Re: Moving on (was Re: URIs quack like a duck)

"Simon St.Laurent" wrote:

> I'd suggest that:
> 1) For now, we leave Namespaces in XML alone - effectively, accepting the
> comparison as string literals already there.  This allows relative URIs
> while not specifying how they should be processed.
>
> 2) If changes to that status quo are needed, the W3C should start a new
> working group devoted to namespace issues, which uses this mailing list as
> a public forum for reviewing comments.  That working group can issue
> 'Namespaces in XML 2.0' in whatever form it finds appropriate, with
> concessions (or not) to backward compatibility.
>
> Does this sound like a plausible solution to our current quandary?

Almost.  The problem I see is that if we follow (1), which at this point I more
or less agree with, then there's a nasty inconsistency between the namespace
spec and the definition of expanded names in the XPath spec.  The working group
you speak of would have to look at, and propose revisions to, all specs that
are impacted by the string-literal interpretation of namespace names, not just
the namespace spec itself.  XPath is an instance but  not the only one.   XBase
is another.

Paul Abrahams

Received on Monday, 29 May 2000 22:02:53 UTC