Wikipedia talk:Talk page guidelines

Follow-up: what to actually do with the over 2,200 instances of undetected vandalism in talk page archives?

The previous discussion on talk page archives closed with no consensus (in either direction) as to what is and is not allowed to be removed from talk page archives. But meanwhile, there is still undetected vandalism sitting around in talk page archives, right now, to the tune of thousands. The issue isn't abstract; since the discussion began I have found over 2,200 instances. (The number isn't exact because I probably made some counting errors.)

For obvious reasons, I would very much like to just fix these instead of posting about them -- I didn't dwell on this but some of this undetected vandalism happened to comments by editors who are still active, and who I'm sure would be surprised to learn that the archives falsely show them using homophobic slurs or discussing their masturbation habits. However, I don't know if I am even allowed to now. Despite what certain people have claimed I do actually listen to other people's input and do not enjoy being scolded for trying to improve the project so I am asking for input.

This is not a complete list of bad edits -- The list is broken out into categories because some people thought certain kinds of vandalism were more worth fixing than others, but it probably under-represents certain categories. (In particular, there's probably a lot of comment blanking/subtle vandalism that I've missed because it's hard to search for, almost all of the blanking on the list was found while tracking down other diffs.) Definitions of the terminology used in the list:

"Undetected vandalism": Refers to vandalism that happened before the talk page was archived. A lot of people seemed confused about this for some reason and I'm not sure why. For instance:

• January 1, 2005: Alice posts a comment to a talk page.
• February 5, 2007: Bob vandalizes Alice's comment.
• March 10, 2007: The talk page is archived, including Bob's vandalized version of Alice's comment.

"To other people's comments": Vandalism that makes it appear like someone said something that they didn't. This encompasses two cases:

• Alice posts a comment, and Bob edits the text to vandalize it.
• Alice posts a comment, and Bob vandalizes the page in a way that makes it appear that his vandalism is part of Alice's original comment. Usually this happens when there's insufficient whitespace and/or no signature.

"Standalone": The disruptive edit is a separate comment, not affecting surrounding comments, and not possible to confuse with someone else.

"Slurs": I don't need to explain what this is.

"Crude vandalism": Vandalism that is profane, scatological, etc.

"Blanking/meaning-changing": Removing other people's constructive comments without explanation or editing them to change their meaning in a substantive way (i.e., not typo fixes or cosmetic changes).

"Nonsense": Keysmashes etc.

"Self-insert vandalism": "Bob wuz here" type stuff.

"Off-topic/WP:NOTFORUM edits": A narrow definition of "off-topic," mostly this means drive-by "I like cats" stuff and other useless edits that would uncontroversially be removed per WP:NOTFORUM.

"Text-to-speech/LLM/homework/Siri edits": I don't know how to describe this but anyone who has been on a talk page in the past few years knows what this is. These edits only started happening in 2021, and I suspect it's people using talk pages as if they're ChatGPT/search/Siri/text-to-speech -- these edits will often put a literal school subject in the "Subject" line or a test question in the body, request the type of speech output ("soft voice") or language to translate to, or outright include prompts. We are getting bombarded with this shit and not catching it all in time.

"Other vandalism/other test edits": Disruptive edits that don't fall into other categories; "other vandalism" implies it's likely bad-faith. Gnomingstuff (talk) 01:13, 30 December 2024 (UTC)

• Thanks for your efforts but leave archives alone. A lot of junk and spam is posted to standard talk pages and you are welcome to remove that. However, an archive is an archive and not a place for people to tweak things. Johnuniq (talk) 02:16, 30 December 2024 (UTC)

  Please tell me when you became my boss. I would love to know. Gnomingstuff (talk) 17:38, 30 December 2024 (UTC)

  No one's claiming to be your boss, this isn't helpful. 331dot (talk) 19:05, 30 December 2024 (UTC)

• I'm with Johnuniq on this. Zero^talk 02:52, 30 December 2024 (UTC)
• Me too. Before we even START talking about this show me five instances of someone saying, "Zounds! Because of archived talk-page boo-boos, something bad happened!" Otherwise this is a solution in search of a problem. EEng 03:16, 30 December 2024 (UTC)

  You can see many more than five instances in the RfC above. Gnomingstuff (talk) 17:36, 30 December 2024 (UTC)

  No. EEng is saying that there is no evidence that something bad has happened as a consequence of junk in archives. By contrast, some of us think that gnoming archives is harmful because it confuses anyone trying to use the archive for its intended purpose of easily examining history. Further, if one person gnomes archives, others will follow to fix what they perceive to be layout and indenting problems, and more. Johnuniq (talk) 05:25, 31 December 2024 (UTC)

  Because of archived vandalism, I've lost various amounts of time trying to figure out older discussions at some points. Ymmv as to whether that is something bad or not. CMD (talk) 09:04, 31 December 2024 (UTC)

  Do you have a couple of examples? Johnuniq (talk) 09:32, 31 December 2024 (UTC)

  It's just one of those things that pops up every now and then. Flicking through my contribs the most recent one I found, although this was not vandalism but a pre-sinebot era, was looking at past discussions at Talk:Montserrat when it had its September RM. The oldest archived conversation was the first place this was discussed, and there a missing signature made it look like someone had written arguments for and against the move together. CMD (talk) 14:53, 31 December 2024 (UTC)

  @EEng and Johnuniq: Not all bad things in archives are because of vandalism. Please see this post by myself, which describes a situation where a number of archived threads became invisible, and although I could trace the problem edit, I could not find any evidence whatsoever that vandalism was applicable. Question: should I have left it alone (several threads remain hidden), or fixed it (so that they're all visible again)? --Redrose64 🦌 (talk) 01:48, 1 January 2025 (UTC)

  Yes, of course important wikitext fixes are wanted. Inconsequential markup should not be "fixed", but stuff like misplaced ref or broken HTML comments that hide text should be fixed. Johnuniq (talk) 01:53, 1 January 2025 (UTC)

  Indeed. But the OP's original laundry list went way past that narrow goal. EEng 21:39, 5 January 2025 (UTC)

• Revert/remove especially when it comes to other people's posts being vandalized. Yes archives are archives but they are supposed to be used to check prior discussions and consensus on the subject page so if they say something someone didn't say it could cause problems where people think a consensus is different to what it actually is. Simiarly why otherwise would we want to keep blatant vandalism in archives? It doesn't look good to us or the subject. I can understand leaving borderline off topic/NOTFORUM cases but we should probably still consider removing as we don't want to clutter the archives with irrelevant content. WP:BEENHERE probably applies. Crouch, Swale (talk) 18:59, 30 December 2024 (UTC)

  Pretty well agree with Crouch, Swale. If the vandalism is that clear-cut, it probably shouldn't be in archives. With the caveat that some vandalism-looking stuff may be intentionally left where it was - I reverted an over-enthusiastic usertalk patroller once when I disagreed with leaving an insult on my page. --SarekOfVulcan (talk) 19:08, 30 December 2024 (UTC)

  I agree with Crouch and Sarek. And I disagree with both the substance and the tone of the replies by Johnuniq, Zero, and EEng. If there's no consensus about something, then there is no consensus that you are forbidden to do it (as well as no consensus that you need to do it) – and people who get indignant over someone doing that thing that is not forbidden, on the basis that doing it is being a busy-body, need to take a look in the mirror. I think a reasonable rule of thumb for you to follow, that I think is consistent with the close, would be to revert any edits that were made after archiving that misleadingly alter what the talk page was at the time of archiving, if you feel like reverting them. As for vandalism that occurred before archiving, but well after the discussion was closed or petered out, same thing, I think. Stuff that was present around the time the discussion was closed, probably best to shrug it off and leave it. --Tryptofish (talk) 22:43, 30 December 2024 (UTC)

  Kindly keep comments about other peoples' "tone" to yourself. The usual thing to do when an RfC gets no consensus is to continue current practice, which in this case is that talk page archives are not edited further except in very limited circumstances (copyvio, blp violation, etc). Declaring that current practice can be changed when the RfC did not conclude that is a violation of the close. I'm particularly against some of Gnomingstuff's examples which involve judgement: slurs, NOTFORUM edits, and Text-to-speech/LLM/homework/Siri edits, none of which can be objectively defined or identified. People can remove NOTFORUM edits on talk pages in full view of other editors who are free to disagree on the removal, but doing it out of sight in an archive just on their own judgement is not acceptable. Zero^talk 02:40, 31 December 2024 (UTC)

  I stand by what I said. There's actually no consensus as to what current practice has been. --Tryptofish (talk) 20:33, 31 December 2024 (UTC)

  We know that generally we can't modify other's talk page comments so reverting vandalism to other's posts does not need a consensus rather there would need to be consensus against it. Similarly we know that we should revert vandalism on sight so there is no reason to require consensus to do so. I agree that there might not be consensus to remove borderline off topic/NOTFORUM cases though. Crouch, Swale (talk) 20:57, 31 December 2024 (UTC)

  Archives and crufts. We are in agreement then that you cannot disparage others. 2603:6000:C305:78DF:682F:E3D4:99D1:3755 (talk) 11:46, 6 January 2025 (UTC)

  I just find this whole thing very frustrating, because I feel as if I am trying to work with people, explain my position, be transparent about its scope, and gather policy-based reasons to support it -- the big ones being WP:VANDALISM (Upon discovering vandalism, revert such edits) and WP:5P3 (any contributions can and may be mercilessly edited). And in response I get endless status quo stonewalling (this whole saga started out as 3.3.3, segued into a lot of 3.3.2 and 3.4.3, and is now 3.3.4 out the wazoo).

  To respond to recent individual comments:

  As far as "current practice," I was reverting vandalism in archives for over a year before anyone complained. This was not because no one noticed -- for over a year, all of the feedback I got was positive. That's part of why the whole thing is so confusing.

  As far as "doing it out of sight," not only is that just literally false (all edits are traceable unless revdelled/oversighted), but it's one more contradiction in the mass of contradictions that this whole thing is. If archives are "out of sight," then you can't have the simultaneous argument that editing them clogs up people's watchlists (which, by definition, means that they're in sight).

  As far as edits being subjective, yes, that's an issue (although I think basically everyone would agree that the hard-R N-word is a slur, like, be serious here). A lot of the vaguer stuff is easier to identify en masse than individually; I've proposed an edit filter to catch some of it but the details still need some work. Gnomingstuff (talk) 20:54, 8 January 2025 (UTC)

  I've been noticing a small percentage of your fixes because they are on my watchlist, and they look fine to me. I'm genuinely sorry that some editors have been criticizing you, but I hope that you will not let it get you down, and I hope that you will keep going. --Tryptofish (talk) 23:38, 8 January 2025 (UTC)

  Gnomingstuff, although we have disagreed about some of the core issues on this topic, I just wanted to quote your comment of 20:54, 8 January above (diff):

  ...I feel as if I am trying to work with people, explain my position, be transparent about its scope, and gather policy-based reasons to support it

  because I just wanted to go on record to say that we are in whole-hearted agreement on this point: that is, you have indeed worked, explained, been transparent, and gathered P&G support as you saw fit, and I have nothing but commendations for your having done so. We are all on the same side here, namely: trying to improve the encyclopedia. It is of course inevitable that with numerous editors, there will not be 100% agreement on how best to do this, but I am certain that improvement is your goal, discussion is how you/we attempted to achieve consensus, and I for one, will be happy to work with you in the future if we happen to land on the same project, whether we are in agreement or not. (Personally, I tend to learn more from a good tussle, than from an echo chamber.) In the meantime, Happy trails! Mathglot (talk) 06:51, 28 January 2025 (UTC)

It seems to me that all this categorization of edits into "slurs" and "vulgar" and whatever is just parsing it way too fine. I would suggest that it's much simpler:

• If an edit attributes text to an editor that that editor did not in fact write, or alters what they did write, then that can be undone. We don't need to examine the content at all.
• If text violates BLP, copyright, or is illegal, then that can also be removed.
• Otherwise leave it alone. It's irrelevant whether it could have been removed pre-archival. It wasn't, so just leave it.

I suppose there might also be narrow exceptions for stuff that somehow messes up the software at a technical level or something, but other than that this seems pretty simple to me. --Trovatore (talk) 05:51, 27 January 2025 (UTC)

In Request for comment: Do the guidelines in WP:TPO also apply to archived talk pages? the closer (voorts) found no consensus, which I don't interpret as a go-ahead. To me it means the current Template:Archive instruction stands, and template texts are backed (I think) by the WP:TMP guideline. Peter Gulutzan (talk) 20:51, 27 January 2025 (UTC)

OK. I'm not arguing legalities. This is the approach that makes sense to me.

If you like, you can take it as a preview of what I would say in the next RFC. But even simpler, if Gnomingstuff were to edit archives in this fashion, I doubt it would be particularly controversial. --Trovatore (talk) 22:29, 27 January 2025 (UTC)

I believe the discussion and decision in Striking comments from banned sockpuppets and modifying archived comments are relevant so I hereby ping all its participants besides myself: Newslinger, XOR'easter, PackMecEng, Drmies, Nosebagbear, Doug Weller, Shar'abSalam, Reyk, Number 57, Serial Number 54129, Davey2010, Hasteur, Only in death, Rhododendrites, Eggishorn, Alanscottwalker, Pawnkingthree, Jayron32, Jauerback, BX, HandThatFeeds, AmandaNP. Peter Gulutzan (talk) 16:53, 17 February 2025 (UTC)

• I'm very much in agreement with Johnuniq and EEng - archives should be left alone irrespective of what's on them (with the exception that people who have modified other peoples comments before/after archive should be reverted), I'm not seeing a reason as to why we need to have issue over comments made 16 years ago, the world was a much different place in 2008 and as such we shouldn't rewrite or remove history just because a few people are offended by what's said.

Off-topic but now knowing 2008 was 16 years ago makes me feel so old, man those were the good ol' days though, life was worth living back then!. –Davey2010^Talk 17:19, 17 February 2025 (UTC)

It was indeed. But I think that serious BLP violations are an exception, in fact probably anything qualifying for ref/del and certainly anything that should be oversighted should be removed. Doug Weller talk 17:22, 17 February 2025 (UTC)

BLP violations should be removed and looked at for revdel. Racist comments/jokes/etc. probably should be removed (User:Davey2010, that's not rewriting/removing history: it's rewriting/removing racist shit). Striking through old sock comments in archives--I guess I'm OK if they are accompanied by a comment indicating why something was done and when. Drmies (talk) 18:21, 17 February 2025 (UTC)

Fair point, "rewrite or remove history" was the wrong choice of words here. Some things/jokes etc were acceptable in 2008 that aren't now but yeah racist comments weren't acceptable even back then so I do agree the racial shit should be removed, –Davey2010^Talk 19:19, 17 February 2025 (UTC)

I'd say if someone wants to edit an archive to correct vandalism, that's fine. Such an edit should link to the original source text, to verify it's being corrected, but that's about the extent of what's necessary. I do not envy anyone who takes up such a task at scale, but if someone had edited my text to make it say something other than what I'd intended, I'd like for the archived record to be correct. — The Hand That Feeds You:^Bite 18:07, 17 February 2025 (UTC)

• Talk page vandalism is talk page vandalism, and it should be undone even if that vandalism has been moved. I don't know why that's controversial, or what the benefit is of enshrining vandalism that managed to last long enough to get archived. There's no good argument to preserve abuse. I was pinged above because I participated in a past discussion about striking sockpuppets in archives. There, I still say we should not be striking in archives. If nothing else, it's preserved as evidence that the sockpuppet influenced consensus (by virtue of not being struck beforehand). I don't see a need to retroactively strike for discussions that aren't ongoing. That's an entirely different subject from vandalism, though. — Rhododendrites ^talk \\ 18:30, 17 February 2025 (UTC)
• I think the linked discussion, and the discussion here, are not in conflict. There seems to be consensus that it's fine to strike disruptive comments in open, active discussions, and it's fine to fix things like BLP violations or post-archiving vandalism in archives, but it's best to leave a disruptive user's comments alone in an archive, if their views had been allowed in the active discussion when it took place, because that keeps an accurate record of the discussion, as it took place at the time. --Tryptofish (talk) 00:22, 18 February 2025 (UTC)

  That's not my understanding of this discussion, which AFAICT is about vandalism that wasn't detected until it was archived, not removing borderline or non-disruptive comments based on the person who made them. Am I wrong? — Rhododendrites ^talk \\ 00:39, 18 February 2025 (UTC)

  You're not wrong. Mathglot (talk) 01:16, 18 February 2025 (UTC)

  Oh, I'm not disagreeing with that; rather, I think I might not have been clear enough in what I said. When I referred to a disruptive user, I was just using that as a way to refer to disruptive edits, but I didn't mean anyone wants to base this on who the person was who made the edit. And I referred to post-archiving vandalism as a way of distinguishing it from vandalism that wasn't detected prior to archiving. --Tryptofish (talk) 23:13, 18 February 2025 (UTC)

  What is being missed is that once one person starts massaging archives, others will join in. I don't want to have to examine archive history to work out if I am looking at an archive or what someone thinks should have been on the original talk page. What if an IP starts fixing archives? Reverting someone when they fix an archive is quick and guaranteed to restore the archive. Checking diffs to see if the fix was desirable or not is wasted time that does not help the encyclopedia. Johnuniq (talk) 01:53, 18 February 2025 (UTC)

  In that hypothetical you have to look at the archive history anyway, as you won't know if it has been edited otherwise. And then you have to trace the base talkpage history anyway to see if that is how the text was actually written as well. CMD (talk) 02:47, 18 February 2025 (UTC)

  We have several potential situations.

