Wikipedia:Village pump (miscellaneous)

 Policy Technical Proposals Idea lab WMF Miscellaneous 

• Table of contents
• First discussion
• Jump to bottom
• All
• New topic

The miscellaneous section of the village pump is used to post messages that do not fit into any other category. Please post on the policy, technical, or proposals sections when appropriate, or at the help desk for assistance. For general knowledge questions, please use the reference desk.

For questions about a wiki that is not the English Wikipedia, please post at m:Wikimedia Forum instead.

Discussions are automatically archived after remaining inactive for 8 days.

« Archives, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84

Wikipedia and open knowledge ideals

Ever since template {{Backwards copy}} was created in 2008 it has garnered only about 1,900 usages. Can that be right? Enwiki has over 7 million articles, a zillion views yearly, and is frequently copied in many ways (sentences, paragraphs, whole-cloth). Some of this is because we don't track backwards copies very well, but still, 1.9k out of 7 million pages beggars belief. Evidently, our Attribution-ShareAlike license is not being honored very often.

The victim is not WMF, who have little to gain through enforcement of licensing, other than a legal bill. Rather Wikipedia authors receive no credit for their creative efforts. The other victim, most seriously, is open knowledge, since the visionary idea was to create an ecosystem of open knowledge producers, consumers and re-users. Wikipedia has abundant producers and consumers, but re-users are not with the program. The open knowledge system exemplified by CC-BY-SA is not working well in practice. Violators face no consequence, and nobody is enforcing it anyway, much less monitoring and reporting.

Some may contend it is beyond the scope of Wikipedia to determine if an external website is in violation of Wikipedia licensing. But we can and do make copyright determinations, frequently, see WP:Media copyright questions for example. Yet, it's not being done for CC-BY-SA violators. There is no template to flag it. No noticeboard for discussions. (I think)

The other argument is this problem exists any place open knowledge is produced, this is true. But I think Wikipedia in particular has a reputation as de-facto Public Domain. Unlike, say, a professionally produced LaTeX document with a DOI and a CC-BY-SA license, is more likely to make reusers think twice about how they reuse it.

Personally as a content creator I love Wikipedia for many reasons as a place to publish my research. However it's a bit Hotel California. Easy to walk into, hard to get out of, and broadly fails to protect creators from those who disregard Attribution/ShareAlike. None of this mattered to me, at first, but looking back over decades of producing content, my oeuvre feels increasingly at risk by a lack of licensing compliance, enforcement, or monitoring. The obscure and little known maintenance template {{Backwards copy}} reveals something. -- GreenC 06:44, 21 September 2025 (UTC)

I don't usually bother with that template unless the Wikipedia article has been incorrectly reported as the copy. WhatamIdoing (talk) 04:14, 23 September 2025 (UTC)

Yup, "Some of this is because we don't track backwards copies very well". Still.. some people will use the template. They are not finding many instances of Wikipedia content being credited. -- GreenC 05:27, 23 September 2025 (UTC)

I don't know of anyone who deliberately seeks out Wikipedia:Mirrors and forks these days, much less the one-off copies of an individual article. Maybe you do? The conclusion I'd reach is that most mirrors get filtered out of search results (so people are very unlikely to see them) and maybe even that small, credited uses (a paragraph in a news article that says "According to Wikipedia, a ___ is a...") aren't what we want to track either at all, or with this template (vs Template:Press). WhatamIdoing (talk) 16:37, 23 September 2025 (UTC)

Yes, I've come across my own work being used and not attributed or licensed properly. I don't actively look for it, but I come across it when updating my older work searching for new sources. I think most people would if they took the time, at least if you are producing good quality+ articles that are well sourced and well written. I've seen my work in a "List of .." article become the basis of a government research report in India - no attribution or CC license. I've seen my work about a food topic form the basis of a newspaper article in an English newspaper - no attribution or license. And why not? If you steal from Wikipedia and obscure it well enough, you can get your assignment done, get paid, and take the credit. -- GreenC 17:04, 23 September 2025 (UTC)

I've had the same experience, both with my own contributions and with other people's contributions (usually via Wikipedia:Citation Hunt). But I don't deliberately seek it out.

A few years ago, a new "medical" journal blatantly copied several articles. (Those have all been tagged.) I think that unlabeled partial re-use is a bigger risk for expanding our Wikipedia:List of citogenesis incidents than whole articles. WhatamIdoing (talk) 18:07, 23 September 2025 (UTC)

Right welcome to the club! I bet this is true for every serious contributor. Partial re-use is usually what a backwards copy consists of. Per the template: where there may be potential for confusion about the copyright status of parts of the text due to its re-use in a mainstream news article or publication. Basically any part of the Wikipedia article that could be mistaken as a copyvio, is a backwards copy. -- GreenC 18:27, 23 September 2025 (UTC)

Funny enough this came up at Talk:Jane_Goodall#Copyright_violations .. I thought there was a copyvio, turns out it was a backwards copy! I found the supposed "copyvio" with the new Copyvio detection tool (history tab). In the end though, there is no place to report the website as violating Wikipedia CC-BY-SA. It would be useful to check other pages where this website is sourced, and build a picture of how reliable the source is for discussion at WP:RSN (a single known violation is not enough for RSN). -- GreenC 17:49, 2 October 2025 (UTC)

I think that ever since the onset of LLMs trained on Wikipedia, we might have a bigger problem than backwards copies. 173.206.37.177 (talk) 18:20, 2 October 2025 (UTC)

I honestly prefer LLMs to be trained on Wikipedia rather than Twitter (or whatever the fuck Musk calls it these days). Mikeycdiamond (talk) 18:33, 2 October 2025 (UTC)

Multiscript detection

A new maintenance page is available at Wikipedia:WikiProject Academic Journals/Journals cited by Wikipedia/Maintenance/Multiscript1. This essential detects the usage of multiple scripts in the |journal= parameter of citations, color coded for ease of browsing (with apologies for the color blind, but that's a thing that could be improved in the future). For example, the entry Abteilung Verhaltensökologie, Institut für Ӧkologie & Evolution, Universität Bern tells you that the Cyrillic character Ӧ is used, instead of the latin Ö in the article Xyleborinus saxesenii. So I went there and fixed it.

There's many cases like this that needs to be handled. Many entries aren't actually problematic, like Բանբեր Հայագիտության = Вестник Арменоведения = Journal of Armenian Studies, but many will be. Any help you can give will be greatly appreciated. Especially on Asian and Arabic alphabets.

Thanks ahead of things to anyone that helps! Headbomb {t · c · p · b} 21:05, 21 September 2025 (UTC)

Somebody worked out a really huge list of homoglyphs that would be a good resource for you; I'll see if I can relocate it. Mathglot (talk) 09:44, 26 September 2025 (UTC)

This isn't it, but pretty big, and more recent, iirc. Mathglot (talk) 09:46, 26 September 2025 (UTC)

Congratulations to Internet Archive

I'd like to congratulate Internet Archive from here, for 2 reasons: first, the lawsuit that could have threatened its existence was finally succesfully solved, and, second, next month, they will be celebrating the first trillion of webpages archived in Wayback Machine.

That said, I hope they are really able to preserve their archives, as they say, for future generations. There is no information about full copies of Archive out of San Francisco Bay Area, and that's a concern that has worried me for years, and that I try to mention whenever I can. More than 175 petabytes (source) is a really insane amount of storage space. I think that, with Archive's budget, not many copies of that can be kept. Ideally, areas with high natural risks should be avoided at all for hosting even a single copy, and, of course, at least 1 copy should be out of them. WMF is also based in San Francisco, but it hosts all content and backups far away from that area, and seismic risk is probably the main reason for it (and WMF has a far bigger budget and a far, far smaller amount of data to store). I hope I am too paranoid, I hope that they have an excellent earthquake-resistant infrastructure, that they keep secret full backups at safe places... but these are only hopes, there is no evidence that most of Archive's data can survive a really strong earthquake.

Internet Archive, and, especially, Wayback Machine, are a vital resource for Wikipedia, since they are the only possible way that defunct webpages used as citations in Wikipedia articles remain accessible. Wayback Machine is also the only way deleted (for example, merged) Wikipedia articles remain publicly accessible. Both things also apply to other WMF wikis. These are good reasons to worry about the risks facing Archive and Wayback Machine. I fear that, after the lawsuit is ended, people relax and think that Archive is fully at safe, now, when it isn't (and has never been). MGeog2022 (talk) 12:13, 22 September 2025 (UTC)

All good points, but largely lost on a guy who gives this round of Western civilization about another hundred years (we almost lost it in 2008, and we didn't seem to learn much from the experience). We have created a world that we can't manage. Right now, there are smart people plotting how best to accelerate the collapse.

