Re: AD Sponsorship of draft-moonesamy-recall-rev

[Michael StJohns <mstjohns@xxxxxxxxxxx>]

On 4/22/2019 10:19 PM, Eric Rescorla wrote:

On Fri, Apr 19, 2019 at 8:33 AM Spencer Dawkins at IETF <spencerdawkins.ietf@xxxxxxxxx> wrote:

I'm actually enjoying this more than I should be ... 

<deleted several levels of response>

The report that design team produced is at https://www.ietf.org/archive/id/draft-dawkins-nomcom-3777-issues-00.txt. It resulted in most of the updates to RFC 3777 before they were all obsoleted by https://datatracker.ietf.org/doc/rfc7437/.

Thanks for pointing us to this document. Useful to avoid reinventing
the issues list.

In that vein, I think it would be useful to get crisp about the
problem statement. It seems to me that this document implicitly makes
two arguments:

- It's too difficult to initiate a recall (because the threshold
  is too high and because members of the leadership who might have
  special knowledge are excluded) [S 2.1, 2.3]

- It's unfair to exclude remote participants [S 2.2]

I'm not sure how persuaded I am by the second argument, but I think
the overall thrust of the first argument has some merit. However,
I tend to agree with Alissa that the main obstacle isn't so much
the number of people or excluding leadership, but rather the
reputational risk to people of signing the petition, especially
if the recall fails [0].

So, if the idea is to make recalls easier, then it seems to me
that the easiest thing to do would be to make the signatures
on the petition anonymous; the Secretariat can still verify them,
of course, and the recall proceedings themselves would still
be conducted in our usual fashion.

-Ekr


[0] "If you come at the king, you best not miss".
[1] It's also possible to build a system that allows third party
    verification that the right number of sigantures were provided
    while not identifying the signers, but the amount of crypto
    required seems somewhat prohibitive.
   

The more I listen to this thread, the more I'm thinking that the whole concept of recall as described in the appropriate documents makes little sense for the IETF/IAB as currently constituted.  It certainly makes no sense for any of the other entities for which the Nomcom proposes candidates (e.g. the LLC and Trust) for which the appropriate remedies implicate various laws or contractual issues. 

The ONLY time we've ever come close to running a recall, it was for the IAOC and it was to deal with the fact that we hadn't created procedures to deal with abandonment of a position or disability of a member especially where we had (with one person AWOL from a normal set of 5 members) a potential quorum problem.

When the problem occurred, it took only about a day to get 20 members to sign the petition once it was decided to handle it via recall.  The member in question ended up resigning quickly enough that the recall committee was not formed. (I think I've got this timing correct...)   It would have taken something like 6 weeks or so to get the recall committee up and running and then more time to actually complete the recall, then another 4-6 weeks to fill the spot - call it 2-3 months to fix a problem.

Instead, I'd like to propose that we move to an expulsion model for the IAB and IESG where the members of the organizations are able to remove a member under certain circumstances:  Behavior inconsistent with a fiduciary (e.g. acting for your company or contracting entity to the detriment of the standards process); behavior that adversely affects the standardization process (IESG) or behavior that adversely affects the general operation of the IAB (e.g. things like harassment); abandonment of the position or  lack of communication from the member.  

2/3 of the permanent members can vote to remove.   For the IETF chair, a majority of both the IESG and IAB (e.g. two groups each with majority approvals) are needed.

This would at least be a lot faster to resolve if there is a great enough issue.

If we still want to talk about recalls, let's talk about *why* we might want to recall someone?  Is there a current or potential (within the next year or so) reasonable scenario that we need to address?  Once we figure out what the possible offenses are, we can figure out if recall is the remedy or if we need to think about something else.  I'm still puzzled as to why we need to make recalls easier to accomplish and whether there is an actual issue with not allowing remote participants to trigger a petition.

Later, Mike