Re: Re: change of version string

On Aug 5, 2004, at 3:55 PM, John Glube wrote:
> You indicated in response to a question from Frank that the
> marid protocol will be set up to allow for checks involving
> either SMTP mail from or PRA.
I wasn't clear.  The MARID Working Group is proceeding with a set of 
drafts that only define checking the PRA.  The group (as a whole) has 
never advanced drafts checking MAIL FROM.  My examples were just 
showing that the change in version string allows one to continue to 
check MAIL FROM using "v=spf1" records.  The MARID drafts will not 
mention "v=spf1" at all, nor checking MAIL FROM.
> I thought EHELO was included in previous versions. Why has
> this been dropped?
There has been some group motion toward a set of drafts that check 
HELO/EHLO.  These drafts are called CSV and have no relation to the 
original SPF work or the current Sender ID work.  In particular, they 
do not use the same record format for publishing a domains 
authorizations.
> Is there any intent, consideration or thought being given
> to adding, "want to test for ehelo go here, want to test
> for sender from go here, do you like unified, consider
> this?"
If this is the where you think the world should be, then you shouldn't 
be pushing for ANY draft to come out of the working group.  The normal 
IETF practice would have been to: Wait until several methods of 
interest have been implemented and deployed for a year or more.  Let 
domains learn about them.  Let them decided what to implement and 
publish.  Wait some more time and gather real data on both public 
interest and operational impact.  THEN form a working group to pick 
only that which really works and hammer it all down into a spec.
In other words, if you think we should be just letting everyone try 
anything they'd like, then we don't need an IETF spec for that.  Just 
develop ideas, write code, disseminate, evangelize, win hearts, minds 
and hosts.
However, many people felt that spam had reached such a critical state, 
and that something like SPF was such a clear win if we could promote 
quick adoption.  Getting it approved as an IETF spec could help do 
that, and so the working group was convened with the specific charter 
to hammer it all out very quickly, without the normal multi-year lets 
see how they all work processes.
> I ask because some service providers are going to want to
> run checks using SPF, Sender ID, CSV and when ready
> DomainKeys.
There is nothing stopping those checks from being performed 
concurrently.  Nor anything stopping a domain from participating in all 
four schemes.  However, the working group as a whole did not show any 
serious interest in a "unified" approach, such as put forth by Meng and 
I.  It isn't clear to the group as a whole that the "correct" solution 
involves all four schemes and "pick and choose" approach for publishers 
and checkers.  Advancing the "unified" approach essentially enshrines 
such a path as standard practice.
        - Mark