Talk:Stewards/Archive 1

From Meta, a Wikimedia project coordination wiki

Technical question

Since Tim proposed that developpers be not allowed any more to do these duties that they were doing before, and that those with this new status will be doing, what will happen on the international wikipedias ?

It seems french people are not willing to have yet another additional status, and do not see developpers holding all powers a serious issue. So, we have one developper, Shaihulud, who may take care of this now.

If developpers can not do this any more, and if we have no one with this status, who will take care of unsysoping a sysop doing something wrong ? I read above that you think 4 people will be enough ? Does that mean Angela and Maverick will be those who will take care of our administrative needs then ? Anthere 18:00, 14 Mar 2004 (UTC)

That's not a technical question. It's also not my decision. But under the proposed rules, you would express your confidence in Shaihulud's ability to administer fr by nominating and voting him in. -- Tim Starling 22:54, 14 Mar 2004 (UTC)
It's probably best to think of these as Wikimedia technical servants, shared among all wikis and replacing Tim and Brion for jobs which don't require a programmer. If you'd ask Tim or Brion today, these are the people you would ask instead. If you think someone at fr or any other wiki has the technical and human skills (mainly a reputation of trying to find and follow consensus, IMO) then please do nominate them. Jamesday 01:12, 17 Mar 2004 (UTC)
To unsysop someone, you need to know how to do the necessary SQL command and enough about linux and the database to get to the place where you do that. If there is a need to see if an account has the same IP address as other accounts, that takes the technical skill to find and look in the logs. This is much lower technical skill than a software developer/programmer needs but it is enough to make it impossible for some people to do this job. If Shaihulud is willing to take requests for projects other than fr Wikipedia (wikisource, wikibooks, Welsh and Latin Wikipedias, and many others), please do nominate Shaihulud. This isn't for just one wiki - it's Wikimedia Foundation support for them all, including those with no developers who are participating in them. Jamesday 10:37, 17 Mar 2004 (UTC)
As I said in my mailing list post "The feature is easy to use and does not carry the security risks of write-access to the database." To unsysop someone, you log in as a developer, go to Special:Makesysop, blank the "rights" box (which contains "sysop" by default), type the name into the name box, then click "set rights". To set user rights across wikis, use e.g. "Anthere@fr". To work around any cross-wiki encoding problems, find out the target user id and use something like "#55@fr". The user will be able to find out their user ID by viewing Special:Preferences. -- Tim Starling 05:06, 22 Mar 2004 (UTC)

From what I know, there are a number of functions that are turned on/off by developers. Some - like access counter, search, and some Special: pages - are turned on/off based on technical considerations such as server load. Others, such as bureaucrat status, "category" function, direct display of external images, etc, are introduced/ stopped more based on human/ social considerations.

It would be good, if technically feasible, to have the latter kind of functions handled by the non-developer people, I suppose. Tomos 19:39, 17 Mar 2004 (UTC)

Sure, sounds like a good idea. -- Tim Starling 01:13, 23 Mar 2004 (UTC)

Practical questions

Finally, is this true or not true  ? To unsysop someone, you need to know how to do the necessary SQL command and enough about linux and the database to get to the place where you do that.

No it's not true. I'll write some documentation when I get time. But if you're competent enough to edit Wikipedia, you're competent enough to desysop people. -- Tim Starling 15:57, 20 Mar 2004 (UTC)

If there is a need to see if an account has the same IP address as other accounts, that takes the technical skill to find and look in the logs. This is much lower technical skill than a software developer/programmer needs but it is enough to make it impossible for some people to do this job.

There are bandwidth issues with allowing people to look at logs over the web. I'm not sure if it's practical. I'm not planning such a feature at the moment, there's a few other things I think are more important. -- Tim Starling 15:57, 20 Mar 2004 (UTC)

If it is, are all the people listed above able to do that job without mistakes ?

Would someone please comment on the target number of people with this access? I'm not convinced we should elect as many people as have been nominated. -- Tim Starling 15:57, 20 Mar 2004 (UTC)
I've voted "oppose" on about half of the people above, on the basis that the number should be closer to 5 than 10. Of course, I don't have anything against any of the people I'm opposing (except maybe Wik). I had hoped to have some discussion on this before the voting started, but you can't have everything. -- Tim Starling 16:06, 20 Mar 2004 (UTC)
I suggest three as ample to start with and have selected who to vote for on that basis - the three I trust most to follow consensus and not act inappropriately based in part on how much experience I have with them. Jamesday
If there are more than three we can trust, I would say there is no harm to have five or ten people serving this position. Indeed, if we have more people, it is easy to revoke the status or allow them to resign. Tomos 00:02, 27 Mar 2004 (UTC)

Opposition by Maximu Rex

Nothing personal with the 'oppose' votes, I just think that 'developer access' should be in the hands of very few people, at least at the beginning. Maximus Rex 11:31, 22 Mar 2004 (UTC)

Opposition by Cimon

Perhaps I am just switching my foot from one cheek to the other, but let me clarify that my supports and oppositions were purely based on the idea that the hon-devs should be "anal retentive" types (sorry Angela, Mav, Martin!), rather than "justice-oriented" types. IMO rules are more important than justice here. -- Cimon Avaro on a pogo stick

Exclusion of arbitrators

Per Erik's comment. -SV We candidates who are also on the arbitration committee hereby promise, if supported for developer priveleges, do accept the nomination with the knowledge that a conflict of interest may result, and thus affirm: Upon received community support for developer priveliges, we will resign from the arbitration cmte.

So then current developers should not be arbitrators? And since when have developers been cops? Admins are cops too... Not that I want this position anyway (too many hats already). --mav
The problem of people having too many roles is exactly the one we are trying to solve here. We shouldn't replace one problem with another.—Eloquence 18:52, 23 Mar 2004 (UTC)
I have already been approved before for shell access so that I could help admin the foundation's website. I also want to eventually help with actual MediaWiki/ Wikimedia server work. Why should I be barred from doing that? --mav
With the new "steward" role, those roles are sufficiently separate to be held in parallel. I do not believe this to be the case for the two roles of "arbitrator" and "steward".—Eloquence 22:20, 23 Mar 2004 (UTC)

A further thought: I do not understand the idea of voting against Arbitrators at all. If I had my druthers, there should be a rule that at least one Arbitrator should be a "hon-dev" at all times. -- Cimon Avaro on a pogo stick 21:33, 22 Mar 2004 (UTC)

Arbitrators wielding substantial direct power reinforces the impression of a clique or power elite, especially by the user who is affected by the ruling. E.g. let's say a user is desysopped for violating protection policy. Martin puts the arbitration findings on the user's talk page. A minute later, Martin desysops the user. Now the user in question will focus their entire reaction on Martin. Shared responsibility makes for shared blame and creates a better impression, one of a working justice system.—Eloquence 18:49, 23 Mar 2004 (UTC)
With a sizable amount of respect, I think you are wrong. Separation of powers is good for a state, but not for a community of self-selected members.
If there is efficient liaison between the different institutions, at some moments it may seem to work seamlessly, and the checks and balances may "appear" to exert a beneficient effect. Most of the time, however, regardless of how good liaison there is, the system will resemble a Kafkaesque nightmare, the inter-connections of the different institutions will be so arcane, that the lack of transparency of the system will counteract any outward impression of conterbalancing forces. If Joe User cannot see and understand how the mechanism works, he will not be appreciative of its fairness. (more later..) -- Cimon Avaro on a pogo stick 20:05, 23 Mar 2004 (UTC)
To some extent I actually agree with you. I would prefer it if we would merge this new role and the old "bureaucrat" role into one, because yet another level is confusing. I do, however, believe that in this instance, separation of powers is essential to avoid the impression of a class of "super-users".—Eloquence 22:22, 23 Mar 2004 (UTC)

Uncle Ed

For the sake of simplicity I will refer to this new access level as "stewards" here.

We have a special situation with Ed Poor, who currently has developer access, but is neither involved with the actual development of the software, nor (as far as I can tell) with the administration of the servers. On the other hand, Ed is an active mediator, and has used his developer/bureaucrat privileges to make sysops on a regular basis (he has never desysopped anyone, I believe).

I propose that Ed should be given the choice to be either an actual developer, with the caveat that in this case we expect him to in fact help with the development of the software and/or the server administration, or to be a steward, in which case his developer access should be revoked.

