Too Much Consensus
by Deanna B. Marcum
Test Database for Digital Visual Resources in Art History
by Anne R. Kenney
Building Digital Libraries: Small Is Beautiful
by Abby Smith
“Localizing” Reference Linking
by Priscilla Caplan, Dale Flecker, and Daniel Greenstein
There are enormous challenges ahead as we attempt to define digital library services. Questions abound: How will digital information be managed, delivered, and preserved? How do we engage scholars to consider their roles as data and knowledge creators, and how do these roles determine the long-term access to and preservation of that knowledge? What organizational structures must be put in place to meet new needs and requirements? How do we help university administrators understand that building a digital library requires a rethinking of the very core of an institution?
These are big questions, and responding to them requires innovation and leadership. In many cases, leadership can be exercised only by carrying out research and development to determine feasible options. Nonetheless, we cannot wait for all the needed answers. To advance digital library development, we are compelled to move quickly with the best practices that can be determined at the moment, recognizing that what constitutes “best” changes quickly and that we must be prepared to adopt better methods as they are identified.
Because change is so rapid, libraries can make real progress only if they learn from the results of other institutions’ work and build them into their own systems. The ability to do this requires libraries to work in quite a different way.
We are caught in a contradictory moment. At a time when libraries and the organizations that serve them are compelled to transform themselves in response to the demands of digital technology, our organizational structures and behaviors are still tied to the social dictates of the 1980s and 1990s. Librarians tend to value consensus above all else, but in doing so, we severely limit our ability to move forward.
It is now simply expected that everyone must be represented at the decision-making table and that everyone’s interests are served by the decisions taken. This approach, I believe, seriously impedes our collective ability to act decisively and to respond to market forces with intellect and vigor.
Instead of spending time reaching a consensus on what approach will reflect well on all, we should be experimenting with new approaches, developing promising new tools, and undertaking serious research that will illuminate the dark corners of our understanding. Being even bolder, we should initiate new services and create organizations that provide what scholars, researchers, and students need. We need to think imaginatively about how to provide services in ways not bounded by library walls. Such approaches require courage, and they require partnerships with organizations unfamiliar to us.
Instead of consensus, we as a library community need to develop trust among ourselves. We should promote new ideas, even radical ideas, instead of expecting all our colleagues to follow a single direction. We need to pledge to one another that we will learn from what each organization is doing and leverage that learning to useful purposes. If we cannot trust other organizations to keep us informed about what they are doing and have learned, we are destined to limit our accomplishments. We cannot be at all the decision tables. Our traditional insistence on being involved in everything simply keeps us busy; it moves us ahead little, if at all.
Librarians have expected the service organizations to develop a consensus on the roles each will play so that individual institutions will not be funding duplicate activities. The desire to avoid duplication is commendable, but I wonder whether this is the appropriate time for such careful coordination. Breakthrough thinking is needed to solve the problems of ensuring access to scholarly resources of the twenty-first century, digital archiving, and digital library development. Libraries need to demand that the organizations that serve them show they are making real progress toward solving some of these problems.
Last spring, the Cornell Institute for Digital Collections (CIDC), under the direction of Peter B. Hirtle and Anne R. Kenney, convened a group of researchers, technologists, archivists, librarians, scholars, and funders to discuss productive ways to develop and use digitized text and images in support of teaching and research.1 The group came up with several recommendations that CLIR will be pursuing over the next year.
Of immediate interest is the group’s strong recommendation to build a rich test database of digital visual resources in art history. Unlike the Academic Image Cooperative, which is developing a collection of curriculum-based digital images to be used for teaching, the database is intended to serve as a testbed for the development and evaluation of technical applications for creating, managing, and exploiting digital image content. It would be constructed to support experiments by the educational, cultural, commercial, and research communities that would yield statistically valid results in several areas. For instance, such a database might support experimentation and evaluations in quality assessment, image processing, content-based image retrieval, image registration, authentication, performance metrics, migration, and descriptive searching. The database could provide a powerful unifying force for comparative analyses that cover a range of disciplines, technological approaches, and longitudinal studies. Such test databases exist in other domains—full text, fingerprints, handwriting, photography, and face recognition—and their availability has driven the development of new processes and products.
