On Thu, 2004-02-26 at 16:39 -0600, Andrew W.Donoho wrote:
This looks like a non-problem. The recipient is choosing to use anSPF
record without using a forwarding mechanism to protect his/her
ownmail. Is that really your problem?
No, it is the problem of the intended recipient who failed to receive a
valid mail, and the problem of my user, whose mail was not delivered.
Consider the following realistic situation:
The intended recipient did _not_ choose to use SPF; an admin at their
receiving site (or an intermediate site) made that choice.
The forwarding site did not _choose_ not to implement SRS; perhaps they
didn't realise that the flag day had come and the world needed to
upgrade, or perhaps they'd investigated and consulted their ISP, who
answered that to implement SRS would be considered network abuse and
lead to disconnection.
In that entirely realistic case, it may not be my _problem_ but it would
be my _fault_, because I chose to publish the SPF record.
Publishing an SPF record which doesn't end in '?all' at this time --
when no complete answer to the forwarding problem even exists, and
certainly hasn't been implemented widely -- would be like starting to
drive a new car before fitting the brakes.
Yes, I can declare that everyone else on the roads should fit big
airbags on their bumpers, and I can do so myself to make the point. And
I can go out driving like that in an attempt to help all those people
out there who are 'poorly using my scheme' to learn. They'd still lock
me up though.
In my very small community mail service, I cannot imaginethat one of
my users would prefer that I protect the ability of anexternal to our
community user's ability to poorly use SPF and acceptjoe-jobbing as
the consequence.
The people out there forwarding their mail have been doing so for years.
They are not 'poorly using SPF'. They are using SMTP, and may have been
doing so for 20 years or more. They may not even have heard of SPF,
until valid mail starts bouncing.
--
dwmw2