Wikipedia talk:Criteria for speedy deletion

G8 conflict?

Imagine that I create User talk:Nyttend/subpage. Assuming that there's no problem with the content I put on the page, it's a perfectly fine page; we wouldn't delete it on G8 merely because User:Nyttend/subpage doesn't exist. Now, imagine that I create User:Nyttend/subpage with problematic content, e.g. blatant spam. Someone comes around and deletes it. Should the talk page be deleted (because it's the talk page of a deleted page), or should it remain (because it's a valid subpage in userspace), or is this something we should just leave up to admin judgement?

Also, imagine that I create User:Nyttend/subpage with the content #REDIRECT ebwtriypnry0tiw5mr4te5, or I create the page with valid content and then replace it with the broken redirect. G8 (redirect to nonexistent page), or keep (established user's subpage), or admin judgement?

This just now came to mind as I was deleting spam in userspace; I deleted User talk:BassettHousePic/sandbox after deleting User:BassettHousePic/sandbox, and I'm not sure this is always right. Here it was (the user's been spamblocked and has no other edits), but that won't always be the case, especially if the userpage is deleted via U1. Nyttend (talk) 23:25, 31 May 2024 (UTC)

I usually end up deleting the user talk subpages, and mostly consider it a bug that the "G8: Talk page of a deleted page" option doesn't appear in the deletion dropdown and that there isn't a link to the talk page from the post-deletion screen. I guess the distinction is that User talk:Nyttend/Archive 42 is primarily a subpage of User talk:Nyttend; while User talk:Nyttend/spam sandbox, which typically consists of wikiproject templates from the article creation wizard, would primarily be the talk page of User:Nyttend/spam sandbox.

For a U1, or other ambiguous cases where the deletion would be unobjectionable to the page owner, I'd probably leave a note on their talk page or maybe in the deletion log to ask for deletion of the user talk subpage separately if desired. It hasn't come up. —Cryptic 00:29, 1 June 2024 (UTC)

I'd perhaps say user talk pages should be an exception (as opposed to only archives of talk pages) to G8. Crouch, Swale (talk) 06:01, 1 June 2024 (UTC)

Hmm, my impression has always been that "User talk pages" in the "not eligible for G8" list only applies to the main user talk page and equivalents, i.e User talk:Nyttend, but not automatically every talk page of a page in the user namespace. Jo-Jo Eumerus (talk) 07:52, 1 June 2024 (UTC)

That's always been my impression too. Obviously there will be some exceptions (e.g. a theoretical user talk:Thryduulf/Subpage/Archive 1) that require admin discretion. The easiest (but non-ideal) ways to resolve this are probably using the {{G8-exempt}} template proactively, and having a very low bar to restoration when asked. Thryduulf (talk) 11:23, 1 June 2024 (UTC)

imo it depends on whether the talk page was actually a talk page for the deleted page. If e.g. User talk:Billy Bob/archive1 is an archive of his talk page, creating random nonsense at User:Billy Bob/archive1 does not change that. jp×g🗯️ 19:48, 8 July 2024 (UTC)

Procedure for rejected draftified articles and declined/rejected AfC submissions improperly moved or copies to mainspace

So, this is something I've seen happen in some cases, usually involving inexperienced editors who have just gotten autoconfirmed. First there is the instance I've dealt with, where an article is draftified (when I've done it, it's usually for lacking references) and the author immediately recreates it by either moving it or, more often, by copy-paste. I haven't seen an exact consensus on what to do in this instance. Some have said db-copypaste might apply, but not all admins will delete in this instance. The other options would be re-draftify, PROD, or AfD.

The other instance would be declined or rejected AfC submissions that are moved or copied to mainspace without improvement. My understanding is that declined AfC pages that are resubmitted without improvement are usually summarily declined. If a submission is declined multiple times, that would indicate consensus that the page, in its present state, is unsuitable to be an article, putting it in a similar position to a page that meets G4. TornadoLGS (talk) 22:01, 1 June 2024 (UTC)

I think straight to AfD is usually the right choice. If it has already been moved to mainspace after a decline, there's strong reason to believe a prod would be contested. And as a contested deletion, it's not really a good candidate for a speedy deletion. That said, AfD is not mandatory. It happens that the draft reviewer can be wrong. Only take to AfD if you still believe it is not notable. —David Eppstein (talk) 22:30, 1 June 2024 (UTC)

That's unfortunate since AfD takes up a lot more editor time than CSD. Quick question on that. At which point does contesting a deletion count? Since I've seen plenty of authors go for the "contest deletion" option on G11, U5, and at least one A7 but it's ignored by the deleting admin. TornadoLGS (talk) 03:43, 2 June 2024 (UTC)

We are not here to delete as much stuff in as little editor time as we can, but to make proper decisions on whether something is or is not suitable to be an encyclopedia article. G11 and A7 are also possible speedies for some drafts moved to mainspace, but not valid for many of them. —David Eppstein (talk) 04:31, 2 June 2024 (UTC)

I guess I thought that a draft that has been declined or rejected (especially multiple times) might be sufficiently unsuitable for mainspace to be speedily deleted. I had been tentatively floating the idea of whether a new CSD criterion might be established for this instance, though it seems like that sort of criterion wouldn't work at this point. TornadoLGS (talk) 22:33, 2 June 2024 (UTC)

Rejected is just the opinion of one editor. On occasion I have moved such pages to mainspace, as clearly the rejection was not appropriate. A speedy delete in the situation you say is controversial, so AFD or other existing speedy delete (eg advertising or no claim of importance) is best. Graeme Bartlett (talk) 01:54, 7 June 2024 (UTC)

(edit conflict) Draftspace and AfC are optional. If someone moves a page from draft to article space that doesn't meet an existing speedy deletion criterion, but you believe should be deleted then PROD and AfD are your only options. AfC rejections represent the opinion of the reviewer(s), not consensus. Thryduulf (talk) 22:31, 1 June 2024 (UTC)

Just noting that if a copy/paste has been performed, unless the original editor is the only content editor a {{histmerge}} will be necessary. Primefac (talk) 01:02, 7 June 2024 (UTC)

• Use AFD, there shouldn't be a criteria for deleting such articles as it would discourage people from using AFD in the first place. Crouch, Swale (talk) 17:17, 7 June 2024 (UTC)
• AfC is not mandatory by any means. Any number of declines is not a consensus of any sort. AfD is the appropriate path in such situations described above; a speedy criterion based on drafts evaluated through a voluntary, non-binding process would be highly inappropriate. — Godsy ^(TALK_CONT) 07:41, 25 July 2024 (UTC)

A small update to U1

Currently, if I want to delete Module:Sandbox/Nickps or one of its subpages, I have to use G7, which means that if another editor edits it, I will have to go to MFD to get it deleted. However, those pages are user sandboxes (see Module:Module sandbox for the communal sandbox) and are only placed in the Module namespace for technical reasons, so U1 should apply instead. So, I propose that the first sentence of U1 is changed into Personal user pages, subpages as well as Module:Sandbox/<the user's name> and its subpages (but not their corresponding talk pages) upon request by their user. Nickps (talk) 23:01, 4 July 2024 (UTC)

Also consider updating U2 in a similar way. Nickps (talk) 23:03, 4 July 2024 (UTC)

Has this ever actually happened? * Pppery * _{it has begun...} 23:41, 4 July 2024 (UTC)

I have no way of knowing. I didn't find anything in TfD or MfD but there might be IAR deletions that admins have done over the years. That probably means rejection on the grounds of NEWCSD#3 anyway, but I still think my suggestion is correct, even if IAR ends up being the justification. Nickps (talk) 23:56, 4 July 2024 (UTC)

Revisiting a criterion for UPE creations

I'd like to revisit the idea of creating a new speedy deletion for criterion for articles created in violation of the Wikimedia Foundation Terms of Use's prohibition of undisclosed paid editing. This was previously proposed in 2017 but narrowly failed to achieve consensus. However, I think that many of the assumptions made in that discussion are no longer valid in 2024, due to the changing nature of undisclosed paid editing. Specifically, the closer wrote the objection to the proposed criterion was based on the following:

• the relative vagueness and/or subjectivity of its definition and/or applicability in practice due to frequency of edge cases – we have become much more organised about handling undisclosed paid editing cases since 2017. There is now an off-wiki reporting mechanism and a group of trusted functionaries actively monitoring it. Blocks for UPE based on nonpublic evidence are documented on-wiki using a process set by ArbCom, similar to checkuser blocks. Adding a clarifying bullet to the new criterion saying that it is only applicable if UPE is demonstrated by solid evidence on-wiki (rare in practice) or endorsed by a functionary should therefore be sufficient to overcome this objection. It wouldn't be any more vague or subjective than G5.
• other criteria already cover most cases (e.g., WP:A7, WP:G11) – UPE operations are smarter than they were seven years ago (or if they didn't get smarter, they couldn't survive). They don't write articles that without a claim of significance, they create the appearance of notability by WP:REFBOMBing citations to articles they have paid to be placed in superficially reliable sources. They don't write blatant advertisements, they just only write positive things and neglect to include anything their client doesn't want mentioned. I cannot remember the last time I came across a UPE creation that was so clear-cut as to meet these criteria.
• the normal, slower deletion processes are sufficient given a purportedly low incidence of pages that would meet the proposed criteria – the functionaries receive reports of undisclosed paid editing almost daily, and it is not unusual for them to reveal dozens of paid creations. Sending them en masse to AfD would flood it, and discussion there is in any case complicated by the nonpublic nature of the grounds for deletion.

Another approach suggested in the past was to expand G5 to cover these. And it is true that the vast majority of the regular UPErs we see were blocked long ago. But I don't think this is a good approach for two reasons. First, we frequently can't link them to a specific blocked account. Second, the more successful UPErs have moved away from sockfarms and instead hire one-off freelancers to post articles for them. This technically doesn't preclude G5 but does make it a lot less clear cut and in practice I think it has become less and less useful for UPE.

I'm not proposing this criterion right now. I'd like to hear whether others think it is viable and workshop the wording a bit first (so please hold the support/opposes). – Joe (talk) 08:10, 5 July 2024 (UTC)

Get UPE listed as a WP:DEL#REASON first.

Show that these lead to SNOW deletions second. SmokeyJoe (talk) 09:13, 5 July 2024 (UTC)

Review Wikipedia talk:Deletion policy/Archive 48#Undeclared Paid Editor (UPE) product. SmokeyJoe (talk) 09:16, 5 July 2024 (UTC)

Thanks for the comments SmokeyJoe. Just as an aside, you might want to consider how structuring them as a series of terse commands comes across. The first WP:DEL#REASON is "content that meets at least one of the criteria for speedy deletion", so in a way trying to get UPE listed there is exactly what I'm doing now :). But I understand that your broader point is that there should be an existing consensus that a CSD is a deletion reason. I believe that already exists for UPE creations. First and foremost, they are explicitly forbidden by the WMF Terms of Use, and that pre-empts local policy. We can also look at practice: in my experience, UPE creations only survive the unmasking of their creator if they have been significantly rewritten by other, uninvolved editors. Otherwise, they are deleted under G5, G11, at AfD, or G13 after being moved to draftspace. These are all long-standing, accepted practices but, as I explained above, they are inconvenient, and a more specific criterion would make combatting paid-for spam much easier. A final point of evidence is the discussion I linked above, where there was nearly consensus for precisely this criterion, and little doubt that it was a valid reason for deletion, just a series of (well-placed) concerns that I believe are now either no longer relevant or can be overcome.

Thanks for pointing out the 2019 DELPOL discussion – I didn't know about it, but I'm familiar with the arguments there. I think TonyBallioni's view, that a combination of WP:SOCK/CheckUser and WP:NOTPROMO/G11 is sufficient to deal with UPE, has long been influential and reflects his firsthand experience in countering the first wave of UPE outfits in the late 2010s. However, I think I remember even Tony himself saying that things it's no longer the case. The big UPE sockfarms were all blocked, they adopted new tactics, and we need to adapt too. Ivanvector and Thryduulf's argument, that we should only delete content for what it is and not where it came from, is more a point of principle, but ultimately I'm more interested in how the deletion process works in practice, i.e. does it do enough to support overstretched volunteer editors in dealing with people who try to spam our encyclopaedia for a living. As the late great DGG put it in that discussion: In practice, over my 12 years here, I have seen almost no satisfactory articles from people who are writing an article for money. – Joe (talk) 10:46, 5 July 2024 (UTC)

Smokeyjoe is right. Before we can even consider making this a speedy deletion criterion you need to show that every time a page is nominated for deletion solely for being a UPE creation by someone G5 does not apply to the consensus is to delete. Unless and until you can do that then any criterion will fail WP:NEWCSD point 2. Thryduulf (talk) 11:03, 5 July 2024 (UTC)

Terse? Sorry if that disturbed you.

