Henry Pisciotta, VIUS Project Coordinator, Penn State University Libraries
Michael Pelikan, Technical Initiatives Librarian, Penn State University Libraries

Penn State's Visual Image User Study (VIUS) is rigorously assessing the requirements for a digital image delivery system at this large and complex university. A broad variety of user studies have been employed during the first 11 months of this 26-month project (including surveys sent to thousands of faculty and students in arts, humanities, and environmental studies disciplines, demographic data for the users of licensed image databases, eight focus group discussions, 22 interviews, and some analysis of transaction logs for image databases). This paper provides a first taste of this "fresh squeezed" data. Some findings: Throughout most of these disciplines use of pictures is widespread and interest in digital images is intense. As with other information services, content is the primary consideration for users. Faculty expect that systematic image delivery will primarily benefit their teaching (although there is an interesting disjuncture between this expectation and the current uses of digital images for other purposes, such as research). Some search techniques are more associable with researchers and independent learners than with teachers. Many teachers and a surprising number of students maintain individual collections of pictures for educational use. Faculty who maintain large picture collections are more likely to use other collections than those who maintain small collections. Users want to be able to easily mix "their" pictures with those supplied by a delivery system. Although users want good retrieval tools and good presentation tools, they do not necessarily want these to be combined in the same system. The support for these and other findings will be described.

Linda J. White, Digital Project Coordinator, Library of Congress

Advances in technology are bringing about a redefinition of the nature of reference services. Trends indicate that patrons are going online for their information needs, challenging librarians to harness new tools to meet patrons at their point of need. One solution for creating and maintaining a digital reference service is QuestionPoint. Begun as a pilot project, QuestionPoint (formerly the Collaborative Digital Reference Service) has matured and moved to a new level of service. The partnership of libraries, Library of Congress and OCLC has allowed the service to build the next generation of collaborative reference tools, providing new functionality such as: a "local" component designed to manage and implement traditional and electronic reference services; chat software; tracking and reporting tools; technical support; standards-based advocacy; and seamless escalation to a global network of online librarians, to provide access to collections and subject experts.

The hallmark of QuestionPoint is migrating the traditional skills of librarianship (reference interviews, controlled vocabularies and source citations, networks and access to online and print resources) to the online environment; to meet patrons' needs in a global community and thereby redefine public service.

This presentation will include a live demonstration of QuestionPoint and first person accounts by librarians using the service in their libraries.

Daniel Greenstein, Director, DLF
Leigh Watson Healy, Outsell

This session will present preliminary findings resulting from a that was conducted to inform us about how university and college faculty and students view and use their libraries as part of their overall scholarly information environment. The study is based on an extensive survey of more than 3,200 faculty and students at research universities and liberal arts colleges.

The report presents selected data gathered through the survey. It is intended to indicate the breadth and depth of the data that have been collected and the potential for their analysis. By presenting selected tables and graphs and lightweight interpretation, the document is also intended to stimulate discussion about the data's significance and meaning, about new and important analyses that need to be tried, and about how the data might be reported in publication. That discussion may also identify gaps that the dataset as collected cannot fill, and in this respect to focus thinking about new or follow-up lines of inquiry.

Aaron Choate, Digital Collections Unit, Digital Library Services Division, The General Libraries, University of Texas at Austin

The General Libraries at the University of Texas at Austin has been working to develop a standards-based framework for sharing many of its digital resources with the world community. Until recently we have focused almost all of our efforts on web-based publishing and digital collection-building. Last year we began exploring ways to aid users in the discovery of information resources by exposing metadata to harvesting services. Topics will include: UT Austin's efforts in developing collaborative metadata repositories using XML; providing access to collection metadata via the Open Archives Initiative (OAI); current plans to develop a cooperative archive/repository for researchers on campus, at partner institutions and in the field.

Chuck Eckman, Principal Government Documents Librarian, Stanford University Libraries
Stuart Snydman, Digital Library Projects Manager, Stanford University Libraries

The Stanford University Libraries are collaborating with the World Trade Organization (WTO) in an effort to digitize the complete historical record of the WTO's predecessor organization, the General Agreement on Tariffs and Trade (GATT). Stanford archivists, technology staff and scholars have conducted selection, digitization and metadata collection for over 2 million pages of print archival records and related documentation at the WTO headquarters in Geneva. The large size and complexity of the archive, coupled with the geographically remote aspect of the operation, present significant challenges for both collection development and technological implementation. The collection development goal is to develop a representative digital collection based on original print archival records, documents and publications. In the course of identifying and incorporating this content within the project, the team has confronted several specific selection, access, and contractual issues. Technology challenges include the creation of a production-level digitization system in a remote location as well as the development of access tools able to accommodate the requirements of both Stanford's emerging digital library environment and the WTO's own document management system. The content and metadata of this collection will be the first to be incorporated within Stanford's planned digital repository, and will constitute a test of its ability to accommodate the idiosyncratic needs of a rare and unique archive, while at the same time interoperating with the digital library environment of non-academic collaborators. The presenters will provide an overview of the project's history, describe lessons learned, and conclude with a discussion of how librarianship, technological innovation and diplomacy converge in the development of the global digital library.

