|
Summary of the Strategic Planning Session Of the Digital Library Federation Capital Hilton Hotel, Washington, D.C. 6-7 February 2003 29 of 30 DLF Partner institutions were represented: Carnegie Mellon University: Gloriana St. Clair Columbia University: James Neal Cornell University: Thomas Hickerson Council on Library and Information Resources: Deanna Marcum Dartmouth College: Richard Lucier Emory University: Joan Gotwals Harvard University: Nancy Cline Indiana University: Suzanne Thorin Johns Hopkins University: Winston Tabb Library of Congress: Laura Campbell National Archives and Records Administration: Kenneth Thibodeau New York Public Library: William Walker New York University: Carol Mandel North Carolina State University: Susan Nutter Pennsylvania State University: Nancy Eaton Princeton University: Karin Trainer Stanford University: Michael Keller University of California, Berkeley: Thomas Leonard University of Chicago: Martin Runkle University of Illinois at Urbana-Champaign: Beth Sandore University of Michigan: William Gosling University of Minnesota: Wendy Pradt Lougee University of Pennsylvania: Paul Mosher University of Southern California: Lynn O'Leary-Archer University of Tennessee: Barbara Dewey University of Texas at Austin: Mark McFarland University of Virginia: Karin Wittenborg University of Washington: Lizabeth Wilson Yale University: Alice Prochaska Absent, with regrets: California Digital Library 3 of 4 DLF Allies were represented: Coalition for Networked Information: Clifford Lynch Online Computer Library Center: Lorcan Dempsey Research Libraries Group: James Michalko Absent, with regrets: Los Alamos National Laboratory Research Library Also attending were DLF Director David Seaman and DLF staff members Cassie Savage and Gerald George. The meeting began at 2 p.m. on Thursday 6 February 2003, with Mr. Keller in the chair. Purpose and Background Papers
Scenario Discussion Several participants individually expressed discomfort with the scenario on the following grounds. The scenario seemed too closely extrapolated from current standard library functions rather than considering possible new roles for libraries in education and publication. The scenario assumed continuity in the legal and policy frameworks of libraries, and treated them as containers of information rather than as supporters of scholarship and learning. It was unmoving and spoke insufficiently about the real business of higher-education institutions, which is to meet needs in education and research. It did not acknowledge new kinds of partnerships that libraries could forge with scholars and others in the future, and it did not deal with the staffing component of libraries. Noting the unpredictability of changes in how data will be used, Mr. Keller said that the development of massive content as envisioned in the scenario would support new functions and meet real needs. Several participants nonetheless called for a different emphasis in the scenario, and expressed the following views. Planners should first determine how digital libraries could create value for research and learning in a changing environment, then work back to what to do. The scenario did not speak to dynamic learning possibilities, to what libraries would do with massive digital content, or to what roles various players would have. To make the content useful would require a superstructure. Planners should not jump to conclusions that libraries will have future roles, but instead should identify what work will need to be done and who best can do it. The DLF should look at points of intersection and tension to determine where it can make a difference. The discussion then turned to changes in higher education and in library roles, dwelling on the following points. Differences among institutions and among fields are expanding (for example, between natural scientists and art historians), reducing points of continuity. A lot of what is now available for scholarship and learning is not in the library, and a digitized humanities environment may not need the library at all. However, libraries may be needed to help users negotiate masses of digitized information, and libraries will want to be the trusted repositories for data. There is a need for people who understand how information is created and know how to package and deliver it for use -- people who can help users get and use data; digital libraries can be information brokers. One participant countered that the concern should not be how to position the library for the future, but how to help educational institutions maximize the benefits of technological development; and another predicted that useful functions and roles would emerge as change unfolded. The agenda then called for the group as a whole to break out into three smaller groups, charged by Mr. Keller to test assumptions in the scenario and report proposals for changing it. First Break-out Group Reports Group #1 Ms. Dewey reported for Group #1, chaired by Ms. Eaton (pending the arrival of Mr. Walker), and containing also Mr. Hickerson, Mr. Leonard, Ms. Mandel, Ms. Marcum, Mr. McFarland, Mr. Tabb, and Ms. Thorin. Group #1 proposed the following, revised scenario. The DLF would continue to take advantage of its ability to be nimble, assemble smart people, move quickly, and get things done. It would adopt three priorities: (1) creation of digital material (2) interlinking (3) preservation Members are already close to content sharing and would move to the creation of a Digital Library Federation collection. To achieve this, digitization budgets would be