Any thoughts?—Eloquence

What developer access does he have? CVS perhaps? Anything else? Jamesday 01:27, 1 Apr 2004 (UTC)

I agree with Eloquence. I haven't written or even looked at a single line of code. I originally was given Developer access by Jimbo, because I offered to help find a better database schema. Brion gave me the ssh password. But I gradually became convinced that it would take far more time than I was willing to devote, for me to become sufficinetly conversant with the source code and database schema to make any significant contributions.
I wound up being the SQL Query Guy. When consensus formed on the mailing list to promote a contributor to Admin, I ran the UPDATE user SET user_rights query to do this. Now Bureaucrats have a webpage programmed for this task.
It might hurt my feelings to be "demoted" from Developer -- especially if I lose the "Steward" election (I'm way below 80% right now). But I have consistently called for a more democratic process. I could have quietly made myself a Bureaucrat via ssh and SQL, but I didn't -- for one very good reason. I believe Wikipedia should be less authoritarian and more democratic. I initiated the process of "voting on bureaucratship", using myself as the test case. I refused to even certify or promote myself, because I wanted everyone to see that I DO HOLD MYSELF ACCOUNTABLE to the community, even if I'm quick on the trigger sometimes.
And if (like Benjamin Franklin) I wind up holding no office in the new republic I will still have my "place in history" as a founding father <pats self on back>. -- Ed Poor, April 8, 2004.

We don't need super-administrator

I really don't see the benefit of removing developer's "powers" to give them to Wikimedia wide super-administrator. If super-administrator role is just to follow instructions of community, why developers can't do it? Otherwise, if super-administrator can give or remove administrator "powers" as they want, I strongly oppose to this new status. Perhaps you had (many?) problems on English Wikipedia with administrators abuse but, for example, we didn't ever have this kind of problem on French one. Moreover, I personally have many more trust in Brion and Tim who helped us a much, than in Mav or Angela I don't know at all (only by reputation).

Imho, Wikimedia wide "powers" are only relevance in case of starting a new Wikipedia to assign first administrators. After that, each community (I mean each language) must be able to administrate itself. We don't need super-administrator especially if they don't speak our language.

With bureaucrat-administrators, each community can evolve without disturbing programmers and except for remove administrator "powers", we are autonomous. I don't think that remove administrator "powers" is something as frequent as it may bore developers. And if you think we must be 100% autonomous from developers; give this "power" to local administrators. No one better understand community choices than people from this community. If you are afraid about this "power", I suppose you can easily limit this use to 1 status remove per day (I hope we will not need it at all).

For me, we may have only 3 statuses on Wikimedia:

  • Developers: who create the software and do low level maintenance on the server (all language).
  • Administrators: who do all maintenance task (language specific).
  • Moderator: who do arbitration in irresolvable dispute (language specific).

Aineko 02:25, 24 Mar 2004 (UTC) (sorry for my poor English, but I did my best)

Who are our moderators ? And since when do moderators do arbitration ? You are confusing roles Aoi. As for this status, I have been trying to raise discussion about it since the 13th of march ([1]). Some understood. I already indicated to Tim that the few french people who had given their opinion on that were not hot with the idea, I also mentionned the issue on wikipedia-l. However, when I tried to promote discussion about it on fr, you moved all the discussion toward the admin status, and I could never succeed to have it refocus on this. I think I already gave as much arguments as I could think of on fr, in french. Perhaps you did not read them ? Here is my suggestion. Fix, let us fix the admin role as we have been trying to do in the past days. Then, let it rest for a while, and let's do start the discussion on power structures again. We already talk about it with Tim a bit; some things are possible, others are not for the moment. There is no reason to hurry. Meanwhile, Shai can take care of this. Anthere 05:38, 24 Mar 2004 (UTC)
I didn't confused, I just spoke about an other kind of moderator (arbitrator?). Sorry, it's really a pain to me to read and wrote English about important subject (I don't found the right words), so I will stop now. I don't think multiply status is a good thing, but do as you want, future will judge. Aineko 08:09, 24 Mar 2004 (UTC)
C'est pas grave. On en reparle en français, sur fr: :-) On a toujours un status d'admin sur le feu... Sur fr, nous n'avons ni modérateur, ni arbitre. La modération se fait au petit bonheur la chance. et la dernière fois que nous avons eu besoin de prendre une décision grave (hmmm, Papotages d'accord ?), cela s'est fait par vote général. Nous sommes encore suffisemment peu nombreux pour que cela marche relativement bien :-)
There are several reasons why removing developers' powers and giving them to elected representatives is a good idea. Elected representatives are chosen on the basis of merit rather than technical ability, and hence are, on average, better suited to make these kinds of decisions. It's very difficult to demote a developer for performing non-technical tasks incorrectly, because by demoting them they lose the ability to effectively test and implement software. And developers have better things to do than to check up on the trustworthiness of people starting wikis, and to follow debates on de-adminship. To require them to perform non-technical tasks distracts them from those tasks which no-one else can do.
I agree with you that local administrators are those best able to make decisions regarding a community. Allowing them to do so remains a challenge, both technically and politically, but be assured I'll continue to work towards it. Separation of developer powers is just one step along the road.
In particular, I would very much like to enable each wiki to choose its own power structure. At the moment I can't implement the power structure that the French people want for fr, because the English people don't want it on en. That clearly has to change. -- Tim Starling 04:10, 24 Mar 2004 (UTC)
The current election is taking place in English, so only English speaker can participate. I strongly doubt about its Wikimedia wide representativeness. If you really want to create international-super-administrator, one of the more important quality must be to the ability to speak many languages, and candidates must be choose to obtain the largest panel of language as possible. I think it's why Brion made a perfect international-super-administrator (even if he was developer). Sure, technical ability is not necessary a quality, but it's not a vice. If a developer abuse of their "powers", why don't simply doom him? Create new layer of power don't solve any problem if higher level people don't respect rules. And I really don't like the idea to hierarchize the administrator! Either an administrator does its job, or he must lose its status. What's the problem? What we need is:
  • Clearly explain an administrator is just for maintenance, he must not act by himself (administrator are not moderator) and male rules.
  • Make a maintenance version of Recent Change page where *everybody* may see all maintenance acts (page block, IP and user block, deleted page and image, admin nomination, etc.).
I'm glad to see that some people care about local community problem, and hope your project to give them some autonomy will rise a day. Aineko 08:03, 24 Mar 2004 (UTC)
Your criticism that the election is taking place in English is a fair one, and your suggestion that these people should be able to speak many languages sounds very reasonable to me. I can only apologise for the current voting situation. I didn't make it clear that the I intended the vote to be a two-phase process: first nominating and discussing, then once we had sorted out all the issues, voting. Of course we could organise for the relevant text to be translated, and we could find multilingual candidates, but it seems a bit late now. Perhaps we could attempt to cancel the vote and start again.
I also agree that super-administrators should not act on their own behalf. If an administrator breaks the rules, on en: we have the arbitration committee to deal with that, and the role for developers is fairly small. It may be a small role, but I still believe it's best off in the hands of a non-technical user. -- Tim Starling 00:04, 25 Mar 2004 (UTC)
I agree. The new "bureaucrat-status" can be expand whit the power to remove special user powers. The idea to include a failsafe like only 1 de-sysoping per 24 hours is a good idea. I also do not see the need for wikipedia wide power. One user whit sysop/bureaucrat powers is all you need for a wikipedia and that can be done by a developer when creating the new wiki. Walter 11:10, 24 Mar 2004 (UTC)
Expansion of bureaucrat powers can be discussed as a separate issue. It's true that the developer role is small, but to give this role to people who do not have security of tenure ensures that it stays small. At the moment, there are no significant balances to developer power, and that is the primary thing I'm trying to change. -- Tim Starling 00:04, 25 Mar 2004 (UTC)

Should the current developers maintain their rights ?

