From: Sandro Hawke (sandro@w3.org)
Date: 08/12/03
> Mike: > would like to do b-nodes; will develop some test cases; > painful not to permit this > > Sandro and Benj: > but that's head existential, and thus not Horn; and permitting just in body > would probably be confusing to users If I said it's not Horn, I misspoke. I meant it's not datalog. While existentials in the head is *technically* not Horn, my understanding is that the conversion to Horn via Skolemization is such a common technique that we might as well consider it Horn. It's funny, I was pestering Benjamin and rdf-logic with this question in March 2001 (before the term "b-node" was even coined): http://lists.w3.org/Archives/Public/www-rdf-logic/2001Mar/0075.html http://lists.w3.org/Archives/Public/www-rdf-logic/2001Mar/0079.html etc I wasn't very clear at the time about the my use case, but I think it was about vocabulary migration, where you want to migrate from a predicate like grandparent to one like parent. It does seem common in RDF modeling to say you could either have (in n3) a) x:sandro y:friend [ y:name "Eric" ]. or b) x:sandro y:friendName "Eric". and we need bNodes in the head of rules (or Horn, of course) in order to express that these two forms are equivalent. Having them for inferring lists is also important, too, of course. -- sandro
This archive was generated by hypermail 2.1.4 : 08/12/03 EST