Closed gaperez64 closed 8 months ago
One thought about item 2.
Perhaps the easiest is to keep a vector with the nonremoved elements and use it as a cache that is dropper after every trim operation that does remove at least one node?
The kd-tree structure is now fully updated. Moving on to implementing the logarithmic method next.
One thought about the logarithmic method: for unions we avoid rebuilding trees, but unbalance the ones already there. For intersections though... we still need to rebuild, so this is less promising. I will instead implement the dynamic version. And on that note, the action points are now:
I started implementing the new kd-tree building and querying algorithms. Two things are missing on the kd-tree front: