03:22 <GitHub> [solvespace] Evil-Spirit commented on issue #225: @whitequark , @jwesthues ... https://github.com/solvespace/solvespace/issues/225#issuecomment-297897186

06:15 <GitHub> [solvespace] jwesthues commented on issue #225: @Evil-Spirit Cool. What's the condition number of the matrix? It should intuitively be not too big, in which case Cholesky should work robustly, though I don't think I ever actually measured... https://github.com/solvespace/solvespace/issues/225#issuecomment-297916392

06:15 <GitHub> [solvespace] jwesthues commented on issue #225: @Evil-Spirit Cool. What's the condition number of the matrix? It should intuitively be not too big, in which case Cholesky should work robustly, though I don't think I ever actually measured... https://github.com/solvespace/solvespace/issues/225#issuecomment-297916392

07:48 <GitHub> [solvespace] whitequark commented on issue #225: @Evil-Spirit Excellent work! https://github.com/solvespace/solvespace/issues/225#issuecomment-297932118

08:38 <GitHub> [solvespace] Evil-Spirit commented on issue #225: @jwesthues , I will try to calculate the condition number of the matrix and do output for it.... https://github.com/solvespace/solvespace/issues/225#issuecomment-297941890

08:52 <GitHub> [solvespace] jwesthues commented on issue #225: SVD rank testing is better but unlikely to be faster. I'm not sure about Eigen, but sparse SVD solvers are well-studied, http://www.netlib.org/svdpack/ or whatever.... https://github.com/solvespace/solvespace/issues/225#issuecomment-297944845

11:08 <GitHub> [solvespace] urmilparikh commented on issue #206: I confirm that `10mmcube-modded3.step` shared by @Jonimoose does show up as solid in 123D Design as well. https://github.com/solvespace/solvespace/issues/206#issuecomment-297971472