Re: missing (and incorrect) RDFS axioms

I thought we decided that rdf:_1 etc. could be attached to things in the 
context of the idea that they "had no semantics" and that specifying 
domains for them didn't really constrain what they could be attached to 
(because domain specs aren't really "constraints" unless some processor 
interprets them that way), but that the "intent" was that they were for 
expressing container membership (if not, why do we call them container 
membership properties?), and if you attached, say, rdf:_1 to a resource, 
you could reasonably infer that that resource was a container.  We seem 
to be going around in circles:

1.  rdf:_01 has domain rdfs:Container
2.  we can't really enforce (1) as a constraint
3.  so we might as well remove it
(?)


Brian McBride wrote:
 >At 14:33 09/11/2002 -0600, pat hayes wrote
 >
 >>> rdfs:member rdfs:domain rdfs:Container .
 >
 >Where did that come from?  I thought we decided that rdf:_1 etc
 >could be attached to things that are not containers.

Whoops. I think I must have read that somewhere.

OK, I can remove that.

Pat

Received on Monday, 11 November 2002 14:02:16 UTC