I agree with all of your intent, except the attempt to creatively wikilawyer a solution in under CSD. This style of management further breaks the community into admins and nonadmins, where admins only deal with UPE. I also disagree with the premise that AfD can’t handle UPE-based, or even -mentioning nominations. XfD should most definitely be used to provide evidence of a need for a CSD. That’s what we did for U5 and G13, and every step of the process worked well and the outcome remains good.

The T&Cs are at odds with policy as written. T&Cs are often ignore rambling gobbledygook. I don’t see where the T&Cs authorise speedy deletion. I remain frustrated that agreement can’t be found to have deletion policy even mention undeclared paid editing. SmokeyJoe (talk) 22:55, 5 July 2024 (UTC)

I agree with you Joe. They are much more sophisticated and some have also started including a minor negative incident to hide UPE and to prevent the article from being tagged as UPE/NPOV type issues. For folks who are not familiar with UPE tactics, it goes well beyond just creating articles. They are also hired to participate in AfDs (different people than the article's creator) who bombard it with Keep votes and various poor sources which makes it extremely time consuming because you have go through all the sources presented and explain/argue why they do not meet the criteria. Only so many editors have the time or are going to take the time and often these are not brand new accounts; they have hundreds, if not thousands of edits, and perms so not SPAs/obvious to those unfamiliar. WP:Articles for deletion/Justin Jin (entrepreneur) is an example. In that instance they were blocked as socks but not until after the AfD and even so the article was not G5 eligible because they were not proven to be socks of an already blocked editor. Had it not been for the efforts of @Jeraxmoira and @Usedtobecool the article may have been been kept at AfD.

They are also hired to remove maintenance tags and of course update articles. There are articles where almost the entire editing history is blocked UPE but not G5 or G11 eligible. Sure, you can nominate it for AfD but again, you might have to deal with UPE participants so you need to be prepared to dedicate time and cross your fingers at least another non-UPE editor participates. S0091 (talk) 16:34, 7 July 2024 (UTC)

Joe is completely correct, and our failure to do basic stuff like make UPE creations automatically CSD-eligible is one of the main reasons there are UPEs -- our policies not only permit it but openly encourage it. Why should they wait eight months fo a draft to be declined when they can spam their slop shit into mainspace and flip a coin on having it stay forever? jp×g🗯️ 19:46, 8 July 2024 (UTC)

My experience with works accused of being UPE is that very often these accusations are made with little or no visible evidence. Sometimes even when the accuser is directly asked they will refuse to provide any evidence on the basis of not spilling beans. If an editor is actually determined to be a UPE-creator and blocked for it we can apply G5. If it's blatantly promotional we can apply G11. If neither, we need a clear and transparent process, not a hidden and unanswerable tribunal of UPE-inquisitors. —David Eppstein (talk) 23:04, 5 July 2024 (UTC)

If I understand Joe's proposal, the account(s) must be blocked for UPE in order for the article(s) to be eligible for CSD and I agree that should be requirement. S0091 (talk) 16:55, 7 July 2024 (UTC)

In which case it's redundant to G5. Thryduulf (talk) 17:47, 7 July 2024 (UTC)

It wouldn't be redundant - a person can be editing for pay without disclosure without also evading blocks. UPE accounts are often also sockpuppets but the more successful UPErs have moved away from sockfarms and instead hire one-off freelancers to post articles for them and the freelancers were never themselves blocked. * Pppery * _{it has begun...} 17:54, 7 July 2024 (UTC)

Or they do end up being blocked for UPE but either they are not sock of an already blocked user or likely a but the master is unknown or it's not until later it is discovered they are the master but masters aren't G5 eligible. Happens all the time. S0091 (talk) 18:13, 7 July 2024 (UTC)

In which case we're back to needing to get consensus at XfD that these creations should always be deleted. Only when that consensus exists can we consider speedily deleting them. Thryduulf (talk) 18:45, 7 July 2024 (UTC)

Did you catch my other response above which addresses AfDs? Editors are hired to participate at AfDs as well. S0091 (talk) 18:55, 7 July 2024 (UTC)

So? If you want to speedily delete a class of pages it is an absolute and non-negotiable requirement that you get community consensus that that class of pages should always be deleted. If you think that AfDs are being tainted by paid editors then you need to get consensus that their comments can and should be excluded from determining the result. Thryduulf (talk) 21:09, 7 July 2024 (UTC)

I believe all the comments on Justin Jin were given equal weightage apart from the SPAs/IPs - How would you get consensus that an editor's comments should be excluded from determining the result when they are not blocked yet? Am I misunderstanding your point? Jeraxmoira🐉 (talk) 07:58, 8 July 2024 (UTC)

To be clear, we are talking about undisclosed paid editing which per the Terms of Use is already prohibited so we do not need consensus their comments should be excluded. Of course, it has to be proven which can be a time-consuming endeavor and often not possible during the course of an AfD because you have prove several accounts are UPE, not just the creator. As it stands now even if the creator is proven to be UPE, G5, G11, etc. are often not applicable. I will also note in the Justin Jin case, a UPE editor also closed it as no consensus. Jeraxmoira had to go to DRV to have it reversed so not only did the community have to spend two weeks in a robust AfD discussion with several UPEs, the community also had to spend a week in a DRV discussion all because of UPE. S0091 (talk) 17:19, 8 July 2024 (UTC)

I don't agree that showing unanimous consensus at XfD is the only way to demonstrate that a proposed new criterion meets point #2 of WP:NEWCSD, and scanning through the archives of previous successfully-proposed criteria, this isn't routinely asked of them. That almost all pages that could be deleted using the rule, should be deleted could also be demonstrated by consensus in an RfC, for example. – Joe (talk) 08:18, 8 July 2024 (UTC)

User:Joe Roe, ease back on “unanimous”. Can you point to several AfDs that show a pattern? User:S0091 pointed to Wikipedia:Articles for deletion/Justin Jin (entrepreneur). SmokeyJoe (talk) 22:50, 9 July 2024 (UTC)

I haven't looked. I can, but in my experience these aren't handled at AfD too often. – Joe (talk) 12:00, 10 July 2024 (UTC)

So where is your evidence of consensus that these pages should be deleted? Thryduulf (talk) 13:24, 10 July 2024 (UTC)

I believe I've already answered that question. – Joe (talk) 13:50, 10 July 2024 (UTC)

You wrote a lot of words without actually answering the question. Where is the consensus that all pages (or all promotional pages) created by undisclosed paid editors that do not have any other problems should always be deleted? Thryduulf (talk) 13:57, 10 July 2024 (UTC)

Ultimately the test will be whether there's a consensus to establish this CSD or not. – Joe (talk) 14:06, 10 July 2024 (UTC)

I mean, like it or not, we do handle UPE through a hidden and unanswerable tribunal of inquisitors. If there's a better way of doing it within the very strict limitations we have on discussing off-wiki evidence on-wiki, I'd be an enthusiastic supporter. But as far as I know nobody has been able to come up with one, and until we do CSD has to work with the processes we have, not the processes we'd like to have. As I've mentioned above, while a combination of G5 and G11 used to work quite well, UPEers have countered this by avoiding blatant promotionalism and subcontracting out creation instead of using block-evading sockfarms. – Joe (talk) 08:24, 8 July 2024 (UTC)

Document this secret unanswerable tribunal. It would have to be covered by a different policy, not creatively squeezed under G5. SmokeyJoe (talk) 22:53, 9 July 2024 (UTC)

Note also that arbcom and the checkusers are neither secret nor unanswerable, and yet are entrusted with private evidence. So there can be no need for any UPE investigators to keep themselves secret and unanswerable. —David Eppstein (talk) 00:03, 10 July 2024 (UTC)

Indeed, two people (Bilby and Extraordinary Writ) are currently publicly applying to become entrusted to see and deal with non-public evidence regarding conflict of interest editing (with which paid editing is an overlapping set), See Wikipedia:Arbitration Committee/Conflict of interest reports/July 2024 appointments - it needs more attention. Thryduulf (talk) 00:21, 10 July 2024 (UTC)

Up until recently it was CheckUsers that handled UPE reports. Now it's CheckUsers, Oversight, and the special appointments (the first two linked above). So this is already documented, at Wikipedia:Functionaries and now Wikipedia:Arbitration Committee/Conflict of interest reports.

But functionaries aren't given special powers when it comes to deletion. That's why CheckUsers need G5, and why we need a new criterion for this. – Joe (talk) 11:55, 10 July 2024 (UTC)

Is there any evidence that those actually handling UPE reports desire additional deletion options? I've never seen it mentioned on the Functionaries list. Thryduulf (talk) 12:01, 10 July 2024 (UTC)

Well, I'm one of them... so that's one data point. @Moneytrees, Spicy, Blablubbs, GeneralNotability, Bilby, and Extraordinary Writ: Do you want to weigh in here? – Joe (talk) 13:42, 10 July 2024 (UTC)

And @DatGuy:, plus apologies to others that I've surely overlooked. – Joe (talk) 13:52, 10 July 2024 (UTC)

I handled a paid- ticket a while ago that had accounts in the same country but with different markers. After further investigation, I concluded one known blocked UPE was paying other people to perform the edits the UPE requests. Ultimately, I still consider this sockpuppetry—or meatpuppetry which fits the Wikipedia:Requests for arbitration/Regarding Ted Kennedy#Sockpuppets bill—and whatever articles they created would be eligible for G5 deletion. Therefore, whether to implement a carte blanche UPE criterion (which I haven't decided on whether I support/oppose yet) shouldn't be because of undisclosed paid editors evading G5, but because of a belief that UPE is inherently not worth fixing. DatGuy^Talk_Contribs 12:06, 14 July 2024 (UTC)

My general position on this is that deletion of UPE work is necessary if we actually want to meaningfully disincentivise this form of abuse. accounts are cheap and plentiful – simply blocking them does nothing to threaten the viability of UPE as a business model. Deleting the work product, however, does, because that's the part people get paid for.

I also agree that in practice, G5 and G11 often don't "cut it" unless construed quite broadly, because the proliferation of technical obfuscation measures deliberately employed to evade CU complicate long-term tracking of farms, and many spammers have figured out to write unbalanced promotional garbage while falling just shy of promotion that is sufficiently overt for G11. A criterion that would allow us to delete stuff that was verifiably bought and paid for would fix that problem. --Blablubbs (talk) 14:38, 10 July 2024 (UTC)

@Thryduulf, @Blablubbs Is bang on in terms of G5 being easily worked around; there are tons of SPIs that likely stem from the same master/firm, but aren't eligiable for G5 because the socks haven't been technically connected together. Moneytrees🏝️(Talk) 22:27, 10 July 2024 (UTC)

Before talking expansion of CSD, they should send cases to AfD to verify that they are aligned with the community. Is this a community run project, or oligarchy? SmokeyJoe (talk) 13:12, 10 July 2024 (UTC)

One requirement should be they are made of aware of WP:PAID/TOU. There are editors who meet the definition of PAID but legitimately do not know about the disclosure requirement but once made aware comply or at least make an AGF effort to comply. S0091 (talk) 18:25, 9 July 2024 (UTC)

I don't think this is written down anywhere, but the usual practice is only to block people if they've previously been made aware or are experienced enough that they can't plausible claim ignorance. When it's just a newbie that didn't know the rules, we just tell them about them, and usually that solves the problem. So if we link this criterion to blocks (which seems necessary), that wouldn't come up. – Joe (talk) 12:00, 10 July 2024 (UTC)

If this is linked to blocks then either your duplicating G5 or deleting pages created before the block, which is what you claim above you don't want to do. Which is it? Thryduulf (talk) 12:02, 10 July 2024 (UTC)

It's to be used to delete pages created before the block. I don't believe I've said otherwise. – Joe (talk) 13:33, 10 July 2024 (UTC)

Unless we're talking about where UPE is discovered and determined by other means, I don't see how this can go anywhere. Sometimes in NPP I see a case where after a thorough review, I'd say I'm 95% sure it's UPE. Wikipedia has no place to go with this information (e.g. for further review) ....the folks who ostensibly we're to report these to do not consider it to be their purview which is only to review submitted off-wiki evidence. And 95% sure is not enough to anything with. North8000 (talk) 18:51, 9 July 2024 (UTC)

