Re: [saag] SSH & Ntruprime

[Eric Rescorla <[e-mail]>]

On Mon, Mar 25, 2024 at 6:53 AM Eliot Lear <[e-mail]> wrote:

>
> On 25.03.2024 14:46, Eric Rescorla wrote:
> > [Citation needed].
>
> No.  No citation needed because it happens during the development
> process of a draft all the time, and you should know this because you've
> written enough code to drafts. The point is, you don't know when a
> draft is "finished".  The benefit of an RFC or something similar is that
> it is a signal that indeed the spec won't change.  We make an exception
> for early allocation for those drafts we know are going to become RFCs.
>

You seem to be confusing two cases:

1. The IETF is working on a draft that will eventually become an RFC.
2. Someone else wants to register a code point.

In the former case, we don't register until the specification is stable,
but in the latter case, the draft is used to document the code point
for assignment.

Obviously, you can't register code points against a specification which
might, but once the code point is assigned, you can simply not change the
draft, just as you don't change the RFC when published. If you're really
concerned about this, however, we could simply register against a
specific draft version, which is every bit as stable as an RFC.


-Ekr




> Eliot
>
>