Re: [OWL-S]Avoiding Lists

> >[Drew McDermott]
>> It's not clear  whether you mean
> > 
> >    What are the advantages ... that will outweigh consistency checking
> >      and classification?
> > or [the other thing Florian might have meant]
> > 
> [Florian Probst]
>  that's what I would like to know.

The current goal is to pull Owl-S back to the Owl-DL corral.  The
details have been entrusted to the capable hands of Bijan Parsia.

> > In either case, you seem to be assuming that consistency checking and
> > inference are unavailable or ineffectual in Owl-Full?  Am I correct?
 
> [Florian Probst]
>  correct :-)
> What I would like to learn is why OWL-S is developed using OWL-Full
> constructs rather than only using OWL-DL constructs.
> Can I infer from your question, that consistency checking and
> classification is possible with OWL-Full?
> regards

I don't know if classification is possible in Owl-Full, because I
rarely think in terms of classification.   I mentally translate Owl to
predicate calculus, and then ask how hard various inference problems
are in that framework.  Apparently the answers are:

Consistency checking: we have at least one example of finding
inconsistencies in a DAML+OIL ontology, by Richard Waldinger and
company using the Snark first-order theorem prover.  Of course
consistency checking in first-order logic is undecidable, but on the
other hand you don't have to be in a big hurry for the results, so if
a theory is inconsistent chances are good that a general-purpose
theorem prover will figure that out in less than a week.

Inference: *Many* kinds of inference can be done reasonably fast using
a first-order inference engine on Owl-Full-style axioms.  What do you
want to do?  

Classification: In its most general form the classification problem is
determining whether for some term R and a predicate Q, Q(R) holds.  Is
that right?  So any problem at all is a classification problem in a
general enough logic.  (The goal G is equivalent to Q(R) if Q(x) is
defined to be G.)  This is not meant to be a criticism of anything,
merely an observation that the term "classification" has little use
outside a DL.

-- 
                                             -- Drew McDermott
                                                Yale University CS Dept.

Received on Friday, 6 February 2004 15:05:37 UTC