Digital Media Services

Open Archives: exposing new digital collections to researchers around the world

The Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH) is beginning to show promise as a tool for libraries to share information about their digitized collections. The Committee for Institutional Cooperation (CIC) has two demonstration projects underway to gather metadata about these collections and build search engines around them.

Most research libraries and many prominent museums have been engaged in some kind of collections digitization during the past decade. At Northwestern University, such projects have included the Siege of Paris online, the World War II Posters project, and the Video Encyclopedia of the Twentieth Century, among others (see http://digital.library.northwestern.edu for a complete list). Libraries and museums digitize material to make their rare or unique collections available to a wider audience of researchers and to expand the methods of investigation available to these scholars through full-text search, high-resolution image view, and other tools. Availability of a wealth of fascinating primary source materials has exploded since the advent of the Web and the technologies that facilitate digital image, text, audio and video capture.

Unlike most traditional, or "analog", library materials, however, it is more likely that descriptions of these digital collections may not be included in a library's online catalog. This may be because the associated descriptions, or metadata, may be in a non-MARC format. MARC is short for Machine-Readable Cataloging, and is the standard most libraries use to create online cataloging records. It is also true that some libraries prefer not to include metadata records for individual images in their catalogs along with records for full monograph and serials titles. In some cases, descriptions may be included in the online catalog, but researchers may find it difficult to focus on and locate materials available directly in digital format using traditional search methods.

It is somewhat frustrating, both for library practictioners and to scholars, that despite the significant effort invested in digitizing and describing these collections, intellectual access has not been significantly enhanced. OAI has emerged as a possible solution to this access problem. The OAI protocol defines roles for data providers and data harvesters. Data providers are required to respond to very simple requests from harvesters to provide copies of metadata records. The data harvester collects these metadata copies from any number of different repositories and then builds services, such as search engines, on top of them.

The University of Illinois at Urbana-Champaign and the University of Michigan are leading a Mellon Foundation-funded project to test OAI-PMH for digital library collections. The mid-term results of their work may be viewed online.
OAISter at Michigan: http://oaister.umdl.umich.edu
OAI at U of I: http://oai.grainger.uiuc.edu/oai/search/

Figure 1: Search of the U of I OAI system for "Feather Headdress"

Figure 2: Image from the Colorado Digitization Project

What are the benefits for faculty at Northwestern? Many objects in these digital library collections are available online because there are no copyright restrictions to prohibit worldwide access. Therefore, they may be a rich source of materials to support university classes, and can be freely linked to from NU course web sites.

Claire Stewart
Head, Digital Media Services