- From: Alexander J. Vincent <jscript@pacbell.net>
- Date: Sun, 11 Nov 2001 16:19:52 -0500 (EST)
- To: xsl-editors@w3.org
- Message-id: <3BEEEB58.6060002@pacbell.net>
Ladies and gentlemen,
The XSLT 1.0 W3C Recommendation provides a non-normative DTD. It
claims, and I quote:
*NOTE: *This DTD Fragment is not normative because XML 1.0 DTDs do
not support XML Namespaces and thus cannot correctly describe the
allowed structure of an XSLT stylesheet.
Work since the XSLT 1.0 Recommendation was published, particularly in
the Modularization of XHTML W3C Recommendation, shows this may no longer
be accurate. By a modularized version of XHTML and MathML being written
as a module, we are able to cleanly intermix XHTML and MathML in the
same document, and validate it against an experimental XHTML 1.1 +
MathML 2.0 DTD the MathML 2.0 Recommendation specifies. Without much
trouble, we are also able to dictate via an XML entity the namespace of
MathML fragments. The SVG working group has targeted their 1.1 version
to be a modularized version of SVG.
With all due respect to the working group, perhaps it is time to
re-examine the issue of a normative DTD for XSLT.
A modularized DTD for XSLT would intermix very well with the XHTML 1.1
W3C Recommendation. An XSLT 1.0.1 + XHTML 1.1 DTD would allow for XSLT
documents containing XHTML templates, which could therein contain XSLT
instructions. XSLT 1.0 already defines this as a useful procedure;
modularization makes it possible to create a DTD which is scalable in
this sense. Modularization also makes it possible to create a normative
DTD which we can use to enforce validation.
I'd appreciate hearing the working group's opinions on this matter.
Most Respectfully Yours,
Alexander J. Vincent
author, JavaScript Developer's Dictionary (forthcoming from Sams Publishing)
Vallejo, CA
Received on Wednesday, 14 November 2001 04:48:48 UTC