The attention that aspect-oriented software (AOS) development has gained recently has prompted the need for a rigorous set of metrics that can be used to quantify the different attributes of AOS. With the long list of metrics proposed, empirical evaluation becomes vital to assess the effectiveness of these metrics to provide guidance to software engineers in selecting the ones to use during software development.
The authors of this paper performed an empirical study to evaluate six selected metrics for AOS using data from ten open-source programs. Each metric is rigorously defined and supported with usage scenarios, empirical data values (mean, standard deviation, ratio), and an analytical evaluation to validate the metric.
The main contribution of the paper is the analytical evaluation of the selected metrics against established criteria for validity. The results serve to recommend the use of these metrics in assessing aspect-oriented software. The authors have provided rigorous definitions of the selected metrics that disambiguate informal descriptions present in the literature. The paper is a good reference for researchers and industry personnel looking to understand aspect-oriented software metrics and their use contexts.