Clare McInerney, Programmer/Analyst, Office for Information Systems, Harvard University Library

This presentation will discuss the planned usage of the Tamino XML database by Harvard University Library in the design of the TED (Templated Database) system.

A number of catalogs have been developed at Harvard that enable creation, maintenance, delivery, and searching of different types of metadata about digital collections. Examples of existing catalogs are Visual Information Access (VIA), Harvard Geospatial Library (HGL), Online Archival Search Information System (OASIS) and HOLLIS. There are a number of digital collections that do not fit these existing catalogs, such as the Milman Parry Collection and the Biomedical Image Laboratory (BIL). The Milman Parry Collection is a collection of text and recordings of oral literature made during 1933-1935 in Yugoslavia, and the Biomedical Image Laboratory (BIL) will be a central catalog of images produced in support of biomedical research.

Existing catalogs have been manually coded using relational database models. It is very difficult to abstract metadata dependencies out of the system design using relational database models, especially when a collection contains a complex hierarchy. XML technology, and XML Schema in particular, provides a more flexible mechanism. The TED system abstracts metadata dependencies to a template layer. This means that the TED system takes a formal description of the metadata as input and automatically creates a customized catalog, including a database schema and query interface. We are using the Tamino native XML database from Software AG to implement TED. While we are still in the early stages of the TED project, I would like to share our experiences so far with Tamino.

MacKenzie Smith, Associate Director for Technology, MIT Libraries

As digital libraries begin to build significant holdings of research publications and other material that is useful in instructional settings, we are beginning to see what new challenges supporting instructional technology will bring, and how traditional boundaries between classrooms and libraries are beginning to blur. The scope of what we collect, the metadata we create to manage digital collections, and the means of delivering digital content are all affected by the requirements of instructional technology. MIT is undertaking significant programs in both instructional technology and digital library development, and is investigating how these two technologies will need to interact for digital libraries to reach their full potential in academic research institutions. The OpenCourseWare and Open Knowledge Initiative projects in instructional technology will be described, as well as the DSpace institutional repository system in development by the MIT Libraries, and how these projects are collaborating to extend the reach of digital libraries into the classroom.

Bradley C. Wheeler, Associate Dean of Teaching, Learning Technologies, Associate Professor of Information Systems, Kelley School of Business, Indiana University

http://portfolio.iu.edu/bwheeler/presentations/DLF_Chicago2002_Wheeler_Winning_The_Right_Game.ppt

Speaker to be advised

Abstract pending

John Kunze, California Digital Library

If there's one thing that distinguishes a digital library from a web site, it's probably reliable, long-term access to information. An important challenge for DLF member institutions is to address the widespread problem of broken URLs - that is, web links that once worked to gain access to vital material, but that now no longer work.

Until recently, the blame and the proposed remedies were typically attached to the URL itself. Why couldn't this volatile identifier be tamed by switching to indirect identifiers, just as DNS introduced symbolic hostnames to replace those unstable numeric IP addresses? Unfortunately, the main proposed remedies -- URNs, DOIs, PURLs -- are predicated on the surprising assumption that provider organizations that currently do not, for whatever reason, maintain URL redirection tables, would suddenly, by using different identifiers, become the kinds of organizations that do maintain such tables.

The California Digital Library is taking a new approach that treats persistence purely as a matter of service. If a persistence claim is to be seen as credible, a provider's identifiers will be backed up by services that take efficient, consistent access to objects as seriously as access to object descriptions and to statements detailing the nature of the provider's commitments to them. Economies of scale are possible to the extent that access methods are standardized. If persistence is a priority, identifiers can be generated in such a way as to avoid several common age-related disorders.

Peter Brantley, New York University

Shibboleth has been proposed by Internet2 as a technology to support inter-institutional authentication and authorization for access to Web pages. Its intent is to support, as much as possible, the heterogeneous security systems in use on campuses today, including Kerberos and PKI.

Shibboleth is developing an architectural framework and an associated open-source software prototype to support inter-institutional resource sharing subject to access restrictions. Shibboleth enables the secure exchange of interoperable authorization information, working hand-in-hand with existing campus-level authentication systems. Shibboleth combines a high level of granularity over the release of user attributes with robust logging support, meeting both academic expectations for privacy and publisher needs for accountability.

Shibboleth has reached a stable architectural stage and coding is commencing. Recently, New York University, CNI, and the DLF invited a cross section of academics, publishers and resource providers to meet with Shibboleth's developers and comment on its design and goals. This talk will report on the dominant themes of this sponsored meeting, and present expected future directions in Shibboleth development and deployment.

