Wikipedia:Village pump (idea lab)/Archive 61

• Village pump
  • Policy
  • Technical
  • Proposals (persistent)
  • Idea lab
  • WMF
  • Miscellaneous

Village pump (idea lab) archive  

This page contains discussions that have been archived from Village pump (idea lab). Please do not edit the contents of this page. If you wish to revive any of these discussions, either start a new thread or use the talk page associated with that topic.

---

< Older discussions · Archives: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62

Adding personal page view history

I think it would be a cool feature to be able to view the history of articles which you have viewed in the past on the web version of Wikipedia (I think this might exist on the app version already?). I go through many Wikilink rabbit holes and I’d like to be able to see a list of all the articles which I have read. Is this a conceivable addition to the Wikipedia software? Cleebadee (talk) 21:10, 12 October 2024 (UTC)

I suggest using your browser history (which may mean finding a browser with a conveninent way to search your history). I think many users would feel uncomfortable with our Wikipedia reading history easily available for anyone to access on the Wikipedia server. (The information can of course be determined from web server logs, but that's less convenient for bad actors to take advantage of.) isaacl (talk) 21:32, 12 October 2024 (UTC)

As an analogy, my local library uses software that has an option of whether to retain a record of books I have checked out, which is turned off by default. With the default, the record of a book I have checked out is erased when I return the book. That way, if a government agency demands a record of the books I have checked out, and I have left the option set to not retain a record, the library has no information to turn over. Government agencies and other parties cannot optain records that do not exist. BTW, I might be paranoid, but I have earned it. I have copies of two FBI reports about me (via the US Freedom of Information Act), one about 200 pages long and the other 22 pages long. Donald Albury 22:05, 12 October 2024 (UTC)

Are you saying that Wikipedia should allow this as an option? It seems like this would be a good way to make it a thing without forcing people to use it. Maybe it could be a tool which is off by default? Cleebadee (talk) 03:28, 13 October 2024 (UTC)

I am saying that one of the considerations on whether to keep a record on what you have looked at on Wikipedia is that if a record exists it can be subpoenaed or hacked. Another consideration is that if such records do not exist, then the Wikimedia Foundation does not have to expend time and money responding to/resisting government requests for access to those records. If I could have a say on whether to have such an option, I would oppose it. I think users deserve the right to read whatever they want to on Wikipedia without worrying about someone looking over their shoulder. We cannot prevent someone from monitoring what a reader looks at from their end, but we can do what we can to preserve privacy within Wikipedia. Donald Albury 13:54, 13 October 2024 (UTC)

Would having the record available on Wikipedia itself make it easier to hack than it being available indirectly through one’s search engine history? Cleebadee (talk) 14:31, 13 October 2024 (UTC)

You can filter your browser history by websites. In the worst case, you can search "Wikipedia" in the history. Aaron Liu (talk) 22:11, 13 October 2024 (UTC)

I feel like using a hypothetical Wikipedia view history and searching Wikipedia in a search engine history are both easy things for a bad actor to do. So I don’t understand the idea that adding the Wikipedia one would make it easier for bad actors. Is it easier for them to hack into Wikipedia compared to search engines? Cleebadee (talk) 20:12, 14 October 2024 (UTC)

If you want the data to be synced across devices, you'd need to upload the information to Wikpedia's servers; meanwhile, history is usually not synced by default. It's also additional complexity to the software with very little benefit. Aaron Liu (talk) 20:49, 14 October 2024 (UTC)

I believe Zotero Connector has this option which you can activate for the wikipedia domain. That is, if you're interested in tracing back your wikipedia history (and of other websites and reading, and of sources for WP editing and general research + writing) beyond simply your browser history, Zotero is a nice thing to have around. SamuelRiv (talk) 22:23, 12 October 2024 (UTC)

Nice, I hadn’t heard of this website before. I use an iPad so I don’t think I’m able to use this though. Cleebadee (talk) 14:29, 13 October 2024 (UTC)

Automatically add archive date

The archive URL comes in this format: www.web.achive.org/web/yyyymmdd/https:www.placeholder.com

The date can be automatically entered when the URL is given, instead of having to add the date. Who am I? / Talk to me! / What have I done? 05:03, 7 October 2024 (UTC)

I know that is true of the Internet Archive, but I seem to remember that other archiving services have different URL formats. Alt.Donald Albury (talk) 12:00, 7 October 2024 (UTC)

Maybe we can detect the link (starting with web.archive.org), and if it is the internet archive, then the format can be used. The other archives are not used significantly, so this will not impact the process much. Who am I? / Talk to me! / What have I done? 12:43, 7 October 2024 (UTC)

This is an excellent idea. Archive.org does not use yyyymmdd but, for example, 20031224105129. I manually have to insert dashes at the appropriate points 2003-12-24105129 and remove the timestamp so I end up with 2003-12-24. It seems very very likely that a bot is already doing this task, but to be sure you can ask over at Meta or WP:VPT. Polygnotus (talk) 23:08, 7 October 2024 (UTC)

Ok, so only the archive URL is needed to fill the citation template. The date can just be filled in automatically

Who am I? / Talk to me! / What have I done? 02:47, 8 October 2024 (UTC)

Pinging @GreenC for archive.org's date format WhatamIdoing (talk) 17:43, 9 October 2024 (UTC)

As an idea, I fully support this. I also used to wonder why the heck the template would nag about the missing date when it's available in the URL, and was particularly annoyed by the timestamp mismatch message. However, I've since learned that this is not as easy to implement as it may seem.

