ITR: Implementation of Functional Logic Languages

Portland State University, Portland OR

Investigators

Abstract

Functional logic (FL) programming languages provide several advantages over declarative languages based on either functions or predicates only. A narrowing engine is an essential component of FL languages. Narrowing engines available to date are all prototypical and almost universally incomplete. The goal of this research is the design and development of a general, efficient and complete narrowing engine for FL languages. This research will define both internal and external representations of executable FL programs---the latter in XML. This approach will decouple the frontends of different compilers and interpreters of FL languages from the narrowing engine. This separation will enable frontends for different languages to use the same backend. By contrast to other contemporary research efforts, the proposed narrowing engine will map narrowing computations to computations in an imperative language. The internal representation of the executable code will be a thin-layer bytecode over the mapping of narrowing computations to computations in an imperative language. This approach promises a good efficiency without sacrificing tasks, such as tracing, debugging and profiling, that are more easily implemented in an interpreter. An integral part of the implementation will be a run-time environment for using the narrowing engine in the backend of a compiler/interpreter of FL languages.