hammax wrote: ↑
Wed Nov 14, 2018 6:44 am
… I see new features updated for GordonSurface in the CurvesWB.
I am trying to fix Gordon surface.
Currently, the algorithm is only partially working, and produces surfaces that have far too much control points.
The smartest and most important part of the TiGL algorithm is about preparing and reparametrizing the curves before feeding them to the actual Gordon algo.
I realized that these optimization functions were actually skipped in my implementation (and broken BTW).
The good news is that I found were the bug is happening.
The bad news is that it is located in the most complex function, dealing with "matrix solving", that is way over my poor math skills.
So I don't know how I will be able to fix that ...