Wikipedia:Administrators/2019 request for comment on inactivity standards

• I have no thoughts one way or the other about adding a midterm notice, but would not replace the end-of-term notice with a mid-term notice. Essentially what would happen here is that people would get a notice saying "you have six months to do X"...so they will not feel any pressure or urgency to do "X". With a one-month notice, at least they understand that they've only got four weeks to get their act together. A very large percentage of the population naturally procrastinates. All that you need to confirm this is by looking at the number of people doing Christmas shopping the days immediately before Christmas. It's not like they didn't know that Christmas was going to happen on December 25 - and they're being reminded constantly on a multiple-times-daily basis from the end of October. (Or you could see how many Arbcom decisions actually follow the schedule, if you want a Wikipedia example.) Risker (talk) 18:54, 28 January 2019 (UTC)[reply]

  Just read this, and it's in line with my thinking that a notification after 3 months wouldn't hurt. Three months, six months, whatever, anything that even marginally has a slight chance of encouraging a 'lapsed' admin to return sooner is a good thing. Fish+Karate 14:15, 29 January 2019 (UTC)[reply]

• I like this idea (as an additional notification). Looking at the shall-we-say extremely infelicitous timing of when my own one-month bot-delivered warning for inactivity was issued, I'm actually rather surprised I didn't delete my user page, scramble my password and walk away from the project permanently. A personalised note from a bureaucrat or admin, conveying the message "Hey EA, how's it going? (By the way, I see you've been inactive for 6 months, do you want to have your permissions temporarily removed while you sort your life out?)" would have felt a great deal more encouraging. Espresso Addict (talk) 05:27, 22 February 2019 (UTC)[reply]

Affirmative permission removal

In the spirit of discussing ideas, what if instead of notifying admins to make one edit to keep the bit, we instead notify them that they're coming up on one year of inactivity and ask if they would like to suspend their privileges? They can then decide to go to BN and request suspension of their mop, or if they do nothing they keep the bit. (In both cases, subject to the long-term inactivity provision)

My thinking is this addresses the view that automatic removal for inactivity is undignified. Inactive users still get a periodic reminder about the project, but they don't have to do anything unless they want to voluntarily (and temporarily) relinquish their access. It removes the security theatre of making people make one nonsense edit once a year. It leaves the condition where a reconfirmation RFA is required after a very long period of total inactivity, which isn't a change but I think that's a good thing to leave in. Anyway, just a late night thought really. Ivanvector (^Talk/_Edits) 05:13, 29 January 2019 (UTC)[reply]

My experience in contacting inactive admins is that they don't respond even when the message is clearly personalized. The current notification messages encourage people to resign voluntarily. Few do (<10%). UninvitedCompany 18:41, 21 February 2019 (UTC)[reply]

How serious is this problem really?

According to the latest Administrator Newsletter, there were six administrators last month who avoided desysoping by making last minute edits. That does not seem to me to amount to a whole hill of beans. And we are still desysoping more than we promote every month; have been now for years. SpinningSpark 02:54, 4 February 2019 (UTC)[reply]

I note that the introduction is a typical case of insufficient data presented as proof of a problem. I am not claiming any bad faith, quite the opposite, I believe in the good faith of whomever started this (I have not checked who). The problems with the introduction, as I see them, are... The sentence "A number of these accounts were almost totally inactive or had not used administrative tools in a long time" is as true as «A number of these accounts were active and have used administrative tools recently», meaning that we must desysop active admins... Then "Some outreach to almost totally inactive admins was attempted but there was little response." meaning the same as «A few admins responded to our outreach», which is a success, no? I know, we read and ear this a lot from politicians, or the news, and friends talking, and so on; but we could do better, no? So, please, first show us the problem, then it will make more sense to find a solution. (disclaimer: I am a almost inactive admin right now, I have been doing some 10, 20, edits per month or so and no admin actions, since I am back to college - but I keep reading the newsletters, I care, and I should be studying right now... :-) - Nabla (talk) 12:20, 4 February 2019 (UTC)[reply]

Not at all serious on security or activity aspects, more serious about lack of respect for others and nickle and diming policies to the detriment of others in the community. There's a set of folk who appear to not want to recognise that life happens and we should bend over backwards to be friendly towards people to whom life may already be delivering troubles. Jamesday (talk) 21:51, 10 February 2019 (UTC)[reply]

I spent a fair bit of time on assessment of this problem and was working on something at User:UninvitedCompany/sandbox, which you can look at if you like. Out of around 1200 administrator accounts, around 300 of them have not been actively engaged in admin work for years, and are either making occasional good-faith edits, or annual token edits. Most of them are people who became admins in 2005, 2006, or 2007. I do think it poses various risks and distorts the way we look at the project in important ways, but I've thought about it -- a lot -- and haven't been able to figure out a policy answer to it that I actually like. UninvitedCompany 23:04, 19 February 2019 (UTC)[reply]

@UninvitedCompany: Is there any way of getting a list of such admins? I'm not really sure what level of disengagement you are talking about (would I have counted, for example?). Espresso Addict (talk) 00:49, 22 February 2019 (UTC)[reply]

Well, there's a list of particularly inactive administrators at Wikipedia:WikiProject_Editor_Retention/inactive_administrators who nonetheless make enough edits to hold onto their privileges. I don't think there's any way to get a complete list without writing a bot. UninvitedCompany 13:33, 22 February 2019 (UTC)[reply]

By "hill of beans", you're referring to the policy of "don't stuff beans up your nose", if I'm not mistaken? Woshiyiweizhongguoren (🇨🇳) 22:31, 29 March 2019 (UTC)[reply]

