Re: Debian's mail daemons
mdorman@calder.med.miami.edu (Michael Alan Dorman) wrote on 15.06.97 in <[🔎] 87lo4bsge4.fsf@calder.med.miami.edu>:
> My two personal reservations:
>
> 1) I think Daniel J. Bernstein (qmail's author) doesn't seem to know
> how to have a technical discussion without seeming as if he's tacking
> an implicit "you stupid idiot" on to the end of each sentence, which
> irks me.
And it's not always implicit, either. Having him on a technical mailing
list does make for frequent flame wars, in every case I've seen.
> 2) I understand that he did measurements, and found that it was
> faster to do multiple SMTP connections than doing multiple RCPT's for
> a single message going to multiple users on a single host, but, I
> think he should have done what he basically did in every other
> instance and said, "tough shit, it's a problem with your MTA, switch
> to qmail, which does it right", and then had qmail do it right (in
> other words, quickly).
Multiple simultaneous SMTP connections are, of course, highly antisocial
to the host at the other end. And if multiple sequential SMTP connections
are faster, then somebody is doing something very wrong here.
> Absent those reservations, which are not serious, he's written quite a
I consider them very serious. Besides, the author's attitude problem
applies to standards, too, another point I find unacceptable.
I find myself completely unable to trust him. So, qmail cannot be an
option for machines I administer.
I should probably mention that I have similar trust problems with
sendmail. Besides the sendCERTadvisories problem, I have trouble trusting
a program when it's author admits that he purposefully inserted a back
door in that program (it's the DEBUG hack that made the Internet Worm
possible - it was meant to give root access to the main sendmail test
machine, because the author didn't have the root password).
> mailing lists here at UM, using raw qmail. It is about 1000% times
> easier to set up than sendmail, seems to have a much lower impact on
> the system, etc.
What isn't about 1000 times easier to set up than sendmail? :-)
MfG Kai
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: