Re: range, domain: Conjunctive AND disjunctive semantics both supportable

At 03:26 PM 10/2/00 +0100, Jan Grant wrote:
>As has been pointed out, my construction gives you exactly what the
>proposed disjunctive interpretation gives you, ie. not a lot! - unless
>your application allows you to make further constraints.

Youu say "not a lot", and in the open world wide web I agree.  But the 
open-world philosophy of RDF doesn't prohibit more constrained "enclaves" 
in which such structures might be of use.  I guess this is one way of 
describing "your application allows you to make further constraints".

As for more richer kinds of constraints, I think it would be good to (a) 
really find out how far we can go with core RDF, and (b) allow that there 
may be multiple frameworks (but hopefully not too many ;-) that build upon 
that core.   I don't think this is at odds with what you would like to see...

>What I would like to see is a general mechanism for expressing more
>complex constraints within RDF. Some of the proposals we've seen in this
>debate (and prior to it) are such useful mechanisms. The next question
>to ask is "do they all belong in RDFS?" or is there room for a rdfs,
>rdfc(onstraint) and rdfl(ogic) layer?
>
>My only concern (hence my use of the slightly perjorative word
>"burden") is that while having a standard mechanism for expressing
>general constraints can only be a good thing, I think mandating such a
>mechanism at too low a level in the RDF family of specs needs to be
>considered carefully. Many people have application-specific semantics
>which are useful, but might not paint the only picture. The trick is to
>(a) get a general consensus as to what the commonalities of the various
>techniques are, and (b) look at (possibly) breaking out the semantics
>into a few layers with more general to more specific application. That
>way we can hopefully maximise the benefit from the not inconsiderable
>intellectual effort that folks are putting into this!


#g
--

------------
Graham Klyne
(GK@ACM.ORG)

Received on Monday, 2 October 2000 12:06:39 UTC