_poem_style"]: + missing_styles.append(node.to_tag_string()) - # First, check if we have an se-lint-ignore.xml file in the ebook root. If so, parse it. For an example se-lint-ignore file, see semos://1.0.0/2.3 - lint_ignore_path = self.path / "se-lint-ignore.xml" - if not skip_lint_ignore and lint_ignore_path.exists(): - lint_config = self.get_dom(lint_ignore_path) + if "z3998:verse" in node.get_attr("epub:type") and not local_css["has_verse_style"]: + missing_styles.append(node.to_tag_string()) - elements = lint_config.xpath("/se-lint-ignore/file") + if "z3998:song" in node.get_attr("epub:type") and not local_css["has_song_style"]: + missing_styles.append(node.to_tag_string()) - if not elements: - messages.append(LintMessage("m-049", "No [path]se-lint-ignore.xml[/] rules. Delete the file if there are no rules.", se.MESSAGE_TYPE_ERROR, lint_ignore_path)) + if "z3998:hymn" in node.get_attr("epub:type") and not local_css["has_hymn_style"]: + missing_styles.append(node.to_tag_string()) - has_illegal_path = False + if "z3998:lyrics" in node.get_attr("epub:type") and not local_css["has_lyrics_style"]: + missing_styles.append(node.to_tag_string()) - if not allowed_messages: - allowed_messages = [] + # Check frontmatter for missing styling + nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'dedication|epigraph')]") + for node in nodes: + if "dedication" in node.get_attr("epub:type") and not local_css["has_dedication_style"]: + missing_styles.append(node.to_tag_string()) - for element in elements: - path = element.get_attr("path").strip() + if "epigraph" in node.get_attr("epub:type") and not local_css["has_epigraph_style"]: + missing_styles.append(node.to_tag_string()) - if path == "*": - has_illegal_path = True # Set a bool so that we set a lint error later, to prevent adding it multiple times + return missing_styles - if path not in ignored_codes: - ignored_codes[path] = [] +def _lint_special_file(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_info: dict, special_file: str) -> list: + """ + Process error checks in “special” .xhtml files - for ignore in element.lxml_element: - if ignore.tag == "ignore": - has_reason = False - for child in ignore: - if child.tag == "code" and child.text.strip() not in allowed_messages: - ignored_codes[path].append({"code": child.text.strip(), "used": False}) + INPUTS + filename: The name of the file being checked + dom: The dom of the file being checked + file_contents: The contents of the file being checked + ebook_info: A dictionary containing several pieces of information about the ebook + special_file: A string identifying the type of special file being checked + self - if child.tag == "reason" and child.text.strip() != "": - has_reason = True + OUTPUTS + A list of LintMessages representing errors found in the special file + """ - if not has_reason: - messages.append(LintMessage("m-046", "Missing or empty [xml][/] element.", se.MESSAGE_TYPE_ERROR, lint_ignore_path)) + messages = []; + source_links = self.metadata_dom.xpath("/package/metadata/dc:source/text()") - if has_illegal_path: - messages.append(LintMessage("m-047", "Ignoring [path]*[/] is too general. Target specific files if possible.", se.MESSAGE_TYPE_WARNING, lint_ignore_path)) + if special_file in ("colophon,imprint"): + if len(source_links) <= 2: + # Check that links back to sources are represented correctly + nodes = dom.xpath("/html/body//a[@href='https://www.pgdp.net' and text()!='The Online Distributed Proofreading Team']") + if nodes: + messages.append(LintMessage("m-071", "DP link must be exactly [text]The Online Distributed Proofreading Team[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + nodes = dom.xpath("/html/body//a[re:test(@href, '^https://www.pgdp.org/ols/') and text()!='Distributed Proofreaders Open Library System']") + if nodes: + messages.append(LintMessage("m-072", "DP OLS link must be exactly [text]Distributed Proofreaders Open Library System[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + nodes = dom.xpath("/html/body//a[re:test(@href, '^https://[^\"]*?hathitrust.org') and re:test(text(), '[Hh]athi') and not(text()='HathiTrust Digital Library')]") + if nodes: + messages.append(LintMessage("m-041", "Hathi Trust link text must be exactly [text]HathiTrust Digital Library[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + nodes = dom.xpath("/html/body//p[re:test(., '\\ba transcription\\b') and ./a[contains(@href, '#transcriptions')]]") + if nodes: + messages.append(LintMessage("t-071", "Multiple transcriptions listed, but preceding text is [text]a transcription[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + if special_file == "colophon": + # Check for illegal Wikipedia URLs + nodes = dom.xpath("/html/body//a[contains(@href, '.m.wikipedia.org')]") + if nodes: + messages.append(LintMessage("m-026", "Illegal [url]https://*.m.wikipedia.org[/] URL. Hint: use non-mobile Wikipedia URLs.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + # Check for non-English Wikipedia URLs + nodes = dom.xpath("/html/body//a[re:test(@href, 'https://(?!en)[a-z]{2,}\\.wikipedia\\.org')]") + if nodes: + messages.append(LintMessage("m-043", "Non-English Wikipedia URL.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes])) + + # Check for wrong grammar filled in from template + nodes = dom.xpath("/html/body//a[starts-with(@href, 'https://books.google.com/') or starts-with(@href, 'https://www.google.com/books/')][(preceding-sibling::text()[normalize-space(.)][1])[re:test(., '\\bthe$')]]") + if nodes: + messages.append(LintMessage("s-016", "Incorrect [text]the[/] before Google Books link.", se.MESSAGE_TYPE_ERROR, filename, ["the
\n" + node.to_string() for node in nodes])) + + se_url = self.generated_identifier.replace("url:", "") + if not dom.xpath(f"/html/body//a[@href='{se_url}' and text()='{se_url.replace('https://', '')}']"): + messages.append(LintMessage("m-035", f"Unexpected S.E. identifier in colophon. Expected: [url]{se_url}[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if self.metadata_dom.xpath("/package/metadata/meta[@property='role' and text()='trl']") and "translated from" not in file_contents: + messages.append(LintMessage("m-025", "Translator found in metadata, but no [text]translated from LANG[/] block in colophon.", se.MESSAGE_TYPE_ERROR, filename)) + + if ebook_info["has_multiple_transcriptions"] and not dom.xpath("/html/body//a[contains(@href, '#transcriptions')]"): + messages.append(LintMessage("m-074", "Multiple transcriptions found in metadata, but no link to [text]EBOOK_URL#transcriptions[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if ebook_info["has_multiple_page_scans"] and not dom.xpath("/html/body//a[contains(@href, '#page-scans')]"): + messages.append(LintMessage("m-075", "Multiple page scans found in metadata, but no link to [text]EBOOK_URL#page-scans[/].", se.MESSAGE_TYPE_ERROR, filename)) + + # Check that the formula changed from the default if we added 'various sources' + if ebook_info["has_multiple_transcriptions"] or ebook_info["has_multiple_page_scans"]: + nodes = dom.xpath("/html/body//a[text() = 'various sources' and not(re:test(preceding-sibling::br[1]/preceding-sibling::node()[1], '(digital scans|transcriptions) from\\s*$'))]") + if nodes: + messages.append(LintMessage("t-072", "[text]various sources[/] link not preceded by [text]from[/].", se.MESSAGE_TYPE_ERROR, filename)) + + # Check if we forgot to fill any variable slots + missing_colophon_vars = [var for var in SE_VARIABLES if regex.search(fr"\b{var}\b", file_contents)] + if missing_colophon_vars: + messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, filename, missing_colophon_vars)) + + # Check that we have
s at the end of lines + # First, check for b or a elements that are preceded by a newline but not by a br + nodes = [node.to_string() for node in dom.xpath("/html/body/section/p/*[name()='b' or name()='a'][(preceding-sibling::node()[1])[contains(., '\n')]][not((preceding-sibling::node()[2])[self::br]) or (normalize-space(preceding-sibling::node()[1]) and re:test(preceding-sibling::node()[1], '\\n\\s*$')) ]")] + # Next, check for text nodes that contain newlines but are not preceded by brs + nodes = nodes + [node.strip() for node in dom.xpath("/html/body/section/p/text()[contains(., '\n') and normalize-space(.)][(preceding-sibling::node()[1])[not(self::br)]]")] + if nodes: + messages.append(LintMessage("s-053", "Colophon line not preceded by [xhtml]
[/].", se.MESSAGE_TYPE_ERROR, filename, nodes)) + + # Is there a comma after a producer name, if there's only two producers? + nodes = dom.xpath("/html/body/section/p/*[name()='b' or name()='a'][(following-sibling::node()[1])[normalize-space(.)=', and']][(preceding-sibling::*[1])[name()='br']]") + if nodes: + messages.append(LintMessage("t-006", "Comma after producer name, but there are only two producers.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and re:test(@epub:type, 'z3998:.*?name')]") + if nodes: + messages.append(LintMessage("s-092", "Anonymous contributor with [val]z3998:*-name[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + # Check for anonymous volunteers misrepresented in the colophon. Note that we only match SE producers and transcribers, because + # a cover art artist can also be anonymous but they're not volunteers. + nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and text() != 'An Anonymous Volunteer' and (preceding-sibling::a[@href='https://standardebooks.org'])]") + if nodes: + messages.append(LintMessage("s-100", "Anonymous digital contributor value not exactly [text]An Anonymous Volunteer[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + # Check for primary contributors misrepresented in the colophon. these differ from ebook production volunteers. + nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and text() != 'Anonymous' and (preceding-sibling::node()[contains(., 'The cover page')] or preceding-sibling::i[contains(@epub:type, 'se:name.publication')])]") + if nodes: + messages.append(LintMessage("s-101", "Anonymous primary contributor value not exactly [text]Anonymous[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) + + # Is there a page scan link in the colophon, but missing in the metadata? + for node in dom.xpath("/html/body//a[re:test(@href, '(gutenberg\\.org/ebooks/[0-9]+|hathitrust\\.org|/archive\\.org|books\\.google\\.com|www\\.google\\.com/books/)')]"): + if not self.metadata_dom.xpath(f"/package/metadata/dc:source[contains(text(), {se.easy_xml.escape_xpath(node.get_attr('href'))})]"): + messages.append(LintMessage("m-059", f"Link to [url]{node.get_attr('href')}[/] found in colophon, but missing matching [xhtml]dc:source[/] element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + + # Are the sources represented correctly? + # We don't have a standard yet for more than two sources (transcription and scan) so just ignore that case for now. + # We can't merge this with the imprint check because imprint doesn't have `
` between `the` + if not ebook_info["has_multiple_transcriptions"] and not ebook_info["has_other_sources"]: + for link in source_links: + if "gutenberg.org" in link and f"Project Gutenberg" not in file_contents: + messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]Project Gutenberg[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if not ebook_info["has_multiple_page_scans"] and not ebook_info["has_other_sources"]: + for link in source_links: + if "hathitrust.org" in link and f"the
\n\t\t\tHathiTrust Digital Library" not in file_contents: + messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]the
HathiTrust Digital Library[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if "archive.org" in link and f"the
\n\t\t\tInternet Archive" not in file_contents: + messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]the
Internet Archive[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if ("books.google.com" in link or "www.google.com/books/" in link) and f"Google Books" not in file_contents: + messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]Google Books[/].", se.MESSAGE_TYPE_ERROR, filename)) + + # If we're in the imprint, are the sources represented correctly? + # We don't have a standard yet for more than two sources (transcription and scan) so just ignore that case for now. + elif special_file == "imprint": + # Check for wrong grammar filled in from template + nodes = dom.xpath("/html/body//a[starts-with(@href, 'https://books.google.com/') or starts-with(@href, 'https://www.google.com/books/')][(preceding-sibling::node()[1])[re:test(., 'the\\s+$')]]") + if nodes: + messages.append(LintMessage("s-016", "Incorrect [text]the[/] before Google Books link.", se.MESSAGE_TYPE_ERROR, filename, ["the " + node.to_string() for node in nodes])) + + # Check if we forgot to fill any variable slots + missing_imprint_vars = [var for var in SE_VARIABLES if regex.search(fr"\b{var}\b", file_contents)] + if missing_imprint_vars: + messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, filename, missing_imprint_vars)) + + if ebook_info["has_multiple_transcriptions"] and not dom.xpath("/html/body//a[contains(@href, '#transcriptions')]"): + messages.append(LintMessage("m-074", "Multiple transcriptions found in metadata, but no link to [text]EBOOK_URL#transcriptions[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if ebook_info["has_multiple_page_scans"] and not dom.xpath("/html/body//a[contains(@href, '#page-scans')]"): + messages.append(LintMessage("m-075", "Multiple page scans found in metadata, but no link to [text]EBOOK_URL#page-scans[/].", se.MESSAGE_TYPE_ERROR, filename)) + + # Check that the formula changed from the default if we added 'various sources' + if ebook_info["has_multiple_transcriptions"] or ebook_info["has_multiple_page_scans"]: + nodes = dom.xpath("/html/body//a[text() = 'various sources' and not(re:test(preceding-sibling::node()[1], '(digital scans|transcriptions) from\\s*$'))]") + if nodes: + messages.append(LintMessage("t-072", "[text]various sources[/] link not preceded by [text]from[/].", se.MESSAGE_TYPE_ERROR, filename)) + + # Check for correctly named links. We can't merge this with the colophon check because the colophon breaks `the` with `
` + if not ebook_info["has_multiple_transcriptions"] and not ebook_info["has_other_sources"]: + for link in source_links: + if "gutenberg.org" in link and f"Project Gutenberg" not in file_contents: + messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]Project Gutenberg[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if not ebook_info["has_multiple_page_scans"] and not ebook_info["has_other_sources"]: + for link in source_links: + if "hathitrust.org" in link and f"the HathiTrust Digital Library" not in file_contents: + messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: the [xhtml]HathiTrust Digital Library[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if "archive.org" in link and f"the Internet Archive" not in file_contents: + messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: the [xhtml]Internet Archive[/].", se.MESSAGE_TYPE_ERROR, filename)) + + if ("books.google.com" in link or "www.google.com/books/" in link) and f"Google Books" not in file_contents: + messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]Google Books[/].", se.MESSAGE_TYPE_ERROR, filename)) + + + # Endnote checks + elif special_file == "endnotes": + # Do we have to replace Ibid.? Only match Ibid. if the endnote does not appear to cite any names. + nodes = dom.xpath("/html/body//li[re:test(@epub:type, '\\bendnote\\b')]//abbr[re:test(., '\\b[Ii]bid\\b') and not(ancestor::li[1]//*[contains(@epub:type, 'se:name')])]") + if nodes: + messages.append(LintMessage("s-039", "[text]Ibid[/] in endnotes. “Ibid” means “The previous reference” which is meaningless with popup endnotes, and must be replaced by the actual thing [text]Ibid[/] refers to, unless it refers to text within the same endnote.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) + + # Check that endnotes have their backlink in the last

element child of the

  • . This also highlights backlinks that are totally missing. + nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')][./p[last()][not(a[contains(@epub:type, 'backlink')])] or not(./p[last()])]") + if nodes: + messages.append(LintMessage("s-056", "Last [xhtml]

    [/] child of endnote missing backlink.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) + + # Make sure the backlink points to the same note number as the parent endnote ID + nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]//a[contains(@epub:type, 'backlink')][not(re:match(@href, '\\-[0-9]+$')=re:match(ancestor::li/@id, '\\-[0-9]+$'))]") + if nodes: + messages.append(LintMessage("s-057", "Backlink noteref fragment identifier doesn’t match endnote number.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) + + # Check that citations at the end of endnotes are in a element. If not typogrify will run the last space together with the em dash. + # This tries to catch that, but limits the match to 20 chars so that we don't accidentally match a whole sentence that happens to be at the end of an endnote. + nodes = dom.xpath(f"/html/body//li[contains(@epub:type, 'endnote')]/p[last()][re:test(., '\\.”?{se.WORD_JOINER}?—[A-Z].{{0,20}}\\s*↩$')]") + if nodes: + messages.append(LintMessage("s-064", "Endnote citation not wrapped in [xhtml][/]. Em dashes go within [xhtml][/] and it is preceded by one space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) + + # Did we forget the endnote semantic on li elements? + nodes = dom.xpath("/html/body/section/ol/li[not(re:test(@epub:type, '\\bendnote\\b'))]") + if nodes: + messages.append(LintMessage("s-099", "List item in endnotes missing [xhtml]endnote[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) + + # Match backlink elements whose preceding node doesn't end with ' ', and is also not all whitespace + nodes = dom.xpath("/html/body//a[@epub:type='backlink'][(preceding-sibling::node()[1])[not(re:test(., ' $')) and not(normalize-space(.)='')]]") + if nodes: + messages.append(LintMessage("t-027", "Endnote backlink not preceded by exactly one space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) + + # Check LoI descriptions to see if they match associated figcaptions + elif special_file == "loi": + for node in dom.xpath("/html/body/nav[contains(@epub:type, 'loi')]//li//a"): + figure_ref = node.get_attr("href").split("#")[1] + chapter_ref = regex.findall(r"(.*?)#.*", node.get_attr("href"))[0] + figcaption_text = "" + loi_text = node.inner_text() + file_dom = self.get_dom(self.content_path / "text" / chapter_ref) - # Done parsing ignore list + try: + figure = file_dom.xpath(f"//*[@id={se.easy_xml.escape_xpath(figure_ref)}]")[0] + except Exception: + messages.append(LintMessage("s-040", f"[attr]#{figure_ref}[/] not found in file [path][link=file://{self.path / 'src/epub/text' / chapter_ref}]{chapter_ref}[/][/].", se.MESSAGE_TYPE_ERROR, filename)) + continue - # Check that the spine has all the expected files in the book - missing_spine_files = files_not_in_spine(self) - if missing_spine_files: - for missing_spine_file in missing_spine_files: - messages.append(LintMessage("f-007", "File not listed in [xml][/].", se.MESSAGE_TYPE_ERROR, missing_spine_file)) + for child in figure.xpath("./*"): + if child.tag == "img": + figure_img_alt = child.get_attr("alt") - # Get the ebook language for later use - try: - language = self.metadata_dom.xpath("/package/metadata/dc:language")[0].text - except se.InvalidXmlException as ex: - raise ex - except Exception as ex: - raise se.InvalidSeEbookException(f"Missing [xml][/] element in [path][link=file://{self.metadata_file_path}]{self.metadata_file_path.name}[/][/].") from ex + if child.tag == "figcaption": + figcaption_text = child.inner_text() - # Check local.css for various items, for later use - try: - self.local_css = self.get_file(local_css_path) - except Exception as ex: - raise se.InvalidSeEbookException(f"Couldn’t open [path]{local_css_path}[/].") from ex + # Replace tabs and newlines with a single space to better match figcaptions that contain
    + figcaption_text = regex.sub(r"(\n|\t)", " ", figcaption_text) + figcaption_text = regex.sub(r"[ ]+", " ", figcaption_text) - # cssutils prints warnings/errors to stdout by default, so shut it up here - cssutils.log.enabled = False + if (figcaption_text != "" and loi_text != "" and figcaption_text != loi_text) and (figure_img_alt != "" and loi_text != "" and figure_img_alt != loi_text): + messages.append(LintMessage("s-041", f"The [xhtml]

    [/] element of [attr]#{figure_ref}[/] does not match the text in its LoI entry.", se.MESSAGE_TYPE_WARNING, self.path / "src/epub/text" / chapter_ref)) - # Get the css rules and selectors from helper function - local_css_rules, duplicate_selectors = _get_selectors_and_rules(self) + return messages - if duplicate_selectors: - messages.append(LintMessage("c-009", "Duplicate CSS selectors. Duplicates are only acceptable if overriding S.E. base styles.", se.MESSAGE_TYPE_WARNING, local_css_path, list(set(duplicate_selectors)))) +def _lint_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, is_titlepage: bool, special_file: str) -> list: + """ + Helper function used in self.lint() + Process typo checks on an .xhtml file - # Store a list of CSS selectors, and duplicate it into a list of unused selectors, for later checks - # We use a regex to remove pseudo-elements like ::before, because we want the *selectors* to see if they're unused. - local_css_selectors = [regex.sub(r"::[\p{Lowercase_Letter}\-]+", "", selector) for selector in local_css_rules] - unused_selectors = local_css_selectors.copy() + INPUTS + filename: The name of the file being checked + dom: A dom tree to check + file_contents: The contents of the file being checked - local_css_has_poem_style = False - local_css_has_verse_style = False - local_css_has_song_style = False - local_css_has_hymn_style = False - local_css_has_lyrics_style = False - local_css_has_elision_style = False - local_css_has_dedication_style = False - local_css_has_epigraph_style = False - short_story_count = 0 - missing_styles: List[str] = [] - directories_not_url_safe = [] - files_not_url_safe = [] - id_values = {} - duplicate_id_values = [] - ebook_has_subtitle = bool(self.metadata_dom.xpath("/package/metadata/meta[@property='title-type' and text()='subtitle']")) - source_links = self.metadata_dom.xpath("/package/metadata/dc:source/text()") - section_tree: List[EbookSection] = [] - ebook_has_multiple_transcriptions = False - ebook_has_multiple_page_scans = False + OUTPUTS + A list of LintMessages representing possible typos found in the file + """ - # Iterate over rules to do some other checks - abbr_with_whitespace = [] - for selector, rules in local_css_rules.items(): - if "z3998:poem" in selector: - local_css_has_poem_style = True + messages = []; - if "z3998:verse" in selector: - local_css_has_verse_style = True + # Check for common typos + if not is_titlepage: + # Don't check the titlepage because it has a standard format and may raise false positives + typos = regex.findall(r"(?“Bah!” he said to the ‘minister.’

    + # We can't do this on xpath because we can't iterate over the output of re:replace(). + # A fully-featured solution would parse quotes to find balanced pairs. This quick-n-dirty + # solution takes the raw XHTML string, strips any seemingly-valid balanced pairs using regex, + # then searches for unmatched `‘`. If found, it returns the dom node of the match; we can't return + # the actual string match because we stripped surrounding text. + + # Start by merging all

    into single lines, including poetry, but exclude blockquotes as they have special quoting rules + temp_xhtml = "" + dom_copy = deepcopy(dom) + node_number = 0 + # Add IDs to the dom so we can find the paragraphs after we run our regex + # Exclude paragraphs in blockquotes, which may have special quoting rules, and "continued" paragraphs, which may be continued dialog without an “ + for node in dom_copy.xpath("/html/body//p[not(ancestor::blockquote) and not(contains(@class, 'continued'))]"): + node.set_attr("id", "lint-" + str(node_number)) + temp_xhtml = temp_xhtml + f"

    " + regex.sub(r"[\s\n]+", " ", node.inner_text(), flags=regex.DOTALL) + "\n" + node_number = node_number + 1 + + replacement_count = 1 + while replacement_count > 0: + # Remove all regular quotes. Run this in a loop because we may need to remove triple-nested quotes. + (temp_xhtml, replacement_count) = regex.subn(r"“[^“]+?”", " ", temp_xhtml) # Remove all regular quotes + + # Remove contractions to reduce rsquo for next regex + temp_xhtml = regex.sub(r"[\p{Letter}]’[\p{Letter}]", " ", temp_xhtml, flags=regex.MULTILINE) + + # Remove all runs of ldquo that are likely to spill to the next

    + replacement_count = 1 + while replacement_count > 0: + (temp_xhtml, replacement_count) = regex.subn(r"“[^“”]+?$", " ", temp_xhtml, flags=regex.MULTILINE) + + # Match problem `‘` using regex, and if found, get the actual node text from the dom to return. + typos = [] + for match in regex.findall(r"""

    .*‘[\p{Letter}\s]""", temp_xhtml): + for node in dom_copy.xpath(f"/html/body//p[@id='lint-{match}' and re:test(., '‘[A-Za-z\\s]')]"): + typos.append(node.inner_text()) + + if typos: + messages.append(LintMessage("y-007", "Possible typo: [text]‘[/] not within [text]“[/]. Hints: Should [text]‘[/] be replaced with [text]“[/]? Is there a missing closing quote? Is this a nested quote that should be preceded by [text]“[/]? Are quotes in close proximity correctly closed?", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for single quotes when there should be double quotes in an interjection in dialog + typos = [node.to_string() for node in dom.xpath("//p[re:test(., '“[^”]+?’⁠—[^”]+?—“')]")] + if typos: + messages.append(LintMessage("y-008", "Possible typo: dialog interrupted by interjection but with incorrect closing quote.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for dialog starting with a lowercase letter. Only check the first child text node of

    , because other first children might be valid lowercase, like or ; + # exclude

    inside or preceded by

    ; and exclude

    inside endnotes, as definitions may start with lowercase letters. + typos = [node.to_string() for node in dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::li[contains(@epub:type, 'endnote')]) and not(preceding-sibling::*[1][name()='blockquote'])][re:test(./node()[1], '^“[a-z]')]")] + if typos: + messages.append(LintMessage("y-009", "Possible typo: dialog begins with lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + typos = [node.to_string() for node in dom.xpath("//p[not(ancestor::blockquote) and not(following-sibling::*[1][name() = 'blockquote' or contains(@class, 'continued')]) and re:test(., ',”$')]")] + if typos: + messages.append(LintMessage("y-010", "Possible typo: comma ending dialogue.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '’{2,}')]")] + if typos: + messages.append(LintMessage("y-011", "Possible typo: two or more [text]’[/] in a row.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '”[a-z]')]")] + if typos: + messages.append(LintMessage("y-012", "Possible typo: [text]”[/] directly followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for period not inside rsquo + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[^”’]+?’[\\.,]”')]")] + if typos: + messages.append(LintMessage("y-013", "Possible typo: punctuation not within [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for period before dialog tag; try to exclude abbrevations that close a quotation, like `“Mr.”`. + typos = [node.to_string() for node in dom.xpath("/html/body//p[(re:test(., '\\.”\\s[a-z\\s]*?(\\bsaid|[a-z]+ed)') or re:test(., '\\.”\\s(s?he|they?|and)\\b')) and not(.//abbr[following-sibling::node()[re:test(., '^”')]])]")] + if typos: + messages.append(LintMessage("y-014", "Possible typo: Unexpected [text].[/] at the end of quotation. Hint: If a dialog tag follows, should this be [text],[/]?", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for mis-curled ‘ or ‘ without matching ’ + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[A-Za-z][^“’]+?”')]")] + if typos: + messages.append(LintMessage("y-015", "Possible typo: mis-curled [text]‘[/] or missing [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for two periods in a row, almost always a typo for one period or a hellip + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\.\\.[^\\.]')]")] + if typos: + messages.append(LintMessage("y-016", "Possible typo: consecutive periods ([text]..[/]).", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for ldquo followed by space + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '“\\s+[^‘’]')]")] + if typos: + messages.append(LintMessage("y-017", "Possible typo: [text]“[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for lsquo followed by space + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘\\s+[^“’]')]")] + if typos: + messages.append(LintMessage("y-018", "Possible typo: [text]‘[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for closing rdquo without opening ldquo. We ignore blockquotes because they usually have unique quote formatting. + # Remove tds in case rdquo means "ditto mark" + typos = regex.findall(r"”[^“‘]+?”", regex.sub(r"[”\s]+?(.+?)?", "", file_contents), flags=regex.DOTALL) + + # We create a filter to try to exclude nested quotations + # Remove tags in case they're enclosing punctuation we want to match against at the end of a sentence. + typos = [match for match in typos if not regex.search(r"(?:[\.!\?;…—]|”\s)’\s", se.formatting.remove_tags(match))] + + # Try some additional matches before adding the lint message + # Search for

    tags that have an ending closing quote but no opening quote; but exclude

    s that are preceded by a

    + # or that have a
    ancestor, because that may indicate that the opening quote is elsewhere in the quotation. + for node in dom.xpath("//p[re:test(., '^[^“]+”') and not(./preceding-sibling::*[1][name() = 'blockquote']) and not(./ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]) and not(./ancestor::blockquote)]"): + typos.append(node.to_string()[-20:]) + + if typos: + messages.append(LintMessage("y-019", "Possible typo: [text]”[/] without opening [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for ,. + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., ',\\.')]")] + if typos: + messages.append(LintMessage("y-020", "Possible typo: consecutive comma-period ([text],.[/]).", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for single quotes instead of double quotes + typos = [node.to_string() for node in dom.xpath("/html/body//p[not(contains(@class, 'continued') or ancestor::blockquote or ancestor::*[re:test(@epub:type, 'z3998:(verse|song|poem|hymn)')]) and re:test(., '^[^“]+?‘')]")] + if typos: + messages.append(LintMessage("y-021", "Possible typo: Opening [text]‘[/] without preceding [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for two quotations in one paragraph + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^”]+?”\\s“[^”]+?”$')]")] + if typos: + messages.append(LintMessage("y-022", "Possible typo: consecutive quotations without intervening text, e.g. [text]“…” “…”[/].", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for incorrectly nested quotation marks + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘”]+“') and not(.//br)]")] + if typos: + messages.append(LintMessage("y-023", "Possible typo: two opening quotation marks in a run. Hint: Nested quotes should switch between [text]“[/] and [text]‘[/]", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for dashes instead of em-dashes + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\s[a-z]+-(the|there|is|and|they|when)\\s')]")] + if typos: + messages.append(LintMessage("y-024", "Possible typo: dash before [text]the/there/is/and/they/when[/] probably should be em-dash.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for comma not followed by space but followed by quotation mark + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '[a-z],[“”‘’][a-z]', 'i')]")] + if typos: + messages.append(LintMessage("y-025", "Possible typo: comma without space followed by quotation mark.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for punctuation missing before conjunctions. Ignore

    with an child starting in a conjunction, as those are probably book titles or non-English languages + typos = [node.to_string() for node in dom.xpath(f"/html/body//p[not(parent::hgroup) and re:test(., '\\b[a-z]+\\s(But|And|For|Nor|Yet|Or)\\b[^’\\.\\?\\-{se.WORD_JOINER}]') and not(./i[re:test(., '^(But|And|For|Nor|Yet|Or)\\b')])]")] + if typos: + messages.append(LintMessage("y-026", "Possible typo: no punctuation before conjunction [text]But/And/For/Nor/Yet/Or[/].", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for extra closing single quote at the end of dialog + typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘]+”\\s*’$')]")] + if typos: + messages.append(LintMessage("y-027", "Possible typo: Extra [text]’[/] at end of paragraph.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for `` preceded or followed by text. Ignore compass directions followed by `ly`, like S.S.W.ly + typos = [node.to_string() for node in dom.xpath("/html/body//abbr[(preceding-sibling::node()[1])[re:test(., '[A-Za-z]$')] or (following-sibling::node()[1])[re:test(., '^[A-Za-z](?[/] directly preceded or followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for misapplied italics. Ignore 's' because the plural is too common. + typos = [node.to_string() for node in dom.xpath("/html/body//*[(name() = 'i' or name() = 'em') and ./following-sibling::node()[1][re:test(., '^[a-z]\\b', 'i') and not(re:test(., '^s\\b'))]]")] + if typos: + messages.append(LintMessage("y-029", "Possible typo: Italics followed by a letter.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for lowercase letters starting quotations after a preceding period + typos = dom.xpath("/html/body//p/child::text()[re:test(., '\\.\\s[‘“][a-z]')]") + if typos: + messages.append(LintMessage("y-030", "Possible typo: Lowercase quotation following a period. Check either that the period should be a comma, or that the quotation should start with a capital.", se.MESSAGE_TYPE_WARNING, filename, typos)) + + # Check for missing punctuation in continued quotations + # ” said Bob “ + nodes = dom.xpath("/html/body//p[re:test(., '”\\s(?:said|[A-Za-z]{2,}ed)\\s[A-Za-z]+?(?Newspapers + nodes = dom.xpath("/html/body//i[@epub:type and ( (following-sibling::node()[1][re:test(., '^[a-z]', 'i') and not(re:test(., '^(s|es|er)'))]) or preceding-sibling::node()[1][re:test(., '[a-z]$')]) ]") + if nodes: + messages.append(LintMessage("y-032", "Possible typo: Italics running into preceding or following characters.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - if "z3998:hymn" in selector: - local_css_has_hymn_style = True + # Check for three-em-dashes not preceded by a space + nodes = dom.xpath(f"/html/body//p[re:test(., '[^>“(\\s{se.WORD_JOINER}]{se.WORD_JOINER}?⸻')]") + if nodes: + messages.append(LintMessage("y-033", "Possible typo: Three-em-dash obscuring an entire word, but not preceded by a space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - if "z3998:lyrics" in selector: - local_css_has_lyrics_style = True + return messages - if "span.elision" in selector: - local_css_has_elision_style = True +def _lint_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str) -> list: + """ + Helper function used in self.lint() + Process XHTML checks on an .xhtml file - if "dedication" in selector: - local_css_has_dedication_style = True + INPUTS + filename: The name of the file being checked + dom: A dom tree to check + file_contents: The contents of the file being checked - if "epigraph" in selector: - local_css_has_epigraph_style = True + OUTPUTS + A list of LintMessages representing XHTML errors found in the file + """ - if "abbr" in selector and "nowrap" in rules: - abbr_with_whitespace.append(selector) + messages = []; - if regex.search(r"\[\s*xml\s*\|", selector, flags=regex.IGNORECASE) and "@namespace xml \"http://www.w3.org/XML/1998/namespace\";" not in self.local_css: - messages.append(LintMessage("c-003", "[css]\\[xml|attr][/] selector in CSS, but no XML namespace declared ([css]@namespace xml \"http://www.w3.org/XML/1998/namespace\";[/]).", se.MESSAGE_TYPE_ERROR, local_css_path)) + # Check for uppercase letters in IDs or classes + nodes = dom.xpath("//*[re:test(@id, '[A-Z]') or re:test(@class, '[A-Z]') or re:test(@epub:type, '[A-Z]')]") + if nodes: + messages.append(LintMessage("x-002", "Uppercase in attribute value. Attribute values must be all lowercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) - # lxml has not implemented the following in cssselect: *:first-of-type, *:last-of-type, *:nth-of-type, *:nth-last-of-type, *:only-of-type - # BUT ONLY WHEN USED ON *. This includes for example: `section [epub|type~="test"]:first-of-type` (note the `*` is implicit) - # Therefore we can't simplify them in build or test against them. - matches = regex.findall(r"(?:^| )[^a-z\s][^\s]+?:(?:first-of-type|last-of-type|nth-of-type|nth-last-of-type|only-of-type)", self.local_css, flags=regex.MULTILINE) + nodes = dom.xpath("//*[re:test(@id, '^[0-9]+')]") + if nodes: + messages.append(LintMessage("x-007", "[attr]id[/] attributes starting with a number are illegal XHTML.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) + + # Check for double greater-than at the end of a tag + matches = regex.findall(r"(>>|>>)", file_contents) if matches: - messages.append(LintMessage("c-001", "Don’t use [css]*:first-of-type[/], [css]*:last-of-type[/], [css]*:nth-of-type[/] [css]*:nth-last-of-type[/], or [css]*:only-of-type[/] on [css]*[/]. Instead, specify an element to apply it to.", se.MESSAGE_TYPE_ERROR, local_css_path, matches)) + messages.append(LintMessage("x-008", "Elements should end with a single [text]>[/].", se.MESSAGE_TYPE_WARNING, filename)) - if regex.search(r"\s+hyphens:.+?;(?!\s+-epub-hyphens)", self.local_css): - messages.append(LintMessage("c-007", "[css]hyphens[/css] CSS property without [css]-epub-hyphens[/css] copy.", se.MESSAGE_TYPE_ERROR, local_css_path)) + # Check for leading 0 in IDs (note: not the same as checking for IDs that start with an integer) + # We only check for *leading* 0s in numbers; this allows IDs like `wind-force-0` in the Worst Journey in the World glossary. + nodes = dom.xpath("//*[re:test(@id, '-0[0-9]')]") + if nodes: + messages.append(LintMessage("x-009", "Illegal leading 0 in [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) - if abbr_with_whitespace: - messages.append(LintMessage("c-005", f"[css]abbr[/] selector does not need [css]white-space: nowrap;[/] as it inherits it from [path][link=file://{self.path / 'src/epub/css/core.css'}]core.css[/][/].", se.MESSAGE_TYPE_ERROR, local_css_path, abbr_with_whitespace)) + # Check for HTML tags in tags + nodes = dom.xpath("/html/head/title/*") + if nodes: + messages.append(LintMessage("x-010", "Illegal element in [xhtml]<title>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) - # If we select on the xml namespace, make sure we define the namespace in the CSS, otherwise the selector won't work - # We do this using a regex and not with cssutils, because cssutils will barf in this particular case and not even record the selector. - matches = regex.findall(r"\[\s*xml\s*\|", self.local_css) - if matches and "@namespace xml \"http://www.w3.org/XML/1998/namespace\";" not in self.local_css: - messages.append(LintMessage("c-003", "[css]\\[xml|attr][/] selector in CSS, but no XML namespace declared ([css]@namespace xml \"http://www.w3.org/XML/1998/namespace\";[/]).", se.MESSAGE_TYPE_ERROR, local_css_path)) + # Check for underscores in attributes, but not if the attribute is href (links often have underscores) or MathMl alttext, which can have underscores as part of math notation + nodes = dom.xpath("//@*[contains(., '_') and name() !='href' and name() !='alttext']/..") + if nodes: + messages.append(LintMessage("x-011", "Illegal underscore in attribute. Use dashes instead of underscores.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) - matches = regex.findall(r"text-align:\s*left\s*;", self.local_css) - if matches: - messages.append(LintMessage("c-016", "[css]text-align: left;[/] found. Use [css]text-align: initial;[/] instead.", se.MESSAGE_TYPE_ERROR, local_css_path)) + # Check for style attributes + nodes = dom.xpath("/html/body//*[@style]") + if nodes: + messages.append(LintMessage("x-012", "Illegal [attr]style[/] attribute. Don’t use inline styles, any element can be targeted with a clever enough selector.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) - matches = regex.findall(r"[0-9\.]\s?rem;", self.local_css) - if matches: - messages.append(LintMessage("c-022", "Illegal [css]rem[/] unit. Use [css]em[/] instead.", se.MESSAGE_TYPE_ERROR, local_css_path)) + # Check for illegal elements in <head> + nodes = dom.xpath("/html/head/*[not(self::title) and not(self::link[@rel='stylesheet'])]") + if nodes: + messages.append(LintMessage("x-015", "Illegal element in [xhtml]<head>[/]. Only [xhtml]<title>[/] and [xhtml]<link rel=\"stylesheet\">[/] are allowed.", se.MESSAGE_TYPE_ERROR, filename, [f"<{node.tag}>" for node in nodes])) - matches = regex.findall(r"font-size\s*:\s*[0-9\.]+(?![0-9\.]|em|ex)", self.local_css) + # Check for xml:lang attribute starting in uppercase + nodes = dom.xpath("//*[re:test(@xml:lang, '^[A-Z]')]") + if nodes: + messages.append(LintMessage("x-016", "[attr]xml:lang[/] attribute with value starting in uppercase letter.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) + + # Check for ID attributes of numbered paragraphs (like `p-44`) that are used as refs in endnotes. + # Make sure their number is actually their correct sequence number in the text. + for unexpected_id in se.formatting.find_unexpected_ids(dom): + messages.append(LintMessage("x-019", f"Unexpected value of [attr]id[/] attribute. Expected: [attr]{unexpected_id[1]}[/].", se.MESSAGE_TYPE_ERROR, filename, [unexpected_id[0].to_tag_string()])) + + return messages + +def _lint_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str) -> list: + """ + Helper function used in self.lint() + Process typography checks on an .xhtml file + + INPUTS + filename: The name of the file being checked + dom: A dom tree to check + file_contents: The contents of the file being checked + special_file: A string containing the type of special file, if any + + OUTPUTS + A list of LintMessages representing typography errors found in the file + """ + + messages = []; + + # Check for punctuation outside quotes. We don't check single quotes because contractions are too common. + matches = regex.findall(fr"[\p{{Letter}}]+”[,\.](?!{se.WORD_JOINER} {se.WORD_JOINER}…)", file_contents) if matches: - messages.append(LintMessage("c-023", "Illegal unit used to set [css]font-size[/]. Hint: Use [css]em[/] units.", se.MESSAGE_TYPE_ERROR, local_css_path)) + messages.append(LintMessage("t-002", "Comma or period outside of double quote. Generally punctuation goes within single and double quotes.", se.MESSAGE_TYPE_WARNING, filename, matches)) + + # Check for ldquo not correctly closed + # Ignore closing paragraphs, line breaks, and closing cells in case ldquo means "ditto mark" + matches = regex.findall(r"“[^‘”]+?“", file_contents) + matches = [match for match in matches if "</p" not in match and "<br/>" not in match and "</td>" not in match] + # xpath to check for opening quote in p, without a next child p that starts with an opening quote or an opening bracket (for editorial insertions within paragraphs of quotation); or that consists of only an ellipses (like an elided part of a longer quotation) + # Matching <p>s can't have a poem/verse ancestor as formatting is often special for those. + matches = matches + [regex.findall(r"“[^”]+</p>", node.to_string())[0] for node in dom.xpath("/html/body//p[re:test(., '“[^‘”]+$')][not(ancestor::*[re:test(@epub:type, 'z3998:(verse|poem|song|hymn|lyrics)')])][(following-sibling::*[1])[name()='p'][not(re:test(normalize-space(.), '^[“\\[]') or re:test(normalize-space(.), '^…$'))]]")] + + # Additionally, match short <p> tags (< 100 chars) that lack closing quote, and whose direct siblings do have closing quotes (to exclude runs of same-speaker dialog), and that is not within a blockquote, verse, or letter + matches = matches + [regex.findall(r"“[^”]+</p>", node.to_string())[0] for node in dom.xpath("/html/body//p[re:test(., '“[^‘”]+$') and not(re:test(., '[…:]$')) and string-length(normalize-space(.)) <=100][(following-sibling::*[1])[not(re:test(., '“[^”]+$'))] and (preceding-sibling::*[1])[not(re:test(., '“[^”]+$'))]][not(ancestor::*[re:test(@epub:type, 'z3998:(verse|poem|song|hymn|lyrics)')]) and not(ancestor::blockquote) and not (ancestor::*[contains(@epub:type, 'z3998:letter')])][(following-sibling::*[1])[name()='p'][re:test(normalize-space(.), '^[“\\[]') and not(contains(., 'continued'))]]")] + if matches: + messages.append(LintMessage("t-003", "[text]“[/] missing matching [text]”[/]. Note: When dialog from the same speaker spans multiple [xhtml]<p>[/] elements, it’s correct grammar to omit closing [text]”[/] until the last [xhtml]<p>[/] of dialog.", se.MESSAGE_TYPE_WARNING, filename, matches)) - matches = regex.findall(r"line-height\s*:\s*[0-9\.]+(?!;|[0-9\.]+)", self.local_css) + # Check for closing dialog without comma + matches = regex.findall(r"[\p{Lowercase_Letter}]+?” [\p{Letter}]+? said", file_contents) if matches: - messages.append(LintMessage("c-024", "Illegal unit used to set [css]line-height[/]. Hint: [css]line-height[/] is set without any units.", se.MESSAGE_TYPE_ERROR, local_css_path)) + messages.append(LintMessage("t-005", "Dialog without ending comma.", se.MESSAGE_TYPE_WARNING, filename, matches)) + + # Check for possessive 's within name italics, but not in ignored files like the colophon which we know have no possessives + # Allow some known exceptions like `Harper's`, etc. + if filename.name not in IGNORED_FILENAMES: + nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.') and re:test(., '’s$') and not(contains(@epub:type, 'se:name.publication.') and re:test(., '(Pearson’s|Harper’s|Blackwood’s|Fraser’s|Baedeker’s)$'))]") + if nodes: + messages.append(LintMessage("t-007", "Possessive [text]’s[/] within name italics. If the name in italics is doing the possessing, [text]’s[/] goes outside italics.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) + + # Check for lsquo not correctly closed + matches = regex.findall(r"‘[^“’]+?‘", file_contents) + matches = [match for match in matches if "</p" not in match and "<br/>" not in match] + if matches: + messages.append(LintMessage("t-004", "[text]‘[/] missing matching [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, matches)) - matches = regex.findall(r"(height|\stop|\sbottom)\s*:\s*[0-9\.]+%", self.local_css) + # Check for repeated punctuation, but first remove `&` so we don't match `&,` + # Remove tds with repeated ” as they are probably ditto marks + matches = regex.findall(r"[,;]{2,}.{0,20}", file_contents.replace("&", "")) + regex.findall(r"(?:“\s*“|”\s*”|’ ’|‘\s*‘).{0,20}", regex.sub(r"<td>[”\s]+?(<a .+?epub:type=\"noteref\">.+?</a>)?</td>", "", file_contents)) + regex.findall(r"[\p{Letter}][,\.:;]\s[,\.:;]\s?[\p{Letter}<].{0,20}", file_contents, flags=regex.IGNORECASE) if matches: - messages.append(LintMessage("c-025", "Illegal percent unit used to set [css]height[/] or positioning property. Hint: [css]vh[/] to specify vertical-oriented properties like height or position.", se.MESSAGE_TYPE_ERROR, local_css_path)) + messages.append(LintMessage("t-008", "Repeated punctuation.", se.MESSAGE_TYPE_WARNING, filename, matches)) - # Done checking local.css - missing_files = [] - if self.is_se_ebook: - root_files = os.listdir(self.path) - expected_root_files = ["images", "src", "LICENSE.md"] - illegal_files = [root_file for root_file in root_files if root_file not in expected_root_files and root_file != "se-lint-ignore.xml"] # se-lint-ignore.xml is optional - missing_files = [expected_root_file for expected_root_file in expected_root_files if expected_root_file not in root_files and expected_root_file != "LICENSE.md"] # We add more to this later on. LICENSE.md gets checked later on, so we don't want to add it twice + # Check for nbsp before times + nodes = dom.xpath(f"/html/body//text()[re:test(., '[0-9][^{se.NO_BREAK_SPACE}]?$')][(following-sibling::abbr[1])[re:test(., '^[ap]\\.m\\.$')]]") + if nodes: + messages.append(LintMessage("t-009", "Required no-break space not found before time and [text]a.m.[/] or [text]p.m.[/].", se.MESSAGE_TYPE_WARNING, filename, [node[-10:] + "<abbr" for node in nodes])) + + # Check for times with periods instead of colons. + # Only check p, because things like tables/td are more likely to contain non-time numbers + # Exclude numbers preceded by equals, or succeeded by some measurements + # Also remove <a> first because they are likely to contain numbered section references + dom_copy = deepcopy(dom) + for node in dom_copy.xpath("/html/body//p/a"): + node.remove() + + nodes = dom_copy.xpath("/html/body//p[re:test(., '[^=]\\s[0-9]{1,2}\\.[0-9]{2}(?![0-9′″°%]|\\.[0-9]|\\scubic|\\smetric|\\smeters|\\smiles|\\sfeet|\\sinches)')]") + matches = [] + for node in nodes: + for time_match in regex.findall(r"(?<=[^=]\s)[0-9]{1,2}\.[0-9]{2}(?![0-9′″°%]|\.[0-9]|\scubic|\smetric|\smeters|\smiles|\sfeet|\sinches)", node.inner_text()): + time = time_match.split(".") + if not time[0].startswith("0") and int(time[0]) >= 1 and int(time[0]) <= 12 and int(time[1]) >= 0 and int(time[1]) <= 59: + matches.append(time_match) - # If we have illegal files, check if they are tracked in Git. - # If they are, then they're still illegal. - # If not, ignore them for linting purposes. - if illegal_files: - try: - illegal_files = self.repo.git.ls_files(illegal_files).split("\n") - if illegal_files and illegal_files[0] == "": - illegal_files = [] - except Exception: - # If we can't initialize Git, then just pass through the list of illegal files - pass + if matches: + messages.append(LintMessage("t-010", "Time set with [text].[/] instead of [text]:[/].", se.MESSAGE_TYPE_WARNING, filename, set(matches))) - for illegal_file in illegal_files: - messages.append(LintMessage("f-001", "Illegal file or directory.", se.MESSAGE_TYPE_ERROR, Path(illegal_file))) + # Check for missing punctuation before closing quotes + # Exclude signatures in footers as those are commonly quoted without ending punctuation + nodes = dom.xpath("/html/body//p[not( (parent::header or parent::hgroup or (parent::footer and contains(@epub:type, 'z3998:signature'))) and position()=last())][re:test(., '[a-z]+[”’]$')]") + if nodes: + messages.append(LintMessage("t-011", "Missing punctuation before closing quotes.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string()[-30:] for node in nodes])) - # Check for repeated punctuation - nodes = self.metadata_dom.xpath("/package/metadata/*[re:test(., '[,;]{2,}.{0,20}')]") + # Check for whitespace before noteref + # Do this early because we remove noterefs from headers later + # Allow noterefs that do not have preceding text/elements (for example a "see note X" ref that is the only child in a <td>) + nodes = dom.xpath("/html/body//a[contains(@epub:type, 'noteref') and preceding-sibling::node()[normalize-space(.)] and re:test(preceding-sibling::node()[1], '\\s+$')]") if nodes: - messages.append(LintMessage("t-008", "Repeated punctuation.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path, [node.to_string() for node in nodes])) + messages.append(LintMessage("t-012", "Illegal white space before noteref.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - # Check for non-English Wikipedia URLs - nodes = self.metadata_dom.xpath("/package/metadata/*[re:test(., 'https://(?!en)[a-z]{2,}\\.wikipedia\\.org')]") + # Check for period following Roman numeral, which is an old-timey style we must fix + # But ignore the numeral if it's the first item in a <p> tag, as that suggests it might be a kind of list item. + nodes = dom.xpath("/html/body//node()[name()='span' and contains(@epub:type, 'z3998:roman') and not(position()=1)][(following-sibling::node()[1])[re:test(., '^\\.\\s*[a-z]')]]") if nodes: - messages.append(LintMessage("m-043", "Non-English Wikipedia URL.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes])) + messages.append(LintMessage("t-013", "Roman numeral followed by a period. When in mid-sentence Roman numerals must not be followed by a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "." for node in nodes])) - # Check the long description for some errors - try: - # Check if there are non-typogrified quotes or em-dashes in metadata descriptions - # lxml unescapes this for us - # Also, remove HTML elements like <a href> so that we don't catch quotation marks in attribute values - long_description = self.metadata_dom.xpath("/package/metadata/meta[@property='se:long-description']")[0].text - matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", regex.sub(r"<[^<]+?>", "", long_description)) - if matches: - messages.append(LintMessage("m-014", "Non-typogrified character in [xml]<meta property=\"se:long-description\">[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) + # Check for two em dashes in a row + nodes = dom.xpath(f"/html/body//*[re:test(text(), '—{se.WORD_JOINER}*—+')]") + if nodes: + messages.append(LintMessage("t-014", "Two or more em-dashes in a row found. Elided words should use the two- or three-em-dash Unicode character, and dialog ending in em-dashes should only end in a single em-dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - # Is the first instance of the author's last name a hyperlink in the metadata? - authors = self.metadata_dom.xpath("/package/metadata/dc:creator") - for author in authors: - author_sort = self.metadata_dom.xpath(f"/package/metadata/meta[@property='file-as'][@refines='#{author.get_attr('id')}']/text()") - if author_sort: - author_last_name = regex.sub(r",.+$", "", author_sort[0]) - author_last_name = author_last_name.replace("'", "’") # Typogrify apostrophes so that we correctly match in the long description - # We can't use xpath here because the long description is escaped; it has no dom to query against. - if author_last_name.lower() in regex.sub(r"https://standardebooks\.org/ebooks/[^/]+?/", "", long_description.lower()) and not regex.search(fr"<a href=\"https://standardebooks\.org/ebooks/.+?\">.*?{author_last_name}.*?</a>", long_description, flags=regex.IGNORECASE): - messages.append(LintMessage("m-056", "Author name present in [xml]<meta property=\"se:long-description\">[/] element, but the first instance of their name is not hyperlinked to their S.E. author page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for money not separated by commas + nodes = dom.xpath("/html/body//*[re:test(text(), '[£\\$][0-9]{4,}')]") + if nodes: + messages.append(LintMessage("t-015", "Numbers not grouped by commas. Separate numbers greater than 1,000 with commas at every three numerals.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Did we mention an SE book in the long description, but without italics? - # Only match if the title appears to contain an uppercase letter. This prevents matches on a non-title link like `<a href>short stories</a>` - matches = regex.search(r"""(?<!<i>)<a href="https://standardebooks\.org/ebooks/[^"]+?/[^"]+?">(?!<i>)([\p{Letter}\s]+)""", long_description) - if matches and regex.search(r"[\p{Uppercase_Letter}]", matches[1]): - messages.append(LintMessage("m-064", "S.E. ebook hyperlinked in long description but not italicized.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for <abbr epub:type="z3998:*-name"> that does not contain spaces + nodes = dom.xpath("/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and re:test(., '[A-Z]\\.[A-Z]\\.')]") + if nodes: + messages.append(LintMessage("t-016", "Initials in [xhtml]<abbr epub:type=\"z3998:*-name\">[/] not separated by spaces.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - if regex.search(r"""<a href="https?://(?!standardebooks\.org)""", long_description): - messages.append(LintMessage("m-067", "Non-S.E. link in long description.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + if special_file != "colophon": + # This file is not the colophon. + # Check for ending punctuation inside italics that have semantics. + # Ignore the colophon because paintings might have punctuation in their names - # xml:lang is correct for the rest of the publication, but should be lang in the long desc - if "xml:lang" in long_description: - messages.append(LintMessage("m-057", "[xml]xml:lang[/] attribute in [xml]<meta property=\"se:long-description\">[/] element should be [xml]lang[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # This xpath matches b or i elements with epub:type="se:name...", that are not stage direction, whose last text node ends in punctuation. + # Note that we check that the last node is a text node, because we may have <abbr> as the last node + matches = [node.to_string() for node in dom.xpath("(//b | //i)[contains(@epub:type, 'se:name') and not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., '[\\.,!\\?;:]$')]]")] - # US -> U.S. - matches = regex.findall(r"\bUS\b", long_description) - if matches: - messages.append(LintMessage("t-047", "[text]US[/] should be [text]U.S.[/]", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) + # Match b or i elements that are not stage directions, and that end in a comma followed by a lowercase letter + matches = matches + [node.to_string() for node in dom.xpath("(//b | //i)[not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., ',$')] and following-sibling::node()[re:test(., '^\\s*[a-z]')] ]")] - # Make sure long-description is escaped HTML - if not regex.search(r"^\s*<p>", long_description) and long_description.strip() != "LONG_DESCRIPTION": - messages.append(LintMessage("m-016", "Long description must be escaped HTML.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) - else: - # Check for malformed long description HTML - try: - etree.parse(io.StringIO(f"<?xml version=\"1.0\"?><html xmlns=\"http://www.w3.org/1999/xhtml\">{long_description}</html>")) - except lxml.etree.XMLSyntaxError as ex: - messages.append(LintMessage("m-015", f"Metadata long description is not valid XHTML. LXML says: {ex}", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # ...and also check for ending punctuation inside em tags, if it looks like a *part* of a clause + # instead of a whole clause. If the <em> is preceded by an em dash or quotes, or if there's punctuation + # and a space before it, then it's presumed to be a whole clause. + matches = matches + [match.strip() for match in regex.findall(r"(?<!.[—“‘>]|[!\.\?…;:]\s)<em>(?:\w+?\s*)+[\.,\!\?;]</em>", file_contents) if match.islower()] - # Check for apostrophes outside links in long description - matches = regex.findall(r"</a>’s", long_description) + regex.findall(r"s</a>’", long_description) if matches: - messages.append(LintMessage("m-044", "Possessive [text]’[/] or [text]’s[/] outside of [xhtml]<a>[/] element in long description.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) + messages.append(LintMessage("t-017", "Ending punctuation inside formatting like bold, small caps, or italics. Ending punctuation is only allowed within formatting if the phrase is an independent clause.", se.MESSAGE_TYPE_WARNING, filename, list(set(matches)))) - # Check for HTML entities in long-description, but allow &amp; - matches = regex.findall(r"&[a-z0-9]+?;", long_description.replace("&", "")) - if matches: - messages.append(LintMessage("m-018", "HTML entities found. Use Unicode equivalents instead.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) + # Check for stage direction that ends in ?! but also has a trailing period + nodes = dom.xpath("/html/body//i[contains(@epub:type, 'z3998:stage-direction') and re:test(., '\\.$') and not((./node()[last()])[name() = 'abbr']) and (following-sibling::node()[1])[re:test(., '^[\\.,:;!?]')]]") + if nodes: + messages.append(LintMessage("t-018", "Stage direction ending in period next to other punctuation. Remove trailing periods in stage direction.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) + + # Check for single words that are in italics, but that have closing punctuation outside italics + # Outer wrapping match is so that .findall returns the entire match and not the subgroup + # The first regex also matches the first few characters before the first double quote; we use those for more sophisticated + # checks below, to give fewer false positives like `with its downy red hairs and its “<i xml:lang="fr">doigts de faune</i>.”` + matches = regex.findall(r"((?:.{1,2}\s)?“<(i|em)[^>]*?>[^<]+?</\2>[\!\?\.])", file_contents) + regex.findall(r"([\.\!\?] <(i|em)[^>]*?>[^<]+?</\2>[\!\?\.])", file_contents) + + # But, if we've matched a name of something, don't include that as an error. For example, `He said, “<i epub:type="se:name.publication.book">The Decameron</i>.”` + # We also exclude the match from the list if: + # 1. The double quote is directly preceded by a lowercase letter and a space: `with its downy red hairs and its “<i xml:lang="fr">doigts de faune</i>.”` + # 2. The double quote is directly preceded by a lowercase letter, a comma, and a space, and the first letter within the double quote is lowercase: In the original, “<i xml:lang="es">que era un Conde de Irlos</i>.” + # 3. The text is a single letter that is not "I" or "a" (because then it is likely a mathematical variable) + matches = [match for match in matches if "epub:type=\"se:name." not in match[0] and "epub:type=\"z3998:taxonomy" not in match[0] and not regex.match(r"^[\p{Lowercase_Letter}’]+\s“", match[0]) and not regex.match(r"^[\p{Lowercase_Letter}’]+,\s“[\p{Lowercase_Letter}]", se.formatting.remove_tags(match[0])) and not regex.match(r"^.*?<.+?>[^Ia]<.+?>", match[0])] + if matches: + messages.append(LintMessage("t-019", "When a complete clause is italicized, ending punctuation except commas must be within containing italics.", se.MESSAGE_TYPE_WARNING, filename, [match[0] for match in matches])) - except Exception: - if self.is_se_ebook: - missing_metadata_elements.append("""<meta id="long-description" property="se:long-description" refines="#description">""") + # Check for punctuation after endnotes + nodes = dom.xpath(f"/html/body//a[contains(@epub:type, 'noteref')][(following-sibling::node()[1])[re:test(., '^[^\\s<–\\]\\)—{se.WORD_JOINER}a-zA-Z0-9]')]]") + if nodes: + messages.append(LintMessage("t-020", "Endnote links must be outside of punctuation, including quotation marks.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes])) + + # Check for correct typography around measurements like 2 ft. + # But first remove href and id attrs because URLs and IDs may contain strings that look like measurements + # Note that while we check m,min (minutes) and h,hr (hours) we don't check s (seconds) because we get too many false positives on years, like `the 1540s` + matches = regex.findall(fr"\b[1-9][0-9]*[{se.NO_BREAK_SPACE}\-]?(?:[mck]?[mgl]|ft|in|min?|h|sec|hr)\.?\b", regex.sub(r"(href|id)=\"[^\"]*?\"", "", file_contents)) + # Exclude number ordinals, they're not measurements + matches = [match for match in matches if not regex.search(r"(st|nd|rd|th)", match)] + if matches: + messages.append(LintMessage("t-021", "Measurement not to standard. Numbers are followed by a no-break space and abbreviated units require an [xhtml]<abbr>[/] element. See [path][link=https://standardebooks.org/manual/1.0.0/8-typography#8.8.5]semos://1.0.0/8.8.5[/][/].", se.MESSAGE_TYPE_WARNING, filename, matches)) - missing_metadata_vars = [] - for node in self.metadata_dom.xpath("/package/metadata/*/text()"): - for var in SE_VARIABLES: - if regex.search(fr"\b{var}\b", node): - missing_metadata_vars.append(var) - # Quit the loop early to save some time - break + # Check for nbsp within <abbr epub:type="z3998:*-name">, which is redundant + nodes = dom.xpath(f"/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and contains(text(), '{se.NO_BREAK_SPACE}')]") + if nodes: + messages.append(LintMessage("t-022", "No-break space found in [xhtml]<abbr epub:type=\"z3998:*-name\">[/]. This is redundant.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - if missing_metadata_vars: - messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, missing_metadata_vars)) + # Check for trailing commas inside <i> tags at the close of dialog + # More sophisticated version of: \b[^\s]+?,</i>” + nodes = dom.xpath("/html/body//i[re:test(., ',$')][(following-sibling::node()[1])[starts-with(., '”')]]") + if nodes: + messages.append(LintMessage("t-023", "Comma inside [xhtml]<i>[/] element before closing dialog.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "”" for node in nodes])) - # Check if there are non-typogrified quotes or em-dashes in the title. - try: - title = self.metadata_dom.xpath("/package/metadata/dc:title")[0].text - matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", title) - if matches: - messages.append(LintMessage("m-012", "Non-typogrified character in [xml]<dc:title>[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) + # Check for quotation marks in italicized dialog + nodes = dom.xpath("/html/body//i[@xml:lang][starts-with(., '“') or re:test(., '”$')]") + if nodes: + messages.append(LintMessage("t-024", "When italicizing language in dialog, italics go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Do we need an dcterms:alternate meta element? - # Match spelled-out numbers with a word joiner, so for ex. we don't print "eight" if we matched "eighty" - matches = regex.findall(r"(?:[0-9]+|\bone\b|\btwo\b|\bthree\b|\bfour\b|\bfive\b|\bsix\b|\bseven\b|\beight\b|\bnine\b|\bten\b|\beleven\b|\btwelve\b|\bthirteen\b|\bfourteen\b|\bfifteen\b|\bsixteen\b|\bseventeen\b|\beighteen\b|\bnineteen\b|\btwenty\b|\bthirty\b|\bforty\b|\bfifty\b|\bsixty\b|\bseventy\b|\beighty|\bninety)", title, flags=regex.IGNORECASE) - if matches and not self.metadata_dom.xpath("/package/metadata/meta[@property='dcterms:alternate']"): - messages.append(LintMessage("m-052", "[xml]<dc:title>[/] element contains numbers, but no [xml]<meta property=\"dcterms:alternate\" refines=\"#title\"> element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) - except Exception: - missing_metadata_elements.append("<dc:title>") + # Check alt attributes on images, except for the logo + nodes = dom.xpath("/html/body//img[not(re:test(@src, '/(logo|titlepage)\\.(svg|png)$'))]") + img_no_alt = [] + img_alt_not_typogrified = [] + img_alt_lacking_punctuation = [] + for node in nodes: + if "titlepage.svg" not in node.get_attr("src"): + has_images = True # Save for a later check - try: - file_as = self.metadata_dom.xpath("/package/metadata/meta[@property='file-as' and @refines='#title']")[0].text - matches = regex.findall(r".(?:['\"]|\-\-|\s-\s).", file_as) - if matches: - messages.append(LintMessage("m-050", "Non-typogrified character in [xml]<meta property=\"file-as\" refines=\"#title\">[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) - except Exception: - missing_metadata_elements.append("<meta property=\"file-as\" refines=\"#title\">") + alt = node.get_attr("alt") - try: - description = self.metadata_dom.xpath("/package/metadata/dc:description")[0].text - matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", description) - if matches: - messages.append(LintMessage("m-013", "Non-typogrified character in [xml]<dc:description>[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) - except Exception: - missing_metadata_elements.append("<dc:description>") + if alt: + # Check for non-typogrified img alt attributes + if regex.search(r"""('|"|--|\s-\s|")""", alt): + img_alt_not_typogrified.append(node.to_tag_string()) + + # Check alt attributes not ending in punctuation + if filename.name not in IGNORED_FILENAMES and not regex.search(r"""[\.\!\?]”?$""", alt): + img_alt_lacking_punctuation.append(node.to_tag_string()) + + # Check that alt attributes match SVG titles + img_src = node.lxml_element.get("src") + if img_src and img_src.endswith("svg"): + title_text = "" + image_ref = img_src.split("/").pop() + try: + svg_path = self.content_path / "images" / image_ref + svg_dom = self.get_dom(svg_path) + try: + title_text = svg_dom.xpath("/svg/title")[0].text + except Exception: + messages.append(LintMessage("s-027", f"{image_ref} missing [xhtml]<title>[/] element.", se.MESSAGE_TYPE_ERROR, svg_path)) + + if title_text != "" and alt != "" and title_text != alt: + messages.append(LintMessage("s-022", f"The [xhtml]<title>[/] element of [path][link=file://{svg_path}]{image_ref}[/][/] does not match the [attr]alt[/] attribute text in [path][link=file://{filename}]{filename.name}[/][/].", se.MESSAGE_TYPE_ERROR, filename)) + + except FileNotFoundError: + missing_files.append(str(Path("images") / image_ref)) - # Set some variables for later - transcription_source_count = 0 - page_scan_source_count = 0 - other_source_count = 0 - sources = self.metadata_dom.xpath("/package/metadata/dc:source") - for source in sources: - if regex.search(r"(gutenberg\.org|wikisource\.org|fadedpage\.com|gutenberg\.net\.au|gutenberg\.ca)", source.inner_text()): - transcription_source_count = transcription_source_count + 1 - elif regex.search(r"(hathitrust\.org|/archive\.org|books\.google\.com|google\.com/books)", source.inner_text()): - page_scan_source_count = page_scan_source_count + 1 else: - other_source_count = other_source_count + 1 + img_no_alt.append(node.to_tag_string()) - ebook_has_multiple_transcriptions = transcription_source_count >= 2 - ebook_has_multiple_page_scans = page_scan_source_count >= 2 - ebook_has_other_sources = other_source_count > 0 + if img_alt_not_typogrified: + messages.append(LintMessage("t-025", "Non-typogrified [text]'[/], [text]\"[/] (as [xhtml]"[/]), or [text]--[/] in image [attr]alt[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, img_alt_not_typogrified)) + if img_alt_lacking_punctuation: + messages.append(LintMessage("t-026", "[attr]alt[/] attribute does not appear to end with punctuation. [attr]alt[/] attributes must be composed of complete sentences ending in appropriate punctuation.", se.MESSAGE_TYPE_ERROR, filename, img_alt_lacking_punctuation)) + if img_no_alt: + messages.append(LintMessage("s-004", "[xhtml]img[/] element missing [attr]alt[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, img_no_alt)) - # Check for double spacing - if self.metadata_dom.xpath(f"/package/metadata/*[re:test(., '[{se.NO_BREAK_SPACE}{se.HAIR_SPACE} ]{{2,}}')]"): - double_spaced_files.append(self.metadata_file_path) + # Check for low-hanging misquoted fruit + matches = regex.findall(r"[\p{Letter}]+[“‘]", file_contents) + regex.findall(r"[^>]+</(?:em|i|b|span)>‘[\p{Lowercase_Letter}]+", file_contents) + if matches: + messages.append(LintMessage("t-028", "Possible mis-curled quotation mark.", se.MESSAGE_TYPE_WARNING, filename, matches)) + + # Check for periods followed by lowercase. + temp_xhtml = regex.sub(r"<title[^>]*?>.+?", "", file_contents) # Remove because it might contain something like <title>Chapter 2: The Antechamber of M. de Tréville + temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Replace things like + temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Remove attributes + temp_xhtml = temp_xhtml.replace("A.B.C.", "X") # Remove A.B.C, which is not an abbreviations. + # Note the regex also excludes preceding numbers, so that we can have inline numbering like: + # "A number of questions: 1. regarding those who make heretics; 2. concerning those who were made heretics..." + matches = regex.findall(r"[^\s0-9]+\.\s+[\p{Lowercase_Letter}](?!’[\p{Uppercase_Letter}])[\p{Lowercase_Letter}]+", temp_xhtml) + # If is in the match, remove it from the matches so we exclude things like et. al. + matches = [match for match in matches if "" not in match] + if matches: + messages.append(LintMessage("t-029", "Period followed by lowercase letter. Hint: Abbreviations require an [xhtml][/] element.", se.MESSAGE_TYPE_WARNING, filename, matches)) - # Check for illegal Wikipedia URLs - nodes = self.metadata_dom.xpath("/package/metadata/*[contains(., '.m.wikipedia.org') or @*[contains(., '.m.wikipedia.org')]]") + # Check for initialisms without periods + nodes = [node.to_string() for node in dom.xpath("/html/body//abbr[contains(@epub:type, 'z3998:initialism') and not(re:test(., '^[0-9]*([a-zA-Z]\\.)+[0-9]*$'))]") if node.text not in INITIALISM_EXCEPTIONS] if nodes: - messages.append(LintMessage("m-026", "Illegal [url]https://*.m.wikipedia.org[/] URL. Hint: use non-mobile Wikipedia URLs.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes])) + messages.append(LintMessage("t-030", "Initialism with spaces or without periods.", se.MESSAGE_TYPE_WARNING, filename, set(nodes))) - # Check for punctuation outside quotes. We don't check single quotes because contractions are too common. - # We can't use xpath's built-in regex because it doesn't support Unicode classes - for node in self.metadata_dom.xpath("/package/metadata/*"): - if node.text and regex.search(r"[\p{Letter}]+”[,\.](?! …)", node.text): - messages.append(LintMessage("t-002", "Comma or period outside of double quote. Generally punctuation goes within single and double quotes.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path)) - break + nodes = dom.xpath("/html/body//*[re:test(text(), '\\bA\\s*B\\s*C\\s*\\b')]") + if nodes: + messages.append(LintMessage("t-031", "[text]A B C[/] must be set as [text]A.B.C.[/] It is not an abbreviation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Check that the word count is correct, if it's currently set - word_count = self.metadata_dom.xpath("/package/metadata/meta[@property='se:word-count']/text()", True) - if self.is_se_ebook: - if word_count is None: - missing_metadata_elements.append("""""") - elif word_count != "WORD_COUNT" and int(word_count) != self.get_word_count(): - messages.append(LintMessage("m-065", "Word count in metadata doesn’t match actual word count.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for abbreviations followed by periods + # But we exclude some SI units, which don't take periods; abbreviations ending in numbers for example in stage directions; abbreviations like `r^o` (recto) that contain ; and some Imperial abbreviations that are multi-word + nodes = dom.xpath("/html/body//abbr[(contains(@epub:type, 'z3998:initialism') or re:test(@epub:type, 'z3998:[^\\s]+?name') or not(@epub:type))][not(re:test(., '[cmk][mgl]')) and not(re:test(., '[0-9]$')) and not(./sup) and not(re:test(@class, '\\b(era|temperature|compound)\\b')) and not(re:test(text(), '^(mpg|mph|hp|TV)$'))][following-sibling::text()[1][starts-with(self::text(), '.')]]") + if nodes: + messages.append(LintMessage("t-032", "Initialism or name followed by period. Hint: Periods go within [xhtml][/]. [xhtml][/]s containing periods that end a clause require the [class]eoc[/] class.", se.MESSAGE_TYPE_WARNING, filename, [f"{node.to_string()}." for node in nodes])) - # Check if we have a subtitle but no fulltitle - if self.metadata_dom.xpath("/package/metadata[./meta[@property='title-type' and text()='subtitle'] and not(./meta[@property='title-type' and text()='extended'])]"): - messages.append(LintMessage("m-011", "Subtitle in metadata, but no full/extended title element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for space after dash + nodes = dom.xpath("/html/body//*[name()='p' or name()='span' or name='em' or name='i' or name='b' or name='strong'][not(self::comment())][re:test(., '[a-zA-Z]-\\s(?!(and|or|nor|to|und|…)\\b)')]") + if nodes: + messages.append(LintMessage("t-033", "Space after dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - # Check for tags that imply other tags - implied_tags = {"Fiction": ["Science Fiction", "Drama", "Fantasy"]} - for implied_tag, tags in implied_tags.items(): - if self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:subject' and text()={se.easy_xml.escape_xpath(implied_tag)}]"): - for tag in tags: - if self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:subject' and text()={se.easy_xml.escape_xpath(tag)}]"): - messages.append(LintMessage("m-058", f"[val]se:subject[/] of [text]{implied_tag}[/] found, but [text]{tag}[/] implies [text]{implied_tag}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches)) + # Check for preceded by em dash + nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[re:test(., '—$')]]") + if nodes: + messages.append(LintMessage("t-034", "[xhtml][/] element preceded by em-dash. Hint: em-dashes go within [xhtml][/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Check for 'comprised of' - if self.metadata_dom.xpath("/package/metadata/*[re:test(., '[Cc]omprised of')]"): - messages.append(LintMessage("m-069", "[text]comprised of[/] in metadata. Hint: Is there a better phrase to use here?", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for without preceding space in text node. (preceding ( or [ are also OK) + nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[not(re:test(., '[\\[\\(\\s]$'))]]") + if nodes: + messages.append(LintMessage("t-035", "[xhtml][/] element not preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Check for illegal em-dashes in - nodes = self.metadata_dom.xpath("/package/metadata/dc:subject[contains(text(), '—')]") + # Check for partially obscured years in which the last year is an em dash + nodes = dom.xpath("/html/body//p[re:test(., '1\\d{2}⁠—[^“’A-Za-z<]')]") if nodes: - messages.append(LintMessage("m-019", "Illegal em-dash in [xml][/] element; use [text]--[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.text for node in nodes])) + messages.append(LintMessage("t-036", "Em-dash used to obscure single digit in year. Hint: Use a hyphen instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - # Check for incorrect 'anonymous' strings in metadata - nodes = self.metadata_dom.xpath("/package/metadata/dc:contributor[re:test(., 'anonymous', 'i') and text() != 'Anonymous'] | /package/metadata/meta[@property='file-as' and re:test(., 'anonymous', 'i') and text() != 'Anonymous']") + # Check for rdquo preceded by space (but not preceded by a rsquo, which might indicate a nested quotation) + nodes = dom.xpath("/html/body//p[re:test(., '[^’]\\s”')]") if nodes: - messages.append(LintMessage("m-073", "Anonymous contributor values must be exactly [text]Anonymous[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes])) + messages.append(LintMessage("t-037", "[text]”[/] preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Check for metadata elements - nodes = self.metadata_dom.xpath("/package/metadata/*[not(name()='link') and not(normalize-space(.)) and not(./*)]") + # Check obviously miscurled quotation marks + nodes = dom.xpath("/html/body//p[re:test(., '“$')]") if nodes: - messages.append(LintMessage("m-022", "Empty element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes])) + messages.append(LintMessage("t-038", "[text]“[/] before closing [xhtml]

    [/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Check for illegal VCS URLs - nodes = self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:url.vcs.github' and not(text()='{self.generated_github_repo_url}')]") + # Check for some known initialisms with incorrect possessive apostrophes + nodes = dom.xpath("/html/body//abbr[text()='I.O.U.'][(following-sibling::node()[1])[starts-with(., '’s')]]") if nodes: - messages.append(LintMessage("m-009", f"[xml][/] value does not match expected: [url]{self.generated_github_repo_url}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + messages.append(LintMessage("t-039", "Initialism followed by [text]’s[/]. Hint: Plurals of initialisms are not followed by [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "’s" for node in nodes])) - # Check for illegal se:subject tags - illegal_subjects = [] - nodes = self.metadata_dom.xpath("/package/metadata/meta[@property='se:subject']/text()") + # Check if a subtitle ends in a text node with a terminal period; or if it ends in an node containing a terminal period. + nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6]/*[contains(@epub:type, 'subtitle')][(./text())[last()][re:test(., '\\.$')] or (./i)[last()][re:test(., '\\.$')]]") if nodes: - for node in nodes: - if node not in SE_GENRES and node != "TAG": - illegal_subjects.append(node) + messages.append(LintMessage("t-040", "Subtitle with illegal ending period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - if illegal_subjects: - messages.append(LintMessage("m-020", "Illegal value for [xml][/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, illegal_subjects)) + matches = regex.findall(r"[^…]\s[!?;:,].{0,10}", file_contents) # If we don't include preceding chars, the regex is 6x faster + if matches: + messages.append(LintMessage("t-041", "Illegal space before punctuation.", se.MESSAGE_TYPE_ERROR, filename, matches)) - if sorted(nodes) != nodes: - messages.append(LintMessage("m-053", "[xml][/] elements not in alphabetical order.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for a list of some common English loan words that used to be italicized as foreign, but are no longer + nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and re:test(., '^(fianc[eé]+|divorc[eé]+|menu|recherch[eé]|tour-de-force|outr[eé]|d[ée]but(ante)?|apropos|[eé]lite|prot[ée]g[ée]+|chef|salon|r[eé]gime|contretemps|[eé]clat|aides?|entr[ée]+|t[eê]te-[aà]-t[eê]tes?|blas[eé]|bourgeoisie)$')]") + if nodes: + messages.append(LintMessage("t-043", "Non-English loan word set in italics, when modern typography omits italics. Hint: A word may be correctly italicized when emphasis is desired, if the word is meant to be pronounced with an accent, or if the word is part of non-English speech.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - elif self.is_se_ebook: - messages.append(LintMessage("m-021", "No [xml][/] element found.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for comma after leading Or in subtitles + nodes = dom.xpath("/html/body//*[contains(@epub:type, 'subtitle') and re:test(text(), '^Or\\s')]") + if nodes: + messages.append(LintMessage("t-044", "Comma required after leading [text]Or[/] in subtitles.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Check that each has a file-as and title-type, if applicable - titles_missing_file_as = [] - titles_missing_title_type = [] - titles = self.metadata_dom.xpath("/package/metadata/dc:title") - for node in titles: - if not self.metadata_dom.xpath(f"/package/metadata/meta[@property='file-as' and @refines='#{node.get_attr('id')}']"): - titles_missing_file_as.append(node) + # Check for personas that are italicized. Use re:test instead of contains to avoid catching z3998:personal-name + nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:persona\\b') and @data-css-font-style = 'italic']") + if nodes: + messages.append(LintMessage("t-045", "Element has [val]z3998:persona[/] semantic and is also set in italics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - # Only check for title-type if there is more than one - if len(titles) > 1 and not self.metadata_dom.xpath(f"/package/metadata/meta[@property='title-type' and @refines='#{node.get_attr('id')}']"): - titles_missing_title_type.append(node) + if "῾" in file_contents: + messages.append(LintMessage("t-046", "[text]῾[/] (U+1FFE) detected. Use [text]ʽ[/] (U+02BD) instead.", se.MESSAGE_TYPE_ERROR, filename)) - if titles_missing_file_as: - messages.append(LintMessage("m-062", "[xml][/] missing matching [xml][/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in titles_missing_file_as])) + # Check for all-caps first paragraphs in sections/articles + # Note that we don't check for small-caps CSS, because there are lots of legitimate cases for that, + # and it would generate too many false positives. + # If we did want to do that, this xpath after the last re:test would help: or ./*[not(preceding-sibling::node()[normalize-space(.)]) and @data-css-font-variant='small-caps' and following-sibling::node()[1][self::text()]] + nodes = dom.xpath("/html/body//*[(name() = 'section' or name() = 'article') and not(contains(@epub:type, 'dedication') or contains(@epub:type, 'z3998:letter'))]/p[1][re:test(normalize-space(.), '^[“’]?(?:[A-Z’]+\\s){2,}')]") + if nodes: + messages.append(LintMessage("t-048", "Chapter opening text in all-caps.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - if titles_missing_title_type: - messages.append(LintMessage("m-068", "[xml][/] missing matching [xml][/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in titles_missing_title_type])) + # Check for two-em-dashes used for elision instead of three-em-dashes + matches = regex.findall(fr"[^{se.WORD_JOINER}\p{{Letter}}”]⸺[^“{se.WORD_JOINER}\p{{Letter}}].*", file_contents, flags=regex.MULTILINE) + if matches: + messages.append(LintMessage("t-049", "Two-em-dash used for eliding an entire word. Use a three-em-dash instead.", se.MESSAGE_TYPE_WARNING, filename, matches)) - # Check for CDATA tags - if "[/] does not match expected: [text]{self.generated_identifier}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) - except Exception: - missing_metadata_elements.append("") + # Check for quotations that carry to the next paragraph, but the next paragraph has no opening quotation mark + # Exclude p in blockquote because often that has different formatting + nodes = dom.xpath("/html/body//p[re:test(., '^“[^”]+$') and not(./ancestor::blockquote) and ./following-sibling::*[1][name() = 'p' and re:test(normalize-space(), '^[^“]')]]") + if nodes: + messages.append(LintMessage("t-051", "Dialog in [xhtml]

    [/] that continues to the next [xhtml]

    [/], but the next [xhtml]

    [/] does not begin with [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Check if se:name.person.full-name matches their titlepage name - duplicate_names = [] - invalid_refines = [] - nodes = self.metadata_dom.xpath("/package/metadata/meta[@property='se:name.person.full-name']") - for node in nodes: - try: - refines = node.get_attr("refines").replace("#", "") - try: - name = self.metadata_dom.xpath(f"/package/metadata/*[@id={se.easy_xml.escape_xpath(refines)}]")[0].text - if name == node.text: - duplicate_names.append(name) - except Exception: - invalid_refines.append(refines) - except Exception: - invalid_refines.append("") + # Check for stage direction without ending punctuation. We only want to consider stage direction that is not an interjection in a parent clause. + # We match if the stage direction ends in lowercase, and if there's no following node, or if there's a following node that doesn't begin in lowercase or an em dash or semicolon (which suggests an interjection). + nodes = dom.xpath(f"/html/body//i[@epub:type='z3998:stage-direction' and re:test(., '[a-z]$') and re:test(., '^[A-Z]') and (not(./following-sibling::node()[normalize-space(.)]) or ./following-sibling::node()[1][re:test(normalize-space(.), '^[^{se.WORD_JOINER}—;a-z]')])]") + if nodes: + messages.append(LintMessage("t-052", "Stage direction without ending punctuation. Note that ending punctuation is optional in stage direction that is an interjection in a parent clause, and such interjections should begin with a lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - if duplicate_names: - messages.append(LintMessage("m-024", "[xml][/] property identical to regular name. If the two are identical the full name [xml][/] element must be removed.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, duplicate_names)) + # Check for stage direction starting in lowercase. We only want to consider stage direction that is not an interjection in a parent clause. + # We match if the stage direction starts with a lowercase, and if there's no preceding node, or if there is a preceding node and it doesn't end in a lowercase letter or a small subset of conjoining punctuation. + nodes = dom.xpath("/html/body//i[@epub:type='z3998:stage-direction' and re:test(., '^[a-z]') and (not(./preceding-sibling::node()[normalize-space(.)]) or ./preceding-sibling::node()[1][re:test(normalize-space(.), '[^a-z:—,;…]$')])]") + if nodes: + messages.append(LintMessage("t-053", "Stage direction starting in lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - if invalid_refines: - messages.append(LintMessage("m-010", "Invalid [xml]refines[/] property.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, invalid_refines)) + # Epigraphs that are entirely non-English should still be in italics, not Roman + nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph') or contains(@epub:type, 'bridgehead')]//p[./i[@xml:lang and @data-css-font-style='normal' and ( (./preceding-sibling::node()='“' and ./following-sibling::node()='”') or not(./preceding-sibling::node()[normalize-space(.)] or ./following-sibling::node()[normalize-space(.)]) ) ]]") + if nodes: + messages.append(LintMessage("t-054", "Epigraphs that are entirely non-English should be set in italics, not Roman.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Check for malformed URLs - messages = messages + _get_malformed_urls(self.metadata_dom, self.metadata_file_path) + # Check for lone acute accents, which should always be combined or swapped for a more accurate Unicode char + if "´" in file_contents: + messages.append(LintMessage("t-055", "Lone acute accent ([val]´[/]). A more accurate Unicode character like prime for coordinates or measurements, or combining accent or breathing mark for Greek text, is required.", se.MESSAGE_TYPE_ERROR, filename)) - if self.metadata_dom.xpath("/package/metadata/*[contains(., 'https://id.loc.gov/')]"): - messages.append(LintMessage("m-066", "[url]id.loc.gov[/] URI starting with illegal https.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for degree confusable + nodes = dom.xpath("/html/body//*[re:test(text(), '[0-9]+º')]") + if nodes: + messages.append(LintMessage("t-056", "Masculine ordinal indicator ([val]º[/]) used instead of degree symbol ([val]°[/]). Note that the masculine ordinal indicator may be appropriate for ordinal numbers read as Latin, i.e. [val]12º[/] reading [val]duodecimo[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - if self.metadata_dom.xpath("/package/metadata/*[re:test(., 'id\\.loc\\.gov/authorities/names/[^\\.]+\\.html')]"): - messages.append(LintMessage("m-008", "[url]id.loc.gov[/] URI ending with illegal [path].html[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # Check for

    starting in lowercase + # Exclude

    starting in constructs like `(a)` or `a.` as they may be numbered lists + nodes = dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::figure or ancestor::ol[not(ancestor::section[contains(@epub:type, 'endnotes')])] or ancestor::ul or ancestor::table or ancestor::hgroup or preceding-sibling::*[1][name() = 'hr']) and not(contains(@class, 'continued') or re:test(@epub:type, 'z3998:(signature|valediction)')) and not(./*[1][name() = 'math' or name() = 'var' or contains(@epub:type, 'z3998:grapheme')]) and re:test(., '^[^A-Za-z0-9]?[a-z]') and not(re:test(., '^\\([a-z0-9]\\.?\\)\\.?')) and not(re:test(., '^[a-z0-9]\\.\\s'))]") - if self.metadata_dom.xpath("/package/metadata/dc:description[text()!='DESCRIPTION' and re:test(., '[^\\.”]$')]"): - messages.append(LintMessage("m-055", "[xml]dc:description[/] does not end with a period.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) + # We have to additionally filter using the regex library, because often a sentence may begin with an uppercase ACCENTED letter, + # and xpath's limited regex library doesn't support Unicode classes + nodes = [node for node in nodes if regex.match(r"^[^\p{Letter}0-9]?[\p{Lowercase_Letter}]", node.inner_text())] - # Does the manifest match the generated manifest? - try: - manifest = self.metadata_dom.xpath("/package/manifest")[0] - if manifest.to_string().replace("\t", "") != self.generate_manifest().to_string().replace("\t", ""): - messages.append(LintMessage("m-042", "[xml][/] element does not match expected structure.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path)) - except Exception: - missing_metadata_elements.append("") + if nodes: + messages.append(LintMessage("t-057", "[xhtml]

    [/] starting with lowercase letter. Hint: [xhtml]

    [/] that continues text after a [xhtml]

