Characterizing Software Activity

Abstract

The health of a software ecosystem reflects the ability of the ecosystem to endure and remain variable and productive over time. Measurements of health in a software ecosystem are applied to inform on how the ecosystem is evolving, evaluate changes, and predict future states. In this study we investigate the influence of software to the overall health of the ecosystem. We propose an approach of measuring the activity of the ecosystem over time and identifying the influence to health. We do this in two ways: (i) we study the evolution of the software network over time to identify changes in the structure of the software network and investigate whether they relate to general changes in the ecosystem. (ii) We propose the identification of the influence of the independent software components to ecosystem health at an activity level. We do so by defining keystone and dominator activities and propose tentative means of measuring them. We apply our proposed approach to the platform of the Apache Cordova ecosystem, an ecosystem with a community-based platform and independent contributions. Our analysis identifies two points in time where the ecosystem is under major change. These points are confirmed independently by both the measures of software network and keystone and dominator activities.