enlarged to support the emergence of a virtual collection for DLF user communities and others. DLF members would help institutions that do not have large enough budgets to develop such resources. In response, individual participants said that -- simultaneous with developing content -- collaborative work would be needed to provide services and analyze user needs. Also, attention should be given to scholars doing new things. Work toward a DLF collection could start with a focusing project. Group #2 Ms. Sandore reported for Group #2, chaired by Ms. Nutter, and containing also Ms. Cline, Mr. Dempsey, Mr. Gosling, Mr. Lynch, Ms. Prochaska, Ms. St. Clair, Ms. Trainer, and Ms. Wilson. Group #2 proposed that missing elements be worked into the original scenario, which focused on content rather than people, noting that information resources are for use. Libraries can co-evolve with faculty and other content creators. Collectively, the DLF would plan not for a fixed point ten years out but for an organic process of transformation, in which libraries would integrate collections, interoperate, build services to help users find information, and work with other creators on curation and navigation. Libraries should translate across disciplines and datasets; they have an important bridge role on which to capitalize. They can add value to information, help different disciplines to communicate, and be trusted repositories for archiving. Describing these as "preferred directions," Group #2 did not propose to throw out the original scenario but to put the library into it in a revised manner. Group #3 Ms. O'Leary-Archer reported for Group #3, chaired by Ms. Wittenborg, and containing also Ms. Gotwals, Mr. Lucier, Ms. Lougee, Mr. Michalko, Mr. Mosher, Mr. Neal, Mr. Runkle, and Mr. Thibodeau. Group #3 proposed the objective of improving the efficiency of the scholarly enterprise. This would include balancing the interests of knowledge creators and consumers in an access management system. Libraries would have a role in this and would need to collaborate to carry it out. There are tensions in the interactions of technology, information organizations, and preservation needs, and there are economic issues that cannot be ignored. Libraries would be not just content builders but enablers of knowledge creation and sources of balance among the tensions. The DLF and its member libraries would break down barriers and promote communication through all phases of the continuum from knowledge creation through knowledge consumption. All groups having reported, the agenda then called for the break-out groups to return to their individual deliberations, charged by Mr. Keller with identifying methods, opportunities, hurdles, and intermediate goals pertinent to their respective scenarios. Second Break-Out Group Reports
Group #1 Mr. McFarland reported for Group #1, chaired by Mr. Walker. To develop a DLF collection, Group #1 said it would be necessary to recruit helpful scholars, identify collections to digitize, and get input from users. DLF constituent institutions would be able to download content into local systems, for cross-searching, annotation, etc. They would leverage what they already have, ask scholars what they want, get communities of scholars to partner in the endeavor, and enable end users to manipulate content, providing tools for doing so instead of just reformatting content. Concerning barriers, collaborative collection sharing would be useful to a broad population, but some librarians are uncomfortable about sharing digital masters. Group #1 reported disagreement about what level of sharing libraries would provide for a DLF collection. Libraries know how to build content that can "behave" differently, content that can look different in different projects; but libraries may not be able to give access to content for anyone to use in any way. Group #2 Ms. Wilson reported for Group #2, chaired by Ms. Nutter. Group #2 advocated setting scenarios aside in favor of preferred directions. The group recommended an organic, iterative approach in which libraries would undergo co-evolution with their universities within overall transformations. Libraries would focus on creating value for users, overseeing and facilitating the continuum from knowledge creation to consumption. Libraries themselves would have no divine right to any role, but would be part of a framework of trusted partnerships and collaborations, interdependent with communities of learning. Barriers include all kinds of silos and fences. They are a big problem because service expectations have skyrocketed. Users give highest value to digital access. Included are unforeseen and transformational users who make use of materials beyond what was expected or intended. In Group #2's view, the DLF would be involved in all elements of the digital world—collections, services, architectures, uses. The DLF would engage in developing standards, best practices, and a shared architectural view. Group #2 described its "picture" as "Collide-a-scope." Group #3 Ms. O'Leary-Archer reported for Group #3, chaired by Ms. Wittenborg. Group #3 recommended three areas for DLF involvement: (1) The DLF would develop a framework for collaborative digital content development; that is, a framework to enable and facilitate collaboration among partners, both to create digital content and to convert traditional resources to digital. (2) The DLF would engage in research and development activities, such as modeling, testing, and problem solving, concentrating on enabling infrastructure and applied developments. (3) The