Re: Dealing with Design issues following Last Call

[Dmitri Tikhonov <[e-mail]>]

On Fri, Dec 18, 2020 at 08:39:58PM +0000, Lucas Pardue wrote:
> On Fri, Dec 18, 2020 at 8:11 PM Dmitri Tikhonov <[e-mail]>
> wrote:
> > And then what?  In other words, should I implement the changes associated
> > with 4257 or not?  The text is in transport-33, yet the issue is still
> > open.  I take it to mean that this change is not as solid as it could be
> > and it may be rolled back (otherwise, why leave the bug open?).  In this
> > case, implementing this may end up to be a waste of time.
> >
> 
> Draft-33 is effectively the proposed resolution to all issues opened during
> the IETF Last Call. These are the documents that the IESG will review to
> ensure we addressed last call comments. It's feasible, albeit hopefully
> unlikely, that changes might be requested to the text that deals
> specifically with those design issues. The IESG may also request other
> changes as a result of their review, of course. The transport, tls,
> recovery and invariants documents are on the agenda for the next IESG
> telechat on 2021-01-07 [1].
> 
> [1] https://datatracker.ietf.org/iesg/agenda/

I see -- thank you for the explanation.  So if IESG requests changes,
there will be a -34?

> I'd treat implementing draft 33 as trying to implement a PR before it was
> merged, while the WG was running its late-stage process. There's a risk it
> could change or be rolled back, it's upon implementers to decide if they
> wish to bear that risk. I'll also highlight that draft 33 includes the note
> "DO NOT DEPLOY THIS VERSION".

Touche!  Although I merely want to *implement*, not deploy this version.

> That doesn't mean don't trial implement the changes but it is a factor
> people should consider before investing effort at this stage.

I am trying to prevent a situation where IESG approves -33 on one day and
the RFC is published two days later and we do not have the changes
implemented yet.  Then, we would really have to scramble!  Am I naive to
expect such speedy RFC turnaround?

  - Dmitri.