Characterizing Project Evolution on a Social Coding Platform

Abstract

Social coding platforms (SCPs) have realized spontaneous software evolution, where new source code and ideas are spontaneously proposed by altruistic developers. Although there are many projects operated by active communities performing spontaneous evolution, it is yet unclear that how such successful projects and communities have been formed and governed. In this paper, we propose a method that can investigate the history of every project in the SCP. Introducing the concept of project as a city, we consider every project in the SCP as a city, where a government and citizens develop a city through collaborative activities. We then identify essential attributes that characterize a state of a city. For each attribute, we develop metrics that quantity the state S(p, t) of a project p at time t. An experimental evaluation investigating GitHub projects of famous code editors shows that the proposed metrics well visualize the history of the projects from essential perspectives of a city.