On Thu, 11 Nov 2004 18:22:52 -0500, Sam Ruby <[EMAIL PROTECTED]> wrote:

> Whether there every have been a need for such things in syndication land
> (or, alternately, whether we can confidently state that they will never
> be a need for such a thing) is a separate question.

Personally I don't see a real demand need for mU-style mandatory
extensions in general in Atom, but they may be the best option for
versioning.

In the current syndication landscape, RSS 2.0 extensions are few and
far between. The "Missing isn't broken" [1] philosophy works well for
RSS 1.0 extensions - applications can understand as much or as little
of them as they want.

I'll pass on lessons learnt from the Userland approach to versioning,
but note that RSS 0.9 and RSS 1.0 are compatible at the RDF level, and
their versioning/interpretation could be dealt with in a globally
meaningful manner in a handful of OWL statements.

What may be useful from the RSS 1.0 approach is that multiple versions
of the vocabulary can be used together without any need for
redefinition. Consider this:

<atom:feed xmlns:atom="http://atom.now.org/";
xmlns:atom11="http://atom.future.org/";>
    <atom:updated>2004-11-11</atom:updated> 
    <atom11:modified>2004-11-12</atom11:modified>
...
</atom:feed>
  
Cheers,
Danny.

[1] http://rdfweb.org/mt/foaflog/archives/000047.html

-- 

http://dannyayers.com

Reply via email to