Then Dark Ages II (Dark Ages I followed the collapse of the Western Roman Empire). Nothing like Wikipedia, or Internet Archive, or other types of computerized archives, will survive anyway. Just imagine a world without computers and internet, and that's just the start of it. Tell your grandchildren: "If you love your kids and their kids, don't have kids. Don't be selfish." Signed, Ted Kaczynski. P.S. This drove me crazy, to the point I was blowing shit up and killing innocent people. Two errors: I cared too much, and I thought I could change the course of history in a positive direction through violence. All I accomplished was to put myself in prison, where I later hanged myself. ―Mandruss ☎ IMO. 13:14, 22 September 2025 (UTC)

When I read "The Coming Dark Age" in 1974 I got worried. I got into a sort of prepper mindset (although I never put much money into prepping). Anyway, Vacca discussed a lot of systems that were failing then, and his argument was very convincing. And yet, here we are, 51 years later, using all kinds of tech that even science fiction writers hadn't thought of then. Donald Albury 15:14, 22 September 2025 (UTC)

There is a re-issue of Vacca's work from 2000 that includes "hindsight" commentary. One of the notable additions is the unexpected appearance of cell phones. -- Reconrabbit 16:39, 22 September 2025 (UTC)

a few people expected it, like Arthur C. Clarke, Robert Heinlein, and Nikola Tesla to name a few. Andre🚐 04:47, 23 September 2025 (UTC)

Yes, apocalyptic thoughts are a constant throughout human history. At some moments during Cold War, many people believed that nuclear war was imminent, but it didn't happen (in fact, most of the world's population now live better than they lived then). When you think about how many supposed "ends of the world" took place in recent history, you realize that it isn't likely to happen in the near future. In fact, one of the biggest risks can be if most people think it will happen, and behave like it's just going to happen, so they don't care at all about the future. In 1929, Western civilization survived a financial crisis far worse than that of 2008, and it even survived World War II after that. There is no reason to believe civilization will collapse soon: current times can't be compared with Roman Empire at all. Recently, obesity surpassed hunger as a risk to global health: that is not what you would expect from a collapsing civilization. Let's care about people living 100 years from now having Wikipedia and Internet Archive, with all their contents well preserved, so humanity doesn't lose its cultural memory. MGeog2022 (talk) 19:05, 22 September 2025 (UTC)

I'll wager two weeks' Wikipedia unpay that you'll be proven wrong. I will come back from the grave to collect. (Of course, Wikipedia unpay will be worthless then, but wth.) ―Mandruss ☎ IMO. Warning: May contain undeclared humor. 00:55, 23 September 2025 (UTC)

When you think about how many supposed "ends of the world" took place in recent history, you realize that it isn't likely to happen in the near future. I believe that's fallacious logic, similar to saying that the dice's next roll is influenced by the previous twenty. Or, that California's Big Quake isn't likely to happen in the near future because a whole lot of predictions (all of them) have been false alarms. Regrettably, I can't link to a Wikipedia article for support.

FWIW, I'm speaking of the end of this round of Western civilization, not the end of the world. Much or most of this precedent you cite was about the total destruction of the planet, or at least extinction of the human species or its subjugation by apes. That's apocalypse; I'm talking about mere catastrophe. The species will survive, and after a few centuries will begin to rebuild, just as we did the last time. But it will have to re-invent all the technology invented since the end of Dark Ages I. Including, eventually, computers, the internet, and Wikipedia. And Facebook. We won't have gained any wisdom that survived Dark Ages II, so we'll make all the same human mistakes again (lust for money and power, etc) and Round 3 will collapse for similar reasons. "Those who cannot remember the past are condemned to repeat it." Rinse. Repeat until scalp is raw. Until we can genetically create wisdom and have the political will to do so on a global scale.

"Post-apocalyptic" is misused in my opinion, as in The Postman (film), simply because we haven't found a better word for it. "Post-catastrophic" wouldn't do. ―Mandruss ☎ IMO. 06:09, 23 September 2025 (UTC)

I understand, but, unlike for California's Big Quake, I don't see a reason for that happening in the medium term. But I see a reason for a possible partial Dark Ages II: a strong quake destroying most of Internet Archive's collections, if they have no copies at safe places, and if their datacenters (including storage media) are not built to withstand any possible earthquake that may occur in that area. So we may partially agree, returning to my starting point :-) MGeog2022 (talk) 12:02, 23 September 2025 (UTC)

It isn't about the past rolls of the dice we have won, it is about the human perseverance and stubbornness that allowed us to win. It is in human nature to want to preserve what we have, and to seek better things. Western civilization isn't going anywhere, it will change but not collapse. Mikeycdiamond (talk) 17:58, 2 October 2025 (UTC)

In my personal opinion, the plaintiffs come in talking big fines, so they can get better terms on the inevitable private settlement agreement. Nobody wants to actually destroy Internet Archive, including the judges. -- GreenC 16:54, 22 September 2025 (UTC)

In re WMF...hosts all content and backups far away from that area, and seismic risk is probably the main reason: Nope. The wikitech:ulsfo data center is in the Bayview–Hunters Point neighborhood of San Francisco. WhatamIdoing (talk) 04:29, 23 September 2025 (UTC)

The point is that (I hope) the WMF hosts at least one copy of everything worth preserving away from San Francisco (or any other single point of failure). It would be nice if the Internet Archive could afford to do so too. Certes (talk) 08:28, 23 September 2025 (UTC)

ulsfo is an edge caching datacenter, not a data storage one (there are only 2 of this type: eqiad and codfw, and none of them is in a seismic area). MGeog2022 (talk) 11:54, 23 September 2025 (UTC)

Perhaps more importantly, every place on the planet has some sort of natural disaster risk. For example, Eqiad is in Virginia; see also Category:Hurricanes in Virginia. Codfw is in Texas; see also Category:Tornadoes in Texas. The goal isn't "avoid the natural disaster that seems scariest" but "have backups". The m:Server switch exercise has really improved switching time, and even if one of the main centers goes down, whether due to a natural disaster or an unnatural one, editing should be feasible within minutes (not hours or days), and many readers won't notice it at all. WhatamIdoing (talk) 18:23, 23 September 2025 (UTC)

Good point. Both WMF main datacenters are also in areas with natural risks, but there are backups at a very distant place (eqiad and codfw are really far away from each other). If we rely on public, verifiable information, that's not the case with Internet Archive, sadly. All main datacenters are in relatively close proximity (less than 30 km, I believe), so they could be damaged by the same natural disaster. MGeog2022 (talk) 19:44, 23 September 2025 (UTC)

What is the interest to talk about "Internet Archive" there ?

The projects of "Wikimedia Foundation" and "Internet Archive" have not a common body because they are independent of each others.

As they are independent of each others , what is the interest to talk about it there ?

I acknowledge that I'm worry about conservation of pages by "Internet Archive". I hope that there are copies around the world. Anatole-berthe (talk) 23:10, 23 September 2025 (UTC)

We are dependent on the Internet Archive because it preserves content that has been cited in articles but is no longer available on the Internet. There have been other archiving services, but some have gone under. If the Internet Archive were to go under and its archives lost, then a lot of content in WP articles would no longer be supported by sources that could be inspected. Donald Albury 23:30, 23 September 2025 (UTC)

What can we do ? I don't know an alternative equivalent to Internet Archive ?

I know others archiving sites but they haven't the same performances.

I fear that it is something for which Wikipedians can't do anything except use more archiving sites while citing a source. Anatole-berthe (talk) 23:40, 23 September 2025 (UTC)

Well, one thing we can do is to make ourselves less dependent on any archiving system, e.g., by citing more books and academic journal articles instead of ephemeral websites. Another thing we can do is to update articles with newer sources that say the same thing. WhatamIdoing (talk) 03:53, 24 September 2025 (UTC)

Your idea please to me. What to do when sources are only news websites concerning a particular sentence in an article ? Anatole-berthe (talk) 06:29, 24 September 2025 (UTC)

Perhaps WMF could openly discuss this with Archive, since Archive's Wayback Machine is a much needed resource that Wikipedia and other WMF wikis rely on for storing sources. Many web sources can't be replaced by paper ones, and for those that can, it would be a formidable work to search for the new sources (visits to paper libraries to find alternative sources for millions of citations, and for many of them there are no other source than the web one). It's not practical at all.

