Sub-Prolog2SQL compiler

Abstract

If both Prolog and SQL originate from first-order logic, there is obviously some common ground between the two. By analyzing what is common and different in the two languages with regards to syntax, semantics and modes of evaluation, this project attempts to come up with a way of coupling the two languages in order to maximize logical-level and minimize physical-level coupling. By analyzing what is common and different in the two languages this project produces a compiler that has two main methods. The first compiles the Prolog program and the second takes a Prolog query and transforms it into an SQL query using the previously compiled program. Importance has been given to the design and evaluation of the mappings between the two languages by ensuring that the semantics of Prolog constructs are still valid after mappings to SQL have been made. The evaluation of this project discusses the design in terms of performance, coverage and quality. The evaluation confirms the claims that Prolog's performance deteriorates as the number of Prolog facts increase and that the Prolog compiler should have tight logical-level and loose physical-level coupling in order to maximize Prolog's expressive syntax and database's efficient storage and retrieval mechanisms.