A mesh is a discretization of a geometric domain. An unstructured mesh is typically a triangulation. Unstructured meshes are catching up especially in the academic community. I'm looking for open questions in 2D mesh generation, specifically outstanding theoretical problems for refinements of triangles: for example, show that angles stay bounded away from zero in sucessive iterations of division methods.
http://www.freeimagehosting.net/>http://www.freeimagehosting.net/uploads/d34a8238de.jpg border=0 alt="5th iteration equilateral triangle trisection">