Re: POP3 protocol question

Ned Freed <> Wed, 19 October 1994 02:17 UTC

Received: from by IETF.CNRI.Reston.VA.US id aa10867; 18 Oct 94 22:17 EDT
Received: from CNRI.Reston.VA.US by IETF.CNRI.Reston.VA.US id aa10863; 18 Oct 94 22:17 EDT
Received: from PO5.ANDREW.CMU.EDU by CNRI.Reston.VA.US id aa24427; 18 Oct 94 22:17 EDT
Received: (from postman@localhost) by (8.6.9/8.6.9) id WAA18149; Tue, 18 Oct 1994 22:00:35 -0400
Received: via switchmail for; Tue, 18 Oct 1994 22:00:33 -0400 (EDT)
Received: from via qmail ID </afs/>; Tue, 18 Oct 1994 21:58:29 -0400 (EDT)
Received: from (SIGURD.INNOSOFT.COM []) by (8.6.9/8.6.9) with ESMTP id VAA09843 for <>; Tue, 18 Oct 1994 21:58:12 -0400
Received: from SIGURD.INNOSOFT.COM by SIGURD.INNOSOFT.COM (PMDF V4.4-0 #2001) id <01HIFD1EJGHC8ZEKJJ@SIGURD.INNOSOFT.COM>; Tue, 18 Oct 1994 18:56:46 -0700 (PDT)
Date: Tue, 18 Oct 1994 16:06:44 -0700 (PDT)
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Ned Freed <>
Subject: Re: POP3 protocol question
In-reply-to: Your message dated "Mon, 10 Oct 1994 12:50:14 -0400 (EDT)" <>
To: Chris Newman <>
Cc: POP3 IETF Mailing List <>, Jerome Chan <>
MIME-version: 1.0
Content-type: text/plain; charset=us-ascii
Content-transfer-encoding: 7BIT
References: <aabeb23a03015006c552@[]>, <aabeb23a03015006c552@[]>

> RFC-821 (SMTP) allows one to _send_ mail.  POP3 has no reason to
> duplicate this functionality.

The only argument for XTND XMIT and similar facilities that I somewhat agree
with is when you're using POP (or IMAP) directly over something like a serial
line (possibly with some sort of lightweight error correction and possibly
not). Like it or not, some environments just don't have the option of using
TCP/IP everywhere, SLIP and PPP notwithstanding.

The counter, of course, is that these are protocols specified for use over
TCP/IP and that issues that arise with other transports and their limitations
are not relevant. Its a shallow argument, in that it appeals to procedure
to justify a technical decision, but it is valid nevertheless.

It would therefore be appealing to have a "one port" protocol to use for email,
but just because its appealing is no real it has to come to pass.