Date: Fri, 05 Apr 1996 19:03:27 -0500
From: John C Klensin <klensin@mail1.reston.mci.net>
Subject: Re: generating RFC 822 addresses
To: Chris Newman <chrisn+@CMU.EDU>
Cc: drums@cs.utk.edu
Message-id: <2.2.16.19960406000327.0a7734a0@mail1.reston.mci.net>
MIME-version: 1.0
Content-type: text/plain; charset="us-ascii"
Content-transfer-encoding: 7BIT

At 22:38 96.04.01 -0500, Chris <can@you,reply;to:\"this?\"> Newman wrote:
>It's clear we need separate advice for generating addresses.  Since it
>may be easier to reach concensus on this stuff, I'm taking a crack at it.
>...

>3) The domain MUST be encoded as a series of "." separated atoms with
>no internal linear-white-space, or as a single domain-literal with no
>internal linear-white-space.  (justification: consistancy, works best
>with installed base)

Also effectively required by the DNS RFCs.  Inconsistency with them is as
bad as inconsistency between 821 and 822.

>
>3) Comments should not be generated. (justification: works best with
>installed base, only common use of comments should be replaced by the
>phrase <route-addr> form)

Assuming that we are talking about addresses here, rather than other sorts
of things, I'm not completely sure about this, although expressing a very
strong preference for phrases over comments seems reasonable to me, as would
forcing only the form
    phrase <route-addr> (comment)
if comments must be used for some reason.  I don't want to force editorial
remarks like "a blowhard" in
    John Klensin <klensin@mci.net> (a blowhard)
or even in
    klensin@mci.net (a blowhard)
into the personal-name phrase.  To the degree to which we are inventing
obscure mailbox names again (anyone else getting mail from AOL?), really
defining semantics for that "phrase" that binds it to personal names and
encouraging it to be used that way, and that way only, seems to me to be a
good idea.  If the price of doing that is getting stuck with comments, then
I think that comments are the lesser evil.

    --john