- From: Jeremy Carroll <jjc@HPLB.HPL.HP.COM>
- Date: Thu, 21 Mar 2002 13:54:28 -0000
- To: "Brian McBride" <bwm@HPLB.HPL.HP.COM>, "Dan Connolly" <connolly@w3.org>
- Cc: "Pat Hayes" <phayes@ai.uwf.edu>, <w3c-rdfcore-wg@w3.org>
> At 22:28 20/03/2002 -0600, Dan Connolly wrote: > [...] > >Oh... but more substantively: > >the hassle of dealing with an infinite number > >of rdf:_1 rdf:_2 things. Brian: > Hmmm. Its been proposed we make all those subproperties of some common > super property, e.g. rdf:member. Is there anything we are likely > to want > to say about any rdf:_nnn that is not true of rdfs:member. Would > that deal > with this, or am I missing the point? The latter :( ... The problem is that the daml:list construction with an explicit end is viewed as safer vis-a-vis a closed world assumption that is being made. e.g. _:a rdf:type daml:oneOf [ <foo> ] . entails _:a daml:same???As <foo> . whereas if the [ <foo> ] can have other members that we don't know about then _:a might be one of them. If you know Prolog this is the contrast between: member(A,[foo]). and member(A,[foo|_]). Jeremy
Received on Thursday, 21 March 2002 08:55:06 UTC