W3C home > Mailing lists > Public > www-webont-wg@w3.org > August 2003

Re: testing that RDF datatypes can have more than one URI name

From: Jos De_Roo <jos.deroo@agfa.com>
Date: Fri, 15 Aug 2003 15:20:15 +0200
To: pfps@research.bell-labs.com
Cc: bwm@hplb.hpl.hp.com, connolly@w3.org, danbri@w3.org, www-webont-wg@w3.org
Message-ID: <OFDCB5B59C.6CD2FA74-ONC1256D83.00480B92-C1256D83.004943E2@agfa.be>


Peter F. Patel-Schneider wrote:
> From: Dan Connolly <connolly@w3.org>
> Subject: testing that RDF datatypes can have more than one URI name
> Date: Thu, 14 Aug 2003 15:59:35 -0500
>
> >
> > Peter, Jeremy, Jos, and everybody,
> >
> > In response to a comment from Peter, RDFCore changed
> > some details about how datatypes work... Peter seems
> > to think this was the right thing to do...
> >
http://lists.w3.org/Archives/Public/www-rdf-comments/2003JulSep/0065.html
> >
> > In a July 11 RDF Core telcon,
> >   http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jul/0173.html
> > we observed that you can't really test that the same datatype
> > can have two URI-names using only RDF machinery, but you probably
> > can with OWL (I think that's the crux of the matter, but it's
> > been so long that I'm not sure). I accepted an action (20030711#4)
> > to ask the OWL WG to add such a test.
> >
> > I meant to actually design the test and propose it to the WG,
> > but it's been several weeks and I haven't managed to do that,
> > so before this action item gets any stinkier, I'm going to
> > just ask. Peter, maybe you can design a test? Jeremy? Jos?
> >
> > This isn't on the critical path for OWL CR publication, FYI.
> >
> > --
> > Dan Connolly, W3C http://www.w3.org/People/Connolly/
>
> It is quite easy in OWL Full, at least it is easy to do what I think you
> want.
>
> Here is a QName triples version (translate as appropriate for RDF/XML)
>
>            xsd:decimal owl:sameAs ex:bar .
>            ex:xx ex:yy "01^^ex:bar .

I had trouble parsing the last triple, but
            ex:xx ex:yy "01"^^ex:bar .
worked better ;-)

> owl full entails
>
>            ex:xx ex:yy "1"^^xsd:decimal .
>
> because ex:bar and xsd:decimal have the same denotation and thus have the
> same L2V mapping.
>
> Note that this only depends on the datatype map mapping xsd:decimal to
the
> XSD decimal type, and does not depend on ex:bar being in the datatype
map.

I again had trouble to infer that, but this time due to
a shortcoming in our implementation of the substitution
of equals-for-equals; I made a trial change and then
it worked (but I have to eventually reconsider it).

jos

> peter


--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/

PS Is there any kind of tool out there that can re-write OWL documents
   in (what I would call) "Intersection Normal Form" (INF) I mean
   re-write sets as intersections of unions of (complements of) sets?
   Is/will Jena 2 have such a mode?
   (then I can throw away many implications...)
Received on Friday, 15 August 2003 09:21:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:02 GMT