Performance engineering is a fast-moving field where advances in technology mean that new issues constantly need to be addressed. In response to this, the UK Computer and Telecommunications Performance Engineering workshops were set up in 1985 to provide a valuable opportunity for the discussion and exchange of ideas. They have subsequently become well established as the focus for academic and industrial practitioners from the UK and Europe with an interest in performance and modelling and analysis. This volume contains the 16 papers which were presented at the 7th annual workshop, held in Edinburgh in July 1991. The workshop highlighted various aspects of parallel computing - an area which is attracting an increasing amount of interest - and the work presented in these papers is of particular interest as the contributors used real analysis to evaluate their models. The papers cover an unusually wide range of topics, both practical and theoretical, including advances in queueing theory, common performance problems and their solutions, assessments of available tools and presentations of new theoretical results. The result is an extremely comprehensive coverage of this important and challenging field. This volume provides an up-to-date assessment of work being carried out by performance engineers in the UK and Europe and will be an invaluable reference book for researchers and practitioners wishing to familiarise themselves with the benefits of performance and analysis.

The Z notation has been developed at the Programming Research Group at the Oxford University Computing Laboratory and elsewhere for over a decade. It is now used by industry as part of the software (and hardware) development process in both Europe and the USA. It is currently undergoing BSI standardisation in the UK, and has been proposed for ISO standardisation internationally. In recent years researchers have begun to focus increasingly on the development of techniques and tools to encourage the wider application of Z and other formal methods and notations. This volume contains papers from the Seventh Annual Z User Meeting, held in London in December 1992. In contrast to previous years the meeting concentrated specifically on industrial applications of Z, and a high proportion of the participants came from an industrial background. The theme is well represented by the four invited papers. Three of these discuss ways in which formal methods are being introduced, and the fourth presents an international survey of industrial applications. It also provides a reminder of the improvements which are needed to make these methods an accepted part of software development. In addition the volume contains several submitted papers on the industrial use of Z, two of which discuss the key area of safety-critical applications. There are also a number of papers related to the recently-completed ZIP project. The papers cover all the main areas of the project including methods, tools, and the development of a Z Standard, the first publicly-available version of which was made available at the meeting. Finally the volume contains a select Z bibliography, and section on how to access information on Z through comp.specification.z, the international, computer-based USENET newsgroup. Z User Workshop, London 1992 provides an important overview of current research into industrial applications of Z, and will provide invaluable reading for researchers, postgraduate students and also potential industrial users of Z.

The Sixth Refinement Workshop took place at City University in London from 5th to 7th January 1994. The present volume includes all of the papers which were submitted and accepted for presentation, together with two papers by invited speakers. The workshops in the series have generally occurred at one year intervals but in this last case a two year period had elapsed. These workshops have established themselves as an important event in the calendar for all those who are interested in progress in the underlying theory of refinement and in the take-up by industry of the methods supported by that theory. One of the proposed themes of the sixth workshop was the reporting of successful adoption in industry of rigorous software development methods. The programme committee was perhaps slightly disappointed by the response from industry to the call in this respect. However, the recent period could be characterised as one of consolidation, when those companies which have made the decision that formal development methods are important to their business have been adopting them where appropriate and finding them to be worthwhile. On the other hand,. the difficult economic climate which exists in most parts of the developed world is perhaps not the context within which companies still dubious about the benefits are goil'\g to opt for making major changes in their working practices.

The Functional Programming Group at the University of Glasgow was started in 1986 by John Hughes and Mary Sheeran. Since then it has grown in size and strength, becoming one of the largest computing science research groups at Glasgow and earning an international reputation. The first Glasgow Functional Programming Workshop was organised in the summer of 1988. Its purpose was threefold: to provide a snapshot of all the research going on within the group, to share research ideas between Glaswegians and colleagues in the U.K. and abroad, and to introduce research students to the art of writing and presenting papers at a semi-formal (but still local and friendly) conference. The success of the first workshop has led to an annual series: Rothesay (1988), Fraserburgh (1989), Ullapool (1990). Portree (1991), Ayr (1992), and the workshop reported in these proceedings: Ayr (1993). Most participants wrote a paper that appeared in the draft proceedings (distributed at the workshop), and each draft paper was presented by one of the authors. The papers were all refereed by several other participants at the workshop, both internal and external, and the programme committee selected papers for these proceedings. Most papers have been revised twice, based firstly on feedback at the workshop, and secondly using the referee reports.

