From Meta, a Wikimedia project coordination wiki

Minimum number of votes

I would like to make a proposal that there must be a minimum of votes before a Bureaucrat can be deemed valid. I was thinking that the number should be 20, which would reflect an minimal activity level that would require the Bureaucrat abilities (admining, changing user names, etc). I would first like to get a little feed back before I make a proposal (with an RfC) sometime after this weekened. Ottava Rima 20:12, 5 February 2010 (UTC)

Where? On Meta or on any project? -- Lavallen 17:19, 7 February 2010 (UTC)
Anything connected to the WMF - i.e. all projects. CU and OS have a minimum of 25 I believe. Ottava Rima 19:26, 7 February 2010 (UTC)
I think the swedish wikipedia is large enough to have local bureaucrats, but it's not always easy to find 20 users who wants to spend time voting for admins/bureaucrats. Only very welknown users can collect 25 voters. (We have re-election every year of all admin/crats/CU/oversight.)
According to the meta-policy, swedish wikisource is not even big enough to have local sysops, but it's difficult to run a wikisource-project at all, without some of the admin-tools. I frequently have to use the move without redirect-tool for example. Vandalism is not a big deal, but the technical problems around the proofreading-tools and djvu-files, makes it's necessary to have these admin-tools and it's of great value that we have the posibility to spread the tools among very active users... -- Lavallen 10:13, 8 February 2010 (UTC)
There is quite a bit of difference between a local crat and a local admin. This is not a proposal to put a minimum on local admin. Regardless, if a project could not find 20 people to elect a crat in 7 days (not 20 supports, 20 total votes), then there is not enough activity to warrant not coming to meta for crat related needs. For reconfirmation, that is a local process and not an election process, so they could create whatever standards they want. Ottava Rima 15:10, 8 February 2010 (UTC)
Pasting here an email I sent to stewards-l before:

Based on the discussions about russian wikiversity, I think it would be wise to add a line to the steward handbook, to avoid future confusions. Int he past, we have, out of the top of my head, discussed thresholds of 7 admins/30 votes. (either of them would do) I still think that is a reasonable and good threashold because below this threshold

  • a) there is no real /need/ for bureaucrats - everything can easily be handled by stewards
  • b) the community is not large enough yet to be sure there is no ongoing manipulation and that the community will stand up if necessary.

The difference of bureaucrats with admins is that bureaucrats are able to grow their supporter base by granting admin rights. that way a small group can potentially unnoticed "take over" a whole wiki, especially when the userbase is still small.

I would like to propose to add the following line to the stewards handbook (under "Administrator and bureaucrat rights"):

  • If the community is not large enough (rule of thumb: 7 admins or 30 votes) it is not advisable to grant bureaucrat rights at all.

I did not add the reasonings in that line since the handbook is no pamphlet nor a political explanation, but merely a howto. Effeietsanders 21:32, 8 February 2010 (UTC)

7 admins or 30 votes vs. 5 admins and 15 votes ? SergeyJ 21:57, 8 February 2010 (UTC)

