Logic programming with goal-directed search

A goal-directed logic programming system is presented, and an experimental implementation is described. It differs from other logic programming languages in that a goal that has just failed has certain of its variables unbound and is immediately re-evaluated. This control mechanism gives priority to the failed goal until it either succeeds or fails for good. The modifications required are considerable since they require freeing the control structure of logic programming from its emphasis on backtracking and on the order of unbinding and rebinding variables that it entails. The backtrack mechanisms normal to logic programming lack goal direction and are therefore not part of the system. Examples are given to show the advantages of the system presented over other approaches to logic programming