- From: Dave Beckett <dave.beckett@bristol.ac.uk>
- Date: Tue, 29 Oct 2002 11:34:55 +0000
- To: Brian McBride <bwm@hplb.hpl.hp.com>
- cc: Jonathan Borden <jonathan@openhealth.org>, w3c-rdfcore-wg <w3c-rdfcore-wg@w3.org>
>>>Brian McBride said:
> At 22:05 28/10/2002 -0500, Jonathan Borden wrote:
> >The current RDF datatypes proposal seems to consider so-called "simple" XML
> >Schema datatypes e.g.
> >
> ><rdf:Description>
> > <ex:foo rdf:datatype="&xsd;integer">10</ex:foo>
> ></rdf:Description>
> >
> >might this also work for "XML" datatypes i.e. fragments of XML that are
> >valid with respect to a "complex" XML Schema datatype e.g.
> >
> ><rdf:Description>
> > <ex:foo rdf:datatype="&foo;bar">
> > <this> is a simple structured datatype</this>
> > </ex:foo>
> ></rdf:Description>
> >
> >where &foo;bar identifies an XML Schema particule analogous to the XML
> >datatype <!ELEMENT this (#PCDATA)>
> >
> >If this is already allowed, I will propose that WebOnt close our issue and
> >if its not something currently allowed, do RDF Core folk consider this
> >useful?
>
> It is not currently allowed. Dave, correct me if I'm wrong. We
> don't allow combining rdf:datatype and parseType="Literal" do we?
No, we don't.
In particular parseType literal is not what I'd like to encourage
using as an extension mechanism, especially for datatypes!
We used to think of the parseType="Literal" as setting an is_XML flag
on the literal to be true, now we think of it as a datatyped literal
where the URI of the datatype is rdfs:XMLLiteral.
As I understand it, the XML Schema WG has an issue to provide URIs
for their more complex datatypes ("particles"?) but hasn't yet
addressed it. I just remember hearing this, I may be wrong.
So if that happens, then it'll be
<rdf:Description>
<ex:foo rdf:datatype="uri-of-complex-xsd-datatype"><foo>...the serialization here ...</foo></ex:foo>
</rdf:Description>
Dave
Received on Tuesday, 29 October 2002 06:36:56 UTC