Re: RDFCore WG: Datatyping documents

On 2002-01-28 17:10, "ext Jonathan Borden" <> wrote:

> Patrick Stickler wrote:
>> If we only used complete URIs and not qnames to refer to XML Schema
>> datatypes, would that work for you?
>> That would be doable, I think. Not always convenient, but doable.
> Certainly, the URIs are defined by XML Schema and so I would not have any
> sort of objection to that.

Taking this as a starting point ;-) let me try to address
your other (valid) points.

> Use of QNames would also be fine just as long
> that the QName is also defined as in XML Schema. The problem is then that
> you have to map the QName to URI as XML Schema does, which RDF does not do.
> Jonathan

If the qname used in the RDF/XML
serialization always resulted in the correct URI in the RDF graph,
in a context where the qname is not being interpreted by an XML
Schema application, would that be OK?

Note that, because of the nature of the RDF/XML content model
(or lack thereof ;-) you cannot validate an RDF instance according
to a DTD or XML Schema very easily, and I don't believe it was
ever expected that RDF instances would be tested for anything other
than well-formedness except by an RDF specific parser.

So, if we're never applying XML Schema interpretation to the qnames,
and those qnames always map to the correct, full URIs in the RDF
graph (where all the fun stuff happens), would that be OK?

This doesn't mean that RDF's treatment of XML Schema qnames is
"correct" or that RDF doesn't have problems with regards to qname
to URI mapping -- only that, at the present time, it seems that we
actually can find a reasonable way to use both RDF and XML Schema
datatypes, and it would be great to do so, while we work on fixing
all that other stuff.

It's kind of a crawl, walk, run progression...



Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email:

Received on Monday, 28 January 2002 10:21:39 UTC