Most are minor but Atom feeds explicitly state content type. This could be a security issue with RSS since there is no reliable way to guess between plain text or escaped HTML.

http://en.wikipedia.org/wiki/Atom_(standard)#Content_model