Re: Properties no longer required to be resources?

On Thu, 2003-07-31 at 10:37, Graham Klyne wrote:


[...]

> and the axiomatic triple:
> 
> (2)  rdf:type rdfs:domain rdfs:Resource .

[...]

> 
> QED (this time?)

No.  This requires RDFS.

> 
> ...
> 
> Or, more informally, I think that although IP and IR may be disjoint in 
> RDF-interpretations, where there is no available vocabulary for this to be 
> visible or otherwise in an RDF expression, the semantic conditions on RDFS 
> interpretation require that IP be a subset of IR.

The question was about RDF entailment.  Pat states that my test
entailment, which just uses RDF vocabulary, holds in RDF.  I think you
are now saying that it doesn't.

Brian

ps: sorry if you get some duplicate emails.  Evolution giving me more
hassle this morning.

B