[saag] Re: New Version Notification for draft-rsalz-crypto-registries-00.txt

[Simon Josefsson <[e-mail]>]

I don’t understand this line of reasoning - if you follow the tangent, then what’s the point of any standardization work at all? By the same argument why doesn’t just $browser define whatever TLS features they want, register them with IANA, and publish some obscure document with (some of) the details? There is no need for any IETF documents at all in that model beyond an IANA registry. Is that a good model? 

I believe the IETF should publish stable reviewed specifications of protocols as RFCs to have good interoperable descriptions of what is deployed out there, as a general service to all Internet users.

/Simon

> 27 nov. 2024 kl. 21:37 skrev Tero Kivinen <[e-mail]>:
> 
> Damien Miller writes:
>>> On Sat, 16 Nov 2024, Tero Kivinen wrote:
>>> 
>>> For the SSH there is already first come first serve policy (all names
>>> that contained @) but for some reason that was not enough. And
>>> implementations and interoperability works fine even with those first
>>> come first served names. There is no reason to require strings without
>>> @-chracter for algorithms, both forms work fine in the
>>> implementations.
>> 
>> The reason that it's not enough is that agreeing on names alone is
>> insufficient for reliable interoperability. At some point someone
>> needs to write down a specification and, at that point, it makes
>> sense to assign a standard name too.
> 
> For long time the ssh-1 protocol was described in the file called RFC
> inside the source distribution package.
> 
> When someone (like openssh) creates a new thing with [e-mail],
> they should include the documentation of that either on their web
> page, or inside the source distribution etc. Not everything needs to
> be an RFC, and most of those things that are defined that way are
> quite simple to document.
> 
> Yes, some of those are more complicated and should most likely be
> document in the RFC, but new ciphers, macs etc should not be that
> complicated. The SSH RFCs do not use that much text to describe them
> anyways.
> 
> We could add IANA registry with lists commonly used @-signed names and
> provides a pointer to the documentation for them (or contact point if
> no public documentation is available).
> 
> On the other hand, as the IANA registries have been changed to have
> different allocation policy, even that is no longer needed.
> --
> [e-mail]