Re: [OEP] using rdfs:isDefinedBy [was "Classes as values" first draft]

As I was trying to come up with a new version of the "classes as 
values" document based on all the feedback, it has occurred to me that 
the solution with rdfs:isDefinedBy is actually no different than the 
more general solution with annotation properties (2b and 4 in the 
original document respectively). The only reason the solution with 
rdfs:isDefinedBy is in OWL DL is because rdfs:isDefinedBy is an 
annotation property. It has a class as a value anyway. Is there any 
reason to keep it there? It seems that if you are going to use 
annotation properties to stay in OWL DL anyway, you might as well not 
create the additional complexity by introducing an extra level of 
instances.

Any thoughts?

Natasha