An evolutionary study of fan-in and fan-out metrics in OSS

Excessive coupling between object-oriented classes is widely acknowledged as a maintenance problem that can result in a higher propensity for faults in systems and a `stored up' future problem. The aim of this paper is to explore the relationship between `fan-in' and `fan-out' coupling metrics over multiple versions of open-source software. More specifically, we explore the relationship between the two metrics to determine patterns of growth in each over the course of time. The JHawk tool was used to extract the two metrics from five open-source systems. Two questions were posed for each system. First, what are the characteristics of classes exhibiting the highest fan-in values? Second, do fan-in and fan-out increase in corresponding and consistent amounts over time? Results show a wide range of traits in the classes to explain both high and low levels of fan-in and fan-out. We also found evidence of certain `key' classes (with both high fan-in and fan-out) and `client' and `server'-type classes with just high fan-out and fan-in, respectively. We provide an explanation of the composition and existence of such classes as well as for disproportionate increases in each of the two metrics over time.