Weighing options for enterprise systems development, the CIO of a large U.S. university recognized the university's legacy systems were practically obsolete. In addition to their costly maintenance, they were failing to keep up with changes in the university's processes. Moreover, their siloed structure raised barriers to information exchange across the university's multiple schools and business units. From the CIO's perspective, a new enterprise system was needed to bring the university's information architecture up to speed. But how to achieve such an ambitious objective?
The decision was far from trivial. As highlighted by Kvavik et al.,10 "Enterprise system implementation is one of the single largest investments higher education institutions ever make." Given the anticipated expense and complexity, the CIO had to choose from three options: revamp the existing systems in-house; acquire and customize a commercially available platform; or take an approach called open-source software development, or OSSD.24 The retooling option would fail to address ongoing maintenance and increasing platform complexity. The proprietary option promised a mature system with extensive features and vendor support, but high costs and limited flexibility were major disadvantages. The OSSD approach promised greater flexibility and cost savings, but limited control over the development process and indefinite support were significant. How could the university weigh the pros and cons of these options? Should other approaches be considered?
A new approach to enterprise application development has emerged, applying the structures of collaborative development mechanism of open source software.1 The collaborative community-based open source model blends aspects of corporate-style software-development projects and open-source individual cooperative development. Investing institutions shape goals, resource commitment, and values through a project charter,26 an approach called community-based open source, or simply "community source," development.
Community source can be viewed as a hybrid developmental model, positioned between commercial software and OSSD. In a metaphor introduced by Raymond,19 commercial software development reflects a "cathedral" approach in which vendors construct massive structures based on detailed blueprints. In contrast, OSSD is built through a "bazaar" model, where programmers function like merchants in a marketplace, harnessing a range of approaches and objectives, taking input from diverse people anywhere in the world while remaining open to new ideas and new participants. Extending the metaphor, community source can be viewed as a "shopping mall" where a virtual organization, consisting of participating institutions, each directing its own human resources, creates a common collaborative infrastructure all participants can leverage.25 Community source thus creates a formal virtual organization to which partnering institutions contribute resources toward developing custom software solutions in different locations at different times.13 Despite drawing on the strengths of both commercial and open source models, community source remains a complex endeavor, as the community must balance the diverse, sometimes conflicting, requirements of the various development partners.12
The community source model represents a promising area for research, as it offers a novel approach to enterprise software development with the potential to disrupt the domination of commercial software vendors in enterprise applications. Here, we describe the experience of a prominent community source project in higher education, the Kuali Initiative (http://www.kuali.org), illustrating the strengths of the community source model and calling attention to the challenges faced in such efforts. We focus on understanding what motivates institutions to join such a project and how it holds together in pursuit of collective action.
In this research, we adopt a longitudinal case study methodology,27 as it enables investigation of a contemporary phenomenon within its natural setting11,27 and is especially appropriate for novel research domains (such as community source development where theory and research are still in their formative stages6). In addition, the longitudinal case study methodology is considered appropriate for addressing how, why, and what questions5,17 while fostering a deeper contextual understanding of the domain.27 Finally, the longitudinal nature of our study allowed us to observe development and deployment as project events unfold, instead of relying on retrospective accounts.5
Kuali pools institutional resources to develop an open source platform, dramatically reducing the cost for any single institution.
We conducted our case inquiries in accordance with prevailing case-study field procedures, including a case-study protocol prior to data collection, triangulation through multiple sources of evidence, and maintenance of a chain of evidence.17,27 The data-collection effort included semi-structured interviews with Kuali participants, direct observation of project events, milestone review presentations, minutes of meetings, and formal project documentation. We coded interview transcripts, observational field notes, and documentary evidence using NVivo, a qualitative data-analysis tool. We conducted data collection and analysis in line with grounded theory methodology, including adherence to the principles of constant comparison and open, axial, and selective coding.10,22
Data collection. We began to look into Kuali in 2005, conducting formal interviews during three different periods of time in which respondents were encouraged to discuss their Kuali participation in an open-ended fashion. We asked them to share their project experience, reflecting on their activities, problems encountered, lessons learned, how they or others interacted, steps taken to resolve problems, and general perceptions of the benefits and challenges of the community source development experience. We further asked for their perceptions of project success and project control. Most questions were open-ended and non-directive. In each interview session, we allowed for flexibility to focus on particular questions, factual information, and evaluative comments while skipping others based on a respondent's personal experience and expertise.
