Re: my 2 cents worth

Michael D'Errico <> Tue, 07 June 1994 22:06 UTC

Received: from by IETF.CNRI.Reston.VA.US id aa19882; 7 Jun 94 18:06 EDT
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa19877; 7 Jun 94 18:06 EDT
Received: from PO3.ANDREW.CMU.EDU by CNRI.Reston.VA.US id aa19224; 7 Jun 94 18:06 EDT
Received: (from postman@localhost) by (8.6.7/8.6.6) id RAA04511; Tue, 7 Jun 1994 17:57:50 -0400
Received: via switchmail for; Tue, 7 Jun 1994 17:57:49 -0400 (EDT)
Received: from via qmail ID </afs/>; Tue, 7 Jun 1994 17:55:40 -0400 (EDT)
Received: from ( []) by (8.6.7/8.6.6) with ESMTP id RAA04348 for <>; Tue, 7 Jun 1994 17:55:33 -0400
Received: from rome ( []) by with ESMTP id AAA3660 for <>; Tue, 7 Jun 1994 14:55:27 -0700
Subject: Re: my 2 cents worth
In-reply-to: <>
Date: Tue, 07 Jun 1994 14:55:25 -0700
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Michael D'Errico <>
Message-ID: <>

Ian Duncan <> wrote:
>On Tue, 7 Jun 1994, Michael D'Errico wrote:
>> I think that characters from Base64 are too restrictive.  Base64 exists
>> because of the need to gateway messages between different mail systems,
>> some of which have horrible constraints.  With POP, you have an 8-bit
>> connection between client and server to work with.
>I think the premise that POP implies an 8 bit path between ends is a
>little dangerous. In the absence of words to the contrary in the RFC I'd 
>assume printable ascii and very selected controls.

Your selective quoting suggests that I wanted to allow 8bit characters
in the unique identifiers.  I made no such statement, and if you quoted
my next sentence that would be clear.  I wrote, "I think that Steve
Dorner's suggestion to limit the uid's to characters in the range 0x21
to 0x7e (printable 7-bit ASCII except space) is a good one."

>I know the origin of base64. I was suggesting we avoid tokens that are
>long strings of ascii and instead use very large numbers, encoded in
>base64 to keep the representation short -- 24 digits of base64 can carry
>2^144 bits in a safe, documented, well understood and easy to decode

My MTA stores messages in separate files with pseudo-message-id's as the
filename.  These are unique for every message that comes into the system
so I would like to be able to use them as the unique identifier in the
UIDL command.  Here is how UIDL currently works on my POP server:

        C: UIDL
        S: +OK
        S: 1 19940603172145.AAA8293@ANDREW.CMU.EDU
        S: 2
        S: 3
        S: 4
        S: .

I don't see why this is so bad.  I'd hate to force myself and others to
encode information into Base64 when there's no need to.

Michael D'Errico