378
Book Review
all material in terms of how to identify and develop potential managers and what other questions need to be answered about the field. A series of appendixes and selected references and index form the end matter of the work. All in all well worth its modest price tag.
C. EDWARD EVANS UClA
Graduate School of Librar) and Information Science Universityof Cul~f~rn~a Los Angeles, CA 900~4
Coding and Information Theory. RICHARD HAMMING Prentice-Hall, 239pp., $19.95, ISBN 0-13-~39139-9
Englewood
Cliffs, New Jersey. 1980
Coding is the process by which a message presented in one set of symbols is translated into another. more convenient set of symbols. Familar codes include the ASCII code for representing text in computer manipulated form, Morse code, and, increasingly, bar codes. A good code is easily translated and is well adapted to the tasks set for it, Creating such optimal codes presents a substantial intellectual challenge,‘and this book covers some of the more significant and interesting responses. a number of which are due to the author of this book. Though some clever, special purpose codes are included (for example, the ingenious Gray code for translating analog data into digital form), the main tasks considered are: (1) how to create a code in which errors are detectable and perhaps even correctable; and (2) how to encode information most compactly. The book begins with a highly readable, essentially non-mathematical discussion of these preblems; the second half of the book devolopes Shannon’s theory of informatjon and particularly its application to the coding problem. Hamming is a highly gifted. intuitive mathenlati~ian. and in this book presentsthe world of coding from his own vantage point. The insights provided nicely complement the more formal, less personalized but more systematic treatments of the standard textbooks. The more mathematical sections ending the book will be most appreciated by someone who already understands the theory. This is a very compact book, with a lot of information, judiciously chosen and contained in a rather small space. The main topic that struck me as appropriate that was missing was cryptology-a type of coding of increasing importance, with substantial intellectual content, and also owing much in its theoretical development to Claude Shannon’s theory of information. I had the impression Hamming must have enjoyed putting this little book together: anyone interested in coding will surely enjoy reading it.
management of Library Networks: Policy Analysis, Impfementati~n, and Control. WILLIAM B. ROUSE and SANDRA H. ROUSE.Wiley, New York, 1980.228 pp., $25.95, JSBN 0-47l-OS~34-4,LC 80-12644, Zh74.7.R68.
The manager of a library network must be able to think big about the many small details which determine the performance of a network and which may be constantly changing. How to manage such changes and handle growth without becoming paralyzed by an overload of information and decision making is the main theme of this book. This theme has influenced the design of the book itself which is written so that some readers can use it to gain a general understanding of the nature and function of library networks, others can probe more deeply into the critical management problems that networks engender, and still others can get a detailed understanding of methods available for the control and monitoring of network performance. The book is based on a five-year project at the Illinois State Library and three case studies are included to show how the methods proposed can be used in different kinds of network situations.
Book Reviews
379
Chapter 1 describes various kinds of existing networks and discusses the motivation for their development from both an economic and a technical perspective. Chapter 2 delineates the management issues and recommends a prototype design for an on-line management information system. Chapters 3 and 4 provide general guidelines for developing quantitative measures of the performance of library networks. In Chapters 5 and 6, the authors present their major thesis that both the performance and structure of a library network depend essentially on the policies used to route requests for service among resource libraries. The authors argue that the determination of optimal routing policies requires an understanding of the status of the entire system and, therefore, policies should be made from a central informational viewpoint. It is in this regard that computers and communication technology make their greatest contributions to network management, and their importance in routing decisions should be the main basis for justifying such equipment. Formulas are derived in Chapter 5 for showing how routing policies affect (a) the probability of delay in service, (b) the average delay time, and (c) the average cost of service. These three measures are combined in Chapter 6 by first considering a multi-attribute utility function and then developing a measure of the requestor’s cost-to-benefit ratio which is used to rank order the available resource libraries for a given request and to determine the optimal routing for each type of request. Chapters 7 and 8 give detailed instructions for applying the proposed routing methodology to real networks, and special consideration is given to data collection problems. In Chapter 9, three large case studies are described which show how the methodology was used in different kinds of network environments with different kinds of management problems. Further applications are discussed in Chapter 10 along with an overview of the more general implications of this study for the management of library networks. The book is well-written and useful in promoting a better understanding of the important interplay between analytical and managerial viewpoints. There is a large up-to-date list of references and both an author and subject index. School of Industrial Engineering Purdue University West Lafayette, IN 47.907, U.S.A.
FERDINANDF. LEIMKUHLER