In the next round, the Digital Libraries Initiative (DLI) will be somewhat differently structured mainly to enrich the program and enhance its flexibility. We're aiming at a $10 million annual budget over four or five years with a combination of different kinds of grants, as recommended in the Santa Fe report [Report of the Santa Fe Planning Workshop on Distributed Knowledge Work Environments: Digital Libraries, (March 9-11, 1997)]. We want to have a modular kind of program -- not a fixed closed thing -- that will enable new sponsors and new kinds of research to plug-in as new resources and capabilities, like the Next Generation Internet, come around, and the research itself evolves. This means that if another agency or a private foundation decides in a year or so to pursue a complementary or convergent goal, we could combine our resources quickly. We also hope the next phase will help create an environment that will foster collaborative, content-rich knowledge networks in the triangular convergence among the content, computing, and communications industries and technologies.
If we think about Doug Engelbart's insight into technological diffusion in terms of infrastructure, we can see steady improvement in physical technologies - telephone networks, roads, cars, tools, computers, and so on. We can also see that people steadily derive benefits from these innovations, but at a slower rate. What we want is a zigzag, where we can use the DLI to accelerate the process, to combine the power of the notion of "digital" with the ubiquity and power of the notion of "libraries" so that the collective impact of these kinds of initiatives accelerates improvement of human life as a function of what we develop. From the perspective of infrastructure -- that is, the long term investment made in physical or human resources that will have long term benefits -- we look at this as an advance and also as investment in human resources.
Infrastructure is one of five dimensions of research that will be emphasized in the initiative's next phase. The other four are fundamental research, experimental testbeds, content of collections, and context. The first round was very much a program of fundamental research and technology development: How could we make it better or faster? In the second round, we are looking to have a more refined technical scope. We have learned, for example, how to handle different kinds of data and integrate technologies -- text, image, video, and so on. We have a better appreciation of cross-disciplinary searches and how to manage distributed environments. But we want to move onto interoperable rather than distributed activities, to expand the research to examine the entire life cycle of digital content from creation to organization to dissemination and broad use. We are creating an environment where everyone is a creator, user, and archivist, and the traditional roles of librarians and others are fused. Addressing the needs of those roles technologically is a very daunting task.
In the first round, the notion of testbeds was very limited to experimental types of things projects would use, although we required that data collected would eventually be made available. In the second round, we will have something similar. That is, we will continue to have large projects but will also have different kinds of awards more tailored for individuals. Individual creativity is very important: many awards that agencies make go to individuals to encourage independent thinkers who will promote innovation and creativity in their fields. So we will have a mixed mode of projects. The larger awards will have testbed environments, but the emphasis in the second round will be on interoperation and on interactivity among people, groups, and community, driven by information content. For example, if we have several testbeds, one aspect of the research will be how they actually can work together. How can they be used in different kinds of environments and by different users with the results integrated from different sources. The broader issues are interoperability and data integration.
In the first round, we also did not have resources to allow digitization of content, so a wide range of collections were simply not available. In the second round, we will make digital content an integral part of the initiative through new sponsorship from agencies that have more and more diverse content, like the National Library of Medicine and the Library of Congress. We are also talking actively with many private organizations such as the foundations represented at the recent leadership colloquium, "Advancing the Public Interest through Knowledge and Distributed Intelligence" (KDI), held at the National Research Council [September 25-26, 1997]. We can imagine potential public agency/private foundation partnerships to create common cause among similar goals and missions. The question is, how do we share visions and coordinate programs? These organizations may not join the DLI as sponsors per se, but shared visions, goals, visions, and clienteles means that we can leverage complementary resources and collectively do more things. Even if only a small percentage of private foundation support goes to technical programs, the entire community, the entire society benefits. With KDI, we are looking for a shared vision and collaboration. The emphasis in DLI is on flexible collaboration, loosely linked relationships that, I think, can create a larger pool of resources, programs and therefore benefits -- that can create a larger pie.
Finally, the first round tended to focus on narrow contexts. There was some attention to user studies and iterative design and to extending the research to educational contexts. But the primary focus was on technological development, on pushing the frontiers of technology. In the second round, we expect to work with broader contexts, to look at different knowledge domains. We will go beyond computing to the social sciences and humanities as well as to other scientific domains with different types of users and different social contexts. This is all part of the human-centered systems program, which involves many different kinds of folks across several agencies who are helping us define our goals better: What are the most important impacts and aspects of human centered systems? What activities? For what users? To society? And how do we do that?
When I look back at the first round, look forward to the second, and ask what have we accomplished, I think that our vision: research, infrastructure, testbeds, content, and context, has allowed us to see that digital libraries are networks of information networks. The DLI has played a connecting role. But that role has to move to an interacting one. We can look at digital library research as a way of developing new enabling technologies that let us look, for example, at multimedia environments or that can deliver new services to classrooms and remote sites. We need to move from connectivity to the notion of public information spaces, where digital libraries are ubiquitous and public and at the same time also private and personal. With the new technologies and services, we can fulfill the dream that everyone can have access to what's in the Library of Congress and can tailor it to his needs. We are imagining a larger, richer, and more complex information space for every person and every society in the world and every part of the globe.
October 9, 1997