Re: A triple is not unique.

Graham Klyne wrote:

> At 10:07 AM 11/21/00 +0100, Pierre-Antoine CHAMPIN wrote:
> >But the quetion you quote maybe hides a deeper one : "does RDF describe
> >only resources, or does it also describe entities".
> >My guess is that when a resources resolves to a unique entity, it is very
> >tempting to describe the entity rather than the resource -- and I'm pretty
> >sure some RDF users already did.
>
> To put it roughly, I think an entity can be a resource, but a resource is
> not necessarily an entity.

I have troubles with "can be a".  I think it should be "can represent a".

> Now, to try and be a little more precise:
>
> I think the conceptual mapping [RFC2396] associated with a resource MAY be
> with a single entity;  other resources MAY have conceptual mappings that
> involve several entities, or none at all.
>
> Therefore, I don't think using RDF to "describe an entity" raises any problems.

Problems arise where there is a confusion between the map and the territory.  Even
people frequently get confused here.  Referring to [1]:   if i say [you]
--destroy-->[Sub1], hopefully I am just talking about the map.  If I say [you]
--destroy-->[~Seth~], I might be speaking of assisted suicide.  Perhaps the best
way to resolve this problem is to designate whether a property operates on the
entity, or its representation.

[1] http://lists.w3.org/Archives/Public/www-rdf-interest/2000Nov/0269.html

Of course this problem comes up more acutely when we talk about statements; because
then the territory is the map .. (or is it?)... that seems to be the essence of
what is in contention here.

Seth Russell

Received on Tuesday, 21 November 2000 12:18:06 UTC