wp:village pump (policy)

I understand this is sort of a perennial proposal, but hear me out for this one:

Instead of censoring wikipedia, which goes against WP:NOTCENSORED, we should have a separate, kid-friendly version of wikipedia called "Wikipedia Kids"(bit like how mobile wikipedia is slightly different). This does not go against WP:NOTCENSORED, and protects children at the same time.

Many children use wikipedia for a variety of purposes(hell, I'm still a teenager) and i would rather not have people seeing some not so kid friendly stuff here.

Here is how i think it should work:

Normal version remains uncensored and has no changes

The Kids version is practically the normal version, but:

1. Sexually explicit articles cannot be accessed and are not available on the kids version(to what extent it should not be available can be debated, such as should we make them unavailable completely or just have a smaller, safe, educational version of the article that focuses on stuff the kids actually need to cover in say, biology).
2. Gory or violent pictures are unavailable. The pages are still available for reading, e.g. we still keep the nanjing massacre article up however the photos will be removed. This ensures we aren't doing stuff like Holocaust or Nanjing massacre denial while still protecting kids.

Overall this is similar in function to WP:CENSORMAIN

Would like to hear your opinion on this. Additionally, to what extent sexually explicit/violent articles is censored, and what counts as "sexually explicit" or "violent" can be debated. Thehistorianisaac (talk) 15:32, 10 April 2025 (UTC)

:Just noting that there are already a number of these in various languages. Sam Walton (talk) 15:35, 10 April 2025 (UTC)

:maybe it could theoretically work on paper as an option that can be toggled (in which case i'd be against having it on by default), but it absolutely wouldn't work out as its own site (even if it was mostly a mirror) due to the sheer size of the wp-en

:even then, i think it'd be way too hard to program, harder to enforce, and even harder to maintain, since how would those filters even work outside of trudging through the entirety of the wmf to filter things on what's effectively a case by case basis?

:lastly, it also depends on conflicting definitions of "for kids", because you know one of those ankle-biters will have to study up on world war 2 at some point, or sex, or that one time the british colonized a place, or that one time the americans killed people and took over their land manifested their destiny, or literally anything even tangentially related to any religion that isn't satirical (nyarlathotep help them if they're in a jw or mormon environment), and keeping them out of it would only really cause easily avoidable headaches consarn _{(prison phone) (crime record)} 17:18, 10 April 2025 (UTC)

::I agree on kids the "for kids" definition. That is why I would suggest for the kids version, sex-related articles with no connections to sex ed be unavailable, while sex-related articles related to sex ed only show diagrams and be reduced. As for violence, I would not suggest censoring anything other than some of the photos, or possibly even limiting it to a "Show photo-Disclaimer: may contain violence". Thehistorianisaac (talk) 17:33, 10 April 2025 (UTC)

:::Why pick on sexually explicit articles? I don't mind my children or grandchildren (the latter of which are aged five, three, and a month) accessing details about sex, but would prefer that they didn't access some other material, such as graphic violence or material about suicide. I'm sure that there are many different views from parents and grandparents. Phil Bridger (talk) 18:03, 10 April 2025 (UTC)

::::it's just the easiest example to name, really consarn _{(prison phone) (crime record)} 18:40, 10 April 2025 (UTC)

:::let's say that happens. how, then, do you know what will be taught in sex ed? how would you attempt to reduce what is shown in order to make it less explicit without touching the text? how wo- actually, having to choose to see the pictures is nice, no complaints there consarn _{(prison phone) (crime record)} 18:45, 10 April 2025 (UTC)

:Did you do some thinking on how this can be implemented and how much workforce will be required and how much bitter squabbling will follow on whether a picture of a buttocks is permitted and whether sucking the dick properly is part of sex education. (You may think the latter was a joke, but I remember seeing on a Disney Channel an episode where two low-teen girls pressed a boy to explain them how to suck the dick properly.) --Altenmann >talk 18:04, 10 April 2025 (UTC)

::i say this as a former child from a country best known for playstation 2 piracy (which is to say i knew about the hot coffee mod when i was 8): nearly anything we could do would at best do absolutely nothing to protect children lmao. if anything, it'd just fan the flames of their curiosity, because they wanna see the buttocks!! hell, even the idea of it working by censorship comes off more as pandering to overly sensitive parents than attempting to "protect" the leeches on their legs. even then, protect from what? from knowing what "fuck" means? from knowing what a peepee (that could potentially be the one in their own lower torso) looks like and does? from knowing about that angry mustache model who hated jews for existing?

::for better or worse, children will find their way into whatever they want, regardless of whether or not they can handle it (though they usually can), and drawing an arbitrary line would only make them want to cross it more than their tiny, evil brains already instinctively urge them to consarn _{(prison phone) (crime record)} 18:39, 10 April 2025 (UTC)

:This is a great idea for a third-party service, as they can select for inclusion whatever materials they feel meets their own sense of restriction. The Wikipedia license gives them the freedom to do so, and there could even be various versions with different perspectives as to what is appropriate.

:It makes a horrible project for Wikipedia itself to do, however, because then we have to establish an Official Standard for what is improper, and that will both lead to endless bickering and complaints from those who want to provide the censored version that we are not censoring the things that they wish to have censored. You can see how we would face massive complaints if we decided, say, that material on drag entertainment was suitable for kids, or if we said that it wasn't. The group control that Wikipedia projects have and our spot at the most visible source of data would just make this too fraught. -- Nat Gertler (talk) 18:10, 10 April 2025 (UTC)

::Completely agree ꧁Zanahary꧂ 05:07, 20 April 2025 (UTC)

:"For kids" versions of reference materials are usually written for a specific audience based on age/intellectual ability. To meet the expectations set up by the name, the articles should be specifically organized and written at a less complex level, which can mean different ways of breaking down topic areas as well as a different language level. :simple:Simple English Wikipedia currently exists to fill that niche, and would be a better starting point for a kids version. As you noted, though, there are a lot of objections from the community to embedding content filtering as a core function that requires altering the underlying base articles. So at present, any filtering would need to be entirely add-on and optional, and using categorization being stored elsewhere, such as on Wikidata. isaacl (talk) 18:14, 10 April 2025 (UTC)

::I was just about to note the existence of the Simple English Language Wikipedia. Isaacl beat me to it. Blueboar (talk) 18:23, 10 April 2025 (UTC)

::: Simple English Language Wikipedia is decidedly not aimed at children in the way contemplated here. It includes sexual topics, for example, and even has entries for [https://simple.wikipedia.org/wiki/Fuck Fuck] and [https://simple.wikipedia.org/wiki/Fucking Fucking] (the latter a disambiguation page), and graphically illustrated articles at [https://simple.wikipedia.org/wiki/Human_penis Human penis] and [https://simple.wikipedia.org/wiki/Vagina Vagina]. BD2412 T 00:11, 29 April 2025 (UTC)

:I could see something like this becoming its own project, similar to simple English wikipedia. I'd even contribute to it, I enjoy the mental challenge of simplifying a difficult concept into something a child could understand mgjertson (talk) (contribs) 19:26, 10 April 2025 (UTC)

::Since this discussion seems to be moving away from child-protective censorship and towards child-centred language simplification, I'll not the existence of b:Wikijunior, a worthy project. Cremastra _talk 19:53, 10 April 2025 (UTC)

:I don't trust anybody saying "but think of the children" to make any sort of rational decision about what is appropriate for kids. --User:Khajidha (talk) (contributions) 15:20, 16 April 2025 (UTC)

:This is just censorship, with all the typical problems that come with the idea (the non-neutrality of determining what is and is not appropriate). ꧁Zanahary꧂ 05:06, 20 April 2025 (UTC)

:Points 1) and 2) skate over the perennial practical issues that confound such initiatives (putting aside philosophical ones), which are: who decides what is appropriate, and who tracks what is (in)appropriate. Saying these "can be debated" is putting the cart before the horse. CMD (talk) 07:39, 20 April 2025 (UTC)

::New further idea:

::Instead of what wikipedia deems "child appropriate", what is shown on the child version can be controlled by the user's parents/guardians or school. Thehistorianisaac (talk) 23:37, 21 April 2025 (UTC)

:::We don't need to do anything to enable that, any school or institution controlling their own internet systems can selectively block urls of their choice. CMD (talk) 00:40, 22 April 2025 (UTC)

:::{{ec}} How? This would require granular tagging of content by someone. For example say a parent doesn't want their child to see articles dealing with "sex". Sexual intercourse would obviously covered by that, but what about secondary sexual characteristics, animal reproduction, sexual reproduction in plants, virginity, sexual exploitation, rape, sexual selection, pregnancy, clitoris, sex reassignment, intersex, the birds and the bees, Mull of Kintyre test, OnlyFans, Story of O, Fifty Shades of Grey, etc, etc, etc? Thryduulf (talk) 00:44, 22 April 2025 (UTC)

