Workshop participants generally agreed that the idea of an NL for SME&T education is sufficiently promising that the NSF should pursue it further, and the Steering Committee concurs. Although workshop participants did not agree on specific next steps, the Steering Committee makes the following recommendations to guide the NSF's planning for an NL initiative and its issuance of one or more Request for Proposals (RFPs). The Committee suggests that these recommendations be acted upon sequentially.

1. Clarify the potential customers of an NL for undergraduate SME&T education

1.1 Because workshop participants were unable to delineate the stakeholders or to specify the content for this proposed NL, the NSF should do so. The level of funding that the agency can devote to this project may dictate the breadth of the proposed NL's users, and that, in turn, may help with content decisions. However, the Steering Committee recommends that, prior to making final decisions about this issue, the NSF should make a concerted effort to bring together in a series of focus groups. representatives from all communities that might be an NL's likely users and service providers. Focus groups should be small and should be structured to encourage participants to discuss freely 1) their requirements for resources and tools that would help them improve teaching and learning of undergraduate SME&T, and 2) the ways in which the digital National Library could address those requirements. At a minimum, participants in these focus groups should include

• College and university SME&T faculty from all types of post-secondary institutions, including two-year colleges, undergraduate liberal arts colleges, predominantly undergraduate and comprehensive universities, and research universities. Special attention should be paid to including in these discussions people from institutions that have limited access to information technology so that an NL can be designed to take these limitations into account.
• College and university SME&T faculty at different stages of their academic careers. Discussions should be held with pre-tenured faculty as well as those who are at various stages beyond tenure (associate and full rank). Faculty who are not in tenure-track positions (e.g., adjunct and parttime faculty) also should be involved.
• College and university faculty involved with research and practice in science and mathematics education, including the preparation of future K-12 teachers. These faculty have knowledge of and familiarity with the research on effective techniques for improving learning and teaching. Their input is essential to inform discussions about content that is appropriate for undergraduate learners and other users of this resource and for the special needs of students who aspire to careers as science and mathematics teachers for Grades K-12.
• SME&T faculty from middle- and high-schools across the United States. These faculty may become important users of this proposed NL, and they will greatly influence the motivation of students to study SME&T both in high school and in college. These faculty also are often more well versed in pedagogy than university SME&T faculty and, therefore, can provide critical input about an NL's organization and user interfaces.
• Undergraduate students from different types of colleges and universities. This group should include both "traditional" students (i.e., those who enter college immediately after graduating from high school) and "non-traditional" students (e.g., adult learners who are returning to higher education).
• Graduate and postdoctoral students who are likely to enter careers in academe also should be consulted since they can help define needs of future faculty users.

• Librarians. Librarians already possess the requisite background and knowledge about users' borrowing habits and the tools that they employ to search for information. They can work with all of the other user groups to help define the roles and the services that an NL for undergraduate SME&T education might assume. Librarians also deal routinely with people who have special needs that any NL should accommodate.
• Social and behavioral scientists with expertise in organizational constructs and in the ways in which people learn new information can provide valuable insight and perspective to the NSF about how to design an NL's technological platforms and also how best to develop teaching and learning tools that would be widely utilized and educationally effective.
• Computer and information system specialists with specific experience with digital libraries. This community will define what is technologically possible for a digital NL for undergraduate SME&T education. Because so many other digital library projects are now in various stages of development and implementation, individuals involved with those projects should be brought into the discussion both to share information about their technological advances and to take back to their communities information about the potential for an NL.
• Directors of college and university information technology services. The groups of users listed above will help define the vision and objectives for an NL. However, if an NL is to be widely accessible, then it must be constructed with an appreciation of the current computing capabilities and limitations of potential users. Directors who work daily with computer systems on college and university campuses will help define the user interfaces that are currently available and that might be developed in the future.
• Representatives from the commercial publishing sector. Representatives from this community can best inform the NSF and other potential users about the kinds of materials that would likely become available for inclusion in an NL and the costs of procuring those materials. Since commercial publishers are intimately involved with the debates concerning intellectual property, copyright, and fair use of materials, their representatives also can update and listen to recommendations from the other players on these legal issues.
• Representatives from professional SME&T societies. Professional societies could serve this project both as providers of information and materials and as catalysts for engaging their members. Many professional societies maintain extensive collections of information related to their disciplines and may be very interested in an NL as a vehicle for disseminating that information. Professional societies also are recognizing their obligation to improving science education. Direct involvement with an NL could help focus the educational mission and activities of professional societies and encourage their members to see the improvement of undergraduate SME&T education as an integral component of scholarly activity and productivity.
• Representatives from the private non-profit sector, such as foundations. These people will likely provide at least some of the funds to sustain an NL for undergraduate SME&T education in future years. Their input will be critical.

1.2 The Steering Committee suggests that two different types of focus group meetings be held. Some focus groups should concentrate on receiving input from single communities, especially SME&T faculty and students. Others should involve people from many or all of the aforementioned sectors in cross-cutting sessions, with the primary objectives of having convenors listen and respond to the ideas and expressed needs of potential users.

1.3 The Steering Committee recommends that NSF also might employ the services of one or more professional organizations to organize these focus groups, to facilitate discussions within the groups and to prepare an independent assessment of user needs and desires based on the group discussions.

