Re: OWL-S Reasoning -- Instances vs. Classes

Hi all,

as we did not get any responses to our first (lengthy) e-mail, here's 
another (much shorter) try.

Reasoners like RACER are able to compute taxonomies between (OWL) concepts 
(ontology classes). As OWL-S itself is an ontology of services - how is it 
possible to reason about OWL-S service descriptions (which are instances in 
that ontology) in order to derive taxonomies of services?

Thanks,
Michael & Arnd

At 09.02.2004 17:12, Michael Lutz wrote:
>Hi all,
>
>we are a group of "geographers-turned-computer scientists" who are 
>starting to get our heads around using OWL-S. Specifically we are planning 
>to use OWL-S (Profiles) for service discovery (and later simple composition).
>
>We have first experimented a little with "pure" OWL(-DL). We have modelled 
>concepts as classes and used a reasoner (RACER) to compute taxonomies 
>between those concepts. The resulting taxonomy can be used for service 
>discovery (e.g. return all sub- or superclasses of the requested 
>input/output classes to the requester).
>
>OWL-S, however, models services as instances, which then refer through 
>their IOPEs to OWL ontology classes. Our main questions in this context are:
>How can we use those instances in reasoning for service discovery?
>How does the service requester have to specify her request (class or 
>instance)?
>What is the role of more specific service types that can be derived from 
>the generic OWL-S service profile? (Is it possible, e.g. to specify a 
>specific service type (as a class) and then retrieve all service 
>descriptions (instances) that are instances of that class?)
>
>All this boild down to the question whether the service profile can  be 
>used for service discovery as a whole or whether only its parts (mainly 
>IOPEs) are to be parsed from the description and then be used in separate 
>reasoning processes (e.g. computing one taxonomy for inputs, one for 
>outputs etc. - this seems to be what is proposed in [1], for example)? If 
>the latter is the case, i.e. if OWL-S only serves as a "fancy XML 
>template" (I'm playing the devil's advocate here) what use is the OWL-S 
>"overhead" for service discovery?
>
>If you have any questions regarding mine, please don't hesitate to ask.
>Any help with this (or pointers if this has already been discussed 
>elsewhere) are greatly appreciated.
>
>Michael Lutz
>
>
>[1] KAWAMURA, T., J.-A. DE BLASIO, T. HASEGAWA, M. PAOLUCCI & K. SYCARA 
>(2003): Preliminary Report of Public Experiment of Semantic Service 
>Matchmaker with UDDI Business Registry.

--
Michael Lutz
fon     +49 251 83 33724
www     http://ifgi.uni-muenster.de/~lutzm

Received on Friday, 13 February 2004 03:55:47 UTC