  • What will be with bureaucrats who have got the flag earlier, without having satisfied these conditions? SergeyJ 21:59, 8 February 2010 (UTC)
    Generally, projects don't set rules ex post facto, so those bureaucrats already in place would be left in place. Future bureaucrat elections, however, would need to abide by the rules. As always, the Foundation (via its sysadmins and staff) can do whatever the heck they want, since they own the machines... they're responsible for any flubs, however, also. Kylu 00:07, 9 February 2010 (UTC)
  • 7 days are not enough, even active editors can be absent 7 days. 7-14 days it are normal. SergeyJ 22:11, 8 February 2010 (UTC)
I plan on offering many choices for a "minimum" amount of votes for an RfC. Personally, I would not want to have a requirement of more votes for a Bureaucrat than for a CU/OS, as the latter two can be far more damaging. However, others may disagree. Ottava Rima 22:40, 8 February 2010 (UTC)
    • This feels a bit like policy creep. Stewards are here to support all wikis that need help, but their existence should not be a reason to keep small wikis from becoming self-sufficient. By preventing small active wikis from learning how 'crat tools work, we would also be effectively discouraging people fluent in the language of those wikis from getting the early experience using those tools that the older wikis have had. Sj+ help translate 23:52, 8 February 2010 (UTC)
      Policies that seem natural on larger wikis (getting 30 votes for an election should be no big deal!) are not always appropriate for small but active and stable wikis. While it's true that a rogue 'crat can theoretically skew the selection/appointment of new admins, the local community can also point this out once it is a problem and ask on meta for help. Soft security is appropriate here. This edge case is much rarer than the common case: that small wikis successfully get 'crats, translate 'crat-request pages, and start handling their own local admin and renaming processes.
      Example: Swahili Wikipedia is stable and active, and has a couple active 'crats. It might satisfy the letter of "7 admins" but not all are active; and it does not generally get more than 10 votes for any specific vote, though there are many more than 10 regular editors. That wiki seems to me competent to be self-sufficient. Sj+ help translate 23:52, 8 February 2010 (UTC)
Would you support getting rid of all minimums for voting amount, which it seems like your statement suggests? They exist already because of the major potential for abuse, especially when all power in a small community is in the hands of a tiny few and there is no possible check. These are not awards or status symbols that show a project's "worth". These are important tools, and I think it would be unfair for a Bureaucrat to pass simply because he got 3 supports and no one else chimed in. Ottava Rima 00:06, 9 February 2010 (UTC)
CU and OS have those restrictions because of the availability of personal information. 'Crats have no access to such dangerous information. 'Crats are really just admins with slightly more tools. No more dangerous really than admins. -Djsasso 14:10, 9 February 2010 (UTC)
From discussions I've had on the matter, it is more certain that CU and OS are in -pairs- because of personal information and the minimum number is based on need/danger. Crats would not need to be in pairs, but since they can rename users, approve bots, and approve admins, there is a need/danger level. Ottava Rima 14:55, 9 February 2010 (UTC)
There is nothing a crat can do that can't somewhat easily be undone by a steward if they go rogue. Whereas with a CU/OS, once they release public information, it can never be unreleased. There is a very big danger difference. -Djsasso 17:06, 9 February 2010 (UTC)
Technically, OS can only get rid of the damaging information once it is already released, and cached pages and the rest could contain the information. Now, for releasing information (not public, as I assume you meant private), they have to disclose their actual identity to the Foundation, which would mean that they would be responsible for anything damaging to another individual. Bureaucrats do not have such a requirement. Ottava Rima 17:55, 9 February 2010 (UTC)
Actually they can't once its been posted and someone has seen it, its out there. They might be able to remove it from the wikik but short of using the memory erasing device from Men in Black its out there. Right Bureaucrats don't have that requirement because nothing they do is permenantly damaging. -Djsasso 18:46, 9 February 2010 (UTC)
