Re: [regrep] ebXML Registry Profile for Web Ontology Language - a few comments

[ewallace@cme.nist.gov]

Nikola Stojanovic wrote:

>Here are some comments on ebXML Registry Profile for Web Ontology
>Language (OWL), Version 1.1, Draft OASIS Profile, June 13, 2006:
>

...

>1.	Generic

...

>	c.	AssociationTypes naming should be consistent and follow
>ebRIM (UpperCase convention, directionability, ...), Also, if
>"HasProperty" is used for rdf:Property should other owl:XXXProperty be
>prefixed in the same way -> "Has"?

A profile like this is trying to capture the metamodel of another 
language in ebRIM.  Thus profiles should have more relaxed naming rules
then other ebRIM uses. In this case, the use of "Has" as a prefix of 
property is already somewhat misleading.  Properties are first class
entities in RDF/OWL.  

I am not sure of the scope of the above suggestion.  The pattern 
owl:XXXProperty seems to only apply to specializations of property such
as FunctionalProperty.  That would be tolerable if undesirable.  However, 
application of the "Has" prefix to subPropertyOf and subClassOf would result 
in very confusing names indeed.  I would strongly object to such names.


>3.	Section 2 - It would be beneficial to state why this profile
>only addresses OWL Lite.

Yes.  I believe that I also asked for this in the earlier round of reviews.

As I understand it, the reason for OWL lite only support is that OWL Lite 
reasoning can almost be simulated with an ebRIM based repository and queries, 
whereas the additional features of OWL DL and OWL Full just to don't have any 
analog in RIM.  This should, indeed, be explained in the profile.


-Evan

Evan K. Wallace