Discussion

Will the current developers maintain their rights to do what they have been doing, or do they need to be nominated here? Angela 09:07, 24 Mar 2004 (UTC)

As far as I understand Tim, only developers who would be specifically approved to also act in this new role would be allowed to do sysoppings and desysoppings. Developers would only use their access for low level maintenance, software updates etc.—Eloquence 16:30, 24 Mar 2004 (UTC)
I'm not sure there's any point voting on developers. They have to be trusted by default as having real developer access gives them far more abilities than this pseudo-developer access will. For the same reason I suggested all sysops be bureaucrats, I suggest all developers are "stewards", or whatever it will be called. It seems pointless to vote on people who already have the power to do things stewards will be able to do. If they can not be trusted to do simple tasks like unsysop people, they shouldn't be given developer access, and if they already have been given it, people need to accept they have that ability and that voting oppose will not change that. Angela 08:58, 25 Mar 2004 (UTC)
The issue is not one of trust but rather one of policy. Many real-world systems have administrators who have the ability, but not the authority, to implement policy changes. In a project of this size, it is important to separate the technical authority from the policymaking authority. While nominally, policy is made by community consensus, there is fairly wide lattitude given to the developers in "reading" the community consensus. This "reading" of policy is best made a role separate from purely technical work. Tim Starling makes some important related points above, and I agree with what he says. Also, there is a brief essay on a closely related topic at MeatBall:TechnoCracy. UninvitedCompany 17:26, 25 Mar 2004 (UTC)
I don't believe the stewards are supposed to have any role in policy making. Policy implementation maybe. I'm just not happy about having to vote on developers. If they are not supported here, or even not nominated here, and someone asks them to do something that would normally be done by a steward, are they expected to refuse to do that? What if it is some sort of emergency, like new sysops need to be made to combat a vandalbot? Angela 19:46, 25 Mar 2004 (UTC)

I support UC on this entirely. Also, see the initial reason given by Tim [2] for setting this up.

I give my entire trust to Tim, Brion and Erik (with the "report" rule set "on"). So the question is more one of policy and principle than trust itself.

yet you haven't voted for them? Angela

Ed is a different matter, because I remember well he was precisely made developer, not to develop himself but to help sysoping people. In short, Ed was our first steward, even if the status did not exist.

I would gladly vote independently on the 4 of them, but by default developers should *not* be steward.


Finally, the emergency point does not hold. Because, there would be enough stewards to do the job. The issue for people is to keep remembering that they have to ask help from another person than before. Oldbies will find it difficult for a while, but newbies will get it quickly.

If someone ask them, they can then direct him to you Angel.

Anthere 06:35, 26 Mar 2004 (UTC)

Thanks for your replies. I think the problem is that I was seeing this as a new user type, which would be in addition to the existing developer one, rather than a replacement of developer rights. I realise Tim said he did not want developers doing these things, but are they actually going to be banned from doing them once this scheme is implemented? Can there be emergency provisions for cases where there are no stewards available? Not voting for developers for here seems to be supporting the removal of their abilities, which is not something I want to vote for, but equally, I don't actually support all current developers; I've never even heard of some of those listed on en:Wikipedia:Developers. Angela 01:08, 27 Mar 2004 (UTC)

In the mailing list post I said: "A policy should be instituted disallowing any developer from using their power for administrative purposes, except where there is no other way to perform the relevant operation." It's possible that an emergency response to a vandalbot could fall into the "no other way" category. If a developer can fight a bot more quickly and effectively than stewards or sysops, I would expect them to do so. -- Tim Starling 02:37, 27 Mar 2004 (UTC)
The simple truth is we don't have time to do these things, and they thus generally go without being done. Further, dealing with sysopping/desysopping and user bans can bring in WikiPolitics, which is a big turn-off, particular for those of us not very involved in whichever wiki the fighting parties are on. So basically, we want to a) take responsibility off of us developers (overworked volunteers trying to concentrate on technical issues) and let the people who'd like to deal with political issues all the time and are more plugged-in to that issue do so without putting them in a position where they could accidentally (or maliciously) abuse the developer powers (ie, installing a stealth block against someone they didn't like, altering edit histories without an audit trace, etc). [Here I use "developer" to refer to people actually maintaining the software and server.] --Brion VIBBER 04:17, 27 Mar 2004 (UTC)

I have no problem with other people taking care of sysoppings, desysoppings etc., as long as the judgment these people are allowed to exercise on their own is limited. These are really just routine tasks, especially now that we have the quickpoll system to make the actual decisions. There is no reason for a real developer to desysop another user within such a framework, such actions were more born out of necessity than anything else. In cases where I desysopped users - 172 or 168 (always the people without real usernames!) - it was because no efficient mechanism existed to do so. With quickpolls this is no longer the case.

I believe that the high level of opposition we see to virtually every user in this selection can be in part explained because the roles of this new class of users are not particularly well defined. I see them as a kind of super-bureaucrat class who do the same things bureaucrats already do on a local level - take care of administrative tasks that we don't entrust every single sysop to do. Any abuse of power could easily be punished, so there's really no reason to be excessively conservative.

In the interest of decentralizing power, I fully support that developers should not simultaneously hold this super-bureaucrat role (although they should be allowed to hold the regular bureaucrat role), as long as we also separate this power from the arbitration committee -- i.e. no arbitrators as super-bureaucrats.

If we do separate these powers, I think there should be no exceptions, period. No voting on developers, because no developers are super-bureaucrats. As such I would support removing all developers from this page, with the exception of Ed Poor, who should be given the choice of either remaining a developer (under the condition that he actually does some development), or becoming a candidate for super-bureaucrat.—Eloquence 15:08, 27 Mar 2004 (UTC)

discussion moved below


Just a quick interjection re: Erik's comment above about Ed Poor. Ed is not and never has been involved in the development of the software or the management of the machines; he is not in any real sense of the word a developer in this organization. For a while at least he had write access to the database because he volunteered to take care of some administrative tasks (exactly the kind of stuff we're talking about with these 'super-bureaucrats'); the marker in the user fields for this was called "developer", but that doesn't make him an actual developer any more than I'm a plant's ingestive organ because I've got "root" access to a number of machines. Ed is not a developer here, and should be considered distinctly from, say, me and Tim and Erik who spend a lot of time building, debugging, installing, and maintaining the software and directly control the configuration of the software and the databases. --Brion VIBBER 03:47, 28 Mar 2004 (UTC)

Ed is listed as a "developer" various places on en. He was doing sysop promotions, and renaming the occasional user, for some time (and I think still does). Pakaran 06:06, 28 Mar 2004 (UTC)
Exactly my point. --Brion VIBBER 06:51, 28 Mar 2004 (UTC)
Thanks for your support, Pakaran, but Brion is 100% right about this. I should either become a real developer, or wear a different hat. I'm willing to become a Steward, if asked (I appreciate mav's nomination). --Ed Poor 17:33, 8 Apr 2004 (UTC)

Policy

If we do separate these powers, I think there should be no exceptions, period. No voting on developers, because no developers are super-bureaucrats. As such I would support removing all developers from this page, with the exception of Ed Poor, who should be given the choice of either remaining a developer (under the condition that he actually does some development), or becoming a candidate for super-bureaucrat.—Eloquence 15:08, 27 Mar 2004 (UTC)

I support all Erik comments, but one. Quickpolls exist and are recognised on en. Not necessarily in other places. And probably not in any other places. So, general rules are important as well. We can't rely on en rules to settle all cases. I suppose there should be general rules, and that local specific rules can impact general rules locally

Anthere 17:25, 27 Mar 2004 (UTC)

I think we should test this and similar systems, and at some point in the future standardize on a general decision making process for cases like this, one which can be used by all Wikimedia projects. Quickpolls haven't proven themselves yet, and the procedure may have to be overhauled a bit before it can be put in wide use. Maybe there's a completely different scheme that works a lot better. But I do think that (super-)bureaucrats should only act on the basis of community input, and that this should be part of the specifications for this new user class -- no unilateral action.—Eloquence 17:33, 27 Mar 2004 (UTC)
probably right. Though, on small wikipedias, quickpolls are just regular voting sessions. That is just a question of perspective. Obviously, the goal should be nothing unilateral; And that should be easy to achieve on en, de, fr ja...However, it is more tricky when it comes to making a sysop on a wiki where there are only 3 users. And only one can communicate in a common language. In this case, we reach the limit of unilateral action. Because there is no other way around. If the only user able to communicate ask to be sysop on a wiki of 3 editors, this will be unilateral action, and this will be ok...till the two others do not agree and find a way to tell us they do not :-( No quickpoll here can solve this equation Erik Anthere 17:43, 27 Mar 2004 (UTC)
True, but at the very least the adminship should be negotiated on Requests for permissions here on Meta, with up to a week of time to raise objections.—Eloquence 18:01, 27 Mar 2004 (UTC)
Nod. Absolutely. Perhaps 1 week on Requests for permissions for sysop. 2 weeks for bureaucrats. Unsysoping only per a certain number of trusted user request or per evidence of critical vandalism upon one person request (to be confirmed later by a vote on the local wikipedia) ? For example ? ant

To clarify what Anthere says above, I don't think "no unilateral action" should be a "goal". I think it should be an <expletive> struck in stone requirement, no ifs and buts. -- Cimon Avaro on a pogo stick 00:17, 28 Mar 2004 (UTC)

質問

質問. どんな言語が話せますか? = Question. Which languages can you speak?

