much more variability in the number of voxels
improvements in terms of speed (sometimes).
There are two big things to do in voxels that will be developed next:
1) keep only the Long encoding. Even though the long ID is used as the hash lookup, all voxels are stored as struct or class objects. This takes up significant space. I believe the space savings in converting to and from Long will be worth it. There may be a slight time increase however.
2) The tessellation to voxelization could be made faster by reworking the face/edge voxel overlap creation. I know this was a pain initially, and so that will require it's own testing.
much more variability in the number of voxels improvements in terms of speed (sometimes). There are two big things to do in voxels that will be developed next: 1) keep only the Long encoding. Even though the long ID is used as the hash lookup, all voxels are stored as struct or class objects. This takes up significant space. I believe the space savings in converting to and from Long will be worth it. There may be a slight time increase however. 2) The tessellation to voxelization could be made faster by reworking the face/edge voxel overlap creation. I know this was a pain initially, and so that will require it's own testing.