From: Robert Elz <hostmaster@munnari.oz.au>
To: djb@koobera.math.uic.edu (D. J. Bernstein)
Cc: drums@cs.utk.edu
Subject: Re: Domain literals 
In-Reply-To: Your message of "26 Mar 1996 22:37:21 GMT."
             <19960326223721.18061.qmail@koobera.math.uic.edu> 
Date: Wed, 27 Mar 1996 10:13:07 +1100
Message-Id: <6547.827881987@munnari.OZ.AU>
Sender: kre@munnari.oz.au

    Date:        26 Mar 1996 22:37:21 GMT
    From:        djb@koobera.math.uic.edu (D. J. Bernstein)
    Message-ID:  <19960326223721.18061.qmail@koobera.math.uic.edu>

    Question: Should 822bis addresses be limited to Internet mail addresses?

Gawd, that would be a radical change.   I actually think I
like it, I'd prefer not to have any addressing restrictions
in 822, just allow 822 to deal with headers, and have the
addressing details someplace else.

I do assume here that you know that you are doing away with
the (822) requirement that addresses have an '@' character
in them, or that any particular characters are restricted in
addresses.

That is

	To: a!b!c!user

would become a perfectly valid 822bis header (that of course
is a uucp address, not Internet), as would

	To:  host::otherhost::user

which is a decnet type address.

Yes, I think I'd approve that.

On the other hand, if we're to remain internet centric, and
retain the '@' requirement, restrict the character set choice,
etc, then I think I'd make it rational, and certainly get rid
of the absurd user@thing.[literal].otherthing meaningless syntax
which has no Internet meaning at all, and other than treating
"[literal]" as a sub-domain of "otherthing", which certainly
could be done (the DNS would not object) never will.

kre