Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!decwrl!ucbvax!VAX.BBN.COM!dkovar
From: dkovar@VAX.BBN.COM
Newsgroups: comp.protocols.appletalk
Subject: Macintosh mail systems.
Message-ID: <Added.0WNx58y00Ui34=3k9_@andrew.cmu.edu>
Date: 16 Apr 88 21:06:32 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 63


  From the discussion that we've been seeing on info-appletalk and
some personal mail message I've received I have the following impressions:

	1) Many people want some way to send mail from their Mac to
	   their favorite mainframe. 

	2) Many people lack the resources to do it on their own.

	3) Many people would like this to be as cheap as possible.

	4) It should be well supported.

  All of this adds up to a system that will support SMTP, POP, and
a few other odd mail protocols, that will run over Appletalk, Ethernet,
and serial lines, that is written by competent programmers who aren't
going to vanish after having written 90% of it, and that is easy to
license and easy to get.

  Various local implementations exist. Stanford just announced that
their Telenet package will include some mail services. CMU is working
on a system to tie Macs into the Andrew mail/messages system, Dartmouth
has a system set up using dropboxes tied into their campus mail system,
and CITI apparently ported some unix code over to the Mac that sits on
top of the TCP/IP drivers. As far as I know, none of these are complete
and only the Stanford one will really satisfy anyone else's needs.

  I suggest two things:

	1) That those people interested in such a system define their
	   own needs in detail. Stating that you'd like your Mac
	   to talk to your local UNIX machine isn't enough. What
	   mail protocols do you use, what network protocols, what
	   machines you're talking to with your Mac and, perhaps
	   most importantly, what do you want your user interface
	   to look like. Should it provide its own editor? How
	   should the local mailbox be handled? How about dropboxes?
	   Go wild as it is usually easier to prune out the far out
	   ideas than add them in after the fact.

	2) That those people interested in getting it done try and
	   figure out how to get it done. The response has been
	   quite impressive so there is definitely a want if not
	   a need. With the backing of all these people we could
	   certainly go to one of the companies selling Macintosh
	   based mail systems and say "Look, these companies and
	   these universities would like, and would pay for, software
	   that provides these services." Something might happen.
	   You'd also end up paying "real money" to get it done.
	   The ideal system would be public domain with source with
	   the normal "you may use it but you cannot sell it and please
	   send any enhancements back to us" restrictions. Is there
	   a good way to get this done? Can I make myself a non-profit
	   organization and do it? Can we find some university that
	   is willing to take this on? Seems like we need a
	   Consortium Development Group. 

  I'm just tossing out ideas at the moment. Please respond to them,
either to the group or to me personally and I'll collect and summarize
them.

-David Kovar
 DKovar@BBN.COM