Wikipedia talk:Arbitration Committee/Requests for comment/AfD at scale

Article creations at scale

For the RfC on article creation at scale, please discuss at Wikipedia talk:Arbitration Committee/Requests for comment/Article creation at scale

Paraphased deletion issues and possible solutions

This is a scratch pad. I've paraphrased the deletion-related issues and possible solutions from Wikipedia talk:Arbitration Committee/Requests for comment/Article creation at scale/Archive 1.

• 3) mass nominations for deletion become train wrecks because editors drill down further by taking different positions on the individual articles nominated
  • possible solutions:
    • define threshold for bulk nomination (3.1)
    • bulk nominations should be based on a logical set of criteria, rather than a set of articles (3.2, 3.3)
      • a venue should be created to help come up with appropriate criteria before nominations (3.2)
    • don't allow piecemeal votes, all-or-nothiung (3.1)
    • keeping a bulk nomination will not prevent a re-nomination with a refined criteria (3.1)

• 5) simultaneously permitting creations based on SNG and deletions based on GNG will cause an endless stream of AfDs and conflict
  • possible solution: harmonize the criteria, SNGs not in accordance with GNG should not be used for mass creation

• 7) presumed notability and what that allows is not consistently interpreted, which causes disagreement
  • possible solution: clarify wp:notability to be more explicit when an SNG is supplements, supplants, or is prohibited

• 13) NSPORT actually requires the subjects of articles meet GNG as well as SSG
• 13a some participations are demanding onerous BEFORE searches despite missing SIGCOV
  • possible solution below 19
• 13b editors who disagree with the NSPORTS2022 consensus are taking anti-consensus positions and muddying the waters
• 13c editors are adding or quoting noncompliant sources in an attempt to keep articles that don't have SIRS sourcing
• 13d editors who disagree with the NSPORTS2022 consensus are taking the fight to DRV in numbers

• 14) absolutist approach to deletion is unhelpful, redirection, merging, and disambiguation are other options
  • possible solution: alternatives to deletion should be engaged first and those who refuse may be topic banned from nominations

• 16) forcing all articles through a strict interpretation of GNG rewrites the notability policy; not all SNGs require GNG to be met

• 19) article creators are not required to do an exhaustive search for sources during creation, which perversely puts the burden on nominators
  • possible solution (also to 13a): modify BEFORE to make it clearly a suggestion rather than a requirement and explain the BURDEN rests upon the person(s) wishing to retain the content

Next step to create a set of questions like this for consideration. –xeno^talk 17:52, 10 September 2022 (UTC)

My replies a bit late but another issue that comes out of 14) is the redirection of articles is reverted without first ensuring that the articles meet GNG/SNG, thereby forcing an AfD discussion. It would help the issues at AfD if there was some clarification about the interaction of WP:BURDEN and undoing redirection of articles not meeting a notability guideline. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 02:09, 13 October 2022 (UTC)

Questions transferred from WP:ACAS

Question 7: Should we adopt a new speedy deletion criterion that relates to mass-created articles that lack any sourced claims of importance?

A12: No reliably sourced indication of importance (mass-created articles).

This criterion applies to any mass-created article that does not have a reliably sourced indication of importance. This would apply to any mass-created article that does not indicate why its subject is important or significant. This is a lower standard than notability. If the sourced claim's importance or significance is unclear, you can improve the article yourself, propose deletion, or list the article at articles for deletion.

Question 7A:Should we instead introduce the following speedy deletion criterion:

A12: Unsourced or obviously unreliably sourced mass-created articles.

This criterion applies if the mass-created article has been unsourced in all of its history, or is only sourced to obviously unreliable or deprecated sources in all of its history.

Question 10: add mass-creation as a reason to WP:BUNDLE

Should we add mass-creation of articles by the same editor using substantially the same sources and format as a reason for bundling multiple articles into a single AFD at WP:BUNDLE?

Timeline for workshopping RfC on article deletion at scale

@Xeno and I have been discussing whether starting workshopping of the next RfC is helpful once input at WP:ACAS has significantly slowed. We see pros and cons, but we don't expect to be able to see all of them ourselves, nor to think of every possible consequence. So if there are opinions, let's have them.

One pro I see is getting to the end of this in our lifetimes. There's a good chance we're going to run into ArbCom elections and various time-and-energy consuming holidays. If we can actually get everything accomplished before that, we don't have to worry about whether we should instead delay something (probably the deletion-at-scale RfC) until maybe January. Valereee (talk) 17:57, 7 October 2022 (UTC)

• I'd encourage you to proceed. I agree that it's important to get it done and feel the concern that we might lose momentum.—S Marshall T/C 22:25, 7 October 2022 (UTC)
• I'd be surprised now if anything positive of significance comes out of the article creation at scale RfC; it would be a pity to delay the next one because of it. Espresso Addict (talk) 02:25, 9 October 2022 (UTC)

  Yep. I think the community is showing us pretty clearly that they don't want bad mass creation fixed. But giving that a chance was the point of the RfC, so if that's what ends up happening, it's still productive. No one will be able to argue that we didn't give the community a chance to fix this major underlying problem. We asked if the community could find a solution to bad mass creation, and the answer may simply be "no". Valereee (talk) 14:59, 9 October 2022 (UTC)

  "I think the community is showing us pretty clearly that they don't want..." Another word for this phenomenon is "consensus." Gnomingstuff (talk) 02:51, 30 October 2022 (UTC)

  Sorry, what's the difference you're perceiving? Valereee (talk) 20:48, 1 November 2022 (UTC)

  If you are going to defer to consensus, then you have to actually defer to consensus. If the community does not see a problem with article creation at scale, then the opinion that it is "bad mass creation" is anti-consensus. You can, of course, take the position that the consensus is wrong -- but then you cannot also criticize others for disagreeing with consensus in other areas. Gnomingstuff (talk) 06:39, 6 November 2022 (UTC)

  I think the community (well, the portion that happens to be participating in the discussion) isn't able to agree if certain modes of article creation are undesirable, which is not the same as saying it doesn't want to fix "bad mass creation", for some meaning of the word bad. I concur though that the present state might be that no solution can be agreed upon. isaacl (talk) 23:39, 30 October 2022 (UTC)

Format for workshopping RfC on article deletion at scale

There've been a lot of complaints about the formatting of the current RfC. I'm open to suggestions for formatting the workshopping for the AfD RfC and seeing how it goes. So if there are opinions, let's have them. Valereee (talk) 14:51, 8 October 2022 (UTC)

• I suspect I made your life harder at the current RfC by leaving too many comments, so let me stick to a simplifying one this time. I think it's in your interest to either have a workshopping session for the questions being presented, or to allow people to add their own proposals. As this stand, you spent a lot of effort trying to hammer out consensus proposals, but given how divergent the views are on this topic, many others are being added anyway, and the "synthesized" ones aren't doing particularly well. Vanamonde (Talk) 22:03, 8 October 2022 (UTC)

  You didn't make things harder! I do want comments, I just don't want long pages of back-and-forth bickering. :)

  Yes, I have to agree that the workshopping doesn't really seem to have helped. Maybe we should just skip it for the deletion-at-scale RfC? I don't want to limit proposals to what is workshopped. I feel like it leaves too many people out of the process, as many won't show up for the workshop but will for the RfC. Which maybe is why the distilled questions aren't being supported -- the workshop was likely full of experts who care deeply. The RfC of course attracts drive-by voters, but we can't do anything about that. Valereee (talk) 14:47, 9 October 2022 (UTC)

  My thoughts exactly. If folks are looking for feedback on proposals before submitting them, you could possibly set up a sub-page for that, but one which isn't compulsory and isn't bound by the need to advertise everything and have a set timeline. Vanamonde (Talk) 15:36, 9 October 2022 (UTC)

  With the added benefit of making things easier for moi. Valereee (talk) 22:19, 9 October 2022 (UTC)

• I'd be inclined to allow polite threaded responses to comments (but not votes) and ditch the word count limit. This comment-only-in-your-own-section format is extremely hard to follow, pings keep failing because people are not signing to keep the word count down, and generally it's stifling positive debate (possibly as well as acrimonious fixed differences of opinion). It would probably mean stricter moderation would be necessary. Espresso Addict (talk) 02:23, 9 October 2022 (UTC)

• I'd can see the point of some threading for comments only, but given experiences with other RfC which have effectively been bludgeoned by a small number of editors, some sort of word limit seems essential to me. Maybe 300 is too low. Signatures should be excluded from word limits if we think they're causing problems. Blue Square Thing (talk) 07:28, 9 October 2022 (UTC)

  Perhaps just moving any discussion that involves more than X exchanges to the talk page? There's really no reason the RfC itself needs long arguments which could be done at the talk. The closers don't need to see the long exchanges between people trying to convince other !voters or bludgeon a discussion. They only need to see the !votes and reasonings for those votes. If people want to have long convos on the talk, I have no opinion on it. Valereee (talk) 14:52, 9 October 2022 (UTC)

• If you guys are really on top of things, that might work, but I'd be really concerned that the deletion RfD is going to be rather different to the creation one - more "at stake" (yes, I know...). The way the Sports RfC ended up earlier this year, as an impenetrable wall of text that a actively put users off even trying to contribute (and that's before you consider the page stats - 50% of edits after it was moved to it's own page from three users...). For that reason alone I'd like to see the word limits kept. Blue Square Thing (talk) 18:33, 9 October 2022 (UTC)

  I'm happy to deputize mods to move back-and-forths to talk during the hours I typically am less likely to be editing (4pm-4am Eastern). :D Valereee (talk) 18:36, 9 October 2022 (UTC)

• Just let us talk in the normal Wikipedian way. This isn't generating the level of acrimony that necessitates the "post only in your own section" approach.—S Marshall T/C 19:23, 9 October 2022 (UTC)

  And if the reason it's not generating that level of acrimony is that we aren't talking in the normal Wikipedian way? Let me be clear: I have no issue with the "normal" Wikipedian way. I love discussion and participate often, sometimes at length and sometimes multiple times within a single discussion. But that normal way sometimes becomes hundreds of thousands of words with back-and-forth bickering that goes on for days, bludgeoning that discourages people from participating, and walls of text that no one wants to wade through. That's the point of unthreaded discussion and of word limits, and it would be the point of moving back-and-forth exchanges to the talk page after X exchanges. Valereee (talk) 20:12, 9 October 2022 (UTC)

  Hmm. I can't tell if that's a decision from on high or point of view for discussion. Would further conversation about this be unwelcome?—S Marshall T/C 21:07, 9 October 2022 (UTC)

  Not at all! All discussion is welcome. Not sure what you mean by a decision from on high? If you mean a moderator decision, any decision I've made is completely open for discussion. Any decision I'm discussing making is open for discussion. I prefer it not be snark directed at me, but even then I try to just ignore it. :D Although sometimes I write really cutting ripostes and leave them sitting unposted overnight so that with the new light of day and a cup of coffee, they make me laugh and appreciate my own brilliance all over again. Valereee (talk) 22:13, 9 October 2022 (UTC)

  Okay, good. Text isn't a good medium for judging others' mood. I find myself very much in Espresso Addict's camp here, frustrated by the difficulty of holding conversations with WhatamIdoing and JoelleJay when the text isn't organized in its sequence. I agree that very long conversations if they occur would be best moved to a subpage or talk page, and then summarized back on the main page when and if they reach a conclusion. I also see why any acrimony would need intervention by a moderator. But, I feel that these are still theoretical or hypothetical concerns, while the communication difficulties that Espresso Addict sums up so well are happening right now.—S Marshall T/C 23:06, 9 October 2022 (UTC)

  In general you'll probably find my mood "able to find humor in most situations, generally assuming good faith, willing to flex when appropriate."

  Are you proposing that we change the format for the current RfC? Valereee (talk) 23:27, 9 October 2022 (UTC)

  Agree with S Marshall; perhaps run the 2nd a little like an RfA? Long back-and-forth discussions get moved to talk page, acrimonious comments get actively moderated. Honestly not sure if/how the 1st can be salvaged now. Espresso Addict (talk) 00:25, 10 October 2022 (UTC)

  Once again I concur with Espresso Addict in every respect.—S Marshall T/C 12:35, 10 October 2022 (UTC)

  So, w/re: running it like an RfA. So each proposal would have its own discussion section, with threaded discussion? Participants open a new subject section for each new conversation, which could be multiple discussions for a given proposal?

  I'm just sitting here thinking about logistics as a moderator, I guess...in an RfA, it's pretty easy to tell when discussion needs to move. It's almost always direct responses to an oppose, for one thing, that goes on after any needed clarification has been made. It's almost always a concern about piling onto opposers. Transferring that to an RfC that could have a couple dozen proposals with multiple concerns about each means...well, as someone mentioned, it's a lot to moderate. One of the upsides of unthreaded discussion is that the moderator doesn't have to moderate a 200,000 word discussion. The fact communication is more difficult is feature as well as bug: if you have something important enough to say, you'll say it even in an unthreaded convo. But if you're sitting there refreshing the convo so you can fire off another wall of text at the person who is pissing you off...ai yi yi. Maybe limit the number of responses and words any editor could add to any discussion? That's a moderation nightmare, that is, I am definitely not up for that kind of tedium. I'd literally end up mostly just responding to people ratting each other out.

  Honestly, I don't know what the answer is. I know everyone hates unthreaded discussion. But from the point of view of the person who is supposed to be keeping it from becoming an impenetrable wall of text, I'm just not sure they aren't the lesser of two evils. Valereee (talk) 13:20, 10 October 2022 (UTC)

  Yes, that would be my suggestion. Each proposal has a section for threaded discussion, but it's not permitted to respond to supports or opposes directly. So say EA proposes all AfDs would go better with doughnuts, and XY opposes on grounds that sugar gets on the keys, EA/XY/anyone else may open a single discussion thread to discuss whether or not doughnut sugar is a real problem. Perhaps each thread should be given a descriptive subheading? All subsequent discussion on this topic gets moved to the thread. Discussion that goes round in circles, degenerates into "mmm doughnuts", or digresses into the spelling of doughnut or the character of doughnut-loving EA is either deleted or moved to the talk page. Espresso Addict (talk) 00:15, 11 October 2022 (UTC)

  @S Marshall, I'm thinking the discussion with JoelleJay and WhatamIdoing would have definitely been better here on this talk. JJ opened them for discussion there, possibly to make sure others did actually see them, and I decided to see how it went. Should they be moved here to allow for a long threaded discussion? Valereee (talk) 16:41, 10 October 2022 (UTC)

  Sure we can move it here. JoelleJay (talk) 17:41, 10 October 2022 (UTC)

  Thank you for your flexibility, JoelleJay. Valereee (talk) 19:58, 10 October 2022 (UTC)

  Oh JFC. Just ridiculous...I forgot where we were. We're at the talk page for the AfD RfC. I need a drink. Valereee (talk) 19:59, 10 October 2022 (UTC)

  @Valereee, I thought about switching to the talk page at one point, but then didn't. I think this is a good model, and that what's needed is a note (even though Wikipedia:Nobody reads the directions) that points out that when you find yourself replying to someone who was replying to you, then it's probably time to copy that to the talk page.

  Any/all of my comments can be moved to any place that makes you happy. WhatamIdoing (talk) 20:07, 10 October 2022 (UTC)

  Nothing is making me happy with myself right this minutes lol...I am however warming some sake and intend to sit in front of the fire and watch Derry Girls with a cat in my lap. This will make me...well, a lot less unhappy with myself. Valereee (talk) 20:10, 10 October 2022 (UTC)

  There's nothing quite like volunteering to shepherd a massive RFC to make you wonder about your life choices.  ;-) WhatamIdoing (talk) 20:38, 10 October 2022 (UTC)

• @Valereee: What about adding an "Evidence" section at the top, adjacent to the statistics, where short items of background information can be added by consensus on this talk page? Ovinus (talk) 06:36, 19 October 2022 (UTC)

  Hm. It sounds reasonable. I like the idea of requiring such a section to be vetted by others here first to make sure it doesn't get out of control. Valereee (talk) 11:08, 19 October 2022 (UTC)

  I've added that at the 2nd RfC. Valereee (talk) 11:11, 19 October 2022 (UTC)

For ease of navigation

I've posted a proposed format at WP:ADAS, feel free to comment here. Valereee (talk) 17:18, 18 October 2022 (UTC)

Thanks, Valereee. Not looked in detail yet, but quick query, do we have any stats on deletion? Espresso Addict (talk) 03:03, 19 October 2022 (UTC)

Not as far as I'm aware. What kinds of statistics would be helpful? Pinging @BilledMammal as they did the last set. Valereee (talk) 10:59, 19 October 2022 (UTC)

Got to go offline (busy day tomorrow) but off the top of my head,

• average number of AfDs per [time unit];
• average number of prods per [time unit];
• proportion of AfDs that result in delete/keep/no consensus/other outcomes;
• proportion of AfDs that are relisted (once, twice, thrice);
• average/range in number of distinct participants in AfDs (preferably excluding those who purely do indexing);
• total pool of editors who (frequently) participate in AfDs;
• number of editors who (frequently) challenge prods;
• number of admins who (frequently) close AfDs;
• number of editors who bring more than [threshold] AfDs in [time period];
• number of editors who bring more than [threshold] prods in [time period];
• number of editors with >[threshold] AfDs in the past year whose AfDs end in (say) >2/3 keep;
• number of editors with >[threshold] prods in the past year >50% of whose prods are contested.

That turned out to be a lot, and I've no idea how feasible any of these queries are, sorry. Any thoughts on how these figures have changed over time would also be welcome; I feel like we're talking in a bit of a vacuum here. Espresso Addict (talk) 00:42, 20 October 2022 (UTC)

No response to this.

If no data are available re deletion, then I think we should delete all the evidence relating to mass creation. It tends to bias the discussion towards repeating the other RfC, which we (desperately) want to avoid. Espresso Addict (talk) 10:31, 2 November 2022 (UTC)

I don't know, EA. If mass creation really is the heart of the problem, then we'll have to discuss it. The fact we couldn't solve it as a contributing issue surely doesn't mean we wouldn't still need the evidence? Valereee (talk) 10:45, 2 November 2022 (UTC)

We certainly can't post a bunch of evidence wholly about creation and no evidence about deletion! We're just asking for a repeat of the other RfC, which -- not to prejudice the unfortunate closers -- seems to me personally to have been a colossal waste of time and effort. As the original quarry-querier seems interested largely in the creation side of things, could someone else be found? Espresso Addict (talk) 10:50, 2 November 2022 (UTC)

Database discussion

Moved to Wikipedia talk:Arbitration Committee/Requests for comment/Article creation at scale/Archive 4#Database discussion

Deletion

Would anyone like to discuss deletion? I have multiple concerns relating to the way in which prods & AfDs work at present but was injured during the Arbcom case that led to all this and haven't the energy to read the walls of text, so not sure what's in scope. Espresso Addict (talk) 22:41, 14 October 2022 (UTC)

I imagine it might be useful to understand the sorts of concerns you have and to figure out if they apply to this or not. Blue Square Thing (talk) 06:18, 15 October 2022 (UTC)

Going back & looking at the ArbCom request, it states "comment on how to handle mass nominations at Articles for Deletion", so would proposed deletion be excluded? I recall a suggestion for a new speedy category was punted to this discussion but seems out of scope?

My fundamental concerns are pretty much the opposite of what a lot of editors have been expressing in my hearing. It's late here, so forgive me for lack of eloquence:

• it is a lot easier to propose an article for deletion than to create (a decent) one;
• it is also much easier to vote for deletion in AfD than to try to find sources or amend the article in response to the demands of those who call for deletion;
• AfD is becoming a very unfriendly place for people who don't like seeing articles deleted unnecessarily, leading to it being dominated by those who don't find this painful;
• proposed deletion in particular appears very underpatrolled; as far as I'm aware the admins who perform ~all of the deletions no longer seem to check out the articles for themselves (as used to be the case back in 2007 when I was working on processing prods).

