|RSS Feeds and lj-specific content
||[Feb. 16th, 2005|03:48 pm]
I'm a researcher at Cornell working on a project called Affective Presence, and I'm interested in using LJ posts to build support vector machines that recognize textual descriptions of moods, based, of course, on the moods tag. The logical way to do this is to read in an rss feed, and if the content contains a 'mood' tag, assign it to that particular class. Spot the problem -- LJ RSS feeds don't contain the mood tag.
Now, I don't want to reopen old wounds but is there any *good* reason the LJ RSS feeds don't contain mood and the other LJ-specific tags? I know evan has spec'd out the LJ RSS Module 1.0 and 2.0. Most of the arguments suggested seem to be about avoiding clutter in the namespace, as mentioned by rho. And in the words of sprote, "And guess what, that is exactly what XML namespaces are for. (Remember what the X stands for?)" Or, as benzado put it, "Wasn't RSS designed so that people could define their own namespaces and fields in such a way that they would be ignored by aggregators that don't support them?"
I'm new to lj_dev, so perhaps people can explain what it would take for the LJ rss tags to start including mood (or ljmood, if you want) tags, etc, as per evan's spec?
(And I know that I can just write a bot (with a nice 1 sec delay or something in there, to be a good boy) to look at the pages themselves and look for
, but it'd be much nicer to have the bulk of the parsing done for me already by using an RSS feed.)