  1. An ongoing discussion is vandalised, which is not noticed at the time, and becomes archived in the vandalised state
  2. A concluded/resolved/hatted discussion is vandalised, which is not noticed at the time, and becomes archived in the vandalised state
  3. A discussion is archived in a "clean" state, and is subsequently vandalised within the archive

  There may be more. Situation 1 is easy to spot, since somebody will be watching the page, possibly there will also be subscribers to that discussion. Situation 2 should be easy to spot, if people are continuing to watch the page; unless all participants unwatch/unsubscribe once the discussion concludes. Situation 3 is much harder to spot, since few people watch discussion page archives, but easy to resolve once spotted - a check of the page history will soon show which edits were archiving edits and which were not. --Redrose64 🌹 (talk) 09:22, 18 February 2025 (UTC)

  I wonder about the feasibility of a new mw wishlist item to add a new Preference > Watchlist feature and checkbox: "If I am watching [ns ][Tt]alk:Foo, then also watch [ns ][Tt]alk:Foo\/Archive_\d+ by default." I would definitely check that box. Mathglot (talk) 09:53, 18 February 2025 (UTC)

  I don't think I missed that; I just don't think it's a big problem. People can already edit an archive in ways we don't allow for, after all. If there's a slippery slope argument that people will start making all sorts of changes in archives, require bot authorization or protect them by default or set up an edit filter. None of that seems like good reason to preserve abuse. — Rhododendrites ^talk \\ 14:17, 18 February 2025 (UTC)

• Per my previous comments, we should not be editing archives unless there is a clear benefit in doing so. Vandalism that is subsequently archived? Clearly not an issue. Archive that gets vandalised? Yes its an issue because it distorts the record of the discussion and should be reverted like any other vandalism edit. Its unclear to me at this point what is the issue here. Only in death does duty end (talk) 14:05, 18 February 2025 (UTC)
• Keep it simple: archives can be corrected on the same basis as not archived, so revert vandalism, strike sock comments, etc. If we are going to have archives at all, they should be maintained. That's not changing archives, its correcting them, so if anyone actually has to consult the archives, they get a maintained archive. Alanscottwalker (talk) 17:16, 18 February 2025 (UTC)

Joseph Fletcher

Hello, let me start in good faith by pointing out that I am a newcomer; please don't bite! I wonder why Fletcher's topic introduction end by saying he was an athiest later in life. This does not seem to be the case, and appears to be a possible instance of original research. I could easily see how one could come to this conclusion if glancing over research notes. And keep in mind, by the nature of his philosophic view of extending love and grace selflessly, and because of the trials he faced during the well-documented Congressional Mccarthy Hearings c. 1950's, uncontextualized translations of such dileberation or from the hearings could leave one to believe a less-than perfectly stated acknowledgment of another's standing akin to an agreement with their Point of View. This is hopefully not too speculative! I could not find refernce to him being an athiest later in life. Can we confirm? Thank you! 136.38.193.203 (talk) 05:34, 11 January 2025 (UTC)

We have numerous pages about different people by that name, so I don't know which one you are talking about. But the place to raise these issues is at the talk page for that particular article. If you go to the article for the particular Joseph Fletcher you are concerned about, there is a blue link near the top, called "Talk". Click on that, and you will be at the right place to discuss this. But not here. --Tryptofish (talk) 01:57, 12 January 2025 (UTC)

Bad guys just wipe out what people say on Talk pages

Wikipedia:Talk page guidelines#Editing others' comments: I find that the bad guys just wipe out what one says on talk pages. E.g., you might not see what I have wrote here soon. Yes, even if what one wrote is within the rules. Jidanni (talk) 07:25, 25 January 2025 (UTC)

One thought is that this is phrased as a declarative statement, not a question or proposal. Sometimes you phrase talk page posts in a form directly related to improving the article, but I could see other editors seeing your posts as general discussion posts.

Owing to the above, I have to ask whether you want advice or remedy or a change to the guidelines or what? Remsense ‥ 论 08:15, 25 January 2025 (UTC)

Comments such as at permalink are not appropriate at Wikipedia. An article talk page is a place to propose changes to an article based on reliable sources, and in compliance with other procedures such as WP:DUE. The comment at your talk which you erased (diff) provided the WP:NOTFORUM explanation. Johnuniq (talk) 08:59, 25 January 2025 (UTC)

Why does black September not show the actions that occurred at the 1972 Olympics

Not showing correct history 172.56.153.186 (talk) 15:01, 26 January 2025 (UTC)

Because it concerns the events of the actual Black September, which happened in 1970-71. The organization that was spawned from those events is at Black September Organization, and covers the Olympic massacre of 1972. The article heading could probably do a better job of highlighting the distinction between the two articles. Acroterion (talk) 15:13, 26 January 2025 (UTC)

What does this have to do with improving the page Wikipedia:Talk page guidelines? See Template:Editnotices/Page/Wikipedia talk:Talk page guidelines. --Redrose64 🌹 (talk) 17:15, 26 January 2025 (UTC)

IP address removal/ replacement

after writing a correction of an article at the wrong place, and being informed about this, I replied; now the used IP address is there instead of my name, and won't disappear even after having logged in. There should be an easy and for everybody simply to understand way, how to replace the ip-address with the user name. Roushanne (talk) 00:04, 27 January 2025 (UTC)

The way to do that is to make a new edit where you replied, and delete the IP signature, and then sign your comment while logged in, as you did here. --Tryptofish (talk) 00:32, 27 January 2025 (UTC)

Semi-protected edit request on 11 February 2025

Add a link to WP:NOTFORUM to the word "platforms" in the second sentence.

−

Article talk pages should not be used by editors as platforms for their personal views on a subject.

+

Article talk pages should not be used by editors as [[WP:NOTFORUM|platforms]] for their personal views on a subject.

Justjourney (talk) 05:04, 11 February 2025 (UTC)

 Note: that seems unnecessary. M.Bitton (talk) 11:55, 11 February 2025 (UTC)

 Done. Although I agree that it's not particularly necessary, I do think it's a reasonable link to make, especially for new editors, so I implemented the edit. --Tryptofish (talk) 22:34, 11 February 2025 (UTC)

"blank"

@Moxy:, I'm confused by your edit that added "blank" here:

Never blank, edit, or move someone's comment to change its meaning, even on your own talk page.

Blanking can't change a comment's meaning, and editors are permitted to blank (remove) comments on their own talk pages. Schazjmd (talk) 16:32, 18 February 2025 (UTC)

Removing someone reply would be the worst example of mocking about with someones comments. This should be clear to all. eg. Trying to change the look of how a talk evolves is very grievous.Moxy🍁 16:36, 18 February 2025 (UTC)

I agree that there should be wording to explicitly prohibit removing others' comments on talk pages (except the user's own talk page and except in the cases listed in the exceptions subsection). I don't agree with adding "blank" to that sentence. Schazjmd (talk) 16:54, 18 February 2025 (UTC)

What do you propose? link like we do with WP:3RRNO. Say ..."With the exception of user talk pages removing others' comments is prohibit if they do not violate scenarios outlined above." Moxy🍁 17:07, 18 February 2025 (UTC)

Thanks for making it a separate sentence; I've tweaked the wording. I'm not sure that it currently captures all of the exceptions though. Schazjmd (talk) 17:28, 18 February 2025 (UTC)

AI for translation and grammar checking

We've had a lot of discussions about AI during the last few months (this incomplete list has links to dozens of discussions), but after this comment, I think we need to amend the Wikipedia:Talk page guidelines to write down one of the few things that we agree about, which is that using an AI tool to "generate" a comment is not the same as using the same tool to "translate" or "correct" your own original words so that they appear in decent English on the talk page.

For example:

• Please do not ask Wikipedia:Large language models to generate de novo or create whole comments for you ("Tell me what I should say to prevent Wikipedia from deleting the article on _____"). However, if you are not a native English speaker or if you struggle to write due to dyslexia or other situations, you are allowed to write your own words yourself and then use machine translation such as Google Translate or AI tools such as ChatGPT for translation and for the correction of spelling and grammar errors ("Please translate the following text into correct English:").
  • If you encounter a comment that you believe to be generated completely (not written by an editor and then translated, copyedited, or grammar-checked) by an AI tool, remain civil and politely ask the editor whether they are using an AI tool.

I think this could go right under the WP:ENGLISHPLEASE bullet point (in which we tell non-native English speakers to provide translations "even when modern browsers have machine translation built-in"). What do you think? WhatamIdoing (talk) 03:56, 5 March 2025 (UTC)

We should be wary of endorsing to use of llm tools to edit grammar. Machine translation has flaws, but people are more aware of them, and some flaws are inherent to the concept of translation itself. However, llms often copyedit in ways that change the meaning of the writing, and sometimes invent brand new sentences and remove others, without informing the user that they have done so. (I do not know if llms also engage in such creativity when asked to translate, but I do not believe most online translators do.) CMD (talk) 04:16, 5 March 2025 (UTC)

How concerned are you about this for a talk page discussion, when the realistic alternative is broken English? WhatamIdoing (talk) 04:19, 5 March 2025 (UTC)

Quite. Broken English is often consistent across different comments, and sometimes a clear meaning can still often be ascertained. On the other hand, llm comments (and it can be difficult to tell how much is generated and how much is edited) will vary if the llm follows a word chain towards a different idea, and often does so in much wordier ways. Further, broken English gives some indication of what sort of writing style might be helpful as a reply, whereas there are no such indicators from llm-edited text. CMD (talk) 04:44, 5 March 2025 (UTC)

We've long accepted the use of machine translation software for people who do not have a good command of English. Increasingly, translation websites are using LLMs underneath the hood, often with better results than the statistical MT or encoder-decoder techniques that came before. So I don't see anything wrong with prompting an LLM to translate, which will allow for more customization of the target text (e.g. style) compared to just using a translation website (which is really just an LLM with a hidden prompt that can't be edited). -- King of ♥ ♦ ♣ ♠ 04:53, 5 March 2025 (UTC)

The more customization part is what I'm worried about, as it's not impossible that it may change the meaning of some sentences or add new ones, as Chipmunkdavis mentioned above. Chaotic Enby (talk · contribs) 07:30, 5 March 2025 (UTC)

Low-skill manual translation has the same risks: unintentionally changing the meaning is just par for the course if you're not fluent in the target language.

This is a problem we deal with every day of the year. I suspect that everyone on this page has accidentally left out a key word (such as 'not') when we're typing. Native English speakers with dyslexia do it more often. WhatamIdoing (talk) 17:00, 5 March 2025 (UTC)

Forgetting a word, or making some small translation errors, is not comparable to changing the entire style of the text or adding new content. Chaotic Enby (talk · contribs) 17:29, 5 March 2025 (UTC)

I agree, and hope you can agree with me that changing the style (e.g., from texting abbreviations to business standard) is a matter of degree, rather than type.

