Re: PI target names unscoped -- why?

James Clark wrote:
> 
> David Brownell wrote:
> 
> > I'd still prefer to see at least a recognition of the problem that
> > the draft's approach only aims for the "trivial subset" of XML (sans
> > validation of DTD based declarations).  Similarly, to see a statement
> > about replacing the existing declaration mechanism with TBD schema
> > mechanisms to regain the DTD based error checking that appears to be
> > getting discarded.
> 
> I think that's too strong.

Actually so do I -- I meant to clarify this.

I was going by something Dan Connolly posted on XML-Dev that seemed to
read that way ... which I actually didn't agree with.  (I've not had
a chance to catch up to quite everything there, as I was off-net for a
week; and I might have misread it.)  If it were actually intended that
namespaces not work with validation, it needs to be explicitly stated.
(And then folk need to recover from the resulting firestorm ... :-)

What you said below, I certainly agree with.

- Dave


> The goal of the namespace draft is to be able to create documents that
> draw from multiple vocabularies.  It allows this to be done in a way
> that is compatible with XML 1.0 validation, that is, I can create a DTD
> that uses element types and attributes from multiple vocabularies and I
> can create documents conforming to that DTD that are valid and make use
> of all the element types and attributes in the DTD.
> 
> What it doesn't do is allow me to create a DTD independently for each
> vocabulary and automatically combine those DTDs so that I can validate a
> document that uses a combination of vocabularies without having to
> create a DTD for that specific combination.  I think the spec should
> say  this and also say that we're working on a schema mechanism that
> will support this.
> 
> James

Received on Monday, 17 August 1998 11:25:53 UTC