W3C home > Mailing lists > Public > public-owl-wg@w3.org > November 2007

RE: ISSUE-82 (Conrad): Metamodel diagrams in Syntax document should be aligned with OMG ODM

From: Conrad Bock <conrad.bock@nist.gov>
Date: Thu, 29 Nov 2007 16:00:25 -0500
To: "'Boris Motik'" <boris.motik@comlab.ox.ac.uk>, "'OWL Working Group WG'" <public-owl-wg@w3.org>
Message-ID: <1af301c832ca$dda79870$b3200681@MEL.NIST.GOV>


 >  To which document are you referring here: the structural 
 >  specification (http://www.w3.org/2007/OWL/wiki/Syntax) or 
 >  the UML metamodel
 >  (http://www.w3.org/2007/OWL/wiki/MOF-Based_Metamodel)?

The syntax.  The mof-based metamodel isn't on the recommendation track,
as I understand it, but I guess it might be of concern if there was some
other way it might appear to have some sort of W3C endorsement.

 >  I believe that the UML metamodel document is the proper 
 >  document to align OWL 1.1 with OMG ODM.

Both would be.

 >  In contrast, I don't think that this should be reflected in the
 >  structural specification document. There, we used UML just as a
 >  simple light-weight mechanism for specifying an object model. The
 >  main reason for having the object-model based specification was to
 >  ease the life for developers. I have had some positive feedback from
 >  the implementers of the new OWL API on this -- they turned this
 >  object model more of less directly into interfaces of the API. Thus,
 >  I believe it is important to keep this object model simple, and I
 >  don't consider alignment with OMG ODM in this document appropriate.

Yes, and you put very well the purpose of all metamodels, including ODM,
and why OMG defines all its standards as metamodels.  It doesn't serve
the community to have more than one structural specification for OWL,
because it breaks the interoperability for those working at the
structural level.

 >  If the name "UML" poses problems as such, we can simply remove all
 >  references to UML from the structural specification document and
 >  simply call it an "object This isn't the issue.  The spec currently 
 >  model".

This isn't the issue, see above.

Received on Thursday, 29 November 2007 21:00:46 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:00 UTC