Abstract

miniKanren, and microKanren, was designed to be an embedded language in a small subset of Scheme. This aids developers in transliterating *Kanren to a great number and variety of other languages (50+), making logic programming available in many non-traditional environments and helping to bridge a gap between distinct families of declarative languages.

This shallow embedding enables a “quick and dirty” explanation of the meanings of our programs – we explain them in terms of our host language’s features. However, advanced features (e.g., a generic constraint system) and varied host languages make this solution unsatisfactory.

As such, we require more rigorous and host-independent characterizations of miniKanren’s meanings – and hopefully at a higher level of abstraction than source code.

In this talk, I will present our first steps along this path, developing and demonstrating a reimplementation of microKanren amenable to a host-independent characterization, and discuss future steps.