    [/] requires the [class]continued[/] class; and use [xhtml]
    [/] to split one clause over many lines.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - if missing_metadata_elements: - messages.append(LintMessage("m-051", "Missing expected element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, missing_metadata_elements)) + # Check for illegal characters + matches = regex.findall(fr"({se.UNICODE_BOM}|{se.SHY_HYPHEN})", file_contents) + if matches: + # Get the keys of a dict in order to create a list without duplicates + messages.append(LintMessage("t-058", "Illegal character.", se.MESSAGE_TYPE_ERROR, filename, list({match.encode("unicode_escape").decode().replace("\\u", "U+").upper():None for match in matches}.keys()))) - # Check for common typos - for node in self.metadata_dom.xpath("/package/metadata/dc:description") + self.metadata_dom.xpath("/package/metadata/meta[@property='se:long-description']"): - matches = regex.findall(r"(?[/] element before endnote backlink.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - nodes = self.metadata_dom.xpath("/package/metadata//*[re:test(., '[,;][a-z]', 'i')]") + # Check for Bible verses in old-style notation + nodes = dom.xpath("/html/body//*[(name()='p' or name()='cite' or re:test(name(), '^h[1-6]$')) and .//node()[re:test(., '(Genesis|Gen\\.|Exodus|Ex\\.|Leviticus|Lev\\.|Numbers|Num\\.|Deuteronomy|Deut\\.|Joshua|Josh\\.|Judges|Ruth|Kings|Chronicles|Chron\\.|Ezra|Nehemiah|Neh\\.|Esther|Esth\\.|Job|Psalm|Psalms|Ps\\.|Proverbs|Prov\\.|Ecclesiastes|Ecc\\.|Eccl\\.|Solomon|Sol\\.|Isaiah|Is\\.|Isa\\.|Jeremiah|Jer\\.|Lamentations|Lam\\.|Ezekiel|Ez\\.|Ezek\\.|Daniel|Dan\\.|Hosea|Hos\\.|Joel|Amos|Obadiah|Obad\\.|Jonah|Jon\\.|Micah|Mic\\.|Nahum|Nah\\.|Habakkuk|Hab\\.|Zephaniah|Zeph\\.|Haggai|Hag\\.|Zechariah|Zech\\.|Malachi|Mal\\.|Tobit|Judith|Sirach|Baruch|Maccabees|Esdras|Manasses|Matthew|Matt\\.|Mark|Luke|John|Acts|Romans|Rom\\.|Corinthians|Cor\\.|Corinth\\.|Galatians|Gal\\.|Ephesians|Eph\\.|Philippians|Phil\\.|Philipp\\.|Colossians|Col\\.|Coloss\\.|Thessalonians|Thes\\.|Thess\\.|Timothy|Tim\\.|Titus|Tit\\.|Philemon|Phil\\.|Hebrews|Heb\\.|James|Jas\\.|Peter|Pet\\.|Jude|Revelation|Revelations|Rev\\.)\\s*$') and following-sibling::node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman') and following-sibling::node()[1][re:test(., '^\\s*[\\.,]?\\s+[0-9]')]]]]") if nodes: - messages.append(LintMessage("y-002", "Possible typo: punctuation followed directly by a letter, without a space.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path, [node.to_string() for node in nodes])) + messages.append(LintMessage("t-060", "Old style Bible citation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Make sure some static files are unchanged - if self.is_se_ebook: - try: - with importlib_resources.path("se.data.templates", "LICENSE.md") as license_file_path: - if not filecmp.cmp(license_file_path, self.path / "LICENSE.md"): - messages.append(LintMessage("f-003", f"File does not match [path][link=file://{license_file_path}]{license_file_path}[/][/].", se.MESSAGE_TYPE_ERROR, self.path / "LICENSE.md")) - except Exception: - missing_files.append("LICENSE.md") + # Check for summary-style bridgeheads not ending in punctuation + # Decide if it's a summary-style bridgehead if it contains two or more em dashes. We use Python regex because + # xpath can't count the number of occurances of a string. + nodes = dom.xpath("/html/body//*[contains(@epub:type, 'bridgehead') and re:test(., '[^\\.\\!\\?”]”?$')]") + nodes = [node for node in nodes if len(regex.findall(r"—", node.to_string())) >= 2] + if nodes: + messages.append(LintMessage("t-061", "Summary-style bridgehead without ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) + + # Check for uppercased am/pm, except for if they have an epub:type, or when uppercased in titles, or in the nav document + if not dom.xpath("//nav[contains(@epub:type, 'toc')]"): + nodes = dom.xpath("/html/body//abbr[re:test(., '^[AP]\\.M\\.$') and not(@epub:type) and not(./ancestor::*[contains(@epub:type, 'title')])]") + if nodes: + messages.append(LintMessage("t-062", "Uppercased [text]a.m.[/] and [text]p.m.[/]", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) + + # Check for some Latinisms that need italics according to the manual. We check ancestor-or-self in case the phrase is set in Roman because it's nested in a parent italic. + # Exclude toto followed by ’ since Toto can be a name. + # Exclude whose entire contents is a matched Latinism as we do not italicize those. + # Ignore the ToC because we have different rules there + nodes = dom.xpath("/html/body//text()[re:test(., '\\b(a (priori|posteriori|fortiori)|(?s in endnotes that are directly followed by the backlink; + # ignore s starting with `Cf.` which means `Compare` and was often used to mean `see` + nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[last()]/cite[following-sibling::node()[normalize-space(.)][1][name() = 'a' and contains(@epub:type, 'backlink')] and re:test(., '^[^—]') and not(re:test(., '^(Cf\\.|\\()'))]") + if nodes: + messages.append(LintMessage("t-068", "Citation not offset with em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - try: - with importlib_resources.path("se.data.templates", "se.css") as core_css_file_path: - if not filecmp.cmp(core_css_file_path, self.content_path / "css/se.css"): - messages.append(LintMessage("f-014", f"File does not match [path][link=file://{self.path / 'src/epub/css/se.css'}]{core_css_file_path}[/][/].", se.MESSAGE_TYPE_ERROR, self.content_path / "css/se.css")) - except Exception: - missing_files.append("css/se.css") + # Check for in
    s that start with a leading em dash + nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '^—')]") + if nodes: + messages.append(LintMessage("t-069", "[xhtml][/] in epigraph starting with an em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - # Before we start, walk the files in spine order to build a tree of sections and heading levels. - # section_tree is a list of EbookSections. - # For example, The Woman in White by Wilkie Collins would have a section_tree that looks like this: - # titlepage (1) - # imprint (2) - # dedication (2) - # preface (2) - # halftitlepage (2) - # part-1 (2) - # division-1-1 (3) - # chapter-1-1-1 (4) - # chapter-1-1-2 (4) - # chapter-1-1-3 (4) - # chapter-1-1-4 (4) - for filename in self.spine_file_paths: - dom = self.get_dom(filename) - for dom_section in dom.xpath("/html/body//*[re:test(name(), '^(section|article|nav)$')][@id]"): - # Start at h2 by default except for the titlepage, which is always h1 - starting_depth = 1 if regex.findall(r"\btitlepage\b", dom_section.get_attr("epub:type") or "") else 2 + # Check for in
    s that end with a period. Try to exclude s that are last children + nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '\\.$') and not( (./node()[last()])[./descendant-or-self::abbr]) ]") + if nodes: + messages.append(LintMessage("t-070", "[xhtml][/] in epigraph ending in a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - section = _find_ebook_section(dom_section.get_attr("id"), section_tree) + # If we have Greek text, try to normalize it + for node in dom.xpath("/html/body//*[@xml:lang='grc' or @xml:lang='el']"): + node_text = node.inner_text() + expected_text = se.typography.normalize_greek(node_text) + if node_text != expected_text: + messages.append(LintMessage("t-073", f"Possible transcription error in Greek. Found: [text]{node_text}[/], but expected [text]{expected_text}[/text]. Hint: Use [bash]se unicode-names[/] to see differences in Unicode characters.", se.MESSAGE_TYPE_WARNING, filename)) - section_parent_id = dom_section.get_attr("data-parent") + return messages - has_header = bool(dom_section.xpath("./*[re:test(name(), '^h[1-6]$')]")) - if not has_header: - has_header = bool(dom_section.xpath(".//*[not(re:test(name(), '^(section|article|nav)$'))]//*[re:test(name(), '^h[1-6]$')]")) +def _lint_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_info: dict, language: str) -> list: + """ + Helper function used in self.lint() + Process syntax checks on an .xhtml file - if not section_parent_id: - # We don't have a data-parent, but do we have a direct parent section in the same file as this section? - direct_parent = dom_section.xpath("./ancestor::*[re:test(name(), '^(section|article|nav)$')][@id][1]") + INPUTS + filename: The name of the file being checked + dom: A dom tree to check + file_contents: The contents of the file being checked + ebook_info: A dictionary containing several pieces of information about an ebook - if direct_parent: - section_parent_id = direct_parent[0].get_attr("id") + OUTPUTS + A list of LintMessages representing syntax errors found in the file + """ - # If we don't have the section in our list yet, and it has a parent, append it to the parent - if not section and section_parent_id: - parent = _find_ebook_section(section_parent_id, section_tree) - if parent: - # Only increment the depth if the parent has a header. If it doesn't then we don't want to increment, - # because we will end up skipping an h# level (like h2 -> h4 instead of h2 -> h3) - parent.children.append(EbookSection(dom_section.get_attr("id"), parent.depth + 1 if parent.has_header else parent.depth, has_header)) + messages = []; - # If we don't have it in our list yet, and it has no parent, append it to the top - elif not section: - section_tree.append(EbookSection(dom_section.get_attr("id"), starting_depth, has_header)) + # Check for numeric entities + matches = regex.findall(r"&#[0-9]+?;", file_contents) + if matches: + messages.append(LintMessage("s-001", "Illegal numeric entity (like [xhtml]Α[/]).", se.MESSAGE_TYPE_ERROR, filename)) - # This block is useful for pretty-printing section_tree should we need to debug it in the future - # def dump(item, char): - # print(f"{char} {item.section_id} ({item.depth}) {item.has_header}") - # for child in item.children: - # dump(child, f" {char}") - # for section in section_tree: - # dump(section, "") - # exit() + # Check nested
    elements, but only if it's the first child of another
    + nodes = dom.xpath("/html/body//blockquote/*[1][name()='blockquote']") + if nodes: + messages.append(LintMessage("s-005", "Nested [xhtml]
    [/] element.", se.MESSAGE_TYPE_WARNING, filename)) - # Now iterate over individual files for some checks - # We use os.walk() and not Path.glob() so that we can ignore `.git` and its children - for root, directories, filenames in os.walk(self.path): - if ".git" in directories: - directories.remove(".git") + # Check to see if we've marked something as poetry or verse, but didn't include a first + # This xpath selects the p elements, whose parents are poem/verse, and whose first child is not a span + nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')]/p[not(./*[name()='span' and position()=1])]") + if nodes: + matches = [] + for node in nodes: + # Get the first line of the poem, if it's a text node, so that we can include it in the error messages. + # If it's not a text node then just ignore it and add the error anyway. + first_line = node.xpath("./descendant-or-self::text()[normalize-space(.)]", True) + if first_line: + match = first_line.strip() + if match: # Make sure we don't append an empty string + matches.append(match) + + messages.append(LintMessage("s-006", "Poem or verse [xhtml]

    [/] (stanza) without [xhtml][/] (line) element.", se.MESSAGE_TYPE_WARNING, filename, matches)) + + # Check for elements that don't have a direct block child + # Allow white space and comments before the first child + # Ignore children of the ToC and landmarks as they do not require p children + nodes = dom.xpath("/html/body//*[(name()='blockquote' or name()='dd' or name()='header' or name()='li' or name()='footer') and not(./ancestor::*[contains(@epub:type, 'toc') or contains(@epub:type, 'landmarks')]) and (node()[normalize-space(.) and not(self::comment())])[1][not(name()='p' or name()='blockquote' or name()='div' or name()='table' or name()='header' or name()='ul' or name()='ol' or name()='footer' or name()='hgroup' or re:test(name(), '^h[0-6]'))]]") + if nodes: + messages.append(LintMessage("s-007", "Element requires at least one block-level child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes])) - for directory in natsorted(directories): - if directory == "META-INF": - continue + # Check for block-level tags that end with
    + nodes = dom.xpath("/html/body//*[self::p or self::blockquote or self::table or self::ol or self::ul or self::section or self::article][br[last()][not(following-sibling::text()[normalize-space()])][not(following-sibling::*)]]") + if nodes: + messages.append(LintMessage("s-008", "[xhtml]
    [/] element found before closing tag of block-level element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) - url_safe_filename = se.formatting.make_url_safe(directory) - if directory != url_safe_filename: - directories_not_url_safe.append(Path(root) / directory) + # Check for hgroup elements with only one child + nodes = dom.xpath("/html/body//hgroup[count(*)=1]") + if nodes: + messages.append(LintMessage("s-009", "[xhtml]

    [/] element with only one child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - for filename in natsorted(filenames): - filename = (Path(root) / filename).resolve() + # Check for empty elements. Elements are empty if they have no children and no non-whitespace text + # Use name() to check for MathML elements, because if the MathML namespace is not declared in a given file and we search by the m: namespace then this xpath will fail entirely. + nodes = dom.xpath("/html/body//*[not(self::br) and not(self::hr) and not(self::img) and not(self::td) and not(self::th) and not(self::link) and not(self::col) and not(self::colgroup) and not(local-name()='none') and not(local-name()='mspace') and not(local-name()='mprescripts')][not(./*)][not(normalize-space())][not(./ancestor::*[local-name()='annotation-xml'])]") + if nodes: + messages.append(LintMessage("s-010", "Empty element. Use [xhtml]
    [/] for thematic breaks if appropriate.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes])) - if filename.stem != "LICENSE": - url_safe_filename = se.formatting.make_url_safe(filename.stem) + filename.suffix - if filename.name != url_safe_filename and not filename.stem.endswith(".source"): - files_not_url_safe.append(filename) + # Check for
    and
    without ID attribute + nodes = dom.xpath("/html/body//*[self::section or self::article][not(@id)]") + if nodes: + messages.append(LintMessage("s-011", "Element without [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes])) - if "-0" in filename.name: - messages.append(LintMessage("f-009", "Illegal leading [text]0[/] in filename.", se.MESSAGE_TYPE_ERROR, filename)) + # Check for 
     tags
+	if dom.xpath("/html/body//pre"):
+		messages.append(LintMessage("s-013", "Illegal [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename))
 
-			if filename.stem == "cover.source":
-				has_cover_source = True
+	# Check for 
     after block-level elements
+	nodes = dom.xpath("/html/body//*[self::p or self::blockquote or self::table or self::ol or self::ul or self::section or self::article][following-sibling::br]")
+	if nodes:
+		messages.append(LintMessage("s-014", "[xhtml]
    [/] after block-level element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-			if filename.suffix == ".jpeg":
-				messages.append(LintMessage("f-011", "JPEG files must end in [path].jpg[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for 
     tags before the end of a section, which is a common PG artifact
+	if dom.xpath("/html/body//hr[count(following-sibling::*)=0]"):
+		messages.append(LintMessage("s-012", "Illegal [xhtml]
    [/] as last child.", se.MESSAGE_TYPE_ERROR, filename))
 
-			if filename.suffix == ".tiff":
-				messages.append(LintMessage("f-012", "TIFF files must end in [path].tif[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for hgroup elements with a subtitle but no title
+	nodes = dom.xpath("/html/body//hgroup[./*[contains(@epub:type, 'subtitle')] and not(./*[contains(concat(' ', @epub:type, ' '), ' title ')])]")
+	if nodes:
+		messages.append(LintMessage("s-015", "Element has [val]subtitle[/] semantic, but without a sibling having a [val]title[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-			# Run some general tests on images, but skip the cover source since it's an exception to all of these rules
-			if filename.suffix in (".jpg", ".jpeg", ".tif", ".tiff", ".png") and "cover.source" not in filename.name:
-				try:
-					image = Image.open(filename)
-				except UnidentifiedImageError as ex:
-					raise se.InvalidFileException(f"Couldn’t identify image type of [path][link=file://{filename}]{filename.name}[/][/].") from ex
+	# Check for deprecated MathML elements
+	# Note we dont select directly on element name, because we want to ignore any namespaces that may (or may not) be defined
+	nodes = dom.xpath("/html/body//*[local-name()='mfenced']")
+	if nodes:
+		messages.append(LintMessage("s-017", "[xhtml][/] is deprecated in the MathML spec. Use [xhtml](...)[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-				# Check the source cover image
-				if self.path / "images" / "cover.jpg" == filename:
-					if image.size != (se.COVER_WIDTH, se.COVER_HEIGHT):
-						messages.append(LintMessage("f-017", f"[path][link=file://{self.path / 'images/cover.jpg'}]cover.jpg[/][/] must be exactly {se.COVER_WIDTH} × {se.COVER_HEIGHT}.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for 
     tags without id attributes
+	nodes = dom.xpath("/html/body//img[@id]")
+	if nodes:
+		messages.append(LintMessage("s-018", "[xhtml][/] element with [attr]id[/] attribute. [attr]id[/] attributes go on parent [xhtml]
    [/] elements. Hint: Images that are inline (i.e. that do not have parent [xhtml]
    [/]s do not have [attr]id[/] attributes.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-				# Run some tests on distributable images in ./src/epub/images/
-				# Once we reach Python 3.9 we can use path.is_relative_to() instead of this string comparison
-				if str(filename).startswith(str(self.content_path / "images")):
-					if os.path.getsize(filename) > 1500000: # 1.5MB
-						messages.append(LintMessage("f-016", "Image more than 1.5MB in size.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for IDs on  tags
+	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][@id]")
+	if nodes:
+		messages.append(LintMessage("s-019", "[xhtml][/] element with [attr]id[/] attribute. [xhtml][/] elements should be wrapped in [xhtml]
    [/] elements, which should hold the [attr]id[/] attribute.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-					# Make sure distributable images have reasonable dimensions
-					# We check SVGs later on in a separate check
-					if image.size[0] * image.size[1] > 4000000:
-						messages.append(LintMessage("f-018", "Image greater than 4,000,000 pixels square in dimension.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for legal cases that aren't italicized
+	# We can't use this because v. appears as short for "volume", and we may also have sporting events without italics.
+	#nodes = dom.xpath("/html/body//abbr[text()='v.' or text()='versus'][not(parent::i)]")
+	#if nodes:
+	#	messages.append(LintMessage("t-xxx", "Legal case without parent [xhtml][/].", se.MESSAGE_TYPE_WARNING, filename, {f"{node.to_string()}." for node in nodes}))
 
-					if filename.suffix == ".png" and not se.images.has_transparency(image):
-						messages.append(LintMessage("f-019", "[path].png[/] file without transparency. Hint: If an image doesn’t have transparency, it should be saved as a [path].jpg[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Only do this check if there's one  or one 
     tag. If there's more than one, then the xhtml file probably requires an overarching title
+	# We merge two xpaths here because /
     can be either a direct child of 
    , or it could be nested in 
    
+	if len(dom.xpath("/html/body/*[name()='section' or name()='article']/*[re:test(name(), '^h[1-6]$') or name()='hgroup'] | /html/body/*[name()='section' or name()='article']/header/*[re:test(name(), '^h[1-6]$') or name()='hgroup']"))==1:
+		title = se.formatting.generate_title(dom)
 
-			if filename.suffix in BINARY_EXTENSIONS or filename.name == "core.css":
-				continue
+		if not dom.xpath(f"/html/head/title[text()={se.easy_xml.escape_xpath(title.replace('&', '&'))}]"):
+			messages.append(LintMessage("s-021", f"Unexpected value for [xhtml][/] element. Expected: [text]{title}[/]. (Beware hidden Unicode characters!)", se.MESSAGE_TYPE_ERROR, filename))
 
-			# Read the file and start doing some serious checks!
-			try:
-				file_contents = self.get_file(filename)
-			except UnicodeDecodeError:
-				# This is more to help developers find weird files that might choke `se lint`, hopefully unnecessary for end users
-				messages.append(LintMessage("f-010", "Problem decoding file as utf-8.", se.MESSAGE_TYPE_ERROR, filename))
-				continue
+	# Check to see if <h#> tags are correctly titlecased
+	# Ignore <h#> tags with an `xml:lang` attribute, as other languages have different titlecasing rules
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') or (name() = 'p' and parent::hgroup)][not(contains(@epub:type, 'z3998:roman')) and not(@xml:lang)]")
+	for node in nodes:
+		node_copy = deepcopy(node)
 
-			# Remove comments before we do any further processing
-			file_contents = regex.sub(r"<!--.+?-->", "", file_contents, flags=regex.DOTALL)
+		# Remove *leading* Roman spans
+		# This matches the first child node excluding white space nodes, if it contains the z3998:roman semantic.
+		for element in node_copy.xpath("./node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman')]"):
+			element.remove()
 
-			matches = regex.findall(r"http://standardebooks\.org[^\"<\s]*", file_contents)
-			if matches:
-				messages.append(LintMessage("m-003", "Non-HTTPS URL.", se.MESSAGE_TYPE_ERROR, filename, matches))
+		# Remove noterefs
+		for element in node_copy.xpath(".//a[contains(@epub:type, 'noteref')]"):
+			element.remove()
 
-			if "UTF-8" in file_contents:
-				messages.append(LintMessage("x-001", "String [text]UTF-8[/] must always be lowercase.", se.MESSAGE_TYPE_ERROR, filename))
+		# Remove hidden elements, for example in poetry identified by first line (keats)
+		for element in node_copy.xpath(".//*[@hidden]"):
+			element.remove()
 
-			if filename.suffix == ".svg":
-				svg_dom = self.get_dom(filename)
+		# Unwrap (not remove) any subelements in subtitles, as they may trip up further regexes
+		for element in node_copy.xpath(".//*[contains(@epub:type, 'subtitle')]//*"):
+			element.unwrap()
 
-				# If we're looking at the cover image, ensure that the producer has built it.
-				# The default cover image is a white background which when encoded to base64 begins with 299 characters and then has a long string of `A`s
-				if self.cover_path and filename.name == self.cover_path.name and svg_dom.xpath("//image[re:test(@xlink:href, '^.{299}A{50,}')]"):
-					messages.append(LintMessage("m-063", "Cover image has not been built.", se.MESSAGE_TYPE_ERROR, filename))
+		title = node_copy.inner_xml()
 
-				# Check for fill: #000 which should simply be removed
-				nodes = svg_dom.xpath("//*[contains(@fill, '#000') or contains(translate(@style, ' ', ''), 'fill:#000')]")
-				if nodes:
-					messages.append(LintMessage("x-004", "Illegal [xml]style=\"fill: #000\"[/] or [xml]fill=\"#000\"[/].", se.MESSAGE_TYPE_ERROR, filename))
+		# Remove leading leftover spacing and punctuation
+		title = regex.sub(r"^[\s\.\,\!\?\:\;]*", "", title)
 
-				# Check for illegal height or width on root <svg> element
-				if filename.name != "logo.svg": # Do as I say, not as I do...
-					if svg_dom.xpath("/svg[@height or @width]"):
-						messages.append(LintMessage("x-005", "Illegal [xml]height[/] or [xml]width[/] attribute on root [xml]<svg>[/] element. Size SVGs using the [xml]viewBox[/] attribute only.", se.MESSAGE_TYPE_ERROR, filename))
+		# Normalize whitespace
+		title = regex.sub(r"\s+", " ", title, flags=regex.DOTALL).strip()
 
-				match = regex.search(r"viewbox", file_contents, flags=regex.IGNORECASE)
-				if match and match[0] != "viewBox":
-					messages.append(LintMessage("x-006", f"[xml]{match}[/] found instead of [xml]viewBox[/]. [xml]viewBox[/] must be correctly capitalized.", se.MESSAGE_TYPE_ERROR, filename))
+		# Do we have a subtitle? If so the first letter of that must be capitalized, so we pull that out
+		subtitle_matches = regex.findall(r"(.*?)<span epub:type=\"subtitle\">(.*?)</span>(.*?)", title, flags=regex.DOTALL)
+		if subtitle_matches:
+			for title_header, subtitle, title_footer in subtitle_matches:
+				title_header = se.formatting.titlecase(se.formatting.remove_tags(title_header).strip())
+				subtitle = se.formatting.titlecase(se.formatting.remove_tags(subtitle).strip())
+				title_footer = se.formatting.titlecase(se.formatting.remove_tags(title_footer).strip())
 
-				# Make images have reasonable dimensions
-				viewbox = svg_dom.xpath("/svg/@viewBox", True)
-				if viewbox:
-					svg_dimensions = viewbox.split()
-					try:
-						if float(svg_dimensions[2]) * float(svg_dimensions[3]) > 4000000:
-							messages.append(LintMessage("f-018", "Image greater than 4,000,000 pixels square in dimension.", se.MESSAGE_TYPE_ERROR, filename))
-					except Exception as ex:
-						raise se.InvalidFileException(f"Couldn’t parse SVG [xhtml]viewBox[/] attribute in [path][link=file://{filename.resolve()}]{filename}[/][/].") from ex
+				titlecased_title = f"{title_header} {subtitle} {title_footer}"
+				titlecased_title = titlecased_title.strip()
 
-				# Check for illegal transform or id attribute
-				nodes = svg_dom.xpath("//*[@transform or @id]")
-				if nodes:
-					invalid_transform_attributes = set()
-					invalid_id_attributes = []
-					for node in nodes:
-						if node.get_attr("transform"):
-							invalid_transform_attributes.add(f"transform=\"{node.get_attr('transform')}\"")
+				title = se.formatting.remove_tags(title).strip()
+				if title != titlecased_title:
+					messages.append(LintMessage("s-023", f"Title [text]{title}[/] not correctly titlecased. Expected: [text]{titlecased_title}[/].", se.MESSAGE_TYPE_WARNING, filename))
 
-						if node.get_attr("id"):
-							invalid_id_attributes.append(f"id=\"{node.get_attr('id')}\"")
+		# No subtitle? Much more straightforward
+		else:
+			titlecased_title = se.formatting.titlecase(se.formatting.remove_tags(title))
+			title = se.formatting.remove_tags(title)
+			if title != titlecased_title:
+				messages.append(LintMessage("s-023", f"Title [text]{title}[/] not correctly titlecased. Expected: [text]{titlecased_title}[/].", se.MESSAGE_TYPE_WARNING, filename))
 
-					if invalid_transform_attributes:
-						messages.append(LintMessage("x-003", "Illegal [xml]transform[/] attribute. SVGs should be optimized to remove use of [xml]transform[/]. Try using Inkscape to save as an “optimized SVG”.", se.MESSAGE_TYPE_ERROR, filename, invalid_transform_attributes))
+	# Check for header elements that are entirely non-English
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./i[@xml:lang][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	if nodes:
+		messages.append(LintMessage("s-024", "Header elements that are entirely non-English should not be set in italics. Instead, the [xhtml]<h#>[/] element has the [attr]xml:lang[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-					if invalid_id_attributes:
-						messages.append(LintMessage("x-014", "Illegal [xml]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, invalid_id_attributes))
+	# Check for z3998:roman elements with invalid values. Roman numerals can occasionally end in `j` as an alias for ending `i`. See _The Worm Ouroboros_.
+	# We also allow the numeral to end in a digit, because that might be an endnote. For example `<h2 epub:type="ordinal z3998:roman">II<a href="..." epub:type="noteref">3</a></h2>`
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:roman') and not(re:test(normalize-space(.), '^[ivxlcdmIVXLCDM]+j?[0-9]*?$'))]")
+	if nodes:
+		messages.append(LintMessage("s-026", "Invalid Roman numeral.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+
+	# Check for <h#> elements that are higher or lower than their expected level based on how deep they are in <section>s
+	# or <article>s. Exclude <nav> in the ToC.
+	# Get all h# elements not preceded by other h# elements (this includes the 1st item of an hgroup but not the next items)
+	# Exclude the ToC and landmarks
+	invalid_headers = []
+	invalid_parent_ids = []
+	for heading in dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') and not(./preceding-sibling::*[re:test(name(), '^h[1-6]$')]) and not(./ancestor::*[re:test(@epub:type, '\\b(toc|landmarks)\\b')])]"):
+		# Get the IDs of the direct parent section
+		parent_section = heading.xpath("./ancestor::*[@id and re:test(name(), '^(section|article|nav)$')][1]")
+
+		if parent_section:
+			section = _find_ebook_section(parent_section[0].get_attr("id"), section_tree)
+
+			if not section:
+				# We can accidentally raise s-029 if the file isn't in the spine. This is technically correct, but s-029 is a
+				# misleading message in that case, and f-007 will also be raised, which is the correct message.
+				if filename in self.spine_file_paths:
+					invalid_parent_ids.append(parent_section[0].to_tag_string())
+			else:
+				if str(section.depth) != heading.tag[1:2]:
+					invalid_headers.append(heading.to_string())
 
-				if f"{os.sep}src{os.sep}" not in root:
-					# Check that cover and titlepage images are in all caps
-					if self.cover_path and filename.name == self.cover_path.name:
-						nodes = svg_dom.xpath("//text[re:test(., '[a-z]')]")
-						if nodes:
-							messages.append(LintMessage("s-002", "Lowercase letters in cover. Cover text must be all uppercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	if invalid_parent_ids:
+		messages.append(LintMessage("s-029", "Section with [attr]data-parent[/] attribute, but no section having that [attr]id[/] in ebook.", se.MESSAGE_TYPE_ERROR, filename, invalid_parent_ids))
+	if invalid_headers:
+		messages.append(LintMessage("s-085", "[xhtml]<h#>[/] element found in a [xhtml]<section>[/] or a [xhtml]<article>[/] at an unexpected level. Hint: Headings not in the title page start at [xhtml]<h2>[/]. If this work has parts, should this header be [xhtml]<h3>[/] or higher?", se.MESSAGE_TYPE_ERROR, filename, invalid_headers))
 
-						# For later comparison with titlepage
-						cover_svg_title = svg_dom.xpath("/svg/title/text()", True).replace("The cover for ", "") # <title> can appear on any element in SVG, but we only want to check the root one
+	# Check for a common typo
+	if "z3998:nonfiction" in file_contents:
+		messages.append(LintMessage("s-030", "[val]z3998:nonfiction[/] should be [val]z3998:non-fiction[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-					if filename.name == "titlepage.svg":
-						nodes = svg_dom.xpath("//text[re:test(., '[a-z]') and not(text()='translated by' or text()='illustrated by' or text()='and')]")
-						if nodes:
-							messages.append(LintMessage("s-003", "Lowercase letters in titlepage. Titlepage text must be all uppercase except [text]translated by[/] and [text]illustrated by[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Run some checks on epub:type values
+	incorrect_attrs = set()
+	unnecessary_z3998_attrs = set()
+	duplicate_attrs = []
 
-						# For later comparison with cover
-						titlepage_svg_title = svg_dom.xpath("/svg/title/text()", True).replace("The titlepage for ", "") # <title> can appear on any element in SVG, but we only want to check the root one
+	for node in dom.xpath("//*[@epub:type]"):
+		attrs = set()
 
-			if filename.suffix == ".xml":
-				xml_dom = self.get_dom(filename)
+		for val in regex.split(r"\s+", node.get_attr("epub:type")):
+			if val in attrs:
+				duplicate_attrs.append(node.to_tag_string())
+			else:
+				attrs.add(val)
 
-				if xml_dom.xpath("/search-key-map") and filename.name != "glossary-search-key-map.xml":
-					messages.append(LintMessage("f-013", "Glossary search key map must be named [path]glossary-search-key-map.xml[/].", se.MESSAGE_TYPE_ERROR, filename))
+			if val.startswith("z3998:"):
+				bare_val = val.replace("z3998:", "")
+				if bare_val not in Z3998_SEMANTIC_VOCABULARY:
+					incorrect_attrs.add(val)
 
-				# Make sure that everything in glossaries are in the rest of the text
-				# We’ll check the files later, and log any errors at the end
-				if filename.name == "glossary-search-key-map.xml":
-					has_glossary_search_key_map = True
-					# Map the glossary to tuples of the values and whether they’re used (initially false)
-					glossary_usage = list(map(lambda node: (node.get_attr("value"), False), xml_dom.xpath(".//*[@value]")))
+				elif bare_val in EPUB_SEMANTIC_VOCABULARY:
+					unnecessary_z3998_attrs.add((val, bare_val))
 
-			if filename.suffix == ".xhtml":
-				# Read file contents into a DOM for querying
-				dom = self.get_dom(filename, True)
+			elif val.startswith("se:"):
+				bare_val = val.replace("se:", "")
+				if bare_val not in SE_SEMANTIC_VOCABULARY:
+					incorrect_attrs.add(val)
 
-				# Apply stylesheets.
-				# First apply the browser default stylesheet
-				dom.apply_css(self.get_file(Path("default")), "default")
+			else:
+				# Regular epub vocabulary
+				if val not in EPUB_SEMANTIC_VOCABULARY:
+					incorrect_attrs.add(val)
+
+	if duplicate_attrs:
+		messages.append(LintMessage("s-031", "Duplicate value in [attr]epub:type[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, duplicate_attrs))
+	if incorrect_attrs:
+		messages.append(LintMessage("s-032", "Invalid value for [attr]epub:type[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, incorrect_attrs))
+	if unnecessary_z3998_attrs:
+		messages.append(LintMessage("s-034", "Semantic used from the z3998 vocabulary, but the same semantic exists in the EPUB vocabulary.", se.MESSAGE_TYPE_ERROR, filename, [attr for (attr, bare_attr) in unnecessary_z3998_attrs]))
+
+	# Check if language tags in individual files match the language in the metadata file
+	if filename.name not in IGNORED_FILENAMES:
+		file_language = dom.xpath("/html/@xml:lang", True)
+		if language != file_language:
+			messages.append(LintMessage("s-033", f"File language is [val]{file_language}[/], but [path][link=file://{self.metadata_file_path}]{self.metadata_file_path.name}[/][/] language is [val]{language}[/].", se.MESSAGE_TYPE_WARNING, filename))
+
+	# Check for endnotes
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[not(preceding-sibling::*)]/cite[not(preceding-sibling::node()[normalize-space(.)]) and (following-sibling::node()[normalize-space(.)])[1][contains(@epub:type, 'backlink')]]")
+	if nodes:
+		messages.append(LintMessage("s-035", "Endnote containing only [xhtml]<cite>[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				messages = messages + _get_malformed_urls(dom, filename)
+	# Check for wrong semantics in frontmatter/backmatter
+	if dom.xpath("/html/body/section[not(@data-parent) and re:test(@epub:type, '\\b(dedication|introduction|preface|foreword|preamble|titlepage|halftitlepage|imprint|epigraph|acknowledgements)\\b') and not(ancestor-or-self::*[contains(@epub:type, 'frontmatter')])]"):
+		messages.append(LintMessage("s-036", "No [val]frontmatter[/] semantic inflection for what looks like a frontmatter file.", se.MESSAGE_TYPE_WARNING, filename))
 
-				# Apply any CSS files in the DOM
-				for node in dom.xpath("/html/head/link[@rel='stylesheet']"):
-					css_filename = (filename.parent / node.get_attr("href")).resolve()
-					dom.apply_css(self.get_file(css_filename), str(css_filename))
+	if dom.xpath("/html/body/section[not(@data-parent) and re:test(@epub:type, '\\b(endnotes|loi|afterword|appendix|colophon|copyright\\-page|lot)\\b') and not(ancestor-or-self::*[contains(@epub:type, 'backmatter')])]"):
+		messages.append(LintMessage("s-037", "No [val]backmatter[/] semantic inflection for what looks like a backmatter file.", se.MESSAGE_TYPE_WARNING, filename))
 
-				typos: List[str] = []
+	# Check for leftover asterisms. Asterisms are sequences of any of these chars: * . • -⁠ —
+	nodes = dom.xpath("/html/body//*[self::p or self::div][re:test(., '^\\s*[\\*\\.•\\-⁠—]\\s*([\\*\\.•\\-⁠—]\\s*)+$')]")
+	if nodes:
+		messages.append(LintMessage("s-038", "Illegal asterism. Section/scene breaks must be defined by an [xhtml]<hr/>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Extract ID attributes for later checks
-				id_attrs = id_attrs + dom.xpath("//*[name() != 'section' and name() != 'article' and name() != 'figure' and name() != 'nav']/@id")
+	# Check for <table> element without a <tbody> child
+	if dom.xpath("/html/body//table[not(tbody)]"):
+		messages.append(LintMessage("s-042", "[xhtml]<table>[/] element without [xhtml]<tbody>[/] child.", se.MESSAGE_TYPE_ERROR, filename))
 
-				# Add to the short story count for later checks
-				short_story_count += len(dom.xpath("/html/body//article[contains(@epub:type, 'se:short-story')]"))
+	# Check that short stories are on an <article> element
+	nodes = dom.xpath("/html/body/section[contains(@epub:type, 'se:short-story') or contains(@epub:type, 'se:novella')]")
+	if nodes:
+		messages.append(LintMessage("s-043", "[val]se:short-story[/] semantic on element that is not [xhtml]<article>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-				# Test against word boundaries to not match `halftitlepage`
-				if dom.xpath("/html/body/section[re:test(@epub:type, '\\btitlepage\\b')]"):
-					# Check if the <title> element is set correctly, but only if there's no heading content.
-					# If there's heading content, then <title> should match the expected generated value from the heading content.
-					if dom.xpath("/html[not(./body//*[re:test(name(), '^h(group|[1-6])$')]) and ./head/title[text()!='Titlepage']]"):
-						messages.append(LintMessage("s-025", "Titlepage [xhtml]<title>[/] elements must contain exactly: [text]Titlepage[/].", se.MESSAGE_TYPE_ERROR, filename))
-				else:
-					# Check for common typos
-					# Don't check the titlepage because it has a standard format and may raise false positives
-					typos = regex.findall(r"(?<!’)\b(and and|the the|if if|of of|or or|as as)\b(?!-)", file_contents, flags=regex.IGNORECASE)
-					typos = typos + regex.findall(r"\ba a\b(?!-)", file_contents)
+	# Check for poetry/verse without a descendent <p> element.
+	# Skip the ToC landmarks because it may have poem/verse semantic children.
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')][not(descendant::p)][not(ancestor::nav[contains(@epub:type, 'landmarks')])]")
+	if nodes:
+		messages.append(LintMessage("s-044", "Element with poem or verse semantic, without descendant [xhtml]<p>[/] (stanza) element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-					if typos:
-						messages.append(LintMessage("y-001", "Possible typo: doubled [text]a/the/and/of/or/as/if[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for <abbr> elements that have two or more letters/periods, that don't have a semantic epub:type
+	# SS. is the French abbreviation for "Saints"
+	nodes = dom.xpath("/html/body//abbr[not(@epub:type)][text() != 'U.S.' and text() != 'SS.'][re:test(., '([A-Z]\\.?){2,}')]")
+	filtered_nodes = []
+	for node in nodes:
+		add_node = True
+
+		if node.text in INITIALISM_EXCEPTIONS:
+			add_node = False
+		elif node.get_attr("class"):
+			for attr_value in node.get_attr("class").split():
+				if attr_value in IGNORED_CLASSES:
+					add_node = False
+					break
+		elif node.text in ("A.M.", "P.M.") and node.xpath("./preceding-sibling::node()[re:test(., '[0-9]\\s$')]"):
+			# Ignore instances of capitalized times, usually in titles. For example:
+			# `From 10:20 <abbr>P.M.</abbr> to 10:47 <abbr>P.M.</abbr>` in a chapter title.
+			# We have to test for a preceding number because `<abbr>P.M.</abbr>` in other contexts may mean `Prime Minister`
+			add_node = False
+
+		if add_node:
+			filtered_nodes.append(node)
+
+	if filtered_nodes:
+		messages.append(LintMessage("s-045", "[xhtml]<abbr>[/] element without semantic inflection like [class]z3998:personal-name[/] or [class]z3998:initialism[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in filtered_nodes]))
+
+	# Check for <p> elems that has some element children, which are only <span> and <br> children, but the parent doesn't have poem/verse semantics.
+	# Ignore spans that have a class, but not if the class is an i# class (for poetry indentation)
+	nodes = dom.xpath("/html/body//p[not(./text()[normalize-space(.)])][*][not(ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][not(*[not(self::span) and not(self::br)])][ (not(span[@epub:type]) and not(span[@class]) ) or span[re:test(@class, '\\bi[0-9]\\b')]]")
+	if nodes:
+		messages.append(LintMessage("s-046", "[xhtml]<p>[/] element containing only [xhtml]<span>[/] and [xhtml]<br>[/] elements, but its parent doesn’t have the [val]z3998:poem[/], [val]z3998:verse[/], [val]z3998:song[/], [val]z3998:hymn[/], or [val]z3998:lyrics[/] semantic. Multi-line clauses that are not verse don’t require [xhtml]<span>[/]s.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				is_colophon = bool(dom.xpath("/html/body/section[contains(@epub:type, 'colophon')]"))
-				is_imprint = bool(dom.xpath("/html/body/section[contains(@epub:type, 'imprint')]"))
+	# For this series of selections, we select spans that are direct children of p, because sometimes a line of poetry may have a nested span.
+	nodes = dom.xpath("/html/body/*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')]/descendant-or-self::*/p/span/following-sibling::*[contains(@epub:type, 'noteref') and name()='a' and position()=1]")
+	if nodes:
+		messages.append(LintMessage("s-047", "[val]noteref[/] as a direct child of element with poem or verse semantic. [val]noteref[/]s should be in their parent [xhtml]<span>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				if (is_colophon or is_imprint):
-					if len(source_links) <= 2:
-						# Check that links back to sources are represented correctly
-						nodes = dom.xpath("/html/body//a[@href='https://www.pgdp.net' and text()!='The Online Distributed Proofreading Team']")
-						if nodes:
-							messages.append(LintMessage("m-071", "DP link must be exactly [text]The Online Distributed Proofreading Team[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for incorrectly applied se:name semantic
+	nodes = dom.xpath("/html/body//*[self::p or self::blockquote][contains(@epub:type, 'se:name.')]")
+	if nodes:
+		messages.append(LintMessage("s-048", "[val]se:name[/] semantic on block element. [val]se:name[/] indicates the contents is the name of something.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-						nodes = dom.xpath("/html/body//a[re:test(@href, '^https://www.pgdp.org/ols/') and text()!='Distributed Proofreaders Open Library System']")
-						if nodes:
-							messages.append(LintMessage("m-072", "DP OLS link must be exactly [text]Distributed Proofreaders Open Library System[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for <header> elements with only h# child nodes
+	nodes = dom.xpath("/html/body//header[./*[re:test(name(), 'h[1-6]') and (count(preceding-sibling::*) + count(following-sibling::*)=0)]]")
+	if nodes:
+		messages.append(LintMessage("s-049", "[xhtml]<header>[/] element whose only child is an [xhtml]<h#>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-						nodes = dom.xpath("/html/body//a[re:test(@href, '^https://[^\"]*?hathitrust.org') and re:test(text(), '[Hh]athi') and not(text()='HathiTrust Digital Library')]")
-						if nodes:
-							messages.append(LintMessage("m-041", "Hathi Trust link text must be exactly [text]HathiTrust Digital Library[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for <span>s that only exist to apply epub:type
+	nodes = dom.xpath("/html/body//*[span[@epub:type][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	if nodes:
+		messages.append(LintMessage("s-050", "[xhtml]<span>[/] element appears to exist only to apply [attr]epub:type[/]. [attr]epub:type[/] should go on the parent element instead, without a [xhtml]<span>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-					nodes = dom.xpath("/html/body//p[re:test(., '\\ba transcription\\b') and ./a[contains(@href, '#transcriptions')]]")
-					if nodes:
-						messages.append(LintMessage("t-071", "Multiple transcriptions listed, but preceding text is [text]a transcription[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for language tags misusing the unk lang code
+	nodes = dom.xpath("//*[@xml:lang='unk']")
+	if nodes:
+		messages.append(LintMessage("s-051", "Element with [xhtml]xml:lang=\"unk\"[/] should be [xhtml]xml:lang=\"und\"[/] (undefined).", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				if is_colophon:
-					# Check for illegal Wikipedia URLs
-					nodes = dom.xpath("/html/body//a[contains(@href, '.m.wikipedia.org')]")
-					if nodes:
-						messages.append(LintMessage("m-026", "Illegal [url]https://*.m.wikipedia.org[/] URL. Hint: use non-mobile Wikipedia URLs.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for title attrs on abbr elements
+	nodes = dom.xpath("/html/body//abbr[@title]")
+	if nodes:
+		messages.append(LintMessage("s-052", "[xhtml]<abbr>[/] element with illegal [attr]title[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-					# Check for non-English Wikipedia URLs
-					nodes = dom.xpath("/html/body//a[re:test(@href, 'https://(?!en)[a-z]{2,}\\.wikipedia\\.org')]")
-					if nodes:
-						messages.append(LintMessage("m-043", "Non-English Wikipedia URL.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
+	nodes = dom.xpath("/html/body//blockquote//p[parent::*[name()='footer'] or parent::*[name()='blockquote']]//cite") # Sometimes the <p> may be in a <footer>
+	if nodes:
+		messages.append(LintMessage("s-054", "[xhtml]<cite>[/] as child of [xhtml]<p>[/] in [xhtml]<blockquote>[/]. [xhtml]<cite>[/] should be the direct child of [xhtml]<blockquote>[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-					# Check for wrong grammar filled in from template
-					nodes = dom.xpath("/html/body//a[starts-with(@href, 'https://books.google.com/') or starts-with(@href, 'https://www.google.com/books/')][(preceding-sibling::text()[normalize-space(.)][1])[re:test(., '\\bthe$')]]")
-					if nodes:
-						messages.append(LintMessage("s-016", "Incorrect [text]the[/] before Google Books link.", se.MESSAGE_TYPE_ERROR, filename, ["the<br/>\n" + node.to_string() for node in nodes]))
+	# Check for <th> element without a <thead> ancestor. However, <th scope="...">  and <th/> are allowed, for use in tables with headers in the middle of tables (url:https://standardebooks.org/ebooks/dorothy-day/the-eleventh-virgin) and vertical table headers
+	# (https://standardebooks.org/ebooks/charles-babbage/passages-from-the-life-of-a-philosopher)
+	if dom.xpath("/html/body//table//th[not(ancestor::thead)][not(@scope)][not(count(node())=0)]"):
+		messages.append(LintMessage("s-055", "[xhtml]<th>[/] element not in [xhtml]<thead>[/] ancestor. Note: [xhtml]<th>[/] elements used as mid-table headings or horizontal row headings require the [attr]scope[/] attribute.", se.MESSAGE_TYPE_ERROR, filename))
 
-					se_url = self.generated_identifier.replace("url:", "")
-					if not dom.xpath(f"/html/body//a[@href='{se_url}' and text()='{se_url.replace('https://', '')}']"):
-						messages.append(LintMessage("m-035", f"Unexpected S.E. identifier in colophon. Expected: [url]{se_url}[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for z3998:stage-direction on elements that are not <i>
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:stage-direction') and name() !='i' and name() !='abbr' and name() !='p']")
+	if nodes:
+		messages.append(LintMessage("s-058", "[attr]z3998:stage-direction[/] semantic only allowed on [xhtml]<i>[/], [xhtml]<abbr>[/], and [xhtml]<p>[/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-					if self.metadata_dom.xpath("/package/metadata/meta[@property='role' and text()='trl']") and "translated from" not in file_contents:
-						messages.append(LintMessage("m-025", "Translator found in metadata, but no [text]translated from LANG[/] block in colophon.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check that internal links don't begin with ../
+	nodes = dom.xpath("/html/body//a[re:test(@href, '^\\.\\./text/')]")
+	if nodes:
+		messages.append(LintMessage("s-059", "Internal link beginning with [val]../text/[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-					if ebook_has_multiple_transcriptions and not dom.xpath("/html/body//a[contains(@href, '#transcriptions')]"):
-						messages.append(LintMessage("m-074", "Multiple transcriptions found in metadata, but no link to [text]EBOOK_URL#transcriptions[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for italics on things that shouldn't be italics
+	nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.music.song') or contains(@epub:type, 'se:name.publication.short-story') or contains(@epub:type, 'se:name.publication.essay')]")
+	if nodes:
+		messages.append(LintMessage("s-060", "Italics on name that requires quotes instead.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-					if ebook_has_multiple_page_scans and not dom.xpath("/html/body//a[contains(@href, '#page-scans')]"):
-						messages.append(LintMessage("m-075", "Multiple page scans found in metadata, but no link to [text]EBOOK_URL#page-scans[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for h# tags followed by header content, that are not children of <header>.
+	# Only match if there is a following <p>, because we could have the case where there's an epigraph after a division title.
+	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][following-sibling::*[contains(@epub:type, 'epigraph') or contains(@epub:type, 'bridgehead')]][following-sibling::p][not(parent::header)]")
+	if nodes:
+		messages.append(LintMessage("s-061", "Title and following header content not in a [xhtml]<header>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-					# Check that the formula changed from the default if we added 'various sources'
-					if ebook_has_multiple_transcriptions or ebook_has_multiple_page_scans:
-						nodes = dom.xpath("/html/body//a[text() = 'various sources' and not(re:test(preceding-sibling::br[1]/preceding-sibling::node()[1], '(digital scans|transcriptions) from\\s*$'))]")
-						if nodes:
-							messages.append(LintMessage("t-072", "[text]various sources[/] link not preceded by [text]from[/].", se.MESSAGE_TYPE_ERROR, filename))
-
-					# Check if we forgot to fill any variable slots
-					missing_colophon_vars = [var for var in SE_VARIABLES if regex.search(fr"\b{var}\b", file_contents)]
-					if missing_colophon_vars:
-						messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, filename, missing_colophon_vars))
-
-					# Check that we have <br/>s at the end of lines
-					# First, check for b or a elements that are preceded by a newline but not by a br
-					nodes = [node.to_string() for node in dom.xpath("/html/body/section/p/*[name()='b' or name()='a'][(preceding-sibling::node()[1])[contains(., '\n')]][not((preceding-sibling::node()[2])[self::br]) or (normalize-space(preceding-sibling::node()[1]) and re:test(preceding-sibling::node()[1], '\\n\\s*$')) ]")]
-					# Next, check for text nodes that contain newlines but are not preceded by brs
-					nodes = nodes + [node.strip() for node in dom.xpath("/html/body/section/p/text()[contains(., '\n') and normalize-space(.)][(preceding-sibling::node()[1])[not(self::br)]]")]
-					if nodes:
-						messages.append(LintMessage("s-053", "Colophon line not preceded by [xhtml]<br/>[/].", se.MESSAGE_TYPE_ERROR, filename, nodes))
-
-					# Is there a comma after a producer name, if there's only two producers?
-					nodes = dom.xpath("/html/body/section/p/*[name()='b' or name()='a'][(following-sibling::node()[1])[normalize-space(.)=', and']][(preceding-sibling::*[1])[name()='br']]")
-					if nodes:
-						messages.append(LintMessage("t-006", "Comma after producer name, but there are only two producers.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-					nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and re:test(@epub:type, 'z3998:.*?name')]")
-					if nodes:
-						messages.append(LintMessage("s-092", "Anonymous contributor with [val]z3998:*-name[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-					# Check for anonymous volunteers misrepresented in the colophon. Note that we only match SE producers and transcribers, because
-					# a cover art artist can also be anonymous but they're not volunteers.
-					nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and text() != 'An Anonymous Volunteer' and (preceding-sibling::a[@href='https://standardebooks.org'])]")
-					if nodes:
-						messages.append(LintMessage("s-100", "Anonymous digital contributor value not exactly [text]An Anonymous Volunteer[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-					# Check for primary contributors misrepresented in the colophon. these differ from ebook production volunteers.
-					nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and text() != 'Anonymous' and (preceding-sibling::node()[contains(., 'The cover page')] or preceding-sibling::i[contains(@epub:type, 'se:name.publication')])]")
-					if nodes:
-						messages.append(LintMessage("s-101", "Anonymous primary contributor value not exactly [text]Anonymous[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-					# Is there a page scan link in the colophon, but missing in the metadata?
-					for node in dom.xpath("/html/body//a[re:test(@href, '(gutenberg\\.org/ebooks/[0-9]+|hathitrust\\.org|/archive\\.org|books\\.google\\.com|www\\.google\\.com/books/)')]"):
-						if not self.metadata_dom.xpath(f"/package/metadata/dc:source[contains(text(), {se.easy_xml.escape_xpath(node.get_attr('href'))})]"):
-							messages.append(LintMessage("m-059", f"Link to [url]{node.get_attr('href')}[/] found in colophon, but missing matching [xhtml]dc:source[/] element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-
-					# Are the sources represented correctly?
-					# We don't have a standard yet for more than two sources (transcription and scan) so just ignore that case for now.
-					# We can't merge this with the imprint check because imprint doesn't have `<br/>` between `the`
-					if not ebook_has_multiple_transcriptions and not ebook_has_other_sources:
-						for link in source_links:
-							if "gutenberg.org" in link and f"<a href=\"{link}\">Project Gutenberg</a>" not in file_contents:
-								messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Project Gutenberg</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
-
-					if not ebook_has_multiple_page_scans and not ebook_has_other_sources:
-						for link in source_links:
-							if "hathitrust.org" in link and f"the<br/>\n\t\t\t<a href=\"{link}\">HathiTrust Digital Library</a>" not in file_contents:
-								messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]the<br/> <a href=\"{link}\">HathiTrust Digital Library</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
-
-							if "archive.org" in link and f"the<br/>\n\t\t\t<a href=\"{link}\">Internet Archive</a>" not in file_contents:
-								messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]the<br/> <a href=\"{link}\">Internet Archive</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
-
-							if ("books.google.com" in link or "www.google.com/books/" in link) and f"<a href=\"{link}\">Google Books</a>" not in file_contents:
-								messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Google Books</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for <dt> elements without exactly one <dfn> child, but only in glossaries
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'glossary')]//dt[not(count(./dfn)=1)]")
+	if nodes:
+		messages.append(LintMessage("s-062", "[xhtml]<dt>[/] element in a glossary without exactly one [xhtml]<dfn>[/] child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				else:
-					# This file is not the colophon.
-					# Check for ending punctuation inside italics that have semantics.
-					# Ignore the colophon because paintings might have punctuation in their names
+	# Check that `z3998:persona` is only on <b> or <td>. We use the contact() xpath function so we don't catch `z3998:personal-name`
+	nodes = dom.xpath("/html/body//*[contains(concat(' ', @epub:type, ' '), ' z3998:persona ') and not(self::b or self::td)]")
+	if nodes:
+		messages.append(LintMessage("s-063", "[val]z3998:persona[/] semantic on element that is not a [xhtml]<b>[/] or [xhtml]<td>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-					# This xpath matches b or i elements with epub:type="se:name...", that are not stage direction, whose last text node ends in punctuation.
-					# Note that we check that the last node is a text node, because we may have <abbr> as the last node
-					matches = [node.to_string() for node in dom.xpath("(//b | //i)[contains(@epub:type, 'se:name') and not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., '[\\.,!\\?;:]$')]]")]
+	# Check for fulltitle semantic on a header not in the half title
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'fulltitle') and name()!='h2' and name()!='hgroup' and not(ancestor::*[contains(@epub:type, 'halftitlepage')])]")
+	if nodes:
+		messages.append(LintMessage("s-065", "[val]fulltitle[/] semantic on element that is not in the half title.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-					# Match b or i elements that are not stage directions, and that end in a comma followed by a lowercase letter
-					matches = matches + [node.to_string() for node in dom.xpath("(//b | //i)[not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., ',$')] and following-sibling::node()[re:test(., '^\\s*[a-z]')] ]")]
+	# Check for header elements that have a label, but are missing the label semantic
+	# Find h# nodes whose first child is a text node matching a label type, and where that text node's next sibling is a semantic roman numeral
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./node()[1][self::text() and not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)$') and following-sibling::*[1][contains(@epub:type, 'z3998:roman')]]]")
+	if nodes:
+		messages.append(LintMessage("s-066", "Header element missing [val]label[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-					# ...and also check for ending punctuation inside em tags, if it looks like a *part* of a clause
-					# instead of a whole clause. If the <em> is preceded by an em dash or quotes, or if there's punctuation
-					# and a space before it, then it's presumed to be a whole clause.
-					matches = matches + [match.strip() for match in regex.findall(r"(?<!.[—“‘>]|[!\.\?…;:]\s)<em>(?:\w+?\s*)+[\.,\!\?;]</em>", file_contents) if match.islower()]
+	# Check for header elements that have a label semantic, but are missing an ordinal sibling
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./span[contains(@epub:type, 'label')]][not(./span[contains(@epub:type, 'ordinal')])]")
+	if nodes:
+		messages.append(LintMessage("s-067", "Header element with a [val]label[/] semantic child, but without an [val]ordinal[/] semantic child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-					if matches:
-						messages.append(LintMessage("t-017", "Ending punctuation inside formatting like bold, small caps, or italics. Ending punctuation is only allowed within formatting if the phrase is an independent clause.", se.MESSAGE_TYPE_WARNING, filename, list(set(matches))))
+	# Check for header elements with a roman semantic but without an ordinal semantic
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][contains(@epub:type, 'z3998:roman') and not(contains(@epub:type, 'ordinal'))]")
+	if nodes:
+		messages.append(LintMessage("s-068", "Header element missing [val]ordinal[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for unused selectors
-				if dom.xpath("/html/head/link[contains(@href, 'local.css')]"):
-					for selector in local_css_selectors:
-						try:
-							if dom.css_select(selector):
-								unused_selectors.remove(selector)
-						except lxml.cssselect.ExpressionError:
-							# This gets thrown on some selectors not yet implemented by lxml, like *:first-of-type
-							unused_selectors.remove(selector)
-							continue
-						except Exception as ex:
-							raise se.InvalidCssException(f"Couldn’t parse CSS in or near this line: [css]{selector}[/]. Exception: {ex}")
+	# Check for body element without child section or article. Ignore the ToC because it has a unique structure
+	nodes = dom.xpath("/html/body[not(./*[name()='section' or name()='article' or (name()='nav' and re:test(@epub:type, '\\b(toc|loi)\\b'))])]")
+	if nodes:
+		messages.append(LintMessage("s-069", "[xhtml]<body>[/] element missing direct child [xhtml]<section>[/] or [xhtml]<article>[/] element.", se.MESSAGE_TYPE_ERROR, filename))
 
-				# Update our list of local.css selectors to check in the next file
-				local_css_selectors = list(unused_selectors)
+	# Check for h# without semantics; h# that have child elements (which likely have the correct semantics) are OK
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(@epub:type)][not(./*[not(name()='a' and contains(@epub:type, 'noteref'))])]")
+	if nodes:
+		messages.append(LintMessage("s-070", "[xhtml]<h#>[/] element without semantic inflection.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Done checking for unused selectors.
+	# Check for sectioning elements with more than one heading element
+	nodes = dom.xpath("/html/body//*[name()='article' or name()='section'][count(./*[name()='header' or name()='hgroup' or re:test(name(), '^h[1-6]$')]) > 1]")
+	if nodes:
+		messages.append(LintMessage("s-071", "Sectioning element with more than one heading element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-				# Check if this is a frontmatter file, but exclude the titlepage, imprint, and toc
-				if dom.xpath("/html//*[contains(@epub:type, 'frontmatter') and not(descendant-or-self::*[re:test(@epub:type, '\\b(titlepage|imprint|toc)\\b')])]"):
-					has_frontmatter = True
+	# Check for some elements that only have a <span> child (including text inline to the parent).
+	# But, exclude such elements that are <p> elements that have poetry-type parents.
+	nodes = dom.xpath("/html/body//*[name() !='p' and not(ancestor-or-self::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][./span[count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	if nodes:
+		messages.append(LintMessage("s-072", "Element with single [xhtml]<span>[/] child. [xhtml]<span>[/] should be removed and its attributes promoted to the parent element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Add new CSS classes to global list
-				if filename.name not in IGNORED_FILENAMES:
-					for node in dom.xpath("//*[@class]"):
-						for css_class in node.get_attr("class").split():
-							if css_class in xhtml_css_classes:
-								xhtml_css_classes[css_class] += 1
-							else:
-								xhtml_css_classes[css_class] = 1
+	# Check for header elements that are missing both label and ordinal semantics
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)\\s+[ixvIXVmcd]+$')]")
+	if nodes:
+		messages.append(LintMessage("s-073", "Header element that requires [val]label[/] and [val]ordinal[/] semantic children.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for whitespace before noteref
-				# Do this early because we remove noterefs from headers later
-				# Allow noterefs that do not have preceding text/elements (for example a "see note X" ref that is the only child in a <td>)
-				nodes = dom.xpath("/html/body//a[contains(@epub:type, 'noteref') and preceding-sibling::node()[normalize-space(.)] and re:test(preceding-sibling::node()[1], '\\s+$')]")
-				if nodes:
-					messages.append(LintMessage("t-012", "Illegal white space before noteref.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for table headers that don't have content, which is an accessibility issue.
+	# Note that @aria-label and @title should (apparently) not be used for table headers
+	nodes = dom.xpath("/html/body//th[not(text())]")
+	if nodes:
+		messages.append(LintMessage("s-074", "[xhtml]<th>[/] element with no text content should be a [xhtml]<td>[/] element instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check that internal links don't begin with ../
-				nodes = dom.xpath("/html/body//a[re:test(@href, '^\\.\\./text/')]")
-				if nodes:
-					messages.append(LintMessage("s-059", "Internal link beginning with [val]../text/[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for incorrect children of <body>
+	nodes = dom.xpath("/html/body/*[name() != 'section' and name() != 'article' and name() != 'nav']")
+	if nodes:
+		messages.append(LintMessage("s-075", "[xhtml]<body>[/] element with direct child that is not [xhtml]<section>[/], [xhtml]<article>[/], or [xhtml]<nav>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-				# Get the title of this file to compare against the ToC later.
-				# We ignore the ToC file itself.
-				# Also ignore files that have more than 3 top-level sections or articles, as these are probably compilation works that will have unique titles.
-				if not dom.xpath("/html/body/nav[contains(@epub:type, 'toc')]") and not dom.xpath("/html/body[count(./section) + count(./article) > 3]"):
-					try:
-						header_text = dom.xpath("/html/head/title/text()")[0]
-					except Exception:
-						header_text = ""
+	# Check for lang="" instead of xml:lang=""
+	nodes = dom.xpath("/html/body//*[@lang]")
+	if nodes:
+		messages.append(LintMessage("s-076", "[attr]lang[/] attribute used instead of [attr]xml:lang[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-					if header_text != "":
-						headings.append((header_text, str(filename)))
+	# Check for <header> preceded by non-sectioning elements
+	nodes = dom.xpath("/html/body//header[./preceding-sibling::*[not(re:test(name(), '^(section|div|article)$'))]]")
+	if nodes:
+		messages.append(LintMessage("s-077", "[xhtml]<header>[/] element preceded by non-sectioning element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for illegal characters
-				matches = regex.findall(fr"({se.UNICODE_BOM}|{se.SHY_HYPHEN})", file_contents)
-				if matches:
-					# Get the keys of a dict in order to create a list without duplicates
-					messages.append(LintMessage("t-058", "Illegal character.", se.MESSAGE_TYPE_ERROR, filename, list({match.encode("unicode_escape").decode().replace("\\u", "U+").upper():None for match in matches}.keys())))
+	# Check for <footer> followed by non-sectioning elements
+	nodes = dom.xpath("/html/body//footer[./following-sibling::*[not(re:test(name(), '^(section|div|article)$'))]]")
+	if nodes:
+		messages.append(LintMessage("s-078", "[xhtml]<footer>[/] element followed by non-sectioning element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for z3998:roman elements with invalid values. Roman numerals can occasionally end in `j` as an alias for ending `i`. See _The Worm Ouroboros_.
-				# We also allow the numeral to end in a digit, because that might be an endnote. For example `<h2 epub:type="ordinal z3998:roman">II<a href="..." epub:type="noteref">3</a></h2>`
-				nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:roman') and not(re:test(normalize-space(.), '^[ivxlcdmIVXLCDM]+j?[0-9]*?$'))]")
-				if nodes:
-					messages.append(LintMessage("s-026", "Invalid Roman numeral.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for elements with no children and only white space contents
+	nodes = dom.xpath("/html/body//*[not(./*) and re:test(., '^\\s+$')]")
+	if nodes:
+		messages.append(LintMessage("s-079", "Element containing only white space.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check that `Internet Archive` and `HathiTrust` are preceded by `the`. They might have an immediate preceding text node that ends in `the`, or they
-				# might be preceded by a white space node, then a <br/>, then a text node ending in `the`.
-				nodes = dom.xpath("/html/body//a[(re:test(text(), '[Hh]athi') and re:test(@href, '^https://[^\"]*?hathitrust.org')) or (re:test(text(), 'Internet Archive') and re:test(@href, 'https://[^\"]*?archive.org'))][(preceding-sibling::node()[2][not(self::br)] and preceding-sibling::node()[1][not(re:test(., '\\sthe $'))]) or (preceding-sibling::node()[2][self::br] and preceding-sibling::node()[3][not(re:test(., '\\sthe$'))]) ]")
-				if nodes:
-					messages.append(LintMessage("m-061", "Link must be preceded by [text]the[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for a play <td> that contains both inline text and block-level elements.
+	# Note that we check any child of <html> because often <body> has the drama semantic.
+	nodes = dom.xpath("/html//*[contains(@epub:type, 'z3998:drama')]//td[./text()[normalize-space(.)] and ./*[name() = 'p' or name() = 'blockquote' or name() = 'div']]")
+	if nodes:
+		messages.append(LintMessage("s-080", "[xhtml]<td>[/] in drama containing both inline text and a block-level element. All children should either be only text, or only block-level elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for uppercase letters in IDs or classes
-				nodes = dom.xpath("//*[re:test(@id, '[A-Z]') or re:test(@class, '[A-Z]') or re:test(@epub:type, '[A-Z]')]")
-				if nodes:
-					messages.append(LintMessage("x-002", "Uppercase in attribute value. Attribute values must be all lowercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for <p> tags that follow figure, blockquote, or table, and start with a lowercase letter, but that don't have the `continued` class. Exclude matches whose first child is <cite>, so we don't match things like <p><cite>—Editor</cite>
+	nodes = dom.xpath("/html/body//p[preceding-sibling::*[1][name() = 'figure' or name() = 'blockquote' or name() = 'table'] and re:test(., '^([a-z]|—|…)') and not(contains(@class, 'continued')) and not(./*[1][self::cite])]")
+	if nodes:
+		messages.append(LintMessage("s-081", "[xhtml]<p>[/] preceded by [xhtml]<figure>[/], [xhtml]<blockquote>[/xhtml], or [xhtml]<table>[/], but without [val]continued[/] class.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				nodes = dom.xpath("//*[re:test(@id, '^[0-9]+')]")
-				if nodes:
-					messages.append(LintMessage("x-007", "[attr]id[/] attributes starting with a number are illegal XHTML.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for language tags transliterated into Latin script but missing `-Latn` suffix
+	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '^(he|ru|el|zh|bn|hi|sa|uk|yi|grc)$') and re:test(., '[a-zA-Z]')]")
+	if nodes:
+		messages.append(LintMessage("s-082", "Element containing Latin script for a non-Latin-script language, but its [attr]xml:lang[/] attribute value is missing the [val]-Latn[/] language tag suffix. Hint: For example Russian transliterated into Latin script would be [val]ru-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for <section> and <article> without ID attribute
-				nodes = dom.xpath("/html/body//*[self::section or self::article][not(@id)]")
-				if nodes:
-					messages.append(LintMessage("s-011", "Element without [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for persona <td>s that have child <p> elements
+	nodes = dom.xpath("/html/body//td[contains(@epub:type, 'z3998:persona') and ./p]")
+	if nodes:
+		messages.append(LintMessage("s-083", "[xhtml]<td epub:type=\"z3998:persona\">[/] element with child [xhtml]<p>[/] element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				for node in dom.xpath("/html/body//*[@id]/@id"):
-					if node in id_values:
-						duplicate_id_values.append(node)
-					else:
-						id_values[node] = True
+	# Check for incorrect semantics on some common publications
+	nodes = dom.xpath("/html/body//i[@epub:type='se:name.publication.book' and re:test(., '^(The )?(Iliad|Odyssey|Aeneid|Metamorphoses|Beowulf|Divine Comedy|Paradise Lost)$')]")
+	if nodes:
+		messages.append(LintMessage("s-084", "Poem has incorrect semantics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for numeric entities
-				matches = regex.findall(r"&#[0-9]+?;", file_contents)
-				if matches:
-					messages.append(LintMessage("s-001", "Illegal numeric entity (like [xhtml]Α[/]).", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for `Op. Cit.` in endnotes. `Op. Cit.` means "the previous reference" which doesn't make sense
+	# in popup endnotes. But, if the endnote has a book reference, then allow it as it might be referring to that.
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote') and (re:test(., '(Loc\\.|Op\\.) Cit\\.', 'i') or re:test(., '\\bl\\.c\\.\\b', 'i')) and not(.//*[contains(@epub:type, 'se:name.publication')])]")
+	if nodes:
+		messages.append(LintMessage("s-086", "[text]Op. Cit.[/] or [text]Loc. Cit.[/] in endnote. Hint: [text]Op. Cit.[/] and [text]Loc. Cit.[/] mean [text]the previous reference[/], which usually doesn’t make sense in a popup endnote. Such references should be expanded.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+
+	# Check for half title pages missing subtitles
+	if ebook_info["has_subtitle"]:
+		# Make sure we exclude <a> because that appears in the ToC landmarks
+		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and not(.//*[contains(@epub:type, 'subtitle')])]")
+		if nodes:
+			messages.append(LintMessage("s-087", "Subtitle in metadata, but no subtitle in the half title page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+
+	else:
+		# Make sure we exclude <a> because that appears in the ToC landmarks
+		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and .//*[contains(@epub:type, 'subtitle')]]")
+		if nodes:
+			messages.append(LintMessage("s-088", "Subtitle in half title page, but no subtitle in metadata.", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Check for MathML without alttext
+	nodes = dom.xpath("/html/body//m:math[not(@alttext)]")
+	if nodes:
+		messages.append(LintMessage("s-089", "MathML missing [attr]alttext[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check nested <blockquote> elements, but only if it's the first child of another <blockquote>
-				nodes = dom.xpath("/html/body//blockquote/*[1][name()='blockquote']")
-				if nodes:
-					messages.append(LintMessage("s-005", "Nested [xhtml]<blockquote>[/] element.", se.MESSAGE_TYPE_WARNING, filename))
+	# Check for common errors in language tags
+	# `gr` is often used instead of `el`, `sp` instead of `es`, and `ge` instead of `de` (`ge` is the Georgian geographic region subtag but not a language subtag itself)
+	nodes = dom.xpath("//*[re:test(@xml:lang, '^(gr|sp|ge)$')]")
+	if nodes:
+		messages.append(LintMessage("s-090", "Invalid language tag.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-				# Check for <hr> tags before the end of a section, which is a common PG artifact
-				if dom.xpath("/html/body//hr[count(following-sibling::*)=0]"):
-					messages.append(LintMessage("s-012", "Illegal [xhtml]<hr/>[/] as last child.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for <span> in poetry not followed by <br/>. Ignore spans that are roman as they might be present in poem headers, or spans with xml:lang.
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]//p/span[not(@epub:type='z3998:roman') and not(@xml:lang) and following-sibling::*[1][name()='span']]")
+	if nodes:
+		messages.append(LintMessage("s-091", "[xhtml]<span>[/] not followed by [xhtml]<br/>[/] in poetry.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for space after dash
-				nodes = dom.xpath("/html/body//*[name()='p' or name()='span' or name='em' or name='i' or name='b' or name='strong'][not(self::comment())][re:test(., '[a-zA-Z]-\\s(?!(and|or|nor|to|und|…)\\b)')]")
-				if nodes:
-					messages.append(LintMessage("t-033", "Space after dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for nested `<abbr>` elements
+	nodes = dom.xpath("/html/body//abbr[./abbr]")
+	if nodes:
+		messages.append(LintMessage("s-093", "Nested [xhtml]<abbr>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for double greater-than at the end of a tag
-				matches = regex.findall(r"(>>|>>)", file_contents)
-				if matches:
-					messages.append(LintMessage("x-008", "Elements should end with a single [text]>[/].", se.MESSAGE_TYPE_WARNING, filename))
-
-				# Check for lone acute accents, which should always be combined or swapped for a more accurate Unicode char
-				if "´" in file_contents:
-					messages.append(LintMessage("t-055", "Lone acute accent ([val]´[/]). A more accurate Unicode character like prime for coordinates or measurements, or combining accent or breathing mark for Greek text, is required.", se.MESSAGE_TYPE_ERROR, filename))
-
-				# Check for periods followed by lowercase.
-				temp_xhtml = regex.sub(r"<title[^>]*?>.+?", "", file_contents) # Remove  because it might contain something like <title>Chapter 2: The Antechamber of M. de Tréville
-				temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Replace things like 
-				temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Remove  attributes
-				temp_xhtml = temp_xhtml.replace("A.B.C.", "X") # Remove A.B.C, which is not an abbreviations.
-				# Note the regex also excludes preceding numbers, so that we can have inline numbering like:
-				# "A number of questions: 1. regarding those who make heretics; 2. concerning those who were made heretics..."
-				matches = regex.findall(r"[^\s0-9]+\.\s+[\p{Lowercase_Letter}](?!’[\p{Uppercase_Letter}])[\p{Lowercase_Letter}]+", temp_xhtml)
-				# If  is in the match, remove it from the matches so we exclude things like et. al.
-				matches = [match for match in matches if "" not in match]
-				if matches:
-					messages.append(LintMessage("t-029", "Period followed by lowercase letter. Hint: Abbreviations require an [xhtml][/] element.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Check for language tags transliterated into Latin script with incorrect `-latn` suffix (lowercase l)
+	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '-latn')]")
+	if nodes:
+		messages.append(LintMessage("s-094", "Element has an [attr]xml:lang[/] attribute that incorrectly contains [val]-latn[/] instead of [val]-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for nbsp before times
-				nodes = dom.xpath(f"/html/body//text()[re:test(., '[0-9][^{se.NO_BREAK_SPACE}]?$')][(following-sibling::abbr[1])[re:test(., '^[ap]\\.m\\.$')]]")
-				if nodes:
-					messages.append(LintMessage("t-009", "Required no-break space not found before time and [text]a.m.[/] or [text]p.m.[/].", se.MESSAGE_TYPE_WARNING, filename, [node[-10:] + "[/] child of [xhtml]
    [/] in poetry/verse does not have [css]text-align: center;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for uppercased am/pm, except for if they have an epub:type, or when uppercased in titles, or in the nav document
-				if not dom.xpath("//nav[contains(@epub:type, 'toc')]"):
-					nodes = dom.xpath("/html/body//abbr[re:test(., '^[AP]\\.M\\.$') and not(@epub:type) and not(./ancestor::*[contains(@epub:type, 'title')])]")
-					if nodes:
-						messages.append(LintMessage("t-062", "Uppercased [text]a.m.[/] and [text]p.m.[/]", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check that half titles without subtitles have the fulltitle semantic
+	nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage')]/*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'fulltitle'))]")
+	if nodes:
+		messages.append(LintMessage("s-096", "Heading element in half title page missing the [val]fulltitle[/] semantic.", se.MESSAGE_TYPE_ERROR, filename))
 
-				# Check for low-hanging misquoted fruit
-				matches = regex.findall(r"[\p{Letter}]+[“‘]", file_contents) + regex.findall(r"[^>]+‘[\p{Lowercase_Letter}]+", file_contents)
-				if matches:
-					messages.append(LintMessage("t-028", "Possible mis-curled quotation mark.", se.MESSAGE_TYPE_WARNING, filename, matches))
-
-				# Check for times with periods instead of colons.
-				# Only check p, because things like tables/td are more likely to contain non-time numbers
-				# Exclude numbers preceded by equals, or succeeded by some measurements
-				# Also remove  first because they are likely to contain numbered section references
-				dom_copy = deepcopy(dom)
-				for node in dom_copy.xpath("/html/body//p/a"):
-					node.remove()
-
-				nodes = dom_copy.xpath("/html/body//p[re:test(., '[^=]\\s[0-9]{1,2}\\.[0-9]{2}(?![0-9′″°%]|\\.[0-9]|\\scubic|\\smetric|\\smeters|\\smiles|\\sfeet|\\sinches)')]")
-				matches = []
-				for node in nodes:
-					for time_match in regex.findall(r"(?<=[^=]\s)[0-9]{1,2}\.[0-9]{2}(?![0-9′″°%]|\.[0-9]|\scubic|\smetric|\smeters|\smiles|\sfeet|\sinches)", node.inner_text()):
-						time = time_match.split(".")
-						if not time[0].startswith("0") and int(time[0]) >= 1 and int(time[0]) <= 12 and int(time[1]) >= 0 and int(time[1]) <= 59:
-							matches.append(time_match)
+	# Check for missing href attributes, sometimes a leftover from PG transcriptions
+	nodes = dom.xpath("/html/body//a[not(@href)]")
+	if nodes:
+		messages.append(LintMessage("s-097", "[xhtml]a[/] element without [attr]href[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				if matches:
-					messages.append(LintMessage("t-010", "Time set with [text].[/] instead of [text]:[/].", se.MESSAGE_TYPE_WARNING, filename, set(matches)))
+	# Check for s with only one child
+	nodes = dom.xpath("/html/body/*[name() = 'section' or name() = 'article']/header[./*[not(name() = 'p') and not(preceding-sibling::* or following-sibling::*)]]")
+	if nodes:
+		messages.append(LintMessage("s-098", "[xhtml]
    [/] element with only one child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Do we have a half title?
-				# Sometimes the half title might not be a section, like in Cane by Jean Toomer
-				if dom.xpath("/html/body//*[contains(@epub:type, 'halftitlepage')]"):
-					has_halftitle = True
+	# Check for @lang attributes
+	nodes = dom.xpath("//*[@lang]")
+	if nodes:
+		messages.append(LintMessage("s-102", "[attr]lang[/] attribute detected. Hint: Use [attr]xml:lang[/] instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-				# Check for leading 0 in IDs (note: not the same as checking for IDs that start with an integer)
-				# We only check for *leading* 0s in numbers; this allows IDs like `wind-force-0` in the Worst Journey in the World glossary.
-				nodes = dom.xpath("//*[re:test(@id, '-0[0-9]')]")
-				if nodes:
-					messages.append(LintMessage("x-009", "Illegal leading 0 in [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for common missing roman semantics for “I”
+	regent_regex = r"(?:Charles|Edward|George|Henry|James|William) I\b"
+	matches = regex.findall(fr"King {regent_regex}", file_contents) + regex.findall(fr"{regent_regex}’s", file_contents)
+	if matches:
+		messages.append(LintMessage("s-103", "Probable missing semantics for a roman I numeral.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-				# Check for underscores in attributes, but not if the attribute is href (links often have underscores) or MathMl alttext, which can have underscores as part of math notation
-				nodes = dom.xpath("//@*[contains(., '_') and name() !='href' and name() !='alttext']/..")
-				if nodes:
-					messages.append(LintMessage("x-011", "Illegal underscore in attribute. Use dashes instead of underscores.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	return messages
 
-				# Check for stage direction that ends in ?! but also has a trailing period
-				nodes = dom.xpath("/html/body//i[contains(@epub:type, 'z3998:stage-direction') and re:test(., '\\.$') and not((./node()[last()])[name() = 'abbr']) and (following-sibling::node()[1])[re:test(., '^[\\.,:;!?]')]]")
-				if nodes:
-					messages.append(LintMessage("t-018", "Stage direction ending in period next to other punctuation. Remove trailing periods in stage direction.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+def _lint_metadata_checks(filename: Path, dom: se.easy_xml.EasyXmlTree) -> list:
+	"""
+	Helper function used in self.lint()
+	Process metadata checks on an .xhtml file
 
-				# Check for h# without semantics; h# that have child elements (which likely have the correct semantics) are OK
-				nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(@epub:type)][not(./*[not(name()='a' and contains(@epub:type, 'noteref'))])]")
-				if nodes:
-					messages.append(LintMessage("s-070", "[xhtml][/] element without semantic inflection.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
 
-				# Check for  element without a  child
-				if dom.xpath("/html/body//table[not(tbody)]"):
-					messages.append(LintMessage("s-042", "[xhtml]
    [/] element without [xhtml][/] child.", se.MESSAGE_TYPE_ERROR, filename))
+	OUTPUTS
+	A list of LintMessages representing metadata errors found in the file
+	"""
 
-				# Check for  ancestor. However, [/] ancestor. Note: [xhtml]", "", file_contents), flags=regex.DOTALL)
-
-				# We create a filter to try to exclude nested quotations
-				# Remove tags in case they're enclosing punctuation we want to match against at the end of a sentence.
-				typos = [match for match in typos if not regex.search(r"(?:[\.!\?;…—]|”\s)’\s", se.formatting.remove_tags(match))]
-
-				# Try some additional matches before adding the lint message
-				# Search for 
     tags that have an ending closing quote but no opening quote; but exclude 
    s that are preceded by a 
    
-				# or that have a 
     ancestor, because that may indicate that the opening quote is elsewhere in the quotation.
-				for node in dom.xpath("//p[re:test(., '^[^“]+”') and not(./preceding-sibling::*[1][name() = 'blockquote']) and not(./ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]) and not(./ancestor::blockquote)]"):
-					typos.append(node.to_string()[-20:])
-
-				if typos:
-					messages.append(LintMessage("y-019", "Possible typo: [text]”[/] without opening [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for ,.
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., ',\\.')]")]
-				if typos:
-					messages.append(LintMessage("y-020", "Possible typo: consecutive comma-period ([text],.[/]).", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for single quotes instead of double quotes
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[not(contains(@class, 'continued') or ancestor::blockquote or ancestor::*[re:test(@epub:type, 'z3998:(verse|song|poem|hymn)')]) and re:test(., '^[^“]+?‘')]")]
-				if typos:
-					messages.append(LintMessage("y-021", "Possible typo: Opening [text]‘[/] without preceding [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for two quotations in one paragraph
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^”]+?”\\s“[^”]+?”$')]")]
-				if typos:
-					messages.append(LintMessage("y-022", "Possible typo: consecutive quotations without intervening text, e.g. [text]“…” “…”[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for incorrectly nested quotation marks
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘”]+“') and not(.//br)]")]
-				if typos:
-					messages.append(LintMessage("y-023", "Possible typo: two opening quotation marks in a run. Hint: Nested quotes should switch between [text]“[/] and [text]‘[/]", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for dashes instead of em-dashes
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\s[a-z]+-(the|there|is|and|they|when)\\s')]")]
-				if typos:
-					messages.append(LintMessage("y-024", "Possible typo: dash before [text]the/there/is/and/they/when[/] probably should be em-dash.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for comma not followed by space but followed by quotation mark
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '[a-z],[“”‘’][a-z]', 'i')]")]
-				if typos:
-					messages.append(LintMessage("y-025", "Possible typo: comma without space followed by quotation mark.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for punctuation missing before conjunctions. Ignore 
     with an  child starting in a conjunction, as those are probably book titles or non-English languages
-				typos = [node.to_string() for node in dom.xpath(f"/html/body//p[not(parent::hgroup) and re:test(., '\\b[a-z]+\\s(But|And|For|Nor|Yet|Or)\\b[^’\\.\\?\\-{se.WORD_JOINER}]') and not(./i[re:test(., '^(But|And|For|Nor|Yet|Or)\\b')])]")]
-				if typos:
-					messages.append(LintMessage("y-026", "Possible typo: no punctuation before conjunction [text]But/And/For/Nor/Yet/Or[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for extra closing single quote at the end of dialog
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘]+”\\s*’$')]")]
-				if typos:
-					messages.append(LintMessage("y-027", "Possible typo: Extra [text]’[/] at end of paragraph.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for `` preceded or followed by text. Ignore compass directions followed by `ly`, like S.S.W.ly
-				typos = [node.to_string() for node in dom.xpath("/html/body//abbr[(preceding-sibling::node()[1])[re:test(., '[A-Za-z]$')] or (following-sibling::node()[1])[re:test(., '^[A-Za-z](?[/] directly preceded or followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for misapplied italics. Ignore 's' because the plural is too common.
-				typos = [node.to_string() for node in dom.xpath("/html/body//*[(name() = 'i' or name() = 'em') and ./following-sibling::node()[1][re:test(., '^[a-z]\\b', 'i') and not(re:test(., '^s\\b'))]]")]
-				if typos:
-					messages.append(LintMessage("y-029", "Possible typo: Italics followed by a letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for lowercase letters starting quotations after a preceding period
-				typos = dom.xpath("/html/body//p/child::text()[re:test(., '\\.\\s[‘“][a-z]')]")
-				if typos:
-					messages.append(LintMessage("y-030", "Possible typo: Lowercase quotation following a period. Check either that the period should be a comma, or that the quotation should start with a capital.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for missing punctuation in continued quotations
-				# ” said Bob “
-				nodes = dom.xpath("/html/body//p[re:test(., '”\\s(?:said|[A-Za-z]{2,}ed)\\s[A-Za-z]+?(?Newspapers
-				nodes = dom.xpath("/html/body//i[@epub:type and ( (following-sibling::node()[1][re:test(., '^[a-z]', 'i') and not(re:test(., '^(s|es|er)'))]) or preceding-sibling::node()[1][re:test(., '[a-z]$')]) ]")
-				if nodes:
-					messages.append(LintMessage("y-032", "Possible typo: Italics running into preceding or following characters.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		if dom.xpath("/html/body/*[contains(@epub:type, 'loi') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='ill']"):
+			messages.append(LintMessage("m-034", "[val]loi[/] semantic inflection found, but no MARC relator [val]ill[/] (Illustrator).", se.MESSAGE_TYPE_WARNING, filename))
 
-				# Check for body element without child section or article. Ignore the ToC because it has a unique structure
-				nodes = dom.xpath("/html/body[not(./*[name()='section' or name()='article' or (name()='nav' and re:test(@epub:type, '\\b(toc|loi)\\b'))])]")
-				if nodes:
-					messages.append(LintMessage("s-069", "[xhtml][/] element missing direct child [xhtml]
    [/] or [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check that `Internet Archive` and `HathiTrust` are preceded by `the`. They might have an immediate preceding text node that ends in `the`, or they
+	# might be preceded by a white space node, then a 
    , then a text node ending in `the`.
+	nodes = dom.xpath("/html/body//a[(re:test(text(), '[Hh]athi') and re:test(@href, '^https://[^\"]*?hathitrust.org')) or (re:test(text(), 'Internet Archive') and re:test(@href, 'https://[^\"]*?archive.org'))][(preceding-sibling::node()[2][not(self::br)] and preceding-sibling::node()[1][not(re:test(., '\\sthe $'))]) or (preceding-sibling::node()[2][self::br] and preceding-sibling::node()[3][not(re:test(., '\\sthe$'))]) ]")
+	if nodes:
+		messages.append(LintMessage("m-061", "Link must be preceded by [text]the[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for header elements that are entirely non-English
-				nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./i[@xml:lang][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
-				if nodes:
-					messages.append(LintMessage("s-024", "Header elements that are entirely non-English should not be set in italics. Instead, the [xhtml][/] element has the [attr]xml:lang[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	return messages
 
-				# Check for half title pages missing subtitles
-				if ebook_has_subtitle:
-					# Make sure we exclude  because that appears in the ToC landmarks
-					nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and not(.//*[contains(@epub:type, 'subtitle')])]")
-					if nodes:
-						messages.append(LintMessage("s-087", "Subtitle in metadata, but no subtitle in the half title page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+def _lint_css_checks(filename: Path, dom: se.easy_xml.EasyXmlTree) -> list:
+	"""
+	Helper function used in self.lint()
+	Process CSS checks on an .xhtml file
 
-				else:
-					# Make sure we exclude  because that appears in the ToC landmarks
-					nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and .//*[contains(@epub:type, 'subtitle')]]")
-					if nodes:
-						messages.append(LintMessage("s-088", "Subtitle in half title page, but no subtitle in metadata.", se.MESSAGE_TYPE_ERROR, filename))
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
 
-				# Check that half titles without subtitles have the fulltitle semantic
-				nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage')]/*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'fulltitle'))]")
-				if nodes:
-					messages.append(LintMessage("s-096", "Heading element in half title page missing the [val]fulltitle[/] semantic.", se.MESSAGE_TYPE_ERROR, filename))
+	OUTPUTS
+	A list of LintMessages representing CSS errors found in the file
+	"""
 
-				# Check for header elements that have a label, but are missing the label semantic
-				# Find h# nodes whose first child is a text node matching a label type, and where that text node's next sibling is a semantic roman numeral
-				nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./node()[1][self::text() and not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)$') and following-sibling::*[1][contains(@epub:type, 'z3998:roman')]]]")
-				if nodes:
-					messages.append(LintMessage("s-066", "Header element missing [val]label[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	messages = [];
 
-				# Check for header elements that are missing both label and ordinal semantics
-				nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)\\s+[ixvIXVmcd]+$')]")
-				if nodes:
-					messages.append(LintMessage("s-073", "Header element that requires [val]label[/] and [val]ordinal[/] semantic children.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Do we have any elements that have specified border color?
+	# `transparent` and `none` are allowed values for border-color
+	if dom.xpath("/html/body//*[attribute::*[re:test(local-name(), 'data-css-border.+?-color') and text() != 'transparent' and text != 'none']]"):
+		messages.append(LintMessage("c-004", "Don’t specify border colors, so that reading systems can adjust for night mode.", se.MESSAGE_TYPE_WARNING, local_css_path, matches))
 
-				# Check for header elements that have a label semantic, but are missing an ordinal sibling
-				nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./span[contains(@epub:type, 'label')]][not(./span[contains(@epub:type, 'ordinal')])]")
-				if nodes:
-					messages.append(LintMessage("s-067", "Header element with a [val]label[/] semantic child, but without an [val]ordinal[/] semantic child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check that footers have the expected styling
+	# Footers may sometimes be aligned with text-align: center as well as text-align: right
+	# We also ignore text-align on any footers whose only child is a postscript, which is typically left-aligned
+	nodes = dom.xpath("/html/body//footer[not(@data-css-margin-top='1em') or (not(@data-css-text-align) and not(./*[contains(@epub:type, 'z3998:postscript') and not(following-sibling::*) and not(preceding-sibling::*) ]))]")
+	if nodes:
+		messages.append(LintMessage("c-010", "[xhtml]
    [/] missing [css]margin-top: 1em; text-align: ;[/]. [css]text-align[/] is usually set to [css]right[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for @lang attributes
-				nodes = dom.xpath("//*[@lang]")
-				if nodes:
-					messages.append(LintMessage("s-102", "[attr]lang[/] attribute detected. Hint: Use [attr]xml:lang[/] instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for elements that have `text-align: center` but also `text-indent: 1em`
+	nodes = dom.xpath("/html/body//*[@data-css-text-align='center' and @data-css-text-indent='1em']")
+	if nodes:
+		messages.append(LintMessage("c-011", "Element with [css]text-align: center;[/] but [css]text-indent[/] is [css]1em[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for sectioning elements with more than one heading element
-				nodes = dom.xpath("/html/body//*[name()='article' or name()='section'][count(./*[name()='header' or name()='hgroup' or re:test(name(), '^h[1-6]$')]) > 1]")
-				if nodes:
-					messages.append(LintMessage("s-071", "Sectioning element with more than one heading element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for sections that don't have header elements, and that do not have margin-top: 8em.
+	# This tries to find the first 
     or 
     whose first child is 
    , and that doesn't
+	# have any ancestor 
    s or s with heading content, and that doesn't have the correct top margin.
+	# Ignore the titlepage, loi, and dedications as they typically have unique styling.
+	nodes = dom.xpath("/html/body//*[name()='section' or name()='article'][1][./p[1][not(preceding-sibling::*)] and not(re:test(@epub:type, '(titlepage|loi|dedication)')) and not(ancestor::*/*[re:test(name(), '^(h[1-6]|header|hgroup)$')]) and not(@data-css-margin-top='20vh')]")
+	if nodes:
+		messages.append(LintMessage("c-012", "Sectioning element without heading content, and without [css]margin-top: 20vh;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-				# Check for some elements that only have a  child (including text inline to the parent).
-				# But, exclude such elements that are 
     elements that have poetry-type parents.
-				nodes = dom.xpath("/html/body//*[name() !='p' and not(ancestor-or-self::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][./span[count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
-				if nodes:
-					messages.append(LintMessage("s-072", "Element with single [xhtml][/] child. [xhtml][/] should be removed and its attributes promoted to the parent element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for padding and margin not in .5 increments, except for table headers/cells which usually need .25em increments
+	nodes = dom.xpath("/html/body//*[not(re:test(name(), '^(h[1-6]|td|th)$'))][attribute::*[re:test(local-name(), 'data-css-(margin|padding)')][re:test(., '^[0-9]*\\.[^5]')]]")
+	if nodes:
+		messages.append(LintMessage("c-013", "Element with margin or padding not in increments of [css].5em[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for s with only one child
-				nodes = dom.xpath("/html/body/*[name() = 'section' or name() = 'article']/header[./*[not(name() = 'p') and not(preceding-sibling::* or following-sibling::*)]]")
-				if nodes:
-					messages.append(LintMessage("s-098", "[xhtml]
    [/] element with only one child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for 
     element without a 
      and  are allowed, for use in tables with headers in the middle of tables (url:https://standardebooks.org/ebooks/dorothy-day/the-eleventh-virgin) and vertical table headers
-				# (https://standardebooks.org/ebooks/charles-babbage/passages-from-the-life-of-a-philosopher)
-				if dom.xpath("/html/body//table//th[not(ancestor::thead)][not(@scope)][not(count(node())=0)]"):
-					messages.append(LintMessage("s-055", "[xhtml][/] element not in [xhtml]
    [/] elements used as mid-table headings or horizontal row headings require the [attr]scope[/] attribute.", se.MESSAGE_TYPE_ERROR, filename))
+	messages = [];
 
-				# Check for money not separated by commas
-				nodes = dom.xpath("/html/body//*[re:test(text(), '[£\\$][0-9]{4,}')]")
-				if nodes:
-					messages.append(LintMessage("t-015", "Numbers not grouped by commas. Separate numbers greater than 1,000 with commas at every three numerals.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
-
-				# Do we have any elements that have specified border color?
-				# `transparent` and `none` are allowed values for border-color
-				if dom.xpath("/html/body//*[attribute::*[re:test(local-name(), 'data-css-border.+?-color') and text() != 'transparent' and text != 'none']]"):
-					messages.append(LintMessage("c-004", "Don’t specify border colors, so that reading systems can adjust for night mode.", se.MESSAGE_TYPE_WARNING, local_css_path, matches))
-
-				# Check for all-caps first paragraphs in sections/articles
-				# Note that we don't check for small-caps CSS, because there are lots of legitimate cases for that,
-				# and it would generate too many false positives.
-				# If we did want to do that, this xpath after the last re:test would help: or ./*[not(preceding-sibling::node()[normalize-space(.)]) and @data-css-font-variant='small-caps' and following-sibling::node()[1][self::text()]]
-				nodes = dom.xpath("/html/body//*[(name() = 'section' or name() = 'article') and not(contains(@epub:type, 'dedication') or contains(@epub:type, 'z3998:letter'))]/p[1][re:test(normalize-space(.), '^[“’]?(?:[A-Z’]+\\s){2,}')]")
-				if nodes:
-					messages.append(LintMessage("t-048", "Chapter opening text in all-caps.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for missing MARC relators
+	# Don't check the landmarks as that may introduce duplicate errors
+	# Only check for top-level elements, to avoid intros in short stories or poems in compilation.
+	if not dom.xpath("/html/body//nav[contains(@epub:type, 'landmarks')]"):
+		if dom.xpath("/html/body/*[contains(@epub:type, 'introduction') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and (text()='aui' or text()='win')]"):
+			messages.append(LintMessage("m-030", "[val]introduction[/] semantic inflection found, but no MARC relator [val]aui[/] (Author of introduction, but not the chief author) or [val]win[/] (Writer of introduction).", se.MESSAGE_TYPE_WARNING, filename))
 
-				# Check that footers have the expected styling
-				# Footers may sometimes be aligned with text-align: center as well as text-align: right
-				# We also ignore text-align on any footers whose only child is a postscript, which is typically left-aligned
-				nodes = dom.xpath("/html/body//footer[not(@data-css-margin-top='1em') or (not(@data-css-text-align) and not(./*[contains(@epub:type, 'z3998:postscript') and not(following-sibling::*) and not(preceding-sibling::*) ]))]")
-				if nodes:
-					messages.append(LintMessage("c-010", "[xhtml]
    [/] missing [css]margin-top: 1em; text-align: ;[/]. [css]text-align[/] is usually set to [css]right[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		if dom.xpath("/html/body/*[contains(@epub:type, 'preface') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='wpr']"):
+			messages.append(LintMessage("m-031", "[val]preface[/] semantic inflection found, but no MARC relator [val]wpr[/] (Writer of preface).", se.MESSAGE_TYPE_WARNING, filename))
 
-				# Check for poetry/verse without a descendent 
     element.
-				# Skip the ToC landmarks because it may have poem/verse semantic children.
-				nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')][not(descendant::p)][not(ancestor::nav[contains(@epub:type, 'landmarks')])]")
-				if nodes:
-					messages.append(LintMessage("s-044", "Element with poem or verse semantic, without descendant [xhtml]
    [/] (stanza) element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+		if dom.xpath("/html/body/*[contains(@epub:type, 'afterword') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='aft']"):
+			messages.append(LintMessage("m-032", "[val]afterword[/] semantic inflection found, but no MARC relator [val]aft[/] (Author of colophon, afterword, etc.).", se.MESSAGE_TYPE_WARNING, filename))
 
-				# Check for elements that have `text-align: center` but also `text-indent: 1em`
-				nodes = dom.xpath("/html/body//*[@data-css-text-align='center' and @data-css-text-indent='1em']")
-				if nodes:
-					messages.append(LintMessage("c-011", "Element with [css]text-align: center;[/] but [css]text-indent[/] is [css]1em[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-				# Check for basic typo, but exclude MathML and endnotes. Some endnotes (like in Ten Days that Shook the World) may start with letters.
-				dom_copy = deepcopy(dom)
-				for node in dom_copy.xpath("//a[contains(@epub:type, 'noteref')]") + dom_copy.xpath("//*[namespace-uri() = 'http://www.w3.org/1998/Math/MathML']"):
-					node.remove()
-
-				typos = [node.to_string() for node in dom_copy.xpath("/html/body//p[re:match(., '[a-z][;,][a-z]', 'i')]")]
-				if typos:
-					messages.append(LintMessage("y-002", "Possible typo: punctuation followed directly by a letter, without a space.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-                # Check for paragraphs ending in lowercase letters, excluding a number of instances where that might be valid.
-				typos = [node.to_string() for node in dom.xpath("//p[re:test(., '[a-z]$') and not(@epub:type) and not(following-sibling::*[1]/node()[1][contains(@epub:type, 'z3998:roman')] or following-sibling::*[1][re:test(., '^([0-9]|first|second|third|fourth|fifth|sixth|seventh|eight|ninth|tenth)', 'i')]) and not(@class or contains(@epub:type, 'z3998:salutation')) and not(following-sibling::*[1][name() = 'blockquote' or name() = 'figure' or name() = 'table' or name() = 'footer' or name() = 'ul' or name() = 'ol'] or ancestor::*[name() = 'blockquote' or name() = 'footer' or name() = 'header' or name() = 'table' or name() = 'ul' or name() = 'ol' or name() = 'li' or name() = 'figure' or name() = 'hgroup' or re:test(@epub:type, '(z3998:drama|z3998:letter|dedication|halftitlepage)')])]")]
-				if typos:
-					messages.append(LintMessage("y-003", "Possible typo: paragraph missing ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				typos = [node.to_string() for node in dom.xpath("//p[re:test(., '-[‘“]')]")]
-				if typos:
-					messages.append(LintMessage("y-004", "Possible typo: mis-curled quotation mark after dash.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for commas or periods following exclamation or question marks. Exclude the colophon because we may have painting names with punctuation.
-				if not is_colophon:
-					typos = [node.to_string() for node in dom.xpath("//p[re:test(., '[!?][\\.,][^”’]')]")]
-					if typos:
-						messages.append(LintMessage("y-005", "Possible typo: question mark or exclamation mark followed by period or comma.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for opening lsquo in quote that doesn't appear to have a matching rsquo. Rsquos must be followed by a space, punctuation (except period/comma), word joiner and thus em dash, or a number for an endnote.
-				typos = dom.xpath(f"re:match(//*, '“\\s*‘[^“]+?”', 'g')/text()[not(re:test(., '’[\\s\\?\\!;<2060{se.WORD_JOINER}0-9]'))]")
-				if typos:
-					messages.append(LintMessage("y-006", "Possible typo: [text]‘[/] without matching [text]’[/]. Hints: [text]’[/] are used for abbreviations;  commas and periods must go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Try to find top-level lsquo; for example, 
    “Bah!” he said to the ‘minister.’
    