I agree that any article created by a proven UPE should be deletable, but very often that accusation is made with no evidence. I remember that some years ago, when I was editing as an IP, the accusation was made against me that I was either a sockpuppet or a UPE (I still don't know which, but was neither) and the accuser (who is still editing) was asked to substantiate the allegations and replied that he had private evidence that he wasn't willing to disclose, even though we have procedures for disclosing such evidence to Arbcom. Yes, any (or at least most) methods should be used to get rid of the scourge of UPEs, but we need to avoid witchhunts. Phil Bridger (talk) 19:53, 9 July 2024 (UTC)

If it only applies to articles, then why can’t it be solved by draftification? Draftification would have the advantage of nonadmins being able to see what’s going on. SmokeyJoe (talk) 13:14, 10 July 2024 (UTC)

That's how it is handled in practice now. But I'm uncomfortable with it because it violates WP:NOTBACKDOOR. These drafts aren't going to be improved; the creator is blocked and nobody else wants to earn their payday for them. Keeping unwanted, ToS-violating content in draftspace for six months before G11 kicks in is a waste of time, and risks encouraging gaming by UPErs, who have been known to try and exploit their client's ignorance of how Wikipedia works by passing off drafts or old versions of articles as live ones. – Joe (talk) 13:24, 10 July 2024 (UTC)

NOTBACKDOOR. That’s WP:Deletion policy. That goes back to my earlier point. Get UPE included as a reason for deletion. Carve it out of NOBACKDOOR. Get ToS violations written into deletion policy. Creative solutions in CSD is the wrong approach. CSD is only for when consensus to delete is established to be obvious, objection, unobjectionable. The reason not even being listed at Deletion policy is a pretty obvious reason to object. SmokeyJoe (talk) 13:35, 10 July 2024 (UTC)

I thought draftspace was created so that AfC drafts could be moved out of Wikipedia space, because PAID editors were getting paid because clients thought Wikipedia space was good enough, and we thought the Draft prefix would not satisfy them. SmokeyJoe (talk) 13:37, 10 July 2024 (UTC)

And you and Thryduulf have made that same point numerous times now; I respectfully disagree. If you are right, then an RfC on this proposed criterion will not gain consensus and you have nothing to worry about. In the mean time, I would really like to focus on workshopping the actual proposal text. – Joe (talk) 13:47, 10 July 2024 (UTC)

You can choose to waste the community's time on an RFC that is doomed to failure if you wish, alternatively you can start listening to the feedback you've solicited and formulate a proposal that meets all the requirements of WP:NEWCSD and thus stands a chance of succeeding. Thryduulf (talk) 14:00, 10 July 2024 (UTC)

You will have noticed that you and SmokeyJoe are not the only people who have responded above. – Joe (talk) 14:04, 10 July 2024 (UTC)

But none of them have actually refuted or contradicted any point we've made. Thryduulf (talk) 14:10, 10 July 2024 (UTC)

What I understand from your and SmokeyJoe's feedback is that you don't think this is a viable proposal and won't be unless we spend a considerable amount of time (years?) taking test cases to AfD. Since I can't really act on that feedback right now, I'm choosing to focus on the feedback from others who were more positive about the proposal's chances. I hope you can see that that's not the same as ignoring your feedback altogether. Since we're at the point of workshopping a proposal rather than trying to get consensus for it, I don't think there is any need to engage in debate. – Joe (talk) 14:21, 10 July 2024 (UTC)

SmokeyJoe's feedback is that you don't think this is a viable proposal and won't be unless we spend a considerable amount of time (years?) taking test cases to AfD.

No, send dozen cases in a week. Seven days. Maybe longer with relists. Not a forever goose chase. We can trust the functionaries word that they have compelling evident the authors are UPE. Let’s test the DGG conjecture, no UPE ever writes a good article.

Undelete or re-mainspace and list at AfD some recent past cases if you don’t have a dozen right now. SmokeyJoe (talk) 23:17, 10 July 2024 (UTC)

If it only applies to articles, then why can’t it be solved by draftification? …

That's how it is handled in practice now. … – Joe (talk) 13:24, 10 July 2024 (UTC)

Can you link to some of these draftified pages please? SmokeyJoe (talk) 23:22, 10 July 2024 (UTC)

[1] and search for "commissioned" or "UPE". – Joe (talk) 08:43, 26 July 2024 (UTC)

Thanks User:Joe Roe. I see you draftified a few articles written by User:Didgeridoo2022, who you blocked on the basis of a ticket.

If someone disagrees with you, they have six months for easy reversal of the draftification and presumably we would then see it debated at AfD. Digeridoo2022 could do that if only they respond reasonably to your block notice.

Why is this method not good enough for dealing with UPE product? SmokeyJoe (talk) 05:47, 28 July 2024 (UTC)

I feel like I've already answered this in this discussion, but in brief: a) the purpose and only policy-supported use of draftification is to 'incubate' articles for improvement, but I don't really expect these articles to be improved because the creator is indefinitely blocked, blocked UPErs are rarely unblocked, and few volunteer editors are interested in helping them get paid by picking it up; b) the ToE forbids undisclosed paid editing, it doesn't say "oh go ahead as long as you use draftspace"; c) I have dealt with many cases sent to the WP:COIVRT queue where UPErs have tried to argue to their ripped-off clients that the continued existence of the draft counts as fulfilling their contract. – Joe (talk) 06:56, 2 August 2024 (UTC)

Well, let me know if I understand this correctly: the official paid-editing policy is that you're allowed to do it and we have to go through a seven-day-long process biased towards keeping the page so long as you include the phrase "is notable for" somewhere in the pile of slop? jp×g🗯️ 17:20, 10 July 2024 (UTC)

If you follow the rules on disclosure then you are allowed to edit for pay, yes. If you want to change that then you're in the wrong place. If you think AfDs should be biased towards deletion in some circumstances then again you're in the wrong place. If you have a proposal that meets all the requirements of WP:NEWCSD then you are in the right place, however all I've seen so far is complaining that it's too difficult to get the evidence required for point 2, rather than any attempts to actually get that evidence. If the community is as against paid editing as you believe then it will be very easy. Thryduulf (talk) 17:57, 10 July 2024 (UTC)

Sorry, I meant "the official undisclosed paid-editing policy" -- considering "we block the throwaway account you took thirty seconds to create but we keep the article you got paid $2,000 to write" to be "you're allowed to do it" (e.g. their goal is to have the article kept, they aren't getting paid $2,000 to have an unblocked sock) jp×g🗯️ 18:39, 10 July 2024 (UTC)

Standard of evidence

Based on the above discussion, the main concern (aside from those who don't want this criterion at all), appears to be how we will ensure that it is only a. I also think I assumed too much prior knowledge of how UPE enforcement and handling of nonpublic evidence currently works, which will also. To address this, I suggest making deletion explicitly conditional on a prior block for UPE, and inserting some more links for context on how these happen. So this would be the proposed text:

AXX. Articles created in violation of the Wikimedia Foundation Terms of Use

This applies to promotional articles created by a user who is indefinitely blocked for violating the Wikimedia Foundation Terms of Use's prohibition of undisclosed paid editing

• This criterion only applies to undisclosed paid editing, not paid editing that has been disclosed or any other types of conflict of interest
• Mere suspicion of undisclosed paid editing is not enough; there must be a consensus documented on-wiki or nonpublic evidence reported through the proper channels (see below), and the user must have been blocked for it
• If the block is based on nonpublic evidence, it must have been placed or endorsed by a functionary or administrator authorised to handle nonpublic evidence for this criterion to apply
• Unlike WP:CSD#G5, this criterion applies to articles created before the block was placed

The insertion of "promotional" is so we don't end up deleting non-paid creations alongside paid ones. I'm not sure if this also needs an explanatory bullet point.

Feedback and suggested modifications welcome, but I'm hopeful we're zeroing in on something that can be put to an RfC soon. – Joe (talk) 12:44, 10 July 2024 (UTC)

• If it applies only to promotional articles then it's entirely redundant to G11. Thryduulf (talk) 12:46, 10 July 2024 (UTC)

  G11 applies to "unambiguous" promotion and "pages that are exclusively promotional and would need to be fundamentally rewritten to serve as encyclopedia articles". The intention is that this would be broader. That's what I was wondering if we might need an extra bullet to explain. – Joe (talk) 12:58, 10 July 2024 (UTC)

  Where is the consensus that such pages should be deleted rather than fixed? Thryduulf (talk) 13:24, 10 July 2024 (UTC)

  I mean, do you want to spend your time fixing something like this, so somebody else can get paid for it? – Joe (talk) 13:31, 10 July 2024 (UTC)

  What either of us wants to do is irrelevant, what matters is that there is a demonstrated community consensus that the pages should always be deleted. If you don't have that then a speedy deletion criterion is impossible. Thryduulf (talk) 13:34, 10 July 2024 (UTC)

Regarding "there must be a consensus documented on-wiki", what qualifies as consensus? For example if I have a discussion with you on-wiki, provide evidence of UPE and you agree so you block the user, does that qualify as consensus? As for "or nonpublic evidence reported through the proper channels", does nonpublic include WP:BEANS where all the evidence is on-wiki but you don't want to give UPEs any ideas about how to circumvent detection or game Wikipedia's processes? Also, reading the 2017 RfC, there was growing consensus for a sticky prod. What is you reasoning for proposing CSD over the sticky prod idea? S0091 (talk) 16:02, 10 July 2024 (UTC)

I'm not convinced that "conditional on a prior block" deletions address the issue, because UPE creators can also create throwaway socks and it could plausibly be much easier to determine that an article is UPE (because it matches the particulars of some paid-editing request) than that it was created by some known sockmaster. —David Eppstein (talk) 19:06, 10 July 2024 (UTC)

I can see pros and cons to both routes. It certainly would be cleaner to focus on articles rather than editors—after all that is what CSD is about—but right now the UPE enforcement system is geared around editors. If I block someone based on off-wiki evidence of UPE, I wouldn't usually list the specific articles that that block was based on, and it is often not possible to do so because it would out the blocked editor. So an article-based criterion might end up being only usable by admins with access to nonpublic information. Maybe we could present both options in an RfC? Or try one first, then the other? – Joe (talk) 08:40, 26 July 2024 (UTC)

I don't think this should be retrospective as good faith editors have sometimes spent a lot of time rewriting and improving a notable UPE article unaware that it would be deleted at a future date. Also the evidence used on UPE blocks is varied. In my experience of UPE investigations some have rock solid evidence while others are only guesswork - for example one suspected UPE denied being so but wouldn't elaborate on further questioning so was blocked despite no real evidence except suspiscions, imv Atlantic306 (talk) 20:32, 12 July 2024 (UTC)

There's no point having it at all if it's not retrospective: after an editor is blocked for UPE, they can't create articles, and if they evade the block to do so then all their creations are eligible for G5. But like G5, we could and probably should add a clause that says only articles that haven't been substantially edited by others are eligible. If admins are making bad UPE blocks, I think that's an issue for ArbCom rather than WT:CSD... – Joe (talk) 08:33, 26 July 2024 (UTC)

Oh and to your question of why not make a new type of PROD: I consider PROD basically a failed experiment at this point. It fell between the cracks of discuss-at-AfD and quietly-remove-things-to-draftspace and so is barely used any more. – Joe (talk) 12:28, 26 July 2024 (UTC)

@S0091: I'd leave it open; the fulcrum point is the block. If there is a level of consensus sufficient for an admin to place a UPE block and not have it immediately reversed, then this CSD criterion would become an option. As above, if we have admins blocking people based on insufficient evidence or consensus, than we have bigger problems than losing some articles. – Joe (talk) 08:35, 26 July 2024 (UTC)

• Not seeing this meeting the frequent and objective criteria, sorry. Stifle (talk) 07:43, 25 July 2024 (UTC)

other solutions to throwaway accounts

There’s a big premise going on that throwaway accounts can’t be limited.

