Re: question: datatype reasoning?

On February 5, Frank van Harmelen writes:
> 
> 
> 
> Ian Horrocks wrote:
> 
> > As for as the question of completeness is concerned, I agree with Jim
> > that demanding complete reasoning of this kind for all XMLS/RDF
> > datatypes may be setting the bar impossibly high. I think we should
> > allow implementors to support subsets of the the available
> > datatypes. Maybe we should, as Jim suggests, specify some minimum set
> > of datatypes that need to be supported. A reasoner could then claim
> > completeness if it was complete *for the datatypes it
> > supported*. Users would be able to choose a reasoner that supported
> > the datatypes they needed in their application.
> 
> I think the last few sentences are a very good suggestion:
> an OWL reasoner can claim to be complete w.r.t. a self-chosen set of 
> datatypes. This has the benefit of (a) including data-type reasoning in the 
> spec's of the reasoners, without (b) setting the bar impossibly high.
> 
> A variation of this option could be to demand support for a minimal set of 
> datatypes (integers and strings were mentioned). I'm neutral on whether to do 
> this or not.
> 
> To repeat a question by Jim: where would such a statement go in our docs?

We agreed at the Manchester f2f that this should go in the test doc:

   RESOLVED: The test document should specify the conformance clauses for
   OWL Lite, DL, and FULL documents

and I note that the Test doc does indeed contain a section on OWL
reasoners. This would be the obvious place to add a statement about
support for datatypes.

One *SERIOUS PROBLEM* is that the existing statement is
incorrect/inadequate. It should say that a reasoner is unsound if it
*either* shows an entailment in a non-entailment test *or* shows a
non-entailment in an entailment test (and similarly for consistency).
A incomplete reasoner is one that may return a "don't know" answer.

Ian


> 
> Frank.
>    ----
> 
>

Received on Thursday, 6 February 2003 11:00:21 UTC