Re: [OWLWG-COMMENT] Punning and the "properties for classes" use case

From: "Michael Schneider" <schneid@fzi.de>
Subject: RE: [OWLWG-COMMENT] Punning and the "properties for classes" use case 
Date: Mon, 5 Nov 2007 10:24:32 +0100

> Hi again, Peter!
> 
> I see that I still have several concrete problems in understanding punning,
> before I can answer your other mails. Here is the next question:
> 
> Peter F. Patel-Schneider wrote:
> 
> >Well, in OWL 1.1, it is true that 
> >
> >	Declaration(Individual(a))
> >	Declaration(OWLClass(a))
> >	Declaration(Individual(b))
> >	Declaration(OWLClass(b))
> >	SameIndividual(a b)
> >
> >does not entail
> >
> >	EquivalentClasses(a b)
> 
> My question is /not/ about this claim, 

My mistake then.

> but about RDF mapping. I want to
> know: Given an OWL-1.1 ontology in functional style syntax, in which punning
> is used. How is such an ontology mapped to RDF? As an example, let's take
> the above six axioms:
> 
>   Declaration(Individual(a))
>   Declaration(OWLClass(a))
>   Declaration(Individual(b))
>   Declaration(OWLClass(b))
>   SameIndividual(a b)
>   EquivalentClasses(a b)
> 
> How do these axioms get mapped to RDF? I can imagine two different mappings:
> 
> A: Each URI ref in the functional syntax version gets directly mapped to the
> same URI ref in the resulting RDF graph:
> 
>   <a> owl11:declaredAs owl11:Individual .
>   <a> owl11:declaredAs owl:Class .
>   <b> owl11:declaredAs owl11:Individual .
>   <b> owl11:declaredAs owl:Class .
>   <a> owl:sameAs <b> .
>   <a> owl:equivalentClass <b> .
>   
> B: The URI refs get explicitly separated:
> 
>   <a_Thing> owl11:declaredAs owl11:Individual .
>   <a_Class> owl11:declaredAs owl:Class .
>   <b_Thing> owl11:declaredAs owl11:Individual .
>   <b_Class> owl11:declaredAs owl:Class .
>   <a_Thing> owl:sameAs <b_Thing> .
>   <a_Class> owl:equivalentClass <b_Class> .

Well, looking at the RDF mapping document gets you the (current) answer,
which is A (roughly).   (There are also typing tripes generated.)

> As far as I understand it: If punning allows to denote semantically
> unrelated individuals and classes by the same URI ref in functional syntax,
> than mapping A would change the original semantics of the ontology, while
> this semantics would be preserved by mapping B. But otherwise, if an
> individual and a class having the same name are intended to be the same
> entity, than mapping A would be a correct mapping, while B would also allow
> interpretations where the individual and the class are semantically
> unrelated.

No.  From the RDF mapping document:

	OWL 1.1 ontologies serialized in RDF/XML are interpreted by
	translating them into the functional-style syntax and applying
	the OWL 1.1 semantics.

So the OWL 1.1 meaning of the triples is defined in the OWL 1.1
Semantics document, just as the semantics of an OWL DL ontology is
defined by the direct model-theoretic semantics for OWL DL no matter
what form it is written in.

So even if the mapping was changed, the semantics would remain the same.

> So only one of these mappings should be allowed. But I cannot find out from
> the "RDF mapping" draft which mapping is correct. Probably, I have only
> overlooked the important bits in the draft?

Well, you have to follow the RDF mapping a bit.  Consider

	EquivalentClasses(a b)

Its transformation generates the triple

	T(a) owl:equivalentClass T(b)

because a and b occur in the functional syntax as owlClassURI, their
transformation, T(a) and T(b), generates the triples

	a rdf:type owl:Class
	b rdf:type owl:Class

and T(a) is a and T(b) is b in the first triple.  A little examination
of the mapping will show that there is no modification of URIs in the
mapping.  (Yes, I agree that the mapping is quite hard to read.)

> Cheers,
> Michael

peter

Received on Monday, 5 November 2007 11:48:53 UTC