Re: revised version of semantics document

> > [...]
> >
> > > > > Why should anyone care at all about the entailment rules?
> > > >
> > > > because
> > > >
> > > [...]
> > >
> > > OK, agreed, entailment can be related to *inference* rules.
> > >
> > > But why should I care whether ``no new existentials are introduced 
in the
> > > [inference] rules''?
> >
> > OK, e.g. we can't have existentials in the conclusion that are
> > in the domain of an owl:InverseFunctionalProperty because that
> > property is simply not defined over it's whole range
> > (lists in the range of e.g. owl:intersectionOf could contain a
> > mix of resources being classes, properties, classes/properties
> > and so many of such intersections simply don't exist and in
> > general we can't give the list constraints in the premis)
>
> Why not?  You mean, perhaps, that you can't.
>
> However, there is nothing to prevent one from writing an inference rule
> like
>
>                 ?x rdf:type owl:Class .
>               ?y rdf:type owl:Class .
>               ?i rdf:type ?x .
>               ?i rdf:type ?y.
>       implies ?i rdf:type _:l .
>               _:l owl:intersectionOf [ ?x ?y ] .

Why not having owl:intersectionOf defined for instances
of rdf:Class and even instances of rdf:Property?
They are sets, aren't they?

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

Received on Sunday, 25 August 2002 20:11:38 UTC