• I don't hate this idea. A separate version of Wikipedia along these lines could serve as an entry point for potential Wikipedians who would mature to engage in other aspects of the project, and could also serve as a place to which to point those who fret about illustrations of mature topics on the main Wikipedia. BD2412 T 00:20, 22 April 2025 (UTC)
• I'm not really sure we should do this under Wikimedia (to start, what's considered generally appropriate for children in one culture may not be in another, and is our hypothetical "kid" 5 or 15?), but if anyone wants to do something like that, Wikipedia is CC-BY-SA for a reason. So if you think "The Children's Encyclopedia" or whatever you'd like to call it is a good idea, go do that, you don't need anyone's permission. (Just remember you can't call it Wikipedia or anything close to that due to it being trademarked.)Seraphimblade ^{Talk to me} 00:32, 22 April 2025 (UTC)
• What if we started something from the opposite direction, beginning with building child-directed articles on things that virtually everyone would agree should be in such a resource (e.g., what is a Lion, what is an Alphabet, what is a Guitar, what is Multiplication, what is Pluto), with near-unanimity required to add or post a topic or image? BD2412 T 18:16, 22 April 2025 (UTC)
• :Why should this be a Wikimedia project? ꧁Zanahary꧂ 18:25, 22 April 2025 (UTC)
• :that'd still be a logistical nightmare from the start, because even the most banal topics could be a little much for some children, and as seraphimblade mentioned, the target audience could be 5 or 15, and we can't really target both, since their tastes and needs are guaranteed to clash. plus, wikipedia is right here, so anything beyond that borders on being a choosing beggar consarn _{(prison phone) (crime record)} 18:54, 22 April 2025 (UTC)
• ::There's a sex photo on the lion article, as well as an evocative description of their penis spines. CMD (talk) 00:28, 23 April 2025 (UTC)
• ::: {{re|Chipmunkdavis}} I'm not suggesting that we should copy-and-paste our current content, just that these are subjects that would be reasonably uncontroversial for inclusion as topics of coverage for a kid's encyclopedia. BD2412 T 18:27, 25 April 2025 (UTC)
• I have obtained a Herit*ge Found*tion document titled Our Real Strategy, which envisages surreptitiously encouraging the creation of Wi/kids, placing obnoxious material in it alongside contentious material that woke hostiles will defend, and the material's eventual discovery by the HF's grass-roots division. They seem confident of destroying all Wikipedia in the ensuing storm. NebY (talk) 18:00, 22 April 2025 (UTC)
• :Is this actually true that that org is encouraging this? Deeply concerning if it is. Sounds up their alley. -1ctinus📝🗨 15:45, 25 April 2025 (UTC)
• ::A few years ago I would have taken NebY's comment as a joke, but these days I'm not so sure. Phil Bridger (talk) 17:44, 25 April 2025 (UTC)
• :::regardless of whether or not it's true, i really hope it's not an implication that the op could be trying something of the sort
• :::...or that this kind of plan can work for that matter. sega didn't kill nintendo, so i don't imagine another wiki has much of a chance of killing wikipedia consarn _{(prison phone) (crime record)} 17:52, 25 April 2025 (UTC)
• ::::See reputational risk. NebY (talk) 18:28, 25 April 2025 (UTC)
• :::::decentraland didn't destroy the concept of reality, if that's anything to go off of consarn _{(prison phone) (crime record)} 18:46, 25 April 2025 (UTC)
• :If this is a joke, please clarify that! 3df (talk) 18:25, 25 April 2025 (UTC)
• ::They are also selling a very nice bridge affiliate link. NebY (talk) 18:33, 25 April 2025 (UTC)
• ::: Poe's law applies here. Anomie⚔ 13:22, 26 April 2025 (UTC)
• Anyone is free to WP:FORK Wikipedia and censor it however they want. I for one won't be a part of that project, but if others want to be, have at it. Headbomb {t · c · p · b} 18:25, 22 April 2025 (UTC)
• I like the idea of how Basque Wikipedia did this: having a second tab next to article. I'm not keen on making any type of censored version for kids, maybe except extreme violence, but see a use for explaining things in much easier terms. For medical content, the tone would be more akin to the NHS than to academic literature. We do lose a large audience on Wikipedia, which is a shame. In terms of culture, I hope that more people learn to write for an appropriately broad audience, and that our normal articles become easier to digest too. But perhaps it'll be used for the opposite ("if you don't understand the default article, go to the kids one"). —Femke 🐦 (talk) 07:17, 23 April 2025 (UTC)
• :{{tpq|a use for explaining things in much easier terms}} this is why the Simple English Wikipedia exists. We should do a much better job of making its existence known - currently it's only linked in the other languages list, where most people wont think to look for it and because this is arranged alphabetically "Simple English" can be several screens down the list on articles that exist in many languages (for me it's right at the bottom of the second page down when starting from the top of the languages section at Aspirin for example). Thryduulf (talk) 11:45, 23 April 2025 (UTC)
• The OP started this discussion to be about having a censored version of Wikipedia, but many people have taken it in the direction of having a version which has language that can be understood by kids, whether targeted at 5-year-olds or 15-year-olds or something between, and commenters have said that that role is fulfilled by the Simple English Wikipedia. These are two very different topics, but any new WMF project should be discussed at Meta, not here. I think the only thing that belongs on this project is Femke's proposal of a separate tab. Maybe such a tab could point to the Simple English version of the article, if it exists. That would also address Thryduulf's point about making the Simple English Wikipedia more visible. Phil Bridger (talk) 12:35, 23 April 2025 (UTC)
• :Exactly lol. Though now I sort of realize that a "censored kids version" is easier said than done. Thehistorianisaac (talk) 12:40, 23 April 2025 (UTC)
• I would oppose on the ground it takes away energy and time from the project we have long been engaged on.--Wehwalt (talk) 04:43, 26 April 2025 (UTC)

:Personally, I would like an opt-in ability to hide/blur sexually explicit and/or gory images. Sometimes I browse Wikipedia in a public place and don't want people around me to think I'm looking at pornography.

:On a technical level, this is very easy with AI image classifiers. Chess (talk) (please mention me on reply) 21:59, 28 April 2025 (UTC)

::@Chess What are the objective definition of "sexually explicit" and "gory" that you are using? It's been a while since I looked at the state of AI image classifiers, but last time I did the reliability was very poor (e.g. Facebook believes [https://scontent.flhr4-3.fna.fbcdn.net/v/t1.6435-9/106908189_174214387419694_8462461105739578679_n.jpg?_nc_cat=109&ccb=1-7&_nc_sid=833d8c&_nc_ohc=N4Fp1GXHDmkQ7kNvwEzsqIX&_nc_oc=AdmP1N8aT_LHCcZ6fUJDz_fcb9dgDXPzWvVakvY3dXwVsjQtMTeZbxho9b95mHSsxtpp7I_DhAtwWCrzWdOiSOGk&_nc_zt=23&_nc_ht=scontent.flhr4-3.fna&_nc_gid=5qIMr9wKooznv7QxLcyZrw&oh=00_AfFBjBo7MnQNbKT57vRwGyXjxgEM_P1lmWr7aaR9bsNMZQ&oe=68378354 this image of a white daisy] contains "violent or graphic content" while not recognising an image containing penetrative sex as pornographic (possibly because both people in the image were essentially fully clothed). I don't know if it is still the case, but distinguishing images of roast chicken from images of naked people was also very tricky for classifiers. Thryduulf (talk) 22:21, 28 April 2025 (UTC)

:::{{re|Thryduulf}} I'm fine with the false positives. Blurring anything with a bunch of flesh tones isn't perfect but it means I don't have to worry about opening something NSFW in class or at work. Chess (talk) (please mention me on reply) 22:24, 28 April 2025 (UTC)

::::Given the wide range of human flesh tones, that is a lot of images getting unnecessarily blurred. Flesh is also far from the only thing that is not safe for some workplaces (but also perfectly safe for others) Thryduulf (talk) 22:26, 28 April 2025 (UTC)

::::Are you looking at porn? No? Then their reaction only serves to identify them as 1) rude and 2) stupid. So why do you care what they think? --User:Khajidha (talk) (contributions) 23:35, 28 April 2025 (UTC)

:I think the simple solution is that concerned parents should monitor their children while they are online. They can censor it themselves if they believe it needs to be done. 2600:1700:8AEC:6810:AC99:E5E9:337:31FA (talk) 13:44, 30 April 2025 (UTC)

::1) That makes far too much sense and 2) you expect the general public to be responsible? --User:Khajidha (talk) (contributions) 19:00, 30 April 2025 (UTC)

::consider that if we make it a policy, or better yet, clarify it in some legal document somewhere, we can then safely refer to any given parent's reaction as an ill skissue. a skiss illue. a slick issun. a silk insure. a- you get the idea, a skill issue on their part, failing to protect their totally not desensitized child from seeing a girl nipple (someone please cover that with a family-friendly male nipple!). why is this important? because it would be really funny consarn _{(prison phone) (crime record)} 19:31, 30 April 2025 (UTC)