Sorry to wrote comment in Japanese, but I wanted to make you picture how it may be sometime frustrating for non English people to see all important discussion taking place in English. Sure, we need a common language to be able to communicate easily between WPs and English is perhaps the best choice, but I think we must promote other languages when it's important. Steward (super-administrator) will have to be in close contact with all communities, so elect English only speaking people will make them approachable only by English speakers. I really don't like the idea that in a community, only English speaker may be able to make them understood.

I propose that each WP (or, for example, each WP with more than 100 active peoples) elect its own "steward" (super-administrator). Of course, the "steward" must speak English, but not only. Like this, all active WP will have "steward" approachable by every one, as well English Wikipedia will have many "stewards" (all, in fact).

Aineko 02:09, 29 Mar 2004 (UTC)

There's a guideline called State your point, don't prove it on the English Wikipedia which is very applicable here. Adding lots of incomprehensible text to an already chaotic page is not going to help anyone.—Eloquence
English is incomprehensible language for main part of humans and meta is not the English Wikipedia. Meanwhile, I sorry if I disturb the debate... feel free to clean all you think that must be cleaned. Aineko 06:40, 29 Mar 2004 (UTC)
English is the most diversely spoken language in the world. It is the most widely learned second language and the second most widely spoken first language. Last but not least, it is the most widely spoken language on the Internet. By any reasonable standard, English is the one and only world language. Learning English is of course only an option for people who can afford the education to do so, and who are still able to learn things. Since Wikipedia depends on both of these conditions, it is not unreasonable to ask people to learn English if they want to participate in certain project-wide decisions, or at the very least to communicate through an appointed ambassador. I was born in and live in Germany, I have only visited the United States once. Yet by most accounts my written English is on par with that of the average native speaker. I am by no means extraordinary. While I recognize that English is easier to learn from an Indo-European basis, I also find the "DON'T SPEAK ENGLISH! NOBODY CAN UNDERSTAND IT!" attitude that is often displayed among some members of the non-English Wikipedias quite absurd. Learning English is free and reasonably easy, it is an essential skill in a globalized world, and it is certainly a practical approach to handle quick communication in a multilingual project.—Eloquence 07:23, 29 Mar 2004 (UTC)
From my previous post : "Sure, we need a common language to be able to communicate easily between WPs and English is perhaps the best choice". I just reacted seeing Japanese treat as incomprehensible language. Aineko 07:47, 29 Mar 2004 (UTC)
Japanese is an island language, a language isolate. Japanese culture was for a long time highly encapsulated, as a result the language is in many ways the exact opposite of English. To almost everyone outside Japan, Japanese is indeed incomprehensible, whereas about half the world population have at least some rudimentary English skills. Add to this the complexity of the Japanese alphabet and you have to seriously wonder if Engrish will gradually take over as the dominant language for practical reasons alone. I personally find it absolutely wonderful that we have one language which we can use to communicate across all borders, and I think we should promote this trend wherever we reasonably can.—Eloquence 08:15, 29 Mar 2004 (UTC)
Rudimentary English skill is unfortunately not enough to explain things precisely and participate in important debate. But, one more time, I never denounced the use of English as an international language. I also find wonderful to be able to use it in many countries (as I find wonderful to be able to use Japanese to communicate with Thai or Indonesian people living in Japan).But I just want to say that speak English must not become a required skill for wikipedians to be heard. I think it's very important, when possible, the people who have some "power" on a community can understand the language of that community. The "when possible" refer the fact that small Wikipedia perhaps don't have (yet) enough people to be able to elect a trusted user. But it's not the case in many other languages (as German, Japanese, French, Dutch, etc.). I don't see any point in the steward description that makes this status impossible to be gave to the more trusted users of each language. So why don't elect steward in each language?
Of course, if steward don't act into "structured" community, the language skill request don't apply (but still a very important quality I think). Aineko 09:28, 29 Mar 2004 (UTC)
Any stewards should be able to speak English, because they need the trust of all Wikipedians. They should at least be basically known in the English-speaking sphere of Wikipedia. That is not the case for bureaucrats -- these can be only locally known. But stewards (I prefer super-bureaucrats) have access to all Wikipedias, and that means they should speak the most widely spoken language on Wikipedia: English. This requirement impedes the election of a steward for the smaller —Eloquence 10:47, 29 Mar 2004 (UTC)
Oki if international-bureaucrats (super-bureaucrats, stewards) act only on Wikipedia where there are no local-bureaucrats available. Aineko 11:09, 29 Mar 2004 (UTC)~
English also lacks phonetic spelling, has tons of irregular verbs, is harder to learn from the Romance languages than they were from the others, etc. As for whether learning English is "essential" - it is to speak with English speakers. Similiarily, learning Cobol is essential for maintaining Cobol programs. I would suggest that learning Cobol is not needed to work on Mozilla, and that perhaps learning English should not be a requirement of Wikipedia, at any level. It would make it more convenient to coordinate with the others here, to be sure. The same can be said of hanging out on IRC. There are sysops and bureauctats who do neither. Pakaran 16:51, 30 Mar 2004 (UTC)
Я понимаю проблему, но я не соглашаюсь, что должны быть множества стюардов -- существует другое решение к проблеме? Возможно there's a way to limit a стюард to one Википедия? -- BCorr|Брайен 04:25, 29 Mar 2004 (UTC)
I do understand problem, but 4 I do not agree that must be many stewards - there is another solution to the problem? Is possible there's a way to limit a steward to one Vikipediya? [SYSTRAN]
Supongo que no hay problema técnico para hacer eso. Aineko 04:39, 29 Mar 2004 (UTC)
I suppose that there is no technical problem to do that. [SYSTRAN]
I only speak English fluently, but I don't think this should be a problem. All the stewards will be able to do is change user statuses, such as creating sysops or bots. We could make up a list of standard phrases that people can use to request this, so people can ask in their own language, and the stewards can compare that on a table, like:
FrançaisEnglish中文
Faites à Utilisateur:Aoineko un Sysop sur le fr: svp Please make User:Aoineko a sysop on en: 请做 User:Aoineko 管理员在 zh:
Angela 05:12, 29 Mar 2004 (UTC)
Users are not bots, they don't speak with this kind of simplistic sentence ;o) And how can you know the validity of the request if you can't read the debates? Local wikipedians are better judge, don't you think? For example, consider the problem of Stuart who tried to ruin the reputation of Anthere. What happened if Anthere was unable to justified herself in English? My question is what is the problem to have "super-administrator" (steward) for each (active) Wikipedia? Aineko 05:37, 29 Mar 2004 (UTC)
The bureaucrat level is the super-sysop level which is local to each wiki. I think any future single-wiki powers should be given to them. The thing that sets the steward level apart from the other levels is its interwiki nature. Unfortunately the power structure at the moment (pre-steward) gives more power to those who can speak English, because it is those people who are able to confidently contact developers and explain their situation. I think empowerment of local users is important in solving this problem. Having stewards able to speak German, French, Japanese and English won't help with this because those wikis already administer themselves. The wikis which need help are those with just a few active users -- that is where developer judgement is often requested. BTW I've added English machine translations above to save people the few clicks required.-- Tim Starling 05:24, 29 Mar 2004 (UTC)


Hm. Aoineko may have a point here. How is the steward supposed to know that he is approached through the normal channels, and not by an individual plugging a single other individual? Preferably, a steward can read a discussion page. Now, knowing some background of the language, one can use babelfish to translate a Japanese discussion into English, but that service does not exist for fi: or ee:! OTOH, I think having a steward from each language would be wrong. It might help if every proposed steward made a list of the languages (s)he can read, thereby answering the question. It may be possible to select a combination of stewards that together understand many different languages. Rob Hooft 05:33, 29 Mar 2004 (UTC)

Tim > If steward are really only there to help manage new Wikipedia and don't interfere in local administration, I suppose that don't speak local language is not (yet) a problem (even if I think it's still a great quality). The problem is with the description I read, I understood the steward status as a level over bureaucrat. Because I understood that only steward may able to remove status of administrators (additionally to have all bureaucrat features). If it's right, they absolutely need to understand the local language to be able to understand if the request is valid or not and need to be able to read debate. If steward will only act in small WP, I don't understand the concern of remove the "power" of help new WP from developers. I really don't want to mess your debate, but I just don't understand what the benefits to have a new status are? The worse is I feel everybody use word steward with a different meaning. Aineko 06:07, 29 Mar 2004 (UTC)

I suggest Fuzheado could help here, because he probably has at least one language we do not :-) Anthere 06:11, 29 Mar 2004 (UTC)~


The problem is not new. While creating stewards will not solve the problem, it will not create one. (Well...maybe fr. has not experienced this problem before because they have its own developper and Brion?)

