Hi! Repeatedly in discussions people raise the issue that the amount of permastubs about villages due to NGEO is not being handled in the best way possible. While WP is not a paper wikipedia, there are various concerns that having a large minority of articles be villages in third-world countries where our own sourcing bias (see Geographical bias on Wikipedia) prevents us from improving many of these articles or preventing misinformation due to very high ratios of articles to geo content patrollers (the Iranian Well issue comes to mind, for example, and I expect there to be other less absurdly flagrant content mistakes). On the other hand, there is no policy against the existence of stubs even in a permanent state. Additionally, I think we can all agree that having information on developing world villages is beneficial both to our readers and companies that use Wikipedia for their services. With the increased success of Wikidata (as much as we might not admit that very often here), this "loss of information" fear from altering how we manage these stubs seems less reasonable (but still valid). I wonder as well if redirecting impacts the accessibility of this information.

With all this in mind, I was wondering if y'all could help by giving your thoughts on the issue, including past discussions I am not aware of or ideas on how to fix the problem. A. C. Santacruz ⁂ Please ping me! 11:23, 23 March 2022 (UTC)[reply]

• If you don't think there is a problem, you're opinion is welcome too! I'm trying to gather a wide range of opinions and ideas on the subject, as I think there's a lot we can do to improve these articles. Sadly these discussions tend to happen in highly contentious AfDs or partisan venues so we tend to lose a lot of the nuance I hope can surface in this discussion. BTW I don't think there's a problem with NGEO. Perhaps it could be highly different opinions on how WP:NOPAGE applies to geo stubs? Not sure. A. C. Santacruz ⁂ Please ping me! 11:40, 23 March 2022 (UTC)[reply]
  • The median length of a page view is maybe long enough to read one sentence. This suggests that very brief articles are not a problem for at least half of our readers ("Oh, the Iranian Well is in Iran. That's all I needed to know. *click*"). WhatamIdoing (talk) 19:00, 23 March 2022 (UTC)[reply]
    • I wish there were stats on specifically the average length of a pageview on articles about locations - that would back up this argument more strongly. However, I do agree with you on this. casualdejekyll 19:20, 23 March 2022 (UTC)[reply]
• I don't see a specific question to comment on. North8000 (talk) 11:49, 23 March 2022 (UTC)[reply]
  • North8000 it's more of a vague brainstorm I guess? Perhaps the question would be "Are there issues with having so many geo permastubs on wiki?" with a follow-up question "If so, how can we fix them?" A. C. Santacruz ⁂ Please ping me! 11:57, 23 March 2022 (UTC)[reply]
    • We could redirect the stubs to lists, either a stand-alone list article ("Villages of Foo") or a list within the parent entity article ("Foo#Villages"), giving the same information as the stub. But that does not improve the quality of the information and may slightly discourage addition of content. Aymatth2 (talk) 13:10, 23 March 2022 (UTC)[reply]
• The problem isn't information loss. There isn't valid information to be lost. It's about huge amounts of outright information error. Merging erroneous information doesn't correct or remove it, and we are still, this year, this month, this week, finding and researching and dealing with articles where the database dumpers gave us boilerplate "X is an unincorporated community" for things that were never communities. This is in the United States, where the aforementioned sourcing bias would have one believe that sourcing is strong. Some of the recent highlights have been "communities" that were actually toxic material handling sites in Utah that were deliberately sited many miles away from populated areas; "communities" that the source used outright said were railway stops and maintenance camps; "communities" that in fact were mines; and a "community" based upon solely the information in GNIS record #1742680 (see https://www.topoquest.com/place-detail.php?id=1742680). I rewrote Bullfrog, Utah (AfD discussion) roughly 3 weeks ago. Bob, West Virginia came up yesterday.

  If there's a problem with villages that cannot be expanded upon because there isn't much to say, which I am not sure there is, it is positively swamped by the massive problem of geographic articles that can never be expanded because they are outright falsehoods. Are there issues? Yes. This is the issue. Still.

  Uncle G (talk) 20:44, 23 March 2022 (UTC)[reply]

  • If the issue is that the article fails WP:V and thereby produces falsehoods, then that isn't an issue with the current policy. WP:DEL-REASON#7 is pretty clear that articles that fail WP:V can be deleted. The very issue you're describing is one of cleanup owing to sloppy uses of unreliable database sources by Wikipedians over many years who were largely acting in good faith.

    The purpose of Wikipedia is, in short, to provide a free, neutrally written, accessible encyclopedia to the world. Removing unverifiable information is part of that. But arguing that no good information would be lost by nuking all geostubs seems naïve; surely it's the case that we're going to wind up deleting valid entries. Much like copyright cleanup investigations, undoing these sorts of mistakes is going to be hard, labor-intensive work. Perhaps a WikiProject, a noticeboard, or a task force of WikiProject Cities could be created to patrol weakly sourced geostubs and to add reliable sources to them. Gamification akin to NPP/AFC drives could also serve as a model to get people invested in cleanup, with barnstars being awarded to participants based upon their actions. We might need a script to do this sort of "geostub patrolling", but it doesn't seem like this is going to be too technically challenging to get started on. — Mhawk10 (talk) 05:22, 24 March 2022 (UTC)[reply]

    • The issue is a huge amount of false information. Systematic false information. A. C. Santacruz's question was "Are there issues with having so many geo permastubs on wiki?", and this is the answer. There are, and it's this. It wasn't a question about policy, or about whether things like whether Wikipedia:WikiProject California/GNIS cleanup task force exist. It was a request for what the issues are, and this is the answer. And it's going to be the answer for a long time. Given the number of long-term participants in GNIS cleanup, versus the number of "unincorporated community" cop-out articles, there's decades of work here. Uncle G (talk) 12:17, 25 March 2022 (UTC)[reply]
  • @Uncle G, is your concern less about whether the place existed (e.g., a toxic waste area, a mine, a railway stop), but instead that you think it false to call these places "communities"? WhatamIdoing (talk) 01:40, 25 March 2022 (UTC)[reply]
    • It's a concern shared by a lot of people, and it's also about whether the place existed, because we've come across many that actually did not, such as the survey corners in Virginia. Independence, Uintah County, Utah (AfD discussion) is an even more extreme example, as not only did the place never exist we even have a history book saying that it never existed. Wikipedia has been proclaiming it an "unincorporated community" for over 4 years.

      "unincorporated community" is an information-free cop-out. Because the GNIS never made distinctions amongst "populated place"s, the database dumpers gave us everything as "unincorporated community". Some even went so far to give us "unincorporated community" for things that the GNIS did not call "populated place", such as "Bone Lick Post Office (historical)" which a database dumper gave to us as an "unincorporated community" of Bone Lick, West Virginia (AfD discussion) 10 years ago, and which Wikipedia has been proclaiming ever since. The consequence of all this is that "unincorporated community" is completely debased now. Every "unincorporated community" could be a database dumper giving us post offices, mines, springs, railway junctions, survey corners, or 3 roads near Salem as "unincorporated community".

      And, worse, people see all this and copy it, thinking that that's normal, even to the extent of countries that don't have the GNIS to mislead them. So Bridgend, Perth and Kinross becomes an "area" instead of the village and burgh of barony that it actually is, for example. Columbia, Tyne and Wear was just "in Washington" for 16 years, without any clue as to what it even was (Shoe shop? Statue? Roundabout?), to the extent that one editor couldn't find it at all.

      We have hundreds of thousands of stub articles that systematically do not give correct factual information about what their subject is, or indeed was, that do not give proper context enabling expansion by editors, and that egregiously mislead readers in many cases. And this is going to be the case for many years.

      Uncle G (talk) 12:17, 25 March 2022 (UTC)[reply]

      There is a perennial problem of some early user fundamentally misunderstanding a particular concept, which then gets propagated out as other users then assume that it's correct - often well beyond English Wikipedia.

      One example that springs to mind is Unitary authorities of England, which is an article that uses the term "unitary authority" to both refer to a kind of local authority (which is correct), as well as the area which it controls (which isn't, but is a common shorthand for "unitary authority area"). This is a bit like using "council" every time you mean "county", and makes for extremely confusing reading. Worse, this has been wrongly repeated on numerous other Wikis - something that I only became aware of due to stumbling across a totally incoherent Wikidata item that was trying to represent both concepts at once. Theknightwho (talk) 16:31, 25 March 2022 (UTC)[reply]

      Looking at the Bone Lick example, how do you know that a US post office was placed in a location that wasn't – according to one or more of the reasonably typical definitions you would find in a dictionary – a community? Rural communities are real things. They might not be the kind of real thing that needs a separate Wikipedia article, but I don't think we should treat post offices and communities as exclusive entities. It seems more likely to me that if a place had enough people there to get a post office, there probably was a community there. WhatamIdoing (talk) 20:48, 25 March 2022 (UTC)[reply]

      It depends on your definition of community, but I think some collection of dwellings is required to designate a populated place. Post offices were often sited in isolated stores that were not near a populated place, as at Dudley Farm, or even in a farmhouse. Without some source telling us that the post office was sited in a populated place, I don't think we can assume notability. "Railroad stations" have a similar problem, as they were often just a shed or an open platform on a siding with no other buildings. Donald Albury 22:59, 25 March 2022 (UTC)[reply]

      Some railway stations were a fair distance from the community that they were named after. This often happened in sparely-populated areas such as northern Scotland, but it happened in England also - for instance Micheldever railway station, two and a half miles away from Micheldever. After the station opened, a new community grew up around the station - named Micheldever Station. --Redrose64 🌹 (talk) 11:44, 26 March 2022 (UTC)[reply]

      That reminds me of a (possibly apocryphal) story that I heard. Someone asked why Tring station was so far from the town, and the answer was that it made more sense to put it on the railway line. Phil Bridger (talk) 13:39, 26 March 2022 (UTC)[reply]

      That's also told of one of the stations on the Settle and Carlisle line, possibly Dent or Garsdale. But in the case of Micheldever, the village is half a mile from the railway. --Redrose64 🌹 (talk) 13:50, 26 March 2022 (UTC)[reply]

      I was thinking of names that appeared on maps along rail-lines that were freight-loading points, and served local farmers rather than any populated place, of which there were many in the US before motor transports and improved roads came along. Donald Albury 13:43, 26 March 2022 (UTC)[reply]

      Right. In much of the US's "breadbasket", the post office would have been near the Co-op grain elevator, and the grain elevator was always next to the rail line. This might or might not be near an officially recognized town. You could still have "a community" (e.g., "the people with common interests living in a particular area") in that location even if you didn't have the kind of "homes placed close by each other" appearance that a city dweller would expect. WhatamIdoing (talk) 21:17, 28 March 2022 (UTC)[reply]

      No grain elevators around here. I have a list of 37 post offices that once existed in this county that I cannot verify were part of a community. I can find many of them on old maps, and a few even have GNIS entries. What I do not have is any other source for any community that may have been associated with a post office. Even if there was a community, it may not have been known by the same name as the post office. The U.S. Post Office would not allow more than one post office with a given name in a state. Many times someone would petition for a post office, only to have their preferred name rejected because another post office using that name already existed in the state. There are even a couple of post offices on that list that changed their names during the few years they existed, for which I haven't found anything about a community under either of the names. I have another 22 places (not P.O.s) that are/were listed in GNIS as populated places, some of which show on old maps, but for which I haven't found anything else. Without other sources, I think it would be wrong to assume that those places are notable enough to have articles in Wikipedia. Donald Albury 22:09, 28 March 2022 (UTC)[reply]

• I don't see it as a significant problem, although it would be good to raise the bar a bit. A 1/2 million geo articles is fine, 5 million wouldn't be. Geographic places are highly encyclopedic topics and also given extra emphasis in the Five Pillars and informally giving some allowance for that per Wikipedia:How Wikipedia notability works is IMO fine. I didn't separate these in my study User:North8000/Display but of the 27% of all articles being "geographic places,broadly construed" there were probably as many less notable things like tiny train stations and tiny bridges as there were small villages.North8000 (talk) 20:52, 23 March 2022 (UTC)[reply]
  • Considering that most of our geostubs cover a limited number of regions, I think this suggests that there is an issue; if we created the same density of articles for Asia, Africa, and South America as we do for the United States, we would have at least those five million geo-articles, most of them similar to the current average geo-article - an under sourced microstub. BilledMammal (talk) 04:48, 27 March 2022 (UTC)[reply]

