[TLS] Re: WG Last Call: draft-ietf-tls-mlkem-05 (Ends 2026-02-27)

[Paul Wouters <[e-mail]>]

On Wed, 25 Feb 2026, Stephen Farrell wrote:

Hi Stephen,

> On 25/02/2026 19:49, Paul Wouters wrote:
>>       The working group last call for pure ML-KEM has concluded, thanks to
>>       those that participated in the discussion. In summary, we do not have
>>       consensus to publish the document as is.
>>
>>       The largest number of participants wanted to publish the document as
>>       is, however there was also a significant number that wanted changes
>>       to the document before publication and a small, but vocal, number of
>>       participants that do not want the document to be published at all.
>>       There were several issues raised, but the main area of contention was
>>       around having a statement on the security and applicability of this
>>       mechanism versus the hybrid key mechanisms.
>>
>>       Given this, the chairs will move the document back to the "WG Document"
>>       state and ask the author to work on resolving the issues brought up on the
>>       list including text to address concerns that there are reasons to prefer
>>       hybrid over the pure approach. The chairs will then redo a working group
>>       last call to see if there is rough consensus for publishing this document.
>
> I'm still at a loss as to how the above doesn't map to the current
> WGLC being just a normal WGLC. And I don't see how that text maps
> to a WGLC limited to any specific aspects of the draft. Maybe there's
> some miscommunication somewhere but I certainly did not interpret
> the above as meaning some part(s) of the draft were considered to have
> passed the first WGLC.

It is a normal WGLC. And opponents and proponents can state their
view. However, as with the adoption call, the 1st WGLC also had a group
of people that were between "yes" and "no" that wanted to see changes
to the text before giving their final view on the document. This 2nd
WGLC started with the proposed changes as the diff between -05 and -07.

So there is a possible rough consensus if those people that wanted to read
new text can agree on the new text. But the current influx of messages
from people (repeatedly) saying yes and no and arguing about it, are
drowning out those people who would like to discuss the textual changes.
This also happened earlier in this 2nd WGLC, and when I sent a reminder to
let people talk about proposed text changes, it just got worse with the
repetitive no/yes voices, new -00 drafts, hypothetical charter discussions
and updating other documents in the RFC Editor queue.

So I am once again asking that if you shared your yes/no view during
this 2nd WGLC, to step back and give those people who wish to talk about
textual changes the chance to actually communicate over the list.

It seems we have some issues to get text for from what I read:

- Updated text in Abstract / Introduction giving a clear preference
   for hybrid.
- Changes in the Security Considerations Section with more technical
   details or references why hybrids are preferred where possible.
- Clarify on the key reuse text.
- Clarify text on regulatory requirements and other SDOs that have
   non-hybrid requirements (and check liaison statements )

I hope this answers your question,

Paul