Big deal

We keep saying being an admin is no big deal, we refer to the tools mops. Yet at every turn from the moment you consider an RfA everything becomes a big deal. Every action an admin does can be reverted by other admins, if one goes rogue there are measures and options in place to take action. After 15 years of throwing admin options around trying to find the best format, why not automate the system with Q&A test around policies for getting the tools and then repeat it every 6 months. All thats needed is a base criteria, then anyone can do the test - Those that are new Crats do a commonsense check start a discussion, those that have the tools and fail crats can review, discuss, or remove, everyone else just keep doing what they are doing. If you arent active when your 6 months comes up tools get taken away, then its truly no big deal. Gnangarra 03:53, 4 February 2019 (UTC)[reply]

In theory, being an admin is not supposed to be a big deal, but in practice it has been made into a big deal, largely by people trying to impose their own requirements at RfAs. Being an artificially big deal has inflated the desire to keep the bit just in case - to avoid another artificially big deal of another RfA. We reap what we have sown. Or maybe we reap what those before us have sown - depends how long we have been here. Cheers, · · · Peter Southwood ^(talk): 06:08, 4 February 2019 (UTC)[reply]

Yep so we take the big deal out of the equation and this isnt an issue Gnangarra 07:51, 5 February 2019 (UTC)[reply]

Mindset

The problem, as amply demonstrated on this page, is the community's fractured attitude towards adminship. Everybody seems to have a different idea of what it's about.

It's not a question of need. No single editor needs the admin tools, because there are hundreds of others who could do that particular job – that how you define 'need'.

It's not a question of competence. Every single admin is trusted to edit Module:String (or refrain from editing it), but I am not allowed to edit it. It's simply a matter of trust, not competence.

It's not a question of status or of levelling-up or of anything else redolent of a MMORPG. For obvious reasons.

What it should be about is one thing only: the trust that the community has in the editor. It's completely binary: the community trusts that editor to make sensible decisions and actions; or it doesn't.

If you define an admin as a user who has received confirmation that the community trusts them, then RfA becomes the process whereby that trust is gauged and acknowledged, and the inactivity standards simply become a security protection against unused admin accounts being hijacked. If you can bring yourself to that mindset, you'll see that the requirement for edits or logged action is just to save a 'crat from removing the bit from any admin who is still committed, albeit lightly at the moment. If we trusted someone with the tools in the past, it's really most likely that we're going to trust them enough to return the tools on request. After all, why wouldn't we trust them to take things easy at first and ease their way back if they've been away for a long time? Most of the proposals here are addressing an issue that doesn't really exist. --RexxS (talk) 22:33, 18 February 2019 (UTC)[reply]

You've seen the following comment elsewhere, RexxS, but I'm saying it here because I said at

WP:BN

that I would make some sort of comment regarding what is happening in this RfC.

there is an open RfC regarding the inactivity requirements. I've read it and the thing is mind-boggling: admins I know well, trust and respect are circling the wagons to stop change. I'm paraphrasing but if someone says "increase by ten", they reply "just as easy to game as if we leave it as it is"; if someone says "increase by 100", they say "woah, this is a volunteer thing and we can't expect people to put in the time like that". I'm sorry, but if someone thinks they have a need for the bits then I'd like to imagine they're reasonably active. I can understand breaks lasting several months but not twelve, and if it does get to twelve then they're out of touch

That's how it appears to me and I strongly suspect it is how it appears to a lot of other people. It is why I initially told Xaosflux that it would be a waste of time for me to comment here. - Sitush (talk) 07:32, 19 February 2019 (UTC)[reply]

"If we trusted someone with the tools in the past, it's really most likely that we're going to trust them enough to return the tools on request." That's not true in my case. I do not trust somebody who passed RfA in 2003 as much as I trust somebody who passed RfA in 2018. The latter has been thoroughly vetted by hundreds of people against established standards; the former hasn't. Ritchie333 ^{(talk) (cont)} 14:59, 21 February 2019 (UTC)[reply]

I think You are talking about competence (that is mostly what RfAs currentlky *try* to assess - do you know most editing issues, do you know protection rules, do you know blocking rules, and so on); while RexxS is talking about trust (do we believe your intention is good, do we believe you will not go blocking or protecting unless you (re)read the rules, and so on). And trust changes very little since 2003. So I'll have more trust on someone which was active from 2003 to 2013 without much trouble, than on some newcommer from 2018 (which as less actions to show) - Nabla (talk) 11:49, 3 March 2019 (UTC)[reply]

@Nabla: The attitude today is that people would prefer a new admin to be one with some—not necessarily massively lengthy—tenure and steady, regular edit count over, say, a legacy admin who has demonstrably found it difficult to hit three figures a month for the last five years. It's because experience is demonstrated, and trust is proven, with both tenure and activity. I suggest reading some recent RfAs... ——SerialNumber54129 12:24, 3 March 2019 (UTC)[reply]

I do read a RfA a once in a long while, I wonder what makes you think you know what others do or not. - Nabla (talk) 23:07, 3 March 2019 (UTC)[reply]

@Serial Number 54129: And it's exactly that attitude that I'm criticising by asking for a change of mindset. The attitude today has led to the deeply flawed process that masquerades as RfA at present – a process that has more in common with a fraternity pledge that with a rational system of deciding whom the community can trust to have extra powers. It seems that contributors can't make their minds up about whether they can trust a candidate without a hundred-and-one proxies like the number of times they mashed an XfD button. --RexxS (talk) 13:21, 3 March 2019 (UTC)[reply]

No worries, RexxS, you can have all the admins you're willing to pay for. I don't disagree with you though; I was alerting the previous poster to the difference in what is and what was. ——SerialNumber54129 14:09, 3 March 2019 (UTC)[reply]

Oh, thanks. I'll have two pointy admins, a big flat crat, and a packet of rollbackers. Cheers --RexxS (talk) 17:11, 3 March 2019 (UTC)[reply]

New proposals?

