Re: [OEP] n-ary relations and topic maps

Rather than using annotation properties, in a project I'm working on we've 
defined a "meta ontology"  that expresses the semantics of the style of 
reification we use.  It is in OWL-full, as it uses the OWL vocabulary 
(like rdfs:property).  We don't use this "meta ontology" for reasoning, 
rather we consider it the formal semantics of reification.

In the application, we have a part of the ontology expressed in pure OWL 
DL, which we do use for reasoning, and a seperate ontology that contains 
the axioms that tie the DL-defined elements (the classes and properties) 
to the "meta ontology" - this includes defining which classes in the DL 
portion are reified relations.  The mapping axioms are also in OWL full, 
as they require classes as instances.

We are finding this a very elegant approach, and in particular we have to 
interoperate with KIF-based reasoning systems that use actual n-ary 
predicates corresponding to reified ones on the OWL side.

I'm pretty sure we'd be willing to share this (I'm checking), and maybe we 
can consider this instead of annotation properties.

-Chris

Dr. Christopher A. Welty, Knowledge Structures Group
IBM Watson Research Center, 19 Skyline Dr., Hawthorne, NY  10532     USA   
 
Voice: +1 914.784.7055,  IBM T/L: 863.7055, Fax: +1 914.784.7455
Email: welty@watson.ibm.com, Web: 
http://www.research.ibm.com/people/w/welty/



Alan Rector <rector@cs.man.ac.uk> 
Sent by: public-swbp-wg-request@w3.org
12/09/2004 01:01 PM

To
Lars Marius Garshol <larsga@ontopia.net>
cc
public-swbp-wg@w3.org
Subject
Re: [OEP] n-ary relations and topic maps







I think this issue is more general than Topic Maps.  In both OWL and RDF, 
the
question is whether there should be a set of annotations to indicate that 
a
particular set of constructs is actually part of a larger pattern.  The 
n-ary
relation case is most obvious because other representations support n-ary
relations natively.  Users consistently ask to see ontologies at a higher 
level
of abstraction.  That's part of what patterns were about.  To achieve 
this, we
need annotations to indicate the patterns.

Would a sensible procedure be to seek to establish a namespace suggestion 
for
such annotation properties?  Is there any mechanism for doing so?

Regards

Alan


Lars Marius Garshol wrote:

> * Bernard Vatant
> |
> | For the record, the idea to add a note about TM to the n-ary
> | document arose before the TM Task Force was formed.
> |
> | I agree with Lars and Fabien.
>
> I'm glad to hear that.
>
> However, I think we now lost the point I *was* making. The n-ary
> document is presumably going to make recommendations about how to
> express n-ary relationships in RDF, and the RDFTM documents should
> note this and
>
>   1) explain how an RDF->TM converter can detect such relationships
>      and correctly convert them to n-ary topic map associations,
>
>   and
>
>   2) explain how a TM->RDF converter can represent n-ary topic map
>      association in RDF.
>
> It now seems like this is mostly something for the RDFTM TF to think
> about, but possibly with some consequences for the n-ary relations
> draft.
>
> --
> Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
> GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >

--
Alan L Rector
Professor of Medical Informatics
Department of Computer Science
University of Manchester
Manchester M13 9PL, UK
TEL: +44-161-275-6188/6149/7183
FAX: +44-161-275-6236/6204
Room: 2.88a, Kilburn Building
email: rector@cs.man.ac.uk
web: www.cs.man.ac.uk/mig
        www.opengalen.org
        www.clinical-escience.org
        www.co-ode.org