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

RE: DTTF: summary (gasp!)

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Fri, 24 May 2002 09:30:08 -0400
To: jos.deroo.jd@belgium.agfa.com
Cc: jjc@hplb.hpl.hp.com, jonathan@openhealth.org, www-webont-wg@w3.org
Message-Id: <20020524093008N.pfps@research.bell-labs.com>
From: "Jos De_Roo" <jos.deroo.jd@belgium.agfa.com>
Subject: RE: DTTF: summary (gasp!)
Date: Fri, 24 May 2002 15:01:41 +0200

> 
> > > [...]
> > >
> > > > In particular the following implication that is valid under RDFS will
> > > > not hold under OWL.
> > > >
> > > > eg:prop rdfs:subClassOf owl:Restriction .
> > > > _:x rdf:type eg:prop .
> > > >
> > > > entails
> > > >
> > > > _:x rdf:type owl:Restriction .
> > >
> > > let's decide on this one
> > > for me this entailment is OK under OWL (as well as under RDFS)
> > >
> > > --
> > > Jos
> > >
> >
> > Making owl restrictions be elements of the domain of discourse is one of
> > the most, if not the most, dangerous things to do.
> >
> > That said, it would be possible to have the above entailment go through
> > (maybe) even if owl restrictions are not elements of the domain of
> > discourse.
> 
> really!?
> can you please elaborate a bit on that?

Sure.  In fact there is no need to have any special treatment here.

SubClassOf(eg:prop owl:Restriction) 

just says that eg:prop is a subclass of owl:Restriction.

Then 

Individual(type=eg:prop)

just says that there is some object that is an instance of eg:prop.

From this, it had better follow that there is an instance of
owl:Restriction, i.e.,

Individual(type=owl:Restriction)


Note that there is no treatment of restrictions here at all.


I realize that this may not be what you are asking.  If you want
owl:Restriction to be dark, i.e., that there is not necessarily a
SubClassOf axiom above, there is a separate way to go.  

In this way the model theory of OWL would turn triples like

	?x rdfs:subClassOf owl:Restriction

into a subclass relationship on its own authority, i.e., not because the
RDFS semantics required it.  The rest would follow.  I do not advocate this
way of proceeding, however,


peter
Received on Friday, 24 May 2002 09:30:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:50 GMT