I support account confirmation requiring verification via a mobile telephone number. Flag accounts to functionaries, where many (>5?) accounts connect to the same number. Require confirmation to participate at AfD. This method of discouraging multiple account works great in Chinese social media. Use a side process for the rare new editor who doesn’t have access to phone, their own, a family member’s, or a friend’s. —SmokeyJoe (talk) 23:31, 10 July 2024 (UTC)

I think this would be bad. jp×g🗯️ 17:17, 25 July 2024 (UTC)

What’s one speculative bad case? SmokeyJoe (talk) 21:50, 25 July 2024 (UTC)

Uh, we can start with the Union of Soviet Socialist Republics and go from there, I guess. jp×g🗯️ 22:32, 25 July 2024 (UTC)

I think this goes against the spirit of "anyone can edit" to an unacceptable degree. Tazerdadog (talk) 19:42, 25 July 2024 (UTC)

“Anyone can edit” is unimpinged. IPs can edit. Non confirmed accounts can edit. Do you know any serious editor who can’t access a mobile phone message, or any editor who can’t do without >5 accounts per mobile phone number accessible? SmokeyJoe (talk) 21:49, 25 July 2024 (UTC)

I know plenty of well-off upper-class web developers who have uninterrupted mobile phone plans with consistent phone numbers and SMS service who keep said phones in good repair, updated, plugged in, turned on, and physically on their person 24/7, yes. jp×g🗯️ 21:57, 25 July 2024 (UTC)

Do these overlap with problem undisclosed PAID editors? Is there any other solution besides giving the small group of functionaries effectively unfettered permission to delete on suspicion? SmokeyJoe (talk) 02:55, 26 July 2024 (UTC)

Chinese social media is just great, huh. Do you really want your phone numbers stored against your username in some database available to some people in San Francisco, some based in other jurisdictions, many potential hackers, and any litigant who can rustle up a convincing subpoena? I certainly don't. -- zzuuzz ^(talk) 22:26, 25 July 2024 (UTC)

Replace "phone number" with "email address" and you describe the current situation with email addresses. But I guess those are optional. * Pppery * _{it has begun...} 22:43, 25 July 2024 (UTC)

I would probably trust WMF to store an encrypted version of my phone number, so that it could be used to link accounts verified with it. SmokeyJoe (talk) 02:58, 26 July 2024 (UTC)

Congratulations, you have just changed "anyone can create an account" to "anyone can create an account as long as you have a working mobile phone, with sufficient signal and credit, that is not shared between multiple people, has a persistent number, that can receive probably-international phone calls/texts and/or access the internet*, and both you and the authorities in your location don't mind a probably-foreign organisation having access to the number and trust them to keep it safe". While that might slow down some sockpuppeteers it would exclude a very significantly greater number of good-faith editors, disproportionately from economically disadvantaged parts of the world. AIUI the current MO of the big groups engaged in those engaged in bad-faith UPE (and remember that is a subset of all UPE) is to farm the editing out to individual freelancers working from home rather than centralised content farms and so your solution would be of limited impact anyway.

*If you don't require verification what you are actually asking for is a random number that matches the regex for mobile phone numbers in at least one country in the world (and AIUI mobile phone numbers are not differentiated from landline numbers in some places, including North American Numbering Plan countries). Spammers are the most likely people to figure this out and to have ways of generating regex-passing fake numbers. Even if near real-time lists of real phone numbers are available in all circa 200 jurisdictions the WMF is not going to be paying to access them (given the desirability of such lists to telemarketing firms I expect they are not cheap). Thryduulf (talk) 03:18, 26 July 2024 (UTC)

You start with an obvious factual error. Why?

Stepping back, we are presented with a choice:

1. Do nothing. All UPE product must be sent to AfD, where it is allegedly subject mass sock AfD keepers, or continue to rely on functionary abuse of CSD to delete what they think is UPE T&C violating pages

2. Expand CSD to encompass CSD practice mentioned in #1.

3. Find a way to limit throwaway accounts, the ones used by regular undisclosed PAID editors, and especially those allegedly used to undermine AfD.

Phone numbers are one known way to authenticate an individual as a probably individual. More than two-thirds of humans have a mobile phone. Of the subset who will ever edit Wikipedia, I bet the fraction is way higher. Phone number use as a method of authentication as a true individual is already being used for this by some, but this is not to say that we should model ourselves on them, it’s just proof that it can work.

Requiring authentication by access to an SMS doesn’t limit account creation, only authentication of that account.

It is easy to keep a phone number unique record safe, if recorded only in an encrypted form.

Of course verification is required. Once. That poor kid in the Bangladeshi swamp, who edits on a library computer, can get his verification code delivered via the town leader who has a phone. If there’s more than five such kids, we might ask the town leader some questions.

If it’s too hard, there’s Wikipedia:Request an account.

Sure, some clever Americans can authenticate through many different phone numbers, but I’ll bet that it’s not these people writing poor quality undisclosed paid promotional articles. SmokeyJoe (talk) 04:49, 26 July 2024 (UTC)

Clever? How much do you imagine a new sim costs? I've been openly offered enough to buy a hundred of them to rewrite a single article that I'd just deleted, and I live in the overpriced corporatocratic hellscape that is the US. The cost of a new phone number isn't going to make even a small dent in paid editing, just inconvenience everybody else. At best. —Cryptic 06:03, 26 July 2024 (UTC)

OK. This says $1 to $10. I didn’t think of that. SmokeyJoe (talk) 06:27, 26 July 2024 (UTC)

I'd like to highlight that in many countries, especially in Indian subcontinent, where a large number of UPEs operate, SIM cards are distributed free of charge. — Saqib (talk I contribs) 09:24, 26 July 2024 (UTC)

I don't think we will ever make creating throwaway accounts costly enough to stop UPErs. The economics simply don't stack up. As JPxG has pointed out above, the going rate for an article that sticks to mainspace on freelance websites is anywhere from USD$500 up to $2000, and there is constant demand because of our prominence in search engine results and the widespread perception that Wikipedia is independent of commercial interests. The only reason we aren't flooded with commissioned spam is that, up until now, we've managed to make it difficult enough for articles to stick that freelancers can only make a fraction of their contracts pay out, which has spread the word in SEO circles that Wikipedia is a risky bet. – Joe (talk) 10:17, 26 July 2024 (UTC)

I think the high value of a UPE article is a major issue here. Economically speaking, there are a couple things we can do about this:

1. Make it cost more to write an article, such that it eats into firms' profit
2. Decrease the value of the finished product

I think both are reasonable, and we can take action on both, but 1 runs into some limitations. The sole fact of PPP disparity means that something like "an hour of work" is of wildly different value between countries -- say we make an autoconfirmed account cost ten edits, and assume they take six minutes each for a new user -- this is $18 in San Francisco, but it's $0.36 in Bangladesh (fifty times less).

I think it is more productive for us to simply enact policies and procedures that diminish the value of a UPE article. Summary deletion is one of these things, or at least it was for UPE creations crappy enough to instafail WP:G11. This diminishes the value of the article sharply: no matter how much you paid for it, the whole thing can just be gone in a second. The thinking that goes into "sure, let's spend $2000 on a Wikipedia article" is that it's a large up-front investment but it's a long-term investment and when it sticks it sticks. Making it more likely that such articles will get discovered and instantly vaporized makes them a lot less valuable.

If I were tasked with coming up with something that would destroy UPE firms, and I was allowed to break one rule in doing so, I wouldn't go with phone verification -- I'd go with making deletion logs publicly visible and indexed by Google, so "dogshit astroturf spam article gets nuked" would be actively bad rather than something Wikipedia went out of its way to avoid reflecting on a company. I mean, I would be opposed to actually doing this, but at least it would work. jp×g🗯️ 20:50, 26 July 2024 (UTC)

That just makes it so you pay people to write deliberately-awful articles about your competitors, with the added bonus that what we currently delete as G11 seems less bad in comparison and we maybe get more lenient overall. —Cryptic 21:17, 26 July 2024 (UTC)

Just an aside, we do get off-wiki reports of people paying freelancers to trash their competitors in articles. It's not common, but it happens. – Joe (talk) 21:34, 26 July 2024 (UTC)

Yeah, that would be the crux of my opposition. But at least the inconvenience would be restricted to silly-valley startups and soundcloud rappers, and not all Wikipedia users across the globe. jp×g🗯️ 22:11, 26 July 2024 (UTC)

I remember seeing allegation a few years ago that a scammer repeatedly created blatantly deletable articles about their (the complainant's) company with the aim of getting the title salted (or possibly threatened that this is what they were going to do) when they (the complainant) refused to pay the scammer for a "good" article. I don't remember what industry they were in particularly but I think they were an established brick-and-mortar local/regional business. Thryduulf (talk) 23:13, 26 July 2024 (UTC)

Blanking

Is blanking a problematic userpage an acceptable alternative to speedy deletion? Ae245 (talk) 10:21, 6 July 2024 (UTC)

Maybe. It depends on the problem. Don’t blank instead of WP:G10 or WP:G12, but you might do better to quietly blank an old problem if you are not sure it reaches the G10 or G12 threshold. If you’re talking WP:U1 or WP:G7, it’s entirely your preference. Note that most CSD don’t apply to userpages.

Blanking is almost always preferable to MfD-ing someone else’s problematic Userpage, especially if they are long inactive. You can use {{Userpage blanked}}. SmokeyJoe (talk) 10:58, 6 July 2024 (UTC)

Okay, thanks. Ae245 (talk) 11:10, 6 July 2024 (UTC)

Proposal to revise CSD R3 for foreign language redirects

I've put together a proposal to revise and extend R3 to better address redirects in languages other than English. Your feedback is welcomed.  — Scott • talk 16:51, 9 July 2024 (UTC)

F8 and keep local

I can understand why quite a number of those uploading files might want to keep their creative content local by adding the template {{Keep local}}, but it seems that this template might also be being used by others with respect to content they didn't create. The thing that started me thinking about this is File:The Penguin (TV series) logo.jpg. It's obvously not the original work of the uploader. It was originally uploaded as non-free but subsequently converted to a PD license and moved to Commons based on c:Commons:Deletion requests/File:The Penguin (TV series) logo.jpg. The uploader then added a {{Keep local}} template to the file. I know some file uploaders have had some bad experiences with Commons, but it seems a bit odd for someone other than the original copyright holder of the content to be able to request such things. There certainly can be time and effort involved in finding content to upload to Wikipedia for use in various articles, but that doesn't really create a claim of ownership over the content for the uploader. So, it seems odd that acknowledgement of such a claim (at least in my opinion) is being given just to whoever adds "Keep Local" to a file, particularly in the case where the file licensed as PD.

FWIW, I'm not trying to single out one particular file or one particular uploader by linking to the file mentioned above; I'm only using it as one example of what are probably lots of similarly tagged files. It seems that there should be some restrictions placed on the use of this template, including perhaps limiting it to original content uploaded locally to Wikipedia in which the uploader/copyright holder is the making the request. For reference, Category:Wikipedia files on Wikimedia Commons for which a local copy has been requested to be kept has almost 6000 entries. How many of these really need to be or should be kept local? -- Marchjuly (talk) 03:00, 12 July 2024 (UTC)

{{keep local}} is simply is a request to retain a local copy, not a claim of ownership. Not seeing any reason to add an ownership component to it. Nikkimaria (talk) 03:14, 12 July 2024 (UTC)

I’m still newb about most issues regarding verification and what are facts and informations that are red flag in layman’s terms and our rules and regulations are evolving continuously as new technologies introduction of ie..AI but we must stand firm to some rules that are foundation to the organization, I introduced a proposal of team confirmation and a editor or contributor page highlighting Green on the statement word link etc… stating Color green as under verification to be factual or within legal limits boundaries including this issue you have encountered. Proposal of a Green Highlight as a flagged feature on wiki. Or any other color for easy visualization that an issu is presentThe Summum Bonum (talk) 12:34, 12 July 2024 (UTC)

@The Summum Bonum: I think you might've mistakenly posted your above comment in the wrong discussion thread or maybe even on the wrong talk page because it doesn't seem to be about what's being discussed here. -- Marchjuly (talk) 23:38, 13 July 2024 (UTC)

It's unlikely that someone "might want to keep their creative content local". A file file would not need to be kept local if the uploader is the copyright holder. If the user is the copyright holder, the file normally must have a free license. So, it can be on Commons without problem. Unless it is out of scope, but then it is likely out of scope on Wikipedia also.

The most usual use of the template "Keep local" would be for some files of which the user who adds the template is not the copyright holder. A typical use can be for files whose public domain status might be considered borderline or disputable or anyway not undoubtably 100% certain for everybody. And consequently there might be some possibility, even if small, that the file might be deleted from Commons. It happens that files are kept on Commons at one point but deleted some months or years later.

The file "The Penguin (TV series) logo.jpg" is a mild example. It was uploaded to Wikipedia on 13 Avril 2023, copied to Commons on 15 April 2023, deleted from Commons on 18 April 2023, undeleted on Commons on 10 May 2023, deleted from Wikipedia per F8 on 11 May 2023, nominated for deletion on Comons on 15 June 2023, undeleted on Wikipedia on 15 June 2023, kept on Commons on 17 October 2023, deleted from Wikipedia per F8 on 17 October 2023, undeleted on Wikipedia on 18 October 2023.

