ietf-smtp
[Top] [All Lists]

Re: 2821bis-03 comments

2007-04-26 05:53:54



--On Thursday, 26 April, 2007 03:27 -0400 Hector Santos
<hsantos(_at_)santronics(_dot_)com> wrote:


Great work John!

Just a few comments.

1) Removal of 1yz in 4.2.1

With the the removal of 1yz, it is probably a good idea to use
a different example for reply codes in 4.2.1:

    For example:

       123-First line
       123-Second line
       123-234 text beginning with numbers
       123 The last line

Fixed in -04 working copy

2) Recording the new FTP text in 4.2.1

    It is worth noting that the file transfer protocol (FTP
[35]) uses a
    very similar code architecture and that the SMTP codes are
based on
    the FTP model.  However, SMTP uses a one-command,
one-response model,
    while FTP is asynchronous and the 1yz codes are not part
of the SMTP
    model.

IMO, I am not sure this is necessary John. Might be best to
not confuse future developers with this history note,
especially since 1yz is not longer part of the spec and even
more since one can argue the that DATA command is a one
command, two response model. In a  sequence of:

     C: DATA
     S: 354
     C: [upload data]  <--- this is technically not a command
     S: 250

The client is [arguably] not sending a new command to 354, but
rather the mail block.

I didn't know whether to include that or not, but thought the
most efficient way to get comments was to put it in.   Comments
from others welcome --it is easily removed or tuned if needed/
desired.  I decided to put it in after you pointed us to Charles
Kozierok's "TCP/IP Guide" piece and a brief exchange of notes
with the author who apparently was, in fact, confused by the FTP
comparison.   So that I don't lose track, Issue 31 is assigned
to this.
 
3) In 2.1, you added a MUST to this line:

    protocol requires that a server MUST accept responsibility
for either
    delivering a message or properly reporting the failure to
do so.

BRAVO, BRAVO!!

However, I am sending you an off-list comment about this.

Keep in mind that this wasn't intended to be a change of
substance, but a simple rewording of "...protocol requires that
a server accept..." per issue 16.   If people believe this is a
change in anything but wording, I should revert the text.
Issue 32.
 
Any who, I think with the MUST above, you probably should
upper case some "shoulds" in section 6.2 and I think if you
want to appease the mail process/AVS service market, I think
it may be appropriate to remove the clause regarding of
"syntax" and "mail content":

I think it can be just:

    Utility and predictability of the Internet mail system
requires that
    messages that can be delivered SHOULD be delivered.  If
    a message cannot be delivered, and cannot be rejected by
the SMTP
    server during the SMTP transaction, the message SHOULD be
"bounced"
    (returned with non-delivery notification messages) as
described
    above.  .........

I will take a look at the "shoulds" -- didn't recheck those on
the last pass.   My, and I hope our, goal here is to clarify
what the spec says and, as appropriate, to reflect existing
practice in the sending and handling of normal mail messages,
rather than to "appease" anyone.  I think this is part of issue
25.
 
and in my opinion, the addition of the sentence would be
helpful:

    If applicable, the SMTP server SHOULD attempt to determine
    delivery status during the SMTP transaction to help
minimize
    undesirable BOUNCE traffic.

thanks for the careful reading.
    john

<Prev in Thread] Current Thread [Next in Thread>