Is it too late to add proposals here? Based on the latest re-sysop request, I'm considering a proposal along the lines of: editors de-sysopped for inactivity must make at least 25 edits or logged actions before (possibly: within 30 days of) re-requesting the bit. Bureaucrats have discretion to decline requests if the edits appear to have been made to game the system and are not constructive improvements to the encyclopedia or its administration.

power~enwiki, go for it! This has yet to be closed. However, I would probably relist the proposal formally for it to get a chance at being seen.:D ―Matthew J. Long -Talk-^☖ 20:55, 18 February 2019 (UTC)[reply

]

There's no such thing as "relisting" a proposal like this.

π, ν) 20:56, 18 February 2019 (UTC)[reply

]

I meant like adding a seperate {{

RFC}} tag to it. ―Matthew J. Long -Talk-^☖ 20:57, 18 February 2019 (UTC)[reply

]

• If you wan't something that's actually going to solve at least part of the problem, rather than add more overly prescriptive rules that can just as well be gamed but with somewhat greater effort, instead add language giving the crats discretion to decide not to resysop in cases where a user is obviously gaming the system. The idea that being a crat means that you can't have a pulse and apply any level of common sense is silly. GMG^talk 23:37, 18 February 2019 (UTC)[reply]
• Well, the stated purpose of the RFC revolves around security, although there isn't a single proposal related to security in this entire RFC, so...why not. However, it probably won't get very much feedback after this length of time, and won't have as clear an outcome. Please at least try to make it relate to security, though. If it doesn't, at least be very, very clear what problem you're trying to solve. Risker (talk) 04:56, 19 February 2019 (UTC)[reply]
  • Didn’t see this section before making my additional proposal above. Basically the same thing, Power if you have suggestions please do pile in. Fish+Karate 05:24, 19 February 2019 (UTC)[reply]
  • WP:BN, where Xaosflux linked this discussion. As for the focus of the RfC revolving around security, yes, that is how it seems and is why I initially told Xaosflux that it seemed somewhat irrelevant to the matter at hand. However, Xaosflux has suggested that this is the correct venue and it is obvious that at least some 'crats are struggling to reconcile current policy and the specific event being discussed there. It is also well-known that there are numerous other admins who "game the system", as is being alleged in the specific BN situation. - Sitush (talk) 08:00, 19 February 2019 (UTC)[reply
    ]
• I like this a lot more than the other new proposal above (checking whether people have truly returned to the editing community before re-sysopping makes more sense than asking them to make promises to do so). If this is adopted, it should come with a friendly message inviting returning ex-admins to re-engage with the community and then ask for the admin bit again. —Kusma (t·c) 13:17, 19 February 2019 (UTC)[reply]

Well, during all of 2018 we had 4 users who were re-sysopped at BN after losing their bit due to inactivity, and three of them have been making a genuine effort to contribute, so I'm not sure that this proposal addresses a large enough share of the problem to be meaningful. The problem is, instead, the ~300 people who are making token or occasional edits -- in most cases often enough that they never get the warning message -- and who have no intention to return to the project. UninvitedCompany 23:07, 19 February 2019 (UTC)[reply]

If I recall correctly, the wave of compromised admin accounts last year were mostly those who almost but not quite met the inactivity standards. Ritchie333 ^{(talk) (cont)} 15:44, 21 February 2019 (UTC)[reply]

• Comment. This should possibly be re-linked at CEN, as the traffic it is receiving at the moment from the Bureaucrat Noticeboard might not be representative of the whole community. Espresso Addict (talk) 04:53, 21 February 2019 (UTC)[reply]
• Some of the discussion here may be relevant. I don't currently have the time to formalise the points I raised there here, but wanted to be sure those participating here are aware of that (short version: voluntary removal of tools used to differ from removal due to inactivity but changes to the policy removed that difference). Those who are actively engaged with the project, but decide to voluntarily give up their tools (e.g. to go on a break), are being treated the same as those who are long-term inactive and not responding to messages and are not engaged with the community. That doesn't seem right to me. If you voluntarily give up your tools, but remain engaged with the community and continue editing, why should there be a limit on when you can ask for the tools back? Carcharoth (talk) 10:15, 21 February 2019 (UTC)[reply]

Elephant in the room

It's been discussed upthread, but I think the root cause of this discussion is the standards to pass RfA and to regain the bit at BN are wildly different. If passing another RfA after inactivity was straightforward, people would suggest that - hand-wringing over "what's inactive and what isn't" wouldn't happen because somebody would simply say "why don't you just go to RfA and see what happens?" and there'd be enough of a view of that being a sensible idea, whereas currently people see it as trial by ordeal. Similarly, admins make one edit a year to avoid losing the tools because they know getting them back at RfA would be too hard. This is basically what all this discussion is about.

