James Holderness is reporting that most RSS readers are still struggling with XML namespace use in Atom. James reported 1/3 of aggregators can't hand proper Atom namespace prefixes and 2/3 can't handler proper XHTML namespace prefixes. So after 2 years, people still can't Atom right. Don't you wish they had made some more simple? And they thought escaping titles was a problem.
http://www.xn--8ws00zhy3a.com/blog/2007/11/atom-namespace-problems
These feeds that are failing are labratory-created feeds built to stress-test aggregators, not the millions of ordinary, working feeds that are being produced by thousands of Atom-enabled everyday apps in use by people.