mindmap cluster
The problem:

Using images on older computers or in special situations often demand reducing the number of colours in the image drastically. Sometimes there are additional constraints on the colours used. The currently available tools for this often lack vital features, or they are designed for being fast on computers designed in the 90ies, not 2013.

The solution:

An advanced tool that will excel at colour reduction, taking as much CPU time as needed to search for an optimal palette while having hooks for imposing limitations on the destination palette.

The general idea is doing the reduction in a custom colourspace, and using an iteration-based solver where each entry in the palette will search for the best solution while competing with the other palette entries. Each colour will get a score depending on a set of rules, and the goal for each colour is to increase its score.