Jérôme M. Berger wrote:
> 	Actually, it is possible in spherical geometry. You just have to
> remember that two points are not sufficient to define an edge on a
> sphere: there are two possibilities, one of which goes around the
> sphere...
> 

Ah yes, I was thinking of the triangles specified by joining the points 
according to the great-circle distance, i.e. the shortest arc between 
the vertices.

Problem is that if you allow to build triangles using the other arcs 
(the ones that go all the way around), you end up also allowing 
triangles that have intersecting sides, which is interesting, I guess, 
but then calling them triangles is perhaps slightly abusive :-)

-- 
Vincent

Post a reply to this message