This volume contains extended versions of papers presented at the Third International Workshop on Logic Program Synthesis and Transformation (LOPSTR 93) held in Louvain-la-Neuve in July 1993. Much of the success of the workshop is due to Yves Deville who served as Organizer and Chair. Many people believe that machine support for the development and evolution of software will play a critical role in future software engineering environments. Machine support requires the formalization of the artifacts and processes that arise during the software lifecycle. Logic languages are unique in providing a uniform declarative notation for precisely describing application domains, software requirements, and for prescribing behavior via logic programs. Program synthesis and transfonnation techniques formalize the process of developing correct and efficient programs from requirement specifications. The natural intersection of these two fields of research has been the focus of the LOPSTR workshops. The papers in this volume address many aspects of software develop ment including: deductive synthesis, inductive synthesis, transforma tions for optimizing programs and exploiting parallelism, program analysis techniques (particularly via abstract interpretation), meta programming languages and tool support, and various extensions to Prolog-like languages, admitting non-Horn clauses, functions, and constraints. Despite the progress represented in this volume, the transition from laboratory to practice is fraught with difficulties.

Logic and object-orientation have come to be recognized as being among the most powerful paradigms for modeling information systems. The term "information systems" is used here in a very general context to denote database systems, software development systems, knowledge base systems, proof support systems, distributed systems and reactive systems. One of the most vigorously researched topics common to all information systems is "formal modeling". An elegant high-level abstraction applicable to both application domain and system domain concepts will always lead to a system design from "outside in"; that is, the aggregation of ideas is around real-life objects about which the system is to be designed. Formal methods \yhen applied with this view in mind, especially during early stages of system development, can lead to a formal reasoning on the intended properties, thus revealing system flaws that might otherwise be discovered much later. Logic in different styles and semantics is being used to model databases and their transactions; it is also used to specify concurrent, distributed, real-time, and reactive systems. ,The notion of "object" is central to the modeling of object oriented databases, as well as object-oriented design and programs in software engineering. Both database and software engineering communities have undoubtedly made important contributions to formalisms based on logic and objects. It is worthwhile bringing together the ideas developed by the two communities in isolation, and focusing on integrating their common strengths.

The papers in this volume were presented at the First International Workshop on Larch, held at MIT Endicott House near Boston on 13-15 July 1992. Larch is a family of formal specification languages and tools, and this workshop was a forum for those who have designed the Larch languages, built tool support for them, particularly the Larch Prover, and used them to specify and reason about software and hardware systems. The Larch Project started in 1980, led by John Guttag at MIT and James Horning, then at Xerox/Palo Alto Research Center and now at Digital Equipment Corporation/Systems Research Center (DEC/SRC). Major applications have included VLSI circuit synthesis, medical device communications, compiler development and concurrent systems based on Lamport's TLA, as well as several applications to classical theorem proving and algebraic specification. Larch supports a two-tiered approach to specifying software and hardware modules. One tier of a specification is wrillen in the Larch Shared Language (LSL). An LSL specification describes mathematical abstractions such as sets, relations, and algebras; its semantics is defined in terms of first-order theories. The second tier is written in a Larch interface language, one designed for a specific programming language. An interface specification describes the effects of individual modules, e.g. state changes, resource allocation, and exceptions; its semantics is defined in terms of first-order predicates over two states, where state is defined in terms of the programming language's notion of state. Thus, LSL is programming language independent; a Larch interface language is programming language dependent.