:::We sort of already do - see Wikipedia:Content disclaimer (indirectly linked from every page - Wikipedia:General disclaimer is linked (as "Disclaimers") in the footer, the content, legal, medical, risk and survey disclaimers are prominently linked from the general one). Thryduulf (talk) 19:56, 30 April 2025 (UTC)

::::how lucky! all we should need to do is point at the sign consarn _{(prison phone) (crime record)} 20:28, 30 April 2025 (UTC)

::I don't expect parents to be responsible for this if they believe all of Wikipedia's articles is for all ages with "no explicit content" Gonna eatpizza (talk) 19:18, 5 May 2025 (UTC)

:::Well if for some reason they got that idea into their heads it's their fault. I wouldn't expect a modern paper encyclopedia to be censored, and there's no reason for anyone to believe that an online one would be either. Cremastra _talk 19:32, 5 May 2025 (UTC)

:::I wouldn't consider a person who believed that to be mentally competent to be a parent, so their opinion really doesn't matter to me. --User:Khajidha (talk) (contributions) 12:19, 7 May 2025 (UTC)

::::unintentionally savage, jesus fuck consarn _{(grave) (obituary)} 12:50, 7 May 2025 (UTC)

:::::Who says it was unintentional? --User:Khajidha (talk) (contributions) 13:02, 7 May 2025 (UTC)

::::::as the absolute arbiter of debatably absolute truth, i do >:3 consarn _{(grave) (obituary)} 13:13, 7 May 2025 (UTC)

: Many years separate me from kidhood, but I still remember one thing pretty clearly: it was totally worth hunting down the stepladder so I could reach those books that were way up there on the top shelf. Sometimes I wonder if they stimulated a germ of an interest in some of the presumptively taboo topics I began to contribute to decades later at Wikipedia. Should we label the tantalizing allure of the top shelf to kids the "Baby Streisand effect"? Mathglot (talk) 20:47, 2 May 2025 (UTC)

:Expect 'the law of the United States' (see WP:NOTCENSORED) to deprecate this discussion.

:Fenke (talk) 11:02, 13 May 2025 (UTC)

:Surely the Simple English Wikipedia is what you are referring to for children, and disabled people. ThemeChat (talk) 10:55, 17 May 2025 (UTC)

:English wikipedia is WP:NOTCENSORED, but in my view there's also nothing on English wikipedia that's inappropriate for children? Some parents or guardians may disagree, and may restrict their own children's access to wikipedia, but many authoritarian governments also disagree, and restrict their citizen's access to wikipedia, and we don't censor wikipedia on their behalf either! This is ultimately a fake problem, there is no solution needed. Psychastes (talk) 22:21, 24 May 2025 (UTC)

Several problems with this; here are two of them: 1. Everyone has their own version/definition/opinion of what kids shouldn't see. And so it gets classified based on politics. 2. So if the uncensored Wikipedia is readily available, what's to stop kids from looking at it? And they will. North8000 (talk) 19:43, 5 May 2025 (UTC)

:Seconded. Kids come in various ages and parents have different opinions on what's appropriate. Waste of effort.  Mr.choppers | ✎  19:59, 5 May 2025 (UTC)

::Discussions like this just make me want to tell everyone involved to read No Future by Lee Edelman. Simonm223 (talk) 13:26, 7 May 2025 (UTC)

I really agree that there should be a kids wiki.cl200 talk — Preceding unsigned comment added by Cl200 (talk • contribs) 17:17, 15 May 2025 (UTC)

• RE: "protects children at the same time", As mentioned in several places above, protect them from what? who decides? And as also mentioned above several times, if they already have access to Wikipedia, a separate version is not going to protect them, if a child has access to click on whatever they want, nothing we do here is going to "Protect the children". Jeepday (talk) 12:05, 23 May 2025 (UTC)
• I would imagine that there are applications which allow parents to set parameters of what websites their children can visit (and the Wikipedia is probably outside of those parameters). BD2412 T 23:27, 24 May 2025 (UTC)
• :There totally are, and the reasons I would see the Wikipedia blocked is if its a whitelist system (which would be manually set by the parent) or it uses keyword parameters. Gonna eatpizza (talk) 16:25, 27 May 2025 (UTC)

Proposal:

Increase the WP:SEMI threshold by raising the minimum edit count to 50–100 edits and the account age to 10-15 days.

Rationale:

I think, the current requirement — just 4 days and 10 edits — is low to effectively deter vandals, sockpuppets, and other bad-faith actors. Disruptive users can easily create accounts, meet these minimal thresholds, and gain access to semi-protected pages, where they cause damage and then abandon or switch accounts.

Raising the bar would introduce a small but meaningful barrier to abuse without significantly affecting good-faith contributors. Most constructive editors will naturally meet the higher threshold over time, while bad actors would need to invest more effort, making large-scale abuse more difficult. Cinaroot (talk) 00:52, 16 May 2025 (UTC)

:Have you seen a lot of problems caused by accounts that met the prior threshold but would not meet the new one? It's not been my experience, but yours may differ. Really, most vandals seem deterred by any additional effort at all. -- Nat Gertler (talk) 01:10, 16 May 2025 (UTC)

::i think you are right about vandals. most first time vandals and IPs seems to back off. But not socks. i think i'm mostly talking about socks. they are persistent and motivated - there are socks who evaded extended confirmed protection by making repeated edits and revert on talk pages and get caught. What i want to see most is increase waiting time 15 days min. Cinaroot (talk) 01:46, 16 May 2025 (UTC)

::LTAs quite commonly just warehouse accounts for 4 days and then race to 10 before moving on to disruption This happens on a daily basis, I'm reluctant to link many examples to avoid glorification, but for just one recent case see [https://en.wikipedia.org/wiki/Special:Contributions/HappyMeal1121]. The real issue is that warehousing accounts for additional days requires no effort at all, while the obsessive motivated LTAs are unlikely to be deterred much by a higher edit number. Meanwhile, AC is tied to several other things, so the impact on good-faith new editors will be substantial; even with the existing low-bar, most accounts never become autoconfirmed. 184.152.65.118 (talk) 01:57, 16 May 2025 (UTC)

:::to be clear - i'm not asking to change WP:AUTOCONFIRM criteria. only WP:SEMI Cinaroot (talk) 02:07, 16 May 2025 (UTC)

::::made a mistake in post haha Cinaroot (talk) 02:08, 16 May 2025 (UTC)

:::::The confusion arises because WP:SEMI has always been linked with AC. I suppose it is possible to unlink them, or to add another level of protection in between that of SEMI and ECP and such proposals might be better for WP:VPI, but implementation would pose challenges both practical and technical, the impact on goo-faith new users is going to be large and the benefits highly uncertain. 184.152.65.118 (talk) 02:17, 16 May 2025 (UTC)

::::::I’m proposing additional restrictions to semi-protection, not changes to autoconfirmed (AC) user rights. When an article is semi-protected, it’s acceptable—and often preferable—that new users and AC cannot edit it. Lets take some examples (terrorist attack or a plane crash) the main article and related pages are frequently targeted by vandalism or edits pushing biased narratives or containing poor-quality contributions.

::::::It’s often difficult to secure extended confirmed protection (ECP) for related articles, so we rely on semi-protection. However, the current threshold of 4 days and 10 edits is too low to deter bad-faith actors or sockpuppets. Increasing the required account age and edit count for semi-protection would give other editors to evaluate suspicious behavior, improving our ability to detect and report disruptive users easily. Cinaroot (talk) 03:21, 16 May 2025 (UTC)

:If the bar were set to 100 edits this would prevent editors with 5-99 edits—approximately 30% of active editors per [https://en.wikiscan.org/ WikiScan]—from editing semi-protected pages. I suspect that most editors in that range are skewed towards the lower end, making even a 50-edit threshold prohibitive. voorts (talk/contributions) 01:55, 16 May 2025 (UTC)

:::I'm not sure it would have a very significant impact in the big scheme of things. There aren't that many indefinitely semi-protected articles, less than 12000 it seems. And if you look at the activity for those articles in the last 30 days, there are a couple of thousand accounts with edit counts <= 100 making about 4800 edits to about 1800 articles. Another way of saying that is that potentially annoying a couple of thousand different people in a month is probably not ideal. Sean.hoyland (talk) 16:52, 20 May 2025 (UTC)

::made a mistake - i was talking about WP:SEMI Cinaroot (talk) 02:10, 16 May 2025 (UTC)

:::Yes. And I've explained why your idea is a bad one. voorts (talk/contributions) 12:02, 16 May 2025 (UTC)

::::do you have other idea's to discourage socks?

