Software Reliability - 40 Years of Avoiding the Question

Summary form only given. During the past years, there has been an explosion of technical complexity of both hardware and software. Hardware, characterized by physics and empirical data, provides the reliability and systems engineers with a capability to estimate the expected reliability. Software has however managed to avoid this type of model development in part because the factors affecting reliability are not measurable by physical data. Software reliability is characterized by data gathered during systems integration and test. This data has attributes and parameters such as defect density, capability of programming of the software engineering team, the experience of the engineering team, the understanding of the application be the designers, the language used and more. Software reliability is more than the processes advocated by CMMI (Capability Maturity Modelreg Integration) and is susceptible to esoteric and infinitely harder parameters to measure. The author discusses some of the elements that affect software reliability and compares some of the differences when trying to estimate reliability of today's systems.