I tend to be paranoid and look for possible dangers, so risks can be identified. I suppose people at Internet Archive are not stupid at all: they still keep webpages from nearly 30 years ago without problem, and they claim to store all content for future generations, and, so far, they are accomplishing that. The proof of fire for this is the one that has not still happened, but, sadly, will happen at some point in the future.

Logic would say that they are well prepared for that (for example, in Japan there are datacenters, including equipment, that are built to resist strong earthquakes without problems, could Archive's ones be built in the same way?; or maybe they don't have primary datacenters out of that area, but they do yearly backups to tape and store that tapes far away), but there is no public proof about that, so it is a major reason for concern. MGeog2022 (talk) 12:40, 24 September 2025 (UTC)

As can be read here, back in 2016, Archive's infrastructure was nearly as bad as it could be. They had meetings with Long Now Foundation, talking about preserving content for people living in year 10,000, they talked about forever and future generations... but they were storing only 2 paired copies of almost all content, a few km apart in a highly seismic area.

The worst part (it can be read in the comments section of the linked post): a person from Archive's staff happily talked about how they didn't make backups at all, and how even storing a third paired copy was impossible for them. If that person said that they were fully conscious about how extremely critical their situation was, and how hard they were working to improve it, I wouldn't be as worried as I am. But when risks (a simple fire in a rack would mean that petabytes of content would be left with only 1 copy) are underestimated in such a way, I must said (and I don't like at all saying this) that I don't trust them at all. If you want to preserve content for future generations, don't store so much content that you can't assure it will preserved for next year, let alone future generations. This seems something like diogenes syndrome.

WMF needs a solution for its links to online references. If Archive is now reliable, or it becomes reliable in the future, human, in addition to financial and technical, changes are very much needed. MGeog2022 (talk) 12:41, 26 September 2025 (UTC)

To highlight the positive aspects about Internet Archive (I don't want this congratulations to end up being just a voracious criticism), I'll add that they do an excellent job in digital preservation for their partner institutions, through Archive-It service, and also providing easy access to Common Crawl archives. The contents preserved by both Archive-It and Common Crawl are very likely to remain there indefinitely, for future generations (both of them are well backuped, and it's publicly known). So IA is vital for the long-term preservation and availability of some content, but, at least for the moment, there is no reason to believe that this applies to all the content they offer. From the available evidence, the vast majority of the content, while very useful for use at the present moment, can't be truly considered as preserved, and that matters if you want the references at Wikipedia articles to be verifiable in the far future. MGeog2022 (talk) 10:20, 27 September 2025 (UTC)

We need our own archive

As much as I love (trust, respect, etc) IA, we really need to have our own archive of documents we use as sources in our articles. While the obvious risk to IA is physical destruction due to an earthquake, there's also external factors to worry about. While it may be unthinkable that some outside entity might buy IA and destroy it, the unthinkable happens. Biodiversity Heritage Library is just climbing out of an existential threat caused by the Smithsonian deciding to no longer host them. MySQL (not to mention Java) is now owned by Oracle. Freenode got eaten by a hostile entity. Deja News got bought by Google. These things happen. They're rare, but they do happen, and for a resource we are so dependent on, we really need to have a better plan than hoping it will never happen to IA. RoySmith (talk) 12:58, 24 September 2025 (UTC)

IA is a non-profit, I think that the situation you talk about is almost as unlikely for IA than it is for WMF. For example, MySQL and Java belonged to commercial companies, that were bought by Oracle (and, despite that, they still are free and open source software). I think that risks where evil ones are present are often overperceived, and those without evil ones are often underperceived.

WMF doesn't seem to want to store non-free content as it would be needed for such an archive, and it relies on Internet Archive for that (in fact, I believe there is some kind of cooperation between them about it). The matter here is how reliable IA is in the long term (it could already be fine, but nobody outside of IA knows it). MGeog2022 (talk) 13:17, 24 September 2025 (UTC)

IA, like the Foundation, is dependent on donations (I have donated several times.) My impression is that they are not as secure financially as the Foundation is. Non-profits get bought up by for-profits all too often as an alternative to going under (pretty much the norm for hospitals in the US, for example). That is why some people have called for the Foundation to explore helping support IA. Donald Albury 15:02, 24 September 2025 (UTC)

I feel like we don't need our own version of IA so much as we need a backup on a hard disk somewhere so that if the worst happens we could always rebuild. Bawolff (talk) 18:36, 24 September 2025 (UTC)

Having a backup of the data is a necessary first step, but anything as large as IA has lots of moving parts. There's a web front end, databases, tools, APIs, indexers, backups, infrastructure, config files, and all sorts of chazzerai that make the difference between a pile of bits and a working system. Not to mention the institutional memory of how all the pieces plug together.

WMF practices data center failover twice a year (I think the last one was just a couple of days ago). If you're not doing that, when the time comes that you need it, all you'll have is a pile of bits and a world of hurt. RoySmith (talk) 20:56, 24 September 2025 (UTC)

When someone asks me how to support Wikipedia, I tell them to donate to the Internet Archive. I am less concerned about someone buying them, but lawsuits are a nontrivial risk and have made the IA worse already. —Kusma (talk) 21:19, 24 September 2025 (UTC)

If we need one (which isn't clear to me), it's probably simpler (and cheaper) to buy one than to build one. French Wikipedia has long used Wikiwix; it is a very small company, and if we want to go that route, maybe we should just buy them. You could ask over at French Wikipedia about how they like it there (try: WP:Questions techniques). Or maybe just ask knowledgeable and friendly French admin Jules* what the general opinion of Wikiwix is by French users. Mathglot (talk) 09:40, 26 September 2025 (UTC)

I'm a contributor on "Wikipedia in French".

I don't use "Wikiwix" for my contributions over there because I prefer "Wayback Machine" founded by the "Internet Archive".

I think that we should use multiple archiving services for all sources used.

A source can be archived on service 1 , service 2 , service 3 , service 4 , service 5 etc...

I think that 5 archiving services should be the minimum for all language versions of Wikipedia. Anatole-berthe (talk) 08:46, 2 October 2025 (UTC)

Maybe a better solution would be not relying on a single entity?

I hate how archive.today and Ghostarchive are often overlooked in such discussions. Sapphaline (talk) 08:38, 1 October 2025 (UTC)

It would be a great idea if the same source is archived at multiple archive sites at the same time. But bear in mind that neither archive.today nor Ghostarchive publicly disclose who is behind them, so they could close tomorrow without previous warning. That's why I am so concerned about this: the most reliable archive is Internet Archive, and, probably, it isn't reliable. MGeog2022 (talk) 12:42, 1 October 2025 (UTC)

It's true that an archive provider can close tomorrow. But I think that we have to think to others risks even those that seems unlilkely like a WWIII (Nuclear or not) or a war implicating many countries in Europe or elsewhere.

As others risks that seems unlikely , I think to big asteroids and the like that can destroy a region in a country. Anatole-berthe (talk) 08:56, 2 October 2025 (UTC)

Nothing and nobody would survive a large scale nuclear war. As for a war implicating many countries in Europe, with NATO having 3.5 million soldiers and personnel, and more than half of global military spending (and growing), it also seems really unlikely.

There has been no such a big asteroid incident in recorded history (Tunguska event would be the most similar case), so the earthquake risk is highly likely to hit Internet Archive far before that kind of incident ever happens. And, even before that, other archives with unknown owners may also close because of an unknown reason. MGeog2022 (talk) 13:34, 2 October 2025 (UTC)

Edit farming

Hopefully, this wouldn't be a bad time to discuss this, but...

Edit farming! What does it mean?

Edit farming means that you are artificially editing pages, sections or talks on Wikipedia to indirectly increase your edit count for gain of relevancy. It is also to help people maliciously gain notoriety as a user or to be given barnstars. Each edit one by one should be unnecessary (especially by me). Instead, just think of one thing that is relevant, reliable and legitimate. Just put your heart into it to just edit one proper section in a better quality and size. Not a small portion one by one to make yourself notable. Not right. Rather than low-quality, why not make an edit here on WP high-quality? That's what I learned.

