Re: URIs / URLs

> You cannot necessarily rely on a URI alone, but must
> evaluate it in some context.

Yep, that is a fundamental axiom of the Semantic Web in general. The
RDF model is opaque in that it does not evaluate any particular
proprietary semantics that one associates with URIs, and that's what
gives it its power, but it's that that makes people conclude that if a
processor runs across:-

     :Blue = :Yellow .

the world will blow up or something. Maybe one particular processor
will, but then to the environment within which this particular
processor operates, this collapse will be useful! In other words, if a
processor needs to recognize that blue isn't yellow, then running
across this particular inconsistency will tell it that something is
up. The system works :-)

It's all about open and closed world systems. cf. [1].

> Does this imply a rdfs:Class of identifier that includes
> context with the URI would be beneficial?

Blargh. URIs are opaque, context is proprietary. You could refer to a
URI as being equivalent to a certin context (see below), but that
doesn't mean that the URI (always has to) identify that context.
Contexts are contexts!

> RDF statements can then be made about instances of
> that class, which in turn identify a resource with a URI
> within some context.

Hmm... I see your resoning here: I think it is best to simply restate
a particular context like so:-

     { :Sky :color :Blue } = :x .
     :y :operatesUnderPretextOf :x .

[1] http://infomesh.net/swnotes/swnotions/

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .

Received on Thursday, 12 April 2001 12:09:55 UTC