Program Generation in the Equivalent Transformation Computation Model Using the Squeeze Method

Abstract

In the equivalent transformation (ET) computation model, a specification provides background knowledge in a problem domain, a program is a set of prioritized rewriting rules, and computation consists in successive reduction of problems by rule application. As long as meaning-preserving rewriting rules, called ET rules, with respect to given background knowledge are used, correct computation results are guaranteed. In this paper, a general framework for program synthesis in the ET model is described. The framework comprises two main phases: (1) equivalent transformation of specifications, and (2) generation of a program from an obtained specification. A method for program generation in the second phase, called the squeeze method, is presented. It constructs a program by accumulation of ET rules one by one on demand, with the goal of producing a correct, efficient, and nonredundant program.