Re: Working Group Last Call: QUIC Ops Drafts

[Ian Swett <[e-mail]>]

I have not read either draft in its entirety recently, but I've read all
the recent issues and most subsequent PRs.

There is currently a huge amount of activity in the manageability draft and
a decent amount in applicability.  These drafts didn't receive this level
of review previously, partially because the other drafts were still
changing.  I don't think these changes will be huge, but they are likely to
take a few weeks to a month and I would prefer another draft and WGLC after
the changes are complete.

Thanks, Ian

On Mon, Feb 15, 2021 at 10:12 PM Martin Thomson <[e-mail]> wrote:

> Hi,
>
> I've read reviewed both of these drafts.
>
> I think that the applicability draft is good.  It's clear and accurate.
> Having followed the changes there, what comes out of WGLC should be in an
> acceptable state for requesting publication.
>
> I would like to see another version of the manageability draft before
> approving it.
>
> Most of the problems are "just" editorial, but there were numerous factual
> inaccuracies that directly relate to the suggestions being made.  I think
> that I caught most of those things that were obviously wrong with issues,
> but I might have missed a few things.
>
> The biggest concern I have is with the way that the draft is not always
> clear about the conditions of statements it makes.  There are several
> critical distinctions that need to be kept very clear in writing something
> like this and this is excellent in some areas, but inconsistently applied.
> The things that I think are important to keep clear always are:
>
> 1. The distinction between the properties of version 1 of QUIC and things
> that apply to all versions of QUIC.  I think that this could be achieved by
> stating up front that text only refers to version 1, but also repeating
> "version 1" to avoid ambiguity, even more than strictly necessary.
> Explicitly calling out those few cases where invariant properties are being
> discussed (which is more often than might be obvious) seems sensible.
>
> 2. The distinction between actions taken unilaterally by path elements and
> actions taken by entities that cooperate with QUIC endpoints in performing
> their functions.  Here, most of the document is written from the
> perspective of not requiring cooperation.  That is stated up front[*], but
> the frequent digressions to talk about endpoint cooperation means that it
> is worth being extra clear, even to the point of redundancy.
>
> [*] The draft says on-path observer in Section 3, but many of the actions
> in Section 4 involve intervention.
>
> I don't have editorial pull requests, because I have to prioritize other
> work.  I think that this requires more work than I can commit to doing
> within the WGLC period.
>
> Cheers,
> Martin
>
> On Fri, Feb 5, 2021, at 05:12, Matt Joras wrote:
> > Hello,
> >
> > Now that the base drafts are cruising down the road to RFC the Chairs
> > and Editors believe it's time to proceed with the standardization of
> > the ops drafts. Therefore, this email announces a Working Group Last
> > Call (WGLC) for the following QUIC documents:
> >
> > * Manageability of the QUIC Transport Protocol
> >   https://tools.ietf.org/html/draft-ietf-quic-manageability-09
> > * Applicability of the QUIC Transport Protocol
> >   https://tools.ietf.org/html/draft-ietf-quic-applicability-09
> >
> > The WGLC will run for two weeks, ending on 18 February 2021.
> >
> > Please review the documents above and open issues for your review
> > comments in our repository at https://github.com/quicwg/ops-drafts. You
> > may also send comments to [e-mail].
> >
> > Thanks,
> >
> > Lars, Lucas, Matt
> > QUIC WG Chairs
>
>