2. Articulate priorities for content, technological considerations, and economic and legal models before committing to the establishment of an NL.

The Steering Committee can offer no specific recommendations about whether the proposed NL should commission the creation and storage of materials vs. developing a sophisticated system of

pointers to materials that reside and are maintained elsewhere (see footnote 8, page 29, on archiving and preserving). Differences in cost between the two systems, evolving legal precedents with respect to copyright and fair use of materials, and the emergence of new technologies that may overcome some of the limitations of pointing to information stored elsewhere all must be factored into the final structure of an NL. Moreover, these parameters are likely to change during the development phase of the project. Ongoing advice from appropriate experts in all of these fields is warranted if the project proceeds.

2.1 The Steering Committee recommends that the proposed NL be viewed primarily as a resource for improving and inspiring learning of undergraduate SME&T rather than merely as a means to promote more effective teaching of these subjects. If an NL is to be a central component of current efforts to reform and improve undergraduate SME&T education, it must offer more than teaching tools. alone. The NSF should appoint a Board of Overseers consisting of acknowledged experts in SME&T education, library sciences, and digital libraries that is charged to work with a broad spectrum of intended users and the other stakeholders before decisions are made about what kinds of materials should be placed into the proposed NL. If an NL initiative cannot afford to support all areas of SME&T, then the Board should decide on the initial areas of focus and look to expand coverage as the project develops.

2.2 Steering Committee members also agree with many workshop participants and recommend that an NL should strive to focus on collecting or pointing to materials that either are inaccessible through other media formats or are so innovative that they are unlikely to be commercially available or viable in the short-term. Because a "critical mass" of materials is vitally important to the success of an NL, the acquisition of such innovative new materials will likely need to be balanced with more traditional materials, at least initially.

2.3 The Steering Committee recommends that the NSF emphasize involvement by professional SME&T societies in developing content that could be appropriate for an NL. Many of these organizations already have produced materials that might be incorporated into an NL at little or no cost. By promoting the development of these kinds of teaching and learning tools and by officially recognizing their members who do so, professional societies could become key catalysts in changing the culture of higher education to embrace as legitimate scholarly activities the promotion and evaluation of teaching and the promotion of effective learning by students.

2.4 The Steering Committee recommends that an NL should provide information about and access to projects in undergraduate SME&T education that the NSF and other agencies have supported financially.

2.5 The Steering Committee recommends that the NSF also seek a new, more encompassing descriptor for this project. Workshop participants recognized, and the Steering Committee concurs, that "Digital National Library" or "National Library"—the terms that have been most commonly used to describe this entity—may be more confusing than enlightening to anyone who envisions the potential stakeholders in this project and the services it may provide. Any NL initiative is likely to transcend the functions of many conventional libraries. A more appropriate descriptor might help to focus the higher education community on the need for such a resource and its importance.

3. Develop and issue one or more RFPs to establish an NL for undergraduate SME&T education

As the NSF receives additional input from stakeholders about the goals of and need for an NL (via Recommendations 1 and 2), the scope and potential cost of the project should become clearer. During the workshop, Steering Committee Chair Jack Wilson charged participants with trying to arrive at answers to the following major cross-cutting questions: 1) Is an NL a good idea for improving undergraduate SME&T education? and 2) Is an NL a better idea than other initiatives that might compete for the same funds? If the NSF is convinced on the basis of its explorations that it can answer these questions in the affirmative, then the question of how to implement this project should become the central focus. Options for proceeding at that point would include

Option 1: Undertaking a single, large initiative that would result in an operational NL within several years.

Option 2: Undertaking several smaller initiatives for shorter periods of time (12-24 months). These initiatives might be competitive and operate independently of each other or they might be components of some larger cooperative agreement. These various models for establishing an NL could then be evaluated against each other, with a final coordination of best practices that might lead to a single, integrated project.

3.1 Given the tremendous complexity of this project and the number of communities that must be directly involved if it is to have any chance for success, the Steering Committee recommends that NSF consider adopting Option 2. Steering Committee members envision that the smaller initiatives suggested in Option 2 might be incorporated into a program similar to those that the NSF's Division of Undergraduate Education has sponsored in recent years to change the ways in which chemistry and calculus are taught. Optimally, this new initiative would incorporate many similar components, including those delineated in Recommendations 3.2 and 3.3 below.

3.2 The Steering Committee recommends that the NSF, in following through with Recommendation 3.1, should develop an RFP articulating the need for and issues involving the establishment of an NL as outlined in this report. The RFP would encourage diverse groups of stakeholders to focus on some subset of the issues. Collaboration among stakeholders and interdisciplinary approaches to address the questions posed here would be encouraged. Preproposals could be sought, with funds then awarded to successful groups to encourage them to develop full proposals. Depending on the funds available, the NSF might then award larger contracts to one or more groups to tackle specific issues or sets of issues. Each of these final awardees would be expected to inform each other of their progress and problems through routine communications, reports, and through meetings of teams convened on a regular basis (at least annually).

3.3 Because the central concern of workshop participants was to define the users of and the need for an NL for undergraduate SME&T education, the Steering Committee recommends that RFPs for preproposals not be formulated until the NSF sponsors the focus groups described above. Feedback and evaluation of information from these groups of users and providers could then serve as the basis for constructing RFPs that would help eventual awardees to address specifically the established needs and requirements of potential NL users.