If UML is OWL's friend, how should we play together?

The subject line of this email is inspired by discussion at the OWLED 
Panel Discussion yesterday, titled "How might OWL fail?"

It was an excellent panel discussion, moderated by Sean Bechhofer, with 
Nick Drummond, Tom Heath, Michel Dumontier and Carsten Lutz as 
panelists.  (http://www.webont.org/owled/2008/program.html)

A key issue identified, which Tom Heath summed up succinctly in his 
wrap-up, is that OWL can fail by failing to show its relevance.  Despite 
nearly five years passing since becoming a W3C recommendation, OWL still 
does not have broad adoption and visibility within the mainstream IT 
community.  It still needs to show its value in order to assure its future.

One of the topics discussed was the possible "competitors" to OWL, 
including Common Logic and UML.  Bijan Parsia argued that both are 
really OWL's friends, as OWL offers additional value to each.

I agree wholeheartedly with Bijan, especially in respect of UML being a 
strong potential ally for OWL.  UML has broad adoption within the 
mainstream community, and in recent years has focused on moving into 
executable UML.  It already has UML Class diagrams which share features 
with ontologies.  It lacks a formal semantics, but has recently worked 
on defining an Ontology Definition Metamodel 
(http://ontolog.cim3.net/file/work/DatabaseAndOntology/2007-01-04_ElisaKendall/ODM--ElisaKendall_20070104.pdf)

which defines how various knowledge representation languages, including 
RDF, OWL and Common Logics may be mapped into its framework.

My personal intuition is that the easiest way for OWL to show its 
relevance, and start to be accepted and used by the mainstream IT 
community quickly, is through building bridges with UML that will allow 
those technologists already familiar with UML  (and there are millions 
of them!) to extend easily into OWL and make use of what it has to 
offer.  I believe OWL's offerings complement UML very nicely, and that 
will not fail to be recognized if we are simply able to provide 
visibility and accessibility into them from the existing UML platforms.  
I don't pretend to be an expert on UML, but I know that many of you are, 
and can share your insights on how OWL and UML could best "play together".

So I would like to invite you, the broader OWL community, to discuss 
whether this is a good idea, and if so, what is the most effective way 
to pursue it?

Please post comments, ideas and suggestions.  If someone would like to 
repost on a suitable UML discussion forum, they would be most welcome.

Over to you.....
Anne Cregan

Received on Tuesday, 28 October 2008 08:50:20 UTC