Currently, developers do not know all the languages of wikipedias they help. True, it would be nice to have a lot of polyglots like Brion, and true, the developers who have been helping japanese wikipedia knew a bit of Japanese, and that made difference (even when our explanations were in awkward English, they could take a look and understand the problem, for example. Another time, they knew language-related problems existed even before japanese people noticed it). But it is simply too much to ask developers to know five languages, etc., I think. Stewards are no different.

A number of solutions exist, although none is perfect.

First, we can try to limit the liaison between one language wikipedia and stewards - by making a rule "unless a request is made through a designated liaison user, who is elected in the language-wiki, the request would not be considered official." Or, stewards can call help from others (trusted users, ambassadors, etc.) to understand the situation.

Second, we can try to have some interlingual interactions so that we can get to know each other and build some trust.

Third, each language wikipedia can make effort to present simply-written request, (along with an easy-to-verify evidences when necessary,). An example I can remember is here: [3] Each language wikipedia may create a page titled "Wikipedia:requests to stewards" and use the page to present requests that achieved concensus within the language wikipedia.

The first one is kind of un-wiki, and the second one may create a cabal-like group. So I'm inclined to think the third is the best, but probably the combination of all of them would be a more realistic approach.

Also, related to the third solution, if stewards are expected to make no judgment of their own, they should receive just requests. It is not steward's job to read a complex discussion and make his/her own judgment.

Tomos 06:26, 29 Mar 2004 (UTC)

> "unless a request is made through a designated liaison user, who is elected in the language-wiki, the request would not be considered official."
Why not simply allow this trusted user to act as a steward? Why must be required to wait an trusted user from an other Wikipedia made it? What happen if a sysop freak out and we can't block him just beacause all distant stewards are sleeping at this time? Aineko 06:37, 29 Mar 2004 (UTC)

Proposition

What about make bureaucrat able to remove administrator's status (for example 1 per day for security purpose) and simply make steward an "international bureaucrat" who can act as bureaucrat in all language where there are no bureaucrats? Aineko 10:37, 29 Mar 2004 (UTC)

That doesn't handle setting bot flags, notably. Pakaran 15:45, 29 Mar 2004 (UTC)
Someone said that bureaucrat may have access to all community specific features, so I suppose it include setting bot as well. Aineko 01:18, 30 Mar 2004 (UTC)
At this point, it doesn't. The interface looks something like this.
Make a user into a sysop
From Wikipedia, the free encyclopedia.
This form is used by bureaucrats to turn ordinary users into administrators. Type the name of the user in the box and press the button to make the user an administrator


Name of the user: _________
Set bureaucrat flag [ ]
[BUTTON]
Pakaran 01:34, 30 Mar 2004 (UTC)



I've noted that there's some concern that some of the proposed developers speak only English. I should point out that most of the en sysops do speak 1-2 other languages, if only from our college courses etc, and could at least handle "please make x our first bureaucrat on foo.wiki" type requests. Pakaran 15:45, 29 Mar 2004 (UTC)

I for example speak English natively, and speak Spanish fairly well (I don't mentally translate anymore) though I can't necessarily listen to people talk at native speed.
Oh, how long do votes stay up on here? Pakaran 17:44, 29 Mar 2004 (UTC)

On a side issue, though I only really speak french and english, I strongly believe on the building of trust relationship along the days/weeks/months. For example, Walter (be:) and I only share weak english, but we shared a bouple of fights together (hummm, we were the first two non english to invest meta for example :-)), and I have known him for some time. I know I can use him as a reference. If some belgian user I do not know was to ask me to unsysop someone, I will gather second opinion from Walter before doing it. This can't be done for all wikipedias probably, but I believe super-bureaucrat would benefit in investing themselves in international relationships (development, articles edition sharing, embassies, pr, whatever). Anthere 17:55, 29 Mar 2004 (UTC)

Need to finish voting

I suggest that we wrap up the first phase of voting soon based on this: Any user which gets at least 80% yes votes should become finalists. Then we can have a separate vote on what user types should be disqualified (such as users who are on a wiki's arbitration committee). Last but not least, each finalist who had not been disqualified for some reason will have to accept the nomination. Yeah, I know this is all backwards, but it should be done nontheless. --mav 04:03, 5 Apr 2004 (UTC)

You can manage it yourselves. I don't see any further need for developer involvement. You might need these: [4] [5] [6] [7] -- Tim Starling 05:07, 5 Apr 2004 (UTC)
See Stewards. I've proposed a week's discussion and a week's voting. Someone should remove the this is only a proposal if there are no objections. What about people wanting to apply now that the first round of voting has finished? Angela 06:41, 5 Apr 2004 (UTC)
I don't think it's appropriate to close the vote, then decide how to interpret the results. I had assumed that a simple majority would be sufficient, and I think many of the candidates and voters had as well. Let's make it clear that 80% is required, then allow one more week for voting on this round. People should also be allowed to change their vote. -- Tim Starling 11:25, 5 Apr 2004 (UTC)

I hate to point this out, but it is completely backasswards to vote on stewards before deciding what they will actually be allowed and able to do.--—Eloquence 20:49, 5 Apr 2004 (UTC)

It is still beter than saying something in a post to the mailinglist, holding a 2-day vote and execute the result. (not personal) Walter 21:26, 5 Apr 2004 (UTC)
Actually, now that you mention it, that sounds like a pretty good idea ...--—Eloquence 21:41, 5 Apr 2004 (UTC)
We already know what they will be able to do in the short term. We vote on sysops without knowing what they will be able to do in the long term, is that "completely backasswards"? -- Tim Starling 00:05, 6 Apr 2004 (UTC)
There are no plans to grant sysops any far reaching powers that I know of. On the other hand you have expressed a quite far reaching vision for what stewards should be able to do in the future. Yes, I think it is backasswards to select these people first and then decide on what exactly they will be allowed to do. We are effectively asking people to vote under the condition that "eventually, this thing or that thing may happen, but we don't know yet". I'm not asking for a detailed roadmap, but we should at least agree on the basics. Oh well.--Eloquence 00:11, 6 Apr 2004 (UTC)
Sysop power has been fundamentally altered over the last two years, such that it is barely recognisable. Once upon a time, sysops could only protect, delete and rollback, now they can block whole ranges of IP addresses, block usernames, change the interface text and retrospectively hide edits from recent changes. It's planned to give them the ability to delete specific revisions, and some users have asked for sysops to be given the ability to forcibly change usernames and to change attribution for edits. The biggest change, in my opinion, was the ability to block logged-in users, which elevated sysops from users with control over articles, to users with the ability to punish other users at their whim. The difference between my idea for steward power and your idea for steward power is insignificant in practical terms compared to the changes to sysop access in recent times.
I am willing to discuss this, however. I just don't have much time for this sort of thing. -- Tim Starling 03:30, 6 Apr 2004 (UTC)
On a technical level, you are correct. On the non-technical level, things haven't changed that much. Sysops are not allowed to "punish users at their whim", and if they do so without reason, their privileges will be revoked. Your steward proposal encompasses both technical and social elements. We don't have to discuss this if we don't go beyond treating stewards as super-bureaucrats. If you do want to go beyond that, I think we should discuss it now.--Eloquence 03:35, 6 Apr 2004 (UTC)

OK, I think our positions are closer than we thought. But first I'll digress into a description of how I see the Wikipedia power structure.