And probably a lot more, but it's late. Espresso Addict (talk) 09:27, 15 October 2022 (UTC)

I agree with your first two points. I have no opinion regarding your third; I don't find AfD particularly unpleasant, but I think I interact with that area less than I used to, so my current feelings might not even be representative of how I would feel if I tried participating more. I have no useful input regarding your fourth point, either, but I would be unsurprised if that turned out to be true. XOR'easter (talk) 19:38, 15 October 2022 (UTC)

On the other hand, it's a lot easier to create an article of any quality or merit than it is to get that article deleted at AfD. All it takes is at most three regular editors (creator, AfC reviewer, NPPer), and more often just one, with no minimum time delay for creation. Whereas AfD always requires at least two editors, including an admin, for soft del and some higher number for full del, and must be discussed for at least 7 days. The per-editor effort load may be smaller for an AfD, but it still necessitates far greater overall time and effort and way more editors than creation. It's also not just keep !voters who are trying to find sources; in fact, before the WP:NSPORTS2022 RfC, delete !voters were pretty much the only editors looking for sources on sportspeople at AfD, since they had to prove a subject didn't have coverage while keep !voters only needed to state they met a sport-specific subguideline. I think one's experience at AfD is really dependent on which area the subject is in. JoelleJay (talk) 02:23, 16 October 2022 (UTC)

JoelleJay To prod an article takes writing "substprod" at the top and optionally notifying the creator, who has often retired. AfD is a little more time-consuming but with automated editing is just one or two clicks away. I've often spent literally weeks or longer writing a single article, sometimes with lots of trips to the library, trekking round with a camera, money spent on purchasing books, and the like, and bothering folk at the Resource Exchange board or wikiprojects for sources.

"I think one's experience at AfD is really dependent on which area the subject is in." This is becoming increasingly obvious, yes. Espresso Addict (talk) 03:10, 16 October 2022 (UTC)

Yeah I don't dispute that PRODs take less effort than most article creations. And by "automated editing" do you mean Twinkle et al? You still need to provide a DELREASON, and are "expected" to do a BEFORE, but sure, nominating can also be almost as trivial. My initial response to what you said about all the time spent writing a well-referenced article was going to be, "well, those aren't the articles that are getting AfD'd". But I'm guessing what you mean is that a stub at AfD could have just as much coverage in just as inaccessible sources as the topics you've worked hard to write good articles on, and that possibility is what is upsetting about AfDs being much less effort in comparison. My opinion is that there's a reason why BEFORE isn't more extreme, and it's tied to the balance between comprehensiveness and WP:NOT as well as to equitable distribution of burden and WP:NORUSH. I don't believe an un(der)-sourced stub on a topic that might be notable is any different from a couple mentions of the topic that might be DUE within another article; the only distinction is that one involved clicking "publish page" and the other "publish changes". Either way, deletion of that content doesn't affect whether the subject actually is notable, and with so little info in the original stub (even without the option of REFUND) it's not like recreating the article will take more than epsilon extra effort should someone have source access in the future. It just might take a little longer for someone to notice WP coverage of the subject is lacking if they don't come across the topic as a link. JoelleJay (talk) 04:42, 16 October 2022 (UTC)

• What's clear to me from those remarks is that EA hasn't recently tried to delete a poorly sourced BLP about a sportsperson. The amount of effort required to do that massively exceeds the effort put in by most sports BLP creators.—S Marshall T/C 08:23, 16 October 2022 (UTC)

Mostly I delete speedies; pretty low effort! But indeed, I don't work much on sports bios. Espresso Addict (talk) 10:20, 16 October 2022 (UTC)

OK. Here is where JoelleJay and I agree. My last AfD nomination of a sports BLP was this one. I've now completely given up trying to delete them: it's impossible.—S Marshall T/C 12:34, 16 October 2022 (UTC)

That looks like an argument primarily based on participation only in a single event? I'm not sure it tracks with the run-of-the-mill sports bios; there's a 1200-word article with 41 references. Espresso Addict (talk) 22:24, 16 October 2022 (UTC)

Yeah that AfD is not the norm for sports bios, which nowadays are primarily the editors who relied on "meets NFOOTY" pre-RfC now attempting to claim GNG is met from routine sports recaps, transactional news, blogs, and press releases; or that we must keep every old or non-Western athlete based on IAR. This, this, this, and this are more representative of the headache we face whenever an athlete who would have met a pre-RfC guideline is taken to AfD. JoelleJay (talk) 22:47, 16 October 2022 (UTC)

Thanks for these examples; looks like they all ended up with delete. I think the AfD discussions exemplify my problem with GNG in practice, which is that one person's significant coverage is another person's passing mention/routine coverage, together with all the arguments over reliability and independence of sources. Espresso Addict (talk) 23:07, 16 October 2022 (UTC)

There are also sports AfD like this, this, this or this where there appears to be a different approach being taken and where, arguably, the nominator is rather in error and might have been better asking someone about the article subject first. But we'll bundle all these and delete them, yes? Blue Square Thing (talk) 16:30, 18 October 2022 (UTC)

While they closed as delete, there are many others that weren't despite failing all notability guidelines (see BilledMammal's evidence in the ArbCom case; also this AfD that was kept on the basis of a deprecated argument and only deleted once editors pointed this and the poor sourcing out in the renomination). Lots of deleted articles also could very easily have been kept if no one had bothered to assess refbombed sources. With sportsperson bios we actually have pretty clear standards on what can't contribute to GNG: ROUTINE match reports, stats, transactional coverage. JoelleJay (talk) 04:51, 19 October 2022 (UTC)

It does strike me that in many cases there are ATD which can be used - perhaps not this one, but the close specifically mentions that there's no consensus to entirely delete. Given the difficulties that I had using REFUND earlier this year, I would suggest that much stronger use of ATD and ATC would probably be a helpful way forward. Blue Square Thing (talk) 09:02, 17 October 2022 (UTC)

Blue Square Thing: Alternatives to deletion would be one measure that we could debate; moving a mass-created set of stubs to draft and encouraging interested editors to improve them individually would be a useful possibility. Sorry, not familiar with the acryonym ATC? Espresso Addict (talk) 01:18, 18 October 2022 (UTC)

moving a mass-created set of stubs to draft and encouraging interested editors to improve them individually would be a useful possibility I would support that, with them being allowed to be individually returned to main space if improved, or returned as a group if there is a consensus approving their mass creation. I would suggest something like:

Articles that were mass-created without consensus are subject to mass-draftification. This is done through a discussion at WP:VPR, where any editor may propose either a list of articles to be draftified, or a quarry query that produces a list of articles. This proposal must meet the following criteria:

1. The articles are limited to a single creator
2. The articles are limited to a single topic area, broadly defined
3. The articles are limited to mass created articles with few (<5%) false positives. "False positives" are articles that were not mass created, or were mass created but later expanded.

Arguments for or against approval of the mass-draftification on grounds other than whether they meet or do not meet this criteria must be ignored by the closer.

Articles may be restored to article space under one of two circumstances:

1. Individually, when they can no longer be considered mass created; this includes both false positives, and articles that were expanded after draftification.
2. As a group, if there is a consensus to retroactively approve the mass-creation of the articles.

few (<5%) false positives in criteria #3 is intended to balance the fact that for some editors who engaged in mass-creation (for example, Lugnuts) a manually curated list is not plausible, and so some articles will be incorrectly included (and less problematically, some articles will be incorrectly excluded). This is balanced by restoration circumstance #1, which allows those articles to be returned to article space without issue. BilledMammal (talk) 02:12, 18 October 2022 (UTC)

I doubt I'd support "are subject", but "may be subject" might be acceptable. It's a pity that G13 means drafting is essentially a six-month-delayed prod but that's an argument for a different forum; perhaps they could be moved to the WP space of a relevant wikiproject instead? I dislike the idea of using the Village Pump, largely because content contributers rarely hang out there; you are far more likely to get content contributors by discussing in wikiprojects. It would also need a working definition of mass creation, of course. Espresso Addict (talk) 03:08, 18 October 2022 (UTC)

My experience is that articles will pretty much always need to be manually checked. In the area I'm most interested in, redirection to a list is a much better option than moving to draft - no one will find draft articles, whereas a redirection to a list is much easier to find and we don't lose the original work and attribution, which, frankly, is actually important. I"d suggest it will be almost impossible to show that articles created in the past didn't have consensus - any raw of AfD which saw numerous articles kept, for example, would suggest that there was consensus for their creation. Blue Square Thing (talk) 16:15, 18 October 2022 (UTC)

There wouldn't be anything stopping you from manually reviewing the draftified articles and creating appropriate redirects, or moving the articles back and redirecting them; I would consider that to be appropriate under when they can no longer be considered mass created, as they are no longer articles.

mass-created without consensus, not created without consensus. If their mass creation was not approved in an appropriate forum (and some mass creations have been) then this process would provide a path to address that mass creation.

I would note the other reason for criteria #3 is to prevent these discussions from turning into trainwrecks; since any articles included in the list that shouldn't be included are easily restored, the discussion shouldn't be derailed because reasonable editors disagree with a few items on the list. BilledMammal (talk) 08:29, 20 October 2022 (UTC)

Espresso Addict Alternatives to Creation - e.g. inclusion in a list. Blue Square Thing (talk) 16:15, 18 October 2022 (UTC)

I suggested something here specific to sports stubs that might be workable if we ironed it out a bit. JoelleJay (talk) 05:04, 19 October 2022 (UTC)

• I think that arguments about the reliability and independence of sources are inherent in writing an encyclopaedia. We've got to evaluate sources in context, case by case.

Obviously the problem is that I, personally, don't have the final say in these things, which means we get a lot of decisions wrong.—S Marshall T/C 12:35, 17 October 2022 (UTC)

• The kind of thing I'm talking about is Wikipedia:Articles for deletion/Elizabeth Otto. Nominated 16 mins after creation without sufficient understanding of the relevant notability guidelines, nor of how to look for sources in this field. The at-scale problem here is that countless articles are nominated for AfD or, worse, prodded with this level of lack of understanding. Espresso Addict (talk) 01:56, 18 October 2022 (UTC)

I agree that this particular AfD was ill-advised especially since the article had just been created. However, aside from the sources which were found, folks seem to be treating WP:NAUTHOR as sufficient to establish notability when in fact it is part of the "additional criteria" section which reads in part "People are likely to be notable if they meet any of the following standards. Failure to meet these criteria is not conclusive proof that a subject should not be included; conversely, meeting one or more does not guarantee that a subject should be included." We really need to harmonize the various SNGs to clarify which ones are actually sufficient to keep an article and which are merely intended to tell us that the subject is likely to be notable. In this case it seems that GNG is met, and the process worked as intended since it appears that the article will be kept. –dlthewave ☎ 05:16, 18 October 2022 (UTC)

It was just one example that I had to hand, and perhaps singling out the nominator, who doesn't seem to make a habit of nominating academics/authors, was unkind. I could unearth a lot more examples if anyone's interested. As I intended to imply, the problem with prods is a great deal more pernicious as there's little/no co-ordinated patrolling. I've deprodded many articles that are more obvious WP:AUTHOR/PROF passes than this one. Articles on anglophone women that are listed at the appropriate delsort do tend to get a decent amount of attention from the Women in Red project.

My understanding of the usual WP:AUTHOR application in AfD is multiple books, each with multiple reviews (with some complicated caveats), which is just a restatement of GNG making it clear that authors are judged on coverage of their works, not their person. Espresso Addict (talk) 06:18, 18 October 2022 (UTC)

None of the notability guidelines guarantee that meeting the standard they express is sufficient to keep an article. Not even the GNG.

My big-picture hot take is that we have drifted away from the questions of "What topics belong in an encyclopedia?" and "How do we organize encyclopedia content into separate-but-interlinked pages?". Instead, we've gone out to the bikeshed and devoted all our energy to harmonizing the various abbreviations that we made up. XOR'easter (talk) 17:29, 19 October 2022 (UTC)

An issue I would like to mention is the trend of people using "success" in terms of number of nominations/!votes that ended up following consensus as an indicator of anything at all. There are just so many problems with this:

1. This is the definition of WP:BATTLEGROUND.

2. It incentivizes groupthink. The reasons for this should be obvious.

3. It erases the distinction between low-effort drive-by comments and ones that cite policy.

4. It ignores the fact that participation is self-selecting. Many people, myself included, are more likely to comment if they have something new to add to the discussion -- which often means a differing opinion -- and less likely to comment if their opinion is just "what that guy said."

Related to this is the skew with which these stats are interpreted. I read someone saying that someone who !votes 50% of the time to keep and 50% of the time to delete is "moderately inclusionist," which boggles the mind. (And also is self-selecting, for the same reason.) Gnomingstuff (talk) 23:41, 28 October 2022 (UTC)

Suggested evidence

Valereee has agreed with the idea of including an evidence section, so I propose below a (partial) skeleton of background info to add. I don't know any of the details here and I know I'm missing plenty of important cases. Ovinus (talk) 19:20, 19 October 2022 (UTC)

I would suggest changing the Lugnuts section to a more general one on athlete bios, e.g.,

=== Sportsperson stubs ===

... statistics, database used ... These stubs were presumed to meet GNG through various WP:NSPORT sport-specific guidelines such as WP:NCRIC and WP:NFOOTY, but many (est. percentage) no longer qualify for that presumption following the 2022 RfC on sportsperson notability criteria. Most of these subjects cannot be quickly identified as notable or non-notable according to new criteria. Many "keep" advocates—who now have the burden to demonstrate sourcing—find the rate overwhelming. At current rates of nomination at AfD (...), the stubs will take ... years to sort out.

JoelleJay (talk) 21:07, 19 October 2022 (UTC)

Adjust as you please. Ovinus (talk) 21:32, 19 October 2022 (UTC)

If they're cricket articles it would take me no more than two minutes to get an idea if there are or aren't likely to be suitable sources notable. Fwiw, if they were early 20th century British cricketers, at least 50% would be likely to have suitable sourcing. If they were Australians or New Zealanders from the same era, that figure's likely to be higher. Naturally the longer a person played for the more likely it is that sources already exist. I would also point out, yet again, that the sources already in an article about a cricketer will need to be checked individually. A number contain extensive prose - as well as links to prose - already and are not simple database entries.

It would take, of course, much longer to actually add the sources and write the articles, but that's different from saying that most of these subjects cannot be quickly identified as notable or non-notable according to new criteria. In other sports that might hold true; it doesn't for cricket. Blue Square Thing (talk) 15:44, 20 October 2022 (UTC)

As a rough guide, we're looking at in excess of 10,000 cricketer biographies (see User:Lugnuts/Cricket#Cricketers), greatly in excess of that when we also include articles created by several other prolific database copiers in this area; most of these articles are sourced exclusively to a cricinfo/cricketarchive database entry with no prose. There are also vast numbers of Olympian biographies (see User:Lugnuts/Olympics) sourced exclusively to the OlyMadMen database (via Olympedia, Olympics.com, sports-reference.com, or elsewhere), and over 3,500 non-Olympic cyclist biographies sourced to Cycling Archives (see User:Lugnuts/Cycling); again, it hasn't just been one individual creating these non-cricket stubs. As such, while "2 mins per article" doesn't sound much, it is a colossal chunk of time. _wjemather^{please leave a message...} 16:30, 20 October 2022 (UTC)

Until you run across an in-depth prose section on the CricInfo or, even worse, the CricketArchive profile. And that's fairly random. In other words, in some cases there is much better sourcing. Every article's going to need checking anyway - there are stubs that have multiple reliable sources in and articles that haven't been re-categorised after their expansion from stub. Fortunately we have lists. From lists we can work through - with a handful of people who know what they're doing we could get the majority of the articles that are likely to be significant dealt with in six months, tops. We'd have to prioritise, but it's doable.

See below for more about subsets. Blue Square Thing (talk) 18:58, 20 October 2022 (UTC)

Ah, I thought quick identification was only applicable to a subset of the articles. Removed. Also, I would appreciate more numbers like this, Wjemather. Maybe have a table or graph of stub type vs. creator, count, and totals? I dunno how to use the Quarry, so like, "Number of stubs in category Cricketers created by Lugnuts/BlackJack/etc." Also @Blue Square Thing: What do you think about other categories of cricketers? Ovinus (talk) 17:38, 20 October 2022 (UTC)

There would be different percentages for different groups - and although those with more appearances are more likely to have sources found on them, that's not entirely certain. As a rule of thumb, for example, British County Championship era cricketers with 20 appearances can almost always have sources found for them. Below 20 it starts to get less certain - but there are plenty of examples of single appearance players who are obviously notable for other stuff.

South Africans, and West Indians are much harder to find online sources for in my experience (but I may not know where to look), and Indian, Pakistani, Sri Lankan and Bangladeshi players harder again, mainly due to the lack of digitised sources and language barriers. Zimbabweans and Irish are more recent in general so likely to be living. They vary - much harder to be certain, but as a rule of thumb I'd want to check every international player in that group. Other countries (UAE, Canada, USA, PNG etc...): harder still.

But in each case a check for a Wisden obituary is the first thing that needs doing. If we find that, there's an argument for looking further. That's one reason why pre-war players are a specific subgroup. Blue Square Thing (talk) 18:58, 20 October 2022 (UTC)

Btw, I should be specific about where I'm getting the 50% from. BilledMammal provided a list of more than 1000 stubs (1,168 by my reckoning) created by BlackJack - almost all of whom were British players for first-class counties. By randomly sampling some of the alphabetic groups that were provided I was able to determine that in every alphabetic group around half had a very good chance of being clearly notable - for example, of the 65 whose forenames began with an A, at least 38 are well worth checking on for more details and there are only 11 which I feel are absolutely unlikely to have any reasonable sources found without too much effort.

As an example, Leslie Wilson (cricketer), an article I've not got to yet, is one of those on the list. So obviously notable - the Carlaw source I've added to the article has significant prose and I've got three books within reaching distance of where I'm sat that will add more, plus his Wisden obit. And that's just the starters.

Of course, look at non-county sides, the minor counties list A chaps and so on and things will look different.

One final point (for now...) - it's easy to bulk out an article with "sources" to avoid it being a stub - but without almost any of those sources being actually about the person or being in any way in-depth. Blue Square Thing (talk) 19:15, 20 October 2022 (UTC)

I disagree with your assessment that 50% are notable, both due to the lack of provided sources for the ones you have reviewed, and due to the sources you have provided often not counting towards GNG due to either not containing significant coverage or due to not being independent. @Valereee: Related to this, I don't think it is appropriate to include editor assertions, made without evidence, such as Blue Square Thing estimates that at least 50% of early-20th-century cricketer stubs are notable in the evidence section. BilledMammal (talk) 08:58, 21 October 2022 (UTC)

Noted. Valereee (talk) 11:55, 21 October 2022 (UTC)

How many would you like me to review and how long can I have to do it? If you want full sourcing adding to the article then I can do two a week. If you want a quick look with a gut feeling, it’s one every two minutes. How long do you want it to take? Or do you just want to delete articles because you can’t wait for someone to actually review the sourced already in the article? Blue Square Thing (talk) 12:48, 21 October 2022 (UTC)

I agree that we shouldn't have that particular piece, but it'd be very nice to know about how many of them are actually notable under current guidelines. Is there a quantitative way to do so? You could look at past AfDs, maybe. Ovinus (talk) 15:11, 21 October 2022 (UTC)

The problem with looking at AfD is that the articles that have been sent there are probably more likely to be sketchy in terms of notability. I can try to do that at some point, but I’m awfully busy just now. And I’m not sure, given some of the stuff that’s gone on at AfD that it’d be awfully helpful tbh. I’ll see what I can do, but can’t promise that it’s a quick job.