I don't have a solution for making RfA easier. Well, I have one idea - drop the pass mark so that a supermajority (66% support) is a straight pass, a majority (50% support) has a crat chat, anything else is down to crat discretion, so even a 45% support could pass if the support rationales are really good and the oppose rationales are bogus. Heck, major changes are happening in the UK on 51.9% support. Also, do more clerking so that any votes that don't meet the spirit of Wikipedia:Advice for RfA voters in support or oppose get struck. This would bring it more in line with what happens in practice at AfD. It means if you really, really want / don't want somebody to be an admin, you need to up your game and talk a better argument - just like AfD. I don't believe it will pass because there's not a widespread agreement on this, so there'll be lots of opposition. But there'll be lots of opposition for a wide variety of interesting reasons. Ritchie333 ^{(talk) (cont)} 15:37, 21 February 2019 (UTC)[reply]

I don't think it's the difference in standards per se; I think it's more that people like to feel that administrators are part of the community. If an administrator isn't participating in the community but then comes in and starts performing administrative actions, people feel like outsiders are trying to tell them what to do. I think if a higher activity threshold is introduced, it should be described as an activity ramp-up period: editors are expected to spend some time re-engaging with the community before they can resume administrative duties. isaacl (talk) 19:00, 21 February 2019 (UTC)[reply]

I have devoted a good deal of thought to this problem and have tried to deal with it constructively through

So, let's name the conflict, that is, define exactly what the problem is here:

The problem is that we have a significant share of the administrator community -- over 25% -- that has not made regular and frequent contributions in an administrative capacity for years, and has not edited regularly outside a few personal areas of interest for years, that does not respond to one-on-one attempts to engage with them, and is unlikely to return to the project. Most of these people joined in 2005, 2006, and 2007 during the rapid growth in numbers of admins and before the era of automated tools that reduced the admin workload and made it less mechanical. And yes, these accounts do pose various problems, quite real ones, to wit:

• The quality of the occasional admin actions taken by these accounts is poor, because they tend to be in topic areas where the admin in question is overly involved.
• It is, in most cases, impossible to determine whether the accounts are still operated by their original user, and not by a friend, cousin, coworker, etc.
• There is a risk that these individuals, since they are now marginally attached to Wikipedia and do not intend to use their accounts, may attempt to sell them to individuals or organizations who do not have Wikipedia's best interests at heart. In particular, I think that there is a risk that paid editing organizations would bid top dollar for accounts that will more readily allow them to subvert WP:COI.
• The ongoing retention of adminship by these individuals is creating a legitimate feeling of unfairness in those who have more recently worked hard to obtain adminship or who are in the process of doing so.
• The risk of password compromise, while small, is real.
• The volume of accounts complicates security solutions that pose dollar costs or developer-time costs on a per user basis, such as 2FA.
• The group is large enough that it confounds other analysis of all admins as a subcommunity.

There are several reasons this is hard to fix and it's important to identify those, too:

• It is impossible to come up with any sort of useful yardstick that defines a minimum level of activity that is difficult to "game" and that does not adversely affect legitimate ongoing contributors. The discussion upthread here shows that.
• Any policy change will draw significant opposition from people who are marginally active but who still identify closely with the project. There are many of these people.
• There are legitimate concerns that returning contributors may not be treated reasonably at RFA.
• Any process we adopt will have to be either a) mechanical, or b) require some sort of reconfirmation. There are legitimate reasons to dislike either alternative.
• Due to the large number of users affected, the rollout of any change will have to be handled carefully, both technically and in a way that is sensitive and respectful.
• It is, at best, difficult to confer recognition on people for their past good work that is as meaningful to them as their retention of a disused admin flag.

I have tried to come up with a complete policy proposal. It appears below. I don't like it, because it's too mechanical, but I offer it for discussion, and maybe it will serve as a springboard to something better.

I haven't worked out a way to handle the transition on this. At a minimum, we should choose a future date for it to go into effect, and notify people who may be affected by it. UninvitedCompany 00:28, 20 February 2019 (UTC)[reply]

UninvitedCompany 00:20, 20 February 2019 (UTC)[reply]

--- proposed page title --

Wikipedia:Inactive administrators

--- cut here ---

History

The first administrators were appointed in 2002. Administrator activity was not seen as a problem when policies were formalized in 2005, and administrative rights could be held indefinitely. In June 2011, an inactivity policy was adopted, but former admins returning to the project could regain their administrative tools "without further discussion" (i.e. demonstrating community support at

Purpose

Inactive accounts with administrative permissions pose several potentially serious risks to the project:

1. The security of the accounts can be compromised. The known incidents have led to vandalism of the
   protection
   ), and disruptive blocks of other administrators. It is possible that other accounts have been compromised but have not been used for egregious violations of policies and have therefore been unidentified. While the accounts of active administrators can be and have been compromised, active administrators are more motivated to take steps to keep their accounts secure, and can report compromise when it occurs.
2. There have been isolated instances of administrators who have utilized their tools disruptively after diverging from the evolving norms of the community during a period of inactivity or low activity
3. While there have been no confirmed cases related to inactivity, there is a risk that former administrators may attempt to monetize their accounts by selling their credentials or utilizing them on behalf of
   paid editing firms
   .
4. Substantial numbers of inactive or low-activity administrative accounts complicate widespread deployment of
   WMF
5. The presence of inactive accounts held by individuals who no longer have a serious interest in Wikipedia distort statistics on the number of administrators serving the project.

The goal of the inactivity policy is to encourage administrators who genuinely plan to re-engage with the project while limiting the risks of inactive accounts.

Definitions

• Administrative action means any logged action that can only be performed by administrators or other user groups with comparable or greater access rights.
• Editing day means any day, beginning and ending at 00:00 UTC, when at least one edit or administrative action is made.
• Active means a user who has five or more editing days and one or more administrative actions in the preceding twelve months.
• Fully active means an active user who has achieved ten or more editing days per month, in each of the preceding six consecutive months.
• Otherwise contributing means a user contributing to Wikipedia in a substantial capacity that does not result in locally logged edits or administrative actions. Examples include service to the
  Wikimedia Commons or other sister projects, serving as a steward, serving on the ombudsman commission
  or substantial and ongoing work building, maintaining, or operating a useful bot, script, or other tool that benefits Wikipedia.

