Re: [Rfced-future] Comments on draft-iab-rfcefdp-rfced-model-12

[Benjamin Kaduk <[e-mail]>]

On Sat, Mar 12, 2022 at 12:07:22PM -0700, Peter Saint-Andre wrote:
> On 3/11/22 7:19 PM, John C Klensin wrote:
> > 
> > 
> > --On Thursday, March 10, 2022 13:40 -0800 Benjamin Kaduk
> > <[e-mail]> wrote:
> > 
> >>>> "This document requires that the RPC document registry value
> >>>> assignments made by IANA."
> >>>
> >>> That's pretty much what it said before, no? ;-)
> >>>
> >>> I suggest this in the "RPC Responsibilities" section:
> >>>
> >>> 14. Ensuring that RFCs accurately document registry value
> >>> assignments made by IANA.
> >>>
> >>> For the avoidance of doubt, we could also say the same thing
> >>> under the  IANA considerations.
> >>
> >> That does remove the bits I was confused about, but to me it
> >> also seems to change the semantics somewhat.  Namely, now the
> >> RPC is just consuming things produced by IANA, which could be
> >> seen as removing the possibility to coordinate on which
> >> allocations are actually to be made, from what range(s), etc.,
> >> that the previous text seems to have implied.  I think I have
> >> seen the RPC notice things in editing that would affect what
> >> IANA does, and thus am not confident that describing this as a
> >> unidirectional flow would be entirely accurate.  (Whether such
> >> coordination could occur between RPC and IANA in an informal
> >> manner so as to get the right thing to happen anyway, is
> >> another question.)
> > 
> > In practice, it has always been a two-way flow, even when the
> > RFC Editor and IANA held discussions by looking into a mirror.
> > And even with that level of coordination, it has often been a
> > bit of a dance because of the long-standing principle that RFCs
> > do not direct IANA to assign, or even request, particular code
> > points.  So please try to write the text along the lines Ben
> > suggests, e.g., as something more like "Coordinate with IANA as
> > necessary to ensure that any registry value assignments that
> > actually appear RFCs are consistent with the registries."
> 
> What I suggested most recently was:
> 
> 14. Coordinating with IANA to ensure that RFCs accurately document
>      registration processes and assigned values for IANA registries.
> 
> Note that this mentions registration processes in addition to assigned 
> values, because many RFCs establish registries and the processes for 
> registering values in those registries need to be correct, too.

Mentioning registration procedures as well is a good catch!
I know that I am behind on the thread, but since I hadn't seen this
specific proposal yet either (like John), let me note here that it looks
good to me.

Thanks,

Ben