- From: Gerd Wagner <wagnerg@tu-cottbus.de>
- Date: Sat, 11 Feb 2006 02:48:06 +0100
- To: "'Francois Bry'" <bry@ifi.lmu.de>
- Cc: "'W3C RIF WG'" <public-rif-wg@w3.org>
> What might well be "more conventional semantics"? There are > two kinds of > model theories used for formalizing the declarative (ie > non-procedural) semantics of programming/rule languages: > > 1. ad hoc definitons fo models eg well-founded and stable model > semantics. I agree that they are not easy to understand. I would not call the stable model semantics an "ad hoc definition", and I would not call the "well-founded semantics" to be a model-theoretic semantics at all (rather it is a proof theory, which is sound with respect to the stable model semantics). The stable model semantics (unlike the well-founded semantics) can be considered a Tarski-style model-theoretic semantics, because it is based on the Tarskian concept of a model. It departs from the standard Tarski semantics in not accepting all models as intended, but only the stable ones which are supported by rules (following the basic intuition of minimal model semantics). This departure is rather natural for a semantics of information and knowledge - as opposed to a semantics of theorems (aspiring for eternal truth). > 2. Tarskian model theories, ie the valuation (= truth value) of a > formula is recursively defined on the structure of the formula. Such > model theories are intuitive and much easier to understand than other > formalisms unsed in formalazing the "meaning" of programs. I'm not so sure if the "direct model-theoretic semantics" of OWL is easier to understand than stable model semantics. We should make a test... -Gerd
Received on Saturday, 11 February 2006 01:51:08 UTC