How to effectively define and measure maintainability

Maintainability and flexibility at the software level are of predominant importance to drive innovation at the business process level. However, existing definitions of maintainability, such as the Halstead Volume, McCabe's Cyclomatic Complexity or the SEI maintainability index provide a very poor understanding of what maintainability is how it can be assessed and ultimately controlled. This paper explains a new and more effective way to construct software product quality models. The key design principle is the strict separation of activities and properties of the system. This separation facilitates the identification of sound quality criteria and allows to reason about their interdependencies and their effects. The application of this quality modelling approach in large scale commercial software organizations helped to effectively reveal important quality shortcomings and raised the awareness for the importance of long-term quality aspects among developers as well as managers.