From: Hector Santos
Sent: Wednesday, July 07, 2004 4:52 AM
From: "Seth Goodman" <sethg(_at_)GoodmanAssociates(_dot_)com>
Sent: Tuesday, July 06, 2004 2:32 PM
<...>
My main point is basically once a level of satisfaction is
reached - that's
it!! No Mas! SPF will not cover all issues. Implementators
are going to
do what it takes, which includes augment SPF whether its MARID variety or
not with other technologies as well that currently exist or in place.
Well, I'm glad you're satisfied with SPF and want to freeze it now, but that
is far from a universal opinion. I don't think it accomplishes the basic
necessities yet so IMO, we need to keep going.
My recommendation is:
Slow down, stabilize it, stop "scaring people" and work out the key issues
which is basically two as I see it:
- Mail forwarding
- HELO validation using a persistent domain concept.
In my view, the last item will remove CSV from consideration in my book.
I don't agree. This is an ideal time to get important changes in. With the
present SPF, we can only validate the return-path on the first hop. That's
just not good enough. Handling mail forwarding by SUBMITTER is not a
terrific solution, IMHO. We're not ready for prime time, yet, even though
what you've got is good enough for your particular product.
I am not suggesting that progress should stop. But if you are going to
continue with new complex logics, please use a "v=SPF2" record
so that they
is no ambiguity. V=SPF2 implementators will make sure it is backward
compatible, but without a doubt, it should be a requirement by trying to
force it into SPF1.
Reasonable people can disagree whether changes should be SPF1 or SPF2.
There are good arguments on both sides. As SPF1 is not yet frozen, in my
mind at least, I'd prefer to see an changes as part of SPF1. It seems that
we'd be shooting ourselves in the foot by coming out of the chute with two
versions of the language and parser, before most people even published
records. Other people will disagree.
--
Seth Goodman