On the Effectiveness of Bug Predictors with Procedural Systems: A Quantitative Study

摘要

Many bug predictors have been proposed, and their main target is object-oriented systems. Although object-orientation is currently the choice for most of the software applications, the procedural paradigm is still being used in many—sometimes crucial—applications, such as operating systems and embedded systems. Consequently, they also deserve attention. We present a study in which we investigated the effectiveness of existing bug prediction approaches with procedural systems. Such approaches use as input static code metrics. We evaluated to what extent they are applicable to our context, and compared their effectiveness using standard metrics, with adaptations when needed. We assessed five approaches, using eight procedural software systems, including open-source and industrial projects. We concluded that lines of code is the metric that plays the key role in our context, and approaches that use of a large set of metrics can introduce noise in the prediction model. In addition, the best results were obtained with open-source systems.