Wikisource talk:Restricted access policy

Majority/super-majority
Before I begin drafting up a voting policy, I've got a question I would like to ask first. This will help me with writing the proposal. On the project page, it says that the outcome of votes of confidences is not determined, and that we are debating whether a majority or super-majority must be established. On a wiki such as WS, where people come and go quite regularly, I think we need to have a clear understanding of what "majority" and "super-majority" mean, as it can't be used in the traditional sense (because we have an always changing number of users here--sometimes it increases, sometimes it decreases).

Does "majority," etc., here mean of the amount of people who actually voted? Or is there a set number of "support" votes required to meet a criteria which will then be construed as being a "majority," etc., of votes?—Zhaladshar (Talk) 18:59, 8 May 2006 (UTC)


 * Sorry, I didn't notice your comment until now. The terms supermajority and simple majority describe the ratio of support to oppose votes. A simple majority is exactly 50% plus one user in support; any number between 50.000~1% and 100% is a simple majority. A supermajority, on the other hand, describes a situation in which the vast majority of votes were in support. Although the exact percentage used varies, Wikipedia considers at least 75% to be a supermajority. // [ admin ] Pathoschild (talk/map) 01:01, 15 May 2006 (UTC)

Majority --> Consensus
The policy says that "Administrator access depends on the continued support of the community", a statement I fully endorse. But it then says that to keep adminship, admins need only "simple majority support (50% + 1 user)". I interpret "support of the community" as consensus, and my view is that if the community is split anywhere close to 50-50, then the admin is too controversial to have community support. I propose that the threshold to stay an admin be made the same as the threshold to become an admin. That would go for the annual votes as well as for user-initiated motions of confidence; in each case, the admin would need consensus to retain adminship. Community-based recall has had a hard time at the English Wikipedia, so I'm glad to see Wikisource being proactive about it. Good work! Tim Smith 07:39, 23 August 2006 (UTC)
 * Well perhaps it is not worder the best, but I always thought the yearly confirmation to be like a vote of no confidence. Consensus is need to show the loss of community support and otherwise the support is retained.  I don't expect there to be any real close calls however.   English Wikipedia's problems are not in my opinion inherent to wiki's.  The reasons that they have contraversial admins should not occur here.--BirgitteSB 13:51, 23 August 2006 (UTC)
 * The thing is, is there is not a definition of consensus. On the obtention portion of adminship, it just says the person will receive access upon a community consensus (pretty vague, to me).  So, likewise, to keep adminship should not be such a stringent procedure, because obtaining it is not a stringent procedure.  If we could come up with a definition of what consensus is, then we should change both parts accordingly.  But, also, WS is a much smaller community, and many of the contributors sometimes miss ongoing votes (not a fault of their own), so in the event where one user is up for renomination and someone votes deny, having a lower threshold will help offset the negative vote since some of the contributors who would have voted approve missed it.


 * I am not saying I disagree with Tim, though, just that at the time being it would not be an apt move to make it a stringent process (even though, like BirgitteSB says, it's doubtful there will be any mixed community support for our admins).—Zhaladshar (Talk) 14:06, 23 August 2006 (UTC)

I agree that the standard to keep should be no more or less stringent than the standard to obtain. In fact, I think the standards should be the same. The wording for obtention is that adminship is granted "following community consensus", and I'm happy to leave it vague so that consensus can be gauged on a case-by-case basis. Similarly, the wording for retention might say that adminship is retained "following community consensus", again with deliberate vagueness to allow case-by-case judgements, like taking into account the possibility Zhaladshar raises that some contributors may have missed the vote. I see the yearly and user-initiated confirmations as motions of confidence rather than motions of no confidence. Since consensus is needed to gain adminship, consensus should be needed to keep it, and a lack of consensus shows a loss of support. The time to make this change is now, while our admins are uncontroversial. Hopefully controversy will not come, but if it does, toughening the policy will be much more difficult (see the English Wikipedia). Tim Smith 15:32, 23 August 2006 (UTC)

Any further comments? I'm just proposing to make the threshold to keep adminship the same as the threshold to obtain adminship. Both would be "consensus", left deliberately vague to allow case-by-case judgements. I do think it's important to enact this change now, while it's easy and harmless. Tim Smith 19:27, 25 August 2006 (UTC)


 * I suggest making a proposal in the Scriptorium. There is a particular section for proposal, I doubt many people are watching this page.--BirgitteSB 20:10, 25 August 2006 (UTC)


 * Done here. Tim Smith 21:17, 25 August 2006 (UTC)

Inactivity
"An 'inactive administrator' is one who has not edited during the past six months and has not made more then 50 edits during the last year."