Bureaucrats are expected to exercise the duties granted by their role while remaining cognizant of relevant community standards concerning their tasks. If a bureaucrat does not participate in bureaucrat activity[2] for over three years, their bureaucrat permissions may be removed. The user must be notified on their talk page and by email one month before the removal, and again a few days prior to the removal. If the user does not return to bureaucrat activity, another bureaucrat may request the removal of permissions at meta:Steward requests/Permissions. Permissions removed for not meeting bureaucrat activity requirements may be re-obtained through a new request for bureaucratship.

# # #

Just a small point, but users sometimes have more than one account. While administrative tools are attached to a single account, adminship belongs to the user. You may want to consider how that may affect some of the statistics, and you ought to reformulate at least the phrase "Account not currently in good standing." Accounts don't have standing; users do. --RexxS (talk) 01:25, 20 February 2019 (UTC)[reply]

UninvitedCompany you say that you are concerned that your proposal is "too mechanical". Does that mean you would prefer to see some discretionary "powers"? The current resysop process, as you know from the BN discussion closed yesterday, is entirely mechanical. - Sitush (talk) 11:01, 20 February 2019 (UTC)[reply]

Perhaps a better phrase is too "mechanically complicated" with counting editing days over various periods. I tried to capture what I saw by manually going through the contribution history and logs for dozens of editors -- some clearly inactive, some trying to contribute -- and that's the best objective set of tests I was able to create. I think that any discretionary policy is going to pose the usual problems and distortions. We especially would not want to give that to the 'crats, as they are chosen for trustworthiness and sound judgment, but are not representative of the community. UninvitedCompany 14:23, 20 February 2019 (UTC)[reply]

OK, thanks. I understand now. However, any mechanical process can be gamed. - Sitush (talk) 14:27, 20 February 2019 (UTC)[reply]

True enough. My goal when drafting was to create a policy where someone "gaming" the policy would end up engaging enough that we would be happy to have them retain the tools. UninvitedCompany 18:28, 21 February 2019 (UTC)[reply]

I think this is a good proposal, which I will be happy to support, but I think the last point (account which has been sanctioned after resigning the tools) needs to be more elaborated: which exactly sanctions will lead to ineligibility of resysop, and whether this ineligibility has a finite time duration (if sanctions have been imposed once does it mean that the user is never eligible for the restoration of the tools without RfA).--Ymblanter (talk) 18:42, 20 February 2019 (UTC)[reply]

I think a decision whether a sanction should prevent re-sysop should come through ArbCom, or failing that, a decision not to resysop should be subject to appeal through ArbCom. —Kusma (t·c) 07:55, 21 February 2019 (UTC)[reply]

• I am surprised to see such a lengthy discussion without any figures to describe the problem what is, perhaps, to solve. My impression is that there are 42 so-called admins that have done exactly zero admin actions from 2016-01-01 together with less than 13 edits from 2016-01-01, i.e. no more than 4 edits per year on the average. In fact, adminstats gives the impression that 381 so-called admins have done exactly zero admin actions in three full years. Something is broken for sure, either adminstats or what adminstats is describing. Pldx1 (talk) 23:09, 20 February 2019 (UTC)[reply]
  • I included some figures up in the "Naming the conflict" section upthread. The problem is much larger when you consider admins who have done very little in addition to those who have done nothing. UninvitedCompany 18:28, 21 February 2019 (UTC)[reply]
    • Once again: adminstats says that 381 so-called admins haven't done any admin action in three full years. Faithul reflect of a bad reality or only the bad result of some bug in the scripts ? Remark: some of the 381 have good reasons for that. But not so many !!! Pldx1 (talk) 19:28, 21 February 2019 (UTC)[reply]
      • There's always the fact that things like editing protected pages, including the main page, isn't logged. ~90% of my admin activity would fall into that category, and the same would be true of many of the main-page regulars. Espresso Addict (talk) 00:58, 22 February 2019 (UTC)[reply]
        • @Espresso Addict: check out this log. — xaosflux ^Talk 01:55, 22 February 2019 (UTC)[reply]
          • @
            Beeblebrox
            on whether such actions should or should not count, but I don't recall whether the discussion ever came to a conclusion.

(Also my log gives a grand total of 8 edits? I might have been inactive, but I wasn't that inactive... I don't think it's collating PotD or OTD or all the other dated subpages that go in and out of protection.) Espresso Addict (talk) 02:43, 22 February 2019 (UTC)[reply]

@Espresso Addict: that log was only enabled ~3 months ago. If you have an example of something not showing please let me know at my talk and I'll look in to it. — xaosflux ^Talk 03:32, 22 February 2019 (UTC)[reply]

FYI for anyone following this, phab:T216827 opened to improve matching cascade protection matches. — xaosflux ^Talk 16:22, 22 February 2019 (UTC)[reply]

• This is the reason why Bureaucrats should be allowed some leeway when deciding that someone is inactive to the point that the bit shouldn't be kept or returned without a full RfA. When seeing
  * Espresso Addict (admin);2010 40;2012 1;2013 50;2014 21;2015 196;2016 2472;2017 66;2018 735;2019 25;
  * Espresso Addict (edits);2006 2075;2007 5378;2008 3438;2009 1625;2010 4193;2011 1;2012 1075;2013 4665;2014 928;2015 2063;2016 4871;2017 1005;2018 3070;2019 663;
  * RfA 2007-09-23 60/0/1 ; Suspend 2019-02-09
  the activity is largely proven and there is no need to ask the opinion of filter 942. When seeing
  * XXX (admin);2013 5;2014 18;2015 1;
  * XXX (edits);2008 3;2009 0;2010 5491;2011 103182;2012 96001;2013 3202;2014 2046;2015 1564;2016 2069;2017 1664;2018 1609;2019 206;
  the activity is largely proven. Moreover the RfA was preventatively explaining the values of adminstats. On the contrary, when seeing:
  Where (admins);2006 469;2007 37;2008 2;2011 3;
  Where (edits);2005 14;2006 5831;2007 166;2008 24;2009 0;2010 3;2011 2;2012 2;2013 1;2014 1;2015 1;2016 1;2017 1;2018 1;2019 0;
  it is not so difficult to see that User:Where is gaming the system. The actual policy forces the bureaucrats to take iteratively the same stupid decision. Time to change, isn't it ? Pldx1 (talk) 10:14, 22 February 2019 (UTC)[reply]

What if...