Right, I think we are now somewhat on the same page. Now its just down to numbers and the rest. There wont be a requirement to be over 18 or to identify to the Wiki. There probably will be a much smaller number in the proposal necessary for OS and CU already. Wouldn't having at least 10 people necessary at least help prevent the temporary damage (as it could cause quite a bit of headache, especially with bot approval and admin approval, and Stewards don't have that much free time to clean up a major mess). Ottava Rima 18:59, 9 February 2010 (UTC)
No, I think the risk is worth the reward of a smaller wiki being able to become self reliant. Being forced to run to stewards is a detriment to a wiki, as users often feel they are the dealt with at the whims of someone who isn't part of their community. This causes editors to leave the wiki disheatened. Thus contributing to stiffling growth of the wiki. If anything gaining crat should be easier not harder. A number of wikis have recently adopted policies which allow for automatic cratship once an admin has been admin long enough and has a couple current crats sponsor him (meta being one). If anything that is the way I would prefer things to go. You mention stewards don't have time or a clean up....but they have time to basically be the crats on all these wikis that would now never be able to get a crat? -Djsasso 20:10, 9 February 2010 (UTC)
I'm sorry, but I cannot find any credibility in your argument. If a wiki has less than 20 people, then there is no way it can become "self reliant" nor would there be need for a Crat as there is no real need to use any of the abilities. Even larger Wikis, like Wikiversity with hundreds of active users, need to rely on Stewards for various tools. As to your last point, a Steward may have to promote up to 5 people on a wiki that would lack a Crat under this idea (per a 4 to 1 ratio of admin to non admin) but a Crat operating bots or oping people with admin privileges to abuse can destroy a lot in a very short amount of time, which would require far more than the few seconds needed to op the five admin. Ottava Rima 21:21, 9 February 2010 (UTC)
What you fail to realize is not every editor on a wiki is going to take part in Rfbs. Finding 20 votes is often not an easy task on wikis that actually have 100 active editors. Alot of people actually just come to the wikis to know...the whole purpose of the wikis? Nevermind that alot of editors might not be active in the 7 day or whatever timeframe. Quite frankly there is no good reason to make it harder and numerous reasons to leave it as is.. None of the scenerios you mention have ever even remotely even come to pass with our current way of doing things, your proposal is a solution in search of a problem.-Djsasso 03:38, 10 February 2010 (UTC)
How do you know? And if less than 20 people are interested in voting in a Crat election, who is to say that they are interested in having a Crat to begin with? I would think that a lack of attendance at a vote would say quite a lot about the community's feelings on the matter. By the way, the current way of doing things is requiring 5 or more admin on a project and quite a few votes already. I only wanted to solidify a standard and make it more than just the current "rule of thumb". Ottava Rima 04:18, 10 February 2010 (UTC)
One minor correction, Ottava: Oversights can not only remove information, but can also see information hidden most times. Specifically, if they use the suppression functions, they can see and un-suppress the information again, but the older oversight function (hiderevision) requires a developer to view/replace. As far as privacy issues go, while checkusers can see IP addresses and have a larger target base (being all users), oversighters have access to more potentially private or damaging information, though only regarding those users who've had it removed via that function. Sukida 21:04, 9 February 2010 (UTC)
Please reread. The point was that OS cannot suppress something that never existed, and once something existed, it can still be seen in various caches or through other matters (even if its for only a limited time). The damage of an oversighter is very limited, especially seeing as how a crat could approve multiple admin bots that could block everyone, delete all pages, etc. Ottava Rima 21:21, 9 February 2010 (UTC)
I understand your point, I'm just more used to thinking in a "reversible vs. irreversible" manner: While annoying, a group of adminbots could be reversed, and most actions of a bureaucrat can be as well (I plead en:WP:BEANS here). Certainly the most potentially damaging user groups would be checkuser and steward, by both our measures, would you agree? Sukida 21:32, 9 February 2010 (UTC)