The corpus of images selected for this initiative would have to provide sufficient breadth, depth, and variety to constitute a statistically valid database for the field. It would need to be promoted as a community resource that would provide a common benchmark against which to measure and compare processes and approaches. Although initially focusing on art history resources, the database would be developed in such a way that it could serve the humanities computing culture in general and accommodate new lines of research as they arise.
CLIR is supporting the work of a small planning group, chaired by Clifford Lynch, executive director of the Coalition for Networked Information, to develop a formal proposal for the test database that can be submitted to funding organizations in 2001. The planning group will consider scope, methodology, one-time and ongoing costs, requirements for an institutional home, and appropriate uses of the database. As part of its charge, the group will examine comparable databases and gauge user demand for a testbed for digital visual resources. It is anticipated that a Request for Proposals for both creating and maintaining the database will be developed as a result of the planning group’s efforts. By supporting this initiative, CLIR is serving as a catalyst for the development of an important resource tool that can serve both the visual resources and the digital library communities over the next decade and beyond.
1CIDC’s mission is to explore the use of emerging technologies for expanding access to cultural and scientific collections. It develops digital resources and supports their effective use for research, teaching, and scholarship. CIDC also conducts applied research to test and evaluate the utility of such resources for the Cornell community as well as diverse global audiences. See http://cidc.library.cornell.edu.
CLIR welcomes Gerald George to its staff as special projects associate. He will work with both CLIR and the Digital Library Federation on a variety of outreach projects, including adapting information from CLIR and DLF projects to specific audiences.
Mr. George formerly served as director of the communications staff of the National Archives and Records Administration, where he was responsible for Congressional affairs, public affairs, Web site content management, four periodical publications, and the preparation of ad hoc communications. He joined NARA in 1991 as executive director of the National Historical Publications and Records Commission, a federal grant-making agency administered within NARA.
Before joining NARA, he served as director of the American Association for State and Local History. He has been a consultant to historical organizations, a newspaper reporter, and a freelance writer. He has published three books, several special reports, and more than 50 articles in journals, magazines, and books. Mr. George has a B.A. in history from Wichita State University, and an M.A. in history from Yale University.
Any library—or individual, for that matter—creating digital resources for online access is building a digital library, a network of resources designed to provide information to the curious. Nearly all libraries, regardless of size, already have the building blocks. If they have an online catalog, for example, they have metadata. Yet among libraries, there is a pervasive sense that size matters: that the big libraries are doing more, doing better, and doing well. Smaller institutions must be doing less, doing worse, maybe even doing poorly. This plays on the assumption that the greater the resources—read: money—the greater the possibility for success.
But success at what? If the mission of libraries is to provide information to their patrons, they now face the challenge of identifying the audiences they serve online and of tailoring their resources to them. In some respects, the smaller the institution, the easier this is. A New England historical society with incomparable whaling collections can develop a highly focused digital collection that can extend the reach of its holdings and services to its prime constituency, a well-defined, if far-flung, group. From New Bedford to Hokkaido, the whaling community will know where to go with queries. In this case, size does matter, and small is beautiful.
The real challenge in building a digital library is that we have no blueprint. Everyone is free to build such a library, of course, but without a vision, even a big, well-funded institution can end up with a structure that is uninhabitable or merely a house of cards. Every institution faces the awkward situation of having to learn as it goes if those that have gone before are unwilling to share their knowledge. The willingness to risk failure in the pursuit of experience and knowledge is not a notable feature of library organizational cultures.
One thing we have learned as we have worked with the new technology is that both its possibilities and its limitations are greater than we had originally thought. Whereas once we might have seen digitization, for example, as a way to serve collections to our regular patrons or to extend their reach to new audiences, or as a superior means of creating preservation surrogates, we now know that it is all that and much more. Now we know that we are doing something radically different by creating new library assets and adding value. We are not realizing cost savings, at least not in significant ways, but we are providing better services in many cases. Moreover, we find ourselves in an environment that is user-driven, not library- or collection-driven, and our users, many of them new to us, also have very high demands. Where once patrons understood the rules—that you get access to resources on the library’s terms, during its hours, and often with delays—now patrons want and expect what has been called “anytime, anywhere” service.
Building digital libraries also puts libraries in a new relationship with one another in the as-yet ill-defined digital commons. One thing this digital commons does not reward is competition among libraries. The digital commons rewards cooperation—in building collections, in sharing resources, and in developing standards for interoperability. Large libraries generally find it much more difficult to solve problems through cooperation than do smaller ones, often because the organizational cultures of the former reward competition. They are used to competing for collections and research grants, for example, and they exist within a larger institution—the research university—that also encourages competition for donations, research funds, and even students.
Research libraries find themselves significantly hampered by the legacy of their collection-building strategies. To support original scholarship in addition to teaching, they have collected “just in case” and so have created very large and inevitably redundant collections that face a problematic future. Research libraries find themselves burdened with the responsibility to preserve collections that are valuable but seldom used. Funds for preservation of those collections are eroding. But in today’s new economy, in which knowledge is touted as the “new currency,” it will be hard to justify the existence of redundant collections. CLIR is working with the Task Force on the Role of the Artifact in Library Collections to reexamine old assumptions about the best ways to support those collections in the library of the future. We must devise new models for sustainability. College libraries and others with collection-development policies that place a premium on fitness for the purpose of teaching will make the transition into the mixed digital and analog realm more efficiently than will libraries that do not have such policies.
Whether big, small, or in between, there is no library that, if given all the funds it wanted, would be guaranteed a successful, let alone smooth, transition to the library that it needs to become. What seems to be missing is a clear vision of the role digital technology should play in helping libraries achieve their core mission: service and stewardship of cultural and information resources. When it comes to meeting that challenge, size does not matter; focus does.
by Priscilla Caplan, Dale Flecker, and Daniel Greenstein
The Digital Library Federation (DLF), with the National Information Standards Organization (NISO) and other bodies, has been involved for some time with the issue of citation linking in the world of commercial electronic journals.1 In recent years, there has been much activity related to the building of persistent links for e-journal articles. The way in which these links are implemented will have profound implications for the freedom of individual libraries to select appropriate services for themselves and for the development of competing commercial systems and services.
The most notable development in recent years has been the creation of the Digital Object Identifier (DOI) and its use in systems such as CrossRef.2 Use of the DOI is filled with enormous scholarly promise; it affords users an opportunity to navigate seamlessly between citations—whether they appear in an abstract and indexing service or as a footnote in an e-journal—and the works themselves, provided they are available in electronic form.
Nonetheless, there are problems with the current implementation of the DOI. For example, it is capable of finding only a single copy of a work. Since the publisher deposits the DOI and the associated location, the copy that the DOI points to is generally the one stored at the publisher’s own site.
Since a great deal of linking will be based on the DOI, this limitation will have significant consequences. In the current environment, it means that links will not be made to articles loaded into local systems such as OhioLINK or the system at the University of Toronto, or in aggregator databases such as OCLC, ProQuest, or Ebsco. Instead, users with access to such alternate copies will be sent to the publisher’s site and either turned away or asked to pay. As our environment becomes more complex with the development of e-print aggregations and archiving services, this limitation will become even more problematic. The effect may well be to create a monopoly for publishers, not only in the original marketing of articles but in their accessibility over time.
Recently, a group of libraries acting under the DLF and NISO umbrella approached CrossRef, the first major application of cross-publisher reference linking that uses the DOI. The libraries urged CrossRef to participate in developing a system that would allow the users to locate an article and that would support multiple copies and systems. A series of meetings and discussions followed, most notably a workshop involving representatives from CrossRef, several major scholarly and commercial publishers, the International DOI Foundation, the digital library research community, NISO, and research libraries. The workshop resulted in a general agreement on a possible solution to allow what is now being called “localization” in linking (that is, linking to a copy that is not held by the publisher). CrossRef, working in collaboration with the DLF and several digital libraries, has agreed to create a prototype for that solution. A report of the workshop is available from http://www.niso.org/CNRI-mtg.html.
Journal publishers have become more aware of the issues involved in linking and seem increasingly inclined to accept the need to open up their linking infrastructures. Should work on the prototype be successful, the case for opening up linking infrastructures should be much stronger. Yet more remains to be done. Pockets of resistance persist within the publisher community. Accordingly, at a time when many publishers are debating what stance to take, it is important that libraries individually and collectively make clear their vital interest in these developments.
