Proof procedures for detecting redundancy of database constraints

Abstract

A production-rule-based notation is introduced to express and manipulate database constraints. Two proof procedures are developed to detect the redundancy of database constraints. The direct proof procedure is human oriented and based on natural deduction. Human-oriented proof procedures are important as most data-model design is still done manually and nonredundancy of database constraints is the major determinant of a good design. The indirect proof procedure is machine oriented. It involves conversion of the production rules to the clausal form of first-order logic and resolution theorem proving. Machine-oriented proof procedures are important as many proofs are cumbersome and beyond the time and complexity tolerance of data-model designers. Using the same notation to facilitate both human and machine proof procedures is useful in facilitating human-machine interaction in data-model design. Both procedures are general enough to be used with a variety of data models, including relational and functional models. The applicability of the procedures to a variety of data models is useful in multidatabase environments and in exploring concepts that are common to many data models.