Can I check btw: are we focussing on stubs sourced only to simple, non-prose databases or on articles people don’t like and want to question the sourcing of? I’m getting lost already. Blue Square Thing (talk) 17:53, 21 October 2022 (UTC)

@Ovinus: @Valereee: So, I looked at AfD. I restricted it to AfD from 2022 - I wanted to ensure that the work was manageable and that the AfD had all happened after the implementation of the NSPORTS RfC - which involved a messy implementation that took some time to work through properly. I took the AfD from this list (note: I found 2 errors in the automated result). Once I'd taken out non-biographies I was left with 115 articles. 4 of these were nominated more than once during the period - I considered only the last nomination in those cases; 4 were only umpires which I ignored. I found:

• 19 bios without any senior appearances (i.e. would have failed all cricket notability guidelines) - all but one of which were deleted
• 3 Olympics only appearances which are "complex"
• This leaves 85 cricketers who there's a chance might meet guidelines

• 3 Australians - 2 keep, 1 redirect
• 1 Bangladeshi - 1 delete
• 16 English - 8 keep, 7 redirect, 1 delete (refunded to redirect)
• 31 India - 4 keep, 14 redirect, 11 delete, 1 no consensus, 1 draft (hasn't been worked on - I contend that draftifying doesn't work as it's impossible to find the articles)
• 2 Irish - 2 redirect
• 6 New Zealanders - 4 kept, 1 redirected, 1 deleted
• 1 Pakistani - 1 redirect
• 2 South Africans - 2 redirect
• 3 Sri Lankans - 1 keep, 2 delete
• 3 West Indians - 2 redirect, 1 delete
• 8 females - 3 keep, 2 redirect, 2 delete, 1 no consensus
• 9 non full-member countries - 1 keep, 8 redirect (these are all internationals)

In total: 22.4% are deleted, 27.1% kept and 45.8% redirected; when you limit it to just full-member countries, 39% are kept - when you exclude sub-continentals that rises to 43%

What does this show? Well, my 50% happens to have been on the money (and that's not deliberate), but more importantly, there are viable ATD in many cases. Why were so many South Asians deleted? We don't have the same number of redirect targets, they play for multiple sides more frequently due to the structure of cricket in that part of the world, it's much, much harder to find sources. I wonder to what extent there's also an element of institutional bias, but that's difficult to prove.

None of the 16 English county cricketer articles have been deleted (the one that was, you need to read the close rationale for frankly - it took bloody ages to get it refunded as well)

I could go back into 2021 but at some point I'd reach the NSPORTS stuff and it would skew more in favour of keeping - a quick glance through the list suggests that this is probably true - minor players get redirected, sub continentals get deleted, very few really significant players get nominated - and the stubs that people are considering bundling contain really significant players where there are clearly a range of sources.

Of course, smallish sample size, my biased nature etc... means that this almost certainly means nothing. But reading through the AfDs from 2022 was actually quite instructive - you can't quantify it, but there's so much value that I got from that which will almost certainly inform my views at this RfC when it gets going - most importantly, that in a number of cases when work was done, sources were found and nominators withdrew their AfD. Blue Square Thing (talk) 09:49, 29 October 2022 (UTC)

Further evidence. BlackJack created c. 104 British cricket biographies in November 2016. All, I think, are of historic cricketers from the 1860s to 1940s; I think all of them are dead. I've looked at each one:

• 16 had already been developed or had sources added. That leaves 88 as stubs;
• of those 88, I judge 36 as being extremely likely to prove notable - i.e. be kept at AfD (40.1%)
• 16 I judge as being worth further research - there's enough evidence to suggest that there may be more sources, but it would take some time to develop them (18.1%)
• 33 I judge as being unlikely to be able to be shown as notable quickly (37.5%)
• 3 I judge as being flat out redirects (3.4%)

This is based on the "quick look" premise - checking what I consider to be the obvious sources and taking no more than 2 minutes per biography. They include people such as Lionel Lister - so obviously notable when you actually look.

Obviously I may be wrong on all of this. It's subjective and based on a really quick look in obvious places and what that's likely to mean in wider terms - i.e. what other sources are likely to be available etc... Any that weren't to be kept at AfD could be redirected very easily. Blue Square Thing (talk) 14:39, 29 October 2022 (UTC)

Can we strike the language "keep" advocates from the sentence: Many "keep" advocates—who now have the burden to demonstrate sourcing—find the rate overwhelming? I don't consider myself a keep advocate (I've voted 9 to 1 to delete sportsperson biographies at AfD) but I always look for sourcing before voting, thus the rate of sportsperson AfDs is a burden for all editors who participate in those AfDs. Jogurney (talk) 18:34, 20 October 2022 (UTC)

@Jogurney, I interpreted "keep advocates now have the burden..." as a reference to the pre-RfC times where the burden was pretty much entirely on delete !voters to prove sources didn't exist for an athlete who met a sport-specific subguideline. Of course, I also agree that all editors at AfD should be looking for or at least assessing provided sources; if anything, I'd say the effort I expend searching subscription-only, non-Google-indexed, and non-English media for each AfD has actually gone up since NSPORTS2022. This is especially true for AfDs with perfunctory nomination rationales that don't demonstrate BEFORE, but (as you know) even in those that are well-reasoned the threat of !voters influencing an outcome in a low-attendance discussion by misrepresenting coverage or refbombing is enough that I end up doing comprehensive source analyses anyway. I don't mind this as much; it's the endless attempts at relitigating global community consensus by the same IDHT users over and over that really wears me down, and I'm hoping this RfC will better define the community's expectations on AfD participation such that that doesn't happen. JoelleJay (talk) 01:20, 22 October 2022 (UTC)

I wanted to add that the lack of participation at sportsperson AfDs (many of them result in soft deletion) is only partly due to the contentious nature of the discussion, and almost certainly driven down by the amount of effort required in reviewing/analyzing sources compared to when a SNG could be relied upon. The greater the volume of these AfD, the lower the participation has been. Jogurney (talk) 18:56, 20 October 2022 (UTC)

I feel that the list of examples is skewed by the fact that it focuses on situations in which mass article creations resulted in catastrophe or ignominy. If these are to form our whole understanding of what mass article creation entails, it seems that quite draconian measures may appear reasonable. For this reason, I propose my own mass creations be admitted as evidence, none of which have been deleted, and a significant portion of which have passed peer review processes. jp×g 09:51, 25 October 2022 (UTC)

Conduct at AfD

Acrimony at Articles for deletion led to an ArbCom case and the (topic-)banning of several editors. A substantial number of editors have stated the atmosphere of AfD has made them avoid the process.

Can confirm I for quite some while also tried to avoid it.Paradise Chronicle (talk) 14:29, 20 October 2022 (UTC)

Can also confirm, I can only dip in so often before the cumulative unpleasantness and effort expended to find sources are too much to take. In my experience WP:CIVIL is selectively enforced. The meta-discussion is even worse. A lot of verbiage has been expended on Lugnuts' uncivil behavior, but comparatively little, at least in the discussions I've read (and I have read too many, truly), about comments comparing him to "animals that mark their territories with bodily fluids" (probably one of the most degrading things I've seen on here), being accused of "[obviously believing] Wikipedia is some geeky RPG where he's out to win Game High Score" (the "obviously" is doing a lot of heavy lifting here), and so on. Gnomingstuff (talk) 00:25, 29 October 2022 (UTC)

Iranian village stubs

Carlossuarez46 created roughly 70,000 articles in the period from about 2008 to 2014—a rate of about 32 per day—from the 2006 Iranian census. Many of the purportedly populated places turned out to be local companies, wells, or other unpopulated places. Following an AN discussion followed by an an ArbCom case, consensus formed to delete 13,157 of their Iranian "village" stubs focusing on those that had zero population according to the articles as written by Carlossuarez46. This left those that were identifiable based simply on the name of the "village" as being likely not villages per se. 2000+ of them have so far been deleted, across numerous bundled AfDs, including those identifiable as pumps (1 2 3), farms (1), numbered subdivisions and counting-places for e.g., nomads (1), and businesses (1).FOARP (talk) 08:58, 20 October 2022 (UTC)

This is mass creation, not deletion. Espresso Addict (talk) 10:32, 2 November 2022 (UTC)

Espresso Addict - This discusses the deletion of more than 15,000 articles and includes numerous links to deletion discussions. How is this not about deletion? FOARP (talk) 11:47, 2 November 2022 (UTC)

FOARP: Can you focus the evidence on the deletion problem, as reading this I am honestly not sure where it lies. Espresso Addict (talk) 14:11, 2 November 2022 (UTC)

Espresso Addict - This is evidence, not argument. I'm not sure it would be helpful or suitable for me to literally spell out that these 70k articles are overwhelmingly garbage that does not survive AFD, but that deleting them has still been a gargantuan task under our present policies. FOARP (talk) 14:18, 2 November 2022 (UTC)

• I should point out for the record that I carried out the technical aspect of these deletions. –xeno^talk 15:17, 2 November 2022 (UTC)

Sportsperson stubs

Lugnuts... created ... statistics, databases used .... These stubs were mostly presumed notable under WP:NCRIC, WP:NOLY and WP:NCYC, since tightened, mainly as a result of a 2022 RfC on sportsperson notability criteria, and sourced exclusively to all-inclusive databases which commonly contain only basic statistical information, meaning many of the article subjects (est. percentage ...) are no longer presumed notable. <Other examples> <Estimations on notability> Nonetheless, sportsperson AfD discussion take substantial editor time searching for sources, are occasionally contentious, and come in a volume that leads to low or low-quality participation. At current rates of nomination at AfD (...), the stubs will take ... years to sort out.

Again, this is creation focused, not deletion focused. Espresso Addict (talk) 10:34, 2 November 2022 (UTC)

@Espresso Addict: Identifying problematic creation is key to identifying the relevant deletion/other discussions for analysis. _wjemather^{please leave a message...} 12:21, 2 November 2022 (UTC)

My point (to respond to the redacted version) is that you are a month too late with these concerns; identifying problematic creation has been ongoing for a month and is due to close ~today. Espresso Addict (talk) 14:18, 2 November 2022 (UTC)

Espresso Addict - Doesn't it literally state that the problem here is the slow speed of deletion under our present processes? Since the creation argument has basically hit the buffers because a substantial portion of !voters are hung up on the issue of definition, is your proposal that we simply continue that impasse into this discussion? FOARP (talk) 14:22, 2 November 2022 (UTC)

@Espresso Addict: This RfC on mass-deletion is only necessary because of mass-creation. The other RfC should have sharply focused on preventing problematic mass creation (but sadly that focus was lacking); one of the things this RfC should deal with how to best clean up the mess left behind by mass creation, which cannot be done without acknowledging and identifying it. It would take decades and thousands of contributor hours to process these articles individually via AfD so, for example, one possible way forward is to treat problematic mass-creators articles as a whole or in large related batches (e.g. South African cricketer biographies created by x, pre-WWII Olympic modern pentathlete biographies created by x, Fooian village articles created by x, etc.). _wjemather^{please leave a message...} 15:15, 2 November 2022 (UTC)

"The Arbitration Committee requests comment on how to handle mass nominations at Articles for Deletion." -- Not how to delete mass creations! Espresso Addict (talk) 15:33, 2 November 2022 (UTC)

Mass nominations are overwhelmingly the result of mass creation. This was inherent in the focus on Lugnuts and the people who had proposed some of their articles for deletion in the ArbCom case. FOARP (talk) 16:03, 2 November 2022 (UTC)

+1. The majority of mass nominations at AfD are unquestionably a direct result of effort to clean-up following misguided and problematic mass creation. They invariably get derailed quickly by 'procedural' objections. Establishing acceptable parameters based on creation is one way of preventing that. Ultimately, the creation aspect cannot simply be dismissed. _wjemather^{please leave a message...} 16:36, 2 November 2022 (UTC)

But not necessarily, and this RfC is tasked with looking beyond that limiting factor, isn't it? We need to make sure that we don't get the "hey, Lugnuts was shit, lets delete all of their articles" movement (for want of a better word/description - we can dress it up differently if you want, but there's an element of that at play) lead us to a "solution" where we say "this is how we're going to change all deletion to allow massive nominations that we haven't bothered checking". That's not to say that I don't think the examples put forward by Wjemather don't have some merit to them - although, obviously, I'd try to fine tune them a little more. Blue Square Thing (talk) 16:31, 2 November 2022 (UTC)

Turkish village stubs

... statistics on rate of creation and of deletion ...

Only deletion is relevant. Espresso Addict (talk) 10:36, 2 November 2022 (UTC)

That isn't so. Mass deletion is almost invariably a response to mass creation. The biggest issue in mass deletion is what to do with the articles that have already been mass-created. Mangoe (talk) 04:38, 5 November 2022 (UTC)

Mass creation from US Census data

... statistics on rate of creation and of deletion ...

Only deletion is relevant. Espresso Addict (talk) 10:36, 2 November 2022 (UTC)

That isn't so. Mass deletion is almost invariably a response to mass creation. The biggest issue in mass deletion is what to do with the articles that have already been mass-created. Mangoe (talk) 04:38, 5 November 2022 (UTC)

Mass creation of island articles from GNIS

In 2020, JPxG created a large number of extremely short articles about islands in California and Michigan, written formulaically based almost entirely off of information from GNIS. After months (and sometimes years) of not being edited, they were expanded considerably when JPxG returned with additional sources (including many obtained through TWL). Subsequently, 24 of them were featured at WP:DYK and successfully taken through the Good Article nomination process. Furthermore, one (Powder House Island) was nominated at FAC and became a Featured Article in 2022. While they were created with minimal referencing, none of JPxG's island articles have been nominated for deletion, and all of them now feature multiple independent sources apart from database entries. jp×g 09:51, 25 October 2022 (UTC)

This is mass creation, not deletion. Espresso Addict (talk) 10:36, 2 November 2022 (UTC)

Mass creation is what leads to mass deletion attempts. Mangoe (talk) 04:39, 5 November 2022 (UTC)

Bundling and trainwrecks

Bundling of large numbers of articles can often lead to so-called trainwrecks, in which editors advocate keeping only some of the articles. Even slightly differing votes of this type can preclude consensus, or at least make it inefficient.

This is not neutrally worded. Espresso Addict (talk) 10:43, 2 November 2022 (UTC)

Articles left as sub-stubs are not developed

The second paragraph of Wikipedia:Abandoned stubs is relevant to this; it is also relevant that mass created articles are typically on more obscure topics than articles that are not mass created. BilledMammal (talk) 04:19, 2 November 2022 (UTC)

I don't think this is true in general. I'd be interested in actual (ie database query) evidence that related to this. Espresso Addict (talk) 10:29, 2 November 2022 (UTC)

Also as I've written above, this is mass creation, not deletion. Espresso Addict (talk) 10:36, 2 November 2022 (UTC)

It's true in 2018; why wouldn't it be true in general?

This is the underlying query. However, to derive the statistics require API calls for the pageviews as well as further analysis.

It relates to both; it is evidence of whether mass created articles are a positive for the encyclopedia. BilledMammal (talk) 10:53, 2 November 2022 (UTC)

But *is* it true in general? Is this a query which can be run and can do the maths for you, or does it need a lot of extra hand-done work? I'd be interested to know if it's true for 2020, 2016, 2012 and 2008, for example. It would be interesting as well to understand if there's a critical time after creation when articles are more likely to be expanded? I'm not sure that anecdotally I could say that there was, but it's possible. The statistics are interesting though, although I don't know if an article not being looked at very much makes it any less notable. Just because something is obscure it doesn't necessarily make it unimportant.

Fwiw, I'm also not sure that I would define "obscure topic" as being in the bottom 50% of page views. That seems a high value to me - it might not be, but I don't know what the distribution is like. If the bottom 50% of page views meant 3 a year, for example, then I'd be happy, but I honestly don't know what it means. Blue Square Thing (talk) 16:37, 2 November 2022 (UTC)

Fwiw, I hit random article (discarding disambiguations) twenty times, and got average daily hits of 0,0,0,0,0,1,1,1,1,2,3,3,5,5,13,19,21,24,72,99 ([1] & [2]). Espresso Addict (talk) 17:20, 2 November 2022 (UTC)

Was thinking about this. Any way to know how many times Special:Random gets used? Maybe they can be represented as a Zipf's law–looking distribution (organic views) plus a tiny distribution from random views. That said, page views only correlates with reader value; some views (like random ones) are trivial, while others will involve staying on the page for a long time. Ovinus (talk) 17:30, 2 November 2022 (UTC)

I would have though web-crawler hits were more common than random; I use random a fair bit to find articles to copy edit, but I don't imagine it's of much value to readers. Espresso Addict (talk) 17:39, 2 November 2022 (UTC)

This is dubious extrapolation, but probably within a factor of two: Wikipedia:Popular pages gives the ratio of random/main page views as 17% since 2007. Scaling that across all articles gives an average of 4 Special:Random views per month. Ovinus (talk) 17:44, 2 November 2022 (UTC)

A lot of extra work needs to be done, particularly to get page views where API access rate is limited.

Just because something is obscure it doesn't necessarily make it unimportant It doesn't, but it does mean it has less opportunity to be expanded by someone other than the creator as it is less likely to be seen.

Bottom 50% of pageviews means ~1300 views over four years, or 0.9 views per day. BilledMammal (talk) 21:05, 2 November 2022 (UTC)

I have a page view analysis program which uses the dumps, so if you have any particular queries I can run them. (I will need to adapt the program a bit to detect categories.) Ovinus (talk) 21:09, 2 November 2022 (UTC)

@Ovinus: If it's not too much work, I would be interested in statistics for the number of views per article for every article over a one year period; perhaps displayed as a line graph and as a table showing the minimum number of views for each percentage point?

Related to that, where do you get the dumps from? I was under the impression the API was the only way. BilledMammal (talk) 02:39, 4 November 2022 (UTC)

[3], which includes all pages on all Wikimedia sites. When I have time I'll decruft my program and get the data you request, probably at month granularity. (Perhaps we could move technical discussion to my talk or elsewhere.) Ovinus (talk) 03:32, 4 November 2022 (UTC)

If it's not too much work, I'd be interested to know the bottom 75% and 25% of page views for comparison. I'm not totally sold on this quantitative measure of obscure yet - and it feels like Wikipedia should be for the obscure - the whole not paper thing etc... Blue Square Thing (talk) 06:40, 3 November 2022 (UTC)

I agree that Wikipedia should include obscure topics, but for obscure topics the facts suggest that it is particularly important for article creators to develop the articles as no one else will.

25% of articles have more than 5074 views, 75% of articles have more than 543 views. BilledMammal (talk) 09:02, 3 November 2022 (UTC)

Focus on deletion

@Valereee and Xeno: To the moderators, the thing we most need to do with this RfC moving forwards is to avoid re-doing the not-very-conclusive RfC just coming to an end. Therefore I would very strongly suggest limiting/slanting all evidence provided in this RfC to deletion, not creation. Espresso Addict (talk) 10:42, 2 November 2022 (UTC)

I don't actually agree. We did the first RfC to see if we could as a community find some consensus somewhere about mass creations before we started in here. Not being able to doesn't mean it's not still a major factor, maybe the major factor. We can't simply ignore it or the evidence, IMO. Valereee (talk) 10:53, 2 November 2022 (UTC)

Thanks for responding, Valereee. I think in retrospect splitting the RfC seems not to have been the correct way of proceeding. As I've written before, I did not participate in the ArbCom case because I was unable to type at the time, and I might well be misunderstanding or misremembering, but my recollection from the time was that TenPoundHammer & Johnpacklambert were brought to ArbCom because of problematic patterns of deletion, which to my recollection were not directly related to efforts to remove mass creations? Espresso Addict (talk) 14:27, 2 November 2022 (UTC)

While I think this RfC will and should lean heavily towards discussing the deletion aspect, it's almost impossible to completely divorce deletion from the original act of creation, so while we could certainly steer things more towards discussion the act and process of deletion, I agree with Valereee that ignoring or prohibiting evidence about creation will not be possible. –xeno^talk 15:20, 2 November 2022 (UTC)

Agree. Espresso: One oft-cited problem at AfD is large numbers of very-similar nominations. My (unoriginal) point, in proposing this evidence, is that if the nominating editors keep it up, it would take at least a decade until they are "done". As to whether that is a problem, that is for you to decide. But it does imply that "just wait, it'll be over in a few years" is incorrect thinking. Ovinus (talk) 17:25, 2 November 2022 (UTC)

Definitions, forums

The article creation discussion is deadlocked in large part due to opposition to literally any proposal to do anything due to disputes over definitions and forums, even when the definition/forum is basically tangential to the actual proposal. I really hope we are not going to see a repeat of that behaviour here. I hope closing admins will see the "Oppose Until we have a definition of/specific forum for XXXX" !votes for what they are and will weigh them accordingly. FOARP (talk) 08:19, 20 October 2022 (UTC)

FOARP, that would be something to note on that RfC as a 'Note to closers'. Valereee (talk) 19:57, 20 October 2022 (UTC)

I don't know how this RfC can address "mass created articles" in any way if we have no working definition and no working process for mass creation. My hope, per what I wrote over here, is that the closers of that discussion can give us those things, even if only an experimental basis. At least then we'll be working from a shared vocabulary and shared premise here. There are other elements of bulk deletion nominations, of course, but it sure seemed like a lot of the impetus for this RfC was "how do we deal with articles created at scale". — Rhododendrites ^talk \\ 13:07, 24 October 2022 (UTC)

@Rhododendrites - If this discussion is just going to end up deadlocked in exactly the same way, what's the point? It looks to me like people are trying to force through their preferred definitions/forums, opposing even when the proposal works with an already-existing and operating policy. FOARP (talk) 07:30, 25 October 2022 (UTC)

Part of the problem *might* be that people have experience of the ways in which other "consensus" decisions have been implemented in the past - perhaps in ways that they weren't expecting them to have been implemented. That might, quite reasonably, make people wary of agreeing something that isn't entirely clear and where the implementation hasn't been spelled out.

