Bottom-up evaluation of HiLog in the context of deductive database systems

Abstract

HiLog is a logic-based language which boasts the expressiveness of a higher-order syntax while retaining the simplicity of a first-order semantics. This work examines the suitability of Horn-clause HiLog as a query language for deductive databases by investigating the feasibility of adapting well-established Datalog evaluation algorithms for the evaluation of HiLog programs. Each of the evaluation algorithms examined in the work is formally described and verified in terms of completeness and correctness. Furthermore, a practical HiLog evaluator based on each algorithm verifies the feasibility of its implementation in a real-world context. It is demonstrated that the Datalog evaluation algorithms do indeed have realistic HiLog analogs. The work also compares the performance of these analogs.