Seventeen chapters and an index compose this book. No bibliography is included, but lists of references are given at the end of some chapters.

Chapter 1 clearly explains what is to be understood by fundamental words in the test arena. Those words are requirement, error, fault, and failure.

Chapter 2 depicts the context of testing from the very beginning, that is, the contract. Royer emphasizes contract changes and demonstration versus qualification. The outcomes of demonstration and qualification, respectively, are the error discovery and the buyer satisfaction report.

In chapter 3, the author considers how testing is to be organized out of contract words and how one translates those into testability attributes, which should be measurable and reachable using the resources dedicated to this activity. Planning, description, procedure development, verification, demonstration, and result reports are described clearly, as is the famous independence of test.

Once the preparation is made, the planning steps are to be established. A detailed process for issuing a sound, workable test program is proposed in chapter 4. The author warns against shortcuts, especially in the requirement definition.

Chapter 5 gives a comprehensive explanation of the IEEE Standard for Software Test Documentation 829, dated 1983. It also provides an overview of the DoD standard 2167A (February 1988) with regard to testing issues.

Chapter 6 is a good discussion of the famous “fuzzy requirements,” that is, those requirements whose meanings are unclear, improperly stated, and incomplete. Real-life examples are given, and for each of them the reason why it should be restated is thoroughly explained.

In chapter 7, the author says that to reach satisfaction, developers and testers should identify all failure modes. Then software fault tree analysis is to be used to concentrate the testing where needed using the definition of the failure category.

Chapter 8 presents all the techniques in use for testing--glass box, black box, and finally customization of hybridization. Royer then discusses verification testing, which includes equivalence partitioning, boundary analysis, cause-effect graphing, error guessing, statement and condition coverage, and mutation testing.

Chapter 9 is devoted to the actions needed to convince the buyer that the product is the one he or she ordered and that it will work in the environment it is going to be embedded in. The purpose of chapter 10 is to provide guidance in constructing the software system from its constituent pieces and verifying that the system will satisfy the users. The strategies that might be used are listed.

Chapter 11 describes the goal of formal demonstrations for either custom or contractual products. Royer’s experience is in government projects, and this chapter shows how to handle government-style demonstrations. The reader can follow the whole operation step by step, through the content of the test report.

Chapter 12 is particularly interesting, since it deals with test tools and enables the reader to select the proper tool. Chapter 13 is dedicated to the testing of artificial intelligence software. The novelty is that the problems solved with AI are by definition ill-defined or open-ended. Therefore, designing the test is drastically different than for conventional programming. An eight-step method is advocated.

Chapter 14 considers reliability from as many viewpoints as possible. Mathematical functions such as component connection, execution frequency, fault density, methodology factor, implementation technique influence, and failure rate computation are given for each viewpoint.

Chapter 15 consists of two main sections, one on cost drivers, the other on cost estimation models. The conclusion is that each development group is different. Therefore you should record your own data for each parameter to be used in the chosen model.

Chapter 16 describes a method of status reporting and accounting in use in the defense contracting business. This method is based on earned value.

In chapter 17, a work breakdown structure is proposed for testing activity. This structure follows the generic breakdown structure given in MIL-STD-881A.

This textbook is based on experience gained in testing products developed under government contract. The style is pleasant, and the scheme is easy to follow for engineering-minded people. The best features are the step-by-step progression used for many topics and the plain writing style. The combination of both characteristics will ensure readers that they understand every bit of the text. The references are mostly to classic books or standards documents, with a few citations of recent works (1990) needed for topics such as object orientation, starvation, critical race analyzers, and engineering software under statistical quality control. This book is handy for anyone dealing with the testing of a software product, and will be a convenient companion for the junior or senior testing manager. To get the most out of it, one should have some knowledge of the software life cycle and some understanding of true engineering discipline.

I wonder if the subtitle is fitting, though, since after reading the book it seems that testing may be outside the critical path. This book is indeed an achievement; it is a wonderful demonstration. Though he may not know it, the author has sold the most awkward product on the market: software testing.