[ietf-dkim] Re: Requirements comment: Bigbank example description

[william(at)elan.net]

On Thu, 10 Aug 2006, Hector Santos wrote:

> ----- Original Message -----
> From: "william(at)elan.net" <william at elan.net>
> To: "Hector Santos" <hsantos at santronics.com>
> Cc: "Scott Kitterman" <ietf-dkim at kitterman.com>; <ietf-
>
>>> What's wrong with checking each one?  I mean, why
>>> allow for a loophole?
>>
>> Wasn't one of the requirements finite number of queries?
>
> As finite as required I guess.
>
>> If not then I'd like to send email with 100 different
>> addresses in From but all within different domains on
>> your dns server to 10,000 random people. Want to
>> guess how many requests you'll receive?
>
> Maybe there is short circuit?
>
> Maybe I won't even bother with such nonsense multi-address from lines, which
> probably is going to break along some down stream anyway.  In all honesty,
> it is rare to encounter this, IMV.

Read my previous message. It maybe rare now but what you ended up
doing is moving 3rd party senders to where they would change From
and add their name into it as 1st address and move or put what 
would have been or was original From address as 2nd one, basically
they would be emulating current use of "Sender" (in fact they
would still use Sender as required by RFC2822 in cases of
multi-address From). I'm not saying its all bad, but its a
change to current email system practices.

BTW - In above scenario I'm actually more concerned with situations
where the intermediate party would >>CHANGE<< From header field -
same parties that currently add Sender. First of all I dont want
that happening to my emails no matter who they go through and
2nd it will break existing signature if it existed.

-- 
William Leibzon
Elan Networks
william at elan.net