Software Quality And Reliability

Abstract

Has the institutionalization of software quality and reliability within industry and government improved software products and systems? And, if so, can this improvement be meaningfully quantified and the recipients of its benefits identified? There is no simple answer to these questions since there may be no quantitative proof that things have improved. However, on the positive side it can be said that software quality has been improved because of: improvements in software technology, increased concern of government and industry resulting in the establishment of quality organizations, increased incidence of formal software quality audits and reviews, improved software documentation, and the use of independent V and V contractors. On the negative side, there are some major obstacles in assessing the improvement or implementation of software quality, among them is the lack of techniques and metrics for measuring software quality. This session will address the above questions, in addition to the problems and accomplishments in software quality assurance from both the government and industry viewpoints, with an emphasis on potential standards and testing criteria applicable on a government-wide basis or within industry as a whole. The principal elements of a quality plan for embedded software will also be highlighted by the panelists. Specific attention will be paid to the need for a well-ordered development methodology and to the review cycle attending it; to the array of hardware and software resources that may be required to assure software quality; and to the manner in which configuration control can be maintained. These matters will be discussed within the context of a synthesis of the disciplines of Software Engineering and Quality Assurance. Finally, experiences with Higher Level Language (HLL) compiler validation will be used as concrete examples to give the panel focus.