Geostubs break

• Personally, I don't think we should have articles on places that are sourced only to a single database or list. If nobody has been able to find anything other than a database or list entry for a place, then, in my opinion, it has not been demonstrated to be notable enough to be included in Wikipedia. I have a lot of names on a subpage (User:Donald Albury/Notes/Alachua County communities) under my account of place names in the county I live in, that appear in a list of post offices, or on old maps, or in the GNIS, or mentioned in some otherwise unrelated source, for which I have not yet been able to find any other usable sources, and for which I therefore will not create articles. - Donald Albury 23:35, 24 March 2022 (UTC)[reply]
  • If nobody has been able to find: One of the things that we don't have is a way to say "I tried to find sources, and I didn't succeed." WhatamIdoing (talk) 01:42, 25 March 2022 (UTC)[reply]
    • I've been saying that quite easily at AFD for many years. The best thing that one can do is double-check and look; and if the result is negative that's what one gets to say. And it's demonstrably easy to say it on worklists like User:Hog Farm/Missouri attention needed. Uncle G (talk) 12:17, 25 March 2022 (UTC)[reply]
    • You, WhatamIdoing, just said it, so where is the problem? A statement like that can just as easily be made on an article talk page or in an AfD discussion or wherever it needs to be said. The real issue seems to be that there is no template for saying it and that it is not part of an automated procedure, which seem to be the only ways in which many editors communicate. Phil Bridger (talk) 18:43, 25 March 2022 (UTC)[reply]

      That's it. There's no structured workflow for that. So you see an article whose current version is below average, but you have no insight into whether anyone else looked at it. This starts with the Special:NewPagesFeed, which is set up so that dozens of NPPers can silently duplicate each others' work, until someone finally clicks a button to mark it as patrolled. (Oh, look, some IP caused Tasmania – yes, the island/state – to appear in the NPP queue. It's now the "oldest article" in the queue, despite having actually been in the queue for 29 minutes as of now.)

      But the problem extends beyond that: Nobody knows how many NPPers looked at an article in the queue and decided that it wasn't worthy of deletion, but for years afterwards, we have no way to know whether anyone tried to find sources, whether anyone tried to expand an article, etc. When Template:Orphan was a thing, we tried to reduce unproductive duplication of effort by marking them as "attempted" with |att=. I don't remember seeing any similar system for any other maintenance tag. WhatamIdoing (talk) 20:38, 25 March 2022 (UTC)[reply]

      I would have thought that anyone remotely competent in editing an English language encyclopedia, and especially anyone qualified to do new page "patrol" (that's a horrible word that should be replaced by "review"), would be able to put together a sentence in English without there being a template or automated procedure to do it for them. Phil Bridger (talk) 20:56, 25 March 2022 (UTC)[reply]

      Unfortunately, "is able to" and "will actually" are two separate considerations. And even if you do, who says that all (or even any) of the subsequent reviewers will notice that you put something on the talk page? WhatamIdoing (talk) 21:20, 28 March 2022 (UTC)[reply]

  Agree there should be some way to clear stubs like "Imaginaryville is a place in California" with no other useful content. Such stubs offer very little information to the reader. 2601:647:5800:1A1F:B45D:61AC:E323:660D (talk) 18:33, 7 April 2022 (UTC)[reply]

• Could we say that all geostubs that meet certain criteria should be flagged for speedy deletion by a bot? Criteria could include some combination of
  • Has not been edited for two years
  • Has only one sentence
  • Does not include coordinates
  • Has no source, or only one database-type source

A fair number of valid entries would be removed, but very little information would be lost. Aymatth2 (talk) 14:30, 27 March 2022 (UTC)[reply]

Deletion without manual review is basically never accepted. Also, how many articles do you think would be in such a list? I took at look at half a dozen articles about unincorporated communities (all in the US) just now, and all of them had been edited in the last 9 months. WhatamIdoing (talk) 21:24, 28 March 2022 (UTC)[reply]

I suppose the gnomish edits tend to obscure real activity. So how about we dropped the "last edit" criterion and PROD the articles? The idea is to clear away stubs that say only "Foo is a place in Finland". Don't know how many of those there are. Aymatth2 (talk) 12:04, 29 March 2022 (UTC)[reply]

Perhaps add "Is more than five year old". Aymatth2 (talk) 13:27, 29 March 2022 (UTC)[reply]

A more general concern, I don't think these articles should be mass-flagged for deletions by a bot. It would be better to create some type of speedy deletion criteria. 2601:647:5800:1A1F:B45D:61AC:E323:660D (talk) 18:35, 7 April 2022 (UTC)[reply]

• Frankly, this seems like a solution in search of a problem. There are certainly a lot of shitty geostubs created from databases without a lot of sources. Like, say, this one. Wow, what a pile of crap! What's going on with this stupid awful geostub now? Oh -- it's a FA. Wowzies. I've written 20 GAs from such geostubs. Thank God there wasn't an army of people running around the project CSDing all of them in the meantime. jp×g 07:03, 11 April 2022 (UTC)[reply]

  I definitely agree with you that a geostub CSD idea (or mass deletion) is terrible, JPxG. Geostubs range from Iranian wells to future FAs and the idea of removing hundreds of thousands of articles because they're short is completely contrary to wiki PAGs. I don't know what you mean by "this seems like a solution in search of a problem", though, as I started this thread just to get the views from many people on geostubs. Or are you responding to Donald? A. C. Santacruz ⁂ Please ping me! 08:42, 11 April 2022 (UTC)[reply]

Proposal

Thank you all for your comments, they have been very instructive. I'll work on a proposal for the next few days to see if there is consensus to change some guidelines to deal with some of the problems identified in this thread. A. C. Santacruz ⁂ Please ping me! 12:24, 29 March 2022 (UTC)[reply]

What are the different types of geostubs, in terms of quantity and type of information? Has this ever been categorized formally? A. C. Santacruz ⁂ Please ping me! 09:22, 30 March 2022 (UTC)[reply]

Geostubs will all be indirectly in Category:Geography stubs. This PetScan query lists all articles in this category to a depth of 5, which probably is not deep enough, giving 682,896 results. It runs very slowly. Probably better to start at the country level, e.g. Category:France geography stubs, A query for France, 5 deep, at https://petscan.wmflabs.org/?psid=21772838, sorted by size gives 40,889 results, starting with the 260-byte Clément, French Guiana, which informs us that Clément is a town in French Guiana, and gives coordinates. OpenStreetMap has never heard of it and Google shows nothing but forest at the coordinates. Aymatth2 (talk) 13:33, 30 March 2022 (UTC)[reply]

Relevant discussion at

→ Wikipedia:Village_pump_(policy)#RfC_on_the_addition_of_a_stand-alone_page_creation_criteria_to_the_geography_notability_guideline

Started proposal based on some of the comments here, thanks everyone! I might propose some other stuff at a later point to improve collaboration in the geo topic area but we'll see how this one goes. A. C. Santacruz ⁂ Please ping me! 11:06, 13 April 2022 (UTC)[reply]

Aguaxima, a plant growing in Brazil and on the islands of South America. This is all that we are told about it; and I would like to know for whom such descriptions are made. It cannot be for the natives of the countries concerned, who are likely to know more about the aguaxima than is contained in this description, and who do not need to learn that the aguaxima grows in their country. It is as if you said to a Frenchman that the pear tree is a tree that grows in France, in Germany, etc . It is not meant for us either, for what do we care that there is a tree in Brazil named aguaxima, if all we know about it is its name? What is the point of giving the name? It leaves the ignorant just as they were and teaches the rest of us nothing. If all the same I mention this plant here, along with several others that are described just as poorly, then it is out of consideration for certain readers who prefer to find nothing in a dictionary article or even to find something stupid than to find no article at all.

— Denis Diderot, Encyclopédie (1751)

My study (which classified 277 of 1,000 articles as "geographic, broadly construed") was not that granular, but I did pick up some impressions from reviewing those 270. Also from my New Page Patrol work:

• One glaring stub example was non-notable train stations. For example, and editor takes a train line and makes an article for each little train station on the line.
• The US geo articles tended to less stubby. Presumably because they are more familiar to English editors and also with ready access to sources in English.
• Probably the most geostub articles on ones that look borderline on notability are from India. Probably because it's a huge country with lots of places and lots of potential editors most of which know some English. The borderline ones there tend to be very small populations, small and abstract units of government, and train stations.

Also, because a deletion at AFD requires "proving a negative" that suitable sources don't exist, and an English speaking editor is typically unable to do that for an article when it's in a country where the primary media are Non-english and so borderline ones tend to stay....not just due to AFD results, but also due to editors never even taking them to AFD for that reason.

Again, I don't consider the maybe 200,000 or 500,000 current geostub articles to be a big problem. Geography is a very enclyclopedic topic. The big problem is that that could easily jump to 5,000,000.

(BTW, just to make sure my big "277" doesn't mislead, "broadly construed" included edge cases like facilities and planets.)

Sincerely, North8000 (talk) 13:45, 30 March 2022 (UTC)[reply]

For Category:United States geography stubs PetScan gives 130,527 results at 5 deep, 132,059 results at 7 deep, probably deep enough to get almost all of them, starting with the 261-byte Makaokahaʻi Point, a jutting headland on the south coast of the island of Kauai in the Hawaiian Islands. I do not know an efficient way to separate human geography stubs like villages, railway stations etc. from physical stubs like streams, hills etc. The United States has 4.25% of the total world population and 6.1% of world landmass. Perhaps it has 5% of world geography, giving 2,610,540 potential geostubs worldwide at the same level of coverage. But see Rivers of Lake County, California: there is potential to greatly increase the number of United States geostubs. We could be looking at over 20 million worldwide. Aymatth2 (talk) 14:46, 30 March 2022 (UTC)[reply]

Aymatth2 and North8000 I've done some initial research into the number of geostubs in proportion to geo articles for various countries at User:A. C. Santacruz/Research/Geostubs. The picture right now does not look very good... I'm not entirely sure how one would find the ratio of stubs per editor actively patrolling stubs for accuracy but I imagine it is unmanageably high based off this first impression. A. C. Santacruz ⁂ Please ping me! 14:34, 31 March 2022 (UTC)[reply]

@A. C. Santacruz: I thought the ratio of stubs to real articles was much worse than you have found, so in a sense this is encouraging. Short geo articles do not bother me if they are accurate and give some useful information. But I suspect that in some areas there are no editors checking the stubs at all. See Clément, French Guiana. Does this place exist? An editor could waste a fair amount of time before finding that the French government Géoportail does show something called Clément at those coordinates. They leave the stub, even though it is useless. Then a few years later another editor repeats the check... Aymatth2 (talk) 15:57, 31 March 2022 (UTC)[reply]

Yeah, the articles so far are from 3 very urbanized countries. I'll add India and Iran later, I expect those to be much worse. A. C. Santacruz ⁂ Please ping me! 16:10, 31 March 2022 (UTC)[reply]

North8000 that study sounds interesting, where can I read it? In terms of quantity of information I was wondering like, if someone had ever gotten some data on like, the distribution of article sizes within geostubs and the such. A. C. Santacruz ⁂ Please ping me! 14:42, 30 March 2022 (UTC)[reply]

@A. C. Santacruz: I have all of the data up at User:North8000/Display and am starting to add summaries etc. Sincerely, North8000 (talk) 14:49, 30 March 2022 (UTC)[reply]

Challenging closures already states that a close will rarely be changed by either the closing editor or a closure review if the complaint is that the closer is not an admin. Given this, requiring non-admins to specify that they are a non-admin when closing discussions is not helpful, and contributes to the perception of adminship being a seen as a big deal. I note also that many of the best closers are not admins, such as Paine Ellsworth for RM's, and Mhawk10 for RFC's.

The possible exception to this is WP:AN and WP:ANI, as given the nature of the board whether an admin closed a discussion may be relevant.

