A Hyperbolic Image Viewer

Tom White, April 1999.

Abstract

This page is devoted to a discussion of the benefits hyperbolic displays can bring to visualisation of large datasets. I have provided some Java Applets to demonstrate some of the ideas.

Java Applets

Escher (applet), (source image), (hyperbolic image)
Stoke Newington (applet), (source image), (hyperbolic image)
Windows Desktop (applet), (source image), (hyperbolic image)

2019 Update: Applets are no longer supported in browsers, so to run, install a JDK and type:
appletviewer escher.html, or
appletviewer stokie.html, or
appletviewer desktop.html.

Hyperbolic Geometry in Computer Displays

Although steadily becoming cheaper screen 'real estate' is still at a premium. Computer applications developers strain to squeeze as much information as possible from each patch of pixels while monitor manufacturers grow their models inexorably - but slowly. Until we have wall-sized display units that allow us to shuffle views as flexibly as we hang pictures now we shall continue to use nifty space-saving measures such as scrolling panels and switchable windows.

It may then come as a surprise that - in the world of pure mathematics at least - an infinite amount of space can be contained in a finite area. The trick of course is that the visible size of objects in such a space, called a hyperbolic geometry, decreases to nothing as you get closer to the 'edge'. To illustrate this we can use the Poincare/ Model which is one such hyperbolic geometry. This model represents a point in hyperbolic space by a point on the unit disk - that is, the numbers in the complex plane with magnitude less than 1. Distances in this disk are not measured using the familiar Euclidean metric, but rather one which attributes the distance to the edge of the disk as infinity. This means that the whole Euclidean plane is 'contained' in the unit disk. The only problem is the edge is a little squashed.

Nice things about this new view:

But:

Promising applications include:

References

A Focus+Context Technique Based on Hyperbolic Geometry for Visualizing Large Hierarchies by John Lamping, Ramana Rao, and Peter Pirolli.
Visualizing the Structure of the World Wide Web in 3D Hyperbolic Space by Tamara Munzner and Paul Burchard.