Think about. Edit one page with only one statement, but then... "oh, I forgot to add sources to reference", then you add a reference, but it's from a non-reliable source after you publish it. Then, you revert the edit? No, you re-edit it again without reverting it. Then, you got to another article and then edit it with a reference, save it. Go back edit with another after another one by one. You think, there's nothing wrong? Guess what, it its.

The repeating low-quality, artificial and low-end edits cannot be overlooked. The edit-farming problem on here, Fandom and Miraheze (which holds the AVID Wiki) can be a serious online struggle for us users and editors. It can lead to serious consequences such as vandalism, bad faith and banishment. So to any user who thinks that edit farming is appropriate here, they are wrong! Be careful!

Maybe we should look into edit farming one day and review. If my discussion isn't good enough, then I could be lying.

Darrion N. Brown 🙂 (my talk page / my sandbox) 06:47, 26 September 2025 (UTC)

Different editors have different styles. There was a time when I got into the habit of making many small edits to an article because when I tried to make a single long edit, my log-in session would sometimes drop before I could finish the edit. That is not a particular problem now, but I still find myself having to go back to make several small edits after a major edit, because I'm terrible at proof-reading my own writing, and some errors (SFN cites, for example) do not show in a preview, but are glaring after saving. Gaming is occasionally a problem, but tends to be obvious to experienced editors. Edit-count is a very rough indicator of contribution levels, and my opinion of other editors depends much more on the quality of edits I see them make, rather than the quantity. Donald Albury 19:36, 26 September 2025 (UTC)

I use to see S.E.'s as normal until I saw the "farming" problem on Fandom. But if any Wikipedian felt like I was wrong for bringing this here, then I owe you all an apology. It wasn't in good faith, it was a bad opinion. Even so, I had no business bringing that up out of spite. I know I keep messing up. Darrion N. Brown 🙂 (my talk page / my sandbox) 04:12, 27 September 2025 (UTC)

I was overthinking certain things. I have no excuse to pull about it. Darrion N. Brown 🙂 (my talk page / my sandbox) 04:13, 27 September 2025 (UTC)

You were not wrong for bringing it up. You saw something that you perceived as a problem, and you raised it in an appropriate forum. I happen to see the situation differently, but that is not a criticism of you. Wikipedia is supposed to be a collaborative effort, with good faith assumed until proven otherwise, and disagreements settled by discussion, hopefully leading to consensus. I have proposed or backed a number of ideas over the years that have been rejected or just ignored by the community. That's just the way we work. Suggestions and complaints are not a problem unless and until they are pushed long after it is clear that the community is not interested. Donald Albury 13:50, 27 September 2025 (UTC)

While what Donald Albury said is overall right, partly contra to him I want to add that for editors that edit mainly in a controversial topic area (or even worse, pivot to a controversial topic area suddenly), making each of their early edits into several smaller ones can be a red flag for malicious intent. Loki (talk) 20:23, 27 September 2025 (UTC)

As I said, gaming is occasionally a problem, but we should AGF until shown otherwise. Donald Albury 21:11, 27 September 2025 (UTC)

Additionally, making several smaller edits make instead of a massive article revamp lets people revert problematic individual edits, rather than have to untangle 2% shit from 98% quality. Headbomb {t · c · p · b} 21:21, 27 September 2025 (UTC)

@DBrown SPS: If it is done in good faith, there is nothing wrong with making many small edits instead of one large edit. In your example of doing this editing in order to gain some user rights or awards, then that is covered by WP:GAMING. RudolfRed (talk) 22:02, 27 September 2025 (UTC)

User:Larry Sanger/Nine Theses

For the interested. Gråbergs Gråa Sång (talk) 21:20, 29 September 2025 (UTC)

I agree with point 1, that "consensus" doesn't really work well at a community else of this size. I won't lower myself to commenting on the other eight theses except to say that grovelling bothesidism in articles isn't a good look. Cremastra (talk · contribs) 01:14, 30 September 2025 (UTC)

[After reading more of the essay] Wow, he really doesn't understand what AGF means in practice, does he? Cremastra (talk · contribs) 01:20, 30 September 2025 (UTC)

A lot of editors don't understand that guideline. WhatamIdoing (talk) 02:23, 30 September 2025 (UTC)

What would you replace consensus with? Consensus is messy, but what would work better? Off hand, I think any other system will either lead to anarchy, or to getting mired down in more bureaucracy than the community could stand. I certainly think Citizendium has shown us the perils of putting too much control in the hands of an elite of experts. Donald Albury 01:36, 30 September 2025 (UTC)

Sangar has started and moved on from other projects since leaving Wikipedia. I fully expect that the only consensus he could ever fully agree with is his own. His idea of what a neutral, his idea of what is balanced, not working with other people to come together and decide on such matters. His latest musings only serve to solidify that opinion. -- LCU ActivelyDisinterested «@» °∆t° 15:21, 1 October 2025 (UTC)

My super brief, and polite as possible, summary: We're hostile to fringe and other views that are not documented in reliable sources, including to those editors that try to promote those views. But in reality, fact-based knowledge is left-leaning, and by that nature are going to be highly suspect of right-leaning views, publications that primarily deal in promoting those views, and editors that try to push those views. Or more fundamentally, it's proposing we should cover all sides of a topic with completely equal weight, which we know just doesn't work over several years of work. Masem (t) 01:36, 30 September 2025 (UTC)

I must object to the statement that fact-based knowledge is left-leaning. My politics are very definitely left/libertarian or left/progressive, according to the on-line political tests I have taken, but I have known professed leftists who were not really committed to "fact-based knowledge". I think it is counter-productive for the future of Wikipedia to tie support for fact-based knowledge to any particular political belief. Similarly, we should not try to tie Know Nothingism to all politically conservative groups. Donald Albury 01:59, 30 September 2025 (UTC)

Fair enough, though I would say most of the time I see the type of arguments (like, being "neutral" by presenting both sides equally) are common talking points with the right, and some of the other aspects, like identifying the major contributors, are ideas frequently raised by right-leaning politicians and also mirrors aspects like the ADL. I agree that some of those points can be raised anywhere on the political spectrum, but most often is coming from those that want to push right-leaning concepts onto WP. Masem (t) 03:37, 30 September 2025 (UTC)

• The way I would put it, instead, is that Wikipedia is aimed at being academic, international, and mainstream. Large swaths of the modern right are to some extent or another hostile to all three of these things. Nationalists, for instance, will rarely be happy with what we say about their nations. People who place their religions on the level of sacred revealed truth and who believe that they should be treated that way in all contexts are likewise going to have trouble with an approach that reduces them to just another primary source. And anyone with an anti-intellectual mindset or who believes that academia or the mainstream media are biased as a whole are going to be unhappy with what we say, since we largely summarize these things. We strive to be neutral but our definition of neutrality is derived from our main purpose as an encyclopedia, and is not going to align with everyone's views; if someone's idea of neutrality is to listen to what the two main political parties in their country say and to try and find a midpoint between them, they're naturally going to be unhappy here. But that's inevitable! That approach is, simply put, not encyclopedic - certainly not when writing an international, academia-focused encyclopedia like this one. It's worth pointing out that our articles do reflect plenty of right-wing ideas (especially on economics or when it comes to more libertarian views), in contexts where those views are treated seriously in academia. And I don't think it's a surprise that those are also the right-wing views that are the most international in character and the least tied to one specific religious faith or the like. --Aquillion (talk) 16:45, 30 September 2025 (UTC)

(edit conflict) IMO it's not so much "fact-based knowledge is left-leaning" as it is that, in current Western politics, groups on the right have set themselves up in opposition to fact-based knowledge (and been politically successful by doing so) far more often than those on the left have. Anomie⚔ 02:05, 30 September 2025 (UTC)

I agree with you there. When I was growing up in the 1960s and 70s it was the conservatives who placed most emphasis on facts, but now it is the left. Phil Bridger (talk) 19:24, 30 September 2025 (UTC)

Encyclopedias are traditionally left-leaning. Recall the Encyclopédie: Because of its sometimes radical contents, the Encyclopédie stirred up controversy in conservative circles, and after the publication of the second volume, it was briefly suspended by a royal edict of 1752 accusing it of "destroying royal authority, fomenting a spirit of independence and revolt, and ... laying the groundwork for error, for the corruption of morals, and for irreligion and atheism." Sound familiar? Our current rulers want to return to the good old days of the Ancien régime, the French Rrevolution was a travesty, and everything since has been a disaster. The few need to rule the many, authority never questioned, and rulers stay in power for life. A lot like China and Russia today, is what they want. -- GreenC 20:07, 30 September 2025 (UTC)

The Encyclopedia Britannica on the other hand, was very conservative for some time. Cremastra (talk · contribs) 21:10, 30 September 2025 (UTC)

So, TL;DR, this entire piece is just a reframing of Sanger's prior complaints about Wikipedia where the fact that we don't allow fringe conspiracy nonsense and false claims to be stated in Wikivoice is the height of impropriety in his view. Especially when it comes to not pushing a conservative POV on everything. Silverseren^C 02:50, 30 September 2025 (UTC)

7 would be an utter disaster, as demonstrated by each of the millions of review bombed topics across the Internet. Saying that Amazon reviews are something we should emulate??? Gnomingstuff (talk) 02:52, 30 September 2025 (UTC)

The mitigations for that are even worse. Using AI for this is a bad idea, as anyone who's seen the volume of promotional text inserted into articles alongside an edit summary likely "Made text compliant with Wikipedia's neutral point of view" would know. Asking users to verify with their credit card or ID is a bad idea especially given that we had a massive data breach only a few months ago. At one point this asks for someone to invent userboxes. Gnomingstuff (talk) 03:00, 30 September 2025 (UTC)

@Gnomingstuff wait, we had a data breach? Gaismagorm (talk) 13:44, 30 September 2025 (UTC)

This is the one I recall: m:Wikimedia Foundation/March 2025 discovery of account compromises. ClaudineChionh (she/her · talk · email · global) 14:15, 30 September 2025 (UTC)

That's not a data breach; it's people using data breaches elsewhere to perform credential stuffing attacks on Wikipedia. I think it's more likely Gnomingstuff is referring to Wikipedia:Village_pump_(technical)/Archive_223#alleged_"massive_leak_on_Wikimedia_sites". * Pppery * _{it has begun...} 15:31, 30 September 2025 (UTC)

Yeah that was it, I hadn't heard of the March incident. Even more reason why we shouldn't ask people for their IDs. Gnomingstuff (talk) 02:25, 1 October 2025 (UTC)

The March incident is actually probably more relevant. The "leak" was all of info that was at some point public; there was no actual security lapse on the WMF's part, although it's a good cautionary tale that sometimes OSing old revisions can increase visibility rather than decrease. On the other hand, with the account compromises, a big part of why that wasn't a huge deal was that there wasn't much PII for attackers to glean from the accounts they compromised: just their emails (which they probably already had), their watchlist contents (usually not that identifying), and in a minority of cases their timezone. 35,893 compromises would be massively more consequential if there were credit cards or IDs tied in there. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 02:08, 2 October 2025 (UTC)

I agree. The average person on the internet is, frankly, a fucking idiot. We already have to put up with a deluge of IPs and new users who don't have a clue how Wikipedia works and just want to complain that it isn't how they personally want it to be. I don't see why we'd make that worse. — Czello ^(music) 14:06, 30 September 2025 (UTC)

Also, don't articles already have pseudo-ratings? Like, A-class, stub-class, good-article, featured-article etc.? Mr. Sanger seems to be suggesting something that already exists. Besides, most articles would never get rated if it was done this way. Gaismagorm (talk) 15:27, 30 September 2025 (UTC)

He wants either populist ratings (i.e. open to offsite vote brigading), or ratings by "experts" (presumably hoping a significant proportion of those "exports" will agree with his viewpoints), or rating by AI (???). Anomie⚔ 17:29, 30 September 2025 (UTC)

From the one person who could make me thankful that we got Jimmy Wales. Just remind me; which of his projects in the last 20-and-a-bit years has had even a small amount of success? Phil Bridger (talk) 08:43, 30 September 2025 (UTC)

Thesis 6 is a devastatingly bad idea. Given the amount of hostility already directed towards Wikipedia (particuarly where politics is concerned), requiring the most prominent Wikipedians to self-dox is, frankly, dangerous. — Czello ^(music) 14:11, 30 September 2025 (UTC)

I'd say thesis 2 is a close second, having multiple "versions" of articles where people POV-push their own points is such a bad idea I found it absurd to read it and think someone honestly thought that would be good to introduce, especially written so matter-of-factly. Sophisticatedevening^(talk) 15:16, 30 September 2025 (UTC)

Thesis 2 would actually be great in an ideal world where the reader would be able enough to decide the truth for themselves and both sides did not regularly engage in 500 page flame wars 95.5.189.54 (talk) 21:34, 1 October 2025 (UTC)

You know, I'm not the hugest fan of Larry Sanger, but #9 makes a good point. It can be a little bit tricky to figure out what is policy and what is opinion. I'm not saying we need some giant codex of rules, but making them easier to access and find would be a really good idea. It's also be nice to have some way of storing major consensus agreements somewhere. All too often, I'll have to skim through a few dozen pages just to figure out how to capitalize something, or whether we should spell something one way or the other way. I don't really like the rest of the suggestions, but #9 at least makes a decent point. Gaismagorm (talk) 15:31, 30 September 2025 (UTC)

Absolutely would support a template for talk pages of P&G to list out all major RFC that impacted the P&G. Often I see key ones highlighted separately from page archives, but realistically a simplate with a collapsible list of all such RFCs could help navigate the past discussions Masem (t) 15:50, 30 September 2025 (UTC)

When I learned that Sanger had found Jesus[1] I wasn't expecting him to unironically go full Martin Luther nor to forget what a questionable figure the historical Luther was (although to be fair half remembering history is very much in Sanger's wheelhouse)... One wonders whether the new 95 thesis will be followed by the new On the Jews and Their Lies? Horse Eye's Back (talk) 15:47, 30 September 2025 (UTC)

the WMF should convene a constitutional convention to create an editorial charter and assembly - Dang I didn't realize it was 1787 still, maybe we should get Thomas Jefferson on the phone and save our crumbling encyclopedia apparently. Sophisticatedevening^(talk) 15:50, 30 September 2025 (UTC)

• It's also worth pointing out that people have, in fact, tried to produce online encyclopedias based on different definitions of truth and neutrality, including the ones Sanger advocates here - nationalist ones under the auspices of individual states, religious ones that enshrine the revealed truth of individual religions, etc. The reason why Wikipedia has won out (and the reason why people like Sanger target it in the first place, rather than just retreating to Conservapedia or wherever) is because our approach works - it has produced the most useful encyclopedia in the world. Wikipedia is freely available to everyone who follows its license; if Sanger's ideas had merit he could create a fork of Wikipedia at any time and govern it the way he wanted. Like, he says he wants competing versions of articles, but nothing actually stops him from creating a competing version of Wikipedia as a whole! Except that it would fail, of course, because nobody outside of his bubble would trust it or find it useful. The teeth-grinding thing to people like Sanger is that Wikipedia is useful to almost everyone, and is therefore broadly trusted. Our policies work. --Aquillion (talk) 16:46, 30 September 2025 (UTC)

  He did, Citizendium. Which still exists, although he left it years ago. A good chunk of these theses seems to be trying (again) to get Wikipedia to adopt the policies he tried to implement there. Anomie⚔ 17:36, 30 September 2025 (UTC)

For the interested Larry hasn't said anything interesting in decades and has been completely disconnected from the community ever since its inception. Save yourself the time and hassle, go read something more edifying, like a chicken noodle soup recipe. Headbomb {t · c · p · b} 17:06, 30 September 2025 (UTC)

An authoritative summary _signed, Rosguill ^talk 19:51, 30 September 2025 (UTC)

@Headbomb: Heh. You may already know this, but as it happens, back when Larry Sanger was editor-in-chief, Wikipedia actually used to have a cookbook with lots of soups, but alas, the only chicken noodle soup didn't get its recipe until August 2003. We got a chicken soup recipe in March 2002, but no noodles in that one. Also see Wikipedia:Cookbook. Graham87 (talk) 09:03, 1 October 2025 (UTC)

Actually I didn't, that's a few years before my time. Thanks for the trivia! Headbomb {t · c · p · b} 11:40, 1 October 2025 (UTC)

I wonder how much dark money this think-tank-like piece cost the foundation (and I don't mean the WM foundation!). -- GreenC 19:00, 30 September 2025 (UTC)

Well, might as well state my opinions on, this. I'm not gonna call it nine theses because that's pretentious.

