Re: A weird question?

From: Dan Brickley <Daniel.Brickley@bristol.ac.uk>

>On Thu, 23 Dec 1999, Sankar Virdhagriswaran wrote:

 >At the W3C spec level, RDF deferred datatyping issues so they could be
>dealt with once and for all across all XML apps by XML Schema. Now we
>(more or less) have this, it is natural to explore a mapping of the
>concepts defined in XML datatypes spec into RDF data graphs. Once some
>mapping has been established, applications should be able to go both
>ways, ie. first we reflect the XML datatype machinery into an
>RDF-processsable representation, _then_ we can (hopefully) reflect
>datatyped XML information into RDF and vice-versa.

The schematron-rdf application (beta at
    http://www.ascc.net/xml/resource/schematron/schematron.html
may do something similar soon.  It takes a schema and an instance and
generates an RDF document labelling each part of the instance according
to which pattern was found.  An XML-schema-to-schematron converter
is definitely on the cards; with that you could generate an RDF document
that shows which parts of a schema apply to which parts of a document.

If anyone is interested in this, a good first stage would be an XSL or
Perl
script to resolve all type references so that we don't need to trace
back
along multiple type references to figure out the facets of a datatype.
I think that might be a useful module for other applications too, for
example
for converting XML-schemas-to-DTDs to allow validation using current
systems.

Rick Jelliffe

Received on Thursday, 23 December 1999 13:48:40 UTC