Have you had experience with an AI tool adding new content to a text you've given it? I don't use any of these tools, so I don't have practical experience with seeing the results. WhatamIdoing (talk) 17:45, 5 March 2025 (UTC)

Yes. Adding text, deleting text, randomly making changes. They do it all the time. I'm surprised you're making a proposal here on llm use if you feel you don't have experience with them. CMD (talk) 22:56, 5 March 2025 (UTC)

I have experience with listening to Wikipedians, who in previous discussions have accepted the use of LLMs for translation and grammar checking. I also have experience with editors making false accusations (on this and other subjects), and refusing to consider the possibility that their hunches are wrong.

I explained above why I'm here, today, with this. Click the second link in the original message if you haven't, and read through the discussion in which two experienced editors insult a newcomer for attempting to give them an explanation, in decent English, of the edit that one of them reverted. (He was right about the content question, BTW.) Note that he stops using any form of machine translation, but they keep insulting his subsequent hand-written, non-AI messages as "AI slop" that isn't worth their time to read.

I know that you wouldn't ever react that way, but apparently we are not all as calm and steady as you. If we were all like our most cool-headed contributors, we could just delete this guideline as unnecessary. But we're not. Some of us need it written down in black and white. WhatamIdoing (talk) 23:27, 5 March 2025 (UTC)

I don't see a clear link between that discussion/reactions and this proposal. If the attempt is to reduce the space for hunches, this proposal would do the opposite, as it expands expected use in a grey rather than black and white manner. If that discussion is correct and that post is the result of an llm translation or grammar edits, then that's suggestive the translations/edits suffer from similar challenges to generated text.

Tangentially, I've been running some recent stuff I've been writing into ChatGPT and asking it to point out spelling and grammar errors. Anecdotally, it seems good on spelling, but on grammar it is much more hit and miss. It also requires specific coaxing to get it to list out changes; it seems to usually want to just churn out a text rewrite with no explanation. CMD (talk) 01:02, 6 March 2025 (UTC)

In that discussion, we see:

• A new editor using ChatGPT for translation and grammar checking of his own original, human-generated words and thoughts.
• Two experienced editors responding to his comments with insults (e.g., "complete and utter BS", "AI garbage", "Either use your own brain or don't bother. Nobody needs to sort through your AI-generated crap", "obvious AI slop") and making up non-existent rules (e.g., "If you use AI to generate a response, you are required to say so").
• At least one experienced editor doubling down on his belief that all of the comments are "at best mildly modified AI slop", even after the new editor stops using AI in any form and discloses details about his original/uncorrected comments vs the grammar checked version. This editor also claims that there is no meaningful distinction between AI-generated and AI-translated content ("using machine-generated text, whether it was "translated" (whatever that is supposed to mean) or otherwise").

In my proposal, you should see:

• explicit permission for people to use ChatGPT for translation and grammar checking of their own original, human-generated words and thoughts;
• a reminder to experienced editors that suspecting that someone has used AI does not grant them an exemption from the Wikipedia:Civility conduct policy; and
• a clear distinction between using an AI tool to generate comments de novo and using an AI tool to translate your own thoughts or correct your spelling and grammar.

Can you see the connection between that conversation and this proposal now? WhatamIdoing (talk) 03:10, 8 March 2025 (UTC)

I suppose it could be framed that way, but that's taking a very strong interpretation of "grammar checking" for heavily modified text that is not what I suspect people think of when they read "grammar checking". CMD (talk) 14:42, 8 March 2025 (UTC)

The new editor gives copies of the original vs ChatGPT output:

−

"just make correct english:<br/>Due to no myth one can wrongly interpret the image. For example, if we did not know about Thor's fishing trip Jörmungandr. We could interpret the Gosforth Cross showing Thor fishing as the god Thor with another god(for example Loki), but we know that the other person is not a god, but the jötunn Hymir. It is easy to mininterpret the figure as there is not enought distinct characteristic for this person. This person could be a human or jötunn for example."

+

LLM:<br/>"Due to the lack of a specific myth, the image can be easily misinterpreted. For example, if we didn’t know about Thor’s fishing trip with Jörmungandr, we might interpret the Gosforth Cross depiction of Thor fishing as showing Thor alongside another god (such as Loki). However, we know that the other figure is not a god but the jötunn Hymir. It’s easy to misinterpret figures like this when there aren’t enough distinct characteristics. In the case of the Böksta Runestone, the skier could just as easily be a human or a jötunn."

I see no changes here that concern me. I wouldn't have produced exactly the same – for example, changing "It is" to "It's" would be a (minor) MOS problem for the mainspace, but this is not in the mainspace; I'm a comma maximalist, and it removes a comma that I would have kept; it retained the word distinct from the original, and I would have modified it to distinctive – but I see no significant change in meaning, no added content, no hallucinated claims, and none of the other problems that editors have been concerned about. Do you see any concerning changes? Or is this a use that you would accept? WhatamIdoing (talk) 18:25, 8 March 2025 (UTC)

It looks pretty good indeed, although I'm curious if the skier at the Böksta Runestone was something that was in an earlier prompt or paragraph, or something that ChatGPT added itself. Otherwise, this looks like a pretty good use case of LLMs where the grammar is improved without the meaning being altered. Thanks for bringing a concrete example! Chaotic Enby (talk · contribs) 19:45, 8 March 2025 (UTC)

I haven't raised any concerns about that bullet point, but as you raise it and taking it as given, you see no significant change or added content in the invention of context for a sentence at the end? I don't think you'd do that in grammar checking, and quite clearly that isn't a use we should accept. It's probably pulling from somewhere else in the chat thread, but such things can also be pulled form other threads, or out of the ether. CMD (talk) 06:58, 9 March 2025 (UTC)

As the (on-wiki) comment indicates that this grammar request was preceded by several other attempts, I also assume that it was pulling it from somewhere else in the chat thread. I also think that editors need to read what they post (to the extent that they are able to read it; we've seen people post – and this guideline recommend that they post – machine translation into English even if they can't read it, and that's usually worked out well). WhatamIdoing (talk) 19:55, 9 March 2025 (UTC)

Pulling things from somewhere else is not grammar checking, nor is it machine translation. CMD (talk) 00:22, 10 March 2025 (UTC)

It probably is basic copyediting, though. WhatamIdoing (talk) 21:25, 10 March 2025 (UTC)

• I think this is a good start, but I think we need to be clearer that it's only relevant to ask if you think they've used AI for more than is permitted, and we need to stress that AGF is important and that that includes believing the answer people give when asked about AI usage, unless you have evidence (not gueses, not even reasons supposition, actual hard evidence) they are not telling the truth. Thryduulf (talk) 10:31, 5 March 2025 (UTC)

  That sounds like a lot of text for a page that mostly has bullet points of one or two sentences in length. Maybe a separate essay (which could be linked here)? WhatamIdoing (talk) 17:01, 5 March 2025 (UTC)

I would prefer that the commenter give both the original text and the ML translations together rather than permitting them to give only the ML translations, which the proposed change seems to allow? ML translations is not much better than writing badly in English in some cases, and there are plenty of bilinguals or polyglots around to make sense from the original text if the ML translations don't. – robertsky (talk) 13:30, 5 March 2025 (UTC)

What about the English speaker who is just trying to get their grammar cleaned up? Grammarly is using AI for grammar checking. MS Word is using AI for grammar checking. Do we really want the original bad grammar posted followed by the corrections? At some level, that just sounds like shaming people with dyslexia and ADHD for being smart enough to realize that they should use a grammar checker if they want to be successful at communicating in text. I don't want to see "Original: 'Thank you, your very helpful'. Grammarly: 'Thank you, you're very helpful". Do you? WhatamIdoing (talk) 17:09, 5 March 2025 (UTC)

I do think there's a difference between using a grammar checker and feeding a comment into a LLM for a full rewrite. Chaotic Enby (talk · contribs) 17:28, 5 March 2025 (UTC)

The difference is increasingly blurred these days, everything is shoved into multi-headed enterprise products and people aren't necessarily going to know the distinction between what is "real AI" and what is "just grammar checking." Gnomingstuff (talk) 03:32, 7 March 2025 (UTC)

I don't like the idea of mandating that editors must make their first drafts available for comparison with the final version of their contributions. The act of writing out one's thoughts can help clarify them in one's mind, and result in changes. If just the right phrasing comes about through some form of revision process, whether on my own or through assistance from others or tools, I don't want people to second-guess the final message just because it started out differently in the first draft. isaacl (talk) 17:54, 5 March 2025 (UTC)

I have the opposite problem. Very often, I will write one thing, and another reader will come along and interpret it in a way that I never intended or meant. If there's any way that AI can fix that, then we should encourage it. In other words, if AI can increase clarity and understanding, it's a good thing. Viriditas (talk) 23:12, 5 March 2025 (UTC)

I think there's a common underlying principle: your first draft of your thoughts may not be the most effective way of communicating your actual viewpoint, and gaining feedback, either through self-review, from others, or from tools can help improve your message's effectiveness. However, what I'm saying is that I think editors should be able to choose to implement this feedback process privately. isaacl (talk) 01:39, 6 March 2025 (UTC)

I agree with Isaacl. Sometimes I write a reply that's quite long, only to realize that my initial reaction was unfair to the other editors, or that my response won't be helpful even if (I believe) I'm right. Isn't there a fairy tale about being able to read other people's thoughts, and discovering that it doesn't make you happy? WhatamIdoing (talk) 03:18, 8 March 2025 (UTC)

