- From: <jos.deroo.jd@belgium.agfa.com>
- Date: Sat, 5 Jan 2002 00:10:21 +0100
- To: pfps@research.bell-labs.com
- Cc: www-webont-wg@w3.org
> A couple of points with respect to the owl rules and euler.
>
> 1/ Where is the formal specification of what Euler is doing? I'm trying to
> figure it out by example, but a formal spec is needed.
that is indeed very needed, but for the moment we don't have one...
however, I can partly refer to [1] chapters 3 and 4
certainly w.r.t.
o Prolog unification (substitutions, unifiers)
o resolution algorithm
o backtracking
o box trace model
this is of course partly and among other stuff, we also have
o Euler path detection (maybe this is close to Loveland's
"Model Elimination" (I still have to check that in detail))
o "existential introduction rule" generation (w.r.t. RDF bNodes)
o unification of variable predicates (highly experimental and
using inspiration from Hayes&Menzel [2])
o as much as possible from DanC's and TimBL's SWAP [3]
> 2/ The rules in owl-rules.n3 don't correspond to the model theory for
> DAML+OIL or to the axiomatization either, as they generate extra lists
> and classes. (This is fine, of course, if you are not trying to capture
> either of these.)
(we still have a long way to go...)
--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
[1] ISBN 3-540-59304-7 Prolog: The Standard - Reference Manual
-- P.Deransart, A. Ed-Dbali, L. Cervoni
[2] http://reliant.teknowledge.com/IJCAI01/HayesMenzel-SKIF-IJCAI2001.pdf
[3] http://dev.w3.org/cvsweb/2000/10/swap/
Received on Friday, 4 January 2002 18:10:49 UTC