0xDECAFBAD

It's all spinning wheels and self-doubt until the first pot of coffee.

Atom Format Versioning After Last Call?

An interesting sidebar is that the new Atom format doesn't have a version attribute and I understand it will sit in the same namespace. I tried a few rss clients with the new format and many crashed.
Source: Randy Charles Morin, in comments on "Format Wars, Episode 3"

I've been pretty much tuned out of the development of Atom, due to a complete lack of time and an assumption that it'll all make sense once it's all been hashed out. My ears have perked up with the "last call" noises coming from that direction, so I'm getting the sense that I should start paying attention again. (And I might need to think about updating my book!)

From what little I've seen in my glimpses at drafts issued since this standards process started, there have of course been changes--some small tweaks like changing the <feed>-level <tagline/> to <subtitle/>, and some subtle tweaks in the content model. I've yet to see anyone chronicle or report on these changes, but I assume that's because it's such a moving target not yet worth tracking. That, or I've just missed something.

But, changes these are, and incompatibilities they make. So, the above comment concerns me--I don't want to read another article about mythical compatibility, this time about Atom. I figured versioning by namespaces was a good thing, versus some arbitrary "version" attribute on the root element of a feed--but in any case, versioning is a must. Unfortunately due to my tune-out, I have no idea whether the above assertion is true, a misunderstanding, or where to find the context surrounding the issue.

Rather than jump onto the mailing list or something and ask what's likely a well-hashed issue, I'll restrain my ignorance to my own blog, then, while I continue googling and idly browsing through list archives as time permits. But, just in case someone knows all about this and feels like tossing me a bone, I'm posting this entry.

Update #1: The following bit I'd missed on first glance in the latest Atom draft makes me think that this may not end up being a problem (anyone feel free to correct me):

This specification uses XML Namespaces [W3C.REC-xml-names-19990114] to uniquely identify XML element names. It uses the following namespace prefix for the indicated namespace URI;

"atom": http://purl.org/atom/ns#draft-ietf-atompub-format-08

[rfc.comment.1: This paragraph to be removed by the RFC Editor. The namespace here is a temporary one and will be changed when the IESG approves this document as a standard. At that time, the namespace will be drawn from W3C URI space. The choice of that namespace will be coordinated between the IETF and W3C through their respective liaisons.]

Archived Comments

  • You found the correct area of the spec. Also related: http://www.intertwingly.net/wiki/pie/PaceRemoveVersionAttr http://www.intertwingly.net/blog/2005/04/21/Atom-Format-Last-Call
  • I don't understand this version by namespace thingy. It would seem to me that versioning by namespace is gonna mean recoding every client for every version of the spec. Unfortunately, like yourself, I dropped out of the Atom noise list long ago.