On Sat, Apr 01, 2000 at 03:51:47PM -0600, Charles Cazabon wrote:
> D. J. Bernstein <[EMAIL PROTECTED]> wrote:
>
> > The total number of servers that have ever run Postfix is smaller than
> > the number of qmail servers added in the past six months.
>
> Out of curiosity, how do you measure the number of qmail servers added in
> a given time period? None of the qmail installations that I have performed
> or use downloaded the source from cr.yp.to, and I have never run the
> "To report success" quickie in the instructions, so presumably you can't
> track them.
>
> Or are you counting only SMTP hosts identified through your automated
> surveys? I have the feeling that this would miss a lot of qmail installations.
Right. But it also misses a lot of postfix systems. Your question can only be:
"is your sampling method statistically valid". Is that what you're a professor
of mathematics?
Mark.
