Combining scheduling strategies in tabled evaluations

Tabled evaluations ensure termination for Datalog programs by distinguishing calls to tabled subgoals. Given several variant subgoals in an evaluation, only the first (the generator) will use program clause resolution, the rest (consumers) must perform answer resolution using answers computed by the original invocation. This use of answer resolution prevents the possibility of infinite looping for Datalog programs, which sometimes occurs in SLD. As variant subgoals can be called at different stages of the evaluation, there is an intrinsic asynchronism between the generation and consumption of answers in SLG. Given this asynchrony, implementations of tabled logic programs face an important scheduling choice not present in traditional top-down evaluation: When to return answers to consumer subgoals.