Re: Assumptions about anonymous resources

From : "McBride, Brian" <bwm@hplb.hpl.hp.com>
> My thoughts are that the second rdf:description element
> does not specify what resource it describes. Standard RDF
> processors will therefore not be able to figure out its 
about
> the element in which it is embedded.
> 
> However, you have specific application knowledge that does
> allow you identify the resource it describes. So you 
could
> write your own processor, that would use that knowledge
> and be aware of the identity of resource being described.
> 
> Whilst you could do that, it doesn't feel like a very 
clean
> solution. What advantage do you get out of using
> RDF in this way?

Because you don't always *know* the URI of that resource.
If I tell you that "my wife is a teacher", what can else 
can you write than :

<rdf:Description about="mailto:pachampi@caramail.com">
 <foo:wife>
 <rdf:Description>
 <foo:job> Teacher </foo:job>
 </rdf:Description>
 </foo:wife>
</rdf:Description>

If, from another source, you get an URI for my wife,
assuming that I have only one, then you can identify the 
anonymous resource (and its description) to that URI.

You can not express that meta-knowledge/rule with RDF as-
is, but RDF is a *framework*, so what is so unclean about 
it ?
You can always add vocabularies to make it expressible.
(in that case, a class UniqueProperty, subclass of 
rdf:Property, would be useful -- I'm sure it exists in some 
extension to RDFS).

 Pierre-Antoine
______________________________________________________
Boîte aux lettres - Caramail - http://www.caramail.com