Re: ACTION-103 are all OWL 1.0 ontologies representable in RDF

From: "Michael Schneider" <schneid@fzi.de>
Subject: RE: ACTION-103 are all OWL 1.0 ontologies representable in RDF
Date: Wed, 19 Mar 2008 22:25:33 +0100

> [cc'ed Alan, since I refer to him]
> 
> Hi Peter!
> 
> Peter F. Patel-Schneider wrote on Wednesday, March 12, 2008:
> 
> >From: "Michael Schneider" <schneid@fzi.de>
> >Subject: RE: ACTION-103 are all OWL 1.0 ontologies representable in RDF
> >Date: Wed, 12 Mar 2008 21:36:49 +0100
> >
> >> [related to ISSUE-100]
> >> 
> >> Hi Peter!
> >> 
> >> Here is a concrete example expression to check my understanding:
> >> 
> >>   ObjectProperty(http://example.org#foo)
> >> 
> >>   Class(http://example.org#foo partial 
> >>         restriction(http://example.org#foo 
> >>                     value(http://example.org#foo) ) )
> >> 
> >> My questions:
> >> 
> >> (A) Is this expression syntactically valid in OWL-1.0 
> >> Abstract Syntax?
> >
> >Close enough.  :-)
> >
> >> (B) Is this expression transformable to RDF by means of the 
> >> OWL-1.0 RDF mapping?
> >
> >Yes, as every OWL 1.0 DL ontology is transformable.  
> 
> Indeed! I was confused by this answer last week. But you are right, of
> course: The RDF mapping actually works on /every/ OWL-DL ontology in
> abstract syntax form.
> 
> I remember that Alan has talked about a kind of "roundtripping problem",
> probably with the idea that it is possible to map from abstract syntax
> to RDF, but not back in every case. But I think this isn't a perfectly
> adequate characterization of the situation here. In OWL-DL, there is no
> reverse mapping, there is just an RDF mapping. And if this RDF mapping
> allows to travel from abstract syntax to RDF, then, trivially, there is
> also a way back again, and the result of this "round trip" will always
> be the original OWL-DL ontology in abstract syntax form.

Except that the mapping is not 1-1, so there can be several
possibilities in the reverse direction.

> However, what this mapping alone doesn't give an answer to is the
> question whether the result of the mapping of an OWL-DL ontology in
> abstract syntax form is in every case also an /OWL-DL ontology in RDF
> graph form/. The latter term is defined in sec. 4.2 of the AS&S:
> 
>   <http://www.w3.org/TR/owl-semantics/mapping.html#4.2>
> 
>   "An RDF graph is an OWL DL ontology in RDF graph form 
>   if it is equal [...] to a result of the transformation 
>   to triples [...] of a collection of OWL DL ontologies 
>   and axioms and facts in abstract syntax form 
>   that has a separated vocabulary."

True.

> >> (C) Is this expression a legal OWL-1.0-DL ontology?
> >
> >If it is syntactically valid, then it is legal.
> 
> So the question, which I should better have been asking for, was
> (following question C):
> 
>   (D) Is there an OWL-DL ontology in RDF graph form,
>       which corresponds (by means of the RDF mapping) 
>       to the above OWL-DL ontology in abstract syntax form?
> 
> And the answer to (D) is "no", simply because my example ontology is an
> OWL-DL ontology in abstract syntax form, which does *not* have a
> separated vocabulary.

Again true.

> The question here is of course: Is this just play on words? Or have
> there been any technical considerations which have eventually lead to
> the situation that non-separated vocabularies (i.e. punning) were
> allowed for "OWL-DL ontologies in abstract syntax form", but not for
> "OWL-DL ontologies in RDF graph form"?

Well, the theorems in the latter parts of the document are only
guaranteed to hold for ontologies with separated vocabularies.

> >> Regards,
> >> Michael
> >
> >peter
> 
> All the best,
> Michael

peter

Received on Wednesday, 19 March 2008 23:50:24 UTC