W3C home > Mailing lists > Public > www-webont-wg@w3.org > March 2002

Re: Layering on what? was: Re: more on a same-syntax extension from RDF(S) to OWL

From: Jos De_Roo <jos.deroo.jd@belgium.agfa.com>
Date: Tue, 5 Mar 2002 02:08:43 +0100
To: pfps@research.bell-labs.com
Cc: "jonathan" <jonathan@openhealth.org>, "www-webont-wg" <www-webont-wg@w3.org>
Message-ID: <OF482B1AF6.D47C5E55-ON41256B73.00011AAF@agfa.be>

> > I am having a hard time following the layering threads. In the interest
of
> > simplifying this enough to penetrate my dense head:
> >
> > 1) Are we talking about layering on RDF _as RDF is defined_ or
> > 2) on N3, which seems to be used _as if_ it were RDF, but it is not at
all
> > clear to me how N3 is itself 'layered' on, or defined as an extension
to,
> > RDF
> >
> > My preference at this point would be to 'layer' OWL on N3, whatever
that
> > means. My reasons are simply that N3 provides syntax for some seemingly
> > important capabilities that are lacking in the current RDF syntax --
and I
> > do not mean that RDF is serialized as XML, as one could create an XML
> > version of N3, rather that the seemingly innocuous rdf:parseType
="log:quote"
> > hides some major advances in the RDF syntax, that (IMHO) would
considerably
> > change (and improve) RDF.
> >
> > So for example, this discussion seems to be in N3, not RDF:
> >
> > >
> > > > > [...]
> > > > >
> > > > > > So, to be more precise it should have been
> > > > > >
> > > > > >
> > > > > > log:entails
> > > > > >   _:1 owl:oneOf ( _:2 ) .
> > > > > >   _:2 a owl:Restriction .
> > > > > >   _:2 owl:onProperty rdf:type .
> > > > > >   _:2 hasClassQ _:1 .
> > > > >         ^owl:
> > > > > >   _:2 maxCardinalityQ "0" .
> > > > >         ^owl:
> > > > >
> > ...
> > >
> > > good point, I've added
> > >   { ?L owl:item ?x } log:implies { ?x a [ owl:oneOf ?L ] } .
> >
> > How would one write the above in RDF?

we can surely look to the premis { _:L owl:item _:x } as an RDF graph
where the bnodes of that graph (luckily) become universally quantified
(reaching to conclusion scope) therefore we write ?L instead of _:L
the premis statements are *not* asserted
we can also look to { ?x a [ owl:oneOf ?L ] } as an RDF graph :c, where
[ owl:oneOf ?L ] is like a Skolem functional term replacement of a bnode
also the conclusion graph is *not* asserted
:p log:implies :c is an RDF statement that *is* asserted

> > This is important, because if we cannot even provide examples in RDF,
how
> > can we properly 'layer' OWL on RDF?
> >
> > In any case, this seems like a fairly important discussion, but I am
lost.

well, let's say it's a way to express SIMPLE, RDFS, OWL, etc. entailment

> > Jonathan
> >
>
> Generally I try to use raw RDF, but occasionally, particularly if someone
> else is so doing, I slip into an informal N-triples or N3 syntax.  My
part
> of the example above uses only parts of N3 that map well into RDF.
>
> The problem with layering on N3 is that you can't really layer on
something
> that has no semantics. N-triples, on the other hand, has both a
reasonable
> syntax and a decent semantics (or soon will).  However, N-triples is
really
> nothing more than RDF.

a small step for you, but a...

--
Jos
Received on Monday, 4 March 2002 22:43:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:48 GMT