::::new user flooding the platform to edit after a terrorist attack or plane crash etc... these user's are also disruptive and 4 days AC restrictions is not enough. Cinaroot (talk) 00:23, 20 May 2025 (UTC)

:::::If you want to discourage socks/vandalism, the best way to do that is to fight it. I suggest checking out WP:CVU. voorts (talk/contributions) 00:56, 20 May 2025 (UTC)

::::::At least for the Arab-Israeli conflict "contentious topic area" I don't think there is any evidence to support the view that ban evading actors employing sockpuppetry are discouraged by our existing countermeasures. There is a large cost-benefit asymmetry, and we appear to be on the wrong side of it. Sean.hoyland (talk) 04:43, 20 May 2025 (UTC)

:I'm not convinced this will bring sufficient benefits to be worth it. The current requirements are already enough to deter most casual vandalism and clueless editing, so increasing the requirements will not have any impact on that at all. It doesn't deter all intentional bad actors, but those people will be motivated to and able to do reach (via gaming or otherwise) the new threshold in the same way they do the current one - the evidence for this is that ECP doesn't stop those people. On the other hand, it does significantly impact good faith new editors, raising a barrier to entry. On balance, I think it would be a net negative. Thryduulf (talk) 10:43, 16 May 2025 (UTC)

::Do you have any other idea's to prevent socks? maybe changing ECP 30 days waiting period - to 30 days where where each day user must have made at-least one edit. this will prevent - socks creating account and waiting 30 days and them making 500 edits to game the system Cinaroot (talk) 00:20, 20 May 2025 (UTC)

:::30 consecutive days of editing is a very high barrier to entry (for example, your user page prominently notes that your longest streak is only 17 days) and just as easy to game for bad actors. While dealing with disruptive socks is important, it is at least equally important that we don't also prevent new good-faith editors from contributing. Thryduulf (talk) 03:48, 20 May 2025 (UTC)

::::New accounts without EC can contribute where EC is required by posting edit requests. That's not really a barrier in practice for bad actors as you note because gaming EC is trivial and Wikipedia provides several tools to help new users do it. Sean.hoyland (talk) 04:33, 20 May 2025 (UTC)

::::Not consecutive - just 30 days of editing ( even separately )... anyway - i guess my idea's are no good :( Cinaroot (talk) 04:49, 20 May 2025 (UTC)

:::::It might be interesting to see, among the few people who ever make 500 edits, how long it usually takes them to do so. Maybe 500 edits/90 days would make more sense than 500/30. Wikipedia:The Wikipedia Library is 500 edits + 6 months. WhatamIdoing (talk) 05:26, 20 May 2025 (UTC)

::::::I've tried to look at that out of curiosity, but also to see whether there is a relationship between the speed of EC grant acquisition and the likelihood of ban evasion/blocking (and there does appear to be a relationship). But to see the results you will need to view files on Google drive, which some people don't want to do. The results are for all of Wikipedia and for accounts that have touched things in PIA topic area subset, from 2018 onwards. That last run was a couple of weeks ago. The link is [https://drive.google.com/drive/folders/1A6tbvvZP0niUxUWfr9aHsLe3v679Aprz?usp=sharing here if interested]. Sean.hoyland (talk) 08:35, 20 May 2025 (UTC)

:::::::@Sean.hoyland, can you give me a two-sentence summary? I'm hoping for like "when you exclude accounts that later ended up blocked, then the median editor took ____ months to make their 500th edit" or "Only ___% reached that within 30 days". WhatamIdoing (talk) 21:06, 21 May 2025 (UTC)

::::::::{{u|WhatamIdoing}}, here are some stats for 2018-JAN-01 to 2025-MAY-06.

::::::::*There were 36824 new EC grants in total. If you exclude accounts that were later blocked, it's 31925 new EC grants. (Update: I should have noted here that these grant counts exclude the 78 accounts where the registration date is unknown because it's not possible to calculate the days from registration to EC grant timespan and include them in the stats. Sean.hoyland (talk) 10:07, 23 May 2025 (UTC))

::::::::*I've included percentiles. The median editor took 922 days to obtain an EC grant. The median becomes 1182 days when accounts that ended up blocked are excluded.

::::::::*1236 accounts were issued an EC grant 30 days after registration. That becomes 787 accounts when accounts that ended up blocked are excluded.

::::::::

::::::::

::::::::New extendedconfirmed grants (includes 78 accounts without registration dates)

::::::::

! {{vertical header|EC year}}

! {{vertical header|non blocked count}}

! {{vertical header|blocked sock count}}

! {{vertical header|blocked non sock count}}

! {{vertical header|total new EC grants}}

! {{vertical header|sock percent}}

|-

| 0

| 2018

| 4846

| 379

| 256

| 5481

| 6.91

|-

| 1

| 2019

| 4307

| 426

| 237

| 4970

| 8.57

|-

| 2

| 2020

| 4541

| 512

| 261

| 5314

| 9.63

|-

| 3

| 2021

| 4556

| 588

| 281

| 5425

| 10.84

|-

| 4

| 2022

| 4016

| 428

| 250

| 4694

| 9.12

|-

| 5

| 2023

| 3952

| 413

| 227

| 4592

| 8.99

|-

| 6

| 2024

| 4296

| 368

| 176

| 4840

| 7.6

|-

| 7

| 2025

| 1488

| 63

| 35

| 1586

| 3.97

|}

::::::::It's good that you asked this. It made me realize that the enwiki database is not a 100% reliable source for account registration dates. There are gaps. Oddly, a small number of the almost 37000 accounts that have acquired an EC grant since 2018 are missing registration dates (Basie and Pnslotero are 2 examples), so you can't always calculate a time from registration to grant value without pulling data out of the centralauth database. Weird. Sean.hoyland (talk) 14:38, 22 May 2025 (UTC)

:::::::::The user creation log seems to have begun in September 2005 so any accounts registered before that don't have an entry. For example my alt Awkward42 was created in January 2005 so when it eventually becomes EC (with only 174 edits in 20 years that's not going to be soon) it won't be possible to calculate a duration for that. That's unlikely not impossible for Baise, but seems very improbable to be the reason for Prslotero - was the account perhaps created at a different wiki? Thryduulf (talk) 16:51, 22 May 2025 (UTC)

::::::::::Yes, Prslotero seems to be a bit more mysterious. The global account data suggests the enwiki was the first account attached. But I'm now realizing that centralauth database doesn't necessarily tell me when an account was actually registered locally. Sean.hoyland (talk) 04:37, 23 May 2025 (UTC)

:::::::::Thanks, @Sean.hoyland. This is awesome. Can you check my interpretation? Right now, I think these sentences are true:

:::::::::* 13% of all EC accounts end up blocked. (4,899 blocked EC accounts, 31,925 non-blocked EC accounts, 36,824 total EC accounts)

:::::::::* 3.3% of all EC accounts achieved EC status on Day 30. (1,236 EC-30 accounts compared to 36,824 total EC grants) However, only 2.5% of non-blocked EC editors (787 non-blocked EC accounts) achieve EC status that soon, compared to 10% of blocked EC editors (449 blocked EC accounts).

:::::::::* 36% of the accounts achieving EC status on Day 30 end up blocked. (449 blocked accounts out of 1,236 EC-30 accounts)

:::::::::* 91% of blocked EC accounts take at least 30 days to achieve EC status.

:::::::::* 90% of all EC editors take at least 60 days to achieve EC status.

:::::::::* 90% of non-blocked EC editors take almost 90 days to achieve EC status.

:::::::::Overall, the chance of an EC-30 account eventually getting blocked is much higher than, e.g., an EC-90 account or an EC-365 account or an EC-1200 account. This probably isn't easy, but I'd be interested in seeing a "survivor" analysis: How long does an EC-30 account have to survive, to have the same chance of being blocked during the next month as an EC-60 or EC-90 (etc.) account? Do the curves ever merge?

:::::::::(I miss seeing Basie's username turn up on my watchlist.) WhatamIdoing (talk) 16:55, 22 May 2025 (UTC)

::::::::::{{u|WhatamIdoing}}, I'll try to have a proper look at this later, but in the meantime, I've uploaded one of the results pngs that are on Google drive to Wikipedia for interest - :File:New extendedconfirmed grants - extendedconfirmed privilege acquisition statistics - all wikipedia - 2025-05-06.png. Sean.hoyland (talk) 04:37, 23 May 2025 (UTC)

:::::::::::Looking at the graphs, the ban evaders decrease significantly around EC-90 or EC-120 days and almost level off around EC-180 days. Non-socks level off sooner, around EC-60. WhatamIdoing (talk) 05:05, 23 May 2025 (UTC)

::::::::::::The registration to EC grant density plot is quite interesting in what it suggests about the probability of ban evasion as a function of EC grant acquisition speed. Another interesting thing that is not apparent from these stats is that the ban evasion rate for accounts with newly acquired EC grants bouncing around in the 6-10% range is substantially higher than for accounts in general. EC protection is a honey-trap for ban evading actors (or maybe it encourages ban evasion, who knows). If you randomly sample 1 million articles and look at the percentage of accounts blocked for ban evasion, it bounces around in the 2-3% range (with the caveat that we can only see apparent ban evasion rates, which might be substantially different from actual rates). So, it seems that editcount/account-age based privileges can concentrate ban evading actors in subpopulations. Sean.hoyland (talk) 07:12, 23 May 2025 (UTC)

{{outdent}}

{{u|WhatamIdoing}}, re {{tq|Can you check my interpretation?}}, they all look correct to me, but I guess 10% is 9%, almost 10%. A fun fact is that some accounts acquire EC after their block because they retain access to their talk page and edits there get them over the EC edit count threshold. Another fun fact is that getting blocked on the same-ish day as acquiring the EC grant happens quite often e.g. accounts with editcount >= 500 that became inactive before the EC privilege existed are compromised, immediately acquire the grant and are blocked. Regarding {{tq|I'd be interested in seeing a "survivor" analysis}}, see below, first attempt. It's for all accounts, including those blocked later.

this plot.