Is this supposed to be automatic? Or just a criterion for de-admining via the normal voting process? If the former, why are we voting on people who have been inactive for longer than 6 months on WS:ADMIN? If the latter, why so specific? --Spangineerwp (háblame) 14:09, 2 November 2006 (UTC)


 * It was originally intended to be an automatic process, but I don't think that is practical. Some users—such as Brion VIBBER (see Development team)—have de facto honorary administrator status, and are not even scheduled for confirmation. Some other users may have good reasons to be inactive for six months; perhaps they're volunteering in Africa for seven months, for example. In such a case they'd have to place a notice on their userpage to that effect, of course.


 * I think that inactivity is too flexible a concept to be applied automatically. It works fairly well as a voting criterion, though. —[ admin ] Pathoschild 18:52, 2 November 2006 (UTC)


 * That makes sense. I'm just surprised to see it so explicitly stated in the policy&mdash;almost as if one can't claim that an admin is "inactive" unless his/her editing habits fit this description.  Would be akin to "you can't be an admin unless you have at 1,000 edits and 2 months experience"&mdash;something everyone basically agrees to, but no one really wants to write in stone (at least, that's my experience on Wikipedia; maybe over here consensus is to be more specific). --Spangineerwp (háblame) 19:04, 2 November 2006 (UTC)


 * That's because it was originally intended to be automatic. If that's no longer the case (it would seem not), we should reword it to be less specific. —[ admin ] Pathoschild 19:17, 2 November 2006 (UTC)


 * Should the entire section be removed? No other criteria are mentioned in the "Votes of Confidence" section (just what the community thinks), so it would be strange to highlight one.  Perhaps introduce a line to the previous section saying that "misuse of administrator tools or inactivity" are common reasons for being recalled, and delete the section on inactivity? --Spangineerwp (háblame) 03:56, 3 November 2006 (UTC)


 * Perhaps we can streamline the process by making it semi-automatic for inactive administrators. Instead of three oppose votes triggering a confirmation vote, the initial discussion would be decisive. —[ admin ] Pathoschild 19:22, 3 November 2006 (UTC)


 * So a confirmation vote for an inactive admin would take place whenever anyone wanted? Seems reasonable to me. This brings up something else though&mdash;should there be some sort of restriction on how much time must pass between confirmation votes for the same person? Wouldn't make sense for the same user to ask for a new confirmation vote every month. --Spangineerwp (háblame) 19:44, 3 November 2006 (UTC)


 * That type of abuse is unlikely to be effective, since at least three established users must support the need for a confirmation vote. The only likely result of such abuse would be to have the request rejected every month, and probably rejected immediately if the user is clearly acting in bad faith. —[ admin ] Pathoschild 03:24, 4 November 2006 (UTC)


 * True, but I only thought of it because an inactivity-based confirmation vote wouldn't require three established users. Probably not worth worrying about though. --Spangineerwp (háblame) 05:40, 4 November 2006 (UTC)

I propose the following rewording of the inactivity section based on this discussion. Aside from minor spelling and grammar improvements, it reduces the inactivity level from six to four months, removes the overly specific '50 edits during last year' requirement, and allows for automatic loss of access unless the voting community decides otherwise on a case-by-case basis.

 It is a common occurrence on wikis to create an account, be active for a certain period of time, then become inactive. While this is to be expected due to the nature of this site, administrator accounts should not be left unattended for a long period of time. Inactive accounts are more open to account hijacking, and may be difficult to detect as compromised if the original user is not there to notice edits he didn't make. Furthermore, the list of current administrators should only contain administrators capable of responding to queries from users within a reasonable period of time.

