On Thu, 18 Aug 2005, johnp wrote:
You're reinventing mailfrom/sender "equivalency" I talked about year ago,
see http://archives.listbox.com/spf-discuss(_at_)v2(_dot_)listbox(_dot_)com/200410/0710.html
and follow comments on that thread.
I've worked further on it privately but has not done a draft because I
believe it would only be good if we actually seriously start working on
SPFng (spf3 I presume) and it has not happened.
http://spfhelp.net/fix-1.php for a similar idea I came up with about the same
time as William came up with his. One consolation is that the concept is now
probably unpatent-able ;-) With refining - it is the obvious way to make SPF
*much* more usable.
I don't have time to fully go through your page right now, but at first
glance, what you wrote does not look same as I what I have or what Seth
is talking about is looks more like SID then anything else (i.e. look
at one field, then another, then another) and suffers from the same
problem of assuming that all identities are equal and there is just one
SPF "pass" or "fail" (which I previous thought was good but after further
look at the problem, I now think its only of limited use). I believe
equivalency modifier or new scope should serve specifically RFC2822 sender
identity but as far as making SPF *much* more usable, I'm not sure that
is quite correct - it would make SPF usable for certain situations
where it does not serve well right now (i.e. anti-phishing) but we should
not equate this with some kind of FUSSP.
--
William Leibzon
Elan Networks
william(_at_)elan(_dot_)net