Similarly, I believe it may be beneficial to allow non-admins to close AFD's as "delete", with a new CSD criteria to support that. This would be similar to the current process for RM's, which allows non-admins to close RM's even if they are unable to implement the move themselves, and would not be a significant change from the current situation where non-admins can close discussions as redirect. BilledMammal (talk) 06:02, 3 April 2022 (UTC)[reply]

As far as I know, there is no formal requirement, and editors close discussions all the time without specifying if they are an administrator. Regarding closing deletion discussions as delete, it's a perennial question where the consensus to-date has been that it duplicates effort that the admin implementing the close would have to do in order to take responsibility for their actions. isaacl (talk) 06:27, 3 April 2022 (UTC)[reply]

There is a requirement to mark RM's that are closed by non admins, but you are right - it seems that is the only area where it is required. I didn't realize the deletion discussion was a perennial discussion (it's not listed at perennial proposals), but it seems that objection could be addressed by making the closure responsible for the deletion, with the admin who responds to the CSD only being engaged in non-controversial maintenance tasks. BilledMammal (talk) 07:32, 3 April 2022 (UTC)[reply]

Is there some XFD backlog that is beyond the capabilities of our current admins? --Redrose64 🌹 (talk) 07:39, 3 April 2022 (UTC)[reply]

It's not really a matter of questioning the capabilities of admins. It's just experienced non-admins who want to help whether or not there's a backlog, or it's about editors who want to gain more experience with an eye on becoming an admin someday. And it's about whether or not it's necesessary to declare oneself a non-admin when that's not really supposed to be much of an issue. P.I. Ellsworth - ed. ^put'r there 08:42, 3 April 2022 (UTC)[reply]

To-date, the response to that has been that administrators are the ones who face consequences for their use of administrative privileges. Also frequently proposed is a user group with permissions to delete pages who could close deletion discussions. So far, community consensus is that the abilities to delete, view deleted pages, and block users form a core set of privileges needed to make decisions involving these actions. isaacl (talk) 07:52, 3 April 2022 (UTC)[reply]

(edit conflict) Yes, that should be made clear. It is the closer who should be accountable for actions that result from the closure. My question, though never asked until now, has always been about the statement at Wikipedia:Non-admin closure that goes, For practical purposes, non-administrators should not take formal action in discussions whose outcome would require the use of administrator tools... – always wondered why? In all these years there has never been an instance where an admin has not helped me with edits, closures and with explanations when they were needed. There are always ways for non-admins to use the tools by getting admins to actually make the edits. That's an excellent way to learn the ropes here. While it is true that admins are always held responsible for their tool usage, it is ultimately the closer who is responsible for their closures and implementations, whether or not they actually have the tools. Thank you, BilledMammal, for bringing this up, because I doubt if I'm the only editor who would agree with this idea yet who wouldn't mention it until and unless someone else does. Guess it's the Wikignome in me. P.I. Ellsworth - ed. ^put'r there 08:01, 3 April 2022 (UTC)[reply]

First off, thank you for the ping and the compliment; I'm flattered by your description of me.

Getting to the substance: there's an RfC that happened around 2013 which ended with the conclusion that the sole reason for overturning a request for comment can never be summarily overturned on the basis that the closer was not an administrator. There's not obviously any WP:IAR exception (if a user wrote X, and X is without substantial flaw, then a then it doesn't matter whether or not that user has a mop). When there's a substantial flaw in the closing summary, the only thing in my mind that matters is the substance of the writing, not the technical privileges that the user has. Even a panel of three of the most experienced administrators on Wikipedia had a close of theirs overturned after non-admin S Marshall succesfully filed a close challenge. This is not to say that the median admin doesn't have a good understanding of policies and guidelines—the median admin is much more knowledgeable about them the average non-admin—but it's to say that sometimes the experienced non-admin can get it right even when a panel of admins get it wrong. It's also that closing RfCs can be hard, even very hard, because of the vast span of Wikipedia policies and how WP:CONLEVEL comes into play. S Marshall is exceptionally adept at closing RfC's (I would say moreso than the majority of admins) and is not representative of the typical user.

With respect to deletion, it's technically possible to create a new usergroup that has access to delete articles but not block people. The current guidance on deletions, per WP:NACD is that Non-administrators should limit their closes to outcomes they have the technical ability to implement; for example, non-admins should not close a discussion as delete, because only admins can delete pages. This seems reasonable to me, since letting people without the ability to delete pages close AfDs as delete would simply add onto the CSD backlog. This doesn't seem like the most time-efficient way to do things; having one editor close a deletion discussion and a second editor do the deletion seems inefficient to me. The question I think that's worth considering in the field of deletion is whether or not we, as a community, want to have a role that allows people to delete pages but not to block people nor perform other sensitive administrator tasks. And, if we do, the extent to which the community should vet these users before they are granted the role seems like a reasonable discussion to have. — Mhawk10 (talk) 07:43, 3 April 2022 (UTC)[reply]

Having two editors involved in editing happens all the time. Can't tell you how many edit requests I've made, how many technical page moves I've requested, it happens quite a lot. I see it as part of being an admin means using the tools to help other editors to make improvements. If efficiency is such a problem, then the solution would be to make every editor an admin so they don't have to bother others to get their editing done. And of course, that's not gonna happen. P.I. Ellsworth - ed. ^put'r there 08:12, 3 April 2022 (UTC)[reply]

I don't think that efficiency is the be-all-end-all, but from a segregation of duties perspective I don't see much of a benefit to having two separate people delete AfDs. In the current deletion workflow, an admin reads the AfD, makes a determination on consensus, and then actually deletes the page. There's even a nice script that lets people do this in one go. With two people, the non-admin reads the AfD and its comments, makes a determination of consensus, and then tags the relevant pages under CSD. A CSD-patrolling admin then comes along, reads the closure of the AfD to make sure it isn't insane, and then performs the deletion. There's a small benefit with the admin doing the whole thing in terms of total time used, since no work is really duplicated (although at the expense of a second set of eyes on the AfD closure).

The main thing that I think would come up is that deletion is one of the most fraught areas of the administrative side of the encyclopedia and the community (from my understanding) tends to want to vet people before they go around deleting articles wholesale in close-call scenarios. I also think that there would be less disruption (i.e. time and effort wasted from challenging bad closes) in the current process relative to one where Randy in Boise decided to close an AfD as delete where a merge or a "no consensus" was the appropriate outcome. And, as S Marshall points out below, there are some other systemic risks that would be amplified in the absence of a vetting process.

Having a role that allows a user to delete pages but doesn't allow them to block people seems to be tailored towards both direct workflow efficiency and preserving the existence of a thorough vetting process. The drawbacks I've heard for this idea are that this would lead to fewer administrators and that the vetting process could itself be inefficient (see: RFA disasters). — Mhawk10 (talk) 08:39, 3 April 2022 (UTC)[reply]

This isn't to say that the vetting process is perfect—it's not even perfect at stopping socks—but it's a risk reduction approach that is aimed at preventing bad people from getting advanced permissions. — Mhawk10 (talk) 08:44, 3 April 2022 (UTC)[reply]

All things considered, the vetting process does seem to work pretty well. In all these years I've only come across two admins who were "questionable", and I learned from them, too, anyway. And I think they learned some things, as well. The vast majority of admins (and probably all of 'em in this present moment) are knowledgable, helpful and trusted. Still though, I don't really understand how non-disclosure would help unscrupulous editors game the system. I don't see how an undisclosed non-admin who closes an AfD and slaps a SD on a page would in any way disrupt the system. One thing I DO understand is that there are some things we don't really want to talk about, because we don't want to give out any "bad ideas", so feel free to ignore that part of my response. My understanding of your and S. Marshall's ideas is not crucial to whether or not I love to edit Wikipedia. P.I. Ellsworth - ed. ^put'r there 09:01, 3 April 2022 (UTC)[reply]

In terms of there are some things we don't really want to talk about, because we don't want to give out any "bad ideas" I'm not really sure what this is supposed to convey except WP:BEANS, but I feel like I'm missing something. Anywho, my comment you appear to responding to focused on closing AfDs as delete rather than the question of how to handle RfC/RM disclosures. I generally think that the tag at the end of the closures serve little utility. The main thing that matters to me is whether or not the summary fails to accurately reflect the discussion and/or appropriately ascertain consensus in light of relevant policies. With respect to these, the closing summary should be able to speak for itself. The only time where the user themself matters is if they've made an involved close. — Mhawk10 (talk) 16:47, 3 April 2022 (UTC)[reply]

For deletion, while it would add the CSD backlog, I don't see it as an inefficient use of time; an admin confirming that an AFD has been closed as delete and deleting the article takes less time than determining consensus, closing the discussion, and deleting the article. BilledMammal (talk) 11:58, 3 April 2022 (UTC)[reply]

Guess the bottom line for me is, just because I'm a non-admin doesn't mean I can't get the job done. When I need the tools, I make a request for an admin's help. I make an edit request, or I slap an SD on a page after closing a deletion discussion. Of course the admin is going to check the deletion discussion to make sure it's proper, and that's expected. That's SOP. The vast majority of edits (thank goodness!) can be made by a non-admin without the help of an admin. And when necessary, a non-admin can, with the help of an admin, use the tools and make any edit on Wikipedia that's possible. As much as I don't mind declaring I'm a non-admin, I really don't see why it's necessary. P.I. Ellsworth - ed. ^put'r there 08:29, 3 April 2022 (UTC)[reply]

• We have some comedians, on Wikipedia, who would love to take it on themselves to close discussions with their own alternative accounts. Don't remove the disclosure requirement. It would be easily gamed.—S Marshall T/C 08:34, 3 April 2022 (UTC)[reply]
  • I don't believe removing the requirement to disclose in RM's that the closer is not an admin will impact that. BilledMammal (talk) 11:58, 3 April 2022 (UTC)[reply]

No thank you. Admins being closers of AFD discussions especially is one of the core things that makes an admin an admin. casualdejekyll 12:45, 4 April 2022 (UTC)[reply]

Yes, I agree and I don't think anybody is trying to take these core functions away from the important things that admins do. Take a look for example at Wikipedia:Miscellany for deletion/Draft:2022 West bengal Municipal election; that's an MfD, and similar things happen at AfD as well. They really aren't something that admins need to spend time on when there are so many much more important and difficult decisions we trust them with. Point is, what difference did my disclosure that I'm a non-admin make? There might be some small advantage to it, and it's just not that big a deal either way, so why make it necessary to disclose it? If there's something I'm missing, I'd love to learn it! P.I. Ellsworth - ed. ^put'r there 17:48, 4 April 2022 (UTC)[reply]

This seems more like an WP:IAR scenario than a scenario that actually requires a changing of the rules. casualdejekyll 18:44, 11 April 2022 (UTC)[reply]

• I'm not particularly keen on this. I've closed a few AFDs when the article had already been deleted and it's just a matter of cleanup, but I don't see the point of having people close AFDs as "delete" when they can't carry out the action. It simply creates a different backlog for admins to monitor. Mangoe (talk) 20:28, 4 April 2022 (UTC)[reply]

  It also doesn't save any work. I'm not going to delete a page just because somebody told me to. By the time I've analyzed enough of the AfD to convince myself that the NAC delete request is valid, I might as well have just closed it myself. -- RoySmith (talk) 22:08, 4 April 2022 (UTC)[reply]

  In some cases it saves work, as in those already mentioned. In other cases, while it doesn't save work, if it's done correctly then it doesn't make extra work for you. This has the benefit of helping an editor prepare to be an admin, or to just gain experience as a closer. If it is done incorrectly, then the benefit is an education for the closer. Admins are crucial to this project; it couldn't be done well without them. One of their important jobs that I've seen most admins take very seriously is to educate me when I've messed up. Such errors and corrections have made me a better editor, a better closer, a better non-admin. Where would we be if all educators took the stand, "I might as well have just closed it myself"? If you've not found any errors, then you can take a second to commend the non-admin closer, and if you did find something amiss, then you can take a minute to educate the closer. Where exactly is the downside? P.I. Ellsworth - ed. ^put'r there 05:09, 5 April 2022 (UTC)[reply]

  @Paine Ellsworth it does take longer than it would otherwise: in the same way that a DRV review takes me longer than an AfD review. I'd need to review the AfD and a close. Sometimes that's 20 seconds extra, but often significantly longer. Any AfD that I needed to revert the close on would take significantly longer than a minute. Much longer, for the double check any reverted action gets, then the time to undo and reclose, then the comments to the editor.

