Software quality from a behavioral perspective

The challenge facing software quality research is to produce a metric that will guide developers in choosing techniques to achieve the desired product quality. "Quality" comprises some set of key behavioral attributes: reliability, performance, fault tolerance, safety, security, availability, testability, and maintainability. Software quality is thus a function of these combined attributes plus an error term that represents quality aspects these eight attributes can\´t define. This article discusses how to quantify these behavioral attributes by working around relative values and formulating schemes from an attribute\´s key indicators. Thus, one can still assign numerical values to various attributes and then normalize them. By having such a system during requirements elicitation, developers can immediately begin to determine what techniques, methodologies, tools, processes, and costs they will need to produce a system with that quality level.