RE: RDFCore Update

> -----Original Message-----
> From: ext Devon Smith [mailto:devon@taller.pscl.cwru.edu]
> Sent: 18 October, 2001 23:46
> To: Stickler Patrick (NRC/Tampere)
> Cc: www-rdf-interest@w3.org
> Subject: Re: RDFCore Update
> 
> 
> On Thu, Oct 18, 2001 at 10:46:49PM +0300, 
> Patrick.Stickler@nokia.com wrote:
> > 
> > > From: "Brian McBride" <bwm@hplb.hpl.hp.com>
> > > 
> > > > One major area of focus for the WG at the moment is 
> > > datatyping, e.g. using
> > > XML
> > > > schema datatypes in RDF.  Now would be a good time to let 
> > > us have your
> > > thoughts
> > > > and ideas on this.
> > > 
> > > I think values should be anonymous nodes in RDF with property arcs
> > > describing them in XML datatypes.
> > > 
> > > Seth Russell
> > 
> > Yes. That's one approach. But not necessarily the only or
> > most optimal approach for all circumstances.
> > 
> > IMO it should also be possible to assign types to values in
> > other ways, and there should be some official equivalence
> > logic defined for these variant methods.
> > 
> > One would presume that all of the following three examples
> > define precisely the same knowledge regarding data types:
> > 
> > --
> > 
> > <rdf:Description rdf:about="urn:foo:bar">
> >   <abc:someProperty>
> >     <rdf:Description>
> >       <rdf:type rdf:resource="x:dataType"/>
> >       <rdf:value>dataValue</rdf:value>
> >     </rdf:Description>
> >   </abc:someProperty>
> > </rdf:Description>
> > 
> > --
> > 
> > <rdf:Description rdf:about="urn:foo:bar">
> >   <abc:someProperty>dataValue</abc:someProperty>
> > </rdf:Description>
> > 
> > <rdf:Description rdf:about="abc:someProperty">
> >   <rdfs:range rdf:resource="x:dataType"/>
> > </rdf:Description>
> 
> wouldn't this use of rdfs:range change it's semantics?
> if it's used as a type specification mechanism, that would seem
> to exclude it from being used as a type verification mechanism.
> that is, if my parser reads in the fragment above, does it reject
> it because 'dataValue' isn't of type x:dataType, or does it decide
> that 'dataValue's type is x:dataType. i've understood rdfs:range to
> mean the former. i guess it could be used for both, where the 
> specification would apply to literals and verification would be used
> with resources. that would exclude verification for any literal
> values though. 

Well, if you don't know what the data type of the literal
is, how can you test if it conforms to a range constraint?

I.e. there really is a fourth possible variant to the three
above, which seems to be what you are expecting. Namely

 <rdf:Description rdf:about="urn:foo:bar">
   <abc:someProperty>
     <rdf:Description>
       <rdf:type rdf:resource="x:dataType"/>
       <rdf:value>dataValue</rdf:value>
     </rdf:Description>
   </abc:someProperty> 
 </rdf:Description>
 
 <rdf:Description rdf:about="abc:someProperty">
   <rdfs:range rdf:resource="x:dataType"/>
 </rdf:Description>

In which case, range validation could take place, since
we can compare the 'locally defined' data type with the
'globally defined' range (data type) of the property.

This ambiguity about whether rdfs:range is descriptive or
prescriptive (or both) was pointed out recently in a post
by Ora Lassila, and perhaps could be addressed in conjunction 
with standardized methodologies/mechanisms for data typing.

Cheers,

Patrick

--
Patrick Stickler                      Phone:  +358 3 356 0209
Senior Research Scientist             Mobile: +358 50 483 9453
Nokia Research Center                 Fax:    +358 7180 35409
Visiokatu 1, 33720 Tampere, Finland   Email:  patrick.stickler@nokia.com

Received on Thursday, 18 October 2001 17:12:51 UTC