  Now you can still feel it's worthwhile, but please don't understate the downsides. Nosebagbear (talk) 19:29, 7 April 2022 (UTC)[reply]

  To editor Nosebagbear: thank you for this! I don't want to do that; don't want to understate the downsides. This is an education for me. I can't possibly know all there is to know about being an admin, never having actually been one. So thanks again. In your mind, with a closer's eye on such downsides, how do they compare with the upsides mentioned here? From my point of view the comparison does still favor the nom's idea, just not quite as favorably as before I read your words above. How do you think the ups compare with the downs? P.I. Ellsworth - ed. ^put'r there 22:56, 7 April 2022 (UTC)[reply]

• Count me as one who thinks that non-admins should not close AFD discussions as delete - deciding on an action that they cannot implement. The person who makes the decision that an article should be deleted, and the person who actually deletes it, should be the same person, and should be someone who has been authorized by the community to make that decision. I agree with Roy Smith that by the time an admin has reviewed the "delete" recommendation and decided whether it is valid, they might just as well have been the primary closer. In any case, I don't see any real need for such a change. There does not appear to be a huge backlog of AfD discussions languishing for want of someone to close them, at least not where articles are concerned. Most discussions are closed on the same day they become eligible for closure, and the few that aren't are far too complex for a non-administrator to attempt. -- MelanieN (talk) 00:41, 8 April 2022 (UTC)[reply]
• I agree with the above in that an AfD close as delete as a non-admin is redundant; currently it is not outright banned for a non-admin to close as delete, it is heavily discouraged, as Roy stated above it doesn't really alleviate administrator workload. The status quo works fine, considering the XfD backlog doesn't get full very often. Curbon7 (talk) 03:16, 8 April 2022 (UTC)[reply]

• I have performed a couple of non-admin delete closures. What happened in those cases was that the article was deleted by an admin for copyvio. The non-admin close was therefore just a formality to close the AfD. Hawkeye7 (discuss) 05:44, 8 April 2022 (UTC)[reply]

Yes, in this case where the article has already been deleted, usually as a speedy delete for some reason or other, then anyone can close the discussion, with appropriate explanation that the article has already been deleted and why. As you say, it is a formality. The admin who deleted it may not even have been aware that there was an AfD. -- MelanieN (talk) 14:41, 8 April 2022 (UTC)[reply]

• On the first half of this, there should not be any requirement to tag a closure by NAC. I'm not sure when that practice begun, but it's just silly; any experienced editor in good standing should be able to close most discussions, and shouldn't have to wear the badge of shame for not being an admin when closing it. On the second part, notice I said "most". This topic comes up every so often, and I still (as I have every time in the past) think that there's no practical reason for a non-admin to close a deletion discussion as delete if they can't actually delete the article. Closing as "keep" is fine; they don't have to do anything. For a purely pragmatic and workflow-related reason, if you can't enact the results, don't close the discussion. Any admin who would delete the article would need to read and assess the consensus anyways (best practice to make sure you're doing it for a good reason) and at that point, they can close it as well. It saves no one any work for a non-admin to close a discussion as delete and then fetch an admin to do the deleting. It's cumbersome and not helpful to anyone. This has nothing to do with trust, it's just a pragmatic issue that it isn't helpful. --Jayron32 14:49, 8 April 2022 (UTC)[reply]

• It seems there may be a consensus for removing the first, but no consensus for the second.

I think an RFC will be needed for to implement the first, and I propose the following question: "Should all requirements for non-admin closers to state that they are not an admin in their closure, such as at WP:RMNAC, be removed?" BilledMammal (talk) 03:05, 23 April 2022 (UTC)[reply]

Hi! Are there any meaningful distinctions between these two assessment grades? What would be the disadvantages of merging them? Is there recent consensus on this matter?A. C. Santacruz ⁂ Please ping me! 08:42, 8 April 2022 (UTC)[reply]

There is no useful distinction in practice. The only downside to merging them today is that the whole system should be redone (separate assessments for different wikiprojects should go, for a start), and this merger just rearranges deckchairs. —Kusma (talk) 10:08, 8 April 2022 (UTC)[reply]

C-class was created through a discussion some years after the rest.

If you define "Stub" as <10 sentences, then "Start" is 11+ sentences.

If you define "B class" as multiple sections, each of which has at least one inline citation, then "C class" is less than that.

The difference between "Start" and "C class" is usually length and the number of citations on the page, but you can see the official definitions at Wikipedia:Content assessment. Some of the rating tools automatically provide estimates based on the Wikipedia:ORES system. WhatamIdoing (talk) 19:14, 8 April 2022 (UTC)[reply]

I think my point is that I'm unaware of any use for the information encoded in the distinction between C and Start. —Kusma (talk) 19:48, 8 April 2022 (UTC)[reply]

C class articles generally only need a bit of work to meet B or GA standard. The projects target them for improvement. Hawkeye7 (discuss) 20:07, 8 April 2022 (UTC)[reply]

I think one of the questions to consider is in how many projects this actually happens. Things like proper B-class and A-class assessments work fine at the MILHIST project but are uncommon or poorly organised elsewhere. —Kusma (talk) 07:25, 9 April 2022 (UTC)[reply]

It makes sense to target the grade preceding B to improve for B level, but how different is that from targetting Start-level articles, Hawkeye7? A. C. Santacruz ⁂ Please ping me! 09:52, 9 April 2022 (UTC)[reply]

The difference is that Start class articles would not have sufficient content, but C class could have enough. It is much easier to grade something as C, as B needs some assessment work and C can easily be determined by its size. Graeme Bartlett (talk) 21:33, 9 April 2022 (UTC)[reply]

Roughly half of all MilHist articles are graded Start class; a quarter are C class and a tenth are B class. C class is very close to B; usually the only problem is a lack of referencing. How much work it is to fix that depends on how many references are missing. Sometimes there is just one; sometimes the whole article is unreferenced. Start class articles cover a wider quality range. Some are little better than stubs; they invariably lack adequate references, but usually also lack content, and often proper structure and supporting materials as well. As Graeme says, they are incomplete, and will require considerable work to bring them up to standard. Hawkeye7 (discuss) 05:08, 11 April 2022 (UTC)[reply]

@Kusma, the rating system exists because of Wikipedia:Version 1.0 Editorial Team. They wanted to create a subset of articles for offline use. Which articles? They decided that it would be good to include all the popular ones (page views), all the heavily linked ones (Special:WhatLinksHere), all the high-quality ones (quality ratings), and all the ones that various groups say are important (WikiProject priority/importance ratings). Consequently, there's a formula for determining which things to include. Two of those are encoded in WikiProject ratings. A top-importance FA will basically always be included. A "low-importance stub" almost never will. Differentiating between multiple quality classes lets them be a little more granular in their choices.

These ratings are used for other purposes. WPMED, for example, periodically checks to see whether we have any high-importance stubs, and brings those up to a better standard. You can see at Wikipedia:WikiProject Medicine/Assessment#Statistics that all of our top-importance articles are currently C-class or better, and all of our high-importance articles are Start-class or better.

What people shouldn't do is assume that this system is really important. It's not useless, but nobody should treat it like a key system. WhatamIdoing (talk) 18:28, 11 April 2022 (UTC)[reply]

@WhatamIdoing: Yes, I am aware of that. I remember when a lot of WikiProjects were founded mostly in order to provide article assessments (some of the big country projects did not exist before WP1.0). A long time ago, I made most of the original "importance scale" decisions for WikiProject Germany. I have also tagged and assessed a few thousand articles. So I have produced a lot of this type of data, but I don't remember using it much, and I am no longer convinced this was worth spending time on. —Kusma (talk) 18:48, 11 April 2022 (UTC)[reply]

One of the admins who wrote an assessment script came to the same decision years ago, and deleted the script I was using at the time. It was temporarily inconvenient.

I think that we should consider ways of automating as much of the Stub- and Start- assessments as feasible. That would reduce the amount of time spent. I can tell you, for example, that nearly everything in Category:Unknown-importance medicine articles could be safely assessed by a bot. Does ORES say that it's a stub? Then it's a stub. Is the article about a person or organization? Then it needs |importance=low|society=yes. That would clear out a huge fraction of the unassessed articles without anyone spending any more time on it. WhatamIdoing (talk) 19:52, 11 April 2022 (UTC)[reply]

Personal opinion: Instead of merging start and C class, merge stub and start class? casualdejekyll 18:59, 10 April 2022 (UTC)[reply]

This is one of my pet peeves. There's too many grades, with too fine gradations between them. What's the difference between "essentially complete" and "mostly complete"? Beats me. I can't for the life of me figure out how A differs from GA. A means "It's passed a grade-A review". GA means "It's passed a GA review". Is A better than GA, or the other way around? It's listed above GA in the chart, but it also says GA is not a requirement. But, B is specifically called out as less than GA. So, beats me.

I can't think of a single time when I've queried somebody who assigned a grade to an article I've written when I got back a useful explanation of why it was one grade instead of another. I'm not saying we shouldn't have a grading system. Just that 7 levels, with vaguely defined criteria, is too many to rationally expect people to accurately sort into. I've been in discussions where one person was arguing for Start and another for B, for the same article. -- RoySmith (talk) 19:35, 10 April 2022 (UTC)[reply]

A is better than GA. Most A class articles could pass FA. The important distinction is that A, B and C are project ratings, whereas GA and FA are Wikipedia-wide. Hawkeye7 (discuss) 04:43, 11 April 2022 (UTC)[reply]

I would suggest that in practice, B and C rating are not linked to projects for the majority of articles. Updates tend to update for all projects at the same time. CMD (talk) 04:48, 11 April 2022 (UTC)[reply]

Again, the only project for which this is true is MILHIST. No other project has the necessary manpower or level of organisation. —Kusma (talk) 13:16, 11 April 2022 (UTC)[reply]

In my experience, most assessments by regular assessors are much lower than the official definitions would suggest, & appear to be based entirely on length, which the criteria don't mention at all as such. They generally take no account of the size of the topic, and the information that is likely to be available. One can tell from edit histories that the "assessment" lasts only a few seconds. Hardly anyone takes any notice of them. Don't be afraid to self-assess. Johnbod (talk) 03:17, 11 April 2022 (UTC)[reply]

On MilHist, new articles tagged as belonging to the project by the creator or the new page patrol are automatically assessed by the project bot. Stub, Start and C grade assessments are based on ORES and our project review checklist, which provides the basis for ratings. Articles assessed as B class by the Bot are reassessed by the project coordinators. (The Bot provides a monthly list.) Hawkeye7 (discuss) 04:43, 11 April 2022 (UTC)[reply]

I'd be very happy to have a bot auto-apply Stub ratings based on the ORES system. I've asked the ORES folks in the past to make lists of WPMED articles that are stubs but rated as something else, and when it says that an article is a Stub, the rating has always been accurate. Sometimes I think it slightly overrates an article on the border between Stub and Start, but it's really quite good overall. Some of the rating scripts have this built in now. WhatamIdoing (talk) 18:34, 11 April 2022 (UTC)[reply]

After studying content quality in communities like Wikipedia and through that developing the machine learning models that were the basis for ORES' article quality model, my take on the distinctions between some of the classes largely involves whether they follow the Manual of Style (MoS) or not. Both FAs and A-class articles are "complete", so an FA is an A-class article that has gone through FAC and through that follows the MoS. From my experience, MILHIST is the only project that really uses A-class, and so there's not enough of them to learn how to predict them well, which is why ORES doesn't have A-class. GA and B-class articles to me share a similar relationship with regards to the MoS.

It's relatively easy to predict the highest and lowest quality classes, something which ORES' training scores also shows. As @WhatamIdoing mentions, if it defines something as a Stub it's almost always correct. I suspect this is because the quality ratings in English Wikipedia are closely tied to the amount of content and references in the article, and with Stubs having little of both it's not a challenging task. The boundaries between many of the other classes are ones where it struggles, and I've also seen Wikipedians struggle with them as well. Whether an article is Start- or C-class isn't necessarily easy to determine, and the same goes for moving from C- to B-class. In one of our research papers we found a set of articles that were predicted to be C-class but labelled as Start-class by humans. Closer inspection of these found that the humans weren't following the criteria as defined at that time, meaning the machine learning model seemed to apply the criteria more appropriately. The criteria has since changed, and while we've also improved the dataset used for training ORES it still struggles with those intermediary classes. I'm not surprised to hear that it's overrating on the boundaries, that also seems to be my impression. Whether humans underrate articles and a "truer" rating is somewhere in between has yet to be determined ;)

I'd also like to mention the human aspect of having multiple quality classes and being able to receive confirmation that one's work is valuable. Improving an article a lot and then seeing the quality rating change is good. However, rating articles is a tedious task and I think everyone agrees that the quality ratings tend to not reflect the actual quality of the article (e.g. we accounted for this when collecting data to train ORES by identifying the first occurrence of a rating on the talk page). Our first paper about the ORES models argues that machine learning models like these should help contributors make decisions about what to work on. We did some initial experimentation with SuggestBot around this, but I think Wiki Ed has since done this much better in their dashboard where one can ask questions like "if I add two images to this article, what happens to its quality rating?" It's then possible to compare that to "what happens if I add a section with two paragraphs and citations?" and make decisions on what to work on. We're not at the stage where that's easily accessible, but the Hackathon is coming up in May and someone could take that idea and build a gadget.

