This award will continue an ongoing research project to explore the implementability and utility of Equational Logic Programming . The project has already produced a compiler for an equational programming language, in which a program is a set of equations, an input is a term, a computation is a derivation of certain logical consequences of the equations in the program, leading to the output of a term in normal form equivalent to the input. Novel features of the language include a total refusal to compromise semantics for performance, leading to a uniform use of "lazy" evaluation-- a technique in which precisely the information demonstrably relevant to the final output is automatically computed-- and a lazy input/output facility, which gives the consumer of a term dynamic control of its computation. The current version of the compiler uses partial evaluation of intermediate code as an optimized compilers for C. Continuing research will include the study of more sophisticated evaluation techniques to improve performance, the performance implications of applying those techniques to other languages, theoretical advances to extend the class of equational programs that can be compiled, the semantic ramifications of introducing modular programming constructs, and experimentation with advanced user interfaces, including structure editor interfaces.
