A truly funny story: I wrote an rss aggregator and one day I discover some feeds had died without me noticing it. I looked at the feed, it was gone, I look at my aggregate and the headlines were all there?!?!

Since I gather a lot of feeds I couldn't help but noticed that a very large amount isn't wellformed. For example, in xml attributes the & (in urls) is suppose to be &, if you do that however many aggregators won't be able to parse it.

Every other month I wrote little bits of code to address the most annoying issues. 1) if I cant find a <link> or <guide> etc I eventually just gather <a>'s and take the href. 2) if I really cant find a title for the item I had it fail back on whatever is in the <a> since I was gathering those anyway. 3) if I cant even find an <item> I just look for the things that are suppose to go in the <item> 4) if I cant find a proper time stamp ill try parse one out of the url 5) if the urls are relative path complete them.

What was actually going on: The feed was gone, it redirected to the home page. In an attempt to parse the "xml" it eventually resorted to gathering the url and title from the <a>'s and build valid time stamps from the urls.