Re: [regrep-semantic] Groups - New Action Item #0000 Issue: Do we need to choose Ontology representation

[ewallace@cme.nist.gov]

For me, the questions below don't restate the issue, they change it!


Carl Mattocks wrote:
>To restate the issue with (perhaps) easier questions.
>
>Does the ebXML Registry need to support all flavors of the OWL ontology ?

Agree that the question above is one component of the issue.  My answer
to that question would be "yes."

>Does the ebXML Registry need to explicitly (via technical notes ?) support
>all 'Institutionally supported Ontologies' e.g. DCML, OWL

It would be helpful to have a more complete list of examples of what you mean 
here.  Not clear if you mean ontologies or ontology languages?

>
>If the ebXML Registry supported 'All Ontologies that can be modeled using
>RDF/S ' would that address all the known 'Use Cases' ?

RDFS doesn't support my use case.  Even supporting all dialects of OWL
may be too limiting.  Some ontologies may be written in languages which
support more expressivity than OWL.  Some such languages are the Knowledge 
Interchange Format (KIF) and Simple Common Logic (SCL).  One effort to 
formalize UBL is indeed using KIF.  An alternative approach to the 
expressivity problem might want to use a rule language to supplement an OWL 
ontology.  This last is, ostensibly, the Semantic Web approach.

-Evan