Currently there are two power structures: the technical and the social. I like to think of the latter metaphorically as the "civilian" power structure. The technical power structure is subservient to the civilian power structure. The technical power structure is vertical, with well-defined levels of access. The civilian power structure is more democratic. Jimbo is the accepted leader, and delegates power to the mediation and arbitration committees. The defining characteristic is consensus-finding. A consensus among users beats just about anything, arguably even Jimbo's opinion. Because of the lack of a strict ladder, the overall opinion of the civilian power structure can be hard to judge.


The pre-steward technical power structure looks something like this:

     Jimbo
       |
   Developers
      |||
   Bureaucrats
      |||
     Sysops
      |||
 Registered users
      |||
Unregistered users

I have drawn Jimbo's connection to the developers with one bar, because it is unlikely that it will ever be used, it's just theoretical. Jimbo's place is that if a developer changes the root password and locks everyone out, Jimbo can go down the server room and sort everything out.

The role of people with technical power is to interpret policy and community consensus. They boldly perform actions which they expect will be uncontroversial and beneficial. For potentially controversial actions, they seek approval from the civilian power structure, or attempt to find a basis in policy. When someone uses technical power which is against the wishes of the civilian power structure, there needs to be technical means in place to allow punishment. Punishment typically comes from a higher rung of the ladder. The higher levels are thus charged with judging the community's desire for punishment.

People with technical power unavoidably have some social power. This social power should be minimised. The power primarily comes from the fact that there may be differing interpretations of community opinion. Those holding technical power may be biased in their choice of interpretation. Technical power also leads to social power if the community is unwilling or unable to punish unilateral, unpopular action. This is undesirable.

My proposal is to change the technical power structure, but not the social power structure. Developers currently must interpret the wishes of the civilian power structure in areas such as:

  • Desysopping
  • Forced name changes
  • Creation of bureaucrats on new wikis
  • Software settings such as:
    • Enabling the proxy blocker
    • Disabling external images
    • Choice of logo
  • Autonomy of wikis versus centralisation

My proposal is to allow stewards, rather than developers, to perform that interpretation. Hence, the power structure will be more like this:

     Jimbo
       |
   Developers
       |
    Stewards
      |||
   Bureaucrats
      |||
     Sysops
      |||
 Registered users
      |||
Unregistered users

Now I know you prefer centralisation (or whatever you want to call it) over autonomy. But that is not a decision for the technical power structure to make. As a developer, the limit to your role in this issue should be to interpret community opinion. Of course, you can wear two hats, and argue your point of view on the mailing list.

At the moment, developers not only have the ability to choose how they interpret community consensus on this issue, but they may also refuse to code a power structure which does not agree with their personal views. I'm suggesting to remove the first power by transferring it to stewards, and to remove the second power by coding a general system which can be used for any purpose.

By coding a general system, I am not supporting some particular application of that system. Instead, I am supporting the principle that interpretation of community opinion should be performed by democratically selected users, rather than people with technical skills.

-- Tim Starling 04:58, 6 Apr 2004 (UTC)

I agree with Tim. --Ed Poor, April 8, 2004

About language that candidate cam speak

Angela:

  1. 質問. どんな言語が話せますか?(英語以外) Aineko 01:29, 29 Mar 2004 (UTC)
  2. Question. Which language can you speak?
    私は英語しか流暢に話さない. (English is the only language I speak fluently) Angela 04:50, 29 Mar 2004 (UTC)
    Thanks for the effort :o) Aineko 06:18, 29 Mar 2004 (UTC)

Anthere:

  1. Question. Which language can you speak?
    Français, Anglais, Suffisemment d'Allemand pour ne pas trop stresser, un peu d'Italien. Et le language des fleurs.

Maveric149:

  1. Question. Which language can you speak? Aineko 01:29, 29 Mar 2004 (UTC)
At one time I was semi-fluent in Spanish and could speak a bit of Latin but I speak only English nowadays. It's been a long time since I've had much practice speaking/writing anything but English but would like to give it a go again. Esperanto looks interesting (as does Interlingua). --mav

Dysprosia:

  1. 質問. どんな言語が話せますか?(英語以外) Aineko 01:29, 29 Mar 2004 (UTC)
  2. Question. Which language can you speak?
Français, mais je ne pas parlez tres facile - mais je comprend Français bien. Dysprosia 05:15, 29 Mar 2004 (UTC)
Merci beaucoup de ta reponse. Aineko 06:18, 29 Mar 2004 (UTC)

Looxix:

  1. Question. Which language can you speak? Aineko 01:29, 29 Mar 2004 (UTC)
フランス語 も 英語 も オランダ語 (french, english and dutch) and can also somewhat decrypt a bit of german, italian and a little bit of dansk. -- Looxix 17:52, 29 Mar 2004 (UTC)
Merci :o) Aineko

User:Shaihulud:

  1. Question. Which language can you speak? Aineko 01:29, 29 Mar 2004 (UTC)
    1. French, english and I can read (a little) German Shaihulud

Wikibooks:User:Karl Wick:

  1. Question. Which language can you speak? Aineko 01:29, 29 Mar 2004 (UTC)
Besides English, I am fluent in Spanish. --KarlWick

André Engels:

  1. ??. ?????????????????? Aineko 01:29, 29 Mar 2004 (UTC)
  2. Question. Which language can you speak?
I am sure André knows quite some, but not yours! Rob Hooft 05:06, 29 Mar 2004 (UTC)
It's not mine ;o) Aineko 06:18, 29 Mar 2004 (UTC)
Actually, I might speak Aoineko's language - that depends on whether you say French or Japanese is (I even speak some Japanese, but only a few dozen words at most, and not the writing system (well, romanji of course, but not hiragana, katakana or kanji)). My French is not that well, but good enough for written discussion (I would not try writing a Wikipedia article in French, though). I speak excellent English, Dutch and German, and have passive knowledge of Plattdüüts.
Apart from that, my experience in Wikipedia interlanguage links has given me quite some experience in scanning texts in languages I do not know. At that level (that is, reading an article on a subject I know, and then knowing what the subject is and have an idea of what is said in most lines) I can read most Germanic and Romanic languages as well as Esperanto. Finally, for languages where I don't even have that level, I use machine translations from the internet, which make Chinese, Japanese, Polish, Russian etcetera quite well scannable. The main exception here (apart from languages for which I have not found machine translation yet) is Korean, which seems to give horrible machine translations. - Andre Engels 10:56, 29 Mar 2004 (UTC)
????????????????????????I ?????????????????????????????????? - Andre Engels 11:26, 29 Mar 2004 (UTC)
Thank you very much. In fact, I'm a French man in Japan :o) Aineko 15:26, 29 Mar 2004 (UTC)

Arno Lagrange:

  1. Question. Which language can you speak? Aineko 01:29, 29 Mar 2004 (UTC)
en : As said above I speak french esperanto german and english and I understand romanid and germanid languages and a few of slavish languages. I read easily latine greek and cyrilic alphabets. I've created for myself accounts in about 40 wikipedias where I'm putting interlanguages links. I am very active in the eo-WP and a little in the fr one. I've programming knowledges.
fr : Comme dit ci-dessus je parle français espéranto allemand et anglais et je comprends la plupart des langues romanes et germaniques et un peu des langues slaves. Je lis sans problème les alphabets latins grecs et cyrilliques. J’ai créé des comptes dans près de quarante wikipédias où je crée des liens interlangues. Je suis très actif sur la Vikipedio (en espéranto) et un peu sur la Wikipédie (francophone). J'ai des connaissances en programmation.
eo : Kiel supre dirite mi parolas France esperante germane kaj angle kaj mi komprenas la romanidajn kaj ĝermanajn lingvojn kaj iom de la slavaj. Mi flue legas latinan grekan kaj kirilan alfabetojn. Mi malfermis kontojn en kvardeko da vikipedioj kie mi ligadas interlingve. Mi estas tre aktiva sur la esperanta Vikipedio kaj iom sur la franclingva. Mi havas programdajn kapablojn.
de : Ich spreche auch Deutsch und bin zu faul um diesen Text einem vierten mal zu schreiben.
ArnoLagrange 07:23, 1 Apr 2004 (UTC)


Discussion

Hmmm, i am not sure I understand completely the difference between a bureaucrat (I'll shorten it to bu) and the new honorary developer (I'll call them hondev for simpicities sake, no, wait, I prefer stewart more, just read it). Well, sure, a bu is only allowed to turn persons to sysops on her wp, and a stewart may change this on all wps, but boths actions are logged, may be easily undone, and both must wait on community consensus to make their job.

