RE: [rfc-i] I-D expiry [was Re: RFCs vs Standards]

[Roman Danyliw <rdd@xxxxxxxx>]

Hi!

 

A new WG to revisit/clarify IANA Considerations, IANAPLUS (https://datatracker.ietf.org/wg/ianabis/about/), is close to final approval– it is through the IESG’s ballot on “do we approve this charter”.  The charter (https://datatracker.ietf.org/doc/charter-ietf-ianabis/00-05/) has the following activity in-scope related to the question of “can an I-D to satisfy a the ‘Specification Required’ requirement”:

 

==[ snip from charter ]==

A process for using the Internet-Drafts system to create permanent references.

==[ snip ]==

 

Regards,

Roman

 

From: Eliot Lear <lear@xxxxxxx>
Sent: Monday, December 9, 2024 2:56 PM
To: Brian E Carpenter <brian.e.carpenter@xxxxxxxxx>; touch@xxxxxxxxxxxxxx
Cc: Christian Huitema <huitema@xxxxxxxxxxx>; Randy Bush <randy@xxxxxxx>; Michael Richardson <mcr+ietf@xxxxxxxxxxxx>; IETF discussion list <ietf@xxxxxxxx>
Subject: Re: [rfc-i] I-D expiry [was Re: RFCs vs Standards]

 

Yes. and now we are repeating the SAAG discussion in its entirety.

Eliot

On 09.12.2024 20:32, Brian E Carpenter wrote:

On 10-Dec-24 05:27, Eliot Lear wrote:

But it is, Joe.  The proof is that the ISE gets requests all the time from people saying that the need the RFC # in order to gain customer adoption.  That doesn't mean don't EVER get deployed but at least then we've done our due diligence.

This does, however, suggests by analogy that allowing an I-D to satisfy a "Specification Required" IANA assignment policy is not acceptable. BCP 26 (RFC 8126) requires "a permanent and readily available public specification" which also implies due diligence. Every I-D states clearly that "It is inappropriate to use Internet-Drafts as reference material" which seems to make things very clear, with or without a mythical 6 months expiry.

   Brian

Eliot

On 09.12.2024 17:17, touch@xxxxxxxxxxxxxx wrote:

On Dec 9, 2024, at 7:20 AM, Eliot Lear <lear@xxxxxxx> wrote:

What I care about is this:

The IETF community (and those generating IRTF and independent submissions) need a way to signal to the community that draft means just that: it's draft work, and not intended for broad deployment.  Otherwise, we end up with all of the support issues I mentioned earlier.

Eliot

There’s no way to avoid that in any public message even it it is stated explicitly

Lots of things get deployed even when they aren’t in such messages.

I.e., that doesn’t seem avoidable.

Joe