RE: [SE] Ontology Driven Architectures

Holger, I agree with your point, but I don't think it is in the scope of 
the SE TF, rather it is in the charge of the WG itself.

-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/



"Holger Knublauch" <holger@SMI.Stanford.EDU> 
Sent by: public-swbp-wg-request@w3.org
01/06/2005 12:23 AM

To
"'SWBPD'" <public-swbp-wg@w3.org>
cc

Subject
RE: [SE] Ontology Driven Architectures







> You should know that this has been tried before and failed.
> The problem was that the ontology, as an independent artifact,
> needed to be maintained to stay relevant, but after requirements,
> specification, and design were complete, the ontology was no
> longer strictly needed for maintenance of the software, and
> like documentation, fell out of synch with the software and became
> less useful.  It is extremely difficult to convince people that
> they need to maintain all these artifacts along with the software.

I think in addition to regarding ontologies as a design artifact for
other software models and code, we should not forget that ontologies
themselves can be development results.  By encoding explicit semantics
in ontologies we make them executable by reasoners.  While ontologies
failed to add value in development processes in the past, the situation
is different in the context of the Semantic Web:  On the Web, anyone is
free to publish ontologies directly as OWL or RDF files, with no need
to derive other artifacts from them.  As an example consider the travel
domain (which I described in [1]): Here, simply by publishing an OWL
ontology, the execution logic of another application is modified.  Yet,
there is a link between the application and the external extension
ontologies by means of a core ontology, which is hard-coded in the
application.  However, anyone is free to extend such core ontologies,
and these are independent from any code in the application, but could
also be used in hundreds of other applications.

All that I am saying is that in addition to Ontology Driven Development,
there is also a notion of "Semantic Web Driven Development" which requires
guidelines and tools.  The idea of freely publishing interlinked models
on the web, instead of using them only as intermediate artifacts, will be 
a
killer argument in favor of ontologies.

Holger

[1] http://smi-web.stanford.edu/people/holger/publications/MDSW2004.pdf