On Mon, Nov 14, 2011 at 4:24 AM, Andrea Canciani wrote:
> Instead of artificially introducing collisions in the step value by
> replacing 0 with 1 (which causes the value 1 to have twice the
> frequency of any other value), the step value can simply be computed
> as an uniformly distributed value
Instead of artificially introducing collisions in the step value by
replacing 0 with 1 (which causes the value 1 to have twice the
frequency of any other value), the step value can simply be computed
as an uniformly distributed value in the range [1, rehash], extremes
included.
This is safe becaus
