On 7/27/06, Erik Moeller <eloquence(a)gmail.com> wrote:
[snip, out of order]
In my view, establishing clear ground rules for votes
to change policy
is a better way to deal with the problem than delegation of authority.
It allows for community consensus processes (and indeed requires them
to be tried first), brings out as many arguments and solutions as
possible, and enables everyone to share the responsibility, credit and
blame for the result.
The current image use policy reflects the position of the active
editing base and is result of a fairly strong consensus, not merely
the mob rule of a majority wins vote. There is no substantial desire
to change our policy.
The challenge is that only a bright line policy can protect us against
a "slow and often intensely frustrating process" for each of tens of
thousands of images per month. But a bright line policy will exclude
things which common sense would permit. I'd like to discuss ways we
can accept such exceptions without breaking the well functioning
policy and without creating a slow and intensely frustrating process.
Voting is currently used as part of our image process... but because
it's such a niche area, deletions tend to be virtually uncontested or
the vote turns into a contest of who can inject more friends.
Elected or appointed expert committees are one way to
deal with this
problem, and they may often work quite well. However, if we accept an
expert committee as a method to achieve consensus about policy
[snip]
I have not proposed an expert based solution. This is a role which
practically any established Wikipedian would be qualified for.
"But," I hear some people say, "you
can't let ordinary people vote on
these complex issues. They do not understand! That's why we need to
have smart people to make these decisions for us. Through enlightened,
reasoned debate, surely they will find the solution that is best for
us all." I'm not sure that's true.
I'm glad you're not sure it's true, because I haven't suggested it..
and I haven't seen anyone else here suggest it.
There isn't any group of people appointed to be experts on enwiki, and
as far as I'm aware there has been no serious advocacy of such a
system any time recently.
There are good votes and there are
bad votes. A good vote is one where voters are presented with a
concise summary of the different arguments that have come up in a
discussion that preceded the vote, where the _options_ in the vote
have been developed through consensus, and where there is a strong
culture that pressures voters to read and understand all arguments
before voting. A bad vote is one that is done ad hoc, out of process,
with poor methodology and no clear prerequisites.
It's not clear that the facts support your position. Perhaps on
dewiki? Experimentation on Enwiki has demonstrated that the majority
of the participants in at least some of our voting process do not read
evidence presented preceding or included in the debate (measured by
placing external links in intros and in individual votes). Instead it
would appear that many voters make their decision based on initial
impression and a passing glance at the standing votes.
While it may be true that acceptable results can be achieved through
good votes, where the participants consist of informed parties, it is
not clear if such votes ever actually happen on enwiki.
I'm not sure if you're aware of [[Dunbar's number]]....
Experience on Enwiki seems to support that and related research.
On single articles (over 99% of which have less than that dunbar upper
bound on community size) decisions are often quickly and supported
made via consensus which is so clear that a poll is unneeded.
These decisions are not merely tolerated with indifference but supported.
In many voting areas and on many policy pages, where we have
substantially more than Dunbar's number of participants, new proposals
are treated with either icy indifference or extreme factionalization.
It sometimes appears that the vote has become nothing more than a
contest to measure which subgroup can draw in the most members. This
is also supported by the overwhelming clustering that happens with
some votes.
I believe that without the control feedback of a functioning social
group our users are too unwilling to engage in the mixture of
compromise and consideration which are required to have a 'good vote'
and instead their behavior appears to be determined more by a desire
to assert their authority (by fighting against something rather than
working with it).
This matter appears to be made substantially worse because partisan
social groups form (i.e. inclusionist vs deletionists) which include
participating in votes as part of their social activity. So rather
than having a big social group which includes people of all
perspectives encouraging their members to work for the good of the
project, we get fights between smaller social groups which include
more myopic subsets who socially reinforce unthinking 'mob like'
behavior. :( ... then we reward this behavior by allowing the count
of people to impact our decisions.