ietf-mxcomp
[Top] [All Lists]

RE: A proposal on identities

2004-04-21 06:43:26



-----Original Message-----
From: owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org
[mailto:owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org]On Behalf Of Arnt 
Gulbrandsen

wayne(_at_)midwestcs(_dot_)com writes:
Really, there two important considerations:

1) Are we breaking existing practices? If so, is it worth it?

The goal of this thing is to break existing practice, 
specifically the 
existing practice of sending mail in someone else's name.

I don't think we break anything.

If we do not achieve a clean zero false-fail test then 
messages that fail the test get put into the spam filter, 
the same way they do today.

What I am interested in here is that we have a zero
false-pass test. If a message passes the DNS authentication
test at RFC 821 or 822 levels it is definitively from the
claimed domain.

We do not necessarily achieve everything in MARID alone
I think it very likely that we end up with:

MARID positive authentication = 100% accurate pass
MARID negative authentication = 90% accurate fail

MARID + NYI = 100% accurate fail

Where NYI is some dohickey that we devise to address the 
forwarding issue, possibly SRS, possibly some other scheme.

2) Will the chairs, or IETF editors, or whoever, let us 
publish an RFC 
without complaining that there is a conflict?

The Area Director is the big bad wolf. The AD is likely to 
ask questions 
like "is this incompatibility a necessary part of MARID's goal, or is 
it some accidental breakage?"

I don't think that what we put in the MARID record has the
slightest bearing here, all we can tell people to do here is
to tell people to list their outgoing edge MTAs.

There is absolutely nothing we can do to alter the pass / fail 
numbers given above they are the consequences of the mail
infrastructure configuration. the only way they can be changed
is if the people operating those infrastructures change their
configurations so that they achieve 100% accurate pass.


What we should probably do here is to write a note that describes
the consequences of a MARID record for each of the Andy
specified identities.

It would also be useful to consider some use cases, here I mean
'forwarding' 'mailing list' and 'postcard site' and for each
of these edge cases suggest a way to get to golden status.


                Phill


<Prev in Thread] Current Thread [Next in Thread>