Overview
of Implementation
This algorithm was implemented in Java 1.4.2 and its performance was
tested with a variety of input points. Some of the general classes of
input points include:
1) A random selection of about three thousand world cities obtained
2) Points arranged on a 10x10 grid and a 20x20 grid
3) Points near the North and South Poles or near the Equator.
4) Points on only the northern, southern, eastern, western hemisphere.
The maximum distortion was generally below ten for a random selection
of points which compares favorably with the distortion from other map
projections. The maximum distortion tended to be higher when the
data-sets included a number of points near both of the poles. Extremely
high distortion at the poles is a major issue for many map projections,
for example, in the commonly used Mercator Projection, two points near
the North Pole will appear to be very far apart on a planar map
resulting in very high distortion since the poles are projected as an
entire line.