DLF would build, support, and nurture alliances and partnerships to advance digital library developments. The DLF would have to overcome economic, technological, and legal barriers; surmount stakeholder differences; and develop staff expertise and business rules. At the end of these reports, the meeting was adjourned until the next morning, Friday 7 February 2003, at 9 am, with Mr. Keller in the chair. Second Day: Recap and "Straw Man” Proposal Mr. Keller began with the following recap of the previous day's session. The DLF has accomplished many good projects devised by its members' technical people, including establishing registries and advancing the Open Archives Initiative (OAI). The DLF has been developing architecture amidst change, and now may need a new, high-level view with functional specifications. The DLF has worked collaboratively and flexibly, forming small groups as needed. The DLF has been action-oriented, responsive to library staffs and the needs of institutions and patrons, and an agent for transformation. The question now is how to leverage what the DLF has done? The previous day's reports, Mr. Keller continued, had provided some ideas. DLF members are ready now to move to sharing resources and services, federating to benefit scholars and learners. They are ready to talk about collaboratively building a collection or some collections, and engaging scholars in the process, recognizing that tools and collections go together because the availability of collections generates demands for new tools, and vice versa. Mr. Keller then asked Mr. Seaman to elucidate a "straw man" proposal derived from the previous day's discussion. Mr. Seaman proposed that as a first step the DLF collaboratively build new content as a test bed, involving collections, services, and users. An individual institution could commit to working with the test bed in any area with which the institution was comfortable. Each institution contributing to the collections track, Mr. Seaman explained, would provide 50 digitized items, using selection criteria to be developed. These items would come from the humanities, from American culture, or from whatever else made sense locally. Each institution would contribute its titles to the shared collection, and get back access to X times that number, determined by the number of participating institutions. Lots of collections already digitized could also go into this DLF library, which would be a prototype for a collaborative body of recombinant content. The content could be collections of images, books, audio, or video. Concerning services and users for the test bed, Mr. Seaman said, an institution that did not feel able to contribute content to the collection could commit to help build context, could work with their faculty on how to mine the content, or could actively partner with users in developing services. Mr. Keller added that institutions could create links from digital objects in the collection to other literature and resources, and could assess the value of services to users. On Day One there was some resistance to federating what libraries already have digitized, he observed, so DLF members would create this initial prototype collection with material newly digitized. Mr. Seaman noted that for some purposes it may be enough to be able to get access to digital content on another institution’s site, but that alone is not enough for some scholars, who want to search material simultaneously from many sites, re-shape it for a particular audience, software tool, or hardware device, or add levels of encoding to it for a particular local project. These types of use are rarely possible now, and their expected high value is an assumption that the prototype project could test. In response to Ms. Lougee's observation that a design process would be needed for so complex an undertaking, Mr. Seaman said that members of the DLF with relevant expertise could be pulled together for that purpose. Mr. Keller posited a need to get the functions down, determine what the collaborators would want to do with the content, and then re-deploy the functions in an access structure. He also described how the design could incorporate descriptive bibliographies as scaffolds for linking items in the collection to pertinent material elsewhere, with variable presentation options for readers. There would be an extensible set of possibilities. Two participants observed that users would have to be consulted on they would want and used. And Mr. Lucier raised questions about success factors. Trying to create a model for federated content would be enormously complex, he said; earlier DLF work had been done by small groups, and such a group would be useful for "reality testing" of the proposed "straw man." Involving everyone could be difficult, he said, because it is hard to uncouple content and services, and might not be desirable; but an expert small group could begin working out arrangements without the need for all institutions to participate right away. The project could develop in phases under Mr. Seaman's management. Other comments from the group in a lively discussion included the following:
