W3C home > Mailing lists > Public > public-owl-wg@w3.org > December 2007

example of change in satisfiability because of rdfs:Class to owl:Class repair Re: ISSUE-55 (owl:class): owl:class vs rdfs:class

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Wed, 19 Dec 2007 12:22:25 -0500 (EST)
Message-Id: <20071219.122225.80011931.pfps@research.bell-labs.com>
To: public-owl-wg@w3.org

From: "Boris Motik" <boris.motik@comlab.ox.ac.uk>
Subject: RE: ISSUE-55 (owl:class): owl:class v. rdfs:class
Date: Tue, 13 Nov 2007 14:08:27 +0900

> Hello,
> 
> Changing rdfs:Class into owl:Class does significantly affect the formal
> meaning of an ontology. [...] If
> necessary, I can come up with examples in which satisfiability of
> an ontology changes if you change rdfs:Class to owl:Class.

[...]

> Regards,
> 
> 	Boris

Here is one such example.  It also doesn't use any syntactic markers
what could indicate that the repair is problematic (e.g., the use of
owl:Thing).


OWL Full ontology [with comments]:

ex:Class rdf:type rdfs:Class
nex:ClassC rdf:type rdfs:Class
ex:R rdf:type owl:SymmetricProperty
ex:P rdf:type owl:SymmetricProperty
ex:spy rdf:type owl:Thing

ex:ClassC owl:complementOf ex:Class

[ex:spy is a spy point for ex:Class and property ex:P]
ex:Class rdfs:subClassOf _:r1
_:r1 rdf:type owl:Restriction
_:r1 owl:onProperty ex:P
_:r1 owl:hasValue ex:spy

[ex:spy is a spy point for ex:ClassC and property ex:R]
ex:ClassC rdfs:subClassOf _:r3
_:r3 rdf:type owl:Restriction
_:r3 owl:onProperty ex:R
_:r3 owl:hasValue ex:spy

[ex:spy forces extension of ex:Class to maximum size 5]
ex:spy rdf:type _:r2
_:r2 rdf:type owl:Restriction
_:r2 owl:onProperty ex:P
_:r2 owl:maxCardinality "5"^^xsd:nonNegativeInteger

[ex:spy forces extension of ex:ClassC to maximum size 6]
ex:spy rdf:type _:r4
_:r4 rdf:type owl:Restriction
_:r4 owl:onProperty ex:R
_:r4 owl:maxCardinality "6"^^xsd:nonNegativeInteger

This is unsatisfiable in OWL Full, but with the rdfs:Class to owl:Class
"repair" it is in OWL DL (modulo missing typing triples, but I think I
have a sufficient set) and is satisfiable in OWL DL.  


peter
Received on Wednesday, 19 December 2007 17:43:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:13:29 GMT