It can be noted that on 15 June 2023 it was not the uploader who added the template "Keep local". The template "Keep local" was added on Wikipedia by the user who nominated the file for deletion on Commons. Adding the template was indeed the wise and logical thing to do in that circumstance, while the file was still on Commons. That deletion discussion on Commons was closed as "kept", but if the file had been deleted from Commons, the template "Keep local" could probably have been replaced with the template "Do not move to Commons".

It can be noted also that the file was deleted from Wikipedia on 17 October 2023 despite the fact that the file was marked with the template "Keep local" at that time, so it was deleted in violation of the policy Wikipedia:Criteria for speedy deletion. Fortunately, that was noticed by another admin the next day and the file was undeleted.

We've seen typical worse examples, where files go through multiple such cycles of deletions and undeletions in a game of ping pong between Wikipedia and Commons. It's a waste of time and effort every time to have to track the deleted files, find an admin to undelete the files, only to have, six monts later, some user come from nowhere and, without thinking, pull a F8 again and destroy all the work and restart another round of the cycle.

The purpose of the template "Keep local" is that the file cannot be speedy deleted only by invoking F8. It's not that the file cannot be deleted. It could be nominated for deletion with a convincing deletion rationale. If the user who added the template is still active on Wikimedia, they should be consulted, to at least know and understand their reason and to check if they think the file must still be kept on Wikipedia. -- Asclepias (talk) 15:05, 12 July 2024 (UTC)

As pointed out below by JPxG, I think someone "might want to keep their creative content local" does happen, and it happens particularly because of the reasons they gave below. There are uploaders who have had really bad experiences with Commons; they, therefore, want to keep local files of their own work just in case. This seems (again at least to me) to be a valid reason for keeping the local file. I also do get that "keep local" is just a precaution against unnecessary speedy deletion and doesn't mean a file can't ever be deleted. FWIW, I haven't gone through each of the entries in Category:Wikipedia files on Wikimedia Commons for which a local copy has been requested to be kept. Maybe they're all like the files JPxG is referring to, but there are almost 6000 files in that category. My guess is that a good lot of them probably have been on Commons long enough to no longer justify a local version also being kept. Of course, removing the "keep local" doesn't mean the corresponding Commons file will never end up deleted. If, however, that happens for a really strong-polciy based reason (not some bot error or personal preference reason), it would also seem to imply that the local file should go as well. -- Marchjuly (talk) 23:38, 13 July 2024 (UTC)

When I upload files locally, it is on purpose, and there is a reason for it. I've often noticed (sometimes months after the fact) the mysterious silent disappearance of stuff I've uploaded to/embedded from Commons. This includes stuff which is unambiguously my own work and duly noted as such. Then I will look back, and see that it was deleted for some nonsensical reason: e.g. a bot failed to parse the license information from the license text I typed into the file description and marked it for deletion as "unlicensed". Sometimes images in active use will be nominated for deletion (with zero reference to policy or guidelines; on the explicit basis that the nominator doesn't like them) and this will go through after a second person agrees with them. At one point, there was an image which someone was slow-motion edit warring to remove from an enwp article, and after one of the removals, it got tagged for deletion at Commons as "unused", it was gone.

The times I've made undeletion requests on Commons or left messages to ask administrators directly about closures, they've been ignored, so I don't really have much choice except to upload the files locally on enwp, where there tend to be fewer frivolous deletion requests in the first place, but in the event there is one I will at least see it on my watchlist and be able to deal with it within a couple days, rather than realizing eight months later that some page looks different, digging into the edit history, and seeing that the image was delinked by a bot after a DR with the text "Hfdjksl". jp×g🗯️ 21:11, 13 July 2024 (UTC)

Can you link specific examples of instances where you've had files deleted without warning and/or had undeleteion requests ignored? Agreed that Commons isn't intuitive to navigate, but it should be possible to get help when you need it. Also if you are in a situation where you are being ignored, please feel free to loop me in, I'll apply pressure in all the right places :) -Fastily 22:58, 22 July 2024 (UTC)

Prior to speedy deletion and or removal of a statement on articles ie… promotional, promotion of violence and discrimination etc…

An additional contributor page of the article like talk with statements or discrepancies to be highlighted Green the whole article/s/ word/s or statement/s stating that the edit or statement or article is under review or under verification being acknowledged by Wikipedia to be considered a true statement or its existence and are/is factual or within legal limits and boundaries and will be finalized by team for compliance and confirmation The Summum Bonum (talk) 12:04, 12 July 2024 (UTC)

@The Summum Bonum Please could you try rephrasing that? I've read it several times and still don't understand what you are trying to say. Thryduulf (talk) 12:31, 12 July 2024 (UTC)

My apologies I’m simply proposing that a statement/sentence or a word/words can be Highlighted by color green fonts or any other colors to visualize that the statement/sentence or words/phrase is under review or being reviewed to confirm or verify that its is a legitimate statement or source etc..prior to deletion or adding another page on article a contributor page or Edit page /article/talk/discussion/editor/ page< Visualized for your convenience hope you got it this time again my apologies.

it wasn’t my intention to speak in an Coded encrypted paragraph an old practice.The Summum Bonum (talk) 12:52, 12 July 2024 (UTC)

If a page is being nominated for speedy deletion, the entire page is problematic, and we would then be highlighting the entire page. If only part of a page is problematic, then it should be dealt with via removal and no CSD is required. Primefac (talk) 19:42, 13 July 2024 (UTC)

R3 and redirects ending with "(disambiguation)"

The final sentence of R3 currently says It also does not apply to [...] redirects ending with "(disambiguation)" that point to a disambiguation page.. My first thought was this should be changed to "...a disambiguation page or a page that performs a disambiguation-like function" to match the language of G14. However I then realised that I don't think that's quite right either. I think the intent is to exclude WP:INTDABLINK redirects from being considered implausible. However, it also excludes redirects that contain very obvious typos (e.g. Bulse (disambiguation) → Blues (disambiguation)) or other clear errors (e.g. British Rail Class 9001 (disambiguation) → Languages of the Congo) which cannot be the intention. I'm not sure what the best alternative wording is, but something along the lines of "...unless the part before the parentheses contains implausible typos or is implausibly related to the target" (along with incorporating the disambiguation-like language from G14) maybe? Thryduulf (talk) 01:05, 16 July 2024 (UTC)

Do we really need it at all? It strikes me as the sort of thing that got stuck onto the end of the policy unnoticed because one day a single admin decided to improve his placement on that awful ADMINSTATS scoreboard by deleting every one of these that he could find, to a universal chorus of "No, of course those aren't implausible." At most it should be stuck down lower in the #Non-criteria section. —Cryptic 01:19, 16 July 2024 (UTC)

Wikiblame says it dates from Wikipedia_talk:Criteria_for_speedy_deletion/Archive_45#"Foo_(disambiguation)"_redirects_created_in_accordance_with_WP:INTDABLINK -> Wikipedia talk:Disambiguation/Archive 35#Speedy deletion of "Foo (disambiguation)" redirects. I also found User_talk:DangerousPanda/Archive_7#Intentional disambiguation redirects (context), User talk:RHaworth/2012 Jan 29#Intentional disambig redirects, User_talk:Cindamuse/Archive_19#Intentional_disambig_redirects, User_talk:Fastily/Archive_4#Intentional_disambig_redirects. Of the four admins listed there two of them have been desysopped due to unrelated misconduct, one of them stopped editing in 2014, and one of them is still an admin but probably knows better a decade later. * Pppery * _{it has begun...} 01:55, 16 July 2024 (UTC)

Well that shows that BD2412 should be invited to express their opinion. Now that I'm more awake I realise that if we do want anything, we could be massively more concise and say something like "it also does not apply to [...] correctly formed WP:INTDABLINK redirects." Thryduulf (talk) 10:21, 16 July 2024 (UTC)

My opinion as expressed in the discussion is correct. A Bar (disambiguation) redirect (or even a BAR (disambiguation) redirect) pointing to the disambiguation page Bar (to which BAR also points) should never be speedily deleted as an "implausible" typo, because such a redirect is not implausible at all, it is policy-supported to have it. Word this as you wish to make it clear. BD2412 T 15:14, 16 July 2024 (UTC)

So many things would be so much better if ALL disambiguation pages ended with " (disambiguation)". There would be no barrier to newcomers understanding what a disambiguation was. Readers going to a page would know upfront that they were going to a disambiguation page. Most of these troublesome pages would never have been created. SmokeyJoe (talk) 11:31, 16 July 2024 (UTC)

There would still be a need for redirects like Languages of the Congo → Languages of the Congo (disambiguation) that I'm sure would cause about the same number of issues with deletion, incorrect bold retargetting and people not knowing/understanding (or disagreeing with) the exception to the usual primary topic is not disambiguated rule. Whether it would be better for readers I don't know, but it wouldn't be significantly better (or worse) for editors. Thryduulf (talk) 11:45, 16 July 2024 (UTC)

Redirects are cheap.

The disambiguation pages would be unambiguously disambiguation pages, and articles would be unambiguously not disambiguation pages. Simple obvious principles, rather than convoluted rules and practices, makes for less issues.

A page title accurately telling the reader what the page is, a dab page or not, is obviously better for the reader, in my personal experience for sure. SmokeyJoe (talk) 13:03, 16 July 2024 (UTC)

Exclude G2 from draftspace

If WP:G2 doesn't cover userspace, why should it cover draftspace? The same reasons apply: experimentation isn't an unreasonable thing to do in draftspace, it's not indexed, and greeting a new user with Template:Db-test-notice is rather bitey. I'm also concerned that G2 is being used as a catch-all criterion to delete things that aren't test edits. I've listed the last 100 draftspace G2s here: hardly any are prototypical editing tests, while a large share are either blank pages (often good-faith placeholders that shouldn't be deleted) or low-quality but non-test efforts at writing an article or user page. The risk of bitey invalid deletions outweighs the handful of valid ones, and at any rate almost none of it needs to be deleted since detritus in draftspace is harmless and cleared out after six months anyway. I would suggest excluding draftspace from G2, just like we've done for userspace. Extraordinary Writ (talk) 07:14, 18 July 2024 (UTC)

Drafts in draftspace are meant to be drafts. Drafts of mainspace content, usually new articles, but drafting for merging to an existing article is perfectly reasonable. It’s not junkspace.

At MfD, we often delete draft pages for not being a genuine draft. This is softly worded delete justification for something that could be more aggressively called a test, vandalism or hoax, or implausible unverifiable material.

If a draftpage is obviously an old test, why not G2 it? Alternatives are to ingore it, or move it to the author’s userspace, Userfy it. Userfication of a test is much less bitey than speedy deletion, and if it was a test, the user will presumably want to look at again.

