[Mod-discuss] Re: WG Last Call: draft-ietf-modpod-group-processes-10 (Ends 2025-09-08)

["Rob Wilton (rwilton)" <[e-mail]>]

Hi Dan,

Please see inline …

From: D. J. Bernstein <[e-mail]>
Date: Thursday, 28 August 2025 at 18:42
To: [e-mail] <[e-mail]>
Subject: [Mod-discuss] Re: WG Last Call: draft-ietf-modpod-group-processes-10 (Ends 2025-09-08)
Rob Wilton (rwilton) writes:
> if someone repeated makes the same point that drowns out other useful
> discussion

Hmmm, really? Can you give a few links to examples of IETF mailing-list
threads that were drowned out, so that people can understand concretely
what happened and consider potential solutions?

No, not publicly. I don’t think that would be appropriate or necessary since most of the people participating in this working group will be aware of such instances, and ‘relitigating’ such examples is unlikely to be fruitful.  I would be willing to send you such information privately, but (1) I get an auto reply from your email address that indicates that you are probably dropping/filtering my emails given that I’m not willing to enter into a contract just to send you emails, and (2) it seems that you would ask that I share it in public, which as per above, I won’t do because I don’t think that is appropriate.


Did the list server run
out of disk space?

Not that I’m aware of, but disk space is cheap and isn’t the problem (except it is allegedly causing water shortages in the UK ;-).  The cost to be concerned about is participant time/energy in reading and responding to those messages.


By default I would think that any repetition of the same objection
reflects a failure by the chairs of whichever WG to maintain a chart of
the objections that had been raised already.



No, not really.  Often in these cases the same issue gets raised repeated, sometimes in very slightly different guises, and responded to.  Eventually, many of the WG participants get fed up and either stop reading the emails, or stop replying to them, since the discussion isn’t progressing, it is just stuck at a point of disagreement.  Often you will see other participants starting to complain on the list about the same point being repeated brought up.  IETF consensus is not, and should not be, either last person standing or whoever shouts the loudest wins.  Sometimes, a participant’s opinion is in the rough and they have to accept it, even if they think the WG is making a technical mistake and move on.  In some cases, the WG may subsequently realise the mistake and rectify it.  Life is imperfect.

In addition, for the key issues I believe that the WG chairs already did this (e.g., please search the mod-discuss archives for emails that Lisa has sent out as chair and you will find various examples where she has summarized open issues.  I also cite such an email later in my reply).  In addition to this, the authors have also been tracking issues on github (available here, https://github.com/larseggert/draft-ietf-modpod-group-processes/issues)  So, I disagree that there is any failure of the chairs here, or that issues haven’t been adequately tracked.


As a concrete example of such a failure, having this WG proceed to last
call suggests to observers that all objections have been addressed---but
that's not true: for example, there was no response to Keith Moore's
point about Challenger. Observers can't find a chart of the objections;
they're forced to dredge through list archives. Here I am suffering
through highlighting the Challenger point again, exactly because the
point was improperly ignored. The root-cause fix for the repetition is
to track objections (as in other standards-development organizations),
not to shoot the messenger.

And this is perhaps a case where you and I have a difference of opinion on how the IETF process works. I don’t agree that all emails must be responded to (e.g., if there are just restating a point that has already been fully debated), and I don’t agree that all issues require changes to the underlying text.   If all issues have been clearly considered by the WG then in many cases, the conclusion may be that no further changes to the text are needed.  Further, WG LC (and IETF LC) is another opportunity to ensure that all remaining issues have been raised.

So, for example, if someone was moderated by the WG chairs (e.g., prevented from posting directly to the mailing list) during WG LC because the chairs perceived that they were being disruptive then, depending on the nature of the disruption, I would anticipate that the chairs (or moderation team) might take 1 of 2 actions:

  1.  Set the mailer so that the WG chairs can manually read/filter the emails from the moderated participant before deciding whether they can be sent on to the WG.
  2.  Allow the moderated person to send a single email to summarize their objections during WG LC, to ensure that their  issues have been heard.

The moderation team is expected to come up with guidance for WG chairs, which may end up being like what I suggest above, or it may be substantially different.  It doesn’t really matter because the key point is that, at least for me, there is an expectation that the moderation team acts to minimize the risk of censorship, as is stated in the draft.


But I don’t really see the point that Keith is making about Challenger having any bearing on a moderation team for the following reasons:

  *   The challenger issue (as summarised by Keith - I’ve not read any details) seemed to be due to managers ignoring (not censoring) the technical input of engineers.  However, the moderation team is made up of other participants, not folks involved within the IETF leadership and the moderation procedures document clearly states that moderation is not to be used to suppress technical input.  In my experience, often contentious decisions in the IETF can boil down to disagreements as to whether A or B is the better approach, or disagreements above whether a particular approach or change will be harmful to the Internet.  Just because someone is in the minority and is objecting loudly doesn’t mean that that they are technically right and their technical input is more valid that the others participants in the discussion.
  *   The challenger decision seemed to happen quickly.  I haven’t read the details, but I would presume within 24 hours of when the disaster happened.  But IETF decisions happen over a period of weeks or months, normally with multiple opportunities to raise the issue, and an appeals process in place.  Often IETF solutions are tested and deployed over a matter of years.  Hence, I’m struggling to see a case where a moderation action is going to cause the Internet to fail in a way that is not spotted and rectified or mitigated.

But I would like to turn this argument around.  Can you please cite examples within the last 5 years (that don’t directly involve yourself because your opinion may not be objective in such cases), where moderation has been used in the IETF to suppress constructive technical input into the standards process.  Ideally, I would be looking for cases where the decision has been appealed and the appeal succeeded.



> However, given that we have a WG with an approved charter I believe
> that we are well past the point of debating or deciding whether
> anything needs to be done.

The charter does _not_ mandate censorship.

Neither does the solution, as best explained by Eliot’s previous reply (https://mailarchive.ietf.org/arch/msg/mod-discuss/OYrqrTbf0Gubb2twyPOj3nGPyjI/)


Everything the charter asks
for can be met by purely advisory actions.

I disagree, and this specific issue was one that was tracked by Lisa, discussed, and I believe rough consensus has been reached (https://mailarchive.ietf.org/arch/msg/mod-discuss/fVxyeW-qN0cDF6OLHgOh7k8DOik/)  That isn’t to say that the consensus cannot be changed during WG LC, but I do think that it is incumbent on the individuals who want to change the consensus to make the argument and persuade the WG that the previously reached consensus should be changed.


This would also eliminate the
concerns about damaging IETF's openness and silencing minorities.

If we are unable to improve the decorum on IETF lists then we just end up silencing a different set of minorities.  In particular, younger folk who seem to be much less willing to put up a bunch of old white guys having very vocal and forthright debate – what was acceptable in the 80’s and 90’s is no longer acceptable today, and the IETF must somewhat move with the times (hopefully in a moderate fashion) to continue to stay relevant.

Kind regards,
Rob



---D. J. Bernstein

_______________________________________________
Mod-discuss mailing list -- [e-mail]
To unsubscribe send an email to [e-mail]