W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > October 2002

Re: rdf:datatype v xsi:type

From: Graham Klyne <Graham.Klyne@MIMEsweeper.com>
Date: Wed, 16 Oct 2002 10:18:18 +0100
Message-Id: <5.1.0.14.2.20021016101235.03b86ec0@127.0.0.1>
To: Jeremy Carroll <jjc@hpl.hp.com>
Cc: w3c-rdfcore-wg@w3.org

At 09:44 AM 10/16/02 +0200, Jeremy Carroll wrote:


> > one attribute that can take qname values and none of the others
>
> >From the parser writers point of view this would be a pain.
>The hard work is in supporting the first, and then we can't give (this useful
>functionality) to the users other than in a very limited scope.
>
>Thus qnames in RDF2 looks like a plausible mantra.
>
>There is also a can of worms that we are currently ducking to do with
>invisibly used namespaces (in XML Literals). I find it harder to duck if we
>(like xsi, xslt xpointer ...) decide that invisibly used namespaces are cool.
>
>(Note xsi:type="xsd:decimal" visibly uses xsi, and invisibly uses xsd).
>
>Jeremy


I seem to recall some time ago some inconclusive TAG discussion of this... 
I think there was some discussion of the idea that QName values (and, 
presumably, other namespace prefix uses) should somehow be detectable at 
the schema level.  (I'm not sure if that's any help, just a comment.)

At the level of RDF concrete syntax, using namespace prefixes (as in 
QNames) might be OK, but I think it would be problematic in an RDF graph 
where the concept of scoping is, at best, very weak.  In this respect, the 
decision to use (just) full URIs as identifiers seems rather sound.

#g


-------------------
Graham Klyne
<GK@NineByNine.org>
Received on Wednesday, 16 October 2002 05:15:53 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:52:26 EDT