Errors-to: owner-drums@CS.UTK.EDU
Message-ID: <Ijn9lZu00WBwQ0wYpB@andrew.cmu.edu>
Date: Wed, 31 May 1995 12:42:13 -0400 (EDT)
From: John Gardiner Myers <jgm+@CMU.EDU>
To: drums@CS.UTK.EDU
Subject: RFC 821 problem list
Beak: Is

The definition of "final delivery" is too SMTP-centric.  The
definition of "final delivery" in 822 and the current notary work is
when the message leaves the transport system.

The document uses BNF instead of ABNF to define the formal syntax,
causing confusion.

The VRFY command uses the fuzzy concept "user", when in practice it is
used to verify local-parts.

The MAIL command is not permitted to return a 503 reply, but there are
situations where a 503 reply to MAIL is appropriate.

The 522 reply code identifies a *transient* condition--the error is
likely to go away if the RCPT command is retried in a separate
transaction.  The condition should instead be assigned a 4XX reply
code.

The grammars for local-part are different in 821 and 822.  821 allows
a local-part of ``Joe\,Smith'', which is not allowed by 822.  822
allows a local-part of ``Joe.",smith"'', which is not permitted by
821.

The grammars for Received: headers are different in 821 and 822 (even
after modification by 1123) and do not reflect common use.

"The mail data may contain any of the 128 ASCII characters" permits
NUL octets, as well as bare CR and LF octets, which are not handled by
the vast majority of the installed base.

Appendices B, C, and D are obsolete.

The "send", "saml", "soml", and "turn" commands are obsolete.

There exist some SMTP servers which refuse incoming connections when
they cannot reverse-lookup the domain for the incoming IP address.
This goes against the spirit, but not the exact letter of RFC 1123
section 5.2.5.

-- 
_.John G. Myers		Internet: jgm+@CMU.EDU
			LoseNet:  ...!seismo!ihnp4!wiscvm.wisc.edu!give!up