-				# We can't do this on xpath because we can't iterate over the output of re:replace().
-				# A fully-featured solution would parse quotes to find balanced pairs. This quick-n-dirty
-				# solution takes the raw XHTML string, strips any seemingly-valid balanced pairs using regex,
-				# then searches for unmatched `‘`. If found, it returns the dom node of the match; we can't return
-				# the actual string match because we stripped surrounding text.
-
-				# Start by merging all 
     into single lines, including poetry, but exclude blockquotes as they have special quoting rules
-				temp_xhtml = ""
-				dom_copy = deepcopy(dom)
-				node_number = 0
-				# Add IDs to the dom so we can find the paragraphs after we run our regex
-				# Exclude paragraphs in blockquotes, which may have special quoting rules, and "continued" paragraphs, which may be continued dialog without an “
-				for node in dom_copy.xpath("/html/body//p[not(ancestor::blockquote) and not(contains(@class, 'continued'))]"):
-					node.set_attr("id", "lint-" + str(node_number))
-					temp_xhtml = temp_xhtml + f"
    " + regex.sub(r"[\s\n]+", " ", node.inner_text(), flags=regex.DOTALL) + "\n"
-					node_number = node_number + 1
-
-				replacement_count = 1
-				while replacement_count > 0:
-					# Remove all regular quotes. Run this in a loop because we may need to remove triple-nested quotes.
-					(temp_xhtml, replacement_count) = regex.subn(r"“[^“]+?”", " ", temp_xhtml) # Remove all regular quotes
-
-				# Remove contractions to reduce rsquo for next regex
-				temp_xhtml = regex.sub(r"[\p{Letter}]’[\p{Letter}]", " ", temp_xhtml, flags=regex.MULTILINE)
-
-				# Remove all runs of ldquo that are likely to spill to the next 
    
-				replacement_count = 1
-				while replacement_count > 0:
-					(temp_xhtml, replacement_count) = regex.subn(r"“[^“”]+?$", " ", temp_xhtml, flags=regex.MULTILINE)
-
-				# Match problem `‘` using regex, and if found, get the actual node text from the dom to return.
-				typos = []
-				for match in regex.findall(r"""
    .*‘[\p{Letter}\s]""", temp_xhtml):
-					for node in dom_copy.xpath(f"/html/body//p[@id='lint-{match}' and re:test(., '‘[A-Za-z\\s]')]"):
-						typos.append(node.inner_text())
-
-				if typos:
-					messages.append(LintMessage("y-007", "Possible typo: [text]‘[/] not within [text]“[/]. Hints: Should [text]‘[/] be replaced with [text]“[/]? Is there a missing closing quote? Is this a nested quote that should be preceded by [text]“[/]? Are quotes in close proximity correctly closed?", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for single quotes when there should be double quotes in an interjection in dialog
-				typos = [node.to_string() for node in dom.xpath("//p[re:test(., '“[^”]+?’⁠—[^”]+?—“')]")]
-				if typos:
-					messages.append(LintMessage("y-008", "Possible typo: dialog interrupted by interjection but with incorrect closing quote.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for dialog starting with a lowercase letter. Only check the first child text node of 
    , because other first children might be valid lowercase, like  or ;
-				# exclude 
     inside or preceded by 
    ; and exclude 
     inside endnotes, as definitions may start with lowercase letters.
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::li[contains(@epub:type, 'endnote')]) and not(preceding-sibling::*[1][name()='blockquote'])][re:test(./node()[1], '^“[a-z]')]")]
-				if typos:
-					messages.append(LintMessage("y-009", "Possible typo: dialog begins with lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				typos = [node.to_string() for node in dom.xpath("//p[not(ancestor::blockquote) and not(following-sibling::*[1][name() = 'blockquote' or contains(@class, 'continued')]) and re:test(., ',”$')]")]
-				if typos:
-					messages.append(LintMessage("y-010", "Possible typo: comma ending dialogue.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '’{2,}')]")]
-				if typos:
-					messages.append(LintMessage("y-011", "Possible typo: two or more [text]’[/] in a row.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '”[a-z]')]")]
-				if typos:
-					messages.append(LintMessage("y-012", "Possible typo: [text]”[/] directly followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for period not inside rsquo
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[^”’]+?’[\\.,]”')]")]
-				if typos:
-					messages.append(LintMessage("y-013", "Possible typo: punctuation not within [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for period before dialog tag; try to exclude abbrevations that close a quotation, like `“Mr.”`.
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[(re:test(., '\\.”\\s[a-z\\s]*?(\\bsaid|[a-z]+ed)') or re:test(., '\\.”\\s(s?he|they?|and)\\b')) and not(.//abbr[following-sibling::node()[re:test(., '^”')]])]")]
-				if typos:
-					messages.append(LintMessage("y-014", "Possible typo: Unexpected [text].[/] at the end of quotation. Hint: If a dialog tag follows, should this be [text],[/]?", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for mis-curled ‘ or ‘ without matching ’
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[A-Za-z][^“’]+?”')]")]
-				if typos:
-					messages.append(LintMessage("y-015", "Possible typo: mis-curled [text]‘[/] or missing [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for two periods in a row, almost always a typo for one period or a hellip
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\.\\.[^\\.]')]")]
-				if typos:
-					messages.append(LintMessage("y-016", "Possible typo: consecutive periods ([text]..[/]).", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for ldquo followed by space
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '“\\s+[^‘’]')]")]
-				if typos:
-					messages.append(LintMessage("y-017", "Possible typo: [text]“[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for lsquo followed by space
-				typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘\\s+[^“’]')]")]
-				if typos:
-					messages.append(LintMessage("y-018", "Possible typo: [text]‘[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-				# Check for closing rdquo without opening ldquo. We ignore blockquotes because they usually have unique quote formatting.
-				# Remove tds in case rdquo means "ditto mark"
-				typos = regex.findall(r"”[^“‘]+?”", regex.sub(r"
    		[”\s]+?(.+?)?
     without set margins. We ignore tables that are for drama structuring, and tables that are ancestors of 
     because 
     has its own margin.
+	# We also ignore tables whose preceding sibling sets a bottom margin that is not 0.
+	nodes = dom.xpath("/html/body//table[not(./ancestor-or-self::*[contains(@epub:type, 'z3998:drama')]) and not(./ancestor::blockquote) and not(./preceding-sibling::*[1][@data-css-margin-bottom != '0']) and (not(@data-css-margin-top) or not(@data-css-margin-right) or not(@data-css-margin-bottom) or not(@data-css-margin-left))]")
+	if nodes:
+		messages.append(LintMessage("c-014", "[xhtml]
    [/] element without explicit margins. Most tables need [css]margin: 1em;[/] or [css]margin: 1em auto 1em auto;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for a list of some common English loan words that used to be italicized as foreign, but are no longer
-				nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and re:test(., '^(fianc[eé]+|divorc[eé]+|menu|recherch[eé]|tour-de-force|outr[eé]|d[ée]but(ante)?|apropos|[eé]lite|prot[ée]g[ée]+|chef|salon|r[eé]gime|contretemps|[eé]clat|aides?|entr[ée]+|t[eê]te-[aà]-t[eê]tes?|blas[eé]|bourgeoisie)$')]")
-				if nodes:
-					messages.append(LintMessage("t-043", "Non-English loan word set in italics, when modern typography omits italics. Hint: A word may be correctly italicized when emphasis is desired, if the word is meant to be pronounced with an accent, or if the word is part of non-English speech.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for elements that follow salutations, but that don't have text-indent: 0;
+	# We have to check both 
     elems that are salutations, and also 
     elems that have a first-child inline element that is a salutation, and that does not have following siblings that are senders/valedictions, as that might indicate a letter in a prose context
+	nodes = dom.xpath("/html/body//*[(contains(@epub:type, 'z3998:salutation') or ./preceding-sibling::*[1][contains(@epub:type, 'z3998:salutation') or (name() != 'blockquote' and count(./node()[normalize-space(.)]) = 1 and ./*[contains(@epub:type, 'z3998:salutation')])] or ./*[1][contains(@epub:type, 'z3998:salutation') and not((./following-sibling::node()[1][self::text()]))]) and @data-css-text-indent != '0']")
+	if nodes:
+		messages.append(LintMessage("c-015", "Element after or containing [val]z3998:salutation[/] does not have [css]text-indent: 0;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for header elements with a roman semantic but without an ordinal semantic
-				nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][contains(@epub:type, 'z3998:roman') and not(contains(@epub:type, 'ordinal'))]")
-				if nodes:
-					messages.append(LintMessage("s-068", "Header element missing [val]ordinal[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for block-level postscripts that don't have margin-top: 1em. Exclude postscripts that are the first child of blockquote or footer, since blockquotes/footers gives the desired margin.
+	nodes = dom.xpath("/html/body//*[(name() = 'div' or name() = 'p') and contains(@epub:type, 'z3998:postscript') and not((./parent::blockquote or ./parent::footer) and count(./preceding-sibling::*) = 0) and @data-css-margin-top != '1em']")
+	if nodes:
+		messages.append(LintMessage("c-017", "Element with [val]z3998:postscript[/] semantic, but without [css]margin-top: 1em;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for deprecated MathML elements
-				# Note we dont select directly on element name, because we want to ignore any namespaces that may (or may not) be defined
-				nodes = dom.xpath("/html/body//*[local-name()='mfenced']")
-				if nodes:
-					messages.append(LintMessage("s-017", "[xhtml][/] is deprecated in the MathML spec. Use [xhtml](...)[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for 
     postscripts that don't have text-indent: 0.
+	nodes = dom.xpath("/html/body//p[contains(@epub:type, 'z3998:postscript') and (not(@data-css-text-align) or @data-css-text-align = 'initial' or @data-css-text-align = 'left') and @data-css-text-indent != '0']")
+	if nodes:
+		messages.append(LintMessage("c-018", "Element with [val]z3998:postscript[/] semantic, but without [css]text-indent: 0;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for period following Roman numeral, which is an old-timey style we must fix
-				# But ignore the numeral if it's the first item in a 
     tag, as that suggests it might be a kind of list item.
-				nodes = dom.xpath("/html/body//node()[name()='span' and contains(@epub:type, 'z3998:roman') and not(position()=1)][(following-sibling::node()[1])[re:test(., '^\\.\\s*[a-z]')]]")
-				if nodes:
-					messages.append(LintMessage("t-013", "Roman numeral followed by a period. When in mid-sentence Roman numerals must not be followed by a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "." for node in nodes]))
+	# Check for signature semantic without small caps
+	# Ignore signatures that are only capital letters
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:signature') and not(@data-css-font-variant = 'small-caps') and not(@data-css-font-style = 'italic') and not(re:test(., '^[A-Z\\.\\s]+$'))]")
+	if nodes:
+		messages.append(LintMessage("c-019", "Element with [val]z3998:signature[/] semantic, but without [css]font-variant: small-caps;[/] or [css]font-style: italic;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				# Check for common missing roman semantics for “I”
-				regent_regex = r"(?:Charles|Edward|George|Henry|James|William) I\b"
-				matches = regex.findall(fr"King {regent_regex}", file_contents) + regex.findall(fr"{regent_regex}’s", file_contents)
-				if matches:
-					messages.append(LintMessage("s-103", "Probable missing semantics for a roman I numeral.", se.MESSAGE_TYPE_WARNING, filename, matches))
-
-				# Check for  elements that have two or more letters/periods, that don't have a semantic epub:type
-				# SS. is the French abbreviation for "Saints"
-				nodes = dom.xpath("/html/body//abbr[not(@epub:type)][text() != 'U.S.' and text() != 'SS.'][re:test(., '([A-Z]\\.?){2,}')]")
-				filtered_nodes = []
-				for node in nodes:
-					add_node = True
-
-					if node.text in INITIALISM_EXCEPTIONS:
-						add_node = False
-					elif node.get_attr("class"):
-						for attr_value in node.get_attr("class").split():
-							if attr_value in IGNORED_CLASSES:
-								add_node = False
-								break
-					elif node.text in ("A.M.", "P.M.") and node.xpath("./preceding-sibling::node()[re:test(., '[0-9]\\s$')]"):
-						# Ignore instances of capitalized times, usually in titles. For example:
-						# `From 10:20 P.M. to 10:47 P.M.` in a chapter title.
-						# We have to test for a preceding number because `P.M.` in other contexts may mean `Prime Minister`
-						add_node = False
-
-					if add_node:
-						filtered_nodes.append(node)
-
-				if filtered_nodes:
-					messages.append(LintMessage("s-045", "[xhtml][/] element without semantic inflection like [class]z3998:personal-name[/] or [class]z3998:initialism[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in filtered_nodes]))
-
-				# Check for 
    s or 
    s that occur more than once in a file, without page break CSS
-				# Exclude the last element because we don't care if it breaks after the last one.
-				nodes = dom.xpath("/html/body[count(./article) > 1 or count(./section) > 1]/*[(name() = 'article' or name() = 'section') and (not(@data-css-break-after) or @data-css-break-after != 'page') and not(position() = last())]")
-				if nodes:
-					messages.append(LintMessage("c-020", "Multiple [xhtml]
    [/]s or [xhtml]
    [/]s in file, but missing [css]break-after: page;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+	# Check for 
    s or 
    s that occur more than once in a file, without page break CSS
+	# Exclude the last element because we don't care if it breaks after the last one.
+	nodes = dom.xpath("/html/body[count(./article) > 1 or count(./section) > 1]/*[(name() = 'article' or name() = 'section') and (not(@data-css-break-after) or @data-css-break-after != 'page') and not(position() = last())]")
+	if nodes:
+		messages.append(LintMessage("c-020", "Multiple [xhtml]
    [/]s or [xhtml]
    [/]s in file, but missing [css]break-after: page;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-				# Check for two em dashes in a row
-				nodes = dom.xpath(f"/html/body//*[re:test(text(), '—{se.WORD_JOINER}*—+')]")
-				if nodes:
-					messages.append(LintMessage("t-014", "Two or more em-dashes in a row found. Elided words should use the two- or three-em-dash Unicode character, and dialog ending in em-dashes should only end in a single em-dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for elements that are set in italics, and whose italic children don't have font-style: normal.
+	nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and ./*[(name()='i' or name()='em') and @data-css-font-style='italic']]")
+	if nodes:
+		messages.append(LintMessage("c-021", "Element with [css]font-style: italic;[/], but child [xhtml][/] or [xhtml][/] does not have [css]font-style: normal;[/]. Hint: Italics within italics are typically set in Roman for contrast; if that’s not the case here, can [xhtml][/] be removed while still preserving italics and semantic inflection?", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				nodes = dom.xpath("/html/body//blockquote//p[parent::*[name()='footer'] or parent::*[name()='blockquote']]//cite") # Sometimes the 
     may be in a 
    
-				if nodes:
-					messages.append(LintMessage("s-054", "[xhtml][/] as child of [xhtml]
    [/] in [xhtml]
    [/]. [xhtml][/] should be the direct child of [xhtml]
    [/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for tables that appear to be summing/listing numbers, but that are missing font-variant-numeric styling
+	nodes = dom.xpath("/html/body//table[count(.//td[not(following-sibling::*) and re:test(., '^[0-9]+$') and not(@data-css-font-variant-numeric='tabular-nums')]) >= 2]")
+	if nodes:
+		messages.append(LintMessage("c-026", "Table that appears to be listing numbers, but without [css]font-variant-numeric: tabular-nums;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-				# Find citations not offset by em dashes. We're only interested in s in endnotes that are directly followed by the backlink;
-				# ignore s starting with `Cf.` which means `Compare` and was often used to mean `see`
-				nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[last()]/cite[following-sibling::node()[normalize-space(.)][1][name() = 'a' and contains(@epub:type, 'backlink')] and re:test(., '^[^—]') and not(re:test(., '^(Cf\\.|\\()'))]")
-				if nodes:
-					messages.append(LintMessage("t-068", "Citation not offset with em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	return messages
 
-				# Check for 
     without set margins. We ignore tables that are for drama structuring, and tables that are ancestors of 
     because 
     has its own margin.
-				# We also ignore tables whose preceding sibling sets a bottom margin that is not 0.
-				nodes = dom.xpath("/html/body//table[not(./ancestor-or-self::*[contains(@epub:type, 'z3998:drama')]) and not(./ancestor::blockquote) and not(./preceding-sibling::*[1][@data-css-margin-bottom != '0']) and (not(@data-css-margin-top) or not(@data-css-margin-right) or not(@data-css-margin-bottom) or not(@data-css-margin-left))]")
-				if nodes:
-					messages.append(LintMessage("c-014", "[xhtml]
    [/] element without explicit margins. Most tables need [css]margin: 1em;[/] or [css]margin: 1em auto 1em auto;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-				# Check for a common typo
-				if "z3998:nonfiction" in file_contents:
-					messages.append(LintMessage("s-030", "[val]z3998:nonfiction[/] should be [val]z3998:non-fiction[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-				# Check for initialisms without periods
-				nodes = [node.to_string() for node in dom.xpath("/html/body//abbr[contains(@epub:type, 'z3998:initialism') and not(re:test(., '^[0-9]*([a-zA-Z]\\.)+[0-9]*$'))]") if node.text not in INITIALISM_EXCEPTIONS]
-				if nodes:
-					messages.append(LintMessage("t-030", "Initialism with spaces or without periods.", se.MESSAGE_TYPE_WARNING, filename, set(nodes)))
+def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = None) -> list:
+	"""
+	Check this ebook for some common SE style errors.
 
-				# Check for padding and margin not in .5 increments, except for table headers/cells which usually need .25em increments
-				nodes = dom.xpath("/html/body//*[not(re:test(name(), '^(h[1-6]|td|th)$'))][attribute::*[re:test(local-name(), 'data-css-(margin|padding)')][re:test(., '^[0-9]*\\.[^5]')]]")
-				if nodes:
-					messages.append(LintMessage("c-013", "Element with margin or padding not in increments of [css].5em[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	INPUTS
+	None
 
-				# Check for  that does not contain spaces
-				nodes = dom.xpath("/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and re:test(., '[A-Z]\\.[A-Z]\\.')]")
-				if nodes:
-					messages.append(LintMessage("t-016", "Initials in [xhtml][/] not separated by spaces.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	OUTPUTS
+	A list of LintMessage objects.
+	"""
 
-				# Check for z3998:stage-direction on elements that are not 
-				nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:stage-direction') and name() !='i' and name() !='abbr' and name() !='p']")
-				if nodes:
-					messages.append(LintMessage("s-058", "[attr]z3998:stage-direction[/] semantic only allowed on [xhtml][/], [xhtml][/], and [xhtml]
    [/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
-
-				# Check for  elements that are higher or lower than their expected level based on how deep they are in 
    s
-				# or 
    s. Exclude 
    [/], but without [val]continued[/] class.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		if not elements:
+			messages.append(LintMessage("m-049", "No [path]se-lint-ignore.xml[/] rules. Delete the file if there are no rules.", se.MESSAGE_TYPE_ERROR, lint_ignore_path))
 
-				# Check for incorrect children of 
-				nodes = dom.xpath("/html/body/*[name() != 'section' and name() != 'article' and name() != 'nav']")
-				if nodes:
-					messages.append(LintMessage("s-075", "[xhtml][/] element with direct child that is not [xhtml]
    [/], [xhtml]
    [/], or [xhtml]
     ancestor. However, [/] ancestor. Note: [xhtml]
     element without a 
      and  are allowed, for use in tables with headers in the middle of tables (url:https://standardebooks.org/ebooks/dorothy-day/the-eleventh-virgin) and vertical table headers
+	# (https://standardebooks.org/ebooks/charles-babbage/passages-from-the-life-of-a-philosopher)
+	if dom.xpath("/html/body//table//th[not(ancestor::thead)][not(@scope)][not(count(node())=0)]"):
+		messages.append(LintMessage("s-055", "[xhtml][/] element not in [xhtml]
    [/] elements used as mid-table headings or horizontal row headings require the [attr]scope[/] attribute.", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Check for z3998:stage-direction on elements that are not 
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:stage-direction') and name() !='i' and name() !='abbr' and name() !='p']")
 	if nodes:
-		messages.append(LintMessage("t-034", "[xhtml][/] element preceded by em-dash. Hint: em-dashes go within [xhtml][/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-058", "[attr]z3998:stage-direction[/] semantic only allowed on [xhtml][/], [xhtml][/], and [xhtml]
    [/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for  without preceding space in text node. (preceding ( or [ are also OK)
-	nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[not(re:test(., '[\\[\\(\\s]$'))]]")
+	# Check that internal links don't begin with ../
+	nodes = dom.xpath("/html/body//a[re:test(@href, '^\\.\\./text/')]")
 	if nodes:
-		messages.append(LintMessage("t-035", "[xhtml][/] element not preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-059", "Internal link beginning with [val]../text/[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for partially obscured years in which the last year is an em dash
-	nodes = dom.xpath("/html/body//p[re:test(.,  '1\\d{2}⁠—[^“’A-Za-z<]')]")
+	# Check for italics on things that shouldn't be italics
+	nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.music.song') or contains(@epub:type, 'se:name.publication.short-story') or contains(@epub:type, 'se:name.publication.essay')]")
 	if nodes:
-		messages.append(LintMessage("t-036", "Em-dash used to obscure single digit in year. Hint: Use a hyphen instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-060", "Italics on name that requires quotes instead.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for rdquo preceded by space (but not preceded by a rsquo, which might indicate a nested quotation)
-	nodes = dom.xpath("/html/body//p[re:test(., '[^’]\\s”')]")
+	# Check for h# tags followed by header content, that are not children of 
    .
+	# Only match if there is a following 
    , because we could have the case where there's an epigraph after a division title.
+	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][following-sibling::*[contains(@epub:type, 'epigraph') or contains(@epub:type, 'bridgehead')]][following-sibling::p][not(parent::header)]")
 	if nodes:
-		messages.append(LintMessage("t-037", "[text]”[/] preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-061", "Title and following header content not in a [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check obviously miscurled quotation marks
-	nodes = dom.xpath("/html/body//p[re:test(., '“$')]")
+	# Check for 
     elements without exactly one  child, but only in glossaries
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'glossary')]//dt[not(count(./dfn)=1)]")
 	if nodes:
-		messages.append(LintMessage("t-038", "[text]“[/] before closing [xhtml]
    [/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-062", "[xhtml]
    [/] element in a glossary without exactly one [xhtml][/] child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for some known initialisms with incorrect possessive apostrophes
-	nodes = dom.xpath("/html/body//abbr[text()='I.O.U.'][(following-sibling::node()[1])[starts-with(., '’s')]]")
+	# Check that `z3998:persona` is only on  or 
    		. We use the contact() xpath function so we don't catch `z3998:personal-name`
+	nodes = dom.xpath("/html/body//*[contains(concat(' ', @epub:type, ' '), ' z3998:persona ') and not(self::b or self::td)]")
 	if nodes:
-		messages.append(LintMessage("t-039", "Initialism followed by [text]’s[/]. Hint: Plurals of initialisms are not followed by [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "’s" for node in nodes]))
+		messages.append(LintMessage("s-063", "[val]z3998:persona[/] semantic on element that is not a [xhtml][/] or [xhtml][/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check if a subtitle ends in a text node with a terminal period; or if it ends in an  node containing a terminal period.
-	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6]/*[contains(@epub:type, 'subtitle')][(./text())[last()][re:test(., '\\.$')] or (./i)[last()][re:test(., '\\.$')]]")
+	# Check for fulltitle semantic on a header not in the half title
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'fulltitle') and name()!='h2' and name()!='hgroup' and not(ancestor::*[contains(@epub:type, 'halftitlepage')])]")
 	if nodes:
-		messages.append(LintMessage("t-040", "Subtitle with illegal ending period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-065", "[val]fulltitle[/] semantic on element that is not in the half title.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	matches = regex.findall(r"[^…]\s[!?;:,].{0,10}", file_contents) # If we don't include preceding chars, the regex is 6x faster
-	if matches:
-		messages.append(LintMessage("t-041", "Illegal space before punctuation.", se.MESSAGE_TYPE_ERROR, filename, matches))
+	# Check for header elements that have a label, but are missing the label semantic
+	# Find h# nodes whose first child is a text node matching a label type, and where that text node's next sibling is a semantic roman numeral
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./node()[1][self::text() and not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)$') and following-sibling::*[1][contains(@epub:type, 'z3998:roman')]]]")
+	if nodes:
+		messages.append(LintMessage("s-066", "Header element missing [val]label[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for a list of some common English loan words that used to be italicized as foreign, but are no longer
-	nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and re:test(., '^(fianc[eé]+|divorc[eé]+|menu|recherch[eé]|tour-de-force|outr[eé]|d[ée]but(ante)?|apropos|[eé]lite|prot[ée]g[ée]+|chef|salon|r[eé]gime|contretemps|[eé]clat|aides?|entr[ée]+|t[eê]te-[aà]-t[eê]tes?|blas[eé]|bourgeoisie)$')]")
+	# Check for header elements that have a label semantic, but are missing an ordinal sibling
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./span[contains(@epub:type, 'label')]][not(./span[contains(@epub:type, 'ordinal')])]")
 	if nodes:
-		messages.append(LintMessage("t-043", "Non-English loan word set in italics, when modern typography omits italics. Hint: A word may be correctly italicized when emphasis is desired, if the word is meant to be pronounced with an accent, or if the word is part of non-English speech.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-067", "Header element with a [val]label[/] semantic child, but without an [val]ordinal[/] semantic child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for comma after leading Or in subtitles
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'subtitle') and re:test(text(), '^Or\\s')]")
+	# Check for header elements with a roman semantic but without an ordinal semantic
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][contains(@epub:type, 'z3998:roman') and not(contains(@epub:type, 'ordinal'))]")
 	if nodes:
-		messages.append(LintMessage("t-044", "Comma required after leading [text]Or[/] in subtitles.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-068", "Header element missing [val]ordinal[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for personas that are italicized. Use re:test instead of contains to avoid catching z3998:personal-name
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:persona\\b') and @data-css-font-style = 'italic']")
+	# Check for body element without child section or article. Ignore the ToC because it has a unique structure
+	nodes = dom.xpath("/html/body[not(./*[name()='section' or name()='article' or (name()='nav' and re:test(@epub:type, '\\b(toc|loi)\\b'))])]")
 	if nodes:
-		messages.append(LintMessage("t-045", "Element has [val]z3998:persona[/] semantic and is also set in italics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-069", "[xhtml][/] element missing direct child [xhtml]
    [/] or [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename))
 
-	if "῾" in file_contents:
-		messages.append(LintMessage("t-046", "[text]῾[/] (U+1FFE) detected. Use [text]ʽ[/] (U+02BD) instead.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for h# without semantics; h# that have child elements (which likely have the correct semantics) are OK
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(@epub:type)][not(./*[not(name()='a' and contains(@epub:type, 'noteref'))])]")
+	if nodes:
+		messages.append(LintMessage("s-070", "[xhtml][/] element without semantic inflection.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for all-caps first paragraphs in sections/articles
-	# Note that we don't check for small-caps CSS, because there are lots of legitimate cases for that,
-	# and it would generate too many false positives.
-	# If we did want to do that, this xpath after the last re:test would help: or ./*[not(preceding-sibling::node()[normalize-space(.)]) and @data-css-font-variant='small-caps' and following-sibling::node()[1][self::text()]]
-	nodes = dom.xpath("/html/body//*[(name() = 'section' or name() = 'article') and not(contains(@epub:type, 'dedication') or contains(@epub:type, 'z3998:letter'))]/p[1][re:test(normalize-space(.), '^[“’]?(?:[A-Z’]+\\s){2,}')]")
+	# Check for sectioning elements with more than one heading element
+	nodes = dom.xpath("/html/body//*[name()='article' or name()='section'][count(./*[name()='header' or name()='hgroup' or re:test(name(), '^h[1-6]$')]) > 1]")
 	if nodes:
-		messages.append(LintMessage("t-048", "Chapter opening text in all-caps.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-071", "Sectioning element with more than one heading element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for two-em-dashes used for elision instead of three-em-dashes
-	matches = regex.findall(fr"[^{se.WORD_JOINER}\p{{Letter}}”]⸺[^“{se.WORD_JOINER}\p{{Letter}}].*", file_contents, flags=regex.MULTILINE)
-	if matches:
-		messages.append(LintMessage("t-049", "Two-em-dash used for eliding an entire word. Use a three-em-dash instead.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Check for some elements that only have a  child (including text inline to the parent).
+	# But, exclude such elements that are 
     elements that have poetry-type parents.
+	nodes = dom.xpath("/html/body//*[name() !='p' and not(ancestor-or-self::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][./span[count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	if nodes:
+		messages.append(LintMessage("s-072", "Element with single [xhtml][/] child. [xhtml][/] should be removed and its attributes promoted to the parent element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check that possessives appear within persona blocks
-	nodes = dom.xpath("/html/body//b[contains(@epub:type, 'z3998:persona') and ./following-sibling::node()[1][re:test(., '^’s?')]]")
+	# Check for header elements that are missing both label and ordinal semantics
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)\\s+[ixvIXVmcd]+$')]")
 	if nodes:
-		messages.append(LintMessage("t-050", "Possessive [text]’s[/] or [text]’[/] outside of element with [val]z3998:persona[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-073", "Header element that requires [val]label[/] and [val]ordinal[/] semantic children.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for quotations that carry to the next paragraph, but the next paragraph has no opening quotation mark
-	# Exclude p in blockquote because often that has different formatting
-	nodes = dom.xpath("/html/body//p[re:test(., '^“[^”]+$') and not(./ancestor::blockquote) and ./following-sibling::*[1][name() = 'p' and re:test(normalize-space(), '^[^“]')]]")
+	# Check for table headers that don't have content, which is an accessibility issue.
+	# Note that @aria-label and @title should (apparently) not be used for table headers
+	nodes = dom.xpath("/html/body//th[not(text())]")
 	if nodes:
-		messages.append(LintMessage("t-051", "Dialog in [xhtml]
    [/] that continues to the next [xhtml]
    [/], but the next [xhtml]
    [/] does not begin with [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-074", "[xhtml]
    		[/] element with no text content should be a [xhtml][/] element instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for stage direction without ending punctuation. We only want to consider stage direction that is not an interjection in a parent clause.
-	# We match if the stage direction ends in lowercase, and if there's no following node, or if there's a following node that doesn't begin in lowercase or an em dash or semicolon (which suggests an interjection).
-	nodes = dom.xpath(f"/html/body//i[@epub:type='z3998:stage-direction' and re:test(., '[a-z]$') and re:test(., '^[A-Z]') and (not(./following-sibling::node()[normalize-space(.)]) or ./following-sibling::node()[1][re:test(normalize-space(.), '^[^{se.WORD_JOINER}—;a-z]')])]")
+	# Check for incorrect children of 
+	nodes = dom.xpath("/html/body/*[name() != 'section' and name() != 'article' and name() != 'nav']")
 	if nodes:
-		messages.append(LintMessage("t-052", "Stage direction without ending punctuation. Note that ending punctuation is optional in stage direction that is an interjection in a parent clause, and such interjections should begin with a lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-075", "[xhtml][/] element with direct child that is not [xhtml]
    [/], [xhtml]
    [/], or [xhtml]
    		 that contains both inline text and block-level elements.
+	# Note that we check any child of  because often  has the drama semantic.
+	nodes = dom.xpath("/html//*[contains(@epub:type, 'z3998:drama')]//td[./text()[normalize-space(.)] and ./*[name() = 'p' or name() = 'blockquote' or name() = 'div']]")
+	if nodes:
+		messages.append(LintMessage("s-080", "[xhtml][/] in drama containing both inline text and a block-level element. All children should either be only text, or only block-level elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
+	# Check for 
     tags that follow figure, blockquote, or table, and start with a lowercase letter, but that don't have the `continued` class. Exclude matches whose first child is , so we don't match things like 
    —Editor
+	nodes = dom.xpath("/html/body//p[preceding-sibling::*[1][name() = 'figure' or name() = 'blockquote' or name() = 'table'] and re:test(., '^([a-z]|—|…)') and not(contains(@class, 'continued')) and not(./*[1][self::cite])]")
 	if nodes:
-		messages.append(LintMessage("t-057", "[xhtml]
    [/] starting with lowercase letter. Hint: [xhtml]
    [/] that continues text after a [xhtml]
    [/] requires the [class]continued[/] class; and use [xhtml]
    [/] to split one clause over many lines.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-081", "[xhtml]
    [/] preceded by [xhtml]
    [/], [xhtml]
    [/xhtml], or [xhtml][/], but without [val]continued[/] class.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for illegal characters
-	matches = regex.findall(fr"({se.UNICODE_BOM}|{se.SHY_HYPHEN})", file_contents)
-	if matches:
-		# Get the keys of a dict in order to create a list without duplicates
-		messages.append(LintMessage("t-058", "Illegal character.", se.MESSAGE_TYPE_ERROR, filename, list({match.encode("unicode_escape").decode().replace("\\u", "U+").upper():None for match in matches}.keys())))
+	# Check for language tags transliterated into Latin script but missing `-Latn` suffix
+	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '^(he|ru|el|zh|bn|hi|sa|uk|yi|grc)$') and re:test(., '[a-zA-Z]')]")
+	if nodes:
+		messages.append(LintMessage("s-082", "Element containing Latin script for a non-Latin-script language, but its [attr]xml:lang[/] attribute value is missing the [val]-Latn[/] language tag suffix. Hint: For example Russian transliterated into Latin script would be [val]ru-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for period in cite in endnotes
-	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]//cite[not((./node()[last()])[name() = 'abbr']) and ./following-sibling::*[1][contains(@epub:type, 'backlink')] and re:test(., '^—') and ( (re:test(., '\\.$') and ./following-sibling::node()[re:test(., '^\\s*$')]) or ./following-sibling::node()[re:test(., '^\\.\\s*$')])]")
+	# Check for persona " not in match]
+	# xpath to check for opening quote in p, without a next child p that starts with an opening quote or an opening bracket (for editorial insertions within paragraphs of quotation); or that consists of only an ellipses (like an elided part of a longer quotation)
+	# Matching 
    s can't have a poem/verse ancestor as formatting is often special for those.
+	matches = matches + [regex.findall(r"“[^”]+
    ", node.to_string())[0] for node in dom.xpath("/html/body//p[re:test(., '“[^‘”]+$')][not(ancestor::*[re:test(@epub:type, 'z3998:(verse|poem|song|hymn|lyrics)')])][(following-sibling::*[1])[name()='p'][not(re:test(normalize-space(.), '^[“\\[]') or re:test(normalize-space(.), '^…$'))]]")]
 
-	# Check to see if we've marked something as poetry or verse, but didn't include a first 
-	# This xpath selects the p elements, whose parents are poem/verse, and whose first child is not a span
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')]/p[not(./*[name()='span' and position()=1])]")
-	if nodes:
-		matches = []
-		for node in nodes:
-			# Get the first line of the poem, if it's a text node, so that we can include it in the error messages.
-			# If it's not a text node then just ignore it and add the error anyway.
-			first_line = node.xpath("./descendant-or-self::text()[normalize-space(.)]", True)
-			if first_line:
-				match = first_line.strip()
-				if match: # Make sure we don't append an empty string
-					matches.append(match)
+	# Additionally, match short 
     tags (< 100 chars) that lack closing quote, and whose direct siblings do have closing quotes (to exclude runs of same-speaker dialog), and that is not within a blockquote, verse, or letter
+	matches = matches + [regex.findall(r"“[^”]+
    ", node.to_string())[0] for node in dom.xpath("/html/body//p[re:test(., '“[^‘”]+$') and not(re:test(., '[…:]$')) and string-length(normalize-space(.)) <=100][(following-sibling::*[1])[not(re:test(., '“[^”]+$'))] and (preceding-sibling::*[1])[not(re:test(., '“[^”]+$'))]][not(ancestor::*[re:test(@epub:type, 'z3998:(verse|poem|song|hymn|lyrics)')]) and not(ancestor::blockquote) and not (ancestor::*[contains(@epub:type, 'z3998:letter')])][(following-sibling::*[1])[name()='p'][re:test(normalize-space(.), '^[“\\[]') and not(contains(., 'continued'))]]")]
+	if matches:
+		messages.append(LintMessage("t-003", "[text]“[/] missing matching [text]”[/]. Note: When dialog from the same speaker spans multiple [xhtml]
    [/] elements, it’s correct grammar to omit closing [text]”[/] until the last [xhtml]
    [/] of dialog.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-		messages.append(LintMessage("s-006", "Poem or verse [xhtml]
    [/] (stanza) without [xhtml][/] (line) element.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Check for closing dialog without comma
+	matches = regex.findall(r"[\p{Lowercase_Letter}]+?” [\p{Letter}]+? said", file_contents)
+	if matches:
+		messages.append(LintMessage("t-005", "Dialog without ending comma.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	# Check for elements that don't have a direct block child
-	# Allow white space and comments before the first child
-	# Ignore children of the ToC and landmarks as they do not require p children
-	nodes = dom.xpath("/html/body//*[(name()='blockquote' or name()='dd' or name()='header' or name()='li' or name()='footer') and not(./ancestor::*[contains(@epub:type, 'toc') or contains(@epub:type, 'landmarks')]) and (node()[normalize-space(.) and not(self::comment())])[1][not(name()='p' or name()='blockquote' or name()='div' or name()='table' or name()='header' or name()='ul' or name()='ol' or name()='footer' or name()='hgroup' or re:test(name(), '^h[0-6]'))]]")
-	if nodes:
-		messages.append(LintMessage("s-007", "Element requires at least one block-level child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for possessive 's within name italics, but not in ignored files like the colophon which we know have no possessives
+	# Allow some known exceptions like `Harper's`, etc.
+	if filename.name not in IGNORED_FILENAMES:
+		nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.') and re:test(., '’s$') and not(contains(@epub:type, 'se:name.publication.') and re:test(., '(Pearson’s|Harper’s|Blackwood’s|Fraser’s|Baedeker’s)$'))]")
+		if nodes:
+			messages.append(LintMessage("t-007", "Possessive [text]’s[/] within name italics. If the name in italics is doing the possessing, [text]’s[/] goes outside italics.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for block-level tags that end with 
    
-	nodes = dom.xpath("/html/body//*[self::p or self::blockquote or self::table or self::ol or self::ul or self::section or self::article][br[last()][not(following-sibling::text()[normalize-space()])][not(following-sibling::*)]]")
+	# Check for lsquo not correctly closed
+	matches = regex.findall(r"‘[^“’]+?‘", file_contents)
+	matches = [match for match in matches if "" not in match]
+	if matches:
+		messages.append(LintMessage("t-004", "[text]‘[/] missing matching [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, matches))
+
+	# Check for repeated punctuation, but first remove `&` so we don't match `&,`
+	# Remove tds with repeated ” as they are probably ditto marks
+	matches = regex.findall(r"[,;]{2,}.{0,20}", file_contents.replace("&", "")) + regex.findall(r"(?:“\s*“|”\s*”|’ ’|‘\s*‘).{0,20}", regex.sub(r"
    ", "", file_contents)) +  regex.findall(r"[\p{Letter}][,\.:;]\s[,\.:;]\s?[\p{Letter}<].{0,20}", file_contents, flags=regex.IGNORECASE)
+	if matches:
+		messages.append(LintMessage("t-008", "Repeated punctuation.", se.MESSAGE_TYPE_WARNING, filename, matches))
+
+	# Check for nbsp before times
+	nodes = dom.xpath(f"/html/body//text()[re:test(., '[0-9][^{se.NO_BREAK_SPACE}]?$')][(following-sibling::abbr[1])[re:test(., '^[ap]\\.m\\.$')]]")
 	if nodes:
-		messages.append(LintMessage("s-008", "[xhtml]
    [/] element found before closing tag of block-level element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-009", "Required no-break space not found before time and [text]a.m.[/] or [text]p.m.[/].", se.MESSAGE_TYPE_WARNING, filename, [node[-10:] + " first because they are likely to contain numbered section references
+	dom_copy = deepcopy(dom)
+	for node in dom_copy.xpath("/html/body//p/a"):
+		node.remove()
+
+	nodes = dom_copy.xpath("/html/body//p[re:test(., '[^=]\\s[0-9]{1,2}\\.[0-9]{2}(?![0-9′″°%]|\\.[0-9]|\\scubic|\\smetric|\\smeters|\\smiles|\\sfeet|\\sinches)')]")
+	matches = []
+	for node in nodes:
+		for time_match in regex.findall(r"(?<=[^=]\s)[0-9]{1,2}\.[0-9]{2}(?![0-9′″°%]|\.[0-9]|\scubic|\smetric|\smeters|\smiles|\sfeet|\sinches)", node.inner_text()):
+			time = time_match.split(".")
+			if not time[0].startswith("0") and int(time[0]) >= 1 and int(time[0]) <= 12 and int(time[1]) >= 0 and int(time[1]) <= 59:
+				matches.append(time_match)
+
+	if matches:
+		messages.append(LintMessage("t-010", "Time set with [text].[/] instead of [text]:[/].", se.MESSAGE_TYPE_WARNING, filename, set(matches)))
 
-	# Check for hgroup elements with only one child
-	nodes = dom.xpath("/html/body//hgroup[count(*)=1]")
+	# Check for missing punctuation before closing quotes
+	# Exclude signatures in footers as those are commonly quoted without ending punctuation
+	nodes = dom.xpath("/html/body//p[not( (parent::header or parent::hgroup or (parent::footer and contains(@epub:type, 'z3998:signature'))) and position()=last())][re:test(., '[a-z]+[”’]$')]")
 	if nodes:
-		messages.append(LintMessage("s-009", "[xhtml]
    [/] element with only one child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-011", "Missing punctuation before closing quotes.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string()[-30:] for node in nodes]))
 
-	# Check for empty elements. Elements are empty if they have no children and no non-whitespace text
-	# Use name() to check for MathML elements, because if the MathML namespace is not declared in a given file and we search by the m: namespace then this xpath will fail entirely.
-	nodes = dom.xpath("/html/body//*[not(self::br) and not(self::hr) and not(self::img) and not(self::td) and not(self::th) and not(self::link) and not(self::col) and not(self::colgroup) and not(local-name()='none') and not(local-name()='mspace') and not(local-name()='mprescripts')][not(./*)][not(normalize-space())][not(./ancestor::*[local-name()='annotation-xml'])]")
+	# Check for whitespace before noteref
+	# Do this early because we remove noterefs from headers later
+	# Allow noterefs that do not have preceding text/elements (for example a "see note X" ref that is the only child in a 
    s that have child 
     elements
+	nodes = dom.xpath("/html/body//td[contains(@epub:type, 'z3998:persona') and ./p]")
 	if nodes:
-		messages.append(LintMessage("t-059", "Period at the end of [xhtml][/] element before endnote backlink.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-083", "[xhtml]
    		[/] element with child [xhtml]
    [/] element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for Bible verses in old-style notation
-	nodes = dom.xpath("/html/body//*[(name()='p' or name()='cite' or re:test(name(), '^h[1-6]$')) and .//node()[re:test(., '(Genesis|Gen\\.|Exodus|Ex\\.|Leviticus|Lev\\.|Numbers|Num\\.|Deuteronomy|Deut\\.|Joshua|Josh\\.|Judges|Ruth|Kings|Chronicles|Chron\\.|Ezra|Nehemiah|Neh\\.|Esther|Esth\\.|Job|Psalm|Psalms|Ps\\.|Proverbs|Prov\\.|Ecclesiastes|Ecc\\.|Eccl\\.|Solomon|Sol\\.|Isaiah|Is\\.|Isa\\.|Jeremiah|Jer\\.|Lamentations|Lam\\.|Ezekiel|Ez\\.|Ezek\\.|Daniel|Dan\\.|Hosea|Hos\\.|Joel|Amos|Obadiah|Obad\\.|Jonah|Jon\\.|Micah|Mic\\.|Nahum|Nah\\.|Habakkuk|Hab\\.|Zephaniah|Zeph\\.|Haggai|Hag\\.|Zechariah|Zech\\.|Malachi|Mal\\.|Tobit|Judith|Sirach|Baruch|Maccabees|Esdras|Manasses|Matthew|Matt\\.|Mark|Luke|John|Acts|Romans|Rom\\.|Corinthians|Cor\\.|Corinth\\.|Galatians|Gal\\.|Ephesians|Eph\\.|Philippians|Phil\\.|Philipp\\.|Colossians|Col\\.|Coloss\\.|Thessalonians|Thes\\.|Thess\\.|Timothy|Tim\\.|Titus|Tit\\.|Philemon|Phil\\.|Hebrews|Heb\\.|James|Jas\\.|Peter|Pet\\.|Jude|Revelation|Revelations|Rev\\.)\\s*$') and following-sibling::node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman') and following-sibling::node()[1][re:test(., '^\\s*[\\.,]?\\s+[0-9]')]]]]")
+	# Check for incorrect semantics on some common publications
+	nodes = dom.xpath("/html/body//i[@epub:type='se:name.publication.book' and re:test(., '^(The )?(Iliad|Odyssey|Aeneid|Metamorphoses|Beowulf|Divine Comedy|Paradise Lost)$')]")
 	if nodes:
-		messages.append(LintMessage("t-060", "Old style Bible citation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-084", "Poem has incorrect semantics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for summary-style bridgeheads not ending in punctuation
-	# Decide if it's a summary-style bridgehead if it contains two or more em dashes. We use Python regex because
-	# xpath can't count the number of occurances of a string.
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'bridgehead') and re:test(., '[^\\.\\!\\?”]”?$')]")
-	nodes = [node for node in nodes if len(regex.findall(r"—", node.to_string())) >= 2]
+	# Check for `Op. Cit.` in endnotes. `Op. Cit.` means "the previous reference" which doesn't make sense
+	# in popup endnotes. But, if the endnote has a book reference, then allow it as it might be referring to that.
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote') and (re:test(., '(Loc\\.|Op\\.) Cit\\.', 'i') or re:test(., '\\bl\\.c\\.\\b', 'i')) and not(.//*[contains(@epub:type, 'se:name.publication')])]")
 	if nodes:
-		messages.append(LintMessage("t-061", "Summary-style bridgehead without ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-086", "[text]Op. Cit.[/] or [text]Loc. Cit.[/] in endnote. Hint: [text]Op. Cit.[/] and [text]Loc. Cit.[/] mean [text]the previous reference[/], which usually doesn’t make sense in a popup endnote. Such references should be expanded.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for uppercased am/pm, except for if they have an epub:type, or when uppercased in titles, or in the nav document
-	if not dom.xpath("//nav[contains(@epub:type, 'toc')]"):
-		nodes = dom.xpath("/html/body//abbr[re:test(., '^[AP]\\.M\\.$') and not(@epub:type) and not(./ancestor::*[contains(@epub:type, 'title')])]")
+	# Check for half title pages missing subtitles
+	if ebook_info["has_subtitle"]:
+		# Make sure we exclude  because that appears in the ToC landmarks
+		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and not(.//*[contains(@epub:type, 'subtitle')])]")
 		if nodes:
-			messages.append(LintMessage("t-062", "Uppercased [text]a.m.[/] and [text]p.m.[/]", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+			messages.append(LintMessage("s-087", "Subtitle in metadata, but no subtitle in the half title page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	# Check for some Latinisms that need italics according to the manual. We check ancestor-or-self in case the phrase is set in Roman because it's nested in a parent italic.
-	# Exclude toto followed by ’ since Toto can be a name.
-	# Exclude  whose entire contents is a matched Latinism as we do not italicize those.
-	# Ignore the ToC because we have different rules there
-	nodes = dom.xpath("/html/body//text()[re:test(., '\\b(a (priori|posteriori|fortiori)|(? because that appears in the ToC landmarks
+		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and .//*[contains(@epub:type, 'subtitle')]]")
+		if nodes:
+			messages.append(LintMessage("s-088", "Subtitle in half title page, but no subtitle in metadata.", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Check for MathML without alttext
+	nodes = dom.xpath("/html/body//m:math[not(@alttext)]")
 	if nodes:
-		messages.append(LintMessage("t-063", "Non-English confusable phrase set without italics.", se.MESSAGE_TYPE_WARNING, filename, nodes))
+		messages.append(LintMessage("s-089", "MathML missing [attr]alttext[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check that all names are correctly titlecased. Ignore titles with xml:lang since non-English languages have different titlecasing rules,
-	# and ignore titles that have children elements, because se.titlecase() can't handle XML-like titles right now.
-	# Ignore titles longer than 150 chars, as long titles are likely old-timey super-long titles that should be mostly sentence-cased
-	incorrectly_cased_titles = []
-	for node in dom.xpath("/html/body//*[contains(@epub:type, 'se:name') and not(contains(@epub:type, 'se:name.legal-case')) and not(@xml:lang) and not(./*) and string-length(.) <= 150]"):
-		# Replace any space that is not a hair space with a regular space. This is because in inline titles, we may correctly
-		# have nbsp for example after `St.`, but titlecase will remove that nbsp.
-		if se.formatting.titlecase(node.inner_text()) != regex.sub(fr"[^\S{se.HAIR_SPACE}]+", " ", node.inner_text()):
-			incorrectly_cased_titles.append(node.to_string())
+	# Check for common errors in language tags
+	# `gr` is often used instead of `el`, `sp` instead of `es`, and `ge` instead of `de` (`ge` is the Georgian geographic region subtag but not a language subtag itself)
+	nodes = dom.xpath("//*[re:test(@xml:lang, '^(gr|sp|ge)$')]")
+	if nodes:
+		messages.append(LintMessage("s-090", "Invalid language tag.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	if incorrectly_cased_titles:
-		messages.append(LintMessage("t-064", "Title not correctly titlecased. Hint: Non-English titles should have an [attr]xml:lang[/] attribute as they have different titlecasing rules.", se.MESSAGE_TYPE_WARNING, filename, incorrectly_cased_titles))
+	# Check for  in poetry not followed by 
    . Ignore spans that are roman as they might be present in poem headers, or spans with xml:lang.
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]//p/span[not(@epub:type='z3998:roman') and not(@xml:lang) and following-sibling::*[1][name()='span']]")
+	if nodes:
+		messages.append(LintMessage("s-091", "[xhtml][/] not followed by [xhtml]
    [/] in poetry.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for headers ending in periods
-	# Allow periods at the end of headers that both start and end with double quotes; but headers that only end in double quotes probably don't need periods
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'z3998:initialism')) and not((./node()[last()])[name() = 'abbr']) and (re:test(., '\\.$') or re:test(., '^[^“].+\\.”$'))]")
+	# Check for nested `` elements
+	nodes = dom.xpath("/html/body//abbr[./abbr]")
 	if nodes:
-		messages.append(LintMessage("t-065", "Header ending in a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-093", "Nested [xhtml][/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for royal names whose roman numeral is preceded by `the`
-	nodes = dom.xpath("/html/body//span[contains(@epub:type, 'z3998:roman') and ./preceding-sibling::node()[1][re:test(., '[A-Z]\\w+ the $')]]/parent::*")
+	# Check for language tags transliterated into Latin script with incorrect `-latn` suffix (lowercase l)
+	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '-latn')]")
 	if nodes:
-		messages.append(LintMessage("t-066", "Regnal ordinal preceded by [text]the[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-094", "Element has an [attr]xml:lang[/] attribute that incorrectly contains [val]-latn[/] instead of [val]-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for plural graphemes without apostrophes
-	nodes = dom.xpath("/html/body//i[re:test(@epub:type, 'z3998:(grapheme|phoneme|morpheme)') and ./following-sibling::node()[1][starts-with(., 's')]]")
+	# Check for poetry/verse that has an hgroup but which does not have the correct text alignment
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(hymn|poem|song|verse)')]//hgroup/p[@data-css-text-align != 'center']")
 	if nodes:
-		messages.append(LintMessage("t-067", "Plural [val]z3998:grapheme[/], [val]z3998:phoneme[/], or [val]z3998:morpheme[/] formed without apostrophe ([text]’[/]).", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-095", "[xhtml]
    [/] child of [xhtml]
    [/] in poetry/verse does not have [css]text-align: center;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Find citations not offset by em dashes. We're only interested in s in endnotes that are directly followed by the backlink;
-	# ignore s starting with `Cf.` which means `Compare` and was often used to mean `see`
-	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[last()]/cite[following-sibling::node()[normalize-space(.)][1][name() = 'a' and contains(@epub:type, 'backlink')] and re:test(., '^[^—]') and not(re:test(., '^(Cf\\.|\\()'))]")
+	# Check that half titles without subtitles have the fulltitle semantic
+	nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage')]/*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'fulltitle'))]")
 	if nodes:
-		messages.append(LintMessage("t-068", "Citation not offset with em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-096", "Heading element in half title page missing the [val]fulltitle[/] semantic.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for  in 
    s that start with a leading em dash
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '^—')]")
+	# Check for missing href attributes, sometimes a leftover from PG transcriptions
+	nodes = dom.xpath("/html/body//a[not(@href)]")
 	if nodes:
-		messages.append(LintMessage("t-069", "[xhtml][/] in epigraph starting with an em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-097", "[xhtml]a[/] element without [attr]href[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for  in 
    s that end with a period. Try to exclude s that are last children
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '\\.$') and not( (./node()[last()])[./descendant-or-self::abbr]) ]")
+	# Check for s with only one child
+	nodes = dom.xpath("/html/body/*[name() = 'section' or name() = 'article']/header[./*[not(name() = 'p') and not(preceding-sibling::* or following-sibling::*)]]")
 	if nodes:
-		messages.append(LintMessage("t-070", "[xhtml][/] in epigraph ending in a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-098", "[xhtml]
    [/] element with only one child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# If we have Greek text, try to normalize it
-	for node in dom.xpath("/html/body//*[@xml:lang='grc' or @xml:lang='el']"):
-		node_text = node.inner_text()
-		expected_text = se.typography.normalize_greek(node_text)
-		if node_text != expected_text:
-			messages.append(LintMessage("t-073", f"Possible transcription error in Greek. Found: [text]{node_text}[/], but expected [text]{expected_text}[/text]. Hint: Use [bash]se unicode-names[/] to see differences in Unicode characters.", se.MESSAGE_TYPE_WARNING, filename))
+	# Check for @lang attributes
+	nodes = dom.xpath("//*[@lang]")
+	if nodes:
+		messages.append(LintMessage("s-102", "[attr]lang[/] attribute detected. Hint: Use [attr]xml:lang[/] instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+
+	# Check for common missing roman semantics for “I”
+	regent_regex = r"(?:Charles|Edward|George|Henry|James|William) I\b"
+	matches = regex.findall(fr"King {regent_regex}", file_contents) + regex.findall(fr"{regent_regex}’s", file_contents)
+	if matches:
+		messages.append(LintMessage("s-103", "Probable missing semantics for a roman I numeral.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
 	return messages
 
-def _lint_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_info: dict, language: str) -> list:
+def _lint_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str) -> list:
 	"""
 	Helper function used in self.lint()
-	Process syntax checks on an .xhtml file
+	Process typography checks on an .xhtml file
 
 	INPUTS
 	filename: The name of the file being checked
 	dom: A dom tree to check
 	file_contents: The contents of the file being checked
-	ebook_info: A dictionary containing several pieces of information about an ebook
+	special_file: A string containing the type of special file, if any
 
 	OUTPUTS
-	A list of LintMessages representing syntax errors found in the file
+	A list of LintMessages representing typography errors found in the file
 	"""
 
-	messages = [];
+	messages = []
 
-	# Check for numeric entities
-	matches = regex.findall(r"&#[0-9]+?;", file_contents)
+	# Check for punctuation outside quotes. We don't check single quotes because contractions are too common.
+	matches = regex.findall(fr"[\p{{Letter}}]+”[,\.](?!{se.WORD_JOINER} {se.WORD_JOINER}…)", file_contents)
 	if matches:
-		messages.append(LintMessage("s-001", "Illegal numeric entity (like [xhtml]Α[/]).", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("t-002", "Comma or period outside of double quote. Generally punctuation goes within single and double quotes.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	# Check nested 
     elements, but only if it's the first child of another 
    
-	nodes = dom.xpath("/html/body//blockquote/*[1][name()='blockquote']")
-	if nodes:
-		messages.append(LintMessage("s-005", "Nested [xhtml]
    [/] element.", se.MESSAGE_TYPE_WARNING, filename))
+	# Check for ldquo not correctly closed
+	# Ignore closing paragraphs, line breaks, and closing cells in case ldquo means "ditto mark"
+	matches = regex.findall(r"“[^‘”]+?“", file_contents)
+	matches = [match for match in matches if "" not in match and "
    		[”\s]+?(.+?)?)
+	nodes = dom.xpath("/html/body//a[contains(@epub:type, 'noteref') and preceding-sibling::node()[normalize-space(.)] and re:test(preceding-sibling::node()[1], '\\s+$')]")
 	if nodes:
-		messages.append(LintMessage("s-010", "Empty element. Use [xhtml]
    [/] for thematic breaks if appropriate.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-012", "Illegal white space before noteref.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for 
     and 
     without ID attribute
-	nodes = dom.xpath("/html/body//*[self::section or self::article][not(@id)]")
+	# Check for period following Roman numeral, which is an old-timey style we must fix
+	# But ignore the numeral if it's the first item in a 
     tag, as that suggests it might be a kind of list item.
+	nodes = dom.xpath("/html/body//node()[name()='span' and contains(@epub:type, 'z3998:roman') and not(position()=1)][(following-sibling::node()[1])[re:test(., '^\\.\\s*[a-z]')]]")
 	if nodes:
-		messages.append(LintMessage("s-011", "Element without [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
-
-	# Check for 
     tags
-	if dom.xpath("/html/body//pre"):
-		messages.append(LintMessage("s-013", "Illegal [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("t-013", "Roman numeral followed by a period. When in mid-sentence Roman numerals must not be followed by a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "." for node in nodes]))
 
-	# Check for 
     after block-level elements
-	nodes = dom.xpath("/html/body//*[self::p or self::blockquote or self::table or self::ol or self::ul or self::section or self::article][following-sibling::br]")
+	# Check for two em dashes in a row
+	nodes = dom.xpath(f"/html/body//*[re:test(text(), '—{se.WORD_JOINER}*—+')]")
 	if nodes:
-		messages.append(LintMessage("s-014", "[xhtml]
    [/] after block-level element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
-
-	# Check for 
     tags before the end of a section, which is a common PG artifact
-	if dom.xpath("/html/body//hr[count(following-sibling::*)=0]"):
-		messages.append(LintMessage("s-012", "Illegal [xhtml]
    [/] as last child.", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("t-014", "Two or more em-dashes in a row found. Elided words should use the two- or three-em-dash Unicode character, and dialog ending in em-dashes should only end in a single em-dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for hgroup elements with a subtitle but no title
-	nodes = dom.xpath("/html/body//hgroup[./*[contains(@epub:type, 'subtitle')] and not(./*[contains(concat(' ', @epub:type, ' '), ' title ')])]")
+	# Check for money not separated by commas
+	nodes = dom.xpath("/html/body//*[re:test(text(), '[£\\$][0-9]{4,}')]")
 	if nodes:
-		messages.append(LintMessage("s-015", "Element has [val]subtitle[/] semantic, but without a sibling having a [val]title[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-015", "Numbers not grouped by commas. Separate numbers greater than 1,000 with commas at every three numerals.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for deprecated MathML elements
-	# Note we dont select directly on element name, because we want to ignore any namespaces that may (or may not) be defined
-	nodes = dom.xpath("/html/body//*[local-name()='mfenced']")
+	# Check for  that does not contain spaces
+	nodes = dom.xpath("/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and re:test(., '[A-Z]\\.[A-Z]\\.')]")
 	if nodes:
-		messages.append(LintMessage("s-017", "[xhtml][/] is deprecated in the MathML spec. Use [xhtml](...)[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-016", "Initials in [xhtml][/] not separated by spaces.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for 
     tags without id attributes
-	nodes = dom.xpath("/html/body//img[@id]")
-	if nodes:
-		messages.append(LintMessage("s-018", "[xhtml][/] element with [attr]id[/] attribute. [attr]id[/] attributes go on parent [xhtml]
    [/] elements. Hint: Images that are inline (i.e. that do not have parent [xhtml]
    [/]s do not have [attr]id[/] attributes.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	if special_file != "colophon":
+		# This file is not the colophon.
+		# Check for ending punctuation inside italics that have semantics.
+		# Ignore the colophon because paintings might have punctuation in their names
 
-	# Check for IDs on  tags
-	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][@id]")
-	if nodes:
-		messages.append(LintMessage("s-019", "[xhtml][/] element with [attr]id[/] attribute. [xhtml][/] elements should be wrapped in [xhtml]
    [/] elements, which should hold the [attr]id[/] attribute.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+		# This xpath matches b or i elements with epub:type="se:name...", that are not stage direction, whose last text node ends in punctuation.
+		# Note that we check that the last node is a text node, because we may have  as the last node
+		matches = [node.to_string() for node in dom.xpath("(//b | //i)[contains(@epub:type, 'se:name') and not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., '[\\.,!\\?;:]$')]]")]
 
-	# Check for legal cases that aren't italicized
-	# We can't use this because v. appears as short for "volume", and we may also have sporting events without italics.
-	#nodes = dom.xpath("/html/body//abbr[text()='v.' or text()='versus'][not(parent::i)]")
-	#if nodes:
-	#	messages.append(LintMessage("t-xxx", "Legal case without parent [xhtml][/].", se.MESSAGE_TYPE_WARNING, filename, {f"{node.to_string()}." for node in nodes}))
+		# Match b or i elements that are not stage directions, and that end in a comma followed by a lowercase letter
+		matches = matches + [node.to_string() for node in dom.xpath("(//b | //i)[not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., ',$')] and following-sibling::node()[re:test(., '^\\s*[a-z]')] ]")]
 
-	# Only do this check if there's one  or one 
     tag. If there's more than one, then the xhtml file probably requires an overarching title
-	# We merge two xpaths here because /
     can be either a direct child of 
    , or it could be nested in 
    
-	if len(dom.xpath("/html/body/*[name()='section' or name()='article']/*[re:test(name(), '^h[1-6]$') or name()='hgroup'] | /html/body/*[name()='section' or name()='article']/header/*[re:test(name(), '^h[1-6]$') or name()='hgroup']"))==1:
-		title = se.formatting.generate_title(dom)
+		# ...and also check for ending punctuation inside em tags, if it looks like a *part* of a clause
+		# instead of a whole clause. If the  is preceded by an em dash or quotes, or if there's punctuation
+		# and a space before it, then it's presumed to be a whole clause.
+		matches = matches + [match.strip() for match in regex.findall(r"(?]|[!\.\?…;:]\s)(?:\w+?\s*)+[\.,\!\?;]", file_contents) if match.islower()]
 
-		if not dom.xpath(f"/html/head/title[text()={se.easy_xml.escape_xpath(title.replace('&', '&'))}]"):
-			messages.append(LintMessage("s-021", f"Unexpected value for [xhtml][/] element. Expected: [text]{title}[/]. (Beware hidden Unicode characters!)", se.MESSAGE_TYPE_ERROR, filename))
+		if matches:
+			messages.append(LintMessage("t-017", "Ending punctuation inside formatting like bold, small caps, or italics. Ending punctuation is only allowed within formatting if the phrase is an independent clause.", se.MESSAGE_TYPE_WARNING, filename, list(set(matches))))
 
-	# Check to see if <h#> tags are correctly titlecased
-	# Ignore <h#> tags with an `xml:lang` attribute, as other languages have different titlecasing rules
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') or (name() = 'p' and parent::hgroup)][not(contains(@epub:type, 'z3998:roman')) and not(@xml:lang)]")
-	for node in nodes:
-		node_copy = deepcopy(node)
+	# Check for stage direction that ends in ?! but also has a trailing period
+	nodes = dom.xpath("/html/body//i[contains(@epub:type, 'z3998:stage-direction') and re:test(., '\\.$') and not((./node()[last()])[name() = 'abbr']) and (following-sibling::node()[1])[re:test(., '^[\\.,:;!?]')]]")
+	if nodes:
+		messages.append(LintMessage("t-018", "Stage direction ending in period next to other punctuation. Remove trailing periods in stage direction.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		# Remove *leading* Roman spans
-		# This matches the first child node excluding white space nodes, if it contains the z3998:roman semantic.
-		for element in node_copy.xpath("./node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman')]"):
-			element.remove()
+	# Check for single words that are in italics, but that have closing punctuation outside italics
+	# Outer wrapping match is so that .findall returns the entire match and not the subgroup
+	# The first regex also matches the first few characters before the first double quote; we use those for more sophisticated
+	# checks below, to give fewer false positives like `with its downy red hairs and its “<i xml:lang="fr">doigts de faune</i>.”`
+	matches = regex.findall(r"((?:.{1,2}\s)?“<(i|em)[^>]*?>[^<]+?</\2>[\!\?\.])", file_contents) + regex.findall(r"([\.\!\?] <(i|em)[^>]*?>[^<]+?</\2>[\!\?\.])", file_contents)
 
-		# Remove noterefs
-		for element in node_copy.xpath(".//a[contains(@epub:type, 'noteref')]"):
-			element.remove()
+	# But, if we've matched a name of something, don't include that as an error. For example, `He said, “<i epub:type="se:name.publication.book">The Decameron</i>.”`
+	# We also exclude the match from the list if:
+	# 1. The double quote is directly preceded by a lowercase letter and a space: `with its downy red hairs and its “<i xml:lang="fr">doigts de faune</i>.”`
+	# 2. The double quote is directly preceded by a lowercase letter, a comma, and a space, and the first letter within the double quote is lowercase: In the original, “<i xml:lang="es">que era un Conde de Irlos</i>.”
+	# 3. The text is a single letter that is not "I" or "a" (because then it is likely a mathematical variable)
+	matches = [match for match in matches if "epub:type=\"se:name." not in match[0] and "epub:type=\"z3998:taxonomy" not in match[0] and not regex.match(r"^[\p{Lowercase_Letter}’]+\s“", match[0]) and not regex.match(r"^[\p{Lowercase_Letter}’]+,\s“[\p{Lowercase_Letter}]", se.formatting.remove_tags(match[0])) and not regex.match(r"^.*?<.+?>[^Ia]<.+?>", match[0])]
+	if matches:
+		messages.append(LintMessage("t-019", "When a complete clause is italicized, ending punctuation except commas must be within containing italics.", se.MESSAGE_TYPE_WARNING, filename, [match[0] for match in matches]))
 
-		# Remove hidden elements, for example in poetry identified by first line (keats)
-		for element in node_copy.xpath(".//*[@hidden]"):
-			element.remove()
+	# Check for punctuation after endnotes
+	nodes = dom.xpath(f"/html/body//a[contains(@epub:type, 'noteref')][(following-sibling::node()[1])[re:test(., '^[^\\s<–\\]\\)—{se.WORD_JOINER}a-zA-Z0-9]')]]")
+	if nodes:
+		messages.append(LintMessage("t-020", "Endnote links must be outside of punctuation, including quotation marks.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-		# Unwrap (not remove) any subelements in subtitles, as they may trip up further regexes
-		for element in node_copy.xpath(".//*[contains(@epub:type, 'subtitle')]//*"):
-			element.unwrap()
+	# Check for correct typography around measurements like 2 ft.
+	# But first remove href and id attrs because URLs and IDs may contain strings that look like measurements
+	# Note that while we check m,min (minutes) and h,hr (hours) we don't check s (seconds) because we get too many false positives on years, like `the 1540s`
+	matches = regex.findall(fr"\b[1-9][0-9]*[{se.NO_BREAK_SPACE}\-]?(?:[mck]?[mgl]|ft|in|min?|h|sec|hr)\.?\b", regex.sub(r"(href|id)=\"[^\"]*?\"", "", file_contents))
+	# Exclude number ordinals, they're not measurements
+	matches = [match for match in matches if not regex.search(r"(st|nd|rd|th)", match)]
+	if matches:
+		messages.append(LintMessage("t-021", "Measurement not to standard. Numbers are followed by a no-break space and abbreviated units require an [xhtml]<abbr>[/] element. See [path][link=https://standardebooks.org/manual/1.0.0/8-typography#8.8.5]semos://1.0.0/8.8.5[/][/].", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-		title = node_copy.inner_xml()
+	# Check for nbsp within <abbr epub:type="z3998:*-name">, which is redundant
+	nodes = dom.xpath(f"/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and contains(text(), '{se.NO_BREAK_SPACE}')]")
+	if nodes:
+		messages.append(LintMessage("t-022", "No-break space found in [xhtml]<abbr epub:type=\"z3998:*-name\">[/]. This is redundant.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		# Remove leading leftover spacing and punctuation
-		title = regex.sub(r"^[\s\.\,\!\?\:\;]*", "", title)
+	# Check for trailing commas inside <i> tags at the close of dialog
+	# More sophisticated version of: \b[^\s]+?,</i>”
+	nodes = dom.xpath("/html/body//i[re:test(., ',$')][(following-sibling::node()[1])[starts-with(., '”')]]")
+	if nodes:
+		messages.append(LintMessage("t-023", "Comma inside [xhtml]<i>[/] element before closing dialog.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "”" for node in nodes]))
 
-		# Normalize whitespace
-		title = regex.sub(r"\s+", " ", title, flags=regex.DOTALL).strip()
+	# Check for quotation marks in italicized dialog
+	nodes = dom.xpath("/html/body//i[@xml:lang][starts-with(., '“') or re:test(., '”$')]")
+	if nodes:
+		messages.append(LintMessage("t-024", "When italicizing language in dialog, italics go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		# Do we have a subtitle? If so the first letter of that must be capitalized, so we pull that out
-		subtitle_matches = regex.findall(r"(.*?)<span epub:type=\"subtitle\">(.*?)</span>(.*?)", title, flags=regex.DOTALL)
-		if subtitle_matches:
-			for title_header, subtitle, title_footer in subtitle_matches:
-				title_header = se.formatting.titlecase(se.formatting.remove_tags(title_header).strip())
-				subtitle = se.formatting.titlecase(se.formatting.remove_tags(subtitle).strip())
-				title_footer = se.formatting.titlecase(se.formatting.remove_tags(title_footer).strip())
+	# Check alt attributes on images, except for the logo
+	nodes = dom.xpath("/html/body//img[not(re:test(@src, '/(logo|titlepage)\\.(svg|png)$'))]")
+	img_no_alt = []
+	img_alt_not_typogrified = []
+	img_alt_lacking_punctuation = []
+	for node in nodes:
+		if "titlepage.svg" not in node.get_attr("src"):
+			has_images = True # Save for a later check
 
-				titlecased_title = f"{title_header} {subtitle} {title_footer}"
-				titlecased_title = titlecased_title.strip()
+		alt = node.get_attr("alt")
 
-				title = se.formatting.remove_tags(title).strip()
-				if title != titlecased_title:
-					messages.append(LintMessage("s-023", f"Title [text]{title}[/] not correctly titlecased. Expected: [text]{titlecased_title}[/].", se.MESSAGE_TYPE_WARNING, filename))
+		if alt:
+			# Check for non-typogrified img alt attributes
+			if regex.search(r"""('|"|--|\s-\s|")""", alt):
+				img_alt_not_typogrified.append(node.to_tag_string())
 
-		# No subtitle? Much more straightforward
-		else:
-			titlecased_title = se.formatting.titlecase(se.formatting.remove_tags(title))
-			title = se.formatting.remove_tags(title)
-			if title != titlecased_title:
-				messages.append(LintMessage("s-023", f"Title [text]{title}[/] not correctly titlecased. Expected: [text]{titlecased_title}[/].", se.MESSAGE_TYPE_WARNING, filename))
+			# Check alt attributes not ending in punctuation
+			if filename.name not in IGNORED_FILENAMES and not regex.search(r"""[\.\!\?]”?$""", alt):
+				img_alt_lacking_punctuation.append(node.to_tag_string())
 
-	# Check for header elements that are entirely non-English
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./i[@xml:lang][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
-	if nodes:
-		messages.append(LintMessage("s-024", "Header elements that are entirely non-English should not be set in italics. Instead, the [xhtml]<h#>[/] element has the [attr]xml:lang[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+			# Check that alt attributes match SVG titles
+			img_src = node.lxml_element.get("src")
+			if img_src and img_src.endswith("svg"):
+				title_text = ""
+				image_ref = img_src.split("/").pop()
+				try:
+					svg_path = self.content_path / "images" / image_ref
+					svg_dom = self.get_dom(svg_path)
+					try:
+						title_text = svg_dom.xpath("/svg/title")[0].text
+					except Exception:
+						messages.append(LintMessage("s-027", f"{image_ref} missing [xhtml]<title>[/] element.", se.MESSAGE_TYPE_ERROR, svg_path))
 
-	# Check for z3998:roman elements with invalid values. Roman numerals can occasionally end in `j` as an alias for ending `i`. See _The Worm Ouroboros_.
-	# We also allow the numeral to end in a digit, because that might be an endnote. For example `<h2 epub:type="ordinal z3998:roman">II<a href="..." epub:type="noteref">3</a></h2>`
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:roman') and not(re:test(normalize-space(.), '^[ivxlcdmIVXLCDM]+j?[0-9]*?$'))]")
-	if nodes:
-		messages.append(LintMessage("s-026", "Invalid Roman numeral.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+					if title_text != "" and alt != "" and title_text != alt:
+						messages.append(LintMessage("s-022", f"The [xhtml]<title>[/] element of [path][link=file://{svg_path}]{image_ref}[/][/] does not match the [attr]alt[/] attribute text in [path][link=file://{filename}]{filename.name}[/][/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for <h#> elements that are higher or lower than their expected level based on how deep they are in <section>s
-	# or <article>s. Exclude <nav> in the ToC.
-	# Get all h# elements not preceded by other h# elements (this includes the 1st item of an hgroup but not the next items)
-	# Exclude the ToC and landmarks
-	invalid_headers = []
-	invalid_parent_ids = []
-	for heading in dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') and not(./preceding-sibling::*[re:test(name(), '^h[1-6]$')]) and not(./ancestor::*[re:test(@epub:type, '\\b(toc|landmarks)\\b')])]"):
-		# Get the IDs of the direct parent section
-		parent_section = heading.xpath("./ancestor::*[@id and re:test(name(), '^(section|article|nav)$')][1]")
+				except FileNotFoundError:
+					missing_files.append(str(Path("images") / image_ref))
 
-		if parent_section:
-			section = _find_ebook_section(parent_section[0].get_attr("id"), section_tree)
+		else:
+			img_no_alt.append(node.to_tag_string())
 
-			if not section:
-				# We can accidentally raise s-029 if the file isn't in the spine. This is technically correct, but s-029 is a
-				# misleading message in that case, and f-007 will also be raised, which is the correct message.
-				if filename in self.spine_file_paths:
-					invalid_parent_ids.append(parent_section[0].to_tag_string())
-			else:
-				if str(section.depth) != heading.tag[1:2]:
-					invalid_headers.append(heading.to_string())
+	if img_alt_not_typogrified:
+		messages.append(LintMessage("t-025", "Non-typogrified [text]'[/], [text]\"[/] (as [xhtml]"[/]), or [text]--[/] in image [attr]alt[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, img_alt_not_typogrified))
+	if img_alt_lacking_punctuation:
+		messages.append(LintMessage("t-026", "[attr]alt[/] attribute does not appear to end with punctuation. [attr]alt[/] attributes must be composed of complete sentences ending in appropriate punctuation.", se.MESSAGE_TYPE_ERROR, filename, img_alt_lacking_punctuation))
+	if img_no_alt:
+		messages.append(LintMessage("s-004", "[xhtml]img[/] element missing [attr]alt[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, img_no_alt))
 
-	if invalid_parent_ids:
-		messages.append(LintMessage("s-029", "Section with [attr]data-parent[/] attribute, but no section having that [attr]id[/] in ebook.", se.MESSAGE_TYPE_ERROR, filename, invalid_parent_ids))
-	if invalid_headers:
-		messages.append(LintMessage("s-085", "[xhtml]<h#>[/] element found in a [xhtml]<section>[/] or a [xhtml]<article>[/] at an unexpected level. Hint: Headings not in the title page start at [xhtml]<h2>[/]. If this work has parts, should this header be [xhtml]<h3>[/] or higher?", se.MESSAGE_TYPE_ERROR, filename, invalid_headers))
+	# Check for low-hanging misquoted fruit
+	matches = regex.findall(r"[\p{Letter}]+[“‘]", file_contents) + regex.findall(r"[^>]+</(?:em|i|b|span)>‘[\p{Lowercase_Letter}]+", file_contents)
+	if matches:
+		messages.append(LintMessage("t-028", "Possible mis-curled quotation mark.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	# Check for a common typo
-	if "z3998:nonfiction" in file_contents:
-		messages.append(LintMessage("s-030", "[val]z3998:nonfiction[/] should be [val]z3998:non-fiction[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for periods followed by lowercase.
+	temp_xhtml = regex.sub(r"<title[^>]*?>.+?", "", file_contents) # Remove  because it might contain something like <title>Chapter 2: The Antechamber of M. de Tréville
+	temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Replace things like 
+	temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Remove  attributes
+	temp_xhtml = temp_xhtml.replace("A.B.C.", "X") # Remove A.B.C, which is not an abbreviations.
+	# Note the regex also excludes preceding numbers, so that we can have inline numbering like:
+	# "A number of questions: 1. regarding those who make heretics; 2. concerning those who were made heretics..."
+	matches = regex.findall(r"[^\s0-9]+\.\s+[\p{Lowercase_Letter}](?!’[\p{Uppercase_Letter}])[\p{Lowercase_Letter}]+", temp_xhtml)
+	# If  is in the match, remove it from the matches so we exclude things like et. al.
+	matches = [match for match in matches if "" not in match]
+	if matches:
+		messages.append(LintMessage("t-029", "Period followed by lowercase letter. Hint: Abbreviations require an [xhtml][/] element.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	# Run some checks on epub:type values
-	incorrect_attrs = set()
-	unnecessary_z3998_attrs = set()
-	duplicate_attrs = []
+	# Check for initialisms without periods
+	nodes = [node.to_string() for node in dom.xpath("/html/body//abbr[contains(@epub:type, 'z3998:initialism') and not(re:test(., '^[0-9]*([a-zA-Z]\\.)+[0-9]*$'))]") if node.text not in INITIALISM_EXCEPTIONS]
+	if nodes:
+		messages.append(LintMessage("t-030", "Initialism with spaces or without periods.", se.MESSAGE_TYPE_WARNING, filename, set(nodes)))
 
-	for node in dom.xpath("//*[@epub:type]"):
-		attrs = set()
+	nodes = dom.xpath("/html/body//*[re:test(text(), '\\bA\\s*B\\s*C\\s*\\b')]")
+	if nodes:
+		messages.append(LintMessage("t-031", "[text]A B C[/] must be set as [text]A.B.C.[/] It is not an abbreviation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		for val in regex.split(r"\s+", node.get_attr("epub:type")):
-			if val in attrs:
-				duplicate_attrs.append(node.to_tag_string())
-			else:
-				attrs.add(val)
+	# Check for abbreviations followed by periods
+	# But we exclude some SI units, which don't take periods; abbreviations ending in numbers for example in stage directions; abbreviations like `r^o` (recto) that contain ; and some Imperial abbreviations that are multi-word
+	nodes = dom.xpath("/html/body//abbr[(contains(@epub:type, 'z3998:initialism') or re:test(@epub:type, 'z3998:[^\\s]+?name') or not(@epub:type))][not(re:test(., '[cmk][mgl]')) and not(re:test(., '[0-9]$')) and not(./sup) and not(re:test(@class, '\\b(era|temperature|compound)\\b')) and not(re:test(text(), '^(mpg|mph|hp|TV)$'))][following-sibling::text()[1][starts-with(self::text(), '.')]]")
+	if nodes:
+		messages.append(LintMessage("t-032", "Initialism or name followed by period. Hint: Periods go within [xhtml][/]. [xhtml][/]s containing periods that end a clause require the [class]eoc[/] class.", se.MESSAGE_TYPE_WARNING, filename, [f"{node.to_string()}." for node in nodes]))
 
-			if val.startswith("z3998:"):
-				bare_val = val.replace("z3998:", "")
-				if bare_val not in Z3998_SEMANTIC_VOCABULARY:
-					incorrect_attrs.add(val)
+	# Check for space after dash
+	nodes = dom.xpath("/html/body//*[name()='p' or name()='span' or name='em' or name='i' or name='b' or name='strong'][not(self::comment())][re:test(., '[a-zA-Z]-\\s(?!(and|or|nor|to|und|…)\\b)')]")
+	if nodes:
+		messages.append(LintMessage("t-033", "Space after dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				elif bare_val in EPUB_SEMANTIC_VOCABULARY:
-					unnecessary_z3998_attrs.add((val, bare_val))
+	# Check for  preceded by em dash
+	nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[re:test(., '—$')]]")
+	if nodes:
+		messages.append(LintMessage("t-034", "[xhtml][/] element preceded by em-dash. Hint: em-dashes go within [xhtml][/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-			elif val.startswith("se:"):
-				bare_val = val.replace("se:", "")
-				if bare_val not in SE_SEMANTIC_VOCABULARY:
-					incorrect_attrs.add(val)
+	# Check for  without preceding space in text node. (preceding ( or [ are also OK)
+	nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[not(re:test(., '[\\[\\(\\s]$'))]]")
+	if nodes:
+		messages.append(LintMessage("t-035", "[xhtml][/] element not preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-			else:
-				# Regular epub vocabulary
-				if val not in EPUB_SEMANTIC_VOCABULARY:
-					incorrect_attrs.add(val)
+	# Check for partially obscured years in which the last year is an em dash
+	nodes = dom.xpath("/html/body//p[re:test(.,  '1\\d{2}⁠—[^“’A-Za-z<]')]")
+	if nodes:
+		messages.append(LintMessage("t-036", "Em-dash used to obscure single digit in year. Hint: Use a hyphen instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	if duplicate_attrs:
-		messages.append(LintMessage("s-031", "Duplicate value in [attr]epub:type[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, duplicate_attrs))
-	if incorrect_attrs:
-		messages.append(LintMessage("s-032", "Invalid value for [attr]epub:type[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, incorrect_attrs))
-	if unnecessary_z3998_attrs:
-		messages.append(LintMessage("s-034", "Semantic used from the z3998 vocabulary, but the same semantic exists in the EPUB vocabulary.", se.MESSAGE_TYPE_ERROR, filename, [attr for (attr, bare_attr) in unnecessary_z3998_attrs]))
+	# Check for rdquo preceded by space (but not preceded by a rsquo, which might indicate a nested quotation)
+	nodes = dom.xpath("/html/body//p[re:test(., '[^’]\\s”')]")
+	if nodes:
+		messages.append(LintMessage("t-037", "[text]”[/] preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check if language tags in individual files match the language in the metadata file
-	if filename.name not in IGNORED_FILENAMES:
-		file_language = dom.xpath("/html/@xml:lang", True)
-		if language != file_language:
-			messages.append(LintMessage("s-033", f"File language is [val]{file_language}[/], but [path][link=file://{self.metadata_file_path}]{self.metadata_file_path.name}[/][/] language is [val]{language}[/].", se.MESSAGE_TYPE_WARNING, filename))
+	# Check obviously miscurled quotation marks
+	nodes = dom.xpath("/html/body//p[re:test(., '“$')]")
+	if nodes:
+		messages.append(LintMessage("t-038", "[text]“[/] before closing [xhtml]
    [/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for endnotes
-	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[not(preceding-sibling::*)]/cite[not(preceding-sibling::node()[normalize-space(.)]) and (following-sibling::node()[normalize-space(.)])[1][contains(@epub:type, 'backlink')]]")
+	# Check for some known initialisms with incorrect possessive apostrophes
+	nodes = dom.xpath("/html/body//abbr[text()='I.O.U.'][(following-sibling::node()[1])[starts-with(., '’s')]]")
 	if nodes:
-		messages.append(LintMessage("s-035", "Endnote containing only [xhtml][/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-039", "Initialism followed by [text]’s[/]. Hint: Plurals of initialisms are not followed by [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "’s" for node in nodes]))
 
