I've done some reading on nav-meshes, and I understand how to generate a path of polygons to reach a goal. However, what I don't understand is how you determine the bearing to follow within each polygon. Without a central node to aim for, what do you aim for? I suppose you could cast a ray to the goal and then head to the point where that ray crosses into the next cell - but that would only work if that next cell is actually on your path. If your ray doesn't cross the edge into the next cell, do you instead plot a path to whichever corner of the edge is closest to the goal? I think that would get you the path shown in the 3rd diagram, but would it work in all cases?
