The first chapter of this book outlines the motivation for the use of associative processing: “With state of the art technology it is possible and desirable to build computer systems based on hierarchies of content-addressable memories]” This is followed by a short chapter classifying storage access philosophies. Chapter 3 contains a comprehensive survey of the literature covering most aspects of associative processing. Chapter 4 introduces some associative concepts into standard PASCAL, enabling the programmer to define and operate upon associatively stored data structures from within a conventional language. The low-level primitives required to implement such a language are outlined in Chapter 5. The actual design of the hierarchical associative processing system is presented in Chapter 6. Chapter 7 contains a morass of implementation detail concerning the low-level primitives defined in Chapter 5. The subject of memory management, a fundamental feature of any hierarchical storage system, is discussed in Chapter 8.

The purpose of this series of books is to report new developments in computer science research and teaching--quickly, informally, and at a high level. This book satisfies these requirements very well. In addition, Chapters 2 and 3 form a useful introductory tutorial and literature survey in the area of associative processing.

The large quantity of implementation detail in Chapter 7 is likely to be of interest to only a small number of readers. In contrast, the problem of memory management in a multilevel associative system is of much wider general interest; unfortunately, it is dealt with in a relatively brief manner.

The best features of this book are the classification of storage access philosophies and the survey of previous work, which, by themselves, would make an excellent introduction to associative processing. However, the design presented in this book is only a “paper” design. Notable among the list of omissions is the lack of simulation studies, and the lack of any estimation of efficiency or performance. Multilevel storage systems are clearly required in any situation where large databases are used. However, this book contains no justification for its inherent assumption that performance will be improved as a result. Conventional multilevel storage systems exploit some form of data-locality to reduce the movement of data between levels. This book does not clearly state how locality is to be exploited in the system under discussion.

Little background is required in order to understand the principles contained in this book, since it begins with a good tutorial introduction. Students of computer architecture would benefit from reading at least the first half of this book. The list of references is comprehensive, although it does not contain any references after 1982. The early chapters also contain quite a large number of typographical errors.