Re: Normative Reference to SGML?

Paul Prescod wrote:
> 
> len bullard wrote:
> > I don't think so.  That would be true only for an instance.
> > It says nothing about the schemata.  That is probably the
> > point.
> 
> A document is composed of an instance and a DTD. For the document to
> conform to SGML, both must conform to SGML.

That is interesting but if it were the case, why would anyone 
have problems with a clear normative reference?

This looks like a legalistic approach that doesn't adequately 
express the intent.  The problem is in the definition of 
"the document".  It's a fascinating problem.  By comparison 
to the way this has been worked out by the VRML Consortium, one 
could infer the W3C behavior is mercurial.  Let's see how 
the ERB works it out.

But without a normative reference, and given the weak position 
of the W3C with respect to member conformance, I'd say XML 
has less than half the same chance HTML had of remaining a 
stable standard.  For a long lifecycle format, that is a failure 
for this process and unfortunately for this membership, the 
acid test of whether it is worth moving to XML from SGML 
except for trivial information.  Without the safeguards 
which only conformance guarantees, it isn't worth the expense.

len

Received on Monday, 30 June 1997 07:07:12 UTC