* Elliotte Harold <[EMAIL PROTECTED]> [2006-11-18 21:00]: > Mark Nottingham wrote: > >Atom has a namespace; that can be use to introduce new > >versions of the format. > > No, no, and no. We've been down this road before in other > specs, and the community wisdom is that you do not rev the > namespace just to introduce a new version. Doing so breaks > a huge amount of the existing processing chain for an > application. > > It is, of course, possible to introduce new elements in other > namespaces; and the Atom spec does a much better than average > job of preparing for this. Indeed there are already many > interesting extensions. > > However, even if fundamental flaws were found in Atom 1.0 that > required revisions and an Atom 1.1 or 2.0, the namespace should > and almost certainly woulds remain the same.
Yes and no. If future changes can be made in a backward-compatible fashion, they will go into a spec that recycles the same namespace. Existing implementations can just ignore the differences. If they cannot, they will go into a spec which revs the namespace, because trying to process documents conforming to the new spec with existing implementations can’t work. In neither case is there a need to identify the version explicitly. This is why Atom does not have a version attribute. All that said, the goal during development was always that Atom would ship when it was done and be stable thereafter. It is now set in stone for the forseeable future. No further development of the core format is expected. (Instead, the spec includes a number of extension points that can be used to repurpose the format to new use cases.) This is another reason why Atom does not have a version attribute. Regards, -- Aristotle Pagaltzis // <http://plasmasturm.org/>