Whatever is allowed, it should be made clear that the human is responsible for the words they put into wp, regardless of how those words arose. Any allowance for machine translation (llm or not) should include instructions to carefully sanity-check the result before posting it. One reason is that every language has built-in ambiguities that are resolved by context. Recently I translated a portion of old Arabic on chatgpt which was a dialogue between "He" and "I", but the result had the two players reversed from how I knew it should be from context. When I replied, "is it possible that 'he' and 'I' could be the other way around"? it said "yes, that is also a valid translation" and gave a new version that made sense. The ability to discuss the translation is the biggest advantage llm tools have over systems like google translate. You can even tell it some of the context before it translates. Another example is that telling chatgpt in advance which common words in the passage are actually place names or people names will stop it from using their common meaning. Zero^talk 03:34, 6 March 2025 (UTC)

I agree that whatever you post, you are responsible for it. I don't think that's actually on this page. Perhaps it's too much of a "common sense" issue for people to have felt like it needs to be written down? Or perhaps we should worry that if we do write it down, it will get weaponized against ordinary mistakes (e.g., if you attribute a quote to the wrong editor in a discussion)? WhatamIdoing (talk) 20:01, 9 March 2025 (UTC)

I don't really care if people use ChatGPT for talk page comments (articles are different, the ongoing problem of people typing ChatGPT prompts into talk pages is also different). Posting the first draft for a comparison -- assuming anyone bothers to do that, which they mostly won't -- is just going to bloat talk pages. Although the AI inquisition linked above is also doing that, so who knows. Gnomingstuff (talk) 03:39, 7 March 2025 (UTC)

Limit

I removed the 75 KB limit. It was added by someone in 2012 because someone in 2009 thought there should be some limit in KBs. It is not helpful, time has moved on and storage has become cheaper and internet faster, and the decision to archive or not archive something is more complicated than simply looking at the amount of bytes. Polygnotus (talk) 02:37, 6 March 2025 (UTC)

Our recommendations should be based on accessibility.... not size. Scrolling although accessible to the average individual is an accessibility nightmare for many with mobile disabilities and visual impairments. Moxy🍁 02:40, 6 March 2025 (UTC)

Good point! On mobile you'll find it much easier to scroll past 100KB in a {{collapse top}} template than 100KB of embedded images. Ideally we should have some JavaScript that determines the height in pixels. I am not sure if visually impaired people (e.g. those using screenreaders) can skip collapsed sections. Polygnotus (talk) 02:47, 6 March 2025 (UTC)

I personally written a few papers on the new layout of Wikipedia and accessibility.... and in conclusion find it odd they decide to go with the style that requires someone to expand the contents to be able to view it. Every click is an accessibility barrier to those with mobility issues. This is also the case for screen readers..... if someone is blind there is no way they're able to access the contents of a page now. Moxy🍁 04:03, 6 March 2025 (UTC)

@Moxy I haven't done as much archive digging as I should to make a definitive statement, but it looks like the Vector 2022 skin was one of the most major setbacks in recent years, and not just for accessibility. Polygnotus (talk) 05:18, 6 March 2025 (UTC)

One of the oddest things is that accessibility concerns were mentioned multiple times during its initial deployment to no avail. Those of us that advocate for accessibility here simply think it's the amount of time and money invested in a new skin ....thus it was simply never going to be undone... we are talking about paid work that quite a few people did to make these changes vs volunteers here at Wikipedia advocating for very small group of individuals. Moxy🍁 05:23, 6 March 2025 (UTC)

@Moxy One can hire disabled accessibility testers online, and there are plenty of WCAG/ARIA compliant web developers. A lot of the things you do for accessibility you would do anyway if you want to make a good design. E.g. not having a cramped interface is not just good for HeadMouse users and those with decreased motor function control, but it also just plain looks better. And the changes made for visual impairment also benefit those who are using a cheap mobile phone in a sunny area. Something like Dark mode makes the site much more usable for a very wide range of people. Polygnotus (talk) 08:31, 6 March 2025 (UTC)

What is the style that requires someone to expand the contents to be able to view it? I've been using Vector 2022 for over a year. I never have to expand anything to view it, unless an editor has specifically put the text in a {{hat}}. Or are you talking about a different "new layout"? The mobile skin does collapse sections, but it's only "new" if you mean "more than a dozen years old". WhatamIdoing (talk) 03:26, 8 March 2025 (UTC)

On a talk page (for me) when not logged in ...both cotents are colasped on my pc and mobile view. Is this not the norm? Moxy🍁 19:03, 8 March 2025 (UTC)

Not the norm. It's been requested but not implemented. My favorite suggestion was to have the sections all open by default, but you could collapse them as you move down the page. WhatamIdoing (talk) 19:18, 8 March 2025 (UTC)

The mobile view looks like something out of the early 2000's. When the WMF where making Vector 2020 I assumed it would be cross desktop and mobile, to at least catch up on the last decade. -- LCU ActivelyDisinterested «@» °∆t° 19:35, 8 March 2025 (UTC)

They might have done that, except that power users here have strenuously objected to that. The usual rules of engagement for skins are:

• Don't change anything noticeably.
• Make mobile and desktop match.
• Especially don't change anything for me.

We give much more power to experienced editors (and much less to ordinary readers) than most websites. The upside is that, with enough effort, I can block or avoid just about any change that I hate. The downside of this approach reminds me of the Edsel: the most survey-tested, user-driven design in automotive history – and everybody hated it. WhatamIdoing (talk) 19:55, 8 March 2025 (UTC)

I enjoy the fact that the WMF either listens too much or too little. But never just enough! Polygnotus (talk) 20:00, 8 March 2025 (UTC)

It might be the right amount overall, but I think the power distribution is wrong. WhatamIdoing (talk) 20:02, 9 March 2025 (UTC)

Unfortunately every common for all development, change rarely makes many friends. -- LCU ActivelyDisinterested «@» °∆t° 20:08, 8 March 2025 (UTC)

I appreciate the accessibility concerns but our problem is in the opposite direction. "Thanks" to archive bots we have a glut of per-month archives with 1 entry each. Gnomingstuff (talk) 16:34, 7 March 2025 (UTC)

There are still a lot of readers or editors who are using old or less powerful devices to edit, so some kind of limit is a good idea. Very large talk pages can become unresponsive and difficult to edit otherwise. -- LCU ActivelyDisinterested «@» °∆t° 21:42, 7 March 2025 (UTC)

One image, or worse yet a video player, will be a thousand time more difficult for an old device than some text. So if there needs to be a limit, it should be based on the actual impact. Polygnotus (talk) 01:20, 8 March 2025 (UTC)

It depends on which kind of difficulty you're envisioning. Is this page loading time? Bandwidth use? How much I have to scroll? A long, high-definition video is a problem for bandwidth, but it's a single screen: it's easier to scroll past that than to scroll past a thousand-word comment. WhatamIdoing (talk) 03:22, 8 March 2025 (UTC)

There is an actual impact of having large amounts of text. 75kb is small, but getting above 500kb can make pages become unresponsive. -- LCU ActivelyDisinterested «@» °∆t° 10:43, 8 March 2025 (UTC)

I agree. Also, the larger the page, the more likely that there will be some other sort of problem (e.g., an unclosed tag, WP:PEIS limits). WP:ANI, which is coming up on 1200 archives soon, currently sets its limit to 800K. These are huge pages. I think that most pages should consider a maximum of half or even a quarter of that. WhatamIdoing (talk) 18:28, 8 March 2025 (UTC)

That might explain why the AN/ANI pages are misbehaving for me. 400K or less is a good idea, as single threads can get huge and make the archives much larger than the archive limit. -- LCU ActivelyDisinterested «@» °∆t° 18:34, 8 March 2025 (UTC)

Or perhaps something like "not normally more than 250K"? WhatamIdoing (talk) 18:42, 8 March 2025 (UTC)

Again, that does not solve the problem because it does not take into account 250K of what. 250K of image embeds or 250K of video embeds or 250K of plain text in a collapse template are completely different. 250K of video embeds is a trillion times harder to render than 250K of text. 250K of image embeds of the largest images on Wikimedia Commons and we can have AD's computer selfdestruct! Polygnotus (talk) 18:44, 8 March 2025 (UTC)

I honestly don't know how many times I can't keep saying this, large qualities of text is the issue. Images and other objects don't really have that much impact. -- LCU ActivelyDisinterested «@» °∆t° 18:54, 8 March 2025 (UTC)

@ActivelyDisinterested I mean that clearly isn't true. Polygnotus (talk) 18:56, 8 March 2025 (UTC)

@Polygnotus, I find that I am easily irritated these days by people who refuse to Believe editors when those editors are telling you about their personal experiences. WhatamIdoing (talk) 18:58, 8 March 2025 (UTC)

@WhatamIdoing 3 decades in IT will make you not trust user feedback. Or 2 days in any medical profession. And if someone makes a claim about observable reality that is false then we can say that. We can't discount their feelings or emotions tho! Polygnotus (talk) 19:00, 8 March 2025 (UTC)

On the contrary: It has made me trust user feedback.

