Re: [apps-discuss] The SMTP 521 reply code, RFC 1846, and nullMX

["Rolf E. Sonneveld" <[e-mail]>]

John,

On 03/02/2015 03:30 PM, John C Klensin wrote:
> Hi.
>
> There was a lot of activity around this in the middle of last
> year, largely because of the nullMX draft
> (draft-ietf-appsawg-nullmx).  According to the tracker, nullMX
> has been sitting in the RFC Editor queue since mid-September, in
> part because it contains a normative reference to
> draft-klensin-smtp-521code.  I think the assumption was that a
> quick Last Call was going to be done on the latter I-D, possibly
> through appsawg, but nothing has ever happened.
>
> According to the Secretariat's automatic system,
> draft-klensin-smtp-521code-02.txt expires next week.
>
> I'm happy to update and repost it if there is a plan to move it
> forward.  Or I can let it expire, thereby either killing the
> nullMX draft or requiring that it be reissued with a different
> strategy and presumably a new Last Call.

Two minor nits: first sentence of Abstract: code should be plural:

s/code/codes

and, section 3:

s/Server/server

In section 3:

> One additional case is
>     covered in the next section.

it is not clear for me:

a) to which this case is referring (code 554 or 521)
b) to which section this is referring.

General question:

Excuse me if this has been discussed before, it is not my intention to 
propose significant changes right before publication. What I would like 
to ask is: is there a specific reason that some paragraphs talk about 
'client' and 'server' where other paragraphs talk about 'system' and 
'host'? I know these terms have different meanings, but wouldn't it be 
more consistent to use one 'set' throughout the document as much as 
possible?

If (I repeat: if) there is no specific reason, I'd suggest to use the 
words 'client' and 'server'. Section 3, for example:

>     This specification adds the 521 reply code to the repertoire
>     specified in SMTP, reserving it for use at connection-opening time to
>     indicate that the host does not accept email under any circumstances.
>     It SHOULD be used for dummy SMTP servers whose sole purpose is to
>     notify systems that attempt to open mail connections that the host
>     never accepts mail.  It MAY be used in other situations where the
>     intent is for the host to indicate that it never accepts email.

would then become:

    This specification adds the 521 reply code to the repertoire
    specified in SMTP, reserving it for use at connection-opening time to
    indicate that the server does not accept email under any circumstances.
    It SHOULD be used for dummy SMTP servers whose sole purpose is to
    notify clients that attempt to open mail connections that the server
    never accepts mail.  It MAY be used in other situations where the
    intent is for the server to indicate that it never accepts email.


and the proposed 'acceptable' message for code 521 would become:

"Server does not accept mail"

/rolf