NB- see this question for a more formal request.

For many many years (honestly, since 1987) I've had my own MMa computational geometry code for dealing with 3D meshes. I principally (there's a joke there somewhere) use it to calculate coordinate frames / curvatures at the discrete vertex positions, etc.

The introduction of MeshRegion[] in 10.x seems to make my some of my code redundant (thankfully!) but I'm not 100% sure how to proceed.

I can imagine a few things, like computing and saving useful information about the vertices (coordinate frame, curvatures, etc) as Property[] for each vertex, but computing that stuff (seems like) I'd just do it the old 'discrete' way. I'd also suppose that something along the lines of using NDSolve over the {x,y,z} in the MeshRegion[] would also be useful here.

Has anyone already done/started to do such a thing?

  • 1
    $\begingroup$ More detail would be useful. Perhaps, you could provide the code for some of your functions that you feel may now be redundant. $\endgroup$ – bbgodfrey Mar 24 '15 at 18:15