I think draft space tests should be userfied, unless “test” is a euphemism. SmokeyJoe (talk) 11:53, 18 July 2024 (UTC)

Support. I can't think of any reason why something in draftspace would be harmful enough to need speedy deletion and not fall under at least one of G1, G3, G9, G10, G11 or G12 and as Smokeyjoe says userfication or just waiting for G13 is going to be more appropriate in most cases anyway. Thryduulf (talk) 11:59, 18 July 2024 (UTC)

Oppose. These is actually a case where a test edit won't be dealt with by G13. If the test page is a redirect, then G13 doesn't apply. WP:R3 won't always apply either due to the "recently created" requirement. Being able to retain a test page forever by sticking a "#REDIRECT Wherever" on top goes against the principle that draftspace is not junkspace. Nickps (talk) 12:11, 18 July 2024 (UTC)

WP:RDRAFT says most redirects in draftspace should be kept. In other cases take it to RfD, from the list compiled by Extraordinary Wit there was only one redirect and that could (probably should) have been deleted under G8 anyway so there doesn't seem like RfD will be unable to handle the few remaining cases where a redirect in draftspace needs to be deleted. Thryduulf (talk) 12:17, 18 July 2024 (UTC)

RDRAFT concerns redirects made by moves. I'm talking about the case where a test page is created and as part of whatever test edits the editor makes, they also happen to make the page a redirect. In that case G13 won't apply, so RfD and userfication are the only ways forward. That was the point I was trying to make. Honestly, userfication is a better approach anyway, so I'll think about retracting my oppose, but G13 is still a flawed argument. Nickps (talk) 12:39, 18 July 2024 (UTC)

In my experience, G2s anywhere are mostly used as a catch-all criterion to delete things that aren't test edits. An old analysis. —Cryptic 12:28, 18 July 2024 (UTC)

Perhaps it's time for someone to write WP:!G2? Nickps (talk) 12:40, 18 July 2024 (UTC)

Maybe the meaning of “test” is not technical, but a test of Wikipedian tolerance for a bad faith contribution. A breaching experiment often involving promotion. SmokeyJoe (talk) 14:30, 18 July 2024 (UTC)

LOL. * Pppery * _{it has begun...} 15:48, 18 July 2024 (UTC)

Agreed with Cryptic. I've declined several G2s for that reason, and yesterday I declined a G2 at Willy Hüttenrauch only to be overridden by another admin who deleted it a minute later in an edit conflict. But my motivation to aggressively patrol the deletion log has been low lately, so not much has gotten done. * Pppery * _{it has begun...} 15:48, 18 July 2024 (UTC)

Whether that was a test or not, it was also a broken redirect. —Kusma (talk) 12:32, 25 July 2024 (UTC)

• If admins could be trusted, I would oppose this (genuine edit tests should be deleted without having to wait 6 months). But Extraordinary Writ's analysis makes it clear they can't, so support. * Pppery * _{it has begun...} 15:48, 18 July 2024 (UTC)
• Strong support - Draftspace deserves a certain amount of leeway that it often has not been afforded as it has matured. It needs less technical maintenance, "cleanup", and deletion. This is a step in the right direction. — Godsy ^(TALK_CONT) 07:51, 25 July 2024 (UTC)
• There is a lot of junk in draft space that violates WP:NOTWEBHOST but has no chance to ever become an encyclopaedia article. Deleting it per G2 is suboptimal, but happens a lot when the only other alternative is G13. I am opposed to restricting G2 unless we extend U5 to draft space. —Kusma (talk) 08:10, 25 July 2024 (UTC)

  If you're deleting non-test pages using G2 then you need to stop immediately as you are abusing your admin tools. If these NOTWEBHOST pages are actually a problem and don't meet an actual speedy criterion (rather than one you would like to exist) then take them to MfD. If they are as frequent as you claim then you will soon have the evidence needed to craft a speedy deletion criterion similar to U5. Thryduulf (talk) 09:06, 25 July 2024 (UTC)

  From my own G2 deletions:

  • Draft:HS content was "HS"
  • Draft:Wow Beck hansen content was "Wow Beck fanpage".
  • Draft:Matilda luck content was a different person's name and "ewwwwwwwwwww". Possibly a G10, but who knows. Deletion as G2 assumed no bad faith.

  The general problem with "test pages" as a speedy criterion is that we are asked to judge intent, not content. For a lot of graffiti pages ("Jim loves Suzie!", "I am the Playstation KING!") it seems the most appropriate of the G criteria (otherwise these will probably get nominated as G11's). —Kusma (talk) 12:28, 25 July 2024 (UTC)

  Why did the first two of those need to be deleted before G13 applied? The third was a G10 - it served no purpose other than to disparage the subject. Thryduulf (talk) 12:31, 25 July 2024 (UTC)

  Maybe we should rename Draft space to "Anything goes for six months" space so people do not work under the mistaken assumption that it is for drafting Wikipedia articles. —Kusma (talk) 12:35, 25 July 2024 (UTC)

  Consensus seems to be that the purpose of draftspace is somewhere between "anything goes" and "strictly only for drafting Wikipedia articles." If you think something in draftspace is actively harmful but doesn't meet a speedy deletion criterion then that's what MfD is for. Thryduulf (talk) 12:40, 25 July 2024 (UTC)

• Support per nom. Draftspace is not a terrible place to do a test. —SmokeyJoe (talk) 09:33, 25 July 2024 (UTC)

Allow creators to remove R4 tags

R4 is a criterion split from G6, but unlike G6, the redirect's creator is not allowed to remove the tag themself. I see no reason that restriction is necessary for R4 in particular and I think that it should be removed, much like it was done for G14. Nickps (talk) 12:32, 20 July 2024 (UTC)

Note that there are valid reasons to contest an R4 like the ones listed at Wikipedia:Criteria for speedy deletion § Other issues with redirects. Nickps (talk) 12:39, 20 July 2024 (UTC)

Has this actually been a problem? R4 is one of those criteria where the "objective" and "uncontestible" dogmas truly apply - either it has the same name as a file on Commons or it doesn't, and if it truly does then the problem is unfixable and it should always be deleted and if it doesn't than an admin will decline. In either case there's no value to the creator removing the tag. * Pppery * _{it has begun...} 15:40, 20 July 2024 (UTC)

I'm not aware of a convenient way to find all contested R4 deletions so I can't answer that. My motivation for this change is mostly to regularise the CSD process and make it less hostile towards new users. For example, up until today, {{db-redircom-notice}} (as well as {{db-talk-notice}}, {{db-disambig-notice}} and {{db-rediruser-notice}}) directed editors to a non existent "Contest this speedy deletion" button. After I fixed that, I also realised that there is really no reason to disallow the creator removing the R4 tag so I brought it here. Since, as Thryduulf has pointed out, R4 isn't entirely objective and there is a valid reason for the creator of a redirect to remove the R4 tag, can we put this to rest now and add R4 to the list? Nickps (talk) 16:31, 20 July 2024 (UTC)

And realistically there will never be file redirects with useful page history so your second comment doesn't apply. * Pppery * _{it has begun...} 15:43, 20 July 2024 (UTC)

If there incoming links then it is not eligible unless they are "clearly intended for the file on Commons" (which is subjective). The implication being that links to the image that is not at this title on Commons need fixing first, and the creator could be highlighting the existence of such links. I don't see a problem with the suggestion. Thryduulf (talk) 15:44, 20 July 2024 (UTC)

In that case they may as well fix them rather than removing the tag and allowing things to remain indefinitely in a state the community has declared verboten. * Pppery * _{it has begun...} 16:38, 20 July 2024 (UTC)

That assumes that the creator can fix all of the incoming links, which will not always be true - for example some may require discussion (e.g. where the intended target is not clear) or be on pages that they cannot edit (e.g. protected pages). The page can be nominated for G4 again when the links are fixed or taken to RfD at any time. Thryduulf (talk) 08:24, 22 July 2024 (UTC)

• I think its fine to allow authors to remove R4 tags, many users dealing with redirects are experienced and unlike R3 which is similar to A7 and A9 R4 doesn't seem like its too much of a problem to allow authors to remove in the rare cases where they object. RFD would be sufficient in such cases. Crouch, Swale (talk) 17:33, 20 July 2024 (UTC)

C4 – unused maintenance categories

Hi all. I am thinking of drafting a new category CSD, which would cover unused maintenance categories. It would cover two related situations. The first is it would split from G6 empty dated maintenance categories from the past, and thereby lessen the load G6 is bearing. The second case is maintenance categories no longer used by a template. As an example, Category:Eiei-year — used with year parameter(s) equals year in page title was at one point populated by {{eiei-year}}, but that template no longer populates that category after a rewrite. It is not a G8 because {{eiei-year}} still exists – it just no longer populates that category. (Note that empty != unused: categories which happen to be empty are not necessarily unused. I am talking about categories a template does not populate under any circumstances.)

NEWCSD checklist (I am only focusing on case two, because case one is already eligible for CSD):

1. Objective: Y Obviously objective: either a category is in use or it is not
2. Uncontestable: Y As a regular CFD closer, I have only seen these get deleted unanimously (see my list below)
3. Frequent: Y case one is the most common reason G6 is used (see Taxonomy of G6 deletions), and see below for case two
4. Nonredundant: Y I guess case two could be a part of G6, but the last thing we need is to shove more deletion reasons into G6. And obviously case one is currently part of G6, but getting this out of G6 is a feature, not a bug.

Is this something people would be inclined to support? Are there other related cases which should be included? If so, we can work on wording, but I wanted to get others' input first. HouseBlaster (talk · he/they) 17:06, 21 July 2024 (UTC)

Just to be pedantic, WP:G8 currently says Categories populated by deleted or retargeted templates, so a redirected template that no longer uses a cat would make the cat eligible for G8. Primefac (talk) 17:08, 21 July 2024 (UTC)

{{eiei-year}} was moved (per WP:TPN), not redirected to a different template. I was just using the shortcut because that is what the categories used in their names. HouseBlaster (talk · he/they) 17:11, 21 July 2024 (UTC)

Ah, my misunderstanding. I would still argue that if the template is not populating the category, it is eligible for G8. Primefac (talk) 17:13, 21 July 2024 (UTC)

• Last discussed at Wikipedia talk:Criteria for speedy deletion/Archive 86#Empty monthly maintenance categories, which more or less petered out due to lack of participation. I'm all in favor, with mild preference towards merging the main case into C1 (without a timeout) rather than a separate C4. —Cryptic 17:16, 21 July 2024 (UTC)
• Support I'd agree this makes sense similar to G14 and R4 splits of G6. Crouch, Swale (talk) 17:17, 21 July 2024 (UTC)
• Support this as well - turns a deletion people (including myself) are doing anyway by stretching G6 and G8 in areas they don't quite into a clear and objective criterion. * Pppery * _{it has begun...} 17:25, 21 July 2024 (UTC)
• I continue to think this is not a good idea, for the reasons I gave in the last discussion: "All it seems to do is shift these deletions from one category to another, and while I agree that G6 is overused, the proper solution is to talk with admins who are using it incorrectly (and go to DRV if necessary), not to make a new criterion for something that undeniably is 'uncontroversial maintenance'" Per the closure there, this would require an RfC. Extraordinary Writ (talk) 18:23, 21 July 2024 (UTC)

  Another thing that I should've made more clear: case two is currently not (at least in my understanding) covered by any CSD criteria. They were in my CSD log because of {{db-xfd}}. I guess that could be described as uncontroversial maintenance, but stretching G6 even further is unappealing. Nor is stretching G8 to cover cases in which the "dependent" page still exists, imo. HouseBlaster (talk · he/they) 00:48, 22 July 2024 (UTC)

• Support per Crouch, Swale and Pppery. Stretching criteria on a regular basis is something that we strive to avoid and adding making G6 larger and more complicated is not a solution we should even contemplate undertaking. The nom presents a good case that a new criterion is needed. Thryduulf (talk) 07:50, 22 July 2024 (UTC)
• I have a bit of a concern with respect to the template rewrite part namely if consensus is required for it or not otherwise for non protected templates anyone could rewrite them and the category ends up being deleted. Maybe the dependant on template part should be added to C1 to allow a week before deletion to allow for objections. Crouch, Swale (talk) 18:06, 22 July 2024 (UTC)

  Has this ever actually been a problem? A careful admin applying this criterion would have to look at the history to see when (and hence why) the category was removed. and if there's some hairy dispute involved they won't delete it. On the other hand non-careful admins will just use G6 for this anyway so in neither case does the waiting period help. And undeletions are cheap anyway. * Pppery * _{it has begun...} 18:14, 22 July 2024 (UTC)

• Alright, before starting a formal RfC here is my draft wording of C4:

  C4. Unused maintenance categories

  This applies to unused maintenance categories, such as empty dated maintenance categories for dates in the past (e.g. Category:Articles lacking sources from July 2004) or tracking categories no longer used by a template after a rewrite. Note that empty maintenance categories are not necessarily unused—this criterion is for categories which will always be empty, not just currently empty. If you are unsure whether a category is still being used by a template, consider asking the creator of the category or at the template's talk page before tagging.