Please find a way to re-write your comment so that it cannot possibly be interpreted as "I think you're lying". It's possible that Wikipedia:By definition and thinking about If a tree falls in a forest and no one is around to hear it, does it make a sound? would help you differentiate between your own experience and their equally true experience. It is possible that bandwidth is your main problem and scrolling size is another editor's main problem. Imagine a world in which the other user isn't wrong. WhatamIdoing (talk) 19:03, 8 March 2025 (UTC)

@WhatamIdoing The claim made is not about their personal experience. I thought you were as nitpicky as I am!Polygnotus (talk) 19:05, 8 March 2025 (UTC)

I encourage you to review the comment that says pages are misbehaving for me and reformulate your reply based on the understanding that when an user says something is broken "for me", then the user is speaking "about their personal experience". WhatamIdoing (talk) 19:11, 8 March 2025 (UTC)

See strawman argument. Polygnotus (talk) 19:17, 8 March 2025 (UTC)

When one person says "The claim made is not about their personal experience", and the next person gives evidence that it actually is, then that's not a strawman argument. WhatamIdoing (talk) 19:20, 8 March 2025 (UTC)

@WhatamIdoing But that isn't what happened. Polygnotus (talk) 19:21, 8 March 2025 (UTC)

That's what I think happened. Perhaps take it to my talk page? Start with explaining how the 19:05 comment says "The claim made is not about their personal experience." and the 19:11 comment provides evidence that the user reported a problem "for me" and therefore about their own personal experience, but this directly related comment about your exact words is a strawman argument. It might amuse my talk page stalkers. WhatamIdoing (talk) 19:34, 8 March 2025 (UTC)

Well, you said you were easily irritated these days. So maybe in a week or so? Polygnotus (talk) 19:35, 8 March 2025 (UTC)

Sure, if you prefer. But since I've been irritated about disbelieving user reports for a long time, I suspect that particular pet peeve may continue for the rest of my life. WhatamIdoing (talk) 19:47, 8 March 2025 (UTC)

@WhatamIdoing Looks like I got unlucky and (was perceived to) hit a sore spot by accident. I have no interest in fighting about this topic. There are other topics we could spar about, far more productively. I am not the enemy. Polygnotus (talk) 19:50, 8 March 2025 (UTC)

I didn't realise you realise you were spying on my personal experiences, I can think of no other way you could say what I experience is untrue. As I said elsewhere, that you don't experience the issues doesn't mean they dot exist. -- LCU ActivelyDisinterested «@» °∆t° 19:28, 8 March 2025 (UTC)

@ActivelyDisinterested If your pc can't handle WP:ANI and it was created in the past decade then you should probably post your specs on WP:VPT and ask them what to do. Upgrading to an SSD would probably help. Reducing the amount of programs that start when your computer starts helps too. Polygnotus (talk) 18:46, 8 March 2025 (UTC)

I don't edit on PC, like the majority Wikipedia's readership I use a phone. It's a bit but not that old, there will be many readers who are using older devices. Also look up sayings about eggs. -- LCU ActivelyDisinterested «@» °∆t° 18:53, 8 March 2025 (UTC)

!@ActivelyDisinterested Which make and model? Polygnotus (talk) 18:58, 8 March 2025 (UTC)

It doesn't matter. No matter what make and model AD has, someone else will have a worse one. WhatamIdoing (talk) 18:58, 8 March 2025 (UTC)

@WhatamIdoing It does matter, which is why I asked. Polygnotus (talk) 19:01, 8 March 2025 (UTC)

No, it doesn't matter. Even if you could say "Oh, if you've got the 2011 Illudium PU-36 spacephone, and you're having trouble finding a specific comment when the text exceeds the length of War and Peace, then you need to go into Preferences > Settings > Text > Obscure and pick "War and Peace mode", then that won't solve the real problem, which is that multiple editors, with multiple devices, most of whom are not posting here, find it difficult to scroll through and edit pages that are the length of a novel. WhatamIdoing (talk) 19:09, 8 March 2025 (UTC)

I do not recommend speaking for others, they might find it annoying. Or helpful. Anyway, I wouldn't ask if it didn't matter. I am not trying to solve the problem for all those people, or even for AD, I am just trying to understand what AD is experiencing on their personal device. Polygnotus (talk) 19:16, 8 March 2025 (UTC)

The way to solve the issue for all editors and readers effected is to set an upper limit on page sizes. If there was a different solution I would have already thought of and implemented it. -- LCU ActivelyDisinterested «@» °∆t° 19:32, 8 March 2025 (UTC)

There is a different solution. A better solution even. I am not a mindreader so it is unclear if you have thought of it or not. Polygnotus (talk) 19:34, 8 March 2025 (UTC)

You don't need to read my mind, my personal situation isn't something you should concern yourself with. This discussion is about the general issue, it exists and a general solution is required. -- LCU ActivelyDisinterested «@» °∆t° 19:40, 8 March 2025 (UTC)

To understand the general issue it would be helpful if we could have an example. Perhaps someone who experiences that issue on their personal device, who could answer some simple question like which make and model and which browser. Oh well. Polygnotus (talk) 20:09, 8 March 2025 (UTC)

There is no need, it's a simple case of keeping large size down. There is nothing else to fix. -- LCU ActivelyDisinterested «@» °∆t° 19:34, 9 March 2025 (UTC)

If we want to do something to fix a problem, we need to understand it first. Polygnotus (talk) 04:34, 10 March 2025 (UTC)

If you have an interest in doing so good for you, but I don't. I just want pages not to be broken, and until another fix is available the way to do that is to limit their size. -- LCU ActivelyDisinterested «@» °∆t° 22:06, 10 March 2025 (UTC)

That will guarantee that no one will ever fix the actual problem until something bad happens. See here. Nerds want to understand stuff, not apply a bandaid and ignore it until it explodes. Polygnotus (talk) 22:10, 10 March 2025 (UTC)

None of your business, and inconsequential to the discussion. Also I've subscribed to the thread, no need to pin me every time. -- LCU ActivelyDisinterested «@» °∆t° 19:29, 8 March 2025 (UTC)

Broken windows theory of the internet. When one person is rude... Not to be confused with the broken Windows theory. Polygnotus (talk) 21:31, 8 March 2025 (UTC)

I have no idea what you're talking about, but it appears unrelated to the discussion. -- LCU ActivelyDisinterested «@» °∆t° 19:35, 9 March 2025 (UTC)

It is not. Polygnotus (talk) 19:23, 13 March 2025 (UTC)

Test

Commons has a category for large images. Loading the image embeds as text (so [[File:filename.jpg]]) surrounded by <nowiki> tags takes less than a second, even when there are 1000. Loading even a couple of those images takes far far longer than a second. QED. I haven't tested videos but they are slower than text and faster than the largest images Commons has to offer. Polygnotus (talk) 20:31, 8 March 2025 (UTC)

Even the smallest images I can find on Commons (icons, because those in Category:Small images are too big) load significantly slower than text. Polygnotus (talk) 21:37, 8 March 2025 (UTC)

Yes, I'm sure that's true. However, have you considered the possibility that a large page could cause problems for some editors even if it loaded quickly? WhatamIdoing (talk) 00:07, 9 March 2025 (UTC)

Yep. Polygnotus (talk) 04:41, 9 March 2025 (UTC)

Do you agree that a large page could cause problems even if it loads quickly, or did you consider it and decide the users claiming they have problems with large pages are just wrong? WhatamIdoing (talk) 19:42, 9 March 2025 (UTC)

could cause problems is a bit vague. A concrete example of a problem that can be caused by a page being very long (even if it loads very quickly) is that it requires a lot of scrolling which is inconvenient for some people (e.g. mobile phone users whose thumbs get RSI scrolling all the way down)

People are asked to post at the bottom of certain pages (e.g. talkpages and noticeboards) and mobile users can't press Ctrl-End. Imagine having to scroll to the bottom of User_talk:Lajmmoore! Polygnotus (talk) 04:12, 10 March 2025 (UTC)

I agree: Very long pages of text are difficult to handle on mobile regardless of page loading time. Scrolling is a problem if you are using the desktop site on a mobile device. This can be solved in part by having shorter pages.

If you are on the mobile site, then scrolling is less of a problem (because the sections are collapsed), but finding individual comments or a string of text is very difficult (also because the sections are collapsed). WhatamIdoing (talk) 21:22, 10 March 2025 (UTC)

@WhatamIdoing Having shorter pages is one approach, but lazy loading of sections is also an option. Why render stuff outside the viewport anyway? Sure, searching does not work in sections that haven't been lazy loaded yet, but searching is also impossible when sections are collapsed. A simple approach is to always lazyload x sections on mobile, but we can also use a slightly more sophisticated approach where you take into account stuff like how difficult it is to render a particular set of sections (how many bytes, how many embeds, height in pixels etc) and lazyload more or less depending on that. Polygnotus (talk) 21:40, 10 March 2025 (UTC)

The short answer sounds a lot like "because editors from the English Wikipedia pitched a fit about lazy loading when it was implemented in mw:Flow". I doubt that will be tried again until there's been complete turnover in the WMF's Product department. I believe that the problem of ⌘F searching was a significant concern (and the people expressing concern were on desktop, so they didn't see collapsed sections). WhatamIdoing (talk) 16:57, 12 March 2025 (UTC)

@WhatamIdoing Lazy loading is the standard approach on much of the internet these days and, when implemented properly, users shouldn't even notice it unless they experience connection problems. Polygnotus (talk) 17:08, 12 March 2025 (UTC)

Except when a significant activity is using ⌘F to find the comment you want to report at WP:ANI. It's not a question of whether users "should" or "shouldn't" notice it; I'm saying that the minority of us who have disproportionate power over the design did notice it, and it didn't work for us. WhatamIdoing (talk) 17:14, 12 March 2025 (UTC)

Understood. Sounds like I should join this minority! Polygnotus (talk) 17:17, 12 March 2025 (UTC)

You already have. WhatamIdoing (talk) 00:22, 13 March 2025 (UTC)

Even a simple {{Skip to bottom}} could save someone from RSI. Note that not all mobile users use the mobile interface. Perhaps most famously User:Cullen328/Smartphone_editing#Mobile_site_vs._desktop_site. And Cullen is not alone, see here. Polygnotus (talk) 21:55, 10 March 2025 (UTC)

