AmphetaDesk has some great documentation on finding RSS channels, including a bookmarklets to do RSS discovery, and a visual guide for icons which commonly imply RSS feeds.

This dovetails in nicely with the work that Mark has done on his ultra liberal RSS locator. The idea is to apply a number of heurisitcs for trying to discover a websites RSS feed, when you only know its URL. An idea codified into python by rssfinder.py. (apparently Straw, a python RSS aggregator, uses rssfinder.py, but bad interactions between Debian’s Python2.1 packages [required by xchat] and my source installed 2.2 means I haven’t played with it yet)

Finally, Mnot (whose RSS parser I use for weblog watch) has put together a gentle but informative RSS tutorial.