-	# Check for wrong semantics in frontmatter/backmatter
-	if dom.xpath("/html/body/section[not(@data-parent) and re:test(@epub:type, '\\b(dedication|introduction|preface|foreword|preamble|titlepage|halftitlepage|imprint|epigraph|acknowledgements)\\b') and not(ancestor-or-self::*[contains(@epub:type, 'frontmatter')])]"):
-		messages.append(LintMessage("s-036", "No [val]frontmatter[/] semantic inflection for what looks like a frontmatter file.", se.MESSAGE_TYPE_WARNING, filename))
+	# Check if a subtitle ends in a text node with a terminal period; or if it ends in an  node containing a terminal period.
+	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6]/*[contains(@epub:type, 'subtitle')][(./text())[last()][re:test(., '\\.$')] or (./i)[last()][re:test(., '\\.$')]]")
+	if nodes:
+		messages.append(LintMessage("t-040", "Subtitle with illegal ending period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	if dom.xpath("/html/body/section[not(@data-parent) and re:test(@epub:type, '\\b(endnotes|loi|afterword|appendix|colophon|copyright\\-page|lot)\\b') and not(ancestor-or-self::*[contains(@epub:type, 'backmatter')])]"):
-		messages.append(LintMessage("s-037", "No [val]backmatter[/] semantic inflection for what looks like a backmatter file.", se.MESSAGE_TYPE_WARNING, filename))
+	matches = regex.findall(r"[^…]\s[!?;:,].{0,10}", file_contents) # If we don't include preceding chars, the regex is 6x faster
+	if matches:
+		messages.append(LintMessage("t-041", "Illegal space before punctuation.", se.MESSAGE_TYPE_ERROR, filename, matches))
 
-	# Check for leftover asterisms. Asterisms are sequences of any of these chars: * . • -⁠ —
-	nodes = dom.xpath("/html/body//*[self::p or self::div][re:test(., '^\\s*[\\*\\.•\\-⁠—]\\s*([\\*\\.•\\-⁠—]\\s*)+$')]")
+	# Check for a list of some common English loan words that used to be italicized as foreign, but are no longer
+	nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and re:test(., '^(fianc[eé]+|divorc[eé]+|menu|recherch[eé]|tour-de-force|outr[eé]|d[ée]but(ante)?|apropos|[eé]lite|prot[ée]g[ée]+|chef|salon|r[eé]gime|contretemps|[eé]clat|aides?|entr[ée]+|t[eê]te-[aà]-t[eê]tes?|blas[eé]|bourgeoisie)$')]")
 	if nodes:
-		messages.append(LintMessage("s-038", "Illegal asterism. Section/scene breaks must be defined by an [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-043", "Non-English loan word set in italics, when modern typography omits italics. Hint: A word may be correctly italicized when emphasis is desired, if the word is meant to be pronounced with an accent, or if the word is part of non-English speech.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for  element without a  child
-	if dom.xpath("/html/body//table[not(tbody)]"):
-		messages.append(LintMessage("s-042", "[xhtml]
    [/] element without [xhtml][/] child.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for comma after leading Or in subtitles
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'subtitle') and re:test(text(), '^Or\\s')]")
+	if nodes:
+		messages.append(LintMessage("t-044", "Comma required after leading [text]Or[/] in subtitles.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check that short stories are on an 
     element
-	nodes = dom.xpath("/html/body/section[contains(@epub:type, 'se:short-story') or contains(@epub:type, 'se:novella')]")
+	# Check for personas that are italicized. Use re:test instead of contains to avoid catching z3998:personal-name
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:persona\\b') and @data-css-font-style = 'italic']")
 	if nodes:
-		messages.append(LintMessage("s-043", "[val]se:short-story[/] semantic on element that is not [xhtml]
    [/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-045", "Element has [val]z3998:persona[/] semantic and is also set in italics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for poetry/verse without a descendent 
     element.
-	# Skip the ToC landmarks because it may have poem/verse semantic children.
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')][not(descendant::p)][not(ancestor::nav[contains(@epub:type, 'landmarks')])]")
+	if "῾" in file_contents:
+		messages.append(LintMessage("t-046", "[text]῾[/] (U+1FFE) detected. Use [text]ʽ[/] (U+02BD) instead.", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Check for all-caps first paragraphs in sections/articles
+	# Note that we don't check for small-caps CSS, because there are lots of legitimate cases for that,
+	# and it would generate too many false positives.
+	# If we did want to do that, this xpath after the last re:test would help: or ./*[not(preceding-sibling::node()[normalize-space(.)]) and @data-css-font-variant='small-caps' and following-sibling::node()[1][self::text()]]
+	nodes = dom.xpath("/html/body//*[(name() = 'section' or name() = 'article') and not(contains(@epub:type, 'dedication') or contains(@epub:type, 'z3998:letter'))]/p[1][re:test(normalize-space(.), '^[“’]?(?:[A-Z’]+\\s){2,}')]")
 	if nodes:
-		messages.append(LintMessage("s-044", "Element with poem or verse semantic, without descendant [xhtml]
    [/] (stanza) element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-048", "Chapter opening text in all-caps.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for  elements that have two or more letters/periods, that don't have a semantic epub:type
-	# SS. is the French abbreviation for "Saints"
-	nodes = dom.xpath("/html/body//abbr[not(@epub:type)][text() != 'U.S.' and text() != 'SS.'][re:test(., '([A-Z]\\.?){2,}')]")
-	filtered_nodes = []
-	for node in nodes:
-		add_node = True
+	# Check for two-em-dashes used for elision instead of three-em-dashes
+	matches = regex.findall(fr"[^{se.WORD_JOINER}\p{{Letter}}”]⸺[^“{se.WORD_JOINER}\p{{Letter}}].*", file_contents, flags=regex.MULTILINE)
+	if matches:
+		messages.append(LintMessage("t-049", "Two-em-dash used for eliding an entire word. Use a three-em-dash instead.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-		if node.text in INITIALISM_EXCEPTIONS:
-			add_node = False
-		elif node.get_attr("class"):
-			for attr_value in node.get_attr("class").split():
-				if attr_value in IGNORED_CLASSES:
-					add_node = False
-					break
-		elif node.text in ("A.M.", "P.M.") and node.xpath("./preceding-sibling::node()[re:test(., '[0-9]\\s$')]"):
-			# Ignore instances of capitalized times, usually in titles. For example:
-			# `From 10:20 P.M. to 10:47 P.M.` in a chapter title.
-			# We have to test for a preceding number because `P.M.` in other contexts may mean `Prime Minister`
-			add_node = False
+	# Check that possessives appear within persona blocks
+	nodes = dom.xpath("/html/body//b[contains(@epub:type, 'z3998:persona') and ./following-sibling::node()[1][re:test(., '^’s?')]]")
+	if nodes:
+		messages.append(LintMessage("t-050", "Possessive [text]’s[/] or [text]’[/] outside of element with [val]z3998:persona[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		if add_node:
-			filtered_nodes.append(node)
+	# Check for quotations that carry to the next paragraph, but the next paragraph has no opening quotation mark
+	# Exclude p in blockquote because often that has different formatting
+	nodes = dom.xpath("/html/body//p[re:test(., '^“[^”]+$') and not(./ancestor::blockquote) and ./following-sibling::*[1][name() = 'p' and re:test(normalize-space(), '^[^“]')]]")
+	if nodes:
+		messages.append(LintMessage("t-051", "Dialog in [xhtml]
    [/] that continues to the next [xhtml]
    [/], but the next [xhtml]
    [/] does not begin with [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	if filtered_nodes:
-		messages.append(LintMessage("s-045", "[xhtml][/] element without semantic inflection like [class]z3998:personal-name[/] or [class]z3998:initialism[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in filtered_nodes]))
+	# Check for stage direction without ending punctuation. We only want to consider stage direction that is not an interjection in a parent clause.
+	# We match if the stage direction ends in lowercase, and if there's no following node, or if there's a following node that doesn't begin in lowercase or an em dash or semicolon (which suggests an interjection).
+	nodes = dom.xpath(f"/html/body//i[@epub:type='z3998:stage-direction' and re:test(., '[a-z]$') and re:test(., '^[A-Z]') and (not(./following-sibling::node()[normalize-space(.)]) or ./following-sibling::node()[1][re:test(normalize-space(.), '^[^{se.WORD_JOINER}—;a-z]')])]")
+	if nodes:
+		messages.append(LintMessage("t-052", "Stage direction without ending punctuation. Note that ending punctuation is optional in stage direction that is an interjection in a parent clause, and such interjections should begin with a lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for 
     elems that has some element children, which are only  and 
     children, but the parent doesn't have poem/verse semantics.
-	# Ignore spans that have a class, but not if the class is an i# class (for poetry indentation)
-	nodes = dom.xpath("/html/body//p[not(./text()[normalize-space(.)])][*][not(ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][not(*[not(self::span) and not(self::br)])][ (not(span[@epub:type]) and not(span[@class]) ) or span[re:test(@class, '\\bi[0-9]\\b')]]")
+	# Check for stage direction starting in lowercase.  We only want to consider stage direction that is not an interjection in a parent clause.
+	# We match if the stage direction starts with a lowercase, and if there's no preceding node, or if there is a preceding node and it doesn't end in a lowercase letter or a small subset of conjoining punctuation.
+	nodes = dom.xpath("/html/body//i[@epub:type='z3998:stage-direction' and re:test(., '^[a-z]') and (not(./preceding-sibling::node()[normalize-space(.)]) or ./preceding-sibling::node()[1][re:test(normalize-space(.), '[^a-z:—,;…]$')])]")
 	if nodes:
-		messages.append(LintMessage("s-046", "[xhtml]
    [/] element containing only [xhtml][/] and [xhtml]
    [/] elements, but its parent doesn’t have the [val]z3998:poem[/], [val]z3998:verse[/], [val]z3998:song[/], [val]z3998:hymn[/], or [val]z3998:lyrics[/] semantic. Multi-line clauses that are not verse don’t require [xhtml][/]s.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-053", "Stage direction starting in lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# For this series of selections, we select spans that are direct children of p, because sometimes a line of poetry may have a nested span.
-	nodes = dom.xpath("/html/body/*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')]/descendant-or-self::*/p/span/following-sibling::*[contains(@epub:type, 'noteref') and name()='a' and position()=1]")
+	# Epigraphs that are entirely non-English should still be in italics, not Roman
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph') or contains(@epub:type, 'bridgehead')]//p[./i[@xml:lang and @data-css-font-style='normal' and ( (./preceding-sibling::node()='“' and ./following-sibling::node()='”') or not(./preceding-sibling::node()[normalize-space(.)] or ./following-sibling::node()[normalize-space(.)]) ) ]]")
 	if nodes:
-		messages.append(LintMessage("s-047", "[val]noteref[/] as a direct child of element with poem or verse semantic. [val]noteref[/]s should be in their parent [xhtml][/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-054", "Epigraphs that are entirely non-English should be set in italics, not Roman.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for incorrectly applied se:name semantic
-	nodes = dom.xpath("/html/body//*[self::p or self::blockquote][contains(@epub:type, 'se:name.')]")
+	# Check for lone acute accents, which should always be combined or swapped for a more accurate Unicode char
+	if "´" in file_contents:
+		messages.append(LintMessage("t-055", "Lone acute accent ([val]´[/]). A more accurate Unicode character like prime for coordinates or measurements, or combining accent or breathing mark for Greek text, is required.", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Check for degree confusable
+	nodes = dom.xpath("/html/body//*[re:test(text(), '[0-9]+º')]")
 	if nodes:
-		messages.append(LintMessage("s-048", "[val]se:name[/] semantic on block element. [val]se:name[/] indicates the contents is the name of something.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-056", "Masculine ordinal indicator ([val]º[/]) used instead of degree symbol ([val]°[/]). Note that the masculine ordinal indicator may be appropriate for ordinal numbers read as Latin, i.e. [val]12º[/] reading [val]duodecimo[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+
+	# Check for 
     starting in lowercase
+	# Exclude 
     starting in constructs like `(a)` or `a.` as they may be numbered lists
+	nodes = dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::figure or ancestor::ol[not(ancestor::section[contains(@epub:type, 'endnotes')])] or ancestor::ul or ancestor::table or ancestor::hgroup or preceding-sibling::*[1][name() = 'hr']) and not(contains(@class, 'continued') or re:test(@epub:type, 'z3998:(signature|valediction)')) and not(./*[1][name() = 'math' or name() = 'var' or contains(@epub:type, 'z3998:grapheme')]) and re:test(., '^[^A-Za-z0-9]?[a-z]') and not(re:test(., '^\\([a-z0-9]\\.?\\)\\.?')) and not(re:test(., '^[a-z0-9]\\.\\s'))]")
+
+	# We have to additionally filter using the regex library, because often a sentence may begin with an uppercase ACCENTED letter,
+	# and xpath's limited regex library doesn't support Unicode classes
+	nodes = [node for node in nodes if regex.match(r"^[^\p{Letter}0-9]?[\p{Lowercase_Letter}]", node.inner_text())]
 
-	# Check for 
     elements with only h# child nodes
-	nodes = dom.xpath("/html/body//header[./*[re:test(name(), 'h[1-6]') and (count(preceding-sibling::*) + count(following-sibling::*)=0)]]")
 	if nodes:
-		messages.append(LintMessage("s-049", "[xhtml]
    [/] element whose only child is an [xhtml][/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-057", "[xhtml]
    [/] starting with lowercase letter. Hint: [xhtml]
    [/] that continues text after a [xhtml]
    [/] requires the [class]continued[/] class; and use [xhtml]
    [/] to split one clause over many lines.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for s that only exist to apply epub:type
-	nodes = dom.xpath("/html/body//*[span[@epub:type][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	# Check for illegal characters
+	matches = regex.findall(fr"({se.UNICODE_BOM}|{se.SHY_HYPHEN})", file_contents)
+	if matches:
+		# Get the keys of a dict in order to create a list without duplicates
+		messages.append(LintMessage("t-058", "Illegal character.", se.MESSAGE_TYPE_ERROR, filename, list({match.encode("unicode_escape").decode().replace("\\u", "U+").upper():None for match in matches}.keys())))
+
+	# Check for period in cite in endnotes
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]//cite[not((./node()[last()])[name() = 'abbr']) and ./following-sibling::*[1][contains(@epub:type, 'backlink')] and re:test(., '^—') and ( (re:test(., '\\.$') and ./following-sibling::node()[re:test(., '^\\s*$')]) or ./following-sibling::node()[re:test(., '^\\.\\s*$')])]")
 	if nodes:
-		messages.append(LintMessage("s-050", "[xhtml][/] element appears to exist only to apply [attr]epub:type[/]. [attr]epub:type[/] should go on the parent element instead, without a [xhtml][/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-059", "Period at the end of [xhtml][/] element before endnote backlink.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for language tags misusing the unk lang code
-	nodes = dom.xpath("//*[@xml:lang='unk']")
+	# Check for Bible verses in old-style notation
+	nodes = dom.xpath("/html/body//*[(name()='p' or name()='cite' or re:test(name(), '^h[1-6]$')) and .//node()[re:test(., '(Genesis|Gen\\.|Exodus|Ex\\.|Leviticus|Lev\\.|Numbers|Num\\.|Deuteronomy|Deut\\.|Joshua|Josh\\.|Judges|Ruth|Kings|Chronicles|Chron\\.|Ezra|Nehemiah|Neh\\.|Esther|Esth\\.|Job|Psalm|Psalms|Ps\\.|Proverbs|Prov\\.|Ecclesiastes|Ecc\\.|Eccl\\.|Solomon|Sol\\.|Isaiah|Is\\.|Isa\\.|Jeremiah|Jer\\.|Lamentations|Lam\\.|Ezekiel|Ez\\.|Ezek\\.|Daniel|Dan\\.|Hosea|Hos\\.|Joel|Amos|Obadiah|Obad\\.|Jonah|Jon\\.|Micah|Mic\\.|Nahum|Nah\\.|Habakkuk|Hab\\.|Zephaniah|Zeph\\.|Haggai|Hag\\.|Zechariah|Zech\\.|Malachi|Mal\\.|Tobit|Judith|Sirach|Baruch|Maccabees|Esdras|Manasses|Matthew|Matt\\.|Mark|Luke|John|Acts|Romans|Rom\\.|Corinthians|Cor\\.|Corinth\\.|Galatians|Gal\\.|Ephesians|Eph\\.|Philippians|Phil\\.|Philipp\\.|Colossians|Col\\.|Coloss\\.|Thessalonians|Thes\\.|Thess\\.|Timothy|Tim\\.|Titus|Tit\\.|Philemon|Phil\\.|Hebrews|Heb\\.|James|Jas\\.|Peter|Pet\\.|Jude|Revelation|Revelations|Rev\\.)\\s*$') and following-sibling::node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman') and following-sibling::node()[1][re:test(., '^\\s*[\\.,]?\\s+[0-9]')]]]]")
 	if nodes:
-		messages.append(LintMessage("s-051", "Element with [xhtml]xml:lang=\"unk\"[/] should be [xhtml]xml:lang=\"und\"[/] (undefined).", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-060", "Old style Bible citation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for title attrs on abbr elements
-	nodes = dom.xpath("/html/body//abbr[@title]")
+	# Check for summary-style bridgeheads not ending in punctuation
+	# Decide if it's a summary-style bridgehead if it contains two or more em dashes. We use Python regex because
+	# xpath can't count the number of occurances of a string.
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'bridgehead') and re:test(., '[^\\.\\!\\?”]”?$')]")
+	nodes = [node for node in nodes if len(regex.findall(r"—", node.to_string())) >= 2]
 	if nodes:
-		messages.append(LintMessage("s-052", "[xhtml][/] element with illegal [attr]title[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-061", "Summary-style bridgehead without ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	nodes = dom.xpath("/html/body//blockquote//p[parent::*[name()='footer'] or parent::*[name()='blockquote']]//cite") # Sometimes the 
     may be in a 
    
+	# Check for uppercased am/pm, except for if they have an epub:type, or when uppercased in titles, or in the nav document
+	if not dom.xpath("//nav[contains(@epub:type, 'toc')]"):
+		nodes = dom.xpath("/html/body//abbr[re:test(., '^[AP]\\.M\\.$') and not(@epub:type) and not(./ancestor::*[contains(@epub:type, 'title')])]")
+		if nodes:
+			messages.append(LintMessage("t-062", "Uppercased [text]a.m.[/] and [text]p.m.[/]", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+
+	# Check for some Latinisms that need italics according to the manual. We check ancestor-or-self in case the phrase is set in Roman because it's nested in a parent italic.
+	# Exclude toto followed by ’ since Toto can be a name.
+	# Exclude  whose entire contents is a matched Latinism as we do not italicize those.
+	# Ignore the ToC because we have different rules there
+	nodes = dom.xpath("/html/body//text()[re:test(., '\\b(a (priori|posteriori|fortiori)|(?[/] as child of [xhtml]
    [/] in [xhtml]
    [/]. [xhtml][/] should be the direct child of [xhtml]
    [/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-063", "Non-English confusable phrase set without italics.", se.MESSAGE_TYPE_WARNING, filename, nodes))
 
-	# Check for 
     ancestor. However, [/] ancestor. Note: [xhtml] ancestor. However, [/] ancestor. Note: [xhtml]
     element without a 
      and  are allowed, for use in tables with headers in the middle of tables (url:https://standardebooks.org/ebooks/dorothy-day/the-eleventh-virgin) and vertical table headers
-	# (https://standardebooks.org/ebooks/charles-babbage/passages-from-the-life-of-a-philosopher)
-	if dom.xpath("/html/body//table//th[not(ancestor::thead)][not(@scope)][not(count(node())=0)]"):
-		messages.append(LintMessage("s-055", "[xhtml][/] element not in [xhtml]
    [/] elements used as mid-table headings or horizontal row headings require the [attr]scope[/] attribute.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check that all names are correctly titlecased. Ignore titles with xml:lang since non-English languages have different titlecasing rules,
+	# and ignore titles that have children elements, because se.titlecase() can't handle XML-like titles right now.
+	# Ignore titles longer than 150 chars, as long titles are likely old-timey super-long titles that should be mostly sentence-cased
+	incorrectly_cased_titles = []
+	for node in dom.xpath("/html/body//*[contains(@epub:type, 'se:name') and not(contains(@epub:type, 'se:name.legal-case')) and not(@xml:lang) and not(./*) and string-length(.) <= 150]"):
+		# Replace any space that is not a hair space with a regular space. This is because in inline titles, we may correctly
+		# have nbsp for example after `St.`, but titlecase will remove that nbsp.
+		if se.formatting.titlecase(node.inner_text()) != regex.sub(fr"[^\S{se.HAIR_SPACE}]+", " ", node.inner_text()):
+			incorrectly_cased_titles.append(node.to_string())
+
+	if incorrectly_cased_titles:
+		messages.append(LintMessage("t-064", "Title not correctly titlecased. Hint: Non-English titles should have an [attr]xml:lang[/] attribute as they have different titlecasing rules.", se.MESSAGE_TYPE_WARNING, filename, incorrectly_cased_titles))
 
-	# Check for z3998:stage-direction on elements that are not 
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:stage-direction') and name() !='i' and name() !='abbr' and name() !='p']")
+	# Check for headers ending in periods
+	# Allow periods at the end of headers that both start and end with double quotes; but headers that only end in double quotes probably don't need periods
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'z3998:initialism')) and not((./node()[last()])[name() = 'abbr']) and (re:test(., '\\.$') or re:test(., '^[^“].+\\.”$'))]")
 	if nodes:
-		messages.append(LintMessage("s-058", "[attr]z3998:stage-direction[/] semantic only allowed on [xhtml][/], [xhtml][/], and [xhtml]
    [/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-065", "Header ending in a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check that internal links don't begin with ../
-	nodes = dom.xpath("/html/body//a[re:test(@href, '^\\.\\./text/')]")
+	# Check for royal names whose roman numeral is preceded by `the`
+	nodes = dom.xpath("/html/body//span[contains(@epub:type, 'z3998:roman') and ./preceding-sibling::node()[1][re:test(., '[A-Z]\\w+ the $')]]/parent::*")
 	if nodes:
-		messages.append(LintMessage("s-059", "Internal link beginning with [val]../text/[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-066", "Regnal ordinal preceded by [text]the[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for italics on things that shouldn't be italics
-	nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.music.song') or contains(@epub:type, 'se:name.publication.short-story') or contains(@epub:type, 'se:name.publication.essay')]")
+	# Check for plural graphemes without apostrophes
+	nodes = dom.xpath("/html/body//i[re:test(@epub:type, 'z3998:(grapheme|phoneme|morpheme)') and ./following-sibling::node()[1][starts-with(., 's')]]")
 	if nodes:
-		messages.append(LintMessage("s-060", "Italics on name that requires quotes instead.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-067", "Plural [val]z3998:grapheme[/], [val]z3998:phoneme[/], or [val]z3998:morpheme[/] formed without apostrophe ([text]’[/]).", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for h# tags followed by header content, that are not children of 
    .
-	# Only match if there is a following 
    , because we could have the case where there's an epigraph after a division title.
-	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][following-sibling::*[contains(@epub:type, 'epigraph') or contains(@epub:type, 'bridgehead')]][following-sibling::p][not(parent::header)]")
+	# Find citations not offset by em dashes. We're only interested in s in endnotes that are directly followed by the backlink;
+	# ignore s starting with `Cf.` which means `Compare` and was often used to mean `see`
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[last()]/cite[following-sibling::node()[normalize-space(.)][1][name() = 'a' and contains(@epub:type, 'backlink')] and re:test(., '^[^—]') and not(re:test(., '^(Cf\\.|\\()'))]")
 	if nodes:
-		messages.append(LintMessage("s-061", "Title and following header content not in a [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-068", "Citation not offset with em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for 
     elements without exactly one  child, but only in glossaries
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'glossary')]//dt[not(count(./dfn)=1)]")
+	# Check for  in 
    s that start with a leading em dash
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '^—')]")
 	if nodes:
-		messages.append(LintMessage("s-062", "[xhtml]
    [/] element in a glossary without exactly one [xhtml][/] child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-069", "[xhtml][/] in epigraph starting with an em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check that `z3998:persona` is only on  or 
    		. We use the contact() xpath function so we don't catch `z3998:personal-name`
-	nodes = dom.xpath("/html/body//*[contains(concat(' ', @epub:type, ' '), ' z3998:persona ') and not(self::b or self::td)]")
+	# Check for  in 
    s that end with a period. Try to exclude s that are last children
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '\\.$') and not( (./node()[last()])[./descendant-or-self::abbr]) ]")
 	if nodes:
-		messages.append(LintMessage("s-063", "[val]z3998:persona[/] semantic on element that is not a [xhtml][/] or [xhtml]
    		[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-070", "[xhtml][/] in epigraph ending in a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for fulltitle semantic on a header not in the half title
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'fulltitle') and name()!='h2' and name()!='hgroup' and not(ancestor::*[contains(@epub:type, 'halftitlepage')])]")
-	if nodes:
-		messages.append(LintMessage("s-065", "[val]fulltitle[/] semantic on element that is not in the half title.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# If we have Greek text, try to normalize it
+	for node in dom.xpath("/html/body//*[@xml:lang='grc' or @xml:lang='el']"):
+		node_text = node.inner_text()
+		expected_text = se.typography.normalize_greek(node_text)
+		if node_text != expected_text:
+			messages.append(LintMessage("t-073", f"Possible transcription error in Greek. Found: [text]{node_text}[/], but expected [text]{expected_text}[/text]. Hint: Use [bash]se unicode-names[/] to see differences in Unicode characters.", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for header elements that have a label, but are missing the label semantic
-	# Find h# nodes whose first child is a text node matching a label type, and where that text node's next sibling is a semantic roman numeral
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./node()[1][self::text() and not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)$') and following-sibling::*[1][contains(@epub:type, 'z3998:roman')]]]")
-	if nodes:
-		messages.append(LintMessage("s-066", "Header element missing [val]label[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	return messages
 
-	# Check for header elements that have a label semantic, but are missing an ordinal sibling
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./span[contains(@epub:type, 'label')]][not(./span[contains(@epub:type, 'ordinal')])]")
-	if nodes:
-		messages.append(LintMessage("s-067", "Header element with a [val]label[/] semantic child, but without an [val]ordinal[/] semantic child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+def _lint_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str) -> list:
+	"""
+	Helper function used in self.lint()
+	Process XHTML checks on an .xhtml file
 
-	# Check for header elements with a roman semantic but without an ordinal semantic
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][contains(@epub:type, 'z3998:roman') and not(contains(@epub:type, 'ordinal'))]")
-	if nodes:
-		messages.append(LintMessage("s-068", "Header element missing [val]ordinal[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
+	file_contents: The contents of the file being checked
 
-	# Check for body element without child section or article. Ignore the ToC because it has a unique structure
-	nodes = dom.xpath("/html/body[not(./*[name()='section' or name()='article' or (name()='nav' and re:test(@epub:type, '\\b(toc|loi)\\b'))])]")
-	if nodes:
-		messages.append(LintMessage("s-069", "[xhtml][/] element missing direct child [xhtml]
    [/] or [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename))
+	OUTPUTS
+	A list of LintMessages representing XHTML errors found in the file
+	"""
 
-	# Check for h# without semantics; h# that have child elements (which likely have the correct semantics) are OK
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(@epub:type)][not(./*[not(name()='a' and contains(@epub:type, 'noteref'))])]")
-	if nodes:
-		messages.append(LintMessage("s-070", "[xhtml][/] element without semantic inflection.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	messages = []
 
-	# Check for sectioning elements with more than one heading element
-	nodes = dom.xpath("/html/body//*[name()='article' or name()='section'][count(./*[name()='header' or name()='hgroup' or re:test(name(), '^h[1-6]$')]) > 1]")
+	# Check for uppercase letters in IDs or classes
+	nodes = dom.xpath("//*[re:test(@id, '[A-Z]') or re:test(@class, '[A-Z]') or re:test(@epub:type, '[A-Z]')]")
 	if nodes:
-		messages.append(LintMessage("s-071", "Sectioning element with more than one heading element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("x-002", "Uppercase in attribute value. Attribute values must be all lowercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for some elements that only have a  child (including text inline to the parent).
-	# But, exclude such elements that are 
     elements that have poetry-type parents.
-	nodes = dom.xpath("/html/body//*[name() !='p' and not(ancestor-or-self::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][./span[count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	nodes = dom.xpath("//*[re:test(@id, '^[0-9]+')]")
 	if nodes:
-		messages.append(LintMessage("s-072", "Element with single [xhtml][/] child. [xhtml][/] should be removed and its attributes promoted to the parent element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("x-007", "[attr]id[/] attributes starting with a number are illegal XHTML.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for header elements that are missing both label and ordinal semantics
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)\\s+[ixvIXVmcd]+$')]")
-	if nodes:
-		messages.append(LintMessage("s-073", "Header element that requires [val]label[/] and [val]ordinal[/] semantic children.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for double greater-than at the end of a tag
+	matches = regex.findall(r"(>>|>>)", file_contents)
+	if matches:
+		messages.append(LintMessage("x-008", "Elements should end with a single [text]>[/].", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for table headers that don't have content, which is an accessibility issue.
-	# Note that @aria-label and @title should (apparently) not be used for table headers
-	nodes = dom.xpath("/html/body//th[not(text())]")
+	# Check for leading 0 in IDs (note: not the same as checking for IDs that start with an integer)
+	# We only check for *leading* 0s in numbers; this allows IDs like `wind-force-0` in the Worst Journey in the World glossary.
+	nodes = dom.xpath("//*[re:test(@id, '-0[0-9]')]")
 	if nodes:
-		messages.append(LintMessage("s-074", "[xhtml]
    		[/] element with no text content should be a [xhtml][/] element instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("x-009", "Illegal leading 0 in [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for incorrect children of 
-	nodes = dom.xpath("/html/body/*[name() != 'section' and name() != 'article' and name() != 'nav']")
+	# Check for HTML tags in  tags
+	nodes = dom.xpath("/html/head/title/*")
 	if nodes:
-		messages.append(LintMessage("s-075", "[xhtml]<body>[/] element with direct child that is not [xhtml]<section>[/], [xhtml]<article>[/], or [xhtml]<nav>[/].", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("x-010", "Illegal element in [xhtml]<title>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for lang="" instead of xml:lang=""
-	nodes = dom.xpath("/html/body//*[@lang]")
+	# Check for underscores in attributes, but not if the attribute is href (links often have underscores) or MathMl alttext, which can have underscores as part of math notation
+	nodes = dom.xpath("//@*[contains(., '_') and name() !='href' and name() !='alttext']/..")
 	if nodes:
-		messages.append(LintMessage("s-076", "[attr]lang[/] attribute used instead of [attr]xml:lang[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("x-011", "Illegal underscore in attribute. Use dashes instead of underscores.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for <header> preceded by non-sectioning elements
-	nodes = dom.xpath("/html/body//header[./preceding-sibling::*[not(re:test(name(), '^(section|div|article)$'))]]")
+	# Check for style attributes
+	nodes = dom.xpath("/html/body//*[@style]")
 	if nodes:
-		messages.append(LintMessage("s-077", "[xhtml]<header>[/] element preceded by non-sectioning element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("x-012", "Illegal [attr]style[/] attribute. Don’t use inline styles, any element can be targeted with a clever enough selector.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for <footer> followed by non-sectioning elements
-	nodes = dom.xpath("/html/body//footer[./following-sibling::*[not(re:test(name(), '^(section|div|article)$'))]]")
+	# Check for illegal elements in <head>
+	nodes = dom.xpath("/html/head/*[not(self::title) and not(self::link[@rel='stylesheet'])]")
 	if nodes:
-		messages.append(LintMessage("s-078", "[xhtml]<footer>[/] element followed by non-sectioning element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("x-015", "Illegal element in [xhtml]<head>[/]. Only [xhtml]<title>[/] and [xhtml]<link rel=\"stylesheet\">[/] are allowed.", se.MESSAGE_TYPE_ERROR, filename, [f"<{node.tag}>" for node in nodes]))
 
-	# Check for elements with no children and only white space contents
-	nodes = dom.xpath("/html/body//*[not(./*) and re:test(., '^\\s+$')]")
+	# Check for xml:lang attribute starting in uppercase
+	nodes = dom.xpath("//*[re:test(@xml:lang, '^[A-Z]')]")
 	if nodes:
-		messages.append(LintMessage("s-079", "Element containing only white space.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("x-016", "[attr]xml:lang[/] attribute with value starting in uppercase letter.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for a play <td> that contains both inline text and block-level elements.
-	# Note that we check any child of <html> because often <body> has the drama semantic.
-	nodes = dom.xpath("/html//*[contains(@epub:type, 'z3998:drama')]//td[./text()[normalize-space(.)] and ./*[name() = 'p' or name() = 'blockquote' or name() = 'div']]")
-	if nodes:
-		messages.append(LintMessage("s-080", "[xhtml]<td>[/] in drama containing both inline text and a block-level element. All children should either be only text, or only block-level elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for ID attributes of numbered paragraphs (like `p-44`) that are used as refs in endnotes.
+	# Make sure their number is actually their correct sequence number in the text.
+	for unexpected_id in se.formatting.find_unexpected_ids(dom):
+		messages.append(LintMessage("x-019", f"Unexpected value of [attr]id[/] attribute. Expected: [attr]{unexpected_id[1]}[/].", se.MESSAGE_TYPE_ERROR, filename, [unexpected_id[0].to_tag_string()]))
 
-	# Check for <p> tags that follow figure, blockquote, or table, and start with a lowercase letter, but that don't have the `continued` class. Exclude matches whose first child is <cite>, so we don't match things like <p><cite>—Editor</cite>
-	nodes = dom.xpath("/html/body//p[preceding-sibling::*[1][name() = 'figure' or name() = 'blockquote' or name() = 'table'] and re:test(., '^([a-z]|—|…)') and not(contains(@class, 'continued')) and not(./*[1][self::cite])]")
-	if nodes:
-		messages.append(LintMessage("s-081", "[xhtml]<p>[/] preceded by [xhtml]<figure>[/], [xhtml]<blockquote>[/xhtml], or [xhtml]<table>[/], but without [val]continued[/] class.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	return messages
 
-	# Check for language tags transliterated into Latin script but missing `-Latn` suffix
-	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '^(he|ru|el|zh|bn|hi|sa|uk|yi|grc)$') and re:test(., '[a-zA-Z]')]")
-	if nodes:
-		messages.append(LintMessage("s-082", "Element containing Latin script for a non-Latin-script language, but its [attr]xml:lang[/] attribute value is missing the [val]-Latn[/] language tag suffix. Hint: For example Russian transliterated into Latin script would be [val]ru-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+def _lint_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, is_titlepage: bool, special_file: str) -> list:
+	"""
+	Helper function used in self.lint()
+	Process typo checks on an .xhtml file
 
-	# Check for persona <td>s that have child <p> elements
-	nodes = dom.xpath("/html/body//td[contains(@epub:type, 'z3998:persona') and ./p]")
-	if nodes:
-		messages.append(LintMessage("s-083", "[xhtml]<td epub:type=\"z3998:persona\">[/] element with child [xhtml]<p>[/] element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
+	file_contents: The contents of the file being checked
 
-	# Check for incorrect semantics on some common publications
-	nodes = dom.xpath("/html/body//i[@epub:type='se:name.publication.book' and re:test(., '^(The )?(Iliad|Odyssey|Aeneid|Metamorphoses|Beowulf|Divine Comedy|Paradise Lost)$')]")
-	if nodes:
-		messages.append(LintMessage("s-084", "Poem has incorrect semantics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	OUTPUTS
+	A list of LintMessages representing possible typos found in the file
+	"""
 
-	# Check for `Op. Cit.` in endnotes. `Op. Cit.` means "the previous reference" which doesn't make sense
-	# in popup endnotes. But, if the endnote has a book reference, then allow it as it might be referring to that.
-	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote') and (re:test(., '(Loc\\.|Op\\.) Cit\\.', 'i') or re:test(., '\\bl\\.c\\.\\b', 'i')) and not(.//*[contains(@epub:type, 'se:name.publication')])]")
-	if nodes:
-		messages.append(LintMessage("s-086", "[text]Op. Cit.[/] or [text]Loc. Cit.[/] in endnote. Hint: [text]Op. Cit.[/] and [text]Loc. Cit.[/] mean [text]the previous reference[/], which usually doesn’t make sense in a popup endnote. Such references should be expanded.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+	messages = []
 
-	# Check for half title pages missing subtitles
-	if ebook_info["has_subtitle"]:
-		# Make sure we exclude <a> because that appears in the ToC landmarks
-		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and not(.//*[contains(@epub:type, 'subtitle')])]")
-		if nodes:
-			messages.append(LintMessage("s-087", "Subtitle in metadata, but no subtitle in the half title page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	# Check for common typos
+	if not is_titlepage:
+		# Don't check the titlepage because it has a standard format and may raise false positives
+		typos = regex.findall(r"(?<!’)\b(and and|the the|if if|of of|or or|as as)\b(?!-)", file_contents, flags=regex.IGNORECASE)
+		typos = typos + regex.findall(r"\ba a\b(?!-)", file_contents)
 
-	else:
-		# Make sure we exclude <a> because that appears in the ToC landmarks
-		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and .//*[contains(@epub:type, 'subtitle')]]")
-		if nodes:
-			messages.append(LintMessage("s-088", "Subtitle in half title page, but no subtitle in metadata.", se.MESSAGE_TYPE_ERROR, filename))
+		if typos:
+			messages.append(LintMessage("y-001", "Possible typo: doubled [text]a/the/and/of/or/as/if[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for MathML without alttext
-	nodes = dom.xpath("/html/body//m:math[not(@alttext)]")
-	if nodes:
-		messages.append(LintMessage("s-089", "MathML missing [attr]alttext[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for basic typo, but exclude MathML and endnotes. Some endnotes (like in Ten Days that Shook the World) may start with letters.
+	dom_copy = deepcopy(dom)
+	for node in dom_copy.xpath("//a[contains(@epub:type, 'noteref')]") + dom_copy.xpath("//*[namespace-uri() = 'http://www.w3.org/1998/Math/MathML']"):
+		node.remove()
 
-	# Check for common errors in language tags
-	# `gr` is often used instead of `el`, `sp` instead of `es`, and `ge` instead of `de` (`ge` is the Georgian geographic region subtag but not a language subtag itself)
-	nodes = dom.xpath("//*[re:test(@xml:lang, '^(gr|sp|ge)$')]")
-	if nodes:
-		messages.append(LintMessage("s-090", "Invalid language tag.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	typos = [node.to_string() for node in dom_copy.xpath("/html/body//p[re:match(., '[a-z][;,][a-z]', 'i')]")]
+	if typos:
+		messages.append(LintMessage("y-002", "Possible typo: punctuation followed directly by a letter, without a space.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for <span> in poetry not followed by <br/>. Ignore spans that are roman as they might be present in poem headers, or spans with xml:lang.
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]//p/span[not(@epub:type='z3998:roman') and not(@xml:lang) and following-sibling::*[1][name()='span']]")
-	if nodes:
-		messages.append(LintMessage("s-091", "[xhtml]<span>[/] not followed by [xhtml]<br/>[/] in poetry.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for paragraphs ending in lowercase letters, excluding a number of instances where that might be valid.
+	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '[a-z]$') and not(@epub:type) and not(following-sibling::*[1]/node()[1][contains(@epub:type, 'z3998:roman')] or following-sibling::*[1][re:test(., '^([0-9]|first|second|third|fourth|fifth|sixth|seventh|eight|ninth|tenth)', 'i')]) and not(@class or contains(@epub:type, 'z3998:salutation')) and not(following-sibling::*[1][name() = 'blockquote' or name() = 'figure' or name() = 'table' or name() = 'footer' or name() = 'ul' or name() = 'ol'] or ancestor::*[name() = 'blockquote' or name() = 'footer' or name() = 'header' or name() = 'table' or name() = 'ul' or name() = 'ol' or name() = 'li' or name() = 'figure' or name() = 'hgroup' or re:test(@epub:type, '(z3998:drama|z3998:letter|dedication|halftitlepage)')])]")]
+	if typos:
+		messages.append(LintMessage("y-003", "Possible typo: paragraph missing ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for nested `<abbr>` elements
-	nodes = dom.xpath("/html/body//abbr[./abbr]")
-	if nodes:
-		messages.append(LintMessage("s-093", "Nested [xhtml]<abbr>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '-[‘“]')]")]
+	if typos:
+		messages.append(LintMessage("y-004", "Possible typo: mis-curled quotation mark after dash.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for language tags transliterated into Latin script with incorrect `-latn` suffix (lowercase l)
-	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '-latn')]")
-	if nodes:
-		messages.append(LintMessage("s-094", "Element has an [attr]xml:lang[/] attribute that incorrectly contains [val]-latn[/] instead of [val]-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for commas or periods following exclamation or question marks. Exclude the colophon because we may have painting names with punctuation.
+	if special_file != "colophon":
+		typos = [node.to_string() for node in dom.xpath("//p[re:test(., '[!?][\\.,][^”’]')]")]
+		if typos:
+			messages.append(LintMessage("y-005", "Possible typo: question mark or exclamation mark followed by period or comma.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for poetry/verse that has an hgroup but which does not have the correct text alignment
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(hymn|poem|song|verse)')]//hgroup/p[@data-css-text-align != 'center']")
-	if nodes:
-		messages.append(LintMessage("s-095", "[xhtml]<p>[/] child of [xhtml]<hgroup>[/] in poetry/verse does not have [css]text-align: center;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for opening lsquo in quote that doesn't appear to have a matching rsquo. Rsquos must be followed by a space, punctuation (except period/comma), word joiner and thus em dash, or a number for an endnote.
+	typos = dom.xpath(f"re:match(//*, '“\\s*‘[^“]+?”', 'g')/text()[not(re:test(., '’[\\s\\?\\!;<2060{se.WORD_JOINER}0-9]'))]")
+	if typos:
+		messages.append(LintMessage("y-006", "Possible typo: [text]‘[/] without matching [text]’[/]. Hints: [text]’[/] are used for abbreviations;  commas and periods must go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Try to find top-level lsquo; for example, <p>“Bah!” he said to the ‘minister.’</p>
+	# We can't do this on xpath because we can't iterate over the output of re:replace().
+	# A fully-featured solution would parse quotes to find balanced pairs. This quick-n-dirty
+	# solution takes the raw XHTML string, strips any seemingly-valid balanced pairs using regex,
+	# then searches for unmatched `‘`. If found, it returns the dom node of the match; we can't return
+	# the actual string match because we stripped surrounding text.
 
-	# Check that half titles without subtitles have the fulltitle semantic
-	nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage')]/*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'fulltitle'))]")
-	if nodes:
-		messages.append(LintMessage("s-096", "Heading element in half title page missing the [val]fulltitle[/] semantic.", se.MESSAGE_TYPE_ERROR, filename))
+	# Start by merging all <p> into single lines, including poetry, but exclude blockquotes as they have special quoting rules
+	temp_xhtml = ""
+	dom_copy = deepcopy(dom)
+	node_number = 0
+	# Add IDs to the dom so we can find the paragraphs after we run our regex
+	# Exclude paragraphs in blockquotes, which may have special quoting rules, and "continued" paragraphs, which may be continued dialog without an “
+	for node in dom_copy.xpath("/html/body//p[not(ancestor::blockquote) and not(contains(@class, 'continued'))]"):
+		node.set_attr("id", "lint-" + str(node_number))
+		temp_xhtml = temp_xhtml + f"<p id=\"lint-{node_number}\">" + regex.sub(r"[\s\n]+", " ", node.inner_text(), flags=regex.DOTALL) + "\n"
+		node_number = node_number + 1
 
-	# Check for missing href attributes, sometimes a leftover from PG transcriptions
-	nodes = dom.xpath("/html/body//a[not(@href)]")
-	if nodes:
-		messages.append(LintMessage("s-097", "[xhtml]a[/] element without [attr]href[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	replacement_count = 1
+	while replacement_count > 0:
+		# Remove all regular quotes. Run this in a loop because we may need to remove triple-nested quotes.
+		(temp_xhtml, replacement_count) = regex.subn(r"“[^“]+?”", " ", temp_xhtml) # Remove all regular quotes
 
-	# Check for <headers>s with only one child
-	nodes = dom.xpath("/html/body/*[name() = 'section' or name() = 'article']/header[./*[not(name() = 'p') and not(preceding-sibling::* or following-sibling::*)]]")
-	if nodes:
-		messages.append(LintMessage("s-098", "[xhtml]<header>[/] element with only one child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Remove contractions to reduce rsquo for next regex
+	temp_xhtml = regex.sub(r"[\p{Letter}]’[\p{Letter}]", " ", temp_xhtml, flags=regex.MULTILINE)
 
-	# Check for @lang attributes
-	nodes = dom.xpath("//*[@lang]")
-	if nodes:
-		messages.append(LintMessage("s-102", "[attr]lang[/] attribute detected. Hint: Use [attr]xml:lang[/] instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Remove all runs of ldquo that are likely to spill to the next <p>
+	replacement_count = 1
+	while replacement_count > 0:
+		(temp_xhtml, replacement_count) = regex.subn(r"“[^“”]+?$", " ", temp_xhtml, flags=regex.MULTILINE)
 
-	# Check for common missing roman semantics for “I”
-	regent_regex = r"(?:Charles|Edward|George|Henry|James|William) I\b"
-	matches = regex.findall(fr"King {regent_regex}", file_contents) + regex.findall(fr"{regent_regex}’s", file_contents)
-	if matches:
-		messages.append(LintMessage("s-103", "Probable missing semantics for a roman I numeral.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Match problem `‘` using regex, and if found, get the actual node text from the dom to return.
+	typos = []
+	for match in regex.findall(r"""<p id="lint-([0-9]+?)">.*‘[\p{Letter}\s]""", temp_xhtml):
+		for node in dom_copy.xpath(f"/html/body//p[@id='lint-{match}' and re:test(., '‘[A-Za-z\\s]')]"):
+			typos.append(node.inner_text())
 
-	return messages
+	if typos:
+		messages.append(LintMessage("y-007", "Possible typo: [text]‘[/] not within [text]“[/]. Hints: Should [text]‘[/] be replaced with [text]“[/]? Is there a missing closing quote? Is this a nested quote that should be preceded by [text]“[/]? Are quotes in close proximity correctly closed?", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-def _lint_metadata_checks(filename: Path, dom: se.easy_xml.EasyXmlTree) -> list:
-	"""
-	Helper function used in self.lint()
-	Process metadata checks on an .xhtml file
+	# Check for single quotes when there should be double quotes in an interjection in dialog
+	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '“[^”]+?’⁠—[^”]+?—“')]")]
+	if typos:
+		messages.append(LintMessage("y-008", "Possible typo: dialog interrupted by interjection but with incorrect closing quote.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	INPUTS
-	filename: The name of the file being checked
-	dom: A dom tree to check
+	# Check for dialog starting with a lowercase letter. Only check the first child text node of <p>, because other first children might be valid lowercase, like <m:math> or <b>;
+	# exclude <p> inside or preceded by <blockquote>; and exclude <p> inside endnotes, as definitions may start with lowercase letters.
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::li[contains(@epub:type, 'endnote')]) and not(preceding-sibling::*[1][name()='blockquote'])][re:test(./node()[1], '^“[a-z]')]")]
+	if typos:
+		messages.append(LintMessage("y-009", "Possible typo: dialog begins with lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	OUTPUTS
-	A list of LintMessages representing metadata errors found in the file
-	"""
+	typos = [node.to_string() for node in dom.xpath("//p[not(ancestor::blockquote) and not(following-sibling::*[1][name() = 'blockquote' or contains(@class, 'continued')]) and re:test(., ',”$')]")]
+	if typos:
+		messages.append(LintMessage("y-010", "Possible typo: comma ending dialogue.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	messages = [];
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '’{2,}')]")]
+	if typos:
+		messages.append(LintMessage("y-011", "Possible typo: two or more [text]’[/] in a row.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for missing MARC relators
-	# Don't check the landmarks as that may introduce duplicate errors
-	# Only check for top-level elements, to avoid intros in short stories or poems in compilation.
-	if not dom.xpath("/html/body//nav[contains(@epub:type, 'landmarks')]"):
-		if dom.xpath("/html/body/*[contains(@epub:type, 'introduction') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and (text()='aui' or text()='win')]"):
-			messages.append(LintMessage("m-030", "[val]introduction[/] semantic inflection found, but no MARC relator [val]aui[/] (Author of introduction, but not the chief author) or [val]win[/] (Writer of introduction).", se.MESSAGE_TYPE_WARNING, filename))
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '”[a-z]')]")]
+	if typos:
+		messages.append(LintMessage("y-012", "Possible typo: [text]”[/] directly followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		if dom.xpath("/html/body/*[contains(@epub:type, 'preface') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='wpr']"):
-			messages.append(LintMessage("m-031", "[val]preface[/] semantic inflection found, but no MARC relator [val]wpr[/] (Writer of preface).", se.MESSAGE_TYPE_WARNING, filename))
+	# Check for period not inside rsquo
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[^”’]+?’[\\.,]”')]")]
+	if typos:
+		messages.append(LintMessage("y-013", "Possible typo: punctuation not within [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		if dom.xpath("/html/body/*[contains(@epub:type, 'afterword') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='aft']"):
-			messages.append(LintMessage("m-032", "[val]afterword[/] semantic inflection found, but no MARC relator [val]aft[/] (Author of colophon, afterword, etc.).", se.MESSAGE_TYPE_WARNING, filename))
+	# Check for period before dialog tag; try to exclude abbrevations that close a quotation, like `“<abbr>Mr.</abbr>”`.
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[(re:test(., '\\.”\\s[a-z\\s]*?(\\bsaid|[a-z]+ed)') or re:test(., '\\.”\\s(s?he|they?|and)\\b')) and not(.//abbr[following-sibling::node()[re:test(., '^”')]])]")]
+	if typos:
+		messages.append(LintMessage("y-014", "Possible typo: Unexpected [text].[/] at the end of quotation. Hint: If a dialog tag follows, should this be [text],[/]?", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		if dom.xpath("/html/body/*[contains(@epub:type, 'endnotes') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='ann']"):
-			messages.append(LintMessage("m-033", "[val]endnotes[/] semantic inflection found, but no MARC relator [val]ann[/] (Annotator).", se.MESSAGE_TYPE_WARNING, filename))
+	# Check for mis-curled ‘ or ‘ without matching ’
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[A-Za-z][^“’]+?”')]")]
+	if typos:
+		messages.append(LintMessage("y-015", "Possible typo: mis-curled [text]‘[/] or missing [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		if dom.xpath("/html/body/*[contains(@epub:type, 'loi') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='ill']"):
-			messages.append(LintMessage("m-034", "[val]loi[/] semantic inflection found, but no MARC relator [val]ill[/] (Illustrator).", se.MESSAGE_TYPE_WARNING, filename))
+	# Check for two periods in a row, almost always a typo for one period or a hellip
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\.\\.[^\\.]')]")]
+	if typos:
+		messages.append(LintMessage("y-016", "Possible typo: consecutive periods ([text]..[/]).", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check that `Internet Archive` and `HathiTrust` are preceded by `the`. They might have an immediate preceding text node that ends in `the`, or they
-	# might be preceded by a white space node, then a <br/>, then a text node ending in `the`.
-	nodes = dom.xpath("/html/body//a[(re:test(text(), '[Hh]athi') and re:test(@href, '^https://[^\"]*?hathitrust.org')) or (re:test(text(), 'Internet Archive') and re:test(@href, 'https://[^\"]*?archive.org'))][(preceding-sibling::node()[2][not(self::br)] and preceding-sibling::node()[1][not(re:test(., '\\sthe $'))]) or (preceding-sibling::node()[2][self::br] and preceding-sibling::node()[3][not(re:test(., '\\sthe$'))]) ]")
-	if nodes:
-		messages.append(LintMessage("m-061", "Link must be preceded by [text]the[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for ldquo followed by space
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '“\\s+[^‘’]')]")]
+	if typos:
+		messages.append(LintMessage("y-017", "Possible typo: [text]“[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	return messages
+	# Check for lsquo followed by space
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘\\s+[^“’]')]")]
+	if typos:
+		messages.append(LintMessage("y-018", "Possible typo: [text]‘[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-def _lint_css_checks(filename: Path, dom: se.easy_xml.EasyXmlTree) -> list:
-	"""
-	Helper function used in self.lint()
-	Process CSS checks on an .xhtml file
+	# Check for closing rdquo without opening ldquo. We ignore blockquotes because they usually have unique quote formatting.
+	# Remove tds in case rdquo means "ditto mark"
+	typos = regex.findall(r"”[^“‘]+?”", regex.sub(r"<td>[”\s]+?(<a .+?epub:type=\"noteref\">.+?</a>)?</td>", "", file_contents), flags=regex.DOTALL)
 
-	INPUTS
-	filename: The name of the file being checked
-	dom: A dom tree to check
+	# We create a filter to try to exclude nested quotations
+	# Remove tags in case they're enclosing punctuation we want to match against at the end of a sentence.
+	typos = [match for match in typos if not regex.search(r"(?:[\.!\?;…—]|”\s)’\s", se.formatting.remove_tags(match))]
 
-	OUTPUTS
-	A list of LintMessages representing CSS errors found in the file
-	"""
+	# Try some additional matches before adding the lint message
+	# Search for <p> tags that have an ending closing quote but no opening quote; but exclude <p>s that are preceded by a <blockquote>
+	# or that have a <blockquote> ancestor, because that may indicate that the opening quote is elsewhere in the quotation.
+	for node in dom.xpath("//p[re:test(., '^[^“]+”') and not(./preceding-sibling::*[1][name() = 'blockquote']) and not(./ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]) and not(./ancestor::blockquote)]"):
+		typos.append(node.to_string()[-20:])
 
-	messages = [];
+	if typos:
+		messages.append(LintMessage("y-019", "Possible typo: [text]”[/] without opening [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Do we have any elements that have specified border color?
-	# `transparent` and `none` are allowed values for border-color
-	if dom.xpath("/html/body//*[attribute::*[re:test(local-name(), 'data-css-border.+?-color') and text() != 'transparent' and text != 'none']]"):
-		messages.append(LintMessage("c-004", "Don’t specify border colors, so that reading systems can adjust for night mode.", se.MESSAGE_TYPE_WARNING, local_css_path, matches))
+	# Check for ,.
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., ',\\.')]")]
+	if typos:
+		messages.append(LintMessage("y-020", "Possible typo: consecutive comma-period ([text],.[/]).", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check that footers have the expected styling
-	# Footers may sometimes be aligned with text-align: center as well as text-align: right
-	# We also ignore text-align on any footers whose only child is a postscript, which is typically left-aligned
-	nodes = dom.xpath("/html/body//footer[not(@data-css-margin-top='1em') or (not(@data-css-text-align) and not(./*[contains(@epub:type, 'z3998:postscript') and not(following-sibling::*) and not(preceding-sibling::*) ]))]")
-	if nodes:
-		messages.append(LintMessage("c-010", "[xhtml]<footer>[/] missing [css]margin-top: 1em; text-align: <value>;[/]. [css]text-align[/] is usually set to [css]right[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for single quotes instead of double quotes
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[not(contains(@class, 'continued') or ancestor::blockquote or ancestor::*[re:test(@epub:type, 'z3998:(verse|song|poem|hymn)')]) and re:test(., '^[^“]+?‘')]")]
+	if typos:
+		messages.append(LintMessage("y-021", "Possible typo: Opening [text]‘[/] without preceding [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for elements that have `text-align: center` but also `text-indent: 1em`
-	nodes = dom.xpath("/html/body//*[@data-css-text-align='center' and @data-css-text-indent='1em']")
-	if nodes:
-		messages.append(LintMessage("c-011", "Element with [css]text-align: center;[/] but [css]text-indent[/] is [css]1em[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for two quotations in one paragraph
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^”]+?”\\s“[^”]+?”$')]")]
+	if typos:
+		messages.append(LintMessage("y-022", "Possible typo: consecutive quotations without intervening text, e.g. [text]“…” “…”[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for sections that don't have header elements, and that do not have margin-top: 8em.
-	# This tries to find the first <section> or <article> whose first child is <p>, and that doesn't
-	# have any ancestor <section>s or <articles>s with heading content, and that doesn't have the correct top margin.
-	# Ignore the titlepage, loi, and dedications as they typically have unique styling.
-	nodes = dom.xpath("/html/body//*[name()='section' or name()='article'][1][./p[1][not(preceding-sibling::*)] and not(re:test(@epub:type, '(titlepage|loi|dedication)')) and not(ancestor::*/*[re:test(name(), '^(h[1-6]|header|hgroup)$')]) and not(@data-css-margin-top='20vh')]")
-	if nodes:
-		messages.append(LintMessage("c-012", "Sectioning element without heading content, and without [css]margin-top: 20vh;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for incorrectly nested quotation marks
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘”]+“') and not(.//br)]")]
+	if typos:
+		messages.append(LintMessage("y-023", "Possible typo: two opening quotation marks in a run. Hint: Nested quotes should switch between [text]“[/] and [text]‘[/]", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for padding and margin not in .5 increments, except for table headers/cells which usually need .25em increments
-	nodes = dom.xpath("/html/body//*[not(re:test(name(), '^(h[1-6]|td|th)$'))][attribute::*[re:test(local-name(), 'data-css-(margin|padding)')][re:test(., '^[0-9]*\\.[^5]')]]")
-	if nodes:
-		messages.append(LintMessage("c-013", "Element with margin or padding not in increments of [css].5em[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for dashes instead of em-dashes
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\s[a-z]+-(the|there|is|and|they|when)\\s')]")]
+	if typos:
+		messages.append(LintMessage("y-024", "Possible typo: dash before [text]the/there/is/and/they/when[/] probably should be em-dash.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for <table> without set margins. We ignore tables that are for drama structuring, and tables that are ancestors of <blockquote> because <blockquote> has its own margin.
-	# We also ignore tables whose preceding sibling sets a bottom margin that is not 0.
-	nodes = dom.xpath("/html/body//table[not(./ancestor-or-self::*[contains(@epub:type, 'z3998:drama')]) and not(./ancestor::blockquote) and not(./preceding-sibling::*[1][@data-css-margin-bottom != '0']) and (not(@data-css-margin-top) or not(@data-css-margin-right) or not(@data-css-margin-bottom) or not(@data-css-margin-left))]")
-	if nodes:
-		messages.append(LintMessage("c-014", "[xhtml]<table>[/] element without explicit margins. Most tables need [css]margin: 1em;[/] or [css]margin: 1em auto 1em auto;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for comma not followed by space but followed by quotation mark
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '[a-z],[“”‘’][a-z]', 'i')]")]
+	if typos:
+		messages.append(LintMessage("y-025", "Possible typo: comma without space followed by quotation mark.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for elements that follow salutations, but that don't have text-indent: 0;
-	# We have to check both <p> elems that are salutations, and also <p> elems that have a first-child inline element that is a salutation, and that does not have following siblings that are senders/valedictions, as that might indicate a letter in a prose context
-	nodes = dom.xpath("/html/body//*[(contains(@epub:type, 'z3998:salutation') or ./preceding-sibling::*[1][contains(@epub:type, 'z3998:salutation') or (name() != 'blockquote' and count(./node()[normalize-space(.)]) = 1 and ./*[contains(@epub:type, 'z3998:salutation')])] or ./*[1][contains(@epub:type, 'z3998:salutation') and not((./following-sibling::node()[1][self::text()]))]) and @data-css-text-indent != '0']")
-	if nodes:
-		messages.append(LintMessage("c-015", "Element after or containing [val]z3998:salutation[/] does not have [css]text-indent: 0;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for punctuation missing before conjunctions. Ignore <p> with an <i> child starting in a conjunction, as those are probably book titles or non-English languages
+	typos = [node.to_string() for node in dom.xpath(f"/html/body//p[not(parent::hgroup) and re:test(., '\\b[a-z]+\\s(But|And|For|Nor|Yet|Or)\\b[^’\\.\\?\\-{se.WORD_JOINER}]') and not(./i[re:test(., '^(But|And|For|Nor|Yet|Or)\\b')])]")]
+	if typos:
+		messages.append(LintMessage("y-026", "Possible typo: no punctuation before conjunction [text]But/And/For/Nor/Yet/Or[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for block-level postscripts that don't have margin-top: 1em. Exclude postscripts that are the first child of blockquote or footer, since blockquotes/footers gives the desired margin.
-	nodes = dom.xpath("/html/body//*[(name() = 'div' or name() = 'p') and contains(@epub:type, 'z3998:postscript') and not((./parent::blockquote or ./parent::footer) and count(./preceding-sibling::*) = 0) and @data-css-margin-top != '1em']")
-	if nodes:
-		messages.append(LintMessage("c-017", "Element with [val]z3998:postscript[/] semantic, but without [css]margin-top: 1em;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for extra closing single quote at the end of dialog
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘]+”\\s*’$')]")]
+	if typos:
+		messages.append(LintMessage("y-027", "Possible typo: Extra [text]’[/] at end of paragraph.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for <p> postscripts that don't have text-indent: 0.
-	nodes = dom.xpath("/html/body//p[contains(@epub:type, 'z3998:postscript') and (not(@data-css-text-align) or @data-css-text-align = 'initial' or @data-css-text-align = 'left') and @data-css-text-indent != '0']")
-	if nodes:
-		messages.append(LintMessage("c-018", "Element with [val]z3998:postscript[/] semantic, but without [css]text-indent: 0;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for `<abbr>` preceded or followed by text. Ignore compass directions followed by `ly`, like S.S.W.ly
+	typos = [node.to_string() for node in dom.xpath("/html/body//abbr[(preceding-sibling::node()[1])[re:test(., '[A-Za-z]$')] or (following-sibling::node()[1])[re:test(., '^[A-Za-z](?<!s\\b)') and not((./preceding-sibling::abbr[1])[contains(@epub:type, 'se:compass')] and re:test(., '^ly\\b'))]]")]
+	if typos:
+		messages.append(LintMessage("y-028", "Possible typo: [xhtml]<abbr>[/] directly preceded or followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for signature semantic without small caps
-	# Ignore signatures that are only capital letters
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:signature') and not(@data-css-font-variant = 'small-caps') and not(@data-css-font-style = 'italic') and not(re:test(., '^[A-Z\\.\\s]+$'))]")
-	if nodes:
-		messages.append(LintMessage("c-019", "Element with [val]z3998:signature[/] semantic, but without [css]font-variant: small-caps;[/] or [css]font-style: italic;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for misapplied italics. Ignore 's' because the plural is too common.
+	typos = [node.to_string() for node in dom.xpath("/html/body//*[(name() = 'i' or name() = 'em') and ./following-sibling::node()[1][re:test(., '^[a-z]\\b', 'i') and not(re:test(., '^s\\b'))]]")]
+	if typos:
+		messages.append(LintMessage("y-029", "Possible typo: Italics followed by a letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for <article>s or <section>s that occur more than once in a file, without page break CSS
-	# Exclude the last element because we don't care if it breaks after the last one.
-	nodes = dom.xpath("/html/body[count(./article) > 1 or count(./section) > 1]/*[(name() = 'article' or name() = 'section') and (not(@data-css-break-after) or @data-css-break-after != 'page') and not(position() = last())]")
-	if nodes:
-		messages.append(LintMessage("c-020", "Multiple [xhtml]<article>[/]s or [xhtml]<section>[/]s in file, but missing [css]break-after: page;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+	# Check for lowercase letters starting quotations after a preceding period
+	typos = dom.xpath("/html/body//p/child::text()[re:test(., '\\.\\s[‘“][a-z]')]")
+	if typos:
+		messages.append(LintMessage("y-030", "Possible typo: Lowercase quotation following a period. Check either that the period should be a comma, or that the quotation should start with a capital.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for elements that are set in italics, and whose italic children don't have font-style: normal.
-	nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and ./*[(name()='i' or name()='em') and @data-css-font-style='italic']]")
+	# Check for missing punctuation in continued quotations
+	# ” said Bob “
+	nodes = dom.xpath("/html/body//p[re:test(., '”\\s(?:said|[A-Za-z]{2,}ed)\\s[A-Za-z]+?(?<!\\bthe)(?<!\\bto)(?<!\\bwith)(?<!\\bfrom)(?<!\\ba\\b)(?<!\\bis)\\s“') or re:test(., '[^\\.]”\\s(\\bhe\\b|\\bshe\\b|I|[A-Z][a-z]+?)\\s(?:said|[A-Za-z]{2,}ed)\\s“') or re:test(., ',” (?:said|[A-Za-z]{2,}ed) [A-Za-z]+? [A-Za-z]+?ly “')]")
 	if nodes:
-		messages.append(LintMessage("c-021", "Element with [css]font-style: italic;[/], but child [xhtml]<i>[/] or [xhtml]<em>[/] does not have [css]font-style: normal;[/]. Hint: Italics within italics are typically set in Roman for contrast; if that’s not the case here, can [xhtml]<i>[/] be removed while still preserving italics and semantic inflection?", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("y-031", "Possible typo: Dialog tag missing punctuation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for tables that appear to be summing/listing numbers, but that are missing font-variant-numeric styling
-	nodes = dom.xpath("/html/body//table[count(.//td[not(following-sibling::*) and re:test(., '^[0-9]+$') and not(@data-css-font-variant-numeric='tabular-nums')]) >= 2]")
+	# Check for italics having epub:type that run in to preceding or following characters
+	# Ignore things like <i>Newspaper</i>s
+	nodes = dom.xpath("/html/body//i[@epub:type and ( (following-sibling::node()[1][re:test(., '^[a-z]', 'i') and not(re:test(., '^(s|es|er)'))]) or preceding-sibling::node()[1][re:test(., '[a-z]$')]) ]")
 	if nodes:
-		messages.append(LintMessage("c-026", "Table that appears to be listing numbers, but without [css]font-variant-numeric: tabular-nums;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("y-032", "Possible typo: Italics running into preceding or following characters.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
 	return messages
 
@@ -3410,7 +3410,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 
 				messages = messages + _lint_metadata_checks(filename, dom)
 
-				messages = messages + _lint_syntax_checks(self, filename, dom, file_contents, ebook_info, language)
+				messages = messages + _lint_syntax_checks(self, filename, dom, file_contents, ebook_info, language, section_tree)
 
 				messages = messages + _lint_typography_checks(filename, dom, file_contents, special_file)
 

From 86c362b0417618fff50c78cc8b22b82ab23a4d49 Mon Sep 17 00:00:00 2001
From: vr8hub <vr_sebooks@letterboxes.org>
Date: Sun, 27 Aug 2023 17:26:25 -0500
Subject: [PATCH 3/9] Include is_titlepage in special_files

---
 se/se_epub_lint.py | 55 +++++++++++++++++++++++-----------------------
 1 file changed, 28 insertions(+), 27 deletions(-)

diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py
index df18019b..f703fc45 100644
--- a/se/se_epub_lint.py
+++ b/se/se_epub_lint.py
@@ -99,6 +99,7 @@
 IGNORED_CLASSES = ["elision", "eoc", "full-page", "continued", "together"]
 BINARY_EXTENSIONS = [".jpg", ".jpeg", ".tif", ".tiff", ".bmp", ".png", ".epub", ".xcf", ".otf", ".jp2"]
 IGNORED_FILENAMES = ["colophon.xhtml", "titlepage.xhtml", "imprint.xhtml", "uncopyright.xhtml", "toc.xhtml", "loi.xhtml"]
+SPECIAL_FILES = ["colophon", "endnotes", "imprint", "loi"]
 
 # These are partly defined in semos://1.0.0/8.10.9.2
 INITIALISM_EXCEPTIONS = ["G", # as in `G-Force`
@@ -808,7 +809,7 @@ def _update_missing_styles(filename: Path, dom: se.easy_xml.EasyXmlTree, local_c
 
 	return missing_styles
 
-def _lint_special_file(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_info: dict, special_file: str) -> list:
+def _lint_special_file_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_info: dict, special_file: str) -> list:
 	"""
 	Process error checks in “special” .xhtml files
 
@@ -827,7 +828,7 @@ def _lint_special_file(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_
 	messages = []
 	source_links = self.metadata_dom.xpath("/package/metadata/dc:source/text()")
 
-	if special_file in = "colophon" or special_file = "imprint":
+	if special_file == "colophon" or special_file == "imprint":
 		if len(source_links) <= 2:
 			# Check that links back to sources are represented correctly
 			nodes = dom.xpath("/html/body//a[@href='https://www.pgdp.net' and text()!='The Online Distributed Proofreading Team']")
@@ -2310,7 +2311,7 @@ def _lint_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_conten
 
 	return messages
 
-def _lint_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, is_titlepage: bool, special_file: str) -> list:
+def _lint_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str) -> list:
 	"""
 	Helper function used in self.lint()
 	Process typo checks on an .xhtml file
@@ -2327,7 +2328,7 @@ def _lint_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_content
 	messages = []
 
 	# Check for common typos
-	if not is_titlepage:
+	if special_file == "titlepage":
 		# Don't check the titlepage because it has a standard format and may raise false positives
 		typos = regex.findall(r"(?<!’)\b(and and|the the|if if|of of|or or|as as)\b(?!-)", file_contents, flags=regex.IGNORECASE)
 		typos = typos + regex.findall(r"\ba a\b(?!-)", file_contents)
@@ -3308,28 +3309,6 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 				# Add to the short story count for later checks
 				short_story_count += len(dom.xpath("/html/body//article[contains(@epub:type, 'se:short-story')]"))
 
-				# Test against word boundaries to not match `halftitlepage`
-				if dom.xpath("/html/body/section[re:test(@epub:type, '\\btitlepage\\b')]"):
-					is_titlepage = True
-					# Check if the <title> element is set correctly, but only if there's no heading content.
-					# If there's heading content, then <title> should match the expected generated value from the heading content.
-					if dom.xpath("/html[not(./body//*[re:test(name(), '^h(group|[1-6])$')]) and ./head/title[text()!='Titlepage']]"):
-						messages.append(LintMessage("s-025", "Titlepage [xhtml]<title>[/] elements must contain exactly: [text]Titlepage[/].", se.MESSAGE_TYPE_ERROR, filename))
-
-				if dom.xpath("/html/body/section[contains(@epub:type, 'colophon')]"):
-					special_file = "colophon"
-				elif dom.xpath("/html/body/section[contains(@epub:type, 'imprint')]"):
-					special_file = "imprint"
-				elif dom.xpath("/html/body/section[contains(@epub:type, 'endnotes')]"):
-					special_file = "endnotes"
-				elif dom.xpath("/html/body/nav[contains(@epub:type, 'loi')]"):
-					special_file = "loi"
-				else:
-					special_file = ""
-
-				if special_file != "":
-					messages = messages + _lint_special_file(self, filename, dom, file_contents, ebook_info, special_file)
-
 				# Check for unused selectors
 				if dom.xpath("/html/head/link[contains(@href, 'local.css')]"):
 					for selector in local_css_selectors:
@@ -3404,6 +3383,28 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 						if glossary_value[1] is False and regex.search(glossary_value[0], source_text, flags=regex.IGNORECASE):
 							glossary_usage[glossary_index] = (glossary_value[0], True)
 
+				# Test against word boundaries to not match `halftitlepage`
+				if dom.xpath("/html/body/section[re:test(@epub:type, '\\btitlepage\\b')]"):
+					special_file = "titlepage"
+					# Check if the <title> element is set correctly, but only if there's no heading content.
+					# If there's heading content, then <title> should match the expected generated value from the heading content.
+					if dom.xpath("/html[not(./body//*[re:test(name(), '^h(group|[1-6])$')]) and ./head/title[text()!='Titlepage']]"):
+						messages.append(LintMessage("s-025", "Titlepage [xhtml]<title>[/] elements must contain exactly: [text]Titlepage[/].", se.MESSAGE_TYPE_ERROR, filename))
+
+				elif dom.xpath("/html/body/section[contains(@epub:type, 'colophon')]"):
+					special_file = "colophon"
+				elif dom.xpath("/html/body/section[contains(@epub:type, 'imprint')]"):
+					special_file = "imprint"
+				elif dom.xpath("/html/body/section[contains(@epub:type, 'endnotes')]"):
+					special_file = "endnotes"
+				elif dom.xpath("/html/body/nav[contains(@epub:type, 'loi')]"):
+					special_file = "loi"
+				else:
+					special_file = ""
+
+				if special_file in SPECIAL_FILES:
+					messages = messages + _lint_special_file_checks(self, filename, dom, file_contents, ebook_info, special_file)
+
 				missing_styles = missing_styles + _update_missing_styles(filename, dom, local_css)
 
 				messages = messages + _lint_css_checks(filename, dom)
@@ -3416,7 +3417,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 
 				messages = messages + _lint_xhtml_checks(filename, dom, file_contents)
 
-				messages = messages + _lint_typo_checks(filename, dom, file_contents, is_titlepage, special_file)
+				messages = messages + _lint_typo_checks(filename, dom, file_contents, special_file)
 
 	if self.cover_path and cover_svg_title != titlepage_svg_title:
 		messages.append(LintMessage("s-028", f"[path][link=file://{self.cover_path}]{self.cover_path.name}[/][/] and [path][link=file://{self.path / 'images/titlepage.svg'}]titlepage.svg[/][/] [xhtml]<title>[/] elements don’t match.", se.MESSAGE_TYPE_ERROR, self.cover_path))

From f6eef16d2432f3d7f41997a9e3eaae20355c00ec Mon Sep 17 00:00:00 2001
From: vr8hub <vr_sebooks@letterboxes.org>
Date: Sun, 27 Aug 2023 17:41:46 -0500
Subject: [PATCH 4/9] Move svg checks to function

---
 se/se_epub_lint.py | 183 +++++++++++++++++++++++++--------------------
 1 file changed, 102 insertions(+), 81 deletions(-)

diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py
index f703fc45..8b47e7f5 100644
--- a/se/se_epub_lint.py
+++ b/se/se_epub_lint.py
@@ -105,7 +105,7 @@
 INITIALISM_EXCEPTIONS = ["G", # as in `G-Force`
 			"1D", "2D", "3D", "4D", # as in `n-dimensional`
 			"MS.", "MSS.", # Manuscript(s)
-			"MM.",  # Messiuers
+			"MM.",	# Messiuers
 			"κ.τ.λ.", # "etc." in Greek, and we don't match Greek chars.
 			"TV",
 			"AC", "DC" # electrical current
@@ -453,7 +453,7 @@
 "y-003”, "Possible typo: paragraph missing ending punctuation."
 "y-004”, "Possible typo: mis-curled quotation mark after dash."
 "y-005”, "Possible typo: question mark or exclamation mark followed by period or comma."
-"y-006”, "Possible typo: [text]‘[/] without matching [text]’[/]. Hints: [text]’[/] are used for abbreviations;  commas and periods must go inside quotation marks."
+"y-006”, "Possible typo: [text]‘[/] without matching [text]’[/]. Hints: [text]’[/] are used for abbreviations;	commas and periods must go inside quotation marks."
 "y-007”, "Possible typo: [text]‘[/] not within [text]“[/]. Hints: Should [text]‘[/] be replaced with [text]“[/]? Is there a missing closing quote? Is this a nested quote that should be preceded by [text]“[/]? Are quotes in close proximity correctly closed?"
 "y-008”, "Possible typo: dialog interrupted by interjection but with incorrect closing quote."
 "y-009”, "Possible typo: dialog begins with lowercase letter."
@@ -549,11 +549,11 @@ def _build_section_tree(self) -> List[EbookSection]:
 	#  preface (2)
 	#  halftitlepage (2)
 	#  part-1 (2)
-	#    division-1-1 (3)
-	#      chapter-1-1-1 (4)
-	#      chapter-1-1-2 (4)
-	#      chapter-1-1-3 (4)
-	#      chapter-1-1-4 (4)
+	#	 division-1-1 (3)
+	#	   chapter-1-1-1 (4)
+	#	   chapter-1-1-2 (4)
+	#	   chapter-1-1-3 (4)
+	#	   chapter-1-1-4 (4)
 	section_tree: List[EbookSection] = []
 
 	for filename in self.spine_file_paths:
@@ -771,7 +771,7 @@ def _update_missing_styles(filename: Path, dom: se.easy_xml.EasyXmlTree, local_c
 	local_css: Dictionary containing several flags concerning the local CSS
 
 	OUTPUTS
-	Set of files not in the spine (typically an empty set)
+	List of styles used in this file but missing from local CSS
 	"""
 
 	missing_styles: List[str] = []
@@ -809,6 +809,84 @@ def _update_missing_styles(filename: Path, dom: se.easy_xml.EasyXmlTree, local_c
 
 	return missing_styles
 
+def _lint_svg_checks(self, filename: Path, file_contents: str, svg_dom: se.easy_xml.EasyXmlTree, root: str) -> list:
+	"""
+	Perform several checks on svg files
+
+	INPUTS
+	filename: The name of the svg file being checked
+	file_contents: The contents of the svg file being checked
+	svg_dom: The dom of the svg file being checked
+	self
+	root
+
+	OUTPUTS
+	A list of LintMessages representing errors found in the svg file
+	"""
+
+	messages = []
+
+	# Make images have reasonable dimensions
+	viewbox = svg_dom.xpath("/svg/@viewBox", True)
+	if viewbox:
+		svg_dimensions = viewbox.split()
+		try:
+			if float(svg_dimensions[2]) * float(svg_dimensions[3]) > 4000000:
+				messages.append(LintMessage("f-018", "Image greater than 4,000,000 pixels square in dimension.", se.MESSAGE_TYPE_ERROR, filename))
+		except Exception as ex:
+			raise se.InvalidFileException(f"Couldn’t parse SVG [xhtml]viewBox[/] attribute in [path][link=file://{filename.resolve()}]{filename}[/][/].") from ex
+
+	# If we're looking at the cover image, ensure that the producer has built it.
+	# The default cover image is a white background which when encoded to base64 begins with 299 characters and then has a long string of `A`s
+	if self.cover_path and filename.name == self.cover_path.name and svg_dom.xpath("//image[re:test(@xlink:href, '^.{299}A{50,}')]"):
+		messages.append(LintMessage("m-063", "Cover image has not been built.", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Check for illegal transform or id attribute
+	nodes = svg_dom.xpath("//*[@transform or @id]")
+	if nodes:
+		invalid_transform_attributes = set()
+		invalid_id_attributes = []
+		for node in nodes:
+			if node.get_attr("transform"):
+				invalid_transform_attributes.add(f"transform=\"{node.get_attr('transform')}\"")
+
+			if node.get_attr("id"):
+				invalid_id_attributes.append(f"id=\"{node.get_attr('id')}\"")
+
+		if invalid_transform_attributes:
+			messages.append(LintMessage("x-003", "Illegal [xml]transform[/] attribute. SVGs should be optimized to remove use of [xml]transform[/]. Try using Inkscape to save as an “optimized SVG”.", se.MESSAGE_TYPE_ERROR, filename, invalid_transform_attributes))
+
+	# Check for fill: #000 which should simply be removed
+	nodes = svg_dom.xpath("//*[contains(@fill, '#000') or contains(translate(@style, ' ', ''), 'fill:#000')]")
+	if nodes:
+		messages.append(LintMessage("x-004", "Illegal [xml]style=\"fill: #000\"[/] or [xml]fill=\"#000\"[/].", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Check for illegal height or width on root <svg> element
+	if filename.name != "logo.svg": # Do as I say, not as I do...
+		if svg_dom.xpath("/svg[@height or @width]"):
+			messages.append(LintMessage("x-005", "Illegal [xml]height[/] or [xml]width[/] attribute on root [xml]<svg>[/] element. Size SVGs using the [xml]viewBox[/] attribute only.", se.MESSAGE_TYPE_ERROR, filename))
+
+	match = regex.search(r"viewbox", file_contents, flags=regex.IGNORECASE)
+	if match and match[0] != "viewBox":
+		messages.append(LintMessage("x-006", f"[xml]{match}[/] found instead of [xml]viewBox[/]. [xml]viewBox[/] must be correctly capitalized.", se.MESSAGE_TYPE_ERROR, filename))
+
+		if invalid_id_attributes:
+			messages.append(LintMessage("x-014", "Illegal [xml]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, invalid_id_attributes))
+
+	if f"{os.sep}src{os.sep}" not in root:
+		# Check that cover and titlepage images are in all caps
+		if self.cover_path and filename.name == self.cover_path.name:
+			nodes = svg_dom.xpath("//text[re:test(., '[a-z]')]")
+			if nodes:
+				messages.append(LintMessage("s-002", "Lowercase letters in cover. Cover text must be all uppercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+
+		if filename.name == "titlepage.svg":
+			nodes = svg_dom.xpath("//text[re:test(., '[a-z]') and not(text()='translated by' or text()='illustrated by' or text()='and')]")
+			if nodes:
+				messages.append(LintMessage("s-003", "Lowercase letters in titlepage. Titlepage text must be all uppercase except [text]translated by[/] and [text]illustrated by[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+
+	return messages
+
 def _lint_special_file_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_info: dict, special_file: str) -> list:
 	"""
 	Process error checks in “special” .xhtml files
@@ -1541,7 +1619,7 @@ def _lint_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file
 	if nodes:
 		messages.append(LintMessage("s-054", "[xhtml]<cite>[/] as child of [xhtml]<p>[/] in [xhtml]<blockquote>[/]. [xhtml]<cite>[/] should be the direct child of [xhtml]<blockquote>[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for <th> element without a <thead> ancestor. However, <th scope="...">  and <th/> are allowed, for use in tables with headers in the middle of tables (url:https://standardebooks.org/ebooks/dorothy-day/the-eleventh-virgin) and vertical table headers
+	# Check for <th> element without a <thead> ancestor. However, <th scope="...">	and <th/> are allowed, for use in tables with headers in the middle of tables (url:https://standardebooks.org/ebooks/dorothy-day/the-eleventh-virgin) and vertical table headers
 	# (https://standardebooks.org/ebooks/charles-babbage/passages-from-the-life-of-a-philosopher)
 	if dom.xpath("/html/body//table//th[not(ancestor::thead)][not(@scope)][not(count(node())=0)]"):
 		messages.append(LintMessage("s-055", "[xhtml]<th>[/] element not in [xhtml]<thead>[/] ancestor. Note: [xhtml]<th>[/] elements used as mid-table headings or horizontal row headings require the [attr]scope[/] attribute.", se.MESSAGE_TYPE_ERROR, filename))
@@ -1814,7 +1892,7 @@ def _lint_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_c
 
 	# Check for repeated punctuation, but first remove `&` so we don't match `&,`
 	# Remove tds with repeated ” as they are probably ditto marks
-	matches = regex.findall(r"[,;]{2,}.{0,20}", file_contents.replace("&", "")) + regex.findall(r"(?:“\s*“|”\s*”|’ ’|‘\s*‘).{0,20}", regex.sub(r"<td>[”\s]+?(<a .+?epub:type=\"noteref\">.+?</a>)?</td>", "", file_contents)) +  regex.findall(r"[\p{Letter}][,\.:;]\s[,\.:;]\s?[\p{Letter}<].{0,20}", file_contents, flags=regex.IGNORECASE)
+	matches = regex.findall(r"[,;]{2,}.{0,20}", file_contents.replace("&", "")) + regex.findall(r"(?:“\s*“|”\s*”|’ ’|‘\s*‘).{0,20}", regex.sub(r"<td>[”\s]+?(<a .+?epub:type=\"noteref\">.+?</a>)?</td>", "", file_contents)) +	 regex.findall(r"[\p{Letter}][,\.:;]\s[,\.:;]\s?[\p{Letter}<].{0,20}", file_contents, flags=regex.IGNORECASE)
 	if matches:
 		messages.append(LintMessage("t-008", "Repeated punctuation.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
@@ -2044,7 +2122,7 @@ def _lint_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_c
 		messages.append(LintMessage("t-035", "[xhtml]<cite>[/] element not preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
 	# Check for partially obscured years in which the last year is an em dash
-	nodes = dom.xpath("/html/body//p[re:test(.,  '1\\d{2}⁠—[^“’A-Za-z<]')]")
+	nodes = dom.xpath("/html/body//p[re:test(.,	 '1\\d{2}⁠—[^“’A-Za-z<]')]")
 	if nodes:
 		messages.append(LintMessage("t-036", "Em-dash used to obscure single digit in year. Hint: Use a hyphen instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
@@ -2120,7 +2198,7 @@ def _lint_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_c
 	if nodes:
 		messages.append(LintMessage("t-052", "Stage direction without ending punctuation. Note that ending punctuation is optional in stage direction that is an interjection in a parent clause, and such interjections should begin with a lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for stage direction starting in lowercase.  We only want to consider stage direction that is not an interjection in a parent clause.
+	# Check for stage direction starting in lowercase.	We only want to consider stage direction that is not an interjection in a parent clause.
 	# We match if the stage direction starts with a lowercase, and if there's no preceding node, or if there is a preceding node and it doesn't end in a lowercase letter or a small subset of conjoining punctuation.
 	nodes = dom.xpath("/html/body//i[@epub:type='z3998:stage-direction' and re:test(., '^[a-z]') and (not(./preceding-sibling::node()[normalize-space(.)]) or ./preceding-sibling::node()[1][re:test(normalize-space(.), '[^a-z:—,;…]$')])]")
 	if nodes:
@@ -2363,7 +2441,7 @@ def _lint_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_content
 	# Check for opening lsquo in quote that doesn't appear to have a matching rsquo. Rsquos must be followed by a space, punctuation (except period/comma), word joiner and thus em dash, or a number for an endnote.
 	typos = dom.xpath(f"re:match(//*, '“\\s*‘[^“]+?”', 'g')/text()[not(re:test(., '’[\\s\\?\\!;<2060{se.WORD_JOINER}0-9]'))]")
 	if typos:
-		messages.append(LintMessage("y-006", "Possible typo: [text]‘[/] without matching [text]’[/]. Hints: [text]’[/] are used for abbreviations;  commas and periods must go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, typos))
+		messages.append(LintMessage("y-006", "Possible typo: [text]‘[/] without matching [text]’[/]. Hints: [text]’[/] are used for abbreviations;	commas and periods must go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
 	# Try to find top-level lsquo; for example, <p>“Bah!” he said to the ‘minister.’</p>
 	# We can't do this on xpath because we can't iterate over the output of re:replace().
@@ -3117,11 +3195,11 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 	section_tree: List[EbookSection] = _build_section_tree(self)
 	# This block is useful for pretty-printing section_tree should we need to debug it in the future
 	# def dump(item, char):
-	# 	print(f"{char} {item.section_id} ({item.depth}) {item.has_header}")
-	# 	for child in item.children:
-	# 		dump(child, f"  {char}")
+	#	print(f"{char} {item.section_id} ({item.depth}) {item.has_header}")
+	#	for child in item.children:
+	#		dump(child, f"	{char}")
 	# for section in section_tree:
-	# 	dump(section, "")
+	#	dump(section, "")
 	# exit()
 
 	# Now iterate over individual files for some checks
@@ -3207,71 +3285,14 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 
 			if filename.suffix == ".svg":
 				svg_dom = self.get_dom(filename)
+				messages = messages + _lint_svg_checks(self, filename, file_contents, svg_dom, root)
+				if self.cover_path and filename.name == self.cover_path.name:
+					# For later comparison with titlepage
+					cover_svg_title = svg_dom.xpath("/svg/title/text()", True).replace("The cover for ", "") # <title> can appear on any element in SVG, but we only want to check the root one
+				elif filename.name == "titlepage.svg":
+					# For later comparison with cover
+					titlepage_svg_title = svg_dom.xpath("/svg/title/text()", True).replace("The titlepage for ", "") # <title> can appear on any element in SVG, but we only want to check the root one
 
-				# Make images have reasonable dimensions
-				viewbox = svg_dom.xpath("/svg/@viewBox", True)
-				if viewbox:
-					svg_dimensions = viewbox.split()
-					try:
-						if float(svg_dimensions[2]) * float(svg_dimensions[3]) > 4000000:
-							messages.append(LintMessage("f-018", "Image greater than 4,000,000 pixels square in dimension.", se.MESSAGE_TYPE_ERROR, filename))
-					except Exception as ex:
-						raise se.InvalidFileException(f"Couldn’t parse SVG [xhtml]viewBox[/] attribute in [path][link=file://{filename.resolve()}]{filename}[/][/].") from ex
-
-				# If we're looking at the cover image, ensure that the producer has built it.
-				# The default cover image is a white background which when encoded to base64 begins with 299 characters and then has a long string of `A`s
-				if self.cover_path and filename.name == self.cover_path.name and svg_dom.xpath("//image[re:test(@xlink:href, '^.{299}A{50,}')]"):
-					messages.append(LintMessage("m-063", "Cover image has not been built.", se.MESSAGE_TYPE_ERROR, filename))
-
-				# Check for illegal transform or id attribute
-				nodes = svg_dom.xpath("//*[@transform or @id]")
-				if nodes:
-					invalid_transform_attributes = set()
-					invalid_id_attributes = []
-					for node in nodes:
-						if node.get_attr("transform"):
-							invalid_transform_attributes.add(f"transform=\"{node.get_attr('transform')}\"")
-
-						if node.get_attr("id"):
-							invalid_id_attributes.append(f"id=\"{node.get_attr('id')}\"")
-
-					if invalid_transform_attributes:
-						messages.append(LintMessage("x-003", "Illegal [xml]transform[/] attribute. SVGs should be optimized to remove use of [xml]transform[/]. Try using Inkscape to save as an “optimized SVG”.", se.MESSAGE_TYPE_ERROR, filename, invalid_transform_attributes))
-
-				# Check for fill: #000 which should simply be removed
-				nodes = svg_dom.xpath("//*[contains(@fill, '#000') or contains(translate(@style, ' ', ''), 'fill:#000')]")
-				if nodes:
-					messages.append(LintMessage("x-004", "Illegal [xml]style=\"fill: #000\"[/] or [xml]fill=\"#000\"[/].", se.MESSAGE_TYPE_ERROR, filename))
-
-				# Check for illegal height or width on root <svg> element
-				if filename.name != "logo.svg": # Do as I say, not as I do...
-					if svg_dom.xpath("/svg[@height or @width]"):
-						messages.append(LintMessage("x-005", "Illegal [xml]height[/] or [xml]width[/] attribute on root [xml]<svg>[/] element. Size SVGs using the [xml]viewBox[/] attribute only.", se.MESSAGE_TYPE_ERROR, filename))
-
-				match = regex.search(r"viewbox", file_contents, flags=regex.IGNORECASE)
-				if match and match[0] != "viewBox":
-					messages.append(LintMessage("x-006", f"[xml]{match}[/] found instead of [xml]viewBox[/]. [xml]viewBox[/] must be correctly capitalized.", se.MESSAGE_TYPE_ERROR, filename))
-
-					if invalid_id_attributes:
-						messages.append(LintMessage("x-014", "Illegal [xml]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, invalid_id_attributes))
-
-				if f"{os.sep}src{os.sep}" not in root:
-					# Check that cover and titlepage images are in all caps
-					if self.cover_path and filename.name == self.cover_path.name:
-						nodes = svg_dom.xpath("//text[re:test(., '[a-z]')]")
-						if nodes:
-							messages.append(LintMessage("s-002", "Lowercase letters in cover. Cover text must be all uppercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-						# For later comparison with titlepage
-						cover_svg_title = svg_dom.xpath("/svg/title/text()", True).replace("The cover for ", "") # <title> can appear on any element in SVG, but we only want to check the root one
-
-					if filename.name == "titlepage.svg":
-						nodes = svg_dom.xpath("//text[re:test(., '[a-z]') and not(text()='translated by' or text()='illustrated by' or text()='and')]")
-						if nodes:
-							messages.append(LintMessage("s-003", "Lowercase letters in titlepage. Titlepage text must be all uppercase except [text]translated by[/] and [text]illustrated by[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-						# For later comparison with cover
-						titlepage_svg_title = svg_dom.xpath("/svg/title/text()", True).replace("The titlepage for ", "") # <title> can appear on any element in SVG, but we only want to check the root one
 
 			if filename.suffix == ".xml":
 				xml_dom = self.get_dom(filename)

