Previously I wrote:

> Well, open surfaces should work without a problem, but there is not an
> automatic "outside" direction that can be calculated for an open surface,
> unlike a closed surface.

However, I should be able to do something automatic for surfaces of revolution, even if they are open.

Since they have a central axis I should be able to point the normals to the outside of that.

So that's one thing that should be not too hard for me to fix up.

- Michael