Re: owl:equivalentClass and rds:subClassOf

On January 12, Bernard Vatant writes:
> 
> 
> Follow-up of a debate in Protégé list.
> 
> Holger Knublauch wrote:
> 
> "... in Protege there is no way to distinguish between
> 
>  :A rdfs:subClassOf :B
>  :B rdfs:subClassOf :A
> 
> and
> 
>  :A owl:equivalentClass :B
> 
> because they are internally mapped into mutual superclasses..."

Semantically, the two are completely equivalent. Whether one would
like to distinguish the two cases in modelling tools is another
(important) issue, and is a special case of a general problem that
Bechhofer et al noted w.r.t. DAML+OIL and its RDF syntax in an SWWS
2001 paper [1]. 

In general, I think that the answer is that mapping the KB into a
logically equivalent but syntactically different form may be very
confusing for modellers who often use syntactic structure to help
organise their thoughts and maintain a consistent modelling style.  I
would therefore suggest that tools use the XML presentation syntax
internally and only export to the RDF syntax when necessary to
"publish" the ontology - the XML syntax does a much better job of
retaining the syntactic structure as actually created by the modeller.

Regards, Ian

p.s. this doesn't really seem to be connected to the sufficient -v-
equivalent argument.

[1] http://www.cs.man.ac.uk/~horrocks/Publications/download/2001/not-enough.pdf

> 
> The reciprocal rdfs:subClassOf declares a logical equivalence of definition
> (intensional), whereas owl:equivalentClass declares an equivalence at
> instance level (extensional).
> 
> Should those declarations be kept distinct or not by a conformant OWL tool?
> And if yes, what would be the logical relationship, if any, between the
> former and the latter?
> 
> Bernard Vatant
> Senior Consultant
> Knowledge Engineering
> Mondeca - www.mondeca.com
> bernard.vatant@mondeca.com
> 
>

Received on Monday, 12 January 2004 08:42:29 UTC