Re: [Asrg] MTA registration means Path registration

[Dave Crocker <dhc at dcrocker.net>]

Alan,

AD>   My thesis is that we have to re-examine that sequence of MTA's, the
AD> roles that they play, and our terminology used to describe those
AD> roles.

ok. however, i have not seen any document that attempts to do that. nor
will such an effort produce anything useful for a very long time.


AD>   If we say "mail traverses multiple MTA's, and therefore each
AD> MTA does relaying", then we are avoiding the exact topic I wish to
AD> discuss.

it does not 'avoid' anything.  it describes reality.

you want to change the nature of the infrastructure. you want to
redefine established terminology. those are understandable and valid
goals. whether they are practical is a separate matter.

in any event, it it does not warrant calling an infrastructure with a
30-year history 'broken'.


>> Since RFC2821.MailFrom derives from the author of the message, this
>> means that the author needs to have a registration that authorizes each
>> sending SMTP client along the path.
AD>   If we ignore the fact those MTA's play *different* roles in handling
AD> the traffic, yes.

when there is a document describing an new architecture for Internet
mail, it will be possible to explore your model.


AD>   If they really do play different roles, then SPF applies only in
AD> certain situations.

i do not recall seeing anything in the spf documentation that specifies
restricted application.  specifications need to be precise both about
the details of the mechanism and the scope of its application.


AD>   The confusion arises because you're trying to treat each MTA as
AD> being identical.  They're not.

there is nothing in the current Internet mail service that supports your
view, technically. there is nothing in the spf specification that
details such differentiation.


>> AD>   My statement was true, even in your re-definition of SPF.
>> 
>> I did not redefine SPF.  Since you think I did, please tell me what
>> statement I made that deviated from SPF semantics.

AD>   You seemed to be saying that SPF validates sender/recipient pairs.

I said that it validates the path between sender and recipients, and
that there is a different path from the sender to each new recipient.


AD> It doesn't.  It permits recipients to validate senders.

sorry, no.

the mta registration are created by the sender.  that means the sender
is doing the validating.


AD>   While I agree
AD> that for each sender there is a recipient, saying that is *very*
AD> different than saying SPF validates sender/recipient pairs.

for this topic, saying that it validates every path is not all that much
different from saying it validates every combination of end-points.

d/
--
 Dave Crocker <mailto:dcrocker at brandenburg.com>
 Brandenburg InternetWorking <http://www.brandenburg.com>
 Sunnyvale, CA  USA <tel:+1.408.246.8253>, <fax:+1.866.358.5301>


_______________________________________________
Asrg mailing list
Asrg at ietf.org
https://www1.ietf.org/mailman/listinfo/asrg