brandonpelfrey
commented
11 years ago Hi Mik,
Sorry it took a million years to get back to you. I think this is a really interesting idea, but I was kind of put off by the upper bound in performance in javascript. I might have time later to look into this, but at the moment I'm busy with work and a project that I'm doing with some friends. I think you have plenty of understanding on how this could be coded. If you have questions though or want some discussion on it, I'm usually free to chat, either through email or chat at brandonpelfrey@gmail.com .
Best, Brandon
Hi Brandon!
I've been trying to build up a decent library of tools for working on meshes in JavaScript, and I was wondering if you'd be willing to make an npm package for your Laplacian operator code?
I've been using the npm style of making many small modules, and I find it works pretty well. For example, here is module to compute topological indices for meshes (or more generally simplicial complexes):
https://github.com/mikolalysenko/simplicial-complex
And here is a simpler module that fuses vertices together in a mesh:
https://github.com/mikolalysenko/fuse-vertices
I am thinking that a highly modular approach could be really successful in geometry processing, where one often has a bunch of loosely couple techniques that are frequently combined together in various ways. Once you have these modules, it is easy to distribute them using npm and use them in the browser with browserify.
Having some code to compute the Laplacian of a mesh (as say a numericjs csr matrix) could be quite useful.
-Mik