John Glube wrote:
Seth,
You write:
"That's really not so bad. Anyone who wants to use SPF
classic will publish a record with v=spf1. Most people
have already done that and there are several free,
unencumbered, open software solutions for them to chose
from. Anyone who wants to use PRA/SUBMITTER, yes, that
means both of you, will publish a record with v=marid1, and
don't forget to apply for your Microsoft license if you
want to touch the code. I think choice is an excellent
thing."
Maybe I missed the boat here, but it is my understanding if
this change in version string goes through the next draft
of the marid protocol will tell folks wanting to publish a
sender policy record:
* you must use v=marid1.
Senders will not have the option of choosing:
* use v=spf1 if you want mail from checked, or v=marid1 if
you want sender from checked.
Hmm. Perhaps we can be more subtle: unless the CallerID system is very
tightly managed and authenticated at the time of purchase of the keys, I
expect spammers to be using it extensively. I'm also concerned that CallerID
for individual hosts will be purchased and used to overwhelm domain level
SPF policies. Preventing that is a huge administrative load for the key
managers.
Moreover, I'd prefer to have the option of enabling or disabling either
CallerID or SPF Classic at the recipient's local MTA, not in the sender
policy.