The sheer number of proposals is, also, dangerous in my opinion - once you get to proposal 10, that's too many to keep focus; we're at proposals 17 are we? That's at the stage where people lose interest and just say, "yeah, OK..." without a clear view of what's actually being agreed. Fewer, more focussed, easier to understand proposals would be more helpful. Blue Square Thing (talk) 08:26, 25 October 2022 (UTC)

@Blue Square Thing - I agree that this discussion was not ready for prime time when it opened (and I did say so at the time). We needed longer in the workshop phase to actually come up with something worth discussing.

On the other hand, there's a lot of catastrophising in the discussion that is very unhelpful. Yes, a consensus might come out later about what X should be that you disagree with - but that's Wikipedia? FOARP (talk) 08:36, 25 October 2022 (UTC)

And then that definition gets held up as a "global consensus"? Nah - we need to get things that people can agree on so that we don't end up with "but this obscure definitions proves that... and is better than your definition because I say so". Come up with some simple things that people can agree on:

• some editors have created many articles using very similar formats
• this isn't always very helpful
• they should be encouraged to do this...
• if they don't, we might do this other thing...
• and we might be able to show the need to delete lots of those articles if their notability is obviously questionable

People catastrophise for what seem to them like good reasons. Blue Square Thing (talk) 08:43, 25 October 2022 (UTC)

Viability of modified deletion processes for batches

I made the general outline of a discussion on this subject at WT:AfD some time ago (Wikipedia_talk:Articles_for_deletion/Archive_76#Viability_of_a_modified_nomination_procedure_for_large_batches_of_articles); at that point, this arbitration case was still open, and it was recommended by Worm That Turned that I hold off on such a proposal until this RfC was opened. Now that it seems like an appropriate time, I will reproduce what I said there here.

Having witnessed the regularly occurring arguments at ANI over large amounts of low-quality articles, it seems like large batch nominations (dozens or even hundreds of articles) are a subject of some interest. Having looked through the archives of this talk page, it doesn't seem like there have been any recent proposals along this vein (the last discussion along this vein I could find here was in 2009). Anyway, here's the shit of it, as of now:

It seems to me like batch AfD nominations tend to cause a lot of headache -- we don't really have a good process for dealing with dozens (or hundreds) of articles in a single go. Basically, we only have two ways to do this.

1. The first method is to make nominations one at a time, which causes a lot of redundant effort from participants, who must make a large number of identical arguments across many pages (as well as monitor all the discussions individually, which is difficult even if you use your own AfD stats page to get a current list).
2. The second method is to make one nomination which includes many articles. This practice of "batch nomination" was created as an alternative to the first method, but it still leaves much to be desired. For example:

• It's hard to !vote on a batch AfD. People who don't have the same opinion about every article in the batch end up having to make awkward !votes ("Keep 1 through 5, Redirect 6, and Delete the rest). This is doubly a pain in the ass for closers -- if ten people comment on a ten-page AfD, that could be as many as a hundred !votes to read through when closing. Of course, it won't be that many, but it's considerably more than a single AfD with ten !votes would have.
• It's hard to discuss things in a batch AfD. There are several conversations happening simultaneously on the same page. "What people have to say about Article #3" is dispersed throughout a gigantic discussion where people are talking about many other articles. It's often unclear whether comments are about one of the articles, several of the articles, or all of the articles.
• Most importantly, they only have seven days (and a relist, if someone decides to relist, but that's still only another seven days). This makes no sense to me, and feels like an oversight rather than a deliberate decison. If we agree that it takes seven days to discuss one article, why the hell would it also take seven days to discuss a dozen articles?

Because of this, I think it may be worth contemplating some kind of supplemental guideline (or even a new process) for batch nominations. I don't know exactly what a solution would look like -- it might just be a couple lines in Wikipedia:Deletion policy saying "batches of more than five articles run for thirty days instead of seven" and a page that links to all currently-running batch deletions. But, who knows: maybe an entire "BfD" process could exist where a nomination page has separate sections for each article (as well as one section for overall discussion).

jp×g 09:57, 25 October 2022 (UTC)

One challenge of our current process is that even batch nominations are expected to be evaluated article-by-article to ensure that no notable topics are being deleted. This may work marginally well for small batches, but it bogs down AfD when dealing with indiscriminate mass-creations. The analogy I like to use is that if someone had written List of people from New York City by copying the entire New York City phone directory, we wouldn't go through the list one-by-one evaluating each name to make sure they're not notable. We would WP:TNT the whole thing and start over with sources that actually do demonstrate notability. Likewise we should have a process to mass-delete very large batches of articles that meet certain criteria and allow folks to re-create any that are proven to be notable. It would make sense to hold this type of mass-nomination at a community venue such as Village Pump with, say, a 30-day time limit. –dlthewave ☎ 12:15, 25 October 2022 (UTC)

This is all covered by WP:BUNDLE, no? Yes, bundled deletions can be complex if there are differences between articles, but for mass-created articles this often isn't such an issue because they've been written with the same sources and the same template, so if that's problematic then it's a problem for all the bundled articles (EDIT: dlthewave's example above is a good one for this). The one thing I'd say is mass-creation should be a stand-alone justification for bundling.

But articles having significant differences between them should still be a good reason to unbundle.

EDIT: also agree with dlthewave that doing WP:BEFORE on every single article in a large mass-created set of articles made using the same sources and template is disproportionate. WP:BEFORE should be proportionate to the work done to create the original article. However, I think AFD is also a perfectly appropriate venue for this kind of discussion and don't see the need really to take it elsewhere - once you explain the issue most people will understand why it needs to be done - see, for example, this bundled deletion of 216 articles. FOARP (talk) 12:21, 25 October 2022 (UTC)

"...if that's problematic then it's a problem for all the bundled articles" Not exactly, since sources often cover a mix of notable and non-notable topics, and there's currently no deletion process that allows "indiscriminate mass creation" as a criteria. GNIS would be a classic example since it marks both notable communities and random railroad sidings, subdivisions, ranches etc as "populated places" and the nominator is expected to assess each entry individually when bundling even if no suitable sources are included. The bundling of Iranian census tracts has been successful for straightforward farms, pumps, factories, etc but eventually we're going to be left with a short list of miscellaneous place names that's going to be very time consuming to sort through if individual BEFORE is required. –dlthewave ☎ 12:54, 25 October 2022 (UTC)

It depends on whether you're going to suggest we have to undertake a fully detailed BEFORE or whether we should simply be checking the sources already in the article super quickly. So, for example, I wouldn't be unhappy bundling Heman Luwang, Akoijam Tenyson Singh, Kumar Aditya, Govind Sharma, Rajesh Singh (cricketer, born 1976), Abhay Joshi etc... - all Indian cricketers who have played less than 5 professional matches for not very prominent teams. I've checked the refs on each of them and there's nothing obvious that raises a flag for me. They took less than 5 minutes to check (along with another 30-odd articles). All are Lugnuts' creations from January 2019 and I could easily add another 10 articles to that list. But then I'd come across Lalit Yadav (Delhi cricketer) (Lugnuts, early February 2019). Doesn't look much more promising does it? Sources look a bit run of the mill, databases, scorecards and lists, don't they? Now click the source in the infobox and scroll down - obviously notable. Not even close.

I've not had to do a detailed BEFORE - I've literally clicked the CricInfo link in the infobox for each of these articles (and another 115-ish - on average less than 30 seconds per article, including clerking time.

We can do that. We're good enough to at least check - and working with wikiprojects we can come up with a way of shortcutting as well. (btw, of the 120 articles, I reckon I'd send about 95 to a set of bundle lists - and that's being really harsh - some of them would probably survive AfD if I'm honest; but I'd not be happy to send 25 or so - those are either obviously notable, have already been developed or deserve a second look. Blue Square Thing (talk) 16:52, 2 November 2022 (UTC)

@Dlthewave:, @FOARP:: What I mean to say here is that WP:BUNDLEs are a pain, and most of the time, we end up having to do them individually, because the entire AfD process is set up for a single article. I know that all three of us have been at the same AfDs often, and the situation generally looks like twenty individual noms in a row named "Oaks Hollow, Missouri" and "Stevens Hollow, Missouri" and "Huxley Hollow, Missouri" and "Adams Hollow, Missouri" -- and for the most part we are all saying the same stuff at all of them. Now, maybe I fire up my TWL accounts and I find some really great coverage for Huxley Hollow -- it shouldn't be some big huge ordeal (for me or for you or for the closer) if I say "I agree that all of these are crap but we should keep Huxley Hollow". Yet it is, in practice, and I think this is largely because bundles are a poorly-implemented system. This is why I say a good step towards making them more logical might be to establish longer listing times for batches, and make a delsort category for them. This wouldn't break or seriously damage the existing system -- they could still go through AfD just like everything else -- but it would allow for people to save effort and time in what's otherwise an extremely tedious process. If there are 500 garbo articles that are all basically the same thing, there should be a realistic non-nuts way to nominate them all at once. jp×g 14:26, 25 October 2022 (UTC)

jp - One things that's a problem with bundling is the need to template every page. Using AWB makes this quicker but it's not exactly and easy tool to use. FOARP (talk) 15:24, 25 October 2022 (UTC)

A longer listing time for batches is a good idea - and the length of that period should probably vary by the size of the batch. The problem comes when 25 of the 500 articles turn out to be about obviously notable people or where there are obvious ATD which apply to the sort of article we're taking about (e.g. schools to the location). The biggest inherent problem with batches that I've come across in the areas I tend to have an interest in is simply that nominators don't actually consider those possibilities. That's going to be an even bigger problem when people start to consider the sorts of articles we're all thinking about here. Blue Square Thing (talk) 08:19, 26 October 2022 (UTC)

Regarding half-way views when bundling: perhaps there could be a "mark a few as possibly notable" approach. Editors would simply plop their signature next to the couple articles they think should be kept, subjected to an appropriate ATD, or at least further discussed. Then, unless they are clearly marking those articles disruptively, those articles would be kept when closing the bundled deletion. In other words, a single editor can mark an article as possibly or likely notable. But those kept articles could then be raised at a separate, smaller AfD. And if an editor thinks too many articles in the bundle shouldn't be deleted, they should simply !vote for a (procedural) keep.

The main problem I foresee would be the appropriate ATDs that Blue Square Thing raises, since for some topics it might be best to redirect a bunch of the articles rather than deleting them. One option is to simply establish, for certain classes of problematic articles, broad criteria where an ATD is presumed over deletion. So, Iranian villages, probably nowhere; cricket stubs, probably to some team or list of contemporaneous cricketers. The thing is that the closer should not be forced to redirect it themselves. Not sure how that would work. Ovinus (talk) 19:25, 3 November 2022 (UTC)

One way might be to discuss a batch at a wiki project with whoever might be interested and come to an agreement to do the AtD thing there rather than take it to AfD. That might help stop AfD getting overwhelmed with discussions. There are people who will cry "local consensus" here. If they want progress on this, they need to stop doing that - it's really not helping. Blue Square Thing (talk) 08:19, 4 November 2022 (UTC)

Except in really obvious cases, it's always going to be a good idea to run large bundles past a few other editors (ideally including subject matter experts) before nominating them. The goal of such a discussion would be solely to determine whether it's a good bundle or not, including removing pages/groups of pages that are obviously notable or should be considered separately (including splitting the bundle into multiple smaller ones). There would be no need to tag the articles or talk pages (although no prejudice against the latter) and should be lightweight explicitly without prejudice (either way) to any actual deletion discussions and should not be seen as form of canvassing.

I would also strongly suggest that whenever a bundle is nominated the nomination statement include a mention of why this set of pages was bundled together. Thryduulf (talk) 19:16, 5 November 2022 (UTC)

Noting two threads at ANI relevant to this discussion

• Wikipedia:Administrators' noticeboard/Incidents#Concerns about articles nominated for deletion
• Wikipedia:Administrators' noticeboard/Incidents#Undiscussed mass article merging and redirection by BilledMammal

Espresso Addict (talk) 04:08, 29 October 2022 (UTC)

Should we still go ahead with this?

Conversation here has dropped off quite a bit and I really don't get the impression that anything has come out of the discussion so far that's going to pass. Just me? FOARP (talk) 10:47, 2 November 2022 (UTC)

Wasn't this the RfC that was actually mandated by the Arbitration Committee? Espresso Addict (talk) 10:52, 2 November 2022 (UTC)

Yes, AfD at scale was the RfC requested by ArbCom. Valereee (talk) 11:07, 2 November 2022 (UTC)

An RFC has already been held and we can all see what the result of it has been. If we go ahead with a further discussion that at this point seems unlikely to produce any meaningful result over what has already been seen, are we really fulfilling ArbCom's intent? FOARP (talk) 11:25, 2 November 2022 (UTC)

@FOARP, I'm a little puzzled as to what you think the alternative is? Valereee (talk) 12:10, 3 November 2022 (UTC)

Valereee - Simply not opening this RFC unless there is a good reason to believe that the issues that hindered the previous RFC are resolved? There seems to be little appetite for it on this page. FOARP (talk) 12:20, 3 November 2022 (UTC)

@FOARP, and you think we can resolve those issues how? I'm sorry, I just have no idea what you're going for, here. Are you trying to make some point? If so, I wish you'd just come right out and say it clearly. Valereee (talk) 13:00, 3 November 2022 (UTC)

My point is that this RFC already shows signs of being headed in exactly the same direction as the previous RFC (i.e., nothing meaningful passes). That being the case, why open it? FOARP (talk) 13:03, 3 November 2022 (UTC)

So we can tell ArbCom, "Nope, the community is unable to solve this. We tried our best, but we are unable to come to any consensus about any of the solutions posed by anyone." If we don't even open it, how can we say we have done our due diligence to try to solve the problems caused at AfD by the issues surrounding deletions at scale? Valereee (talk) 13:07, 3 November 2022 (UTC)

OK, but let's have a more clearly put together RFC with way fewer option than last time out, otherwise the same thing is going to happen again. I'm not pretending this is easy but I really think the previous RFC failed because it came out of the gate too early and not prepared. I understand why the discussion was split but in retrospect it seems to have been the wrong move since it's left questions being bounced into this RFC and people disputing whether some issues shouldn't be excluded entirely. I'm not blaming anyone specifically for this - it's just the way things came out. FOARP (talk) 14:26, 3 November 2022 (UTC)

I think the split was a net pos, myself. I think we probably are going to find we need another intermediate RfC on defining mass creation. This is a huge problem. It's not going to be easily solved, and I think it could take a long time and multiple discussions. Valereee (talk) 01:44, 4 November 2022 (UTC)

I'll just say that requested ≠ mandated. Nobody has to do this. – Joe (talk) 15:00, 3 November 2022 (UTC)

Perhaps start by focussing on things that **can** be agreed. We know where people stand on this - some people want to delete shed loads of stuff, others want to keep everything. OK, so where can we meet to agree some stuff? We know bundling causes problems, so can we agree that there are any ways that we can handle that? Is it possible to agree that there are ever any times that it's acceptable to bundle, say 250 articles? (yes, in my view - so long as the list has been curated carefully) How long should we give people to check those over (2 weeks, rather than 1, perhaps), what do we do if there's an obvious problem with the creation and more than 5% are notable (stop the process and start again, being more careful with curation perhaps?), can we place a hard upper limit on this stuff - or do we need to? Even if we can't get agreement on bigger stuff, can we agree on some of the smaller stuff?

If the proposals can be things people can agree on, I think there's a chance. If it's just going to be extreme ends of the spectrum, maybe it does end the same way. Blue Square Thing (talk) 15:59, 3 November 2022 (UTC)

Your comment makes me think we might be able to agree on two things:

Large bundles are not required to be 100% accurate; editors objecting to specific elements need to be objecting to at least 5% of them for their !vote to be considered by the closer. As a consequence of this, large bundles should not be deleted but sent to draft space, so that editors who object to a few examples within the group can easily restore the content. BilledMammal (talk) 16:49, 3 November 2022 (UTC)

@BilledMammal: I'll bullet this to make it easier, perhaps, to follow...

• The 5% thing is workable - that would throw out the entire bundle at that point and, perhaps, after multiple attempts like that deny the curator the opportunity to bundle that many? Not automatically, but as an option.
• We do need to curate lists carefully - I know there's a lot of articles people want to delete, but in >99% of cases there's no real rush - and what we agree here needs to last and apply to as wide a range of articles and situations as possible
• Indvidual items should still be removable - I think we need to work at the level of "there's nothing here that's even vaguely obvious and all these items are really similar". The odd one might then sneak through, but those can be removed on their own basis and without in depth argument - even a reasonable hint that they might have decent sources should mean we go to AfD individually or in a small bundle
• We really do need to think about the time element - if there are shed loads, it'll overwhelm and won't work
• Draft == deletion for me - unless they're in an individuals user space; I'd much rather promote a range AtD, working with wikiprojects, as part of the way forward - you know I prefer redirection for the sorts of articles I'm interested in; I know that works well for schools as well, for example. Don't mandate draft, but give it as an option

There is hope here, but we need to consider what will work most of the time. Blue Square Thing (talk) 07:59, 4 November 2022 (UTC)

• There's also the possibility of a new speedy criterion, or a new type of prod. I'd be more comfortable with deletions at scale with limited oversight if we could find a way of getting proper consideration from the appropriate wikiprojects in advance. Mass move to draft is certainly worth discussion. Espresso Addict (talk) 17:04, 3 November 2022 (UTC)

I am just waiting for the RfC to start in which I would like to suggest the deletion of all one two-line stubs that haven't been edited in the last x years per WP:TNT in order to encourage the creation of new and expanded articles.Paradise Chronicle (talk) 22:22, 3 November 2022 (UTC)

Ah, yes and sorry if I do not read all the comments of the talk page, there of course exists the possibility that this has been suggested before in the long discussion, but I didn't find such a suggestion by now.Paradise Chronicle (talk) 22:24, 3 November 2022 (UTC)

I don't think it has, but I would suggest we open a section on the talk page to propose and workshop RfC questions before the RfC is opened. I think it would help avoid some of the issues with the creation RfC. BilledMammal (talk) 02:05, 4 November 2022 (UTC)

@BilledMammal, we workshopped the questions for the first RfC for a couple of weeks and tried to distill them. We still ended up with 20+ questions. I am of the opinion that the workshopping didn't contribute much to the process. We still ended up with revised questions being added, and several contributors referenced questions that weren't included.

Re: the start of the next RfC. Current thinking for Xeno and me is that we're going to need an RfC on defining mass creation, and with ArbCom elections starting in a month, we maybe should hold the RfC on mass deletions until after the end of year holidays. We know that's not ideal. Valereee (talk) 02:32, 4 November 2022 (UTC)

That's fair enough, although I think the option to workshop questions would still be useful.

