why is sendmail the adopted mta?

[Miguel Filipe]

On Sun, 2 Jan 2005 13:03:26 +0100, Simon 'corecode' Schubert
<corecode at xxxxxxxxxxxx> wrote:
> On 02.01.2005, at 02:15, Miguel Filipe wrote:
> > After talking about this on #dragonflybsd@ efnet, and with some
> > agreace that there are more pleasent, easy and secure mtas out there,
> > ppl suggested I talked about this in the MLs.
> 
> ... about the 10th time or so, now

It was I who asked about it...

> 
> > One thing that allways itches in my head is why almost all bsds use
> > sendmail has the default MTA.
> >
> > Why do I have gripes with sendmail?
> > 1) Its difficult to configure
> > 2) it has a lousy security track record
> > 3) there are very good (superior) alternatives
> 
> these are your opinions, i gather. other people might disagree.
> 

about 2)
Sorry, but sendmail's track record on security isn't my opinion.. its
a know historical fact!
Postfix is more secure _by design_, and experience shows that on
field.... the security track record of postfix is simply _very good_.



> > But there are also other choices: exim; qmail; ....
> > And the same way that I'd prefer that sendmail wasn't forced on me
> > when I install a  dragonflybsd... I don't mind having the choise of
> > mta on install.. instead of a default one.
> 
> we're not there yet. not even half way; for that we'd need to package
> the base system which is something that is highly controversial, too.
> 
> > A also good solution, instead of adopting some other mta, having the
> > user too choose the mta on install is also better than having sendmail
> > installed by default.
> 
> that's the same as above. but where do you draw the line? maybe i
> myself don't want to use bsd ls and instead use gnu ls
> --with-fancy-colors=yes. should we give the user a choice there, too? i
> don't think so. but you have a choice when compiling world: set
> NO_SENDMAIL (or however it may be called) and sendmail won't get built.
> install the postfix port and you're happy.

and why not the other way around?

> 
> > Why do I bring  this up?
> > Because I see no good reason to keep using sendmail ... and because I
> > believe that dragonflybsd is doing a great work adopting and creating
> > new software solutions that look better than the existing ones....
> > pf
> > open ntpd
> > dragonfly itself .. when comparing to freebsd.
> 
> as you might see no ``good'' reason, there might be hundreds of people
> seeing a very good reason to keep it.
> 

Okay, speak about it, what are those reasons?

> > In conclusion, I  thought about this because  I think that replacing
> > sendmail with postfix is a reasonable, racionnal and inovative move
> > that plays well with the DragonFlyBSD spirit!
> 
> in conclusion i think that keeping sendmail in base is a reasonable
> point, too.

Why is it reasonable again?
Because changing a base aplication drives you to wonder if you
can/should do that to some oter application? thats a good thing! Its
called thinking and not accepting something just because its there for
ages...


> 
> i don't use sendmail, too. i'm using postfix, from ports, and i'm happy
> with it. you got the choice.

I've got the choice of ditching the allready installed sendmail and
installing postfix.
Yes, but what are the raccional reasons to use sendmail by default
instead of "having the choice" to ditch the default smtpd and install
sendmail from ports?


> 
> cheers
>    simon
> 
> keywords: better MTA sendmail postfix switch change remove base system
> neverending story

About never ending stories... 
I think is good that ppl argue and question what is and what's not,
and most importantly, WHY.
Questioning stuff is good, and makes us _think_ the whys and why
nots.. that way you can raccionally make a point and decide
something... Instead of just accepting whats allready being used
without thinking about it.

Whithout questionning ourselves, there isn't change, and there isn't progress...

> 
> --
> /"\
> \ /
>   \     ASCII Ribbon Campaign
> / \  Against HTML Mail and News
> 
> 
> 

Before this gets out of control, if this idea isn't that absurd, and
there are more important things to be done, thats OK, each thing at
its time...

lets finish up VFS work first, 
work on software packaging after that...
and after that about things like
default mta.. 


keep up the good work

-- 
Miguel Sousa Filipe