An inactive administrator is one who has not edited during the past four months. Inactive administrators automatically lose their access in their next scheduled confirmation vote unless at least half of the voting community supports continued access. Any administrator who has lost access due to inactivity may reapply through the regular processes. —[ admin ] Pathoschild 03:23, 5 November 2006 (UTC)

Remember that someone who is un-sysoped can always be re-sysoped if they come back and go through the process. While that may be somewhat of a hassle, I do not think it is a really big deal. --Benn Newman 03:49, 5 November 2006 (UTC)


 * Do these paragraphs say anything different from the general policy? Doesn't any admin "automatically" lose access at their next scheduled confirmation vote unless half the voting community supports continued access? --Spangineerwp (háblame) 22:42, 5 November 2006 (UTC)


 * Nope (to the second question).
 * Under the normal process, a vote is automatically held one year after the previous vote; if at least three established users oppose continued adminship, a second vote is held in which a simple majority is required. This second vote is brought to the attention of the community through a post to the Scriptorium and an article in Wikisource News. The second step is intended to ensure fairness by not requiring that the entire community show up every single month to make sure the administrators they support are not being quietly voted out.
 * The adjusted process for inactive users is much simpler. Unless 50% of voting users object in the scheduled vote, they lose their access without that extra step or the requirement for three established users.
 * —[ admin ] Pathoschild 00:25, 6 November 2006 (UTC)
 * Got it. Thanks! --Spangineerwp (háblame) 00:36, 6 November 2006 (UTC)

Implemented. —[ admin ] Pathoschild 22:36, 14 November 2006 (UTC)

Obtention -> Obtaining administrator access
I made the subject change to the first subheading as obtention wasn't in Wiktionary!!! (until I added it :) ) It's a legitimate word, but one that may be a bit obscure for those who don't speak English as a first language. It's a very minor change but I felt I should mention it since this is after all, accepted policy. Revert and discuss if you feel strongly about it. :) ++Lar: t/c 21:50, 14 October 2007 (UTC)

Copied from Wikisource talk:CheckUser policy
I really don't see the need for a separate reconfirmation for Checkuser access if the user is going for admin. Why not make one confirmation annually for both, clearly stated in the instructions. The process may require two support votes (one for each user right), or simply allow the user to support on the whole, object to one user right, or object to both. - Mtmelendez 21:41, 4 April 2008 (UTC)
 * Is there a global checkuser log? I thought that there used to be, but now checkusers can only see their local log.  The custom (though not the policy) on en:wq is that a note of the results of checkusers is placed on the administrators' noticeboard.  Obviously, nothing disclosive is said.-- Poetlister  17:44, 7 April 2008 (UTC)

Vote versus debate
"The vote will he held for at least one week. Although any user is welcome to discuss the nomination, only established editors may cast a vote with weight." I don't think that an RfA should be a vote. A bureaucrat assesses the remarks and the arguments and determines whether there is consensus.-- Poetlister 17:46, 7 April 2008 (UTC)


 * Our traditional approach on Wikisource been a simple vote for user elections, since we're measuring whether they are trusted by members of the community (rather than how well-reasoned that trust or distrust is). In the peculiar case of checkuser access, a vote is explicitly required by Foundation policy. Since this policy is only intended to update the administrator policy for relevance to non-administrator access, I think we should discuss other changes separately (that wording is copied from the administrator policy).


 * By the way, I'm likely the only person who'll see your comment. If you want to discuss this more publicly, please do so at . — {admin} Pathoschild 17:54:15, 07 April 2008 (UTC)


 * Well, John saw it! But I'll copy it over.-- Poetlister  12:37, 10 April 2008 (UTC)


 * Until we have a definition of "established editors", there is still an element of 'crat discretion.  John Vandenberg (chat) 06:38, 8 April 2008 (UTC)


 * Short history lesson: This whole practice grew out of our deletion policy where we used to regularly have a large group of people show up just to vote against the deletion of text that was "dumped" here, even though they had no interest in becoming regular contributors to Wikisource.  Rather than go the "not a vote" route; we instead for whatever reason went the route of "only established contributors get suffrage".  It quickly solved the issue of the voice of actual contributers being overwhelmed by one-issue outsiders at deletion.  I don't know if prevented such problems from happening over adminship, as all non-established editor votes have agreed with the established editors as far I remember.  We have never made it a practice to strike the votes of such people even at deletion.  Personally I believe it is important that they are allowed to say their piece, as it makes them feel better and there is always a possibility the established editors will learn something useful from what is said.  It hasn't really been an issue here, but I can't imagine the community would feel very differently about an influx of people not invested in this wiki opposing the community at RfA rather than at a deletion.  And even though it is a vote, everyone understands it is best to try and reach agreement. We don't just toss in our ballot, but rather try to convince one another when opinions differ. So far it has worked out well. Personally I think systems that use !voting end up diluting what consensus means.  I would rather vote over people and require true consensus on other issues.  I think it would not be right to shelve a person who is dedicated to this project the way we shelve issues which have no consensus.  And I also think redefining consensus to mean something with a deadline would weaken the whole concept and lead to the sort of problems that exist at other wikis. --Birgitte SB  01:29, 15 April 2008 (UTC)

Oversight
Nice work, but should probably include mention of Oversight as well. Hopefully WS will never get to where it needs someone to locally have it, but it would be best to include it anyway for consistency. Oversight, like CU, has a foundationwide policy too. ++Lar: t/c 11:48, 8 April 2008 (UTC)


 * I don't think we need a local oversight policy, since we don't have anyone with oversight access. If we ever do, we just need to slip in a reference to the Foundation policy. — {admin} Pathoschild 16:46:34, 08 April 2008 (UTC)


 * Why not just slip the ref in now (with a note that we have no local oversighters but we'd like the stewards to abide by standard policy)? Won't hurt, will make things clear. Support, otherwise. ++Lar: t/c 02:40, 14 April 2008 (UTC)


 * Might be easier to do it now than to have an argument if and when there's an oversighter (overseer?) vote.-- Poetlister 11:19, 14 April 2008 (UTC)


 * Done. — {admin} Pathoschild 22:40:35, 24 April 2008 (UTC)

Inactive admins
I think this policy is good, but I have considered a new parameter (taken from my now-closed wiki's adminship criteria):
 * Admins who declare their inactivity on their user/user talk pages (e.g. X Admin is editing sporadically but will check his user talk every so often) are exempt from recall.

Just a suggestion. Thanks, AC aka --Sunstar NW XP 15:02, 5 September 2008 (UTC)
 * Sounds like a great way to avoid recall. Someone can say that and still do a handful of edits a week.-- Poetlister  16:08, 5 September 2008 (UTC)


 * I don't see any reason to change to current policy. Active admins are not "exempt from recall", and inactive admins are desysoped regardless of merit when their confirmation period comes up.  So I am not sure what the change you suggest is meant to accomplish.  If someone attempted to start a recall of merit on an inactive admin, I am sure that most people would feel the recall should wait either till activity was resumed or for the pre-ordained confirmation period.  It would be difficult to get the number of people needed to start a recall to agree to start one for a moderate reason when the subject is inactive and unable or unwilling to defend themself given the other safety valves available.  It would only be easy to get enough people together for this when the admin in question did something beyond the pale.  And leaving that option open for extreme cases is a desirable safety valve.--Birgitte SB  17:13, 5 September 2008 (UTC)

We dont have recall here. With yearly reconfirmations, inactivity being reason for desysop, and emergency votes of confidence, there is simply no need for recall, and I hope we never have it in any form resembling what has grown legs on English Wikipedia. Simple put, on English Wikisource all admins are open for recall, and it is conducted in accordance with a site wide policy rather than admins being able to pick their own terms and process. John Vandenberg (chat) 22:46, 5 September 2008 (UTC)


 * Sorry, my lack of understanding here. I should try to understand these policies a bit more... --Sunstar NW XP (talk) 10:11, 6 September 2008 (UTC)

