On Fri, 6 Aug 2004, Roger Moser wrote:
v=spf1 a/24 mx/24 ptr ?all
What exactly were you using? Sure, the recommended best guess might
not give 'pass' for all legitimate mail, but when it doesn't, the
'neutral' result should be treated the same as 'none'. So why
would you have to remove it?
I used "v=spf1 a/24 mx/24 ptr -all" because "?all" wouldn't have made any
sense (all mail would have been accepted).
It is a little early to refuse to accept mail sites with no SPF record.
When a site *does* have an SPF record, a 'pass' result either
bypasses content checking, or changes bayesian weights (automatic when
Received-SPF header is present), or otherwise has some effect.
Even if your SPF checker is only rejecting 'fail' results, you
should still use best guess and insert a Received-SPF header. If
nothing else, that header will be very nutritional bayesian fodder for your
end-users' spam folder in Mozilla/Thunderbird/Outlook.
Here are examples of the two possible best_guess results from my system:
Received-SPF: pass (spidey.bmsi.com: guessing domain of hylafax.org designates
216.152.199.7 as permitted sender) client-ip=216.152.199.7;
envelope-from=hylafax-devel-bounce(_at_)hylafax(_dot_)org;
helo=polaris.ifax.com;
and
Received-SPF: neutral (spidey.bmsi.com: guessing 207.46.104.72 is neither
permitted nor denied by domain of msn.com)
Neither one is rejected by SPF, but the 'pass' token in a Received-SPF header
is a significant bayes discriminator - especially if you keep a domain
blacklist.
--
Stuart D. Gathman <stuart(_at_)bmsi(_dot_)com>
Business Management Systems Inc. Phone: 703 591-0911 Fax: 703 591-6154
"Confutatis maledictis, flamis acribus addictis" - background song for
a Microsoft sponsored "Where do you want to go from here?" commercial.