Diane Nester Kresh
In 1990, the Library of Congress began its American Memory Project in response to Librarian of Congress James H. Billington's initiative to find ways for the Library to disseminate in electronic form its unique special collections of photographs, maps, manuscripts, and other special format materials. With the rise of the World Wide Web, American Memory became part of the Library's National Digital Library (NDL) Program and in the five years since the NDL began in 1995, the Library has added 70 online collections of Americana, has collaborated with other institutions to create standards and best practices for the capture and retention of digital images, and has begun to experiment with and define selection criteria for the capture of research materials that exist only in electronic form. While the challenges of managing and sustaining a digital library are significant, they are not the subject of this paper. Though digital libraries have the potential to increase learning and knowledge, to date developers of digital libraries, the Library of Congress among them, have focused almost exclusively on the provision of content. What is a harder task is to take that same content and make it usable and searchable in ways that can engender discovery, creativity and invention.
At no other time in history has the emergence of technology affected so significantly the core mission of a library. These technological advances have created new opportunities for libraries, information managers, researchers and library patrons of all kinds. The Internet is expanding at a dramatic rate and is creating a fundamental change in the way people collect information and acquire knowledge. It is an exciting and eclectic research tool and its power is limitless. A review of some statistics demonstrates how prevalent use of the Internet is and how broad based are its users:
With the emergence of Web help services, libraries are no longer the lone providers of information. At this time, the numerous web-based reference services search only the Internet, not the vast collections found in libraries nor the thousands of library online catalogs that describe and manage the collections. Ask Jeeves (http://www.askjeeves.com/), a commercial site that searches the Internet, reports receiving 20 million questions per day and to date has answered more than 150 million questions. Webhelp.com, a service that boasts that it is the only search service on the Internet to provide "real answers from real people in real time," reports 5 million hits per day. And the list of expert sites keeps growing. According to a recent New York Times article, "Suddenly Everybody's an Expert" (February 3, 2000), "an expert, it seems, is now an ordinary person sitting at home, beaming advice over the Internet to anyone who wants help." The founders of these sites argue that they are providing the missing link to the millions of pages of information. What Internet users need, they say, is human intervention to locate answers that are fast, personalized, easy to find, and free, at least for the time being. Of course, what they are describing sounds a lot like what happens in a library. Research is currently being done on the accuracy and effectiveness of these "expert" services, and the findings, reflecting mixed reviews for the most part, are being published in library literature. The Library of Congress and the University of Washington have undertaken their own study of Ask-A Services (Ask-A Services are primarily expert services serving the K-12 community, e.g., Ask Dr. Math), and the findings will be made available in the next several months. One consortial system in northern California conducted an informal test of Ask Jeeves by sending it 12 questions its libraries had answered. There were no trick questions, none were arcane, just questions typically received by those libraries. Jeeves was unable to answer any of the questions.
In summary, the growth of the Internet is exploding; it is located everywhere and is open for business 24x7. Its content is rich and varied, but easy access is undermined by several factors, including the absence of traditional means of cataloging or organizing information and the absence of distinctions between credible sources and sources created by self-styled experts.
Librarians Add Value
Today's researchers need to find quickly information that is usable, relevant, authoritative and verifiable. To meet that need, libraries must adapt our traditional strengths of acquiring, describing, and serving information to an environment that is not bound by time or physical place, the virtual library without walls. So how do we take the reference desk to cyberspace?
We can start with what libraries are good at. Libraries are good at organizing information using controlled vocabularies and other standards tools to make materials accessible. Libraries evaluate materials carefully before selecting them and according to documented policy statements and guidelines. Libraries have both digital and analog collections which are regularly mined by highly skilled and trained subject, language and navigation experts. Finally, the communications options for libraries have been increasing in number and variety. People can conduct research and ask questions in person, in writing, by phone and fax, and online by email, and experiments using chat rooms and video conferencing have begun.