From 703eda020cbf7189a152e2f986e68db639e9d433 Mon Sep 17 00:00:00 2001
From: vr8hub <vr_sebooks@letterboxes.org>
Date: Sun, 27 Aug 2023 23:06:33 -0500
Subject: [PATCH 5/9] Move ignore file processing to function

---
 se/se_epub_lint.py | 4187 ++++++++++++++++++++++----------------------
 1 file changed, 2130 insertions(+), 2057 deletions(-)

diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py
index 8b47e7f5..b1b254d9 100644
--- a/se/se_epub_lint.py
+++ b/se/se_epub_lint.py
@@ -607,2552 +607,2702 @@ def _find_ebook_section(section_id: str, sections: List[EbookSection]) -> Union[
 
 	return None
 
-def _get_malformed_urls(dom: se.easy_xml.EasyXmlTree, filename: Path) -> list:
+def _lint_metadata_checks(self):
 	"""
-	Helper function used in self.lint()
-	Get a list of URLs in the epub that don't match SE standards.
+	Process main metadata checks
 
 	INPUTS
-	dom: A dom tree to check
+	self
 
 	OUTPUTS
-	A list of LintMessages representing any malformed URLs in the XHTML string
+	A list of LintMessage objects
 	"""
 
 	messages = []
+	metadata_xml = self.metadata_dom.to_string()
+	missing_metadata_elements = []
 
-	# Check for non-https URLs
-	search_regex = r"(?<!www\.)gutenberg\.org"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-001", "gutenberg.org URL missing leading [text]www.[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-	search_regex = r"www\.archive\.org"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	# Check for non-English Wikipedia URLs
+	nodes = self.metadata_dom.xpath("/package/metadata/*[re:test(., 'https://(?!en)[a-z]{2,}\\.wikipedia\\.org')]")
 	if nodes:
-		messages.append(LintMessage("m-002", "archive.org URL should not have leading [text]www.[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("m-043", "Non-English Wikipedia URL.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
 
-	search_regex = r"www\.gutenberg\.net\.au"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-076", "gutenberg.net.au URL should not have leading [text]www.[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check the long description for some errors
+	try:
+		# Check if there are non-typogrified quotes or em-dashes in metadata descriptions
+		# lxml unescapes this for us
+		# Also, remove HTML elements like <a href> so that we don't catch quotation marks in attribute values
+		long_description = self.metadata_dom.xpath("/package/metadata/meta[@property='se:long-description']")[0].text
+		matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", regex.sub(r"<[^<]+?>", "", long_description))
+		if matches:
+			messages.append(LintMessage("m-014", "Non-typogrified character in [xml]<meta property=\"se:long-description\">[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
 
-	search_regex = r"http://(gutenberg\.org|gutenberg\.net\.au|gutenberg\.ca|www\.fadedpage\.com|archive\.org|pgdp\.net|catalog\.hathitrust\.org|en\.wikipedia\.org|standardebooks\.org)"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-003", "Non-HTTPS URL.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# Is the first instance of the author's last name a hyperlink in the metadata?
+		authors = self.metadata_dom.xpath("/package/metadata/dc:creator")
+		for author in authors:
+			author_sort = self.metadata_dom.xpath(f"/package/metadata/meta[@property='file-as'][@refines='#{author.get_attr('id')}']/text()")
+			if author_sort:
+				author_last_name = regex.sub(r",.+$", "", author_sort[0])
+				author_last_name = author_last_name.replace("'", "’") # Typogrify apostrophes so that we correctly match in the long description
+				# We can't use xpath here because the long description is escaped; it has no dom to query against.
+				if author_last_name.lower() in regex.sub(r"https://standardebooks\.org/ebooks/[^/]+?/", "", long_description.lower()) and not regex.search(fr"<a href=\"https://standardebooks\.org/ebooks/.+?\">.*?{author_last_name}.*?</a>", long_description, flags=regex.IGNORECASE):
+					messages.append(LintMessage("m-056", "Author name present in [xml]<meta property=\"se:long-description\">[/] element, but the first instance of their name is not hyperlinked to their S.E. author page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	search_regex = r"https?://books\.google\.com/books\?id=.+?[&#]"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-004", "Non-canonical Google Books URL. Google Books URLs must look exactly like [url]https://books.google.com/books?id=<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# Did we mention an SE book in the long description, but without italics?
+		# Only match if the title appears to contain an uppercase letter. This prevents matches on a non-title link like `<a href>short stories</a>`
+		matches = regex.search(r"""(?<!<i>)<a href="https://standardebooks\.org/ebooks/[^"]+?/[^"]+?">(?!<i>)([\p{Letter}\s]+)""", long_description)
+		if matches and regex.search(r"[\p{Uppercase_Letter}]", matches[1]):
+			messages.append(LintMessage("m-064", "S.E. ebook hyperlinked in long description but not italicized.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	search_regex = r"https?://www\.google\.com/books/edition/[^/]+?/[^/?#]+/?[&#?]"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-060", "Non-canonical Google Books URL. Google Books URLs must look exactly like [url]https://www.google.com/books/edition/<BOOK-NAME>/<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		if regex.search(r"""<a href="https?://(?!standardebooks\.org)""", long_description):
+			messages.append(LintMessage("m-067", "Non-S.E. link in long description.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	search_regex = r"https?://(babel\.hathitrust\.org|hdl\.handle\.net)"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-005", "Non-canonical HathiTrust URL. HathiTrust URLs must look exactly like [url]https://catalog.hathitrust.org/Record/<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# xml:lang is correct for the rest of the publication, but should be lang in the long desc
+		if "xml:lang" in long_description:
+			messages.append(LintMessage("m-057", "[xml]xml:lang[/] attribute in [xml]<meta property=\"se:long-description\">[/] element should be [xml]lang[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	search_regex = r"https?://.*?gutenberg\.org/(files|cache)"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-006", "Non-canonical Project Gutenberg URL. Project Gutenberg URLs must look exactly like [url]https://www.gutenberg.org/ebooks/<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# US -> U.S.
+		matches = regex.findall(r"\bUS\b", long_description)
+		if matches:
+			messages.append(LintMessage("t-047", "[text]US[/] should be [text]U.S.[/]", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
 
-	search_regex = r"^https?://.*?archive\.org/(stream|details/.+?/page.+)"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-007", "Non-canonical archive.org URL. Internet Archive URLs must look exactly like [url]https://archive.org/details/<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# Make sure long-description is escaped HTML
+		if not regex.search(r"^\s*<p>", long_description) and long_description.strip() != "LONG_DESCRIPTION":
+			messages.append(LintMessage("m-016", "Long description must be escaped HTML.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+		else:
+			# Check for malformed long description HTML
+			try:
+				etree.parse(io.StringIO(f"<?xml version=\"1.0\"?><html xmlns=\"http://www.w3.org/1999/xhtml\">{long_description}</html>"))
+			except lxml.etree.XMLSyntaxError as ex:
+				messages.append(LintMessage("m-015", f"Metadata long description is not valid XHTML. LXML says: {ex}", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	search_regex = r"https?://standardebooks\.org[^\s]*/$"
-	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
-	if nodes:
-		messages.append(LintMessage("m-054", "Standard Ebooks URL with illegal trailing slash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# Check for apostrophes outside links in long description
+		matches = regex.findall(r"</a>’s", long_description) + regex.findall(r"s</a>’", long_description)
+		if matches:
+			messages.append(LintMessage("m-044", "Possessive [text]’[/] or [text]’s[/] outside of [xhtml]<a>[/] element in long description.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
 
-	return messages
+		# Check for HTML entities in long-description, but allow &amp;
+		matches = regex.findall(r"&[a-z0-9]+?;", long_description.replace("&", ""))
+		if matches:
+			messages.append(LintMessage("m-018", "HTML entities found. Use Unicode equivalents instead.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
 
-def _get_selectors_and_rules (self) -> tuple:
-	"""
-	Helper function used in self.lint()
-	Construct a set of CSS selectors and rules in local.css
+	except Exception:
+		if self.is_se_ebook:
+			missing_metadata_elements.append("""<meta id="long-description" property="se:long-description" refines="#description">""")
 
-	INPUTS
-	None
+	missing_metadata_vars = []
+	for node in self.metadata_dom.xpath("/package/metadata/*/text()"):
+		for var in SE_VARIABLES:
+			if regex.search(fr"\b{var}\b", node):
+				missing_metadata_vars.append(var)
+				# Quit the loop early to save some time
+				break
 
-	OUTPUTS
-	2-tuple (local_css_rules, duplicate_selectors) to be used by the lint function
-	"""
+	if missing_metadata_vars:
+		messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, missing_metadata_vars))
 
-	def _recursive_helper(rules: cssutils.css.cssstylesheet.CSSStyleSheet, local_css_rules: dict, duplicate_selectors: list, single_selectors: list, top_level: bool):
-		"""
-		Because of the possibilty of nested @supports and @media at-rules, we
-		need to use recursion to get to rules and selectors within at-rules.
-		This function is the helper for _get_selectors_and_rules and does the
-		actual work.
+	# Check if there are non-typogrified quotes or em-dashes in the title.
+	try:
+		title = self.metadata_dom.xpath("/package/metadata/dc:title")[0].text
+		matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", title)
+		if matches:
+			messages.append(LintMessage("m-012", "Non-typogrified character in [xml]<dc:title>[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
 
-		INPUTS
-		rules: A CSSStyleSheet with the rules
-		local_css_rules: A dictionary where key = selector and value = rules
-		duplicate_selectors: Selectors which are counted as duplicates and will be warned about
-		single_selectors: Not multiple selectors separated by comma
-		top_level: Boolean set to True on first level of recursion and False thereafter
+		# Do we need an dcterms:alternate meta element?
+		# Match spelled-out numbers with a word joiner, so for ex. we don't print "eight" if we matched "eighty"
+		matches = regex.findall(r"(?:[0-9]+|\bone\b|\btwo\b|\bthree\b|\bfour\b|\bfive\b|\bsix\b|\bseven\b|\beight\b|\bnine\b|\bten\b|\beleven\b|\btwelve\b|\bthirteen\b|\bfourteen\b|\bfifteen\b|\bsixteen\b|\bseventeen\b|\beighteen\b|\bnineteen\b|\btwenty\b|\bthirty\b|\bforty\b|\bfifty\b|\bsixty\b|\bseventy\b|\beighty|\bninety)", title, flags=regex.IGNORECASE)
+		if matches and not self.metadata_dom.xpath("/package/metadata/meta[@property='dcterms:alternate']"):
+			messages.append(LintMessage("m-052", "[xml]<dc:title>[/] element contains numbers, but no [xml]<meta property=\"dcterms:alternate\" refines=\"#title\"> element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	except Exception:
+		missing_metadata_elements.append("<dc:title>")
 
-		OUTPUTS
-		2-tuple (local_css_rules, duplicate_selectors) to be used by the lint function
-		"""
-		for rule in rules:
-			# i.e. @supports or @media
-			if isinstance(rule, cssutils.css.CSSMediaRule):
-				# Recurisive call to rules within CSSMediaRule.
-				new_rules = _recursive_helper(rule.cssRules, local_css_rules, duplicate_selectors, single_selectors, False)
+	try:
+		file_as = self.metadata_dom.xpath("/package/metadata/meta[@property='file-as' and @refines='#title']")[0].text
+		matches = regex.findall(r".(?:['\"]|\-\-|\s-\s).", file_as)
+		if matches:
+			messages.append(LintMessage("m-050", "Non-typogrified character in [xml]<meta property=\"file-as\" refines=\"#title\">[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	except Exception:
+		missing_metadata_elements.append("<meta property=\"file-as\" refines=\"#title\">")
 
-				# Then update local_css_rules. The duplicate and single selector lists aren't updated
-				# because anything within CSSMediaRules isn't counted as a duplicate
-				local_css_rules.update(new_rules[0])
+	try:
+		description = self.metadata_dom.xpath("/package/metadata/dc:description")[0].text
+		matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", description)
+		if matches:
+			messages.append(LintMessage("m-013", "Non-typogrified character in [xml]<dc:description>[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	except Exception:
+		missing_metadata_elements.append("<dc:description>")
 
-			# Recursive end condition
-			if isinstance(rule, cssutils.css.CSSStyleRule):
-				for selector in rule.selectorList:
-					# Check for duplicate selectors.
-					# We consider a selector a duplicate if it's a TOP LEVEL selector (i.e. we don't check within @supports)
-					# and ALSO if it is a SINGLE selector (i.e. not multiple selectors separated by ,)
-					# For example abbr{} abbr{} would be a duplicate, but not abbr{} abbr,p{}
-					if "," not in rule.selectorText:
-						if top_level:
-							if selector.selectorText in single_selectors:
-								duplicate_selectors.append(selector.selectorText)
-							else:
-								single_selectors.append(selector.selectorText)
+	# Check for double spacing
+	if self.metadata_dom.xpath(f"/package/metadata/*[re:test(., '[{se.NO_BREAK_SPACE}{se.HAIR_SPACE} ]{{2,}}')]"):
+		double_spaced_files.append(self.metadata_file_path)
 
-					if selector.selectorText not in local_css_rules:
-						local_css_rules[selector.selectorText] = ""
+	# Check for illegal Wikipedia URLs
+	nodes = self.metadata_dom.xpath("/package/metadata/*[contains(., '.m.wikipedia.org') or @*[contains(., '.m.wikipedia.org')]]")
+	if nodes:
+		messages.append(LintMessage("m-026", "Illegal [url]https://*.m.wikipedia.org[/] URL. Hint: use non-mobile Wikipedia URLs.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
 
-					local_css_rules[selector.selectorText] += rule.style.cssText + ";"
+	# Check for punctuation outside quotes. We don't check single quotes because contractions are too common.
+	# We can't use xpath's built-in regex because it doesn't support Unicode classes
+	for node in self.metadata_dom.xpath("/package/metadata/*"):
+		if node.text and regex.search(r"[\p{Letter}]+”[,\.](?! …)", node.text):
+			messages.append(LintMessage("t-002", "Comma or period outside of double quote. Generally punctuation goes within single and double quotes.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path))
+			break
 
-		return (local_css_rules, duplicate_selectors)
+	# Check that the word count is correct, if it's currently set
+	word_count = self.metadata_dom.xpath("/package/metadata/meta[@property='se:word-count']/text()", True)
+	if self.is_se_ebook:
+		if word_count is None:
+			missing_metadata_elements.append("""<meta property="se:word-count">""")
+		elif word_count != "WORD_COUNT" and int(word_count) != self.get_word_count():
+			messages.append(LintMessage("m-065", "Word count in metadata doesn’t match actual word count.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	local_css_rules: Dict[str, str] = {} # A dict where key = selector and value = rules
-	duplicate_selectors: List[str] = []
-	single_selectors: List[str] = []
+	# Check if we have a subtitle but no fulltitle
+	if self.metadata_dom.xpath("/package/metadata[./meta[@property='title-type' and text()='subtitle'] and not(./meta[@property='title-type' and text()='extended'])]"):
+		messages.append(LintMessage("m-011", "Subtitle in metadata, but no full/extended title element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	# cssutils doesn't understand @supports, but it *does* understand @media, so do a replacement here for the purposes of parsing
-	all_rules = cssutils.parseString(self.local_css.replace("@supports", "@media"), validate=False)
+	# Check for tags that imply other tags
+	implied_tags = {"Fiction": ["Science Fiction", "Drama", "Fantasy"]}
+	for implied_tag, tags in implied_tags.items():
+		if self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:subject' and text()={se.easy_xml.escape_xpath(implied_tag)}]"):
+			for tag in tags:
+				if self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:subject' and text()={se.easy_xml.escape_xpath(tag)}]"):
+					messages.append(LintMessage("m-058", f"[val]se:subject[/] of [text]{implied_tag}[/] found, but [text]{tag}[/] implies [text]{implied_tag}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
 
-	# Initial recursive call
-	_recursive_helper(all_rules, local_css_rules, duplicate_selectors, single_selectors, True)
+	# Check for 'comprised of'
+	if self.metadata_dom.xpath("/package/metadata/*[re:test(., '[Cc]omprised of')]"):
+		messages.append(LintMessage("m-069", "[text]comprised of[/] in metadata. Hint: Is there a better phrase to use here?", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	return (local_css_rules, duplicate_selectors)
+	# Check for illegal em-dashes in <dc:subject>
+	nodes = self.metadata_dom.xpath("/package/metadata/dc:subject[contains(text(), '—')]")
+	if nodes:
+		messages.append(LintMessage("m-019", "Illegal em-dash in [xml]<dc:subject>[/] element; use [text]--[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.text for node in nodes]))
 
-def files_not_in_spine(self) -> set:
-	"""
-	Check the spine against the actual files.
+	# Check for incorrect 'anonymous' strings in metadata
+	nodes = self.metadata_dom.xpath("/package/metadata/dc:contributor[re:test(., 'anonymous', 'i') and text() != 'Anonymous'] | /package/metadata/meta[@property='file-as' and re:test(., 'anonymous', 'i') and text() != 'Anonymous']")
+	if nodes:
+		messages.append(LintMessage("m-073", "Anonymous contributor values must be exactly [text]Anonymous[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
 
-	INPUTS
-	None
+	# Check for metadata elements
+	nodes = self.metadata_dom.xpath("/package/metadata/*[not(name()='link') and not(normalize-space(.)) and not(./*)]")
+	if nodes:
+		messages.append(LintMessage("m-022", "Empty element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
 
-	OUTPUTS
-	Set of files not in the spine (typically an empty set)
-	"""
+	# Check for illegal VCS URLs
+	nodes = self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:url.vcs.github' and not(text()='{self.generated_github_repo_url}')]")
+	if nodes:
+		messages.append(LintMessage("m-009", f"[xml]<meta property=\"se:url.vcs.github\">[/] value does not match expected: [url]{self.generated_github_repo_url}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	xhtml_files = set(self.content_path.glob("**/*.xhtml"))
-	spine_files = set(self.spine_file_paths + [self.toc_path])
-	return xhtml_files.difference(spine_files)
+	# Check for illegal se:subject tags
+	illegal_subjects = []
+	nodes = self.metadata_dom.xpath("/package/metadata/meta[@property='se:subject']/text()")
+	if nodes:
+		for node in nodes:
+			if node not in SE_GENRES and node != "TAG":
+				illegal_subjects.append(node)
 
-def _update_missing_styles(filename: Path, dom: se.easy_xml.EasyXmlTree, local_css: dict) -> list:
-	"""
-	Identify any missing CSS styles in the file being checked
+		if illegal_subjects:
+			messages.append(LintMessage("m-020", "Illegal value for [xml]<meta property=\"se:subject\">[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, illegal_subjects))
 
-	INPUTS
-	filename: the name of the file being checked
-	dom: The dom of the file being checked
-	local_css: Dictionary containing several flags concerning the local CSS
+		if sorted(nodes) != nodes:
+			messages.append(LintMessage("m-053", "[xml]<meta property=\"se:subject\">[/] elements not in alphabetical order.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	OUTPUTS
-	List of styles used in this file but missing from local CSS
-	"""
+	elif self.is_se_ebook:
+		messages.append(LintMessage("m-021", "No [xml]<meta property=\"se:subject\">[/] element found.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	missing_styles: List[str] = []
+	# Check that each <dc:title> has a file-as and title-type, if applicable
+	titles_missing_file_as = []
+	titles_missing_title_type = []
+	titles = self.metadata_dom.xpath("/package/metadata/dc:title")
+	for node in titles:
+		if not self.metadata_dom.xpath(f"/package/metadata/meta[@property='file-as' and @refines='#{node.get_attr('id')}']"):
+			titles_missing_file_as.append(node)
 
-	if not local_css["has_elision_style"]:
-		missing_styles += [node.to_tag_string() for node in dom.xpath("/html/body//span[contains(@class, 'elision')]")]
+		# Only check for title-type if there is more than one <dc:title>
+		if len(titles) > 1 and not self.metadata_dom.xpath(f"/package/metadata/meta[@property='title-type' and @refines='#{node.get_attr('id')}']"):
+			titles_missing_title_type.append(node)
 
-	# Check to see if we included poetry or verse without the appropriate styling
-	if filename.name not in IGNORED_FILENAMES:
-		nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')][./p/span]")
-		for node in nodes:
-			if "z3998:poem" in node.get_attr("epub:type") and not local_css["has_poem_style"]:
-				missing_styles.append(node.to_tag_string())
+	if titles_missing_file_as:
+		messages.append(LintMessage("m-062", "[xml]<dc:title>[/] missing matching [xml]<meta property=\"file-as\">[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in titles_missing_file_as]))
 
-			if "z3998:verse" in node.get_attr("epub:type") and not local_css["has_verse_style"]:
-				missing_styles.append(node.to_tag_string())
+	if titles_missing_title_type:
+		messages.append(LintMessage("m-068", "[xml]<dc:title>[/] missing matching [xml]<meta property=\"title-type\">[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in titles_missing_title_type]))
 
-			if "z3998:song" in node.get_attr("epub:type") and not local_css["has_song_style"]:
-				missing_styles.append(node.to_tag_string())
+	# Check for CDATA tags
+	if "<![CDATA[" in metadata_xml:
+		messages.append(LintMessage("m-017", "[xml]<!\\[CDATA\\[[/] found. Run [bash]se clean[/] to canonicalize [xml]<!\\[CDATA\\[[/] sections.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-			if "z3998:hymn" in node.get_attr("epub:type") and not local_css["has_hymn_style"]:
-				missing_styles.append(node.to_tag_string())
+	# Check that our provided identifier matches the generated identifier
+	if self.is_se_ebook:
+		try:
+			identifier = self.metadata_dom.xpath("/package/metadata/dc:identifier")[0].text
+			if identifier != self.generated_identifier:
+				messages.append(LintMessage("m-023", f"[xml]<dc:identifier>[/] does not match expected: [text]{self.generated_identifier}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+		except Exception:
+			missing_metadata_elements.append("<dc:identifier>")
 
-			if "z3998:lyrics" in node.get_attr("epub:type") and not local_css["has_lyrics_style"]:
-				missing_styles.append(node.to_tag_string())
+	# Check if se:name.person.full-name matches their titlepage name
+	duplicate_names = []
+	invalid_refines = []
+	nodes = self.metadata_dom.xpath("/package/metadata/meta[@property='se:name.person.full-name']")
+	for node in nodes:
+		try:
+			refines = node.get_attr("refines").replace("#", "")
+			try:
+				name = self.metadata_dom.xpath(f"/package/metadata/*[@id={se.easy_xml.escape_xpath(refines)}]")[0].text
+				if name == node.text:
+					duplicate_names.append(name)
+			except Exception:
+				invalid_refines.append(refines)
+		except Exception:
+			invalid_refines.append("<meta property=\"se:name.person.full-name\">")
 
-		# Check frontmatter for missing styling
-		nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'dedication|epigraph')]")
-		for node in nodes:
-			if "dedication" in node.get_attr("epub:type") and not local_css["has_dedication_style"]:
-				missing_styles.append(node.to_tag_string())
+	if duplicate_names:
+		messages.append(LintMessage("m-024", "[xml]<meta property=\"se:name.person.full-name\">[/] property identical to regular name. If the two are identical the full name [xml]<meta>[/] element must be removed.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, duplicate_names))
 
-			if "epigraph" in node.get_attr("epub:type") and not local_css["has_epigraph_style"]:
-				missing_styles.append(node.to_tag_string())
+	if invalid_refines:
+		messages.append(LintMessage("m-010", "Invalid [xml]refines[/] property.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, invalid_refines))
 
-	return missing_styles
+	if self.metadata_dom.xpath("/package/metadata/*[contains(., 'https://id.loc.gov/')]"):
+		messages.append(LintMessage("m-066", "[url]id.loc.gov[/] URI starting with illegal https.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-def _lint_svg_checks(self, filename: Path, file_contents: str, svg_dom: se.easy_xml.EasyXmlTree, root: str) -> list:
+	if self.metadata_dom.xpath("/package/metadata/*[re:test(., 'id\\.loc\\.gov/authorities/names/[^\\.]+\\.html')]"):
+		messages.append(LintMessage("m-008", "[url]id.loc.gov[/] URI ending with illegal [path].html[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+
+	if self.metadata_dom.xpath("/package/metadata/dc:description[text()!='DESCRIPTION' and re:test(., '[^\\.”]$')]"):
+		messages.append(LintMessage("m-055", "[xml]dc:description[/] does not end with a period.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+
+	# Does the manifest match the generated manifest?
+	try:
+		manifest = self.metadata_dom.xpath("/package/manifest")[0]
+		if manifest.to_string().replace("\t", "") != self.generate_manifest().to_string().replace("\t", ""):
+			messages.append(LintMessage("m-042", "[xml]<manifest>[/] element does not match expected structure.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	except Exception:
+		missing_metadata_elements.append("<manifest>")
+
+	if missing_metadata_elements:
+		messages.append(LintMessage("m-051", "Missing expected element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, missing_metadata_elements))
+
+	# Check for common typos in description
+	for node in self.metadata_dom.xpath("/package/metadata/dc:description") + self.metadata_dom.xpath("/package/metadata/meta[@property='se:long-description']"):
+		matches = regex.findall(r"(?<!’)\b(and and|the the|if if|of of|or or|as as)\b(?!-)", node.text, flags=regex.IGNORECASE)
+		matches = matches + regex.findall(r"\ba a\b(?!-)", node.text)
+		if matches:
+			messages.append(LintMessage("y-001", "Possible typo: doubled [text]a/the/and/of/or/as/if[/].", se.MESSAGE_TYPE_WARNING, self.metadata_file_path, matches))
+
+	nodes = self.metadata_dom.xpath("/package/metadata//*[re:test(., '[,;][a-z]', 'i')]")
+	if nodes:
+		messages.append(LintMessage("y-002", "Possible typo: punctuation followed directly by a letter, without a space.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path, [node.to_string() for node in nodes]))
+
+	return messages
+
+def _get_malformed_urls(dom: se.easy_xml.EasyXmlTree, filename: Path) -> list:
 	"""
-	Perform several checks on svg files
+	Helper function used in self.lint()
+	Get a list of URLs in the epub that don't match SE standards.
 
 	INPUTS
-	filename: The name of the svg file being checked
-	file_contents: The contents of the svg file being checked
-	svg_dom: The dom of the svg file being checked
-	self
-	root
+	dom: A dom tree to check
+	filename: The filename being processed
 
 	OUTPUTS
-	A list of LintMessages representing errors found in the svg file
+	A list of LintMessages objects.
 	"""
 
 	messages = []
 
-	# Make images have reasonable dimensions
-	viewbox = svg_dom.xpath("/svg/@viewBox", True)
-	if viewbox:
-		svg_dimensions = viewbox.split()
-		try:
-			if float(svg_dimensions[2]) * float(svg_dimensions[3]) > 4000000:
-				messages.append(LintMessage("f-018", "Image greater than 4,000,000 pixels square in dimension.", se.MESSAGE_TYPE_ERROR, filename))
-		except Exception as ex:
-			raise se.InvalidFileException(f"Couldn’t parse SVG [xhtml]viewBox[/] attribute in [path][link=file://{filename.resolve()}]{filename}[/][/].") from ex
-
-	# If we're looking at the cover image, ensure that the producer has built it.
-	# The default cover image is a white background which when encoded to base64 begins with 299 characters and then has a long string of `A`s
-	if self.cover_path and filename.name == self.cover_path.name and svg_dom.xpath("//image[re:test(@xlink:href, '^.{299}A{50,}')]"):
-		messages.append(LintMessage("m-063", "Cover image has not been built.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for non-https URLs
+	search_regex = r"(?<!www\.)gutenberg\.org"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	if nodes:
+		messages.append(LintMessage("m-001", "gutenberg.org URL missing leading [text]www.[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for illegal transform or id attribute
-	nodes = svg_dom.xpath("//*[@transform or @id]")
+	search_regex = r"www\.archive\.org"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
 	if nodes:
-		invalid_transform_attributes = set()
-		invalid_id_attributes = []
-		for node in nodes:
-			if node.get_attr("transform"):
-				invalid_transform_attributes.add(f"transform=\"{node.get_attr('transform')}\"")
+		messages.append(LintMessage("m-002", "archive.org URL should not have leading [text]www.[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-			if node.get_attr("id"):
-				invalid_id_attributes.append(f"id=\"{node.get_attr('id')}\"")
+	search_regex = r"www\.gutenberg\.net\.au"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	if nodes:
+		messages.append(LintMessage("m-076", "gutenberg.net.au URL should not have leading [text]www.[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-		if invalid_transform_attributes:
-			messages.append(LintMessage("x-003", "Illegal [xml]transform[/] attribute. SVGs should be optimized to remove use of [xml]transform[/]. Try using Inkscape to save as an “optimized SVG”.", se.MESSAGE_TYPE_ERROR, filename, invalid_transform_attributes))
+	search_regex = r"http://(gutenberg\.org|gutenberg\.net\.au|gutenberg\.ca|www\.fadedpage\.com|archive\.org|pgdp\.net|catalog\.hathitrust\.org|en\.wikipedia\.org|standardebooks\.org)"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	if nodes:
+		messages.append(LintMessage("m-003", "Non-HTTPS URL.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for fill: #000 which should simply be removed
-	nodes = svg_dom.xpath("//*[contains(@fill, '#000') or contains(translate(@style, ' ', ''), 'fill:#000')]")
+	search_regex = r"https?://books\.google\.com/books\?id=.+?[&#]"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
 	if nodes:
-		messages.append(LintMessage("x-004", "Illegal [xml]style=\"fill: #000\"[/] or [xml]fill=\"#000\"[/].", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("m-004", "Non-canonical Google Books URL. Google Books URLs must look exactly like [url]https://books.google.com/books?id=<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for illegal height or width on root <svg> element
-	if filename.name != "logo.svg": # Do as I say, not as I do...
-		if svg_dom.xpath("/svg[@height or @width]"):
-			messages.append(LintMessage("x-005", "Illegal [xml]height[/] or [xml]width[/] attribute on root [xml]<svg>[/] element. Size SVGs using the [xml]viewBox[/] attribute only.", se.MESSAGE_TYPE_ERROR, filename))
+	search_regex = r"https?://www\.google\.com/books/edition/[^/]+?/[^/?#]+/?[&#?]"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	if nodes:
+		messages.append(LintMessage("m-060", "Non-canonical Google Books URL. Google Books URLs must look exactly like [url]https://www.google.com/books/edition/<BOOK-NAME>/<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	match = regex.search(r"viewbox", file_contents, flags=regex.IGNORECASE)
-	if match and match[0] != "viewBox":
-		messages.append(LintMessage("x-006", f"[xml]{match}[/] found instead of [xml]viewBox[/]. [xml]viewBox[/] must be correctly capitalized.", se.MESSAGE_TYPE_ERROR, filename))
+	search_regex = r"https?://(babel\.hathitrust\.org|hdl\.handle\.net)"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	if nodes:
+		messages.append(LintMessage("m-005", "Non-canonical HathiTrust URL. HathiTrust URLs must look exactly like [url]https://catalog.hathitrust.org/Record/<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-		if invalid_id_attributes:
-			messages.append(LintMessage("x-014", "Illegal [xml]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, invalid_id_attributes))
+	search_regex = r"https?://.*?gutenberg\.org/(files|cache)"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	if nodes:
+		messages.append(LintMessage("m-006", "Non-canonical Project Gutenberg URL. Project Gutenberg URLs must look exactly like [url]https://www.gutenberg.org/ebooks/<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	if f"{os.sep}src{os.sep}" not in root:
-		# Check that cover and titlepage images are in all caps
-		if self.cover_path and filename.name == self.cover_path.name:
-			nodes = svg_dom.xpath("//text[re:test(., '[a-z]')]")
-			if nodes:
-				messages.append(LintMessage("s-002", "Lowercase letters in cover. Cover text must be all uppercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	search_regex = r"^https?://.*?archive\.org/(stream|details/.+?/page.+)"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	if nodes:
+		messages.append(LintMessage("m-007", "Non-canonical archive.org URL. Internet Archive URLs must look exactly like [url]https://archive.org/details/<BOOK-ID>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-		if filename.name == "titlepage.svg":
-			nodes = svg_dom.xpath("//text[re:test(., '[a-z]') and not(text()='translated by' or text()='illustrated by' or text()='and')]")
-			if nodes:
-				messages.append(LintMessage("s-003", "Lowercase letters in titlepage. Titlepage text must be all uppercase except [text]translated by[/] and [text]illustrated by[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	search_regex = r"https?://standardebooks\.org[^\s]*/$"
+	nodes = dom.xpath(f"/package/metadata/*[re:test(., '{search_regex}')] | /html/body//a[re:test(@href, '{search_regex}')]")
+	if nodes:
+		messages.append(LintMessage("m-054", "Standard Ebooks URL with illegal trailing slash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
 	return messages
 
-def _lint_special_file_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_info: dict, special_file: str) -> list:
+def _get_selectors_and_rules (self) -> tuple:
 	"""
-	Process error checks in “special” .xhtml files
+	Helper function used in self.lint()
+	Construct a set of CSS selectors and rules in local.css
 
 	INPUTS
-	filename: The name of the file being checked
-	dom: The dom of the file being checked
-	file_contents: The contents of the file being checked
-	ebook_info: A dictionary containing several pieces of information about the ebook
-	special_file: A string identifying the type of special file being checked
-	self
+	None
 
 	OUTPUTS
-	A list of LintMessages representing errors found in the special file
+	2-tuple (local_css_rules, duplicate_selectors) to be used by the lint function
 	"""
 
-	messages = []
-	source_links = self.metadata_dom.xpath("/package/metadata/dc:source/text()")
+	def _recursive_helper(rules: cssutils.css.cssstylesheet.CSSStyleSheet, local_css_rules: dict, duplicate_selectors: list, single_selectors: list, top_level: bool):
+		"""
+		Because of the possibilty of nested @supports and @media at-rules, we
+		need to use recursion to get to rules and selectors within at-rules.
+		This function is the helper for _get_selectors_and_rules and does the
+		actual work.
 
-	if special_file == "colophon" or special_file == "imprint":
-		if len(source_links) <= 2:
-			# Check that links back to sources are represented correctly
-			nodes = dom.xpath("/html/body//a[@href='https://www.pgdp.net' and text()!='The Online Distributed Proofreading Team']")
-			if nodes:
-				messages.append(LintMessage("m-071", "DP link must be exactly [text]The Online Distributed Proofreading Team[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		INPUTS
+		rules: A CSSStyleSheet with the rules
+		local_css_rules: A dictionary where key = selector and value = rules
+		duplicate_selectors: Selectors which are counted as duplicates and will be warned about
+		single_selectors: Not multiple selectors separated by comma
+		top_level: Boolean set to True on first level of recursion and False thereafter
 
-			nodes = dom.xpath("/html/body//a[re:test(@href, '^https://www.pgdp.org/ols/') and text()!='Distributed Proofreaders Open Library System']")
-			if nodes:
-				messages.append(LintMessage("m-072", "DP OLS link must be exactly [text]Distributed Proofreaders Open Library System[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		OUTPUTS
+		2-tuple (local_css_rules, duplicate_selectors) to be used by the lint function
+		"""
+		for rule in rules:
+			# i.e. @supports or @media
+			if isinstance(rule, cssutils.css.CSSMediaRule):
+				# Recurisive call to rules within CSSMediaRule.
+				new_rules = _recursive_helper(rule.cssRules, local_css_rules, duplicate_selectors, single_selectors, False)
 
-			nodes = dom.xpath("/html/body//a[re:test(@href, '^https://[^\"]*?hathitrust.org') and re:test(text(), '[Hh]athi') and not(text()='HathiTrust Digital Library')]")
-			if nodes:
-				messages.append(LintMessage("m-041", "Hathi Trust link text must be exactly [text]HathiTrust Digital Library[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+				# Then update local_css_rules. The duplicate and single selector lists aren't updated
+				# because anything within CSSMediaRules isn't counted as a duplicate
+				local_css_rules.update(new_rules[0])
 
-		nodes = dom.xpath("/html/body//p[re:test(., '\\ba transcription\\b') and ./a[contains(@href, '#transcriptions')]]")
-		if nodes:
-			messages.append(LintMessage("t-071", "Multiple transcriptions listed, but preceding text is [text]a transcription[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+			# Recursive end condition
+			if isinstance(rule, cssutils.css.CSSStyleRule):
+				for selector in rule.selectorList:
+					# Check for duplicate selectors.
+					# We consider a selector a duplicate if it's a TOP LEVEL selector (i.e. we don't check within @supports)
+					# and ALSO if it is a SINGLE selector (i.e. not multiple selectors separated by ,)
+					# For example abbr{} abbr{} would be a duplicate, but not abbr{} abbr,p{}
+					if "," not in rule.selectorText:
+						if top_level:
+							if selector.selectorText in single_selectors:
+								duplicate_selectors.append(selector.selectorText)
+							else:
+								single_selectors.append(selector.selectorText)
 
-	if special_file == "colophon":
-		# Check for illegal Wikipedia URLs
-		nodes = dom.xpath("/html/body//a[contains(@href, '.m.wikipedia.org')]")
-		if nodes:
-			messages.append(LintMessage("m-026", "Illegal [url]https://*.m.wikipedia.org[/] URL. Hint: use non-mobile Wikipedia URLs.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+					if selector.selectorText not in local_css_rules:
+						local_css_rules[selector.selectorText] = ""
 
-		# Check for non-English Wikipedia URLs
-		nodes = dom.xpath("/html/body//a[re:test(@href, 'https://(?!en)[a-z]{2,}\\.wikipedia\\.org')]")
-		if nodes:
-			messages.append(LintMessage("m-043", "Non-English Wikipedia URL.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
+					local_css_rules[selector.selectorText] += rule.style.cssText + ";"
 
-		# Check for wrong grammar filled in from template
-		nodes = dom.xpath("/html/body//a[starts-with(@href, 'https://books.google.com/') or starts-with(@href, 'https://www.google.com/books/')][(preceding-sibling::text()[normalize-space(.)][1])[re:test(., '\\bthe$')]]")
-		if nodes:
-			messages.append(LintMessage("s-016", "Incorrect [text]the[/] before Google Books link.", se.MESSAGE_TYPE_ERROR, filename, ["the<br/>\n" + node.to_string() for node in nodes]))
+		return (local_css_rules, duplicate_selectors)
 
-		se_url = self.generated_identifier.replace("url:", "")
-		if not dom.xpath(f"/html/body//a[@href='{se_url}' and text()='{se_url.replace('https://', '')}']"):
-			messages.append(LintMessage("m-035", f"Unexpected S.E. identifier in colophon. Expected: [url]{se_url}[/].", se.MESSAGE_TYPE_ERROR, filename))
+	local_css_rules: Dict[str, str] = {} # A dict where key = selector and value = rules
+	duplicate_selectors: List[str] = []
+	single_selectors: List[str] = []
 
-		if self.metadata_dom.xpath("/package/metadata/meta[@property='role' and text()='trl']") and "translated from" not in file_contents:
-			messages.append(LintMessage("m-025", "Translator found in metadata, but no [text]translated from LANG[/] block in colophon.", se.MESSAGE_TYPE_ERROR, filename))
+	# cssutils doesn't understand @supports, but it *does* understand @media, so do a replacement here for the purposes of parsing
+	all_rules = cssutils.parseString(self.local_css.replace("@supports", "@media"), validate=False)
 
-		if ebook_info["has_multiple_transcriptions"] and not dom.xpath("/html/body//a[contains(@href, '#transcriptions')]"):
-			messages.append(LintMessage("m-074", "Multiple transcriptions found in metadata, but no link to [text]EBOOK_URL#transcriptions[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# Initial recursive call
+	_recursive_helper(all_rules, local_css_rules, duplicate_selectors, single_selectors, True)
 
-		if ebook_info["has_multiple_page_scans"] and not dom.xpath("/html/body//a[contains(@href, '#page-scans')]"):
-			messages.append(LintMessage("m-075", "Multiple page scans found in metadata, but no link to [text]EBOOK_URL#page-scans[/].", se.MESSAGE_TYPE_ERROR, filename))
+	return (local_css_rules, duplicate_selectors)
 
-		# Check that the formula changed from the default if we added 'various sources'
-		if ebook_info["has_multiple_transcriptions"] or ebook_info["has_multiple_page_scans"]:
-			nodes = dom.xpath("/html/body//a[text() = 'various sources' and not(re:test(preceding-sibling::br[1]/preceding-sibling::node()[1], '(digital scans|transcriptions) from\\s*$'))]")
-			if nodes:
-				messages.append(LintMessage("t-072", "[text]various sources[/] link not preceded by [text]from[/].", se.MESSAGE_TYPE_ERROR, filename))
+def files_not_in_spine(self) -> set:
+	"""
+	Check the spine against the actual files.
 
-		# Check if we forgot to fill any variable slots
-		missing_colophon_vars = [var for var in SE_VARIABLES if regex.search(fr"\b{var}\b", file_contents)]
-		if missing_colophon_vars:
-			messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, filename, missing_colophon_vars))
+	INPUTS
+	None
 
-		# Check that we have <br/>s at the end of lines
-		# First, check for b or a elements that are preceded by a newline but not by a br
-		nodes = [node.to_string() for node in dom.xpath("/html/body/section/p/*[name()='b' or name()='a'][(preceding-sibling::node()[1])[contains(., '\n')]][not((preceding-sibling::node()[2])[self::br]) or (normalize-space(preceding-sibling::node()[1]) and re:test(preceding-sibling::node()[1], '\\n\\s*$')) ]")]
-		# Next, check for text nodes that contain newlines but are not preceded by brs
-		nodes = nodes + [node.strip() for node in dom.xpath("/html/body/section/p/text()[contains(., '\n') and normalize-space(.)][(preceding-sibling::node()[1])[not(self::br)]]")]
-		if nodes:
-			messages.append(LintMessage("s-053", "Colophon line not preceded by [xhtml]<br/>[/].", se.MESSAGE_TYPE_ERROR, filename, nodes))
+	OUTPUTS
+	Set of files not in the spine (typically an empty set)
+	"""
 
-		# Is there a comma after a producer name, if there's only two producers?
-		nodes = dom.xpath("/html/body/section/p/*[name()='b' or name()='a'][(following-sibling::node()[1])[normalize-space(.)=', and']][(preceding-sibling::*[1])[name()='br']]")
-		if nodes:
-			messages.append(LintMessage("t-006", "Comma after producer name, but there are only two producers.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	xhtml_files = set(self.content_path.glob("**/*.xhtml"))
+	spine_files = set(self.spine_file_paths + [self.toc_path])
+	return xhtml_files.difference(spine_files)
 
-		nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and re:test(@epub:type, 'z3998:.*?name')]")
-		if nodes:
-			messages.append(LintMessage("s-092", "Anonymous contributor with [val]z3998:*-name[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+def _lint_css_checks(self, local_css_path: Path, abbr_with_whitespace: list) -> list:
+	"""
+	Process main CSS checks
 
-		# Check for anonymous volunteers misrepresented in the colophon. Note that we only match SE producers and transcribers, because
-		# a cover art artist can also be anonymous but they're not volunteers.
-		nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and text() != 'An Anonymous Volunteer' and (preceding-sibling::a[@href='https://standardebooks.org'])]")
-		if nodes:
-			messages.append(LintMessage("s-100", "Anonymous digital contributor value not exactly [text]An Anonymous Volunteer[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	INPUTS
+	self
+	local_css: Dictionary of local CSS flags
+	local_css_path: Path to local.css file
+	local_css_rules: Dictionary of the local CSS rules
 
-		# Check for primary contributors misrepresented in the colophon. these differ from ebook production volunteers.
-		nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and text() != 'Anonymous' and (preceding-sibling::node()[contains(., 'The cover page')] or preceding-sibling::i[contains(@epub:type, 'se:name.publication')])]")
-		if nodes:
-			messages.append(LintMessage("s-101", "Anonymous primary contributor value not exactly [text]Anonymous[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	OUTPUTS
+	messages
+	"""
+	messages = []
 
-		# Is there a page scan link in the colophon, but missing in the metadata?
-		for node in dom.xpath("/html/body//a[re:test(@href, '(gutenberg\\.org/ebooks/[0-9]+|hathitrust\\.org|/archive\\.org|books\\.google\\.com|www\\.google\\.com/books/)')]"):
-			if not self.metadata_dom.xpath(f"/package/metadata/dc:source[contains(text(), {se.easy_xml.escape_xpath(node.get_attr('href'))})]"):
-				messages.append(LintMessage("m-059", f"Link to [url]{node.get_attr('href')}[/] found in colophon, but missing matching [xhtml]dc:source[/] element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	# lxml has not implemented the following in cssselect: *:first-of-type, *:last-of-type, *:nth-of-type, *:nth-last-of-type, *:only-of-type
+	# BUT ONLY WHEN USED ON *. This includes for example: `section [epub|type~="test"]:first-of-type` (note the `*` is implicit)
+	# Therefore we can't simplify them in build or test against them.
+	matches = regex.findall(r"(?:^| )[^a-z\s][^\s]+?:(?:first-of-type|last-of-type|nth-of-type|nth-last-of-type|only-of-type)", self.local_css, flags=regex.MULTILINE)
+	if matches:
+		messages.append(LintMessage("c-001", "Don’t use [css]*:first-of-type[/], [css]*:last-of-type[/], [css]*:nth-of-type[/] [css]*:nth-last-of-type[/], or [css]*:only-of-type[/] on [css]*[/]. Instead, specify an element to apply it to.", se.MESSAGE_TYPE_ERROR, local_css_path, matches))
 
-		# Are the sources represented correctly?
-		# We don't have a standard yet for more than two sources (transcription and scan) so just ignore that case for now.
-		# We can't merge this with the imprint check because imprint doesn't have `<br/>` between `the`
-		if not ebook_info["has_multiple_transcriptions"] and not ebook_info["has_other_sources"]:
-			for link in source_links:
-				if "gutenberg.org" in link and f"<a href=\"{link}\">Project Gutenberg</a>" not in file_contents:
-					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Project Gutenberg</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	# If we select on the xml namespace, make sure we define the namespace in the CSS, otherwise the selector won't work
+	# We do this using a regex and not with cssutils, because cssutils will barf in this particular case and not even record the selector.
+	matches = regex.findall(r"\[\s*xml\s*\|", self.local_css)
+	if matches and "@namespace xml \"http://www.w3.org/XML/1998/namespace\";" not in self.local_css:
+		messages.append(LintMessage("c-003", "[css]\\[xml|attr][/] selector in CSS, but no XML namespace declared ([css]@namespace xml \"http://www.w3.org/XML/1998/namespace\";[/]).", se.MESSAGE_TYPE_ERROR, local_css_path))
 
-		if not ebook_info["has_multiple_page_scans"] and not ebook_info["has_other_sources"]:
-			for link in source_links:
-				if "hathitrust.org" in link and f"the<br/>\n\t\t\t<a href=\"{link}\">HathiTrust Digital Library</a>" not in file_contents:
-					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]the<br/> <a href=\"{link}\">HathiTrust Digital Library</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	if abbr_with_whitespace:
+		messages.append(LintMessage("c-005", f"[css]abbr[/] selector does not need [css]white-space: nowrap;[/] as it inherits it from [path][link=file://{self.path / 'src/epub/css/core.css'}]core.css[/][/].", se.MESSAGE_TYPE_ERROR, local_css_path, abbr_with_whitespace))
 
-				if "archive.org" in link and f"the<br/>\n\t\t\t<a href=\"{link}\">Internet Archive</a>" not in file_contents:
-					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]the<br/> <a href=\"{link}\">Internet Archive</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	if regex.search(r"\s+hyphens:.+?;(?!\s+-epub-hyphens)", self.local_css):
+		messages.append(LintMessage("c-007", "[css]hyphens[/css] CSS property without [css]-epub-hyphens[/css] copy.", se.MESSAGE_TYPE_ERROR, local_css_path))
 
-				if ("books.google.com" in link or "www.google.com/books/" in link) and f"<a href=\"{link}\">Google Books</a>" not in file_contents:
-					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Google Books</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	matches = regex.findall(r"text-align:\s*left\s*;", self.local_css)
+	if matches:
+		messages.append(LintMessage("c-016", "[css]text-align: left;[/] found. Use [css]text-align: initial;[/] instead.", se.MESSAGE_TYPE_ERROR, local_css_path))
 
-	# If we're in the imprint, are the sources represented correctly?
-	# We don't have a standard yet for more than two sources (transcription and scan) so just ignore that case for now.
-	elif special_file == "imprint":
-		# Check for wrong grammar filled in from template
-		nodes = dom.xpath("/html/body//a[starts-with(@href, 'https://books.google.com/') or starts-with(@href, 'https://www.google.com/books/')][(preceding-sibling::node()[1])[re:test(., 'the\\s+$')]]")
-		if nodes:
-			messages.append(LintMessage("s-016", "Incorrect [text]the[/] before Google Books link.", se.MESSAGE_TYPE_ERROR, filename, ["the " + node.to_string() for node in nodes]))
+	matches = regex.findall(r"[0-9\.]\s?rem;", self.local_css)
+	if matches:
+		messages.append(LintMessage("c-022", "Illegal [css]rem[/] unit. Use [css]em[/] instead.", se.MESSAGE_TYPE_ERROR, local_css_path))
 
-		# Check if we forgot to fill any variable slots
-		missing_imprint_vars = [var for var in SE_VARIABLES if regex.search(fr"\b{var}\b", file_contents)]
-		if missing_imprint_vars:
-			messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, filename, missing_imprint_vars))
+	matches = regex.findall(r"font-size\s*:\s*[0-9\.]+(?![0-9\.]|em|ex)", self.local_css)
+	if matches:
+		messages.append(LintMessage("c-023", "Illegal unit used to set [css]font-size[/]. Hint: Use [css]em[/] units.", se.MESSAGE_TYPE_ERROR, local_css_path))
 
-		if ebook_info["has_multiple_transcriptions"] and not dom.xpath("/html/body//a[contains(@href, '#transcriptions')]"):
-			messages.append(LintMessage("m-074", "Multiple transcriptions found in metadata, but no link to [text]EBOOK_URL#transcriptions[/].", se.MESSAGE_TYPE_ERROR, filename))
+	matches = regex.findall(r"line-height\s*:\s*[0-9\.]+(?!;|[0-9\.]+)", self.local_css)
+	if matches:
+		messages.append(LintMessage("c-024", "Illegal unit used to set [css]line-height[/]. Hint: [css]line-height[/] is set without any units.", se.MESSAGE_TYPE_ERROR, local_css_path))
 
-		if ebook_info["has_multiple_page_scans"] and not dom.xpath("/html/body//a[contains(@href, '#page-scans')]"):
-			messages.append(LintMessage("m-075", "Multiple page scans found in metadata, but no link to [text]EBOOK_URL#page-scans[/].", se.MESSAGE_TYPE_ERROR, filename))
+	matches = regex.findall(r"(height|\stop|\sbottom)\s*:\s*[0-9\.]+%", self.local_css)
+	if matches:
+		messages.append(LintMessage("c-025", "Illegal percent unit used to set [css]height[/] or positioning property. Hint: [css]vh[/] to specify vertical-oriented properties like height or position.", se.MESSAGE_TYPE_ERROR, local_css_path))
 
-		# Check that the formula changed from the default if we added 'various sources'
-		if ebook_info["has_multiple_transcriptions"] or ebook_info["has_multiple_page_scans"]:
-			nodes = dom.xpath("/html/body//a[text() = 'various sources' and not(re:test(preceding-sibling::node()[1], '(digital scans|transcriptions) from\\s*$'))]")
-			if nodes:
-				messages.append(LintMessage("t-072", "[text]various sources[/] link not preceded by [text]from[/].", se.MESSAGE_TYPE_ERROR, filename))
+	return messages
 
-		# Check for correctly named links. We can't merge this with the colophon check because the colophon breaks `the` with `<br/>`
-		if not ebook_info["has_multiple_transcriptions"] and not ebook_info["has_other_sources"]:
-			for link in source_links:
-				if "gutenberg.org" in link and f"<a href=\"{link}\">Project Gutenberg</a>" not in file_contents:
-					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Project Gutenberg</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+def _update_missing_styles(filename: Path, dom: se.easy_xml.EasyXmlTree, local_css: dict) -> list:
+	"""
+	Identify any missing CSS styles in the file being checked
 
-		if not ebook_info["has_multiple_page_scans"] and not ebook_info["has_other_sources"]:
-			for link in source_links:
-				if "hathitrust.org" in link and f"the <a href=\"{link}\">HathiTrust Digital Library</a>" not in file_contents:
-					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: the [xhtml]<a href=\"{link}\">HathiTrust Digital Library</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	INPUTS
+	filename: the name of the file being checked
+	dom: The dom of the file being checked
+	local_css: Dictionary containing several flags concerning the local CSS
 
-				if "archive.org" in link and f"the <a href=\"{link}\">Internet Archive</a>" not in file_contents:
-					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: the [xhtml]<a href=\"{link}\">Internet Archive</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	OUTPUTS
+	List of styles used in this file but missing from local CSS
+	"""
 
-				if ("books.google.com" in link or "www.google.com/books/" in link) and f"<a href=\"{link}\">Google Books</a>" not in file_contents:
-					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Google Books</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	missing_styles: List[str] = []
 
+	if not local_css["has_elision_style"]:
+		missing_styles += [node.to_tag_string() for node in dom.xpath("/html/body//span[contains(@class, 'elision')]")]
 
-	# Endnote checks
-	elif special_file == "endnotes":
-		# Do we have to replace Ibid.? Only match Ibid. if the endnote does not appear to cite any names.
-		nodes = dom.xpath("/html/body//li[re:test(@epub:type, '\\bendnote\\b')]//abbr[re:test(., '\\b[Ii]bid\\b') and not(ancestor::li[1]//*[contains(@epub:type, 'se:name')])]")
-		if nodes:
-			messages.append(LintMessage("s-039", "[text]Ibid[/] in endnotes. “Ibid” means “The previous reference” which is meaningless with popup endnotes, and must be replaced by the actual thing [text]Ibid[/] refers to, unless it refers to text within the same endnote.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check to see if we included poetry or verse without the appropriate styling
+	if filename.name not in IGNORED_FILENAMES:
+		nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')][./p/span]")
+		for node in nodes:
+			if "z3998:poem" in node.get_attr("epub:type") and not local_css["has_poem_style"]:
+				missing_styles.append(node.to_tag_string())
 
-		# Check that endnotes have their backlink in the last <p> element child of the <li>. This also highlights backlinks that are totally missing.
-		nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')][./p[last()][not(a[contains(@epub:type, 'backlink')])] or not(./p[last()])]")
-		if nodes:
-			messages.append(LintMessage("s-056", "Last [xhtml]<p>[/] child of endnote missing backlink.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+			if "z3998:verse" in node.get_attr("epub:type") and not local_css["has_verse_style"]:
+				missing_styles.append(node.to_tag_string())
 
-		# Make sure the backlink points to the same note number as the parent endnote ID
-		nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]//a[contains(@epub:type, 'backlink')][not(re:match(@href, '\\-[0-9]+$')=re:match(ancestor::li/@id, '\\-[0-9]+$'))]")
-		if nodes:
-			messages.append(LintMessage("s-057", "Backlink noteref fragment identifier doesn’t match endnote number.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+			if "z3998:song" in node.get_attr("epub:type") and not local_css["has_song_style"]:
+				missing_styles.append(node.to_tag_string())
 
-		# Check that citations at the end of endnotes are in a <cite> element. If not typogrify will run the last space together with the em dash.
-		# This tries to catch that, but limits the match to 20 chars so that we don't accidentally match a whole sentence that happens to be at the end of an endnote.
-		nodes = dom.xpath(f"/html/body//li[contains(@epub:type, 'endnote')]/p[last()][re:test(., '\\.”?{se.WORD_JOINER}?—[A-Z].{{0,20}}\\s*↩$')]")
-		if nodes:
-			messages.append(LintMessage("s-064", "Endnote citation not wrapped in [xhtml]<cite>[/]. Em dashes go within [xhtml]<cite>[/] and it is preceded by one space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+			if "z3998:hymn" in node.get_attr("epub:type") and not local_css["has_hymn_style"]:
+				missing_styles.append(node.to_tag_string())
 
-		# Did we forget the endnote semantic on li elements?
-		nodes = dom.xpath("/html/body/section/ol/li[not(re:test(@epub:type, '\\bendnote\\b'))]")
-		if nodes:
-			messages.append(LintMessage("s-099", "List item in endnotes missing [xhtml]endnote[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+			if "z3998:lyrics" in node.get_attr("epub:type") and not local_css["has_lyrics_style"]:
+				missing_styles.append(node.to_tag_string())
 
-		# Match backlink elements whose preceding node doesn't end with ' ', and is also not all whitespace
-		nodes = dom.xpath("/html/body//a[@epub:type='backlink'][(preceding-sibling::node()[1])[not(re:test(., ' $')) and not(normalize-space(.)='')]]")
-		if nodes:
-			messages.append(LintMessage("t-027", "Endnote backlink not preceded by exactly one space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		# Check frontmatter for missing styling
+		nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'dedication|epigraph')]")
+		for node in nodes:
+			if "dedication" in node.get_attr("epub:type") and not local_css["has_dedication_style"]:
+				missing_styles.append(node.to_tag_string())
 
-	# Check LoI descriptions to see if they match associated figcaptions
-	elif special_file == "loi":
-		for node in dom.xpath("/html/body/nav[contains(@epub:type, 'loi')]//li//a"):
-			figure_ref = node.get_attr("href").split("#")[1]
-			chapter_ref = regex.findall(r"(.*?)#.*", node.get_attr("href"))[0]
-			figcaption_text = ""
-			loi_text = node.inner_text()
-			file_dom = self.get_dom(self.content_path / "text" / chapter_ref)
+			if "epigraph" in node.get_attr("epub:type") and not local_css["has_epigraph_style"]:
+				missing_styles.append(node.to_tag_string())
 
-			try:
-				figure = file_dom.xpath(f"//*[@id={se.easy_xml.escape_xpath(figure_ref)}]")[0]
-			except Exception:
-				messages.append(LintMessage("s-040", f"[attr]#{figure_ref}[/] not found in file [path][link=file://{self.path / 'src/epub/text' / chapter_ref}]{chapter_ref}[/][/].", se.MESSAGE_TYPE_ERROR, filename))
-				continue
+	return missing_styles
 
-			for child in figure.xpath("./*"):
-				if child.tag == "img":
-					figure_img_alt = child.get_attr("alt")
+def _lint_image_checks(self, filename: Path) -> list:
+	"""
+	Process image checks
 
-				if child.tag == "figcaption":
-					figcaption_text = child.inner_text()
+	INPUTS
+	self
+	filename: The name of the file being processed
 
-					# Replace tabs and newlines with a single space to better match figcaptions that contain <br/>
-					figcaption_text = regex.sub(r"(\n|\t)", " ", figcaption_text)
-					figcaption_text = regex.sub(r"[ ]+", " ", figcaption_text)
+	OUTPUTS
+	A list of LintMessage objects.
+	"""
+	messages = []
 
-			if (figcaption_text != "" and loi_text != "" and figcaption_text != loi_text) and (figure_img_alt != "" and loi_text != "" and figure_img_alt != loi_text):
-				messages.append(LintMessage("s-041", f"The [xhtml]<figcaption>[/] element of [attr]#{figure_ref}[/] does not match the text in its LoI entry.", se.MESSAGE_TYPE_WARNING, self.path / "src/epub/text" / chapter_ref))
+	if filename.suffix == ".jpeg":
+		messages.append(LintMessage("f-011", "JPEG files must end in [path].jpg[/].", se.MESSAGE_TYPE_ERROR, filename))
+	elif filename.suffix == ".tiff":
+		messages.append(LintMessage("f-012", "TIFF files must end in [path].tif[/].", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Run some general tests on images, but skip the cover source since it's an exception to all of these rules
+	if "cover.source" not in filename.name:
+		try:
+			image = Image.open(filename)
+		except UnidentifiedImageError as ex:
+			raise se.InvalidFileException(f"Couldn’t identify image type of [path][link=file://{filename}]{filename.name}[/][/].") from ex
+
+		# Check the source cover image
+		if self.path / "images" / "cover.jpg" == filename:
+			if image.size != (se.COVER_WIDTH, se.COVER_HEIGHT):
+				messages.append(LintMessage("f-017", f"[path][link=file://{self.path / 'images/cover.jpg'}]cover.jpg[/][/] must be exactly {se.COVER_WIDTH} × {se.COVER_HEIGHT}.", se.MESSAGE_TYPE_ERROR, filename))
+
+		# Run some tests on distributable images in ./src/epub/images/
+		# Once we reach Python 3.9 we can use path.is_relative_to() instead of this string comparison
+		if str(filename).startswith(str(self.content_path / "images")):
+			if os.path.getsize(filename) > 1500000: # 1.5MB
+				messages.append(LintMessage("f-016", "Image more than 1.5MB in size.", se.MESSAGE_TYPE_ERROR, filename))
+
+			# Make sure distributable images have reasonable dimensions
+			# We check SVGs later on in a separate check
+			if image.size[0] * image.size[1] > 4000000:
+				messages.append(LintMessage("f-018", "Image greater than 4,000,000 pixels square in dimension.", se.MESSAGE_TYPE_ERROR, filename))
+
+			if filename.suffix == ".png" and not se.images.has_transparency(image):
+				messages.append(LintMessage("f-019", "[path].png[/] file without transparency. Hint: If an image doesn’t have transparency, it should be saved as a [path].jpg[/].", se.MESSAGE_TYPE_ERROR, filename))
 
 	return messages
 
-def  css_checks(filename: Path, dom: se.easy_xml.EasyXmlTree) -> list:
+def _lint_svg_checks(self, filename: Path, file_contents: str, svg_dom: se.easy_xml.EasyXmlTree, root: str) -> list:
 	"""
-	Helper function used in self.lint()
-	Process CSS checks on an .xhtml file
+	Perform several checks on svg files
 
 	INPUTS
-	filename: The name of the file being checked
-	dom: A dom tree to check
+	filename: The name of the svg file being checked
+	file_contents: The contents of the svg file being checked
+	svg_dom: The dom of the svg file being checked
+	self
+	root: The top-level directory
 
 	OUTPUTS
-	A list of LintMessages representing CSS errors found in the file
+	A list of LintMessage objects.
 	"""
 
 	messages = []
 
-	# Do we have any elements that have specified border color?
-	# `transparent` and `none` are allowed values for border-color
-	if dom.xpath("/html/body//*[attribute::*[re:test(local-name(), 'data-css-border.+?-color') and text() != 'transparent' and text != 'none']]"):
-		messages.append(LintMessage("c-004", "Don’t specify border colors, so that reading systems can adjust for night mode.", se.MESSAGE_TYPE_WARNING, local_css_path, matches))
-
-	# Check that footers have the expected styling
-	# Footers may sometimes be aligned with text-align: center as well as text-align: right
-	# We also ignore text-align on any footers whose only child is a postscript, which is typically left-aligned
-	nodes = dom.xpath("/html/body//footer[not(@data-css-margin-top='1em') or (not(@data-css-text-align) and not(./*[contains(@epub:type, 'z3998:postscript') and not(following-sibling::*) and not(preceding-sibling::*) ]))]")
-	if nodes:
-		messages.append(LintMessage("c-010", "[xhtml]<footer>[/] missing [css]margin-top: 1em; text-align: <value>;[/]. [css]text-align[/] is usually set to [css]right[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
-
-	# Check for elements that have `text-align: center` but also `text-indent: 1em`
-	nodes = dom.xpath("/html/body//*[@data-css-text-align='center' and @data-css-text-indent='1em']")
-	if nodes:
-		messages.append(LintMessage("c-011", "Element with [css]text-align: center;[/] but [css]text-indent[/] is [css]1em[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-	# Check for sections that don't have header elements, and that do not have margin-top: 8em.
-	# This tries to find the first <section> or <article> whose first child is <p>, and that doesn't
-	# have any ancestor <section>s or <articles>s with heading content, and that doesn't have the correct top margin.
-	# Ignore the titlepage, loi, and dedications as they typically have unique styling.
-	nodes = dom.xpath("/html/body//*[name()='section' or name()='article'][1][./p[1][not(preceding-sibling::*)] and not(re:test(@epub:type, '(titlepage|loi|dedication)')) and not(ancestor::*/*[re:test(name(), '^(h[1-6]|header|hgroup)$')]) and not(@data-css-margin-top='20vh')]")
-	if nodes:
-		messages.append(LintMessage("c-012", "Sectioning element without heading content, and without [css]margin-top: 20vh;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	if f"{os.sep}src{os.sep}images{os.sep}" not in root:
+		if self.cover_path and filename.name == self.cover_path.name:
+			# Check that cover is in all caps
+			nodes = svg_dom.xpath("//text[re:test(., '[a-z]')]")
+			if nodes:
+				messages.append(LintMessage("s-002", "Lowercase letters in cover. Cover text must be all uppercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for padding and margin not in .5 increments, except for table headers/cells which usually need .25em increments
-	nodes = dom.xpath("/html/body//*[not(re:test(name(), '^(h[1-6]|td|th)$'))][attribute::*[re:test(local-name(), 'data-css-(margin|padding)')][re:test(., '^[0-9]*\\.[^5]')]]")
-	if nodes:
-		messages.append(LintMessage("c-013", "Element with margin or padding not in increments of [css].5em[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		# Check that titlepage is in all caps
+		if filename.name == "titlepage.svg":
+			nodes = svg_dom.xpath("//text[re:test(., '[a-z]') and not(text()='translated by' or text()='illustrated by' or text()='and')]")
+			if nodes:
+				messages.append(LintMessage("s-003", "Lowercase letters in titlepage. Titlepage text must be all uppercase except [text]translated by[/] and [text]illustrated by[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for <table> without set margins. We ignore tables that are for drama structuring, and tables that are ancestors of <blockquote> because <blockquote> has its own margin.
-	# We also ignore tables whose preceding sibling sets a bottom margin that is not 0.
-	nodes = dom.xpath("/html/body//table[not(./ancestor-or-self::*[contains(@epub:type, 'z3998:drama')]) and not(./ancestor::blockquote) and not(./preceding-sibling::*[1][@data-css-margin-bottom != '0']) and (not(@data-css-margin-top) or not(@data-css-margin-right) or not(@data-css-margin-bottom) or not(@data-css-margin-left))]")
-	if nodes:
-		messages.append(LintMessage("c-014", "[xhtml]<table>[/] element without explicit margins. Most tables need [css]margin: 1em;[/] or [css]margin: 1em auto 1em auto;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	if self.cover_path and filename.name == self.cover_path.name:
+		# Ensure the producer has built the cover.
+		# The default cover image is a white background which when encoded to base64 begins with 299 characters and then has a long string of `A`s
+		if svg_dom.xpath("//image[re:test(@xlink:href, '^.{299}A{50,}')]"):
+			messages.append(LintMessage("m-063", "Cover image has not been built.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for elements that follow salutations, but that don't have text-indent: 0;
-	# We have to check both <p> elems that are salutations, and also <p> elems that have a first-child inline element that is a salutation, and that does not have following siblings that are senders/valedictions, as that might indicate a letter in a prose context
-	nodes = dom.xpath("/html/body//*[(contains(@epub:type, 'z3998:salutation') or ./preceding-sibling::*[1][contains(@epub:type, 'z3998:salutation') or (name() != 'blockquote' and count(./node()[normalize-space(.)]) = 1 and ./*[contains(@epub:type, 'z3998:salutation')])] or ./*[1][contains(@epub:type, 'z3998:salutation') and not((./following-sibling::node()[1][self::text()]))]) and @data-css-text-indent != '0']")
-	if nodes:
-		messages.append(LintMessage("c-015", "Element after or containing [val]z3998:salutation[/] does not have [css]text-indent: 0;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Make images have reasonable dimensions
+	viewbox = svg_dom.xpath("/svg/@viewBox", True)
+	if viewbox:
+		svg_dimensions = viewbox.split()
+		try:
+			if float(svg_dimensions[2]) * float(svg_dimensions[3]) > 4000000:
+				messages.append(LintMessage("f-018", "Image greater than 4,000,000 pixels square in dimension.", se.MESSAGE_TYPE_ERROR, filename))
+		except Exception as ex:
+			raise se.InvalidFileException(f"Couldn’t parse SVG [xhtml]viewBox[/] attribute in [path][link=file://{filename.resolve()}]{filename}[/][/].") from ex
 
-	# Check for block-level postscripts that don't have margin-top: 1em. Exclude postscripts that are the first child of blockquote or footer, since blockquotes/footers gives the desired margin.
-	nodes = dom.xpath("/html/body//*[(name() = 'div' or name() = 'p') and contains(@epub:type, 'z3998:postscript') and not((./parent::blockquote or ./parent::footer) and count(./preceding-sibling::*) = 0) and @data-css-margin-top != '1em']")
+	# Check for illegal transform or id attribute
+	nodes = svg_dom.xpath("//*[@transform or @id]")
 	if nodes:
-		messages.append(LintMessage("c-017", "Element with [val]z3998:postscript[/] semantic, but without [css]margin-top: 1em;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		invalid_transform_attributes = set()
+		invalid_id_attributes = []
+		for node in nodes:
+			if node.get_attr("transform"):
+				invalid_transform_attributes.add(f"transform=\"{node.get_attr('transform')}\"")
 
-	# Check for <p> postscripts that don't have text-indent: 0.
-	nodes = dom.xpath("/html/body//p[contains(@epub:type, 'z3998:postscript') and (not(@data-css-text-align) or @data-css-text-align = 'initial' or @data-css-text-align = 'left') and @data-css-text-indent != '0']")
-	if nodes:
-		messages.append(LintMessage("c-018", "Element with [val]z3998:postscript[/] semantic, but without [css]text-indent: 0;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+			if node.get_attr("id"):
+				invalid_id_attributes.append(f"id=\"{node.get_attr('id')}\"")
 
-	# Check for signature semantic without small caps
-	# Ignore signatures that are only capital letters
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:signature') and not(@data-css-font-variant = 'small-caps') and not(@data-css-font-style = 'italic') and not(re:test(., '^[A-Z\\.\\s]+$'))]")
-	if nodes:
-		messages.append(LintMessage("c-019", "Element with [val]z3998:signature[/] semantic, but without [css]font-variant: small-caps;[/] or [css]font-style: italic;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		if invalid_transform_attributes:
+			messages.append(LintMessage("x-003", "Illegal [xml]transform[/] attribute. SVGs should be optimized to remove use of [xml]transform[/]. Try using Inkscape to save as an “optimized SVG”.", se.MESSAGE_TYPE_ERROR, filename, invalid_transform_attributes))
 
-	# Check for <article>s or <section>s that occur more than once in a file, without page break CSS
-	# Exclude the last element because we don't care if it breaks after the last one.
-	nodes = dom.xpath("/html/body[count(./article) > 1 or count(./section) > 1]/*[(name() = 'article' or name() = 'section') and (not(@data-css-break-after) or @data-css-break-after != 'page') and not(position() = last())]")
-	if nodes:
-		messages.append(LintMessage("c-020", "Multiple [xhtml]<article>[/]s or [xhtml]<section>[/]s in file, but missing [css]break-after: page;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+		if invalid_id_attributes:
+			messages.append(LintMessage("x-014", "Illegal [xml]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, invalid_id_attributes))
 
-	# Check for elements that are set in italics, and whose italic children don't have font-style: normal.
-	nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and ./*[(name()='i' or name()='em') and @data-css-font-style='italic']]")
+	# Check for fill: #000 which should simply be removed
+	nodes = svg_dom.xpath("//*[contains(@fill, '#000') or contains(translate(@style, ' ', ''), 'fill:#000')]")
 	if nodes:
-		messages.append(LintMessage("c-021", "Element with [css]font-style: italic;[/], but child [xhtml]<i>[/] or [xhtml]<em>[/] does not have [css]font-style: normal;[/]. Hint: Italics within italics are typically set in Roman for contrast; if that’s not the case here, can [xhtml]<i>[/] be removed while still preserving italics and semantic inflection?", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("x-004", "Illegal [xml]style=\"fill: #000\"[/] or [xml]fill=\"#000\"[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for tables that appear to be summing/listing numbers, but that are missing font-variant-numeric styling
-	nodes = dom.xpath("/html/body//table[count(.//td[not(following-sibling::*) and re:test(., '^[0-9]+$') and not(@data-css-font-variant-numeric='tabular-nums')]) >= 2]")
-	if nodes:
-		messages.append(LintMessage("c-026", "Table that appears to be listing numbers, but without [css]font-variant-numeric: tabular-nums;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+	# Check for illegal height or width on root <svg> element
+	if filename.name != "logo.svg": # Do as I say, not as I do...
+		if svg_dom.xpath("/svg[@height or @width]"):
+			messages.append(LintMessage("x-005", "Illegal [xml]height[/] or [xml]width[/] attribute on root [xml]<svg>[/] element. Size SVGs using the [xml]viewBox[/] attribute only.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for three-em-dashes not preceded by a space
-	nodes = dom.xpath(f"/html/body//p[re:test(., '[^>“(\\s{se.WORD_JOINER}]{se.WORD_JOINER}?⸻')]")
-	if nodes:
-		messages.append(LintMessage("y-033", "Possible typo: Three-em-dash obscuring an entire word, but not preceded by a space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	match = regex.search(r"viewbox", file_contents, flags=regex.IGNORECASE)
+	if match and match[0] != "viewBox":
+		messages.append(LintMessage("x-006", f"[xml]{match}[/] found instead of [xml]viewBox[/]. [xml]viewBox[/] must be correctly capitalized.", se.MESSAGE_TYPE_ERROR, filename))
 
 	return messages
 
-def _lint_metadata_checks(filename: Path, dom: se.easy_xml.EasyXmlTree) -> list:
+def _lint_special_file_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_flags: dict, special_file: str) -> list:
 	"""
-	Helper function used in self.lint()
-	Process metadata checks on an .xhtml file
+	Process error checks in “special” .xhtml files
 
 	INPUTS
 	filename: The name of the file being checked
-	dom: A dom tree to check
+	dom: The dom of the file being checked
+	file_contents: The contents of the file being checked
+	ebook_flags: A dictionary containing ebook information
+	special_file: A string identifying the type of special file being checked
+	self
 
 	OUTPUTS
-	A list of LintMessages representing metadata errors found in the file
+	A list of LintMessage objects.
 	"""
 
 	messages = []
+	source_links = self.metadata_dom.xpath("/package/metadata/dc:source/text()")
 
-	# Check for missing MARC relators
-	# Don't check the landmarks as that may introduce duplicate errors
-	# Only check for top-level elements, to avoid intros in short stories or poems in compilation.
-	if not dom.xpath("/html/body//nav[contains(@epub:type, 'landmarks')]"):
-		if dom.xpath("/html/body/*[contains(@epub:type, 'introduction') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and (text()='aui' or text()='win')]"):
-			messages.append(LintMessage("m-030", "[val]introduction[/] semantic inflection found, but no MARC relator [val]aui[/] (Author of introduction, but not the chief author) or [val]win[/] (Writer of introduction).", se.MESSAGE_TYPE_WARNING, filename))
+	if special_file in ("colophon", "imprint"):
+		if len(source_links) <= 2:
+			# Check that links back to sources are represented correctly
+			nodes = dom.xpath("/html/body//a[@href='https://www.pgdp.net' and text()!='The Online Distributed Proofreading Team']")
+			if nodes:
+				messages.append(LintMessage("m-071", "DP link must be exactly [text]The Online Distributed Proofreading Team[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-		if dom.xpath("/html/body/*[contains(@epub:type, 'preface') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='wpr']"):
-			messages.append(LintMessage("m-031", "[val]preface[/] semantic inflection found, but no MARC relator [val]wpr[/] (Writer of preface).", se.MESSAGE_TYPE_WARNING, filename))
+			nodes = dom.xpath("/html/body//a[re:test(@href, '^https://www.pgdp.org/ols/') and text()!='Distributed Proofreaders Open Library System']")
+			if nodes:
+				messages.append(LintMessage("m-072", "DP OLS link must be exactly [text]Distributed Proofreaders Open Library System[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-		if dom.xpath("/html/body/*[contains(@epub:type, 'afterword') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='aft']"):
-			messages.append(LintMessage("m-032", "[val]afterword[/] semantic inflection found, but no MARC relator [val]aft[/] (Author of colophon, afterword, etc.).", se.MESSAGE_TYPE_WARNING, filename))
+			nodes = dom.xpath("/html/body//a[re:test(@href, '^https://[^\"]*?hathitrust.org') and re:test(text(), '[Hh]athi') and not(text()='HathiTrust Digital Library')]")
+			if nodes:
+				messages.append(LintMessage("m-041", "Hathi Trust link text must be exactly [text]HathiTrust Digital Library[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-		if dom.xpath("/html/body/*[contains(@epub:type, 'endnotes') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='ann']"):
-			messages.append(LintMessage("m-033", "[val]endnotes[/] semantic inflection found, but no MARC relator [val]ann[/] (Annotator).", se.MESSAGE_TYPE_WARNING, filename))
+		nodes = dom.xpath("/html/body//p[re:test(., '\\ba transcription\\b') and ./a[contains(@href, '#transcriptions')]]")
+		if nodes:
+			messages.append(LintMessage("t-071", "Multiple transcriptions listed, but preceding text is [text]a transcription[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-		if dom.xpath("/html/body/*[contains(@epub:type, 'loi') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='ill']"):
-			messages.append(LintMessage("m-034", "[val]loi[/] semantic inflection found, but no MARC relator [val]ill[/] (Illustrator).", se.MESSAGE_TYPE_WARNING, filename))
+	if special_file == "colophon":
+		# Check for illegal Wikipedia URLs
+		nodes = dom.xpath("/html/body//a[contains(@href, '.m.wikipedia.org')]")
+		if nodes:
+			messages.append(LintMessage("m-026", "Illegal [url]https://*.m.wikipedia.org[/] URL. Hint: use non-mobile Wikipedia URLs.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check that `Internet Archive` and `HathiTrust` are preceded by `the`. They might have an immediate preceding text node that ends in `the`, or they
-	# might be preceded by a white space node, then a <br/>, then a text node ending in `the`.
-	nodes = dom.xpath("/html/body//a[(re:test(text(), '[Hh]athi') and re:test(@href, '^https://[^\"]*?hathitrust.org')) or (re:test(text(), 'Internet Archive') and re:test(@href, 'https://[^\"]*?archive.org'))][(preceding-sibling::node()[2][not(self::br)] and preceding-sibling::node()[1][not(re:test(., '\\sthe $'))]) or (preceding-sibling::node()[2][self::br] and preceding-sibling::node()[3][not(re:test(., '\\sthe$'))]) ]")
-	if nodes:
-		messages.append(LintMessage("m-061", "Link must be preceded by [text]the[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# Check for non-English Wikipedia URLs
+		nodes = dom.xpath("/html/body//a[re:test(@href, 'https://(?!en)[a-z]{2,}\\.wikipedia\\.org')]")
+		if nodes:
+			messages.append(LintMessage("m-043", "Non-English Wikipedia URL.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
 
-	return messages
+		# Check for wrong grammar filled in from template
+		nodes = dom.xpath("/html/body//a[starts-with(@href, 'https://books.google.com/') or starts-with(@href, 'https://www.google.com/books/')][(preceding-sibling::text()[normalize-space(.)][1])[re:test(., '\\bthe$')]]")
+		if nodes:
+			messages.append(LintMessage("s-016", "Incorrect [text]the[/] before Google Books link.", se.MESSAGE_TYPE_ERROR, filename, ["the<br/>\n" + node.to_string() for node in nodes]))
 
-def _lint_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_info: dict, language: str, section_tree: List[EbookSection]) -> list:
-	"""
-	Helper function used in self.lint()
-	Process syntax checks on an .xhtml file
+		se_url = self.generated_identifier.replace("url:", "")
+		if not dom.xpath(f"/html/body//a[@href='{se_url}' and text()='{se_url.replace('https://', '')}']"):
+			messages.append(LintMessage("m-035", f"Unexpected S.E. identifier in colophon. Expected: [url]{se_url}[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	INPUTS
-	filename: The name of the file being checked
-	dom: A dom tree to check
-	file_contents: The contents of the file being checked
-	ebook_info: A dictionary containing several pieces of information about an ebook
+		if self.metadata_dom.xpath("/package/metadata/meta[@property='role' and text()='trl']") and "translated from" not in file_contents:
+			messages.append(LintMessage("m-025", "Translator found in metadata, but no [text]translated from LANG[/] block in colophon.", se.MESSAGE_TYPE_ERROR, filename))
 
-	OUTPUTS
-	A list of LintMessages representing syntax errors found in the file
-	"""
+		if ebook_flags["has_multiple_transcriptions"] and not dom.xpath("/html/body//a[contains(@href, '#transcriptions')]"):
+			messages.append(LintMessage("m-074", "Multiple transcriptions found in metadata, but no link to [text]EBOOK_URL#transcriptions[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	messages = []
+		if ebook_flags["has_multiple_page_scans"] and not dom.xpath("/html/body//a[contains(@href, '#page-scans')]"):
+			messages.append(LintMessage("m-075", "Multiple page scans found in metadata, but no link to [text]EBOOK_URL#page-scans[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for numeric entities
-	matches = regex.findall(r"&#[0-9]+?;", file_contents)
-	if matches:
-		messages.append(LintMessage("s-001", "Illegal numeric entity (like [xhtml]Α[/]).", se.MESSAGE_TYPE_ERROR, filename))
+		# Check that the formula changed from the default if we added 'various sources'
+		if ebook_flags["has_multiple_transcriptions"] or ebook_flags["has_multiple_page_scans"]:
+			nodes = dom.xpath("/html/body//a[text() = 'various sources' and not(re:test(preceding-sibling::br[1]/preceding-sibling::node()[1], '(digital scans|transcriptions) from\\s*$'))]")
+			if nodes:
+				messages.append(LintMessage("t-072", "[text]various sources[/] link not preceded by [text]from[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check nested <blockquote> elements, but only if it's the first child of another <blockquote>
-	nodes = dom.xpath("/html/body//blockquote/*[1][name()='blockquote']")
-	if nodes:
-		messages.append(LintMessage("s-005", "Nested [xhtml]<blockquote>[/] element.", se.MESSAGE_TYPE_WARNING, filename))
+		# Check if we forgot to fill any variable slots
+		missing_colophon_vars = [var for var in SE_VARIABLES if regex.search(fr"\b{var}\b", file_contents)]
+		if missing_colophon_vars:
+			messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, filename, missing_colophon_vars))
 
-	# Check to see if we've marked something as poetry or verse, but didn't include a first <span>
-	# This xpath selects the p elements, whose parents are poem/verse, and whose first child is not a span
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')]/p[not(./*[name()='span' and position()=1])]")
-	if nodes:
-		matches = []
-		for node in nodes:
-			# Get the first line of the poem, if it's a text node, so that we can include it in the error messages.
-			# If it's not a text node then just ignore it and add the error anyway.
-			first_line = node.xpath("./descendant-or-self::text()[normalize-space(.)]", True)
-			if first_line:
-				match = first_line.strip()
-				if match: # Make sure we don't append an empty string
-					matches.append(match)
+		# Check that we have <br/>s at the end of lines
+		# First, check for b or a elements that are preceded by a newline but not by a br
+		nodes = [node.to_string() for node in dom.xpath("/html/body/section/p/*[name()='b' or name()='a'][(preceding-sibling::node()[1])[contains(., '\n')]][not((preceding-sibling::node()[2])[self::br]) or (normalize-space(preceding-sibling::node()[1]) and re:test(preceding-sibling::node()[1], '\\n\\s*$')) ]")]
+		# Next, check for text nodes that contain newlines but are not preceded by brs
+		nodes = nodes + [node.strip() for node in dom.xpath("/html/body/section/p/text()[contains(., '\n') and normalize-space(.)][(preceding-sibling::node()[1])[not(self::br)]]")]
+		if nodes:
+			messages.append(LintMessage("s-053", "Colophon line not preceded by [xhtml]<br/>[/].", se.MESSAGE_TYPE_ERROR, filename, nodes))
 
-		messages.append(LintMessage("s-006", "Poem or verse [xhtml]<p>[/] (stanza) without [xhtml]<span>[/] (line) element.", se.MESSAGE_TYPE_WARNING, filename, matches))
+		# Is there a comma after a producer name, if there's only two producers?
+		nodes = dom.xpath("/html/body/section/p/*[name()='b' or name()='a'][(following-sibling::node()[1])[normalize-space(.)=', and']][(preceding-sibling::*[1])[name()='br']]")
+		if nodes:
+			messages.append(LintMessage("t-006", "Comma after producer name, but there are only two producers.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for elements that don't have a direct block child
-	# Allow white space and comments before the first child
-	# Ignore children of the ToC and landmarks as they do not require p children
-	nodes = dom.xpath("/html/body//*[(name()='blockquote' or name()='dd' or name()='header' or name()='li' or name()='footer') and not(./ancestor::*[contains(@epub:type, 'toc') or contains(@epub:type, 'landmarks')]) and (node()[normalize-space(.) and not(self::comment())])[1][not(name()='p' or name()='blockquote' or name()='div' or name()='table' or name()='header' or name()='ul' or name()='ol' or name()='footer' or name()='hgroup' or re:test(name(), '^h[0-6]'))]]")
-	if nodes:
-		messages.append(LintMessage("s-007", "Element requires at least one block-level child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and re:test(@epub:type, 'z3998:.*?name')]")
+		if nodes:
+			messages.append(LintMessage("s-092", "Anonymous contributor with [val]z3998:*-name[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for block-level tags that end with <br/>
-	nodes = dom.xpath("/html/body//*[self::p or self::blockquote or self::table or self::ol or self::ul or self::section or self::article][br[last()][not(following-sibling::text()[normalize-space()])][not(following-sibling::*)]]")
-	if nodes:
-		messages.append(LintMessage("s-008", "[xhtml]<br/>[/] element found before closing tag of block-level element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		# Check for anonymous volunteers misrepresented in the colophon. Note that we only match SE producers and transcribers, because
+		# a cover art artist can also be anonymous but they're not volunteers.
+		nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and text() != 'An Anonymous Volunteer' and (preceding-sibling::a[@href='https://standardebooks.org'])]")
+		if nodes:
+			messages.append(LintMessage("s-100", "Anonymous digital contributor value not exactly [text]An Anonymous Volunteer[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for hgroup elements with only one child
-	nodes = dom.xpath("/html/body//hgroup[count(*)=1]")
-	if nodes:
-		messages.append(LintMessage("s-009", "[xhtml]<hgroup>[/] element with only one child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# Check for primary contributors misrepresented in the colophon. these differ from ebook production volunteers.
+		nodes = dom.xpath("/html/body//b[re:test(., 'anonymous', 'i') and text() != 'Anonymous' and (preceding-sibling::node()[contains(., 'The cover page')] or preceding-sibling::i[contains(@epub:type, 'se:name.publication')])]")
+		if nodes:
+			messages.append(LintMessage("s-101", "Anonymous primary contributor value not exactly [text]Anonymous[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for empty elements. Elements are empty if they have no children and no non-whitespace text
-	# Use name() to check for MathML elements, because if the MathML namespace is not declared in a given file and we search by the m: namespace then this xpath will fail entirely.
-	nodes = dom.xpath("/html/body//*[not(self::br) and not(self::hr) and not(self::img) and not(self::td) and not(self::th) and not(self::link) and not(self::col) and not(self::colgroup) and not(local-name()='none') and not(local-name()='mspace') and not(local-name()='mprescripts')][not(./*)][not(normalize-space())][not(./ancestor::*[local-name()='annotation-xml'])]")
-	if nodes:
-		messages.append(LintMessage("s-010", "Empty element. Use [xhtml]<hr/>[/] for thematic breaks if appropriate.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		# Is there a page scan link in the colophon, but missing in the metadata?
+		for node in dom.xpath("/html/body//a[re:test(@href, '(gutenberg\\.org/ebooks/[0-9]+|hathitrust\\.org|/archive\\.org|books\\.google\\.com|www\\.google\\.com/books/)')]"):
+			if not self.metadata_dom.xpath(f"/package/metadata/dc:source[contains(text(), {se.easy_xml.escape_xpath(node.get_attr('href'))})]"):
+				messages.append(LintMessage("m-059", f"Link to [url]{node.get_attr('href')}[/] found in colophon, but missing matching [xhtml]dc:source[/] element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	# Check for <section> and <article> without ID attribute
-	nodes = dom.xpath("/html/body//*[self::section or self::article][not(@id)]")
-	if nodes:
-		messages.append(LintMessage("s-011", "Element without [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		# Are the sources represented correctly?
+		# We don't have a standard yet for more than two sources (transcription and scan) so just ignore that case for now.
+		# We can't merge this with the imprint check because imprint doesn't have `<br/>` between `the`
+		if not ebook_flags["has_multiple_transcriptions"] and not ebook_flags["has_other_sources"]:
+			for link in source_links:
+				if "gutenberg.org" in link and f"<a href=\"{link}\">Project Gutenberg</a>" not in file_contents:
+					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Project Gutenberg</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for <pre> tags
-	if dom.xpath("/html/body//pre"):
-		messages.append(LintMessage("s-013", "Illegal [xhtml]<pre>[/] element.", se.MESSAGE_TYPE_ERROR, filename))
+		if not ebook_flags["has_multiple_page_scans"] and not ebook_flags["has_other_sources"]:
+			for link in source_links:
+				if "hathitrust.org" in link and f"the<br/>\n\t\t\t<a href=\"{link}\">HathiTrust Digital Library</a>" not in file_contents:
+					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]the<br/> <a href=\"{link}\">HathiTrust Digital Library</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for <br/> after block-level elements
-	nodes = dom.xpath("/html/body//*[self::p or self::blockquote or self::table or self::ol or self::ul or self::section or self::article][following-sibling::br]")
-	if nodes:
-		messages.append(LintMessage("s-014", "[xhtml]<br/>[/] after block-level element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+				if "archive.org" in link and f"the<br/>\n\t\t\t<a href=\"{link}\">Internet Archive</a>" not in file_contents:
+					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]the<br/> <a href=\"{link}\">Internet Archive</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for <hr> tags before the end of a section, which is a common PG artifact
-	if dom.xpath("/html/body//hr[count(following-sibling::*)=0]"):
-		messages.append(LintMessage("s-012", "Illegal [xhtml]<hr/>[/] as last child.", se.MESSAGE_TYPE_ERROR, filename))
+				if ("books.google.com" in link or "www.google.com/books/" in link) and f"<a href=\"{link}\">Google Books</a>" not in file_contents:
+					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Google Books</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for hgroup elements with a subtitle but no title
-	nodes = dom.xpath("/html/body//hgroup[./*[contains(@epub:type, 'subtitle')] and not(./*[contains(concat(' ', @epub:type, ' '), ' title ')])]")
-	if nodes:
-		messages.append(LintMessage("s-015", "Element has [val]subtitle[/] semantic, but without a sibling having a [val]title[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# If we're in the imprint, are the sources represented correctly?
+	# We don't have a standard yet for more than two sources (transcription and scan) so just ignore that case for now.
+	elif special_file == "imprint":
+		# Check for wrong grammar filled in from template
+		nodes = dom.xpath("/html/body//a[starts-with(@href, 'https://books.google.com/') or starts-with(@href, 'https://www.google.com/books/')][(preceding-sibling::node()[1])[re:test(., 'the\\s+$')]]")
+		if nodes:
+			messages.append(LintMessage("s-016", "Incorrect [text]the[/] before Google Books link.", se.MESSAGE_TYPE_ERROR, filename, ["the " + node.to_string() for node in nodes]))
 
