Re: picking a new random seed

On Jan 28, 5:09 pm, "Curtis A. Jones" <curtis_jo...@[EMAIL PROTECTED]
> wrote:
> Phil,
> On this Windows XP system with Workstation APL2 the clock "ticks" seem
> to be much more than 1 millisecond.  i.e. Successive values of
> {quadTS} are incremented by something closer to 10 ms than 1 ms.  (I
> ran {quadTS} 5000 times and got 15 unique millisecond values: 756 766
> 776 786 796 806 816 826 836 846 856 866 876 886 896.)
>
> This may not be a problem for your approach to finding a number of
> random numbers to generate if the millisecond element of {quadTS} can
> take on at some time every one of the integer values from 0 to 999.
> Curtis Jones

Curtis,
I tried what you suggest and more than agree.
=E2=88=AA{0=E2=8A=A5=E2=8E=95TS}=C2=A8=E2=8D=B35000 =E2=8D=9D unique{0
decod=
e quadTS}each iota 5000
being equivalent to what you said on Dyalog gives me, each time I run
it either one or two values, rather less than 15, which of course
depends merely on the speed of the machine, algorithm or interpreter
but each time I do it again I get a different one or two. The pairs
are always about 15 or 16 ms apart being the tick speed, apparently
1/60 or 1/64 sec. (Didn't it used to be 1/18 sec?)
But as you hint it isn't a problem if the intention is just to run it
once in a while.
Phil