By applying the best of what libraries and librarians have to offer (structure and organization, in-depth subject expertise, and analog collections) to the labyrinthine universe of unstructured and unverified information on the Internet, we can begin to bridge the gulf that exists between providers and users of information. The Collaborative Digital Reference Service (CDRS), currently being launched by the Library of Congress and its partner libraries, provides just such an opportunity to connect users with accurate, timely, and credible information anytime anywhere. As originally envisioned, it would start small and grow into a vast international service that would allow libraries to help each other serve all their users, no matter where the users would be. From the beginning, we had expressions of interest from all types of libraries around the world.
Why collaborate? Libraries, including the Library of Congress, have a rich tradition of collaborating to get work done. Institutions have collaborated to preserve collections, to catalog materials and make them accessible, and to create virtual libraries. They have borrowed collection items from one another and used one another as service models. By linking libraries for reference services, the CDRS would combine the power of local collections and staff strengths with the diversity and availability of libraries and librarians everywhere, 24 hours a day, 7 days a week. There would always be a librarian available to provide to users located anywhere the interchange and experience of trained assistance in providing access to collections and resources both analog and digital. The graphic below provides an idea of what this network system could look like:
For researchers to be productive and be able to use information around the clock from anywhere, information needs to be organized and available. For information to have intellectual and economic value, it needs constant refreshing with new information, and needs the hands-on touch of the skilled reference librarian to provide context and added value.
The CDRS provides these.
How Does CDRS Work?
Once the Library of Congress announced that it was willing to initiate a project of this scope, a range of libraries volunteered to work with us to shape and define it. These included special, academic, public and national libraries, some of which belonged to local consortia. The CDRS group also received expressions of interest from libraries outside the United States including libraries in Europe, Asia and Australia. The partnership has been enormously beneficial on many levels as each library brings its special experience, knowledge of user behavior and needs, and subject expertise to bear on the project. Following a series of planning meetings in which we defined goals and responsibilities of the participants, we established an implementation schedule which included pilot-testing key elements of the Service before officially launching it later this year. Although the vision is to reach any researcher anywhere, we have begun with libraries "talking to" libraries on behalf of end users so that we can define the parameters of the Service, determine what works and what does not work, and create a Service that is scalable and maximally responsive to user needs. It is exciting to consider that one day, CDRS will be able to provide enhanced reference services to the on-site scholar who conducts research in the physical confines of a library building, resources to a teacher sitting creating lesson plans in Topeka, Kansas, while sitting at a PC, and information to a businesswoman on her way to an international conference, sitting in seat 4C, thousands of feet in the air, who will be able to query the Service using a wireless connection to the Internet.
As envisioned, CDRS includes three component parts: set-up, submission of question and answer, and follow-up and archiving of the answer for future use. The workflow will look like this: An end user will request information through a CDRS member institution. The member institution will send the query to the online Request Manager software for processing and routing. The Request Manager will then search a database of CDRS Member institution profiles looking for the member institution best suited to answer the question. Matches will be made on the basis of such data elements as hours of service, including time zones, subject strengths, scope of collections, types of patrons served, etc. The "match making" will happen in milliseconds. Once a match on an institution has been made, the query will be sent to that institution for answering. Once the query has been answered, it is routed back to the original CDRS requesting library via the Request Manager to allow for closing out the case and completing other administrative tasks. Again, all post-answering processing by the Request Manager should take no more than mere milliseconds. The conceptual model below provides a graphic view of the proposed workflow.
In March of 2000, we began the series of pilot tests to define and test key elements of the Service. During three planned pilot phases, we are examining the set-up component, the question and answer component, the post question and answer component, and the administrative threads that run through all three. We are incrementally testing the exchange of questions and answers among pilot libraries based on library profiles that enable us to deliver an incoming question to the most appropriate and available library, worldwide. Library profiles were created by each library and include addresses (including email), hours of service, collection strengths, staff strengths, what is out of scope, the geographic location of users served, whether there are special services provided and what they are, and average number of questions received. This information will be captured in a table to be used later by the online Request Manager to sort, route and track incoming questions and to deliver credible answers to the end user that are edited and stored in a separately searchable knowledge base of information. This knowledge base, to be populated with diverse and authentic information provided by librarians, will eventually be available to anyone having access to the Web.
In addition to planning the pilots, the Library of Congress hosted a series of workshops with representatives of the participating institutions to define workflow and network requirements, establish business rules, and define a data structure to capture the questions and answers so that they could be separately searched later. All of these issues inform the operations plan that has been created for CDRS.
