Message-Id: <199612190413.XAA25248@ig.cs.utk.edu>
From: Keith Moore <moore@cs.utk.edu>
To: cos@leftbank.com
cc: drums@cs.utk.edu, moore@cs.utk.edu
Subject: Re: The string after HELO 
In-reply-to: Your message of "Wed, 18 Dec 1996 21:07:56 EST."
             <199612190207.VAA28405@zax.leftbank.com> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Wed, 18 Dec 1996 23:13:57 -0500
Sender: moore@cs.utk.edu

> So, back to my original objection:  Allowing something other than a
> domain after HELO now, would break existing compliant
> implementations.

I suggest:

a. a client SHOULD send its fully-qualified domain name, if it is
available.

b. otherwise, a client MAY send any string which can identify that
particular network node

c. servers SHOULD include the EHLO/HELO argument in Received fields
and log messages. 

Explanation: If a nondelivery report or other complaintis forwarded to
a postmaster at the client site, the string used by the CLIENT SHOULD
be sufficient to identify the network node of that client to the
postmaster.

d. In all cases, the HELO/EHLO argument MUST be in correct domain name
syntax.  If the client does not know its fully-qualified domain name,
it may append ".unknown" to any atom used to identify the client node.

Keith