Hackers,
while reading code written by my GSoC student for knn-spgist I faced many questions about calculation distance from point to box.
1) dist_pb calls close_pb which calls on_pb, dist_ps_internal, close_ps and so on. So, it's very complex way to calculate very simple value. I see this way has some mathematical beauty, but is it acceptable in practice?
2) computeDistance in knn-gist uses it's own quite simplified implementation of this calculation. But why has it own implementation instead on simplifying dist_pb?
3) computeDistance implementation looks still very complex for me. Right way (simple and fast) to calculate point to box distance seems for me to be so:
double dx = 0.0, dy = 0.0;
if (point->x < box->low.x)
dx = box->low.x - point->x;
if (point->x > box->high.x)
dx = point->x - box->high.x;
if (point->y < box->low.y)
dy = box->low.y - point->y;
if (point->y > box->high.y)
dy = point->y - box->high.y;
return HYPOT(dx, dy);
I feel myself quite tangled.
Could anybody clarify it for me? Did I miss something? Thanks.
------
With best regards,
Alexander Korotkov.