Re: Datatype test cases: important ones (please have a look)

[Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com]


----- Original Message ----- 
From: "ext Brian McBride" <bwm@hplb.hpl.hp.com>
To: "Jan Grant" <Jan.Grant@bristol.ac.uk>; "jeremy Carroll" <jjc@hplb.hpl.hp.com>
Cc: "RDFCore Working Group" <w3c-rdfcore-wg@w3.org>
Sent: 20 November, 2002 18:11
Subject: Re: Datatype test cases: important ones (please have a look)


> 
> At 15:55 20/11/2002 +0000, Jan Grant wrote:
> >On Wed, 20 Nov 2002, Brian McBride wrote:
> >
> > > The way this is put suggests that
> > >
> > >   <a> <b> "10"^^xsd:integer .
> > >
> > > entails all other datatype representations of the same value.
> >
> >Within _one_ datatype, that's true - this is rule rdfD 2 in the current
> >MT.
> >
> >For multiple datatypes, the MT says this:
> >
> >[[
> >These rules do not support any entailments based on identity between
> >values of different datatypes. An obvious generalization of the second
> >rule would permit such conclusions, but questions of identity between
> >items in value spaces of two different datatypes should be referred to
> >the authorities who defined the datatypes.
> >]]
> 
> Oh yes, you are right.  We are straying into specifying things about xsd 
> that are not up to us.
> But I do think we have to do that for the 10 and 010 cases.  I do think 
> those need to be changed, but would want a second opinion.

I think the entailment should hold for all other datatype representations
of the same value where the datatype is a supertype of the xsd:integer.

I.e., if we are talking about RDF only entailments, no it doesn't hold.
But if we are talking about RDFS entailments, then we have to take into
account rdfs:subClassOf relations and the fact that for any datatype X that
is a subclass of some datatype Y, the value space of X is a subset of
that of Y. So the entailment would hold.

So there are four tests here:

RDF                    -> does not hold
RDF        + datatypes -> does not hold
RDF + RDFS             -> does not hold 
RDF + RDFS + datatypes -> holds for xsd:integer and all 
                          superclasses of xsd:integer

If the MT does not capture the above, it should. As should the
test cases.

Patrick

Received on Thursday, 21 November 2002 02:19:46 UTC