Adrian,
--On Thursday, March 21, 2019 08:52 +0000 Adrian Farrel
<adrian@xxxxxxxxxxxx> wrote:
> People signing a recall petition are not making judgements.
> Those judgements are made by the recall committee.
>
> The recall petition simply requests that a situation is looked
> into.
>
> Let's take an extreme case.
>
> Suppose an AD made a decision that was hugely detrimental to
> remote participants while favouring people who attend meetings
> in person. This would only be something subject to a recall
> petition if the beneficiaries of the decision decided to issue
> the recall petition. This detail disenfranchises the
> increasing number of remote participants.
> Now, we all know that such situations are unlikely. We also
> know that we haven't seen a lot of recall petitions let alone
> actual recalls. But that is not the point.
It actually is part of the point if the recall process is
designed (whether intentionally or not) to discourage recall
initiations. There have been analyses of that, and proposals to
try to adjust things, before (see below) but they have been
dropped because it was clear that getting them seriously
considered was either nearly impossible or that the problem
wouldn't justify the amount of energy it would cost and/or the
amount of rancor it would cause. That issue may or not
discriminate against remote participants -- it is possible that
it has negative effects on everyone (see response to Harald
below).
> There are two key points to these rules:
>
> - To be able to handle extreme and exceptional conditions
> correctly
> - To present a face that is open, welcoming, and fair
> We fail the second of these, and the first looks shaky.
Indeed, and not a new observation.
> The rule was written before registering for remote
> participation was a thing.
The observation was made years ago that we couldn't do anything
with remote participants and recalls because remote participants
didn't register and because registering them would be an
intrusion on the privacy of those who just wanted to quietly
lurk. The response was to suggest a change with an uncanny
resemblance to where, for other reasons, we have ended up today:
those who wanted to participate remotely needed to register
while those who just wanted to listen and/or watch didn't.
That idea went nowhere, leading some people to conclude that the
IESG didn't want a change (possibly because they believed that
considering it would be more trouble than it was worth) and that
the "no registration" story was just an excuse.
> It's an easy fix. Allow petitions to be signed by people who
> have registered or attended. Change no other parts of the
> rules.
Again, been proposed before. Went nowhere. While I blame a
lot of the problme of lack of progress in areas like this on the
IESG (and have since the NEWTRK meltdown), there is also the
matter that, AFAICT, the vast majority of the IETF just wants to
get the things done they come to participate for, thinks that
are mostly technical, and has little enthusiasm for process
issues. That, as we have seen (and continue to see) with the
small number of people who have been actively involved in the
IASA2 process and contributing to the WG of that name, makes it
hard to get momentum and easy for the IESG to (accurately) say
"not enough people care" when a proposed process change is
something that does not originate within it IESG or within a
close circle of friends of at least some ADs. Have things
changed enough to make a new proposal likely to succeed? I
believe so, but have not seen a lot of evidence.
And...
--On Thursday, March 21, 2019 12:58 +0100 Harald Alvestrand
<harald@xxxxxxxxxxxxx> wrote:
> when we revisited this 10 years ago, the consensus was that if
> a recall petition didn't find enough signatures among the
> people who were nomcom-eligible, it was likely not a
> reasonable recall to pursue anyway.
Yep. And a reasonable concern if there was reason to fear
assorted edge personalities and trolls using the recall process
to paralyze the IETF generally and the IESG in particular. The
observation that appeals were far easier to mount, even under
the original, RFC 2027, "Anyone may request...", and had never
caused such paralysis was legitimate, but recalls require far
more community resources. So a change was made to raise the
bar, which, IMO, was completely reasonable.
However, the decision to raise it to 20 and to tie that to
nomcom-eligibility were questionable then and remain
questionable today. The latter was, IIR, largely because it
seemed at the time like a plausible criterion and no one had
much energy to develop and consider other options. Whether
that was reasonable then or not, increases in remote
participation (and IETF commitments to facilitating and
encouraging remote participation) make it unreasonable now and,
as Adrian effectively points out, unreasonable whether we
actually believe more recalls are needed or just because of the
importance of keeping the process perceived of as open,
welcoming, and fair.
> Unlike with nomcoms, a
> recall petition only needs a few signatures - most people who
> are active in the IETF know someone who's nomcom-eligible and
> could try to convince them to sign.
> The bar for initiating a recall is pretty low, given the size
> of the nomcom-eligible population.
Might have been true a decade or so ago although I thought it
was questionable even then because we had a few people who were
attending few meetings f2f but who were much more clearly
contributing in obvious and important ways that most of the
people who showed up at meetings to enjoy the social contacts,
enjoy fine lunches and dinners, contribute registration fees to
the pot, and otherwise warm seats and do email.
20 is actually a lot of signatures, especially for someone who
is mostly participating remotely and/or is actively involved
only in a single WG or work topic (both makes it worse). It is
an even higher barrier if people are worried about retaliation
--from either the subject of the recall or from others in the
leadership trying to protect one of their own or avoid the work
and disruption that recalls cause. That is a concern I've heard
often in recent years but can't remember anyone's being worried
about ten or fifteen years ago (maybe I have just forgotten or
was too close to the leadership myself for anyone to tell me).
There is a further deterrent to anyone who can look at the
required timeline (especially if there were foot-dragging
anywhere in the system), consult a calendar, and count
backwards, especially if they are aware that recall efforts have
been discouraged in the past by pointing out that, by the time
any recall could get someone out, it would be so close to their
coming up on the normal Nomcom cycle that the recall would be a
pointless waste of time and resources. That situation brings us
pretty close to a situation in which it is reasonable to believe
that the recall process, as now specified, was designed to be
unusable (again, that is not an assertion of intent -- it could
have been, and probably was, designed that way as the result of
a series of accidents).
Sadly, while I'd like to see a different criterion than
nomcom-eligibility used for recalls, if only because it would
improve the optics somewhat, getting to a recall (or other
exceptional removal) procedure that was perceived as usable and
likely to be effective in "extreme and exceptional conditions"
and having it perceived as a useful element of "open, welcoming,
and fair" would require more than that bit of tuning.
best,
john
