On 27/9/2013 12:10, Denis McMahon wrote: > On Fri, 27 Sep 2013 12:08:33 +0000, Dave Angel wrote: > >> i recall >> writing a shuffle function in C decades ago, which took an array of (52) >> unique items and put them in random order. > > Whenever I tried to write shuffles I came up against a fairly fundamental > limit: > > 52! > prng states > > Granted prngs seem to be better on the importing entropy from elsewhere > front these days. >
in 1978, I implemented a random number generator that the OS sequenced through between scheduling processes, so that any individual process would be relatively safe from that phenomenon. Of course I was limited to the precision of the math package - 10**13 -- DaveA -- https://mail.python.org/mailman/listinfo/python-list
