Saturday, January 25, 2003

Parsing RSS At All Costs 2:45:08 AM • • comment [] • Google It! Mark Pilgrim writes about liberally parsing non-well-formed RSS feeds. We've discussed the issue before (see, for example, my own shaming them into submission article, as well as Mark's own comments on his liberal parser on his site). Reading the comments people made is quite interesting, BTW. As I've said in that article, an aggregator cannot simply ignore broken feeds. Once it does, it stops answering users needs, and is destined to be ditched. It helps to review what happened during those five months. The RSS parser I wrote for Aggie RC5 -- soon to be release -- has indeed been complaining for quite some time about broken feeds. I've personally sent quite a few emails to RSS authors whose feed Aggie complained about, and in all cases I got excellent results. I'm certain others has as well. The quality of RSS feeds has been improving over the last few months; largely, I think, because people who noticed complained. To sum it all up, I think the approach we preached certainly paid off. On one hand, we tolerated broken feeds and gave our users tools that satisfied their needs -- reading RSS feeds. On the other hand, we raised a warning flag that helped people improve as they went along. Unlike HTML, the RSS compliance scene is improving over time, all because we took the middle ground.

© Copyright 2003 Ziv Caspi.