THE PROGRAMS

Resources for Scholarship

Scholarship is dynamic; research trends are emerging that could not have been predicted even a decade ago. Digital technology is transforming scholarship, and new methods of inquiry, scholarly communication, and teaching are accelerating the rate of change in research trends. The increase in use of electronic sources such as e-journals and digitized collections is, however, accompanied by a continued reliance on original sources for historical research. Seeing these trends and anticipating the pressures that they will bring to libraries and archives, CLIR has brought together librarians and scholars in several programs to assess current and future needs both for electronic access to information and for access to original collections in all genres and formats.

The Task Force on the Artifact in Library Collections

In 2001, a task force of 15 scholars, librarians, and archivists, assembled by CLIR in 1999 to investigate the demands for original source materials, produced its draft report, The Evidence in Hand: Report of the Task Force on the Artifact in Library Collections. The task force engaged the research community in six public reviews of its findings and recommendations and posted the draft report on CLIR’s Web site for public comment. The chief findings of the report—that scholarship will continue to interrogate original sources for many reasons, and that the broadest-possible access to unreformatted sources therefore best serves scholarship—were by no means controversial. But the breadth of sources identified—from traditional print collections to recorded sound, broadcast media, moving and still images, and all manner of digital information—greatly expanded the universe of materials that libraries, as well as archives, historical museums, and museums, should be collecting and preserving.

Acknowledging that funds for preservation are limited, the task force proposed a number of approaches. Many rely on cooperation among institutions to reduce unnecessary duplication of preservation responsibilities and engender cost efficiencies that will ensure the appropriate types of redundancies through distributed repositories of artifactual collections and, when possible, broadened access to digital and other surrogates.

The public reviews confirmed that the task force was addressing issues deemed critical by the research community. The section on digital resources and the novel, often perplexing, responsibilities that face librarians and scholars when creating and using digital sources drew special interest. The task force’s recommendations to ensure the creation of preservable digital objects while preserving the sources for digitized materials have already influenced developments in the Digital Library Federation libraries and their many partners.

Creating a Test Database for Digital Visual Resources

Computer searching is at its best when it works with standard, replicable characters, such as numbers and letters of the alphabet. But what are the defining components of an image? The quest to create reliable means for searching databases of images has given rise to a range of solutions that focus either on the development of appropriate descriptive metadata or on the creation of effective visual searching software (e.g., content-based image retrieval). Nonetheless, creating a means by which generic images can be searched broadly remains an elusive goal.

CLIR and the Coalition for Networked Information (CNI) are investigating the value and feasibility of developing a test database for digital visual resources. Such a database could serve as a means of measuring the capabilities of various technical applications for creating, managing, and exploiting digital image content.

Project Director Clifford Lynch is preparing scenarios depicting two approaches to developing the database. In the first approach, the database would be designed to support fundamental long-term research into image-retrieval techniques. This would be most useful to computer science researchers interested in content-based retrieval.

In the second approach, the database would be designed to support the assessment of metadata for retrieving and using image files. Special emphasis would be placed on understanding the costs and benefits of investing in metadata creation. This approach would require research into how users are querying image databases. Two test collections would be used in this approach: one in art and another in history or historic photographs. For both types of collections, metadata would be created that could support multidisciplinary use.

Jennifer Trant will serve as a consultant on this project. Working from these two scenarios, she will develop a report on the resource implications and processes necessary to assemble content and metadata. Her report and the scenarios will be circulated for comment. CLIR will issue a final report on these proposals in early 2002.