Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Prefer httpwg.org over rfc-editor.org for IETF specs when possible #933

Closed
tidoust opened this issue Apr 21, 2023 · 3 comments
Closed

Prefer httpwg.org over rfc-editor.org for IETF specs when possible #933

tidoust opened this issue Apr 21, 2023 · 3 comments

Comments

@tidoust
Copy link
Member

tidoust commented Apr 21, 2023

Via #920 (comment), where @sideshowbarker argues that httpwg.org would be better than rfc-editor.org:

[[
I propose considering to revert this change.

I don’t have any objection in principle for this change to be made based on weighing all the pros and cons — but I want to point out a statement in the issue description that appears to not be completely accurate:

the rfc-editor.org URLs are just as readable

I’m not completely sure I’m correctly understanding what’s meant by “URLs are just as readable”, but assuming what’s meant is that the published rfc-editor.org specs have readability for users equal to the readability of the httpwg.org versions, then I think that may not really be correct.

For example, compare these two:

Comparing those based on readability, I guess what jumps out right away is that the Accept-Language production in the HTTPWG version has syntax highlighting while the rfc-editor.org version doesn’t. That seems like a significant difference.

But something else that to me at least is even more important for readability and usability — but that doesn’t jump out right away — is: the ABNF blocks for the productions in the HTTPWG version also have hyperlinks to the related section for that production.

For example, if you click on the word weight in the HTTPWG version’s ABNF block, it jumps you to related section at for the weight production at https://httpwg.org/specs/rfc9110.html#quality.values.

To me at least, that readability/usability difference between the HTTPWG versions and the rfc-editor.org versions is quite significant — given that it seems like a very-common information need/task for users of those specs is to be able to understand those productions and trace the references easily.

So I don’t object to this change being made on weighing all the factors and deciding that there are other factors that outweigh the readability deficiency of the rfc-editor.org versions — but I do object to this change being made on the basis of any claim that the rfc-editor.org URLs are just as readable.
]]

@tidoust
Copy link
Member Author

tidoust commented Apr 21, 2023

To revert the change that reverts the use of httpwg.org URLs for recent IETF specs, I think we need to:

  1. Make sure that Specref adopts the same convention, not to end up with links targeting httpwg.org versions while references target rfc-editor.org versions. That's the main blocking factor from a browser-specs perspective.
  2. Update specs that currently expect fragment IDs defined in the rfc-editor.org versions. There should be a handful of them.

tidoust added a commit to tidoust/specref that referenced this issue Nov 23, 2023
The list was no longer complete as the IETF HTTP WG published a few other RFCs.

Also, the https://httpwg.org/specs page is not such a good source for the list
because it fails to list a few specs, and also contains one RFC that is
actually not published under https://httpwg.org/specs. The comment explains how
to extract the list from the underlying GitHub repository instead.

In terms of updates, this will swap the URL of a few additional RFC entries
from `www.rfc-editor.org` to `httpwg.org/specs`, notably RFC9110, RFC9111,
RFC9112 and RFC9113. That's intended, the `httpwg.org/specs` versions are more
readable, as discussed in: w3c/browser-specs#933
tidoust added a commit to tidoust/specref that referenced this issue Nov 23, 2023
The list was no longer complete as the IETF HTTP WG published a few other RFCs.

Also, the https://httpwg.org/specs page is not such a good source for the list
because it fails to list a few specs, and also contains one RFC that is
actually not published under https://httpwg.org/specs. The comment explains how
to extract the list from the underlying GitHub repository instead.

In terms of updates, this will swap the URL of a few additional RFC entries
from `www.rfc-editor.org` to `httpwg.org/specs`, notably RFC9110, RFC9111,
RFC9112 and RFC9113. That's intended, the `httpwg.org/specs` versions are more
readable, as discussed in: w3c/browser-specs#933
tidoust added a commit that referenced this issue Nov 23, 2023
Take 3 :)

PR #1135 actually had a couple of issues that made the code essentially useless
because it only ran on a handful of IETF specs:
- the code favored info from Specref over info from IETF
- the code only really applied to drafts due to a buggy RegExp

Fixing these problems yielded a new issue: the assumption that HTTP WG specs
are always available under `httpwg.org` turns out to be wrong. Also, there are
other specs that are not published by the HTTP WG but that still have an
`httpwg.org` version. The code now looks at the actual list of specs in the
underlying GitHub repository: https://github.com/httpwg/httpwg.github.io.

As a result, the nightly URL of all IETF specs that have an `httpwg.org`
version now targets that version, implementing the suggestion in #933 (see
that issue for the list of affected specs). A companion PR was sent to Specref
to implement a similar switch there:
tobie/specref#766

The code also looks at the obsolescence data in datatracker and sets the
`standing` and `obsoletedBy` properties accordingly. This fixes #327.
dontcallmedom pushed a commit to tobie/specref that referenced this issue Nov 23, 2023
The list was no longer complete as the IETF HTTP WG published a few other RFCs.

Also, the https://httpwg.org/specs page is not such a good source for the list
because it fails to list a few specs, and also contains one RFC that is
actually not published under https://httpwg.org/specs. The comment explains how
to extract the list from the underlying GitHub repository instead.

In terms of updates, this will swap the URL of a few additional RFC entries
from `www.rfc-editor.org` to `httpwg.org/specs`, notably RFC9110, RFC9111,
RFC9112 and RFC9113. That's intended, the `httpwg.org/specs` versions are more
readable, as discussed in: w3c/browser-specs#933
@tidoust tidoust closed this as completed Nov 24, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant