[Following the good example of Paul Vixie in the NANOG list today I'm
going to implore people to *not* use "reply-all" and instead just reply
to the lists as I've done. I'm quite tired of getting at least three
copies of every message.]
[ On Sat, November 1, 1997 at 11:15:57 (-0500), Evan Leibovitch wrote: ]
> Subject: Re: Goodbye Smail
>
> Given the limited resources naturally available on freeware projects, the
> massive effort needed to maintain a state-of-the-art MTA, and the immense
> similarity between exim and smail, does it really make sense to have two
> such closely parallel projects re-inventing each others' wheels?
Shared volunteer run projects are far more difficult to manage than
those with one central leader. I'm not saying I couldn't work with
Philip (nor probably would he say that either, though we would have lots
of fun debating coding styles, etc. ;-) but at this point we're both
very much working on our own turf.
Of course if you believe the various so-called experts in the industry
it's impossible for a software project to succeed if there's no strong
leadership in control at the top (I had a quote from some book about
this, and I thought it was Philip Metzger's "Managing Programming
People", but I can't find it just now).
In terms of the history I was well aware of Philip's efforts all along
and I hope I've supplied him with only encouragment, ideas, and
constructive critisism along the way. At the time I took over the
maintenance of Smail-3 from Nigel Exim was still not quite ready for
prime time, never mind a full replacement for most of Smail's
capabilities. I was still in need of a working mailer so chose to stick
with Smail for the time being. (I was seriously looking at other mailer
too, such as Rayan's Zmailer and even the possibility of a freed version
of the Bell Labs UPAS mailer.)
> I don't understand Greg's point of "slightly different user bases". Are
> they different in their needs, or merely their makeup? What are their
> separate requirements, such that a pooling of talent between smail and
> exim could not give them what they need?
Well, there the things others have already said about various features
unique to one or the other and the legacy support in Smail-3. However
what I really had in mind was one other point. My goal with Smail ever
since well before the 3.2-alpha process began has been to build a mailer
that installs out of the box and just does the right thing for the
majority of simple cases (modulo some of my decisions about what's best
for default security) without the need for *any* manual configuration.
I don't think the simple case of being able to send and receive mail
should require any configuration beyond what can be learned at compile
and run time directly from the system itself. My current impression of
Exim is that it has far too many knobs to twist and not yet enough of a
body of experience to show what the best setting for each is for the
common cases (though I may be mistaken about this).
When it comes time for me to give up on Smail too I don't think Exim
will be my primary choice despite its advanced capabilities and high
quality. It follows the same over-all monolithic model sendmail and
Smail-3 do and I think I'll soon be ready to try something new. Of
course this might not happen until after I've had time to get Smail to
it's next major release (never mind the upcoming patch release), so
don't hold your breath! ;-)
--
Greg A. Woods
+1 416 443-1734 VE3TCP <gwoods@???> <robohack!woods>
Planix, Inc. <woods@???>; Secrets of the Weird <woods@???>
--
* This is sent by the exim-users mailing list. To unsubscribe send a
mail with subject "unsubscribe" to exim-users-request@???
* Exim information can be found at http://www.exim.org/