Both positions don't have any power, only the technical ability to do something. So, their only reason for existence is to unload the developers from this burden (and that surely is a great reason). But why this double structure?

Well, maybe we have too many bus already, so the risk is to great to give them all stewarts power, but in this case, rather than introducing a new layer, I'd un-bu some (I'll rather give up my bu status than have more complexity than needed).

What risk is there? As far as I understood, stewarts can even de-stewart other stewarts, and block Jimbo and Brion and all the others, so, a well-organized stewart may even start a really damaging attack on the wp. Or will all her actions be reversible easily?

So, the rules will state, that they, in their role as stewarts (as are bus already), will be mindless slaves. They are not allowed to do anything that has no prior consensus. So why not merge these two roles? I mean, if we were able to install a bot who recognizes consensus and then performs the appropriate action, wouldn't we just do it? The only reason we (will) have humans in this position is the lack of such a bot. (Please, I don't want to lessen the persons who fulfill this role, neither their work and time, I appreciate both highly; I just state my perception of the role itself).

So, no extra layer seems necessary, I think.

--denny 01:59, 20 Mar 2004 (UTC)

By the way, are bus or stewarts allowed to decline performing an action that received community consensus?

I suppose bureaucrats are language-specific, whereas this hondevs are not. That's why the selection is made here, at meta, as opposed to each language-wikipedias.
Regarding the decline, I suppose hon-developer is not supposed to, but if that happens, people can talk to the mailinglist, etc., and discuss/ review the performance of the hondev, and possibly demote him/her. Tomos 03:49, 20 Mar 2004 (UTC)
I do not think so. If these people will have rights on international wikipedias, there is a *serious* problem to discuss. Do Wikipedias have a right to different power structures, or do we necessarily have to follow this one ? If so, why should english people have power over international wikipedias ? Anthere
As far as I understood, hondevs won't have rights on international wikipedias, but power. English people will (and do) have power (but no rights, except for Jimbo) over international wikipedias, and ever will. The question is, shall, by implementing international hondevs, also users from non-english speaking wikipedia have the power over the english wikipedia (but alas, still no right), or shall only users from the english wikipedia have power over the non-english speaking wikipedias. --denny 15:25, 20 Mar 2004 (UTC)

Umm.. You are against this idea? There is already such a power held by developers. Tim's recent proposals and contributions, MediaWiki namespace, Bureaucrat, and this hon-dev are all in the direction of reducing the non-programming duties (and administrative privilages) of developers.

No, *I* agree with Tim idea very much. But most french wikipedians seem to not agree with this idea *at all*. So, that is embarassing; I proposed myself, but the reaction was "not again an additional layer, no way". We discussed this the whole week, with admin candidacy at the same time. And most do not like the idea of another level. So, I do not know very much what to do. I would like us to use this very much, but if people there do not want, what do we do if non-french come to do the job nonetheless ? Anthere

Replacing developers for those duties are sometimes language-specific appointees (like admins and bureaucrats who can now change wordings in the interface, make someone admin). But not all the duties can be performed locally - some small wikipedias need someone like Brion who creates the first admin. Hondev is for that purpose, and therefore an interlingual position. It is like Brion's ToDoList will be taken care of by these people, and we do not have to feel very sorry to burden Brion everytime we need to report some problem or request to him. (Though I would certainly miss Brion and Tim, since they both understand Japanese and that helped Japanese Wikipedia many times).

As I mentioned earlier, it is also better if certain functions can be turned on/off by these people, as opposed to developer, I think. Bureaucrat functions, direct display of external images (when URL is written), blocking of registered users, and many other functions are now turned on/off by developers. Some functions, such as search and page counter, should be controlled that way, but others could be better handled by hondevs. (Well, if that is technically feasible).

And certainly, people active at non-English wikipedia can apply for the position and seek for support. You have my support if you accept the nomination. :-) Tomos 05:01, 20 Mar 2004 (UTC)

thanks. Well, I put myself back, with the point that I will not do it on fr, as long as people do not want anyone there but a developer to do it. I will also copy your comment on the fr pump. It is perhaps better that what I explained. It seems clearer that what I told them. Thanks Anthere
I'm glad to hear that! :) Tomos

I think we don't need super-administrator. Aineko 02:26, 24 Mar 2004 (UTC)

Target number of developers

Does anyone have any thoughts on the number of people we should aim for to have developer powers? I would have thought three or four would be enough. We might set up a quick opinion poll later on, if necessary.


Who should be disqualified from being a steward?

Some of the people with more than 80% of their votes in favour, didn't have very many votes in total. Someone who has 3 in support and none against is in my opinion less qualified than someone who has 30 support and 7 against. Perhaps there should be a quorum, say, 10 votes?

This seems reasonable to me. --mav

Arbitrators

The English Wikipedia has a powerful arbitration committee, which can make decisions about permanent user bans without consulting Jimbo. Arbitrators were appointed by Jimbo and will possibly be elected in the future.

I propose that arbitrators cannot also hold this role. Arbitrators wielding substantial direct power reinforces the impression of a clique or power elite, especially by the user who is affected by the ruling. E.g. let's say a user is desysopped for violating protection policy. Martin (an arbitrator) puts the arbitration findings on the user's talk page. A minute later, Martin desysops the user. Now the user in question will focus their entire reaction on Martin. Shared responsibility makes for shared blame and creates a better impression, one of a working justice system.

Separation of powers is essential to avoid the impression of a class of "super-users".

The whole point of this proposal is to decentralize power, to share responsibilities. I oppose the whole steward system if this only leads from one class of user being specially empowered to another.--—Eloquence 06:53, 5 Apr 2004 (UTC)

There is wisdom in your willingness to separate decision from action Erik.
But you are also forgetting something. Stewards are not for english work only. They are here to help any wikipedia, regardless of languages and regardless of their own participation.

The english wikipedia is only one wikipedia among others, even if the bigger and most in need of unsysoping *now*. And the arbitration committee has power only on english wikipedia decisions. Consequently, there is no reason why Maveric, because he is arbitrator on english wikipedia, could not help on all the other wikipedias with regards to sysoping, bureaucraticizing and unsysoping. Help is needed, and he is trusted to do it. So why refusing his help ? He is speaking spanish, he could very easily help spanish wikipedia for example, if a case erupt and he needs to check what is going on (Aoineko is right on this, this is important. We need to know which language each steward speak so he can help best).

We need Mav. With Angela and I (excuse me there), he is one of those frequently around, so he will be available to answer requests from non english. And non english often complain that their requests are not considered quickly enough. So, why trying to remove one of those who can help *here* just because he is arbitrator on the english wikipedia ?

In case there is need to separate powers (which I agree is important), then let's have Mav agrees not to unsysop people on the english wikipedias after arbitration judgement in which he has been involved.

If necessary, *if* similar structures than arbitration committee develop on other wikipedias, any steward part of any such committee on one wikipedia, can agree not to do any unsysoping himself, but to let another steward to it, thus separating decision of sanction from application of sanction.

Erik, I think these are current processes of unsysoping people

  • on en:, 2 cases : urgent unsysoping by developers until recently (now, by stewards), or desysoping decided by arbitration committee (and now done by stewards)
  • on big wikipedias process most likely, unsysoping is probably decided by all the community by vote (perhaps with some restriction for those voting...on fr, currently no restriction is set, but to have a user name)
  • on small wikipedias, process is likely to be "request by the other sysops", such as the recent case on pt, which Tim took action for

Separation of decision from application is really only a problem in the en: power structure. On small wikipedias, the sysops requesting unsysoping are not stewards. On bigger wikipedia with democratic process, the decision is taken by the community, not by a committee. So, the steward is just doing the decision taken by the community. In case there is really a risk of accusation, bigger wikipedias can decide to ask to another steward from another language to do the unsysoping. And that is about it. I know I will probably do this. If I vote for the unsysoping of someone on the french wikipedia, I will refuse to take the action myself. Someone else will have to do it. In exchange, I can do it for any other languages.