We know that's not ideal. I don't believe there is any urgency here; after the end of year holidays sounds reasonable to me. BilledMammal (talk) 02:40, 4 November 2022 (UTC)

@BilledMammal, oh, definitely anyone should go ahead and workshop questions here! I'd suggest opening a new section for each proposed question to be workshopped. I'm just not going to try to do that work myself, as I spent just ridiculous amounts of time trying to distill questions which then were simply overwhelmed by those opposing for being too much plus those opposing for being too little.

And @Paradise Chronicle, yes, if you have a proposed solution, it's always a good idea to get the reactions of others before finalizing it. What BST is saying is extremely likely, IMO. Valereee (talk) 14:47, 4 November 2022 (UTC)

Thanks, I received and understand your concern. Paradise Chronicle (talk) 20:43, 5 November 2022 (UTC)

I believe it would be helpful to have a discussion on the specific problems/pain points that editors want to alleviate, such as new page patrol workload, minimum stub content (such as minimum sourcing requirements), ways to encourage stubs to be developed further, and so forth. From this, we can better understand how to define mass article creation in a way that will help address the agreed-upon issues. isaacl (talk) 04:32, 4 November 2022 (UTC)

At which point the discussion is in danger of getting overwhelming. I don't have a solution to this, but if we take things in small steps and focus on what we can agree on, we've a much better chance of actually agreeing on something. I do agree that there's no rush fwiw - better to get things right(ish). Blue Square Thing (talk) 08:05, 4 November 2022 (UTC)

I feel the first small step should be agreeing on what specific issues are being addressed, as I think a lack of established consensus on this aspect is why it's been hard to agree on what are problematic modes of creation. Figuring out what to do with complex problems is challenging with English Wikipedia's decision-making traditions, as it requires focus and engagement from a sufficiently broad spectrum of editors over an extended period of time. (Most organizations would delegate to a subcommittee and then ratify the results.) Out of a desire to minimize time spent, many discussions end up being very solution-focused. This has its advantages, but does make it harder to discuss problems with many different factors and editors who weigh them differently, as the weighting doesn't get discussed directly. isaacl (talk) 15:27, 4 November 2022 (UTC)

@Paradise Chronicle: to me that's an example that hasn't got a chance of passing - it's far too blunt an instrument and the creation at scale RfC suggests to me very strongly that those simply aren't going to pass. Blue Square Thing (talk) 08:05, 4 November 2022 (UTC)

I also guess it is sort of difficult. But we are discussing "at scale" not article by article. Happy to have received your feedback, probably need to develop a bit more. Paradise Chronicle (talk) 10:02, 6 November 2022 (UTC)

@Espresso Addict: certainly I think wikiprojects *have* to be heavily involved and a full consideration of different AtD need to be considered (as above, I have draft and much prefer redirection in the area I work in...). One of the problems with the Sports RfC is that it didn't have the obvious support of many people actually working in the area of sports and the results came as something of a surprise to an awful lot of people. Sure, if a wiki project simply refuses to cooperate then break out big sticks, but in my experience many will and getting people onside will make any mass deletion process a lot easier Blue Square Thing (talk) 08:05, 4 November 2022 (UTC)

The thing I've found most chilling in some of the recent discussions is the idea that we (whoever "we" is, in this context) should attempt to circumvent the relevant wikiprojects because they are "involved". I can see that something's gone wrong with the area of sports bios, and in the past I know there have been problems with (off the top of my head) fictional characters, porn actors and beauty pageant contestents, but generally wikiprojects attract experts who possess offline sources and the willingness to work on unpromising-looking articles. Espresso Addict (talk) 13:40, 4 November 2022 (UTC)

I agree that the views of editors interested in a topic shouldn't be arbitrarily discarded. I understand why some think they might have an overly inclusive view of what is important to include in Wikipedia, but they're also the ones best positioned to distinguish between insignificant and significant topics within their area of interest. Many industries have trade publications, for example, where a lot of publicity-inspired items are published. Subject-matter expertise is needed to identify actual notable coverage versus product announcement-type coverage. isaacl (talk) 15:41, 4 November 2022 (UTC)

It also is a major factor in why things get so adversarial -- what else can be expected when people are viewed, by default, as the adversary? There are undertones of it here -- a WikiProject might "refuse to cooperate" because they do not believe, based on their subject-matter knowledge, that mass deletion of articles in their coverage zone is warranted. Framing this as them being automatically wrong, and the resulting deletions as punishment ("break out big sticks"), is one of many immensely frustrating things here. Gnomingstuff (talk) 06:58, 6 November 2022 (UTC)

@Gnomingstuff: My apologies for that; it's not intended to be quite read that way, but I've got experience of working within a wiki project where historically a number of people have been unwilling to compromise on a "bright line criteria" whereby "all X are, absolutely have to be, automatically notable and we absolutely must have an article on all of them". Seriously. Getting any form of compromise took years of nudging and I'm fairly certain that there are still people who think it's a really, really bad idea to move away from their absolutionist perspective. That's the experience I'm coming from; I'd hope that all projects are able to see some form of sensible solution, in the same way that I'd hope that those who advocate to delete tens of thousands of articles could also see that some form of non-abolitionist solution is more workable. I'm more than happy to discuss specifics if it helps, here or elsewhere. Blue Square Thing (talk) 08:27, 6 November 2022 (UTC)

@Joe Roe, it doesn't have to be mandated for me to think it's worth trying to solve. For me, a request is enough. I think it's worth it to try to solve this or at minimum prove we can't. I'm not actually sure why anyone would not want to try to solve a major problem that many people are dealing with. This is a problem that is causing major dysfunction in one of our most basic processes. My feeling is that of course we should try to help our colleagues, but it's completely fine if you don't want to contribute to that effort. Valereee (talk) 01:39, 4 November 2022 (UTC)

If there's a problem, sure. – Joe (talk) 06:42, 4 November 2022 (UTC)

Lots of people seem to think there's a problem. Valereee (talk) 11:00, 4 November 2022 (UTC)

YES we should still go ahead with this. Levivich (talk) 14:46, 4 November 2022 (UTC)

I agree with Levicich - let's proceed. No reason to change the process half-way through. (though surely the other AFC should close first) Nfitz (talk) 02:52, 5 November 2022 (UTC)

Also, have we even begun workshopping this? I've been watching the top of WP:Arbitration Committee/Requests for comment/Article creation at scale and WP:Arbitration Committee/Requests for comment/AfD at scale where it notes the timeline for the workshopping is still TBA. 02:56, 5 November 2022 (UTC) — Preceding unsigned comment added by Nfitz (talk • contribs)

Nfitz, yes, we're workshopping now. Anyone with a suggested question should start a section to workshop. Valereee (talk) 13:41, 6 November 2022 (UTC)

Someone should change the headers then User:Valereee - they still indicate it's not started yet. Nfitz (talk) 17:27, 6 November 2022 (UTC)

It was changed this morning, I think? LMK if you still aren't seeing it. Valereee (talk) 21:44, 6 November 2022 (UTC)

Workshopping

Several people have indicated an interest in workshopping proposed questions. I'm not going to try to do this myself, but I definitely think it's a good idea for participants who want to propose a question to first workshop it here. So please feel free to open a section and get feedback. Valereee (talk) 19:42, 5 November 2022 (UTC)

Assessment of consensus for multi-part questions

The closers have requested for multipart questions on future RfCs in this series that we provide instructions for determining consensus.

My initial preference would be something like this:

IRV counting flowchart

1. Assign ordinal rankings of n-1 (most preferred) to n-n (least preferred of n rankings), with unrated responses also assigned a 0.
2. Eliminate the less-favored responses iteratively using something like the flowchart to the other left.

Does anyone have a better solution? Pinging Vanamonde93, with whom I've had similar discussions. Valereee (talk) 12:02, 6 November 2022 (UTC)

Your "flowchart to the left" is on the right. Just saying. Scolaire (talk) 12:57, 6 November 2022 (UTC)

hahahaha Valereee (talk) 13:07, 6 November 2022 (UTC)

• My preference is for IRV. It was used successfully here. --Scolaire (talk) 13:04, 6 November 2022 (UTC)

  Fancy! I like it. I'd have to have someone set that up for me lol... Valereee (talk) 13:10, 6 November 2022 (UTC)

  @Valereee: There's a calculator tool at https://petertheone.github.io/IRV/. --Scolaire (talk) 14:55, 6 November 2022 (UTC)

  Wow, that is so cool! That seems like the way to go. Valereee (talk) 15:17, 6 November 2022 (UTC)

• My preference would be to not ask multi-part questions. I don't think they're effective at generating consensus. Blue Square Thing (talk) 14:18, 6 November 2022 (UTC)
• I have a strong preference for the instant runoff algorithm; it does not depend on arbitrary numerical values. It's also been used successfully in real-world scenarios, including by the WMF. I agree multi-part questions are confusing, but sometimes they're unavoidable, especially when the question we're looking at doesn't have a clean binary outcome. Vanamonde (Talk) 16:05, 6 November 2022 (UTC)
• Instant runoff makes sense now that I've seen an explanation. There was a lot of confusion/objection to the requirement that editors list all options in order of preference; it will probably go more smoothly if the reasoning and counting method is explained up front. –dlthewave ☎ 16:46, 6 November 2022 (UTC)
• Instant runoff is good for votes where one of the options must be selected. Voters are able to specify their relative preferences for options they do support, and for options they don't. For cases where it should be possible to reject all options, I suggest combining it with approval voting: in addition to ranking the options, voters specify all the choices they support, and the choices that have consensus support are identified based on this. The ranking of preferences is then used to select the chosen option out of those with consensus support. (Personally I prefer something like ranked pairs be used, but instant runoff will probably work well enough.) isaacl (talk) 17:26, 6 November 2022 (UTC)
• IRV def the way to go. Levivich (talk) 22:18, 6 November 2022 (UTC)

Workshopping 1

Question 1

Should we create a speedy deletion criteria for articles sourced to a source (or several sources) only mentioning information that do(es) not meet GNG? Articles nominated for a speedy deletion should have been created by experienced editors and the criteria would be able to be applied retroactively.Paradise Chronicle (talk) 17:51, 6 November 2022 (UTC)

PC, apologies for misreading if that's what I'm doing...you seem to be limiting this question to articles created by experienced editors? Valereee (talk) 22:23, 6 November 2022 (UTC)

I also calculated the concern of biting the newcomers. But you are right, it can be worded better. My aim are the basic stub articles of (mass creating/by bot or manually) editors that were/are on wikipedia for/since years.Paradise Chronicle (talk) 23:07, 6 November 2022 (UTC)

If this has any hope of being workable it needs to be much more objective - what is the definition of "experienced editor"? Whose opinion about whether a source meets the GNG matters? Thryduulf (talk) 00:02, 7 November 2022 (UTC)

The same point about sources could be applied to all of the proposals put forward at this point. Blue Square Thing (talk) 07:44, 7 November 2022 (UTC)

Broad community consensus; editors will be expected to be aware of community norms and conform to them per WP:CIR. In the context of proposal 2, this would mean that editors repeatedly prodding articles that do have a source that plausibly contributes to GNG, and editors repeatedly removing such prods from articles that do not have a source that plausibly contributes to GNG, can be taken to ANI and sanctioned, similar to an editor who repeatedly makes bad AfD nominations. BilledMammal (talk) 09:43, 7 November 2022 (UTC)

You keep using "plausibly contributes to GNG" but there is no "broad community consensus" on what that means. We have AfDs where people in good faith argue the same source very clearly does and very clearly does not meet the GNG. Thryduulf (talk) 09:55, 7 November 2022 (UTC)

If it isn't a fringe position that the source meets GNG then it plausibly contributes to GNG; I don't consider this complicated. I will add that very few of our policies have strict bright-line definitions and the community is able to handle those without issue; this won't be any different. BilledMammal (talk) 10:51, 7 November 2022 (UTC)

This is a proposal to allow speedy deletion, which requires objective definitions. Evaluating all the sources in an article to determine whether at least one of them has enough coverage that "it is not a fringe position that it meets the GNG" (whatever that actually means in practice) is not something that can be reliably done by a single admin patrolling speedy deletion categories. Thryduulf (talk) 12:53, 7 November 2022 (UTC)

I would note that my reply was more in response to Blue Square Thing saying that the same point applied to the other proposals. For this specific proposal I'm not experienced in CSD procedures but a brief review of them suggest that non-objective definitions have precedent; for example G4 uses the subjective phrasing "sufficiently identical copies". BilledMammal (talk) 13:21, 7 November 2022 (UTC)

A limited amount of subjectivity is inevitable, but G4 works because a reviewing admin can quickly and easily compare two versions of the same page. I very strongly recommend becoming familiar with CSD procedures, especially WP:NEWCSD, before proposing new criteria as they require very careful wording (this comment is addressed generally, not just to you). Thryduulf (talk) 13:46, 7 November 2022 (UTC)

isn't a fringe position? Who's going to define that? Or does it just depend on someone saying "no, such and such an author is not an expert on the subject because I say so". Blue Square Thing (talk) 16:07, 11 November 2022 (UTC)

That's not really what WP:CIR means though, is it? Especially in the context of AfD etc... Blue Square Thing (talk) 16:07, 11 November 2022 (UTC)

I would argue that anyone who is extended confirmed or has created more than 25 articles is enough experienced to be included in the criteria. This is enough time for the AfC and NPP teams to have prepared the once new editor for article creation. Paradise Chronicle (talk) 18:16, 7 November 2022 (UTC)

Extended confirmed seems a reasonable way of defining this to me. Blue Square Thing (talk) 16:07, 11 November 2022 (UTC)

This doesn't really have any chance. You want to speedy delete articles that may be notable and verifiable based on sourcing -- and not even "mass created" articles? — Rhododendrites ^talk \\ 14:10, 21 November 2022 (UTC)

Proposal 2

(Based on Wikipedia:Proposed deletion#Proposed deletion of biographies of living people, and conditional on the relevant proposal from the mass creation RfC being passed)

All mass created articles (except those not required to meet GNG) must be cited to at least one source which would plausibly contribute to GNG: that is, which constitutes significant coverage in an independent reliable secondary source, or else it can be proposed for deletion. The {{prod masscreate/dated}} tag may not be removed until such a source is added and if none is forthcoming within seventhirty days the article may be deleted. This does not affect the regular prod process, which may still be used on mass created articles, including mass created articles from which the mass create prod has been legitimately removed. BilledMammal (talk) 00:41, 7 November 2022 (UTC)

Seven days isn't very many if there are thousands of such tags in play in a project area at any one time. You might want to be more generous or limit the number of tags which can be applied at one time. Blue Square Thing (talk) 07:43, 7 November 2022 (UTC)

I don't think the number of tags can be practically limited, so I've increased it to a month. BilledMammal (talk) 09:33, 7 November 2022 (UTC)

A month makes sense - thank you. I tend to think that, with there not being a deadline and so on, that it's not unreasonable to suggest a limit on tags - or perhaps once tags reach a particular level the time allowed is increased

I've thought about this, and I'd very strongly object to the need to physically add a source if this is going to be applied retrospectively to huge numbers of articles. If I can show the such a source exists, that should be good enough - per WP:NEXIST. It takes far too long to add sources - and this is placing a huge burden of work on those who review such lists, especially if there are large numbers of articles requiring review within a relatively short timeframe. As I've said in other places here, the idea of reducing the requirement for a full and detailed BEFORE is reasonable; in the same spirit, I should only have to be clear that sources exist - a job that can take seconds - rather than add them - a job that requires longer as well as direct access to the actual source (which can be tricky, especially with offline sources). Blue Square Thing (talk) 16:14, 11 November 2022 (UTC)

Proposal 3

Articles which were mass created without formal consensus from the broader community, or whose consensus was later revoked, are subject to mass-draftification. This is done through a discussion at WP:VPR, where any editor may propose either a list of articles to be draftified, or a quarry query that produces a list of articles. This proposal must meet the following criteria:

1. The articles are limited to a single creator
2. The articles are limited to a single topic area, broadly defined
3. The articles are limited to mass created articles with few (<5%) false positives. "False positives" are articles that were not mass created, or were mass created but later expanded.

Arguments for or against approval of the mass-draftification on grounds other than whether they meet or do not meet this criteria must be ignored by the closer.

Articles may be restored to article space under one of two circumstances:

1. Individually, when they can no longer be considered mass created; this includes false positives, articles that were expanded after draftification, and articles that have been converted into redirects
2. As a group, if there is there is a consensus to approve the mass creation

BilledMammal (talk) 00:41, 7 November 2022 (UTC)

False positives is intended to avoid the discussions turning into train wrecks; I do not believe its inclusion will cause issues, as any editor will be empowered to move those articles back into article space. BilledMammal (talk) 00:41, 7 November 2022 (UTC)

What does "later expanded" mean in this context? How do you propose to deal with clearly notable articles which have not been expanded or only expanded marginally and already have sources in the article which demonstrate this? Wikiprojects should clearly be informed and time provided to check through lists. As an aside, it strikes me that "false positive" means "nominator made an error" or "nominator didn't bother checking", not actually false positive. Or are we simply going to use queries to generate lists and not bother checking them? Blue Square Thing (talk) 07:41, 7 November 2022 (UTC)

What does "later expanded" mean in this context? After the initial mass creation, an editor - either the creator or another editor - came along and further developed the article.

How do you propose to deal with clearly notable articles which have not been expanded or only expanded marginally and already have sources in the article which demonstrate this? Interested editors will be encouraged to expand those articles and return them to article space.

Wikiprojects should clearly be informed and time provided to check through lists. That is part of the reason the proposal is to draftify, not delete, the articles; to give Wikiprojects time to do that.

As an aside, it strikes me that "false positive" means "nominator made an error" or "nominator didn't bother checking", not actually false positive. False positives can mean one of three things; that the nominator made an error, that the current status of an article is ambiguous, or that the article was expanded after being nominated. It is intended to ensure that nominations aren't derailed by a small number of false positives, since the fact that the articles are draftified, and editors are empowered to move individual articles back to article space, mean that the consequences of false positives are minimal. BilledMammal (talk) 09:39, 7 November 2022 (UTC)

I'll bullet for clarity:

• so, *anyone* has to have added *something* to the (body of the???) article? Even the original creator? And even if it's just quite limited? Or are you considering that there needs to be another way to define this?
• similar to my point in proposal 2, there can't be an emphasis on a requirement to expand before returning an article from draft space if it is really obvious that an article is clearly notable (I'm not talking about fringe cases here - but cases such as Lalit Yadav (Delhi cricketer), an article that could easily show up on such a list but where quality sources clearly and obviously exist). Again, if we're going to remove the need for a detailed BEFORE - or any form of BEFORE - then we need to allow articles like that to be returned to article space where any editor can develop them;
• I think you overestimate the ability of editors to find things in draft space or use it in any way. I suspect you'll simply see articles being re-created, perhaps with slightly different names - it'd be dead easy for someone who doesn't really know what they're doing to create Habib Ahmed (cricketer) and not even spot Habib Ahmed has Draft:Habib Ahmed
• There are alternatives to deletion beyond drafts.

More general issue below. Blue Square Thing (talk) 16:34, 11 November 2022 (UTC)

1. Any non-trivial content added to the body of the article would count as expansion; editors would be expected to not game this, such as by adding content based on a boilerplate. This would not include copyediting, adding an infobox, etc. BilledMammal (talk) 11:07, 14 November 2022 (UTC)
2. Editors would be free to develop the articles in draft space, project space, or user space. Once expanded notable topics can be returned to article space.
3. Editors wouldn't be expected to expand these articles if they prefer to create a new article. I don't see them not noticing the draft and creating a new article as an issue.
4. Editors would be free to implement those alternatives; the proposal would not prevent editors merging these articles, redirecting them, etc.