As a note - OS and CU fall under Ombudsman check if there was ever a big incident, but there doesn't really seem to be any real authority to deal with an abusive Crat, except if a group of people could rally together on meta. But if a project has 10 people and 6 dissent, then 6 would not seem like enough to de-Crat a Crat. There is no real check there to the power. If a project has 10 people and they all decide that they want a Crat and the Crat to have a bot that blocks anyone but the original ten from editing and the rest, what is there to really stop such a thing? Etc etc etc. The hope is that a project with at least 20 people is large enough to organize and deal with problems. It is far easier to tyranize over a group of 5 people than a group of 20. Ottava Rima 21:54, 9 February 2010 (UTC)
Stewards would act in both those situations. We have them for a reason. And we hope they have the ability to look at what is happening and come to a reasonable solution of decratting if the situation warrants it. Tyranizing is not as easy as you seem to think it is. -Djsasso 03:41, 10 February 2010 (UTC)
Instead of restrictions offered here it is necessary: to allow to vote in any cases for removal of a flag of the bureaucrat. It clean all suspicions. If the bureaucrat does not arrange - always it is possible to re-elect it! SergeyJ 01:33, 10 February 2010 (UTC)
It is a problem not only small projects, but also big. It is necessary to give simply freedom to participants to re-elect those whom it wants. All other criteria are not effective and decided. Why simply not to give the chance to express freely will (1) for assignment of a flag (2) for flag removal SergeyJ 01:40, 10 February 2010 (UTC)
Free will? I'm confused as to what you are trying to say. Simply put, my opinion is that a project that has less than 20 active people is far too small to be a project in any regard and if that many is unwilling to vote in an election, then they do not have the dedication to the project. It would seem odd to have projects floating about with 5 people who are all crats and basically dominate some obscure place no one else wants to go to. It definitely does not seem like a good use of resources and the imbalance of the structure would make it less inviting for new people. Ottava Rima 03:14, 10 February 2010 (UTC)
I would actually say if the people can't be bothered to vote in elections and stick only to editing, then they are actually more dedicated to the project than those who sit around spending all their time in the bureaucracy of running the wiki. After all the whole point of the wikis are to edit. -Djsasso 03:42, 10 February 2010 (UTC)
Then I think we are at a philosophical divide that will probably not be bridged in the short time that this discussion will last. Ottava Rima 18:47, 10 February 2010 (UTC)
It seems to me you have not understood. All other criteria are not effective and decided. Why simply not to give the chance to express freely will (1) for assignment of a flag (2) for flag removal . Free will = elections, But with any number of participants. Possibility presence at any moment to vote for flag removal eliminates errors which could be earlier. SergeyJ 15:02, 10 February 2010 (UTC)
I do not see free will found in a group of 10. A small group is easily dominated by one person. Ottava Rima 18:47, 10 February 2010 (UTC)
Why? And when 30? What changes? Even if so happens. When there is 30 community can vote and remove a flag itself! SergeyJ 20:42, 10 February 2010 (UTC)
As I said before - 5 people are active in a community and 5 people vote to have a Crat. That Crat falls out of favor and 2 people complain. They could either represent 40% of the community or be "two bad apples". If 30 people are active and 30 people vote to have a Crat, and that Crat falls out of favor and 12 people complain, chances are it is not 12 bad apples. Furthermore, if the two people were complaining legitimately and nothing was done, then that project could easily drop down to 3 people. Definitely not helpful in any way, nor are projects to be "owned" by anyone, which a Crat could easily do. Ottava Rima 22:01, 10 February 2010 (UTC)
I can is bad I explain because of not knowledge of English. In time t0 there are 5 participants. They we will assume have chosen the bureaucrat because of 3 bad apples. In due course (t1) new participants 25 more come. 15 from them complain - the bureaucrat remove. In what harm? While bad apples (what not the fact that such in general is) liked the bureaucrat - it was. As soon as the society became more and it is not pleasant it have removed. SergeyJ 02:07, 11 February 2010 (UTC)

