Re: rdfs:Class vs. daml:Class ?

>Pat,
>
>I don't have the background to discuss most of the fallacies you 
>mention , but I
>would like to talk about one of them:
>
>Pat Hayes wrote:
>
>>  21. The fact that rdfs:Resource is both an instance and a superclass
>>  of rdfs:Class is a "problem".
>
>One of the things I was working on when I started looking at RDF and 
>DAML was to be
>able to generate RDF/DAML from UML system models. I thought it might 
>be useful to
>consider RDF and DAML from a metamodeling perspective, so that there was a
>correspondence to OMG's layered metamodeling architecture. Then I 
>could translate
>from the MOF in the XMI representation of the UML model, to an 
>equivalent set of
>instances of an RDF metamodel.
>
>But my effort foundered on points exactly like the one you mention: the way I
>understand layered metamodel architectures, for "class A" to be an instance of
>"class B", B must be on a higher metamodel layer than A. But for 
>class A to be a
>superclass of class B, both classes must be on the same layer. This 
>is true in the
>metamodeling world, but apparently not a consideration in the RDF world.

RDF does not impose layering. But I don't quite see why your effort 
foundered. Nothing in the RDF world *prevents* you from imposing such 
a layered view of the world, if you wish to do so. (You could define 
a notion of uml:subClassOf, uml:Instance, and so on, which would be 
subProperties of the corresponding rdf: or rdfs: properties, and so 
could have more restrictive conditions. To ensure metamodel layering, 
you could impose appropriate range and domain conditions. )

The UML kind of view is not incompatible with RDF, but RDF allows 
other perspectives to exist as well. So even if you wish to work from 
such a limited, constraining and arbitrary perspective as that used 
by OMG (sorry, but I couldn't resist), RDF does not prevent you doing 
so. So I see no 'problem' here, unless of course you (like, 
apparently, Pan and Horrocks) believe that the entire world should be 
restricted to your (or, better, OMG's) narrow perspective.

>So whether this is a problem or not depends on what we're trying to 
>accomplish.
>Which is true of everything, I suppose.

Indeed.

Pat Hayes

-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes

Received on Tuesday, 16 April 2002 19:22:44 UTC