RE: Using XMLSchema-instance attributes in RDF/XML Syntax

> How about using a post-schema validation infoset, or whatever it is
> called.  Wouldn't that solve the problems?

Yes, and create others.

I'm very tempted to weigh in with support for this, because I like XML
Schema. However it would create a few problems:

1. It would add another layer between XML and RDF/XML, which would require
XML Schema validating parsers to be used rather than any XML parser, with
ensuing difficulties in implementation.

2. It would tie RDF/XML to XML Schema.
	a. There are several schema systems for XML, XML Schema isn't even the only
W3C-endorsed one, fans of the others are going to be penalised when they try
to use RDF/XML.
	b. Future development of both RDF/XML (and hence to some extent of RDF) and
XML Schema will have to keep an eye on what each other is doing so as not to
damage each others standards.

3. The usefulness of XML Schema for RDF/XML is relatively limited unless you
either have a very complicated schema, or else a limited "dialect" of
RDF/XML which adds additional limits on what is allowed on top of the rules
of RDF/XML.

4. If OWL makes use of the Schema datatypes then the usefulness of XML
Schema will become more limited still.

Received on Friday, 23 August 2002 08:25:49 UTC