W3C home > Mailing lists > Public > semantic-web@w3.org > May 2008

Re: Ontologies with standard behavior of an information domain

From: Aldo Gangemi <aldo.gangemi@istc.cnr.it>
Date: Thu, 29 May 2008 18:05:34 +0200
Cc: Aldo Gangemi <aldo.gangemi@cnr.it>, Semantic Web Interest Group <semantic-web@w3.org>
Message-Id: <75B8DFDB-7666-48BC-BEE9-16DCFB1F5B4E@cnr.it>
To: Alan Ruttenberg <alanruttenberg@gmail.com>, Zille Huma <zille.huma@upb.de>

Hi Zille, Alan,

I have made some work on representing design patterns for processes,  
services, etc. in the context of extensions to DOLCE. You can check  
the ontologies at:

http://www.loa-cnr.it/ontologies/DUL.owl
http://www.loa-cnr.it/ontologies/PlansLite.owl

A paper at WWW2004 explains how they can be applied to service  
representation as distinct from actual process representation. You can  
find other references from the EU WonderWeb, Metokis and NeOn projects  
on that work.
BTW, consider this is an area where a lot of work has been made before  
the Semantic Web, and it is known that capturing behavioral  
information in an ontology gets to grips with the fact that reasoners  
for ontologies do not deal with dynamics. In OWL you can encode the  
structure of a process or a plan/service (see also specialized  
ontologies like OWL-S, WSMO, etc.), but then you need additional  
components for the dynamic part, e.g. pi-calculus.

Zille's original questions points to domain activity types, which are  
not very frequently represented in ontologies. This is probably due to  
the fact that information like searchResort or bookHotel is usually  
encoded as object properties, not as individuals or classes.

Still this is a very interesting design issue, whose resolution  
depends on what Alan correctly refers to "competency questions": do  
you want to represent/reason over booking relationships between e.g.  
agents and hotels, or on behaviors that consist in some agent that  
books a hotel? or even on individual services that are implemented to  
allow hotel booking?
Zille's question is probably the second or the third, but one might  
also think that the three questions are interrelated, e.g. for  
interoperability.

Aldo

Il giorno 28/mag/08, alle ore 15:06, Alan Ruttenberg ha scritto:

>
> On May 28, 2008, at 8:36 AM, Zille Huma wrote:
>
>> Actually, My interest is to use ontologies in business domain and  
>> then define the semantics of web services on the basis of these  
>> business ontologies. For example, in the given example, the  
>> semantics of a HotelBooking web service can be defined more  
>> precisely with an underlying tourism ontology. Thanks for  
>> mentioning the ontologies that also contain standard behavior  
>> information. I am more curious about how the behavior can be  
>> captured in ontology, i.e., what is the structure of any behavioral  
>> node in an ontology.  What in your opinion is a better way to  
>> capture behavioral information in an ontology, e.g., behavior may  
>> be captured in the form of business process or stand alone  
>> activities, etc.
>
> Hi Zille,
>
> In my own work, I've been using the Basic Formal Ontology (http://ifomis.org/bfo 
> ) as the upper level ontology, which defines processes as distinct  
> from things that are not processes (continuants).  
> Processes(occurents) are dependent, via the has_participant  
> relation, on continants. They have parts, which are other processes  
> that occupy a piece of the space time of the whole process.
>
> The underlying philosophy of representation is called "Realism",  
> which I can best describe as an attempt, when defining terms, to  
> make clear an "audit trail", if you will, to entities in the real  
> world, i.e. an understandable correspondence between what is being  
> defined in the ontology to things that exist or happen actually. If  
> you are interested in reading more, check out http://ontology.buffalo.edu/smith/
>
> In some ways this avoids the question of what is better, since the  
> comparison is of what is represented to what is out in the world.  
> But of course this doesn't answer the full question in practice. In  
> practice you would first want to make clear what you want to be able  
> to say, and then determine what you will need to be able to ask and  
> have answered using your ontology.  Answers to such questions might  
> determine the formalism, or level of detail at which you represent  
> your processes.
>
> As an example, if all you want to do is record something in an  
> Ontology and then read it out, then there is little constraint. If  
> you want it to be able to be merged with other people's work, then  
> there are some. If you want to be able to state general temporal  
> relations between activities and have consistency of your ontology  
> checked, then you can't even do this within the framework of the  
> current  web ontology languages.
>
> Experience in the OBI project suggests that you work early on  
> outlining such "competency questions" for your ontology.
>
> If you give some such competency questions, I could see if I have  
> any experience that might be relevant to your representation issues,  
> or perhaps point you at people that do.
>
> -Alan
>


_________________________________

Aldo Gangemi

Senior Researcher
Laboratory for Applied Ontology
Institute for Cognitive Sciences and Technology
National Research Council (ISTC-CNR)
Via Nomentana 56, 00161, Roma, Italy
Tel: +390644161535
Fax: +390644161513
aldo.gangemi@cnr.it

http://www.loa-cnr.it/gangemi.html

icq# 108370336

skype aldogangemi
Received on Thursday, 29 May 2008 16:05:57 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:22 GMT