Systems Analysis Documentation: Computer-Aided Data Dictionary Definition

Abstract

A need exists for the development of standard techniques and formal notations for use in systems design and analysis. Their value lies in expediting work, facilitating work sharing by aiding communication between team members, assisting supervision, and providing a clear and generally understood design/implementation interface. Canning (1966) states '. . . our present tools are just too cumbersome to undertake the more complex studies in the average business environment'. The need for such methods is referred to in Gibson et al. (1967) and has been cogently argued by Grindley (1966) whose suggestions are further discussed in King (1967). An attempt has been made to establish a formal theoretical basis for this type of work described by Bosak et al. (1962) and briefly by Clippinger (1962). This, however, is at present far removed from the realm of day to day practical application. A scheme designed for the documenting and study of existing data systems called AUTOSATE has been developed by Butler et al. (1964). Variants of these ideas under other names, e.g. Dataflo, Crowther-Watson (1967), also exist. The basic concepts are those of 'station' and 'event chain'. A data 'system' is divided into a number of separate 'stations', and information flowing between stations and data retention at stations is recorded. This is processed to produce documentation and identify 'event chains' or sequences of derivative connected activity. The basic philosophy is that we record the present system with a view to improving on it. It seems likely that outside the large complex data systems of the military area for which Autosate was designed, the approach will be more appropriate in an industrial context than in business data processing. In an industrial context the data system, though essential, is ancillary to the manufacturing activity, and any data system redesign is constrained by the nature of the existing plant and machinery. In the context of business data processing there is normally more scope for data systems redesign, and it is frequently better to identify basic requirements and design a new system based on these than to attempt to record and relate all details of every existing clerical process, although the essential functioning and purpose of these activities must be well understood.