2CrossRef is a nonprofit, independent organization comprising leading scientific, technical, and medical publishers established as a central source for reference linking. See http://www.crossref.org/.
CLIR has just issued its 1999-2000 annual report. It is available as a .pdf file on the Web at www.clir.org. Print copies are also available from CLIR at no charge. To order, please direct your request to firstname.lastname@example.org.
Librarians rely on usage statistics to inform a range of decisions, from acquisitions to storage. With the advent of electronic publishing, however, usage data reside with publishers, and fewer than half of them currently generate such statistics for libraries.
The task of providing librarians with needed statistics represents a significant additional cost for publishers, and many fear that librarians may use the information to cancel subscriptions. Assuming a publisher is willing to make statistics available, however, the primary challenge will be to provide meaningful data. Guidelines for data collection are just emerging, as are definitions of what constitutes “use.” Librarians and publishers must work closely to agree on which statistics should be collected and how they are collected to ensure that the data are comparable.
To provide a basis for dialog, CLIR commissioned Judy Luther, president of the consulting firm Informed Strategies, to develop a white paper on usage statistics for electronic journals. Her findings, submitted to CLIR in September, are based on extensive interviews with publishers, information aggregators, and librarians.
The white paper reveals that publishers and librarians share a significant number of concerns: a lack of comparable data, a lack of context for understanding data, incomplete usage data, lack of certainty about effective economic models, and issues of user privacy. The author also notes that publishers who have provided usage statistics to librarians are discovering that librarians are not using the data to cancel subscriptions. In fact, as publishers come to terms with the costs of developing the capability to collect and analyze statistics, they are expanding the use of the data for their own purposes. For example, systems staff use data to budget for new hardware as usage increases, and product-development staff analyze how users access content. The marketing department is interested in how users find the site. Sales staff want to know about the level of activity of their customers, and editorial staff want data on the most requested search items.
The author of the white paper notes that currently, “associations involved in creating standards and guidelines on data collection are focused on defining the data elements and determining what is currently being done. No one is working directly with the publishers who have developed data, understand the variables, and are in a position to provide guidance.” Because it is a period of discovery for both librarians and publishers, the author recommends that a forum be held at which invited representatives from the publisher, vendor, and library communities could advance the dialog about providing data that can be compared.
The White Paper on Electronic Journal Usage Statistics will be available shortly on CLIR’s Web site (www.clir.org). Print copies of the report will be available for ordering.
Leading the Information Revolution in 21st Century Higher Education
The Frye Leadership Institute is now accepting applications for its 2001 session.
The challenges and changes of contemporary higher education have created the need for campus leaders with new competencies and perspectives—people who can bring a new framework to our historic enterprise. The purpose of the Institute is to bring to tomorrow’s higher education leadership the insights and understanding of the issues that will inform this framework, including academic, technology, economic, public-policy, student, and constituent-relations dynamics.
Through presentations by recognized leaders in higher education and society, seminars, group projects, and reading, participants will study and analyze the leadership challenges stemming from the changing context and complexity of higher education. The program will pay special attention to the implications of the growing power of information technology to tranform the means of research, teaching, and scholarly communication.
The Institute is an intensive, two-week, residential program held at Emory University. The 2001 session will be held June 3-15, 2001. Participants will be selected competitively from among nominees and applicants who have a commitment to, and talent for, leadership within higher education. The group as a whole will be chosen to reflect the variety of backgrounds and diverse expertise that constitute higher education.
The Institute is supported by a grant from the Robert W. Woodruff Foundation and is sponsored by CLIR, EDUCAUSE, and Emory University.
Nominations are due by November 15, 2000. Completed applications must be submitted by December 15, 2000. Additional information, including application instructions, is available through the Institute Web site at www.fryeinstitute.org. The Institute can also be contacted by e-mail at email@example.com or phone 303-449-4430.
|Council on Library and Information Resources|
1755 Massachusetts Avenue NW, Suite 500
The Council on Library and Information Resources (CLIR) grew out of the 1997 merger of the Commission on Preservation and Access and the Council on Library Resources. CLIR identifies the critical issues that affect the welfare and prospects of libraries and archives and the constituencies they serve, convenes individuals and organizations in the best position to engage these issues and respond to them, and encourages institutions to work collaboratively to achieve and manage change.