Re: summary of current position with respect to semantics proposals (was Re: WOWG: agenda Aug 15 telecon)

From: "Jos De_Roo" <jos.deroo.jd@belgium.agfa.com>
Subject: Re: summary of current position with respect to semantics proposals 		(was  Re: WOWG: agenda Aug 15 telecon)
Date: Thu, 15 Aug 2002 13:41:36 +0200

[...]

> I've actually redone above testcase to better reflect
> that particular aspect of more than one ontology (*)
> 
>   http://www.w3.org/2002/03owlt/intersectionOfP
>   http://www.w3.org/2002/03owlt/intersectionOfX
> OWL-entails
>   http://www.w3.org/2002/03owlt/intersectionOfC

How?  What definition of OWL entailment did you use?  What system did you
use for the test?

> using a resolution based algorithm we have
> found the proof of that as part of the proof
>   http://www.agfa.com/w3c/euler/etc5-proof.n3
> (actually at the end of that file)

Ahh.  You are saying that you have a system that (conceptually) takes two
sorts of input, 

1/ control input in the form of n3 statements
2/ a problem description in the form of n3 statements plus some other
   information 

and returns a trace showing whether the problem description is a
consequence of the control input.

The question here is what relationships the system has to the various
proposals for OWL entailment.  I'm not asking for a proof that the system
is sound and complete wrt some proposal for OWL entailment.  A reasonable
argument that the system was correct for some portions of
some version of OWL entailment would be very useful.   

To take an easy example, and one that should be the same for all the
proposals for OWL entailment, is there a reasonable argument to show that
the system is correct, i.e., both sound and complete, for "type"
entailments, i.e., entailments of the form 
	KB |= Q
where KB is an OWL KB that corresponds to the abstract syntax
and Q is a query of the form 
	i rdf:type c .
where i appears in the KB, but not as a class or property, 
and c appears in the KB as a class.

> -- ,
> Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/

Peter F. Patel-Schneider
Bell Labs Research

Received on Thursday, 15 August 2002 09:24:12 UTC