As another data point on this, the following is a direct quote from the UK 
NCSC’s guidance on PQC migration:

> These algorithms will need to be implemented in protocols before they can be 
> used on the internet and in other networks. The IETF is in the process of 
> updating the most widely-used security protocols. This includes ensuring that 
> PQC algorithms can be incorporated into key exchange and signature mechanisms 
> in existing protocols such as TLS and IPsec. IETF implementations of 
> post-quantum protocols are subject to change until they are published as 
> RFCs. The NCSC strongly advises that operational systems should use protocol 
> implementations based on RFCs, not on Internet Drafts.

https://www.ncsc.gov.uk/whitepaper/next-steps-preparing-for-post-quantum-cryptography,
 (emphasis theirs)

I have a mild preference for not spending lots of time on every possible code 
point that people might require, but I do strongly feel that we need an “end 
state” for “this-draft-exists-just-for-a-code-point” that is better than 
“expired I-D”. Otherwise we’re not believable to external SDOs and regulators.

Regards,

Thom Wiggers


> Op 2 mrt 2026, om 18:13 heeft David Benjamin <[email protected]> het 
> volgende geschreven:
> 
> > That the endorsement of the IETF matters to so many continues to surprise 
> > me.  Not a whole lot, but the way that symbolic measures like this act as a 
> > substitute for good judgment is one of the great mysteries.
> 
> I'd echo Ben's comments earlier in the thread. While there are certainly 
> silly reasons, the reason Ben cites is more than symbolic. For someone 
> implementing and deploying something, an immutable codepoint isn't 
> sufficient. It's also important to know that the feature is "done" and that 
> the TLS community isn't going to next week define a new codepoint that flips 
> the order of the key shares or whatever. (E.g. the many iterations that led 
> to the final version of ECH.)
> 
> This is critical for the health of the TLS ecosystem. Sure, the big 
> deployments and early movers can move fast, ship experiments, unship them, 
> etc. But if we want TLS to be accessible to the broader internet, we need to 
> keep others in mind. Others are less able to unship things they once shipped. 
> But if every draft iteration of ECH had to survive in TLS libraries because 
> people can't unship them, we would never get anywhere. That means there's a 
> crucial milestone when a TLS feature is safe to ship in more durable 
> contexts, and we need to actually be able to get to those milestones.
> 
> For that to happen, for TLS to serve more than just the major players, the 
> TLS community needs a venue to coordinate on work, roughly agree on when this 
> milestone has been reached, and actually reach that milestone. Historically, 
> that venue has been the TLS WG.
> 
> Moreover, when the milestone is reached, there also needs to be a clear 
> indication of this. Those of us here, paying attention to the group and 
> sifting the through the atmosphere, know the status of things. But TLS 
> touches more people than this group. I work with many, many groups who 
> integrate TLS into their projects. They have the rest of their project to 
> worry about and can't follow the WG. But they need to know, when evaluating 
> X, whether X is ready or not for them. Historically, that indicator has been 
> document publication.
> 
> All this is certainly imperfect, even historically. (E.g. the long time 
> between something being "done" and a final RFC increases the window of 
> uncertainty for folks. And of course this venue can be a bit abrasive.) 
> Unfortunately, this WG has trended even more towards those imperfections, so 
> here we are. These kinds of cryptography drafts seem to be the extreme case, 
> where there is some coordination needed, but it is minimal in comparison to 
> the problems it triggers here.
> 
> Perhaps this WG needs to have fewer responsibilities, as this draft suggests, 
> before it can reverse this trend. But that won't change the need for the 
> coordination work somewhere, and I don't think we should be proud of 
> producing an atmosphere where we need to actively drive that work to other 
> venues.
> 
> On Mon, Mar 2, 2026 at 8:46 AM Salz, Rich <[email protected] 
> <mailto:[email protected]>> wrote:
>> - TLS Exporter Labels
>> - TLS SSLKEYLOGFILE Labels
>> - TLS Certificate Types
>> - TLS Application-Layer Protocol Negotiation (ALPN) Protocol IDs
>> - TLS Certificate Compression Algorithm IDs
>> 
>> Most of the entries for registrations in these tables have come from 
>> documents that were NOT adopted by the TLS WG.  Sometimes other WG drafts, 
>> sometimes (mainly ALPN) just using the IANA form.
>> _______________________________________________
>> TLS mailing list -- [email protected] <mailto:[email protected]>
>> To unsubscribe send an email to [email protected] 
>> <mailto:[email protected]>
> _______________________________________________
> TLS mailing list -- [email protected]
> To unsubscribe send an email to [email protected]

_______________________________________________
TLS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to