The importance of life cycle modeling to defect detection and prevention

In many low mature organizations dynamic testing is often the only defect detection method applied. Thus, defects are detected rather late in the development process. High rework and testing effort, typically under time pressure, lead to unpredictable delivery dates and uncertain product quality. This paper presents several methods for early defect detection and prevention that have been in existence for quite some time, although not all of them are common practice. However, to use these methods operationally and scale them to a particular project or environment, they have to be positioned appropriately in the life cycle, especially in complex projects. Modeling the development life cycle, that is the construction of a project-specific life cycle, is an indispensable first step to recognize possible defect injection points throughout the development project and to optimize the application of the available methods for defect detection and prevention. This paper discusses the importance of life cycle modeling for defect detection and prevention and presents a set of concrete, proven methods that can be used to optimize defect detection and prevention. In particular, software inspections, static code analysis, defect measurement and defect causal analysis are discussed. These methods allow early, low cost detection of defects, preventing them from propagating to later development stages and preventing the occurrence of similar defects in future projects.