A Rewriting Machine and Optimization of Strategies of Term Rewriting

Currently, the paradigm of algebraic programming becomes one of the most important paradigms of declarative programming. Algebraic programming is based on methods of term rewriting [1], which are widely used independently and in programming systems based on other paradigms (in functional, logical, object-oriented, and agent-based programming, in computer algebra and systems of theorem proving). Among systems of algebraic programming that appeared in recent years, noteworthy are Maude, Elan, Cafe-obj, and APS [2–7]. The theory of rewriting systems is developed in detail and is presented, for example, in [8, 9]. The system of algebraic programming (APS) was created in V. M. Glushkov Cybernetics Institute of the National Academy of Sciences of Ukraine in the early nineties. The system is a professionally oriented tool for developing applied systems and is based on algebraic and logic models of object domains. The basic means of programming are systems of rewriting rules. The algebraic programming language APLAN used in the system integrates fundamental paradigms of programming and its semantics can be easily extended with the help of metaprogramming, which is also based on rewriting. Fast term rewriting is conditioned by the use of efficient parallel matching for realization of the rewriting machine considered in this article. The corresponding program was developed by S. V. Konozenko and was realized in the language C in the first version of APS. This article deals with a new version of the rewriting machine developed for a new version of the APS system. The machine is represented in the language APLAN used in the capacity of the language of executable specifications. Such a representation allows one to construct the proof of correctness of the corresponding program, to quickly develop a reliable implementation based on the programming language (C++) being used, and to obtain new modifications of the machine that support various special rewriting strategies.