So, I recommand that Mav just agrees not to unsysop people after agreement by the arbitration committee if he has been part of the decision (but he should still do unsysoping when there is urgent unsysoping to do), so that decision and action are not confused.

The whole point of this proposal is to decentralize power, to share responsibilities, *and* to help people regardless of their language wikipedia.

And let's move on now. We have work to do.

User:anthere

If Mav really wants to be a steward, I can accept that if he only acts on the non-English Wikipedias. None of this "decisions in which he is involved" business. The problem is the appearance of a power elite, even to bystanders, and that problem will appear even if a person acts as an arbitrator today, as a steward tomorrow, and as an arbitrator again on Wednesday.--—Eloquence 16:03, 5 Apr 2004 (UTC)

I forgot, the french people are not very keen on that status thingy. Too complicated do they think. They think that it is more simple to have developers go on do the job. We have one french developer, Shaihulud. Shai was not made steward, but he is trusted by french people to do steward things if needed (he is bureaucrat). That means, Shai, a developer, will do unsysoping as well as steward if needed on the french wikipedia. It is not because he was opposed by some people who do not even know him, that he will stop doing it, because we trust him to do it if needed.

User:anthere

No, no, no. Either all developers don't act in this function, or none do. There will be no special privileges for anyone. That is part of the whole deal. We can talk about giving bureaucrats the ability to desysop people on their wiki, which I have already suggested. But developers should not have special privileges, not on one wiki, not on all wikis.--—Eloquence 16:03, 5 Apr 2004 (UTC)
note Eloquence, that I personally do not approve this. But french wikipedians have not been asked whether they accepted that steward status or not before proposed. And when it has been discussed (not widely, we had other urgent matters on the fire), some people opposed the idea. I just indicate the current situation (while adding at the same time that several french approved loo and I, so I guess it is not so much a general feeling; others mostly felt not concerned). In any case, I just report the opinion there, and I won't discuss it anymore :-) Anthere 16:49, 5 Apr 2004 (UTC)
That's OK. At the end of the steward votes there will be a final ratification vote. I think this should be a collected vote for all projects, rather than an individual one for each project. If there really are such sentiments among the French, I bet they also exist elsewhere.--—Eloquence 20:19, 5 Apr 2004 (UTC)
Does fr expect to be able to decide who Jimmy Wales sends to work on the servers, who maintains the server software and who develops mediawiki programs? I think no. The stewards are Wikimedia people, just like the others I listed. However, stewards should be discouraged from acting on fr because fr is one of the wikis with its own established processes and stewards are mainly for doing tasks at wikis which do not have their own process yet. It's usually unnecessarily disruptive to act when there is an existing process at a wiki and all steward candidates probably already know that. Jamesday 17:31, 13 Apr 2004 (UTC)

Anthere is very wise - yes I will, in general, defer steward actions on the English Wikipedia to other Stewards. But I reserve the right to act in emergency situations. Responsiveness to requests on the non-English Wikipedias is more important, so I will concentrate attention there. --mav 21:17, 5 Apr 2004 (UTC)

=


How did Martin get on the list of stewards? He had only 75% support - I thought the threshold was 80%. --217.228.164.118 18:19, 22 Apr 2004 (UTC)

Original threshold proposed by Tim was simple majority. Eloquence later claimed that the threshold was 80%.
So whatever Tim proposes will be done? Simple majority is absurd for something as important as this. It should be at least as high as for admins, and 80% is what most people agree on there. --217.228.164.118 19:21, 22 Apr 2004 (UTC)
At least three of Martin's opposing votes were related to him being an arbitrator, but this was discussed at talk:developer access where the consensus was reached that as long as arbitrators did not use their steward access on their own project, this would not be an issue. As the discussion occured afte the vote and there were no objections to this, I think the results of that would have more weight than the votes that were made without taking this into consideration. Angela 19:58, 22 Apr 2004 (UTC)
Then those three people should have changed their votes. It's not for you to reinterpret opposing votes. --217.228.164.118 20:13, 22 Apr 2004 (UTC)
The same doesn't apply to Tarquin though, so perhaps he shouldn't be one if people do want 80%. There seemed to be no agreement on that being the case though and I added him to list simply because his percentage was so close to Martin's, but that doesn't take into account the fact that the reasons for Martin's opposition were not the same as Tarquin's. Angela 20:10, 22 Apr 2004 (UTC)

About 80% seems to be the current standard of support needed to be made an admin on the english wikipedia. I don't see how someone can be placed into a position of even more power/responsibility with less support than needed to currently become an admin. Maximus Rex 20:18, 22 Apr 2004 (UTC)

Ok, this morning on IRC, everyone seemed to be agreeing with it but that obviously isn't the case now. I don't have any strong opinion either way so I've reverted my stewarding of Martin. Someone else can make the decision. Tarquin was never made one as he didn't have an account on Test at the time. Angela 22:14, 22 Apr 2004 (UTC)

I didn't get here in time to vote, so here goes:
# Oppose. Martin is busy being an arbitrator. Martin spends too much time on Wikipedia already. Martin doesn't speak foreign languages well enough to handle non-English issues, and English issues may be restricted by virtue of being an arbitrator (cf talk:developer access). --Martin
Hopefully I get some say in the matter! :) Also, I think Tarquin would make an excellent steward, so I hope the dispute is resolved in the correct direction. :) Martin

Ok, I guess that resolves the Martin issue. :) Can some other than me do something about Tarquin please; ie- set his access or remove the "disputed status" section from the stewards page. I don't like the way people have tried to change the rules after the vote started. Originally it was 50%, then 80% and now there 4 different "close of vote" dates. By the original date, even Anthere didn't have 80% so removing Tarquin now and saying that changing the date is ok, but changing the percentage required isn't doesn't seem entirely the right way to go about things. Angela 23:37, 22 Apr 2004 (UTC)


note that at original date, the percentage mentionned had never been 80%. It suddenly popped up, while till then, the requirements were 50%. This was the reason why the vote was extended, since people thought the requirements were 50 rather than 80. Two ways to solve the issue would be 1) to ask Tarquin (I think he never voted, so might not even be aware of that vote) and 2) ask the ones opposing what they think about it. Or drop the whole matter. Anthere

I had no idea there was a vote, and in fact, I didn't even know about "stewards" at all! -- I've been off the mailing list for months now. I'm curious though, did people give a reason to object against me? -- tarquin

I think the only objections to you were based on people wanting a limited number of people to be stewards, so were supporting that number and opposing the rest. See [8] and talk:developer access where people explain that their reasons for opposing were not personal. Angela 22:59, 27 Apr 2004 (UTC)

Je suis d accord avec Angela. Tu as été listé au début, et au début le principe suivi était plutot de limiter le nombre de personne. D'ou bcp de votes d'opposition de principes, pas sur la personne, mais sur le nombre total. Je crois qu'il ne faut surtout pas te frapper par les oppositions que tu as eu. En effet, quand il y a eu de vrai oppositions (sur la personne) cela a été le plus généralement indiqué. Amicalement Anthere 23:47, 27 Apr 2004 (UTC)

Move to Meta?

If you make yourself a developer on Meta, the steward interface is now enabled here as well. So, should we move all sysoppings here instead of Test? There are three advantages to this:

  1. When the software is being tested and Test is unstable, we can still make sysops.
  2. The database on Test is occasionally wiped, meaning we would lose the Bureaucrat log; something that is far less likely to happen on Meta.
  3. The log is more visible on Meta. People view recent changes here far more often than on Test, so will notice any mistakes or controversial sysoppings.

Angela 01:15, 16 Jul 2004 (UTC)

I did not understand from the start why we had this function on test... (maybe for test-purpose? ;-)
So, if there is no reason for test, meta is perfect :-) Fantasy 12:20, 16 Jul 2004 (UTC)
yes, meta is more suited. -- Looxix 17:42, 16 Jul 2004 (UTC)
of course. Anthere 21:56, 16 Jul 2004 (UTC)
I agree. - Andre Engels 08:47, 17 Jul 2004 (UTC)