Visualization of the Relevance: Using Physics Simulations for Encoding Context

The task of organizing and retrieving knowledge is often elaborative and involves different types of media including digital or analog. In this paper we describe a system that is based on related research in the fields of spatial hypertext, information retrieval, and visualization. It utilizes a 2D space on which users can add, remove, or manipulate information entities (so-called user nodes) visually. A spatial parser recognizes the evolving structure and queries a knowledge base for helpful other information entities (so-called suggestions nodes). Similar to user nodes, those suggestions are presented as visual objects in the space. We propose a physics model to simulate their behavior. Their characteristics encode the relevance of suggestions to user nodes and to each other. This enables human recipients to interpret the given visual clues and, thus, identify information of interest. The way users organize nodes spatially influences the parsed spatial structures, i.e., the placement of suggestion nodes. This allows the creation of complex queries without any prior knowledge, yet the users do not have to be aware of that, because they can express their thoughts implicitly by manipulating their nodes. We discuss the strengths of a physics based simulation to encode context visually and point to open issues and potential solutions. On the basis of an implemented demonstrator we show the benefits compared to similar and related applications in the field of information visualization, especially when it comes to tasks where a high portion of creativity is involved and the information space is not well known.