Software quality evaluation: state of the art, practice and directions

The aim of the chapter is to provide an integrated view of the approaches, technologies, and problems related to the quality evaluation of software products. The research and application area includes a variety of methods: quality models and related standards, metrics, testing, inspection methods, and model-based approaches. Great research effort has been allocated to software metrics, even if they have had a limited impact in engineering practice. All of these methods may be useful in evaluating the quality of software products, but an integrated view is not easily available from the literature. Moreover, from the practitioner point of view, the key point is not the identification of “the best metric” but the ability to exploit in a coherent framework every available measure (quantitative or qualitative, static or execution-based), depending on the existing technical and managerial constraints. The chapter organizes the knowledge of the area through a conceptual framework taken from other engineering disciplines. The framework is composed of the following parts: a tree of quality attributes; a definition of quality requirements; a set of measures; an algorithm able to generate values of high-level attributes from measures; a number of product models (abstractions of the software components to be measured); and a process model (the product at various development stages in time). Finally, the evolution of the state of the art and the application to current professional practice are discussed.