Industrial practitioners require constant improvements in the software development process and the quality of the resulting product in order to satisfactorily build larger and more complex software systems. Academia praises formal specification techniques as a means to achieve these goals, yet formal specification has not been widely adopted by industry. The focus of this research is to study the disparity between industry and academia in their experience with formal specification methods. During the specification of a significant software system, a control system for a nuclear reactor, it became clear that the use of formal specification methods had potential benefits, but there were practical requirements that were not being met. Previous evaluations of formal specification failed to identify many of these flaws and a new comprehensive approach based on the requirements of the current software development process is needed. A comprehensive approach to evaluation was developed as pa...

A new method of writing Haskell programs from Z specifications is proposed, whose use is illustrated elsewhere (Goodman, 1995). This method can be regarded as a kind of `tool-kit', consisting essentially of three discrete parts: a set of high-level guidelines for the programmer; a framework for handling input/output and state; and a set of lower-level rules, indicating how to translate various Z constructs, in particular, those of Z's `mathematical tool-kit', into Haskell. This paper presents the Z-into-Haskell tool-kit itself, as it currently stands, in the form of a combined tutorial and reference manual. It is intended that the reader, with some knowledge of Z and Haskell, can use this to begin to develop his/her own programs from specifications. (Note: A better typeset version of this paper, set in Monotype Bembo and printed at a resolution of 600 d.p.i., is available by post from the author.) CSR-95-1 April 1995 Contents 1 A strategy for producing Haskell programs f...