A conservative approach to meta-programming in constraint logic programming

Abstract

Constraint Logic Programming [4] extends Logic Programming by generalizing the notion of unification to constraint solving. This is achieved by fixing the interpretation of some of the symbols in the language. The two alternative mechanisms used in the currently implemented CLP systems to achieve this operation are: (1) fix the interpretation before the program executes or (2) fix the interpretation at a point during program execution when it is used in a constraint. CLP(R) [5] and Prolog-III [1] take the first approach whereas CHIP [2] takes the second approach. The problem with the first approach is that interpreted terms cannot be manipulated syntactically. The problem with the second approach is that all constraint operations have to be made explicit and this increases the difficulty of programming. We propose a synthesis of both approaches that overcomes their individual difficulties. Our method is implemented in the ECLⁱPS^e compiler system.