Independence-reducible database schemes

Reviewer: Daniel C. Wahl

Chan and Herna´ndez define key equivalent independence-reducible database schemes, a generalized form of Sagiv-independent database schemes. In addition, they demonstrate that this class of schemes is highly desirable with respect to query answering and constraint enforcement. The paper is well organized. The introduction begins with background material starting with Codd's pioneering work, then provides a quick overview of the paper's contents and major findings. Definitions follow in Section 2. Sections 3, 4, 5, and 6 form the heart of the paper. In Section 3, the authors define the class of key equivalent database schemes and show that it is bounded and algebraic maintainable. In Section 4, they define the class of independence-reducible schemes and prove its properties, including boundedness and algebraic maintainability. Section 5 presents a means of determining whether a given database scheme is independence reducible. In Section 6, Chan and Herna´ndez present a number of significant properties of independence-reducible database schemes. One is that this set of schemes is the same as the class of schemes obtained from decomposing relation schemes in a Sagiv-independent scheme in a dependency-preserving manner. Another property is that the class of independence-reducible database schemes is the largest known class of schemes that have favorable query answering and constraint enforcement qualities when key dependencies are specified. This paper builds carefully on a broad base of two decades of work on relational database theory. All but 2 of the 50 references are used directly in the text.