Sean.hoyland (talk) 05:37, 25 May 2025 (UTC)

:Thanks! WhatamIdoing (talk) 08:34, 25 May 2025 (UTC)

::I have taken this information and turned it into an proposal to increase the requirements for extended-confirmed status. Please see Wikipedia:Requests for comment/Extended confirmed definition#Survey. WhatamIdoing (talk) 20:12, 25 May 2025 (UTC)

:I don't think this is the way forward. If there are continued problems, then extended confirmed should be used. If not, then not. Maybe some sort of script (if it doesn't exist already) to highlight non-extended confirmed editors or newer accounts. This way its easier for (more experienced) editors to see that they need to take a closer look at an edit or not. I think blocking good-faith edits ultimately is harmful for the project as a whole and should only be reserved for the extreme cases. We all were new accounts at some point. AC/SEMI is a low-bar to stop the obviously bad-faith vandalism, while keeping a balance for newer editors to contribute and edit. Raising the bar is a bad move imo JackFromWisconsin (talk | contribs) 16:54, 16 May 2025 (UTC)

::{{tq|Maybe some sort of script (if it doesn't exist already) to highlight non-extended confirmed editors or newer accounts. This way its easier for (more experienced) editors to see that they need to take a closer look at an edit or not.}}{{pb}}Such script already exists! It marks user's with an emoji that indicates their highest user role. User:Bugghost/Scripts/UserRoleIndicator Tarlby ^{(t) (c)} 16:58, 16 May 2025 (UTC)

:I also agree thresholds shouldn't be increased, after reading all the above -- we don't want newcomers to be deterred, and sockpuppets are determined. Mrfoogles (talk) 19:30, 22 May 2025 (UTC)

I've started an RfC about what guidance, if any, there should be for bonus and alternative tracks in album articles in MOS:ALBUM: Wikipedia_talk:WikiProject_Albums#RfC_on_bonus_and_alternate_track_listings. Thanks.--3family6 (Talk to me|See what I have done) 11:56, 16 May 2025 (UTC)

Starting this thread as a spin-off from WP:BN (permalink). At the very least, I think two things need to happen:

• Exclude gaming inactivity or inactivity as a valid criterion for using the newly established recall process.
• Align current inactivity standards to what the wider community expects of "active" administrators.

I believe both of these are required for a more constructive recall process, such that eventually the question if someone has gamed the process does not need to arise. qedk (t 愛 c) 12:49, 25 May 2025 (UTC)

:Words cannot express how strongly I oppose the first half of this, qedk, at least as applies to gaming. Are you actually, with a straight face, proposing that we amend the procedure meant to hold admins accountable for violating policies, in order to exempt violations of a policy we routinely block non-admins for violating? I'd strongly encourage any other admins to consider what supporting that would convey about whether they support holding admins to the same standard as non-admins. {{pb}} Forbidding petitions based only on inactivity without a gaming element... meh. Let's worry about that if it ever happens. Upping inactivity rules, sure. Put me down in support of anything more strict than the status quo. But it absolutely should not be paired with amnesty for admins who violate a specific policy. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 12:59, 25 May 2025 (UTC)

::I am, with a straight face, saying that. The community should align expectations from admins and make it sufficiently clear such that the question if someone is gaming does not need to be a problem that the community needs to solve. It's a waste of the community's time and energy that is better spent elsewhere. I also do not appreciate the comparison to the other actually disruptive applications of gaming the system (a guideline) - that is a strawman argument that is not at all pertinent to the problem we're seeking to solve. --qedk (t 愛 c) 13:13, 25 May 2025 (UTC)

:::If your proposal is based on the idea that admins making de minimis edits to evade the inactivity rules for years, as they drift farther and farther out of touch with community norms, is not disruptive, then I think that's a better rebbutal to the proposal than anything I can say. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 13:45, 25 May 2025 (UTC)

::::The incorrect assumption being made here is that just because they're inactive (or gaming the system, which is not ethical to be clear), they're unaware of community norms. No one is perfect, and like every other editor, it's also unfair to expect admins to be subject to assumptions of gaming and non-awareness of community norms. Similarly, inactivity itself can be due to a multitude of reasons, we don't care about why they are away but what we should care about is - what level of inactivity is unacceptable from a currently standing admin - such that, we're sure that they are not aware of present-day community norms? --qedk (t 愛 c) 17:19, 25 May 2025 (UTC)

::::I would argue that an admin that does nothing, by definition, can't be disrupting anything. If that admin then does something bad/wrong, whether due to being out of touch with the rules or otherwise, then that would be disruptive, yes. Or if the admin's account got hacked, that would be disruptive, too. Or people bringing up conversation about that admin (via recall or otherwise), then that person is causing a disruption about the admin, but the admin didn't cause the disruption. A number of disruptive things can stem from the admin's inactivity, sure, but the inactivity itself cannot be a disruption. Useight (talk) 16:46, 28 May 2025 (UTC)

: Fiddling with various thresholds to trying to have rules that can't be gamed doesn't seem worthwhile to me; it reminds me of the adage about how trying to make something foolproof just leads to the universe inventing a bigger fool. If you make the requirements stricter, that just increases the incentive for people to game it.{{pb}}As to the counterproposal of having the rules say "no gaming", meh, discussion will still be needed to determine whether there's actually gaming or not. Since we already have Wikipedia:Administrator recall, why would it be better to hold discussions over whether an admin is gaming or not at WP:BN instead of there? Do we trust 'crats to do a better job than whatever mob shows up at the recall petition? If so, then perhaps the recall process needs to be revisited as a whole instead of trying to make carve-outs. Anomie⚔ 13:22, 25 May 2025 (UTC)

:There are problems with administrator recall, as I've expressed in multiple discussions about it in various fora, but this is not the way to solve them.

:What the admin activity requirements should be is a separate question. I don't think there is a way of avoiding thresholds completely but they are evidently not quite aligned with what the community desires, which is for admins to remain engaged with the project and to keep in touch with evolving policies and norms. If we want to continue measuring this by edit count (which is easy to measure but imperfectly reliable) then something closer to the following might work:

::{{green|Admins are expected to demonstrate that they remain in touch with the community. Ideally admins should be making multiple non-trivial edits or logged actions most months, but must average at least 10 edits or logged actions every six months over a two year period. Administrators who consistently fall below this activity level or who attempt to game the requirements may be subject to recall no sooner than three months after being alerted.}}

::{{green|Administrators will be procedurally desysopped if they fail to make at least one edit or logged action in any 12 month period or 100 edits in any five-year period. }}

::{{green|Administrators who appear to be gaming activity requirements and/or who appear to have not kept up to date with policies or community norms may be subject to recall no sooner than three months after being alerted (this does not preclude recall petitions being initiated over other matters, including egregiously bad departures from contemporary community norms). }}

:This is explicitly not a proposal at this stage but an idea presented for feedback. Thryduulf (talk) 15:31, 25 May 2025 (UTC)

::On a personal level, if I correctly understand your proposal that {{tq|admins...must average 10 edits or logged actions every six months over a two year period}}, and it had been in force in the past, I would have been forced out on more than one occasion, and yet I do not feel that I have ever lost touch with policies and community norms. Now, I am not a particularly active admin, although, to my continuing surprise, the number of my logged actions is in the middle range of all admins, and losing the bit would merely free my time up a bit, I cannot support such a strict rule. Donald Albury 16:10, 25 May 2025 (UTC)

::I would prefer not to speculate on admin motivations by using "game" (as I suggested on the admin recall talk page). I think it's sufficient to discuss considerations of ongoing familiarity with current community guidance and standard practices. isaacl (talk) 16:12, 25 May 2025 (UTC)

:As I stated on the admin recall talk page, I do think the community should have a discussion on the expectations on admin activity. I feel the community wants its admins to have some ongoing connection with the community, and uses recent edits as a metric to determine this. It's imperfect, but I think it's a reasonable approach, with corner cases able to be handled separately. I also made a suggestion in that same discussion that one possible approach is to shift the emphasis to no-fault removal of administrative privileges to reduce the scope for security vulnerabilities, with the current restoration process available for admins to regain their privileges. This would help smooth the way for a much higher activity threshold, which can better satisfy community expectations. isaacl (talk) 16:07, 25 May 2025 (UTC)

