On Sat, Mar 14, 2009 at 04:26:35PM -0200, Fernando Cacciola wrote:
IMO, any geometric library should provide a toolbox of exact predicates (which shoud be exact independetly of the input type)
Steve M. Robbins wrote:
I agree.
I've always thought that the adaptive floating-point arithmetic of Douglas Priest [1] and Jonathan Shewchuk [2] would be a good way to do this for floats and doubles. Any interest in developing a boost version of this code?
I'm aware of Shewchuk's work. He is a friend of one of my team mates. His 2D Delauney library (Triangles) is considered definitive. However, his software license is not free. I think providing in boost a library such as the adaptive floating point arithmetic Shewchuk's Triangles is based on is a great idea, but it can't be a derivitive work of Shewchuk's code. To be sucessful such a library should be developed along side an application for it such as his Delauney triangulation that will drive its development. It would be hard to develop in isolation. Luke