Data Abstraction and Programme Design

This book is intended to familiarise the reader with an interrelated cluster of concepts including data abstraction, abstract data types, information hiding, object-oriented design, and so forth within a specifically software engineering context. It is not intended as an exhaustive reference work in any of the subjects covered. Rather, it is intended to modify the "mind-set" of its intended readership, assumed to be undergraduates who have completed an introductory "programming methodology" course, probably based on the use of a language like Pascal. (A complementary readership is composed of practising software engineers who have imbibed some of these concepts "on the job" and who wish to see them explored and established within a coherent perspective). The motivation underlying its writing was the recognition of the difficulty that the average student (and the above average student for that matter) has in coming to terms with the concept of "programming in the large", or "architectural" software design. This difficulty, which is endemic to the technology, is often exacerbated by the typical programming methodology course, with its emphasis on "programming in the small", or the detailed implementation of algorithms - providing a set of skills largely different from that involved in high-level, structural design. The importance of structure, of the separation of interface specification from implementation detail, with the primacy of the former over the latter, of data structures as the major structural features of software systems, are some of the priorities that the book attempts to justify and clarify, building from a base compounded of common sense and the technical knowledge that the intended readership might reasonably be expected to have. These issues are developed through a progression from a Pascal context, through recognisable developments of Pascal, to the fairly remote areas of algebraic specification and object-oriented programming environments, with the common, unifying theme of data abstraction.