A Digest of Research on Software Quality

Successfully competing in the software development market requires controlling the quality of the product. This not only involves identifying what software attributes are valued by the target market, but also measuring and managing those attributes over the course of the software development cycle. This represents an emerging point of view in software engineering, as explicit considerations of cost and value over time are largely missing from usual software development methods. This report attempts to help software engineers achieve those goals. It briefly summarizes the relevant aspects of general economic theory and ties those considerations to the challenges of software development. Then, it discusses research surrounding four evaluation techniques (ATAM, Heuristic Evaluation, COCOMO II, and Code Complexity Metrics) which can help software engineers deliver the right quality attributes through their product. Finally, this report comments on where organizations can expect to head from here, in terms of process optimization and current areas of research. 0.0 Introduction p 2 1.0 Microeconomics 101 p 3 1.1 Factor Exchange p 3 1.2 Attributes and Functions p 5 1.3 Economic Choices p 8 1.4 Economic Optimizations p 9 2.0 Application to Software Engineering p 11 2.1 Software as a Valuable Factor p 11 2.2 Software Attributes p 12 2.3 Evaluation Techniques p 14 2.4 Example: An Attribute Evaluation Gone Wrong p 19 3.0 Controlling Software Quality p 21 3.1 Example: Qualitative Techniques Early in the Project – ATAM p 21 3.2 Example: Qualitative Techniques Late in the Project – Heuristic Evaluation p 25 3.3 Example: Quantitative Techniques Early in the Project – COCOMO II Early Design p 29 3.4 Example: Quantitative Techniques Late in the Project – Code Complexity Metrics p 34 4.0 Beyond Today’s Product p 39 4.1 Optimizing the Producer p 39 4.2 Areas of Active Research p 42 4.3 Annotated Bibliography p 43