I think this is some great work from UninvitedCompany. The "fully active" section intrigued me, and I think it might be the key to solving this puzzle. What if, instead of regranting sysop rights on request, bureaucrats regrant them on request only if the requesting editor is fully active, and otherwise ask the requester to come back once they have fully re-engaged with the project? That eliminates the problem of people just appearing out of nowhere, getting the bit back, and disappearing again for another year (or more), which has led to such acrimonious discussions in the past. And no one would be denied a resysopping they were told they would be entitled to: all they have to do is demonstrate that they're willing to be active enough to use it. Thoughts? 28bytes (talk) 11:08, 21 February 2019 (UTC)[reply]

Would it be difficult to get statistics on say how many current admins are fully active now?--Ymblanter (talk) 11:22, 21 February 2019 (UTC)[reply]

Or even how many users are fully active now?--Ymblanter (talk) 11:28, 21 February 2019 (UTC)[reply]

Active, semi-active, inactive. Carcharoth (talk) 11:32, 21 February 2019 (UTC)[reply]

Thank you, but the above definition of fully active (edits on 10 days per month during six months) is different from the definition of active administrators (30 edits in two months). What I would like to understand is how reasonable the requirements of fully active are. If there are hundreds or thousands of users like that it is one story, if there are dozens, it is a completely different tstory. (I know that I was fully active by this definition always since six month after the registration of the account - before six month it is impossible to conform to the definition - but on the other hand I am in 400 most active users of all times and am hardly representative).--Ymblanter (talk) 11:41, 21 February 2019 (UTC)[reply]

I do share your concerns and believe that some sort of assessment should be made, which would require automated tools to perform at scale. If there is support for this policy otherwise, someone should write a bot. If no one else steps forward, I could do it. UninvitedCompany 18:28, 21 February 2019 (UTC)[reply]

Fully active under this definition only requires editing activity. There is no requirement for taking any admin actions. It is entirely possible for people to stay fully active and engaged with the project under this proposal without ever using the tools. However, the current rules state that if you have not taken any logged admin actions for five years, you have to go back through RfA, even if you voluntarily give up the tools. The presumption being made in all this is that someone who has not taken any admin actions for five years is not active as an editor. That is not necessarily the case. I would support this proposal, but only if a clear distinction is drawn between: (i) those admins who are totally inactive; (ii) those admins who are active as editors but not as admins; (iii) those admins who are active as both editors and admins or just active as admins. I get that those who are totally inactive may no longer be conversant with current community norms, but those who are active as editors (depending where they edit) may still be aware of what is needed to use the tools to current standards. A distinction should also be drawn between those who are desysopped for inactivity, and those who give up the tools voluntarily. Carcharoth (talk) 11:49, 21 February 2019 (UTC)[reply]

Substantially all of the sister projects and other-language Wikipedias require admins to actually use their tools to retain their admin rights. I believe this is fundamentally sound policy. The relatively few highly active editors who hold admin rights but do not use them will, I imagine, find something to do with their mop from time to time, should we move forward with something like this. UninvitedCompany 18:28, 21 February 2019 (UTC)[reply]

You may believe that this is fundamentally sound policy, but until you made this change to the policy, that was not the case on the English Wikipedia. Until you made that change, it was still possible for admins in good faith to take a break from being an admin, relinquish the tools, remain active as an editor, and then ask for the tools back, with no time limits. The wording that you removed stated: "Administrators in good standing who were not considered to be in controversial circumstances when voluntarily requesting removal, may request at any time that their administrator

status be restored by a bureaucrat". I am not that active myself these days, but I have been active in the past few days to try and put my views across. At the moment, the current status is that resigned former admins who remain active as editors can request the tools back unless they have been inactive as an editor for three years. That is in line with both this RfC and this RfC. I hope that won't be changed further without a clear consensus at a new RfC. Carcharoth (talk) 13:08, 22 February 2019 (UTC)[reply

]

That is actually not true, the vast majority have no activity requirements at all and the few that do are usually just about editing not admin tool use. -DJSasso (talk) 18:47, 22 February 2019 (UTC)[reply]

I think it is difficult to have a higher activity threshold for regaining administrative privileges than to attain them in the first place, so I think any standards along these lines would have to be applied to new requests for administrative privileges as well. I think some of the thresholds are too high (ten years of ten days of activity every month? Most people aren't around for ten years, and that's a long time to have sustained activity at any thing), but I don't have an objection to the general principle of requiring some kind of ongoing activity. I think a ramp-up period prior to attaining or re-gaining administrative privileges may be desirable, where editors must demonstrate some degree of ongoing engagement with the community. isaacl (talk) 19:06, 21 February 2019 (UTC)[reply]

Perhaps the way it's written is unclear. The intent is that at some point in the last ten years, there has to have been a period of near-daily engagement lasting six months. Every successful RFA in recent history would meet that. I do like the idea that people have to engage with the project before asking for the bit back, and it's mentioned as a stand-alone proposal somewhere up the page, I believe. UninvitedCompany 19:19, 21 February 2019 (UTC)[reply]

OK, that makes more sense. "Not fully active at any point in the preceding ten years" reads like at any given point in the preceding ten years, the petitioner must have been fully active. isaacl (talk) 19:25, 21 February 2019 (UTC)[reply]

I'll just note that there are dozens of administrators who almost everyone on this page would consider to be active editors and administrators who do not meet the "fully active" criterion even now. This is particularly true of administrators whose participation includes off-wiki activities, global activites, or non-public activities. Risker (talk) 03:41, 22 February 2019 (UTC)[reply]

Going backwards?

