A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
From: Ed Popko
Date: 2018 Feb 19, 03:44 -0800
Randomizing points on a sphere is harder than it looks if you want all areas, including the poles, treated the same.
An effective way to randomly distribute them is to first randomize xyz points (three axis) and then normalize their vector lengths to some desired and consistent length. The original set of points are now distributed on a sphere with a radius equal to the normalization value you used. Make sense?