william(at)elan.net wrote:
trying to overshoot is not necessarily bad.
At least better than hitting your own foot.
The protocol may work for some but not for others.
No, it's certainly not a pseudo-random generator.
It just works as designed.
Of course it doesn't work if you fail to do the
right things. You cannot test random identities
at random places, SPF is an awful PRG.
forwarding being fairly common for senders
Senders will either list all their mailouts, or
make vague ?all statements. Not recommended,
but allowed.
chances of it not working for those domains are
good and that is why they are choosing "~all".
Sooner or later they'll find all their mailouts.
It can't be that difficult. The difficulty is to
educate users that they can't abuse the MAIL FROM
over foreign routes as some kind of "BOUNCES TO".
Fortunately ordinary users just configure their
MUA, and then it works automagically as designed.
Ordinary users have no idea what a MAIL FROM is,
therefore they also don't abuse it intentionally.
[removal of -all recommendation]
That is reasonable thought and would like to see
discussed further on this list.
That's not reasonable. It's like removing all odd
numbers from the primes. What you end up is 2, and
then all efforts to generate other primes is wasted.
one possibility s to replace "NOT RECOMMENDED" with
"MAY" use PASS result only for checking other
identities but MUST NOT use FAIL result.
This is madness. We know that checkhost( MAIL FROM )
is different from checkhost( PRA ). Like cubes are
different from squares. It's "possible" to say that
1*1 = 1*1*1 and 0*0 = 0*0*0, and therefore cubes and
squares are not always different, but that's neither
a "possible compromise" nor "engineering".
It's madness. An incorrect PASS is not really better
than an incorrect FAIL. An MSA enforcing submission
rights (2476bis 6.1) would allow me to create bogus
PASS results for all other users allowed to use this
MSA, all I have to do is to forge their PRA.
Madness, complete technical madness. And if the IESG
would be that stupid they don't deserve the right to
live, let alone review proposed standards.
allowing use of positive result for whitelisting.
It would allow phishing with bogus PASS results. It
would be far worse than no SPF at all. Using PRA on
v=spf1 is criminal abuse, no matter how it's done.
The only "possible way" would be explicit consent, as
in theory "v=spf1 op=pra" or "spf2.0/mfrom,pra" resp.
Anything else is madness, bogus, abuse. Pick all.
I'm personally against SIDF going even for experimental
status, simply put it is incompatible with existing
standards and this should be enough to squash it
Then don't propose weird ways to apply FUBAR on v=spf1.
And a FUBAR PASS is not less harmful than a FUBAR FAIL.
Quite the contrary, a FUBAR PASS could be a successful
phish. The damage is limited only by your worst case
fantasy. This is _NOT_ a possible compromise.
We're not on some oriental bazaar, it's about the IESG
and _engineering_ No SPF RfC would be much better for
the net at large than this foul compromise, Obviously
our ethical standards are different.
Bye, Frank