BilledMammal (talk) 11:07, 14 November 2022 (UTC)

Thought about this a lot recently, and I don't think that it's sensible to limit the proposal to either a list of articles to be draftified, or a quarry query that produces a list of articles. Other methods of producing a list are either possible or may become possible. It would be better to simply say a list of articles to be and at that point I'd replace draftified to considered for... and at that point I'm not sure about the wording. Drafting isn't always the solution - in cases such as schools or hamlets there is a long established consensus that redirecting to the parent location is a good AtD that deals with all sorts of issues. The same may well become true of railway stations.

As a further point, I would specifically object to the inclusion of a quarry query that produces a list of articles unless you can show me that that's easy to access in Wiki-space and that the list is clickable. I've got very little experience with such queries, but any list absolutely must be accessible and have clickable links to articles - and if that places the emphasis on the person writing the query to get the output sorted in that fashion then I'd be perfectly happy with that.

I'd also want to feel that any automated query methodology was actually reliable. If it included articles that clearly shouldn't be on such a list at an unacceptable rate (lets take Charles Kettle (cricketer) as an example) then I would think that the whole list should be rejected and the operator might find that they have the ability to submit such lists removed via ANI or similar. We can't get this wrong - this isn't just about removing 93,000 Lugnuts articles, it's about AfD at scale going forward. Blue Square Thing (talk) 16:34, 11 November 2022 (UTC)

Maybe if we should add a note, saying that while a quarry query is acceptable the proposing editor would be required to post a copy of the list in the proposal for ease of access?

I want to ensure the process to handle these articles is kept simple; we move them to draft space, and interested editors can handle them as appropriate. This would include returning them to article space as a redirect pointing to the parent location. There is no need to complicate the process when once draftified WP:BOLD actions will be allowed and encouraged.

included articles that clearly shouldn't be on such a list at an unacceptable rate; agreed. That is why the 5% "false positive" limit is there. A few errors, given the scale, are acceptable; unacceptable rates would result in the list being rejected, and repeated nominations of lists with unacceptable rates would become a behavioural issue. BilledMammal (talk) 11:07, 14 November 2022 (UTC)

Proposal 4

Should we have a separate draftspace (with different time constraints) for holding mass-created stubs that do not meet some minimum criterion? (Adapted from the August workshop): Set up a watchlistable pseudo-draft space with a longer/indefinite time limit where mass-created articles that do not meet [CRITERION] can be moved. Then, have those drafts be eligible for AfD, but only some defined number are allowed to be nominated per week. Editors can also move articles out of pseudo-draftspace, but only if they meet the minimum requirements, the moved drafts go into NPP, and only some defined number can be moved per week beyond anything kept at AfD. Pseudo-draftspace articles would be categorized like we do with DELSORT and wikiprojects could transclude the current lists of drafts in their categories. JoelleJay (talk) 01:13, 7 November 2022 (UTC)

Ping @Flatscan who looked into the technical feasibility of this at the workshop. JoelleJay (talk) 01:15, 7 November 2022 (UTC)

Thanks for the ping. I mentioned WP:Controlling search engine indexing in my previous post that you linked. It happens automatically for all of User: space, but Wikipedia: will need something, probably a template. Flatscan (talk) 05:37, 11 November 2022 (UTC)

As per current practice, users can work on articles on subpages of their user page. If more than one person is working on an article, it can be hosted as a subpage under an appropriate WikiProject. The lifetime of the pages can be managed by the editor/WikiProject. isaacl (talk) 01:47, 7 November 2022 (UTC)

Yes, but for mass draftification it would be infeasible to host articles in userspace, and we would still need some way of categorizing such articles for them to be automatically put into appropriate project subpages. JoelleJay (talk) 20:31, 7 November 2022 (UTC)

Historically, keeping a work-in-progress draft around has been most fruitful when one or more editors are actively interested in working on the draft. If user X is the only one interested in working on the drafts in question, I don't see why it's infeasible for the drafts to be subpages of user X's user page. Yes, some kind of categorization is needed if the drafts are being given to active WikiProjects to work on, whether or not the pages are moved to subpages of the WikiProjects. To me, responsibility for the drafts is more clearly delineated by having them under the WikiProject, or the specific user working on them. isaacl (talk) 21:54, 7 November 2022 (UTC)

I don't think this is necessary; WikiProjects and users are already permitted to move articles from draftspace to their space. Noting this may be useful, however. BilledMammal (talk) 00:09, 8 November 2022 (UTC)

This would really be for instances where there are dozens or hundreds of articles that ought to be draftified, and so single users hosting the drafts in userspace wouldn't be particularly viable. The number one complaint people have against draftification is that it's a "backdoor to deletion", number two is that it's hard to collaborate since draftspace isn't very searchable/browsable. Any proposals in this RfC that even faintly loosen criteria around deletion/draftification will be opposed heavily by certain crowds, so having something like this as an option might be a compromise. JoelleJay (talk) 03:18, 8 November 2022 (UTC)

In my view, putting works-in-progress in user space or Wikipedia space solves both those problems, without any new code or rules. The pages will no longer be subject to the G13 criterion for deletion, and the editors who have shown interest in working on the potential articles can assume responsibility of making the works browsable by other editors interested in the relevant domains. isaacl (talk) 23:21, 8 November 2022 (UTC)

Yes, I agree, which is why I'm suggesting we develop some way of doing this at a larger scale than what is currently in practice. "Pseudo-draftspace" needn't be a new dedicated -space, per se, but we would at least need to create the framework that would make mass projectification/userfication viable system-wide. Like, make it an option for NPPers or an outcome at AfD. We'd also need some language governing when these "drafts" can be moved back to mainspace. JoelleJay (talk) 00:18, 9 November 2022 (UTC)

My intent with proposal three is to establish such a framework, though I am not certain whether it will find consensus and a more scaled-down option may be appropriate. BilledMammal (talk) 12:56, 14 November 2022 (UTC)

4.1

Make a template projects could place on drafts they are interested in working on that would delay/prevent G13 deletion. This would automatically add the drafts to a projectspace list. JoelleJay (talk) 20:33, 7 November 2022 (UTC)

If such a template is present, then the relevant WikiProject should be notified in the same way that the draft creator is for G13, etc. Thryduulf (talk) 23:44, 7 November 2022 (UTC)

Proposal 5

Proposal: add mass-creation as an example reason to WP:BUNDLE

Should we add creation of articles by the same editor using substantially the same sources and format as an example reason for bundling the resulting articles into a single AFD at WP:BUNDLE? FOARP (talk) 07:46, 7 November 2022 (UTC)

A series of articles on nearly identical manufactured products. is already included in BUNDLE. Also I do not believe that the mass creation alone would come through. I suggest a reason additional to that of mass creation to be included in the proposal. Paradise Chronicle (talk) 08:35, 7 November 2022 (UTC)

Paradise Chronicle - That's right, we already have a heading under which certain kinds of mass-created articles can be bundled. This is merely a proposal to generalise it. I honestly don't know why anyone would oppose this, it is only a proposal to give a specific heading under which articles made in the same format and using the same sources can be bundled. It does not prevent unbundling. FOARP (talk) 08:45, 7 November 2022 (UTC)

I think I'm with PC here: I'm not certain this is necessarily needed. It might actually be harmful - this might give the impression of limiting it to articles created by the same user etc... whereas all we need now is to have them "nearly identical", which I think is going to be the case in enough cases if they're produced by the same person with the same sorts of sources, aren't they? So I wouldn't have an issue bundling articles like Eugeniusz Waszkiewicz, Albert Langereis, Francisco António Real etc... right now - assuming the sources in the articles at least had been checked of course. I'm not sure it needs to clarified, does it? Blue Square Thing (talk) 17:47, 11 November 2022 (UTC)

Proposal 6

Proposal: Amend WP:BEFORE to clarify that "the expected amount of work that goes into doing a WP:BEFORE by the nominator is proportionate to the amount of work done to substantiate the notability of the subject of the article concerned by other editors (such as the creator) prior to it being brought to AFD, as evidenced by the sources present in the article". FOARP (talk) 12:37, 7 November 2022 (UTC)

The idea here is that a failure to, for example, search dead-tree archives in Norwegian, should not longer be used as a reason to cast aspersions on the WP:BEFORE of someone nominating a series of articles created solely from sports-reference.com listings. This, of course, does not prevent people with access to archives doing their own search and finding support for notability where it may exist. FOARP (talk) 12:37, 7 November 2022 (UTC)

I am generally in favour of reducing some of the demand with BEFORE searches in mass-bundle cases (see elsewhere on this page), but isn't this impossible to judge? I'm not sure the wording works, does it? I suspect we'd be better off saying that only the major sources included in the article need to be checked - so, to use my favourite example on this page, Lalit Yadav (Delhi cricketer) has seven sources on it, three of which are scorecards so I'd immediately ignore. Four might be interesting, but it only takes seconds to check the profile one, scroll down and see that there's no doubt at all as to his notability - multiple, coverage etc... I just see that as more practical than the specific proposal here. Blue Square Thing (talk) 17:54, 11 November 2022 (UTC)

This is simply to give freedom to closers and people reviewing at AFD not to, in every case, have to do the full BEFORE where the amount of work done by the article creator was so little that it is simply pointless. I do not think the kind of rigid proposals made elsewhere on this page are appropriate as they are in some cases too much whilst in others not enough. FOARP (talk) 18:26, 12 November 2022 (UTC)

I get the intention, and I'm happy to agree that there are many cases where we should be reducing the requirement, so long as there are safeguards in place (someone gets to check the list in a reasonable timeframe and can ping any out that need further consideration etc...). That's all consistent with the requirement to reach a compromise position and makes sense. I'm just not sure, with the experience of the creation RfC, that this is structured enough to pass is it? The idea's good, but I can see people opposing it as unworkable, undefined etc... Blue Square Thing (talk) 08:19, 13 November 2022 (UTC)

Blue Square Thing - The full AFD process (which typically runs 2-3 weeks BTW, and Wikiprojects are automatically pinged on) is still gone through and can still be DELREV'd where needed (nb., the protections and review for RFCs are not so well-developed, which makes me wonder about proposals to use RFCs). My experience of the creation RFC is (assuming good faith here that the opposes weren't just finding any reason to oppose) that detail only served to draw resistance/nit-picking, and anyway that rigid proposals are unlikely to be workable where an all-new system is being proposed. I think good proposals still leave things in the hands of the editors. FOARP (talk) 09:41, 14 November 2022 (UTC)

I think you need to say that the nominator should at least check the major sources already included in the article. I suppose this is probably implied anyway: if you looked at an article such as Edwin Anderson (sport shooter) (currently at AfD) I think you *must* check that either link doesn't have something obvious in it to raise a concern that there may be more obvious notability. I'd argue that the creator must have checked at least one of those in order to get the details for the article. If this is what you mean, then isn't it clearer to say that? Blue Square Thing (talk) 10:45, 14 November 2022 (UTC)

This is intended more as a guide to what is and is not a reasonable BEFORE in all circumstances including (but not only) bundled nominations. Some have raised the idea that, because dead-tree/foreign-language archives can in some cases contain relevant material, that not searching these means that BEFORE has been failed. This is to clarify that where the creator (and any others who worked on the article) did not bother to do that kind of searching (e.g., where the only sources are electronic database ones) then there is no expectation on the nominator to go beyond this - they can, but there is no need to. Of course, where the article is one that more effort went into, and which includes sources that show more work was done in, then a BEFORE might normally be expected to have gone further.

I take your point that the original wording did not quite do this and have added some new bits in underline. FOARP (talk) 11:43, 14 November 2022 (UTC)

My problem with this is that it could be taken as endorsing the point of view that the search suggested at WP:BEFORE is obligatory rather than just a suggestion, which contradicts WP:BURDEN. The only people required to perform searches during or before deletion discussions are the people who wish to retain the article, fullstop. There is no requirement - none - to perform even the most cursory search before nominating an article for deletion, nor should there ever be; and I would strenuously oppose any wording that could be further misconstrued to imply that such a burden exists. --Aquillion (talk) 09:35, 21 November 2022 (UTC)

Thanks Aquillion. Two responses to this:

1) People are already being regularly dragged to ANI for allegedly "failing" to do BEFORE, and,

2) This amendment does nothing to change the degree to which BEFORE is mandatory or not - it only gives an amount of work expected when it is done. Obviously, where no work was ever done to establish the notability of an article, I would say the BEFORE should be proportionate to that, and zero multiplied by anything is still zero.

Your proposal 13 is helpful and appears to offer a way forward to clarifying whether BEFORE is obligatory before nomination or not. I intend to !vote that it should not be mandatory, but if the result turns out to be the opposite then something like this proposal is needed. FOARP (talk) 11:20, 21 November 2022 (UTC)

1) People are already being regularly dragged to ANI for allegedly "failing" to do BEFORE - never, so far as I know, successfully. If we're going to hash out the debates over WP:BEFORE we should start with the basic question of whether it's obligatory or not. I do not think there's a consensus to make it obligatory (and I categorically reject any arguments that that's the way it already is or somehow represents the status quo; a consensus would be needed to make it a hard requirement, which has never been produced. AFAIK it has clearly been contested every time it has been seriously discussed.) --Aquillion (talk) 11:29, 21 November 2022 (UTC)

Aquillion - As I said, I don't think it should be/is obligatory, just strongly advisable, but that doesn't touch on this issue, which is just to describe an expected (i.e., not obligatory) level of effort when doing it. Your proposal 13 will decide whether it is or isn't, and if it isn't this proposal will be governed by that. On the other hand if it is found to be obligatory, but this proposal passes, then the bar for mass-created articles will have been set suitably low since the level of effort put in by the creator will have been simply to look at a single database for the whole set of articles and the level of BEFORE expected will be only similar levels of effort (e.g., showing that the database does not sustain notability). FOARP (talk) 09:59, 23 November 2022 (UTC)

Proposal 7

Proposal: All nominations for the deletion of a bundle of 10 or more pages must be preceded by a discussion involving at least two uninvolved editors (excluding the nominator) about whether it is an appropriate bundle. Discussions of smaller bundles are allowed but optional. The discussion is explicitly without prejudice to whether the articles should be kept or deleted.

Nominating a bundle of 10 or more pages for deletion that does not have consensus at such a discussion is grounds for the discussion to be closed as speedy keep without prejudice to smaller nominations. An editor repeatedly nominating such bundles without consensus that the bundle is appropriate may be sanctioned, which may include (but is not limited to) a ban from nominating (bundles of) pages (of a certain type) for deletion. Thryduulf (talk) 13:20, 7 November 2022 (UTC)

How would we determine if the two editors were "uninvolved"? Levivich (talk) 16:49, 8 November 2022 (UTC)

I'd say that in this context involved editors would be

• Editors who have have significantly contributed to (a substantial portion of) the articles
• The editor who wishes to nominate them for deletion
• Anyone working with either of the above to create or review a set of articles that includes (a significant portion of) the articles in the proposed bundle.

The idea is to get an unbiased opinion about whether or not the proposed bundle would be likely or unlikely to be a trainwreck if nominated at AfD. ie. can they be meaningfully and fairly judged as a set at a 7-day AfD? The ideal people to do this are those who are uninvolved with the specific articles but familiar with the topic area. Thryduulf (talk) 22:23, 8 November 2022 (UTC)

This seems pretty unrealistic - AFD just isn't so well attended and many mass-deletions are uncontroversially bundled (see the Iranian village examples). This excludes people who want the articles deleted from the discussion so, seems that just a couple of people who want to keep them can block deletion regardless of how many favour deletion? Inventing an entirely new process especially for large bundles is also pretty WP:CREEP-esque. FOARP (talk) 11:54, 10 November 2022 (UTC)

Nobody would be able to block deletion discussions, indeed such a discussion would facilitate deletion because it would prevent trainwrecks. The discussion is explicitly not about keep or delete (or alternatives to deletion), solely about "will nominating this bundle at AfD result in a trainwreck?". It is not a replacement for AfD but a precursor to it - think of it like a pre-screen for AfD, breaking up the clogs that would block the system into manageable chunks that can pass through cleanly. If the answer is "no, this wont be a trainwreck" then it will proceed to AfD without an issue, if the answer is "this is likely to be a trainwreck" then it will be split into multiple AfDs that wont.

Something like the Iranian villages example, where a very large bundle was uncontroversial would not be hindered by this process - it would be uncontroversially agreed that it was a good bundle and then the nominator could take it on to AfD, where the discussion can focus on whether they should all be kept/deleted/merged/etc rather than that and whether they can be fairly treated as a set

The reason for excluding the nominator and someone working with them is that the goal of the discussion is getting independent assessment of whether the bundle will lead to a trainwreck or not, those proposing the bundle are independent and they aren't excluded from the discussion, just excluded from the count of independent voices (because they aren't independent).

The reason we are here is because AfD can't handle things as they are, so something needs to change - it is not creepy to suggest a change to fix something that is broken. Thryduulf (talk) 03:02, 11 November 2022 (UTC)

I can see the aim here, although I'm not sure that the number (10) isn't too small. I might be able to go to at least 25 without worrying too much. Or make this a suggestion? Provide a service - a noticeboard or so on perhaps? - so that a sensible bundle can be arrived at? My biggest concern with bundles is that really diverse articles get bundled together at times - anything that encourages sensible bundling of really similar articles is a good thing. But I'm not sure that this will necessarily get us there - and I imagine that a common sense proposal starting with "editors are encouraged..." will get rejected very quickly as not being something that can be enforced... Blue Square Thing (talk) 17:59, 11 November 2022 (UTC)

Proposal 8

Proposal: Where a single nomination at AfD bundles 10 or more articles, and there agreement prior to the nomination from at least two uninvolved editors other than the nominator that the articles are sufficiently similar that bundling is appropriate, the requirement for a WP:BEFORE is reduced to (a) examining the sources present in the articles, and (b) attempting to find sources for a representative sample of them. A representative sample is defined as the greater of (i) 5 articles, or (ii) 10% of the number nominated (rounded to the nearest whole number). Thryduulf (talk) 13:20, 7 November 2022 (UTC)

I can certainly support the a) and b) bits - it seems a reasonable compromise (yeah, *that* word). As above, I might place the value higher than 10 (25 perhaps - I don't know what bundle says; I'm not averse to starting with a number like that and tweaking it with experience). I would think that if we removed the bit about agreement from two editors - which I think might be problematic - that this might be workable in a cut down form? Blue Square Thing (talk) 18:02, 11 November 2022 (UTC)

• The search suggested at WP:BEFORE is not and has never been a requirement (per WP:BURDEN, which is very clear about where the formal obligation for sourcing resides), and I would strenuously oppose any wording that could be misconstrued as making it one. We cannot "reduce" a requirement that does not exist. --Aquillion (talk) 09:37, 21 November 2022 (UTC)

Proposal 9

Where editors want to discuss a bundle of 100 or more articles, they should use RfC rather than AfD. A full 30-day RfC should take place, advertised on the relevant Wikiprojects.—S Marshall T/C 16:11, 7 November 2022 (UTC)

9.1

Only one such RfC per Wikiproject can be open at a time.

9.2

Where the disputed articles are BLPs, a "no consensus" outcome at the RfC means they are moved to project-space (not draftified, but moved to subpages of the RfC page). Any good faith editor in good standing can move such an article back to mainspace on their own authority.

9.3

Where editors want to discuss a large bundle, tentatively defined as 500 or more articles, this RfC must be advertised on WP:CENT as well. No more than one large bundle RfC can be open at a time.