Point 1: No lol. Consensus is vital to Wikipedia. The idea of a single committee regulating all articles issues would slow Wikipedia down due to beaurocracy. Could it make stuff more clear and very slightly more precise? Maybe. Will it drastically slow progress down to a halt. Absolutely it will. I have no doubt about that. The cons outweigh the pros.

Point 2: Absolutely not. This might be the weirdest idea here. We don't want people to go to Wikipedia and turn on a switch so they only see stuff that validates their preconceived notions on a topic. Our goal should be to remain objective and report on the facts. Also, the idea of the article's original author choosing who can write it might be the worst idea I've heard when it comes to Wikipedia (well, at least the worst one when it comes to opinions by major editors).

Point 3: Nope. For starters, it discusses sources like Fox News, which isn't even blacklisted. Pretty much all sources that are blacklisted are sources that have absolutely zero place on this Wikipedia, or any wikipedia. We report with facts, not some website made by a conspiracy theorist.

Point 4: Probably not. I'm gonna be honest, I do think the NPOV policy could do some work, but most of it is at least interpreted in a way that seems fair. Sure, it should be worded better, but I think it's important to make one thing clear. Wikipedia reports on the facts. The only side Wikipedia should be taking is the objective truth. Sometimes, however, that means that WE DO HAVE TO TAKE A SIDE. Sometimes, the facts might disagree with a side. We can and should take a stance towards misinformation and false claims, and label them as such. We should declare the objectively false as the objectively false.

Point 5: no. The ignore all rules rule is great. It allows progress here by making it clear that you can do something different, so long as it's good. That's, like, pretty sweet and wholesome. I do agree, that it probably is sometimes cited incorrectly, but all rules are.

Point 6: NO. Well, I do agree that it would be cool if Wikimedia provided help with those who are being harassed, so I like that idea. But that's the only good idea here. I mean, come on. Revealing the names of wikipedia's top brass is a horrible idea. Especially in this climate. Also, paying them? Excuse me, but I thought we wanted less corruption. part 5 of this point is reliant on the others. Point 6 is horrible. It'll just be people complaining that people are reporting on bad things they did. Also, it's disgusting for him to list these people, and tell them to reveal themselves. This is honestly horrible, and I think an admin should go ahead and delete that chunk from that point. It's doing no good, and is just gonna cause problems.

Point 7: It's already done. There are article ratings. It's not a normal voting system, but that makes it more immune to campaigns to rate articles low or high.

Point 8: Maybe. Honestly, I think that making it a last resort could make sense. Larry makes some good points. But some people need to be indefinitely blocked. But, honestly, this isn't a terrible idea. But only if done for just immature vandalism. Things like harassment, racism, stuff like that should be grounds for a perma-ban.

Point 9: Sure. It should have it's limits, but I often struggle figuring out what the policies for a certain thing are, and I could see this kinda being helpful.

.

One final note. Do not use AI in wikipedia. Just no. Any suggestions made by Larry involving AI is terrible. End of discussion.

.

These are just my thoughts from skimming through the essay. I obviously am not reading all of it, but overall, pretty bad ideas. Except, like 2 of them. Gaismagorm (talk) 23:28, 30 September 2025 (UTC)

On point 2, I think there was wiki-encyclopedia that had that as goal, it would look at your browser history or whatever and give you the version of, say, the Barack Obama article you were ideologically comfortable with. I don't remember the name, might have been Infogalactic. Also, we sort of have this sometimes, what you do is that you change to another language version. Gråbergs Gråa Sång (talk) 10:28, 1 October 2025 (UTC)

Knowing some of the weird stuff that went on places like Serbian Wikipedia, not to far off honestly. Gaismagorm (talk) 10:42, 1 October 2025 (UTC)

And stuff like "The Hebrew and Arabic versions of the "2014 Israel–Gaza conflict" page have, quite unsurprisingly, differing editorial angles. Even leaving language aside, the two pages give contrasting accounts in picture form." And that's a 10-year old article. Gråbergs Gråa Sång (talk) 10:55, 1 October 2025 (UTC)

Additional coverage of he.wiki's POV problems. _signed, Rosguill ^talk 15:25, 1 October 2025 (UTC)

IIRC it was the Croatian project, not the Serbian project, that was subject to extensive project capture. Azerbaijani Wikipedia has also had documented issues with coverage of the Armenian genocide. _signed, Rosguill ^talk 15:24, 1 October 2025 (UTC)

Croatian WP went very far, but there are other stuff that has been commented on in media, like Japanese WP on some WWII-topics, and how different languages write the history of the airplane [2] (for quite natural reasons). To quote one journalist, "Unlike many Wikipedias in languages with a global span, like English, Spanish or Arabic, Hebrew Wikipedia resembles its Polish or Hungarian counterparts in being more of an "Israeli Wikipedia." Gråbergs Gråa Sång (talk) 15:38, 1 October 2025 (UTC)

Monitoring our practices on this fundamental policy across the language editions is what we have to do in the future.  — 魔琴 (Zauber Violino) [ talk contribs ] 12:28, 2 October 2025 (UTC)

Electroimpact odd business

I know that this does not warrant a report at WP:ANI, but it is a little strange.

• Electroimpact (edit | talk | history | links | watch | logs)
• Pollluxo (talk · contribs · deleted contribs · logs · filter log · block user · block log)

In March 2025, an AFD was submitted for Electroimpact, and the article was deleted. On 29 September 2025, a new account pops up, and on 30 September 2025, the new account files a request at Deletion Review, saying that they have developed a new article in their sandbox, and asking to review the sandbox and overturn the deletion. They knew what Deletion Review is, and didn't know that DRV should not be used to request review of draft articles that replace previously deleted articles. An editor at DRV said that the article was not salted, and that the submitter should go ahead and create a new article subject to possible AFD. Another editor at DRV moved the sandbox to draft space and submitted it for AFC review, noting that this was probably what the submitter wanted. A third editor at DRV said that the draft appeared to have been written by artificial intelligence. The submitter then blanked the draft.

What was the purpose of this request? (I am not asking about any of the three DRV editors, all of whom acted reasonably.) Why did a new account pop up to request Deletion Review if they didn't want the sandbox reviewed? Was this an attempt to game the system? If so, what were they trying to game? Robert McClenon (talk) 17:57, 30 September 2025 (UTC)

It looks like the submitter was trying to get the deleted Electroimpact article restored without going through the proper new article process I think that was why the submitter created a new account probably to avoid the history of the original deletion, filing a Deletion Review request asking reviewers to look at their sandbox, why???? Essentially, they were trying to use Deletion Review to bypass AFC and get the article reinstated.

The blanking shows that, they realized their approach wouldn’t work or that the draft might be rejected. They tryna game DRV procedures surly. Thilio^{R O B O T🤖 talk} 18:48, 30 September 2025 (UTC)

That is what I sort of thought. Requests to DRV to look at new sources are common. That isn't how DRV is supposed to be used, because the draft should either be moved into article space subject to AFD or reviewed through AFC, but usually the requests seem to be in mistaken good faith. The blanking of the draft did not seem like good faith. This seems to have been an attempted end run, but an end run is something that I would rather see on television when I am watching American football. Deletion Review is working, in that it isn't allowing the system to be gamed. Robert McClenon (talk) 20:15, 30 September 2025 (UTC)

I don't think that's true. Once you have a finding at AFD of non-notability (as opposed to, e.g., a {{db-nocontent}} speedy deletion), you really aren't supposed to recreate an article on that subject. Wikipedia:Deletion review says "Deletion review may be used...if significant new information has come to light since a deletion that would justify recreating the deleted page", and "Look at these new sources" can constitute "significant new information".

Therefore, if an article was deleted at AFD, and (months or a few years later) you find new sources, you are allowed to go to DRV and ask someone to look at your new sources. If DRV agrees, the old article can be undeleted and you can expand it. There is no requirement to use either the Draft: space or AFC.

I don't think that we should look at blanking a draft as "not good faith". Thilio's probably correct that it's a sign of them accepting that the draft is unlikely to be accepted. That's not gaming; that's communication, and what it communicates is that you should see whether {{db-author}} applies. WhatamIdoing (talk) 01:27, 1 October 2025 (UTC)

I respectfully disagree with User:WhatamIdoing as to whether this was a case for Deletion Review and as to their interpretation of a deletion for non-notability. They are referring to WP:DRVPURPOSE3, which does say:

Deletion Review may be used … if significant new information has come to light since a deletion that would justify recreating the deleted page;

.