Votes of confidence - clarifications
I just made three minor updates to the section on votes of confidence. The first simply makes the language clearer as the prior language was awkward. The second separates the archiving section from the sentence about suspension of privileges during a called vote, as these are unrelated. The third updates the section on archiving to acknowledge that any uninvolved Bureaucrat can close the vote. Current practice is for User:Billinghurst to close discussions he hasn't participated in and we have only two crats and it would be an excessive burden for them to have to close these discussions. Archiving can be done by anyone.--Doug.(talk • contribs) 10:59, 28 September 2011 (UTC)


 * I see John has just reverted me. I agree if we require a crat for called votes that we should specify that but the section already confuses the two very badly.  We need to be clear as there are no clear instructions on a regular vote.--Doug.(talk • contribs) 11:01, 28 September 2011 (UTC)
 * I see now what John means. Clearly I didn't understand the process which means it was even more confusing than I thought.  I suggest the following:--Doug.(talk • contribs) 11:37, 28 September 2011 (UTC)
 * "Restricted access depends on the continued support of the community. This may be tested by a vote of confidence, in which a simple majority (50%+1) must support the user's continued access for it to be retained. (What access a discussion concerns should be explicitly noted in the discussion's introduction.) If a vote of confidence is called, the holder of the restricted rights may not use the restricted access for any non-trivial action at any time until the vote is closed. A bureaucrat will eventually close the discussion and, if so decided, request removal of restricted access by a steward.  Votes of confidence must first be proposed and at least three established users must support the need for a vote of confidence before one can be called. Any user may propose a vote of confidence at any time.  Additionally, at least once annually, a vote of confidence will be automatically proposed to confirm that the community still wants the user to hold restricted access.  This vote should occur approximately one year from the last proposed vote of confidence or the initial grant of rights.  Users with multiple restricted rights need only have a single annual confirmation.  As soon as there are three votes in support of a vote of confidence, the proposal should be closed by any party and a vote of confidence initiated. After a proposal, whether annual or otherwise, has been open for a reasonable time without three votes in favor of a vote of confidence, any uninvolved bureaucrat or administrator, may close the discussion as 'confirmed'.  Ordinarily, in an annual confirmation vote, without further explanation, a vote of 'support' means 'support reconfirmation' and a vote of 'oppose' is a vote to hold a vote of confidence."

In the above text there are a few changes to the policy as I understand it, however some of them are minor or perhaps even unintentional. I think every change needs to be discussed separately, with input from Pathoschild (policy writer), the crats, and other community members who have been involved in implementation/use of this part of the policy. This part of our policy is infrequently used, and there are not many other similar policies in the wiki world. We need to get it right, and have strong consensus for even small changes. John Vandenberg (chat) 23:29, 28 September 2011 (UTC)

votes of confidence proposal - use of restricted rights
The current text is:


 * In the case of an unscheduled (called) proposal, the user may not use the restricted access for any non-trivial action at any time until the vote is closed.

Doug has suggested:


 * If a vote of confidence is called, the holder of the restricted rights may not use the restricted access for any non-trivial action at any time until the vote is closed.

This is a significant shift in when the restricted rights should not be used for non-trivial actions.

Currently a sysop
 * 1) can continue to use tools during a vote of confidence is the result of a scheduled/annual proposal; i.e. even after there are three people calling for a vote of confidence.
 * 2) cannot continue to use tools once an unscheduled proposal for a vote of confidence is initiated; i.e. even if there is only one person calling for a vote of confidence.

Doug's change would mean that a sysop cannot continue to use tools during a vote of confidence, irrespective of whether it was scheduled or unscheduled. i.e. once there are three people calling for a vote of confidence, the vote of confidence begins and the sysop needs to stop using the tools for non-trivial actions.

My initial impression is that Doug's suggested change are good. John Vandenberg (chat) 23:42, 28 September 2011 (UTC)

Boldly merged the practice that we had stated as policy
Adminship has been quoted as policy, and if it isn't it certainly has been clearly leading the practice that we had been using for judging activity/inactivity and especially around duration to classify inactivity. Following conversation elsewhere, Doug had removed some text from the aforementioned page to remove confusion. Some of that removed text, I have added into this page. [ Changed text] — billinghurst  sDrewth  06:36, 8 December 2012 (UTC)