New Metrics for Assessing Aspect Coupling
Previous studies provided metrics to show different types of aspect coupling. However, no existing metric adequately described coupling in a way that combined the invocation of aspect advices, aspect methods, and class methods. This study introduces an invocation coupling framework that defines method-method, advice-method, and advice-advice coupling. This new framework is the basis for two new aspect coupling metrics-interference potential (IP) and interference causality potential (ICP)-to account for each type of invocation coupling. The work validates these metrics by analyzing two AspectJ programs: AspectTetris and AJHotDraw. Two versions of each program, an unmodified version and a version containing a SeedAspect designed to increase interactions, were tested. Findings suggest that low metric values are likely since most modules have few interactions. Abnormally high values indicate high coupling and an increased risk for aspect interference due to the complex interactions. Introducing the SeedAspect caused a small shift in IP, but a significant shift in ICP. Thus, ICP was most successful at detecting aspect interference risk due to complex interactions.
Bennett, Brian T.; and Mitropoulos, Frank J.. 2016. New Metrics for Assessing Aspect Coupling. Conference Proceedings - IEEE SOUTHEASTCON. Vol.2016-July https://doi.org/10.1109/SECON.2016.7506722 ISSN: 0734-7502 ISBN: 9781509022465