Between raw and cooked II: Are? DTDs are just for validation

[Rick Jelliffe]

 From: Marc.McDonald at Design-Intelligence.com
<Marc.McDonald at Design-Intelligence.com>

>The argument is made that once a document has been validated, there is
>no need to validate it again in a parser.

One aim of XML was that documents should be parsed without DTDs. But it
is useful that constants can be removed to a header.

But a document is a living and organic thing: one of the key insights,
to me, from SGML, is that a document's type can also include its future
allowed values (in a basic domain, namely element structure). So
removing content models from whatever header format is used is fine if
you have a terminal document and you are only interested in the
structure of that particular document, but it is not OK if you assume
that the document may be altered at various stages and that you need to
constrain the structures of all possible documents (of that type) to
some extent (e.g., to prevent duplicated element types, to enforce
controlled vocabularies for element type names, to disallow pathalogical
structures).

The asssumption of editability or non-editability changes everything.
Perhaps this is another manifestation of the great
literature-versus-database  divide.  (SGML's problem was that computer
science theory did not (seem to) have enough to say about how to handle
content modelsy: Fuji-Xerox's Murata Makoto's paper at XTech suggested
that, in fact, there is relevant cs theory which may be very useful in
bringing us forward. )

Rick Jelliffe



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)