Empowering Wikiprojects to block deletion, when Wikiprojects are the source of much mass-creation (see particularly the whole WP:NOLY and WP:NCRIC farragoes) seems unwise. FOARP (talk) 11:57, 10 November 2022 (UTC)

Is notifying the Wikiproject your only objection to this idea?—S Marshall T/C 14:40, 10 November 2022 (UTC)

No, actually I'm perfectly OK with that. It's only having one discussion per Wikiproject open at a time that's the specific issue I'm talking about here - something that gives a Wikiproject the power to act as a jamming point - but not the only one. To expand on other issues: there's the WP:CREEP issue, the use of RFC rather than AFD where people have more experience with deletion issues, the potential to bog down CENT which only really has space for one of these discussions at any one time, the need for a full 30 day discussion where many large-scale deletions are uncontroversial (see the Iranian village cases), etc. Happy to discuss further. FOARP (talk) 14:49, 10 November 2022 (UTC)

Ok. The intention here is to make a workable plan for AfD *at scale*, i.e. one that could realistically review 93,000 Lugnuts BLPs in a somewhat acceptable timescale. I'm thinking it's realistic to allow maybe 5 years for that project? Which means we have to be able to talk about 1,500 to 2,000 Lugnuts articles a month. AfD cannot and will not cope with that kind of volume, so we have to use an alternative. The limit per Wikiproject is there because if you put a bundle of 1,000 Lugnuts-created Olympian stubs up for deletion on Tuesday, it's not then fair for me to put a bundle of 1,000 Lugnuts-created Olympian stubs up for deletion on Wednesday: we don't have the people to do the reviewing.—S Marshall T/C 16:12, 10 November 2022 (UTC)

If only Lugnuts was the only ex-editor whose stubs needed clean up, and Olympics the only Wikiproject that would need to be pinged on this work! Both Lugnuts and Carlossuarez46 made thousands of Geostubs, all of which are going to involve pinging multiple country-projects, country-projects that would also get pinged for other issues because of e.g., sportspeople from them. Lugnuts' Olympics stubs will involve dozens of Wikiprojects (the Wikiprojects of each country, the Olympics and sports wikiprojects, and the Wikiproject for each sport). Cleaning up sports bios for Turkey this month? OK, so no cleaning up Turkish village stubs then? I get the intent, but it seems unworkably rigid and requires a slew of new procedures.

For pure sanity's sake, mass TNT is just a much better way to go, not treating cookie-cutter stubs created in a matter of seconds based on the same source as being each an individual special snowflake. We've already seen this approach being deployed successfully and without any great controversy against the worst of the Iranian "village" stubs - once people understood that the source was bad then even people who would normally balk at mass-deletion were OK with it. Obviously there's other areas where people are much more invested in the subject, but that shouldn't stop us from going more quickly where possible. FOARP (talk) 16:42, 10 November 2022 (UTC)

Okay, well, this RfC is about AfD at scale. If we can get consensus for mass TNT, then, great, and AfD at scale is no longer needed. I think that might be a bit unrealistic though.

I propose that where several Wikiprojects are in scope, the consultation requirements apply only to the most relevant one.—S Marshall T/C 22:16, 10 November 2022 (UTC)

@FOARP: Part of the issue here is that we're not just limiting this to any set of editors. This is a way forward for the long term - and to apply to sets of articles where sources are reliable, just not very detailed, rather than "bad" (I don't know the details, but I can imagine). I'm not sure that TNT is good enough here - we have things like attribution and so on to consider Blue Square Thing (talk) 18:11, 11 November 2022 (UTC)

A specific point to note, is that there seems to have been a wide-ranging and long-lasting consensus that sorts of articles created by editors involved in these sorts of projects were, at the time, notable - certainly there are AfD going back years where articles were kept and I think most people would agree that a consensus had formed. So, I can find examples of AfD from 2008, 2010, 2012, 2014 etc... and right up to this (note Lugnuts' contribution) and then this, err, discussion in 2018. Consensus has changed since then of course, but I'm not sure that using the fact that some editors at some wikiprojects in the past worked with the consensus that seemed to exist at the time is really a justification for calling them a farrago is really a fair representation of current views (represented by, such as, this recent unanimous AfD). We should inform wikiprojects. If editors are obstructive unreasonably, ANI is there. We should use it. Blue Square Thing (talk) 18:32, 11 November 2022 (UTC)

Blue Square Thing - NCRIC and NOLY were examples of where a local consensus amongst Wikiproject enthusiasts was laundered into something that it wasn't (an overarching policy/guideline) through repetition at AFD. We should be moving towards wider involvement, not giving Wikiprojects special powers. FOARP (talk) 09:51, 14 November 2022 (UTC)

There are lots of examples where it's true, though, that AfD over the years produced a much broader consensus than a couple of people agreeing something at a wiki project. Schools, railway stations, populated places etc... Sure, wikiprojects shouldn't have special powers, but you'll get better guidance on what might or might not be sourceable if you ask them. And asking people to check through the lists produced is a really sensible idea - we're not aiming to delete obviously notable articles here, are we? Projects are more likely to have people who can judge quickly and who are interested in checking these lists of, potentially, thousands of articles. Blue Square Thing (talk) 10:39, 14 November 2022 (UTC)

Blue Square Thing - Things got out of control. Somehow things that were only ever supposed to be guidelines for what would *likely* be considered notable became "sure, create thousands and thousands of articles that only just meet this guideline but not the underlying policy, and no-one can ever challenge their notability, indeed if you do so repeatedly we're going to try to take you to ANI". The same thing happened on GEOLAND as well, which was, again, only ever supposed to be that articles about settlements were "typically presumed to be notable" and never intended to be just a licence to create tens of thousands of articles about, say, minor geographical features in Antartica (this is definitely ripe for a NOLY-style dethroning BTW). FOARP (talk) 13:47, 17 November 2022 (UTC)

I could get behind something like this. In terms of the numbers, frequency of nomination etc... I think we're better off starting with a lower number, such as the ones proposed here, and then seeing, with experience, how we go in different areas and with different sets of articles. We might also be able to reduce the 30-day time period in some cases (remind me, someone can call for an RfC to be closed if there's obvious agreement, yes?) or with experience. Blue Square Thing (talk) 18:11, 11 November 2022 (UTC)

Proposal 10

Proposal - We agree with the statement that: "there are certain sets of articles for which WP:TNT en masse is an appropriate response, and a consensus in favour of doing so at a suitably-widely advertised AFD discussion listing those articles is sufficient grounds to do so". FOARP (talk) 16:55, 10 November 2022 (UTC)

The purpose of this is to simply protect ourselves against the time-sink that mass-created stubs of a particularly-low quality represent. Some people have cast doubt on the idea that mass-deletion even could be done. This affirms that, where there is a consensus to do so in a widely-advertised discussion, then this can be done. FOARP (talk) 16:55, 10 November 2022 (UTC)

In regards to WP:TNT I agree with FOARP as I believe, a redlink will encourage more likely for a creation of the originally intended article than a redirect. With the creation of an article from a redlink your name gets added to the article creators and the editor also receives the notifications connected with the article creation.Paradise Chronicle (talk) 07:28, 11 November 2022 (UTC)

Thanks Paradise Chronicle. It would be great if there were some data either way on this, though my instinct is that red-links are more attractive than stubs (but I've expanded/created both). Either way, regardless of whatever other more intensive processes are created for some sorts of articles, a simple TNT of an arbitrarily large set of articles should always be an option. WP:DELREV is always present as a back up. FOARP (talk) 08:34, 11 November 2022 (UTC)

I'd be concerned about redlinks simply leading to the re-creation of probably not very notable articles actually. In that sense, redirection strikes me as better - people have to have something to add rather than just do a similar sort of thing to what was done before. And potentially worse (look at some of the AfCs we get...). And in some areas of the project redirection is a well established AtD. Best not to lose sight of that. Blue Square Thing (talk) 18:17, 11 November 2022 (UTC)

The suggestion that there are certain sets of articles where an approach based on really rapid mass deletion is appropriate is totally reasonable. I'm not sure it's necessary to actually agree this - the issue will be on what grounds I imagine! Blue Square Thing (talk) 18:17, 11 November 2022 (UTC)

It is, unfortunately, necessary, I've seen numerous AFDs closed by closers who simply refused to carry out the clear consensus in favour of mass-deletion/redirecting due to formalities. Here's an example - note how 95%+ have been red-linked since that RFC, with half of the remaining blue-links being redirects (e.g., Camp Spaulding, which is now a link to Lake Spaulding Dam, Canebrake (former town) which now redirects to the actual town of Canebrake, Crystal Springs which now redirects to Sanitarium etc.) and even the ones that were kept were basically re-written and might as well have been new articles. This would act as a reminder that, yes, mass-deletion is an appropriate response to some sets of articles, and yes that deletion should be carried out where there is a clear consensus to do so. FOARP (talk) 09:41, 14 November 2022 (UTC)

As I say, I imagine that the issue is more likely to be the practicalities of dealing with this sort of thing - I think you might fall down on simply saying "at AfD" or if someone raises the concern that this can be applied too broadly. Yes, there are sets of articles that most people could agree about. But what if someone brings 3,000 to AfD all at once? How do we deal with *that*? Blue Square Thing (talk) 10:32, 14 November 2022 (UTC)

The same way you'd deal with any other AFD - assess the consensus, and if it is in favour of deletion, then delete. There should be no set of articles too big for deletion if there is a sufficiently-strong consensus in favour of doing so. The technicalities of carrying out the deletion would be an issue for the closing Admin to work out with support from the AN board and others. DELREV is always there as a forum to challenge the deletion. FOARP (talk) 10:40, 14 November 2022 (UTC)

And that's the problem: we're being asked to discuss deletion at scale. That suggests that there are concerns or issues with nominating too many articles at the same time. If proposals ignore the fact that there are concerns, I don't think they have a chance of passing. Wait till we have 10,000 articles about one subject at AfD at once and you'll end up with the nominators at ANI being told to calm down and spread them out. The point here, I **think**, is that we need to be finding ways forward that allow us to deal with that, not simply say, "yeah, it's fine, nom as many as you like without doing a proper BEFORE". I might be wrong, but I thought that was the intention. Blue Square Thing (talk) 10:51, 14 November 2022 (UTC)

There is no magical wand that can be waved over people who simply do not think that deletion should happen in a specific case - only a consensus one way or the other can settle the issue. If someone does nom 10k articles (which would still be less than the largest deletion made in the Iranian "villages" case) then that should just be handled as a normal AFD case. FOARP (talk) 11:00, 14 November 2022 (UTC)

The key with the Iranian villages was that it was a clearly defined set that pretty much everybody agreed was appropriate to discuss as a single unit (whether they favoured deletion or not), at least in part because articles where other considerations apply (e.g. significant content added by others) had been explicitly noted and removed from the discussion. This is the sort of thing I'm trying to get at in Proposal 7 - if you do the preparation right and get someone to check your work, then the actual nomination will go much more smoothly. Thryduulf (talk) 15:43, 14 November 2022 (UTC)

So, sounds like there are such article-sets for which a straight TNT is appropriate. This is not about smooth or rough nominations: this is about closers hesitating to close as "delete all" AFDs where there is a clear consensus to delete, and giving an instruction from the community that, yes, that is what they should. FOARP (talk) 09:30, 21 November 2022 (UTC)

Proposal 11

Proposal - Where one or more editors feel it to be appropriate to nominate a set of articles for deletion, it is not reasonable to expect other editors to assess that set in a shorter timespan than that over which they were created, nor in any circumstances, to expect them to be assessed in fewer than 7 days. For example if a set of articles was created over 6 months then it is unreasonable to require editors to assess them all over less than six months; if the set was created over six years then taking 6 years to evaluate them all is reasonable.

This is a statement of principal to set reasonable expectations of speed. It does not set any rules or limits, including no minimum or maximum times - the 7 days just says that even if the articles were created in 24 hours its reasonable to allow a full AfD, but it doesn't preclude early closures (or even speedy deletion for those articles that meet a criterion) if the consensus is clear. Some evaluations will take longer (e.g. 5000 articles created in 10 days will likely take longer), others will take shorter (e.g. 20 articles created in 30 days), but an approximate 1:1 timescale is not too slow. If there are some articles within a set that need to be dealt with quicker than that for some reason, then nominate them first. Thryduulf (talk) 03:20, 11 November 2022 (UTC)

Some sets of contentless stub articles were created over ~10 years. The content of the articles is far more important than however many days the creator spent cutting and pasting content from a database into Wikipedia. FOARP (talk) 08:38, 11 November 2022 (UTC)

Thryduulf, do you feel this should apply to biographies of living people as well as the geostubs?—S Marshall T/C 09:35, 11 November 2022 (UTC)

As noted this is intended as a general guideline that sets general expectations for everything. It does not set limits and does not impose restrictions. If there are sets of articles created over ~10 years then taking ~10 years to assess them all is not unreasonable. If it happens quicker than that then there is no problem, the problem is demanding other editors work quicker on assessing them than it took to create them. If there are articles within a set that are higher priority to assess then assess them first. Thryduulf (talk) 11:50, 11 November 2022 (UTC)

Saying 10 years is a reasonable timeframe for assessment because 1000 such articles were created (commonly at a clip of <2 minutes per article) over a 10 year span makes no sense because overall creation time is less than a couple of days. For the articles concerned, reasonably matching the time for assessment to the time spent creating would allow for little more than glancing at the article and the database source it was copied from, and assessing against the relevant SNG/GNG/other – frankly, this is sufficient. _wjemather^{please leave a message...} 13:13, 11 November 2022 (UTC)

Even as a statement of principle with no limit in practice, I would still oppose this. The task of reviewing all these articles is monumental. The project's failures of recruitment and retention mean that our resources of editor time are very limited nowadays. I feel that we need to be very realistic about how many poorly sourced biographies of living people we're dealing with here.—S Marshall T/C 13:15, 11 November 2022 (UTC)

Proposal 12

Disruptive mass creation in violation of consensus, and of the policy on bot-like editing, is added to Wikipedia's list of reasons for deletion. The reasoning is the same as for proposal 6: cleaning up each non-notable article shouldn't require more effort than that employed in creating it. Avilich (talk) 21:42, 18 November 2022 (UTC) edited Avilich (talk) 02:38, 22 November 2022 (UTC)

Simply being mass created (with or without consensus) should never be a reason for deletion - it must also be the case that the articles are about clearly non-notable topics and/or are of such poor quality that improvement in place is not realistic. As has been repeatedly noted, it is possible to mass-create quality articles about notable topics. Thryduulf (talk) 14:01, 21 November 2022 (UTC)

That's why I explicitly added 'in violation of the policy on bot-like editing'. If the mass-article is already notable by virtue of an SNG, then the meatbot policy wasn't violated because the guideline explicitly allows the creation of such articles. On the other hand, if the mass-created articles in question are about unverifiable Iranian pseudo-villages or non-notable sportspeople, then this proposal creates a policy basis for dealing with the situation. Avilich (talk) 14:44, 21 November 2022 (UTC)

The policy on bot-like editing contains no references to notability, SNGs, the GNG or anything similar. Thryduulf (talk) 16:18, 21 November 2022 (UTC)

the policy on bot-like editing

Bot-like editing Human editors are expected to pay attention to the edits they make, and ensure that they do not sacrifice quality in the pursuit of speed or quantity. For the purpose of dispute resolution, it is irrelevant whether high-speed or large-scale edits that a) are contrary to consensus or b) cause errors an attentive human would not make are actually being performed by a bot, by a human assisted by a script, or even by a human without any programmatic assistance. No matter the method, the disruptive editing must stop or the user may end up blocked. However, merely editing quickly, particularly for a short time, is not by itself disruptive. Editors who choose to use semi-automated tools to assist their editing should be aware that processes which operate at higher speeds, with a higher volume of edits, or with less human involvement are more likely to be treated as bots. If there is any doubt, you should make a bot approval request. In such cases, the Bot Approvals Group will determine whether the full approval process and a separate bot account are necessary.

It's right there where it says "ensure that they do not sacrifice quality in pursuit of speed or quantity", "contrary to consensus", and "the disruptive editing must stop". It covers disruptive mass creations, and what is disruptive or not is defined by other policies and guidelines which have been sanctioned by community consensus, naturally including notability. The final draft could make explicit reference to "Disruptive mass creation" in order to make it clearer. Avilich (talk) 21:29, 21 November 2022 (UTC)

"Contrary to consensus" is clearly only referring to consensus to make rapid edits and neither the reference to quality nor "disruptive editing must stop" imply anything to do with notability, let alone SNGs. If you want your proposal to be in any way related to notability it must specify that. Thryduulf (talk) 00:18, 22 November 2022 (UTC)

"Consensus" and "disruptive editing" are unqualified, and so apply to whatever global rules exist regarding consensus and disruptive editing. Avilich (talk) 02:28, 22 November 2022 (UTC)

Since it apparently can't be helped that the linked policy will be reinterpreted on the basis of nonexistent wording, I added the words 'disruptive' and 'consensus' to the original proposal. Avilich (talk) 02:38, 22 November 2022 (UTC)

Proposal 13

Proposal: Should we make the search for sourcing that WP:BEFORE describes (and which WP:NEXIST strongly encourages) strictly obligatory, and should WP:BURDEN and WP:NEXIST be updated to reflect this? Or should WP:BEFORE be updated to make it more clear that that the search for sourcing is merely strongly suggested, as the text at WP:NEXIST suggests and as WP:BURDEN currently requires?

I have seen a number of people implying they believe it already is obligatory; but I do not think that that is a settled matter, so we should ask it as part of the RFC. See the discussion in the ArbCom case that this grew out of and my comments on the evidence page for reference; it is clear that there is currently no consensus supporting the interpretation that WP:BEFORE searches are obligatory, so it is plainly one of the core questions we need to resolve. This question is obviously vital to mass-deletions, as some of the discussions above imply (since the people who argue that WP:BEFORE searches are already a hard requirement would say that it is one of the primary impediments), but I, at least, would flatly and unconditionally oppose any wording that might imply that WP:BEFORE searches are obligatory; this is a question we need to settle first before we can refine the more detailed aspects of what BEFORE should say about deletions (mass- or otherwise.) --Aquillion (talk) 09:46, 21 November 2022 (UTC)

• It's important I think to note the difference between some sort of BEFORE being mandatory (which I believe it currently is and should be) and specifics of what constitutes a sufficient BEFORE (about which there is definitely no consensus). Thryduulf (talk) 14:05, 21 November 2022 (UTC)

• If you think you have the consensus to make WP:BEFORE mandatory, you should be able to demonstrate it here. If you think a previous consensus made it mandatory, you should be able to point to that consensus (but repeated discussions have failed to do so, and have constantly shown that despite the strident assertions by people who want to make it mandatory their proposal remains controversial.) But no, I flatly disagree with your assertion that it is currently mandatory; it is simply not, nor has it ever been, nor can it be without a consensus to rewrite WP:BURDEN (which I do not believe exists.) If you believe this discussion will reach a consensus that it is mandatory sufficient to modify WP:BURDEN and want to discuss precisely what that would mean, please make a separate point for that - I feel we need a simple, straightforward "is it mandatory" up-or-down discussion to determine if a consensus to make it mandatory exists so we can avoid these circular discussions about an aspect of policy on which there is clearly no consensus and where the text of multiple policies and guidelines seem to conflict. --Aquillion (talk) 20:05, 22 November 2022 (UTC)

The main argument for WP:BEFORE being made mandatory, if it isn't already, is that in order to claim that an article does not meet WP:GNG (or any other guideline), one must actually show that it does not meet WP:GNG. Otherwise, the claim is based on nothing; it might as well just be a !vote consisting solely of "delete." This does not conflict with WP:BURDEN -- it is the same line of logic, applied to a different space. Gnomingstuff (talk) 22:35, 22 November 2022 (UTC)

Notability is established by presenting reliable sources that support notability. If such sources have not been presented, then notability has not been established. It is the responsibility of anyone claiming a subject is notable to produce the reliable sources. We are not in the business of proving a subject is not notable, we are in the business of determining whether or not reliable sources exist that establish the notability of a subject. Proving that such sources do not exist is a logical fallacy. All you can show is that such sources have not been found. And until such sources are found, we don't shouldn't create an article for the subject. Donald Albury 22:53, 22 November 2022 (UTC) Edited 23:12, 22 November 2022 (UTC)

AfD isn't in the business of creating articles, either. It's in the business of reviewing ones that already exist. Notability is not determined by the state of the article itself (WP:ARTN, which is cited in WP:DELETE), but by the existence of reliable sources. And there is no way to determine whether or not reliable sources exist than to actually go looking for them and seeing whether they exist -- i.e., doing a WP:BEFORE search. Proving without a doubt that they don't exist is indeed a logical fallacy, which is why it is impossible to mandate any "sufficient" standard of WP:BEFORE. But in order to make any kind of credible argument, one has to present evidence. Gnomingstuff (talk) 23:22, 22 November 2022 (UTC)