Apologies for the long post, but this is my research area and it was fun thinking about this and writing it up. Cheers, Nettrom (talk) 09:43, 12 April 2022 (UTC)[reply]

I've seen obvious cases of human under-rating; once, an editor marked a long, sourced article as "Stub" because there was no content related to a specific sub-subject. But I don't think that happens very often. Usually, I think it's a case of editors thinking that nobody will yell if they underrate an article, but someone might accuse them of Having Weak Standards if they overrate it.

@Nettrom, if you're feeling bored some day, I'd love to have a list of WPMED articles whose rating mismatches the ORES prediction by two or more classes. It might be difficult to differentiate between "Start" and "C", but if the page says "Start" and ORES predicts "B" (or the other way around), then the existing rating is probably wrong. WhatamIdoing (talk) 18:27, 12 April 2022 (UTC)[reply]

The tooltips feature is awesome when needing to quickly understand a word or topic. But what if a tooltip's text also contains something that needs to be understood? Enter nested tooltips. When a user hovers over a link and a tooltip expands, if there is another link within that tooltip the user can hover over that creating a new nested tooltip. A user can ctrl+click a nested tooltip to keep it from automatically disappearing when the user moves their mouse. This feature is exactly like the innovative expanded/nested tooltips in the game Crusader Kings 3 ^[1]. — Preceding unsigned comment added by Contramonk (talk • contribs) 13:07, 12 April 2022 (UTC)[reply]

@Contramonk: What you are proposing is not possible in HTML5. A tooltip is generated by your browser using the value of an element's title= attribute (try hovering your mouse over this text for example), and such values may not contain markup of any kind - least of all links. So it's a non-starter. --Redrose64 🌹 (talk) 22:04, 12 April 2022 (UTC)[reply]

Nested popup help can't be done with the title attribute, but could be done via Javascript. But I wouldn't recommend it for general purpose use. Popups activated by mouse motion is a problem for those who have trouble with fine motor movements, and needs a different triggering mechanism on touch devices that doesn't affect the ability to select and follow links. On a site like Wikipedia that tries to be interoperable with as many devices and environments as possible, I also think less is more: if you find yourself wanting to nest popup help, you should re-think how the whole help text is written. isaacl (talk) 22:32, 12 April 2022 (UTC)[reply]

I find the popups to generally be annoying on my chromebook touchpad. Donald Albury 23:20, 12 April 2022 (UTC)[reply]

@Contramonk, I think you can get the functionality that you want by enabling WP:NAVPOPS. Whatamidoing (WMF) (talk) 19:39, 13 April 2022 (UTC)[reply]

Background on my thinking: I've been thinking about this topic for months, informed by creating 200+ articles.

Background on the topic: The world favors certain demographics, and disfavors others. Race, gender, nationality, sexuality and most social factors make it easier for certain people to get noticed, to get media attention. I assume I don't need to provide citations and links and that we're all agreed that there is bias in Wikipedia that is replicating or mirroring the bias in the world.

To define the problem: Wikipedia seems to, for the most part, see itself as neutral. The world is unequal, it's not Wikipedia that is creating inequity, we're just reflecting it. There are exceptions, such as the Women in Red project that is making a great effort to reduce gender bias. But modern thinking on equity is that being neutral is to be part of the problem. It's not enough to be not biased, we need to work to be anti-biased.

A solution?: So here's my concept of an idea that I really want editors to critique and hopefully improve: there should be guidance that allows notability criteria to be considered through a lens of equity. For example if WP:ARTIST says that to be notable, an artist needs to be in the permanent collection of multiple museums, but the artist is an Aboriginal women in Australia, maybe we could consider that being in the permanent collection of one museum could be a fairer measure.

If WP:AUTHOR needs multiple reviews in journals, but the author is a woman in a country that rank in the worst 10% for gender equity, we might relax the rules for her.

These examples are both inspired by recent articles for discussion debates that I've been in where we're holding everyone to exactly the same standard, utterly blind to the much higher bar that many people face based on their sexuality, gender, race or luck of birth.

Is there a way we can introduce equity guidance for a fairer wikipedia?

This is my first proposal, so please be kind if I have erred in any way. CT55555 (talk) 12:46, 13 April 2022 (UTC)[reply]

• So, there are three different issues here. 1) Ignore the SNGs. They are beyond worthless, and give little to no useful guidance on what should and shouldn't be created as a stand-alone articles. You've highlighted just one reason, but the problems with the various SNGs are legion, and beyond the scope of this discussion to get into. 2) WP:GNG is a better measure of when to create a stand-alone article. It says, in essence "We need high-quality source material from which we can research a subject in order to build an article". Let's take your aboriginal artist. Pretend Wikipedia doesn't exist for a second. Where can I learn about them, their works, and their life story? If the answer is "I've got a bunch of great texts for you to read from", then yes, you have the basis to create an article. If your answer is "Basically nowhere", then how do we build a Wikipedia article about that person? 3) You are essentially correct on the equity issues facing Wikipedia, but there are better ways to approach this than writing about subjects for which there is no source text to refer to when building a Wikipedia article. The equity and bias issues with regard to Wikipedia article creation is that lots of great underrepresented topics do have lots of source text we can go to. Those topics are ignored in favor of, say, the standard DWEM subjects we often focus on. My advice is, instead of seeking to lower Wikipedia's standards so we can include poorly referenced articles containing lots of text that we have no means of verifying, we should instead focus on seeking out subjects that we could already be building articles about, but are not. --Jayron32 14:19, 13 April 2022 (UTC)[reply]

  Trying to write articles about people without good source material is obviously a silly endeavor, because it would be difficult or impossible. I think the sort of logic I'm suggestion can only work with more shades of grey examples than someone with no body of works to draw upon. And I think my argument really is only applicable if we don't take your advice to ignore SNGs, because they do form the basis of many discussions/decisions at Articles for Deletion.

  So maybe an example where this could be useful is where an artist, for example, is mentioned a lot in good sources, mainstream media, or good books has lots of brief mentions about their poetry, book, film or play, but none are the depth of coverage we look for. So there is enough to make a profile, but it's pieced together from verifiable details. In that scenario at Articles for Deletion, the SNGs are the benchmarks and if someone has a piece of art in 1 or 2 national galleries, if someone's book is reviewed by 1 or 2 book reviewers, are the decision points on which articles get kept or deleted on.

  So if you and I could agree that SNGs are flawed but we should try to make them better, would this be an improvement to them, for this kind of scenario? CT55555 (talk) 14:38, 13 April 2022 (UTC)[reply]

  So, the situation is this with the GNG, notability, and AFDs. It has rarely happened where an article is written that is well-referenced to the proper sources, is of sufficient depth and length, and then gets nominated for deletion, and even less rarely does it actually get deleted. 99% of the time, the situation happens that a poorly referenced, stubby article about someone, like say your Aboriginal artist, gets nominated for deletion, and then people who think we should keep the article either a) have to scramble to find source material or b) level accusations of bad faith and bias against people who nominated it for deletion. Neither are particularly helpful ways to build the encyclopedia, and it all stems from building an article in the wrong order. First, gather the sources, then write the article while citing the sources the whole way. rarely, if ever, does anyone complain when that happens. --Jayron32 14:44, 13 April 2022 (UTC)[reply]

  Just for the record, none of these examples are about articles I have created. But I have been often been the person who scrambles to try to save articles. About bias, my reading of this is that all have bias, although I'd consider it counter-productive to point that out to an AfD nominator. Maybe I should make a proposal that points out the over/under representation of articles at AfD based on demographic data of biographical subjects :-)

  Anyway, although I think we're not agreeing on my main point, I do agree with your suggesting about just creating more articles about people on topics that are missing. I do welcome any examples of topics that you have to suggest. CT55555 (talk) 14:53, 13 April 2022 (UTC)[reply]

  I agree that equity is a major problem. What I don't agree is that the problem is solved by allowing the creation of substandard articles. We solve the problem by creating better articles about under-represented subjects. Letting articles slide on lower standards of referencing and verifiability is not useful, and would only exacerbate equity problems, not ameliorate them. --Jayron32 15:50, 13 April 2022 (UTC)[reply]

  I hadn't seen my idea thought that lens, but I take your point. I wonder if I'm looking in the wrong direction then and the answer is to incentivize or motivate others towards the creation of such articles. CT55555 (talk) 16:01, 13 April 2022 (UTC)[reply]

  Incentivizing the creation of articles is the realm of various Wikipedia-adjacent organizations. As just one really good example, we have Wikipedia:WikiProject Women in Red, a group that works towards correcting gender imbalance in Wikipedia by creating high-quality biographical articles about women. There are various other things like Wikimedia chapters that work locally to build up Wikipedia editors from local cultures; by increasing representation of editors at Wikipedia, we increase the number of articles these editors will be interested in writing on under-represented topics. There are various edit-a-thons that are organized to help improve articles from many (often under-represented) topics. Ultimately, insofar as the problem of under-representation in Wikipedia articles is one of culture at Wikipedia, and not of policy, then we can only fix the problem through solutions of a cultural nature, like the ones I listed here (among others). Policy by itself can't fix culture. My argument is that the policy is sound policy when our goal is to create a high-quality, well-referenced, verifiable encyclopedia. The fact that that the culture among the Wikipedia community undervalues certain topics is not fixed by changing that policy. It's fixed by adding additional culture around inclusion and expansion of access and deliberately correcting the imbalance within the sound policy. --Jayron32 16:16, 13 April 2022 (UTC)[reply]

  I am now wondering if you read all of my initial post... CT55555 (talk) 16:29, 13 April 2022 (UTC)[reply]

  I have read your post multiple times. What parts of my responses have made you feel misunderstood? --Jayron32 16:39, 13 April 2022 (UTC)[reply]

  When you told me about Women in Red when after starting post mentioned them. CT55555 (talk) 16:40, 13 April 2022 (UTC)[reply]

  Gotcha. Yes, we both agree they are a good example. Sorry that I reached for the same example you did. It is hard to find a better group, and perhaps that could serve as a model for starting other ones. I did not mean to give the impression that you had not already recognized them, but I clearly did do just that, and that is all on me. Sorry that I did that. I should not have. Regardless, I still stand by the rest of my post; culture and policy are different things, and at Wikipedia we have broken culture. Policy cannot fix broken culture, and the hard work of fixing culture requires a different tactic. --Jayron32 16:48, 13 April 2022 (UTC)[reply]

  By the way, Women in Red (which I am an proud and active member of), is producing biographical articles about women, using this month as an example, at rate of:

  5 per day (general, discounting upgrades), 0.5 per day (climate) 8 per day (via translation) and less than 0.5 per day (French overseas territories) Totaling about 14 articles per day.

  Of the 1.5 million biographies, 19% are about women. An equal world would mean about 765,000 biographical articles were about women, but only 285,000 are. We're 480,000 short.

  If Women in Red is our best fix, at the rate Women in Red produces articles, it will solve the problem in 93.9 years. That's assuming AfD stops nominating articles about woman at over double the rate it nominates men, despite the low representation of women on wikipedia, 41% of biographical articles about women nominated for deletion are about women.

  My hypothesis: we need to do a more, we need to move a faster. To me, a 94 year plan isn't good enough.

  Yes, culture beats policy, but culture changes as rules and standards are introduced. People didn't start wearing seatbelts until the law required them. We probably need to move the culture needle and the policy needle. CT55555 (talk) 21:03, 13 April 2022 (UTC)[reply]

  As you led off, we should never expect 50/50 balance on male/female articles, because we know historically the coverage of women was far less then men, and that we should not be trying to correct the ratio by including poor quality articles. We can improve the ratio but it will never get a where close to 50%. --Masem (t) 21:48, 13 April 2022 (UTC)[reply]

  Do you have a suggestions for how to get it higher than the status quo? CT55555 (talk) 21:52, 13 April 2022 (UTC)[reply]

  @CT55555 I would also just note that That's assuming AfD stops nominating articles about woman at over double the rate it nominates men, despite the low representation of women on wikipedia, 41% of biographical articles about women nominated for deletion are about women. is not sufficiently clear. Heavily due to the work of both WIR and those who act in a similar vein without being part of the org, of articles created per day, a majority are about women. AfDs are always more likely to be about newer articles.

