Community structure of modules in the Apache

The relationships among modules in a software project of a certain size can give us much information about its internal organization and a way to control and monitor development activities and evolution of large libre software projects. In this paper, we show how information available in CVS repositories can be used to study the structure of the modules in a project when they are related by the people working in them, and how techniques taken from the social networks fields can be used to highlight the characteristics of that structure. As a case

example, we also show some results of applying this methodology to the Apache project in several points in time. Among other facts, it is shown how the project evolves and is self-structuring, with developer communities of modules corresponding to semantically related families of modules. 1. Introduction Large libre software1 projects are usually organized as a set of modules. Each one of them can correspond to a given program, library, or any other unit identified by the project as distinct. It is common that developers work in several modules, according to their interests, skills, and constraints, and that those modules to which they contribute change over time. The relationships among modules due to the people working in them constitutes a sort of social structure of the project. In some sense, those human relationships are the glue that maintain the whole project together, and the chains that contribute to spread information and uses from one part of the project to others. In this paper, we explore how those developer connections contribute to the making of the community structure of the project, how it can be identified and visualized and how it evolves over time. The study and characterization of complex systems is a very fruitful research area nowadays with may interesting open problems. Special attention has been paid recently to complex networks, where graph and 1 Through this position paper we mean by libre software any program which is either open source software (according to the Open Source Initiative definition) or free software (according to the Free Software Foundation definition). network analysis plays and important role and is gaining great popularity due to its intrinsic power to reduce a particular system to its simple components and relationships. Thanks to this, network characterization is widely used in many scientific and technological disciplines as neurobiology [1], computer networks [2] [3], linguistics [4], etc. Among complex networks, social networks appear in a quite natural way as a method for analyzing the structure and interactions of people and groups of people within complex organizations [5][6][7][8][9]. To understand the structure of those networks, we are interested in determining how the different nodes interact and form groups that, in turn, interact with each other giving rise to higher order groups. The set of groups obtained, as well as their relationships, is which we call the community structure of the network. All the information we need for such an...

Website: www.little-idiot.de | Filesize: -
No of Page(s): 5
Download Community structure of modules in the Apache.pdf