Please clarify: Why are random pivots not practical? And why do you say there are no truly random quicksorts? Because pseudorandom numbers aren't really random, or for some other reason?
The cost of true randomness is really high (in execution time)
> And why do you say there are no truly random quicksorts?
Because pseudo-random numbers aren't really random.
The only reason this is important is that the pseudo-randomness means that its possible the construct an adversarial input, which could drive the quicksort to take N^2 time.