A characterized QFD for software quality attribute to reinforce ATAM

Abstract

Over time, the customer requires software excellent on function and quality. In order to meet their requirements, software engineers should equip more sophisticated methodologies to support this trend. However, the software technology still falls behind the market demands. One of main the reasons is the lack of proven methodologies. It is largely due to short history of software engineering. Software engineers proactively need to seek proven technology used in other engineering fields over a long time. Borrowing the concept of architectural engineering, a software architecture is a good example. In a short period, software engineering has quantum-leaped in both of quality and quantity. However the request of the market is still far beyond these achievements of software engineering. If software engineers can adopt proven technology of other engineering into software development life cycle, software community would make a big progress. QFD is a proven tool as one of total quality methods. QFD can reinforce ATAM, which is the principle technique for designing of a software architecture. ATAM is a useful tool for a software architecture, however, it is not sufficient. ATAM has little analysis on software quality elements, which are composing quality attribute scenarios. In addition, ATAM has a feeble decision-making process to determine the priority among scenarios. QFD can contribute in supplementing the weakness of ATAM, such as providing basis data for the analysis of quality elements and the deciding on the priority among scenarios. However, QFD is specialized to the manufacturing industry. In order to apply to ATAM, QFD needs to tailor for its software. As suggested by this paper, SQFD is tailored QFD for software. SQFD aims to reflect software quality attributes to a software architecture efficiently. Through two projects, MSE point project and a global embedded software project, SQFD could reinforce ATAM was demonstrated in this study.