This book is the second of two volumes that present the main results which emerged from the project CIP - Computer-Aided, Intuition-Guided Programming - at the Technical University of Munich. Its central theme is program development by transformation, a methodology which is becoming more and more important. Whereas Volume I contains the description and formal specification of a wide spectrum language CIP-L particularly tailored to the needs of transformational programming, Volume II serves a double purpose: First, it describes a system, called CIP-S, that is to assist a programmer in the method of transformational programming. Second, it gives a non-toy example for this very method, since it contains a formal specification of the system core and transformational developments for the more interesting system routines. Based on a formal calculus of program transformations, the informal requirements for the system are stated. Then the system core is formally specified using the algebraic data types and the pre-algorithmic logical constructs of the wide spectrum language CIP-L. It is demonstrated how executable, procedural level programs can be developed from this specification according to formal rules. The extensive collection of these rules is also contained in the book; it can be used as the basis for further developments using this method. Since the system has been designed in such a way that it is parameterized with the concrete programming language to be transformed, the book also contains a guide how to actualize this parameter; the proceeding is exemplified with a small subset of CIP-L.

Proceedings of the NATO Advanced Research Workshop on Program Transformation and Programming Environments

"Specification and transformation of programs" is short for a methodology of software development where, from a formal specification of a problem to be solved, programs correctly solving that problem are constructed by stepwise application of formal, semantics-preserving transformation rules. The approach considers programming as a formal activity. Consequently, it requires some mathematical maturity and, above all, the will to try something new. A somewhat experienced programmer or a third- or fourth-year student in computer science should be able to master most of this material - at least, this is the level I have aimed at. This book is primarily intended as a general introductory textbook on transformational methodology. As with any methodology, reading and understanding is necessary but not sufficient. Therefore, most of the chapters contain a set of exercises for practising as homework. Solutions to these exercises exist and can, in principle, be obtained at nominal cost from the author upon request on appropriate letterhead. In addition, the book also can be seen as a comprehensive account of the particular transformational methodology developed within the Munich CIP project.

Algebraic specification, nondeterminism and term rewriting are three active research areas aiming at concepts for the abstract description of software systems: Algebraic specifications are well-suited for describing data structures and sequential software systems in an abstract way. Term rewriting methods are used in many prototyping systems and form the basis for executing specifi cations. Nondeterminism plays a major role in formal language theory; in programming it serves for delaying design decisions in program development and occurs in a "natural" way in formalisations of distributed processes. Heinrich Hussmann presents an elegant extension of equational specification and term rewriting to include nondeterminism. Based on a clean modeltheoretic semantics he considers term rewriting systems without confluence restrictions as a specification language and shows that fundamental properties such as the existence of initial models or the soundness and completeness of narrowing, the basic mechanism for executing equational specifications, can be extended to nondeterministic computations. The work of Heinrich Hussmann is an excellent contribution to Algebraic Programming; it gives a framework that admits a direct approach to program verification, is suitable for describing concurrent and distributed processes, and it can be executed as fast as Prolog.

This volume constitutes the refereed proceedings of the 1995 Asian Computing Science Conference, ACSC 95, held in Pathumthani, Thailand in December 1995. The 29 fully revised papers presented were selected from a total of 102 submissions; clearly the majority of the participating researchers come from South-East Asian countries, but there is also a strong international component. The volume reflects research activities, particularly by Asian computer science researchers, in different areas. Special attention is paid to algorithms, knowledge representation, programming and specification languages, verification, concurrency, networking and distributed systems, and databases.