Errors-to: owner-drums@CS.UTK.EDU
To: Keith Moore <moore@CS.UTK.EDU>
Cc: perry@imsi.com, John Gardiner Myers <jgm+@CMU.EDU>, drums@CS.UTK.EDU
Subject: Re: RFC 821 problem list 
In-Reply-To: Your message of "Wed, 31 May 1995 14:13:33 -0400."
             <199505311813.OAA11913@wilma.cs.utk.edu> 
Date: Thu, 01 Jun 1995 05:28:37 +1000
Message-Id: <810.801948517@munnari.OZ.AU>
From: Robert Elz <kre@munnari.oz.au>

I am confused.  What is 522?   My copy of 821 mentions 522
nowhere.   Was this perhaps intended to be 552 (over quota,
or whatever), in which case I'm not sure I'd like to see it
as a 4xx code.   Sure, if retried later there may be available
space, but how much later?   This isn't like a 450 (mailbox
locked) which should be transient, a mailbox may be over
quote for weeks, or months.  In fact, if a system thought that
an over quote mailbox was likely to be a transient condition,
it could issue a 450 code now, and be within spec, we don't
need a new 4xx code for that.

If the idea were to move every error that might eventually be
corrected into the 4xx (try again) codes, then we should also
be moving 502 (command not implemented - after all, the dest
site may install a new mailer tomorrow where it is implemented)
551 (user not local - they may become local tomorrow), and
others.   About all that would remain 5xx codes would be syntax
errors.

I much prefer to bounce these potential long term problems,
which includes over quota, so the sender can contact the
administration of the dest site, and see if they can correct
the problem.  Over quota usually equates to "user on holiday
and not reading mail", which gets fixed when the admin moves
the users pending mail to some other place (or deletes the
user because the holiday turned permanent, and no-one had
noticed - this happens with students all the time).

kre