Barrett and Beerel have put together a practical and useful guide to the design, construction, and application of expert systems. They intend to answer three questions for the business professional who is considering expert systems: (1) how are expert systems useful; (2) how can profitable applications be identified; and (3) how should expert systems be designed, built, and used. They answer the last question effectively, but Beerel wrote a book in 1987 [1] that answers the first question better, and although the authors address the second question, they do not answer it completely. Readers should have had some exposure to business information systems development or use.

The authors do not cover any new ground. Many good books on expert systems are available today, including this one, but the standard is still Donald Waterman’s classic [2]. The authors’ presentation is as good as any I have seen to date. They include the requisite chapters on the inner workings of expert systems, knowledge acquisition, knowledge representation, tools, and hardware. They also provide a glossary, an information source list, an index, and several application lists, which include a small amount of development cost information. The book contains a useful section on building expert systems. Generally, Barrett and Beerel have presented lots of practical knowledge and wisdom in a straightforward and honest fashion. They make their points crisply and concisely through effective use of checklists and sentence fragments. Again, they include nothing really new, but they present all of their material well. The authors obviously know what they are talking about.

One of the more novel things the authors do is to confine the typical discourse on the XCON and MYCIN expert systems to a small section in the appendix on historic applications. Unfortunately, their section on other, more current applications is also a small appendix. If you do not yet have a friendly treatise on expert systems, this book would be a great candidate, especially if you combined it with Beerel’s previous work [1].