RFC1225 / POP3

John C Klensin <KLENSIN@infoods.mit.edu> Wed, 15 May 1991 08:27 UTC

Received: from INFOODS.MIT.EDU by NRI.NRI.Reston.VA.US id aa01788; 15 May 91 4:27 EDT
Date: Wed, 15 May 1991 04:29:14 -0400
From: John C Klensin <KLENSIN@infoods.mit.edu>
Subject: RFC1225 / POP3
To: gvaudre@NRI.Reston.VA.US
Message-ID: <674296154.814441.KLENSIN@INFOODS.MIT.EDU>
Mail-System-Version: <VAX-MM(279)+TOPSLIB(151)@INFOODS.MIT.EDU>

More little bits of diddly that impact the architecture doc...

This probably does not need to be fixed immediately, but I note that, in
RFC1225 Marshall has used UA and MTA terminology somewhat differently
than we do.  In particular, he talks about a message-originating UA on 
the client machine using SMTP as a protocol, while the message receiving 
one uses POP3 and, consequently, he describes POP3 as a "split-UA" 
model.

I don't think there is anything wrong with the "split-UA" terminology, 
but I think that it is important in our language that only MTAs use 
SMTP.  I.e. Marshall's model/terminology would be something like

Sending:       Client              |     Server
                        UA  ->   SMTP  ->  Relay (MTA) ->
Receiving:
                       UA  <-    POP3  <- Depository  <- receiver-MTA

our version would be, more or less:

Sending:                           |
        composer -> UA - MTA  -> SMTP   -> Relay MTA ->
Receiving:
        reader <- UA     <-      POP3   <- Depository <- receiver-MTA

It would be nice to get him to fix the POP3 description to be consistent 
before it becomes a Standard.  Otherwise, we are going to spawn no end 
of confusion as people look at the descriptions in the two documents.
    --john

RFC1225 / POP3 John C Klensin