"Notability is not determined by the state of the article itself" This is a principle that's unique to Wikipedia and may need to be reassessed as well. –dlthewave ☎ 13:54, 23 November 2022 (UTC)

Why does it matter whether a principle is unique to Wikipedia or not? We're only dealing with Wikipedia here. Reassessing that principle would be out of scope for this discussion, but if you do propose it somewhere you'll need to explain in detail what benefits it will bring to the encyclopaedia as I'm at a loss to think of any. Thryduulf (talk) 14:14, 23 November 2022 (UTC)

Per WP:BURDEN, the obligation to provide reliable sources to support their claim that an article is notable is entirely and solely on people who believe it to be notable; that is the burden to demonstrate verifiability lies with the editor who adds or restores material, - the article creator and people who want to keep it. Nothing there could reasonably be construed as allowing you to shift that burden to someone who wants to remove it or nominate it for deletion. Someone who wants to nominate an article for deletion has absolutely no obligation to "prove" a negative, nor could or should they ever have such an obligation; the simple fact that the article contains insufficient sourcing is already sufficient. The system you are proposing - where someone could create an article with no sourcing, then demand that other people do the footwork of looking for sourcing before their actions can be reversed, or indeed before they're even allowed to propose reversing it - is contrary to WP:FAIT and WP:BURDEN and is completely unworkable. If you think sources can be found, you're free to produce them in the AFD, but nominators have no obligation to WP:SATISFY your demands that they do your footwork for you. --Aquillion (talk) 19:52, 23 November 2022 (UTC)

The question is who has the burden to demonstrate that the sources support a page's claim of notability (a page that makes no claim of notability generally falls under WP:A7 or one of the variations on it and can be speedy-deleted without even going through AFD.) WP:BURDEN is clear that that burden falls on the people who add the claim or the people who want to retain the material in question, which means that it's entirely reasonable to send an article to AFD because they've failed to do so; someone doing so has no obligation to do that search for them. Otherwise, people could add uncited articles, with uncited claims of notability, then flatly refuse any attempts to remove them, while making no attempt to source their claims, and insist that other people do that search for them. In the most extreme cases people have actually tried to argue that editors can be sanctioned - and attempted, unsuccessfully, to seek sanctions at ANI - for not doing such a search before nominating things deletion, even when there's no indication that the article creators or any of the people who want to keep the article did such a search for their own claims of notability. That is plainly absurd and contrary to policy; the burden for that search is ultimately on the people who add or want to retain content, never on the people who want to remove or delete it. This is extremely relevant to AFDs at scale because, historically, people have added large numbers of articles at once with no or minimal efforts to find sourcing for their claims of notability, then falsely implied that WP:BURDEN means that anyone who wants to nominate their creations for deletion has some sort of obligation search for sources for them, individually, before nominating any of those articles for deletion; this makes article creation into a WP:FAIT situation where people can make take action and make wildly unsourced claims and then demand other people provide evidence before their uncited additions can be reversed. Things that are asserted without evidence can be dismissed without evidence; if someone creates an article, without providing sources that demonstrate notability, it is 100% correct and proper for me to be able to send that article to AFD without burdening me with a requirement to do that search on their behalf. In some situations it might be a good idea for me to do such a brief search to avoid wasting time on trivial cases, but I have no obligation to do so and we can't create an obligation to do so without creating WP:FAIT situations that turn WP:BURDEN on its head. --Aquillion (talk) 19:46, 23 November 2022 (UTC)

Proposal 14

No mass deletion process should be initiated, and certainly not approved, without the prior involvement and review of at least one person who is knowledgeable about the subject matter and can provide access to sources during, and preferably before, the duration of the AfD. If the subject matter is primarily covered in non-English-language media, the process should not be initiated or approved without the involvement someone who speaks that language.

Specifics: This does not imply any sort of request for formal credentials. It does not necessarily include the articles' creator -- if that person is even reachable anymore. It may include the WikiProjects on the subject, and these are probably the best place to find such people, but they don't have to come from there. It may, and probably should where relevant, include the non-English-language Wikipedia communities.

The rationale for this is simple: the more sweeping a change one wishes to make to the encyclopedia at once in a short period of time, the more critical it is to have some level of oversight by someone who knows what they are doing. AfD is populated largely by English-speaking generalists, and there is a great deal of misinformation that can be perpetuated by generalists. For example, on multiple AfDs on cricketers, John Pack Lambert attempted to use as a valid argument "he can't be important, we don't even know what his name is," when it is exceedingly common for people with patronymic names to go by initials only. The fact that, so far, many of these mass deletions have involved non-Western topics (Iranian villages, non-English-speaking cricketers), scientific subject matter (species), and other subjects that Wikipedia is regularly criticized in the media for neglecting means that there is a high possibility of the project looking really bad, perhaps in public, if this is not done right. Gnomingstuff (talk) 23:11, 22 November 2022 (UTC)

Similar to some of the other proposals here, this is "people who really love this topic-area should have a veto over deleting anything in it", which is a recipe for disaster. Typically these articles have been created by people who had little knowledge of the subject area (e.g., Carlossuarez46's Iranian villages) so one might ask why it is possible for someone having little knowledge to create the articles, but deletion requires approval from the fan-base of such articles.

The entire reason why these mass-deletions tend to occur in more obscure areas for the average editor is because mass-creations also tend to occur in them as these areas are less well-policed and editors are less likely to face opposition to negligent mass-creation from non-SIGCOV databases within them. FOARP (talk) 09:40, 23 November 2022 (UTC)

Contrary to FOARP, I like this idea. I don't think the wording is perfect yet the but the principle is solid. This is not about anybody having a veto over deletions, it's about ensuring that the basis for any nomination is factually correct. You (FOARP) bring up non-SIGCOV databases, this just ensures that someone with subject specific knowledge checks the nomination to ensure that the coverage is actually not-significant and that no articles about subjects that are very obviously notable get mixed in with articles about subjects that are truly not-notable. The goal is to delete articles that are either incorrect or unexpandable, not to delete as much as possible. Thryduulf (talk) 11:06, 23 November 2022 (UTC)

Determining whether something is or is not SIGCOV does not require subject-specific knowledge. I do not see why barriers should be placed in the way of deletion that do not apply to creation - anyone can create an article, anyone can nominate it for deletion. FOARP (talk) 12:17, 23 November 2022 (UTC)

Determining whether something is or is not significant coverage is very much subject-dependent knowledge in some cases - e.g. please show me what significant coverage of a 17th century Polish poet, a Polynesian religious rite, an Ethiopian cultural historian, a Peruvian plant geneticist, a town council in Uttar Pradesh and an oral tradition in Botswana looks like in all cases. The only "barriers" being placed in the way of deletion are those that will ensure that time is not wasted discussing the deletion of things that clearly should not be deleted. Say for example you find an article about a Georgian academic sourced only to a database - clearly delete you think, but then someone who can read Georgian sees that one of the entries in the database is actually an extensive list of the press coverage about the subject, all in Georgian, that demonstrates he was the pre-eminent scholar of the history of the Georgian language of his generation and meets the GNG in spades. If there are multiple such entries in the batch you nominate for deletion, then the discussion is going to be a train wreck having wasted the time of multiple editors. If you'd just checked beforehand with someone with subject specific knowledge they'd have separated out the notable from the non-notable, you could then nominate only the non-notable ones for deletion and they'd be gone in a week having generated no bad feeling, done no harm to the encyclopaedia and wasted nobody's time or effort. Thryduulf (talk) 14:35, 23 November 2022 (UTC)

• This seems like it would create an imbalance between the effort required for creation (no expertise needed, no need to understand what the name of the topic even means and in many cases, no demonstration of notability) and deletion (must track down a subject matter expert who reads the local language). As others have pointed out, you don't have to be an expert to assess whether or not the sources meet SIGCOV. –dlthewave ☎ 13:50, 23 November 2022 (UTC)

RfC format

Reading for the first time WP:ACAS (I'd not realised it had ever got from the planning stage to the discussion stage - which seems to have not lasted very long compared to the lead up to it!), I think we may be putting the cart before the horse here.

Shouldn't we be ironing out the scope, definitions, and format before getting into questions? Or perhaps having questions in different phases. At ACAS they seem to have failed to clearly define "at scale" or "mass creation" before spending a lot of energy on how to address. As far as I can see, the scope of the creations discussion comes simply down to the creation of articles by Bots. So what are we discussing here - the deletion of articles created by bots? The deletion of articles using bots?

Does the scope include specific issues and behaviours at AFD. Or even the failure of AFD itself, which is grossly underattended, and the results often depend which way the wind is blowing. Are we looking for, or at root causes? Why is AFD failing? Why is it underattended? How do we fix that? Do we need more clearlines, and firmer enforcement to end a lot of AFD debates? Do we need less? How do we evaluate all this through the lens of our 5 pillars - and some AFDs have failed ALL five pillars.

Sorry if this is stepping backwards - my understanding is that the workshopping only began today - and surely this is the first step. Nfitz (talk) 23:38, 6 November 2022 (UTC)

• Nfitz is right to point out that we need the scope, definitions and format. We've failed to reach consensus on what "article creation at scale" is, so I expect it to be hard to reach consensus on what "article deletion at scale" is. The one should be counterpart to the other. But the context and history give us clear parameters.
  Lugnuts created 93,000 and odd articles, and the community can't review them. As I see it, the two main considerations are:
  1) The community expects a full WP:BEFORE on each article. Well, even if you define a full WP:BEFORE as an online-only search, it still takes about ten minutes to do thoroughly. So if we allocate an editor to work full time, by which I mean 40 hours a week for 50 weeks of the year, it's going to take them just under eight years even to do the WP:BEFORE.
  2) AfDs soak up a bloody lot longer than 10 minutes of editor time. If someone does start nominating an article for deletion every ten minutes like this, the editors active in the relevant Wikiprojects will cry uncle, and who can blame them? Our system is buckling under the impact.
  As I see it, we need to generate a set of simplified rules and procedures that let us work through these backlogs in a reasonable and proportionate amount of editor time. We need rules of thumb that enable us to evaluate each article as quickly as Lugnuts could make them. This definitely means the rules of thumb will be quick and dirty. There will be collateral damage. But the alternative is to allow undersourced unreviewed articles in the mainspace, and it's policy that we can't do that with BLPs (as so many of Lugnuts' creations were).
  This principle of proportionality is the key. When articles have been mass-created in this way we should be able to review and where appropriate, remove, them with equal or less input of time and effort than it took Lugnuts to create them. If we don't come up with a workable way of doing this, we're unable to comply with the policies we enacted in 2004-5 whose purpose was to manage real harm being done to living people.—S Marshall T/C 00:18, 7 November 2022 (UTC)

I would suggest that one way to deal with the number of articles is to remove the need for a "full" before search and to limit it to checking the main references already in the article. There are plenty of cases where this would throw up an article where there are clearly sources available and which should be kept and expanded at some point - for example Lalit Yadav (Delhi cricketer) (hit the link in the infobox, scroll down - obviously in depth coverage exists). From working through a few hundred, this seems to take <30 seconds with this sort of article. Blue Square Thing (talk) 07:49, 7 November 2022 (UTC)

Thank you, that's an immensely helpful suggestion. I agree that when dealing with mass deletions we should streamline WP:BEFORE in this way.—S Marshall T/C 09:00, 7 November 2022 (UTC)

I'm concerned about waiving BEFORE without prior discussion. There are several retired/permanently blocked editors that (mass) created articles under a SNG long before WP:NSPORTS2022 which are only sourced to a database entry. When I skim their contributions, I can see that many of these articles are likely about non-notable sportspersons, but certainly not all of them. If BEFORE is waived, every single article they created can be nominated for deletion at once - which will completely overwhelm our AfD processes. We ought to find a process which gives interested editors time to evaluate those creations before hundreds (or thousands) of articles are pushed in AfD en masse (and isn't this the very concern that prompted this portion of the RfC?). Jogurney (talk) 23:12, 15 November 2022 (UTC)

@Jogurney: See #Proposal 3. It is intended to give interested editors time to review and improve these creations without overloading any process. BilledMammal (talk) 23:18, 15 November 2022 (UTC)

Thank you for the ping. I agree that something along the lines of Proposal 3 is a far superior approach to dealing with these types of creations than waiving BEFORE and using AfD. Jogurney (talk) 23:27, 15 November 2022 (UTC)

Deletionists paradise? Or am I dreaming?

Is this all about Lugnuts' pages or something else? If Lugnuts, he's poisoned the well pretty good. But, since I've just become aware of this entire process, through the NFL Wikiproject of all things wanting to get rid of perfectly good articles (and is this what that thing about destroying hundreds of Moon and Martian crater articles is all about?), and just scrolled through this page, and that's all I can say. Are we really talking about how to easily delete thousands or tens of thousands of pages or am I, hopefully, understanding my quick scroll entirely wrong? Thanks. Randy Kryn (talk) 03:08, 23 November 2022 (UTC)

You have not misunderstood - that's exactly what they're discussing here. I'm sure someone will be along presently to point out to you exactly why this is a Very Good Thing that no right-thinking person could possibly object to. Ingratis (talk) 07:27, 23 November 2022 (UTC)

There are some editors here who want to be able to mass-delete thousands of articles without even doing a basic BEFORE, there are some editors here who are trying to ensure that any process for mass deletions includes a large number of safeguards to ensure that as few mistakes get made as is possible. Thryduulf (talk) 11:13, 23 November 2022 (UTC)

Jesus Flubbin Christ (1858-1918). I had objected to the proposed deletion of the lunar and Martian craters and the like, but didn't know that was based on this kind of thing. Also seeing that thousands of pages about baseball players and people from other sports are either leaving us one by one, hundred by hundred, or being proposed to be deleted, and not understanding why. The crater pages, for example, are usually well done even at the stub level, with basic information and usually a good image or two, and to my untrained eye there is nothing broken about them. They and others are cited by the proper authorities in their field (baseball, for example, a player will usually have a reference to a baseball almanac or official listing of some kind). Not to be too long winded, but why doesn't someone in some kind of authority just say enough is enough and end this, one of those super admins who would dare to ignore all rules and tell everyone no. Lugnuts pages, that's another topic, which I supposed is discussed here as well, but mixing Lugnuts starters with anything else seems an end-around putting undue pressure on other regular long-existing articles. Randy Kryn (talk) 11:56, 23 November 2022 (UTC)

• Randy Kryn - It would be great if Lugnuts was the only example. Unfortunately he is not. Instead we have Carlossuarez46, Dr. Blofeld, and a number of others, their mass-creations simply remain around because they simply cannot be cleaned up under out present processes. The WP:NASTRO fails you mention are just another example - the decision on them was taken in 2012 yet they are still being cleaned up.

It was decided, long ago and according to very-well trafficked consensuses, that Wikipedia was not a database. This is the basis of all of this - Wikipedia is an encyclopaedia, not a hobby-site. FOARP (talk) 12:12, 23 November 2022 (UTC)

What hobby? An article on a lunar crater is about a thing, an off-world geographical site. An article about a ballplayer who played one professional game when they were 22 still covers its encyclopedic niche. All of those type of pages should be universally protected from deletion. The Lugnuts and other creations are a different topic, but this page seems to have mixed up the concerns and solutions with regular long-term pages which don't have specific editor-centric problems. Thanks for the reply, please understand that it's kind of a shock to the Wikipedian system to come across this discussion. Randy Kryn (talk) 12:31, 23 November 2022 (UTC)

Yeah, I don't get it either. Lunar craters are among the things I expect to find when I look in a general-purpose encyclopedia. For example, when I open volume 12 of my trusty print Britannica, I find an article on Tycho (pp. 82–83). Wikipedia having more articles about craters than Britannica isn't being "indiscriminate", it's being comprehensive. People say that we're not a "database", but the policy only mentions databases of song lyrics, and mostly argues for excluding them on copyright grounds. It looks like a couple bad experiences with articles sourced to untrustworthy databases has tarnished the reputation of any source to which that label can be applied. (I suppose any website with MySQL on the backend is disqualified now, too...) XOR'easter (talk) 15:08, 23 November 2022 (UTC)

Apparently we don't bother with "comprehensive" any longer - deletionist-Nuspeak has replaced it with ""completionist fantasy". Ingratis (talk) 18:52, 23 November 2022 (UTC)

• Sports biographies (created by Lugnuts and many others) are just one specific example of an area where tens of thousands of articles have been created according to a very inclusive SNG, but that SNG has since changed and the vast majority of the articles that were created no longer meet any notability guideline. These are also not "perfectly good articles", they are pseudo-articles – essentially database entries (which is unsurprising since they were simply copied from a database as the only source). _wjemather^{please leave a message...} 12:21, 23 November 2022 (UTC)

• An article about a crater on the Moon for example, if it doesn't have editor-specific problems, is an article about a geographical site, which by itself meets notability standards. What is a pseudo-article? Like the undead or something less/more scary? Randy Kryn (talk) 12:31, 23 November 2022 (UTC)

• A database entry converted into a sentence (or two) is not a genuine article; more so when they do not, or no longer, meet any notability guideline. The identity of the creator(s) of such quasi-articles is of no relevance when discussing the set, although problem sets are often created (or dominated by) by a single editor. _wjemather^{please leave a message...} 13:20, 23 November 2022 (UTC)

  A brief description of a sentence or two, on a notable subject is indeed "a genuine article", there is nothing false about it. In size, it may be classified as a stub, but it is still a legitimate and (at least so far) an accepted article. Loopy30 (talk) 16:31, 23 November 2022 (UTC)

  You are missing the point. _wjemather^{please leave a message...} 16:57, 23 November 2022 (UTC)

• There's no presumed notability for geographic features, so a crater on the moon would need to meet GNG. –dlthewave ☎ 13:39, 23 November 2022 (UTC)

Wikipedia:Notability (geographic features)#Geographic regions, areas and places (i.e., GEOLand) states that,

• "Populated, legally recognized places are typically presumed to be notable,...",
• "Populated places without legal recognition are considered on a case-by-case basis in accordance with the GNG."
• "Named natural features are often notable, provided information beyond statistics and coordinates is known to exist. This includes mountains, lakes, streams, islands, etc. The number of known sources should be considered to ensure there is enough verifiable content for an encyclopedic article."

Donald Albury 18:41, 23 November 2022 (UTC)

A moon crater is, by definition, not a "geographical site", and it is appropriately excluded by GEONAMES. Lunar features must meet GNG, otherwise we're just hosting a worse version of whichever NASA directory the information came from. JoelleJay (talk) 20:01, 23 November 2022 (UTC)

• See, my impression was that sports biographies are a problem because they are BLPs and craters on the Moon aren't. Nobody will sue or get hurt if a Moon crater article is inaccurate, but someone might on a biography. Thus 10,000 unmaintained Moon crater articles aren't necessarily a big problem, but 10,000 unmaintained biographies are. And consequently, you'll are more likely to see mass deletion nominations for the biographies than the Moon craters. Jo-Jo Eumerus (talk) 13:42, 23 November 2022 (UTC)

• Yes, this place has become a deletionists paradise (a.k.a. a wreck) since that wrongly-closed massive sports rfc earlier this year (at least in my area). It seems all one needs to do to get rid of an article is say "Delete - Fails GNG" whether it does or not. And to try to get an article kept, you have to spend lots of your time doing in-depth searches for SIGCOV (and adding them to the article, since apparently someone can't be notable unless sigcov is on their page) and if you find it, the deletionists can still win by saying "not sigcov" and the admin will side with them. Only if the coverage you find is accepted by the deletionists, which is very difficult to do, then the article may be kept (but only if you add the coverage to the article!) BeanieFan11 (talk) 15:44, 23 November 2022 (UTC)

  Can you give some examples for this type of behaviour? Lurking shadow (talk) 20:58, 23 November 2022 (UTC)