  Thus while there may well be an issue on the AfD side, what we'd need to consider it is if you can determine and write the like for like comparisons. E.g, for articles made in the last year, how's do the creations/nominations/actual deletions stack up, and how does that change in the 2-3 year bracket, and 3+ year bracket. Nosebagbear (talk) 09:36, 19 April 2022 (UTC)[reply]

  I think the point is that, while articles on males are given the benefit of the doubt, and never make it to AFD, articles on women are much more likely to be scrutinized at an unreasonable level. The fact that 41% of the biographical articles at AFD are about women, and yet a much smaller number of newly created articles are about women, means that on the balance, the same quality article about a woman will end up at AFD more often than if it were about a man. We had a particularly nasty case a few years back where a nobel laureate's article was sitting at AFD while they were winning the Nobel Prize. And even afterwards, people were still trying to argue that wasn't a problem... C'mon now...--Jayron32 11:47, 19 April 2022 (UTC)[reply]

  Here is a link to a peer reviewed paper that analysed a large data set of AfD nominations and concluded that that AfD process is incorrectly used for women's biographies significantly more than men's.

  https://journals.sagepub.com/doi/10.1177/14614448211023772

  I found this at: https://en.wikipedia.org/wiki/Gender_bias_on_Wikipedia

  I think we're way past the point of debating if this is a problem and encourage you to join me in trying to imagine solutions. CT55555 (talk) 15:46, 19 April 2022 (UTC)[reply]

  Regarding that particular article, you might want to read Wikipedia:Wikipedia Signpost/2021-07-25/Recent research. Anomie⚔ 21:52, 19 April 2022 (UTC)[reply]

  If we were discussing an article, that response would considered a primary source, whereas mine was a published academic paper. So while I respect the points contained within it, I again plea that we move forward from the "is gender bias a problem" phase and into the "can we please work on solutions" phase. CT55555 (talk) 00:43, 20 April 2022 (UTC)[reply]

  Well, here's one idea for you, @CT55555: Many AFD nominations use Twinkle. Could we make Twinkle 'aware' of which articles are about women? If someone nominates a BLP about a woman, could Twinkle add an extra step/dialog box that says something like "Please make sure that you have done a proper WP:BEFORE search before sending this article to AFD"? (I don't know if it needs to have an explanation like "Articles about women are nominated for deletion about twice as often as articles about men, but they are less likely to be deleted.")

  My idea is that if there are an above-average level of nominations but a below-average level of deletions at AFD, then editors are sending some notable subjects to AFD. If we could encourage them to be a little more careful, we might get closer to the average (which means fewer articles at AFD, but a higher percentage of those articles being deleted). WhatamIdoing (talk) 07:04, 20 April 2022 (UTC)[reply]

  "Here's a paper" "That paper has already been looked at and received significant criticism" "But it's a published paper!" 🤷 Sure. People can get all sorts of crap published, especially in particular topic areas. Without better understanding of the actual parameters of the problem, I think we'll have trouble coming up with (non-politician's) solutions more specific than "interested editors should watch AfD to rescue articles they think can actually pass WP:N". But feel free to run with that one if you want. Anomie⚔ 11:24, 20 April 2022 (UTC)[reply]

  I would remind you that I am here asking people to suggest better solutions, soliciting critique, which seems quite different from the activities described in the politicians solutions article.

  Your quotes are not accurate and I request that you score them out. CT55555 (talk) 13:23, 20 April 2022 (UTC)[reply]

  Those "quotes" appear to represent a hypothetical dialog between two people, not the exact words said by anyone in reality.

  Of course, "received significant criticism" is not evidence of actual problems. Sometimes criticism looks more like the advice given in The Greasy Pole: "Some of the main conclusions have been questioned. (If they haven’t, question them yourself; then they have)." Even when a study isn't perfect, it can sometimes give us an approximate idea of reality. WhatamIdoing (talk) 18:24, 20 April 2022 (UTC)[reply]

  I took a quick look at the ration of male to female subjects of articles at Category:AfD debates (Biographical) right now. I was able to determine the value of the P21 property (sex or gender) in Wikidata of 185 of them. 47 are female, 138 male. This is a gross oversimplification of course, but it looks like articles about women are still nominated for AfD more frequently than those about men. Worth looking into how those ratios work out for only living people, recently created articles, etc. Anyone interested in working on that? Vexations (talk) 22:24, 19 April 2022 (UTC)[reply]

  ^I couldn't agree more with what Jayron32 has saved here. Mysterious Whisper (talk) 16:00, 13 April 2022 (UTC)[reply]

  I think that the bit about lots of brief mentions about their poetry, book, film or play, but none are the depth of coverage we look for. So there is enough to make a profile, but it's pieced together from verifiable details is one of the open questions with the GNG. If you have many short independent sources, but it all adds up to enough material to write a fair article, is that okay? Editors have different opinions, and we have never settled the question.

  Obviously, if you have a dozen sources, all of which contain only the fact that "Alice Expert is a cryptanalyst", then that's not enough to write a whole article about Alice. But if you have enough short sources to write a decent, non-stub article, then is it really a problem that one source provides "Alice Expert is a cryptanalyst" and the next source provides "Alice Expert wrote a book" and a third source provides the basis for "Alice Expert is an American", and so forth? Is it really necessary to have a single source providing significant coverage by itself? (WP:CORP claims it is.) WhatamIdoing (talk) 19:53, 13 April 2022 (UTC)[reply]

  To follow onto what Jayron said, remember that the Various SNG provide presumptions of notability. If an SNG allows for an article based on an artist having works displayed permanently in multiple galleries, we are still going to expect in time that secondary sources with significant coverage will be found and added...if they can't be found, we'll still end up deleting the article. So this means it would be reasonable to consider slight allowances of weaker SNG passage for underrepresented people, but with the understanding we still ultimately are looking for significant coverage too, which is applied equally to all topics. --Masem (t) 16:09, 13 April 2022 (UTC)[reply]

• Have you reviewed WP:WHYN? The notability guidelines aren't just some arbitrary standard, they're the direct result of the nature of Wikipedia as an encyclopedia. Changing the notability guidelines won't make it suddenly possible to write an article without sources to base it on, or to put it another way, changing the words on the page at WP:N won't change the fundamental nature of notability on Wikipedia. The general notability guideline is descriptive, not prescriptive.
  You seem to understand that it's impossible to write a decent article about a subject that doesn't have at least a small body of source material which the article can be based on. In your response to Jayron32's comments, you speak of "shades of grey examples" where there are enough reliable secondary sources to write an article about a subject, but that subject somehow still doesn't pass the notability guidelines. I'm willing to believe, as you seem to, that this is a thing that happens, although I'm not convinced that it's a major problem or that your proposal idea is in any way a viable solution. Can you provide some examples of this, ie, subjects with complete and well-written articles, based on reliable secondary sources, getting deleted at AfD on notability grounds? If you can establish, with evidence, that this is a thing that happens, and that it often happens to subjects that are underrepresented, and that it's a major reason why those topics are underrepresented, then you'll have my full support in finding a solution. As of now, while we can (mostly) all agree that there is a problem here, it seems like a lot of questionable guesswork is being done in ascertaining the exact nature of the problem and in proposing solutions to it.
  I'll also add that it's usually possible to include information from reliable sources on topics that don't warrant a standalone article somewhere in the encyclopedia, as part of a larger article. That way, the information gets included and published for all to see, Wikipedia doesn't get littered with perpetually-incomplete stub articles, and the reader experience is improved, because they get complete and detailed articles rather than many disjointed tiny ones. Mysterious Whisper (talk) 15:33, 13 April 2022 (UTC)[reply]

  I thought this was the place to get a ball rolling and have other people build on it, suggest better ways. Have I misunderstood this space? CT55555 (talk) 15:41, 13 April 2022 (UTC)[reply]

  You're in the right place: you have a the makings of a proposal here, but it's not yet ready for WP:Village pump (proposals). I've given my feedback on your idea. Specifically, I think a proposal to change the notability guidelines will be much more likely to succeed, and accomplish something worthwhile, if its presented in relation to WP:WHYN. I think it would be more likely to succeed if you present a strong case based on verifiable evidence rather than speculation, and I've suggested one form that evidence might take. I have also suggested another way to accomplish some of your goals. I now feel as though this isn't going to go anywhere, however, so I'll cease my involvement. Good luck. Mysterious Whisper (talk) 16:00, 13 April 2022 (UTC)[reply]

  A body of evidence might take me a little while. If you are curious about the anecdotal story that was the trigger to me posting today:

  https://en.wikipedia.org/wiki/Wikipedia:Articles_for_deletion/Jeannie_Pwerle CT55555 (talk) 21:53, 13 April 2022 (UTC)[reply]

  And another example https://en.wikipedia.org/wiki/Wikipedia:Articles_for_deletion/Munira_Al-Fadhel CT55555 (talk) 05:39, 14 April 2022 (UTC)[reply]

  A third example https://en.wikipedia.org/wiki/Wikipedia:Articles_for_deletion/Louise_Crisp_(2nd_nomination) CT55555 (talk) 05:41, 14 April 2022 (UTC)[reply]

I think you may be on to something. There are other measures of "notability" one might conceive of than the number of articles in newspapers that are "about" the subject. I think of that as "critical engagement by domain experts". For me, an artist's inclusion in a significant collection, is more important than a "profile" in Tatler, like something I just saw in an article: https://www.tatler.com/gallery/robyn-ward-exhibition-launch for example. Tatler knows fuck all about art, but the curators at the National Gallery of Australia are highly qualified professionals. We should consider their opinion more significant than that of a celebrity gossip writer. Another thing you could look at is how we consider interviews. We write often write them off as primary sources, but they can provide, and are sometimes the only source of important biographical details. Sometimes piecing together a biography from snippets and primary is possible, and common sense can prevail over blind application of guidelines. It's good those discussions can take place at AfD, and the outcome of those discussions should inform the guidelines that reflect practice, not the other way around. Vexations (talk) 16:54, 13 April 2022 (UTC)[reply]

• I would like to see some discussion on how WP:RIGHTGREATWRONGS factors into all of this (both positively and negatively). Blueboar (talk) 12:21, 20 April 2022 (UTC)[reply]

  I see that guidance as directed towards editors making decisions about individual articles, not systemic issues. CT55555 (talk) 12:26, 20 April 2022 (UTC)[reply]

  The positive side of RGW is that the closer we bring our biographical article ratio to 1:1 male to female (roughly), the better. There's always going to be a systematic bias from how historical reporting has gone one that will limit that we will never reach 1:1 but we can do better than we do now. Negatively, we do not want to weaken notability to try to improve our equity ratios to a point that we have stubs and start-class articles on women that show no likely chance of improvement and were only created because one source mentioned the person's name with some iffy signs of being important. There's the ideal equity limit we would like to reach, and then there's going to be a practical one, the one that we know we can actually create encyclopedic-quality articles even if short simply due to that systematic bias from sourcing that we cannot change. --Masem (t) 12:39, 20 April 2022 (UTC)[reply]

Here's an idea: Improve participation at AfD. Here's a recent AfD I stumbled upon when looking at the list of women's biographies at AfD that I mentioned yesterday: Wikipedia:Articles for deletion/Shawna Hamic. I have no idea who this is, but one participant? One? Not even the creator of the article cared enough, or knew about the nomination to do anything about it. For starters, new contributors need to be told about watchlisting the pages they create. Then, they need to find a peer group that can help them either fix the article, and present valid arguments at AfD for keeping the article without being accused of canvassing. This too, might be something we'd want to look at a bit more closely. How many AfDs close as delete with minimal participation? Do we need to introduce a quorum? Vexations (talk) 14:51, 20 April 2022 (UTC)[reply]

OK, this is fantastic. Combined with the suggestions about some prompts about gender equity when nominating, perhaps something about doing a real WP:BEFORE this could be good. Thanks for this. CT55555 (talk) 14:56, 20 April 2022 (UTC)[reply]

