172 JIT 4,3 - September 1989
much practical enhancement of their security's effectiveness. The book is well written, easily referenced, well informed, very practical and forms a very useful introduction to the subject area. L.W.
Intelligent Robotics by Mark H. Lee. Open University Press, Milton Keynes. 1989. ISBN 0335154204. £16.00. 210 pp. What techniques will enable robots to become more intelligent, and what are the main problems and barriers to such development? The author's research field usefully lies between that of the Artificial Intelligence (AI) practitioner and the robotics engineer. He addresses the major challenge of how knowledge can be acquired, retained and used effectively in ways that will create robotic systems embodying the intelligent connection of perception to action. Lee discusses the necessary tension between making Al techniques viable in real-world applications by pr?viding maximum knowledge about the process being automated (thus cutting down search, perception and reasoning), while at the same time using computing techniques that have transfer value into different robot applications, and that can handle harder, unexpected situations in the same task. The chapters move through describing the concepts of sensor processing, perception, knowledge bases, planning and expert systems, and the author shows these to be crucial to future robot systems in manufacturing. He also deals with the understanding and synthesis of natural speech. There are examples throughout, drawn from industrial robot tasks. Lee is also concerned to show how Al techniques will benefit the entire process of Computer Integrated Manufacturing from product design through to automated manufacture, and can also be applied to harder problems in applying mobile robots to less structured environments outside the factory. The book is well organised, remarkably free of jargon, and the author has a talent for fluent, interesting exposition of his subject. This is an upto-date book, well illustrated, and of interest in different ways to the expert and non-technical reader alike. L.W.
Managing the Software Process by Watts S. Humphrey. Addison-Wesley; Wokingham. ISBN 0201180952. 494PI). Humphrey is concerned with providing practical guidance for improving the software development
and maintenance process. He argues that it is here that organisations encounter the most serious problems in information systems implementation, and also where the best opportunities lie for significant improvement. The author is on safe ground here; there is much research evidence from both the UK and USA to support how pervasive and important software difficulties are. 'rhe. book is in five parts, each dealing with a particular aspect of the software development process. The topics are organised in the sequence in which organisations will want to make improvements, and include how to initiate software process change; how to establish basic control over the software process; methods for specifying the development process; methods for quantitative software process control and how to shift the focus from problem fixing to problem prevention. The conceptual framework for the book is partly dependent on the work ofW.E. Deming andj.M. j uran on statistical quality control, here adapted to control of the software development process. Thus the initial phase is to get the process under statistical control, the second to make this repeatable, the third to define the process as a basis for consistent implementation, the fourth to achieve comprehensive process measurement and analysis the fifth to establish continuing improvement 'and optimization of the software process.
The book is well illustrated, obviously based on a wealth of experience and practical in its approach. There is much here that can be applied directly to org-anisations dependent on software development. L.W.
Taming Your Company Computer by Colin Corder. j\1cGraw-Hill, London. ISBN 0077071670. £13.95.253 pp. This is an updated version of the author's 'Ending the Computer Conspiracy'. I t retains the humour, the determination to remain jargon free and the iconoclastic tone of the original publication. Corder deals essentially with information systems project management, including systems design development and implementation stages. The book's refreshingly easy style is aimed at the manager who needs to control a computer project but has little technical knowledge of computers. The author would seem to be largely successful in serving the needs of his market. He establishes the basic need for user managers to specify requirements and describe in logical terms the set of business rules required. Thereafter physical implementation can be left to technicians, though the