The CDRS will also supplement the services that libraries offer. Libraries and their local users will be the main beneficiaries of a service that works globally while enabling locally. End users will work through their local libraries, and questions that cannot be answered locally will be sent to CDRS. Membership Service Level Agreements (SLA) will define the nature of the local library's relationship to the CDRS and will be codified in the aforementioned library profiles. A variety of SLAs will be possible and will be limited or expanded depending upon the strengths or limitations of the individual library. For example, a library may agree to: ask and answer questions, only ask questions, ask or answer questions only during specified periods, serve as an editor for the knowledge base or serve as the on-call library if the automatic request manager function is inoperable. In addition to defining roles and responsibilities among the partner libraries, the SLAs will ultimately be used to determine what it will cost a library to be a member of CDRS. The planners have been examining a variety of funding options with the goal of being as flexible as possible, both to allow for the broadest participation among libraries and to ensure that no one library or group of libraries has to bear all of the costs of establishing and sustaining the Service. It is fair to say that we are encouraging maximum flexibility in developing the many component parts of CDRS. For a library to want to participate in CDRS, CDRS has to be perceived to have value. If the participating library defines the terms of that value, that library will have greater incentive to make the arrangement work, for itself and for CDRS. Just as there are no "one size fits all" libraries, so too are there no "one size fits all" SLAs. Libraries are structured and organized differently, they have different local audiences and they have different policies and procedures for ensuring quality control. When a library defines its own terms and level of participation, it wins and CDRS wins.
CDRS Pilots and Launch
The Pilot 1 libraries were: Library of Congress, National Agricultural Library, National Library of Australia, National Library of Canada, Smithsonian American Art Museum, University of Texas at Austin, Cornell University, Santa Monica Public Library, Morris County (NJ) Public Library, and the Peninsula Library System of the bay area in northern California. All contributed edited sample questions and answers which were sent through the system according to a scripted schedule. Pilot 1 had two principal goals: to test the effectiveness of the library profiles and to test a web form for submitting questions. Results indicated that more standardization of the data elements was needed (e.g., agreement on use of a standardized tool such as LCSH to describe a library's subject strengths).
Pilot 2 will begin June 19, and will run until July 17. During this phase, we will raise the bar by adding more institutions worldwide, increase the number of questions, revise the profile database, and begin to experiment with software packages to serve as the online Request Manager. Institutions added to this phase of the pilot include: the University of Washington, Vanderbilt University, University of Southern California, Metropolitan Cooperative Library System (Los Angeles area), Ask ERIC, and EARL Ask-a- Librarian (UK public libraries), for a total of 17 participants. On the administrative end of the project, we have begun to develop a variety of SLAs, to identify staff training needs, and to identify the roles of a CDRS collaborative governing board.
Pilot 3, to begin August 14, will focus on scaling up the workflow, determining the needs for manual and automated back-up systems, and developing a true question and answer archive that becomes a network resource in its own right. The Pilots will end in September and the project will launch officially on October 1.
The CDRS will deliver the direct benefits of quality reference service to a broad spectrum of users anytime, anywhere, a reliable and authoritative knowledge navigation service, a large searchable archive of authoritative answers, and increased visibility and support for libraries everywhere. The service will deliver the indirect benefits of quality support for education and research, the promotion of productivity, commerce, and scientific endeavor, and the basic coin of democracy: value-added information. As we build the service, we are performing a number of behind the scenes analyses to ensure economic sustainability, creating a marketing plan to attract new customers and determining the most cost-effective means of administering the collaborative network. While all of these elements might not be in place by October 1, they are essential to the long-term success of the endeavor.
For the Library of Congress, the CDRS helps meet its mission to provide quality reference services through its international collections of broad subject, language, and format scope, as well as the traditional cataloging systems that support electronic reference. But the Library of Congress does not and cannot do this alone. By enabling local libraries everywhere, LC and its partners bring control and context to the global and world of information.
|Top | Contents
Search | Author Index | Title Index | Monthly Issues
Previous story | In Brief
Home | E-mail the Editor
D-Lib Magazine Access Terms and Conditions