[spring] Re: New Version Notification for draft-ietf-spring-srv6-yang-base-01.txt

["Kamran Raza (skraza)" <[e-mail]>]

Hello Alvaro,

Please see my responses below tagged as [skraza2]

From: Alvaro Retana <[e-mail]>
Date: Thursday, May 7, 2026 at 7:46 PM
To: Kamran Raza (skraza) <[e-mail]>; spring Chairs <[e-mail]>
Cc: Pingping Yu <[e-mail]>; Xufeng Liu <[e-mail]>; Satoru Matsushima <[e-mail]>; [e-mail] <[e-mail]>; SPRING WG List <[e-mail]>; Jaganbabu Rajamanickam (jrajaman) <[e-mail]>
Subject: Re: New Version Notification for draft-ietf-spring-srv6-yang-base-01.txt

On May 6, 2026 at 8:57:37 PM, Kamran Raza wrote:

> > 2. Several of the behaviors use a person (Pablo) as a reference. The
> > reference must be the draft in which the behavior is defined.
>
> [skraza]: I used the IANA SRv6 endpoint registry
> (https://www.iana.org/assignments/segment-routing/segment-routing.xhtml#srv6-endpoint-behaviors)
> to align our model and references.

Right. Use the Reference column, not the Change Controller.


> I need guidance on two scenarios:
>
> There are some code-points (e.g. 98, 99) where Pablo is listed both as the
> “Reference” as well as the “Change Controller”, and hence I used Pablo as the
> reference.
>
> Since there is no specific draft linked to this allocation, shall I just
> leave the reference just as empty ?

When creating the YANG model, you're pointing at behaviors, and the references are the place where you would go to look to see the definition of that behavior: the RFC or the draft. That's an easy thing, but what happens if I'm implementing 98 or 99? Do I go ask Pablo? That doesn't scale or make any sense.

For 98 and 99, I have no idea why they were not included in rfc9800. It looks like everything else around them in the registry was included. This issue can be easily fixed with a quick draft that defines them.
[skraza2]: Yes, a new quick draft could be added.


I think those two are problematic if they're not defined anywhere, which defeats the purpose of a specification: interoperability.


The other point you raise is that Pablo is listed as a Change Controller in some cases. Usually, that's okay for drafts, but in the cases of RFCs, the IETF, and not a specific person, should be the Change Controller. I don't know what happened with the code points for rfc9800 where Pablo is listed as the Change Controller.  I'll talk to IANA about them.


> Morever, there are cases where the allocated code point has a non-WG draft as
> the reference and the draft has also expired.
>
> Can I put a non-WG draft as a reference ? Assuming answer is a “no”, then do
> I also leave the reference just as empty in this case as well ? Please
> advise.

The answer is yes. As I said before, the reference indicates where the behavior is defined. Ideally, it would be something that is already published like an RFC, but it can be a draft, either a working group draft or a non-working group draft.


> > 3. All the drafts and RFCs that are referenced inside the models need to be
> > referenced outside the model as well. And listed as Normative references.
>
> [skraza]: I believe this is the case - All the “reference” drafts/RFCs are
> part of our Normative references.
> I will double check.

There are drafts referenced inside the model that are not on the final list. Only I-D.ietf-spring-sr-service-programming is a reference right now, and is listed as Informative. They all need to be Normative.

We need an explicit statement right before the module definition starts:

   This module references [RFCabc], [RFCxyz], [I-D...], and [I-D...].

[skraza2]: I’ve cross-checked and added all the missing references in the statement - please see attached diffs.


Going back to non-working-group drafts. Yes, they become normative. That is, of course, an issue with the fast publication of the model because we will have to wait for those to progress.

I've been working with the authors of I-D.ietf-spring-sr-service-programming because I know your model depends on it.

We will have to either make the decision to leave everything in the model as it is right now and wait for the drafts. That doesn't mean we have to wait for the drafts to last call the model, but we won't get an RFC number until those clear. The other option is to remove those points and do an augmentation later.
[skraza2]: I’d go with the latter option - remove the non-wg drafts and augment the model later if/as they become WG / RFC. I’d still keep the behavior defined in the existing WG drafts though.

[skraza2]: I’ve updated the diffs based on this - please see attached herewith. Thank you.


Alvaro.