A Method to Derive Metric Thresholds for Software Product Lines

Abstract

A software product line (SPL) is a set of software systems that share a common and variable set of components (features). Software metrics provide basic means to quantify several quality aspects of SPL components. However, the effectiveness of the SPL measurement process is directly dependent on the definition of reliable thresholds. If thresholds are not properly defined, it is difficult to actually know whether a given metric value indicates a potential problem in the component implementation. There are several methods to derive thresholds for software metrics. However, there is little understanding about their appropriateness for the context of SPLs. This paper aims to propose a method to derive thresholds in the SPL context. Our method is evaluated in terms of recall and precision using two code smells (God Class and Lazy Class) detection strategies. The evaluation of our method is performed based on a benchmark of 33 SPLs and the results were compared with a method (baseline) with the same purpose used in the context of SPLs (not proposed). The results show that our method has better recall when compared with baseline.