I haven't read through every single comment on this page, but some of this feels like it's going backwards. It is known that there are a number of administrators who have edited very little, basically annually, and they retain their adminship according to current policy. There have been a number of RFCs over the years trying to prevent that. Assuming that "annually-editing admins" is a problem (which I'm not claiming it is or isn't), then a policy that makes it harder to get the admin tools back after they're gone (e.g., "meeting fully active requirements"), will result in the intermittently-editing admins holding their tools even tighter. Useight (talk) 23:50, 21 February 2019 (UTC)[reply]

Also, incremental changes to the policy makes it harder to be sure it is possible to take a meaningful break (whether a complete break or one where the former admin still edits) and be able to come back and be sure that the rules at the time you took the break will be honoured. Some people will conclude that rather than voluntarily relinquish the tools and ask for them back later, it would be better to keep the tools and make token edits and token admin actions, thus making worse the problem some people here are trying to solve. Carcharoth (talk) 12:49, 22 February 2019 (UTC)[reply]

The elephant in the room

There have been about 100 editors who have participated in this RFC. The overwhelming majority are administrators and, of those, a disproportionate number are also bureaucrats. Now, this isn't surprising - all of the proposals here affect mainly administrators and bureaucrats in one way or another. But let's look at the reality on the ground.

We're seriously debating a proposal by

We're kidding ourselves. In the abstract, many people here may think that administrators who are minimally active for an extended period are gaming the system somehow or other. But when we look at individual cases, the same people rarely express concern about the very administrators that these proposals are intended to affect; in fact, if they come back, they're usually treated as a returning prodigal son. I think it's time to stop pretending otherwise. Risker (talk) 03:11, 22 February 2019 (UTC)[reply]

• On a somewhat related note, as per my comments on the first § Elephant in the room, it makes me uneasy when someone who has been away for an extended period of time makes a proposal for a significant policy change. Now certainly someone coming at a problem fresh may introduce new ideas, but there can also be a bit of white knight perception (that is, someone coming in to save the day) that can be offputting for those who have lived through the problem in question for years. (In this particular instance, it's also changing the rules just a few months after the proposer benefited from the current ones.) Now of course this doesn't mean the proposal can't be a good one, and communities should be open to adopting and evolving ideas from anyone. But I think the desire to feel that we're in this together is a significant reason why the community would often like to see a period of prolonged engagement prior to someone taking on administrative tasks, or proposing big process changes. isaacl (talk) 03:48, 22 February 2019 (UTC)[reply]

• dissenting elephant. It seems that:
  AGK (admin);2012 2896;2013 897;2014 535;2015 98;2017 2;2018 237;2019 5;
  AGK (edits);2006 1161;2007 7199;2008 7869;2009 2441;2010 2415;2011 4033;2012 10493;2013 3576;2014 2537;2015 313;2016 1;2017 10;2018 1145;2019 199;
  UninvitedCompany (admin);2005 802;2006 515;2007 55;2008 7;2009 484;2010 348;2011 22;2012 3;2013 13;2014 2;2017 8;2018 107;
  UninvitedCompany (edits);2003 131;2004 2052;2005 2059;2006 1667;2007 1059;2008 230;2009 392;2010 586;2011 83;2012 53;2013 17;2014 13;2015 7;2016 4;2017 52;2018 1807;2019 147;
  Did I miss something ? Pldx1 (talk) 12:04, 22 February 2019 (UTC)[reply]
  Some graphs about RfA's can be seen at How old were the crocodiles when entering the bayou. Pldx1 (talk) 12:15, 22 February 2019 (UTC)[reply]
  • What are you dissenting from, Pldx1? Could you please present the pile of numbers you've inserted above into a more readable format? I genuinely can't tell what your point is. Risker (talk) 12:25, 22 February 2019 (UTC)[reply]
    • Someone said would have had his administrator permissions removed if this proposal had been in place... twice. I have checked and found otherwise after collecting some pairs (year, number). I have no doubt that you, Risker, have also worked your exercise and I would be glad to see the data you could have collected to sustain your own assertion. Best regards. Pldx1 (talk) 13:23, 22 February 2019 (UTC)[reply]
      • I genuinely can't figure out what those numbers are, they're not readable to me. Anyone can look at the edit counts of those admins. Anyone can look at the admin actions of those admins, if that is what those numbers are. But I still don't see what your point is. Are you agreeing with me that, had UC's proposal been in place, the two admins I pointed out would have been desysopped, or are you dissenting from my assessment? And perhaps more importantly, are you dissenting from my key assertion that the community itself didn't care that each of those admins had extensive periods of extremely low activity? Risker (talk) 13:53, 22 February 2019 (UTC)[reply]
        • Dear User:Risker. Anyone can look at the edit counts of those admins, Anyone can look at the admin actions of those admins. Yes, indeed. And I have looked. And your three posts give me the impression you haven't, but an impression is only an impression. Did you look into the details or not ? Are you agreeing with me that, had UC's proposal been in place, the two admins I pointed out would have been desysopped. Do you really have the slightest doubt that I disagree with you on this point ? Could you, please, do your own maths and post the details, not only a stratospherical overview ? Obviously, User:UninvitedCompany's message just below answers the key point of the problem. My own message was only answering the fact that you were suggesting that elephant-UC just not perceived he was breaking porcelan-UC. He was not. Pldx1 (talk) 15:30, 22 February 2019 (UTC)[reply]