Yes. That provision has long been interpreted in different ways. However, it is clarified by not provision 10 below, which adds:

Deletion review should not be used … to ask for permission to write a new version of a page which was deleted, unless it has been protected against creation. In general you don't need anyone's permission to recreate a deleted page, and if your new version does not qualify for deletion then it will not be deleted.

I think that contradicts their statement:

Once you have a finding at AFD of non-notability (as opposed to, e.g., a {{db-nocontent}} speedy deletion), you really aren't supposed to recreate an article on that subject.

You are allowed to create an article on the subject, as long as it isn't the same article. DRV does get a lot of requests to review new sources, and tells the applicant that they can either submit a draft for review or create a new article in article space subject to AFD.

This disagreement illustrates that our guidelines on recreating articles after deletion need clarification. Reasonable editors interpret them differently, and that is a problem. I asked one question, and discovered another question that needs to be addressed. Robert McClenon (talk) 04:46, 1 October 2025 (UTC)

AFD normally deletes articles because the real world doesn't have sources. Therefore "as long as it isn't the same article" is irrelevant; AFD is essentially banning "the subject", not "the sentences and paragraphs most recently used to describe the subject".

So here are the options I can see. The first is creating the article:

• The article is deleted at AFD in September because it is non-notable/the real world doesn't have the right sources.
• Excellent new sources appear in October.
• I create a "new" article with those sources – and get accused of editing against the current/recent consensus.

The second is involving DRV:

• The article is deleted at AFD in September because it is non-notable/the real world doesn't have the right sources.
• Excellent new sources appear in October.
• I go to DRV because (a) it says to ask them about "significant new information" that came to light after deletion and (b) I want to make sure there's consensus – and they yell at me for wasting their time.

It looks like I can't win: If I don't ask for permission, I'm "editing against consensus", which is a blockable offense. IF I do ask for permission, I'm "wasting their time". WhatamIdoing (talk) 15:21, 1 October 2025 (UTC)

Yes, that is the question that needs resolution. Robert McClenon (talk) 17:45, 1 October 2025 (UTC)

The solution here, is to go with 1, and when "accused of editing against the current/recent consensus.", asssuming the article was deleted for notability reasons, to point to the new sources that weren't present in the article. Headbomb {t · c · p · b} 18:23, 1 October 2025 (UTC)

Two-Part Question

So I have a two-part question. First, is WhatamIdoing correct that Once you have a finding at AFD of non-notability (as opposed to, e.g., a {{db-nocontent}} speedy deletion), you really aren't supposed to recreate an article on that subject.? I have never heard that as a principle, although that doesn't mean it isn't correct. I see attempts to recreate improved versions of deleted articles all the time.

The second question is: What is the proper procedure if an article has been deleted by AFD, and an editor wants to recreate an article? Should the editor make a request at DRV to review the new sources, or should the editor simply create a new article in article space, subject to AFD? WhatamIdoing is concerned that the editor will be editing against consensus if they recreate the deleted article, because there was a consensus that the topic is not notable.

So maybe a third question is whether the AFD close of Delete means that the topic has been found not to be notable, or only that the article does not establish notability. Robert McClenon (talk) 03:24, 2 October 2025 (UTC)

@Robert McClenon According to my knowledge and understanding on how Articles for deletion works the first answer to the first question simply its yes and no because it depends on what exactly the AFD decided. Lemme demonstrate....

A Delete close at AFD usually means there was consensus that the article as it stood did not demonstrate notability not necessarily that the topic is intrinsically non notable some AFD discussions do conclude that the subject itself is not notable regardless of article quality also In those cases recreating the page without significant new evidence is considered against consensus.

So WhatamIdoing is partly right its not a blanket rule but recreating an AFD deleted article without substantial new sources or arguments can indeed be seen as ignoring consensus.

What is the proper procedure if an article has been deleted by AFD, and an editor wants to recreate an article? Should the editor make a request at DRV to review the new sources, or should the editor simply create a new article in article space, subject to AFD? according to my understanding like If you have new significant sources that were not available or considered in the original AFD you should (a) Write the draft in Draft space or your sandbox. (b) Then submit it through Articles for creation or (c) Request undeletion at REFUND or DRV if you think the sources could change consensus. If you just recreate it in article space surly another editor may speedily delete it as a G4 recreation of a deleted AfD article thats the rule against recreating substantially the same article deleted via Articles for deletion, unless circumstances have changed… So probably DRV is the correct venue if the new evidence changes the situation. so In that way consensus can be reassessed formally.

whether the AFD close of Delete means that the topic has been found not to be notable, or only that the article does not establish notability Simply means... (a) The article failed to establish notability... (b) It does not necessarily mean this subject can never be notable... And also (c) AFD decisions are about the article as it was presented though sometimes the closer will state that consensus was that the subject itself lacks notability. Thats why G4 requires that the recreation be substantially the same if you bring new strong sources its no longer the same article. Thilio^{R O B O T🤖 talk} 18:04, 2 October 2025 (UTC)

Maybe the top of DRV needs to be clarified, so that people won't do the right (but not necessarily required) thing and then be told by editors that they're doing the wrong thing. WhatamIdoing (talk) 19:40, 2 October 2025 (UTC)

"Grokipedia"

Elon Musk Plans to Take on Wikipedia With 'Grokipedia' (archived). Sapphaline (talk) 08:05, 1 October 2025 (UTC)

He plans to take on Wikipedia every year Gaismagorm (talk) 12:26, 1 October 2025 (UTC)

Yup. Note also the bit about 'Macrohard' at the end of the article. Musk regularly announces all sorts of things that never see the light of day, and sometimes it's hard to distinguish from trolling. If and when 'Grokipedia' ever appears, it might be worth taking notice of, though I can't see it being much of a threat to Wikipedia: different intended audience. AndyTheGrump (talk) 12:35, 1 October 2025 (UTC)

I'm sure it would do as well as any of the other POV forks of Wikipedia that've been attempted. -- LCU ActivelyDisinterested «@» °∆t° 15:39, 1 October 2025 (UTC)

Conservapedia but with AI, absolutely brilliant idea, nothing could go wrong here jolielover♥talk 17:52, 1 October 2025 (UTC)

Honestly, I'm kinda curious is conservapedia suffers from an AI problem. On one hand, a lot of AI has a slight left-lean in politics, but on the other hand conservapedia also seems like the place to have a ton of tech bros. Gaismagorm (talk) 17:57, 1 October 2025 (UTC)

Who wants to volunteer to help Musk make even more money? It would be worse than having banner adds on Wikipedia. It's creepy. Then there is the problem of Copyvio. Few sue the MWF over copyright violations because it's a non-profit, no money changed hands, and the responsible party is a single anonymous editor who is not worth suing. However Musk has 100 billion dollars and it's a for-profit entity, any AI-generated copyvio has measurable damages (page views), it will be a honey trap for every lawyer in the country. -- GreenC 18:26, 1 October 2025 (UTC)

It 100% won't be ad-free. So yeah, we are gonna be fine. Gaismagorm (talk) 18:30, 1 October 2025 (UTC)

I don't think Musk's plan includes volunteers. He's expecting Grok to produce articles unaided. And I don't think that copyvio is going to be the only legal issue if he does that, given the (now mathematically proven to be unavoidable [3]) pattern of hallucinated content LLM's generate. The thing is essentially guaranteed to generate libellous statements at some point. Maybe not often, but enough to make the whole thing problematic. If, that is, this isn't just more of Musk's trolling, and he is genuinely planning to release it. AndyTheGrump (talk) 18:36, 1 October 2025 (UTC)

Hey, I didn't know about the hallucination thing! Honestly, quite relieved to know that AI will still suck for quite a long time. Gaismagorm (talk) 18:40, 1 October 2025 (UTC)

It will be AI produced and human editable. How exactly that interacts is still uncertain. -- GreenC 19:05, 1 October 2025 (UTC)

Probably just:

Fork Wikipedia to fit a mildly specific narrative

Say how dope is Elon

Let people edit it

Call anything that's genuinely neutral "vandalism" and revert it

And now it can go in a various amount of ways, the most possible is:

NO ONE takes the Wikipedia fork seriously

Everyone just forgets it and it dies out Brickguy276 (talk) 19:11, 1 October 2025 (UTC)

