A Uni–Dimensional Model Of Design
Keith J. Laidler · 1993
Abstract Synopsis In the absence of a clearly articulated model, software measurement is a fruitless exercise. In the preceding chapter we sought to remedy this situation by proposing a method to aid the development and evaluation of software metrics, together with their underlying models. This was illustrated by an extremely simple example. This chapter describes a larger software engineering problem for which we develop a model to relate software design to the quality factors of implementability, reliability, and maintainability. A formal model is developed and evaluated using the apparatus of the method outlined in the preceding chapter. Each step in the development of the model is outlined: problem identification, construction of an informal model, statement of the formal model, and axiomatization of the model. This is followed by theoretical and empirical validations. The chapter concludes that more than one model input – or multi–dimensional modelling – is necessary to improve the model performance and provide useful support for the software engineer carrying out design tasks. It is important to point out that in this chapter we are not attempting to promote one system design metric, but, rather, are using a particular metric as an example of the metric development process that was described in the preceding chapter.