- From: Dan Connolly <connolly@w3.org>
- Date: Tue, 21 Aug 2001 14:33:35 -0500
- To: Aaron Swartz <aswartz@upclink.com>
- CC: Dave Beckett <dave.beckett@bristol.ac.uk>, w3c-rdfcore-wg@w3.org
Aaron Swartz wrote:
>
> On Tuesday, August 21, 2001, at 11:21 AM, Dave Beckett wrote:
>
> > Pretty clearly when rdf:type is used as a property attribute, it is
> > defined to take a resource as a value (this is in the grammar).
>
> It is my opinion that since the grammar refers to an unprefixed
> 'type', and we have disallowed that irregularity, this
> irregularity should be removed also. It is likely that
> processors dealing with valid documents (i.e. prefixed with
> rdf:) will already be in line with recent RDF Core decisions. I
> think we should also remove the irregularity that causes type to
> correspond to a resource and not a literal, thus simplifying the
> grammmar, with little impact on backwards compatibility.
I regret to report that when I first implemented RDF syntax,
I completely missed this exception for rdf:type. My
XSL-based RDF parser does as Aaron suggests: treats
rdf:type just like any other propAttr.
http://www.w3.org/XML/2000/04rdf-parse/
lemme check sax2rdf... bad news; it gets rdf:type wrong too:
$ python2 ../cwm.py --rdf typeAttr.rdf --triples
<file:/home/connolly/w3ccvs/WWW/2000/10/swap/test/typeAttr.rdf#Animal>
<http://www.w3.org/1999/02/22-rdf-syntax-ns#type> "http://example/xyz" .
typeAttr.rdf:
<rdf:RDF
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
<rdf:Description rdf:ID="Animal" rdf:type="http://example/xyz"/>
</rdf:RDF>
--
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Tuesday, 21 August 2001 15:33:41 UTC