On Sun, Oct 03, 2004 at 01:08:04PM -0700, James Couzens wrote:
|
| Yep! And you will see I'm doing just that. For the record, and for
| anyone here who isn't aware the SES proposal involves a callback,
| something that John Levine is very set against, unfortunately with no
| really solid argument. In fact, the only real downside to a UDP or TCP
| based callback operating as a standalone service on a different port is
| adoption, however I wager it would be easier to get people running this
| service uses as much as or less bandwidth than DNS, on a non-root port (
| > 1024), and adoption would probably kick ass over something as
| encumbered as DK.
|
if people want to avoid the mandatory callback, they can use
a public-key scheme rather than a single signing secret.
that means that we'd have to add a new modifier, eg.
ses.pubkey=dnstxt:_seskey.%{d}
which would mean that the pubkey for SES signing could be
found at _seskey.%{d}.
then a receiver could download the SES pubkey and verify
localparts without the callback.