My thought, if anyone cares, is pretty much the same as Tim Bray’s thoughts on Atom:
An Atom feed is going to be defined as an XML document, which means that if it’s not well-formed then it’s not Atom.
I say this because I’ve been working with RSS for the last 4 or 5 years, and I’ve too often dealt with RSS as a bunch of text between tags rather than as XML. Regular expressions have been used rather than nice XML toolkits. Liberal feed parsers were probably a good idea for RSS at the time, as RSS was often not XML, well, poorly formed and broken XML perhaps, but not quite XML. I look forward to getting an Atom feed and using an XML library to parse it and do as I wish. I think that’s only going to happen if we ensure that Atom is XML, and I think we do that by creating good tools to create Atom feeds with.
Recently I had to create a syndication feed for a site using ASP, and while I would have preferred an Atom feed, I wasn’t confident enough that the code we had would always produce well formed XML, so we ended up creating an RSS feed instead for now. I mean, sure, it’ll have errors and not validate sometimes, but that’s just how RSS is. I want more from Atom, and once the tools are there to easily create a valid Atom feed every time, then that’s what we’ll do.