In addition to these largely unstructured impressions, we probed specific topics about the Kuali initiative. These more formal inquiry elements varied somewhat across interview periods. We conducted the first phase of interviewing in November 2008 with questions focusing largely on development issues from Kuali projects. We asked respondents about their role in a project, the tasks in which they were involved, and the deliverables for which they were responsible. We conducted the second round of interviews in November 2011, with greater emphasis on Kuali deployment issues. We conducted the third round in May and July 2012 with questions focusing on perceived trends and anticipated future of community source effort.
Interviews varied in length from 30 minutes to 90 minutes. We consolidated our interview notes within 24 hours following an interview. We also took extensive field notes of our observations.17 Respondents included senior executives, project managers, developers, and other IS staff members from participating Kuali institutions. Most respondents had extensive involvement in the development of community source projects.
Data analysis. Our data analysis consisted of multiple rounds of coding. The first involved thematic analysis of the data using the open-coding approach of the grounded-theory approach.4,10 In this initial process, we were immersed in the data, employing constant comparisons to identify persistent patterns in respondent experiences and perceptions. The second involved axial coding4,10 to discern core themes around Kuali participant interactions, challenges encountered, project-control mechanisms, and changes in the Kuali experience. The final round involved selective coding4,10 to target modes of project control and comparisons with other development environments.
We conducted open and axial coding following each of the first two data-collection periods. We used the coding structure we developed following the first round of data collection as a foundational framework for analyzing the second data set. In line with the principle of constant comparison, we modified the coding structure in light of the new responses. Data analysis following the third round of data collection involved only axial and selective coding. Two of the three researchers collaboratively executed open and axial coding of interview transcripts and field notes, with one conducting subsequent selective coding, and the resulting code structure and coded segments being reviewed by all members of the research team.
The Case of Kuali
Kuali is a comprehensive administrative software suite targeting the needs of institutions of higher education. The project began in 2004 as a partnership between Indiana University and the National Association of College and University Business Officers with support from the Andrew W. Mellon Foundation. The driving force was the realization that existing university systems were outdated and too difficult to maintain but that commercial software involves high cost, increased complexity, and inflexibility to the particular needs of institutions. Indiana found itself in a position like the scenario outlined earlier, with the university's leadership finding all standard options unsatisfactory. As a result, it formed the partnership and set out toward Kuali. Other institutions, including the University of Arizona, Cornell University, the University of Hawaii, and Michigan State University, quickly joined up.
The challenges they encountered were far from unique. In higher education, commercial software is often viewed as too expensive and difficult to customize. Even when such platforms are selected, many institutions decide to operate costly "shadow systems" to provide desired features missing in vendors' enterprise resource planning packages. On the other hand, building a financial system in-house is equally daunting and often feasible only for the largest universities. The Kuali project was an attractive alternative to the make-or-buy dilemma. Kuali pools institutional resources to develop an open source platform, dramatically reducing the cost for any single institution. It also enables institutions to address both individual and shared requirements due to the open source nature of the platform code. This emphasis on versatility is reflected in the project's nameKualithe Malaysian word for "humble utensil which plays the most important role in a successful kitchen and is used for frying, steaming, braising, blanching, and many more cooking techniques and styles" (http://www.kuali.org).
The Kuali effort has grown dramatically. In 2004, it had five development partners, including one commercial affiliate.a As of November 2013, it had 72 development partners, including 12 commercial affiliates, all united through a system of shared values, including collaboration, open source licensing, transparency, modularity, reusability, scalability, and accessibility and by projects driven by functional stakeholders. Adherence is what keeps the Kuali partners together.
From its initial focus on a baseline system for financial management, Kuali has expanded to include multiple system elements, including modules for research administration (Kuali Coeus), student information management (Kuali Student), library management (Kuali Open Library Environment), business continuity (Kuali Ready), human capital management (Kuali People Management for the Enterprise), mobile device integration (Kuali Mobility), and middleware/integration (Kuali Rice) (see Figure 1).
Although Kuali software is all open source, the project does not rely on volunteer labor at the level of the individual developer. The development team operates under a formal organizational structure within a virtual environment. The development partners are organized in a project structure. The Kuali board makes final decisions concerning directing and accepting all development work. For each Kuali module, a functional council, technical council, and project manager report to the Kuali board.