Musk is a guy that doesn’t shy away from throwing money. He rarely if ever aborts projects. The project not being closed and instead continuing to exist for at least the next hundred years seems more plausible. Also maybe Grok will also do vandalism reverting and maybe we will geta fully automated wiki 95.5.189.54 (talk) 21:19, 1 October 2025 (UTC)

You are joking, right? https://elonmusk.today/ is a list of all the thing Musk has promised and not yet delivered. -- GreenC 04:49, 2 October 2025 (UTC)

You americnas are weird, though I looked at the list and it seems to be largely matters not related to money, and not all of them seem to be promises. I stand corrected 95.5.189.54 (talk) 07:51, 2 October 2025 (UTC)

Yeah. I wonder though, where do these LLMs get their info from? Oh that's right, they either get it from Wikipedia making them reliant on us or they get it from every website making them a complete mess. Gaismagorm (talk) 21:28, 1 October 2025 (UTC)

Besides, we already have a vandalism bot. We are fine. I also doubt many people would trust grokipedia, cause, you know, it's grok, an AI somehow worse than Gemini. Gaismagorm (talk) 21:33, 1 October 2025 (UTC)

Gemini is horrible, but Grok is worse - simple. However Gemini is still funnier. Brickguy276 (talk) 19:35, 2 October 2025 (UTC)

There is a possibility he is also trying to "flood the zone" with cheap garbage to dislodge Wikipedia's search and AI-training dominance. -- GreenC 18:35, 1 October 2025 (UTC)

"yo grok copy wikipedia but no leftist woke bullshit and stuff and say that im really dope"(Yes, this is kind of stolen from a dude on the Polish Discord.) Brickguy276 (talk) 18:57, 1 October 2025 (UTC)

Page count

Hello, I'd like to ask a few silly questions about Wikipedia's page count. Out of curiosity about Wikipedia I just want to know about these things:

1. Do disambiguation pages and/or lists count to the article count?
2. Do subpages(like Wikipedia:Village pump (miscellaneous)/Archive _ (an example linked)) count to the page count(not the article count, the page count)?
3. Where can I see the list of Wikipedia page count milestones(such as 7 million articles, 800 million pages etc)
4. This should be pretty obvious, but to be sure; do: user talk pages, user pages and user subpages count to the overrall page count?

Brickguy276 (talk) 12:24, 1 October 2025 (UTC)

Brickguy276, For #3, see the navbox at the foot of Wikipedia:Seven million articles. Mathglot (talk) 13:02, 1 October 2025 (UTC)

The article count is documented at mw:Manual:Article count; redirects are never counted, but the rest can be configured per-wiki. The English Wikipedia is configured to count every mainspace page that has at least one wikilink. In practice, mainspace pages without a wikilink are few and quickly fixed, so that last condition doesn't significantly change the count. For the overall page count, every page is counted, regardless of namespace and whether it is a subpage. jlwoodwa (talk) 13:19, 1 October 2025 (UTC)

@Brickguy276: You might be interested in the database report listing page count by namespace. Graham87 (talk) 13:43, 1 October 2025 (UTC)

@Brickguy276 For #1, the total includes disambiguation pages. There are 370,648 disambiguation pages plus a further 120,848 set index articles (so just under half a million total). Identifying a count of lists is a bit trickier, but I would guess somewhere in the range of 50-100,000 - it is muddied because there are a lot of things that we would generally consider "real articles" that are also categorised as lists in some way, because they contain a list element. Andrew Gray (talk) 19:50, 1 October 2025 (UTC)

What about lists of lists like Lists of deaths by year? Brickguy276 (talk) 05:44, 2 October 2025 (UTC)

They'll be counted in the estimate for lists. Again there's some difficulty identifying them clearly from the category, but maybe a thousand? Andrew Gray (talk) 08:34, 2 October 2025 (UTC)

@Brickguy276 and Andrew Gray: As for counting lists, I just realised that there's a table at Wikipedia:Content assessment#Statistics that's relevant. At the time of writing, it says that there are 383,190 articles assessed as lists, not including 2,460 featured lists, giving a total of 385,650 articles assessed as some kind of list. In addition to the caveats listed in the above message about counting lists, not all articles are assessed, as noted in the table. Furthermore, the table at Wikipedia:WikiProject Lists lists 213,096 pages, for what that's worth. All the figures noted here would include lists of lists. Graham87 (talk) 09:02, 2 October 2025 (UTC)

Very interesting! I didn't think to check the talkpage assessments. Will dig into that one tonight, as it's a surprisingly large discrepancy.

There are a handful of GA lists as well, I think, so presumably they would also need to be added to our total. Andrew Gray (talk) 11:29, 2 October 2025 (UTC)

Input needed: Lady Gaga Rio concert attendance figures

Hi everyone, there is an ongoing discussion at Talk:Lady Gaga regarding the reported attendance of Lady Gaga's free Rio concert at Copacabana Beach. Some reliable sources (CNN, Guardian, etc.) cite "2.1–2.2M" from City Hall, while others (Pollstar, Billboard, Variety, Rolling Stone, etc.) report "2.5M" based on organizers (Live Nation). We’re debating the best way to reflect this in the article. Additional input from uninvolved editors would be appreciated. CHr0m4tiko0 (talk) 13:47, 1 October 2025 (UTC)

utm_source=chatgpt.com

We have over 4000 links that match a search for insource: utm_source=chatgpt.com ([4])

Should we use an edit filter to block, or warn for, edits that do? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:45, 2 October 2025 (UTC)

We could. It would least eliminate the edits where the editor thought that they were doing the right thing by posting LLM-generated content, but the worst edits are those where the editor knows it is wrong but still persists in doing them. In that case (I don't think I'm revealing any beans here) the editor can simply change the URL. Phil Bridger (talk) 11:54, 2 October 2025 (UTC)

Sure. Sapphaline (talk) 12:55, 2 October 2025 (UTC)

We have filter 1346, but it only adds a tag. (I haven't gone through these search results yet because I'm more focused on finding the AI text that isn't already obvious.)

The problem is -- and I keep saying this because we are 2 years overdue for doing something about this -- we have no policy preventing the use of AI, so we have no policy-based grounds for blocking people solely for using it. Gnomingstuff (talk) 13:14, 2 October 2025 (UTC)

Hence "or warn". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:19, 2 October 2025 (UTC)

Warn for doing something currently within policy? Gnomingstuff (talk) 14:16, 2 October 2025 (UTC)

Warn that LLM-generated text may include hallucinations. HTH. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:47, 2 October 2025 (UTC)

I think we should. Many new editors seem genuinely unaware that LLMs hallucinate citations, and the edit filter could warn them about this. Helpful Cat {talk} 13:22, 2 October 2025 (UTC)

Much as I'd like to see LLM-generated stuff to be nuked at the door when it comes to WP, I believe many of these tags are merely leftovers in reference links that were found by using ChatGPT as a search engine. Which naturally still comes with a certain baggage, but the beast does perform more commendably as a search engine than as a prose generator, so it's probably not quite as much an indication of dark deeds as it may seem. --Elmidae (talk · contribs) 14:02, 2 October 2025 (UTC)

Agree that these tags are more useful than not having them. If there was an automatic "please remove the tag", or such tags led to automatic reversion and therefore actively brought about removal, then all that would mean is less ability to track the edits. CMD (talk) 14:36, 2 October 2025 (UTC)

I think edits with the ChatGPT tag should automatically become a proposed edit and need an external editor to review the edit. Mikeycdiamond (talk) 19:07, 2 October 2025 (UTC)

I'm not understanding the problem. I just asked ChatGPT for information about the House of Lords, and then followed up with a request for sources. Clicking on one of the supplied links got me to https://www.nationalarchives.gov.uk/help-with-your-research/research-guides/parliament/?utm_source=chatgpt.com. If I read that page and wrote in an article:

The British National Archives contains records created by the House of Lords.

it seems like I'm doing exactly what we want editors to do. I used a tool to research a topic and find reliable sources, and then I cited one of those sources in an article. The only thing I did "wrong" was fail to chop off the silly utm stuff in the URL, which honestly is something our automated referencing tools should just do automatically. Sure, we don't want people copy-pasting LLM output into articles, but why is utm_source=chatgpt.com in a reference URL evidence they did that? RoySmith (talk) 15:23, 2 October 2025 (UTC)

References