-	# Check for deprecated MathML elements
-	# Note we dont select directly on element name, because we want to ignore any namespaces that may (or may not) be defined
-	nodes = dom.xpath("/html/body//*[local-name()='mfenced']")
-	if nodes:
-		messages.append(LintMessage("s-017", "[xhtml]<m:mfenced>[/] is deprecated in the MathML spec. Use [xhtml]<m:mrow><m:mo fence=\"true\">(</m:mo>...<m:mo fence=\"true\">)</m:mo></m:mrow>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		# Check if we forgot to fill any variable slots
+		missing_imprint_vars = [var for var in SE_VARIABLES if regex.search(fr"\b{var}\b", file_contents)]
+		if missing_imprint_vars:
+			messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, filename, missing_imprint_vars))
 
-	# Check for <figure> tags without id attributes
-	nodes = dom.xpath("/html/body//img[@id]")
-	if nodes:
-		messages.append(LintMessage("s-018", "[xhtml]<img>[/] element with [attr]id[/] attribute. [attr]id[/] attributes go on parent [xhtml]<figure>[/] elements. Hint: Images that are inline (i.e. that do not have parent [xhtml]<figure>[/]s do not have [attr]id[/] attributes.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		if ebook_flags["has_multiple_transcriptions"] and not dom.xpath("/html/body//a[contains(@href, '#transcriptions')]"):
+			messages.append(LintMessage("m-074", "Multiple transcriptions found in metadata, but no link to [text]EBOOK_URL#transcriptions[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for IDs on <h#> tags
-	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][@id]")
-	if nodes:
-		messages.append(LintMessage("s-019", "[xhtml]<h#>[/] element with [attr]id[/] attribute. [xhtml]<h#>[/] elements should be wrapped in [xhtml]<section>[/] elements, which should hold the [attr]id[/] attribute.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+		if ebook_flags["has_multiple_page_scans"] and not dom.xpath("/html/body//a[contains(@href, '#page-scans')]"):
+			messages.append(LintMessage("m-075", "Multiple page scans found in metadata, but no link to [text]EBOOK_URL#page-scans[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for legal cases that aren't italicized
-	# We can't use this because v. appears as short for "volume", and we may also have sporting events without italics.
-	#nodes = dom.xpath("/html/body//abbr[text()='v.' or text()='versus'][not(parent::i)]")
-	#if nodes:
-	#	messages.append(LintMessage("t-xxx", "Legal case without parent [xhtml]<i>[/].", se.MESSAGE_TYPE_WARNING, filename, {f"{node.to_string()}." for node in nodes}))
+		# Check that the formula changed from the default if we added 'various sources'
+		if ebook_flags["has_multiple_transcriptions"] or ebook_flags["has_multiple_page_scans"]:
+			nodes = dom.xpath("/html/body//a[text() = 'various sources' and not(re:test(preceding-sibling::node()[1], '(digital scans|transcriptions) from\\s*$'))]")
+			if nodes:
+				messages.append(LintMessage("t-072", "[text]various sources[/] link not preceded by [text]from[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Only do this check if there's one <h#> or one <hgroup> tag. If there's more than one, then the xhtml file probably requires an overarching title
-	# We merge two xpaths here because <h#>/<hgroup> can be either a direct child of <section>, or it could be nested in <header>
-	if len(dom.xpath("/html/body/*[name()='section' or name()='article']/*[re:test(name(), '^h[1-6]$') or name()='hgroup'] | /html/body/*[name()='section' or name()='article']/header/*[re:test(name(), '^h[1-6]$') or name()='hgroup']"))==1:
-		title = se.formatting.generate_title(dom)
+		# Check for correctly named links. We can't merge this with the colophon check because the colophon breaks `the` with `<br/>`
+		if not ebook_flags["has_multiple_transcriptions"] and not ebook_flags["has_other_sources"]:
+			for link in source_links:
+				if "gutenberg.org" in link and f"<a href=\"{link}\">Project Gutenberg</a>" not in file_contents:
+					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Project Gutenberg</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-		if not dom.xpath(f"/html/head/title[text()={se.easy_xml.escape_xpath(title.replace('&', '&'))}]"):
-			messages.append(LintMessage("s-021", f"Unexpected value for [xhtml]<title>[/] element. Expected: [text]{title}[/]. (Beware hidden Unicode characters!)", se.MESSAGE_TYPE_ERROR, filename))
+		if not ebook_flags["has_multiple_page_scans"] and not ebook_flags["has_other_sources"]:
+			for link in source_links:
+				if "hathitrust.org" in link and f"the <a href=\"{link}\">HathiTrust Digital Library</a>" not in file_contents:
+					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: the [xhtml]<a href=\"{link}\">HathiTrust Digital Library</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check to see if <h#> tags are correctly titlecased
-	# Ignore <h#> tags with an `xml:lang` attribute, as other languages have different titlecasing rules
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') or (name() = 'p' and parent::hgroup)][not(contains(@epub:type, 'z3998:roman')) and not(@xml:lang)]")
-	for node in nodes:
-		node_copy = deepcopy(node)
+				if "archive.org" in link and f"the <a href=\"{link}\">Internet Archive</a>" not in file_contents:
+					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: the [xhtml]<a href=\"{link}\">Internet Archive</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-		# Remove *leading* Roman spans
-		# This matches the first child node excluding white space nodes, if it contains the z3998:roman semantic.
-		for element in node_copy.xpath("./node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman')]"):
-			element.remove()
+				if ("books.google.com" in link or "www.google.com/books/" in link) and f"<a href=\"{link}\">Google Books</a>" not in file_contents:
+					messages.append(LintMessage("m-037", f"Transcription/page scan source link not found. Expected: [xhtml]<a href=\"{link}\">Google Books</a>[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-		# Remove noterefs
-		for element in node_copy.xpath(".//a[contains(@epub:type, 'noteref')]"):
-			element.remove()
+	# Endnote checks
+	elif special_file == "endnotes":
+		# Do we have to replace Ibid.? Only match Ibid. if the endnote does not appear to cite any names.
+		nodes = dom.xpath("/html/body//li[re:test(@epub:type, '\\bendnote\\b')]//abbr[re:test(., '\\b[Ii]bid\\b') and not(ancestor::li[1]//*[contains(@epub:type, 'se:name')])]")
+		if nodes:
+			messages.append(LintMessage("s-039", "[text]Ibid[/] in endnotes. “Ibid” means “The previous reference” which is meaningless with popup endnotes, and must be replaced by the actual thing [text]Ibid[/] refers to, unless it refers to text within the same endnote.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		# Remove hidden elements, for example in poetry identified by first line (keats)
-		for element in node_copy.xpath(".//*[@hidden]"):
-			element.remove()
+		# Check that endnotes have their backlink in the last <p> element child of the <li>. This also highlights backlinks that are totally missing.
+		nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')][./p[last()][not(a[contains(@epub:type, 'backlink')])] or not(./p[last()])]")
+		if nodes:
+			messages.append(LintMessage("s-056", "Last [xhtml]<p>[/] child of endnote missing backlink.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-		# Unwrap (not remove) any subelements in subtitles, as they may trip up further regexes
-		for element in node_copy.xpath(".//*[contains(@epub:type, 'subtitle')]//*"):
-			element.unwrap()
+		# Make sure the backlink points to the same note number as the parent endnote ID
+		nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]//a[contains(@epub:type, 'backlink')][not(re:match(@href, '\\-[0-9]+$')=re:match(ancestor::li/@id, '\\-[0-9]+$'))]")
+		if nodes:
+			messages.append(LintMessage("s-057", "Backlink noteref fragment identifier doesn’t match endnote number.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-		title = node_copy.inner_xml()
+		# Check that citations at the end of endnotes are in a <cite> element. If not typogrify will run the last space together with the em dash.
+		# This tries to catch that, but limits the match to 20 chars so that we don't accidentally match a whole sentence that happens to be at the end of an endnote.
+		nodes = dom.xpath(f"/html/body//li[contains(@epub:type, 'endnote')]/p[last()][re:test(., '\\.”?{se.WORD_JOINER}?—[A-Z].{{0,20}}\\s*↩$')]")
+		if nodes:
+			messages.append(LintMessage("s-064", "Endnote citation not wrapped in [xhtml]<cite>[/]. Em dashes go within [xhtml]<cite>[/] and it is preceded by one space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		# Remove leading leftover spacing and punctuation
-		title = regex.sub(r"^[\s\.\,\!\?\:\;]*", "", title)
+		# Did we forget the endnote semantic on li elements?
+		nodes = dom.xpath("/html/body/section/ol/li[not(re:test(@epub:type, '\\bendnote\\b'))]")
+		if nodes:
+			messages.append(LintMessage("s-099", "List item in endnotes missing [xhtml]endnote[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-		# Normalize whitespace
-		title = regex.sub(r"\s+", " ", title, flags=regex.DOTALL).strip()
+		# Match backlink elements whose preceding node doesn't end with ' ', and is also not all whitespace
+		nodes = dom.xpath("/html/body//a[@epub:type='backlink'][(preceding-sibling::node()[1])[not(re:test(., ' $')) and not(normalize-space(.)='')]]")
+		if nodes:
+			messages.append(LintMessage("t-027", "Endnote backlink not preceded by exactly one space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		# Do we have a subtitle? If so the first letter of that must be capitalized, so we pull that out
-		subtitle_matches = regex.findall(r"(.*?)<span epub:type=\"subtitle\">(.*?)</span>(.*?)", title, flags=regex.DOTALL)
-		if subtitle_matches:
-			for title_header, subtitle, title_footer in subtitle_matches:
-				title_header = se.formatting.titlecase(se.formatting.remove_tags(title_header).strip())
-				subtitle = se.formatting.titlecase(se.formatting.remove_tags(subtitle).strip())
-				title_footer = se.formatting.titlecase(se.formatting.remove_tags(title_footer).strip())
+	# Check LoI descriptions to see if they match associated figcaptions
+	elif special_file == "loi":
+		for node in dom.xpath("/html/body/nav[contains(@epub:type, 'loi')]//li//a"):
+			figure_ref = node.get_attr("href").split("#")[1]
+			chapter_ref = regex.findall(r"(.*?)#.*", node.get_attr("href"))[0]
+			figcaption_text = ""
+			loi_text = node.inner_text()
+			file_dom = self.get_dom(self.content_path / "text" / chapter_ref)
 
-				titlecased_title = f"{title_header} {subtitle} {title_footer}"
-				titlecased_title = titlecased_title.strip()
+			try:
+				figure = file_dom.xpath(f"//*[@id={se.easy_xml.escape_xpath(figure_ref)}]")[0]
+			except Exception:
+				messages.append(LintMessage("s-040", f"[attr]#{figure_ref}[/] not found in file [path][link=file://{self.path / 'src/epub/text' / chapter_ref}]{chapter_ref}[/][/].", se.MESSAGE_TYPE_ERROR, filename))
+				continue
 
-				title = se.formatting.remove_tags(title).strip()
-				if title != titlecased_title:
-					messages.append(LintMessage("s-023", f"Title [text]{title}[/] not correctly titlecased. Expected: [text]{titlecased_title}[/].", se.MESSAGE_TYPE_WARNING, filename))
+			for child in figure.xpath("./*"):
+				if child.tag == "img":
+					figure_img_alt = child.get_attr("alt")
 
-		# No subtitle? Much more straightforward
-		else:
-			titlecased_title = se.formatting.titlecase(se.formatting.remove_tags(title))
-			title = se.formatting.remove_tags(title)
-			if title != titlecased_title:
-				messages.append(LintMessage("s-023", f"Title [text]{title}[/] not correctly titlecased. Expected: [text]{titlecased_title}[/].", se.MESSAGE_TYPE_WARNING, filename))
+				if child.tag == "figcaption":
+					figcaption_text = child.inner_text()
 
-	# Check for header elements that are entirely non-English
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./i[@xml:lang][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
-	if nodes:
-		messages.append(LintMessage("s-024", "Header elements that are entirely non-English should not be set in italics. Instead, the [xhtml]<h#>[/] element has the [attr]xml:lang[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+					# Replace tabs and newlines with a single space to better match figcaptions that contain <br/>
+					figcaption_text = regex.sub(r"(\n|\t)", " ", figcaption_text)
+					figcaption_text = regex.sub(r"[ ]+", " ", figcaption_text)
 
-	# Check for z3998:roman elements with invalid values. Roman numerals can occasionally end in `j` as an alias for ending `i`. See _The Worm Ouroboros_.
-	# We also allow the numeral to end in a digit, because that might be an endnote. For example `<h2 epub:type="ordinal z3998:roman">II<a href="..." epub:type="noteref">3</a></h2>`
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:roman') and not(re:test(normalize-space(.), '^[ivxlcdmIVXLCDM]+j?[0-9]*?$'))]")
-	if nodes:
-		messages.append(LintMessage("s-026", "Invalid Roman numeral.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+			if (figcaption_text != "" and loi_text != "" and figcaption_text != loi_text) and (figure_img_alt != "" and loi_text != "" and figure_img_alt != loi_text):
+				messages.append(LintMessage("s-041", f"The [xhtml]<figcaption>[/] element of [attr]#{figure_ref}[/] does not match the text in its LoI entry.", se.MESSAGE_TYPE_WARNING, self.path / "src/epub/text" / chapter_ref))
 
-	# Check for <h#> elements that are higher or lower than their expected level based on how deep they are in <section>s
-	# or <article>s. Exclude <nav> in the ToC.
-	# Get all h# elements not preceded by other h# elements (this includes the 1st item of an hgroup but not the next items)
-	# Exclude the ToC and landmarks
-	invalid_headers = []
-	invalid_parent_ids = []
-	for heading in dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') and not(./preceding-sibling::*[re:test(name(), '^h[1-6]$')]) and not(./ancestor::*[re:test(@epub:type, '\\b(toc|landmarks)\\b')])]"):
-		# Get the IDs of the direct parent section
-		parent_section = heading.xpath("./ancestor::*[@id and re:test(name(), '^(section|article|nav)$')][1]")
+	return messages
 
-		if parent_section:
-			section = _find_ebook_section(parent_section[0].get_attr("id"), section_tree)
+def _lint_xhtml_css_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, local_css_path: Path) -> list:
+	"""
+	Helper function used in self.lint()
+	Process CSS checks on an .xhtml file
 
-			if not section:
-				# We can accidentally raise s-029 if the file isn't in the spine. This is technically correct, but s-029 is a
-				# misleading message in that case, and f-007 will also be raised, which is the correct message.
-				if filename in self.spine_file_paths:
-					invalid_parent_ids.append(parent_section[0].to_tag_string())
-			else:
-				if str(section.depth) != heading.tag[1:2]:
-					invalid_headers.append(heading.to_string())
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
 
-	if invalid_parent_ids:
-		messages.append(LintMessage("s-029", "Section with [attr]data-parent[/] attribute, but no section having that [attr]id[/] in ebook.", se.MESSAGE_TYPE_ERROR, filename, invalid_parent_ids))
-	if invalid_headers:
-		messages.append(LintMessage("s-085", "[xhtml]<h#>[/] element found in a [xhtml]<section>[/] or a [xhtml]<article>[/] at an unexpected level. Hint: Headings not in the title page start at [xhtml]<h2>[/]. If this work has parts, should this header be [xhtml]<h3>[/] or higher?", se.MESSAGE_TYPE_ERROR, filename, invalid_headers))
+	OUTPUTS
+	A list of LintMessages representing CSS errors found in the file
+	"""
 
-	# Check for a common typo
-	if "z3998:nonfiction" in file_contents:
-		messages.append(LintMessage("s-030", "[val]z3998:nonfiction[/] should be [val]z3998:non-fiction[/].", se.MESSAGE_TYPE_ERROR, filename))
+	messages = []
 
-	# Run some checks on epub:type values
-	incorrect_attrs = set()
-	unnecessary_z3998_attrs = set()
-	duplicate_attrs = []
-
-	for node in dom.xpath("//*[@epub:type]"):
-		attrs = set()
-
-		for val in regex.split(r"\s+", node.get_attr("epub:type")):
-			if val in attrs:
-				duplicate_attrs.append(node.to_tag_string())
-			else:
-				attrs.add(val)
-
-			if val.startswith("z3998:"):
-				bare_val = val.replace("z3998:", "")
-				if bare_val not in Z3998_SEMANTIC_VOCABULARY:
-					incorrect_attrs.add(val)
-
-				elif bare_val in EPUB_SEMANTIC_VOCABULARY:
-					unnecessary_z3998_attrs.add((val, bare_val))
-
-			elif val.startswith("se:"):
-				bare_val = val.replace("se:", "")
-				if bare_val not in SE_SEMANTIC_VOCABULARY:
-					incorrect_attrs.add(val)
-
-			else:
-				# Regular epub vocabulary
-				if val not in EPUB_SEMANTIC_VOCABULARY:
-					incorrect_attrs.add(val)
-
-	if duplicate_attrs:
-		messages.append(LintMessage("s-031", "Duplicate value in [attr]epub:type[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, duplicate_attrs))
-	if incorrect_attrs:
-		messages.append(LintMessage("s-032", "Invalid value for [attr]epub:type[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, incorrect_attrs))
-	if unnecessary_z3998_attrs:
-		messages.append(LintMessage("s-034", "Semantic used from the z3998 vocabulary, but the same semantic exists in the EPUB vocabulary.", se.MESSAGE_TYPE_ERROR, filename, [attr for (attr, bare_attr) in unnecessary_z3998_attrs]))
-
-	# Check if language tags in individual files match the language in the metadata file
-	if filename.name not in IGNORED_FILENAMES:
-		file_language = dom.xpath("/html/@xml:lang", True)
-		if language != file_language:
-			messages.append(LintMessage("s-033", f"File language is [val]{file_language}[/], but [path][link=file://{self.metadata_file_path}]{self.metadata_file_path.name}[/][/] language is [val]{language}[/].", se.MESSAGE_TYPE_WARNING, filename))
+	# Do we have any elements that have specified border color?
+	# `transparent` and `none` are allowed values for border-color
+	if dom.xpath("/html/body//*[attribute::*[re:test(local-name(), 'data-css-border.+?-color') and text() != 'transparent' and text != 'none']]"):
+		messages.append(LintMessage("c-004", "Don’t specify border colors, so that reading systems can adjust for night mode.", se.MESSAGE_TYPE_WARNING, local_css_path, matches))
 
-	# Check for endnotes
-	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[not(preceding-sibling::*)]/cite[not(preceding-sibling::node()[normalize-space(.)]) and (following-sibling::node()[normalize-space(.)])[1][contains(@epub:type, 'backlink')]]")
+	# Check that footers have the expected styling
+	# Footers may sometimes be aligned with text-align: center as well as text-align: right
+	# We also ignore text-align on any footers whose only child is a postscript, which is typically left-aligned
+	nodes = dom.xpath("/html/body//footer[not(@data-css-margin-top='1em') or (not(@data-css-text-align) and not(./*[contains(@epub:type, 'z3998:postscript') and not(following-sibling::*) and not(preceding-sibling::*) ]))]")
 	if nodes:
-		messages.append(LintMessage("s-035", "Endnote containing only [xhtml]<cite>[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
-
-	# Check for wrong semantics in frontmatter/backmatter
-	if dom.xpath("/html/body/section[not(@data-parent) and re:test(@epub:type, '\\b(dedication|introduction|preface|foreword|preamble|titlepage|halftitlepage|imprint|epigraph|acknowledgements)\\b') and not(ancestor-or-self::*[contains(@epub:type, 'frontmatter')])]"):
-		messages.append(LintMessage("s-036", "No [val]frontmatter[/] semantic inflection for what looks like a frontmatter file.", se.MESSAGE_TYPE_WARNING, filename))
-
-	if dom.xpath("/html/body/section[not(@data-parent) and re:test(@epub:type, '\\b(endnotes|loi|afterword|appendix|colophon|copyright\\-page|lot)\\b') and not(ancestor-or-self::*[contains(@epub:type, 'backmatter')])]"):
-		messages.append(LintMessage("s-037", "No [val]backmatter[/] semantic inflection for what looks like a backmatter file.", se.MESSAGE_TYPE_WARNING, filename))
+		messages.append(LintMessage("c-010", "[xhtml]<footer>[/] missing [css]margin-top: 1em; text-align: <value>;[/]. [css]text-align[/] is usually set to [css]right[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for leftover asterisms. Asterisms are sequences of any of these chars: * . • -⁠ —
-	nodes = dom.xpath("/html/body//*[self::p or self::div][re:test(., '^\\s*[\\*\\.•\\-⁠—]\\s*([\\*\\.•\\-⁠—]\\s*)+$')]")
+	# Check for elements that have `text-align: center` but also `text-indent: 1em`
+	nodes = dom.xpath("/html/body//*[@data-css-text-align='center' and @data-css-text-indent='1em']")
 	if nodes:
-		messages.append(LintMessage("s-038", "Illegal asterism. Section/scene breaks must be defined by an [xhtml]<hr/>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
-
-	# Check for <table> element without a <tbody> child
-	if dom.xpath("/html/body//table[not(tbody)]"):
-		messages.append(LintMessage("s-042", "[xhtml]<table>[/] element without [xhtml]<tbody>[/] child.", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("c-011", "Element with [css]text-align: center;[/] but [css]text-indent[/] is [css]1em[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check that short stories are on an <article> element
-	nodes = dom.xpath("/html/body/section[contains(@epub:type, 'se:short-story') or contains(@epub:type, 'se:novella')]")
+	# Check for sections that don't have header elements, and that do not have margin-top: 8em.
+	# This tries to find the first <section> or <article> whose first child is <p>, and that doesn't
+	# have any ancestor <section>s or <articles>s with heading content, and that doesn't have the correct top margin.
+	# Ignore the titlepage, loi, and dedications as they typically have unique styling.
+	nodes = dom.xpath("/html/body//*[name()='section' or name()='article'][1][./p[1][not(preceding-sibling::*)] and not(re:test(@epub:type, '(titlepage|loi|dedication)')) and not(ancestor::*/*[re:test(name(), '^(h[1-6]|header|hgroup)$')]) and not(@data-css-margin-top='20vh')]")
 	if nodes:
-		messages.append(LintMessage("s-043", "[val]se:short-story[/] semantic on element that is not [xhtml]<article>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("c-012", "Sectioning element without heading content, and without [css]margin-top: 20vh;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for poetry/verse without a descendent <p> element.
-	# Skip the ToC landmarks because it may have poem/verse semantic children.
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')][not(descendant::p)][not(ancestor::nav[contains(@epub:type, 'landmarks')])]")
+	# Check for padding and margin not in .5 increments, except for table headers/cells which usually need .25em increments
+	nodes = dom.xpath("/html/body//*[not(re:test(name(), '^(h[1-6]|td|th)$'))][attribute::*[re:test(local-name(), 'data-css-(margin|padding)')][re:test(., '^[0-9]*\\.[^5]')]]")
 	if nodes:
-		messages.append(LintMessage("s-044", "Element with poem or verse semantic, without descendant [xhtml]<p>[/] (stanza) element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
-
-	# Check for <abbr> elements that have two or more letters/periods, that don't have a semantic epub:type
-	# SS. is the French abbreviation for "Saints"
-	nodes = dom.xpath("/html/body//abbr[not(@epub:type)][text() != 'U.S.' and text() != 'SS.'][re:test(., '([A-Z]\\.?){2,}')]")
-	filtered_nodes = []
-	for node in nodes:
-		add_node = True
-
-		if node.text in INITIALISM_EXCEPTIONS:
-			add_node = False
-		elif node.get_attr("class"):
-			for attr_value in node.get_attr("class").split():
-				if attr_value in IGNORED_CLASSES:
-					add_node = False
-					break
-		elif node.text in ("A.M.", "P.M.") and node.xpath("./preceding-sibling::node()[re:test(., '[0-9]\\s$')]"):
-			# Ignore instances of capitalized times, usually in titles. For example:
-			# `From 10:20 <abbr>P.M.</abbr> to 10:47 <abbr>P.M.</abbr>` in a chapter title.
-			# We have to test for a preceding number because `<abbr>P.M.</abbr>` in other contexts may mean `Prime Minister`
-			add_node = False
-
-		if add_node:
-			filtered_nodes.append(node)
-
-	if filtered_nodes:
-		messages.append(LintMessage("s-045", "[xhtml]<abbr>[/] element without semantic inflection like [class]z3998:personal-name[/] or [class]z3998:initialism[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in filtered_nodes]))
+		messages.append(LintMessage("c-013", "Element with margin or padding not in increments of [css].5em[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for <p> elems that has some element children, which are only <span> and <br> children, but the parent doesn't have poem/verse semantics.
-	# Ignore spans that have a class, but not if the class is an i# class (for poetry indentation)
-	nodes = dom.xpath("/html/body//p[not(./text()[normalize-space(.)])][*][not(ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][not(*[not(self::span) and not(self::br)])][ (not(span[@epub:type]) and not(span[@class]) ) or span[re:test(@class, '\\bi[0-9]\\b')]]")
+	# Check for <table> without set margins. We ignore tables that are for drama structuring, and tables that are ancestors of <blockquote> because <blockquote> has its own margin.
+	# We also ignore tables whose preceding sibling sets a bottom margin that is not 0.
+	nodes = dom.xpath("/html/body//table[not(./ancestor-or-self::*[contains(@epub:type, 'z3998:drama')]) and not(./ancestor::blockquote) and not(./preceding-sibling::*[1][@data-css-margin-bottom != '0']) and (not(@data-css-margin-top) or not(@data-css-margin-right) or not(@data-css-margin-bottom) or not(@data-css-margin-left))]")
 	if nodes:
-		messages.append(LintMessage("s-046", "[xhtml]<p>[/] element containing only [xhtml]<span>[/] and [xhtml]<br>[/] elements, but its parent doesn’t have the [val]z3998:poem[/], [val]z3998:verse[/], [val]z3998:song[/], [val]z3998:hymn[/], or [val]z3998:lyrics[/] semantic. Multi-line clauses that are not verse don’t require [xhtml]<span>[/]s.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("c-014", "[xhtml]<table>[/] element without explicit margins. Most tables need [css]margin: 1em;[/] or [css]margin: 1em auto 1em auto;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# For this series of selections, we select spans that are direct children of p, because sometimes a line of poetry may have a nested span.
-	nodes = dom.xpath("/html/body/*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')]/descendant-or-self::*/p/span/following-sibling::*[contains(@epub:type, 'noteref') and name()='a' and position()=1]")
+	# Check for elements that follow salutations, but that don't have text-indent: 0;
+	# We have to check both <p> elems that are salutations, and also <p> elems that have a first-child inline element that is a salutation, and that does not have following siblings that are senders/valedictions, as that might indicate a letter in a prose context
+	nodes = dom.xpath("/html/body//*[(contains(@epub:type, 'z3998:salutation') or ./preceding-sibling::*[1][contains(@epub:type, 'z3998:salutation') or (name() != 'blockquote' and count(./node()[normalize-space(.)]) = 1 and ./*[contains(@epub:type, 'z3998:salutation')])] or ./*[1][contains(@epub:type, 'z3998:salutation') and not((./following-sibling::node()[1][self::text()]))]) and @data-css-text-indent != '0']")
 	if nodes:
-		messages.append(LintMessage("s-047", "[val]noteref[/] as a direct child of element with poem or verse semantic. [val]noteref[/]s should be in their parent [xhtml]<span>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("c-015", "Element after or containing [val]z3998:salutation[/] does not have [css]text-indent: 0;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for incorrectly applied se:name semantic
-	nodes = dom.xpath("/html/body//*[self::p or self::blockquote][contains(@epub:type, 'se:name.')]")
+	# Check for block-level postscripts that don't have margin-top: 1em. Exclude postscripts that are the first child of blockquote or footer, since blockquotes/footers gives the desired margin.
+	nodes = dom.xpath("/html/body//*[(name() = 'div' or name() = 'p') and contains(@epub:type, 'z3998:postscript') and not((./parent::blockquote or ./parent::footer) and count(./preceding-sibling::*) = 0) and @data-css-margin-top != '1em']")
 	if nodes:
-		messages.append(LintMessage("s-048", "[val]se:name[/] semantic on block element. [val]se:name[/] indicates the contents is the name of something.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("c-017", "Element with [val]z3998:postscript[/] semantic, but without [css]margin-top: 1em;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for <header> elements with only h# child nodes
-	nodes = dom.xpath("/html/body//header[./*[re:test(name(), 'h[1-6]') and (count(preceding-sibling::*) + count(following-sibling::*)=0)]]")
+	# Check for <p> postscripts that don't have text-indent: 0.
+	nodes = dom.xpath("/html/body//p[contains(@epub:type, 'z3998:postscript') and (not(@data-css-text-align) or @data-css-text-align = 'initial' or @data-css-text-align = 'left') and @data-css-text-indent != '0']")
 	if nodes:
-		messages.append(LintMessage("s-049", "[xhtml]<header>[/] element whose only child is an [xhtml]<h#>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("c-018", "Element with [val]z3998:postscript[/] semantic, but without [css]text-indent: 0;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for <span>s that only exist to apply epub:type
-	nodes = dom.xpath("/html/body//*[span[@epub:type][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	# Check for signature semantic without small caps
+	# Ignore signatures that are only capital letters
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:signature') and not(@data-css-font-variant = 'small-caps') and not(@data-css-font-style = 'italic') and not(re:test(., '^[A-Z\\.\\s]+$'))]")
 	if nodes:
-		messages.append(LintMessage("s-050", "[xhtml]<span>[/] element appears to exist only to apply [attr]epub:type[/]. [attr]epub:type[/] should go on the parent element instead, without a [xhtml]<span>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("c-019", "Element with [val]z3998:signature[/] semantic, but without [css]font-variant: small-caps;[/] or [css]font-style: italic;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for language tags misusing the unk lang code
-	nodes = dom.xpath("//*[@xml:lang='unk']")
+	# Check for <article>s or <section>s that occur more than once in a file, without page break CSS
+	# Exclude the last element because we don't care if it breaks after the last one.
+	nodes = dom.xpath("/html/body[count(./article) > 1 or count(./section) > 1]/*[(name() = 'article' or name() = 'section') and (not(@data-css-break-after) or @data-css-break-after != 'page') and not(position() = last())]")
 	if nodes:
-		messages.append(LintMessage("s-051", "Element with [xhtml]xml:lang=\"unk\"[/] should be [xhtml]xml:lang=\"und\"[/] (undefined).", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("c-020", "Multiple [xhtml]<article>[/]s or [xhtml]<section>[/]s in file, but missing [css]break-after: page;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for title attrs on abbr elements
-	nodes = dom.xpath("/html/body//abbr[@title]")
+	# Check for elements that are set in italics, and whose italic children don't have font-style: normal.
+	nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and ./*[(name()='i' or name()='em') and @data-css-font-style='italic']]")
 	if nodes:
-		messages.append(LintMessage("s-052", "[xhtml]<abbr>[/] element with illegal [attr]title[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("c-021", "Element with [css]font-style: italic;[/], but child [xhtml]<i>[/] or [xhtml]<em>[/] does not have [css]font-style: normal;[/]. Hint: Italics within italics are typically set in Roman for contrast; if that’s not the case here, can [xhtml]<i>[/] be removed while still preserving italics and semantic inflection?", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	nodes = dom.xpath("/html/body//blockquote//p[parent::*[name()='footer'] or parent::*[name()='blockquote']]//cite") # Sometimes the <p> may be in a <footer>
+	# Check for tables that appear to be summing/listing numbers, but that are missing font-variant-numeric styling
+	nodes = dom.xpath("/html/body//table[count(.//td[not(following-sibling::*) and re:test(., '^[0-9]+$') and not(@data-css-font-variant-numeric='tabular-nums')]) >= 2]")
 	if nodes:
-		messages.append(LintMessage("s-054", "[xhtml]<cite>[/] as child of [xhtml]<p>[/] in [xhtml]<blockquote>[/]. [xhtml]<cite>[/] should be the direct child of [xhtml]<blockquote>[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
-
-	# Check for <th> element without a <thead> ancestor. However, <th scope="...">	and <th/> are allowed, for use in tables with headers in the middle of tables (url:https://standardebooks.org/ebooks/dorothy-day/the-eleventh-virgin) and vertical table headers
-	# (https://standardebooks.org/ebooks/charles-babbage/passages-from-the-life-of-a-philosopher)
-	if dom.xpath("/html/body//table//th[not(ancestor::thead)][not(@scope)][not(count(node())=0)]"):
-		messages.append(LintMessage("s-055", "[xhtml]<th>[/] element not in [xhtml]<thead>[/] ancestor. Note: [xhtml]<th>[/] elements used as mid-table headings or horizontal row headings require the [attr]scope[/] attribute.", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("c-026", "Table that appears to be listing numbers, but without [css]font-variant-numeric: tabular-nums;[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for z3998:stage-direction on elements that are not <i>
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:stage-direction') and name() !='i' and name() !='abbr' and name() !='p']")
+	# Check for three-em-dashes not preceded by a space
+	nodes = dom.xpath(f"/html/body//p[re:test(., '[^>“(\\s{se.WORD_JOINER}]{se.WORD_JOINER}?⸻')]")
 	if nodes:
-		messages.append(LintMessage("s-058", "[attr]z3998:stage-direction[/] semantic only allowed on [xhtml]<i>[/], [xhtml]<abbr>[/], and [xhtml]<p>[/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("y-033", "Possible typo: Three-em-dash obscuring an entire word, but not preceded by a space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check that internal links don't begin with ../
-	nodes = dom.xpath("/html/body//a[re:test(@href, '^\\.\\./text/')]")
-	if nodes:
-		messages.append(LintMessage("s-059", "Internal link beginning with [val]../text/[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	return messages
 
-	# Check for italics on things that shouldn't be italics
-	nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.music.song') or contains(@epub:type, 'se:name.publication.short-story') or contains(@epub:type, 'se:name.publication.essay')]")
-	if nodes:
-		messages.append(LintMessage("s-060", "Italics on name that requires quotes instead.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+def _lint_xhtml_metadata_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree) -> list:
+	"""
+	Helper function used in self.lint()
+	Process metadata checks on an .xhtml file
 
-	# Check for h# tags followed by header content, that are not children of <header>.
-	# Only match if there is a following <p>, because we could have the case where there's an epigraph after a division title.
-	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][following-sibling::*[contains(@epub:type, 'epigraph') or contains(@epub:type, 'bridgehead')]][following-sibling::p][not(parent::header)]")
-	if nodes:
-		messages.append(LintMessage("s-061", "Title and following header content not in a [xhtml]<header>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
 
-	# Check for <dt> elements without exactly one <dfn> child, but only in glossaries
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'glossary')]//dt[not(count(./dfn)=1)]")
-	if nodes:
-		messages.append(LintMessage("s-062", "[xhtml]<dt>[/] element in a glossary without exactly one [xhtml]<dfn>[/] child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	OUTPUTS
+	A list of LintMessages representing metadata errors found in the file
+	"""
 
-	# Check that `z3998:persona` is only on <b> or <td>. We use the contact() xpath function so we don't catch `z3998:personal-name`
-	nodes = dom.xpath("/html/body//*[contains(concat(' ', @epub:type, ' '), ' z3998:persona ') and not(self::b or self::td)]")
-	if nodes:
-		messages.append(LintMessage("s-063", "[val]z3998:persona[/] semantic on element that is not a [xhtml]<b>[/] or [xhtml]<td>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	messages = []
 
-	# Check for fulltitle semantic on a header not in the half title
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'fulltitle') and name()!='h2' and name()!='hgroup' and not(ancestor::*[contains(@epub:type, 'halftitlepage')])]")
-	if nodes:
-		messages.append(LintMessage("s-065", "[val]fulltitle[/] semantic on element that is not in the half title.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for missing MARC relators
+	# Don't check the landmarks as that may introduce duplicate errors
+	# Only check for top-level elements, to avoid intros in short stories or poems in compilation.
+	if not dom.xpath("/html/body//nav[contains(@epub:type, 'landmarks')]"):
+		if dom.xpath("/html/body/*[contains(@epub:type, 'introduction') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and (text()='aui' or text()='win')]"):
+			messages.append(LintMessage("m-030", "[val]introduction[/] semantic inflection found, but no MARC relator [val]aui[/] (Author of introduction, but not the chief author) or [val]win[/] (Writer of introduction).", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for header elements that have a label, but are missing the label semantic
-	# Find h# nodes whose first child is a text node matching a label type, and where that text node's next sibling is a semantic roman numeral
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./node()[1][self::text() and not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)$') and following-sibling::*[1][contains(@epub:type, 'z3998:roman')]]]")
-	if nodes:
-		messages.append(LintMessage("s-066", "Header element missing [val]label[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		if dom.xpath("/html/body/*[contains(@epub:type, 'preface') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='wpr']"):
+			messages.append(LintMessage("m-031", "[val]preface[/] semantic inflection found, but no MARC relator [val]wpr[/] (Writer of preface).", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for header elements that have a label semantic, but are missing an ordinal sibling
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./span[contains(@epub:type, 'label')]][not(./span[contains(@epub:type, 'ordinal')])]")
-	if nodes:
-		messages.append(LintMessage("s-067", "Header element with a [val]label[/] semantic child, but without an [val]ordinal[/] semantic child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		if dom.xpath("/html/body/*[contains(@epub:type, 'afterword') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='aft']"):
+			messages.append(LintMessage("m-032", "[val]afterword[/] semantic inflection found, but no MARC relator [val]aft[/] (Author of colophon, afterword, etc.).", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for header elements with a roman semantic but without an ordinal semantic
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][contains(@epub:type, 'z3998:roman') and not(contains(@epub:type, 'ordinal'))]")
-	if nodes:
-		messages.append(LintMessage("s-068", "Header element missing [val]ordinal[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		if dom.xpath("/html/body/*[contains(@epub:type, 'endnotes') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='ann']"):
+			messages.append(LintMessage("m-033", "[val]endnotes[/] semantic inflection found, but no MARC relator [val]ann[/] (Annotator).", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for body element without child section or article. Ignore the ToC because it has a unique structure
-	nodes = dom.xpath("/html/body[not(./*[name()='section' or name()='article' or (name()='nav' and re:test(@epub:type, '\\b(toc|loi)\\b'))])]")
-	if nodes:
-		messages.append(LintMessage("s-069", "[xhtml]<body>[/] element missing direct child [xhtml]<section>[/] or [xhtml]<article>[/] element.", se.MESSAGE_TYPE_ERROR, filename))
+		if dom.xpath("/html/body/*[contains(@epub:type, 'loi') and not(@data-parent)]") and not self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text()='ill']"):
+			messages.append(LintMessage("m-034", "[val]loi[/] semantic inflection found, but no MARC relator [val]ill[/] (Illustrator).", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for h# without semantics; h# that have child elements (which likely have the correct semantics) are OK
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(@epub:type)][not(./*[not(name()='a' and contains(@epub:type, 'noteref'))])]")
+	# Check that `Internet Archive` and `HathiTrust` are preceded by `the`. They might have an immediate preceding text node that ends in `the`, or they
+	# might be preceded by a white space node, then a <br/>, then a text node ending in `the`.
+	nodes = dom.xpath("/html/body//a[(re:test(text(), '[Hh]athi') and re:test(@href, '^https://[^\"]*?hathitrust.org')) or (re:test(text(), 'Internet Archive') and re:test(@href, 'https://[^\"]*?archive.org'))][(preceding-sibling::node()[2][not(self::br)] and preceding-sibling::node()[1][not(re:test(., '\\sthe $'))]) or (preceding-sibling::node()[2][self::br] and preceding-sibling::node()[3][not(re:test(., '\\sthe$'))]) ]")
 	if nodes:
-		messages.append(LintMessage("s-070", "[xhtml]<h#>[/] element without semantic inflection.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("m-061", "Link must be preceded by [text]the[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for sectioning elements with more than one heading element
-	nodes = dom.xpath("/html/body//*[name()='article' or name()='section'][count(./*[name()='header' or name()='hgroup' or re:test(name(), '^h[1-6]$')]) > 1]")
-	if nodes:
-		messages.append(LintMessage("s-071", "Sectioning element with more than one heading element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	return messages
 
-	# Check for some elements that only have a <span> child (including text inline to the parent).
-	# But, exclude such elements that are <p> elements that have poetry-type parents.
-	nodes = dom.xpath("/html/body//*[name() !='p' and not(ancestor-or-self::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][./span[count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
-	if nodes:
-		messages.append(LintMessage("s-072", "Element with single [xhtml]<span>[/] child. [xhtml]<span>[/] should be removed and its attributes promoted to the parent element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+def _lint_xhtml_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_flags: dict, language: str, section_tree: List[EbookSection]) -> list:
+	"""
+	Helper function used in self.lint()
+	Process syntax checks on an .xhtml file
 
-	# Check for header elements that are missing both label and ordinal semantics
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)\\s+[ixvIXVmcd]+$')]")
-	if nodes:
-		messages.append(LintMessage("s-073", "Header element that requires [val]label[/] and [val]ordinal[/] semantic children.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
+	file_contents: The contents of the file being checked
+	ebook_flags: A dictionary containing several pieces of information about an ebook
 
-	# Check for table headers that don't have content, which is an accessibility issue.
-	# Note that @aria-label and @title should (apparently) not be used for table headers
-	nodes = dom.xpath("/html/body//th[not(text())]")
-	if nodes:
-		messages.append(LintMessage("s-074", "[xhtml]<th>[/] element with no text content should be a [xhtml]<td>[/] element instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	OUTPUTS
+	A list of LintMessages representing syntax errors found in the file
+	"""
 
-	# Check for incorrect children of <body>
-	nodes = dom.xpath("/html/body/*[name() != 'section' and name() != 'article' and name() != 'nav']")
-	if nodes:
-		messages.append(LintMessage("s-075", "[xhtml]<body>[/] element with direct child that is not [xhtml]<section>[/], [xhtml]<article>[/], or [xhtml]<nav>[/].", se.MESSAGE_TYPE_ERROR, filename))
+	messages = []
 
-	# Check for lang="" instead of xml:lang=""
-	nodes = dom.xpath("/html/body//*[@lang]")
-	if nodes:
-		messages.append(LintMessage("s-076", "[attr]lang[/] attribute used instead of [attr]xml:lang[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# This block is useful for pretty-printing section_tree should we need to debug it in the future
+	# def dump(item, char):
+	#	print(f"{char} {item.section_id} ({item.depth}) {item.has_header}")
+	#	for child in item.children:
+	#		dump(child, f"	{char}")
+	# for section in section_tree:
+	#	dump(section, "")
+	# exit()
 
-	# Check for <header> preceded by non-sectioning elements
-	nodes = dom.xpath("/html/body//header[./preceding-sibling::*[not(re:test(name(), '^(section|div|article)$'))]]")
-	if nodes:
-		messages.append(LintMessage("s-077", "[xhtml]<header>[/] element preceded by non-sectioning element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for numeric entities
+	matches = regex.findall(r"&#[0-9]+?;", file_contents)
+	if matches:
+		messages.append(LintMessage("s-001", "Illegal numeric entity (like [xhtml]Α[/]).", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for <footer> followed by non-sectioning elements
-	nodes = dom.xpath("/html/body//footer[./following-sibling::*[not(re:test(name(), '^(section|div|article)$'))]]")
+	# Check nested <blockquote> elements, but only if it's the first child of another <blockquote>
+	nodes = dom.xpath("/html/body//blockquote/*[1][name()='blockquote']")
 	if nodes:
-		messages.append(LintMessage("s-078", "[xhtml]<footer>[/] element followed by non-sectioning element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-005", "Nested [xhtml]<blockquote>[/] element.", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for elements with no children and only white space contents
-	nodes = dom.xpath("/html/body//*[not(./*) and re:test(., '^\\s+$')]")
+	# Check to see if we've marked something as poetry or verse, but didn't include a first <span>
+	# This xpath selects the p elements, whose parents are poem/verse, and whose first child is not a span
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')]/p[not(./*[name()='span' and position()=1])]")
 	if nodes:
-		messages.append(LintMessage("s-079", "Element containing only white space.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		matches = []
+		for node in nodes:
+			# Get the first line of the poem, if it's a text node, so that we can include it in the error messages.
+			# If it's not a text node then just ignore it and add the error anyway.
+			first_line = node.xpath("./descendant-or-self::text()[normalize-space(.)]", True)
+			if first_line:
+				match = first_line.strip()
+				if match: # Make sure we don't append an empty string
+					matches.append(match)
 
-	# Check for a play <td> that contains both inline text and block-level elements.
-	# Note that we check any child of <html> because often <body> has the drama semantic.
-	nodes = dom.xpath("/html//*[contains(@epub:type, 'z3998:drama')]//td[./text()[normalize-space(.)] and ./*[name() = 'p' or name() = 'blockquote' or name() = 'div']]")
-	if nodes:
-		messages.append(LintMessage("s-080", "[xhtml]<td>[/] in drama containing both inline text and a block-level element. All children should either be only text, or only block-level elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-006", "Poem or verse [xhtml]<p>[/] (stanza) without [xhtml]<span>[/] (line) element.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	# Check for <p> tags that follow figure, blockquote, or table, and start with a lowercase letter, but that don't have the `continued` class. Exclude matches whose first child is <cite>, so we don't match things like <p><cite>—Editor</cite>
-	nodes = dom.xpath("/html/body//p[preceding-sibling::*[1][name() = 'figure' or name() = 'blockquote' or name() = 'table'] and re:test(., '^([a-z]|—|…)') and not(contains(@class, 'continued')) and not(./*[1][self::cite])]")
+	# Check for elements that don't have a direct block child
+	# Allow white space and comments before the first child
+	# Ignore children of the ToC and landmarks as they do not require p children
+	nodes = dom.xpath("/html/body//*[(name()='blockquote' or name()='dd' or name()='header' or name()='li' or name()='footer') and not(./ancestor::*[contains(@epub:type, 'toc') or contains(@epub:type, 'landmarks')]) and (node()[normalize-space(.) and not(self::comment())])[1][not(name()='p' or name()='blockquote' or name()='div' or name()='table' or name()='header' or name()='ul' or name()='ol' or name()='footer' or name()='hgroup' or re:test(name(), '^h[0-6]'))]]")
 	if nodes:
-		messages.append(LintMessage("s-081", "[xhtml]<p>[/] preceded by [xhtml]<figure>[/], [xhtml]<blockquote>[/xhtml], or [xhtml]<table>[/], but without [val]continued[/] class.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-007", "Element requires at least one block-level child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for language tags transliterated into Latin script but missing `-Latn` suffix
-	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '^(he|ru|el|zh|bn|hi|sa|uk|yi|grc)$') and re:test(., '[a-zA-Z]')]")
+	# Check for block-level tags that end with <br/>
+	nodes = dom.xpath("/html/body//*[self::p or self::blockquote or self::table or self::ol or self::ul or self::section or self::article][br[last()][not(following-sibling::text()[normalize-space()])][not(following-sibling::*)]]")
 	if nodes:
-		messages.append(LintMessage("s-082", "Element containing Latin script for a non-Latin-script language, but its [attr]xml:lang[/] attribute value is missing the [val]-Latn[/] language tag suffix. Hint: For example Russian transliterated into Latin script would be [val]ru-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-008", "[xhtml]<br/>[/] element found before closing tag of block-level element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for persona <td>s that have child <p> elements
-	nodes = dom.xpath("/html/body//td[contains(@epub:type, 'z3998:persona') and ./p]")
+	# Check for hgroup elements with only one child
+	nodes = dom.xpath("/html/body//hgroup[count(*)=1]")
 	if nodes:
-		messages.append(LintMessage("s-083", "[xhtml]<td epub:type=\"z3998:persona\">[/] element with child [xhtml]<p>[/] element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-009", "[xhtml]<hgroup>[/] element with only one child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for incorrect semantics on some common publications
-	nodes = dom.xpath("/html/body//i[@epub:type='se:name.publication.book' and re:test(., '^(The )?(Iliad|Odyssey|Aeneid|Metamorphoses|Beowulf|Divine Comedy|Paradise Lost)$')]")
+	# Check for empty elements. Elements are empty if they have no children and no non-whitespace text
+	# Use name() to check for MathML elements, because if the MathML namespace is not declared in a given file and we search by the m: namespace then this xpath will fail entirely.
+	nodes = dom.xpath("/html/body//*[not(self::br) and not(self::hr) and not(self::img) and not(self::td) and not(self::th) and not(self::link) and not(self::col) and not(self::colgroup) and not(local-name()='none') and not(local-name()='mspace') and not(local-name()='mprescripts')][not(./*)][not(normalize-space())][not(./ancestor::*[local-name()='annotation-xml'])]")
 	if nodes:
-		messages.append(LintMessage("s-084", "Poem has incorrect semantics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-010", "Empty element. Use [xhtml]<hr/>[/] for thematic breaks if appropriate.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for `Op. Cit.` in endnotes. `Op. Cit.` means "the previous reference" which doesn't make sense
-	# in popup endnotes. But, if the endnote has a book reference, then allow it as it might be referring to that.
-	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote') and (re:test(., '(Loc\\.|Op\\.) Cit\\.', 'i') or re:test(., '\\bl\\.c\\.\\b', 'i')) and not(.//*[contains(@epub:type, 'se:name.publication')])]")
+	# Check for <section> and <article> without ID attribute
+	nodes = dom.xpath("/html/body//*[self::section or self::article][not(@id)]")
 	if nodes:
-		messages.append(LintMessage("s-086", "[text]Op. Cit.[/] or [text]Loc. Cit.[/] in endnote. Hint: [text]Op. Cit.[/] and [text]Loc. Cit.[/] mean [text]the previous reference[/], which usually doesn’t make sense in a popup endnote. Such references should be expanded.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
-
-	# Check for half title pages missing subtitles
-	if ebook_info["has_subtitle"]:
-		# Make sure we exclude <a> because that appears in the ToC landmarks
-		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and not(.//*[contains(@epub:type, 'subtitle')])]")
-		if nodes:
-			messages.append(LintMessage("s-087", "Subtitle in metadata, but no subtitle in the half title page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-
-	else:
-		# Make sure we exclude <a> because that appears in the ToC landmarks
-		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and .//*[contains(@epub:type, 'subtitle')]]")
-		if nodes:
-			messages.append(LintMessage("s-088", "Subtitle in half title page, but no subtitle in metadata.", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("s-011", "Element without [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for MathML without alttext
-	nodes = dom.xpath("/html/body//m:math[not(@alttext)]")
-	if nodes:
-		messages.append(LintMessage("s-089", "MathML missing [attr]alttext[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for <pre> tags
+	if dom.xpath("/html/body//pre"):
+		messages.append(LintMessage("s-013", "Illegal [xhtml]<pre>[/] element.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for common errors in language tags
-	# `gr` is often used instead of `el`, `sp` instead of `es`, and `ge` instead of `de` (`ge` is the Georgian geographic region subtag but not a language subtag itself)
-	nodes = dom.xpath("//*[re:test(@xml:lang, '^(gr|sp|ge)$')]")
+	# Check for <br/> after block-level elements
+	nodes = dom.xpath("/html/body//*[self::p or self::blockquote or self::table or self::ol or self::ul or self::section or self::article][following-sibling::br]")
 	if nodes:
-		messages.append(LintMessage("s-090", "Invalid language tag.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("s-014", "[xhtml]<br/>[/] after block-level element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for <span> in poetry not followed by <br/>. Ignore spans that are roman as they might be present in poem headers, or spans with xml:lang.
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]//p/span[not(@epub:type='z3998:roman') and not(@xml:lang) and following-sibling::*[1][name()='span']]")
-	if nodes:
-		messages.append(LintMessage("s-091", "[xhtml]<span>[/] not followed by [xhtml]<br/>[/] in poetry.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for <hr> tags before the end of a section, which is a common PG artifact
+	if dom.xpath("/html/body//hr[count(following-sibling::*)=0]"):
+		messages.append(LintMessage("s-012", "Illegal [xhtml]<hr/>[/] as last child.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for nested `<abbr>` elements
-	nodes = dom.xpath("/html/body//abbr[./abbr]")
+	# Check for hgroup elements with a subtitle but no title
+	nodes = dom.xpath("/html/body//hgroup[./*[contains(@epub:type, 'subtitle')] and not(./*[contains(concat(' ', @epub:type, ' '), ' title ')])]")
 	if nodes:
-		messages.append(LintMessage("s-093", "Nested [xhtml]<abbr>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-015", "Element has [val]subtitle[/] semantic, but without a sibling having a [val]title[/] semantic.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for language tags transliterated into Latin script with incorrect `-latn` suffix (lowercase l)
-	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '-latn')]")
+	# Check for deprecated MathML elements
+	# Note we dont select directly on element name, because we want to ignore any namespaces that may (or may not) be defined
+	nodes = dom.xpath("/html/body//*[local-name()='mfenced']")
 	if nodes:
-		messages.append(LintMessage("s-094", "Element has an [attr]xml:lang[/] attribute that incorrectly contains [val]-latn[/] instead of [val]-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-017", "[xhtml]<m:mfenced>[/] is deprecated in the MathML spec. Use [xhtml]<m:mrow><m:mo fence=\"true\">(</m:mo>...<m:mo fence=\"true\">)</m:mo></m:mrow>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for poetry/verse that has an hgroup but which does not have the correct text alignment
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(hymn|poem|song|verse)')]//hgroup/p[@data-css-text-align != 'center']")
+	# Check for <figure> tags without id attributes
+	nodes = dom.xpath("/html/body//img[@id]")
 	if nodes:
-		messages.append(LintMessage("s-095", "[xhtml]<p>[/] child of [xhtml]<hgroup>[/] in poetry/verse does not have [css]text-align: center;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-018", "[xhtml]<img>[/] element with [attr]id[/] attribute. [attr]id[/] attributes go on parent [xhtml]<figure>[/] elements. Hint: Images that are inline (i.e. that do not have parent [xhtml]<figure>[/]s do not have [attr]id[/] attributes.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check that half titles without subtitles have the fulltitle semantic
-	nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage')]/*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'fulltitle'))]")
+	# Check for IDs on <h#> tags
+	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][@id]")
 	if nodes:
-		messages.append(LintMessage("s-096", "Heading element in half title page missing the [val]fulltitle[/] semantic.", se.MESSAGE_TYPE_ERROR, filename))
+		messages.append(LintMessage("s-019", "[xhtml]<h#>[/] element with [attr]id[/] attribute. [xhtml]<h#>[/] elements should be wrapped in [xhtml]<section>[/] elements, which should hold the [attr]id[/] attribute.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for missing href attributes, sometimes a leftover from PG transcriptions
-	nodes = dom.xpath("/html/body//a[not(@href)]")
-	if nodes:
-		messages.append(LintMessage("s-097", "[xhtml]a[/] element without [attr]href[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for legal cases that aren't italicized
+	# We can't use this because v. appears as short for "volume", and we may also have sporting events without italics.
+	#nodes = dom.xpath("/html/body//abbr[text()='v.' or text()='versus'][not(parent::i)]")
+	#if nodes:
+	#	messages.append(LintMessage("t-xxx", "Legal case without parent [xhtml]<i>[/].", se.MESSAGE_TYPE_WARNING, filename, {f"{node.to_string()}." for node in nodes}))
 
-	# Check for <headers>s with only one child
-	nodes = dom.xpath("/html/body/*[name() = 'section' or name() = 'article']/header[./*[not(name() = 'p') and not(preceding-sibling::* or following-sibling::*)]]")
-	if nodes:
-		messages.append(LintMessage("s-098", "[xhtml]<header>[/] element with only one child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Only do this check if there's one <h#> or one <hgroup> tag. If there's more than one, then the xhtml file probably requires an overarching title
+	# We merge two xpaths here because <h#>/<hgroup> can be either a direct child of <section>, or it could be nested in <header>
+	if len(dom.xpath("/html/body/*[name()='section' or name()='article']/*[re:test(name(), '^h[1-6]$') or name()='hgroup'] | /html/body/*[name()='section' or name()='article']/header/*[re:test(name(), '^h[1-6]$') or name()='hgroup']"))==1:
+		title = se.formatting.generate_title(dom)
 
-	# Check for @lang attributes
-	nodes = dom.xpath("//*[@lang]")
-	if nodes:
-		messages.append(LintMessage("s-102", "[attr]lang[/] attribute detected. Hint: Use [attr]xml:lang[/] instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		if not dom.xpath(f"/html/head/title[text()={se.easy_xml.escape_xpath(title.replace('&', '&'))}]"):
+			messages.append(LintMessage("s-021", f"Unexpected value for [xhtml]<title>[/] element. Expected: [text]{title}[/]. (Beware hidden Unicode characters!)", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for common missing roman semantics for “I”
-	regent_regex = r"(?:Charles|Edward|George|Henry|James|William) I\b"
-	matches = regex.findall(fr"King {regent_regex}", file_contents) + regex.findall(fr"{regent_regex}’s", file_contents)
-	if matches:
-		messages.append(LintMessage("s-103", "Probable missing semantics for a roman I numeral.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Check to see if <h#> tags are correctly titlecased
+	# Ignore <h#> tags with an `xml:lang` attribute, as other languages have different titlecasing rules
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') or (name() = 'p' and parent::hgroup)][not(contains(@epub:type, 'z3998:roman')) and not(@xml:lang)]")
+	for node in nodes:
+		node_copy = deepcopy(node)
 
-	return messages
+		# Remove *leading* Roman spans
+		# This matches the first child node excluding white space nodes, if it contains the z3998:roman semantic.
+		for element in node_copy.xpath("./node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman')]"):
+			element.remove()
 
-def _lint_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str) -> list:
-	"""
-	Helper function used in self.lint()
-	Process typography checks on an .xhtml file
+		# Remove noterefs
+		for element in node_copy.xpath(".//a[contains(@epub:type, 'noteref')]"):
+			element.remove()
 
-	INPUTS
-	filename: The name of the file being checked
-	dom: A dom tree to check
-	file_contents: The contents of the file being checked
-	special_file: A string containing the type of special file, if any
+		# Remove hidden elements, for example in poetry identified by first line (keats)
+		for element in node_copy.xpath(".//*[@hidden]"):
+			element.remove()
 
-	OUTPUTS
-	A list of LintMessages representing typography errors found in the file
-	"""
+		# Unwrap (not remove) any subelements in subtitles, as they may trip up further regexes
+		for element in node_copy.xpath(".//*[contains(@epub:type, 'subtitle')]//*"):
+			element.unwrap()
 
-	messages = []
+		title = node_copy.inner_xml()
 
-	# Check for punctuation outside quotes. We don't check single quotes because contractions are too common.
-	matches = regex.findall(fr"[\p{{Letter}}]+”[,\.](?!{se.WORD_JOINER} {se.WORD_JOINER}…)", file_contents)
-	if matches:
-		messages.append(LintMessage("t-002", "Comma or period outside of double quote. Generally punctuation goes within single and double quotes.", se.MESSAGE_TYPE_WARNING, filename, matches))
+		# Remove leading leftover spacing and punctuation
+		title = regex.sub(r"^[\s\.\,\!\?\:\;]*", "", title)
 
-	# Check for ldquo not correctly closed
-	# Ignore closing paragraphs, line breaks, and closing cells in case ldquo means "ditto mark"
-	matches = regex.findall(r"“[^‘”]+?“", file_contents)
-	matches = [match for match in matches if "</p" not in match and "<br/>" not in match and "</td>" not in match]
-	# xpath to check for opening quote in p, without a next child p that starts with an opening quote or an opening bracket (for editorial insertions within paragraphs of quotation); or that consists of only an ellipses (like an elided part of a longer quotation)
-	# Matching <p>s can't have a poem/verse ancestor as formatting is often special for those.
-	matches = matches + [regex.findall(r"“[^”]+</p>", node.to_string())[0] for node in dom.xpath("/html/body//p[re:test(., '“[^‘”]+$')][not(ancestor::*[re:test(@epub:type, 'z3998:(verse|poem|song|hymn|lyrics)')])][(following-sibling::*[1])[name()='p'][not(re:test(normalize-space(.), '^[“\\[]') or re:test(normalize-space(.), '^…$'))]]")]
+		# Normalize whitespace
+		title = regex.sub(r"\s+", " ", title, flags=regex.DOTALL).strip()
 
-	# Additionally, match short <p> tags (< 100 chars) that lack closing quote, and whose direct siblings do have closing quotes (to exclude runs of same-speaker dialog), and that is not within a blockquote, verse, or letter
-	matches = matches + [regex.findall(r"“[^”]+</p>", node.to_string())[0] for node in dom.xpath("/html/body//p[re:test(., '“[^‘”]+$') and not(re:test(., '[…:]$')) and string-length(normalize-space(.)) <=100][(following-sibling::*[1])[not(re:test(., '“[^”]+$'))] and (preceding-sibling::*[1])[not(re:test(., '“[^”]+$'))]][not(ancestor::*[re:test(@epub:type, 'z3998:(verse|poem|song|hymn|lyrics)')]) and not(ancestor::blockquote) and not (ancestor::*[contains(@epub:type, 'z3998:letter')])][(following-sibling::*[1])[name()='p'][re:test(normalize-space(.), '^[“\\[]') and not(contains(., 'continued'))]]")]
-	if matches:
-		messages.append(LintMessage("t-003", "[text]“[/] missing matching [text]”[/]. Note: When dialog from the same speaker spans multiple [xhtml]<p>[/] elements, it’s correct grammar to omit closing [text]”[/] until the last [xhtml]<p>[/] of dialog.", se.MESSAGE_TYPE_WARNING, filename, matches))
+		# Do we have a subtitle? If so the first letter of that must be capitalized, so we pull that out
+		subtitle_matches = regex.findall(r"(.*?)<span epub:type=\"subtitle\">(.*?)</span>(.*?)", title, flags=regex.DOTALL)
+		if subtitle_matches:
+			for title_header, subtitle, title_footer in subtitle_matches:
+				title_header = se.formatting.titlecase(se.formatting.remove_tags(title_header).strip())
+				subtitle = se.formatting.titlecase(se.formatting.remove_tags(subtitle).strip())
+				title_footer = se.formatting.titlecase(se.formatting.remove_tags(title_footer).strip())
 
-	# Check for closing dialog without comma
-	matches = regex.findall(r"[\p{Lowercase_Letter}]+?” [\p{Letter}]+? said", file_contents)
-	if matches:
-		messages.append(LintMessage("t-005", "Dialog without ending comma.", se.MESSAGE_TYPE_WARNING, filename, matches))
+				titlecased_title = f"{title_header} {subtitle} {title_footer}"
+				titlecased_title = titlecased_title.strip()
 
-	# Check for possessive 's within name italics, but not in ignored files like the colophon which we know have no possessives
-	# Allow some known exceptions like `Harper's`, etc.
-	if filename.name not in IGNORED_FILENAMES:
-		nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.') and re:test(., '’s$') and not(contains(@epub:type, 'se:name.publication.') and re:test(., '(Pearson’s|Harper’s|Blackwood’s|Fraser’s|Baedeker’s)$'))]")
-		if nodes:
-			messages.append(LintMessage("t-007", "Possessive [text]’s[/] within name italics. If the name in italics is doing the possessing, [text]’s[/] goes outside italics.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+				title = se.formatting.remove_tags(title).strip()
+				if title != titlecased_title:
+					messages.append(LintMessage("s-023", f"Title [text]{title}[/] not correctly titlecased. Expected: [text]{titlecased_title}[/].", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for lsquo not correctly closed
-	matches = regex.findall(r"‘[^“’]+?‘", file_contents)
-	matches = [match for match in matches if "</p" not in match and "<br/>" not in match]
-	if matches:
-		messages.append(LintMessage("t-004", "[text]‘[/] missing matching [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, matches))
+		# No subtitle? Much more straightforward
+		else:
+			titlecased_title = se.formatting.titlecase(se.formatting.remove_tags(title))
+			title = se.formatting.remove_tags(title)
+			if title != titlecased_title:
+				messages.append(LintMessage("s-023", f"Title [text]{title}[/] not correctly titlecased. Expected: [text]{titlecased_title}[/].", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for repeated punctuation, but first remove `&` so we don't match `&,`
-	# Remove tds with repeated ” as they are probably ditto marks
-	matches = regex.findall(r"[,;]{2,}.{0,20}", file_contents.replace("&", "")) + regex.findall(r"(?:“\s*“|”\s*”|’ ’|‘\s*‘).{0,20}", regex.sub(r"<td>[”\s]+?(<a .+?epub:type=\"noteref\">.+?</a>)?</td>", "", file_contents)) +	 regex.findall(r"[\p{Letter}][,\.:;]\s[,\.:;]\s?[\p{Letter}<].{0,20}", file_contents, flags=regex.IGNORECASE)
-	if matches:
-		messages.append(LintMessage("t-008", "Repeated punctuation.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Check for header elements that are entirely non-English
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./i[@xml:lang][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	if nodes:
+		messages.append(LintMessage("s-024", "Header elements that are entirely non-English should not be set in italics. Instead, the [xhtml]<h#>[/] element has the [attr]xml:lang[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for nbsp before times
-	nodes = dom.xpath(f"/html/body//text()[re:test(., '[0-9][^{se.NO_BREAK_SPACE}]?$')][(following-sibling::abbr[1])[re:test(., '^[ap]\\.m\\.$')]]")
+	# Check for z3998:roman elements with invalid values. Roman numerals can occasionally end in `j` as an alias for ending `i`. See _The Worm Ouroboros_.
+	# We also allow the numeral to end in a digit, because that might be an endnote. For example `<h2 epub:type="ordinal z3998:roman">II<a href="..." epub:type="noteref">3</a></h2>`
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:roman') and not(re:test(normalize-space(.), '^[ivxlcdmIVXLCDM]+j?[0-9]*?$'))]")
 	if nodes:
-		messages.append(LintMessage("t-009", "Required no-break space not found before time and [text]a.m.[/] or [text]p.m.[/].", se.MESSAGE_TYPE_WARNING, filename, [node[-10:] + "<abbr" for node in nodes]))
+		messages.append(LintMessage("s-026", "Invalid Roman numeral.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for times with periods instead of colons.
-	# Only check p, because things like tables/td are more likely to contain non-time numbers
-	# Exclude numbers preceded by equals, or succeeded by some measurements
-	# Also remove <a> first because they are likely to contain numbered section references
-	dom_copy = deepcopy(dom)
-	for node in dom_copy.xpath("/html/body//p/a"):
-		node.remove()
+	# Check for <h#> elements that are higher or lower than their expected level based on how deep they are in <section>s
+	# or <article>s. Exclude <nav> in the ToC.
+	# Get all h# elements not preceded by other h# elements (this includes the 1st item of an hgroup but not the next items)
+	# Exclude the ToC and landmarks
+	invalid_headers = []
+	invalid_parent_ids = []
+	for heading in dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') and not(./preceding-sibling::*[re:test(name(), '^h[1-6]$')]) and not(./ancestor::*[re:test(@epub:type, '\\b(toc|landmarks)\\b')])]"):
+		# Get the IDs of the direct parent section
+		parent_section = heading.xpath("./ancestor::*[@id and re:test(name(), '^(section|article|nav)$')][1]")
 
-	nodes = dom_copy.xpath("/html/body//p[re:test(., '[^=]\\s[0-9]{1,2}\\.[0-9]{2}(?![0-9′″°%]|\\.[0-9]|\\scubic|\\smetric|\\smeters|\\smiles|\\sfeet|\\sinches)')]")
-	matches = []
-	for node in nodes:
-		for time_match in regex.findall(r"(?<=[^=]\s)[0-9]{1,2}\.[0-9]{2}(?![0-9′″°%]|\.[0-9]|\scubic|\smetric|\smeters|\smiles|\sfeet|\sinches)", node.inner_text()):
-			time = time_match.split(".")
-			if not time[0].startswith("0") and int(time[0]) >= 1 and int(time[0]) <= 12 and int(time[1]) >= 0 and int(time[1]) <= 59:
-				matches.append(time_match)
+		if parent_section:
+			section = _find_ebook_section(parent_section[0].get_attr("id"), section_tree)
 
-	if matches:
-		messages.append(LintMessage("t-010", "Time set with [text].[/] instead of [text]:[/].", se.MESSAGE_TYPE_WARNING, filename, set(matches)))
+			if not section:
+				# We can accidentally raise s-029 if the file isn't in the spine. This is technically correct, but s-029 is a
+				# misleading message in that case, and f-007 will also be raised, which is the correct message.
+				if filename in self.spine_file_paths:
+					invalid_parent_ids.append(parent_section[0].to_tag_string())
+			else:
+				if str(section.depth) != heading.tag[1:2]:
+					invalid_headers.append(heading.to_string())
 
-	# Check for missing punctuation before closing quotes
-	# Exclude signatures in footers as those are commonly quoted without ending punctuation
-	nodes = dom.xpath("/html/body//p[not( (parent::header or parent::hgroup or (parent::footer and contains(@epub:type, 'z3998:signature'))) and position()=last())][re:test(., '[a-z]+[”’]$')]")
-	if nodes:
-		messages.append(LintMessage("t-011", "Missing punctuation before closing quotes.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string()[-30:] for node in nodes]))
+	if invalid_parent_ids:
+		messages.append(LintMessage("s-029", "Section with [attr]data-parent[/] attribute, but no section having that [attr]id[/] in ebook.", se.MESSAGE_TYPE_ERROR, filename, invalid_parent_ids))
+	if invalid_headers:
+		messages.append(LintMessage("s-085", "[xhtml]<h#>[/] element found in a [xhtml]<section>[/] or a [xhtml]<article>[/] at an unexpected level. Hint: Headings not in the title page start at [xhtml]<h2>[/]. If this work has parts, should this header be [xhtml]<h3>[/] or higher?", se.MESSAGE_TYPE_ERROR, filename, invalid_headers))
 
-	# Check for whitespace before noteref
-	# Do this early because we remove noterefs from headers later
-	# Allow noterefs that do not have preceding text/elements (for example a "see note X" ref that is the only child in a <td>)
-	nodes = dom.xpath("/html/body//a[contains(@epub:type, 'noteref') and preceding-sibling::node()[normalize-space(.)] and re:test(preceding-sibling::node()[1], '\\s+$')]")
-	if nodes:
-		messages.append(LintMessage("t-012", "Illegal white space before noteref.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	# Check for a common typo
+	if "z3998:nonfiction" in file_contents:
+		messages.append(LintMessage("s-030", "[val]z3998:nonfiction[/] should be [val]z3998:non-fiction[/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for period following Roman numeral, which is an old-timey style we must fix
-	# But ignore the numeral if it's the first item in a <p> tag, as that suggests it might be a kind of list item.
-	nodes = dom.xpath("/html/body//node()[name()='span' and contains(@epub:type, 'z3998:roman') and not(position()=1)][(following-sibling::node()[1])[re:test(., '^\\.\\s*[a-z]')]]")
-	if nodes:
-		messages.append(LintMessage("t-013", "Roman numeral followed by a period. When in mid-sentence Roman numerals must not be followed by a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "." for node in nodes]))
+	# Run some checks on epub:type values
+	incorrect_attrs = set()
+	unnecessary_z3998_attrs = set()
+	duplicate_attrs = []
 
-	# Check for two em dashes in a row
-	nodes = dom.xpath(f"/html/body//*[re:test(text(), '—{se.WORD_JOINER}*—+')]")
-	if nodes:
-		messages.append(LintMessage("t-014", "Two or more em-dashes in a row found. Elided words should use the two- or three-em-dash Unicode character, and dialog ending in em-dashes should only end in a single em-dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+	for node in dom.xpath("//*[@epub:type]"):
+		attrs = set()
 
-	# Check for money not separated by commas
-	nodes = dom.xpath("/html/body//*[re:test(text(), '[£\\$][0-9]{4,}')]")
-	if nodes:
-		messages.append(LintMessage("t-015", "Numbers not grouped by commas. Separate numbers greater than 1,000 with commas at every three numerals.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		for val in regex.split(r"\s+", node.get_attr("epub:type")):
+			if val in attrs:
+				duplicate_attrs.append(node.to_tag_string())
+			else:
+				attrs.add(val)
 
-	# Check for <abbr epub:type="z3998:*-name"> that does not contain spaces
-	nodes = dom.xpath("/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and re:test(., '[A-Z]\\.[A-Z]\\.')]")
-	if nodes:
-		messages.append(LintMessage("t-016", "Initials in [xhtml]<abbr epub:type=\"z3998:*-name\">[/] not separated by spaces.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+			if val.startswith("z3998:"):
+				bare_val = val.replace("z3998:", "")
+				if bare_val not in Z3998_SEMANTIC_VOCABULARY:
+					incorrect_attrs.add(val)
 
-	if special_file != "colophon":
-		# This file is not the colophon.
-		# Check for ending punctuation inside italics that have semantics.
-		# Ignore the colophon because paintings might have punctuation in their names
+				elif bare_val in EPUB_SEMANTIC_VOCABULARY:
+					unnecessary_z3998_attrs.add((val, bare_val))
 
-		# This xpath matches b or i elements with epub:type="se:name...", that are not stage direction, whose last text node ends in punctuation.
-		# Note that we check that the last node is a text node, because we may have <abbr> as the last node
-		matches = [node.to_string() for node in dom.xpath("(//b | //i)[contains(@epub:type, 'se:name') and not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., '[\\.,!\\?;:]$')]]")]
+			elif val.startswith("se:"):
+				bare_val = val.replace("se:", "")
+				if bare_val not in SE_SEMANTIC_VOCABULARY:
+					incorrect_attrs.add(val)
 
-		# Match b or i elements that are not stage directions, and that end in a comma followed by a lowercase letter
-		matches = matches + [node.to_string() for node in dom.xpath("(//b | //i)[not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., ',$')] and following-sibling::node()[re:test(., '^\\s*[a-z]')] ]")]
+			else:
+				# Regular epub vocabulary
+				if val not in EPUB_SEMANTIC_VOCABULARY:
+					incorrect_attrs.add(val)
 
-		# ...and also check for ending punctuation inside em tags, if it looks like a *part* of a clause
-		# instead of a whole clause. If the <em> is preceded by an em dash or quotes, or if there's punctuation
-		# and a space before it, then it's presumed to be a whole clause.
-		matches = matches + [match.strip() for match in regex.findall(r"(?<!.[—“‘>]|[!\.\?…;:]\s)<em>(?:\w+?\s*)+[\.,\!\?;]</em>", file_contents) if match.islower()]
+	if duplicate_attrs:
+		messages.append(LintMessage("s-031", "Duplicate value in [attr]epub:type[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, duplicate_attrs))
+	if incorrect_attrs:
+		messages.append(LintMessage("s-032", "Invalid value for [attr]epub:type[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, incorrect_attrs))
+	if unnecessary_z3998_attrs:
+		messages.append(LintMessage("s-034", "Semantic used from the z3998 vocabulary, but the same semantic exists in the EPUB vocabulary.", se.MESSAGE_TYPE_ERROR, filename, [attr for (attr, bare_attr) in unnecessary_z3998_attrs]))
 
-		if matches:
-			messages.append(LintMessage("t-017", "Ending punctuation inside formatting like bold, small caps, or italics. Ending punctuation is only allowed within formatting if the phrase is an independent clause.", se.MESSAGE_TYPE_WARNING, filename, list(set(matches))))
+	# Check if language tags in individual files match the language in the metadata file
+	if filename.name not in IGNORED_FILENAMES:
+		file_language = dom.xpath("/html/@xml:lang", True)
+		if language != file_language:
+			messages.append(LintMessage("s-033", f"File language is [val]{file_language}[/], but [path][link=file://{self.metadata_file_path}]{self.metadata_file_path.name}[/][/] language is [val]{language}[/].", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for stage direction that ends in ?! but also has a trailing period
-	nodes = dom.xpath("/html/body//i[contains(@epub:type, 'z3998:stage-direction') and re:test(., '\\.$') and not((./node()[last()])[name() = 'abbr']) and (following-sibling::node()[1])[re:test(., '^[\\.,:;!?]')]]")
+	# Check for endnotes
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[not(preceding-sibling::*)]/cite[not(preceding-sibling::node()[normalize-space(.)]) and (following-sibling::node()[normalize-space(.)])[1][contains(@epub:type, 'backlink')]]")
 	if nodes:
-		messages.append(LintMessage("t-018", "Stage direction ending in period next to other punctuation. Remove trailing periods in stage direction.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-035", "Endnote containing only [xhtml]<cite>[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for single words that are in italics, but that have closing punctuation outside italics
-	# Outer wrapping match is so that .findall returns the entire match and not the subgroup
-	# The first regex also matches the first few characters before the first double quote; we use those for more sophisticated
-	# checks below, to give fewer false positives like `with its downy red hairs and its “<i xml:lang="fr">doigts de faune</i>.”`
-	matches = regex.findall(r"((?:.{1,2}\s)?“<(i|em)[^>]*?>[^<]+?</\2>[\!\?\.])", file_contents) + regex.findall(r"([\.\!\?] <(i|em)[^>]*?>[^<]+?</\2>[\!\?\.])", file_contents)
+	# Check for wrong semantics in frontmatter/backmatter
+	if dom.xpath("/html/body/section[not(@data-parent) and re:test(@epub:type, '\\b(dedication|introduction|preface|foreword|preamble|titlepage|halftitlepage|imprint|epigraph|acknowledgements)\\b') and not(ancestor-or-self::*[contains(@epub:type, 'frontmatter')])]"):
+		messages.append(LintMessage("s-036", "No [val]frontmatter[/] semantic inflection for what looks like a frontmatter file.", se.MESSAGE_TYPE_WARNING, filename))
 
-	# But, if we've matched a name of something, don't include that as an error. For example, `He said, “<i epub:type="se:name.publication.book">The Decameron</i>.”`
-	# We also exclude the match from the list if:
-	# 1. The double quote is directly preceded by a lowercase letter and a space: `with its downy red hairs and its “<i xml:lang="fr">doigts de faune</i>.”`
-	# 2. The double quote is directly preceded by a lowercase letter, a comma, and a space, and the first letter within the double quote is lowercase: In the original, “<i xml:lang="es">que era un Conde de Irlos</i>.”
-	# 3. The text is a single letter that is not "I" or "a" (because then it is likely a mathematical variable)
-	matches = [match for match in matches if "epub:type=\"se:name." not in match[0] and "epub:type=\"z3998:taxonomy" not in match[0] and not regex.match(r"^[\p{Lowercase_Letter}’]+\s“", match[0]) and not regex.match(r"^[\p{Lowercase_Letter}’]+,\s“[\p{Lowercase_Letter}]", se.formatting.remove_tags(match[0])) and not regex.match(r"^.*?<.+?>[^Ia]<.+?>", match[0])]
-	if matches:
-		messages.append(LintMessage("t-019", "When a complete clause is italicized, ending punctuation except commas must be within containing italics.", se.MESSAGE_TYPE_WARNING, filename, [match[0] for match in matches]))
+	if dom.xpath("/html/body/section[not(@data-parent) and re:test(@epub:type, '\\b(endnotes|loi|afterword|appendix|colophon|copyright\\-page|lot)\\b') and not(ancestor-or-self::*[contains(@epub:type, 'backmatter')])]"):
+		messages.append(LintMessage("s-037", "No [val]backmatter[/] semantic inflection for what looks like a backmatter file.", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Check for punctuation after endnotes
-	nodes = dom.xpath(f"/html/body//a[contains(@epub:type, 'noteref')][(following-sibling::node()[1])[re:test(., '^[^\\s<–\\]\\)—{se.WORD_JOINER}a-zA-Z0-9]')]]")
+	# Check for leftover asterisms. Asterisms are sequences of any of these chars: * . • -⁠ —
+	nodes = dom.xpath("/html/body//*[self::p or self::div][re:test(., '^\\s*[\\*\\.•\\-⁠—]\\s*([\\*\\.•\\-⁠—]\\s*)+$')]")
 	if nodes:
-		messages.append(LintMessage("t-020", "Endnote links must be outside of punctuation, including quotation marks.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
-
-	# Check for correct typography around measurements like 2 ft.
-	# But first remove href and id attrs because URLs and IDs may contain strings that look like measurements
-	# Note that while we check m,min (minutes) and h,hr (hours) we don't check s (seconds) because we get too many false positives on years, like `the 1540s`
-	matches = regex.findall(fr"\b[1-9][0-9]*[{se.NO_BREAK_SPACE}\-]?(?:[mck]?[mgl]|ft|in|min?|h|sec|hr)\.?\b", regex.sub(r"(href|id)=\"[^\"]*?\"", "", file_contents))
-	# Exclude number ordinals, they're not measurements
-	matches = [match for match in matches if not regex.search(r"(st|nd|rd|th)", match)]
-	if matches:
-		messages.append(LintMessage("t-021", "Measurement not to standard. Numbers are followed by a no-break space and abbreviated units require an [xhtml]<abbr>[/] element. See [path][link=https://standardebooks.org/manual/1.0.0/8-typography#8.8.5]semos://1.0.0/8.8.5[/][/].", se.MESSAGE_TYPE_WARNING, filename, matches))
+		messages.append(LintMessage("s-038", "Illegal asterism. Section/scene breaks must be defined by an [xhtml]<hr/>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for nbsp within <abbr epub:type="z3998:*-name">, which is redundant
-	nodes = dom.xpath(f"/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and contains(text(), '{se.NO_BREAK_SPACE}')]")
-	if nodes:
-		messages.append(LintMessage("t-022", "No-break space found in [xhtml]<abbr epub:type=\"z3998:*-name\">[/]. This is redundant.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	# Check for <table> element without a <tbody> child
+	if dom.xpath("/html/body//table[not(tbody)]"):
+		messages.append(LintMessage("s-042", "[xhtml]<table>[/] element without [xhtml]<tbody>[/] child.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for trailing commas inside <i> tags at the close of dialog
-	# More sophisticated version of: \b[^\s]+?,</i>”
-	nodes = dom.xpath("/html/body//i[re:test(., ',$')][(following-sibling::node()[1])[starts-with(., '”')]]")
+	# Check that short stories are on an <article> element
+	nodes = dom.xpath("/html/body/section[contains(@epub:type, 'se:short-story') or contains(@epub:type, 'se:novella')]")
 	if nodes:
-		messages.append(LintMessage("t-023", "Comma inside [xhtml]<i>[/] element before closing dialog.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "”" for node in nodes]))
+		messages.append(LintMessage("s-043", "[val]se:short-story[/] semantic on element that is not [xhtml]<article>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for quotation marks in italicized dialog
-	nodes = dom.xpath("/html/body//i[@xml:lang][starts-with(., '“') or re:test(., '”$')]")
+	# Check for poetry/verse without a descendent <p> element.
+	# Skip the ToC landmarks because it may have poem/verse semantic children.
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')][not(descendant::p)][not(ancestor::nav[contains(@epub:type, 'landmarks')])]")
 	if nodes:
-		messages.append(LintMessage("t-024", "When italicizing language in dialog, italics go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-044", "Element with poem or verse semantic, without descendant [xhtml]<p>[/] (stanza) element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check alt attributes on images, except for the logo
-	nodes = dom.xpath("/html/body//img[not(re:test(@src, '/(logo|titlepage)\\.(svg|png)$'))]")
-	img_no_alt = []
-	img_alt_not_typogrified = []
-	img_alt_lacking_punctuation = []
+	# Check for <abbr> elements that have two or more letters/periods, that don't have a semantic epub:type
+	# SS. is the French abbreviation for "Saints"
+	nodes = dom.xpath("/html/body//abbr[not(@epub:type)][text() != 'U.S.' and text() != 'SS.'][re:test(., '([A-Z]\\.?){2,}')]")
+	filtered_nodes = []
 	for node in nodes:
-		if "titlepage.svg" not in node.get_attr("src"):
-			has_images = True # Save for a later check
+		add_node = True
 
-		alt = node.get_attr("alt")
+		if node.text in INITIALISM_EXCEPTIONS:
+			add_node = False
+		elif node.get_attr("class"):
+			for attr_value in node.get_attr("class").split():
+				if attr_value in IGNORED_CLASSES:
+					add_node = False
+					break
+		elif node.text in ("A.M.", "P.M.") and node.xpath("./preceding-sibling::node()[re:test(., '[0-9]\\s$')]"):
+			# Ignore instances of capitalized times, usually in titles. For example:
+			# `From 10:20 <abbr>P.M.</abbr> to 10:47 <abbr>P.M.</abbr>` in a chapter title.
+			# We have to test for a preceding number because `<abbr>P.M.</abbr>` in other contexts may mean `Prime Minister`
+			add_node = False
 
-		if alt:
-			# Check for non-typogrified img alt attributes
-			if regex.search(r"""('|"|--|\s-\s|")""", alt):
-				img_alt_not_typogrified.append(node.to_tag_string())
+		if add_node:
+			filtered_nodes.append(node)
 
-			# Check alt attributes not ending in punctuation
-			if filename.name not in IGNORED_FILENAMES and not regex.search(r"""[\.\!\?]”?$""", alt):
-				img_alt_lacking_punctuation.append(node.to_tag_string())
+	if filtered_nodes:
+		messages.append(LintMessage("s-045", "[xhtml]<abbr>[/] element without semantic inflection like [class]z3998:personal-name[/] or [class]z3998:initialism[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in filtered_nodes]))
 