Even though Kuali received a startup grant ($2.5 million in 2005) from the Andrew W. Mellon Foundation, the project is funded mainly by partner institutions. As the suites of systems and number of participants have grown, Kuali has evolved three distinct forms of partnership: development, deployment, and sustaining.
The most significant effect of the Kuali effort is perhaps best captured by statements from the chair of the board of the Kuali Foundation: "I don't think the real story is about open source or particular products. What is really momentous is the pace at which the higher education community is coming together to meet its own needs." See the sidebar "Colorado State University Early Experience with Kuali" for motivations and effects of Kuali participation of this development partner, which joined the Kuali community in 2007 and implemented its first module, the Kuali Financial System, in 2009.
Since a community source project results in development of open source code, one might expect institutions to take a wait-and-deploy approach, letting other organizations develop the software, then acquire it at no cost when the functionality is mature.14 Insights from the Colorado State experience suggest this concern is unwarranted. Institutions adopting Kuali software overwhelmingly opt for formal participation in the broader community, raising the question of what actually motivates institutions to participate in the development process; Table 1 outlines the primary drivers.
Parallels with OSSD
The community source approach followed by Kuali participants draws explicitly on the emergence and success of the broader OSSD movement. Not surprisingly, our analysis of Kuali case-study data identified several points of commonality that persist between community source and OSSD, including commitment to open source code, the integrative role of shared values, widespread distribution of knowledge and expertise, and openness to multiple approaches.
The Kuali effort aligns with traditional OSSD through its commitment to open source code for all Kuali modules. The various Kuali platforms are available at no charge to institutions interested in deploying them. However, as with OSSD, adoption of one or more Kuali platforms alone does not imply engagement with, or participation in, the overall Kuali community. For Kuali, commitment to keeping the source code open is part of the overarching emphasis on openness of exploration and providing value to the higher-education marketplace. Moreover, the open source nature of the code is seen by participants as essential to the learning that occurs within the community. One leader of Kuali Financial System development said, "The developers' knowledge is critical in this type of environment because it's open source. The developers can actually go in and look at the code and say, 'Well, if you set it like this, this is what happens.' Without it being open source, you don't get that."
Ljungberg15 wrote, "An open source project is a loosely coupled community kept together by strong common values." In Kuali, common values are seen as the critical enabler of collective action. Kuali members frequently discuss the importance of the "Kuali value system." To a significant extent, Kuali values reflect the broader values of collaboration and sharing from the context of higher education and scientific pursuit. A senior administrator at Indiana University said, "Staff collaboration in Kuali really does reflect the inherent values of the university, a model where teams of researchers from multiple institutions collaborate on the frontiers of medicine, physics, and many other areas of research and scholarship." Other values of the Kuali Foundation include commitment to open-source licensing and a community source development model, transparency, modularity, reusability, projects driven by the functional stakeholders, and community diversity.
In OSSD, knowledge is inherently distributed across individuals and organizations. In line with this model, the Kuali community is characterized by widespread distribution of functional and technical knowledge. No single person has all the knowledge needed to move a Kuali project forward; rather, knowledge is accessed through engagement with other members of the community. The members perceive multiple benefits from this distribution, including access to insight, diversity of possible solutions to a given problem, and resilience of the community in the face of project breakdowns. This principle was reflected in a comment by the chair of the board of the Kuali Foundation, who said, "It's not that the knowledge and expertise for support are concentrated in one particular ZIP code of a firm but rather that they are distributed across the community itself. As someone who paid a very, very large maintenance bill for a tier-one support contract with another vendor, I will tell you, unequivocally, the Kuali community is faster and more resilient in responding to and solving a problem than a tier-one service provider."
The open source movement is characterized by diverse approaches to solving problems.8 Rather than mandate specific strategies for creating desired functionality, open source communities create a market dynamic in which multiple approaches are welcome. Likewise, the Kuali community appreciates multiple approaches to project configuration and management. While the Kuali community is defined by a common commitment to the broader vision of software created "by higher education for higher education," no single approach dictates how Kuali projects are executed. Indeed, the Kuali stakeholders have identified the diversity of approaches as a core strength of the Kuali network. Rather than advocate for a specific approach to project configuration, the Kuali Foundation empowers member institutions to determine what methodology and configuration of stakeholders are best for their collective purposes.
