We see the Sakai Project, which here refers to the work done by the project founders under the Mellon grant as well as the contributions of others, including the Hewlett sponsored SEPP and its partners and the SCA affiliates, as providing the best potential for capturing innovation in higher education learning and research practices and getting that innovation rapidly into the hands of a large community of faculty, student and staff users. This is based on the large and expanding group of adopters and developers that can contribute ideas, expertise, and experience, resulting in a large number of differentially coordinated development efforts. Sometimes these development efforts will be tightly coupled, and sometimes they will be relatively autonomous, displaying varying levels of collaboration around a specific goal. We see open source models of licensing and social organization as a key component of the underpinnings of such a large and distributed effort. We also see these models as still being in the process of creation. Discovering just what is the best mix of tightly and loosely coupled development is part of the work we have to do in the Sakai Project. Discovering just what forms of organization will help to provide enduring motivations to the institutions and individuals who go to make up this effort, and at the same time provide sufficient coordination of the resulting work on a complex set of components that go to make up our product, is part of figuring out our own model of community source software development.
We are working from a relatively simple set of ideas as we proceed:
0) The success of this project rests on the ability to mobilize resources from a large number of institutions and individuals. There is a huge untapped reservoir of resources and creative energy that have up to now been dissipated in uncoordinated and often competing efforts to build learning, collaboration, and research software. We can change that.
1) There is a need for a tightly coupled effort and set of resources to construct the software framework and move it rapidly forward. This is the foundation for what follows and the solid ground on which we build the distributed efforts for tool construction and integration.
2) We will need to continue to rapidly innovate on the framework for the foreseeable future. For instance, our goals of interoperability, or portal integration, to pick two things, are still unmet, and will need considerable effort to achieve. Also, the rate of change in software development methods and models is not slowing down, and we need to be leaders in the application of emerging technologies that help ease development and bring new ideas to fruition quickly.
3) The development of tools must be a widely distributed, loosely coupled effort that the software and organizational structures and community practices of the project provide coordination for. We will not realize the potential of this project if we do not succeed in enabling, indeed exciting, a global community of users, designers and developers into contributing their efforts and creativity. This can not be a tightly coupled, centrally coordinated activity.
4) We will need to continuously provide a significant level of support and assistance to a large distributed community of users and developers, adopters and designers, individuals and organizations. We need to provide resources for such a persistent effort.
5) We have a number of ways to develop resources for our central efforts, including those already demonstrated, such as donating staff, paying membership fees, and getting grants, and have already seen large autonomous efforts contribute to the common code base.
6) The ideas of a central body to coordinate the tightly coupled efforts, and a large community of autonomous, locally motivated design and development activities, that are supported by community resources, provides the best model so far for what we need to accomplish. We must be successful at both types of 'organization' to be successful as a project. We must use the production and coordination methods of both centrally run projects and open source projects, and learn how to meld the two.