Re: dane-openpgp 2nd LC resolution

[John C Klensin <[e-mail]>]

--On Sunday, March 13, 2016 1:48 PM -0400 Viktor Dukhovni
<[e-mail]> wrote:

>> On Mar 13, 2016, at 1:11 PM, John Levine <[e-mail]>
>> wrote:
>> 
>>> Given that the DNS RR in question is something the end user
>>> has to  explicitly request, ...
>> 
>> Uh, what?  The DNS is under control of the domain owner, not
>> the end users. 
> 
> A misreading of the comment.  The "end-user" in question is
> the one doing the lookup, not the one whose key is published.
> Paul is making no claim about how the published key got
> there...

I understood that, and I assume John L. did too.  The problem,
again, is that we are conflating several issues, including
whether the right key is going to be found to correspond to a
given address and whether and how it can be trusted.  A
problematic domain owner (and, unless the nominal domain owner
is paying a lot of attention, a problematic registrar or other
third-party domain administrator) can provide bogus,
self-serving keys.

"Making no claim about how the key got there" is almost
certainty true, but that misses the point.  The document more or
less claims that, if one finds a key in the DNS associated with
a particular mailbox string, then that key has some association
with the person who owns/controls (not necessarily the same
thing) that mailbox.  -07 was actually more clear about the
issues with that than -08 is, but neither goes far enough, IMO,
in detailing the risks that the community perfectly well knows
about.

The requirement is still that the I-D be clear about either
known risks, restricting the experiment to those who are very
familiar with those risks and accept them, or both.

    john