Browse Books

Reviewer: MohammadReza Mousavi

The design of concurrent and distributed systems has been a challenging task since the early ages of computing. This challenging task has gained increasing importance due to the ubiquity of concurrent and distributed systems, ranging from systems and networks on a single chip, to distributed embedded systems, to the Internet. Process algebras have been proposed since more than three decades ago to respond to this challenge, thanks to the pioneering work of people like Tony Hoare and the late Robin Milner. This book is about modeling, verifying, and programming concurrent systems using the communicating sequential processes (CSP) methodology pioneered by Hoare. CSP is more than a plain formalism; it comes equipped with a rich background of design principles, verification methods, and programming paradigms. Most, if not all, of these aspects are covered in the four parts of the book. Part 1, which includes chapters 1 to 8, provides an easygoing introduction to the language of CSP by discussing its operators and specifying several examples in its syntax. The examples range from classical examples, such as dining philosophers and alternating bit protocol, to more creative and fun examples, such as sudoku and road-traffic control. This section also offers a flavor of different semantic models of CSP so that the reader can gain an intuitive idea about the verification methods that rely on the semantic models. This part concludes with an introduction to the possibilities provided by the verification toolset for failure-divergence refinement (FDR). Part 2, which includes chapters 9 to 13, spells out the underlying theory of CSP. First, the operational semantics of the language are presented in both the structural operational semantics (SOS) style of Plotkin, as well as a "combinator" style of semantics (a textual coding of de Simone rules in the SOS style). It is shown that if termination is set aside, any operator with the combinator style of semantics can be defined using a CSP expression (up to strong bisimilarity). Different semantic models for CSP are defined formally in subsequent sections, and finally the algebra of CSP is presented. Part 3, which includes chapters 14 to 17, is dedicated to various application areas for CSP. It is shown that CSP can be used to model discrete-time systems; this is achieved by designating a certain action called "tock" (for the tick of a clock) in the specification. This section also explains the verification possibilities in FDR, the semantic properties of tock-CSP, and its relationship to continuous-timed extensions of CSP. The subsequent chapter presents verification techniques for dealing with large, infinite, and parameterized systems. In Part 4, it is first shown that the shared-variable model of concurrency can be given a CSP semantics. This section then presents a corresponding tool for analyzing shared-variable concurrent programs. The final chapter in this part deals with priority (particularly in the semantics of statecharts) and mobility. The book covers a very wide range of topics, which are not all equally easy to grasp for a graduate student of computer science. Hence, a careful selection of topics in this book should be made for a graduate course; even in that case, using some more reading material may help. Fortunately, the book does provide a few alternative selections and comes with a Web site (http://www.comlab.ox.ac.uk/ucs/) that contains several pointers to the older textbooks and papers cited in the book. For many years, Roscoe's book [1] served as the major reference for learning and teaching CSP. I believe this book will serve as its extended, updated, and improved successor for many years to come. Online Computing Reviews Service