  I think we should also allow the creator to remove the CSD tag, because they are the person who best understands whether the category is being used or not. Comments? Suggestions? Typos? HouseBlaster (talk · he/they) 03:06, 24 July 2024 (UTC)

  Does that really meet NEWCSD #1? Proving that a category will always be empty can be challenging, and I'm not convinced we should just let admins figure it out unilaterally. If someone tagged Category:Technology articles with topics of unclear notability, for instance, what would be due diligence for me as a reviewing admin? Examining the source code and history of Template:Notability in detail might help, but even that doesn't rule out that some other template somewhere is using it. I suppose I could run a search like this one, but it's not realistic to expect a deleting admin to do that. If it's possible to be unsure whether a category is still being used by a template, this probably isn't straightforward enough for CSD. Extraordinary Writ (talk) 03:41, 24 July 2024 (UTC)

  Can't the deletion tag have an insource search link built into it so all you have to do is click it? Gonnym (talk) 08:50, 24 July 2024 (UTC)

  We can (and maybe do?) encourage categories that may be sometimes empty to have the {{Empty category}} notice on them. Additionally we should strongly encourage category descriptions to link to all the templates that populate that category. These won't quite solve the issue completely but it very nearly will. Thryduulf (talk) 12:01, 24 July 2024 (UTC)

  We do use {{empty category}} on all potentially empty categories because it stops the category from appearing in the C1 report. And I agree that both including the insource search link on the tag and an encouragement to note which templates use the category should be sufficient. The worst case scenario is a category needs to be REFUNDed, and at that point we can make a note on the category itself saying something like This category is used by {{foo}}. HouseBlaster (talk · he/they) 14:09, 24 July 2024 (UTC)

  In which case this has my support. Thryduulf (talk) 14:14, 24 July 2024 (UTC)

• Support, splitting G6 is generally a good idea if we want to clarify that it is not "clearly unnecessary pages that should be deleted". Category pages almost never have any interesting history; everything about categories is on other pages, so there is very little harm in deleting any empty categories as long as there is no limit on undeletion or recreation. —Kusma (talk) 08:24, 24 July 2024 (UTC)
• I've left a note about this discussion at Wikipedia talk:WikiProject Templates. Thryduulf (talk) 14:22, 24 July 2024 (UTC)

RfC: enacting C4 (unused maintenance categories)

Should C4 (unused maintenance categories) be enacted as a new criterion for speedy deletion? 03:09, 25 July 2024 (UTC)

Proposed text:

C4. Unused maintenance categories

This applies to unused maintenance categories, such as empty dated maintenance categories for dates in the past (e.g. Category:Articles lacking sources from July 2004) or tracking categories no longer used by a template after a rewrite. Note that empty maintenance categories are not necessarily unused—this criterion is for categories which will always be empty, not just currently empty. If you are unsure whether a category is still being used by a template, consider asking the creator of the category or at the template's talk page before tagging.

• Support as proposer. There are two benefits that I see from this change. One, it lessens the load that WP:G6 is carrying. However, the primary reason that I came here is to allow for speedy deletion of additional unused maintenance categories. As an example, Category:EstcatCountry — used with year parameter(s) equals year in page title was previously populated by {{EstcatCountry}}, but no longer did so after a rewrite. It is not a G8 because {{EstcatCountry}} still exists. As a regular closer at CfD, I have only ever seen them get unanimously deleted and it is a fairly regular occurance (you can see the collapsed list above at List of entries in just my own CSD log since June 1 for unused maintenance categories (it is possible I missed some, but I think I got them all); n.b. they were only CSD candidates as {{db-xfd}}s). The WP:NEWCSD checklist:
  1. Objective: Y Obviously objective: either a category is in use or it is not
  2. Uncontestable: Y As a regular CFD closer, I have only seen these get deleted unanimously (see my list below)
  3. Frequent: Y ~20 in the past two months at CFD, and many more which are currently handled by G6
  4. Nonredundant: Y I guess it could be a part of G6, but the last thing we need is to shove more deletion reasons into G6. And obviously dated maintence categories are already part of G6, but G6 is already overloaded and decreasing that burden is a feature, not a bug.
• In sum, this would decrease the burden on both CFD and G6 while also saving editor time rubber-stamping pro forma discussions. One note that did come up in the above discussion is that we can program the {{db-c4}} tag to include an insource: search to make it easy for the patrolling admin to double-check that the category is in use. Best, HouseBlaster (talk · he/they) 03:09, 25 July 2024 (UTC)

Notified: T:CENT, Wikipedia:Village pump (policy), and WT:CFD. Pinging participants in the above discussion: @Crouch, Swale, Cryptic, Extraordinary Writ, Gonnym, Kusma, Pppery, Primefac, and Thryduulf. HouseBlaster (talk · he/they) 03:15, 25 July 2024 (UTC)

• What's wrong with leaving it at G6? G6 says "This is for uncontroversial maintenance" – including, but not limited to, empty dated maintenance categories. Permanently empty undated maintenance categories look like the definition of "uncontroversial maintenance". This will add extra complexity (more CSD cats to watch) without changing the end result (the cats always get deleted). I'm not sure that adopting this would solve any problems. WhatamIdoing (talk) 03:33, 25 July 2024 (UTC)

  Per above, the "what's wrong" is that G6 is overloaded and splitting some would make it easier for reviewers. It also lessens the potential hassle of typing a reason. Support Aaron Liu (talk) 04:00, 25 July 2024 (UTC)

  @Aaron Liu, Category:Candidates for technical speedy deletion currently contains two (2) pages. What makes you think that two pages is "overloaded"? WhatamIdoing (talk) 04:15, 31 July 2024 (UTC)

  @WhatamIdoing it is overloaded in the sense of doing too many different things. Thryduulf (talk) 11:10, 31 July 2024 (UTC)

  • (edit conflict) All speedy deletions are uncontroversial maintenance by definition. Deleting empty nonmaintenance categories is uncontroversial maintenance. Deleting user subpages when their user says they're done with them is uncontroversial maintenance. Deleting disambiguation pages after everything linked from them has already been deleted is uncontroversial maintenance.
    What makes having different criteria meaningful is that A) you can show that a given criterion is known to be uncontroversial; B) users can easily tell why a given page was deleted, and if they care, find the underlying discussions why it's uncontroversial; C) you can list what kinds of things that, while superficially meeting a criterion (such as "AfricA" not being a plausible misspelling of Africa), are nonetheless widely considered to be controversial; D) you can find specific instances where deletion is controversial, and E) you can find the sorts of deletions you're willing to make: I don't act on third party {{db-move}}s anymore because I always get blamed when the move turns out to be controversial, and sometimes my brain's too fried to deal with G11s but I have some spare time I could spend deleting empty maintenance cats if they weren't mixed in with those untouchable db-moves.
    Worse, the controversial and incorrect uses of G6 are getting drowned out by formulaic, well-defined subcriteria of G6 that are only included in G6 because it was too much of a hassle to get an independent criterion passed. Empty maintenance categories are the single largest identifiable group of G6s, accounting for more than one in six out of every deletion mentioning "G6" anywhere in the deletion summary. There's so many that they make it near-impossible to find genuine abuses. Show me an admin who's never declined a {{db-test}} where the closest thing to a test was "OK, so you tried to create an autobiography of yourself without anybody noticing. Your test failed, and I noticed and now I'm going to get an admin to delete it!" and I'll show you an admin who hasn't performed enough speedy deletions to talk knowledgeably about the subject. Or maybe one who just doesn't give a fuck and will happily twinkle-delete anything you put in front of them.
    That, of course, is a (terrible) example of a G2, not a G6; but I've declined {{db-error}}s that amount to the same thing. Lumping them in with automatic deletions made as part of a page move that don't require you to push a delete button or provide a deletion summary or even be an admin, and with unarguably-non-speedy deletions where a specific consensus was formed at TFD, and "Hooray, we've finally cleared the backlog of unreferenced pages! (up until December 2004, anyway)", in a speedy deletion criterion that was created for temporary deletions made as part of history merges, makes it impossible to find such abuses. —Cryptic 04:24, 25 July 2024 (UTC)

    I've made several attempts to do what Cryptic suggested at https://quarry.wmcloud.org/query/61527/Wikipedia_talk:Criteria for speedy deletion/Archive 83#New taxonomy. I found a very large number of different uses, some of which IMO met the criteria and some of which didn't, and then an unclassified "other" which made up a third of the entire set. I also made an attempt once at https://quarry.wmcloud.org/query/62274 to find questionable G8s largely out of spite at someone arguing G8 applied in a case where I felt it clearly didn't. Neither went anywhere because there were too many "other" for me to have the will to look at, and today I limit WP:Database reports/Possibly out-of-process deletions#G6 to old mainspace pages. In the unlikely event I decide to attempt either of those again, having this subcriterion split out won't make the G6 classification runs any easier (dated maintenance category runs were among the easiest to filter out there), and won't make the G8 classification runs that much easier (the biggest problem I had there is that there's no way to find out that a redirect was broken without looking at the deleted history). * Pppery * _{it has begun...} 05:20, 25 July 2024 (UTC)

• Support per my comments in the previous section. * Pppery * _{it has begun...} 05:20, 25 July 2024 (UTC)
• Support per proposer and my previous comments. Thryduulf (talk) 08:47, 25 July 2024 (UTC)
• Support. No one seems to know how to nominate this type of page (C1, G6, G8, etc) so putting these into a dedicated and specific group will ease the burden not only on the nominator but also on the patrolling admin. Primefac (talk) 12:07, 25 July 2024 (UTC)
• Support per nom and Primefac. —pythoncoder (talk | contribs) 19:28, 25 July 2024 (UTC)
• While we're at it, let's explicitly include empty "Wikipedia sockpuppets of Example"/"Suspected Wikipedia sockpuppets of examples" categories. Last time I did a G6 taxonomy they were one of the more common kinds of G6, and, assuming we're fine with them being deleted instantly when they become empty, should put them here. * Pppery * _{it has begun...} 20:36, 25 July 2024 (UTC)

  I'd support that, too. HouseBlaster (talk · he/they) 22:20, 25 July 2024 (UTC)

  Support. I think that falls under "unused maintenance categories" but there is no harm and possible benefit to making it explicit. Thryduulf (talk) 22:51, 25 July 2024 (UTC)

  inb4 someone argues WP:CREEP Aaron Liu (talk) 04:44, 26 July 2024 (UTC)

  For reference, these categories are generated via Template:Sockpuppet. —⁠andrybak (talk) 13:20, 28 July 2024 (UTC)

• Support, anything that reduces the "catch-all" status of G6 is a win. Yes, technically we could have literally nothing but G6 as criterion, and argue that all other criteria count as "uncontroversial maintenance", but having precise criteria allows for more specificity, and spells out what kind of stuff is already known to be uncontroversial deletions. This is a very good example of a criterion that is both uncontroversial (these categories are never going to be populated again) and precise enough to be formulated as a criterion of its own. Chaotic Enby (talk · contribs) 00:30, 27 July 2024 (UTC)
• This is the sort of thing that would be uncontroversial post-enactment, but as long as we're still discussing, can we link WP:Maintenance category here to make it crystal clear which cats are covered and which need to go through the week-long C1 process? —Cryptic 00:48, 27 July 2024 (UTC)
• Separately, suggest "categories which are expected to always be empty", rather than "categories which will always be empty". It's not at all rare for empty dated maintenance cats to get temporarily repopulated, such as when a redirection is reverted or somebody recreates an article with a cut-and-paste of a predeletion version. —Cryptic 00:48, 27 July 2024 (UTC)
• Support per my previous comments. Crouch, Swale (talk) 05:59, 27 July 2024 (UTC)
• Support - By explicitly limiting this to maintenance categories that will never again be populated, i.e. dated maintenance categories that have become empty, this criterion allows for specificity and clear guidance to any admins. — Jkudlick ⚓ (talk) 19:28, 27 July 2024 (UTC)
• Support per nom and Primefac. C F A 💬 22:11, 28 July 2024 (UTC)
• support: seems like it'd be useful per nom, Primefac, & Chaotic Enby. ... sawyer * he/they * talk 07:53, 1 August 2024 (UTC)
• Support In addition to the ones that are currently handled in CfD, such as Wikipedia:Categories_for_discussion/Log/2024_July_24#Category:Technology_articles_with_topics_of_unclear_notability, dated maintenance categories (such as the monthly ones in Category:Articles_that_need_to_differentiate_between_fact_and_fiction) are routinely deleted per G6 when they become empty. –LaundryPizza03 (dc̄) 06:35, 3 August 2024 (UTC)

What criteria would illegal things fall under

Child corn

If someone uploaded something that is illegal (e.g. child corn) to Wikipedia, what speedy deletion criteria would it fall under? I'm just wondering, I know it's an odd question but I've had it on my mind for a few days and I want to know BombCraft8 (talk) 21:32, 21 July 2024 (UTC)

That very much likely falls under WP:OVERSIGHT more than CSD. Headbomb {t · c · p · b} 21:40, 21 July 2024 (UTC)

Ahh, makes sense. BombCraft8 (talk) 21:41, 21 July 2024 (UTC)

However, I do think that in the case of illegal stuff, that kind of stuff should be deleted in its entirety from Wikipedia, because just suppression would still allow the oversighters to continue to be able to see that stuff even after it was suppressed. BombCraft8 (talk) 21:47, 21 July 2024 (UTC)