Parallels with Commercial Development
Along with important points shared with the OSSD approach, our analysis also found several characteristics that distinguish community source development from general OSSD methods and greater parallels with commercial development, including formal partnerships between participating institutions and developers, institutional control of development activities, and competition for access to development skills.
Unlike most OSSD efforts in which individual developers participate voluntarily, Kuali projects rely on formal partnerships between organizations. The functioning of the Kuali community depends on the shared commitment of member institutions. Kuali represents "institutions coming together" to collectively address shared needs. The perception of the Kuali participants is the effort would fail without the sustained commitment of institutions to one another achieved through a tiered membership structure of development, deployment, and sustaining partners. The explicit institutional commitment established through Kuali enables the initiative to avoid the challenges of "forking," or the splitting of a project into multiple separate strands of development, and diminishing participation that undermine the majority of open source projects.
As with institutional commitment, community source development incorporates a layer of institutional control. As in commercial firms, Kuali developers are directed to work on specific design requirements, not on whatever elements they wish, and deliver the functionality required by partnering institutions. Community source projects like Kuali involve an institutional-control element lacking in most open source efforts but that is inherent in commercial-development organizations. While supporting the progress of various Kuali projects, institutional control poses certain coordination challenges for Kuali in terms of integrating new developers and managing projects in a coordinated way. Developers must reconcile their dual commitments to the university that employs them and to the Kuali projects to which they contribute.
Community members consistently invoke Kuali values, reflecting the most important criteria for Kuali participation.
Unlike standard OSSD, the Kuali community does not have unlimited access to software development and integration talent. While Kuali members support one another (enabling access to lessons or insight garnered from earlier efforts), they are limited to the talent pools within their institutions. Ensuring an adequate supply of people with the requisite skills has been a persistent challenge for Kuali partners, including access to developers, high rates of turnover, and enhanced skill development. On the access front, member institutions have struggled to find developers with the needed skills. Access to talent differentiates community source development from ordinary OSSD efforts; Table 2 outlines the elements among community source, ordinary open source, and commercial software.
Mechanisms of Control
Control is a critical distinction among commercial, OSSD, and community source development models. Ouchi's16 widely cited model of organizational control mechanisms helps clarify the differences, defining three primary mechanisms for directing individuals and groups toward desired objectives: markets, bureaucracies, and clans. Markets reflect domains of exchange between two or more parties in which transactions are mediated by a price mechanism. In competitive markets, price mechanisms embody the informational requirements necessary to ensure an equitable exchange for all parties. Bureaucracies imply creation of a corporate structure or hierarchy through which contributions of individual actors are orchestrated. Unlike the price mechanism of markets, bureaucracies depend on acceptance of legitimate authority reflected in an organization's hierarchy. Finally, in clan mechanisms the goals of individuals are aligned through socialization and indoctrination into a system of shared beliefs and values. While these mechanisms maintain the legitimate authority of bureaucracies, the control they reflect is grounded in the goal congruence among participants.
Applying this framework to software-development models shows how commercial development reflects bureaucratic control. Software vendors orchestrate the work of their developers through an employment relationship. Such bureaucratic control enables a vendor to direct the evolution of its platforms and manage projects with respect to distinct timelines and platform functionality. However, the cost of bureaucratic control is reduced ability to pursue multiple solutions simultaneously and limiting of talent to firm employees.
The OSSD model reflects elements of both market and clan mechanisms. As with a market, OSSD developers determine their own willingness to participate based on perceived benefits; no one is obliged to engage in an exchange or a project. Rather, individuals evaluate the benefit their participation will provide (such as access to desired functionality and professional recognition) against the cost of participation (such as time and effort), creating a proxy price mechanism. OSSD projects also involve some elements of clan control. OSSD participants believe in software freedom and common values (such as sharing, assisting others, and learning by doing).21
The community source environment is characterized by a combination of control mechanisms. There is clear evidence of bureaucratic control in Kuali, with projects managed in an integrated fashion and tasks directed through a formal chain of relationships. However, the dynamics of clan control are equally apparent. Shared values, both individual and institutional, are central to the Kuali process. Kuali empowers its members to adopt whatever methods of development they deem appropriate, provided they adhere to the values of the community. This stance implies ceding some bureaucratic control in favor of a clan mechanism. Community members consistently invoke Kuali values, reflecting the most important criteria for Kuali participation; for example, if a commercial affiliate is not "stepping up to community values," it may be asked to leave the community.
