Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!cs.utexas.edu!uunet!ncis.tis.llnl.gov!MIRSA.INRIA.FR!Christian.Huitema
From: Christian.Huitema@MIRSA.INRIA.FR (Christian Huitema)
Newsgroups: comp.protocols.iso.x400.gateway
Subject: Fw: Re: RFC 987 V 2.0
Message-ID: <8907171535.AA14655@jerry.inria.fr>
Date: 17 Jul 89 15:34:58 GMT
References: <RFC-822*854.616423072@UK.AC.UCL.CS>
Sender: root@ncis.tis.llnl.gov
Reply-To: kehres@tis.llnl.gov
Distribution: inet
Organization: The Internet
Lines: 59
Approved: post-x400-gateway@tis.llnl.gov

Steve,

I am just reading your revised document, and I found a little wording
problem. The general tone of section "4.2.1.  Encoding of Personal
Name", page 40, is very "implementor friendly", with sentences like:
"It seems desirable to utilise the "human" conventions for encoding
these components.  A syntax is proposed here.".

You have to be a lot more specific. Just look at the following log:

220 Welcome to IDA Sendmail 5.59++/IDA-1.2.8 running on mirsa.inria.fr
HELO pussy.inria.fr
250 Hello pussy.inria.fr, pleased to meet you
MAIL From: <huitema>
250 <huitema>... Sender ok
RCPT To: <Christian.Huitema@mirsa.inria.fr>
250 <Christian.Huitema@mirsa.inria.fr>... Recipient ok
RCPT To: </G=Christian/S=Huitema/@mirsa.inria.fr>
554 </G=Christian/S=Huitema/@mirsa.inria.fr>... Cannot mail directly to files
QUIT
221 mirsa.inria.fr closing connection

or, if you prefer an exemple which is clearly not related to X.400:

220 sering.cwi.nl Sendmail 5.45/4.8 ready at Mon, 17 Jul 89 17:26:38 +0200
HELO mirsa.inria.fr
250 sering.cwi.nl Hello mirsa.inria.fr, pleased to meet you
MAIL From: <huitema@mirsa.inria.fr>
250 <huitema@mirsa.inria.fr>... Sender ok
RCPT To: <Daniel.Karrenberg@cwi.nl>
250 <Daniel.Karrenberg@cwi.nl>... Recipient ok
RCPT To: </G=Daniel/S=Karrenberg/@cwi.nl>
554 </G=Daniel/S=Karrenberg/@cwi.nl>... Cannot mail directly to files
RCPT To: <no_such_user@cwi.nl>
550 <no_such_user@cwi.nl>... User unknown
QUIT
221 sering.cwi.nl closing connection

I guess it is mostly a problem of wording, but it is clear that what
you call the EBNF mapping should be absolutely avoided if the personal
name mapping is possible. This would probably have much worse effects
than the "OU ordering" misfit! The same is indeed true for all other mappings, like:

220 sering.cwi.nl Sendmail 5.45/4.8 ready at Mon, 17 Jul 89 17:30:04 +0200
HELO mirsa.inria.fr
250 sering.cwi.nl Hello mirsa.inria.fr, pleased to meet you
MAIL From: <huitema@mirsa.inria.fr>
250 <huitema@mirsa.inria.fr>... Sender ok
RCPT To: <dfk@sering.cwi.nl>
250 <dfk@sering.cwi.nl>... Recipient ok
RCPT To: <dfk/OU=sering/@cwi.nl>
550 <dfk/OU=sering/@cwi.nl>... User unknown
QUIT
221 sering.cwi.nl closing connection

I have observed that, when mapping the envelope, using any EBNF
encoding is in practice equivalent to bouncing the mail...

Christian Huitema