- From: McBride, Brian <bwm@hplb.hpl.hp.com>
- Date: Thu, 21 Dec 2000 15:09:26 -0000
- To: ML RDF-interest <www-rdf-interest@w3c.org>
> > No offense, but that sound twisted to me : > the last sentence uses two different semantics for the > property "occursIn"... > > If "occursIn" applies to the Statement nature of its subject > (which is suggested by its other use, with RS as subject) > then your last triple is true. > > If "occursIn" applies to the Stating nature of its subject, > then your last triple is false, and by the way the triple T > is meaningless ! > > If "occursIn" has both semantics (applies sometimes to > Statements, sometimes to Statings), > then yes, the last triple is ambiguous, but it comes from the > predicate, not from the subject :) > > Pierre-Antoine If I understand correctly, then the idea is that the predicate determines which 'mode' of the resource is being referred to. In essence, a contradiction would be possible if there were a predicate which could apply to both 'modes' of the resource, but would yield a different outcome for each mode. To avoid that, the hypothesis is that we disallow such predicates - they are bad - their semantics are not clear precisely because they do not identify which 'mode' of the resource they apply to. Hmmm. I'm going to have to think about this some more. It reminds me a bit of Xeno's paradox. I'm instinctively uncomfortable with it. I want to think about whether there is a logical flaw, whether it violates web principles or whether its just violates my preferred modeling style. A puzzle to think about over the holidays. Merry Christmas Brian
Received on Thursday, 21 December 2000 10:09:32 UTC