Future database designers using Oracle, IBM, or Microsoft (to name just the most important) may well get their first taste of what a database management system (DBMS) is and how it works from this book. It comes with a complete DBMS called MINIBASE, which can be downloaded from its home page at http://www.cs.wisc.edu/˜dbbook. The goal of MINIBASE is to provide students with a DBMS whose individual components can be studied and played with. Other free online supporting material includes lecture slides and solutions to chapter exercises. All chapters are followed by biographical notes.

This is a systems-oriented text that can be used in an introductory database course for undergraduates, and in a second database course at the undergraduate or graduate level. The material is divided into seven parts, which are further divided into chapters. A diagram shows the dependencies among them.

The introduction presents the relational model and uses SQL constructs to make these concepts concrete. They are the core from which the whole course is developed. The abstract principles of DBMSs are always examined in light of how they can be implemented in a relational model or an extension of it.

Part A covers storage media, buffer management, and various kinds of indexes. Ramakrishnan shows how they are specified in SQL. Part B is about query languages. The coverage of SQL is extensive. In addition to the basics, the author discusses such important concepts as authorization, views, and referential integrity. Part C goes into query optimization and evaluation. The techniques discussed in this context will become more and more important as DBMSs have to deal with program-generated SQL statements.

The topics of file organization, query languages, and query optimization lead naturally to Part D, which deals with database design and tuning. In this part, the entity-relationship model is presented and discussed from a practical perspective. The presentation of each topic is followed by a section about the translation of the concept to the relational model.

Part E deals with concurrency control. To my mind, it is especially suitable to dedicate a special chapter to this topic, since most students get their first exposure to computers through single-user systems. A chapter on crash recovery using the ARIES recovery algorithm completes this part.

The last part covers several advanced topics, including parallel and distributed databases, deductive and active databases, object-oriented database extensions, and decision support. In the chapter about object-database systems, the author directs our attention to the pitfalls of extending a relational DBMS with objects. He observes that the introduction of types using type constructors “changes a fundamental characteristic of relational databases: that all fields contain atomic values. A relation that contains a constructed type object is not in first normal form!” Other things that may go wrong include dangling references if object-IDs are used to refer to objects and “buggy or malicious ADT (user-defined abstract data type) methods.” This chapter proved to be particularly useful as a background to the corresponding chapter of Bobrowski’s book [1] that I read in parallel. It was interesting to see how a commercial DBMS tries to prevent those problems.

Students who have worked their way through this book will be well equipped not only to understand the basic concepts and industry-standard techniques of DBMSs, but to participate in new developments.