We present in this paper a general inference machine for building a large class of meta-interpreters. In particular, this machine is suitable for implementing extensions of Prolog with non-classical logics. We give the description of the abstract machine model and an implementation of this machine in a fast language (ADA), along with a discussion on why and how parallelism can easily increase speed, with numerical results of sequential and parallel implementation.
http://www.irit.fr/~Andreas.Herzig