• I think we need to step back and figure out what we're trying to accomplish here, because the best solution depends more on the root goal than the intermediate one of "desysop inactive admins". Some possibilities:
• Are we trying to improve security, by minimizing the chances that a compromise of a long-inactive administrator account will go unnoticed by its owner?
• Then the initial system of zero edits in a year was the right answer all along. If that proved insufficient - say, there were multiple compromises of minimally-active admins that were used to slant content, as opposed to the usual "Hey, I've just deleted the main page and blocked Jimbo, obviously I need to be blocked" - then the proper refinement is to shorten the time period, not to introduce a longer-term one like our current 100 edits/5-year rule.
• Are we trying to ensure each admin is still using their bit to benefit the encyclopedia?
• Then you measure inactivity by the number of logged actions in mainspace only, with the understanding that admins who are close to the threshold will be flagged for manual review, and clearly-frivolous actions like creating a page just so you can immediately speedy it as G7 don't qualify.
• Do we want to show that every admin maintains the community's confidence?
• Then the most direct way is to periodically run every admin through the recall process, or skip it and go directly to re-RFAs. {{Deityname}} have mercy on the souls of the first few dozen.
• Is it most important that whatever inactivity rule we have is hard to game?
• Then solely basing it on a hard limit of some sort is a bad idea, because any hard limit, no matter how absurd, can be gamed. Grade on a curve on instead: three or four times a year, find the 1% or 5% or 10% or whatever admins who've been the least active by whatever measure - edit count or logged actions or support percentage in reRFAs or whatever - and desysop those. Maybe combine with a moderately-absurd hard limit on the order of 100 edits per month to protect us from the "best" case where even that would fall into the lowest 1%.
• —Cryptic 16:25, 25 May 2025 (UTC)
• :I am reminded of the old Microsoft rule that the bottom 5% (or whatever it was) of productive programmers were routinely sacked to go work for Boeing.
• :Having said that, it's an egregious rule (like so much early Microsoft thinking - we can argue 'early'), especially for a volunteer project. And I do note that becoming an admin these days is actually quite hard - the RfA process these days is pretty rigorous. It wasn't always so - WP has changed a lot.
• :So maybe you have a review process, say a 5-year review. No hard bright lines, just consensus that x is still a productive admin, broadly construed. It could even be a peer review - admins agreeing on admins rather than making it open to all editors. As a non-admin, I'd support that, TBH... Best Alexandermcnabb (talk) 17:06, 25 May 2025 (UTC)
• ::{{ec}} The problem with a temporary tenure before was that RfA was atrocious and it would be unfair to subject editors to it more than once. I think with elections, that alleviates the issue, if we make elections more frequent, I do not think that's a bad idea. This would also make adminship not a big deal as it was originally intended. --qedk (t 愛 c) 17:25, 25 May 2025 (UTC)
• :Re your 3rd point, Cryptic, in a perfect world I'd like to see proper admin reconfirmations here. Nothing heavyweight, something like a straw poll where the two options are "reconfirm" and "send to re-RfA", max 50 words' rationale per user, 50% to pass. But I don't expect that to happen anytime soon, so, re your 4th point, I think the easiest solution is to create a discretionary zone, just like we have at RfA and with resysop activity. This could address both inactivity-gaming and the loophole that allowed Gimmetrow to be resysopped in the first place. Something like, {{xt|Bureaucrats may at their discretion deny resysop if they feel the requester would not have met the WP:RESYSOP criteria but for actions taken solely for de minimis compliance with those criteria. Similarly, bureaucrats may at their discretion include an administrator in the monthly inactivity desysop list if they feel that the administrator would meet the criteria for an inactivity desysop but for such actions. The resysop procedures for such an ex-administrator are the same as for any other administrator desysopped for inactivity.}} In an eventuality where bureaucrats have such power, I would be okay with inactivity-gaming being exempt from RECALL, although I wouldn't support the idea either. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 17:23, 25 May 2025 (UTC)

:A few thoughts:

:*I don't want a minority who want a stricter admin activity threshold than community consensus supports to be able to do an end-run around consensus by recalling admins who are likely too disengaged to desire running a new RRFA promptly. I don't think the two most recent recalls are a problem, but I do want caution.

:*I don't think our current admin activity policy is significantly misaligned, but I do wish that crats used discretion more aggressively for any account just skirting the exact fringes of our brightline policy. A good analogy is edit warring versus the 3 revert rule - our inactivity policy is the analogue of the 3RR as a brightline where if you cross it you are almost certainly edit warring/inactive, and we can take action in most cases accordingly. However, you absolutely can edit war while staying within the letter of the 3RR, and you absolutely can be more inactive than the community is willing to tolerate while staying inside the letter of the inactivity policy.

:*I dislike the idea of carving out reasons why an admin explicitly may not be recalled, which feels inelegant.

:*Perhaps a good solution for less active admins would be to apply the "grace thresholds" to the next RRFA/admin election they run in, even if it falls outside the 1 month window. The tools would still be removed after a month. That way, an admin recalled in this way would be able to return to activity, make a number of edits to show that they are in touch with community norms, and then make a compelling case for getting the bit back on their schedule.

:Tazerdadog (talk) 18:04, 25 May 2025 (UTC)

:I disagree with the premise that the current activity requirements are out of sync with community expectations. The cases that have made their way to successful recalls were based on pretty clear gaming and unmet promises to return to activity. Enforcement of inactivity requirements is better handled through bureaucrats. Recall is a poor fit for this type of policy enforcement because it's a heavyweight community process and it adds unnecessary delays to resolve what should be a narrow and factual question. I think it's also extremely unlikely for any re-RFA to be successful in cases of inactivity so it's just a bad fit in every respect.

:I think Tamzin's proposal for a discretionary zone would also enhance the bureaucrats' ability to handle these exceptional cases quickly and fairly, specifically by allowing bureaucrats to desysop in cases where activity barely meets the technical threshold but clearly falls short in substance. A discretionary zone also allows for bureaucrats to consider evidence of activity that may not be visible in the logs.

:Finally, it might help to add a requirement that administrators being resysopped after a period of low activity (e.g., less than 1,000 edits or actions in the previous two years) must meet a somewhat higher threshold to retain their tools (perhaps 1,000 edits and 100 actions in the following year). This also seems like something appropriately handled by bureaucrats. Daniel Quinlan (talk) 18:27, 25 May 2025 (UTC)

:I would agree with the above views that the recent "inactivity" RECALLs were more accurately "gaming" and "communication" (in different ways) recalls. Neither demonstrated a clear community desire to move inactivity requirements, and even less so raise any great ideas about what new requirements could be. As for excluding admins from guidelines/policies, that seems the opposite direction to what the RECALLs wanted, which is admins in touch with community expectations. CMD (talk) 01:54, 26 May 2025 (UTC)