-			# Check that alt attributes match SVG titles
-			img_src = node.lxml_element.get("src")
-			if img_src and img_src.endswith("svg"):
-				title_text = ""
-				image_ref = img_src.split("/").pop()
-				try:
-					svg_path = self.content_path / "images" / image_ref
-					svg_dom = self.get_dom(svg_path)
-					try:
-						title_text = svg_dom.xpath("/svg/title")[0].text
-					except Exception:
-						messages.append(LintMessage("s-027", f"{image_ref} missing [xhtml]<title>[/] element.", se.MESSAGE_TYPE_ERROR, svg_path))
+	# Check for <p> elems that has some element children, which are only <span> and <br> children, but the parent doesn't have poem/verse semantics.
+	# Ignore spans that have a class, but not if the class is an i# class (for poetry indentation)
+	nodes = dom.xpath("/html/body//p[not(./text()[normalize-space(.)])][*][not(ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][not(*[not(self::span) and not(self::br)])][ (not(span[@epub:type]) and not(span[@class]) ) or span[re:test(@class, '\\bi[0-9]\\b')]]")
+	if nodes:
+		messages.append(LintMessage("s-046", "[xhtml]<p>[/] element containing only [xhtml]<span>[/] and [xhtml]<br>[/] elements, but its parent doesn’t have the [val]z3998:poem[/], [val]z3998:verse[/], [val]z3998:song[/], [val]z3998:hymn[/], or [val]z3998:lyrics[/] semantic. Multi-line clauses that are not verse don’t require [xhtml]<span>[/]s.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-					if title_text != "" and alt != "" and title_text != alt:
-						messages.append(LintMessage("s-022", f"The [xhtml]<title>[/] element of [path][link=file://{svg_path}]{image_ref}[/][/] does not match the [attr]alt[/] attribute text in [path][link=file://{filename}]{filename.name}[/][/].", se.MESSAGE_TYPE_ERROR, filename))
+	# For this series of selections, we select spans that are direct children of p, because sometimes a line of poetry may have a nested span.
+	nodes = dom.xpath("/html/body/*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')]/descendant-or-self::*/p/span/following-sibling::*[contains(@epub:type, 'noteref') and name()='a' and position()=1]")
+	if nodes:
+		messages.append(LintMessage("s-047", "[val]noteref[/] as a direct child of element with poem or verse semantic. [val]noteref[/]s should be in their parent [xhtml]<span>[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-				except FileNotFoundError:
-					missing_files.append(str(Path("images") / image_ref))
+	# Check for incorrectly applied se:name semantic
+	nodes = dom.xpath("/html/body//*[self::p or self::blockquote][contains(@epub:type, 'se:name.')]")
+	if nodes:
+		messages.append(LintMessage("s-048", "[val]se:name[/] semantic on block element. [val]se:name[/] indicates the contents is the name of something.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-		else:
-			img_no_alt.append(node.to_tag_string())
+	# Check for <header> elements with only h# child nodes
+	nodes = dom.xpath("/html/body//header[./*[re:test(name(), 'h[1-6]') and (count(preceding-sibling::*) + count(following-sibling::*)=0)]]")
+	if nodes:
+		messages.append(LintMessage("s-049", "[xhtml]<header>[/] element whose only child is an [xhtml]<h#>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	if img_alt_not_typogrified:
-		messages.append(LintMessage("t-025", "Non-typogrified [text]'[/], [text]\"[/] (as [xhtml]"[/]), or [text]--[/] in image [attr]alt[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, img_alt_not_typogrified))
-	if img_alt_lacking_punctuation:
-		messages.append(LintMessage("t-026", "[attr]alt[/] attribute does not appear to end with punctuation. [attr]alt[/] attributes must be composed of complete sentences ending in appropriate punctuation.", se.MESSAGE_TYPE_ERROR, filename, img_alt_lacking_punctuation))
-	if img_no_alt:
-		messages.append(LintMessage("s-004", "[xhtml]img[/] element missing [attr]alt[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, img_no_alt))
+	# Check for <span>s that only exist to apply epub:type
+	nodes = dom.xpath("/html/body//*[span[@epub:type][count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
+	if nodes:
+		messages.append(LintMessage("s-050", "[xhtml]<span>[/] element appears to exist only to apply [attr]epub:type[/]. [attr]epub:type[/] should go on the parent element instead, without a [xhtml]<span>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for low-hanging misquoted fruit
-	matches = regex.findall(r"[\p{Letter}]+[“‘]", file_contents) + regex.findall(r"[^>]+</(?:em|i|b|span)>‘[\p{Lowercase_Letter}]+", file_contents)
-	if matches:
-		messages.append(LintMessage("t-028", "Possible mis-curled quotation mark.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Check for language tags misusing the unk lang code
+	nodes = dom.xpath("//*[@xml:lang='unk']")
+	if nodes:
+		messages.append(LintMessage("s-051", "Element with [xhtml]xml:lang=\"unk\"[/] should be [xhtml]xml:lang=\"und\"[/] (undefined).", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for periods followed by lowercase.
-	temp_xhtml = regex.sub(r"<title[^>]*?>.+?", "", file_contents) # Remove  because it might contain something like <title>Chapter 2: The Antechamber of M. de Tréville
-	temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Replace things like 
-	temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Remove  attributes
-	temp_xhtml = temp_xhtml.replace("A.B.C.", "X") # Remove A.B.C, which is not an abbreviations.
-	# Note the regex also excludes preceding numbers, so that we can have inline numbering like:
-	# "A number of questions: 1. regarding those who make heretics; 2. concerning those who were made heretics..."
-	matches = regex.findall(r"[^\s0-9]+\.\s+[\p{Lowercase_Letter}](?!’[\p{Uppercase_Letter}])[\p{Lowercase_Letter}]+", temp_xhtml)
-	# If  is in the match, remove it from the matches so we exclude things like et. al.
-	matches = [match for match in matches if "" not in match]
-	if matches:
-		messages.append(LintMessage("t-029", "Period followed by lowercase letter. Hint: Abbreviations require an [xhtml][/] element.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Check for title attrs on abbr elements
+	nodes = dom.xpath("/html/body//abbr[@title]")
+	if nodes:
+		messages.append(LintMessage("s-052", "[xhtml][/] element with illegal [attr]title[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for initialisms without periods
-	nodes = [node.to_string() for node in dom.xpath("/html/body//abbr[contains(@epub:type, 'z3998:initialism') and not(re:test(., '^[0-9]*([a-zA-Z]\\.)+[0-9]*$'))]") if node.text not in INITIALISM_EXCEPTIONS]
+	nodes = dom.xpath("/html/body//blockquote//p[parent::*[name()='footer'] or parent::*[name()='blockquote']]//cite") # Sometimes the 
     may be in a 
    
 	if nodes:
-		messages.append(LintMessage("t-030", "Initialism with spaces or without periods.", se.MESSAGE_TYPE_WARNING, filename, set(nodes)))
+		messages.append(LintMessage("s-054", "[xhtml][/] as child of [xhtml]
    [/] in [xhtml]
    [/]. [xhtml][/] should be the direct child of [xhtml]
    [/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	nodes = dom.xpath("/html/body//*[re:test(text(), '\\bA\\s*B\\s*C\\s*\\b')]")
+	# Check for 
    		 element without a 
    	and  are allowed, for use in tables with headers in the middle of tables (url:https://standardebooks.org/ebooks/dorothy-day/the-eleventh-virgin) and vertical table headers
+	# (https://standardebooks.org/ebooks/charles-babbage/passages-from-the-life-of-a-philosopher)
+	if dom.xpath("/html/body//table//th[not(ancestor::thead)][not(@scope)][not(count(node())=0)]"):
+		messages.append(LintMessage("s-055", "[xhtml][/] element not in [xhtml]
    [/] elements used as mid-table headings or horizontal row headings require the [attr]scope[/] attribute.", se.MESSAGE_TYPE_ERROR, filename))
+
+	# Check for z3998:stage-direction on elements that are not 
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'z3998:stage-direction') and name() !='i' and name() !='abbr' and name() !='p']")
 	if nodes:
-		messages.append(LintMessage("t-031", "[text]A B C[/] must be set as [text]A.B.C.[/] It is not an abbreviation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-058", "[attr]z3998:stage-direction[/] semantic only allowed on [xhtml][/], [xhtml][/], and [xhtml]
    [/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for abbreviations followed by periods
-	# But we exclude some SI units, which don't take periods; abbreviations ending in numbers for example in stage directions; abbreviations like `r^o` (recto) that contain ; and some Imperial abbreviations that are multi-word
-	nodes = dom.xpath("/html/body//abbr[(contains(@epub:type, 'z3998:initialism') or re:test(@epub:type, 'z3998:[^\\s]+?name') or not(@epub:type))][not(re:test(., '[cmk][mgl]')) and not(re:test(., '[0-9]$')) and not(./sup) and not(re:test(@class, '\\b(era|temperature|compound)\\b')) and not(re:test(text(), '^(mpg|mph|hp|TV)$'))][following-sibling::text()[1][starts-with(self::text(), '.')]]")
+	# Check that internal links don't begin with ../
+	nodes = dom.xpath("/html/body//a[re:test(@href, '^\\.\\./text/')]")
 	if nodes:
-		messages.append(LintMessage("t-032", "Initialism or name followed by period. Hint: Periods go within [xhtml][/]. [xhtml][/]s containing periods that end a clause require the [class]eoc[/] class.", se.MESSAGE_TYPE_WARNING, filename, [f"{node.to_string()}." for node in nodes]))
+		messages.append(LintMessage("s-059", "Internal link beginning with [val]../text/[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for space after dash
-	nodes = dom.xpath("/html/body//*[name()='p' or name()='span' or name='em' or name='i' or name='b' or name='strong'][not(self::comment())][re:test(., '[a-zA-Z]-\\s(?!(and|or|nor|to|und|…)\\b)')]")
+	# Check for italics on things that shouldn't be italics
+	nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.music.song') or contains(@epub:type, 'se:name.publication.short-story') or contains(@epub:type, 'se:name.publication.essay')]")
 	if nodes:
-		messages.append(LintMessage("t-033", "Space after dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-060", "Italics on name that requires quotes instead.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for  preceded by em dash
-	nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[re:test(., '—$')]]")
+	# Check for h# tags followed by header content, that are not children of 
    .
+	# Only match if there is a following 
    , because we could have the case where there's an epigraph after a division title.
+	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6][following-sibling::*[contains(@epub:type, 'epigraph') or contains(@epub:type, 'bridgehead')]][following-sibling::p][not(parent::header)]")
 	if nodes:
-		messages.append(LintMessage("t-034", "[xhtml][/] element preceded by em-dash. Hint: em-dashes go within [xhtml][/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-061", "Title and following header content not in a [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for  without preceding space in text node. (preceding ( or [ are also OK)
-	nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[not(re:test(., '[\\[\\(\\s]$'))]]")
+	# Check for 
     elements without exactly one  child, but only in glossaries
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'glossary')]//dt[not(count(./dfn)=1)]")
 	if nodes:
-		messages.append(LintMessage("t-035", "[xhtml][/] element not preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-062", "[xhtml]
    [/] element in a glossary without exactly one [xhtml][/] child.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for partially obscured years in which the last year is an em dash
-	nodes = dom.xpath("/html/body//p[re:test(.,	 '1\\d{2}⁠—[^“’A-Za-z<]')]")
+	# Check that `z3998:persona` is only on  or 
    		. We use the contact() xpath function so we don't catch `z3998:personal-name`
+	nodes = dom.xpath("/html/body//*[contains(concat(' ', @epub:type, ' '), ' z3998:persona ') and not(self::b or self::td)]")
 	if nodes:
-		messages.append(LintMessage("t-036", "Em-dash used to obscure single digit in year. Hint: Use a hyphen instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-063", "[val]z3998:persona[/] semantic on element that is not a [xhtml][/] or [xhtml][/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for rdquo preceded by space (but not preceded by a rsquo, which might indicate a nested quotation)
-	nodes = dom.xpath("/html/body//p[re:test(., '[^’]\\s”')]")
+	# Check for fulltitle semantic on a header not in the half title
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'fulltitle') and name()!='h2' and name()!='hgroup' and not(ancestor::*[contains(@epub:type, 'halftitlepage')])]")
 	if nodes:
-		messages.append(LintMessage("t-037", "[text]”[/] preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-065", "[val]fulltitle[/] semantic on element that is not in the half title.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check obviously miscurled quotation marks
-	nodes = dom.xpath("/html/body//p[re:test(., '“$')]")
+	# Check for header elements that have a label, but are missing the label semantic
+	# Find h# nodes whose first child is a text node matching a label type, and where that text node's next sibling is a semantic roman numeral
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./node()[1][self::text() and not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)$') and following-sibling::*[1][contains(@epub:type, 'z3998:roman')]]]")
 	if nodes:
-		messages.append(LintMessage("t-038", "[text]“[/] before closing [xhtml]
    [/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-066", "Header element missing [val]label[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for some known initialisms with incorrect possessive apostrophes
-	nodes = dom.xpath("/html/body//abbr[text()='I.O.U.'][(following-sibling::node()[1])[starts-with(., '’s')]]")
+	# Check for header elements that have a label semantic, but are missing an ordinal sibling
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][./span[contains(@epub:type, 'label')]][not(./span[contains(@epub:type, 'ordinal')])]")
 	if nodes:
-		messages.append(LintMessage("t-039", "Initialism followed by [text]’s[/]. Hint: Plurals of initialisms are not followed by [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "’s" for node in nodes]))
+		messages.append(LintMessage("s-067", "Header element with a [val]label[/] semantic child, but without an [val]ordinal[/] semantic child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check if a subtitle ends in a text node with a terminal period; or if it ends in an  node containing a terminal period.
-	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6]/*[contains(@epub:type, 'subtitle')][(./text())[last()][re:test(., '\\.$')] or (./i)[last()][re:test(., '\\.$')]]")
+	# Check for header elements with a roman semantic but without an ordinal semantic
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][contains(@epub:type, 'z3998:roman') and not(contains(@epub:type, 'ordinal'))]")
 	if nodes:
-		messages.append(LintMessage("t-040", "Subtitle with illegal ending period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-068", "Header element missing [val]ordinal[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	matches = regex.findall(r"[^…]\s[!?;:,].{0,10}", file_contents) # If we don't include preceding chars, the regex is 6x faster
-	if matches:
-		messages.append(LintMessage("t-041", "Illegal space before punctuation.", se.MESSAGE_TYPE_ERROR, filename, matches))
+	# Check for body element without child section or article. Ignore the ToC because it has a unique structure
+	nodes = dom.xpath("/html/body[not(./*[name()='section' or name()='article' or (name()='nav' and re:test(@epub:type, '\\b(toc|loi)\\b'))])]")
+	if nodes:
+		messages.append(LintMessage("s-069", "[xhtml][/] element missing direct child [xhtml]
    [/] or [xhtml]
    [/] element.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for a list of some common English loan words that used to be italicized as foreign, but are no longer
-	nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and re:test(., '^(fianc[eé]+|divorc[eé]+|menu|recherch[eé]|tour-de-force|outr[eé]|d[ée]but(ante)?|apropos|[eé]lite|prot[ée]g[ée]+|chef|salon|r[eé]gime|contretemps|[eé]clat|aides?|entr[ée]+|t[eê]te-[aà]-t[eê]tes?|blas[eé]|bourgeoisie)$')]")
+	# Check for h# without semantics; h# that have child elements (which likely have the correct semantics) are OK
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(@epub:type)][not(./*[not(name()='a' and contains(@epub:type, 'noteref'))])]")
 	if nodes:
-		messages.append(LintMessage("t-043", "Non-English loan word set in italics, when modern typography omits italics. Hint: A word may be correctly italicized when emphasis is desired, if the word is meant to be pronounced with an accent, or if the word is part of non-English speech.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-070", "[xhtml][/] element without semantic inflection.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for comma after leading Or in subtitles
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'subtitle') and re:test(text(), '^Or\\s')]")
+	# Check for sectioning elements with more than one heading element
+	nodes = dom.xpath("/html/body//*[name()='article' or name()='section'][count(./*[name()='header' or name()='hgroup' or re:test(name(), '^h[1-6]$')]) > 1]")
 	if nodes:
-		messages.append(LintMessage("t-044", "Comma required after leading [text]Or[/] in subtitles.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-071", "Sectioning element with more than one heading element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for personas that are italicized. Use re:test instead of contains to avoid catching z3998:personal-name
-	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:persona\\b') and @data-css-font-style = 'italic']")
+	# Check for some elements that only have a  child (including text inline to the parent).
+	# But, exclude such elements that are 
     elements that have poetry-type parents.
+	nodes = dom.xpath("/html/body//*[name() !='p' and not(ancestor-or-self::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn|lyrics)')])][./span[count(preceding-sibling::node()[normalize-space(.)]) + count(following-sibling::node()[normalize-space(.)])=0]]")
 	if nodes:
-		messages.append(LintMessage("t-045", "Element has [val]z3998:persona[/] semantic and is also set in italics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-072", "Element with single [xhtml][/] child. [xhtml][/] should be removed and its attributes promoted to the parent element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	if "῾" in file_contents:
-		messages.append(LintMessage("t-046", "[text]῾[/] (U+1FFE) detected. Use [text]ʽ[/] (U+02BD) instead.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for header elements that are missing both label and ordinal semantics
+	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$')][not(./*) and re:test(normalize-space(.), '^(Part|Book|Volume|Section|Act|Scene)\\s+[ixvIXVmcd]+$')]")
+	if nodes:
+		messages.append(LintMessage("s-073", "Header element that requires [val]label[/] and [val]ordinal[/] semantic children.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for all-caps first paragraphs in sections/articles
-	# Note that we don't check for small-caps CSS, because there are lots of legitimate cases for that,
-	# and it would generate too many false positives.
-	# If we did want to do that, this xpath after the last re:test would help: or ./*[not(preceding-sibling::node()[normalize-space(.)]) and @data-css-font-variant='small-caps' and following-sibling::node()[1][self::text()]]
-	nodes = dom.xpath("/html/body//*[(name() = 'section' or name() = 'article') and not(contains(@epub:type, 'dedication') or contains(@epub:type, 'z3998:letter'))]/p[1][re:test(normalize-space(.), '^[“’]?(?:[A-Z’]+\\s){2,}')]")
+	# Check for table headers that don't have content, which is an accessibility issue.
+	# Note that @aria-label and @title should (apparently) not be used for table headers
+	nodes = dom.xpath("/html/body//th[not(text())]")
 	if nodes:
-		messages.append(LintMessage("t-048", "Chapter opening text in all-caps.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-074", "[xhtml]
    		[/] element with no text content should be a [xhtml][/] element instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for two-em-dashes used for elision instead of three-em-dashes
-	matches = regex.findall(fr"[^{se.WORD_JOINER}\p{{Letter}}”]⸺[^“{se.WORD_JOINER}\p{{Letter}}].*", file_contents, flags=regex.MULTILINE)
-	if matches:
-		messages.append(LintMessage("t-049", "Two-em-dash used for eliding an entire word. Use a three-em-dash instead.", se.MESSAGE_TYPE_WARNING, filename, matches))
+	# Check for incorrect children of 
+	nodes = dom.xpath("/html/body/*[name() != 'section' and name() != 'article' and name() != 'nav']")
+	if nodes:
+		messages.append(LintMessage("s-075", "[xhtml][/] element with direct child that is not [xhtml]
    [/], [xhtml]
    [/], or [xhtml]
    		 that contains both inline text and block-level elements.
+	# Note that we check any child of  because often  has the drama semantic.
+	nodes = dom.xpath("/html//*[contains(@epub:type, 'z3998:drama')]//td[./text()[normalize-space(.)] and ./*[name() = 'p' or name() = 'blockquote' or name() = 'div']]")
 	if nodes:
-		messages.append(LintMessage("t-054", "Epigraphs that are entirely non-English should be set in italics, not Roman.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-080", "[xhtml][/] in drama containing both inline text and a block-level element. All children should either be only text, or only block-level elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for lone acute accents, which should always be combined or swapped for a more accurate Unicode char
-	if "´" in file_contents:
-		messages.append(LintMessage("t-055", "Lone acute accent ([val]´[/]). A more accurate Unicode character like prime for coordinates or measurements, or combining accent or breathing mark for Greek text, is required.", se.MESSAGE_TYPE_ERROR, filename))
+	# Check for 
     tags that follow figure, blockquote, or table, and start with a lowercase letter, but that don't have the `continued` class. Exclude matches whose first child is , so we don't match things like 
    —Editor
+	nodes = dom.xpath("/html/body//p[preceding-sibling::*[1][name() = 'figure' or name() = 'blockquote' or name() = 'table'] and re:test(., '^([a-z]|—|…)') and not(contains(@class, 'continued')) and not(./*[1][self::cite])]")
+	if nodes:
+		messages.append(LintMessage("s-081", "[xhtml]
    [/] preceded by [xhtml]
    [/], [xhtml]
    [/xhtml], or [xhtml][/], but without [val]continued[/] class.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for degree confusable
-	nodes = dom.xpath("/html/body//*[re:test(text(), '[0-9]+º')]")
+	# Check for language tags transliterated into Latin script but missing `-Latn` suffix
+	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '^(he|ru|el|zh|bn|hi|sa|uk|yi|grc)$') and re:test(., '[a-zA-Z]')]")
 	if nodes:
-		messages.append(LintMessage("t-056", "Masculine ordinal indicator ([val]º[/]) used instead of degree symbol ([val]°[/]). Note that the masculine ordinal indicator may be appropriate for ordinal numbers read as Latin, i.e. [val]12º[/] reading [val]duodecimo[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-082", "Element containing Latin script for a non-Latin-script language, but its [attr]xml:lang[/] attribute value is missing the [val]-Latn[/] language tag suffix. Hint: For example Russian transliterated into Latin script would be [val]ru-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for 
     starting in lowercase
-	# Exclude 
     starting in constructs like `(a)` or `a.` as they may be numbered lists
-	nodes = dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::figure or ancestor::ol[not(ancestor::section[contains(@epub:type, 'endnotes')])] or ancestor::ul or ancestor::table or ancestor::hgroup or preceding-sibling::*[1][name() = 'hr']) and not(contains(@class, 'continued') or re:test(@epub:type, 'z3998:(signature|valediction)')) and not(./*[1][name() = 'math' or name() = 'var' or contains(@epub:type, 'z3998:grapheme')]) and re:test(., '^[^A-Za-z0-9]?[a-z]') and not(re:test(., '^\\([a-z0-9]\\.?\\)\\.?')) and not(re:test(., '^[a-z0-9]\\.\\s'))]")
+	# Check for persona 
    " not in match]
+	# xpath to check for opening quote in p, without a next child p that starts with an opening quote or an opening bracket (for editorial insertions within paragraphs of quotation); or that consists of only an ellipses (like an elided part of a longer quotation)
+	# Matching 
    s can't have a poem/verse ancestor as formatting is often special for those.
+	matches = matches + [regex.findall(r"“[^”]+
    ", node.to_string())[0] for node in dom.xpath("/html/body//p[re:test(., '“[^‘”]+$')][not(ancestor::*[re:test(@epub:type, 'z3998:(verse|poem|song|hymn|lyrics)')])][(following-sibling::*[1])[name()='p'][not(re:test(normalize-space(.), '^[“\\[]') or re:test(normalize-space(.), '^…$'))]]")]
 
-	# Check for double greater-than at the end of a tag
-	matches = regex.findall(r"(>>|>>)", file_contents)
+	# Additionally, match short 
     tags (< 100 chars) that lack closing quote, and whose direct siblings do have closing quotes (to exclude runs of same-speaker dialog), and that is not within a blockquote, verse, or letter
+	matches = matches + [regex.findall(r"“[^”]+
    ", node.to_string())[0] for node in dom.xpath("/html/body//p[re:test(., '“[^‘”]+$') and not(re:test(., '[…:]$')) and string-length(normalize-space(.)) <=100][(following-sibling::*[1])[not(re:test(., '“[^”]+$'))] and (preceding-sibling::*[1])[not(re:test(., '“[^”]+$'))]][not(ancestor::*[re:test(@epub:type, 'z3998:(verse|poem|song|hymn|lyrics)')]) and not(ancestor::blockquote) and not (ancestor::*[contains(@epub:type, 'z3998:letter')])][(following-sibling::*[1])[name()='p'][re:test(normalize-space(.), '^[“\\[]') and not(contains(., 'continued'))]]")]
 	if matches:
-		messages.append(LintMessage("x-008", "Elements should end with a single [text]>[/].", se.MESSAGE_TYPE_WARNING, filename))
+		messages.append(LintMessage("t-003", "[text]“[/] missing matching [text]”[/]. Note: When dialog from the same speaker spans multiple [xhtml]
    [/] elements, it’s correct grammar to omit closing [text]”[/] until the last [xhtml]
    [/] of dialog.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	# Check for leading 0 in IDs (note: not the same as checking for IDs that start with an integer)
-	# We only check for *leading* 0s in numbers; this allows IDs like `wind-force-0` in the Worst Journey in the World glossary.
-	nodes = dom.xpath("//*[re:test(@id, '-0[0-9]')]")
+	# Check for closing dialog without comma
+	matches = regex.findall(r"[\p{Lowercase_Letter}]+?” [\p{Letter}]+? said", file_contents)
+	if matches:
+		messages.append(LintMessage("t-005", "Dialog without ending comma.", se.MESSAGE_TYPE_WARNING, filename, matches))
+
+	# Check for possessive 's within name italics, but not in ignored files like the colophon which we know have no possessives
+	# Allow some known exceptions like `Harper's`, etc.
+	if filename.name not in IGNORED_FILENAMES:
+		nodes = dom.xpath("/html/body//i[contains(@epub:type, 'se:name.') and re:test(., '’s$') and not(contains(@epub:type, 'se:name.publication.') and re:test(., '(Pearson’s|Harper’s|Blackwood’s|Fraser’s|Baedeker’s)$'))]")
+		if nodes:
+			messages.append(LintMessage("t-007", "Possessive [text]’s[/] within name italics. If the name in italics is doing the possessing, [text]’s[/] goes outside italics.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+
+	# Check for lsquo not correctly closed
+	matches = regex.findall(r"‘[^“’]+?‘", file_contents)
+	matches = [match for match in matches if "" not in match]
+	if matches:
+		messages.append(LintMessage("t-004", "[text]‘[/] missing matching [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, matches))
+
+	# Check for repeated punctuation, but first remove `&` so we don't match `&,`
+	# Remove tds with repeated ” as they are probably ditto marks
+	matches = regex.findall(r"[,;]{2,}.{0,20}", file_contents.replace("&", "")) + regex.findall(r"(?:“\s*“|”\s*”|’ ’|‘\s*‘).{0,20}", regex.sub(r"
    ", "", file_contents)) +	 regex.findall(r"[\p{Letter}][,\.:;]\s[,\.:;]\s?[\p{Letter}<].{0,20}", file_contents, flags=regex.IGNORECASE)
+	if matches:
+		messages.append(LintMessage("t-008", "Repeated punctuation.", se.MESSAGE_TYPE_WARNING, filename, matches))
+
+	# Check for nbsp before times
+	nodes = dom.xpath(f"/html/body//text()[re:test(., '[0-9][^{se.NO_BREAK_SPACE}]?$')][(following-sibling::abbr[1])[re:test(., '^[ap]\\.m\\.$')]]")
 	if nodes:
-		messages.append(LintMessage("x-009", "Illegal leading 0 in [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-009", "Required no-break space not found before time and [text]a.m.[/] or [text]p.m.[/].", se.MESSAGE_TYPE_WARNING, filename, [node[-10:] + " tags
-	nodes = dom.xpath("/html/head/title/*")
+	# Check for times with periods instead of colons.
+	# Only check p, because things like tables/td are more likely to contain non-time numbers
+	# Exclude numbers preceded by equals, or succeeded by some measurements
+	# Also remove  first because they are likely to contain numbered section references
+	dom_copy = deepcopy(dom)
+	for node in dom_copy.xpath("/html/body//p/a"):
+		node.remove()
+
+	nodes = dom_copy.xpath("/html/body//p[re:test(., '[^=]\\s[0-9]{1,2}\\.[0-9]{2}(?![0-9′″°%]|\\.[0-9]|\\scubic|\\smetric|\\smeters|\\smiles|\\sfeet|\\sinches)')]")
+	matches = []
+	for node in nodes:
+		for time_match in regex.findall(r"(?<=[^=]\s)[0-9]{1,2}\.[0-9]{2}(?![0-9′″°%]|\.[0-9]|\scubic|\smetric|\smeters|\smiles|\sfeet|\sinches)", node.inner_text()):
+			time = time_match.split(".")
+			if not time[0].startswith("0") and int(time[0]) >= 1 and int(time[0]) <= 12 and int(time[1]) >= 0 and int(time[1]) <= 59:
+				matches.append(time_match)
+
+	if matches:
+		messages.append(LintMessage("t-010", "Time set with [text].[/] instead of [text]:[/].", se.MESSAGE_TYPE_WARNING, filename, set(matches)))
+
+	# Check for missing punctuation before closing quotes
+	# Exclude signatures in footers as those are commonly quoted without ending punctuation
+	nodes = dom.xpath("/html/body//p[not( (parent::header or parent::hgroup or (parent::footer and contains(@epub:type, 'z3998:signature'))) and position()=last())][re:test(., '[a-z]+[”’]$')]")
 	if nodes:
-		messages.append(LintMessage("x-010", "Illegal element in [xhtml][/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-011", "Missing punctuation before closing quotes.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string()[-30:] for node in nodes]))
 
-	# Check for underscores in attributes, but not if the attribute is href (links often have underscores) or MathMl alttext, which can have underscores as part of math notation
-	nodes = dom.xpath("//@*[contains(., '_') and name() !='href' and name() !='alttext']/..")
+	# Check for whitespace before noteref
+	# Do this early because we remove noterefs from headers later
+	# Allow noterefs that do not have preceding text/elements (for example a "see note X" ref that is the only child in a <td>)
+	nodes = dom.xpath("/html/body//a[contains(@epub:type, 'noteref') and preceding-sibling::node()[normalize-space(.)] and re:test(preceding-sibling::node()[1], '\\s+$')]")
 	if nodes:
-		messages.append(LintMessage("x-011", "Illegal underscore in attribute. Use dashes instead of underscores.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-012", "Illegal white space before noteref.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for style attributes
-	nodes = dom.xpath("/html/body//*[@style]")
+	# Check for period following Roman numeral, which is an old-timey style we must fix
+	# But ignore the numeral if it's the first item in a <p> tag, as that suggests it might be a kind of list item.
+	nodes = dom.xpath("/html/body//node()[name()='span' and contains(@epub:type, 'z3998:roman') and not(position()=1)][(following-sibling::node()[1])[re:test(., '^\\.\\s*[a-z]')]]")
 	if nodes:
-		messages.append(LintMessage("x-012", "Illegal [attr]style[/] attribute. Don’t use inline styles, any element can be targeted with a clever enough selector.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-013", "Roman numeral followed by a period. When in mid-sentence Roman numerals must not be followed by a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "." for node in nodes]))
 
-	# Check for illegal elements in <head>
-	nodes = dom.xpath("/html/head/*[not(self::title) and not(self::link[@rel='stylesheet'])]")
+	# Check for two em dashes in a row
+	nodes = dom.xpath(f"/html/body//*[re:test(text(), '—{se.WORD_JOINER}*—+')]")
 	if nodes:
-		messages.append(LintMessage("x-015", "Illegal element in [xhtml]<head>[/]. Only [xhtml]<title>[/] and [xhtml]<link rel=\"stylesheet\">[/] are allowed.", se.MESSAGE_TYPE_ERROR, filename, [f"<{node.tag}>" for node in nodes]))
+		messages.append(LintMessage("t-014", "Two or more em-dashes in a row found. Elided words should use the two- or three-em-dash Unicode character, and dialog ending in em-dashes should only end in a single em-dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for xml:lang attribute starting in uppercase
-	nodes = dom.xpath("//*[re:test(@xml:lang, '^[A-Z]')]")
+	# Check for money not separated by commas
+	nodes = dom.xpath("/html/body//*[re:test(text(), '[£\\$][0-9]{4,}')]")
 	if nodes:
-		messages.append(LintMessage("x-016", "[attr]xml:lang[/] attribute with value starting in uppercase letter.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+		messages.append(LintMessage("t-015", "Numbers not grouped by commas. Separate numbers greater than 1,000 with commas at every three numerals.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for ID attributes of numbered paragraphs (like `p-44`) that are used as refs in endnotes.
-	# Make sure their number is actually their correct sequence number in the text.
-	for unexpected_id in se.formatting.find_unexpected_ids(dom):
-		messages.append(LintMessage("x-019", f"Unexpected value of [attr]id[/] attribute. Expected: [attr]{unexpected_id[1]}[/].", se.MESSAGE_TYPE_ERROR, filename, [unexpected_id[0].to_tag_string()]))
+	# Check for <abbr epub:type="z3998:*-name"> that does not contain spaces
+	nodes = dom.xpath("/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and re:test(., '[A-Z]\\.[A-Z]\\.')]")
+	if nodes:
+		messages.append(LintMessage("t-016", "Initials in [xhtml]<abbr epub:type=\"z3998:*-name\">[/] not separated by spaces.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	return messages
+	if special_file != "colophon":
+		# This file is not the colophon.
+		# Check for ending punctuation inside italics that have semantics.
+		# Ignore the colophon because paintings might have punctuation in their names
 
-def _lint_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str) -> list:
-	"""
-	Helper function used in self.lint()
-	Process typo checks on an .xhtml file
+		# This xpath matches b or i elements with epub:type="se:name...", that are not stage direction, whose last text node ends in punctuation.
+		# Note that we check that the last node is a text node, because we may have <abbr> as the last node
+		matches = [node.to_string() for node in dom.xpath("(//b | //i)[contains(@epub:type, 'se:name') and not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., '[\\.,!\\?;:]$')]]")]
 
-	INPUTS
-	filename: The name of the file being checked
-	dom: A dom tree to check
-	file_contents: The contents of the file being checked
+		# Match b or i elements that are not stage directions, and that end in a comma followed by a lowercase letter
+		matches = matches + [node.to_string() for node in dom.xpath("(//b | //i)[not(contains(@epub:type, 'z3998:stage-direction'))][(text()[last()])[re:test(., ',$')] and following-sibling::node()[re:test(., '^\\s*[a-z]')] ]")]
 
-	OUTPUTS
-	A list of LintMessages representing possible typos found in the file
-	"""
+		# ...and also check for ending punctuation inside em tags, if it looks like a *part* of a clause
+		# instead of a whole clause. If the <em> is preceded by an em dash or quotes, or if there's punctuation
+		# and a space before it, then it's presumed to be a whole clause.
+		matches = matches + [match.strip() for match in regex.findall(r"(?<!.[—“‘>]|[!\.\?…;:]\s)<em>(?:\w+?\s*)+[\.,\!\?;]</em>", file_contents) if match.islower()]
 
-	messages = []
+		if matches:
+			messages.append(LintMessage("t-017", "Ending punctuation inside formatting like bold, small caps, or italics. Ending punctuation is only allowed within formatting if the phrase is an independent clause.", se.MESSAGE_TYPE_WARNING, filename, list(set(matches))))
 
-	# Check for common typos
-	if special_file == "titlepage":
-		# Don't check the titlepage because it has a standard format and may raise false positives
-		typos = regex.findall(r"(?<!’)\b(and and|the the|if if|of of|or or|as as)\b(?!-)", file_contents, flags=regex.IGNORECASE)
-		typos = typos + regex.findall(r"\ba a\b(?!-)", file_contents)
+	# Check for stage direction that ends in ?! but also has a trailing period
+	nodes = dom.xpath("/html/body//i[contains(@epub:type, 'z3998:stage-direction') and re:test(., '\\.$') and not((./node()[last()])[name() = 'abbr']) and (following-sibling::node()[1])[re:test(., '^[\\.,:;!?]')]]")
+	if nodes:
+		messages.append(LintMessage("t-018", "Stage direction ending in period next to other punctuation. Remove trailing periods in stage direction.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		if typos:
-			messages.append(LintMessage("y-001", "Possible typo: doubled [text]a/the/and/of/or/as/if[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for single words that are in italics, but that have closing punctuation outside italics
+	# Outer wrapping match is so that .findall returns the entire match and not the subgroup
+	# The first regex also matches the first few characters before the first double quote; we use those for more sophisticated
+	# checks below, to give fewer false positives like `with its downy red hairs and its “<i xml:lang="fr">doigts de faune</i>.”`
+	matches = regex.findall(r"((?:.{1,2}\s)?“<(i|em)[^>]*?>[^<]+?</\2>[\!\?\.])", file_contents) + regex.findall(r"([\.\!\?] <(i|em)[^>]*?>[^<]+?</\2>[\!\?\.])", file_contents)
 
-	# Check for basic typo, but exclude MathML and endnotes. Some endnotes (like in Ten Days that Shook the World) may start with letters.
-	dom_copy = deepcopy(dom)
-	for node in dom_copy.xpath("//a[contains(@epub:type, 'noteref')]") + dom_copy.xpath("//*[namespace-uri() = 'http://www.w3.org/1998/Math/MathML']"):
-		node.remove()
+	# But, if we've matched a name of something, don't include that as an error. For example, `He said, “<i epub:type="se:name.publication.book">The Decameron</i>.”`
+	# We also exclude the match from the list if:
+	# 1. The double quote is directly preceded by a lowercase letter and a space: `with its downy red hairs and its “<i xml:lang="fr">doigts de faune</i>.”`
+	# 2. The double quote is directly preceded by a lowercase letter, a comma, and a space, and the first letter within the double quote is lowercase: In the original, “<i xml:lang="es">que era un Conde de Irlos</i>.”
+	# 3. The text is a single letter that is not "I" or "a" (because then it is likely a mathematical variable)
+	matches = [match for match in matches if "epub:type=\"se:name." not in match[0] and "epub:type=\"z3998:taxonomy" not in match[0] and not regex.match(r"^[\p{Lowercase_Letter}’]+\s“", match[0]) and not regex.match(r"^[\p{Lowercase_Letter}’]+,\s“[\p{Lowercase_Letter}]", se.formatting.remove_tags(match[0])) and not regex.match(r"^.*?<.+?>[^Ia]<.+?>", match[0])]
+	if matches:
+		messages.append(LintMessage("t-019", "When a complete clause is italicized, ending punctuation except commas must be within containing italics.", se.MESSAGE_TYPE_WARNING, filename, [match[0] for match in matches]))
 
-	typos = [node.to_string() for node in dom_copy.xpath("/html/body//p[re:match(., '[a-z][;,][a-z]', 'i')]")]
-	if typos:
-		messages.append(LintMessage("y-002", "Possible typo: punctuation followed directly by a letter, without a space.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for punctuation after endnotes
+	nodes = dom.xpath(f"/html/body//a[contains(@epub:type, 'noteref')][(following-sibling::node()[1])[re:test(., '^[^\\s<–\\]\\)—{se.WORD_JOINER}a-zA-Z0-9]')]]")
+	if nodes:
+		messages.append(LintMessage("t-020", "Endnote links must be outside of punctuation, including quotation marks.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for paragraphs ending in lowercase letters, excluding a number of instances where that might be valid.
-	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '[a-z]$') and not(@epub:type) and not(following-sibling::*[1]/node()[1][contains(@epub:type, 'z3998:roman')] or following-sibling::*[1][re:test(., '^([0-9]|first|second|third|fourth|fifth|sixth|seventh|eight|ninth|tenth)', 'i')]) and not(@class or contains(@epub:type, 'z3998:salutation')) and not(following-sibling::*[1][name() = 'blockquote' or name() = 'figure' or name() = 'table' or name() = 'footer' or name() = 'ul' or name() = 'ol'] or ancestor::*[name() = 'blockquote' or name() = 'footer' or name() = 'header' or name() = 'table' or name() = 'ul' or name() = 'ol' or name() = 'li' or name() = 'figure' or name() = 'hgroup' or re:test(@epub:type, '(z3998:drama|z3998:letter|dedication|halftitlepage)')])]")]
-	if typos:
-		messages.append(LintMessage("y-003", "Possible typo: paragraph missing ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for correct typography around measurements like 2 ft.
+	# But first remove href and id attrs because URLs and IDs may contain strings that look like measurements
+	# Note that while we check m,min (minutes) and h,hr (hours) we don't check s (seconds) because we get too many false positives on years, like `the 1540s`
+	matches = regex.findall(fr"\b[1-9][0-9]*[{se.NO_BREAK_SPACE}\-]?(?:[mck]?[mgl]|ft|in|min?|h|sec|hr)\.?\b", regex.sub(r"(href|id)=\"[^\"]*?\"", "", file_contents))
+	# Exclude number ordinals, they're not measurements
+	matches = [match for match in matches if not regex.search(r"(st|nd|rd|th)", match)]
+	if matches:
+		messages.append(LintMessage("t-021", "Measurement not to standard. Numbers are followed by a no-break space and abbreviated units require an [xhtml]<abbr>[/] element. See [path][link=https://standardebooks.org/manual/1.0.0/8-typography#8.8.5]semos://1.0.0/8.8.5[/][/].", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '-[‘“]')]")]
-	if typos:
-		messages.append(LintMessage("y-004", "Possible typo: mis-curled quotation mark after dash.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for nbsp within <abbr epub:type="z3998:*-name">, which is redundant
+	nodes = dom.xpath(f"/html/body//abbr[re:test(@epub:type, '\\bz3998:[^\\s\"]+name\\b') and contains(text(), '{se.NO_BREAK_SPACE}')]")
+	if nodes:
+		messages.append(LintMessage("t-022", "No-break space found in [xhtml]<abbr epub:type=\"z3998:*-name\">[/]. This is redundant.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for commas or periods following exclamation or question marks. Exclude the colophon because we may have painting names with punctuation.
-	if special_file != "colophon":
-		typos = [node.to_string() for node in dom.xpath("//p[re:test(., '[!?][\\.,][^”’]')]")]
-		if typos:
-			messages.append(LintMessage("y-005", "Possible typo: question mark or exclamation mark followed by period or comma.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for trailing commas inside <i> tags at the close of dialog
+	# More sophisticated version of: \b[^\s]+?,</i>”
+	nodes = dom.xpath("/html/body//i[re:test(., ',$')][(following-sibling::node()[1])[starts-with(., '”')]]")
+	if nodes:
+		messages.append(LintMessage("t-023", "Comma inside [xhtml]<i>[/] element before closing dialog.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "”" for node in nodes]))
 
-	# Check for opening lsquo in quote that doesn't appear to have a matching rsquo. Rsquos must be followed by a space, punctuation (except period/comma), word joiner and thus em dash, or a number for an endnote.
-	typos = dom.xpath(f"re:match(//*, '“\\s*‘[^“]+?”', 'g')/text()[not(re:test(., '’[\\s\\?\\!;<2060{se.WORD_JOINER}0-9]'))]")
-	if typos:
-		messages.append(LintMessage("y-006", "Possible typo: [text]‘[/] without matching [text]’[/]. Hints: [text]’[/] are used for abbreviations;	commas and periods must go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for quotation marks in italicized dialog
+	nodes = dom.xpath("/html/body//i[@xml:lang][starts-with(., '“') or re:test(., '”$')]")
+	if nodes:
+		messages.append(LintMessage("t-024", "When italicizing language in dialog, italics go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Try to find top-level lsquo; for example, <p>“Bah!” he said to the ‘minister.’</p>
-	# We can't do this on xpath because we can't iterate over the output of re:replace().
-	# A fully-featured solution would parse quotes to find balanced pairs. This quick-n-dirty
-	# solution takes the raw XHTML string, strips any seemingly-valid balanced pairs using regex,
-	# then searches for unmatched `‘`. If found, it returns the dom node of the match; we can't return
-	# the actual string match because we stripped surrounding text.
+	# Check alt attributes on images, except for the logo
+	nodes = dom.xpath("/html/body//img[not(re:test(@src, '/(logo|titlepage)\\.(svg|png)$'))]")
+	img_no_alt = []
+	img_alt_not_typogrified = []
+	img_alt_lacking_punctuation = []
+	for node in nodes:
+		if "titlepage.svg" not in node.get_attr("src"):
+			ebook_flags["has_images"] = True # Save for a later check
 
-	# Start by merging all <p> into single lines, including poetry, but exclude blockquotes as they have special quoting rules
-	temp_xhtml = ""
-	dom_copy = deepcopy(dom)
-	node_number = 0
-	# Add IDs to the dom so we can find the paragraphs after we run our regex
-	# Exclude paragraphs in blockquotes, which may have special quoting rules, and "continued" paragraphs, which may be continued dialog without an “
-	for node in dom_copy.xpath("/html/body//p[not(ancestor::blockquote) and not(contains(@class, 'continued'))]"):
-		node.set_attr("id", "lint-" + str(node_number))
-		temp_xhtml = temp_xhtml + f"<p id=\"lint-{node_number}\">" + regex.sub(r"[\s\n]+", " ", node.inner_text(), flags=regex.DOTALL) + "\n"
-		node_number = node_number + 1
+		alt = node.get_attr("alt")
 
-	replacement_count = 1
-	while replacement_count > 0:
-		# Remove all regular quotes. Run this in a loop because we may need to remove triple-nested quotes.
-		(temp_xhtml, replacement_count) = regex.subn(r"“[^“]+?”", " ", temp_xhtml) # Remove all regular quotes
+		if alt:
+			# Check for non-typogrified img alt attributes
+			if regex.search(r"""('|"|--|\s-\s|")""", alt):
+				img_alt_not_typogrified.append(node.to_tag_string())
 
-	# Remove contractions to reduce rsquo for next regex
-	temp_xhtml = regex.sub(r"[\p{Letter}]’[\p{Letter}]", " ", temp_xhtml, flags=regex.MULTILINE)
+			# Check alt attributes not ending in punctuation
+			if filename.name not in IGNORED_FILENAMES and not regex.search(r"""[\.\!\?]”?$""", alt):
+				img_alt_lacking_punctuation.append(node.to_tag_string())
 
-	# Remove all runs of ldquo that are likely to spill to the next <p>
-	replacement_count = 1
-	while replacement_count > 0:
-		(temp_xhtml, replacement_count) = regex.subn(r"“[^“”]+?$", " ", temp_xhtml, flags=regex.MULTILINE)
+			# Check that alt attributes match SVG titles
+			img_src = node.lxml_element.get("src")
+			if img_src and img_src.endswith("svg"):
+				title_text = ""
+				image_ref = img_src.split("/").pop()
+				try:
+					svg_path = self.content_path / "images" / image_ref
+					svg_dom = self.get_dom(svg_path)
+					try:
+						title_text = svg_dom.xpath("/svg/title")[0].text
+					except Exception:
+						messages.append(LintMessage("s-027", f"{image_ref} missing [xhtml]<title>[/] element.", se.MESSAGE_TYPE_ERROR, svg_path))
 
-	# Match problem `‘` using regex, and if found, get the actual node text from the dom to return.
-	typos = []
-	for match in regex.findall(r"""<p id="lint-([0-9]+?)">.*‘[\p{Letter}\s]""", temp_xhtml):
-		for node in dom_copy.xpath(f"/html/body//p[@id='lint-{match}' and re:test(., '‘[A-Za-z\\s]')]"):
-			typos.append(node.inner_text())
+					if title_text != "" and alt != "" and title_text != alt:
+						messages.append(LintMessage("s-022", f"The [xhtml]<title>[/] element of [path][link=file://{svg_path}]{image_ref}[/][/] does not match the [attr]alt[/] attribute text in [path][link=file://{filename}]{filename.name}[/][/].", se.MESSAGE_TYPE_ERROR, filename))
 
-	if typos:
-		messages.append(LintMessage("y-007", "Possible typo: [text]‘[/] not within [text]“[/]. Hints: Should [text]‘[/] be replaced with [text]“[/]? Is there a missing closing quote? Is this a nested quote that should be preceded by [text]“[/]? Are quotes in close proximity correctly closed?", se.MESSAGE_TYPE_WARNING, filename, typos))
+				except FileNotFoundError:
+					missing_files.append(str(Path("images") / image_ref))
 
-	# Check for single quotes when there should be double quotes in an interjection in dialog
-	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '“[^”]+?’⁠—[^”]+?—“')]")]
-	if typos:
-		messages.append(LintMessage("y-008", "Possible typo: dialog interrupted by interjection but with incorrect closing quote.", se.MESSAGE_TYPE_WARNING, filename, typos))
+		else:
+			img_no_alt.append(node.to_tag_string())
 
-	# Check for dialog starting with a lowercase letter. Only check the first child text node of <p>, because other first children might be valid lowercase, like <m:math> or <b>;
-	# exclude <p> inside or preceded by <blockquote>; and exclude <p> inside endnotes, as definitions may start with lowercase letters.
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::li[contains(@epub:type, 'endnote')]) and not(preceding-sibling::*[1][name()='blockquote'])][re:test(./node()[1], '^“[a-z]')]")]
-	if typos:
-		messages.append(LintMessage("y-009", "Possible typo: dialog begins with lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	if img_alt_not_typogrified:
+		messages.append(LintMessage("t-025", "Non-typogrified [text]'[/], [text]\"[/] (as [xhtml]"[/]), or [text]--[/] in image [attr]alt[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, img_alt_not_typogrified))
+	if img_alt_lacking_punctuation:
+		messages.append(LintMessage("t-026", "[attr]alt[/] attribute does not appear to end with punctuation. [attr]alt[/] attributes must be composed of complete sentences ending in appropriate punctuation.", se.MESSAGE_TYPE_ERROR, filename, img_alt_lacking_punctuation))
+	if img_no_alt:
+		messages.append(LintMessage("s-004", "[xhtml]img[/] element missing [attr]alt[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, img_no_alt))
 
-	typos = [node.to_string() for node in dom.xpath("//p[not(ancestor::blockquote) and not(following-sibling::*[1][name() = 'blockquote' or contains(@class, 'continued')]) and re:test(., ',”$')]")]
-	if typos:
-		messages.append(LintMessage("y-010", "Possible typo: comma ending dialogue.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for low-hanging misquoted fruit
+	matches = regex.findall(r"[\p{Letter}]+[“‘]", file_contents) + regex.findall(r"[^>]+</(?:em|i|b|span)>‘[\p{Lowercase_Letter}]+", file_contents)
+	if matches:
+		messages.append(LintMessage("t-028", "Possible mis-curled quotation mark.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '’{2,}')]")]
-	if typos:
-		messages.append(LintMessage("y-011", "Possible typo: two or more [text]’[/] in a row.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for periods followed by lowercase.
+	temp_xhtml = regex.sub(r"<title[^>]*?>.+?", "", file_contents) # Remove  because it might contain something like <title>Chapter 2: The Antechamber of M. de Tréville
+	temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Replace things like 
+	temp_xhtml = regex.sub(r"]*?>", "", temp_xhtml) # Remove  attributes
+	temp_xhtml = temp_xhtml.replace("A.B.C.", "X") # Remove A.B.C, which is not an abbreviations.
+	# Note the regex also excludes preceding numbers, so that we can have inline numbering like:
+	# "A number of questions: 1. regarding those who make heretics; 2. concerning those who were made heretics..."
+	matches = regex.findall(r"[^\s0-9]+\.\s+[\p{Lowercase_Letter}](?!’[\p{Uppercase_Letter}])[\p{Lowercase_Letter}]+", temp_xhtml)
+	# If  is in the match, remove it from the matches so we exclude things like et. al.
+	matches = [match for match in matches if "" not in match]
+	if matches:
+		messages.append(LintMessage("t-029", "Period followed by lowercase letter. Hint: Abbreviations require an [xhtml][/] element.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '”[a-z]')]")]
-	if typos:
-		messages.append(LintMessage("y-012", "Possible typo: [text]”[/] directly followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for initialisms without periods
+	nodes = [node.to_string() for node in dom.xpath("/html/body//abbr[contains(@epub:type, 'z3998:initialism') and not(re:test(., '^[0-9]*([a-zA-Z]\\.)+[0-9]*$'))]") if node.text not in INITIALISM_EXCEPTIONS]
+	if nodes:
+		messages.append(LintMessage("t-030", "Initialism with spaces or without periods.", se.MESSAGE_TYPE_WARNING, filename, set(nodes)))
 
-	# Check for period not inside rsquo
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[^”’]+?’[\\.,]”')]")]
-	if typos:
-		messages.append(LintMessage("y-013", "Possible typo: punctuation not within [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+	nodes = dom.xpath("/html/body//*[re:test(text(), '\\bA\\s*B\\s*C\\s*\\b')]")
+	if nodes:
+		messages.append(LintMessage("t-031", "[text]A B C[/] must be set as [text]A.B.C.[/] It is not an abbreviation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for period before dialog tag; try to exclude abbrevations that close a quotation, like `“Mr.”`.
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[(re:test(., '\\.”\\s[a-z\\s]*?(\\bsaid|[a-z]+ed)') or re:test(., '\\.”\\s(s?he|they?|and)\\b')) and not(.//abbr[following-sibling::node()[re:test(., '^”')]])]")]
-	if typos:
-		messages.append(LintMessage("y-014", "Possible typo: Unexpected [text].[/] at the end of quotation. Hint: If a dialog tag follows, should this be [text],[/]?", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for abbreviations followed by periods
+	# But we exclude some SI units, which don't take periods; abbreviations ending in numbers for example in stage directions; abbreviations like `r^o` (recto) that contain ; and some Imperial abbreviations that are multi-word
+	nodes = dom.xpath("/html/body//abbr[(contains(@epub:type, 'z3998:initialism') or re:test(@epub:type, 'z3998:[^\\s]+?name') or not(@epub:type))][not(re:test(., '[cmk][mgl]')) and not(re:test(., '[0-9]$')) and not(./sup) and not(re:test(@class, '\\b(era|temperature|compound)\\b')) and not(re:test(text(), '^(mpg|mph|hp|TV)$'))][following-sibling::text()[1][starts-with(self::text(), '.')]]")
+	if nodes:
+		messages.append(LintMessage("t-032", "Initialism or name followed by period. Hint: Periods go within [xhtml][/]. [xhtml][/]s containing periods that end a clause require the [class]eoc[/] class.", se.MESSAGE_TYPE_WARNING, filename, [f"{node.to_string()}." for node in nodes]))
 
-	# Check for mis-curled ‘ or ‘ without matching ’
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[A-Za-z][^“’]+?”')]")]
-	if typos:
-		messages.append(LintMessage("y-015", "Possible typo: mis-curled [text]‘[/] or missing [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for space after dash
+	nodes = dom.xpath("/html/body//*[name()='p' or name()='span' or name='em' or name='i' or name='b' or name='strong'][not(self::comment())][re:test(., '[a-zA-Z]-\\s(?!(and|or|nor|to|und|…)\\b)')]")
+	if nodes:
+		messages.append(LintMessage("t-033", "Space after dash.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for two periods in a row, almost always a typo for one period or a hellip
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\.\\.[^\\.]')]")]
-	if typos:
-		messages.append(LintMessage("y-016", "Possible typo: consecutive periods ([text]..[/]).", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for  preceded by em dash
+	nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[re:test(., '—$')]]")
+	if nodes:
+		messages.append(LintMessage("t-034", "[xhtml][/] element preceded by em-dash. Hint: em-dashes go within [xhtml][/] elements.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for ldquo followed by space
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '“\\s+[^‘’]')]")]
-	if typos:
-		messages.append(LintMessage("y-017", "Possible typo: [text]“[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for  without preceding space in text node. (preceding ( or [ are also OK)
+	nodes = dom.xpath("/html/body//cite[(preceding-sibling::node()[1])[not(re:test(., '[\\[\\(\\s]$'))]]")
+	if nodes:
+		messages.append(LintMessage("t-035", "[xhtml][/] element not preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for lsquo followed by space
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘\\s+[^“’]')]")]
-	if typos:
-		messages.append(LintMessage("y-018", "Possible typo: [text]‘[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for partially obscured years in which the last year is an em dash
+	nodes = dom.xpath("/html/body//p[re:test(.,	 '1\\d{2}⁠—[^“’A-Za-z<]')]")
+	if nodes:
+		messages.append(LintMessage("t-036", "Em-dash used to obscure single digit in year. Hint: Use a hyphen instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for closing rdquo without opening ldquo. We ignore blockquotes because they usually have unique quote formatting.
-	# Remove tds in case rdquo means "ditto mark"
-	typos = regex.findall(r"”[^“‘]+?”", regex.sub(r"", "", file_contents), flags=regex.DOTALL)
+	# Check for rdquo preceded by space (but not preceded by a rsquo, which might indicate a nested quotation)
+	nodes = dom.xpath("/html/body//p[re:test(., '[^’]\\s”')]")
+	if nodes:
+		messages.append(LintMessage("t-037", "[text]”[/] preceded by space.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# We create a filter to try to exclude nested quotations
-	# Remove tags in case they're enclosing punctuation we want to match against at the end of a sentence.
-	typos = [match for match in typos if not regex.search(r"(?:[\.!\?;…—]|”\s)’\s", se.formatting.remove_tags(match))]
+	# Check obviously miscurled quotation marks
+	nodes = dom.xpath("/html/body//p[re:test(., '“$')]")
+	if nodes:
+		messages.append(LintMessage("t-038", "[text]“[/] before closing [xhtml]
    [/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Try some additional matches before adding the lint message
-	# Search for 
     tags that have an ending closing quote but no opening quote; but exclude 
    s that are preceded by a 
    
-	# or that have a 
     ancestor, because that may indicate that the opening quote is elsewhere in the quotation.
-	for node in dom.xpath("//p[re:test(., '^[^“]+”') and not(./preceding-sibling::*[1][name() = 'blockquote']) and not(./ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]) and not(./ancestor::blockquote)]"):
-		typos.append(node.to_string()[-20:])
+	# Check for some known initialisms with incorrect possessive apostrophes
+	nodes = dom.xpath("/html/body//abbr[text()='I.O.U.'][(following-sibling::node()[1])[starts-with(., '’s')]]")
+	if nodes:
+		messages.append(LintMessage("t-039", "Initialism followed by [text]’s[/]. Hint: Plurals of initialisms are not followed by [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() + "’s" for node in nodes]))
 
-	if typos:
-		messages.append(LintMessage("y-019", "Possible typo: [text]”[/] without opening [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-	# Check for ,.
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., ',\\.')]")]
-	if typos:
-		messages.append(LintMessage("y-020", "Possible typo: consecutive comma-period ([text],.[/]).", se.MESSAGE_TYPE_WARNING, filename, typos))
-
-	# Check for single quotes instead of double quotes
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[not(contains(@class, 'continued') or ancestor::blockquote or ancestor::*[re:test(@epub:type, 'z3998:(verse|song|poem|hymn)')]) and re:test(., '^[^“]+?‘')]")]
-	if typos:
-		messages.append(LintMessage("y-021", "Possible typo: Opening [text]‘[/] without preceding [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check if a subtitle ends in a text node with a terminal period; or if it ends in an  node containing a terminal period.
+	nodes = dom.xpath("/html/body//*[self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6]/*[contains(@epub:type, 'subtitle')][(./text())[last()][re:test(., '\\.$')] or (./i)[last()][re:test(., '\\.$')]]")
+	if nodes:
+		messages.append(LintMessage("t-040", "Subtitle with illegal ending period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for two quotations in one paragraph
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^”]+?”\\s“[^”]+?”$')]")]
-	if typos:
-		messages.append(LintMessage("y-022", "Possible typo: consecutive quotations without intervening text, e.g. [text]“…” “…”[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+	matches = regex.findall(r"[^…]\s[!?;:,].{0,10}", file_contents) # If we don't include preceding chars, the regex is 6x faster
+	if matches:
+		messages.append(LintMessage("t-041", "Illegal space before punctuation.", se.MESSAGE_TYPE_ERROR, filename, matches))
 
-	# Check for incorrectly nested quotation marks
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘”]+“') and not(.//br)]")]
-	if typos:
-		messages.append(LintMessage("y-023", "Possible typo: two opening quotation marks in a run. Hint: Nested quotes should switch between [text]“[/] and [text]‘[/]", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for a list of some common English loan words that used to be italicized as foreign, but are no longer
+	nodes = dom.xpath("/html/body//*[@data-css-font-style='italic' and re:test(., '^(fianc[eé]+|divorc[eé]+|menu|recherch[eé]|tour-de-force|outr[eé]|d[ée]but(ante)?|apropos|[eé]lite|prot[ée]g[ée]+|chef|salon|r[eé]gime|contretemps|[eé]clat|aides?|entr[ée]+|t[eê]te-[aà]-t[eê]tes?|blas[eé]|bourgeoisie)$')]")
+	if nodes:
+		messages.append(LintMessage("t-043", "Non-English loan word set in italics, when modern typography omits italics. Hint: A word may be correctly italicized when emphasis is desired, if the word is meant to be pronounced with an accent, or if the word is part of non-English speech.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for dashes instead of em-dashes
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\s[a-z]+-(the|there|is|and|they|when)\\s')]")]
-	if typos:
-		messages.append(LintMessage("y-024", "Possible typo: dash before [text]the/there/is/and/they/when[/] probably should be em-dash.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for comma after leading Or in subtitles
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'subtitle') and re:test(text(), '^Or\\s')]")
+	if nodes:
+		messages.append(LintMessage("t-044", "Comma required after leading [text]Or[/] in subtitles.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for comma not followed by space but followed by quotation mark
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '[a-z],[“”‘’][a-z]', 'i')]")]
-	if typos:
-		messages.append(LintMessage("y-025", "Possible typo: comma without space followed by quotation mark.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for personas that are italicized. Use re:test instead of contains to avoid catching z3998:personal-name
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:persona\\b') and @data-css-font-style = 'italic']")
+	if nodes:
+		messages.append(LintMessage("t-045", "Element has [val]z3998:persona[/] semantic and is also set in italics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for punctuation missing before conjunctions. Ignore 
     with an  child starting in a conjunction, as those are probably book titles or non-English languages
-	typos = [node.to_string() for node in dom.xpath(f"/html/body//p[not(parent::hgroup) and re:test(., '\\b[a-z]+\\s(But|And|For|Nor|Yet|Or)\\b[^’\\.\\?\\-{se.WORD_JOINER}]') and not(./i[re:test(., '^(But|And|For|Nor|Yet|Or)\\b')])]")]
-	if typos:
-		messages.append(LintMessage("y-026", "Possible typo: no punctuation before conjunction [text]But/And/For/Nor/Yet/Or[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+	if "῾" in file_contents:
+		messages.append(LintMessage("t-046", "[text]῾[/] (U+1FFE) detected. Use [text]ʽ[/] (U+02BD) instead.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for extra closing single quote at the end of dialog
-	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘]+”\\s*’$')]")]
-	if typos:
-		messages.append(LintMessage("y-027", "Possible typo: Extra [text]’[/] at end of paragraph.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for all-caps first paragraphs in sections/articles
+	# Note that we don't check for small-caps CSS, because there are lots of legitimate cases for that,
+	# and it would generate too many false positives.
+	# If we did want to do that, this xpath after the last re:test would help: or ./*[not(preceding-sibling::node()[normalize-space(.)]) and @data-css-font-variant='small-caps' and following-sibling::node()[1][self::text()]]
+	nodes = dom.xpath("/html/body//*[(name() = 'section' or name() = 'article') and not(contains(@epub:type, 'dedication') or contains(@epub:type, 'z3998:letter'))]/p[1][re:test(normalize-space(.), '^[“’]?(?:[A-Z’]+\\s){2,}')]")
+	if nodes:
+		messages.append(LintMessage("t-048", "Chapter opening text in all-caps.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for `` preceded or followed by text. Ignore compass directions followed by `ly`, like S.S.W.ly
-	typos = [node.to_string() for node in dom.xpath("/html/body//abbr[(preceding-sibling::node()[1])[re:test(., '[A-Za-z]$')] or (following-sibling::node()[1])[re:test(., '^[A-Za-z](?[/] directly preceded or followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for two-em-dashes used for elision instead of three-em-dashes
+	matches = regex.findall(fr"[^{se.WORD_JOINER}\p{{Letter}}”]⸺[^“{se.WORD_JOINER}\p{{Letter}}].*", file_contents, flags=regex.MULTILINE)
+	if matches:
+		messages.append(LintMessage("t-049", "Two-em-dash used for eliding an entire word. Use a three-em-dash instead.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	# Check for misapplied italics. Ignore 's' because the plural is too common.
-	typos = [node.to_string() for node in dom.xpath("/html/body//*[(name() = 'i' or name() = 'em') and ./following-sibling::node()[1][re:test(., '^[a-z]\\b', 'i') and not(re:test(., '^s\\b'))]]")]
-	if typos:
-		messages.append(LintMessage("y-029", "Possible typo: Italics followed by a letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check that possessives appear within persona blocks
+	nodes = dom.xpath("/html/body//b[contains(@epub:type, 'z3998:persona') and ./following-sibling::node()[1][re:test(., '^’s?')]]")
+	if nodes:
+		messages.append(LintMessage("t-050", "Possessive [text]’s[/] or [text]’[/] outside of element with [val]z3998:persona[/] semantic.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for lowercase letters starting quotations after a preceding period
-	typos = dom.xpath("/html/body//p/child::text()[re:test(., '\\.\\s[‘“][a-z]')]")
-	if typos:
-		messages.append(LintMessage("y-030", "Possible typo: Lowercase quotation following a period. Check either that the period should be a comma, or that the quotation should start with a capital.", se.MESSAGE_TYPE_WARNING, filename, typos))
+	# Check for quotations that carry to the next paragraph, but the next paragraph has no opening quotation mark
+	# Exclude p in blockquote because often that has different formatting
+	nodes = dom.xpath("/html/body//p[re:test(., '^“[^”]+$') and not(./ancestor::blockquote) and ./following-sibling::*[1][name() = 'p' and re:test(normalize-space(), '^[^“]')]]")
+	if nodes:
+		messages.append(LintMessage("t-051", "Dialog in [xhtml]
    [/] that continues to the next [xhtml]
    [/], but the next [xhtml]
    [/] does not begin with [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for missing punctuation in continued quotations
-	# ” said Bob “
-	nodes = dom.xpath("/html/body//p[re:test(., '”\\s(?:said|[A-Za-z]{2,}ed)\\s[A-Za-z]+?(?Newspapers
-	nodes = dom.xpath("/html/body//i[@epub:type and ( (following-sibling::node()[1][re:test(., '^[a-z]', 'i') and not(re:test(., '^(s|es|er)'))]) or preceding-sibling::node()[1][re:test(., '[a-z]$')]) ]")
+	# Check for stage direction starting in lowercase.	We only want to consider stage direction that is not an interjection in a parent clause.
+	# We match if the stage direction starts with a lowercase, and if there's no preceding node, or if there is a preceding node and it doesn't end in a lowercase letter or a small subset of conjoining punctuation.
+	nodes = dom.xpath("/html/body//i[@epub:type='z3998:stage-direction' and re:test(., '^[a-z]') and (not(./preceding-sibling::node()[normalize-space(.)]) or ./preceding-sibling::node()[1][re:test(normalize-space(.), '[^a-z:—,;…]$')])]")
 	if nodes:
-		messages.append(LintMessage("y-032", "Possible typo: Italics running into preceding or following characters.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("t-053", "Stage direction starting in lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	return messages
+	# Epigraphs that are entirely non-English should still be in italics, not Roman
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph') or contains(@epub:type, 'bridgehead')]//p[./i[@xml:lang and @data-css-font-style='normal' and ( (./preceding-sibling::node()='“' and ./following-sibling::node()='”') or not(./preceding-sibling::node()[normalize-space(.)] or ./following-sibling::node()[normalize-space(.)]) ) ]]")
+	if nodes:
+		messages.append(LintMessage("t-054", "Epigraphs that are entirely non-English should be set in italics, not Roman.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
+	# Check for lone acute accents, which should always be combined or swapped for a more accurate Unicode char
+	if "´" in file_contents:
+		messages.append(LintMessage("t-055", "Lone acute accent ([val]´[/]). A more accurate Unicode character like prime for coordinates or measurements, or combining accent or breathing mark for Greek text, is required.", se.MESSAGE_TYPE_ERROR, filename))
 
+	# Check for degree confusable
+	nodes = dom.xpath("/html/body//*[re:test(text(), '[0-9]+º')]")
+	if nodes:
+		messages.append(LintMessage("t-056", "Masculine ordinal indicator ([val]º[/]) used instead of degree symbol ([val]°[/]). Note that the masculine ordinal indicator may be appropriate for ordinal numbers read as Latin, i.e. [val]12º[/] reading [val]duodecimo[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = None) -> list:
-	"""
-	Check this ebook for some common SE style errors.
+	# Check for 
     starting in lowercase
+	# Exclude 
     starting in constructs like `(a)` or `a.` as they may be numbered lists
+	nodes = dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::figure or ancestor::ol[not(ancestor::section[contains(@epub:type, 'endnotes')])] or ancestor::ul or ancestor::table or ancestor::hgroup or preceding-sibling::*[1][name() = 'hr']) and not(contains(@class, 'continued') or re:test(@epub:type, 'z3998:(signature|valediction)')) and not(./*[1][name() = 'math' or name() = 'var' or contains(@epub:type, 'z3998:grapheme')]) and re:test(., '^[^A-Za-z0-9]?[a-z]') and not(re:test(., '^\\([a-z0-9]\\.?\\)\\.?')) and not(re:test(., '^[a-z0-9]\\.\\s'))]")
 
-	INPUTS
-	None
+	# We have to additionally filter using the regex library, because often a sentence may begin with an uppercase ACCENTED letter,
+	# and xpath's limited regex library doesn't support Unicode classes
+	nodes = [node for node in nodes if regex.match(r"^[^\p{Letter}0-9]?[\p{Lowercase_Letter}]", node.inner_text())]
 
-	OUTPUTS
-	A list of LintMessage objects.
-	"""
+	if nodes:
+		messages.append(LintMessage("t-057", "[xhtml]
    [/] starting with lowercase letter. Hint: [xhtml]
    [/] that continues text after a [xhtml]
    [/] requires the [class]continued[/] class; and use [xhtml]
    [/] to split one clause over many lines.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	local_css_path = self.content_path / "css/local.css"
-	messages: List[LintMessage] = []
-	is_titlepage = False
-	has_halftitle = False
-	has_frontmatter = False
-	has_cover_source = False
-	cover_svg_title = ""
-	titlepage_svg_title = ""
-	xhtml_css_classes: Dict[str, int] = {}
-	headings: List[tuple] = []
-	double_spaced_files: List[Path] = []
-	unused_selectors: List[str] = []
-	id_attrs: List[str] = []
-	missing_metadata_elements = []
-	abbr_elements_requiring_css: List[se.easy_xml.EasyXmlElement] = []
-	metadata_xml = self.metadata_dom.to_string()
-	has_glossary_search_key_map = False
-	glossary_usage = []
-	has_images = False
+	# Check for illegal characters
+	matches = regex.findall(fr"({se.UNICODE_BOM}|{se.SHY_HYPHEN})", file_contents)
+	if matches:
+		# Get the keys of a dict in order to create a list without duplicates
+		messages.append(LintMessage("t-058", "Illegal character.", se.MESSAGE_TYPE_ERROR, filename, list({match.encode("unicode_escape").decode().replace("\\u", "U+").upper():None for match in matches}.keys())))
 
-	# This is a dict with where keys are the path and values are a list of code dicts.
-	# Each code dict has a key "code" which is the actual code, and a key "used" which is a
-	# bool indicating whether or not the code has actually been caught in the linting run.
-	ignored_codes: Dict[str, List[Dict]] = {}
+	# Check for period in cite in endnotes
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]//cite[not((./node()[last()])[name() = 'abbr']) and ./following-sibling::*[1][contains(@epub:type, 'backlink')] and re:test(., '^—') and ( (re:test(., '\\.$') and ./following-sibling::node()[re:test(., '^\\s*$')]) or ./following-sibling::node()[re:test(., '^\\.\\s*$')])]")
+	if nodes:
+		messages.append(LintMessage("t-059", "Period at the end of [xhtml][/] element before endnote backlink.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Cache the browser default stylesheet for later use
-	with importlib_resources.open_text("se.data", "browser.css", encoding="utf-8") as css:
-		self._file_cache["default"] = css.read() # pylint: disable=protected-access
+	# Check for Bible verses in old-style notation
+	nodes = dom.xpath("/html/body//*[(name()='p' or name()='cite' or re:test(name(), '^h[1-6]$')) and .//node()[re:test(., '(Genesis|Gen\\.|Exodus|Ex\\.|Leviticus|Lev\\.|Numbers|Num\\.|Deuteronomy|Deut\\.|Joshua|Josh\\.|Judges|Ruth|Kings|Chronicles|Chron\\.|Ezra|Nehemiah|Neh\\.|Esther|Esth\\.|Job|Psalm|Psalms|Ps\\.|Proverbs|Prov\\.|Ecclesiastes|Ecc\\.|Eccl\\.|Solomon|Sol\\.|Isaiah|Is\\.|Isa\\.|Jeremiah|Jer\\.|Lamentations|Lam\\.|Ezekiel|Ez\\.|Ezek\\.|Daniel|Dan\\.|Hosea|Hos\\.|Joel|Amos|Obadiah|Obad\\.|Jonah|Jon\\.|Micah|Mic\\.|Nahum|Nah\\.|Habakkuk|Hab\\.|Zephaniah|Zeph\\.|Haggai|Hag\\.|Zechariah|Zech\\.|Malachi|Mal\\.|Tobit|Judith|Sirach|Baruch|Maccabees|Esdras|Manasses|Matthew|Matt\\.|Mark|Luke|John|Acts|Romans|Rom\\.|Corinthians|Cor\\.|Corinth\\.|Galatians|Gal\\.|Ephesians|Eph\\.|Philippians|Phil\\.|Philipp\\.|Colossians|Col\\.|Coloss\\.|Thessalonians|Thes\\.|Thess\\.|Timothy|Tim\\.|Titus|Tit\\.|Philemon|Phil\\.|Hebrews|Heb\\.|James|Jas\\.|Peter|Pet\\.|Jude|Revelation|Revelations|Rev\\.)\\s*$') and following-sibling::node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman') and following-sibling::node()[1][re:test(., '^\\s*[\\.,]?\\s+[0-9]')]]]]")
+	if nodes:
+		messages.append(LintMessage("t-060", "Old style Bible citation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# First, check if we have an se-lint-ignore.xml file in the ebook root. If so, parse it. For an example se-lint-ignore file, see semos://1.0.0/2.3
-	lint_ignore_path = self.path / "se-lint-ignore.xml"
-	if not skip_lint_ignore and lint_ignore_path.exists():
-		lint_config = self.get_dom(lint_ignore_path)
+	# Check for summary-style bridgeheads not ending in punctuation
+	# Decide if it's a summary-style bridgehead if it contains two or more em dashes. We use Python regex because
+	# xpath can't count the number of occurances of a string.
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'bridgehead') and re:test(., '[^\\.\\!\\?”]”?$')]")
+	nodes = [node for node in nodes if len(regex.findall(r"—", node.to_string())) >= 2]
+	if nodes:
+		messages.append(LintMessage("t-061", "Summary-style bridgehead without ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		elements = lint_config.xpath("/se-lint-ignore/file")
+	# Check for uppercased am/pm, except for if they have an epub:type, or when uppercased in titles, or in the nav document
+	if not dom.xpath("//nav[contains(@epub:type, 'toc')]"):
+		nodes = dom.xpath("/html/body//abbr[re:test(., '^[AP]\\.M\\.$') and not(@epub:type) and not(./ancestor::*[contains(@epub:type, 'title')])]")
+		if nodes:
+			messages.append(LintMessage("t-062", "Uppercased [text]a.m.[/] and [text]p.m.[/]", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		if not elements:
-			messages.append(LintMessage("m-049", "No [path]se-lint-ignore.xml[/] rules. Delete the file if there are no rules.", se.MESSAGE_TYPE_ERROR, lint_ignore_path))
+	# Check for some Latinisms that need italics according to the manual. We check ancestor-or-self in case the phrase is set in Roman because it's nested in a parent italic.
+	# Exclude toto followed by ’ since Toto can be a name.
+	# Exclude  whose entire contents is a matched Latinism as we do not italicize those.
+	# Ignore the ToC because we have different rules there
+	nodes = dom.xpath("/html/body//text()[re:test(., '\\b(a (priori|posteriori|fortiori)|(?s in endnotes that are directly followed by the backlink;
+	# ignore s starting with `Cf.` which means `Compare` and was often used to mean `see`
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[last()]/cite[following-sibling::node()[normalize-space(.)][1][name() = 'a' and contains(@epub:type, 'backlink')] and re:test(., '^[^—]') and not(re:test(., '^(Cf\\.|\\()'))]")
+	if nodes:
+		messages.append(LintMessage("t-068", "Citation not offset with em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-						if child.tag == "reason" and child.text.strip() != "":
-							has_reason = True
+	# Check for  in 
    s that start with a leading em dash
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '^—')]")
+	if nodes:
+		messages.append(LintMessage("t-069", "[xhtml][/] in epigraph starting with an em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-					if not has_reason:
-						messages.append(LintMessage("m-046", "Missing or empty [xml][/] element.", se.MESSAGE_TYPE_ERROR, lint_ignore_path))
+	# Check for  in 
    s that end with a period. Try to exclude s that are last children
+	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '\\.$') and not( (./node()[last()])[./descendant-or-self::abbr]) ]")
+	if nodes:
+		messages.append(LintMessage("t-070", "[xhtml][/] in epigraph ending in a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-		if has_illegal_path:
-			messages.append(LintMessage("m-047", "Ignoring [path]*[/] is too general. Target specific files if possible.", se.MESSAGE_TYPE_WARNING, lint_ignore_path))
+	# If we have Greek text, try to normalize it
+	for node in dom.xpath("/html/body//*[@xml:lang='grc' or @xml:lang='el']"):
+		node_text = node.inner_text()
+		expected_text = se.typography.normalize_greek(node_text)
+		if node_text != expected_text:
+			messages.append(LintMessage("t-073", f"Possible transcription error in Greek. Found: [text]{node_text}[/], but expected [text]{expected_text}[/text]. Hint: Use [bash]se unicode-names[/] to see differences in Unicode characters.", se.MESSAGE_TYPE_WARNING, filename))
 
-	# Done parsing ignore list
+	return (messages)
 
-	# Check that the spine has all the expected files in the book
-	missing_spine_files = files_not_in_spine(self)
-	if missing_spine_files:
-		for missing_spine_file in missing_spine_files:
-			messages.append(LintMessage("f-007", "File not listed in [xml][/].", se.MESSAGE_TYPE_ERROR, missing_spine_file))
+def _lint_xhtml_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str) -> list:
+	"""
+	Helper function used in self.lint()
+	Process XHTML checks on an .xhtml file
 
-	# Get the ebook language for later use
-	try:
-		language = self.metadata_dom.xpath("/package/metadata/dc:language")[0].text
-	except se.InvalidXmlException as ex:
-		raise ex
-	except Exception as ex:
-		raise se.InvalidSeEbookException(f"Missing [xml][/] element in [path][link=file://{self.metadata_file_path}]{self.metadata_file_path.name}[/][/].") from ex
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
+	file_contents: The contents of the file being checked
 
-	# Check local.css for various items, for later use
-	try:
-		self.local_css = self.get_file(local_css_path)
-	except Exception as ex:
-		raise se.InvalidSeEbookException(f"Couldn’t open [path]{local_css_path}[/].") from ex
+	OUTPUTS
+	A list of LintMessages representing XHTML errors found in the file
+	"""
 
-	# cssutils prints warnings/errors to stdout by default, so shut it up here
-	cssutils.log.enabled = False
+	messages = []
 
-	# Get the css rules and selectors from helper function
-	local_css_rules, duplicate_selectors = _get_selectors_and_rules(self)
+	# Check for uppercase letters in IDs or classes
+	nodes = dom.xpath("//*[re:test(@id, '[A-Z]') or re:test(@class, '[A-Z]') or re:test(@epub:type, '[A-Z]')]")
+	if nodes:
+		messages.append(LintMessage("x-002", "Uppercase in attribute value. Attribute values must be all lowercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	if duplicate_selectors:
-		messages.append(LintMessage("c-009", "Duplicate CSS selectors. Duplicates are only acceptable if overriding S.E. base styles.", se.MESSAGE_TYPE_WARNING, local_css_path, list(set(duplicate_selectors))))
+	nodes = dom.xpath("//*[re:test(@id, '^[0-9]+')]")
+	if nodes:
+		messages.append(LintMessage("x-007", "[attr]id[/] attributes starting with a number are illegal XHTML.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Store a list of CSS selectors, and duplicate it into a list of unused selectors, for later checks
-	# We use a regex to remove pseudo-elements like ::before, because we want the *selectors* to see if they're unused.
-	local_css_selectors = [regex.sub(r"::[\p{Lowercase_Letter}\-]+", "", selector) for selector in local_css_rules]
-	unused_selectors = local_css_selectors.copy()
+	# Check for double greater-than at the end of a tag
+	matches = regex.findall(r"(>>|>>)", file_contents)
+	if matches:
+		messages.append(LintMessage("x-008", "Elements should end with a single [text]>[/].", se.MESSAGE_TYPE_WARNING, filename))
 
-	local_css = {
-		"has_poem_stye": False,
-		"has_verse_style": False,
-		"has_song_style": False,
-		"has_hymn_style": False,
-		"has_lyrics_style": False,
-		"has_elision_style": False
-		"has_dedication_style": False
-		"has_epigraph_style": False
-	}
-	ebook_info = {
-		"has_subtitle": bool(self.metadata_dom.xpath("/package/metadata/meta[@property='title-type' and text()='subtitle']")),
-		"has_multiple_transcriptions": False,
-		"has_multiple_page_scans": False,
-		"has_other_sources": False
-	}
-	short_story_count = 0
-	missing_styles: List[str] = []
-	directories_not_url_safe = []
-	files_not_url_safe = []
-	id_values = {}
-	duplicate_id_values = []
+	# Check for leading 0 in IDs (note: not the same as checking for IDs that start with an integer)
+	# We only check for *leading* 0s in numbers; this allows IDs like `wind-force-0` in the Worst Journey in the World glossary.
+	nodes = dom.xpath("//*[re:test(@id, '-0[0-9]')]")
+	if nodes:
+		messages.append(LintMessage("x-009", "Illegal leading 0 in [attr]id[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Iterate over rules to do some other checks
-	abbr_with_whitespace = []
-	for selector, rules in local_css_rules.items():
-		if "z3998:poem" in selector:
-			local_css["has_poem_style"] = True
+	# Check for HTML tags in  tags
+	nodes = dom.xpath("/html/head/title/*")
+	if nodes:
+		messages.append(LintMessage("x-010", "Illegal element in [xhtml]<title>[/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-		if "z3998:verse" in selector:
-			local_css["has_verse_style"] = True
+	# Check for underscores in attributes, but not if the attribute is href (links often have underscores) or MathMl alttext, which can have underscores as part of math notation
+	nodes = dom.xpath("//@*[contains(., '_') and name() !='href' and name() !='alttext']/..")
+	if nodes:
+		messages.append(LintMessage("x-011", "Illegal underscore in attribute. Use dashes instead of underscores.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-		if "z3998:song" in selector:
-			local_css["has_song_style"] = True
+	# Check for style attributes
+	nodes = dom.xpath("/html/body//*[@style]")
+	if nodes:
+		messages.append(LintMessage("x-012", "Illegal [attr]style[/] attribute. Don’t use inline styles, any element can be targeted with a clever enough selector.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-		if "z3998:hymn" in selector:
-			local_css["has_hymn_style"] = True
+	# Check for illegal elements in <head>
+	nodes = dom.xpath("/html/head/*[not(self::title) and not(self::link[@rel='stylesheet'])]")
+	if nodes:
+		messages.append(LintMessage("x-015", "Illegal element in [xhtml]<head>[/]. Only [xhtml]<title>[/] and [xhtml]<link rel=\"stylesheet\">[/] are allowed.", se.MESSAGE_TYPE_ERROR, filename, [f"<{node.tag}>" for node in nodes]))
 
-		if "z3998:lyrics" in selector:
-			local_css["has_lyrics_style"] = True
+	# Check for xml:lang attribute starting in uppercase
+	nodes = dom.xpath("//*[re:test(@xml:lang, '^[A-Z]')]")
+	if nodes:
+		messages.append(LintMessage("x-016", "[attr]xml:lang[/] attribute with value starting in uppercase letter.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-		if "span.elision" in selector:
-			local_css["has_elision_style"] = True
+	# Check for ID attributes of numbered paragraphs (like `p-44`) that are used as refs in endnotes.
+	# Make sure their number is actually their correct sequence number in the text.
+	for unexpected_id in se.formatting.find_unexpected_ids(dom):
+		messages.append(LintMessage("x-019", f"Unexpected value of [attr]id[/] attribute. Expected: [attr]{unexpected_id[1]}[/].", se.MESSAGE_TYPE_ERROR, filename, [unexpected_id[0].to_tag_string()]))
 
-		if "dedication" in selector:
-			local_css["has_dedication_style"] = True
+	return messages
 
-		if "epigraph" in selector:
-			local_css["has_epigraph_style"] = True
+def _lint_xhtml_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str) -> list:
+	"""
+	Helper function used in self.lint()
+	Process typo checks on an .xhtml file
 
-		if "abbr" in selector and "nowrap" in rules:
-			abbr_with_whitespace.append(selector)
+	INPUTS
+	filename: The name of the file being checked
+	dom: A dom tree to check
+	file_contents: The contents of the file being checked
+	special_file: A string containing the type of special file the current file is, if any
 
-		if regex.search(r"\[\s*xml\s*\|", selector, flags=regex.IGNORECASE) and "@namespace xml \"http://www.w3.org/XML/1998/namespace\";" not in self.local_css:
-			messages.append(LintMessage("c-003", "[css]\\[xml|attr][/] selector in CSS, but no XML namespace declared ([css]@namespace xml \"http://www.w3.org/XML/1998/namespace\";[/]).", se.MESSAGE_TYPE_ERROR, local_css_path))
+	OUTPUTS
+	A list of LintMessages representing possible typos found in the file
+	"""
 
-	# lxml has not implemented the following in cssselect: *:first-of-type, *:last-of-type, *:nth-of-type, *:nth-last-of-type, *:only-of-type
-	# BUT ONLY WHEN USED ON *. This includes for example: `section [epub|type~="test"]:first-of-type` (note the `*` is implicit)
-	# Therefore we can't simplify them in build or test against them.
-	matches = regex.findall(r"(?:^| )[^a-z\s][^\s]+?:(?:first-of-type|last-of-type|nth-of-type|nth-last-of-type|only-of-type)", self.local_css, flags=regex.MULTILINE)
-	if matches:
-		messages.append(LintMessage("c-001", "Don’t use [css]*:first-of-type[/], [css]*:last-of-type[/], [css]*:nth-of-type[/] [css]*:nth-last-of-type[/], or [css]*:only-of-type[/] on [css]*[/]. Instead, specify an element to apply it to.", se.MESSAGE_TYPE_ERROR, local_css_path, matches))
+	messages = []
+	typos: List[str] = []
 
-	if regex.search(r"\s+hyphens:.+?;(?!\s+-epub-hyphens)", self.local_css):
-		messages.append(LintMessage("c-007", "[css]hyphens[/css] CSS property without [css]-epub-hyphens[/css] copy.", se.MESSAGE_TYPE_ERROR, local_css_path))
+	if special_file != "titlepage":
+		# Don't check the titlepage because it has a standard format and may raise false positives
+		typos = regex.findall(r"(?<!’)\b(and and|the the|if if|of of|or or|as as)\b(?!-)", file_contents, flags=regex.IGNORECASE)
+		typos = typos + regex.findall(r"\ba a\b(?!-)", file_contents)
 
-	if abbr_with_whitespace:
-		messages.append(LintMessage("c-005", f"[css]abbr[/] selector does not need [css]white-space: nowrap;[/] as it inherits it from [path][link=file://{self.path / 'src/epub/css/core.css'}]core.css[/][/].", se.MESSAGE_TYPE_ERROR, local_css_path, abbr_with_whitespace))
+		if typos:
+			messages.append(LintMessage("y-001", "Possible typo: doubled [text]a/the/and/of/or/as/if[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# If we select on the xml namespace, make sure we define the namespace in the CSS, otherwise the selector won't work
-	# We do this using a regex and not with cssutils, because cssutils will barf in this particular case and not even record the selector.
-	matches = regex.findall(r"\[\s*xml\s*\|", self.local_css)
-	if matches and "@namespace xml \"http://www.w3.org/XML/1998/namespace\";" not in self.local_css:
-		messages.append(LintMessage("c-003", "[css]\\[xml|attr][/] selector in CSS, but no XML namespace declared ([css]@namespace xml \"http://www.w3.org/XML/1998/namespace\";[/]).", se.MESSAGE_TYPE_ERROR, local_css_path))
+	# Check for basic typo, but exclude MathML and endnotes. Some endnotes (like in Ten Days that Shook the World) may start with letters.
+	dom_copy = deepcopy(dom)
+	for node in dom_copy.xpath("//a[contains(@epub:type, 'noteref')]") + dom_copy.xpath("//*[namespace-uri() = 'http://www.w3.org/1998/Math/MathML']"):
+		node.remove()
 
-	matches = regex.findall(r"text-align:\s*left\s*;", self.local_css)
-	if matches:
-		messages.append(LintMessage("c-016", "[css]text-align: left;[/] found. Use [css]text-align: initial;[/] instead.", se.MESSAGE_TYPE_ERROR, local_css_path))
+	typos = [node.to_string() for node in dom_copy.xpath("/html/body//p[re:match(., '[a-z][;,][a-z]', 'i')]")]
+	if typos:
+		messages.append(LintMessage("y-002", "Possible typo: punctuation followed directly by a letter, without a space.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	matches = regex.findall(r"[0-9\.]\s?rem;", self.local_css)
-	if matches:
-		messages.append(LintMessage("c-022", "Illegal [css]rem[/] unit. Use [css]em[/] instead.", se.MESSAGE_TYPE_ERROR, local_css_path))
+	# Check for paragraphs ending in lowercase letters, excluding a number of instances where that might be valid.
+	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '[a-z]$') and not(@epub:type) and not(following-sibling::*[1]/node()[1][contains(@epub:type, 'z3998:roman')] or following-sibling::*[1][re:test(., '^([0-9]|first|second|third|fourth|fifth|sixth|seventh|eight|ninth|tenth)', 'i')]) and not(@class or contains(@epub:type, 'z3998:salutation')) and not(following-sibling::*[1][name() = 'blockquote' or name() = 'figure' or name() = 'table' or name() = 'footer' or name() = 'ul' or name() = 'ol'] or ancestor::*[name() = 'blockquote' or name() = 'footer' or name() = 'header' or name() = 'table' or name() = 'ul' or name() = 'ol' or name() = 'li' or name() = 'figure' or name() = 'hgroup' or re:test(@epub:type, '(z3998:drama|z3998:letter|dedication|halftitlepage)')])]")]
+	if typos:
+		messages.append(LintMessage("y-003", "Possible typo: paragraph missing ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	matches = regex.findall(r"font-size\s*:\s*[0-9\.]+(?![0-9\.]|em|ex)", self.local_css)
-	if matches:
-		messages.append(LintMessage("c-023", "Illegal unit used to set [css]font-size[/]. Hint: Use [css]em[/] units.", se.MESSAGE_TYPE_ERROR, local_css_path))
+	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '-[‘“]')]")]
+	if typos:
+		messages.append(LintMessage("y-004", "Possible typo: mis-curled quotation mark after dash.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	matches = regex.findall(r"line-height\s*:\s*[0-9\.]+(?!;|[0-9\.]+)", self.local_css)
-	if matches:
-		messages.append(LintMessage("c-024", "Illegal unit used to set [css]line-height[/]. Hint: [css]line-height[/] is set without any units.", se.MESSAGE_TYPE_ERROR, local_css_path))
+	# Check for commas or periods following exclamation or question marks. Exclude the colophon because we may have painting names with punctuation.
+	if special_file != "colophon":
+		typos = [node.to_string() for node in dom.xpath("//p[re:test(., '[!?][\\.,][^”’]')]")]
+		if typos:
+			messages.append(LintMessage("y-005", "Possible typo: question mark or exclamation mark followed by period or comma.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	matches = regex.findall(r"(height|\stop|\sbottom)\s*:\s*[0-9\.]+%", self.local_css)
-	if matches:
-		messages.append(LintMessage("c-025", "Illegal percent unit used to set [css]height[/] or positioning property. Hint: [css]vh[/] to specify vertical-oriented properties like height or position.", se.MESSAGE_TYPE_ERROR, local_css_path))
+	# Check for opening lsquo in quote that doesn't appear to have a matching rsquo. Rsquos must be followed by a space, punctuation (except period/comma), word joiner and thus em dash, or a number for an endnote.
+	typos = dom.xpath(f"re:match(//*, '“\\s*‘[^“]+?”', 'g')/text()[not(re:test(., '’[\\s\\?\\!;<2060{se.WORD_JOINER}0-9]'))]")
+	if typos:
+		messages.append(LintMessage("y-006", "Possible typo: [text]‘[/] without matching [text]’[/]. Hints: [text]’[/] are used for abbreviations;	commas and periods must go inside quotation marks.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Done checking local.css
-	missing_files = []
-	if self.is_se_ebook:
-		root_files = os.listdir(self.path)
-		expected_root_files = ["images", "src", "LICENSE.md"]
-		illegal_files = [root_file for root_file in root_files if root_file not in expected_root_files and root_file != "se-lint-ignore.xml"] # se-lint-ignore.xml is optional
-		missing_files = [expected_root_file for expected_root_file in expected_root_files if expected_root_file not in root_files and expected_root_file != "LICENSE.md"] # We add more to this later on. LICENSE.md gets checked later on, so we don't want to add it twice
+	# Try to find top-level lsquo; for example, <p>“Bah!” he said to the ‘minister.’</p>
+	# We can't do this on xpath because we can't iterate over the output of re:replace().
+	# A fully-featured solution would parse quotes to find balanced pairs. This quick-n-dirty
+	# solution takes the raw XHTML string, strips any seemingly-valid balanced pairs using regex,
+	# then searches for unmatched `‘`. If found, it returns the dom node of the match; we can't return
+	# the actual string match because we stripped surrounding text.
 
