As a locker owner, i want to browse the links in my locker
- Make sure current collection works with twitter and facebook sourced links
- I can view my links
- I can search my links by
- URL
- Page Title (when available)
- Original context (tweet/share, sender)
- Page contents (future/bonus)
- If a link is shortened, I can view it unshortened
- The favicon for a link is displayed
- My links are organized chronologically, grouped by date
- I can click on a link and it will open in a new window
- I can see what service and person my link came from
- This might be more than one source
Future (could all be done in either a detail view, or expanded link item within the list):
- Links that contain pictures have a thumbnail preview
- Links that contain videos (youtube, vimeo, dailymotion) have a play button to allow playing within the page
- Links that contain music (mp3) have a play button to allow playing within the page
Note the following mockup is here to serve as a guide for the functionality & layout. The treatment (styling) is not necessary for acceptance.
- Tom
- Jason
- Link events from facebook/twitter
- how to handle a lot of links loading efficiently (paging)
- possible traffic surges (HEAD checking, and/or favicon bursts)
- spammy links (games/apps from Facebook)
- HTML matches Wireframes (very roughly)
- Browser support
- Chrome (latest stable)
- Firefox (latest stable)
- Safari (latest stable)
- Selenium Test(s)?
- Backend Test(s)
- User Test with working prototype with 5 social web users
- Requires clucene 0.1.1
- Depends on service manifest / me.json cleanup
- Search is different/deeper than global would be (highlighting matching snippet, rank sorting), can/should be merged with search collection at some point
- Faceted searching, specifically be able to search by author, title, original text, page content
- Content-based dedup, detect various urls for one page (some sites have this, some of this is "wrapper" content aggregation sites, can be a real PITA
- Authenticated pages (facebook), how should we handle these now and ideally?
- Login pages (nytimes), have to detect actual url vs. the redirected login page
- Re-crawling, when should we re-fetch a link? Either overall or if we got an error?
- Auto-fetching url impact? (info leak?)
- Charset/encoding issues (non-utf8), titles and page contents can have this issue, can be involved
- Optimization of search result paging (currently clucene returns everything, need numeric field support in it)
- Events/Types use different names and are in different formats (need to clean this up someday, cost 2-4hrs of pain)
- Painful dealing with me.json data mess (should almost eliminate me.json except pointer back to manifest or any local state)