- From: Jonas Liljegren <jonas@paranormal.o.se>
- Date: Sat, 18 Dec 1999 15:11:24 +0100
- To: Pierre-Antoine CHAMPIN <champin@cpe.fr>, RDF Intrest Group <www-rdf-interest@w3.org>
Pierre-Antoine CHAMPIN wrote: > > But what if this is to be read as a statement ?? > What if I read both statements > > [person:uk:nx9300001]--boss--> [person:uk:nx9300002]--car--> [vehicle:uk:ABC1234] > [person:uk:nx9300001]--boss--> []--car--> []--price--> "30,000 ukp" > > if I know that I (nx9300001) have only one boss, and that person:uk:nx9300002 has only one car, > then I can infer [vehicle:uk:ABC1234]--price--> "30,000 ukp", > but that's a lot of IFs ! > > And neither property-value-unicity nor node-unification are adressed by RDF recommandation. > > That's a really interesting point, but still a big issue, IMHO Yes. These things can wait until later. But we could prepare for this by handling unknown URIs diffrent from known URIs. That has an implication on how to generate URIs from anonymous resources and how to handle those URIs with foreign origins. This leads to that the generated URIs doesnt have to be globaly consistent. -- / Jonas - http://paranormal.o.se/perl/proj/rdf/schema_editor/
Received on Saturday, 18 December 1999 09:11:49 UTC