Re: possible changes to abstract syntax and direct semantics to support annotations and fix problem with imports

From: "Jos De_Roo" <jos.deroo@agfa.com>
Subject: Re: possible changes to abstract syntax and direct semantics to support annotations and fix problem with imports
Date: Sun, 9 Feb 2003 16:20:36 +0100

> 
> [...]
> 
> > > but then that would mean that to have a non empty individual
> > > all appropriate annotations have to in the KB, right?
> >
> > It means that to satisfy
> >
> > Individual(annotation(x y)  annotation(x z) type(Person))
> >
> > there has to be annotation triples of the form <a,I(x),I(y)><a,I(x),I(y)>
> > and <a,I(x),I(z)> for some a.
> 
> Maybe I'm confused...
> Suppose the naming authority of na: creates an ontology in
> which there is a individual na:x that has 2 annotation triples.

Well, I'm not really interesting in anything that starts with a presumption
of a ``nameing authority'', as this has no place in either the RDF or OWL
model theory.

> Suppose we are now given a graph with the triple na:x :y :z.
> but not the annotation triples.
> Would that graph entail na:x :y :z ?

Further, I don't see nearly enough information here to determine what it
going on.  If you want to pose entailment questions, you are going to have
to put them in the form of one RDF graph entailing another or an OWL
abstract ontology or set of directives entailing another. 

> > > > > On the other hand, the so called annotation "triples"
> > > > > (NOT to confuse with RDF triples) are so
> > > > > weak that no entailement other than belonging to A
> > > > > can be done with them (seems to me).
> > > >
> > > > Well, this was the entire idea.
> > >
> > > Assuming that is the idea, then when such
> > > triples are exchanged in RDF/XML how can
> > > we ever know that they are special?
> >
> > You wouldn't have to.
> 
> I don't understand that.
> Suppose we have an annotation triple :x :y :z
> and we know that :y is a subproperty of :v
> would that entail :x :v :z ?

Again, what is the desired entailment?  I don't think that this is even
allowable in the abstract syntax.

> > > I also guess that sameAs would apply to
> > > their URIreference/dataLiteral arguments.
> >
> > The elements of the triples are elements of the domain of discourse, and
> > the arguments of the ontology constructs are run through the denotation
> > function, so
> >
> >          Individual(john annotation(hi john))
> >          Individual(jack)
> >          SameIndividual(john jack)
> >
> > would entail
> >
> >          Individual(john annotation(hi jack))
> 
> That I understand.
> 
> [...]
> 
> -- ,
> Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/

peter

Received on Sunday, 9 February 2003 12:46:06 UTC