For legal reasons, you can't fully expunge these things (for example, Bob upload child porn. If you expunge the database, you now no longer have evidence that Bob uploaded child porn). That's why Oversight exists and why the requirements to be an oversighter are close to ARBCOM-levels of trust, and must sign NDAs with the WMF and so on. Headbomb {t · c · p · b} 21:56, 21 July 2024 (UTC)

That makes sense. BombCraft8 (talk) 21:58, 21 July 2024 (UTC)

(edit conflict) In my understanding some truly illegal stuff (like child pornography uploaded to Commons) truly is hard-deleted from the database. But otherwise we require oversighters to sign a NDA and trust them not to be fools. Given that there's illegall porn etched forever into the Bitcoin blockchain for example I don't think we need to care. * Pppery * _{it has begun...} 21:58, 21 July 2024 (UTC)

Also note that at that point WP:OFFICE would get involved, and the lawyers can direct when things are clear to be fully purged from our records. Headbomb {t · c · p · b} 21:59, 21 July 2024 (UTC)

My understanding is that child pornography specifically needs to be reported to [email protected] and probably oversight too as per above. It's above the pay grade, so to speak, of normal admins/editors, and thus not really suitable for a CSD criterium - although WP:CSD#G9 would presumably apply. Jo-Jo Eumerus (talk) 06:32, 22 July 2024 (UTC)

If anyone finds material on a WMF project that is truly illegal for the WMF to host (such as child pornography) you should report it immediately to the oversight team of that project (or the stewards if there are no local oversighters) and to WMF legal. Thryduulf (talk) 08:05, 22 July 2024 (UTC)

Indeed. An admin who finds these things should RD it in the meantime to keep non-admins from seeing it while waiting for an OSer to suppress and the Office to memory-hole it. Primefac (talk) 16:54, 22 July 2024 (UTC)

ok BombCraft8 (talk) 19:18, 22 July 2024 (UTC)

We, the community, including admins and even oversighters, have no responsibility to deal with legal issues. We do deal with policies (including copyright policy) and preventing major real-world harm. Anything beyond this is for the Foundation, specifically its legal department, to handle. The law explicitly exempts the Foundation from the need to notice legal violations itself, it only has to handle cases reported to them. Animal lover |666| 08:11, 25 July 2024 (UTC)

I have two questions regarding this:

• If a user comes across such an image and immediately reports it to all relevant parties (legal-reports, oversighters, law enforcement, and anyone I missed, let me know if I did), is there any possibility of criminal liability on the user for accessing it? Is it enough to ensure that my caches are purged and permanently deleted according to my computer, or must I do more (such as overwrite the space the files once contained or physically destroy the computer)?
• Are there any other types of images or content that require this third layer of deletion? If so, what level of liability exists or could exist for random users accessing such content?

-BRAINULATOR9 (TALK) 22:25, 26 July 2024 (UTC)

What liability you would have depends on your jurisdiction. IANAL (and nobody here can give you legal advice even if they are), but iIn the United Kingdom my understanding is that if you were not looking specifically for that content, you had no reason to expect to find that content at that location (and this would apply to a Wikipedia article) and you reported it (or attempted to report it) to the relevant people as soon as it was reasonably practical for you to so then you would have no liability at all, even if you do not delete your cache. Thryduulf (talk) 23:04, 26 July 2024 (UTC)

I'm American - Virginian, to be more precise. -BRAINULATOR9 (TALK) 01:15, 27 July 2024 (UTC)

No, really, what category does illegal stuff fall under?

Lots of very scintillating conversation above, unfortunately none of it seems to address a rather simple core issue: what actual category is it under?

If I see something illegal, such as "child corn" ^[sic] — not asking whether I should contact the functionaries — which of these do I select from the dropdown?

The actual options that I get in the dropdown menu when I delete a page (e.g. the contents of MediaWiki:Deletereason-dropdown) are:

• General (can apply to any page)
  • G1: Patent nonsense, meaningless, or incomprehensible
  • G2: Test page
  • G3: Vandalism
  • G3: Blatant hoax
  • G4: Recreation of a page that was deleted per a deletion discussion
  • G5: Creation by a blocked or banned user in violation of block or ban
  • G6: Housekeeping and routine (non-controversial) cleanup
  • G7: One author who has requested deletion – to retrieve it, see WP:REFUND
  • G8: Page dependent on a deleted or nonexistent page
  • G10: Attack page or negative unsourced BLP
  • G11: Unambiguous advertising or promotion
  • G12: Unambiguous copyright infringement
  • G13: Abandoned draft or Articles for Creation submission – to retrieve it, see WP:REFUND/G13
  • G14: Unnecessary disambiguation page
• Talk pages
  • G8: Talk page of a nonexistent or deleted page
• Redirect pages
  • R2: Cross-namespace redirect from mainspace
  • R3: Recently created, implausible redirect
  • R4: Redirects in the "File:" namespace with the same name as a file or redirect at Wikimedia Commons
  • G8: Redirect to a deleted or nonexistent page
  • X3: Redirect with no space before a parenthetical disambiguation
• Other criteria
  • Listed at Wikipedia:Copyright problems for over seven days
  • WP:PROD: Nominated for seven days with no objection
  • WP:BLPPROD: Nominated for seven days with no reliable sources present in the article

I would say that there is, in fact, no CSD category for them, since every category is very rigorously defined and does not include illegal content. All possible options smell strongly of bullshit rules-lawyering.

• "It's a G4 because, uh, if you think about it, man, we've already had deletion discussions that reached consensus to delete all videos of men having sex with horses"
• "It's a G5 because, uh, if you think about it, man, anybody who posts a video of a man having sex with a horse is blocked."
• "It's a G6 because, uh, if you think about it, man, it's not very controversial to delete that, is it?"
• "It's a G10 because, uh, if you think about it, man, it's defamatory to the horse."
• "It's a G11 because, uh, if you think about it, man, it's trying to convince me to have sex with a horse."
• "It's a G12 because, uh, if you think about it, man, horses can't release content as CC-BY-SA."

The only remotely plausible thing I can think of is to manually enter it in as a G9, which I am pretty sure creates an actual urgent issue for WMF Legal, so I do not think it is a good idea to do this. jp×g🗯️ 02:27, 31 July 2024 (UTC)

It's not a WP:CSD case, it's WP:CRD (any of RD1-4 would apply) then WP:OVERSIGHT. Headbomb {t · c · p · b} 02:18, 31 July 2024 (UTC)

Hmm -- but the revdel criteria don't show up in the dropdown on Special:Delete and WP:REVDEL does not explicitly mention actual page deletions anywhere that I can see. jp×g🗯️ 02:43, 31 July 2024 (UTC)

You do know you can type something there without selecting anything from the dropdown, right? Revision deletion doesn't work if there isn't a good version to revert to anyway; and the particular cases raised are likely to be files, which also don't really play nicely with revision deletion.

JPxG, you were made an admin because people had confidence in your good judgment, not in how well you're able to ruleslawyer in order to argue you're permitted to take an action that makes the encyclopedia better. This is the sort of thing WP:Ignore all rules really is for, for all that it's usually poorly-regarded when it comes to deletion.

That said, you don't want to have something like "02:42, 31 July 2024 JPxG (talk | contribs | block) deleted page File:Me and Joey at Disneyland.jpg (omg iar child corn HALP!)" in your deletion log, for the same reason as the bolded text midway through Wikipedia:Revision deletion#Hiding oversightable material prior to Oversight - you pick something banal for the log comment like G7 or F5. (And if the file was at File:Omg_child_corn.jpg or whatever, go ahead and revdelete your deletion log too.) Then you mail oversight and block the uploader and it's not your problem anymore, at least once they're out of sleeper socks.

And if you insist on playing Nomic, G3 is closest. WP:CSD#G3 → Further information: Wikipedia:Vandalism and... → WP:Vandalism#Image vandalism → "using any image in a way that is disruptive". Genuinely illegal imagery can't fall under the "if they have encyclopedic value" clause. —Cryptic 03:10, 31 July 2024 (UTC)

Crypic has it almost spot on. It is arguable that an image of child pornography could serve an encyclopaedic use on the Child pornography article it definitely cannot elsewhere, but uploading an image that you know it is illegal for the WMF to host is unarguably vandalism.

Certainly in terms of child pornography and similar, anything that is not the uploader's own work will almost certainly either be a copyright violation (the normal laws around copyright are not impacted by it being illegal), or the free license claimed will be unverifiable. Claims that material of this nature is the uploader's own work will also almost certainly be unverifiable (by us, the relevant law enforcement body may be interested though) - I would also argue that it is not credible someone would openly claim images that are illegal to create were created by them if it were true.

All of this means that such material is covered under G3 and G12. Thryduulf (talk) 11:55, 31 July 2024 (UTC)

Incredibly problematic ("illegal") content should not be CSD'd. The Oversight team should be contacted directly. If a page must be quickly and expediently deleted, contact an admin directly (IRC, Discord, email to an admin you know is active, etc). That being said, it's not the end of the world if a page is tagged (as Thryduulf suggests above) for G3, but keep in mind that throws it into multiple well-viewed categories so it will likely draw more attention. As much as it might seem like a good idea, {{db-reason|Child corn}} as suggested above is a bad idea, primarily because it increases the chances of that showing up in the deletion log itself (and therefore requiring more hiding). Primefac (talk) 12:11, 31 July 2024 (UTC)

I don't think Thryduulf was suggesting that non-admins tag material like this {{db-g3}}, let alone {{db|child corn}}. I know I wasn't; I was responding directly to and advising another admin. For a non-admin, yes, tagging oversightable material is a bad idea - not only is CAT:CSD highly visible on Wikipedia, there are some... fine... projects that preferentially mirror the pages in it, the better to fight the Evils of Rampant Deletionism. —Cryptic 13:54, 31 July 2024 (UTC)

I mean, assuming I send somebody a message or whatever, it's still going to be a few minutes before they get it; am I just supposed to refrain from deleting it during this interval and leave the goat sex pics/etc sitting there untouched until they get around to formally OSing it? This feels like it cannot be the case. jp×g🗯️ 18:03, 31 July 2024 (UTC)

If you are an admin delete or revdel it with a bland summary. If you aren't and it's a busy page revert the edit/blank the page with a bland summary. If you aren't an admin and it's a page with few likely readers just leave it - especially if you aren't autopatrolled. The goal is to avoid drawing attention, and most OS requests get actioned in much less than 5 minutes. Thryduulf (talk) 18:42, 31 July 2024 (UTC)

Perhaps a G15 that is similar to WP:RD4 should be created so it's obvious to admins that they should do what Thryduulf wrote above. I'm thinking something like:

G15. Oversightable information
This applies if every revision of a page is eligible for suppression. See WP:REVDEL#Hiding oversightable material prior to Oversight for when and how to use this criterion.

Nickps (talk) 23:35, 31 July 2024 (UTC)

Absolutely not! The goal is to avoid drawing attention to oversightable material, not putting up a bright red arrow pointing to it! Thryduulf (talk) 23:50, 31 July 2024 (UTC)

I don't see how that CSD would draw any attention to the oversightable material. Much like RD4, admins would never invoke it by name, they would put some other bland reason in the field and contact oversight, just like you said they should. Nickps (talk) 23:56, 31 July 2024 (UTC)

G▉? jp×g🗯️ 00:28, 1 August 2024 (UTC)

It will draw attention because clueless admins will use it in their log comments even if it says not to. Source: there's one deletion, two log deletions, and 135 revision deletions that mention "RD4". (With the false positives like "prod contested by Richard44306 at WP:REFUND" and "G5: Created by a banned or blocked user (Ford489) in violation of ban or block" filtered out, but I haven't looked at most of the actual deletions except for their log comments. All the ones labeled like "RD4, serious BLP vio" and "RD4/WP:YOUNG" and "RD4: Personal and non-public information: real name and harassment" that I've spot-checked were the real deal though.) —Cryptic 00:59, 1 August 2024 (UTC)

Damn. I didn't know RD4 was misused so much. Maybe just adding some text that says "admins can delete oversightable material while waiting for oversight per the instructions at WP:REVDEL#Hiding oversightable material prior to Oversight" to the lead without assigning a criterion number would work, but honestly, it might be better to have it be an unspoken rule, so some admin doesn't delete with reason "oversight". Nickps (talk) 02:06, 1 August 2024 (UTC)

I think at most we need something (a hatnote perhaps) that points to the instructions elsewhere, perhaps "For material that needs to be oversighted see ...". Thryduulf (talk) 10:05, 1 August 2024 (UTC)

Yeah, that. jp×g🗯️ 11:25, 1 August 2024 (UTC)