My problem can be best explained starting from a 2D example:
Imagine having a circle and wanting to discretize N points on the circumference of the circle so that the difference of the cosine of each pair of adjacent points is constant.
This is relatively easy, as it would be sufficient to discretize the diameter of the unit circle, and then apply $acos()$ to each cosine sample.
What I have been unsuccessfully looking for is a solution for the transposition of this problem onto a small circle on a sphere. To visualize it, imagine we are talking about the Another Small Circle setup in the image below (the right one). And let's assume, unlike the illustration, that the small circle that has been cut away does not contain the pole (although it might make no difference).
(source: slideplayer.com)
What i'm looking for is a discretization of the circumference of this small circle so that the delta (or difference) of the projection of each pair of adjacent samples onto the parallel going through the center of the small circle is constant. The projection of each sample onto the parallel (of latitude) is done by a meridian (that is a great circle going through the poles and the sample on the parallel)