[
Lists Home |
Date Index |
Thread Index
]
On Tue, Jan 13, 2004 at 01:31:08PM -0800, Joshua Allen wrote:
> > I view the XML Information Set spec as defining a vocabulary for
> > other specs to use. It's not a serialization spec. The extent
>
> Well, it certainly wouldn't be a serialization spec. Many people,
> however, treat it as a data model spec. I agree with your point that
> this was not necessarily the original intent, though. OTOH, the
> treatment of XML as a data model is now well-ingrained in W3C specs.
> XSLT 1.0 and XSD both rely on the infoset-as-data model, and XQuery
> defines a new take on XML data model.
The XSL and XQuery Working Groups have defined a data model together
that is intended for XPath 2, which is in turn used by both XSLT 2
and XML Query 1.0; there is no intent that the XPath data model be
suitable for all uses of XML, and I don't personally believe it
makes sense to try and define such a data model for all uses.
> You can't cause the problems to vanish by declaring that infoset is
> not a data model, unless you also get rid of XSLT and XSD.
I'm not trying to say there are no problems. We live in a complex
world full of challenges and problems :-) I suspect that what
we're seeing is a move from a syntactic approach to XML, in
which (for example) XML Schema validation operates on sets of
lexical forms, into a world of data models, in which, for
example, XSLT stylesheets transform sets of values, and XML Queries
operate over sets of forests of trees of values, which may or
may not ever have experienced life with pointy brackets.
Liam
--
Liam Quin, W3C XML Activity Lead, http://www.w3.org/People/Quin/
http://www.holoweb.net/~liam/
|