The clan-control mechanisms in the Kuali community are central to understanding the commitment of individuals and organizations to the initiative. In line with social identity theory,3,23 Kuali creates community cohesiveness through a common social identity, or feelings of belonging with associated value significance in group membership, among participants. Perceptions of social identity influence work motivation and organizational loyalty,2 with participating individuals and institutions making their best effort because it contributes to the shared vision of the broader group. This shared identity engenders a culture of collaboration and "family spirit."18 Indeed, many Kuali participants express feelings of a family bond within the initiative. One of our respondents commented, "The KFS development team is really a family to me. We spent five years together to develop KFS, and we experienced most of things that a family experiences." Such spirit is reinforced by the annual meeting of the Kuali Foundation, or "Kuali Days," which feels more like a family reunion than a professional conference.
Much of the perceived strength of the community source model is its capacity for combining platform creation and evolution in a deliberate manner (bureaucratic control) with the flexibility and innovative potential enabled by values-based collaboration and interpersonal commitment, or clan control (see Figure 2).
While commercial development is within a bureaucratic mode of control, OSSD operates largely through a market-control mechanism, augmented by elements of clan control. In the community source model, we see strong elements of both bureaucratic and clan control. In this way, community source combines significant benefits from both commercial development and OSSD, delivering the coherence and flexibility of coordination based on shared values and principles while maintaining the institutional capacity for long-term planning and directed task completion.
Community Source Development
The ultimate value of community source development is pursuit of shared value creation. The community source approach suggests a fundamental shift from the prevailing practice through which organizations invest vast amounts of talent and resources in IT without considering similar efforts among kindred institutions. Indeed, it is appropriate that higher education should be a primary proving ground for this collaborative exploration of new approaches to knowledge work.
In creating a balance between commercial and open source development, the community source model inherits notable strengths and weaknesses from the other forms. One obvious benefit is significant reduction in costs due to pooling resources. In addition, because users drive development of functionality, community source participants get solutions tailored to their needs. Returning to our hypothetical university CIO, he is no longer forced to choose among a high-cost vendor platform that may not be suited to his institution, a resource-sapping in-house development effort, or an open source solution with gaps in desired functionality. By joining with like-minded institutions in a community source effort, his university can achieve the functionality that addresses its needs at a much lower cost than an isolated development project.
The community source approach addresses the primary concerns vexing the CIO, even though concerns persist. Challenges associated with the community source model include limited access to development talent, managing levels of institutional commitment, coordinating across multiple regulatory regimes, and determination of pricing mechanisms. One particularly salient challenge in the Kuali case is the management of growth. Due to the institutionally based nature of the projects, Kuali members see the need for managing the community's growth proactively. Unlike the organic growth of OSS projects, Kuali must attend to the ways member institutions interact as the community grows.
Our study also found that, despite its significant promise, community source development is no panacea. We are thus eager to see where community source goes in the future. There is evidence the model is taking root in industries other than higher education,7,20 even though questions remain around innovation and control, management of growth and evolution, and new modes of organizational interdependence. In our research, community evolution and modular design are emerging as central themes. We are thus exploring ways community source projects can employ the principles of modularity to achieve sustainability and a predictable evolutionary arc. However, as a research community, we have only scratched the surface of this novel approach to systems development.
8. Garzarelli, G. Open source software and the economics of organization. In Markets, Information and Communication: Austrian Perspectives on the Internet Economy, J. Birner and P. Garrouste, Eds. Routledge, London, 2004.
10. Kvavik, R.B., Katz, R.N., Beecher, K., Caruso, J., King, P., Voloudakis, J., and Williams, L.-A. The Promise and Performance of Enterprise Systems for Higher Education. ECAR Research Study, EDUCAUSE Center for Applied Research, Boulder, CO, 2002.
12. Liu, M., Wang, H.J., and Zhao, J.L. Technology flexibility as enabler of robust application development in community source: The case of Kuali and Sakai. Journal of Systems and Software 85, 12 (Dec. 2012), 29212928.
©2014 ACM 0001-0782/14/05
Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and full citation on the first page. Copyright for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or fee. Request permission to publish from [email protected] or fax (212) 869-0481.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2014 ACM, Inc.