Watson provides the reader with extensive and rather complete fragments of Common Lisp code listings for various activities that are common in artificial intelligence research and practice. The book covers topics from neural networks, natural language processing, expert systems, search, chaos theory, and fractal images. As the title suggests, the book is primarily targeted at people who develop their programs and systems in Common Lisp. The book is not just a collection of code printouts, though; it does attempt to describe the theories and techniques that underlie the code.

Part 1 describes and provides some basic printing and plotting graphics functions that will be used by code that follows. The code presented here is for a version of Common Lisp running on a Macintosh. Do not expect it to work verbatim on any other system, although the bulk of it will indeed work on any Common Lisp platform. Part 2 covers artificial neural networks. Various techniques are described, including delta rule networks (feed-forward and backward error propagation), Hopfield neural networks, and adaptive neural networks. Sections describe techniques for speech recognition and handwriting recognition, using the delta rule network model. This part of the book is comprehensive in terms of both the technical descriptions and the code listings that are provided.

Part 3 covers natural language processing. The focus is on augmented transition networks (ATNs) and their use in building programs that can process natural language. Watson gives a full description and accompanying code for an animated ATN parser, followed by an example of using this parser to build a natural language interface to a library database. Part 4 discusses topics in expert systems. The emphasis is on forward chaining and backward chaining production system (rule) interpreters; a primitive pattern matching capability is also provided. Part 5 is on heuristic search algorithms. The A* search algorithm is described and applied in a chess playing program. Again, extensive code listings complement the descriptions. Parts 6 and 7 cover chaos theory and fractal images. The intent of these last two sections seems to be to provide an exposure to two vast new areas of computational research. The descriptions are brief, and the code provided is essentially graphics generating code to illustrate the potential of real chaos theory and fractal image programs.

The topics best covered and done justice to are those in Part 2, “Artificial Neural Networks,” and to a certain extent those in Parts 3, “Natural Language Processing,” and 5, “Heuristic Search Algorithms.” The other parts of this book are more introductory. Nonetheless, the book is peppered with thought-provoking exercises throughout, which makes reading it enjoyable. Undeniably, the code in the book is useful, especially if you program in Common Lisp. For readers who do, this would be a valuable reference work.