RE: [rfc-i] 3rd party SDO cross-referencing of IETF work (was: Re: Chair/datatracker tracking expired WG documents ?)

["Pascal Thubert (pthubert)" <[e-mail]>]

Hello Brian:

We're in sync for most of your words, many thanks for taking the time. With one caveat:

I never asked to allow normative references to I-Ds; The problem is that the text in RFC 2026 forbids even academic papers to reference a draft. I wrote a number and failed to obey the law. Academia loved WIP and tries to build on it, propose ideas, test them, and that really fair.

You'll find that IEEE references all external document as informational anyway, even our std track RFCs. So in my book it would not hurt to reference informationally a WIP as long as it is clear that the work is not final. 

Note also that there's a work that we decide to keep as a draft and never publish; that work cannot be referenced either. E.g., https://datatracker.ietf.org/doc/html/draft-ietf-roll-protocols-survey was key to establish that we needed a new routing protocol for low power IoT, and distinguish the expected properties that were missing in existing protocols. That document was never to be made RFC - at the time the IESG had the foot on the pedal. With the terms in RFC 2026 no document can reference it and tell the story that lead to RPL.

The IETF are the first to disobey RFC 2026 since we do reference drafts in other drafts, in published RFCs and even in normative documents. 

The proposal was:

"ANY standard in or out IETF MAY reference non-normatively an I-D but MUST NOT reference them normatively. "

To which I may had:

"ANY non-standard document MAY reference an I-D, indicating that it is a WIP."

And yes this would synchronize the Tao and BCP 9 with the boiler plate.

Keep safe;

Pascal

> -----Original Message-----
> From: Brian E Carpenter <[e-mail]>
> Sent: vendredi 1 avril 2022 23:03
> To: Pascal Thubert (pthubert) <[e-mail]>om>; Salz, Rich
> <[e-mail]>om>; BRUNGARD, DEBORAH A <[e-mail]>om>; Eric Rescorla
> <[e-mail]>om>; Joel Halpern Direct <[e-mail]>
> Cc: [e-mail]; [e-mail]
> Subject: Re: [rfc-i] 3rd party SDO cross-referencing of IETF work (was:
> Re: Chair/datatracker tracking expired WG documents ?)
> 
> On 02-Apr-22 02:27, Pascal Thubert (pthubert) wrote:
> > Hello Rich
> >
> > The piece that's still confusing in the reference to BCP 9 is whether
> and where the original text from RFC 2026 that is still currently quoted
> on the IETF site was amended. Removing from the latest online version is
> not the same thing as if it was never there. People looking that text up
> on the internet have a good chance to find one of the TAO RFCs and live
> with it, or get stuck like I am searching through BCP 9, none of which
> particularly satisfying.
> >
> > The trouble: since RFC 2026, a lot happened in https://www.rfc-
> editor.org/info/bcp9. I tried to find if the text was amended somewhere
> but got lost.
> 
> The answer is no, that text is still valid. Of course we can't control
> what third parties do, which is why the text of every I-D states that it
> is work in progress. Fair warning to all readers.
> 
> As we know, *significant* changes can occur in a standards-track I-D up
> to and even beyond the announcement of IESG approval. (Beyond, in the
> form of a note to the RFC Editor from the IESG, and very rarely if a
> significant error is found at AUTH48.) So it is truly never completely
> safe to make a normative reference before the RFC is out. It's IMHO a
> matter of judgment for the relying party at what point they consider it
> safe enough - IESG approval, for example - but it's our responsibility
> to tell the truth, which is that I-Ds are always work in progress.
> 
> 
> > This is a situation where our own "update" work becomes quite opaque
> till we finally "obsolete" to retrofit. The IEEE has an interesting
> maintenance practice for that.
> 
> The IETF has been repeatedly resistant to a full revision of BCP9. I
> tried and failed. It would consume a great deal of energy that might be
> better spent on technical work.
> 
> > So no, I do not think that removing is enough. Since the quote was
> there, we probably have to maintain an up-to-date state there too.
> 
> We agree on that.
> 
>      Brian
> 
> >
> > Keep safe;
> >
> > Pascal
> >
> >
> >> -----Original Message-----
> >> From: Salz, Rich <[e-mail]>
> >> Sent: mardi 29 mars 2022 23:48
> >> To: BRUNGARD, DEBORAH A <[e-mail]>om>; Brian E Carpenter
> >> <[e-mail]>om>; Pascal Thubert (pthubert)
> >> <[e-mail]>om>; Eric Rescorla <[e-mail]>om>; Joel Halpern
> >> Direct <[e-mail]>
> >> Cc: [e-mail]; [e-mail]
> >> Subject: Re: [rfc-i] 3rd party SDO cross-referencing of IETF work
> (was: Re:
> >> Chair/datatracker tracking expired WG documents ?)
> >>
> >>>     These couple of sentences in the Tao are actually from RFC2026
> >>> and about
> >> the transient state of I-Ds.
> >>
> >> In fact, the current version of the Tao (maintained at
> >> https://github.com/ietf/tao) doesn't have that language, but instead
> >> just points to BCP 9. Phrases like "work in progress" are also
> >> removed, pointing to BCP 9.
> >>
> >> We only just finished editing/merging before IETF 113, and there
> >> wasn't time to update the website.  Please take a look at the GitHub
> >> repo and see if the language there needs to be fixed.
> >>
> >> I think Valery's point got lost :(
> >>
> >