Closed twadleigh closed 10 years ago
I am fine with that scope.
Usually my meshes have holes. When simplifying, this sometimes causes trouble.
Bindings to any external library probably should live in a package of its own. That seems to be the Julian way.
Agreed. The Slicer I work with is a big application for medical imaging (including 3d models), but I am not making bindings for that.
However, this (very WIP) project might be of interest with respect to visualization of meshes and other things: https://github.com/ihnorton/VTK.jl There is quite a ways to go still, but the proof-of-concept uses all of the required calling conventions, so I'm starting to be convinced that it may be viable.
What should the scope for this library be?
I would personally be OK with saying that it is for 2D meshes embedded in 3D space. That's all it currently extends to, but I wanted to ensure that others are OK with that.
Another question: do we further restrict it to (orientable) manifolds? This is a property of all surfaces of 3D objects, and is satisfied by any result of the isosurface extraction algorithm. Being able to make this assumption simplifies some algorithms, but it would also be hard to check/enforce with certain mesh representations...