Join Evaluation Schemata for Constraint Handling Rules (chr)

CHR is a committed-choice language consisting of guarded rules that rewrite constraints into simpler ones until they are solved. CHR can de-ne both simpliication of and propagation over user-deened constraints. Simpliication replaces constraints by simpler constraints while preserving logical equivalence. Propagation adds new constraints which are logically redundant but may cause further simpliication. The constraints acted upon by CHR are represented in a constraint store, much like tuples in a conventional data base. The evaluation of the condition part of CHR consists of nding sets of tuples in the constraint store which match the heads of a rule, resembling the relational algebra join operation. In the binding environments of these tuple sets, the applicability of the rules is decided by evaluating the guards. Rule applications may add and remove constraints. The anticipated interaction under an immediate update view leads to a small number of prototypical join evaluation code templates. Their Prolog realization is discussed in this paper.