Refinement is the term used to describe systematic and formal methods of specifying hard- and software and transforming the specifications into designs and implementations. The value of formal methods in producing reliable hard- and software is widely appreciated by academics and workers in industry, despite the fact that certain research areas, such as the application to industrial-scale problems, are still in their infancy. This volume contains the papers presented at the 5th Refinement Workshop held in London, 8-10 January 1992. Its theme was the theory and practice of software specifications, which is the transformation of formal software specifications into more correct specifications, designs and codes. This has been an important area of research for the last 5 years and the workshop addressed specific issues and problems related to it. Among the topics discussed in this volume are: the role of refinement in software development, parallel designs and implementations, methods and tools for verification of critical properties, refinement and confidentiality, concurrent processes as objects, the compliance of Ada programs with Z specifications and a tactic driven refinement tool. This is the latest refinement workshop proceedings to be published in the "Workshops in Computing" series (the 3rd and 4th workshops having appeared in 1990 and 1991 respectively). It will be of interest to academic and industrial researchers, postgraduate students and research-oriented developers in the computer industry.

This volume contains the proceedings ofthe 4th Refinement Workshop which was organised by the British Computer Society specialist group in Formal Aspects of Computing Science and held in Wolfson College, Cambridge, on 9-11 January, 1991. The term refinement embraces the theory and practice of using formal methods for specifying and implementing hardware and software. Most of the achievements to date in the field have been in developing the theoretical framework for mathematical approaches to programming, and on the practical side in formally specifying software, while more recently we have seen the development of practical approaches to deriving programs from their speCifications. The workshop gives a fair picture of the state of the art: it presents new theories for reasoning about software and hardware and case studies in applying known theory to interesting small-and medium-scale problems. We hope the book will be Of interest both to researchers in formal methods, and to software engineers in industry who want to keep abreast of possible applications of formal methods in industry. The programme consisted both of invited talks and refereed papers. The invited speakers were Ib S0rensen, Jean-Raymond Abrial, Donald MacKenzie, Ralph Back, Robert Milne, Mike Read, Mike Gordon, and Robert Worden who gave the introductory talk. This is the first refinement workshop that solicited papers for refereeing, and despite a rather late call for papers the response was excellent.