Mr. Michalko described efforts in which the RLG has been involved to federate already digitized materials from some libraries. The need is to make scattered collections into something with utility, he said. The hard part is developing an applied policy model. People from projects already going on would need to come together to figure out where technological, business, and institutional policies intersect. A monumental project is not necessarily needed. Questions include whether to federate virtually or physically, how to create effective user experiences, and how to reach consensus on policies when libraries operate so differently, using different page-turning mechanisms, for example. Institutional imperatives, he continued, reduce the will to collect collaboratively, making it hard to get resources diverted in a common direction. And service manifestations need to be different for different disciplines, which use material in different ways. But subsets of people within the DLF could work on such issues. In what areas could we break through local imperatives to meet a collaborative need? The ensuing discussion produced the following points. Current budget problems for libraries -- not uncooperativeness -- inhibit collaboration; but libraries cannot afford to build large digital collections on their own, so they must engage each other in spite of the difficulties and try something, hoping for support from scholars who will say that the results are really good for them. The more faculty members who say they want what libraries develop, the more librarians can justify it in their budgets. One approach might be to deconstruct successful existing collections (Making of America, etc) to see whether and how they might develop differently today. A common collection sponsored by DLF would be an attractive alternative to more and more titles coming in library doors. Proposals to help build a DLF collection could be requested, and the collection could draw on digitizing work that libraries are already doing. Each library's collection could become everybody's. But a goal should be to get this project embedded in normal library budgets, not as an add-on. Mr. Seaman explained that getting collaborative collection building into the regular ongoing budgets of multiple institutions is one of two ways in which the straw-man proposal would differ from previous kinds of collection sharing. The other difference is that material in the collaborative collection would be re-combinable for users' own purposes rather than only usable on the terms (functional and aesthetic) of the producers. Mr. Keller added that the collaborative collection would need a simple publishing environment allowing users to get texts they can reformat. But he reiterated that not all DLF libraries would have to take part in all elements of developing the collection. The discussion continued with the following points. Selectivity in content for the collection is not of great importance because whatever is digitized in bulk gets used. Experience has shown that how useful a digitized collection may turn out to be cannot be predicted. High-Wire Press has found a pattern in which the use of newly accessible digital material soars, then descends through a series of plateaus, but remains greater than the use of print versions of the same material. However, can libraries ensure that digital material will be around forever? The agenda then called for the resumption of deliberations in the break-out groups, which Mr. Keller charged to consider the following questions: What basic functions of libraries need to be accounted for in planning a collaborative collection? What functional features would be desirable in the beginning? What types and genres of material would a prototype include? What will need to be built that digital libraries do not already have? What opportunities are there for new partnerships? What is best for storage and access? What are the necessary commonalities? Groups should assume, Mr. Keller said, that funding can be obtained for the prototype. Third Break-Out Group Reports Group #1 Mr. McFarland reported for Group #1, which was chaired again by Mr. Walker. Group #1 recommended the establishment of an "Initiative Committee" to develop a framework within the DLF for moving forward on a collaborative collection. The committee would consist of Mr. Seaman, as DLF director, and five directors of DLF-member libraries, only one of whom was on the Executive Committee. The committee would report back to the DLF Steering Committee in a short time. A coherent theme or focus for the collaborative would be desirable, something with a compelling vision to capture public attention. One possibility, the group suggested, would be a Digital Library for the Humanities. But the initiative should start with something scaleable, some piece of the humanities that is interdisciplinary. Group #2 Ms. Wilson reported for Group #2, which continued to be chaired by Ms. Nutter. Group #2 recommended that the DLF "leverage the interdependence of federated collections," without the unsustainable overhead of a "central bucket." DLF institutions wishing to make their digital materials available would do so in ways that would make possible recombinant uses of material, in exhibits, lab notebooks, course projects, courseware, and other forms of customization. Archiving, registries, and other services would be developed. Development would be phased and allow for different levels of engagement by DLF institutions. "DLF commitments" would be to adhere to minimum standards, develop content as a "circle of gifts (COG)," and migrate content, keeping responsibilities local and in alignment with individual institutional incentives and missions. Group #3 Ms. O'Leary-Archer reported for Group #3, chaired again by Ms. Wittenborg. Group #3 described two tracks that might be followed: (1) a federation of existing content in DLF libraries, which would require design work and money for migration; and (2) a transformational new digital library.The latter would have several attributes. One attribute would be an integrated effort that would be experienced as a whole and seem a single enterprise from the user perspective; yet this digital library would recognize the individual contributions of institutions in the collaborative. This digital library would be collaborative at a high level, allowing for ingestion, discovery, archiving, and user tools, but in sustainable ways. It would use open-source tools but be commerce-compatible. Financial and governance models would be needed, particularly concerning how non-DLF participants would give to and take from the digital library. Group #3 proposed to consider two projects of this kind, involving the sciences as well as the humanities, but focusing on subjects needing interdisciplinary collections for study. Conclusions Mr. Keller summarized what he perceived to be the outcome of the deliberations. 1) The DLF needs an Initiative Committee to frame a collaborative digital library project and produce a focused plan for its first phase. Members of the committee would be Mr. Seaman and five DLF-member university librarians, no more than one of whom would be serving also on the Executive Committee. 2) The library to be constructed would have an "umbrella collection" covering something compelling to the public and broad enough to permit growth. It would be a multi-media collection. To create a Digital Library for the Humanities might be too limiting; the choice of collection should not limit the model. The collection would embrace particular components under a generic, extensible title that identifies goals; "JSTOR" provides an example. The Initiative Committee would identify specifications for a functional architecture, with help, if desired, from other committees of specialists. Mr. Keller said that a chart devised by Group #2 [Appendix 1] summarized much of the meeting's discussion. It called for a distributed rather than a central library. However, Mr. Keller said, a determination of what the functionality would be should precede decisions about the degree of federation; there could even be variations such as a central collection with federated services. However it was arranged, there was common agreement that it had to appear to be a unified body of material – a single library -- to the user. In response, some participants expressed concern about the financial sustainability of centralization and about the difficulties of enforcing legal, ownership, and branding restrictions on contributed content. They advocated leveraging their libraries' interdependence, with recognition of sources of contributions. Mr. Keller concurred with Group #3's proposal to use both old and new digital material in the collaborative project, but observed that ingestion is a huge task and that shared (but not necessarily "open source") tools would be needed. Mr. Keller noted that the plan was for federated collections, not a single one, qualifying the "umbrella collection" idea (#2 above) to mean an umbrella sheltering many collections in many different digital libraries across the DLF membership. The discussion included observations that DLF members should begin the project, but then open it to participation by others. Global access would be the ultimate goal, and the project might be used to seed an international group. Mr. Keller continued his summary of the deliberations with the following points. (1) For this initiative to be credible, a key shared assumption would be the need for true, digital archives at the project's foundation. (2) Also there will be various registries, freely accessible. (3) The test phase will need some kind of access-authorization control. (4) There will be a common platform for user access, and common coding and formats for various kinds of objects. (5) Tests should be conducted on some groups as the project goes along to see if users can use the content creatively, pulling material into their own environments. A brief discussion followed about the desirability -- impractical though it would be at the outset -- of developing an enhancement to allow scholars' creative uses of the collection to be brought back into it. Next Steps As the next step, Mr. Keller requested from each participant in the strategic planning meeting a statement of what that person and her or his institution would be willing to do in line with the plans discussed. Examples included contributing collections, recruiting users, and/or serving on committees to work on particular elements of the project. Mr. Keller asked to receive these statements within a week. The Executive Committee and Mr. Seaman will look at the responses in preparation for setting up the Initiative Committee, which will be asked to report within two to three months. Calling attention to the first paragraph of the DLF's current charter <www.diglib.org/about/dlfcharter.htm>, Mr. Keller concluded that no charter modification would be needed to accommodate plans as discussed at this meeting. In terms of publicity, Mr. Keller counseled that no press release or public announcement be made until an appealing name for the project has been found, potential funders have been approached, and assets have been gathered and made ready for use. However, the project's agenda could be advanced at the next DLF Forum. At approximately 2 p.m., the meeting was adjourned. Respectfully submitted, Gerald George, recorder Appendix I Text Rendition of a Chart Devised by Group 2 |
Copyright © 2004 by the Digital Library Federation
Last updated:
