Models, methods, parameters, characteristics and metrics used in expert systems of evaluating the quality of the design and development of innovative software projects

Abstract

This paper contains two sections relating to software quality issues. First, the various definitions of software quality are examined and an alternative suggested. It continues with a review of the quality model as defined by McCall, Richards and Walters in 1977 and mentions the later model of Boëhm published in 1978. Each of McCall's quality factors is reviewed and the extent to which they still apply in the late 1990s is commented on. The factors include, integrity, reliability, usability, accuracy, efficiency, maintainability, testability, flexibility, interface facility (interoperability), re-usability and transferability (portability). They are subdivided into external and internal quality factors. Interrelationships between the different factors are shown in Perry's model. Issues of quality management and the prioritising of these factors are included. The second section examines the strategic impact of quality from both the supplier's and the purchaser's point of view. In particular product differentiation, tendering and estimating, system acquisition and employee productivity are considered. Product differentiation is mapped to Porter's generic business strategy. The COCOMO model for software costing and estimating is used to show that quality factors influence the cost of a product. As quality impacts on all classes of people in systems, human resources and the consequences for productivity are explored. Finally, system evaluation and selection techniques involve quantitative (weighting and rating) techniques and Robson's example and the influence of quality are examined.4