> Martyn Johnson (private communication) has suggested a much simpler way
> of doing this, which I will probably adopt. In brief: you multiply each
> MX by N, add a random number in the range 0-N, and then sort as before.
> A simple pseudo-random number generator is used, seeded with the time.
> It doesn't really matter if it is a lousy generator. I suppose I could
> exclusive-or the process number into the seed as well.
>
>
> --
> Philip Hazel University Computing Service,
> ph10@??? New Museums Site, Cambridge CB2 3QG,
> P.Hazel@??? England. Phone: +44 1223 334714
>
