suggestions (voting)

[avri doria]

Marc Blanchet wrote:
> Hi,
>  it seems to me that I'm not clear on what I'm saying. Maybe the word
> "voting" has too much overloaded meaning. Let start again:
> - wg meet physically every ~4 months.
> - public non-voting is used during meetings to judge wg concensus on a topic
> 
> problem: judging wg concensus every 4 months is long, which makes the wg
> not moving on topics.
> solution: replicate on-line the public non-voting process, so it can be
> used anytime and don't wait for the next IETF to resolve topics.

I think one misunderstanding here is that the WG meeting is not the
time for checking on consensus. A WG chair only gets a sense of the
room at a meeting.  This is not an indication of consensus.

Consensus is a much more specific thing and can only come from the
mailing list.

Rough consensus is something derived from listening to the people who
speak up.  Silence is assumed to be agreement or at least acquiescense.
But a WG chair must give sufficient  time for a dispute
to be discussed and must clearly state what the issues are.  Once a
WG chair thinks that rough consensus has been reached, or for that
matter can't be reached on a topic, she must state that understanding
to the mailing list and must again give everyone a chance to comment on
that understanding.

One cannot do anything about those who are silent other then ask
the list repeatedly for a response. Even in a voting situation, most
would be silent.  I do some work with an industry fora that does
work on the basis of votes, and the silence there is no different then the
silence of an IETF mailing list: there are those who are active and then
there are the rest.  Voting of any sort, brings in great complexity that
I think is best avoided.  And while the rough part of rough consensus
can be a complicated thing to judge, in most cases it is pretty
straightforward.

I believe that when properly executed, the rough consensus process works.
It takes at least 2 weeks, or even 3, for it to work, and it takes
focus on the part of the WG chair to execute it.

Rough consensus involves time and an interplay between the WG chair(s)
and the participants in a WG.

> 
> Question is:
> - do we agree about the problem?

i don't think we do.

(now if i am the only one who doesn't then you may
have rough consensus on the problem, if on the other
hand people were to agree with me, or were to have
still other opinions on what the problem is, then we
would not have rough consensus on the problem.)

i think the problem is that an insufficient number of
people really understand what rough consensus is and the
responsibility they play in achieving that rough consensus,
and that all WG chairs do not execute the process correctly.

> - if yes, then 
>     if you agree with the solution, 
>       fine
>     else 
>       what do you propose.
>   else 
>     then we don't need to talk about the solution, since we don't agree
> with the problem.
> 
> Marc.
> 
> -- lundi, novembre 25, 2002 01:15:21 +0100 Leif Johansson <leifj@it.su.se>
> wrote/a écrit:
> 
> 
>>Marc Blanchet wrote:
>>
>>>--- dimanche, novembre 24, 2002 18:27:36 +0100 Leif Johansson
>>><leifj@it.su.se> wrote/a écrit:
>>
>><snip>
>>
>>>That is what I'm proposing: it is not to change the procedure, but to
>>>setup and offer a "voting (or named it as we want)" tool so that chairs
>>>can ask questions on the mailing list that would give them the
>>>equivalent of the "public non-voting" but on-line.
>>>
>>
>>There is a big difference: when you are conducting a vote you can't turn
>>to those who have voted and ask them to motivate their position. I agree
>>with Kurt Zeilengas formulation of this principle: contention is out of
>>scope. I.e if you need a tool to judge the sense of a wg then you don't
>>have clear consensus and you need to think about what the wg is doing.
>>
>>Remember: the wg members have to basically agree -- there is no way for
>>the ietf to create standards with a 51% majority vote and there is no
>>stick with which to beat companies who don't implement RFCs. Ultimately
>>the market decides what solutions get deployed. This is (imho) why the
>>ietf works (when it does): since we almost only standardize (not
>>counting all published RFCs mind you) what we agree on the acceptance
>>of ietf standards in the market is almost a tautology.
>>
>>Sorry if I was rambling a bit there ;-)
>>
>>	Cheers Leif
>>
> 
> 
> 


-- 
Avri Doria
http://www.sm.luth.se/~avri/