>>>>> "PW" == Paul Wouters <[email protected]> writes:

PW> John brought up three issues.

PW> 1) Lowwercase might fail for non-ascii 2) mailbox localpart guessing
PW> not allowed
PW> 3) mailbox/local part of different case might be different person

But as was mentioned OpenPGP is relevant, not SMTP.  And given that the
most common OpenPGP and keyserver – and probably all – implementations
are already caseless there is nothing wrong with following their lead.

This isn't just about email.  A large percentage of the sigs I verify
are obtained though methods unrelated to smtp.  Detatched sigs are quite
common, and OpenPGP is or can be used when sending short messages, such
as via xmpp, sip or even ss7.

Using dnssec as an additional – or in many cases the only – trust path
when verifying a sig can only improve things.  It is not a perfect trust
path of course, but it is better than nothing.

Given all of that, taking (base32 (utf8 (lc (localpart address)))) and
splitting it by – working with presentation format — putting a . every
N chars from the right seems like the ideal plan.  Unless reversibility
is undesirable.

I've written otherwise in the past, but only because I forgot that the
OpenPGP world is caseless.  And I did so because I only ever use and
only remember seeing miniscule addresses when dealing with OpenPGP.

-JimC
-- 
James Cloos <[email protected]>         OpenPGP: 0x997A9F17ED7DAEA6


_______________________________________________
dane mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dane

Reply via email to