NCSTRL: Design and deployment of a globally distributed digital library

The World Wide Web provides unprecedented access to globally distributed content. The extent and uniform accessibility of the Web has proven beneficial for research, education, commerce, entertainment, and numerous other uses. Ironically, the fact that the Web is an information space without boundaries has also proven its biggest flaw. Key aspects of libraries, such as selectivity of content, customization of tools and services relative to collection and patron characteristics, and management of content and services are noticeably absent. Over the past four years, we researched the technology and deployment of a digital library architecture that makes it possible to create managed information spaces, digital libraries, within the World Wide Web. Our work has taken place in the context of NCSTRL,1a digital library of computer science research reports. The technical foundation of NCSTRL is Dienst, a protocol and architecture for distributed digital libraries that we developed as part of the DARPA‐funded Computer Science Technical Reports Project. At the time of the writing of this paper, the NCSTRL collection consisted of papers from more than 100 research institutions residing in servers distributed across the United States, Europe, and Asia. In addition, the Dienst protocol and implementation has been successfully adopted by a number of other distributed collections. In this paper, we review our experiences with NCSTRL and Dienst, describe some of the lessons we have learned from the deployment experience, and define some directions for the future.