Advogato Status Report
I improved the ATOM feed handling of the aggregator code. Feeds that include only a <summary> tag and no <content> tag are now handled correctly. Also, feeds that include an <updated> tag but no <published> tag are handled correctly. Both these variations, while technically legal according to RFC 4287, seem to be very rare in the real world (not to mention a bit odd). Why include the datestamp of the last update but not the original publication date? Why include the full content of the blog but call it the summary instead of the content? Both these weird-but-legal annoyances were apparently generated by a "django powered" site. Not sure if that means the problem stems from django or just how it was used in this case.
The last few sections of mod_virgule still using hard-coded pages now use templates. This allowed another nasty chunk of hard-coded, site specific markup to be removed from the mod_virgule codebase. It was nice to see the code and binary get smaller for a change! Even though you probably won't notice any huge change in how the site looks, this is a major milestone for mod_virgule. It's finally possible to use it for a new website without having to modify the C source to remove Advogato or robots.net specific HTML. A few more changes are needed to group all the templates together with the CSS files to create an easily themable layout.
Despite the report that Advogato has failed, things continue to look better each month. We've set new records for user logins three months running (at least since I started keeping records six months ago). More than 70 Advogato users have returned to the recentlog via blog aggregation so far. The founding gato himself even stopped by this week to post an article on the new browser wars.
Advogato got a positive mention in a recent comparison of sites for software developers in John Manoogian's blog Inventing What's Next.