Thought I'd throw this out again to see if anyone has any better ideas. If I want to search for a person on Wikipedia regardless of which language wikipedia they might be on, I have to go to google. Any better Ideas? For example, searching for "Noël Ottavi", there is no page at Noël Ottavi but is one at fr:Noël Ottavi Naraht (talk) 13:17, 13 April 2022 (UTC)[reply]

Help:Searching#Other search tools claims that searching from the Main Page does that, and I remember seeing non-English results in the past, but it no longer seems to work. Certes (talk) 13:26, 13 April 2022 (UTC)[reply]

There is also Global search tool if you need some more options —TheDJ (talk • contribs) 14:10, 13 April 2022 (UTC)[reply]

For your Noël Ottavi example, I think the "missing" feature is a cross-project link to Wikidata. There it technically exists with an English-language label (wikidata:Q3345865), seems like it should come up on the right-hand side ("Results from sister projects") during a Wikipedia search. Then you'd click on that, and be given a list of Wikipedias with an article about that subject. Obviously this wouldn't solve every issue. Brightgalrs (/braɪtˈɡæl.ərˌɛs/)^[ᴛ] 22:55, 13 April 2022 (UTC)[reply]

Hi, an infobox is the source of structured data about a concept and is both for humans and machines (in semantic web). For humans, infoboxes should be as abstract as possible and convey only main information to render them to humans "at the first glance". And for machines, it should be as complete as possible, so that machines can answer semantic queries. These purposes seem contradictory. But I have an idea for infoboxes that makes them ideal both for humans and machines

I think the correct policy is: 1-Infobox of an article contains full data as complete as possible (that is for machines), 2-All infobox items are integrated with an additional property, that is, "only for machines"/"only for humans"/"for both" 3-Infobox of an article is shown in "for humans mode" in the default way, to be abstract enough, but it has the capability of changing its mode: i.e. a human can read full infobox data by changing its mode to "for machines mode". When an article is loaded, a full infobox is loaded too, but only some of its items are visible to humans (i.e., those with "only for humans" or "for both" property) at the default mode, but there is an option for viewing it in the "for machines mode".

This way an infobox is appropriate both for humans (convey only main information at first glance), and for machines (contain complete data to answer their queries in semantic web). Please discuss that idea. Thanks, Hooman Mallahzadeh (talk) 10:02, 14 April 2022 (UTC)[reply]

@Hooman Mallahzadeh, take a look at the information at Wikidata and Wikipedia:Wikidata. Wikidata is a database for such information and each Wikipedia article has an entry. Just click on Wikidata item in the menu at the left of the article. StarryGrandma (talk) 20:00, 14 April 2022 (UTC)[reply]

• I have some quibbles about the premise to this question… while the way information is displayed in an infobox might make it easy for machines to scan and compile our information, that definitely is not the “purpose” of an infobox. I would disagree with saying that our infoboxes are intended to be “for machines”. Blueboar (talk) 21:27, 14 April 2022 (UTC)[reply]

An infobox is one source of structured data, but Wikidata usually provides it in a more machine-readable way, especially if you want to consider large numbers of articles (or even topics with no article on English Wikipedia). Tools such as SPARQL can query Wikidata efficiently to produce, for example, a list of German-speaking composers with an OBE. Certes (talk) 21:48, 14 April 2022 (UTC)[reply]

@Blueboar@Certes@StarryGrandma These semantics are metadata about that article data, and according to

These embedded semantics offer significant advantages such as reasoning over data and operating with heterogeneous data sources.

and

Specifications such as eRDF and RDFa allow arbitrary RDF data to be embedded in HTML pages.

Structured data should be «««embedded»»» within non-structured data, and ««not be redirected»» (e.g., from Wikidata). As I know Tim Berners Lee (the inventor of web) says "embed" metadata for machines within data for humans. Because this convention is not only for Wikipedia articles, but also for the total Web, to promote it to Web 3.0 or Semantic Web. Hooman Mallahzadeh (talk) 03:11, 15 April 2022 (UTC)[reply]

In addition, Wikidata item of an article is not always for a single concept. In current Wikipedia, some articles have more than one concept, e.g. Sine and cosine article. This way, "Wikidata metadata fetching" policy, should be for one of them either Sine or Cosine, and not both of them. But by infobox approach, we can create two infoboxes for each of these concepts, and show/hide one of them (or merge their "for humans" data into a a new infobox). Hooman Mallahzadeh (talk) 03:32, 15 April 2022 (UTC)[reply]

Wikidata solves that problem more simply, by having separate sine and cosine items. It also deals neatly with topics such as archaeologist, where the link to Wikipedia is a redirect to an article on a related topic. That article has no infobox and, if it did, it would describe the topic of archaeology, not of archaeologist. Certes (talk) 10:48, 15 April 2022 (UTC)[reply]

@Certes Does there exist a mechanism in Wikipedia to fetch Wikidata information and embed that data in a machine readable way (e.g., via "Web Ontology Language" (OWL)) in the articles written in different languages for the same concept? Hooman Mallahzadeh (talk) 11:01, 15 April 2022 (UTC)[reply]

Wikipedia:Wikidata#Infoboxes describes the current use in general, and Category:Infobox templates using Wikidata lists the relevant types of infobox. The code tends to be complex, calling subtemplates such as {{Wikidata entity link}} or Lua modules. However, the complexity is mainly aimed at display formatting; the pure data is easily accessible directly from Wikidata via tools such as SPARQL via the Wikidata Query Service (tutorial). Certes (talk) 11:15, 15 April 2022 (UTC)[reply]

@Certes Aside from embedding problem, do you agree with me that, till now, Wikipedia writers are not eager to fill Wikidata items? For example in Factorial Wikidata item, that is, https://www.wikidata.org/wiki/Q120976 there exists no item for «parity, date of invention, growth, approximation, its extensions, and related sequences and functions» that there exists in a unstructured way in the main article. We should use a mechanism to make users eager for filling these parts. Today these items are not filled, so we should alarm users fill these parts in Wikidata as complete as you can, to make the article machine readable.

Let me correct the above scenario, at each article, a machine readable infobox is shown from Wikidata to the user that some of its parts is not filled, and the users easily inspect them and make the machine readable infobox as fill as he can. That is we create a "machine infobox" different from "User infobox" that is hidden by default but users can view that, extract the information they need and fill empty parts, and possibly modify incorrect data. Hooman Mallahzadeh (talk) 11:49, 15 April 2022 (UTC)[reply]

Yes, Wikipedia writers are not eager to fill Wikidata items. But Wikipedia is not a database and, if it were, we wouldn't want to duplicate the information in every language with all the gaps and discrepancies. One way forward is to create a tool to scrape information from infoboxes of English and non-English wikipedias and add it to Wikidata with appropriate attribution. That way, Wikidata will be up to date for everyone, including any infobox designers who wish to draw on it for English Wikipedia. (Such a tool may already exist: this idea is unlikely to be original.) Certes (talk) 11:55, 15 April 2022 (UTC)[reply]

@Certes You are right, "Machine readable infobox" should have items with "label" and "data" parts. Label part is in English or is translated to other languages, but its data part should be some "id" or "url" (not in English).

If you agree with the benefits of creating a separate infobox for "machines" and embedding that machine readable infobox in articles of the same concept in different languages, please alert that to the developers, to implement that. Thanks, Hooman Mallahzadeh (talk) 12:05, 15 April 2022 (UTC)[reply]

@Blueboar@Certes@StarryGrandma As a conclusion, a sperate Machine readable infobox is created and added to all existing articles that its purpose is:

Convert «unstructured data» into «structured data» by that infobox, as much as you can.

Hooman Mallahzadeh (talk) 13:27, 15 April 2022 (UTC)[reply]

Meh. Adding a second infobox would just confuse our editors with no benefit for our readers. I don’t really care about the machines. Blueboar (talk) 13:35, 15 April 2022 (UTC)[reply]

@Blueboar No, it has many benefits for human readers. They know: "what parts of the infobox is free", and "what parts are wrong to modify that". Additionally a normal person (not a machine) may only want, for example, domain of the factorial function and nothing else, which exists only in the second infobox not the first infobox (after making it human readable by a simple process). Note that redirecting to Wikidata is a bad solution becasuse they are not in an infobox and is hard to read, also is not embedded to articles for machines, discussed above. Hooman Mallahzadeh (talk) 13:47, 15 April 2022 (UTC)[reply]

@Blueboar Note that second infobox is «hidden» by default, but the user can view that, so makes no confusion. Hooman Mallahzadeh (talk) 13:50, 15 April 2022 (UTC)[reply]

Um… The entire point of WP is that our editors can modify anything (as long as there is consensus to do so). This includes the information in infoboxes. Blueboar (talk) 14:09, 15 April 2022 (UTC)[reply]

Hi there. I'm not sure if this is the right place for this but I'll post anyways. A few months ago, I created a template in my sandbox. It takes a list of items and creates a list where the terms are bolded, but not the commas, for example: foo, bar, or 123 . The idea being, this saves editors from having to type '''foo''', '''bar''' etc. during source editing, or having to manually precisely select each comma to unbold in visual editor. An example where the template would be used is in article leads, where bolded alternate names are given. I haven't moved it to template namespace because 1) I don't know whether the template would be useful, 2) I don't know if a way to do this already exists, and 3) I don't know if I've implemented it efficiently, since I'm not very familiar with template code. Asking for thoughts on the template. Thanks! — Mcguy15 (talk, contribs) 23:05, 20 April 2022 (UTC)[reply]

@Mcguy15 That could probably be implemented for efficiently with a few lines code in a Lua module (and it would work for an unlimited number of list entries), something like:

return { main = function(frame)
	args = {}
	for i, v in ipairs(frame.args) do table.insert(args, v)	end
	conj = "'''" .. (#args > 2 and ", " or " ") .. (args.conj or "or") .. " '''"
	return "'''"..mw.text.listToText(args, "''', '''", conj).."'''"
end }

The code can be simpler if you don't need it to produce Oxford commas. --Ahecht (TALK
PAGE) 03:54, 23 April 2022 (UTC)[reply]

Wow, that's super cool, thanks! Definitely better than my bodged way of doing it. I'll try to see if I can figure out how add that to {{bold list}} later. — Mcguy15 (talk, contribs) 04:22, 23 April 2022 (UTC)[reply]

Oh! This was unarchived. I followed the advice of the person who archived this thread and created {{Bold list}}. — Mcguy15 (talk, contribs) 04:19, 23 April 2022 (UTC)[reply]

What are some ways we could improve the options available for editors to indicate they are open to going through an RfA? — Ixtal ⁂ (talk) 19:52, 21 April 2022 (UTC)[reply]

Change the name from RFA to VFA (Volunteers for Adminship). Like some other fields, the best candidates are characterized by "Willing to serve" rather than "I want it". North8000 (talk) 20:20, 21 April 2022 (UTC)[reply]

North8000 WOW! That's actually such a great idea I hadn't even considered it. I imagine people might be opposed to change just because it's not status quo but I think your suggestion makes a lot of sense. "Request" feels very uh can I pretty please have tools. — Ixtal ⁂ (talk) 22:06, 21 April 2022 (UTC)[reply]

That's a pretty brilliant reframing, North8000. Schazjmd (talk) 22:11, 21 April 2022 (UTC)[reply]

That renaming neatly moves the emphasis from gaining a privilege to taking on a duty, which might help to encourage good candidates. Certes (talk) 22:21, 21 April 2022 (UTC)[reply]

No thank you. Why? Well then we'd have to move all the current "Wikipedia:Requests for adminship/USERNAME" pages to "Wikipedia:Volunteers for adminship/USERNAME". And it isn't a problem worth solving. We were doing fine with RfA, we don't need to change it. We have promotional articles on this website, and here we're worrying about the name of a venue for discussion??? This just wastes time that could be used to like, improve the encyclopedia? If this was on the main proposals area this thing would be spammed with opposes, I bet you. Cranloa12n / talk / contribs / 00:28, 22 April 2022 (UTC)[reply]

CAT:RFA informs us that there have been roughly 5,000 RfAs throughout wikihistory, successful and unsuccessful. That's a lot for a single person to move around and maybe enough to just get a bot to do it, but hardly grounds for outrage on its own. Dr. Duh 🩺 (talk) 06:48, 22 April 2022 (UTC)[reply]

