Carol C. Henderson - Executive Director,
Washington Office, American Library Association

You have the comments of the ALA on the NGI concept paper; I will not repeat them but simply use them to make some related observations sparked by developments this spring.

The biggest problem the NGI initiative faces is the confusion in the minds of many who must approve and fund the activities envisioned. Clarifying this confusion is critically important. I'm not a technology expert, but I represent a constituency that includes both cutting edge superhighway applications, and what I might characterize as "bike-lane" or "foot-path" applications. I've been working information technology issues with Congress and other policy makers for some 20 years, and have developed a feel for policy maker reactions and some sympathy for the non-technologists who must nevertheless make key decisions on these issues.

First, it would lessen the confusion to emphasize the differences among the NGI initiative, the Internet 2 project, and the NSF high performance connections program. If the federal NGI activities are thought of as an infrastructure program rather than an R&D program, and they are, there will be demands for representation and equity that can only unnecessarily muddle the needed first steps to a next generation infrastructure.

Let me suggest that it would help congressional offices to think of the NSF high performance connections program as an existing and ongoing attempt to position a number of research institutions with "generation and a half" connectivity. For the Internet 2 project, emphasize that it is an activity dominated by university and private sector investment in infrastructure. The NGI initiative, in contrast, constitutes the federal research and development catalyst necessary for the infrastructure of the next century.

Now this may indeed be an oversimplification, but I fear these three activities have been oversimplified in other unproductive and confusing directions, especially since many of the same institutions may be involved in all three activities. So if my attempts to clarify the differences are not fully accurate, correct them. But please, simplify accurately for those who are not as immersed in the details as you are, but who must make decisions, who are legitimately confused, and whose job it is to question whether this initiative justifies spending scarce federal dollars.

I believe ALA's recommendations will help you. Let's look at them just briefly.

If interoperability is clearly a fundamental principle of R&D for the NGI, it will help reassure policy makers that the results will benefit all those who have made and will continue to make considerable investments in current technology.

Emphasize applications in two ways. First, it has never been nor is it now, possible to design an infrastructure architecture, nor the "middleware" capabilities that may be embedded in it, in a vacuum. Whether we're talking about the original need to connect a handful of then large-scale computing facilities for specific projects, or whether it's the NGI, you can't design a facility or conduct research on advanced infrastructure without having some idea of what it's going to be used for.

Further, the NGI initiative should include applications R&D. What kinds of new or improved "middleware" capabilities are needed? What new and better items will users want in the NGI tool kit? For instance, NSF and other agencies made major investments in large-scale digital library projects; there are more lessons to be learned. Other digital library projects should be individually or jointly supported by participating agencies. Consider testbeds for access to preserved digital resources, as the National Research Council and others have suggested.

This leads to the points ALA made about broadening the number of agencies involved, including an applications R&D budget and plan, and increasing the civilian focus. Again, doing so will help the policy maker envision the future grassroots benefits.

Finally, develop a strong and specific diffusion plan to help alleviate fears of a two-tiered access system. Incorporate parts of this plan into NGI activities and budget. Show how other projects beyond the NGI will be part of the diffusion plan, and encourage that to happen through specific activities. You might conduct diffusion workshops for other federal agency personnel, and for constituencies that will benefit downstream, such as the library, education, health care, social service, state and local government, and cultural communities.

Look at specific programs that will be affected by the results of the NGI R&D, that will make use of the results, and that will help diffuse the results. Just a few examples that all include some federal funding include:

• AgNic, the Agriculture Network Information Center, a collaborative effort to link discipline specific or subject oriented resources of agricultural information service providers to improve access to and dissemination of these resources (USDA, universities, national and international associations are among the partners)
  
  

• Institute of Museum and Library Services National Leadership Grant program, including model joint museum and library projects
  
  

• Internet Scout Project at the University of Wisconsin-Madison's Computer Sciences Department (partially NSF supported)
  
  

• Library of Congress National Digital Library project, with its emphasis on unique American Memory collections (in 1996, NDLF vastly increased the number of collections on the WWW -- now more than 350,000 digital files available, with 1.7 million digital files in production or under contract for digitization)
  
  

• National Digital Library Federation, a cooperative program among 16 public and academic institutions, including the Library of Congress, each contributing to the building of global digital libraries