Re: [TLS] Publication of draft-rhrd-tls-tls13-visibility-00

[Stephen Farrell <[e-mail]>]

On 25/10/17 23:58, Peter Bowen wrote:
> So, to be completely clear, no one is arguing that Nick's three
> options (quoted below) are wrong or do not work.  The objection is
> that the IETF should not be publishing a RFC that documents them, is
> that right?

No, it's not that simple.

For myself, I disagree with some aspect's of Nick's analysis, which
we can go into as needed. I don't think that's needed in this mail.

On your second point...

The IETF has a range of policies (BCPs etc) that call for use of
strong and not-weakened cryptography in protocols. Some of that
is mentioned in places at [1] but I'm sure a more complete job
could be done. There are sound technical reasons why the IETF has
consensus on a bunch of those positions. For some of those, it
is also true that the consensus has always been rough, e.g. I
think it's true there have always been IETFers who would actually
like to MitM security protocols for what they consider good
reasons. (That doesn't make those people either bad or correct.)

One particular relevant RFC (2804) does explicitly envisage that
people who want to do snooping might document their ways of doing
that in independent stream RFCs (which are not IETF RFCs despite
almost no RFC-readers grokking any difference there;-). But in
this case the authors say they want standards-track. And in the
previous case, from talking with Russ, he didn't see any benefit
in the independent stream route (which I didn't understand at the
time tbh).

So the situation is actually sort-of clear, but not simple. It's
true that appreciating the clarity requires quite a bit of IETF
lore ;-)

S.

[1] https://github.com/sftcd/tinfoil