On Sat, Apr 19, 2014 at 08:47:37AM +1200, Brian E Carpenter wrote:
So, if the From says
From: goodguy(_at_)yahoo(_dot_)com <haha(_at_)badguy(_dot_)example(_dot_)com>
many UAs would show only goodguy(_at_)yahoo(_dot_)com as the sender,
but badguy could have passed DMARC, no?
This would not exactly enhance goodguy's reputation,
or Yahoo's for that matter. I realise it isn't the exploit
that Yahoo is trying to stop, but it suggests to me that
DMARC is only plugging one small hole in a very leaky dam.
Iif the problem is trying to protect goodguy or yahoo.com's
reputation, I wonder if a better approach would have been to have
yahoo.com issue all of its users S/MIME certificates, and then had a
DMARC-like policy requesting recipients: "if the e-mail has the From:
field of yahoo.com, and it's not an S/MIME-signed e-mail with a
yahoo.com certificate, reject the e-mail".
After all, we know S/MIME successfully passes through mailing lists,
and if in fact the message was appropriately signed using an S/MIME
cert, it would be quite natural to have the UA's display the
information from the Common Name field of the cert.
That would solve a host of problems, including the hand-wringing
around how S/MIME has lots of deployed users, but very few deployed
certs.
- Ted