- From: Michael Kifer <kifer@cs.sunysb.edu>
- Date: Wed, 24 Aug 2005 19:59:38 -0400
- To: jos.deroo@agfa.com
- Cc: public-rule-workshop-discuss@w3.org
> >> [...] > >> > >>> Allow me to set things straight please. The inference rules > >>> that are often (incorrectly) referred to as NAF ***always*** > >>> have scope. I am not familiar with any notion of NAF that > >>> doesn't refer to a scope. Typically the scope is IMPLICIT, > >>> but it is ALWAYS WELL DEFINED. So, NAF is a form of SNAF > >>> where the scope is defined implicitly, but always rigorously. > >>> All Prolog systems that I am aware of are like that. > >> > >> Suppose that in my Prolog program I use lots of consult of > >> resources on the web and also lots of assert and retract all > >> conditioned by the state of the web, then how can you possibly > >> say that that scope in which I'm deriving evidence while using > >> negation as failure is ALWAYS WELL DEFINED?? > > > > At any moment when NAF is computed the set of rules and facts > > is known to the inference engine. What is your problem? > > It is indeed known internally to the engine, no doubt. > Would like to see that "ALWAYS WELL DEFINED" *outside* > the engine to support "proof supporting a conclusion". First, what you are talking about is way ahead of what W3C has in mind. Knowledge base dynamics is not even on the horizon as far as the current discussion is concerned. Second, what do you mean by "well defined outside the engine"? If your KB is dynamic then you can still provide explanations to proofs (if this is what you meant), but such things become more difficult, of course. If you want to determine the state of a dynamic KB at a particular point of execution ahead of time then it is undecidable for reasonably expressive languages (recursion is all you need for that). --michael
Received on Thursday, 25 August 2005 00:00:10 UTC