Re: Submitted for your review: random sampling filters in gawk

On Jan 3, 4:28=A0am, "Luuk" <l...@[EMAIL PROTECTED]
> wrote:

> At 1st i was amazed by the randomness of this script,
> but i think i need an expert on randomness to tell why above is
happening.=
...
> (but still, thanks for sharing the scripts!)

srand() seeds the random number generator with the time in seconds, so
if your computer is fast enough that behavior of repeated selection
could happen. How does it work if you put a sleep 1 in your scripts?

On the other hand there looks to be an edge case with sample size n=3D1
as well. I would not be surprised at all. I can look into it a bit
more and post a fix if I find the bug in my implementation.

I think your script labels may be switched. I am seeing funny behavior
with the Vitter algorithm and sample size n=3D1.  The simpler one seems
to be fine.

Thanks!

-- Steve