Hyperdatabases - Infrastructure for the Information Space

Abstract

The amount of stored information is exploding while, at the same time, tools for accessing relevant information are rather under-developed. Usually, all users have a pre-defined view on a global information space and have to access data by the same primitive means. However, a more convenient solution from a user's point of view considers her/his individual context and interests by mapping the global information space to a personal one. Yet, the organization and personalization of information spaces induces a set of tightly related problems: First, user interfaces have to present personalized information in a user-friendly way and have to be enriched by sophisticated, context-sensitive navigation techniques. Second, the personal information space has to be organized automatically, by exploiting similarities between multimedia documents. Third, in order to allow the user to influence the automatic organization of her/his information space, relevance feedback techniques for multimedia similarity search have to be provided. Finally, taken into account that information is replicated at several sources and is subject to modification, sophisticated coordination mechanisms have to guarantee consistent views on the global information space. In this paper, we introduce the vision of hyperdatabases as the core infrastructure to support these requirements in a holistic way. Moreover, we present the ETHWorld project and its sub-projects in which we apply hyperdatabase concepts for managing and organizing the information space of a virtual campus.