- From: Jos De_Roo <jos.deroo.jd@belgium.agfa.com>
- Date: Fri, 3 May 2002 14:20:53 +0200
- To: jjc@hplb.hpl.hp.com
- Cc: "<www-webont-wg" <www-webont-wg@w3.org>
[very well done Jeremy] > Conclusions: > ============ > I think Pat and Peter exaggerate this issue. > Finite set theory is trivial, countable set theory is not much harder. > It's only when things get really big that life gets tough and we don't go > there. > > Pat's proposal does involve explicit comprehension, illustrating that it's > no big deal. > > Peter's proposal does not have explicit comprehension and does not help the > implementator avoid paradox as a result. > > In practice implementators will have comprehension rules, and it is a > helpful clarification for us to specify them. If I may indeed quote http://www.agfa.com/w3c/euler/owl-rules.n3 We currently have it that we both assert list triples in the N3 context of the graph level and in the N3 context of their own subject. Leaving out the former is I think close to Pat's dark list proposal, but, because owl lists are not circular, those top level triples are not in the way. For the rest, we are not able to (dis)prove a paradox and have not seen trouble in case of their presence (we have done tests with merging in paradoxes). -- Jos
Received on Friday, 3 May 2002 08:21:34 UTC