Tom,

For all practical purposes, using random numbers to sort does the jo
but still on the object of academic interest, what happens if th
random numbers generated are *not* 100% unique?

Put another way, if it is -conceivable - that two random number
returned by the RND function could be exactly the same, we can d
violence to the facts and logically assume that in a very very rar
moment of aberration, all the random numbers returned in a range wil
be equal. That being the case, the sorting mechanimsm will b
stullified.

I do concede that the probability of getting 2 random numbers the sam
in a range of 65536 rows should be infinistemally close to zero, an
the Prob(E) of having all 65536 instances even more so. However, fo
as long as Prob(E) <0, we cannot guarantee *100% accuracy* in sortin
-in all cases -and -at all times-. In the context of pure mathematica
accuracy, Randomization could yet fail us.

Food for thought.


Davi

--
david
-----------------------------------------------------------------------
davidm's Profile: http://www.excelforum.com/member.php...fo&userid=2064
View this thread: http://www.excelforum.com/showthread.php?threadid=48472