View Coordination Architecture for Information Visualisation.

A view is a particular visual representation of a data set. Complex data sets typically require multiple views, each revealing a different aspect of the data. Coordinating the behaviour of these views is known to expedite tasks such as information seeking, and has been used to facilitate the exploration of large and complex data sets. The design and implementation of multiple view coordinations is challenging; poorly designed coordinations may in fact detract from task performance, while the failure to make coordinations apparent could lead to unneccessarily complex mental models of a data set. In order to facilitate the exploration of this design space, we present here an architecture for the implementation of view coordinations. In contrast with many existing tools and prototypes, which focus on specific view coordination techniques, this architecture provides generic support for view coordination. The implementation of several standard coordination techniques within the architecture are discussed by way of illustration, and the architecture compared with the Snap view coordination architecture. Examples of its initial implementation within the InVision component-based framework for information visualisation are also presented.