I think that in a well-functioning community with sensible policy, AGK and I both would have lost our bits when we left (or, more likely, given them up voluntarily), asked for them back at RFA after returning, and received them at that point. One might hope that a culture at RFA would arise where people would be given their bits back routinely and without drama, when it is clear that they intend to rejoin the community in earnest. UninvitedCompany 13:25, 22 February 2019 (UTC)[reply]

@Risker: I'm going to add this contribution history as an example. While not all cases are this obvious, there are dozens if not hundreds of users who are doing this:

(change visibility) 16:32, June 14, 2018 diff hist +18‎ m Rye Brook, New York ‎ Undid revision 845102718 by 2604:2000:6B10:8000:9D45:B493:9088:F5C0 (talk) current [rollback: 1 edit] Tag: Undo [rollback] [vandalism]

(change visibility) 11:05, August 9, 2017 diff hist -425‎ User talk:Where ‎

(change visibility) 20:35, August 31, 2016 diff hist +11‎ User:Where/Header ‎ Mark Wherebot as inactive current [rollback: 3 edits] [rollback] [vandalism]

(change visibility) 06:53, August 20, 2015 diff hist +117‎ User talk:Where ‎ respond ot worm message

(change visibility) 20:13, July 31, 2014 diff hist +8‎ User talk:Where ‎

(change visibility) 11:43, August 1, 2013 diff hist -20‎ User talk:Where ‎

(change visibility) 15:11, August 8, 2012 diff hist -5‎ m Commerce minister ‎ fixed error in stub classification

(change visibility) 15:11, August 8, 2012 diff hist +12‎ Commerce minister ‎ more specific stub type

(change visibility) 20:35, August 9, 2011 diff hist +73‎ Kyar Ba Nyein ‎ Start to copy-edit the article

(change visibility) 20:22, August 9, 2011 diff hist -62‎ 1973 raid on Egyptian missile bases ‎ References had been added. Does not appear to be in need of copy-editing

(change visibility) 22:03, August 9, 2010 diff hist 0‎ Ingeborg of Denmark, Queen of France ‎ reorganize, fix spelling error

(change visibility) 16:32, August 9, 2010 diff hist -4,535‎ Ingeborg of Denmark (1175–1238) ‎ made into a redirect current [rollback: 1 edit] [rollback] [vandalism]

(change visibility) 16:32, August 9, 2010 diff hist +1,094‎ Ingeborg of Denmark, Queen of France ‎ copied some facts from Ingeborg of Denmark to this article for a merge. THis may have added some redundancy to this article, which I will fix later.

(change visibility) 22:39, November 16, 2008 diff hist +265‎ Inkjet printable DVD ‎ question about notability

That's Where (talk · contribs · blocks · protections · deletions · page moves · rights · RfA). There are other examples. I think that some of these examples make a mockery of our policy, and my many attempts to reach out individually to these users on their talk pages and via email have been ignored. That's why I believe there's a problem. UninvitedCompany 13:42, 22 February 2019 (UTC)[reply]

• Nothing stopped you at any time from handing in your tools voluntarily, UninvitedCompany. There's nothing stopping you from doing it right now, for that matter, except that I don't know what point you'd be proving. The fact of the matter is that there are activity standards for admins on a small minority of the 800+ Wikimedia wikis, and there are plenty of "well-functioning" communities that don't have any policy one way or another on this issue. I'm not going to dispute that there are dysfunctional elements in this community - but I'm hard-pressed to say this is anywhere near a significant issue to the vast majority of community members. It bugs about 100 people out of the 30,000 monthly active editors. They're just unusually vocal people. There's no evidence that the broad community cares about this. Risker (talk) 13:53, 22 February 2019 (UTC)[reply]

  Actually, I believe that there are Meta default admin activity standards, and the projects which explicitly did not opt out have their inactive admins desysopped by stewards. There are a couple of stewards who I believe watch this page, I would appreciate their clarification. (Not that it is important for the general point). As far as I know we are the only project where an admin can lose a flag in any manner and then get it back without a new RfA.--Ymblanter (talk) 14:20, 22 February 2019 (UTC)[reply]

  • • Thanks for that, Ymblanter, the page is here. The global policy on admin inactivity is defined as zero edits and zero admin actions over two years (i.e., either a single edit or admin action over a 2-year period is sufficient to maintain admin permissions). In other words, the current English Wikipedia policy, which requires a minimum single edit/action per year, is stricter than the global one. Risker (talk) 14:35, 22 February 2019 (UTC)[reply]

  It is not unknown for people to hand in the bit and run a reconfirmation RfA, or to do so after a lengthy period of time away. I can remember at least one (see Wikipedia:Requests for adminship/HJ Mitchell 3), but I think there have been others (including maybe a bureaucrat somewhere along the way?). I am mildly surprised no-one has maintained a list of these reconfirmation RfAs. Having said that, I found a list, Wikipedia:List of resysopped users, which includes about 30 're-RfA' links, though the reasons for the re-RfA are varied, some are genuine inactivity ('wanting a new consensus') ones, but mostly from the early years. The bureaucrat one I remembered was Wikipedia:Requests for bureaucratship/Cecropia2, but that was more the bureaucrat equivalent of a re-sysop request. The other reconfirmation RfA I am aware of was Wikipedia:Requests for adminship/SarekOfVulcan 2. The two most recent 'after lengthy inactivity' RfAs were Wikipedia:Requests for adminship/Opabinia regalis 2 and Wikipedia:Requests for adminship/Abecedare 2. Carcharoth (talk) 14:24, 22 February 2019 (UTC)[reply]