> ... you have a two-dimensional surface that resembles a sphere. To estimate its curvature, first draw three points on the surface and connect them with the shortest (and therefore straightest) possible paths. This gives you a triangle. Then draw another triangle with the same side lengths on a flat plane.
I wonder if this is even possible.