• I don't see the purpose of this. The recalls occurred because of a clear end run around existing rules. The rules don't need to be changed just because someone tried to get around them. Community action when someone is end-running around edit warring, as another example, is more than sufficient to deal with the problem rather than trying to make edit warring policy more extensive. Similarly, the recall process has worked perfectly as a community action system for someone end-running around the inactivity rules (and other admin problems re prior recall cases). Recall is already an existing method to deal with such edge cases and has worked perfectly well so far. I see no reason to change inactivity rules at this time. And I certainly see no need to alter recall rules until we actually have examples of inappropriate cases going through, which hasn't occurred even a single time thus far. Silverseren^C 02:20, 26 May 2025 (UTC)
• : That is exactly the point of concern. Gaming is not a bright-line policy that we can apply to inactivity because it's inherently impossible to figure out the actual situation that these real people are in and even whether if someone's actual intention is to game the system. When we talk about people not being allowed to game the system we should first and foremost be talking about disincentivizing gaming rather than punishing people who we think are gaming the system. If it was true that the current numbers are aligned with our inactivity standards, why do we care so much that they're gaming to exactly keep their rights? Is it true that they engaged in misconduct or showed that they were somehow out of touch with community norms? The answer is a resounding no. As I said before, the current example is one of recency bias where there is certain validity to the claims of gaming - but when we use for-cause processes to conduct processes that effectively sidestep procedural policy, it's only a question of when, rather than if, things will go wrong. --qedk (t 愛 c) 08:44, 26 May 2025 (UTC)
• This may not be very helpful, but I looked at some data regarding how many accounts are (even possibly) out of line with community standards and have or could have admin flags. 1) I assume that the activity of the 438 accounts at Wikipedia:List of administrators/Active is not in question. 2) There are 307 semi-active admin accounts. Some of these could be making minimal edits, but they would eventually hit the 100-edit/5-year cutoff in that case. 3) There are 94 inactive admin accounts. Changing the standards for these would really only matter if they actually lose their flags for inactivity and then ask for them back. 4) There are about 200 former admin accounts that were desysopped in the last five years due to inactivity or resignation (not "for cause" or other reasons that would make the account ineligible for re-sysopping) and haven't hit the logged action rule. Of these, Wikipedia:Former administrators/reason/resigned#Active and Wikipedia:Former administrators/reason/inactive#Active seem to indicate there are 78 (58+20) active accounts that would normally be eligible for re-sysopping after a hold to review the case. That leaves the list at Wikipedia:Former administrators/reason/inactive#Admins desysopped due to having made fewer than 100 edits in five years (criterion 2), who would need to establish a return to activity, a list of 123 accounts. If that's the list we are considering, then there is no ongoing risk of damage and a relatively small number of accounts that would be the target of any new rule; it seems like a case-by-case evaluation is possible since few of them are likely to ever ask for the tools back. If we are really trying to redefine inactivity to encompass more of the 307 semi-active and 94 inactive admin accounts, then I wonder what evidence there is of an actual problem under the current standards. How many of those 401 admin accounts are actually out of line with community standards and unable to be taken care of eventually under the current inactivity rules? (Open to correction on any classifications I've gotten incorrect here.) Dekimasuよ! 06:08, 26 May 2025 (UTC)
• I should clarify that even proposals that are adjacent or in opposition to what I think are quite doable - it's community consensus after-all. Currently, I see three (most of which are disagreed to mine, some agreement on BN, but whatever):
• # Tamzin's suggestion to allow crats to determine gaming - and have them determine if they should resysop. Theoretically, we can also extend resysop periods beyond 24 hours to allow for community input - this serves similar to recall with lesser requirement of community input but crats make the final call.
• # Thryduulf's suggestion of strengthening the criteria and enshrine not gaming as part of inactivity rules.
• # 28bytes' suggestion (from BN) to simply enshrine gaming inactivity rules as part of the recall process.
• I think I am somewhat aligned with the first two. --qedk (t 愛 c) 08:44, 26 May 2025 (UTC)
• :24 hours is already just a minimum. NaomiAmethyst's resysop request, which fell into the narrow discretionary zone that already exists on resysops, took over a week, including substantial community input and a 59-hour 'crat chat. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 08:58, 26 May 2025 (UTC)
• ::I think it's a middling example of how it could be. I suspect it can be made more procedural like a crat chat, so the community aspect is more reserved. --qedk (t 愛 c) 14:33, 26 May 2025 (UTC)
• This is conceptually wrong. The whole point of things like Admin Recall is precisely to handle vague situations like gaming. The system worked fine, there's no need to change anything, other than perhaps convincing people that the system doing what it was intended to do is not some accidental slip-up. Inactivity rules are in place for a good reason. Enforcing them is a good thing, not a bad thing. The recall process worked and was "constructive" in this case. As edit count is a very weak signal, the standards are intentionally on the low side so as to avoid ensnaring innocent bystanders (e.g. still active admins but who are mostly active on other Wikimedia projects or other languages). Things like recall cover the case where the edit count is overstating their contribution, rather than understating it, because the 100 edits were all of an irrelevant "fixed a typo" kind. Or, put another way: the hard limits should be lenient. But that doesn't mean the subjective limit is lenient. This is, again, good practice, and not a mistake, but an intentional choice. SnowFire (talk) 16:33, 27 May 2025 (UTC)
• :And that's exactly why it's a bad idea. You create a rule - and then you provide a backdoor around the rule that does subsumes exactly what the rule was supposed to be for in the first place. How would we expect inactive or semi-active admins to automatically figure out what our soft limits are, is that not an utterly unrealistic expectation per se? --qedk (t 愛 c) 07:33, 28 May 2025 (UTC)
• ::No, it is not unrealistic for admins to stay in touch with community expectations. There's also no backdoor here, the hard limits are triggers for automatic deadminship, no more, and they are certainly not a target to aim for. CMD (talk) 08:53, 28 May 2025 (UTC)
• ::Re QEDK: Nothing is unrealistic here. This isn't rocket science. All of the recalls-for-inactivity so far have been the kind where the inactivity was inarguable and obvious, and that the only conceivable argument against the recalls is "I disagree inactivity should be a criterion for de-adminship" (which, well, sure, but just say that then). Any attempted recalls on more arguable cases aren't likely to start to begin with, and aren't likely to succeed if they do.
• ::Or, and feel free to correct me if I misinterpreted you: you phrase it as "create a rule." But 100 edits per 5 years isn't the operative concern; inactivity is the concern. 100 edits per 5 years is just the extreme level where the process becomes automatic because there's no argument otherwise. Everywhere else has wiggle room, which is good not bad to repeat, that means we're using human judgment rather than gameable rules. Just 200 edits but is caught up doing tech work and MediaWiki patch requests? Sure, no problem. 200 edits but they're substantial ones including admin work? No problem. 200 edits that are fixed a typo or editing their own user talk page to avoid automatic de-adminship? That's still inactive. Anyway, don't take it from me, ask someone you trust in real life, but change the domain and don't say which side you're on. e.g. is it acceptable for the aviation club to revoke admin-level access from some guy who just checks into the hangar for 2 minutes twice a year and does nothing but is over the bare minimum standard for automatic revocation. SnowFire (talk) 13:49, 28 May 2025 (UTC)

:Are two inactivity recalls too many inactivity recalls, such that the rules should be modified to reduce the number of inactivity recalls? Levivich (talk) 17:21, 27 May 2025 (UTC)

::No, but two recalls in the span of a month both related to inactivity instead of actual conduct issues is too many. --qedk (t 愛 c)

