On Mon, Apr 17, 2006 at 11:39:36PM -0500, Ryan Malayter wrote: > > Do you think a simpler calculation would work well enough to sort > hundreds of servers by "closeness"?
Divide the world into areas (e.g., every 2 deg. lat/long). Pre-compute the closest servers for each of those areas. Now the math is just a few compares and a hash table lookup. [Honestly, I suspect this problem is not CPU-constrained] _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
