Measuring and characterizing crosscutting in aspect-based programs: Basic metrics and case studies

Abstract

Aspects are defined as well-modularized crosscutting concerns. Despite being a core tenet of Aspect Oriented Programming, little research has been done in characterizing and measuring crosscutting concerns. Some of the issues that have not been fully explored are: What kinds of crosscutting concerns exist? What language constructs do they use? And what is the impact of crosscutting in actual Aspect Oriented programs? In this paper we present basic code metrics that categorize crosscutting according to the number of classes crosscut and the language constructs used. We applied the metrics to four non-trivial open source programs implemented in AspectJ. We found that for these systems, the number of classes crosscut by advice per crosscutting is small in relation to the number of classes in the program. We argue why we believe this result is not atypical for Aspect Oriented programs and draw a relation to other non-AOP techniques that provide crosscutting. © Springer-Verlag Berlin Heidelberg 2007.