DSpace is a joint project by MIT Libraries and the Hewlett Packard Company to build a stable and sustainable long-term digital system that captures, preserves and communicates the intellectual output of MIT's faculty and researchers. It provides web access to articles, technical reports, working papers, conference papers, images, datasets and rich media works produced at MIT.

Accommodating the diverse needs of different disciplines and groups within the Institute is one of Dspace's most challenging tasks. The DSpace system addresses the issue of multi-disciplinarity by subdividing DSpace into "communities", each of which serve a particular discipline or type of information. Each community will be self-managed, creating its own contributor, access, and rights policies, following its own workflow patterns, and implementing metadata standards appropriate to its needs. These diverse requirements have led to interesting design solutions, especially in the area of metadata architecture.

The Dspace long-term vision of a federated group of research institutions sharing each other's intellectual work through use of a common system has presented the team with the challenge of selecting low cost or open source components, as well as selecting solutions that can be easily installed, implemented and maintained.

Planning for maximum longevity has presented DSpace with several challenges. It must be flexible enough to accommodate replacement of software components in the future while maintaining the integrity of the data and metadata store. Dealing with data format and data representation issues with a view to long-term preservation presents its own set of challenges as libraries plunge into the digital future.

This presentation will give an overview of the DSpace customer and user needs, and the system design, architecture, and current beta release of DSpace.