- From: Seth Russell <seth@robustai.net>
- Date: Fri, 9 Aug 2002 10:28:55 -0700
- To: "Seth Ladd" <seth@brivo.net>, <www-rdf-interest@w3.org>
From: "Seth Ladd" <seth@brivo.net> > User hasMembership Bag (content, content, content) > into > User hasMembership content. > User hasMembership content. > User hasMembership content. > That would make programming against the model graph a lot simpler. Problem is, i am told, this is not always the case. For example if {W dc:creator (A, B)} then we cannot in good faith say that {W dc:creator A. W dc:creator B.}. There are lots of relationships that simply do not distribute over sequences; but ,OTOH, there are lots of relationships that do. We discussed this in #rdfig the other day. Somebody suggested that we need a new subClass of properties ... maybe call it DistributesOverBags .... then we say {hasMembership rdfs:subPropertyOf DistributesOverBags} and your inferences should come out ok. .... hmmm ... what namespace wants to claim such a term ? Seth Russell http://robustai.net/sailor/
Received on Friday, 9 August 2002 13:29:28 UTC