-		# If we have illegal files, check if they are tracked in Git.
-		# If they are, then they're still illegal.
-		# If not, ignore them for linting purposes.
-		if illegal_files:
-			try:
-				illegal_files = self.repo.git.ls_files(illegal_files).split("\n")
-				if illegal_files and illegal_files[0] == "":
-					illegal_files = []
-			except Exception:
-				# If we can't initialize Git, then just pass through the list of illegal files
-				pass
+	# Start by merging all <p> into single lines, including poetry, but exclude blockquotes as they have special quoting rules
+	temp_xhtml = ""
+	dom_copy = deepcopy(dom)
+	node_number = 0
+	# Add IDs to the dom so we can find the paragraphs after we run our regex
+	# Exclude paragraphs in blockquotes, which may have special quoting rules, and "continued" paragraphs, which may be continued dialog without an “
+	for node in dom_copy.xpath("/html/body//p[not(ancestor::blockquote) and not(contains(@class, 'continued'))]"):
+		node.set_attr("id", "lint-" + str(node_number))
+		temp_xhtml = temp_xhtml + f"<p id=\"lint-{node_number}\">" + regex.sub(r"[\s\n]+", " ", node.inner_text(), flags=regex.DOTALL) + "\n"
+		node_number = node_number + 1
 
-		for illegal_file in illegal_files:
-			messages.append(LintMessage("f-001", "Illegal file or directory.", se.MESSAGE_TYPE_ERROR, Path(illegal_file)))
+	replacement_count = 1
+	while replacement_count > 0:
+		# Remove all regular quotes. Run this in a loop because we may need to remove triple-nested quotes.
+		(temp_xhtml, replacement_count) = regex.subn(r"“[^“]+?”", " ", temp_xhtml) # Remove all regular quotes
 
-	# Check for repeated punctuation
-	nodes = self.metadata_dom.xpath("/package/metadata/*[re:test(., '[,;]{2,}.{0,20}')]")
-	if nodes:
-		messages.append(LintMessage("t-008", "Repeated punctuation.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path, [node.to_string() for node in nodes]))
+	# Remove contractions to reduce rsquo for next regex
+	temp_xhtml = regex.sub(r"[\p{Letter}]’[\p{Letter}]", " ", temp_xhtml, flags=regex.MULTILINE)
 
-	# Check for non-English Wikipedia URLs
-	nodes = self.metadata_dom.xpath("/package/metadata/*[re:test(., 'https://(?!en)[a-z]{2,}\\.wikipedia\\.org')]")
-	if nodes:
-		messages.append(LintMessage("m-043", "Non-English Wikipedia URL.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
+	# Remove all runs of ldquo that are likely to spill to the next <p>
+	replacement_count = 1
+	while replacement_count > 0:
+		(temp_xhtml, replacement_count) = regex.subn(r"“[^“”]+?$", " ", temp_xhtml, flags=regex.MULTILINE)
 
-	# Check the long description for some errors
-	try:
-		# Check if there are non-typogrified quotes or em-dashes in metadata descriptions
-		# lxml unescapes this for us
-		# Also, remove HTML elements like <a href> so that we don't catch quotation marks in attribute values
-		long_description = self.metadata_dom.xpath("/package/metadata/meta[@property='se:long-description']")[0].text
-		matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", regex.sub(r"<[^<]+?>", "", long_description))
-		if matches:
-			messages.append(LintMessage("m-014", "Non-typogrified character in [xml]<meta property=\"se:long-description\">[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	# Match problem `‘` using regex, and if found, get the actual node text from the dom to return.
+	typos = []
+	for match in regex.findall(r"""<p id="lint-([0-9]+?)">.*‘[\p{Letter}\s]""", temp_xhtml):
+		for node in dom_copy.xpath(f"/html/body//p[@id='lint-{match}' and re:test(., '‘[A-Za-z\\s]')]"):
+			typos.append(node.inner_text())
 
-		# Is the first instance of the author's last name a hyperlink in the metadata?
-		authors = self.metadata_dom.xpath("/package/metadata/dc:creator")
-		for author in authors:
-			author_sort = self.metadata_dom.xpath(f"/package/metadata/meta[@property='file-as'][@refines='#{author.get_attr('id')}']/text()")
-			if author_sort:
-				author_last_name = regex.sub(r",.+$", "", author_sort[0])
-				author_last_name = author_last_name.replace("'", "’") # Typogrify apostrophes so that we correctly match in the long description
-				# We can't use xpath here because the long description is escaped; it has no dom to query against.
-				if author_last_name.lower() in regex.sub(r"https://standardebooks\.org/ebooks/[^/]+?/", "", long_description.lower()) and not regex.search(fr"<a href=\"https://standardebooks\.org/ebooks/.+?\">.*?{author_last_name}.*?</a>", long_description, flags=regex.IGNORECASE):
-					messages.append(LintMessage("m-056", "Author name present in [xml]<meta property=\"se:long-description\">[/] element, but the first instance of their name is not hyperlinked to their S.E. author page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	if typos:
+		messages.append(LintMessage("y-007", "Possible typo: [text]‘[/] not within [text]“[/]. Hints: Should [text]‘[/] be replaced with [text]“[/]? Is there a missing closing quote? Is this a nested quote that should be preceded by [text]“[/]? Are quotes in close proximity correctly closed?", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		# Did we mention an SE book in the long description, but without italics?
-		# Only match if the title appears to contain an uppercase letter. This prevents matches on a non-title link like `<a href>short stories</a>`
-		matches = regex.search(r"""(?<!<i>)<a href="https://standardebooks\.org/ebooks/[^"]+?/[^"]+?">(?!<i>)([\p{Letter}\s]+)""", long_description)
-		if matches and regex.search(r"[\p{Uppercase_Letter}]", matches[1]):
-			messages.append(LintMessage("m-064", "S.E. ebook hyperlinked in long description but not italicized.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	# Check for single quotes when there should be double quotes in an interjection in dialog
+	typos = [node.to_string() for node in dom.xpath("//p[re:test(., '“[^”]+?’⁠—[^”]+?—“')]")]
+	if typos:
+		messages.append(LintMessage("y-008", "Possible typo: dialog interrupted by interjection but with incorrect closing quote.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		if regex.search(r"""<a href="https?://(?!standardebooks\.org)""", long_description):
-			messages.append(LintMessage("m-067", "Non-S.E. link in long description.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	# Check for dialog starting with a lowercase letter. Only check the first child text node of <p>, because other first children might be valid lowercase, like <m:math> or <b>;
+	# exclude <p> inside or preceded by <blockquote>; and exclude <p> inside endnotes, as definitions may start with lowercase letters.
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[not(ancestor::blockquote or ancestor::li[contains(@epub:type, 'endnote')]) and not(preceding-sibling::*[1][name()='blockquote'])][re:test(./node()[1], '^“[a-z]')]")]
+	if typos:
+		messages.append(LintMessage("y-009", "Possible typo: dialog begins with lowercase letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		# xml:lang is correct for the rest of the publication, but should be lang in the long desc
-		if "xml:lang" in long_description:
-			messages.append(LintMessage("m-057", "[xml]xml:lang[/] attribute in [xml]<meta property=\"se:long-description\">[/] element should be [xml]lang[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	typos = [node.to_string() for node in dom.xpath("//p[not(ancestor::blockquote) and not(following-sibling::*[1][name() = 'blockquote' or contains(@class, 'continued')]) and re:test(., ',”$')]")]
+	if typos:
+		messages.append(LintMessage("y-010", "Possible typo: comma ending dialogue.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		# US -> U.S.
-		matches = regex.findall(r"\bUS\b", long_description)
-		if matches:
-			messages.append(LintMessage("t-047", "[text]US[/] should be [text]U.S.[/]", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '’{2,}')]")]
+	if typos:
+		messages.append(LintMessage("y-011", "Possible typo: two or more [text]’[/] in a row.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		# Make sure long-description is escaped HTML
-		if not regex.search(r"^\s*<p>", long_description) and long_description.strip() != "LONG_DESCRIPTION":
-			messages.append(LintMessage("m-016", "Long description must be escaped HTML.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-		else:
-			# Check for malformed long description HTML
-			try:
-				etree.parse(io.StringIO(f"<?xml version=\"1.0\"?><html xmlns=\"http://www.w3.org/1999/xhtml\">{long_description}</html>"))
-			except lxml.etree.XMLSyntaxError as ex:
-				messages.append(LintMessage("m-015", f"Metadata long description is not valid XHTML. LXML says: {ex}", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '”[a-z]')]")]
+	if typos:
+		messages.append(LintMessage("y-012", "Possible typo: [text]”[/] directly followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		# Check for apostrophes outside links in long description
-		matches = regex.findall(r"</a>’s", long_description) + regex.findall(r"s</a>’", long_description)
-		if matches:
-			messages.append(LintMessage("m-044", "Possessive [text]’[/] or [text]’s[/] outside of [xhtml]<a>[/] element in long description.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	# Check for period not inside rsquo
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[^”’]+?’[\\.,]”')]")]
+	if typos:
+		messages.append(LintMessage("y-013", "Possible typo: punctuation not within [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		# Check for HTML entities in long-description, but allow &amp;
-		matches = regex.findall(r"&[a-z0-9]+?;", long_description.replace("&", ""))
-		if matches:
-			messages.append(LintMessage("m-018", "HTML entities found. Use Unicode equivalents instead.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	# Check for period before dialog tag; try to exclude abbrevations that close a quotation, like `“<abbr>Mr.</abbr>”`.
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[(re:test(., '\\.”\\s[a-z\\s]*?(\\bsaid|[a-z]+ed)') or re:test(., '\\.”\\s(s?he|they?|and)\\b')) and not(.//abbr[following-sibling::node()[re:test(., '^”')]])]")]
+	if typos:
+		messages.append(LintMessage("y-014", "Possible typo: Unexpected [text].[/] at the end of quotation. Hint: If a dialog tag follows, should this be [text],[/]?", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	except Exception:
-		if self.is_se_ebook:
-			missing_metadata_elements.append("""<meta id="long-description" property="se:long-description" refines="#description">""")
+	# Check for mis-curled ‘ or ‘ without matching ’
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘[A-Za-z][^“’]+?”')]")]
+	if typos:
+		messages.append(LintMessage("y-015", "Possible typo: mis-curled [text]‘[/] or missing [text]’[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	missing_metadata_vars = []
-	for node in self.metadata_dom.xpath("/package/metadata/*/text()"):
-		for var in SE_VARIABLES:
-			if regex.search(fr"\b{var}\b", node):
-				missing_metadata_vars.append(var)
-				# Quit the loop early to save some time
-				break
+	# Check for two periods in a row, almost always a typo for one period or a hellip
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\.\\.[^\\.]')]")]
+	if typos:
+		messages.append(LintMessage("y-016", "Possible typo: consecutive periods ([text]..[/]).", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	if missing_metadata_vars:
-		messages.append(LintMessage("m-036", "Variable not replaced with value.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, missing_metadata_vars))
+	# Check for ldquo followed by space
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '“\\s+[^‘’]')]")]
+	if typos:
+		messages.append(LintMessage("y-017", "Possible typo: [text]“[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check if there are non-typogrified quotes or em-dashes in the title.
-	try:
-		title = self.metadata_dom.xpath("/package/metadata/dc:title")[0].text
-		matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", title)
-		if matches:
-			messages.append(LintMessage("m-012", "Non-typogrified character in [xml]<dc:title>[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	# Check for lsquo followed by space
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '‘\\s+[^“’]')]")]
+	if typos:
+		messages.append(LintMessage("y-018", "Possible typo: [text]‘[/] followed by space.", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-		# Do we need an dcterms:alternate meta element?
-		# Match spelled-out numbers with a word joiner, so for ex. we don't print "eight" if we matched "eighty"
-		matches = regex.findall(r"(?:[0-9]+|\bone\b|\btwo\b|\bthree\b|\bfour\b|\bfive\b|\bsix\b|\bseven\b|\beight\b|\bnine\b|\bten\b|\beleven\b|\btwelve\b|\bthirteen\b|\bfourteen\b|\bfifteen\b|\bsixteen\b|\bseventeen\b|\beighteen\b|\bnineteen\b|\btwenty\b|\bthirty\b|\bforty\b|\bfifty\b|\bsixty\b|\bseventy\b|\beighty|\bninety)", title, flags=regex.IGNORECASE)
-		if matches and not self.metadata_dom.xpath("/package/metadata/meta[@property='dcterms:alternate']"):
-			messages.append(LintMessage("m-052", "[xml]<dc:title>[/] element contains numbers, but no [xml]<meta property=\"dcterms:alternate\" refines=\"#title\"> element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
-	except Exception:
-		missing_metadata_elements.append("<dc:title>")
+	# Check for closing rdquo without opening ldquo. We ignore blockquotes because they usually have unique quote formatting.
+	# Remove tds in case rdquo means "ditto mark"
+	typos = regex.findall(r"”[^“‘]+?”", regex.sub(r"<td>[”\s]+?(<a .+?epub:type=\"noteref\">.+?</a>)?</td>", "", file_contents), flags=regex.DOTALL)
 
-	try:
-		file_as = self.metadata_dom.xpath("/package/metadata/meta[@property='file-as' and @refines='#title']")[0].text
-		matches = regex.findall(r".(?:['\"]|\-\-|\s-\s).", file_as)
-		if matches:
-			messages.append(LintMessage("m-050", "Non-typogrified character in [xml]<meta property=\"file-as\" refines=\"#title\">[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
-	except Exception:
-		missing_metadata_elements.append("<meta property=\"file-as\" refines=\"#title\">")
+	# We create a filter to try to exclude nested quotations
+	# Remove tags in case they're enclosing punctuation we want to match against at the end of a sentence.
+	typos = [match for match in typos if not regex.search(r"(?:[\.!\?;…—]|”\s)’\s", se.formatting.remove_tags(match))]
 
-	try:
-		description = self.metadata_dom.xpath("/package/metadata/dc:description")[0].text
-		matches = regex.findall(r"(?:['\"]|\-\-|\s-\s)", description)
-		if matches:
-			messages.append(LintMessage("m-013", "Non-typogrified character in [xml]<dc:description>[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
-	except Exception:
-		missing_metadata_elements.append("<dc:description>")
+	# Try some additional matches before adding the lint message
+	# Search for <p> tags that have an ending closing quote but no opening quote; but exclude <p>s that are preceded by a <blockquote>
+	# or that have a <blockquote> ancestor, because that may indicate that the opening quote is elsewhere in the quotation.
+	for node in dom.xpath("//p[re:test(., '^[^“]+”') and not(./preceding-sibling::*[1][name() = 'blockquote']) and not(./ancestor::*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]) and not(./ancestor::blockquote)]"):
+		typos.append(node.to_string()[-20:])
 
-	# Set some variables for later
-	transcription_source_count = 0
-	page_scan_source_count = 0
-	other_source_count = 0
-	sources = self.metadata_dom.xpath("/package/metadata/dc:source")
-	for source in sources:
-		if regex.search(r"(gutenberg\.org|wikisource\.org|fadedpage\.com|gutenberg\.net\.au|gutenberg\.ca)", source.inner_text()):
-			transcription_source_count = transcription_source_count + 1
-		elif regex.search(r"(hathitrust\.org|/archive\.org|books\.google\.com|google\.com/books)", source.inner_text()):
-			page_scan_source_count = page_scan_source_count + 1
-		else:
-			other_source_count = other_source_count + 1
+	if typos:
+		messages.append(LintMessage("y-019", "Possible typo: [text]”[/] without opening [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	ebook_info["has_multiple_transcriptions"] = transcription_source_count >= 2
-	ebook_info["has_multiple_page_scans"] = page_scan_source_count >= 2
-	ebook_info["has_other_sources"] = other_source_count > 0
+	# Check for ,.
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., ',\\.')]")]
+	if typos:
+		messages.append(LintMessage("y-020", "Possible typo: consecutive comma-period ([text],.[/]).", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for double spacing
-	if self.metadata_dom.xpath(f"/package/metadata/*[re:test(., '[{se.NO_BREAK_SPACE}{se.HAIR_SPACE} ]{{2,}}')]"):
-		double_spaced_files.append(self.metadata_file_path)
+	# Check for single quotes instead of double quotes
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[not(contains(@class, 'continued') or ancestor::blockquote or ancestor::*[re:test(@epub:type, 'z3998:(verse|song|poem|hymn)')]) and re:test(., '^[^“]+?‘')]")]
+	if typos:
+		messages.append(LintMessage("y-021", "Possible typo: Opening [text]‘[/] without preceding [text]“[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
 
-	# Check for illegal Wikipedia URLs
-	nodes = self.metadata_dom.xpath("/package/metadata/*[contains(., '.m.wikipedia.org') or @*[contains(., '.m.wikipedia.org')]]")
+	# Check for two quotations in one paragraph
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^”]+?”\\s“[^”]+?”$')]")]
+	if typos:
+		messages.append(LintMessage("y-022", "Possible typo: consecutive quotations without intervening text, e.g. [text]“…” “…”[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for incorrectly nested quotation marks
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘”]+“') and not(.//br)]")]
+	if typos:
+		messages.append(LintMessage("y-023", "Possible typo: two opening quotation marks in a run. Hint: Nested quotes should switch between [text]“[/] and [text]‘[/]", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for dashes instead of em-dashes
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '\\s[a-z]+-(the|there|is|and|they|when)\\s')]")]
+	if typos:
+		messages.append(LintMessage("y-024", "Possible typo: dash before [text]the/there/is/and/they/when[/] probably should be em-dash.", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for comma not followed by space but followed by quotation mark
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '[a-z],[“”‘’][a-z]', 'i')]")]
+	if typos:
+		messages.append(LintMessage("y-025", "Possible typo: comma without space followed by quotation mark.", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for punctuation missing before conjunctions. Ignore <p> with an <i> child starting in a conjunction, as those are probably book titles or non-English languages
+	typos = [node.to_string() for node in dom.xpath(f"/html/body//p[not(parent::hgroup) and re:test(., '\\b[a-z]+\\s(But|And|For|Nor|Yet|Or)\\b[^’\\.\\?\\-{se.WORD_JOINER}]') and not(./i[re:test(., '^(But|And|For|Nor|Yet|Or)\\b')])]")]
+	if typos:
+		messages.append(LintMessage("y-026", "Possible typo: no punctuation before conjunction [text]But/And/For/Nor/Yet/Or[/].", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for extra closing single quote at the end of dialog
+	typos = [node.to_string() for node in dom.xpath("/html/body//p[re:test(., '^“[^‘]+”\\s*’$')]")]
+	if typos:
+		messages.append(LintMessage("y-027", "Possible typo: Extra [text]’[/] at end of paragraph.", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for `<abbr>` preceded or followed by text. Ignore compass directions followed by `ly`, like S.S.W.ly
+	typos = [node.to_string() for node in dom.xpath("/html/body//abbr[(preceding-sibling::node()[1])[re:test(., '[A-Za-z]$')] or (following-sibling::node()[1])[re:test(., '^[A-Za-z](?<!s\\b)') and not((./preceding-sibling::abbr[1])[contains(@epub:type, 'se:compass')] and re:test(., '^ly\\b'))]]")]
+	if typos:
+		messages.append(LintMessage("y-028", "Possible typo: [xhtml]<abbr>[/] directly preceded or followed by letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for misapplied italics. Ignore 's' because the plural is too common.
+	typos = [node.to_string() for node in dom.xpath("/html/body//*[(name() = 'i' or name() = 'em') and ./following-sibling::node()[1][re:test(., '^[a-z]\\b', 'i') and not(re:test(., '^s\\b'))]]")]
+	if typos:
+		messages.append(LintMessage("y-029", "Possible typo: Italics followed by a letter.", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for lowercase letters starting quotations after a preceding period
+	typos = dom.xpath("/html/body//p/child::text()[re:test(., '\\.\\s[‘“][a-z]')]")
+	if typos:
+		messages.append(LintMessage("y-030", "Possible typo: Lowercase quotation following a period. Check either that the period should be a comma, or that the quotation should start with a capital.", se.MESSAGE_TYPE_WARNING, filename, typos))
+
+	# Check for missing punctuation in continued quotations
+	# ” said Bob “
+	nodes = dom.xpath("/html/body//p[re:test(., '”\\s(?:said|[A-Za-z]{2,}ed)\\s[A-Za-z]+?(?<!\\bthe)(?<!\\bto)(?<!\\bwith)(?<!\\bfrom)(?<!\\ba\\b)(?<!\\bis)\\s“') or re:test(., '[^\\.]”\\s(\\bhe\\b|\\bshe\\b|I|[A-Z][a-z]+?)\\s(?:said|[A-Za-z]{2,}ed)\\s“') or re:test(., ',” (?:said|[A-Za-z]{2,}ed) [A-Za-z]+? [A-Za-z]+?ly “')]")
 	if nodes:
-		messages.append(LintMessage("m-026", "Illegal [url]https://*.m.wikipedia.org[/] URL. Hint: use non-mobile Wikipedia URLs.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("y-031", "Possible typo: Dialog tag missing punctuation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for punctuation outside quotes. We don't check single quotes because contractions are too common.
-	# We can't use xpath's built-in regex because it doesn't support Unicode classes
-	for node in self.metadata_dom.xpath("/package/metadata/*"):
-		if node.text and regex.search(r"[\p{Letter}]+”[,\.](?! …)", node.text):
-			messages.append(LintMessage("t-002", "Comma or period outside of double quote. Generally punctuation goes within single and double quotes.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path))
-			break
+	# Check for italics having epub:type that run in to preceding or following characters
+	# Ignore things like <i>Newspaper</i>s
+	nodes = dom.xpath("/html/body//i[@epub:type and ( (following-sibling::node()[1][re:test(., '^[a-z]', 'i') and not(re:test(., '^(s|es|er)'))]) or preceding-sibling::node()[1][re:test(., '[a-z]$')]) ]")
+	if nodes:
+		messages.append(LintMessage("y-032", "Possible typo: Italics running into preceding or following characters.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check that the word count is correct, if it's currently set
-	word_count = self.metadata_dom.xpath("/package/metadata/meta[@property='se:word-count']/text()", True)
-	if self.is_se_ebook:
-		if word_count is None:
-			missing_metadata_elements.append("""<meta property="se:word-count">""")
-		elif word_count != "WORD_COUNT" and int(word_count) != self.get_word_count():
-			messages.append(LintMessage("m-065", "Word count in metadata doesn’t match actual word count.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	return messages
+
+def _lint_image_metadata_checks(self, has_images: bool) -> list:
+	messages = []
+
+	# Check for some accessibility metadata regarding images
+	has_visual_accessmode = len(self.metadata_dom.xpath("/package/metadata/meta[@property='schema:accessMode' and text() = 'visual']")) > 0
+	has_accessibility_feature_alt = len(self.metadata_dom.xpath("/package/metadata/meta[@property='schema:accessibilityFeature' and text() = 'alternativeText']")) > 0
+	has_wat_role = len(self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text() = 'wat']")) > 0
+
+	if has_images:
+		if not has_visual_accessmode:
+			messages.append(LintMessage("m-028", "Images found in ebook, but no [attr]schema:accessMode[/] property set to [val]visual[/] in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+
+		if not has_accessibility_feature_alt:
+			messages.append(LintMessage("m-029", "Images found in ebook, but no [attr]schema:accessibilityFeature[/] property set to [val]alternativeText[/] in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+
+		if not has_wat_role:
+			messages.append(LintMessage("m-040", "Images found in ebook, but no [attr]role[/] property set to [val]wat[/] in metadata for the writer of the alt text.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+
+	if not has_images:
+		if has_visual_accessmode:
+			messages.append(LintMessage("m-038", "[attr]schema:accessMode[/] property set to [val]visual[/] in metadata, but no images in ebook.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+
+		if has_accessibility_feature_alt:
+			messages.append(LintMessage("m-039", "[attr]schema:accessibilityFeature[/] property set to [val]alternativeText[/] in metadata, but no images in ebook.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+
+	return messages
+
+def _lint_process_ignore_file(self, skip_lint_ignore: bool, messages: list) -> list:
+	# This is a dict with where keys are the path and values are a list of code dicts.
+	# Each code dict has a key "code" which is the actual code, and a key "used" which is a
+	# bool indicating whether or not the code has actually been caught in the linting run.
+	ignored_codes: Dict[str, List[Dict]] = {}
+
+	# First, check if we have an se-lint-ignore.xml file in the ebook root. If so, parse it. For an example se-lint-ignore file, see semos://1.0.0/2.3
+	lint_ignore_path = self.path / "se-lint-ignore.xml"
+	if not skip_lint_ignore and lint_ignore_path.exists():
+		lint_config = self.get_dom(lint_ignore_path)
+
+		elements = lint_config.xpath("/se-lint-ignore/file")
+
+		if not elements:
+			messages.append(LintMessage("m-049", "No [path]se-lint-ignore.xml[/] rules. Delete the file if there are no rules.", se.MESSAGE_TYPE_ERROR, lint_ignore_path))
+
+		has_illegal_path = False
+
+		if not allowed_messages:
+			allowed_messages = []
+
+		for element in elements:
+			path = element.get_attr("path").strip()
+
+			if path == "*":
+				has_illegal_path = True # Set a bool so that we set a lint error later, to prevent adding it multiple times
+
+			if path not in ignored_codes:
+				ignored_codes[path] = []
+
+			for ignore in element.lxml_element:
+				if ignore.tag == "ignore":
+					has_reason = False
+					for child in ignore:
+						if child.tag == "code" and child.text.strip() not in allowed_messages:
+							ignored_codes[path].append({"code": child.text.strip(), "used": False})
+
+						if child.tag == "reason" and child.text.strip() != "":
+							has_reason = True
+
+					if not has_reason:
+						messages.append(LintMessage("m-046", "Missing or empty [xml]<reason>[/] element.", se.MESSAGE_TYPE_ERROR, lint_ignore_path))
+
+		if has_illegal_path:
+			messages.append(LintMessage("m-047", "Ignoring [path]*[/] is too general. Target specific files if possible.", se.MESSAGE_TYPE_WARNING, lint_ignore_path))
+
+	# Done parsing ignore list
+
+	# Now that we have our lint messages, we filter out ones that we've ignored.
+	if ignored_codes:
+		# Iterate over a copy of messages, so that we can remove from them while iterating.
+		for message in messages[:]:
+			for path, codes in ignored_codes.items():
+				for code in codes:
+					try:
+						# fnmatch.translate() converts shell-style globs into a regex pattern
+						if regex.match(fr"{translate(path)}", str(message.filename.name) if message.filename else "") and message.code == code["code"]:
+							messages.remove(message)
+							code["used"] = True
+
+					except ValueError:
+						# This gets raised if the message has already been removed by a previous rule.
+						# For example, chapter-*.xhtml gets t-001 removed, then subsequently *.xhtml gets t-001 removed.
+						pass
+					except Exception as ex:
+						raise se.InvalidInputException(f"Invalid path in [path][link=file://{lint_ignore_path}]se-lint-ignore.xml[/][/] rule: [path]{path}[/].") from ex
+
+		# Check for unused ignore rules
+		unused_codes: List[str] = []
+		for path, codes in ignored_codes.items():
+			for code in codes:
+				if not code["used"]:
+					unused_codes.append(f"{path}, {code['code']}")
+
+		if unused_codes:
+			messages.append(LintMessage("m-048", f"Unused [path][link=file://{lint_ignore_path}]se-lint-ignore.xml[/][/] rule.", se.MESSAGE_TYPE_ERROR, lint_ignore_path, unused_codes))
+
+	return messages
+
+
+def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = None) -> list:
+	"""
+	Check this ebook for some common SE style errors.
+
+	INPUTS
+	self
+	skip_lint_ignore: Flag indicating whether ignore file should be used
+
+	OUTPUTS
+	A list of LintMessage objects.
+	"""
+
+	local_css_path = self.content_path / "css/local.css"
+	messages: List[LintMessage] = []
+	cover_svg_title = ""
+	titlepage_svg_title = ""
+	xhtml_css_classes: Dict[str, int] = {}
+	headings: List[tuple] = []
+	double_spaced_files: List[Path] = []
+	unused_selectors: List[str] = []
+	id_attrs: List[str] = []
+	abbr_elements_requiring_css: List[se.easy_xml.EasyXmlElement] = []
+	glossary_usage = []
+	short_story_count = 0
+	missing_styles: List[str] = []
+	directories_not_url_safe = []
+	files_not_url_safe = []
+	id_values = {}
+	duplicate_id_values = []
+	local_css = {
+		"has_poem_style": False,
+		"has_verse_style": False,
+		"has_song_style": False,
+		"has_hymn_style": False,
+		"has_lyrics_style": False,
+		"has_elision_style": False,
+		"has_dedication_style": False,
+		"has_epigraph_style": False
+	}
+	ebook_flags = {
+		"has_cover_source": False,
+		"has_frontmatter": False,
+		"has_glossary_search_key_map": False,
+		"has_halftitle": False,
+		"has_subtitle": bool(self.metadata_dom.xpath("/package/metadata/meta[@property='title-type' and text()='subtitle']")),
+		"has_images": False,
+		"has_multiple_transcriptions": False,
+		"has_multiple_page_scans": False,
+		"has_other_sources": False
+	}
+
+	# Cache the browser default stylesheet for later use
+	with importlib_resources.open_text("se.data", "browser.css", encoding="utf-8") as css:
+		self._file_cache["default"] = css.read() # pylint: disable=protected-access
+
+	# Check that the spine has all the expected files in the book
+	missing_spine_files = files_not_in_spine(self)
+	if missing_spine_files:
+		for missing_spine_file in missing_spine_files:
+			messages.append(LintMessage("f-007", "File not listed in [xml]<spine>[/].", se.MESSAGE_TYPE_ERROR, missing_spine_file))
 
-	# Check if we have a subtitle but no fulltitle
-	if self.metadata_dom.xpath("/package/metadata[./meta[@property='title-type' and text()='subtitle'] and not(./meta[@property='title-type' and text()='extended'])]"):
-		messages.append(LintMessage("m-011", "Subtitle in metadata, but no full/extended title element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	# Get the ebook language for later use
+	try:
+		language = self.metadata_dom.xpath("/package/metadata/dc:language")[0].text
+	except se.InvalidXmlException as ex:
+		raise ex
+	except Exception as ex:
+		raise se.InvalidSeEbookException(f"Missing [xml]<dc:language>[/] element in [path][link=file://{self.metadata_file_path}]{self.metadata_file_path.name}[/][/].") from ex
 
-	# Check for tags that imply other tags
-	implied_tags = {"Fiction": ["Science Fiction", "Drama", "Fantasy"]}
-	for implied_tag, tags in implied_tags.items():
-		if self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:subject' and text()={se.easy_xml.escape_xpath(implied_tag)}]"):
-			for tag in tags:
-				if self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:subject' and text()={se.easy_xml.escape_xpath(tag)}]"):
-					messages.append(LintMessage("m-058", f"[val]se:subject[/] of [text]{implied_tag}[/] found, but [text]{tag}[/] implies [text]{implied_tag}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, matches))
+	# Check local.css for various items, for later use
+	try:
+		self.local_css = self.get_file(local_css_path)
+	except Exception as ex:
+		raise se.InvalidSeEbookException(f"Couldn’t open [path]{local_css_path}[/].") from ex
 
-	# Check for 'comprised of'
-	if self.metadata_dom.xpath("/package/metadata/*[re:test(., '[Cc]omprised of')]"):
-		messages.append(LintMessage("m-069", "[text]comprised of[/] in metadata. Hint: Is there a better phrase to use here?", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	# cssutils prints warnings/errors to stdout by default, so shut it up here
+	cssutils.log.enabled = False
 
-	# Check for illegal em-dashes in <dc:subject>
-	nodes = self.metadata_dom.xpath("/package/metadata/dc:subject[contains(text(), '—')]")
-	if nodes:
-		messages.append(LintMessage("m-019", "Illegal em-dash in [xml]<dc:subject>[/] element; use [text]--[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.text for node in nodes]))
+	# Get the css rules and selectors from helper function
+	local_css_rules, duplicate_selectors = _get_selectors_and_rules(self)
 
-	# Check for incorrect 'anonymous' strings in metadata
-	nodes = self.metadata_dom.xpath("/package/metadata/dc:contributor[re:test(., 'anonymous', 'i') and text() != 'Anonymous'] | /package/metadata/meta[@property='file-as' and re:test(., 'anonymous', 'i') and text() != 'Anonymous']")
-	if nodes:
-		messages.append(LintMessage("m-073", "Anonymous contributor values must be exactly [text]Anonymous[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
+	if duplicate_selectors:
+		messages.append(LintMessage("c-009", "Duplicate CSS selectors. Duplicates are only acceptable if overriding S.E. base styles.", se.MESSAGE_TYPE_WARNING, local_css_path, list(set(duplicate_selectors))))
 
-	# Check for metadata elements
-	nodes = self.metadata_dom.xpath("/package/metadata/*[not(name()='link') and not(normalize-space(.)) and not(./*)]")
-	if nodes:
-		messages.append(LintMessage("m-022", "Empty element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in nodes]))
+	# Store a list of CSS selectors, and duplicate it into a list of unused selectors, for later checks
+	# We use a regex to remove pseudo-elements like ::before, because we want the *selectors* to see if they're unused.
+	local_css_selectors = [regex.sub(r"::[\p{Lowercase_Letter}\-]+", "", selector) for selector in local_css_rules]
+	unused_selectors = local_css_selectors.copy()
 
-	# Check for illegal VCS URLs
-	nodes = self.metadata_dom.xpath(f"/package/metadata/meta[@property='se:url.vcs.github' and not(text()='{self.generated_github_repo_url}')]")
-	if nodes:
-		messages.append(LintMessage("m-009", f"[xml]<meta property=\"se:url.vcs.github\">[/] value does not match expected: [url]{self.generated_github_repo_url}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	# Iterate over rules to do some other checks
+	abbr_with_whitespace = []
+	for selector, rules in local_css_rules.items():
+		if "z3998:poem" in selector:
+			local_css["has_poem_style"] = True
 
-	# Check for illegal se:subject tags
-	illegal_subjects = []
-	nodes = self.metadata_dom.xpath("/package/metadata/meta[@property='se:subject']/text()")
-	if nodes:
-		for node in nodes:
-			if node not in SE_GENRES and node != "TAG":
-				illegal_subjects.append(node)
+		if "z3998:verse" in selector:
+			local_css["has_verse_style"] = True
 
-		if illegal_subjects:
-			messages.append(LintMessage("m-020", "Illegal value for [xml]<meta property=\"se:subject\">[/] element.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, illegal_subjects))
+		if "z3998:song" in selector:
+			local_css["has_song_style"] = True
 
-		if sorted(nodes) != nodes:
-			messages.append(LintMessage("m-053", "[xml]<meta property=\"se:subject\">[/] elements not in alphabetical order.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+		if "z3998:hymn" in selector:
+			local_css["has_hymn_style"] = True
 
-	elif self.is_se_ebook:
-		messages.append(LintMessage("m-021", "No [xml]<meta property=\"se:subject\">[/] element found.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+		if "z3998:lyrics" in selector:
+			local_css["has_lyrics_style"] = True
 
-	# Check that each <dc:title> has a file-as and title-type, if applicable
-	titles_missing_file_as = []
-	titles_missing_title_type = []
-	titles = self.metadata_dom.xpath("/package/metadata/dc:title")
-	for node in titles:
-		if not self.metadata_dom.xpath(f"/package/metadata/meta[@property='file-as' and @refines='#{node.get_attr('id')}']"):
-			titles_missing_file_as.append(node)
+		if "span.elision" in selector:
+			local_css["has_elision_style"] = True
 
-		# Only check for title-type if there is more than one <dc:title>
-		if len(titles) > 1 and not self.metadata_dom.xpath(f"/package/metadata/meta[@property='title-type' and @refines='#{node.get_attr('id')}']"):
-			titles_missing_title_type.append(node)
+		if "dedication" in selector:
+			local_css["has_dedication_style"] = True
 
-	if titles_missing_file_as:
-		messages.append(LintMessage("m-062", "[xml]<dc:title>[/] missing matching [xml]<meta property=\"file-as\">[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in titles_missing_file_as]))
+		if "epigraph" in selector:
+			local_css["has_epigraph_style"] = True
 
-	if titles_missing_title_type:
-		messages.append(LintMessage("m-068", "[xml]<dc:title>[/] missing matching [xml]<meta property=\"title-type\">[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, [node.to_string() for node in titles_missing_title_type]))
+		if "abbr" in selector and "nowrap" in rules:
+			abbr_with_whitespace.append(selector)
 
-	# Check for CDATA tags
-	if "<![CDATA[" in metadata_xml:
-		messages.append(LintMessage("m-017", "[xml]<!\\[CDATA\\[[/] found. Run [bash]se clean[/] to canonicalize [xml]<!\\[CDATA\\[[/] sections.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+		if regex.search(r"\[\s*xml\s*\|", selector, flags=regex.IGNORECASE) and "@namespace xml \"http://www.w3.org/XML/1998/namespace\";" not in self.local_css:
+			messages.append(LintMessage("c-003", "[css]\\[xml|attr][/] selector in CSS, but no XML namespace declared ([css]@namespace xml \"http://www.w3.org/XML/1998/namespace\";[/]).", se.MESSAGE_TYPE_ERROR, local_css_path))
 
-	# Check that our provided identifier matches the generated identifier
+	messages = messages + _lint_css_checks(self, local_css_path, abbr_with_whitespace)
+
+	missing_files = []
 	if self.is_se_ebook:
-		try:
-			identifier = self.metadata_dom.xpath("/package/metadata/dc:identifier")[0].text
-			if identifier != self.generated_identifier:
-				messages.append(LintMessage("m-023", f"[xml]<dc:identifier>[/] does not match expected: [text]{self.generated_identifier}[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-		except Exception:
-			missing_metadata_elements.append("<dc:identifier>")
+		root_files = os.listdir(self.path)
+		expected_root_files = ["images", "src", "LICENSE.md"]
+		illegal_files = [root_file for root_file in root_files if root_file not in expected_root_files and root_file != "se-lint-ignore.xml"] # se-lint-ignore.xml is optional
+		missing_files = [expected_root_file for expected_root_file in expected_root_files if expected_root_file not in root_files and expected_root_file != "LICENSE.md"] # We add more to this later on. LICENSE.md gets checked later on, so we don't want to add it twice
 
-	# Check if se:name.person.full-name matches their titlepage name
-	duplicate_names = []
-	invalid_refines = []
-	nodes = self.metadata_dom.xpath("/package/metadata/meta[@property='se:name.person.full-name']")
-	for node in nodes:
-		try:
-			refines = node.get_attr("refines").replace("#", "")
+		# If we have illegal files, check if they are tracked in Git.
+		# If they are, then they're still illegal.
+		# If not, ignore them for linting purposes.
+		if illegal_files:
 			try:
-				name = self.metadata_dom.xpath(f"/package/metadata/*[@id={se.easy_xml.escape_xpath(refines)}]")[0].text
-				if name == node.text:
-					duplicate_names.append(name)
+				illegal_files = self.repo.git.ls_files(illegal_files).split("\n")
+				if illegal_files and illegal_files[0] == "":
+					illegal_files = []
 			except Exception:
-				invalid_refines.append(refines)
-		except Exception:
-			invalid_refines.append("<meta property=\"se:name.person.full-name\">")
-
-	if duplicate_names:
-		messages.append(LintMessage("m-024", "[xml]<meta property=\"se:name.person.full-name\">[/] property identical to regular name. If the two are identical the full name [xml]<meta>[/] element must be removed.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, duplicate_names))
-
-	if invalid_refines:
-		messages.append(LintMessage("m-010", "Invalid [xml]refines[/] property.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, invalid_refines))
-
-	# Check for malformed URLs
-	messages = messages + _get_malformed_urls(self.metadata_dom, self.metadata_file_path)
-
-	if self.metadata_dom.xpath("/package/metadata/*[contains(., 'https://id.loc.gov/')]"):
-		messages.append(LintMessage("m-066", "[url]id.loc.gov[/] URI starting with illegal https.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+				# If we can't initialize Git, then just pass through the list of illegal files
+				pass
 
-	if self.metadata_dom.xpath("/package/metadata/*[re:test(., 'id\\.loc\\.gov/authorities/names/[^\\.]+\\.html')]"):
-		messages.append(LintMessage("m-008", "[url]id.loc.gov[/] URI ending with illegal [path].html[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+		for illegal_file in illegal_files:
+			messages.append(LintMessage("f-001", "Illegal file or directory.", se.MESSAGE_TYPE_ERROR, Path(illegal_file)))
 
-	if self.metadata_dom.xpath("/package/metadata/dc:description[text()!='DESCRIPTION' and re:test(., '[^\\.”]$')]"):
-		messages.append(LintMessage("m-055", "[xml]dc:description[/] does not end with a period.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	# Check for repeated punctuation
+	nodes = self.metadata_dom.xpath("/package/metadata/*[re:test(., '[,;]{2,}.{0,20}')]")
+	if nodes:
+		messages.append(LintMessage("t-008", "Repeated punctuation.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path, [node.to_string() for node in nodes]))
 
-	# Does the manifest match the generated manifest?
-	try:
-		manifest = self.metadata_dom.xpath("/package/manifest")[0]
-		if manifest.to_string().replace("\t", "") != self.generate_manifest().to_string().replace("\t", ""):
-			messages.append(LintMessage("m-042", "[xml]<manifest>[/] element does not match expected structure.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-	except Exception:
-		missing_metadata_elements.append("<manifest>")
+	# Set some variables for later
+	transcription_source_count = 0
+	page_scan_source_count = 0
+	other_source_count = 0
+	sources = self.metadata_dom.xpath("/package/metadata/dc:source")
+	for source in sources:
+		if regex.search(r"(gutenberg\.org|wikisource\.org|fadedpage\.com|gutenberg\.net\.au|gutenberg\.ca)", source.inner_text()):
+			transcription_source_count = transcription_source_count + 1
+		elif regex.search(r"(hathitrust\.org|/archive\.org|books\.google\.com|google\.com/books)", source.inner_text()):
+			page_scan_source_count = page_scan_source_count + 1
+		else:
+			other_source_count = other_source_count + 1
 
-	if missing_metadata_elements:
-		messages.append(LintMessage("m-051", "Missing expected element in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path, missing_metadata_elements))
+	ebook_flags["has_multiple_transcriptions"] = transcription_source_count >= 2
+	ebook_flags["has_multiple_page_scans"] = page_scan_source_count >= 2
+	ebook_flags["has_other_sources"] = other_source_count > 0
 
-	# Check for common typos
-	for node in self.metadata_dom.xpath("/package/metadata/dc:description") + self.metadata_dom.xpath("/package/metadata/meta[@property='se:long-description']"):
-		matches = regex.findall(r"(?<!’)\b(and and|the the|if if|of of|or or|as as)\b(?!-)", node.text, flags=regex.IGNORECASE)
-		matches = matches + regex.findall(r"\ba a\b(?!-)", node.text)
-		if matches:
-			messages.append(LintMessage("y-001", "Possible typo: doubled [text]a/the/and/of/or/as/if[/].", se.MESSAGE_TYPE_WARNING, self.metadata_file_path, matches))
+	messages = messages + _lint_metadata_checks(self)
 
-	nodes = self.metadata_dom.xpath("/package/metadata//*[re:test(., '[,;][a-z]', 'i')]")
-	if nodes:
-		messages.append(LintMessage("y-002", "Possible typo: punctuation followed directly by a letter, without a space.", se.MESSAGE_TYPE_WARNING, self.metadata_file_path, [node.to_string() for node in nodes]))
+	# Check for malformed URLs
+	messages = messages + _get_malformed_urls(self.metadata_dom, self.metadata_file_path)
 
 	# Make sure some static files are unchanged
 	if self.is_se_ebook:
@@ -3220,49 +3370,19 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 			filename = (Path(root) / filename).resolve()
 
 			if filename.stem != "LICENSE":
-				url_safe_filename = se.formatting.make_url_safe(filename.stem) + filename.suffix
-				if filename.name != url_safe_filename and not filename.stem.endswith(".source"):
-					files_not_url_safe.append(filename)
+				if filename.stem == "cover.source":
+					ebook_flags["has_cover_source"] = True
+				else:
+					url_safe_filename = se.formatting.make_url_safe(filename.stem) + filename.suffix
+					if filename.name != url_safe_filename:
+						files_not_url_safe.append(filename)
 
 			if "-0" in filename.name:
 				messages.append(LintMessage("f-009", "Illegal leading [text]0[/] in filename.", se.MESSAGE_TYPE_ERROR, filename))
 
-			if filename.stem == "cover.source":
-				has_cover_source = True
-
-			if filename.suffix == ".jpeg":
-				messages.append(LintMessage("f-011", "JPEG files must end in [path].jpg[/].", se.MESSAGE_TYPE_ERROR, filename))
-
-			if filename.suffix == ".tiff":
-				messages.append(LintMessage("f-012", "TIFF files must end in [path].tif[/].", se.MESSAGE_TYPE_ERROR, filename))
-
-			# Run some general tests on images, but skip the cover source since it's an exception to all of these rules
-			if filename.suffix in (".jpg", ".jpeg", ".tif", ".tiff", ".png") and "cover.source" not in filename.name:
-				try:
-					image = Image.open(filename)
-				except UnidentifiedImageError as ex:
-					raise se.InvalidFileException(f"Couldn’t identify image type of [path][link=file://{filename}]{filename.name}[/][/].") from ex
-
-				# Check the source cover image
-				if self.path / "images" / "cover.jpg" == filename:
-					if image.size != (se.COVER_WIDTH, se.COVER_HEIGHT):
-						messages.append(LintMessage("f-017", f"[path][link=file://{self.path / 'images/cover.jpg'}]cover.jpg[/][/] must be exactly {se.COVER_WIDTH} × {se.COVER_HEIGHT}.", se.MESSAGE_TYPE_ERROR, filename))
-
-				# Run some tests on distributable images in ./src/epub/images/
-				# Once we reach Python 3.9 we can use path.is_relative_to() instead of this string comparison
-				if str(filename).startswith(str(self.content_path / "images")):
-					if os.path.getsize(filename) > 1500000: # 1.5MB
-						messages.append(LintMessage("f-016", "Image more than 1.5MB in size.", se.MESSAGE_TYPE_ERROR, filename))
-
-					# Make sure distributable images have reasonable dimensions
-					# We check SVGs later on in a separate check
-					if image.size[0] * image.size[1] > 4000000:
-						messages.append(LintMessage("f-018", "Image greater than 4,000,000 pixels square in dimension.", se.MESSAGE_TYPE_ERROR, filename))
-
-					if filename.suffix == ".png" and not se.images.has_transparency(image):
-						messages.append(LintMessage("f-019", "[path].png[/] file without transparency. Hint: If an image doesn’t have transparency, it should be saved as a [path].jpg[/].", se.MESSAGE_TYPE_ERROR, filename))
-
 			if filename.suffix in BINARY_EXTENSIONS or filename.name == "core.css":
+				if filename.suffix in (".jpg", ".jpeg", ".tif", ".tiff", ".png"):
+					messages = messages + _lint_image_checks(self, filename)
 				continue
 
 			# Read the file and start doing some serious checks!
@@ -3293,7 +3413,6 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 					# For later comparison with cover
 					titlepage_svg_title = svg_dom.xpath("/svg/title/text()", True).replace("The titlepage for ", "") # <title> can appear on any element in SVG, but we only want to check the root one
 
-
 			if filename.suffix == ".xml":
 				xml_dom = self.get_dom(filename)
 
@@ -3303,7 +3422,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 				# Make sure that everything in glossaries are in the rest of the text
 				# We’ll check the files later, and log any errors at the end
 				if filename.name == "glossary-search-key-map.xml":
-					has_glossary_search_key_map = True
+					ebook_flags["has_glossary_search_key_map"] = True
 					# Map the glossary to tuples of the values and whether they’re used (initially false)
 					glossary_usage = list(map(lambda node: (node.get_attr("value"), False), xml_dom.xpath(".//*[@value]")))
 
@@ -3315,14 +3434,12 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 				# First apply the browser default stylesheet
 				dom.apply_css(self.get_file(Path("default")), "default")
 
-				messages = messages + _get_malformed_urls(dom, filename)
-
 				# Apply any CSS files in the DOM
 				for node in dom.xpath("/html/head/link[@rel='stylesheet']"):
 					css_filename = (filename.parent / node.get_attr("href")).resolve()
 					dom.apply_css(self.get_file(css_filename), str(css_filename))
 
-				typos: List[str] = []
+				messages = messages + _get_malformed_urls(dom, filename)
 
 				# Extract ID attributes for later checks
 				id_attrs = id_attrs + dom.xpath("//*[name() != 'section' and name() != 'article' and name() != 'figure' and name() != 'nav']/@id")
@@ -3350,12 +3467,12 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 
 				# Check if this is a frontmatter file, but exclude the titlepage, imprint, and toc
 				if dom.xpath("/html//*[contains(@epub:type, 'frontmatter') and not(descendant-or-self::*[re:test(@epub:type, '\\b(titlepage|imprint|toc)\\b')])]"):
-					has_frontmatter = True
+					ebook_flags["has_frontmatter"] = True
 
 				# Do we have a half title?
 				# Sometimes the half title might not be a section, like in Cane by Jean Toomer
 				if dom.xpath("/html/body//*[contains(@epub:type, 'halftitlepage')]"):
-					has_halftitle = True
+					ebook_flags["has_halftitle"] = True
 
 				# Add new CSS classes to global list
 				if filename.name not in IGNORED_FILENAMES:
@@ -3395,7 +3512,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 					abbr_elements_requiring_css += dom.xpath("/html/body//abbr[re:test(@epub:type, '\\b(se:temperature|se:era|z3998:acronym)\\b')]")
 
 				# Check and log missing glossary keys
-				if has_glossary_search_key_map and filename.name not in IGNORED_FILENAMES:
+				if ebook_flags["has_glossary_search_key_map"] and filename.name not in IGNORED_FILENAMES:
 					source_text = dom.xpath("/html/body")[0].inner_text()
 					if dom.xpath("/html/body//section[contains(@epub:type, 'glossary')]"):
 						nodes = dom.xpath("/html/body//dd[contains(@epub:type, 'glossdef')]")
@@ -3421,34 +3538,37 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 				elif dom.xpath("/html/body/nav[contains(@epub:type, 'loi')]"):
 					special_file = "loi"
 				else:
-					special_file = ""
+					special_file = None
 
 				if special_file in SPECIAL_FILES:
-					messages = messages + _lint_special_file_checks(self, filename, dom, file_contents, ebook_info, special_file)
+					messages = messages + _lint_special_file_checks(self, filename, dom, file_contents, ebook_flags, special_file)
 
 				missing_styles = missing_styles + _update_missing_styles(filename, dom, local_css)
 
-				messages = messages + _lint_css_checks(filename, dom)
+				messages = messages + _lint_xhtml_css_checks(filename, dom, local_css_path)
 
-				messages = messages + _lint_metadata_checks(filename, dom)
+				messages = messages + _lint_xhtml_metadata_checks(self, filename, dom)
 
-				messages = messages + _lint_syntax_checks(self, filename, dom, file_contents, ebook_info, language, section_tree)
+				messages = messages + _lint_xhtml_syntax_checks(self, filename, dom, file_contents, ebook_flags, language, section_tree)
 
-				messages = messages + _lint_typography_checks(filename, dom, file_contents, special_file)
+				(typography_messages) = _lint_xhtml_typography_checks(filename, dom, file_contents, special_file, ebook_flags)
+				if typography_messages:
+					messages = messages + typography_messages
 
-				messages = messages + _lint_xhtml_checks(filename, dom, file_contents)
+				messages = messages + _lint_xhtml_xhtml_checks(filename, dom, file_contents)
 
-				messages = messages + _lint_typo_checks(filename, dom, file_contents, special_file)
+				messages = messages + _lint_xhtml_typo_checks(filename, dom, file_contents, special_file)
 
 	if self.cover_path and cover_svg_title != titlepage_svg_title:
 		messages.append(LintMessage("s-028", f"[path][link=file://{self.cover_path}]{self.cover_path.name}[/][/] and [path][link=file://{self.path / 'images/titlepage.svg'}]titlepage.svg[/][/] [xhtml]<title>[/] elements don’t match.", se.MESSAGE_TYPE_ERROR, self.cover_path))
 
-	if has_frontmatter and not has_halftitle:
+	if ebook_flags["has_frontmatter"] and not ebook_flags["has_halftitle"]:
 		messages.append(LintMessage("s-020", "Frontmatter found, but no half title page. Half title page is required when frontmatter is present.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	if self.is_se_ebook and not has_cover_source:
+	if self.is_se_ebook and not ebook_flags["has_cover_source"]:
 		missing_files.append("images/cover.source.jpg")
 
+	# check for classes used but not in CSS, and classes only used once
 	missing_selectors = []
 	single_use_css_classes = []
 
@@ -3524,7 +3644,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 			if directories_not_url_safe and directories_not_url_safe[0] == "":
 				directories_not_url_safe = []
 
-			# Git doesn't story directories, only files. So the above output will be a list of files within a badly-named dir.
+			# Git doesn't store directories, only files. So the above output will be a list of files within a badly-named dir.
 			# To get the dir name, get the parent of the file that Git outputs.
 			for index, filepath in enumerate(directories_not_url_safe):
 				directories_not_url_safe[index] = str(Path(filepath).parent.name)
@@ -3568,27 +3688,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 			if f"[epub|type~=\"{value}\"]" not in self.local_css:
 				missing_styles.append(element.to_tag_string())
 
-	# Check for some accessibility metadata regarding images
-	has_visual_accessmode = len(self.metadata_dom.xpath("/package/metadata/meta[@property='schema:accessMode' and text() = 'visual']")) > 0
-	has_accessibility_feature_alt = len(self.metadata_dom.xpath("/package/metadata/meta[@property='schema:accessibilityFeature' and text() = 'alternativeText']")) > 0
-	has_wat_role = len(self.metadata_dom.xpath("/package/metadata/meta[(@property='role') and text() = 'wat']")) > 0
-
-	if has_images:
-		if not has_visual_accessmode:
-			messages.append(LintMessage("m-028", "Images found in ebook, but no [attr]schema:accessMode[/] property set to [val]visual[/] in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-
-		if not has_accessibility_feature_alt:
-			messages.append(LintMessage("m-029", "Images found in ebook, but no [attr]schema:accessibilityFeature[/] property set to [val]alternativeText[/] in metadata.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-
-		if not has_wat_role:
-			messages.append(LintMessage("m-040", "Images found in ebook, but no [attr]role[/] property set to [val]wat[/] in metadata for the writer of the alt text.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-
-	if not has_images:
-		if has_visual_accessmode:
-			messages.append(LintMessage("m-038", "[attr]schema:accessMode[/] property set to [val]visual[/] in metadata, but no images in ebook.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
-
-		if has_accessibility_feature_alt:
-			messages.append(LintMessage("m-039", "[attr]schema:accessibilityFeature[/] property set to [val]alternativeText[/] in metadata, but no images in ebook.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
+	messages = messages + _lint_image_metadata_checks(self, ebook_flags["has_images"])
 
 	if missing_styles:
 		messages.append(LintMessage("c-006", f"Semantic found, but missing corresponding style in [path][link=file://{local_css_path}]local.css[/][/].", se.MESSAGE_TYPE_ERROR, local_css_path, sorted(set(missing_styles))))
@@ -3605,7 +3705,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 	if short_story_count and not self.metadata_dom.xpath("//meta[@property='se:subject' and text() = 'Shorts']"):
 		messages.append(LintMessage("m-027", "[val]se:short-story[/] semantic inflection found, but no [val]se:subject[/] with the value of [text]Shorts[/].", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	if has_glossary_search_key_map:
+	if ebook_flags["has_glossary_search_key_map"]:
 		entries = []
 		for glossary_value in glossary_usage:
 			if glossary_value[1] is False:
@@ -3614,34 +3714,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 		if entries:
 			messages.append(LintMessage("m-070", "Glossary entry not found in the text.", se.MESSAGE_TYPE_ERROR, self.content_path / "glossary-search-key-map.xml", entries))
 
-	# Now that we have our lint messages, we filter out ones that we've ignored.
-	if ignored_codes:
-		# Iterate over a copy of messages, so that we can remove from them while iterating.
-		for message in messages[:]:
-			for path, codes in ignored_codes.items():
-				for code in codes:
-					try:
-						# fnmatch.translate() converts shell-style globs into a regex pattern
-						if regex.match(fr"{translate(path)}", str(message.filename.name) if message.filename else "") and message.code == code["code"]:
-							messages.remove(message)
-							code["used"] = True
-
-					except ValueError:
-						# This gets raised if the message has already been removed by a previous rule.
-						# For example, chapter-*.xhtml gets t-001 removed, then subsequently *.xhtml gets t-001 removed.
-						pass
-					except Exception as ex:
-						raise se.InvalidInputException(f"Invalid path in [path][link=file://{lint_ignore_path}]se-lint-ignore.xml[/][/] rule: [path]{path}[/].") from ex
-
-		# Check for unused ignore rules
-		unused_codes: List[str] = []
-		for path, codes in ignored_codes.items():
-			for code in codes:
-				if not code["used"]:
-					unused_codes.append(f"{path}, {code['code']}")
-
-		if unused_codes:
-			messages.append(LintMessage("m-048", f"Unused [path][link=file://{lint_ignore_path}]se-lint-ignore.xml[/][/] rule.", se.MESSAGE_TYPE_ERROR, lint_ignore_path, unused_codes))
+	messages = _lint_process_ignore_file(self, skip_lint_ignore, messages)
 
 	messages = natsorted(messages, key=lambda x: ((str(x.filename.name) if x.filename else "") + " " + x.code), alg=ns.PATH)
 

From c707077dd024212204eec404e260e5cb39a135c5 Mon Sep 17 00:00:00 2001
From: vr8hub <vr_sebooks@letterboxes.org>
Date: Tue, 12 Sep 2023 18:42:34 -0500
Subject: [PATCH 6/9] Add self to _lint_xhtml_typography_checks parameters

---
 se/se_epub_lint.py | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py
index b1b254d9..0087b49f 100644
--- a/se/se_epub_lint.py
+++ b/se/se_epub_lint.py
@@ -2238,7 +2238,7 @@ def _lint_xhtml_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree
 
 	return messages
 
-def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str, ebook_flags: dict) -> list:
+def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: Optional[str], ebook_flags: dict, self) -> list:
 	"""
 	Helper function used in self.lint()
 	Process typography checks on an .xhtml file
@@ -2248,6 +2248,7 @@ def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree,
 	dom: A dom tree to check
 	file_contents: The contents of the file being checked
 	special_file: A string containing the type of special file the current file is, if any
+	self
 
 	OUTPUTS
 	A list of LintMessages representing typography errors found in the file
@@ -2720,7 +2721,7 @@ def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree,
 		if node_text != expected_text:
 			messages.append(LintMessage("t-073", f"Possible transcription error in Greek. Found: [text]{node_text}[/], but expected [text]{expected_text}[/text]. Hint: Use [bash]se unicode-names[/] to see differences in Unicode characters.", se.MESSAGE_TYPE_WARNING, filename))
 
-	return (messages)
+	return messages
 
 def _lint_xhtml_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str) -> list:
 	"""
@@ -3551,9 +3552,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 
 				messages = messages + _lint_xhtml_syntax_checks(self, filename, dom, file_contents, ebook_flags, language, section_tree)
 
-				(typography_messages) = _lint_xhtml_typography_checks(filename, dom, file_contents, special_file, ebook_flags)
-				if typography_messages:
-					messages = messages + typography_messages
+				messages = messages + _lint_xhtml_typography_checks(filename, dom, file_contents, special_file, ebook_flags, self)
 
 				messages = messages + _lint_xhtml_xhtml_checks(filename, dom, file_contents)
 

From 821c2ee20883529c75bf286b1182787f7f8e0089 Mon Sep 17 00:00:00 2001
From: vr8hub <vr_sebooks@letterboxes.org>
Date: Tue, 12 Sep 2023 19:50:57 -0500
Subject: [PATCH 7/9] Move definition of allowed_messages, add it to ignore
 file function parameter list

---
 se/se_epub_lint.py | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py
index 0087b49f..2c2441b1 100644
--- a/se/se_epub_lint.py
+++ b/se/se_epub_lint.py
@@ -3052,7 +3052,7 @@ def _lint_image_metadata_checks(self, has_images: bool) -> list:
 
 	return messages
 
-def _lint_process_ignore_file(self, skip_lint_ignore: bool, messages: list) -> list:
+def _lint_process_ignore_file(self, skip_lint_ignore: bool, messages: list, allowed_messages: list) -> list:
 	# This is a dict with where keys are the path and values are a list of code dicts.
 	# Each code dict has a key "code" which is the actual code, and a key "used" which is a
 	# bool indicating whether or not the code has actually been caught in the linting run.
@@ -3070,9 +3070,6 @@ def _lint_process_ignore_file(self, skip_lint_ignore: bool, messages: list) -> l
 
 		has_illegal_path = False
 
-		if not allowed_messages:
-			allowed_messages = []
-
 		for element in elements:
 			path = element.get_attr("path").strip()
 
@@ -3713,7 +3710,10 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 		if entries:
 			messages.append(LintMessage("m-070", "Glossary entry not found in the text.", se.MESSAGE_TYPE_ERROR, self.content_path / "glossary-search-key-map.xml", entries))
 
-	messages = _lint_process_ignore_file(self, skip_lint_ignore, messages)
+	if not allowed_messages:
+		allowed_messages = []
+
+	messages = _lint_process_ignore_file(self, skip_lint_ignore, messages, allowed_messages)
 
 	messages = natsorted(messages, key=lambda x: ((str(x.filename.name) if x.filename else "") + " " + x.code), alg=ns.PATH)
 

From 415996c4714dbe041485d78d3abeaa4cb27a144a Mon Sep 17 00:00:00 2001
From: vr8hub <vr_sebooks@letterboxes.org>
Date: Tue, 12 Sep 2023 23:22:20 -0500
Subject: [PATCH 8/9] Add and standardize function documentation

---
 se/se_epub_lint.py | 80 +++++++++++++++++++++++++++++++---------------
 1 file changed, 55 insertions(+), 25 deletions(-)

diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py
index 2c2441b1..1c8fb9b8 100644
--- a/se/se_epub_lint.py
+++ b/se/se_epub_lint.py
@@ -594,7 +594,14 @@ def _build_section_tree(self) -> List[EbookSection]:
 def _find_ebook_section(section_id: str, sections: List[EbookSection]) -> Union[EbookSection, None]:
 	"""
 	Find an ebook section in a tree of sections.
-	"""
+
+	INPUTS
+	section_id: The name of the section to be found
+	sections: A list of EbookSections
+
+	OUTPUTS
+	The entry from sections, if any, that contains section_id
+ 	"""
 
 	for section in sections:
 		if section.section_id == section_id:
@@ -607,7 +614,7 @@ def _find_ebook_section(section_id: str, sections: List[EbookSection]) -> Union[
 
 	return None
 
-def _lint_metadata_checks(self):
+def _lint_metadata_checks(self) -> list:
 	"""
 	Process main metadata checks
 
@@ -905,7 +912,7 @@ def _get_malformed_urls(dom: se.easy_xml.EasyXmlTree, filename: Path) -> list:
 	filename: The filename being processed
 
 	OUTPUTS
-	A list of LintMessages objects.
+	A list of LintMessage objects.
 	"""
 
 	messages = []
@@ -1061,7 +1068,7 @@ def _lint_css_checks(self, local_css_path: Path, abbr_with_whitespace: list) ->
 	local_css_rules: Dictionary of the local CSS rules
 
 	OUTPUTS
-	messages
+	A list of LintMessage objects
 	"""
 	messages = []
 
@@ -1518,15 +1525,15 @@ def _lint_special_file_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree
 
 def _lint_xhtml_css_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, local_css_path: Path) -> list:
 	"""
-	Helper function used in self.lint()
 	Process CSS checks on an .xhtml file
 
 	INPUTS
 	filename: The name of the file being checked
-	dom: A dom tree to check
+	dom: The dom tree of the file being checked
+	local_css_path: The path to the local CSS file
 
 	OUTPUTS
-	A list of LintMessages representing CSS errors found in the file
+	A list of LintMessage objects
 	"""
 
 	messages = []
@@ -1614,15 +1621,15 @@ def _lint_xhtml_css_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, local_c
 
 def _lint_xhtml_metadata_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree) -> list:
 	"""
-	Helper function used in self.lint()
 	Process metadata checks on an .xhtml file
 
 	INPUTS
 	filename: The name of the file being checked
-	dom: A dom tree to check
+	dom: The dom of the file being checked
+	self
 
 	OUTPUTS
-	A list of LintMessages representing metadata errors found in the file
+	A list of LintMessage objects
 	"""
 
 	messages = []
@@ -1656,17 +1663,17 @@ def _lint_xhtml_metadata_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTr
 
 def _lint_xhtml_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, ebook_flags: dict, language: str, section_tree: List[EbookSection]) -> list:
 	"""
-	Helper function used in self.lint()
 	Process syntax checks on an .xhtml file
 
 	INPUTS
 	filename: The name of the file being checked
-	dom: A dom tree to check
+	dom: The dom tree of the file being checked
 	file_contents: The contents of the file being checked
-	ebook_flags: A dictionary containing several pieces of information about an ebook
+	ebook_flags: A dictionary containing several flags about an ebook
+	language: The language identified in the metadata
 
 	OUTPUTS
-	A list of LintMessages representing syntax errors found in the file
+	A list of LintMessage objects
 	"""
 
 	messages = []
@@ -2240,18 +2247,18 @@ def _lint_xhtml_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree
 
 def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: Optional[str], ebook_flags: dict, self) -> list:
 	"""
-	Helper function used in self.lint()
 	Process typography checks on an .xhtml file
 
 	INPUTS
 	filename: The name of the file being checked
-	dom: A dom tree to check
+	dom: The dom tree of the file being checked
 	file_contents: The contents of the file being checked
 	special_file: A string containing the type of special file the current file is, if any
+	ebook_flags: A dictionary containing several flags about an ebook
 	self
 
 	OUTPUTS
-	A list of LintMessages representing typography errors found in the file
+	A list of LintMessage objects
 	"""
 
 	messages = []
@@ -2725,16 +2732,15 @@ def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree,
 
 def _lint_xhtml_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str) -> list:
 	"""
-	Helper function used in self.lint()
 	Process XHTML checks on an .xhtml file
 
 	INPUTS
 	filename: The name of the file being checked
-	dom: A dom tree to check
+	dom: The dom tree of the file being checked
 	file_contents: The contents of the file being checked
 
 	OUTPUTS
-	A list of LintMessages representing XHTML errors found in the file
+	A list of LintMessage objects
 	"""
 
 	messages = []
@@ -2793,17 +2799,16 @@ def _lint_xhtml_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_
 
 def _lint_xhtml_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str) -> list:
 	"""
-	Helper function used in self.lint()
 	Process typo checks on an .xhtml file
 
 	INPUTS
 	filename: The name of the file being checked
-	dom: A dom tree to check
+	dom: The dom tree of the file being checked
 	file_contents: The contents of the file being checked
 	special_file: A string containing the type of special file the current file is, if any
 
 	OUTPUTS
-	A list of LintMessages representing possible typos found in the file
+	A list of LintMessage objects
 	"""
 
 	messages = []
@@ -3026,6 +3031,17 @@ def _lint_xhtml_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_c
 	return messages
 
 def _lint_image_metadata_checks(self, has_images: bool) -> list:
+	"""
+	Process metadata checks on an image file
+
+	INPUTS
+	self
+	has_images: Flag indicating whether the ebook has images
+
+	OUTPUTS
+	A list of LintMessage objects
+	"""
+
 	messages = []
 
 	# Check for some accessibility metadata regarding images
@@ -3052,7 +3068,20 @@ def _lint_image_metadata_checks(self, has_images: bool) -> list:
 
 	return messages
 
-def _lint_process_ignore_file(self, skip_lint_ignore: bool, messages: list, allowed_messages: list) -> list:
+def _lint_process_ignore_file(self, skip_lint_ignore: bool, allowed_messages: list, messages: list) -> list:
+	"""
+	Parse a lint ignore file if pressent and if applicable remove its ignored messages
+
+	INPUTS
+	self
+	skip_lint_ignore: Flag indicating whether the lint ignore file should be respected
+	messages: The list of LintMessages that have been generated so far
+	allowed_messages: A list of messages from the lint ignore file to allow for this run
+
+	OUTPUTS
+	A list of LintMessage objects
+	"""
+
 	# This is a dict with where keys are the path and values are a list of code dicts.
 	# Each code dict has a key "code" which is the actual code, and a key "used" which is a
 	# bool indicating whether or not the code has actually been caught in the linting run.
@@ -3136,6 +3165,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 	INPUTS
 	self
 	skip_lint_ignore: Flag indicating whether ignore file should be used
+	allowed_messages: Optional list of messages from lint ignore file to allow this run
 
 	OUTPUTS
 	A list of LintMessage objects.
@@ -3713,7 +3743,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 	if not allowed_messages:
 		allowed_messages = []
 
-	messages = _lint_process_ignore_file(self, skip_lint_ignore, messages, allowed_messages)
+	messages = _lint_process_ignore_file(self, skip_lint_ignore, allowed_messages, messages)
 
 	messages = natsorted(messages, key=lambda x: ((str(x.filename.name) if x.filename else "") + " " + x.code), alg=ns.PATH)
 

From f95e60352cdb70308aed0bf15f14dfc932a2691a Mon Sep 17 00:00:00 2001
From: vr8hub <vr_sebooks@letterboxes.org>
Date: Thu, 26 Oct 2023 23:30:29 -0500
Subject: [PATCH 9/9] Resolve few mypy issues: needed variables not passed to
 functions, extraneous parameter to LintMessage

---
 se/se_epub_lint.py | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py
index 1c8fb9b8..0089cb59 100644
--- a/se/se_epub_lint.py
+++ b/se/se_epub_lint.py
@@ -739,10 +739,6 @@ def _lint_metadata_checks(self) -> list:
 	except Exception:
 		missing_metadata_elements.append("<dc:description>")
 
-	# Check for double spacing
-	if self.metadata_dom.xpath(f"/package/metadata/*[re:test(., '[{se.NO_BREAK_SPACE}{se.HAIR_SPACE} ]{{2,}}')]"):
-		double_spaced_files.append(self.metadata_file_path)
-
 	# Check for illegal Wikipedia URLs
 	nodes = self.metadata_dom.xpath("/package/metadata/*[contains(., '.m.wikipedia.org') or @*[contains(., '.m.wikipedia.org')]]")
 	if nodes:
@@ -1541,7 +1537,7 @@ def _lint_xhtml_css_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, local_c
 	# Do we have any elements that have specified border color?
 	# `transparent` and `none` are allowed values for border-color
 	if dom.xpath("/html/body//*[attribute::*[re:test(local-name(), 'data-css-border.+?-color') and text() != 'transparent' and text != 'none']]"):
-		messages.append(LintMessage("c-004", "Don’t specify border colors, so that reading systems can adjust for night mode.", se.MESSAGE_TYPE_WARNING, local_css_path, matches))
+		messages.append(LintMessage("c-004", "Don’t specify border colors, so that reading systems can adjust for night mode.", se.MESSAGE_TYPE_WARNING, local_css_path))
 
 	# Check that footers have the expected styling
 	# Footers may sometimes be aligned with text-align: center as well as text-align: right
@@ -2245,7 +2241,7 @@ def _lint_xhtml_syntax_checks(self, filename: Path, dom: se.easy_xml.EasyXmlTree
 
 	return messages
 
-def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: Optional[str], ebook_flags: dict, self) -> list:
+def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: Optional[str], ebook_flags: dict, missing_files: list, self) -> tuple:
 	"""
 	Process typography checks on an .xhtml file
 
@@ -2255,10 +2251,11 @@ def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree,
 	file_contents: The contents of the file being checked
 	special_file: A string containing the type of special file the current file is, if any
 	ebook_flags: A dictionary containing several flags about an ebook
+	missing_files: A list of missing files
 	self
 
 	OUTPUTS
-	A list of LintMessage objects
+	tuple
 	"""
 
 	messages = []
@@ -2728,7 +2725,7 @@ def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree,
 		if node_text != expected_text:
 			messages.append(LintMessage("t-073", f"Possible transcription error in Greek. Found: [text]{node_text}[/], but expected [text]{expected_text}[/text]. Hint: Use [bash]se unicode-names[/] to see differences in Unicode characters.", se.MESSAGE_TYPE_WARNING, filename))
 
-	return messages
+	return (messages, missing_files)
 
 def _lint_xhtml_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str) -> list:
 	"""
@@ -2797,7 +2794,7 @@ def _lint_xhtml_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_
 
 	return messages
 
-def _lint_xhtml_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str) -> list:
+def _lint_xhtml_typo_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: Optional[str]) -> list:
 	"""
 	Process typo checks on an .xhtml file
 
@@ -3173,6 +3170,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 
 	local_css_path = self.content_path / "css/local.css"
 	messages: List[LintMessage] = []
+	typography_messages: List[LintMessage] = []
 	cover_svg_title = ""
 	titlepage_svg_title = ""
 	xhtml_css_classes: Dict[str, int] = {}
@@ -3328,6 +3326,9 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 	ebook_flags["has_other_sources"] = other_source_count > 0
 
 	messages = messages + _lint_metadata_checks(self)
+	# Check for double spacing (done here so double_spaced_files doesn't have to be passed to function)
+	if self.metadata_dom.xpath(f"/package/metadata/*[re:test(., '[{se.NO_BREAK_SPACE}{se.HAIR_SPACE} ]{{2,}}')]"):
+		double_spaced_files.append(self.metadata_file_path)
 
 	# Check for malformed URLs
 	messages = messages + _get_malformed_urls(self.metadata_dom, self.metadata_file_path)
@@ -3579,7 +3580,9 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 
 				messages = messages + _lint_xhtml_syntax_checks(self, filename, dom, file_contents, ebook_flags, language, section_tree)
 
-				messages = messages + _lint_xhtml_typography_checks(filename, dom, file_contents, special_file, ebook_flags, self)
+				(typography_messages, missing_files) = _lint_xhtml_typography_checks(filename, dom, file_contents, special_file, ebook_flags, missing_files, self)
+				if typography_messages:
+					messages = messages + typography_messages
 
 				messages = messages + _lint_xhtml_xhtml_checks(filename, dom, file_contents)
 

    s that have child 
     elements
+	nodes = dom.xpath("/html/body//td[contains(@epub:type, 'z3998:persona') and ./p]")
+	if nodes:
+		messages.append(LintMessage("s-083", "[xhtml]
    		[/] element with child [xhtml]
    [/] element.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# We have to additionally filter using the regex library, because often a sentence may begin with an uppercase ACCENTED letter,
-	# and xpath's limited regex library doesn't support Unicode classes
-	nodes = [node for node in nodes if regex.match(r"^[^\p{Letter}0-9]?[\p{Lowercase_Letter}]", node.inner_text())]
+	# Check for incorrect semantics on some common publications
+	nodes = dom.xpath("/html/body//i[@epub:type='se:name.publication.book' and re:test(., '^(The )?(Iliad|Odyssey|Aeneid|Metamorphoses|Beowulf|Divine Comedy|Paradise Lost)$')]")
+	if nodes:
+		messages.append(LintMessage("s-084", "Poem has incorrect semantics.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
+	# Check for `Op. Cit.` in endnotes. `Op. Cit.` means "the previous reference" which doesn't make sense
+	# in popup endnotes. But, if the endnote has a book reference, then allow it as it might be referring to that.
+	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote') and (re:test(., '(Loc\\.|Op\\.) Cit\\.', 'i') or re:test(., '\\bl\\.c\\.\\b', 'i')) and not(.//*[contains(@epub:type, 'se:name.publication')])]")
 	if nodes:
-		messages.append(LintMessage("t-057", "[xhtml]
    [/] starting with lowercase letter. Hint: [xhtml]
    [/] that continues text after a [xhtml]
    [/] requires the [class]continued[/] class; and use [xhtml]
    [/] to split one clause over many lines.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-086", "[text]Op. Cit.[/] or [text]Loc. Cit.[/] in endnote. Hint: [text]Op. Cit.[/] and [text]Loc. Cit.[/] mean [text]the previous reference[/], which usually doesn’t make sense in a popup endnote. Such references should be expanded.", se.MESSAGE_TYPE_WARNING, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for illegal characters
-	matches = regex.findall(fr"({se.UNICODE_BOM}|{se.SHY_HYPHEN})", file_contents)
-	if matches:
-		# Get the keys of a dict in order to create a list without duplicates
-		messages.append(LintMessage("t-058", "Illegal character.", se.MESSAGE_TYPE_ERROR, filename, list({match.encode("unicode_escape").decode().replace("\\u", "U+").upper():None for match in matches}.keys())))
+	# Check for half title pages missing subtitles
+	if ebook_flags["has_subtitle"]:
+		# Make sure we exclude  because that appears in the ToC landmarks
+		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and not(.//*[contains(@epub:type, 'subtitle')])]")
+		if nodes:
+			messages.append(LintMessage("s-087", "Subtitle in metadata, but no subtitle in the half title page.", se.MESSAGE_TYPE_ERROR, self.metadata_file_path))
 
-	# Check for period in cite in endnotes
-	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]//cite[not((./node()[last()])[name() = 'abbr']) and ./following-sibling::*[1][contains(@epub:type, 'backlink')] and re:test(., '^—') and ( (re:test(., '\\.$') and ./following-sibling::node()[re:test(., '^\\s*$')]) or ./following-sibling::node()[re:test(., '^\\.\\s*$')])]")
-	if nodes:
-		messages.append(LintMessage("t-059", "Period at the end of [xhtml][/] element before endnote backlink.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+	else:
+		# Make sure we exclude  because that appears in the ToC landmarks
+		nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage') and .//*[contains(@epub:type, 'subtitle')]]")
+		if nodes:
+			messages.append(LintMessage("s-088", "Subtitle in half title page, but no subtitle in metadata.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Check for Bible verses in old-style notation
-	nodes = dom.xpath("/html/body//*[(name()='p' or name()='cite' or re:test(name(), '^h[1-6]$')) and .//node()[re:test(., '(Genesis|Gen\\.|Exodus|Ex\\.|Leviticus|Lev\\.|Numbers|Num\\.|Deuteronomy|Deut\\.|Joshua|Josh\\.|Judges|Ruth|Kings|Chronicles|Chron\\.|Ezra|Nehemiah|Neh\\.|Esther|Esth\\.|Job|Psalm|Psalms|Ps\\.|Proverbs|Prov\\.|Ecclesiastes|Ecc\\.|Eccl\\.|Solomon|Sol\\.|Isaiah|Is\\.|Isa\\.|Jeremiah|Jer\\.|Lamentations|Lam\\.|Ezekiel|Ez\\.|Ezek\\.|Daniel|Dan\\.|Hosea|Hos\\.|Joel|Amos|Obadiah|Obad\\.|Jonah|Jon\\.|Micah|Mic\\.|Nahum|Nah\\.|Habakkuk|Hab\\.|Zephaniah|Zeph\\.|Haggai|Hag\\.|Zechariah|Zech\\.|Malachi|Mal\\.|Tobit|Judith|Sirach|Baruch|Maccabees|Esdras|Manasses|Matthew|Matt\\.|Mark|Luke|John|Acts|Romans|Rom\\.|Corinthians|Cor\\.|Corinth\\.|Galatians|Gal\\.|Ephesians|Eph\\.|Philippians|Phil\\.|Philipp\\.|Colossians|Col\\.|Coloss\\.|Thessalonians|Thes\\.|Thess\\.|Timothy|Tim\\.|Titus|Tit\\.|Philemon|Phil\\.|Hebrews|Heb\\.|James|Jas\\.|Peter|Pet\\.|Jude|Revelation|Revelations|Rev\\.)\\s*$') and following-sibling::node()[normalize-space(.)][1][contains(@epub:type, 'z3998:roman') and following-sibling::node()[1][re:test(., '^\\s*[\\.,]?\\s+[0-9]')]]]]")
+	# Check for MathML without alttext
+	nodes = dom.xpath("/html/body//m:math[not(@alttext)]")
 	if nodes:
-		messages.append(LintMessage("t-060", "Old style Bible citation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-089", "MathML missing [attr]alttext[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for summary-style bridgeheads not ending in punctuation
-	# Decide if it's a summary-style bridgehead if it contains two or more em dashes. We use Python regex because
-	# xpath can't count the number of occurances of a string.
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'bridgehead') and re:test(., '[^\\.\\!\\?”]”?$')]")
-	nodes = [node for node in nodes if len(regex.findall(r"—", node.to_string())) >= 2]
+	# Check for common errors in language tags
+	# `gr` is often used instead of `el`, `sp` instead of `es`, and `ge` instead of `de` (`ge` is the Georgian geographic region subtag but not a language subtag itself)
+	nodes = dom.xpath("//*[re:test(@xml:lang, '^(gr|sp|ge)$')]")
 	if nodes:
-		messages.append(LintMessage("t-061", "Summary-style bridgehead without ending punctuation.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
-
-	# Check for uppercased am/pm, except for if they have an epub:type, or when uppercased in titles, or in the nav document
-	if not dom.xpath("//nav[contains(@epub:type, 'toc')]"):
-		nodes = dom.xpath("/html/body//abbr[re:test(., '^[AP]\\.M\\.$') and not(@epub:type) and not(./ancestor::*[contains(@epub:type, 'title')])]")
-		if nodes:
-			messages.append(LintMessage("t-062", "Uppercased [text]a.m.[/] and [text]p.m.[/]", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-090", "Invalid language tag.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# Check for some Latinisms that need italics according to the manual. We check ancestor-or-self in case the phrase is set in Roman because it's nested in a parent italic.
-	# Exclude toto followed by ’ since Toto can be a name.
-	# Exclude  whose entire contents is a matched Latinism as we do not italicize those.
-	# Ignore the ToC because we have different rules there
-	nodes = dom.xpath("/html/body//text()[re:test(., '\\b(a (priori|posteriori|fortiori)|(? in poetry not followed by 
    . Ignore spans that are roman as they might be present in poem headers, or spans with xml:lang.
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(poem|verse|song|hymn)')]//p/span[not(@epub:type='z3998:roman') and not(@xml:lang) and following-sibling::*[1][name()='span']]")
 	if nodes:
-		messages.append(LintMessage("t-063", "Non-English confusable phrase set without italics.", se.MESSAGE_TYPE_WARNING, filename, nodes))
-
-	# Check that all names are correctly titlecased. Ignore titles with xml:lang since non-English languages have different titlecasing rules,
-	# and ignore titles that have children elements, because se.titlecase() can't handle XML-like titles right now.
-	# Ignore titles longer than 150 chars, as long titles are likely old-timey super-long titles that should be mostly sentence-cased
-	incorrectly_cased_titles = []
-	for node in dom.xpath("/html/body//*[contains(@epub:type, 'se:name') and not(contains(@epub:type, 'se:name.legal-case')) and not(@xml:lang) and not(./*) and string-length(.) <= 150]"):
-		# Replace any space that is not a hair space with a regular space. This is because in inline titles, we may correctly
-		# have nbsp for example after `St.`, but titlecase will remove that nbsp.
-		if se.formatting.titlecase(node.inner_text()) != regex.sub(fr"[^\S{se.HAIR_SPACE}]+", " ", node.inner_text()):
-			incorrectly_cased_titles.append(node.to_string())
+		messages.append(LintMessage("s-091", "[xhtml][/] not followed by [xhtml]
    [/] in poetry.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	if incorrectly_cased_titles:
-		messages.append(LintMessage("t-064", "Title not correctly titlecased. Hint: Non-English titles should have an [attr]xml:lang[/] attribute as they have different titlecasing rules.", se.MESSAGE_TYPE_WARNING, filename, incorrectly_cased_titles))
+	# Check for nested `` elements
+	nodes = dom.xpath("/html/body//abbr[./abbr]")
+	if nodes:
+		messages.append(LintMessage("s-093", "Nested [xhtml][/] element.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for headers ending in periods
-	# Allow periods at the end of headers that both start and end with double quotes; but headers that only end in double quotes probably don't need periods
-	nodes = dom.xpath("/html/body//*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'z3998:initialism')) and not((./node()[last()])[name() = 'abbr']) and (re:test(., '\\.$') or re:test(., '^[^“].+\\.”$'))]")
+	# Check for language tags transliterated into Latin script with incorrect `-latn` suffix (lowercase l)
+	nodes = dom.xpath("/html/body//*[re:test(@xml:lang, '-latn')]")
 	if nodes:
-		messages.append(LintMessage("t-065", "Header ending in a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-094", "Element has an [attr]xml:lang[/] attribute that incorrectly contains [val]-latn[/] instead of [val]-Latn[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for royal names whose roman numeral is preceded by `the`
-	nodes = dom.xpath("/html/body//span[contains(@epub:type, 'z3998:roman') and ./preceding-sibling::node()[1][re:test(., '[A-Z]\\w+ the $')]]/parent::*")
+	# Check for poetry/verse that has an hgroup but which does not have the correct text alignment
+	nodes = dom.xpath("/html/body//*[re:test(@epub:type, 'z3998:(hymn|poem|song|verse)')]//hgroup/p[@data-css-text-align != 'center']")
 	if nodes:
-		messages.append(LintMessage("t-066", "Regnal ordinal preceded by [text]the[/].", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-095", "[xhtml]
    [/] child of [xhtml]
    [/] in poetry/verse does not have [css]text-align: center;[/].", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for plural graphemes without apostrophes
-	nodes = dom.xpath("/html/body//i[re:test(@epub:type, 'z3998:(grapheme|phoneme|morpheme)') and ./following-sibling::node()[1][starts-with(., 's')]]")
+	# Check that half titles without subtitles have the fulltitle semantic
+	nodes = dom.xpath("/html/body//*[name()!='a' and contains(@epub:type, 'halftitlepage')]/*[re:test(name(), '^h[1-6]$') and not(contains(@epub:type, 'fulltitle'))]")
 	if nodes:
-		messages.append(LintMessage("t-067", "Plural [val]z3998:grapheme[/], [val]z3998:phoneme[/], or [val]z3998:morpheme[/] formed without apostrophe ([text]’[/]).", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-096", "Heading element in half title page missing the [val]fulltitle[/] semantic.", se.MESSAGE_TYPE_ERROR, filename))
 
-	# Find citations not offset by em dashes. We're only interested in s in endnotes that are directly followed by the backlink;
-	# ignore s starting with `Cf.` which means `Compare` and was often used to mean `see`
-	nodes = dom.xpath("/html/body//li[contains(@epub:type, 'endnote')]/p[last()]/cite[following-sibling::node()[normalize-space(.)][1][name() = 'a' and contains(@epub:type, 'backlink')] and re:test(., '^[^—]') and not(re:test(., '^(Cf\\.|\\()'))]")
+	# Check for missing href attributes, sometimes a leftover from PG transcriptions
+	nodes = dom.xpath("/html/body//a[not(@href)]")
 	if nodes:
-		messages.append(LintMessage("t-068", "Citation not offset with em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-097", "[xhtml]a[/] element without [attr]href[/] attribute.", se.MESSAGE_TYPE_ERROR, filename, [node.to_string() for node in nodes]))
 
-	# Check for  in 
    s that start with a leading em dash
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '^—')]")
+	# Check for s with only one child
+	nodes = dom.xpath("/html/body/*[name() = 'section' or name() = 'article']/header[./*[not(name() = 'p') and not(preceding-sibling::* or following-sibling::*)]]")
 	if nodes:
-		messages.append(LintMessage("t-069", "[xhtml][/] in epigraph starting with an em dash.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-098", "[xhtml]
    [/] element with only one child.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
 
-	# Check for  in 
    s that end with a period. Try to exclude s that are last children
-	nodes = dom.xpath("/html/body//*[contains(@epub:type, 'epigraph')]//cite[re:test(., '\\.$') and not( (./node()[last()])[./descendant-or-self::abbr]) ]")
+	# Check for @lang attributes
+	nodes = dom.xpath("//*[@lang]")
 	if nodes:
-		messages.append(LintMessage("t-070", "[xhtml][/] in epigraph ending in a period.", se.MESSAGE_TYPE_WARNING, filename, [node.to_string() for node in nodes]))
+		messages.append(LintMessage("s-102", "[attr]lang[/] attribute detected. Hint: Use [attr]xml:lang[/] instead.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
 
-	# If we have Greek text, try to normalize it
-	for node in dom.xpath("/html/body//*[@xml:lang='grc' or @xml:lang='el']"):
-		node_text = node.inner_text()
-		expected_text = se.typography.normalize_greek(node_text)
-		if node_text != expected_text:
-			messages.append(LintMessage("t-073", f"Possible transcription error in Greek. Found: [text]{node_text}[/], but expected [text]{expected_text}[/text]. Hint: Use [bash]se unicode-names[/] to see differences in Unicode characters.", se.MESSAGE_TYPE_WARNING, filename))
+	# Check for common missing roman semantics for “I”
+	regent_regex = r"(?:Charles|Edward|George|Henry|James|William) I\b"
+	matches = regex.findall(fr"King {regent_regex}", file_contents) + regex.findall(fr"{regent_regex}’s", file_contents)
+	if matches:
+		messages.append(LintMessage("s-103", "Probable missing semantics for a roman I numeral.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
 	return messages
 
-def _lint_xhtml_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str) -> list:
+def _lint_xhtml_typography_checks(filename: Path, dom: se.easy_xml.EasyXmlTree, file_contents: str, special_file: str, ebook_flags: dict) -> list:
 	"""
 	Helper function used in self.lint()
-	Process XHTML checks on an .xhtml file
+	Process typography checks on an .xhtml file
 
 	INPUTS
 	filename: The name of the file being checked
 	dom: A dom tree to check
 	file_contents: The contents of the file being checked
+	special_file: A string containing the type of special file the current file is, if any
 
 	OUTPUTS
-	A list of LintMessages representing XHTML errors found in the file
+	A list of LintMessages representing typography errors found in the file
 	"""
 
 	messages = []
 
-	# Check for uppercase letters in IDs or classes
-	nodes = dom.xpath("//*[re:test(@id, '[A-Z]') or re:test(@class, '[A-Z]') or re:test(@epub:type, '[A-Z]')]")
-	if nodes:
-		messages.append(LintMessage("x-002", "Uppercase in attribute value. Attribute values must be all lowercase.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for punctuation outside quotes. We don't check single quotes because contractions are too common.
+	matches = regex.findall(fr"[\p{{Letter}}]+”[,\.](?!{se.WORD_JOINER} {se.WORD_JOINER}…)", file_contents)
+	if matches:
+		messages.append(LintMessage("t-002", "Comma or period outside of double quote. Generally punctuation goes within single and double quotes.", se.MESSAGE_TYPE_WARNING, filename, matches))
 
-	nodes = dom.xpath("//*[re:test(@id, '^[0-9]+')]")
-	if nodes:
-		messages.append(LintMessage("x-007", "[attr]id[/] attributes starting with a number are illegal XHTML.", se.MESSAGE_TYPE_ERROR, filename, [node.to_tag_string() for node in nodes]))
+	# Check for ldquo not correctly closed
+	# Ignore closing paragraphs, line breaks, and closing cells in case ldquo means "ditto mark"
+	matches = regex.findall(r"“[^‘”]+?“", file_contents)
+	matches = [match for match in matches if "" not in match and "
    		[”\s]+?(.+?)?[”\s]+?(.+?)?