RE: Service Discovery

>From what I understood about querying services defined in OWL-S is
basically that a query should be similar to a profile. I feel that
this is not that realistic, as I've pointed out in earlier mails to
this list.
I've also asked whether its possible to change the profile to reflect
differences between service capabilities and goals, similar to what
other initiatives are doing, for eg. WSMO. Such seperation together
with existing constructs such as service category, taxonomy etc can
help in creating more realistic queries.
As u pointed out, in real situations, client requesters and
matchmakers will not necessarily use the same ontologies, and as
Adrian pointed out this is quite a major hurdle. Adding to this, the
defintion of a request based solely on input and outputs, possibly
also preconditions and results, makes it even worst.

I may be way out in my argumenation of this issue, and would also
really like to hear the OWL-S coalition' members' view on this.

Regards

Charlie

-----Original Message-----
From: public-sws-ig-request@w3.org
[mailto:public-sws-ig-request@w3.org]On Behalf Of Jyotishman Pathak
Sent: 28 January 2005 21:19
To: Adrian Walker
Cc: public-sws-ig@w3.org; jyotishman@gmail.com
Subject: Re: Service Discovery


Hi Adrian,

Thanks for your reply. But, I am still confused about how to discover
a service based on IOPE's.

For example, assume that the following are the inputs for the Congo
service (in its profile) as part of an advertisement :-

<profile:hasInput rdf:resource="&congProcess;#ExpressCongoBuyBookISBN"/>
<profile:hasInput rdf:resource="&congoProcess;#ExpressCongoBuySignInInfo"/>

Here, "&congProcess;#" refers to the name space of the "Process" of
Congo service.

Having this, can you please give me an example of a service requesting
query (based on the inputs) ?

In other words, if I were to specify a request query to discover the
Congo service, how would it look like ?

Please feel free to correct me if my question itself is wrong. I will
also appreciate inputs from others in the mailing list.

Regards,
Jyoti.




On Fri, 28 Jan 2005 13:16:39 -0500, Adrian Walker <adrianw@snet.net> wrote:
>
> Jyoti --
>
> You wrote...
>
> >can you give me an example of a service request &
> >advertisement which uses different ontologies ?
>
> Cross-ontology reasoning is complex.  There's an approach to this in the
> retailer-to-manufacturer example in [1].  The examples [2,3,4,5] may also
> provide food for thought.
>
> They can be run in an online system at the same URL.
>
> HTH,  - Adrian
>
>   [1]
> http://www.reengineeringllc.com/Internet_Business_Logic_and_Semantic_Web_Presentation.pdf
>
> [2]  http://www.reengineeringllc.com/demo_agents/DataModelling1.agent
>
> [3]  http://www.reengineeringllc.com/demo_agents/MergeOntologies1.agent
>
> [4]  http://www.reengineeringllc.com/demo_agents/OntologyInterop2.agent
>
> [5]  http://www.reengineeringllc.com/demo_agents/RDFQueryLangComparison1.agent
>
> Adrian Walker
> Reengineering LLC
> PO Box 1412
> Bristol
> CT 06011-1412 USA
>
> Phone: USA 860 583 9677
> Cell:    USA  860 830 2085
> Fax:    USA  860 314 1029
>
>
> At 09:58 AM 1/28/2005 -0600, you wrote:
>
> >Hi,
> >
> >I am a beginner with the SWS technologies and have been doing some
> >reading about the service discovery mechanims, mostly focussing on the
> >OWL-s approach.
> >
> >What I can decipher is that, the "profile" (which is used by the
> >service provider to describe the capabilities & service requestor to
> >specify the request) forms an important part for the service discovery
> >mechanism. Based on this, I have a couple of questions: --
> >
> >1.) Do we assume that the service requestor and the provider use the
> >same ontology to describe the IOPE's ?
> >
> >2.) If yes, then do we achieve "real" interoperability ? I think we
> >are constraining ourselves as it is unrealistic for requestors &
> >providers to have the same/shared ontologies.
> >
> >3.) If no, can you give me an example of a service request &
> >advertisement which uses different ontologies ?
> >
> >4.) I believe the WSMO approach tries to address few things about
> >mediation between ontologies. Should we do something like this for
> >OWL-s too ?
> >
> >It might be possible that my questions are naive and infact may be
> >wrong too. But, I will really appreciate some feedback on these
> >issues.
> >
> >Regards,
> >Jyoti.
>
>

Received on Saturday, 29 January 2005 04:22:44 UTC