Re:Hashing to estimate the size of a stream
by jsnell on Wednesday January 06, 2016 @03:38PM
(#10853571)
It shouldn't, the algorithm as described works for arbitrary data. The whole point of the hashing is to transform the stream of arbitrary data to a stream of uniformly distributed random integers, so that you can then apply the reasoning from that paragraph.
This needs to be in the problem statement.