Large pages load just fine, but then misbehave when interacting with them. Showing this isn't an issue with images and objects, but rather other all page size and text volumes. There seems to be an issue with understanding that we need to cater to the lower common denominator when it comes to what readers may be using to access content. -- LCU ActivelyDisinterested «@» °∆t° 19:44, 9 March 2025 (UTC)

Honestly I suspect this is an issue with mediawiki and how it handles such pages. -- LCU ActivelyDisinterested «@» °∆t° 19:44, 9 March 2025 (UTC)

If it is a problem with MediaWiki, then it may be wise to attack the problem at its root instead of only addressing the symptoms. So we should investigate what causes the problem. That starts by asking a few questions about what people do when they experience the problem. What device do they use. Which page do they visit. Et cetera. Polygnotus (talk) 04:55, 10 March 2025 (UTC)

If you wish to help with that issue this isn't the place. I would guess it's phabricator but I don't have knowledge of the exact workings. -- LCU ActivelyDisinterested «@» °∆t° 21:39, 10 March 2025 (UTC)

In what sense do they misbehave when you interact with large pages? Can you describe the problem? Is WP:ANI, which is often very very long, a good example of a page where this problem occurs? Polygnotus (talk) 04:25, 10 March 2025 (UTC)

All talk pages over 500kish in size, nearly always pages wlonly with text, Wikipedia is the only site that exhibited the issue. While there is no other solution I would suggest we don't impact other editors just so pages are more convenient for some. -- LCU ActivelyDisinterested «@» °∆t° 21:41, 10 March 2025 (UTC)

That sounds very annoying! What problems do you experience on those pages? You said they became unresponsive, can you still scroll up and down? Does clicking links still work? Is the problem that the page responds very slowly, or that it does not respond at all? Polygnotus (talk) 21:47, 10 March 2025 (UTC)

Why is this being discussed here, unless you are going to then go to phabricator it won't have any effect. -- LCU ActivelyDisinterested «@» °∆t° 21:53, 10 March 2025 (UTC)

@ActivelyDisinterested Well, if you are using VisualEditor for example then there is a bug filed on Phabricator saying that long pages are very very slow. See T126348. In order to report such things on Phabricator we need more information. So if you would be so kind to answer my questions then I can try to find similar Phabricator tickets. If Wikipedians don't tell the WMF that a problem exists then they are unlikely to fix it. Polygnotus (talk) 22:00, 10 March 2025 (UTC)

Nope, every edit is in source or using the reply tool. Both ways have the same issues. I have no interest in dealing with fixing the mediawiki software, there is a simple way of ensuring the issues don't occur until it is fixed. -- LCU ActivelyDisinterested «@» °∆t° 22:05, 10 March 2025 (UTC)

The problem with fixing the symptoms is that there is no longer a pressing need to find and eliminate the root cause. This has led to the worst disasters in computing and also in other fields like aviation.

You suspect it is an issue with MediaWiki but I suspect it is a problem with your device. I'd estimate a 99.99% probability. I am happy to help you with problems with your device and there are people on WP:VPT who are 5 times smarter and more knowledgeable than I am (but perhaps not as adorable). Polygnotus (talk) 22:07, 10 March 2025 (UTC)

If there is no need to fix the problem, the there is no need to fix the probem. There are always far more pressing things to do. It is not a problem with my device, you are now entering into "I don't want to here you" territory. It only happens with Wikipedia pages, and it I have seen the same issue reported by other editors. Believe me, call me a lie, or drop it. -- LCU ActivelyDisinterested «@» °∆t° 22:22, 10 March 2025 (UTC)

There are no successful botches in prod because a timebomb is not a succesful botch. Learned that the hard way, as have many others. Do you have links to those reports by other editors? I don't see any on Phabricator that are unrelated to the VisualEditor. Polygnotus (talk) 22:26, 10 March 2025 (UTC)

I've seen such complaints, but I never bothered reporting them in Phab. The problems are generally greater for editors in developing countries, where older hardware and slow connections are common, though by no means universal, for desktop users (so they have page weight concerns) and Wikipedia editors using feature phones (e.g., JioPhone) are not uncommon.

The Editing team did some formal work with VisualEditor to test under various conditions. I believe they got some time in a simulation lab from one of the huge tech companies. That kicked off the first of several rounds of optimization, so of course there will be documentation for that work. WhatamIdoing (talk) 17:07, 12 March 2025 (UTC)

The Boeing 737 MAX MCAS system seemed like a great idea at the time. Polygnotus (talk) 22:51, 10 March 2025 (UTC)

This discussion is about the need for limiting page sizes, there is a need. I have no interest in the underlying issue. -- LCU ActivelyDisinterested «@» °∆t° 23:09, 10 March 2025 (UTC)

A 128MB page would be a bit much. If you are not interested in the underlying issues then we'll have to wait until someone else comes along whose device(s) have a problem with large pages which aren't complicated to render and don't require many requests who is willing to answer some simple questions. That may take a while. And if it turns out that the cause is MediaWiki and not their device then we can file a bugreport on Phabricator. Polygnotus (talk) 06:42, 11 March 2025 (UTC)

And if the cause is "big page, no matter which website it comes from", then what? If 50,000 words in plain text on a MediaWiki page is just as much trouble as 50,000 words in plain text on Gutenberg.org, are you calling that "their device's problem"? We need editors to be able to use our site even if they have device limitations. WhatamIdoing (talk) 17:09, 12 March 2025 (UTC)

Break

@WhatamIdoing A device limitation sounds far far more likely than a undiscovered bug in MediaWiki itself in this context. The approach to finding and fixing a bug in MediaWiki (or in the server clusters) is completely different than how we should approach dealing with device limitations.

Lets say we have user X with device Y and we can test it and there is nothing weird going on (e.g. a misconfiguration) but it is just somehow limited to 50.000 words per page for some reason by the manufacturer.

Lets say that there is a significant amount of people using this device or a device with similar capabilities and limitations. The JioPhone has sold 25m in India according to [1] and ~228m people there speak English.

In that case it may be possible to use a bit of JavaScript that tries to figure out how capable a device is (a more sophisticated method than using user agent detection) and serve a "lite" version of Wikipedia when it thinks that is necessary.

• Lazyloaded images/graphs/videos et cetera only when you tap/click them.
• Less reliance on modern CSS tricks.
• Lazy loading sections on talkpages and archives (maybe even on articles).

Of course users should still be able to switch to the full-fat version. A bit like the Wireless Application Protocol back in the day (I hate AMP).

Perhaps the user experience can be like switching themes (even though in reality there is more that happens than loading a CSS file). Or it would be l.wikipedia.org for the lite version and m.wikipedia.org for the mobile version. Maybe the minority who has disproportionate power over the design would care less if they don't have to experience it? Polygnotus (talk) 17:38, 12 March 2025 (UTC)

Maybe the minority who has disproportionate power over the design would care less if they don't have to experience it: This is definitely true, especially at the English and German Wikipedias. WhatamIdoing (talk) 00:21, 13 March 2025 (UTC)

And in response to your question If 50,000 words in plain text on a MediaWiki page is just as much trouble as 50,000 words in plain text on Gutenberg.org, are you calling that "their device's problem"? Yes, of course, if the problem is not the server but the client then that would be the device's problem. That doesn't mean that we should therefore shrug and ignore it, but it does mean that it is clear that it is not the server hiccuping after 50.000 words but, for example, the clients memory cache being full after 50.000 words. A different problem with a different solution. Polygnotus (talk) 17:47, 12 March 2025 (UTC)

And what is the obvious solution for that problem? WhatamIdoing (talk) 00:20, 13 March 2025 (UTC)

Well that problem does not exist as far as we know; it is a hypothetical example for this discussion. So I am going to assume that you mean "What should we do to help users with older devices/slow internet connections". The answer is basically the same. The WMF should do some research, maybe buy some of those devices and/or interview some users, and check how much impact "lazyloading video/images/graphs only when you tap/click them" and "Lazy loading sections on talkpages and archives (maybe even on articles)" has and if there any other ideas they should add to that list. Then they should build a lite version of Wikipedia (and perhaps stick a cache of pre-rendered pages on a CDN since most people are readers not writers).

I know which answer you want, but a workaround is not a solution. And it isn't even a good workaround, as I explained previously. What we should not do is refusing to even try to understand the actual problem. And we should certainly not start infighting, even if that is the culture here on enwiki. Polygnotus (talk) 07:23, 13 March 2025 (UTC)

Or we could ...just not have huge pages. WhatamIdoing (talk) 17:55, 13 March 2025 (UTC)

@WhatamIdoing But that wouldn't fix the problem. All these extra features are nice for people like you and me who have fiber and CPUs with dozens of cores, but the JioPhone costs around 12 dollar. And this kinda stuff is worth researching for the WMF considering the fact that not everyone lives in western Europe. And not everyone buys the latest iPhone every year. And some people need accessibility features. Polygnotus (talk) 18:00, 13 March 2025 (UTC)

If simply restricting page size to 50k words would be a solution then I would be all for it. But I am not because it isn't. If it was then GTMetrix would simply say "reduce the amount of words on the page and you are done". Website performance optimization is quite a bit more complicated. Polygnotus (talk) 18:20, 13 March 2025 (UTC)

"Website performance optimization" is about page weight again.