I definitely agree we need something set down as a policy, this has been brought up to often and there is always a fight (understandably) where users do not understand why they can not get what they voted for. While I totally understand the idea of trying to avoid crats who could "rule" a project (I have seen it happen on at least one very small project) I also think we really need to keep the requirements as small and flexible as possible. Personally I would tend to shy away from vote requirements for 2 reasons:

  • 1. there will always be groups that basically deserve to get one but don't meet the vote requirement for some reason or another, wasting time or effort.
  • 2. It makes the system very inflexable. There are at least a couple wikis I know of who do not have "normal" crat systems. Meta and Simple for example allow any sysop who has been active for 6 months and is approved by at least 2 crats to get promoted (without a ton of bureaucracy and more waste of time voting) esWiki I think does something similar and possibly others. I think whatever we do should allow things like that.

I think a better idea would be a requirement for a certain (smallish) amount of sysops already permanently elected. The easiest number off the top of my head would be 4-5 but that could be up for debate. While we are on the topic we need to come up with a policy number for how many votes you need to elect permanent sysops (5?) that also comes up frequently and we don't really have a policy written down anywhere. James (T C) 03:43, 10 February 2010 (UTC)

Meta and Simple have pre-existing bureaucrats who can implement such systems, stewards granting bureaucrat status are going to be doing so for projects which have not previously had active bureaucrats. It's perhaps not a popular opinion, but I think projects wanting to have alternative bureaucrat promotion systems should wait until they've at least got a couple bureaucrats who can implement it; This has the benefit of preventing stewards from trying to interpret rules for such promotion which may well be in a language they are not fluent in, and therefore subject to ambiguous interpretation by machine translators or overly enthusiastic locals. Sukida 17:58, 10 February 2010 (UTC)
  • dear SergeyJ, do you mean to discuss possible policies on your (or other) wikis in general, or policies regarding stewards promoting bureaucrats on them? to me it sounded like the first, but i want to make sure. oscar 23:43, 10 February 2010 (UTC)
policies regarding stewards promoting bureaucrats on them - Here and only it I discuss it SergeyJ 02:12, 11 February 2010 (UTC)
However differences between (1)wikis in general, or (2)policies regarding stewards promoting bureaucrats on them I do not see. SergeyJ 02:16, 11 February 2010 (UTC)
once wikis have bureaucrats, these can promote admins but also other colleague bureaucrats, according to their own policies, stewards are basically not needed nor involved in this. only in the cases of the first bureaucrats or demotion, stewards are involved. so there is a difference, you see? oscar 02:25, 11 February 2010 (UTC)
Or wait. You wish to tell, what (1) you can think up criteria and agree it to give out a flag of the bureaucrat (2) but cannot oblige re-elections of bureaucrats? SergeyJ 02:25, 11 February 2010 (UTC)
all i wanted was to ask you for more clarity, to make sure we talk about the same thing and understand each other. do you now see the difference i was asking about? very best, oscar 02:32, 11 February 2010 (UTC)
The first (1) that I do not understand - according to their own policies. How so can be? We speak about infringements? . (2) You can remove bureaucrats for infringements? (3) You can oblige all projects to re-elect bureaucrats? .. If your answer (1) yes (2) no (3) no - then I understand a difference. Otherwise it is not present. SergeyJ 02:43, 11 February 2010 (UTC)
within the framework of the wikimedia projects, as long as its mission is served, each project is basically free to develop its own customs and guidelines, also about appointing, reconfirming and resigning of bureaucrats. doesn't your "homeproject" have such pages already? oscar 23:42, 11 February 2010 (UTC)
Well, I have understood you. The problem of stewards only to begin the project. That occurs then it not important. Through what that time stewards do not supervise projects. There is only a question since what moment local projects become not controllable stewards? SergeyJ 06:02, 12 February 2010 (UTC)
stewards and bureaucrats are experienced and knowledgeable all-round helpers of the community/communities, they are not "controllers with power". stewards nor bureaucrats are allowed to "rule" projects, these projects are joint efforts at all times. may i ask again please: can you please give a link to the pages concerning these matters on your homewikiproject, as it would interest me to read them. thanks in advance and very best, oscar 02:31, 27 February 2010 (UTC)
I have not understood, what exactly you wish to read? SergeyJ 12:03, 27 February 2010 (UTC)
your local versions of pages such as Administrator, Bureaucrat, Meta:Requests for adminship, Meta:Requests for help from a sysop or bureaucrat, and especially Meta:Policies and guidelines and similar pages. thanks in advance, oscar 12:21, 27 February 2010 (UTC)
in ru.Wikiversity or in ru.Wikipedia ? SergeyJ 12:41, 27 February 2010 (UTC)
on ru.wikiversity please, since your request concerning this project prompted this (part of the) discussion. all the best, oscar 13:01, 27 February 2010 (UTC)
К нам едет ревизор :) (To us the auditor goes :) ) ... But I as spoke about ru.Wikipedia . Now I have not enough time, more detailed I will answer you later. While Administrator = [1] [2], Bureaucrat = [3], Policies and guidelines = [4] Further under references SergeyJ 13:13, 27 February 2010 (UTC)
If there will be still questions - set them here - [5] I with pleasure will help you to understand. SergeyJ 15:10, 27 February 2010 (UTC)
We're getting some feedback here, but little policy-making decision. I've opened Requests for comment/Minimum voting requirements to do so. Kylu 01:37, 28 February 2010 (UTC)