RE: Cutting the Patrician datatype knot

So we have, if I understand well, 3 propositions for representing typed
data:

<p xsi:type="du">x</p>       <!-- by Peter F. Patel-Schneider-->
<p rdf:parseType="du">x</p>  <!-- by Geoff Chappell-->
<p rdf:resource="xsd:du:x"/> <!-- by Patrick Stickler-->

all resulting to the same RDF triple, most simply represented by
Patrick's version.

On Thu, 2001-11-22 at 17:34, Peter F. Patel-Schneider wrote:
> > If you're going to put the mapping in the parsing, why not just use
> > 'parseType=' to make clear it's a parser directive?
> 
> Precisely because XML Schema has a perfectly good way of doing it, so why
> not use that way?

Because the XML syntax of RDF, as proposed in the RDF M&S
recommendation, does *not* rely on XML schema (neither does it have a
DTD, btw).
Hence it needs another *explicit* way of expressing the type of a value.

Patrick's version is the most simple, though not very intuitive.
I also like Geoff's proposal because it does not need to extend the
syntax (only the parsers), and I do think that rdf:parseType works well
for that purpose.


On the other hand, I can imagine an RDF *harvester* (by contrast to a
*parser*) able to produce RDF triple from XML schema-compliant
documents, provided that
 - we define some general rules about XMLS to RDF mapping
   (in a way similar to the approach used in
      http://www.w3.org/TR/xlink2rdf/ )
 - the schema contains some additionnal RDF "hints", like default values
for rdf:parseType in the appropriate elements.


  Pierre-Antoine