The aims of this book are ambitious. As stated in the preface, McCabe intends “to address some of the deficiencies of logic programming languages, particularly with regard to the design and construction of large programs” and “to answer such questions as ‘What is an object?’ and ‘What is inheritance?’ in the context of logic programming.” It is good to be ambitious, and a book relating logic programming and object-oriented programming is timely.

What the author presents is one approach to accommodating object-oriented programming within Prolog. Specifically, he describes the language L&O, which adds classes, instances, inheritance, and self messages to Prolog. After an introductory chapter, the basic concepts and constructs of L&O are given, followed by three example applications of programming with L&O--graphics, the traveling salesperson problem, and a generic packing and planning program. A semantics of L&O is sketched, as is a preprocessor for compiling L&O programs into Prolog. This last chapter is perhaps the most interesting in the book.

The book is stronger on programming than on philosophy. Numerous code examples are provided, so that the main ideas are easy to follow and to copy. Four appendices give details of the programs, which the reader can reconstruct, though that is unnecessary as L&O is available by anonymous ftp from Imperial College.

The author feels compelled to justify that L&O is a logic programming language. To set this up, the first chapter introduces criteria for a logic programming language, then argues that BASIC is not a logic programming language even though it is possible to translate BASIC programs into Prolog. The discussion is mildly interesting, but irrelevant. Another example is the claim that “an object is what we know to be true of it” in contrast to an object being a type. McCabe gives no real justification of this statement, just illustrations of L&O code.

On the whole, I enjoyed the book. What is present is clearly written. It is an account of only one approach to integrating objects into logic programming, however. The connection with object-oriented deductive databases and objects as types is not addressed in sufficient detail to give the reader a view of the alternatives.

Does the book meet its ambitions? That is an exercise for the reader.