1 Tree visualisation and navigation clues for information visualisation Ivan Herman † , Maylis Delest ‡ , Guy Melançon † † Centrum voor Wiskunde en Informatica (CWI) Kruislaan 413, 1098 SJ Amsterdam, The Netherlands email: {Ivan.Herman,Guy.Melancon}@cwi.nl and ‡ Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université Bordeaux I 351, cours de la Libération, 33405 Talence Cedex, France email: Maylis.Delest@labri.u-bordeaux.fr} Abstract Information visualisation very often requires good navigation aids on large trees, which represent the under- lying abstract information. Using trees for information visualisation requires novel user interface techniques, visual clues, and navigational aids. This paper describes a visual clue for trees as well as an automatic folding (clustering) technique, both based on some mathematical concepts and results in combinatorics. Examples are shown on how these techniques can be used, and what the further challenges in this area are. Keywords: information visualisation, tree visualisation, graph visualisation, user interfaces 1. Introduction The problem of displaying and interacting with abstract in- formation, which is a central task of information visualisa- tion, can very often be abstracted to the problem of displaying and interacting with graphs. This is the case when attempting to visualise file spaces, hypermedia docu- ment structures, internal data structure of computer pro- grams, etc. Graph drawing is extremely complex, and the available results draw on such diverse fields as topological graph the- ory, computational geometry, combinatorics, graphical user interfaces, etc. A comprehensive bibliography on graph drawing algorithms 1 cites more than 300 papers published before 1993; specialised conferences (the so–called Graph Drawing meetings) take place every year, and special issues of various journals (e.g., a special issue of the journal Algo- rithmica, in 1996) are published on the subject. The reason for this high level of activity is the fact that the simply phrased problem of drawing a graph in a plane (or in space) turns out to be, in general, extremely demanding. Some of the questions can lead to NP hard problems; a solution may involve significant computing resources, and may be totally unsuitable for interactive purposes. Roughly speaking, there are two major, albeit inter- twined, problem areas: • How to draw a graph on a plane (or in space)? This line of research tries to find answers to questions like: is it possible to draw the graph on a plane without edge intersections? How to draw the graph by minimising, for example, edge lengths, or by making the edge length proportional to some application–dependent measure? How to draw a graph by respecting some aes- thetic constraints? • How to interact with (the visual representation of) a graph? This raises questions like: how to access infor- mation stored in a node? How to “move around” in a graph in search for some specific pattern or node? How to change the visual representation as an answer to user interaction?