Tim: The toulouse inference machine for non-classical logic programming

Abstract

We present an environment for logic programming languages called Toulouse Inference Machine (TIM). Its meta-level architecture permits the user to define how to compute a new goal from a given one. Our aim is to define a frame as general as possible for creating extensions of Prolog and, in particular, to provide a general methodology to implement non-classical logics. There are three basic assumptions on which our frame is built: first, to keep as a base the fundamental logic programming mechanisms that are backward chaining, depth first strategy, backtracking, and unification; second, to parametrize the inference step, and finally, to select clauses “by hand”. Applications in logic programming and, in particular, in non-classic logic programming are presented: we specify with a few TIM inference rules various extensions of Prolog by non-classical concepts which have been proposed in the literature.