commercial model, exemplified by companies such as ESRI, Intergraph, and Autodesk, is market driven with a business model that reflects the need to balance the costs of software development and support against income from the open market. The educational market plays little or no role in this model. The collaborative model views GIS software development as a collaborative process, underpinned by an open foundation of standards and basic functions. Thus, in the 1980s the U.S. Army Corps of Engineers developed the Geographic Resources Analysis Support System (GRASS ) package and fostered a community of users who contributed extensions to the package. In this model, there is no distinction between users and developers. GRASS was built as open software, with no proprietary restrictions on access or use. Despite its success, it was seen as competing unfairly with the commercial market and, therefore, its support was terminated in the early 1990s although a residual community continues to use it.
These three models offer distinct options for the redesign of GIS software for the K–12 context. For the collaborative mechanism to succeed, a community would have to be identified, comprising specialists with sufficient technical skills to share the development of appropriate software, and with sufficient scientific understanding of the needs in the K–12 context. An organization such as the University Consortium for Geographic Information Science (UCGIS) might be appropriate to facilitate collaboration, with sufficient funding from an appropriate federal agency. UCGIS has access to technical and intellectual expertise at each of its more than 60 member institutions and has sufficient experience in organizing large, distributed projects.
For the commercial mechanism to succeed, it would be necessary for an appropriate federal agency to request proposals and select a suitable developer. One major advantage of this mechanism is that much of the software foundation for a new GIS already exists in each vendor’s products. Because educational support for spatial thinking is somewhat distant from the normal domain of commercial applications, there would have to be strong and robust mechanisms for oversight of the design and implementation of the software, and for practical testing. Contracts would have to deal with issues of long-term maintenance, intellectual property, and long-term support. A careful study would be necessary to determine whether the K–12 sector could generate sufficient funds to pay for long-term maintenance.
For the academic model to succeed there would have to be a similar proposal solicitation process: it would require an appropriate federal agency to select and contract with one or more academic institutions for the basic software development. This mechanism might be more successful than the commercial one in ensuring the appropriate intellectual content, but stringent oversight would be needed to ensure quality control in software development and to ensure that the designs were practical and scalable in the educational context.
All three mechanisms appear to have merit, as well as potential pitfalls. The choice between them, therefore, should be made by the appropriate funding agency. In the committee’s view, the collaborative model appears to offer the most promise because it would involve all parties—software developers, government, academia, and the K–12 user community.
Based on the levels of investment being made by commercial vendors and on experience from many GIS development projects, it would be reasonable to assume that a suitable GIS could be developed over a period of three years by a team of ten programmers. Allowing for oversight and other costs, the initial development might require a total investment of $3 million to $5 million.
To coordinate the development of GIS software using the collaborative model, the committee recommends the creation of a “Federation of GIS Education Partners.” A federation is a bottom-up association of autonomous partners that agree to abide by specified interface standards, business practices, and expectations to achieve a common goal (Handy, 1992). Federations, such as the National Center for Atmospheric Research (NCAR), NASA’s Earth Science Information Partners (ESIPs), and the Association of Research Libraries, provide a means for representing the multiple interests of broad interest communities. In an ideal federation, partners come together to achieve