Interrelated multimedia content and metadata are available in a large number of diverse, distributed, and incompatible data sources. The goal of the MediaSpaces project is to provide tools, software components, and strategies to enable integrated access to these sources.

A "mediaspace" is a metaphor for a Web-accessible multimedia repository storing contents of any kind. Often there are relationships between content items: an MP3 is the soundtrack of a certain movie, a text document contains the lyrics for a song available as WAV file, or an image depicts an actor of a certain movie. A mediaspace can represent such relationships within and across system boundaries.

MediaSpaces is also the name of a tool suite. First, it provides components that support users and organizations in creating Web-accessible mediaspaces from their legacy multimedia repositories. Second, it will provide the integration services for connecting interrelated mediaspaces and reconciling the incompatibilities among them.

The MediaSpaces architecture follows that of a federated system on a Web basis. All data are virtually integrated, meaning that they reside in their original location. Each system component is built on open standards, mainly on Web protocols such as HTTP and Semantic Web technologies such as RDF/S, OWL, and SPARQL.

© 2007-2009 Mediaspaces.info