- From: Jos De_Roo <jos.deroo.jd@belgium.agfa.com>
- Date: Sat, 24 Aug 2002 16:32:45 +0200
- To: Sandro Hawke <sandro@w3.org>
- Cc: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, sean@mysterylights.com, seth@robustai.net, www-rdf-interest@w3.org
> > Well, I prefer Peter's option 1 as we already have it in N3 e.g. > > P log:implies C . > > which is actually > > ~P log:or C . > > and indeed neither ~P nor C are asserted, just their disjunction. > > I really don't see any problem in selfreference as long as you > > are not asserting your own truth-conditions, or as Pat once wrote > > You can do that in N3 because it has a nested syntax, which allows you > to use formulas which are guaranteed to be non-self-referencial (since > they appear as {...} syntactic structures, which clearly cannot > contain themselves). But if you flatten those syntactic structures > out into triples, so we can put it in RDF, then you're allowing people to > say > P is_the_sentence (P log:implies C) > and > P is_the_sentence (P is false) > which is a problem. (I think the problem is addressable, in a way > which doesn't hamper ordinary use, as I suggest in this message's > grandparent. But it does require some complexity in ones > definitions.) or as long as you are not asserting your own truth-conditions which could be easily detected (e.g. what we do in Euler and in that case we simply strike those and so become incomplete) -- , Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Saturday, 24 August 2002 10:33:23 UTC