• I've seen only a few situations where it became a problem; in both cases it was incompetence. Probably a combination of the lower bar from "got in back when it was easy" combined with inactivity. Ironically, these both because problems when they did something. Ironically, those with zero activity do no harm. On one the incompetent one was given the bit back out of courtesy with just a simple request. How about raise the bar a bit and when and when an inactive admin runs into it the need to make a self-assessment statement that they feel that are current enough to do do the job properly in which case it gets waived? North8000 (talk) 18:24, 27 May 2025 (UTC)
• I propose the opposite: add a mention to WP:INACTIVITY that these standards exist because we expect inactive admins to hand in their tools, and that doing the bare minimum to keep admin tools may be seen as gaming the system to subvert these expectations. Thebiguglyalien (talk) 🛸 19:20, 27 May 2025 (UTC)
• :I think if the community wants inactive admins to relinquish administrative privileges, then the most straightforward approach is to just remove the privileges, rather than say the community might decide to remove the privileges at some point. isaacl (talk) 00:29, 28 May 2025 (UTC)
• ::That's what WP:INACTIVITY is. The problem is that we have people cheating the system so they don't get the removal when they're supposed to. The solution is to say that they still get the removal if it's determined that they're cheating the system. Thebiguglyalien (talk) 🛸 00:45, 28 May 2025 (UTC)
• :::You're suggesting to add a mention that inactive admins are expected to relinquish administrative privileges, without defining the level of activity when that expectation starts. I'm saying the community should discuss if its expectations on when administrative privileges should be removed matches the current removal procedure, and if not, then change the criteria so they do. There are reasonable security concerns for doing so, and I don't feel it's a good idea to set up a sword of Damocles. If the community wants to make removal (and restoration when activity resumes) a normal occurrence, with no fault attached, it should just do it, and skip having a discussion about whether or not a particular case should be considered insincere. isaacl (talk) 01:05, 28 May 2025 (UTC)
• ::::Exactly. --qedk (t 愛 c) 07:33, 28 May 2025 (UTC)
• ::::We have done this. If an admin comes back within 3 years of their last edit they can simply request restoration with no fault attached. CMD (talk) 08:56, 28 May 2025 (UTC)
• :::::Unless they get recalled you mean, which is also exactly the problem we're talking about. Nothing simple about it. --qedk (t 愛 c) 10:30, 28 May 2025 (UTC)
• ::::::It sounds like the problem you refer to is that consensus can change, in which case I suppose there is no way around it and nothing simple as that's a building block of Wikipedia. However, it is not particularly or uniquely relevant to the point at hand, and I assume that on average admins are more aware of WP:CCC than most editors. CMD (talk) 11:13, 28 May 2025 (UTC)
• :::::::That is an absurd reduction of what I'm trying to say. What we're doing is basically drawing a line in the sand, pointing admins that way and then saying "welp! there's actually an invisible line you didn't know about (because we didn't tell you), sorry!" and showing them out the door. --qedk (t 愛 c) 11:40, 28 May 2025 (UTC)
• ::::::::This is exactly how edit warring works. You can comply with 3RR or even 1RR and still be blocked for edit warring. No one is promising that editors who comply with 3RR won't be blocked for edit warring. The rule isn't "don't cross 3RR," the rule is "don't edit war." Similarly, you can comply with inactivity requirements and still be recalled for inactivity. The rule is "be active." Also, unlike edit warring blocks, a desysop is not showing an admin out the door. Levivich (talk) 11:58, 28 May 2025 (UTC)
• :::::::::Indeed it is not a particularly unique issue, and admins are expected to be familiar with this concept, which is covered by our guidelines. CMD (talk) 14:04, 28 May 2025 (UTC)
• :::::::::This is not "exactly how edit warring works" - editwarring, 3RR and 1RR are all very explicit that editwarring is not (just) a bright line issue. Admin inactivity, prior to the recent recalls, has been an explicitly bright line issue. When admins have become out of touch through inactivity they have been brought before ANI/Arbcom/Recall for the disruption their being out of touch has caused, not for being inactive. Thryduulf (talk) 16:51, 28 May 2025 (UTC)
• :::::We have (mostly) not had a discussion yet about whether or not the community's expectations on inactivty matches the current criteria. (The discussion has been mostly about whether we should have a discussion.) We are not encouraging admins to request the restoration of administrative privileges by having a recall discussion using terms that bring their sincerity into doubt when removing privileges. If there is a community consensus for a much higher level of ongoing activity, then let's put it into effect without any recall discussion to trigger removal. isaacl (talk) 17:22, 28 May 2025 (UTC)
• {{tq|What we're doing is basically drawing a line in the sand, pointing admins that way and then saying "welp! there's actually an invisible line you didn't know about (because we didn't tell you), sorry!" and showing them out the door}}: Yes, we are, and that's a feature, not a bug. The entire point of the inactivity requirements is that we shouldn't need to tell admins about "invisible lines" (read: community norms). They should already be familiar with those lines, because they're active in the community, and if they're not, they should turn in their adminship. If someone is gaming the deliberately low bar to keep their admin status, that feels like a perfect use case for recall. I don't see what problem any of this is trying to solve. Writ Keeper ⚇♔ 14:22, 28 May 2025 (UTC)
• :{{tq|The entire point of the inactivity requirements is that we shouldn't need to tell admins about "invisible lines" (read: community norms). They should already be familiar with those lines, because they're active in the community, and if they're not, they should turn in their adminship.}} I've been trying to think of how to phrase my thoughts on this, and Writ Keeper nailed it. ScottishFinnishRadish (talk) 14:46, 28 May 2025 (UTC)
• ::Except we're changing the rules on them without telling them. We've spent the past quite a few years now telling admins that "being active in the community" means making at least N edits in (period of time). Now we are saying that's not true and "being active in the community" means doing that and a bunch of nebulous other things that are undefined and unknowable but if you don't do them someone will recall you. If you happen not to be paying attention to Wikipedia during a particular 24-hour period (if you're lucky) you could be out on your ear before you even know anybody is upset at you not doing something you didn't know you were meant to be doing. Thryduulf (talk) 15:26, 28 May 2025 (UTC)
• :::Is that what has happened? Or have admins that have clearly gamed the activity requirements been recalled? WP:GAME has been a guideline long enough that it can legally vote in a couple months, so I don't think the community enforcing it is sudden or unexpected. The only difference is now the community has a path to address activity gaming by admins. ScottishFinnishRadish (talk) 15:43, 28 May 2025 (UTC)
• :::Also, what we actually tell admins, in {{tl|inactive admin}} and {{tl|inactive admin 2}}, is: {{tq|Inactive administrators are encouraged to rejoin the project in earnest rather than to make token edits to avoid loss of administrative permissions.}} So even if we accept your silly premise that admins must be spoon-fed their standards and expectations like a five-year-old, we are already doing that in this case. Writ Keeper ⚇♔ 15:46, 28 May 2025 (UTC)
• ::::It's not a {{tpq|silly premise}} at all and I absolutely reject the bad faith associated with your comment, both towards me and towards those admins who are maintaining activity according the explicit standards we set for them. If we want to change those standards, that's fine, but what is not fine is pretending that we are doing something other than changing the standards and holding people to those standards before they have been changed. Thryduulf (talk) 16:04, 28 May 2025 (UTC)
• :::::Nah, I don't think you're in bad faith, I just think you're being silly. We at least nominally elect admins based on their cluefulness, which implies--perhaps even is defined by--the ability to read and apply these "unwritten rules" of community norms. If admins needed everything spelled out for them, then we could just replace RfA with a basic reading comprehension test. After all, it's not like the concept of "the written rule is a deliberately low bar that doesn't define the minimum acceptable behavior" is a foreign concept on Wikipedia; several others have already pointed out 3RR as an example. If we can't expect admins to pick up on both the community's feelings and the strong hints in the notification templates about activity, why do we expect non-admins to grasp the nuances of edit-warring beyond the bright-line 3RR? Writ Keeper ⚇♔ 16:19, 28 May 2025 (UTC)
• ::::::Those unwritten rules would be fine in theory if they didn't directly contradict the written rules. 3RR and edit warring are both explicit that edit warring isn't (just) a bright line issue. The activity reminders say that you should be engaged with the community - which is what the written rules say meeting the activity requirements is. Thryduulf (talk) 16:37, 28 May 2025 (UTC)
• :::::::It doesn't, though. WP:INACTIVITY defines the threshold at which one is subject to procedural desysopping due to inactivity. It doesn't define what the meaning of "engaged with the community" vs. "inactive" is. (Nor should it, because such a definition would never be agreed upon by everyone--which is why we now have a community process in admin recall to handle it on a case-by-case basis, rather than a semi-automated process to decide it). Nowhere in there does it say "an active admin is considered to be..." or "an inactive admin is one that hasn't...". Which, again, shouldn't matter regardless, because we expect admins to be able to grasp the spirit of a rule without being slavishly devoted to the letter of a rule. Like, there's also no rule that explicitly says I shouldn't blank Jimbo's user page and replace it with a wikilink to Bomis Babe--he outright welcomes people to edit his page, and it's not a non-sequitur or anything. But I don't, because "reading the room" is a skill we expect admins to possess, and that applies here just as much as anywhere else. Writ Keeper ⚇♔ 16:57, 28 May 2025 (UTC)
• :(de-indent, re Thryduulf above) What would it take to convince you that, even if you disagree as a matter of preference, that there is no "contradiction" here? Inactivity has always been a matter of human judgment. You can argue that it shouldn't be, but right now it is, and that's what the community wants. We want a system that can be easy on an admin with a low edit count but who is obviously active, and be harsh on an admin who made 1,000 edits to user page essays or something in the past 5 years and nothing else. If you disagree that's fine, but that's different from claiming that the standard was unclear. It's not unclear, it's just that the standard isn't 100% down to raw edit count. The exact same issue comes up with gaming extended confirmed status via pointless edits. It's not controversial that not all 500 edits are created alike, and the solution to accounts gaming the extended-confirmed requirements is not to increase the threshold to 1000 edits, but to have Actual Humans look at those 500 edits and say "hey, 497 of these were to a Sandbox page." SnowFire (talk) 17:08, 28 May 2025 (UTC)
• Would it be more productive to skip all this discussion and just RFC a slightly higher inactivity requirement? –Novem Linguae (talk) 16:30, 28 May 2025 (UTC)
• :Word. Best Alexandermcnabb (talk) 16:35, 28 May 2025 (UTC)
• ::Nope, because then we'll be right back here arguing about _those_ requirements. :( --SarekOfVulcan (talk) 16:36, 28 May 2025 (UTC)
• :Yes. @SarekOfVulcan debating what the requirements should be is a lot more productive than trying to argue that we should be upholding requirements in excess of the ones that have consensus because some people think the ones with consensus aren't strict enough. Thryduulf (talk) 16:38, 28 May 2025 (UTC)
• ::That strawman is about beaten to chaff at this point. This issue is WP:PGAME, not the requirements themselves. ScottishFinnishRadish (talk) 16:45, 28 May 2025 (UTC)
• :::Firstly it's not a straw man, and I'm far from the only person who disagrees with you about whether following the rules as written is or is not gaming those rules. Please stop trying to characterise it otherwise. Thryduulf (talk) 16:48, 28 May 2025 (UTC)
• ::::"because some people think the ones with consensus aren't strict enough." is a strawman though. People aren't starting recalls for every admin who just exceeded those minimum requirements, only for those who actually game things by e.g. making false promises which result in a positive crat chat. We have a panic among some people after nearly every recall procedure it seems, but so far none of the succesful recalls have been shown to have been done incorrectly (recalling an admin who still had the support of the community, or a recall done by some coordinated cabal). Neither the inactivity requirements nor the recall requirements need adjusting, or at least for neither has such a need been demonstrated. Fram (talk) 17:19, 28 May 2025 (UTC)
• ::::To Fram's point, while you've made it quite clear that you think the inactivity requirements can't be meaningfully gamed, here are 25 to 50 users that disagree with you. Writ Keeper ⚇♔ 17:23, 28 May 2025 (UTC)