The traditional software process models are used in the development of design support software (DSS). They are typically multi phased, and consist of requirements specification, design, implementation, testing and operation as the main phases. One of the goals of phased development is to minimize faults in the delivered codes. Many of the faults can be traced back to the pre implementation phases of requirements specification and design. Reviews are typically performed to remove faults before the requirements or designs are passed to the subsequent phase. The problem is that activities in the phases of software process models, in particular the design phase are coarsely defined and do not scale to precisely match the underlying needs. We have developed a computer based pre-implementation testing methodology called abstract prototyping. It extends the current practices by defining the steps of the design phase of the DSS development process as definition of features, creation or selection of theories, formulation of methods, building of algorithms and implementation of pilot prototypes. Based on this methodology, faults can be traced back to the requirements, theories, methods, algorithms, or pilot prototypes rather than exclusively to the requirements or designs. Under this procedure, reviews are performed to remove faults before the requirements, theories, methods, algorithms, or pilot prototypes are passed to the subsequent stage or phase. Prototypes are used to provide the feel and the look of these work products and specially designed measures are used to estimate the extent to which they fulfill their respective specification requirements. We have applied our methodology to several application cases. The results show that the levels of fulfillment of requirements for theories, methods, algorithms, or pilot prototypes can effectively be estimated and in so doing faults detected sufficiently. In this paper, we first review the problem and present the related works. We then briefly introduce the abstract prototyping concept and present its application case studies.

This content is only available via PDF.
You do not currently have access to this content.