[Foundation-l] breaking English Wikipedia apart

[FT2]

On Mon, Mar 14, 2011 at 12:51 PM, Andre Engels <andreengels at gmail.com>wrote:

> (snip)
> First, let me apologize beforehand for sounding too cynical, but I
> have many years of experience with Wikipedia, and I have seen many
> attempts to deal with trolls, POV pushers and otherwise substandard
> editors (I even initiated one or two myself), and I have not seen a
> sign of any of them actually working.
>

Efforts are incremental. This isn't aiming to "deal with trolls" per se, but
to improve the quality of the editorial community as editors. Right now we
leave people to themselves, and possible benefits that could be gained from
recognizing accomplished proven editors aren't gained. The conclusion was
that if this is given a leg-up, then it will have positive effects on other
areas including issues of poor editing. Most approaches to improvement are
narrow-focus, they fix a specific issue. This one tries the flip side: - if
general quality increases and we recognize the users that we believe won't
POV war or be biased and will act well even on difficult subjects, then we
have ways to approach disputed areas that we didn't have before and we could
remotivate and encourage users who got fed up, as well as providing
something positive for people to aim to once they get "into" editing.

>
> > Newcomers can't
> > distinguish those who edit well and those not shown to edit well.
>
> Why would a newcomer be supposed to care about that? Does it matter
> whether my article gets edited by a 'good' editor or by a 'bad'
> editor? Am I supposed to revert a bad editor but leave a good editor
> alone if he makes the same edit? Or should I better leave the edits of
> the bad editor alone, because he's probably a troll who will chase me
> away if I revert him?
> Even more so - the 'bad' editor may be an excellent editor who just
> has not yet had the time to prove him- or herself.
>
> Newcomers care because they look for easy ways to identify users who can
help them, or whose advice in discussions is more reliable. If I have a
technical problem my first issue as a newcomer is "who of all these mass of
people should I ask? Who might know? Whose answers on policy are more likely
accurate? It makes it easier for informal coaching to take place (user can
post "I see you are experienced here, I have a question" or can I come back
to you if I need more help") Making participants of proven quality easier to
identify helps in other areas of life. There is every reason to expect it is
helpful here.

There is a big difference between "not yet proven" and "bad", so your last
point isn't an issue. A means to formally recognize such users does not mean
others are "bad" any more than the existence of a "junior swimmer's" badge
for children means all other children can't swim.  Given users don't seem to
have a problem with reverting admins, I think they wouldn't have a problem
reverting any other user :)


> In edit wars it
> > provides a bias towards endorsement of probably better edits.
>


>  Actually, no it doesn't. The way to behave in an edit war to avoid
> being singled out as a bad editor is to stay away from it. Is that the
> way we want our editors to act? Be afraid to revert, not because they
> might be wrong, but because there might be people who think they're
> wrong?
>
> Disagree.  The point is that if there is an edit war, users arriving (eg
from noticeboard discussions) initially have to work out what's going on.
Often they don't know any (or most) of the users concerned at the start and
it's all a big pile-on mess to them.  This discourages experienced editors
from getting into complex disputes and in some cases lets edit warriors get
away with tendentiousness longer than is needed. If it were easier to see
that participants XYZ are recognized by the community as competent level
headed editors whose editing is consistently good quality, then it
immediately orients the new visitors to the dispute. It may be the other
"side" are also good editors, or that the respected users are acting badly.
But in most cases this is useful information and helpful. As a side effect
it makes it easy to see when someone who gained community recognition for
their editing quality is acting badly, and so it will tend to discourage
them doing so. Right now there's no consequence so we often see
otherwise-good users behaving badly in this or that dispute.



> > In the case of
> > massively disputed topics such as ethnic wars it provides a dispute
> > resolution tool - editing might be restricted for a time to those editors
> > considered "proven" by the community.
>
> Currently such pages tend to be locked to all but admins. That doesn't
> work either - people just keep on their fighting on the talk page
> until someone gives up, after which the page is unlocked and their
> opponent can declare their victory on the page. Or the fight simply
> moves to the next page.
>

Yes. Now imagine we had 3000 users whom the community has scrutinized their
editing and conduct and feels they act well and edit well across the board.
We don't have to lock down the article to admins, we can restrict editing to
any of those 3000 users, and anyone else who wishes to edit can seek
community recognition and then do so as well. The editing is then open, but
undertaken by users who have proven they know how to cite, discuss, seek
NPOV, etc. Give it 3 or 6 months the article (or topic area) will probably
be in good order and can tentatively be unrestricted again.

Note I have very severe "mass participation" edit wars such as ethnic topics
in mind here, where we have tried for years to bring good editing.  We don;t
have any solution. This could help. It means that instead of locking
articles down to admins (who mustn't edit anything contentious), or snails
pace development, we can restrict it to proven editors of which we have
thousands, and anyone gaining that recoignition can join them. The article
stays open, but poor conduct and mis-citing or tendentiousness vanishes -
because any "proven" editor who does act up will be in a peer group of users
who are overwhelmingly good quality, proven, who know bad conduct when they
see it, and know how to deal with it appropriately.

>
> > Finally it is egalitarian (or at least
> > as much so as anything on the wikis) -- it is a recognition anyone can
> > achieve from the community by editing and behaving well, and anyone can
> lose
> > by editing or behaving to a visibly poor standard. It carries no formal
> > powers, but by peer pressure alone encourages improvement generally.
>
> So we are supposed to add a load of work to the editors' workload in
> judging the edits of prospective proven editors, but then don't even
> make that choice have any real effect? I don't feel safe in voting for
> someone to be considered a 'good' editor in this sense unless first
> checking a few hundred of their edits. And definitely in the beginning
> there will be several such applicants per week, certainly if we are
> going to make this something for 'everyone' to aim for.


I would expect we can find a way to do it that is fairly straightforward.
How it works is a separate issue, I think solvable, but for now I'm looking
at the principle of it.  In the beginning I would suggest we "Grandfather"
in groups that are broadly trustworthy, such as "all users who have written
2 GAs or one FA and also passed RFA - between these criteria we can be
fairly sure they can edit well and also know policy and conduct norms well.



> Again, forgive me if I sound too cynical, but I do get the feeling
> that such a system might well be a nice thing to have, but would be as
> effective in promoting good editing behaviour as a Barnstar.
>
> Not quite. A barnstar means one person, somewhere, wanted to say "well
done". It doesn't mean the user's work generally or their conduct generally
is good quality, that they generally edit neutrally or cite well, or treat
others well, or that the wider community has reached agreement they are of
proven competence and approach. Once you have that, you can do good things
like motivate and coach newcomers, provide goals to head to, or develop new
dispute handling methods. If you can draw on a pool of a few thousand users
whose capability and appropriate behavior can largely be assumed.

FT2