First of all, I am, at least in this context, far less concerned
about explicit behavior that can be identified, raised with ADs,
and appealed than I am about more subtle problems.
I generally agree. Appeals work okay (though they are burdensome) when
there's a simple and obvious purely technical problem that can be easily
fixed. But for a wide variety of reasons they're not a satisfactory
solution to a wide range of problems.
Just to cite a simple example, appeals aren't likely to fix a problem
that's fundamentally caused by nomcom's biases in selecting IESG and IAB
members (for example, a bias that such members should be employees of
large corporations who are willing to pay employees to do those jobs and
fund their travel). And the most likely effect of an attempt to
resolve such a problem with an appeal is to increase those biases.
This is really a systemic problem in IETF that can only be addressed by
significant changes in the structure of the organization.
Putting
aside Carsten's argument for specifics for a moment (will come
back to it; see next note) and exaggerating (I hope) a somewhat
abstracted example, suppose someone repeatedly pushes a
particular proposal in a WG that the WG (or at least its chairs)
conclude is completely unfeasible. Perhaps the idea is even
unfeasible at the level of our classic "that would be a great
idea if we could make some small adjustments to the speed of
light" or "yes, if the earth were flat" positions.
The proposals that would require revision of laws of physics to be
successful are relatively easy to deal with, because many engineers are
already aware of those limits. It's much harder to deal with proposals
that would require people to change their assumptions, and/or business
practices. And yet, very many ideas are defeated, in various ways,
without a WG really considering their feasibility. One way this can
happen is by IESG chartering a WG with such a narrow scope that it's
forbidden to consider other ideas. Another is for the ADs to appoint WG
chairs who already have their minds made up and who can be counted on to
discourage discussion of them.
I'm not saying that this is entirely a bad thing. Some ideas really are
unworkable, and a WG that exhausts itself discussing unworkable ideas
may never manage to produce a workable solution. ADs know that if a
WG's charter can cite running code that does approximately what seems to
be needed, the WG is much more likely to converge in finite time. But
it's hard to make such a process objectively fair, or to claim with a
straight face that the result is a rough consensus. And sometimes
workable ideas are dismissed out-of-hand because of these. And
sometimes a WG's failure to resolve important design questions results
in tussles that do harm to the Internet.
(i) We have, IMO, managed to turn appeals into a Big Deal,
treating them less as "hey, IESG, this needs another look" and
more into a quasi-judicial process in which we think about
decisions (and often even people) being overridden or overruled.
There have even been undertones suggesting that people who file
appeals, especially ones that don't end up changing anything,
may be at a disadvantage in the IETF later. Some reports from
the IESG to the community can even be read as
self-congratulatory about there being no appeals in the previous
reporting period. That climate makes the process rather
intimidating for a newcomer or occasional participant to use and
hence not as realistic an option as many of us -- Carsten, you
(Brian), and myself included-- would like to believe.
In my experience IESG tends to treat appeals as DoS attacks, which is
perhaps understandable given the amount of time and effort IESG
typically (and IMO appropriately) devotes to such considerations.
I suspect that IESG should be relieved of the burden of handling
appeals, especially since many appeals are more-or-less questioning IESG
decisions anyway (either in considering a document prior to
approval/publication, or in writing charters or picking chairs). An
old saying about foxes and henhouses comes to mind, but a more
fundamental problem is that IESG is already overworked and appeals are
significant disruptions to IESG's schedule and sometimes WGs' ability to
finish their work and make room for new work. I suspect appeals should
be handled by an independent panel that doesn't already have an
investment in the decisions it's (re)considering.
However, I am concerned about a different set of cases. As a
sort-of example, of one, many of us look at the brief
descriptions of new WGs, decide on that basis whether to read
the charters, and then decide whether to participate actively or
not. The "not" decisions can include lack of interest, lack of
expertise, or just a conclusion that, given the goals and
constraints of the charter, our time is better spent elsewhere.
And then we ignore the WG and its work until a document shows up
on IETF Last Call or something else, perhaps an Area review
request or even an idle hour during an IETF meeting, brings it
back onto the radar (we presume temporarily). No active effort
to exclude open participation there. Then, attention having
been drawn to a particular issue, we attempt to watch that issue
and perhaps contribute some comments. Attempting to watch is
complicated because we typically lack context. In theory, we
could get that context by reading over a few years or WG mailing
list archives, watching meeting videos, and untangling Github
updates. In practice, that is nearly impossible and wouldn't be
enough even if it were possible. Now parts of what happens are
nearly inevitable: a significant fraction of WG participants,
perhaps even the chairs, are somewhat annoyed by the late input
and what they see as an attempt to disrupt or even sink their
smoothly and tranquilly sailing ship. In many cases, WGs get
past that, are supportive of the effort, and consider the input
on its merit even if doing so disrupts planned schedules. In
others, the input is dismissed, its source treated roughly,
and/or small concessions are made (in discussion or documents)
to the unimportant parts of the input while the rest of it is
ignored. In the process, the person with the input (often not
just their input) is made to feel extremely unwelcome (even if
that was not intentional).
Concur. I think we have too many WGs. I also think that WGs should
keep a record of decisions made and the reasons cited to support those
decisions, so that when such a decision is called into question,
newcomers can catch up without the whole group having to reiterate the
discussion. Also, many times a decision is made without much group
awareness of the consequences, and it might make sense to revisit those
decisions later. But it's also fair to ask those who propose that the
group change its mind, explicitly address the reasons cited for the
original decision.
The last two messages seem to imply that there are corners in
the IETF that are not as tranquil as I'm experiencing.
If there are any such issues, they cannot be discussed in the
abstract. Unless we know what actually happened, we cannot
discuss remedies.
Correct. An appeal (and the eventual IESG or IAB response) is
a manifestation of this.
Except when an appeal is impossible (or so easily dismissed on
procedural grounds to be a waste of time) and the alternative is
public laundry washing on this list -- something that would
probably be unhealthy for the IETF, probably could not be done
without an exemption from the Code of Conduct, and would
certainly be disruptive of the WG and the community.
I actually think that the IETF list should be used for this more often,
that there is currently far too much suppression of useful discussion on
this list, to the detriment of the IETF and the Internet itself. While
I'd like to cut the (new) moderators a bit of slack since they've
inherited some really unhelpful baggage, fundamentally IETF is supposed
to be a volunteer organization, and the volunteer participants should be
the organization's conscience. In recent past the leadership has tried
to assert control over the volunteers whose consensus they are claiming
to represent. The original idea of the SAAs was to get rid of the
worst misbehavior so as to permit useful discussion, but the role has
been re-interpreted much more expansively than either originally
intended or appropriate. This suppression of discussion drives
conscientious engineers away from IETF. Because who wants to beat their
head against a wall (and maybe also spend many thousands of dollars
attending meetings) trying to improve an organization that's become too
stubborn and ossified to improve?
Keith
