Understanding the Role of Core Developers in Open Source Software Development
Long, Ju, Journal of Information, Information Technology, and Organizations
The model of open source software (OSS) represents a disruptive paradigm in the software industry. Compared with the traditional proprietary software development, OSS is a radically new paradigm (Moody, 2001; Raymond, 2005; Sharma, Sugumaran, & Rajagopalan, 2002). With OSS, software source code is freely available for anyone to view, download, modify and redistribute as long as it is under the same open source license (see http://www.opensource.org). Most open source software projects rely entirely on the voluntary efforts of a community of developers (although some projects are coordinated and led by commercial entities). Such a voluntary community process keeps the cost of development and testing low. The nearly zero total cost of ownership (TCO) gives open source software a strong competitive edge.
A few projects initiated by the OSS community, such as GNU, Linux, Apache, MySQL and PHP, have achieved extraordinary success. However, except for these few successful projects, the majority of the open source projects have had lackluster performance, with little development momentum behind them (Thomas & Hunt, 2004).
Why do some open source software projects achieve success while others do not? What are the factors that influence the success or failure of open source software projects? Several researchers have tried to address these questions and have generated very insightful results (e.g. Crowston, Annabi, Howison, & Masango, 2004; Hann, Slaughter, Roberts, & Fielding, 2002; Stewart, Ammeter, & Maruping, 2005). Our research contributes to this prior literature in several important aspects. First, we have separated the OSS development workforce into two distinct groups--core developers and non-core developers--and studied their roles in the OSS development process. Prior research on OSS development has rarely distinguished between core and non-core developers, nor systematically examined the impact of core developers on the fate of an OSS project. Second, our research is based on a dataset of 300 open source projects. These projects not only include successful projects but also less successful ones. By examining projects from both realms, we can potentially gain a better understanding of the entire OSS ecosystem.
Our research addresses three important pillars of information systems research: information, information technology, and organizations. The open source model is a radically new model to gather information, create new information technology products, and to generate new knowledge. In the open source model, information flows in a loosely connected social network that is formed by thousands of developers. New knowledge is generated through volunteer work from these developers. The organizational structure is drastically different from the structure in traditional software development organizations. By analyzing the organizational structure of OSS projects, we have demonstrated several important factors that are related to OSS project success. Our research could shed some light on the sustainability of OSS as a new model of knowledge creation.
The remainder of this paper is organized as follows. First, we develop research hypotheses based on previous theories on OSS development and theories from organizational science. Next, we discuss the study's methodology and present results of the study. Last, we conclude the paper and examine future research topics.
Theoretical Framework and Hypotheses
Compared with the proprietary software development, open source developers have distinctive organizational structure, development process, and culture (Sharma et al., 2002). In proprietary software development process, there is only one development entity, being the software developers who are paid to work for the company. In the open source development process, however, there are two development entities: a small number of core developers (usually less than 15 people), and a large number of anonymous developers from the community (Mockus, Fielding, & Herbsleb, 2000, Schmidt & Porter, 2001). …