Solved: Sendmail question. Problem with yahoo.

[Ben Scott]

On Fri, Apr 18, 2008 at 1:54 AM, Bill McGonigle <bill at bfccomputing.com> wrote:
>  Yeah?  I've seen benchmarks with postfix spanking sendmail on
>  performance, and exim handing it right to postfix.

  I've seen benchmarks that say just about anything.  Even when
they're not designed to produce a certain desired outcome, benchmarks
are very often biased by factors that aren't always immediately
obvious.  For one, they tend to reflect the knowledge base of the
people configuring the software under test.  If Acme Consulting Group
compares Postfix and Sendmail and finds Sendmail performs better,
what's more likely: That Sendmail beats Postfix, or that Acme's staff
simply knows Sendmail better than they know Postfix?

  High-performance benchmarking can also reveal performance
characteristics in the OS or hardware that happen to favor a
particular program.  And maybe the OS wasn't tuned best for this or
that program.  When you're talking high-performance, the whole system
really does matter.

  If you're doing the benchmarks are for your own use, then most of
these concerns don't matter, because of course it's what you're using
and what you know that counts.  But taking someone else's benchmarks
and generalizing them to other situations can be a very misleading
thing to do.

  This is not to defend Sendmail in this respect.  I honestly have no
idea.  Just... "Lies, damn lies, and statistics."

-- Ben