Per Dr. Duh. It's a matter of hours if done by a bot, which wouldn't waste anyone's time. From what I see in the latest draft of Administrators' newsletter, we are on track to losing 5 Admins to inactivity this month, and gaining only 1 new Admin. A net decrease of 4. So far in 2022, we've had a net loss of 4, 6 and 1 Admins per month. Back in the 2006–08 period there were an average of more than 1 RfA every day to replenish the ranks. Now it's about 2 every month. Administrators do a very important job in maintaining the project. *If* a name change or anything else has the potential to change the picture, we should seriously consider that. —CX Zoom[he/him] ^{(let's talk • C • L)} 07:55, 22 April 2022 (UTC)[reply]

No. Rfa will not become better just because we decide to change its name. It will still be the same vetting, the same opposes and supports. Again, if it ain't broke, don't fix it. Cranloa12n / talk / contribs / 13:05, 22 April 2022 (UTC)[reply]

Well, will it be worse, Cranloa12n? — Ixtal ⁂ (talk) 13:09, 22 April 2022 (UTC)[reply]

No, but it's a pointless change. Cranloa12n / talk / contribs / 13:11, 22 April 2022 (UTC)[reply]

Cranloa12n if you wish to oppose a future proposal feel free to do so, but IL is to draft ideas and brainstorm. Not very useful to discourage editors from an idea because you don't see benefit nor harm from the idea in my opinion. — Ixtal ⁂ (talk) 13:39, 22 April 2022 (UTC)[reply]

There is nothing to draft here. The proposal to change the name is simply redundant. Cranloa12n / talk / contribs / 13:41, 22 April 2022 (UTC)[reply]

User:Ahecht/Scripts/massmove could move them all in about 5 hours if you just pasted in the list and let it run. --Ahecht (TALK
PAGE) 03:13, 23 April 2022 (UTC)[reply]

Half a Scorsese movie, Ahecht! :D — Ixtal ⁂ (talk) 09:02, 23 April 2022 (UTC)[reply]

There's nothing in the idea that requires retrofitting history or renaming everything that happened in the past. Schazjmd (talk) 15:42, 22 April 2022 (UTC)[reply]

@Cranloa12n Why would we have to move past RFAs? ~ ONUnicorn^{(Talk|Contribs)}problem solving 20:20, 22 April 2022 (UTC)[reply]

It's hard to indicate you are open to an RfA without appearing to be a hat collector. I seem to recall one prominent editor had a list of things that would disqualify someone from being an admin, one of them was having a userbox such as {{User wikipedia/Administrator someday}} on their user page. --Ahecht (TALK
PAGE) 22:22, 21 April 2022 (UTC)[reply]

VfA could be either of two things. It could answer Ixtal's question by being a better precursor to RfA than the userbox. Alternatively, it could be a rebranding of RfA, which might solve an important related problem. Certes (talk) 23:19, 21 April 2022 (UTC)[reply]

I like the idea of the whole process being rebranded. Even though it doesn't (of itself) change anything (yet) about how admin permissions are assigned, that one word difference in the name casts it in a different perspective. And maybe that different perspective might lead to more of us being more open to other improvements to the process. Schazjmd (talk) 23:34, 21 April 2022 (UTC)[reply]

In the dark ages, people who were ready for adminship could just self-nominate and pass RfA. —Kusma (talk) 07:02, 22 April 2022 (UTC)[reply]

Some editors are in the odd position that they could have passed RfA 5–10 years ago but, despite continuing to improve gradually, wouldn't now. I'd like to think that is due to increasing numbers of better-qualified candidates, but the admin count doesn't support that theory. Certes (talk) 20:06, 22 April 2022 (UTC)[reply]

My theory is that some highly qualified non-admins (who should probably be admins) subconsciously apply the "I am better than the candidate and I am not an admin" standard in the wrong way: instead of running for RfA, they make it hard for others to pass. —Kusma (talk) 08:41, 23 April 2022 (UTC)[reply]

The "unsolvable" no-new-admin problem could be solved easily by three relatively easy changes:

• Something like that above that shifts the ground so that it makes it clear that volunteers are saying "I'm willing to serve" rather than "I want it"
• Organize RFA/VFA discussions to shift discussions more towards reviewing for needed qualities rather than the random gauntlet that we have now
• Strengthen the advice that is already weakly in admin guidance that it's expected that only more experienced admins should be doing the really heavy duty stuff like blocking experienced editors. Without that, the mop really IS a big deal and so reviewers place a very high bar accordingly.

North8000 (talk) 01:22, 22 April 2022 (UTC)[reply]

Given how hard we've been trying to fix RfA for a decade now, I feel like there's no way the fix is going to be easy. The main problem of RfA recruitment at the moment is not that editors are "willing to serve" but afraid of looking like they "want it"—rather, the problem is precisely that many qualified editors are not "willing to serve" in the first place because of the intense and stressful scrutiny you have to endure at RfA in order to serve. It seems like your second bullet point is trying to offer a solution to that, but at the moment, it's a little vague. The English Wikipedia currently has no official standards for what qualities are needed to be an administrator—RfA right now is mostly a free-for-all of editors applying their own views of what the necessary qualifications should be. Defining the necessary qualifications more rigorously does not strike me as an easy task—I suspect the RfC will result in no consensus, at best. Anyway, to answer the question presented by the thread, if you are interested in becoming an administrator, the best way to indicate that interest at the moment is to reach out to someone you trust on this list (either on their user talk page or by email) and start a conversation. Mz7 (talk) 08:28, 22 April 2022 (UTC)[reply]

Mz7 I wonder if you could also contact people in that list at random? It might be intimidating having to "pick" one (or a few) of the names on the list. ORCP has a similar vibe but also has some hat collector stigma (at least if you read some of the comments on leeky's recent RfA). — Ixtal ⁂ (talk) 11:13, 22 April 2022 (UTC)[reply]

@Ixtal: Hmm, yeah, I can definitely see how it can be a little intimidating. My recommendation would be to see if there's anyone on the list that you've had good interactions with—even if only once or twice. Alternatively, if you've never interacted with any of them, try to see if there's someone on the list that you've seen around and have a generally good impression of. I would also recommend reaching out over email. Messaging privately could make it easier for the person to provide more candid feedback, and even if they don't offer to nominate you now, this would put you on their radar for the future. Mz7 (talk) 00:41, 23 April 2022 (UTC)[reply]

Oh Mz7 I don't plan to run, hope my comments above didn't lead to that interpretation. I still have a few years of learning and improving to do, and adminship is not necessary for me to contribute productively. I know for me at least there's a few people in the list who I interact frequently with so I'd ask them, but other editors that are not as big of a metapedian as I am might be intimidated by all the big names. — Ixtal ⁂ (talk) 06:41, 23 April 2022 (UTC)[reply]

Mz7 All good points but IMO so are mine. BTW, I was not thinking of trying to define qualifications and turn RFA/VFA into an RFC. I was imagining a softer change. Just list some needed qualities and provide suggestions to comment on those overall. The real test case isn't recent RFA's which is mostly people who have 100% kept their head low their whole wiki-life. The real test case will be someone who has been experienced and active and has not kept their head low. If we ever get one..... North8000 (talk) 13:13, 22 April 2022 (UTC)[reply]

I disagree that recent successful candidates have kept their heads low their whole wiki-life. Looking at the last five successful candidates, I think that's somewhat fair for 2 of them - Colin and Modus. But Blablubbs is/was all over SPI, Firefly is/was an arbcom clerk, and Sdrqaz has never been shy about expressing their opinions. I am not sure what conclusion to draw from the "near unanimous support" or "withdraw" that we've seen in the last 8 or so months but I think it's a bit too glib to just say they've kept their heads low and that's what distinguishes them from those who didn't pass. FWIW, the idea of having formal qualities for admin was discussed a fair amount in WP:RFA2021 and ran into some real opposition I think you'd find that difficult to pass. The RFA/VFA piece is the most intriguing thought here but also feels like something that would have an impact at the margins, to the extent it actually does any change, in terms of how the reframe would impact voting.

If people want to change RfA, the idea of temporary adminship and admin elections both were closed at RFA2021 as having the chance to pass in a modified form. I hope someone would think about trying to get one of those ideas across the finish line. Best, Barkeep49 (talk) 01:05, 23 April 2022 (UTC)[reply]

Moved from WP:VPPR

Ping specific user groups, for example: @Administrators (pings a random administrator, in case administrator assistance is needed on a specific page) Viewer⁷¹⁹/_Contribs! 10:03, 22 April 2022 (UTC)[reply]

Pinging a random administrator won't necessarily help much. You might ping an administrator who isn't very active, for example, or somebody who doesn't work with that particular area. You'd be better off posting something on the admin noticeboard, posting something on a forum best suited to the topic (e.g. if you need a page moved then try WP:RM/TR), or using the recently active admins tool. Hut 8.5 11:53, 22 April 2022 (UTC)[reply]

Plus, there is already the {{adminhelp}} template to alert any admins monitoring that category. Regards SoWhy 12:09, 22 April 2022 (UTC)[reply]

I was recently thinking about the feasibility of a template that pings whoever's the most recently active admin, both from a technical perspective and a "is this a good idea" perspective. Nice that I stumbled over here at the time I did. For one, this is definitely the sort of thing that would require community approval - and probably moving User:Enterprisey/recently-active-opt-out.json to projectspace. For two, the current list uses API calls - Templates can really only work with stuff in the wikitext, though Modules do some ~other stuff~ which I can't comprehend even remotely.

All this is made moot by just.. going to the page and pinging the unlucky mop-wielder at the top? casualdejekyll 19:14, 22 April 2022 (UTC)[reply]

Sending an echo notification to a random user would almost always send to a completely inactive user. — xaosflux ^Talk 12:22, 22 April 2022 (UTC)[reply]

 Administrator note @Viewer719: I moved this from VPR to VPI as it isn't really an actionable proposal as is, and would get shut down there. The idea and other possible work arounds can be further disucssed here, maybe some other ideas will come up. You mentioned a specific technical process - but can you explain what it is you would want to accomplish? — xaosflux ^Talk 12:24, 22 April 2022 (UTC)[reply]

We already have some ping templates for certain groups of specialists (DYK admins or FA coordinators). —Kusma (talk) 12:29, 22 April 2022 (UTC)[reply]

See Special:PrefixIndex/Template:@ for examples. —Kusma (talk) 12:34, 22 April 2022 (UTC)[reply]

Hi! I want to make an RfC suggesting that vote counts for ongoing requests not be displayed at Wikipedia:Requests for adminship as it influences editors before reading the nominations and questions, and similarly propose that the tally included at the top of every RfX be added below the Q&A of the candidate (as opposed to the top, as is currently done). How can I best word such an RfC? Should two RfCs be made rather than one, and if so should they happen concurrently? — Ixtal ⁂ (talk) 09:22, 23 April 2022 (UTC)[reply]

Page version of last RfA showing what I mean. — Ixtal ⁂ (talk) 09:24, 23 April 2022 (UTC)[reply]

Honestly, I don't see how this will accomplish much. People react to the content of other people's votes, not to where the tally is. If you want the tally to have no great effect, you need secret voting. —Kusma (talk) 09:34, 23 April 2022 (UTC)[reply]

I don't think it is some ground-shaking proposal that will completely change RfX, but I do think that it can have a strong subconscious influence on editors. Yes, editors will still be affected by the tally when they vote, but I think they should at least read the nomination and Q&A with as little influence from groupthink as we can. It's a small change, really. But I think it's a positive one that doesn't take much effort to implement so I want to get the community's opinion on this. I'd just like some help with wording the proposal before I bring it up ^u^ — Ixtal ⁂ (talk) 09:56, 23 April 2022 (UTC)[reply]

On the other hand, showing the vote totals helps respect editors' time. Often, I'll glance at an RFA, see that my input isn't needed, and save myself some time and effort. ScottishFinnishRadish (talk) 13:21, 23 April 2022 (UTC)[reply]

ScottishFinnishRadish by vote totals do you mean a count of all votes made or the current division of S/O/N? — Ixtal ⁂ (talk) 14:38, 23 April 2022 (UTC)[reply]

Both, as well as the time left in running. Ratio is important, as well as the total votes. I prefer to look towards the end to read both sets of arguments and see if I have anything constructive to add or say, which I normally don't. ScottishFinnishRadish (talk) 15:04, 23 April 2022 (UTC)[reply]

I don't think removing the "remaining time" is a good idea. Some editors are only on every few days and it will help them know if they have time to come back to it. — xaosflux ^Talk 15:21, 23 April 2022 (UTC)[reply]