Re: Proposal and Test cases

From: Michael Smith <msmith@clarkparsia.com>
Subject: Re: Proposal and Test cases (Re: skolems: visible differences?)
Date: Thu, 17 Jan 2008 10:21:53 -0500

> 
> 
> On Thu, 2008-01-17 at 14:55 +0000, Jeremy Carroll wrote:
> 
> > so entailment with existential semantics is undecidable.
> > 
> > I don't see this as a problem if we do not specify a conformance label 
> > for entailment. We perhaps ought to have a disclaimer concerning 
> > entailment and non-entailment tests to that effect.
> 
> This suggests to me that you consider interoperability with respect to
> consistency tests more important than interoperability with respect to
> entailment tests.  Is that in fact the case?  If so, why?
> 
> I understand that the 1.0 test document only defined semantic
> conformance with respect to consistency, but don't know what motivated
> that decision.  Can you (or others with webont history) provide some
> pointers to background?

One reason for this is that a sound and complete OWL DL consistency
checker can easily be used to build a sound and complete OWL DL
entailment reasoner.  A move to allow non-tree bnode structures in
consequents in OWL 1.1 DL under the existential reading of bnodes would
mean that this is no longer true, I think, and would require a separate
conformance label for entailment.

Peter F. Patel-Schneider
Bell Labs Research

Received on Thursday, 17 January 2008 15:28:59 UTC