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

On 25. Mar 2022, at 21:55, Brian E Carpenter <brian.e.carpenter@gmail.com> wrote:
> 
> Why should we care if some other SDO is happy to publish a "standard" with a normative dependency on a document whose preamble includes these words:
> 
> 'Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."'
> 
> This is not our problem.

Process confabulation at its finest.

This *SHOULD NOT* be our problem, in what some of us might consider an ideal world.  FTFY.

But it very often is.

When CoAP was done in 2013, the RFC got stuck on normative references to two security documents(*) that inexplicably took another year to emerge.
Referencing the approved I-D was *exactly* the right thing for an SDO to do during this time.

Beyond that, often documents are at a 95 % state for a *long* time before being approved, but can’t get approved until the last i is dotted.
An SDO may have schedules that make them publish documents that are fine with a reference to a 95 % or 98 % ready state.

(Other SDOs may also simply want to use work before *our* quality standards consider it ready — there is a reason other SDOs have rather varying quality objectives, often trading quality for velocity.  We may consider that unfortunate or even detrimental, but we are better off knowing that this has happened than ignoring an emerging installed base.  And I’d rather have them *reference* our I-D than copy-paste text.)

> Trying to maintain cited-by pointers is approximately impossible. TimBL knew that, which is why the Web succeeded where other distributed hypertext projects failed. While I see why Toerless would like it, I don't think we can have it in the real world.

That is a separate discussion that I think derailed a bit here.

I believe that collecting incoming references (whether I-Ds or RFCs) is a worthwhile effort, even if the result will have only quite partial coverage in many cases (sometimes to an extent that it is not useful, say for RFC 793).  Giving SDOs or alliances a way to indicate such a reference early (even while the referenced document is still I-D) can help us with planning and should be encouraged.  The technical process for this might be modeled after the IPR disclosure process.

Grüße, Carsten

(*) The delay also all but killed attention to the approaches provided by these security documents, so we should have excised the reference to them/put them into a separate document and published the referencing document with a gaping hole.
We need better process tools for dealing with such delays — instead of letting a document perish in the RFC editor queue, we need to be able to do emergency publications based on emerging (emergency?) delays (**).
But that is yet another discussion we can’t ignore — at least this one is partially relevant for Rfc-interest…

(**) The average MISSREF age in queue is > 40 weeks [1] — i.e. the average sojourn time once caught in that limbo is likely beyond a year and a half.

[1]: https://www.rfc-editor.org/reports/CurrQstats.txt

_______________________________________________
rfc-interest mailing list
rfc-interest@rfc-editor.org
https://mailman.rfc-editor.org/mailman/listinfo/rfc-interest