RE: New W3C XML Schema facet ? [was: RDDL Schematron Schema]

[Jonathan Borden <jborden@mediaone.net>]

Eric van der Vlist wrote:
>
> Jonathan Borden wrote:
> >
> > I've added a Schematron schema to RDDL (i probably should make
> y'all look at
> > http://www.rddl.org/ to find it but its at:
> http://www.rddl.org/rddl.sch).
> > Schematron is interesting in that it allows constraints which
> are otherwise
> > very difficult to express in for example RELAX or TREX. For example, an
> > xlink:arcrole and xlink:role must have a value of a URI
> reference, but not a
> > relative URI reference. (This could also be expressed in an XML Schema
> > pattern).
>
> This doesn't seem obvious to fully test since it's depending on the URI
> scheme.
>
> For instance, shouldn't "file:myrole.xml" be considered as a relative
> reference even though it will pass through your schematron rule ?

I am using the EBNF productions of RFC 2396 and leaving it at that. A better
implementation would have knowledge of each individual scheme.

I am always happy to include additional implementations, schemata,
stylesheets and transforms that apply to RDDL. Indeed the only reason RDDL
already has a DTD and RDF, RELAX, Schematron and TREX schemata is largely
through the help of the authors of these schemata. For example Murata Makoto
largely wrote the RELAX Namespace module (both the XHTML Basic and RDDL
modules), James Clark provided XHTML Basic TREX grammar to which I was able
to easily add the RDDL module and Rick Jellife was very helpful in helping
me write the Schematron schema (the updated version now handles the Flow.mix
pattern).

You may notice that an XML Schema for RDDL is now conspicuously absent. This
isn't because I "don't like" XML Schema, rather because I don't have access
to an XHTML Basic XSD from which to add a RDDL module. Hint. Hint. Hint :-)

>
> The nature of the URI (absolute|relative) would make an interesting
> facet for the W3C XML Schema uriReference datatype [1]...
>
> This example is also showing how useful it would be to be able to "hook"
> Schematron like rules into other schema languages...
>

I agree. One of the very very cool things about XML Schema is the use of
Regular Expression patterns to define datatypes. Adding Schematron/XPath
type assertions would make XML Schema's complexity well worth it (IMHO).

-Jonathan