Re: A couple of comments on the open issues...

--On Saturday, 15 February, 2003 20:41 +0100 Marc Mutz 
<[email protected]> wrote:

> On Saturday 15 February 2003 02:12, John C Klensin wrote:
> <snip>
>> There are no standards at all for interpreting the local part.
>> The standard is "no one but the delivery MTA can try to
>> interpret the local part in any way".
> <snip>
>
> MIXER (rfc 2156) specifies a "sub-syntax" for local-parts. The
> interpretation is done by the X.400<->rfc(2)822 gateways, not
> the final  delivery MTA.
>
> MIXER is standards-track.

Marc, we could get into an interesting, and quite lengthy, 
discussion of hairsplitting definitions here.  I recommend 
against it.  Just as Dan did not understand "interpreting" the 
way I intended it, you are oversimplifying the concept of a 
delivery MTA.  I perhaps should not have said "no one", or I 
should have made the (intended) contextual restriction to the 
Internet/SMTP mail transport environment more explicit, but I am 
(believe it or not) trying to keep these notes from running on 
forever.

The relevant text from 2821 is section 2.3.8, part of which says

	  A "gateway" SMTP system (usually referred to just as a
	  "gateway") receives mail from a client system in one
	  transport environment and transmits it to a server system
	  in another transport environment.  Differences in
	  protocols or message semantics between the transport
	  environments on either side of a gateway may require that
	  the gateway system perform transformations to the message
	  that are not permitted to SMTP relay systems.

That is an uncomfortable, but necessary, blanket exception for 
gateway systems to do anything at all that they need to do. 
Mixer is, at least, a gateway between an Internet/SMTP mail 
transport environment and an X.400 one, so it is covered by that 
exception.

If one _really_ wants  to split hairs, it probably doesn't even 
need the exemption: The assortment of BITNET, ccMAIL, FidoNet, 
Profs, UUCP mail, etc., etc., gateways we had floating around 
several years ago (I suspect that few, if any, of them have gone 
completely away) were typically defined as SMTP<->foo gateways, 
regardless of what they did once they got hold of the mail. 
But MIXER is, as you say, defined largely as an [2]822 <-> X.400 
gateway.  As such, it is acting either before the first 
("originating") SMTP MTA gets hold of the message, or it is 
acting after handoff from an MTA at the receiving end.   And the 
MTA that hands things off to an MUA or a message store so that 
the 822 processing becomes relevant _is_ a "delivery MTA".

On a side note, I think I can assure you that the much-abused 
editor of 2821 was quite aware of the gateway situation 
generally and MIXED, and RFC1327 and other predecessors, when 
that section was written, as was most of the WG, the relevant 
ADs, etc.  You might even say "painfully aware".  If the text 
isn't clear enough, I'd welcome suggestions for clarification -- 
the list of clarification suggestions I still have to retrofit 
before posting the first 2821bis I-D is shrinking, and I'd much 
rather deal with suggestions now then later.   (A note to anyone 
else reading this: messages suggesting changes that will make 
the text more clear, especially when they propose specific text, 
are being incorporated to the degree possible.  Messages 
explaining that 2821, or SMTP generally, are broken, or 
insisting that it should work differently, or claiming that the 
2821 text didn't actually represent WG consensus, are being 
filed for passing off to the ADs and former WG Chair, whom I 
intend to let make the decisions as to which of the issues are 
worth [re]opening.  But my filing systems are not good, and some 
of those messages may get lost (or might have been lost in one 
of a series of disk crashes last year) or not noticed when the 
recipients go through the pile, so I would really suggest 
holding them until the AD's and Chair are ready for them -- or 
sending them to them, not me. :-(

     john