jsms: sendmail must die completes itself

Ivan Kohler ivan@sisd.com
Thu, 8 Jul 1999 02:42:05 -0400


On Mon, Jul 05, 1999 at 02:56:40AM -0700, Jonathan Walther wrote:
> 
> Im writing a new MTA because Im not "smart".  Im a dumb user.  Im too stupid
> to understand sendmail postfix and exim in the first place.

As a sysadmin deciding what mail system to install, I wouldn't be
reassured by that.  I'd probably skip your MTA as quickly as possible.

Postfix's blurb is "Postfix attempts to be fast, easy to administer, and
secure, while at the same time being sendmail compatible enough to not
upset existing users".  Qmail's blurb is "qmail is a secure, reliable,
efficient, simple message transfer agent. It is meant as a replacement for
the entire sendmail-binmail system on typical Internet-connected UNIX
hosts.". 

Perhaps you should rething the approach you're taking to advocacy here.

May I also suggest that you post something web-accessable besides a .tgz
file?  I'd be interested in browsing your specs and such.

>  On the other
> hand, I read RFC 821 and 822 and they made a lot of sense to me.  If I can't
> even make sendmail etc work for myself, why would I write front ends for
> them?

I think you're loading way too much into that `etc'.

>  I find it simpler to write a new mailer, plus it won't have any
> "baggage".  Sendmail doesn't just handle SMTP, it handles about 15 other
> types of mailer accumulated over 20 years as well.  The configuration models
> of the other mailers are baroque and overfeatured.  Every other MTA out 
> there was written by someone who already understood sendmail and
> succumbed to its brain damage.

I disagree.  The next gneration MTAs, qmail and postfix, sound to me very
close to what you've described you're trying to write. 

No argument that sendmail is crufty as can be.

>  Thats about the only reasons.  If you think
> these are poor reasons, Im interested to hear why.

Far be it from me to say you shouldn't work on whatever project you
desire.  But your main criticism is the configuration step, and you've
been stressing that you are a user interface designre, and that elegance,
security and efficiancy are secondary goals.  I was suggesting, perhaps,
that you work on that specifically, and backend to a known elegent, secure
and efficient mailer.  *shrug*

> As a final reason, its to "scratch an itch".  None of the current mailers
> fit my multihoming needs, or are simple enough to easily handle all my
> virtual domains as well as system mail plus a bunch of mailing lists,
> relays, etc.  By "being able to handle", I mean, are understandable enough
> to me that I can make them do those things.  I know that Sendmail can do
> ANYTHING.  But like the automatic driver who can't drive a stick shift,  
> that means I can't even go from 0 to 30, let alone 60 to 100.  Later.
> 
> Jonathan Walther
> 
> -- sig --
> Witches are fucked up manhating mind perverts.
> 
> On Mon, 5 Jul 1999, Ivan Kohler wrote:
> 
> > I don't understand why you feel the need to write a new MTA.  Why not
> > simply apply your user interface design and write config generators for a
> > couple popular MTAs?  Best of both worlds.
> 

-- 
Ivan Kohler <ivan@sisd.com> - finger for PGP key - <moc.dsis@navi> Relhok Navi
Open-source billing and administration for ISPs - http://www.sisd.com/freeside
20 4,16 * * * saytime # please don't be surprised if you find me dreaming too