Lists Home |
Date Index |
It doesn't matter where it is spec'd;
it matters if it adds compelling value and
we all want it bad enough. RELAX NG + AFNG
has a lot of bang for the buck. Push the vendors.
Pushing the W3C isn't necessarily the only way.
DTD modularization is clunky, that is so. Namespaces
make it clunkier because namespaces are not content.
They are hidden properties the system adds to enable
system wide name scoping, a control imposed into content
to support a system functionality.
DTDs are for structuring content, not system-wide
name scoping. DTDs in the core made it possible
to define application languages such as XML Schema
without defining them in terms of themselves. XML
Schema and RELAX NG are content as far as a DTD is
The uglyness of what is at the end of a namespace is that names and
addresses are conflated in the WWW application as URIs. It's useful but a
source of never ending conundrums. So now we put something
at the end, but to do that, we end up accepting yet another
system language, RDF, inside XHTML, and inside RDF, URIs.
It's like a snake consuming itself. Maybe it feels good
to the snake but it is ugly to contemplate.
From: Michael Brennan [mailto:Michael_Brennan@Allegis.com]
Well, I can certainly the importance of supporting DTDs because of the
installed base of users. But DTD-based modularization has always seemed
inordinately clunky and unweildy to me -- especially when employed with
grammars that use namespaces. Maybe its just me, but I've never been able to
get used to this.
I'd love to see the W3C embrace a more robust approach to doing this sort of
thing, such as the RELAX NG-based approach to XHTML modularization that
James Clark has demonstrated .