Does it matter whether Atom IDs are URIs or simple strings? Does it matter whether the creation date of Atom feed entries is published? How about the last modification date? Does it matter whether all kinds of links go into the a "link" element, or into various elements of different names? Interestingly enough, after intense debate on these and other issues, the conclusion (reached by some, though not all) is often that it doesn't make a whole lot of difference. That raises the question of whether we're all just wasting our time on this project--whether somebody should just write up a specification with some sort of solution for all the things about RSS that have caused issues, and we should all just adopt it.

I think the answer to that question is that there are a lot of ways to do things wrong, and a lot of ways to do things right. RSS did a lot of things right, and can be forgiven for the things that it didn't quite get right. RSS was the pioneer, after all. The challenge for Atom isn't necessarily to find the right way to do everything, just to find a right way to do things. As long as we don't do anything stupid, and as long as we do a good job of writing the specification, we should be fine.

...until we discover that we missed some important points after all. Then we'll get to work on the next version of Atom.

"Perfect is the worst enemy of good enough."