Lists Home |
Date Index |
From: "Dare Obasanjo" <firstname.lastname@example.org>
> I don't believe anyone can honestly claim that an XML based syntax is simpler
> than DTDs especially when one compares the complex mess that is W3C XML
> Schemas to DTDs although I'll be the first to admit that it is
> _more_consistent_. However, consistency and simplicity are not the same thing
> by any stretch of the imagination.
The aspect of "consistency" echoes James Clark's calling DTD "a bit of a mess"
(or whatever his words were, sorry James).
That DTDs include public entity declarations (e.g. for standard entity sets and
boilerplate text) and notation declarations makes complete sense if the idea is to
bundle together the various things that make a "document type".
That DTDs should also include declarations for instance-based entities can
be justified because it is all the same mechanism as standard public entities,
and because entity references can go into attribute values.
I still think there is a difference between documents between entity references
and links: it is not only the "inclusion" aspect, it also that a general parseable
entity referred to in an instance forms part-of the document. I don't think
the other linking mechanisms such as XLink (and xml Inclusions?) provide
a way to warrant that if the entity cannot be resolved then the document
It is interesting to speculate that if "document type" (or "document architecture",
Steve) is in fact more useful (in some significant application domain) than "namespace",
and the warrant that retrieving and parseing the link at the end of a resource
is important, then the attempts to get rid of DTDs might not succeed.
I wonder if anyone involved in the latest XInclude draft could comment
on what the error status is when an XInclude resource cannot be included?