Measurement as an alternative to bureaucracy for the achievement of software quality

In this paper we discuss the two dominant modes of thought on the problem of software quality: the bureaucratic approach and the technical approach. The aim of the paper is to pursue the issues of how these approaches affect people, in our case software developers, and how we can assess the effectiveness of each approach in actually achieving quality software. An outline of the main aspects of the organizational approach — process modelling and software standards — is given. In order to address the effects of each of these modes of thought on people we take a psychological perspective. The implications for software engineering, in terms of formalism and mechanization, are discussed. The paper argues that overzealous structure, control and automation can negatively affect the creative process of developing software. A perspective on each of the components of the technical approach in terms of their effectiveness in achieving quality objectives is provided. From recent work we find that process modelling and many software standards do not provide an adequate assessment of the benefits derived from their application. The paper argues that measurement goes some way to addressing these problems. Measurement should form the core of the software development process, especially with respect to product quality assessment. The main conclusion is that by focusing on measurable objectives and results we can best achieve quality software products and highlight the processes which are likely to create them in a repeatable and manageable fashion. Additionally, by adopting measurement in practice we should expect to increase the freedom and creativity of software developers.