When editors tell us that making the pages shorter solves their problems, then it actually does solve their problem. WhatamIdoing (talk) 18:27, 13 March 2025 (UTC)

"Website performance optimization" is about page weight again. False. Like I said, it is quite a bit more complicated. Polygnotus (talk) 18:28, 13 March 2025 (UTC)

When editors tell us that making the pages shorter solves their problems, then it actually does solve their problem. That appears to be your philosophy. You can believe what you want and I can believe what I want. Polygnotus (talk) 18:35, 13 March 2025 (UTC)

If you want me to I can easily make a page in my userspace that will take ages to load while being relatively small. I have a list of all filenames in the Commons category "Large images". Can you explain why that page would take ages to load despite not containing much text (fewer than 50k words)? Polygnotus (talk) 18:38, 13 March 2025 (UTC)

When editors tell you that pages with lots of words on them are difficult for them to edit, then you should believe that pages with lots of words on them are difficult for them to edit.

It's possible that other editors have other (or additional) problems. For example, Alice might have difficulty editing a page with 50K words, and Bob might have difficulty loading a page full of videos. But we do not have to solve Bob's page-loading problem to address Alice's length-of-text problem. WhatamIdoing (talk) 17:30, 14 March 2025 (UTC)

WAID, stop digging this hole. You can believe that people experience what they say they experience (including emotions). I do not because I know multiple people who have told me that they have seen ghosts. Not everything everyone says is true, people often lie or are misinformed. I know someone who believes he had a conversation with god. It is silly to pretend that endusers of a website always know the cause of the problems they experience and can always accurately diagnose problems with websites/devices as if they are some kind of techwizard. Even real techwizards cannot always accurately diagnose the source of a problem based on what they experience when using a website/device. They can make educated guesses tho. Polygnotus (talk) 17:48, 14 March 2025 (UTC)

Poly, stop complicating simple reports. Editors don't need to know "the cause". What we need to know is "the result". Lots of editors, especially those on mobile devices and sometimes on older hardware, for many years, have indicated that lengthy pages don't work for them, and that shorter ones do.

We don't need a dev team and five years to address this. We just need a rule that says to keep pages from getting too long. We've had that rule for years. (ANI flouts it, and mostly we're okay with that.) The goal here is to figure out where to draw the line. WhatamIdoing (talk) 18:04, 14 March 2025 (UTC)

If you call me Poly then that may be misinterpreted these days.

Lots of editors, especially those on mobile devices and sometimes on older hardware, for many years, have indicated that lengthy pages don't work for them, and that shorter ones do. Where is a list of those reports? They should be on Phabricator so that the WMF can take a look.

The goal here is to figure out where to draw the line. Then why are you talking about everything but the thing you want to talk about? I disagree that that is the goal because drawing an arbitrary line does not actually fix the problem.

You said not normally more than 250K and I said A 128MB page would be a bit much.. There must be a number in between we can agree on?

Making a lite version of Wikipedia is a good idea, and it would actually solve the problem for people who run into device limitations. Adding an artificial limit on this page does not. ANI is currently 300K. List of chiropterans is close to 750KB and it works fine but Wikipedia:Arbitration_Committee_Elections_December_2018/Coordination/Mass_message times out at almost 4MB. So let's say that articles should normally not be more than 1MB in text. Polygnotus (talk) 18:14, 14 March 2025 (UTC)

As I have already told you, there is no list. Also, this problem, which has a clear and known workaround, is low enough in the WMF's priorities that there is no practical point it making the list, because they are not going to do anything about it during the next decade.

And I think at this point, you should probably take a look at User:WhatamIdoing#Notice. WhatamIdoing (talk) 20:06, 14 March 2025 (UTC)

@WhatamIdoing Why should I read that again? I am aware you worked for the WMF at some point.

there is no list Then we should create one. Specifically you, since you wrote: I have seen the same issue reported by other editors.

low enough in the WMF's priorities that there is no practical point it making the list If no one reports an alleged problem to the WMF, and no one bothers to even investigate if an alleged problem actually exists (and if so in what context), and if people react this weirdly when someone tries to get some answers about in what context the alleged problem occurs, then you can't really blame the WMF for not prioritizing handling the nonexistent report.

I don't think they are perfect, probably far from it, but some people's criticism of the WMF is clearly unreasonable. Polygnotus (talk) 20:21, 14 March 2025 (UTC)

The WMF knows that some editors have trouble with large pages, because I told them that for ten years. They do not feel themselves to need any more user reports. Making a list will waste your time (yours, because I'm certainly not going to do it) and not result in the problem being solved.

In the meantime, this is a case of "Doctor, it hurts when I do this", and the workaround is "So don't do that". WhatamIdoing (talk) 21:07, 14 March 2025 (UTC)

@WhatamIdoing So, if I understand you correctly, the WMF ignores even its own employees for 10 years when they report a serious bug/problem that impacts a lot of people. If that is the case then the solution is a revolution, and either moving to another website or getting rid of the WMF. Is Jimbo part of the problem? If not then he may be able to help. If the money the community generates is not being used correctly then revolution is our only option. Polygnotus (talk) 21:11, 14 March 2025 (UTC)

A problem that has an obvious workaround is not "serious", and the number of people affected globally is not what the WMF would necessarily call "a lot".

And even if one pretends that resources are unlimited (which is never true), organizations have to make choices about what to do first, second, and so forth, and other problems actually are serious and actually do affect many, many more people. WhatamIdoing (talk) 22:06, 14 March 2025 (UTC)

But it seems unlikely that you spent 10 years telling the WMF about a problem they already knew of that only impacted a tiny group of people and was not a serious problem. Polygnotus (talk) 22:08, 14 March 2025 (UTC)

I reported all kinds of problems, repeatedly. Sometimes seemingly unimportant reports were big problems. Sometimes reports that leave you concerned about the user's blood pressure turn out to be pure PEBKAC. Sometimes trends matter, and sometimes they don't. For example, back in the day, WP:SIZE set a hard limit of 32K for articles, because some browsers simply wouldn't load any more than that, but that's no longer a problem for anyone. MediaWiki didn't have to do anything except wait for browsers to grow out of the stone age. In other cases, the trend isn't something that devs can fix. For example, younger people want more short-form video. Most of the community here does not. We're losing audience but remaining an encyclopedia. Even if you believe that's a problem (and I can argue both sides), that's not a software problem. WhatamIdoing (talk) 22:15, 14 March 2025 (UTC)

You are confusing me.

A problem that has an obvious workaround is not "serious" I can think of many problems that have obvious workarounds (and solutions), yet are still very serious problems.

the number of people affected globally is not what the WMF would necessarily call "a lot" On this planet (like most I am aware of) the people who have both fast internet and fast modern computers/smartphones are a small minority.

Do you agree that a "lite" version of Wikipedia, as described above, would be a good idea for the many millions of people who do not have fast internet and fast computers/smartphones? Polygnotus (talk) 22:52, 14 March 2025 (UTC)

There have been various "lite" versions in the past, e.g., via Google's AMP and Wikipedia Zero. Each addressed a different aspect of user problems. They wouldn't (and didn't) solve the problem raised in this thread. WhatamIdoing (talk) 22:59, 14 March 2025 (UTC)

@WhatamIdoing According to the article Wikipedia Zero was zero-rated, but not a lite version. AMP does not have the functionality I described above. Polygnotus (talk) 23:02, 14 March 2025 (UTC)

Yes, I know. You have identified your preferred solution, and nothing else matters. WhatamIdoing (talk) 23:07, 14 March 2025 (UTC)

Yes, I know. That seems unlikely. You have identified your preferred solution, and nothing else matters That sounds like projection. I considered the suggestion, explained why it is a bad idea and won't fix the problem, and proposed a solution. Polygnotus (talk) 23:09, 14 March 2025 (UTC)

See Quarry 91712 and Quarry 91714. People often think that something they don't understand is simple, and those assumptions are often wrong. https://xkcd.com/1425/ Polygnotus (talk) 18:42, 14 March 2025 (UTC)

Trolling on talk page

Any suggestions on how to deal with IP editors trolling in a talk page for a serious scientific topic, such as suggesting new scientific terminology will be based on an obscure made-up language? Jc3s5h (talk) 02:37, 13 March 2025 (UTC)

It's difficult. There will always be some who say there may be a golden light hidden in a young editor and we (that is, you) should spend half an hour per day encouraging them in the right direction. However, if you think it has gone too far, give me a link and I'll have a look. Johnuniq (talk) 04:22, 13 March 2025 (UTC)

Please consider going to Wikipedia:Requests for page protection to request WP:SEMI protection for a few days.

Alternatively, if it is sporadic, consider a "gray rock" strategy: Don't engage with the content, box up the comment to discourage others from engaging, and post the same boring boilerplate every time (you might find inspiration in Template:Uw-chat1 or Template:Not a forum). WhatamIdoing (talk) 18:03, 13 March 2025 (UTC)

An aside: we don't have Gray rock strategy (but the internet has plenty, and maybe someone should write it, or at least add something to Trolling#Responses). Thanks for mentioning it, and keep it up: always enjoy learning about new concepts, and you have a bountiful supply. Mathglot (talk) 22:30, 14 March 2025 (UTC)

wrong information

good afternoon.My name is Vitaliy Polyanskyy January 26,1981 Olympian judo players of Ukrainian Olympic team 2004.i found on my page false information .could you remove kind of this information about me. I am not a member of the lGBTQ community. thank you 2607:FEA8:3B80:74C:6551:F098:77DF:3C98 (talk) 17:04, 24 March 2025 (UTC)

This isn't the correct place to ask for help, but I've removed all of the unsourced claims that an IP editor added to that article about 2 weeks ago. Schazjmd (talk) 17:10, 24 March 2025 (UTC)