Applying software reliability engineering in the 1990s

摘要

This paper reviews the progress in software-reliability since 1975, and discusses the best tools and practices that can be applied today. The best development practices are recommended herein, for managing the reliability of software. The development of software-reliability models and user-friendly tool kits is described. These tools allow software-reliability to be measured, tracked, and improved to meet the customer-specified reliability. The ability to measure software-reliability promotes development focus, and consequently, its improvement. The 1990s have seen much growth in the software content of products. The problems have grown as the software mushrooms both in size and complexity. Further, software is increasingly important in safety critical applications. To counter these concerns, there has been a great collective effort to improve the reliability and quality of software, through better development-focus exemplified by the Software Engineering Institute's Capability-Maturing-Model levels. Software-reliability-engineering has shown best practices for developing and measuring software-reliability. The 1990s have brought collective focus on managing reliability. Such collaboration has been in packaging software-reliability models in user-friendly packages such as SMERFS and CASRE.