Bradley Westbrook, Manuscripts Librarian and University Archivist, Mandeville Special Collections Library, University of California, San Diego

This presentation will offer a summary of the Archivists' Workbench meeting, but it will focus on several high-level design criteria for an archivists' toolkit derived from that meeting. The Archivists' Workbench meeting, sponsored by the Digital Library Federation and the California Digital Library, took place in La Jolla, Calif., on Februrary 3-4. It was attended by twenty-one archivists and information technologists, all of whom were interested in the feasibility of designing a digital toolkit supporting archival work and capable of general deployment for use among a wide range of archival repositories. During the meeting, many key objectives for a digital toolkit were agreed to, namely, that a workbench should create efficiences in archival work, reduce barriers to repository participation in consortial or union databases, lower needs and cost for technological training across a repository's archival staff, and promote data content and structure standarization. The meeting discussion revealed several high level design considerations necessary for realizing the objectives. A toolkit would be strongly related to the life cyle of archival materials. It would capture and manage information beyond the description of a collection. It would need to be flexible and capable of being used equally well in different workflows. It would have to be modular, with a clean separation between input, storage, and output components. Finally, a digital toolkit would require a well thought out service plan that would make facilitate implementation of the toolkit and its modification to meet changes in technology and archival work.

Daniel P. Johnson, Digital Library Technical Analyst, Indiana University Digital Library Program

Open systems could be considered the nirvana of technological interoperablity. This type of framework allows for cross-platform semantic homogeny, fostering interface development and storage that is functional across the system. Technology exists today that will allow for a digital library infrastructure that is truly open in this sense. The entire collection can be stored in object-oriented database systems which will permit the development of interfaces written in numerous languages to access the data. The need for file translators and conversion is no longer necessary once such a setting is achieved. Additionally, behaviours can be added to the objects allowing for further operability. Once such a system is in place, specific to the collection, new and improved methods of combining the data and displaying it might be realised. Current techonologies which promote openness as well as future trends will be discussed.

John Helly, San Diego Super Computer Center
Susan Starr, University of California, San Diego Libraries
Dawn Talbot, University of California, San Diego Libraries

The UCSD Libraries, the Geological Data Center at Scripps Institution of Oceanography, and the San Diego Supercomputer Center are collaborating on an NSF-funded NSDL project to 'bridge-the-gap' between collections of oceanographic data collected by ship-based expeditions and physical archives of older, pre-digital, expeditions held within library archives in a manner consistent with modern and emerging library standards and practices. Our project aims to create a site that can deliver research data to the scientist while at the same time communicating the excitement of oceanographic discovery to the non-scientific public. Data sources range from bathymetry and geochemical data collected on scientific expeditions to diaries of expedition participants and photographs of scientific instruments. With tools developed by the San Diego Supercomputer Center, we are using a geospatial approach to unite these diverse data streams and the disparate vocabularies and metadata used by the participants. Our project is forcing us to confront a wide range of interoperability issues, ranging from the technical to the cultural. This paper will present our approach and expose the decisions we have made in attempting to build a highly interoperable and scalable method of publishing scientific data and supporting archival materials.

Tim Jewell, University of Washington

During the last year or two, developers of local systems for managing electronic resources have shared information about their functions and data elements with one another and other interested parties. An interest in developing standards for such work has grown out of these discussions, which led to the organization of a NISO/DLF sponsored half-day workshop on this topic planned for May 10th as an adjunct to the DLF Spring Forum (see http://www.niso.org/news/events_workshops/NISO-DLF-wkshp.html). This presentation will summarize the workshop presentations and discussions for those unable to attend, and provide an opportunity for further discussion of "next steps."

Kat Hagedorn, OAIster Librarian, Digital Library Production Service, University of Michigan
Martin Halbert, Director for Library Systems, Emory University
Beth Sandore, Associate University Librarian for Information Technology, Planning and Policy, University of Illinois at Urbana-Champaign

The Libraries of the University of Michigan, the University of Illinois at Urbana-Champaign, and Emory University have received support from the Andrew W. Mellon Foundation to implement a suite of OAI-based metadata harvesting services, search services, and tools designed to facilitate discovery and retrieval of scholarly information. Two distinct metadata search services are being developed utilizing shared infrastructure components and software tools that will be made available under an Open Source Initiative license.

The University of Michigan search service, called OAIster (http://oaister.umdl.umich.edu/), is under development by the Digital Library Production Service (http://www.umdl.umich.edu/). It is intended to be global in scope, providing cross-repository searching of metadata describing publicly available digital objects. The Michigan service is relatively "lightweight" (e.g., without duplication or thesauri) and is being designed to answer the pressing need for opening the "hidden web" information resources of the scholarly community.

The University of Illinois has developed a vertical, domain-specific portal designed to search metadata describing manuscript archives and digital cultural heritage information resources (http://oai.grainger.uiuc.edu/). Metadata describing non-digital resources and resources of restricted availability is included along with metadata describing publicly available digital objects.

At Emory University, two grant projects - AmericaSouth.org (http://www.AmericaSouth.org/) and MetaArchive.org (http://www.MetaArchive.org/) - have been collaboratively conjoined and are being carried forward in cooperation with partner institu