A Meta-Level Extension of Prolog

Kenneth A. Bowen, Syracuse University
Tobias Weinberg



Prolog has many attractive features as a programming tool for artificial intelligence. These include code that is easy to understand, programs that are easy to modify, and a clear relation between its logical and procedural semantics. Moreover, it has proved possible to create clear and efficient implementations. Nonetheless, we perceive several shortcomings. Chief among these is difficulty representing dynamic databases (databases which change in time) and an apparent restriction to backward chaining, backtracking, and depth-first search. Our intent in this paper is to present an extension to Prolog, called metaProlog, which preserves the virtues of Prolog while introducing powerful constructions to attack these problems. This work is a direct continuation of the investigation into meta-level programming in logic begun by Bowen and Kowalski [1982].