[
Lists Home |
Date Index |
Thread Index
]
Right. The cat can always tell you if thing one or thing
two are what you want, or even that you may have to ask
thing 3. You are layering the resources such that they
resolve to "something" even if only to return a document
telling you where to look next or not to look there.
That is fine. Now the system is behaviorally consisent
and that is what is wanted.
It is only Schrodinger's Cat if it is in the box
and can't be asked or if the energy/time required
to answer changes the answer. It is the nature
of the resource that determines that.
len
From: Bill de hOra [mailto:bill.dehora@propylon.com]
> -----Original Message-----
> From: Bullard, Claude L (Len) [mailto:clbullar@ingr.com]
>
> A more contentious issue is that they don't have a unique
> referent and therefore, are troublesome for RDF.
Just build the semantic web on top of RDDL.
The problem is never going to go away in a distributed naming system.
But we can just do simple behavioural layering; we allow RDF reasoners
who need more info or are confused to throw exceptions that get handled
by web machinery. Check out
<http://citeseer.nj.nec.com/muller93agent.html> or even Peter Stone's
robocup to ways of doing it.
|
