Re: [saag] IETF 93 Agenda Request - Key Discovery

["John Levine" <[e-mail]>]

>> I don't know why you think MXs or MTAs have any role in this at all.
>> The major point of this stuff is to look up keys for e2e, ...

Because MTAs are the only ones that know what account goes with what
addresses.  (Remember that when you log in for SUBMIT, your credentials
as often as not are not your address.)  

As I said yesterday in a related thread, here's my usage model:

I set up my online access at Bigbank.  One of the things I do is to
give them the addresss [e-mail].  The bank looks up my
PGP key and sends all subsequent mail encrypted to my key, which in
this case happens to be the same one as for [e-mail].

My MTA already knows where it routes that address and can find the
appropriate account and the corresponding keys.  Webfinger or anything
else would either need a side channel into the MTA or try to match the
local address resolution logic.  Ugh.

R's,
John

PS: To summarize some lengthy discussions in DANE, the response to
any suggestion starting "You can just canonicalize the address by ..."
is "No, that may work on your MTA, but it doesn't work in general."