Mark Lentczner wrote:
example.com. IN TXT "v=spf1 redirect=_spf1.%{d}"
example.com. IN TXT "spf2.0/pra redirect=_spf2.%{d}"
_spf1.example.com. IN TXT "v=spf1 ... some very long [...]"
_spf2.example.com. IN TXT "spf2.0/pra some different [...]"
Nice idea.
[wildcards]
See the protocol-03 draft for an example and discussion of
just this point.
Where can I get protocol-03 ?
3-Summary)
For sites where records for Sender-ID and SPF are large,
similar though not the same, none of these methods is a win
over the method in (1).
True.
For sites where the records are the same (claimed by many to
be the vast majority of sites), (3a) is the biggest win.
Looking at your summary (3b) could be more flexible than (3a),
but (3a) is probably clearer. At the moment "SPF classic" and
"SenderId / PRA" still use almost the same "protocol" (= format
of the sender policy), although I'm not sure what results like
UNKNOWN or SOFTFAIL _mean_ in conjuction with "SenderId / PRA".
In fact I don't understand SOFTFAIL at all. But assuming that
this is only my problem it would be nice to have _one_ format
for the three kinds of sender policies:
"SPF classic" (mailfrom) only, "SenderId / PRA" only, and both.
Either (3a) or (3b) offer this feature. And in theory it's
still possible to fix the PRA algorithm for compatibility with
RfC 2476 (when an MSA "forgot" to add a Sender), or when a
mailing list uses Errors-To instead of Sender.
Bye, Frank