List of colors and subcolors ad infinitum
Given a number n (integer between 3 and 10), this generates an object containing a list of n "equally spaced" colors. Then for each color, it generates a sublist of n colors that remain in that color palette. This can be repeated recursively, although the colors will converge fairly fast.
With n=3, the first layout would generate a shade of Green, Red and Blue. At the second layer, the Green will be associated with 3 other greens (that are distinct from the parent). The red will be associated with 3 new shades of red etc.
Assume you represent data within a tree and need each node on the tree to be matched to a color. It would be nice that when going down the tree, there is a color convergence.
Using the HSL model for colors on a screen, this method makes sure the entire hue Spectrum is used, as well as almost all the Lightness spectrum. The Saturation is left constant to keep a unified feelings on all colors. Choosing equidistant points within a closed 2d square is hard (NP hard), and is called the Packing Problem. To avoid this, precomputed values of equidistant points in a square are used. These values are pulled from: http://www.ime.usp.br/~egbirgin/packing/packing_by_nlp/numerical.php?table=csq-mina