- From: Jim Hendler <hendler@cs.umd.edu>
- Date: Wed, 24 Aug 2005 23:51:31 -0400
- To: Dieter Fensel <dieter.fensel@deri.org>, public-rule-workshop-discuss@w3.org
At 17:17 +0200 8/23/05, Dieter Fensel wrote: >Dear Jim, > >everybody that is not foolish wants to have maximal interoperability >of the rule language with OWL. Unfortunately there are some laws of >computation in place that prevent us from including more than a slight >subset of OWL-Lite [1] in a rule language without loosing all the reasons >for using a rule language at all. Rule languages have evolved because >of their efficiency in instance reasoning over large instance population. >Injecting too heavy language constructs stemming from Description >Logic will kill this and therefore any justification in using rules at all. >Indeed, than you can use directly FOL with equality instead since you >are no longer staying a computational interesting subset of FOL. > >More pragmatic, do you really try to enforce all vendors of rule engines >to reimplement their rule inference engines on top of a Description >Logic reasoner. Whereas one company may like this idea all other >vendors may be a bit skeptical about this? > > -- dieter > >[1] Basically Deductive Logic Programming I think I'm beginning to understand where a number of the disconnects are coming from - really depends if you include among "rule language" something like CLIPS and the other descendants of OPS5 and the expert systems days (whcih are widely used), whether some of the transaction monitoring sorts of business rules used in large scale databases count, whether the sort of rule sets for policies that some of the folks at the Workshop used ... these are different in general and some of the folks at the Workshop said in their papers they need some of this (so the things like procedural extensions come from the needs of these sorts of systems) Let me be clear on something - I have not advocated the use of FOL - that doesn't come from me. I also don't oppose it either - from my point of view I'm not terribly concerned with the underlying semantic model of the rules but more with what is expected to be done, and what sort of rules must be covered - use cases motivate me a lot more than this, and I thought the use cases Sandro had in the charter draft were a good start... -- Professor James Hendler Director Joint Institute for Knowledge Discovery 301-405-2696 UMIACS, Univ of Maryland 301-314-9734 (Fax) College Park, MD 20742 http://www.cs.umd.edu/users/~hendler
Received on Thursday, 25 August 2005 03:52:03 UTC