I'm following this ptwiki captcha case closely and I think I can bring some elements for this conversation.
The portuguese community was surprised by the removal of catcha promoted by the global (technical) community. This situation led then to react and a long discussion was held not only about captcha but also about community autonomy. Following Nemo sugestion on bug 41745 the Portuguese community voted a new proposal to enable the captcha again.
When we look to the stats of ptwiki in May and June it is certain that the number of reversions has increased a lot, but it is also true that the number of non-reverted editions raised even more. But some editors claim that sysop and rollbackers are overloaded and this numbers are showing us that more "garbage" (revisions that should be reverted) is staying in ptwiki and not that we are having more good faith editions. I believe that it is very hard for us to tell which theory is right without performing an A/B test.
Besides this debate, it is notorious that sysops and rollbackers are felling overloaded and the growth of gross number of edits that should be reverted (independent of the number of good editions) is a real concern for the ptwiki quality. The community felt that it wasn't ready to deal with this new amount of edits and decided that they should do something in the short term to avoid demotivation of vandalism fighters, and the solution they came up with was enabling the captcha again.
But, as Oona said, they know it cannot be a definite measure and the "AntiVadalism Project" ( http://pt.wikipedia.org/wiki/Wikip%C3%A9dia_Discuss%C3%A3o:Projetos/AntiVand...) was reactivated willing to create mid and long term measure to fight vandalism enabling the community to deactivate captcha. In this context, one think proposed is the use of the captcha associated with AbuseFilter (as Steven suggested), but for this to happen bug 18110 should be implemented. Also, they proposed bug 41522 to evaluated which kind of editions are being kept away by the captcha (trying Marc Pelletier's theory) and are working on improving bots, filters and asking for new users to became rollbackers.
So, I think ptwiki community is leading many efforts to create sustainable growth, and right now the captcha is a contextualized temporary part of the plan.
Best,
Henrique Andrade
On Sun, Jul 7, 2013 at 4:21 PM, Samuel Klein meta.sj@gmail.com wrote:
On Sat, Jul 6, 2013 at 8:39 PM, Oona Castro ocastro@wikimedia.org wrote:
For now, it's worth mentioning the Portuguese Wikipedia community has
been
working on this antivandalism project
http://pt.wikipedia.org/wiki/Wikip%C3%A9dia_Discuss%C3%A3o:Projetos/AntiVand...
order to build alternative measures to deal with vandalism and inappropriate edits with a very small portion of the community reverting edits - considering the short and mid terms.
They are already aware that even the return of emergency CAPTCHA won't
be a
definite measure (lasting no more than one year, as per what was agreed)
As much as I dislike captchas, this seems like a considered decision by the Portuguese Wikipedians. We should support local wiki communities in making choices for themselves -- and help them to run short-term experiments, evaluate the results, and correct mistakes.
Letting communities make and learn form their own mistakes is more important than always being 'right' for one definition of rightness: we can learn from many independent communities, each with their own standards. Of course we all want to improve editor engagement + retention, and overall quality + coverage - the pt:wp community does too! The question is how to trade off between these.
One requirement for making a controversial configuration change - or for continuing it beyond a short initial test period - might be the ability of the requesting community to evaluate its effect.
Sam.
Wikimedia-l mailing list Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe