Lists Home |
Date Index |
- From: Rick JELLIFFE <firstname.lastname@example.org>
- Date: Thu, 03 Aug 2000 09:32:46 +0800
Aaron Skonnard wrote:
> Once XSD and XInclude are done, DTDs are dead weight for the Infoset. It's
> really about the fundamental abstractions that we can infer from XML 1.0 +
> Namespaces in light of our current direction.
I don't agree.
XInclude is no replacement for entity referneces in attribute values or
in text-only data content. Neither does it adequately substitute for
the ISO public entity sets for special characters: HTML advocates have
made it quite clear during the XHTML discussions that the idea that
entity references will not go away quickly. (XInclude is a bad idea, if
considered alone, because it will fragment XML into XML+XInclude and
XML: it would be better to bundle it as part of namespaces and XBase.)
A DTD is carted around with a document when the document is considered
by its originater to have some life. The DTD represents constraints on
the future of the document not just its present.
By not including DTDs as part of the infoset or DOM, the W3C groups are
making the simplification that the document has reached its terminal
state: if it is transformed, it is transformed into another structure.
This agrees with XML's premise that it is a resolved document format,
rather than a format for documents in progress. SGML gives many
features suitable for documents-in-progress: it leaves XML for dead in
this because XML was designed for distribution not authoring.
XML Schemas may even lead to a resurgence of the use of DTDs: this is
because once schema software is in place as a layer, the particular
syntax becomes less important. (Of course, where there are features that
cannot be supported, that is different. No magic.)