Many practical problems do not have algorithmic solutions. Many practical problems require sifting among large numbers of facts to arrive at solutions. Computer programs that solve problems in the intersection of these two sets must employ heuristic strategies such as selecting and refining hypotheses and be able to efficiently access and manipulate large data stores. These programs are variously called expert systems, knowledge base management systems, and deductive databases. (I do not mean to imply that these terms are completely synonymous, but that the systems they represent are more alike than different.)

Two approaches to achieving knowledge base management systems have been advocated: evolutionary, using existing database and expert systems technologies, and revolutionary, starting from the beginning [1]. While acknowledging the attraction of the latter, Vassiliou “considers tight coupling of a KBS with a DBMS as the most attractive short- and long-term solution” [2]. Following this advice, Robert Lucas has combined Prolog with Mimer, a relational DBMS, to produce an environment for developing expert systems.

Lucas’s exposition of his project is well organized and readable. After an introductory chapter he provides an overview of Prolog, relational DBMSs, and expert systems. He then discusses, in detail, the interface he implemented between Prolog and Mimer. The last half of the book contains descriptions and evaluations of three prototype expert systems implemented with Prolog-Mimer. I recommend the book for general readers who have some knowledge and interest in computers and to those looking for information on viable implementations of expert systems.