xml-dev - Re: Does DTD validation work with namespaces?

["Winchel 'Todd' Vincent, III" <winchel@mindspring.com>]

This message, sent last night, appears not to have made it to the list
because of email problems.  I am reposting it.  My apologies if anyone gets
it twice.

<MarcusCarr>
> Idle musings only...
>
> Given that DTDs are for parsing for validity, it might not be unreasonable
to
> expect that the entire document will be checked against their various
DTDs. Since
> a DOCTYPE declaration is used to signal the start of a new document, it
would be
> ideal if that concept could just be extended, allowing a number of
DOCTYPEs to
> coexist in a single compound document. Each time a new DOCTYPE was
encountered,
> processing of the parent document would be suspended and the documents
processed
> recursively. The existence of a nested document would have no impact on
the
> structure of the parent document - it wouldn't even need to be anticipated
in the
> DTD.
</MarcusCarr>

Marcus:

Some idle musings on my part . . .

Why couldn't you use this same idea, but instead of relying on finding a
DOCTYPE (which would mean a change in XML 1.0), you simply pay attention to
the declared namespace and use the namespace URI to go fetch a DTD.  I may
be way off base on this, but it seems to me, this would not require a change
in DTD syntax.  It would not require a change in Namespace syntax.  It would
simply mean a change in the rule that the resource at the end of the URI is
irrelevant and it would mean extra work (and new behavior rules) for a DTD
validating parser.  I mentioned in an earlier post the difference between
(1) non-validating parser (2) validating parser (which everyone understands)
and (3) a proposed DTD-Namespace-aware validating parser.

Again, I could be completely wrong about the analysis here.  But, your ideas
match very closely what I think a solution could be.

Any feedback would be welcome.

Thanks,

Todd