• {{cite}} templates literally can't add the missing parameter or alter it, because templates can only alter their output (i.e. the displayed HTML), not change the wikitext input. I guess strictly speaking this would be possible with subst:-ing, but forcing everyone to do that with {{cite}} isn't going to fly.
• The editor (software) certainly could make this happen automagically, but I'm pretty sure the devs would (understandably) be disinclined to introduce special treatment for specific templates on a specific wiki. Also, while the visual editor could hide the necessary wikitext change behind the scenes, for the source editor this would have to mean auto-correcting the submitted wikitext after the fact, and AFAIK this is currently not done for any reason.
• Since {{cite}} is usually placed inside <ref>...</ref>, mw:Extension:Cite could conceivably handle this, but again, good luck convincing the devs.

So, I think "fill in |archive-date= from |archive-url= while the user is editing" is probably going to be a hard sell. There are other considerations, though.

• Since a missing or mismatched |archive-date= automatically lands the page in Category:CS1 errors: archive-url, is it necessary to nag editors about it (and, especially, put red error messages in the displayed page)? fixemptyarchive and fixdatemismatch functionality of WP:WAYBACKMEDIC can easily fix these errors, so why not have the bot run more frequently than the current every 2-3 months? Not necessarily with its full functionality, just against this particular category. I don't really see why this couldn't happen daily.
• And to put this more radically: is a separate |archive-date= needed at all when its content merely duplicates that of another parameter? Since CS1 Lua code already has some special handling for archive.org, why not just add taking the date from the URL as well? CS1 predates the availability of Lua, so perhaps this was too inconvenient to implement with the old template functionality and requiring editors to always include this parameter was unavoidable. This no longer seems to be the case, though. Archive information doesn't seem to be needed for COinS (which wouldn't be taken directly from wikitext anyway), nor do I know of any other reason to keep it mandatory for archive.org links (or any others that include the date).

I'm very possibly missing something important regarding the status quo, so hopefully people like @Trappist the monk and @GreenC can chime in.

Gamapamani (talk) 12:29, 12 October 2024 (UTC)

Thanks a lot for refining the proposal. The second consideration makes a lot of sense. Adding an archive URL is always manual, even if you use the automatic tool, so I think it makes sense to remove the parameter and, using Lua, find the date from the URL. Who am I? / Talk to me! / What have I done? 14:23, 12 October 2024 (UTC)

--> Category:CS1 errors: archive-url currently has 33 pages. I clear it every 15 days. Last time it was cleared was around October 1. Not a big problem. -- GreenC 19:47, 12 October 2024 (UTC)

Sure, I didn't mean to imply it should be cleared more often as things stand right now. What I was trying to say was that if |archive-date= nagging in the editor and on the resulting page were dropped for archive.org, the category would presumably fill somewhat faster, but the bot could still handle it. Gamapamani (talk) 02:59, 13 October 2024 (UTC)

Yes, The CS1 error need not be shown. Who am I? / Talk to me! / What have I done? 03:53, 17 October 2024 (UTC)

URL expansion bots

I agree that short URLs are undesirable. However, wouldn't it be better if a bot auto-expanded those URLs instead of them being blacklisted? For example youtu.be into youtube.com/watch?v= , tinyurl.com/example into example.com (that is its actual target). Elominius (criticize | contributions) 09:19, 30 September 2024 (UTC)

The reason they're blacklisted is not because they're "undesirable", it's because they can be (and have been) used to get around the spam blacklist and other anti-spam measures. If the link is legitimate, there's no reason the user trying to add it can't expand it. Anomie⚔ 11:25, 30 September 2024 (UTC)

@Anomie: there's no reason the user trying to add it can't expand it The reason is that the enduser does not know that they have to do that, and the message does not explain that that is what should be done (or how). And several sites use shorter URLs when you use the Share button, which is a type of URL shortener that only goes to one domain, like a domain alias and therefore cannot be abused for malicious purposes. Polygnotus (talk) 23:00, 7 October 2024 (UTC)

I can certainly see a justification for treating shorteners linked to a single domain (like youtu.be) differently to generic ones like tinyurl. Thryduulf (talk) 23:22, 7 October 2024 (UTC)

@Thryduulf: See meta:Talk:Spam_blacklist#youtu.be. Polygnotus (talk) 23:43, 7 October 2024 (UTC)

For that case, as I had written in VPT, the bot could do blacklist checking and reversion if the expanded link hits the blacklist; or we just integrate the automatic expansion mechanism into MediaWiki. MilkyDefer 08:34, 10 October 2024 (UTC)

This misses the point that people could use a URL shortener to link to spam or malware even if the target is not blocked. A quick check of a diff of someone doing that shows the shortener link and the person checking would need to be fully motivated to examine what happens if they click it. If the original edit adds a link to a marketplace or a dubious site, it is a lot easier to revert. We do not want a bot which blesses external links by expanding them. Johnuniq (talk) 01:32, 11 October 2024 (UTC)

Possibly the bot could discover who added the link, if the account passes some heuristic tests of trustworthiness, it would be considered reliable-enough for expansion. Lot of work though. How big of a problem are short URLs? -- GreenC 02:50, 11 October 2024 (UTC)

It depends what problem you are referring to. According to the Meta discussion they are used to post spam links, but they also prevent the addition of good links and may (I don't know of any statistics) discourage the inclusion of sources and/or lead to abandoned edits both of which could negatively impact editor retention (directly or indirectly) - all of which are problems, but different ones.

I think the ideal would be for MediaWiki to perform a pre-save transform to expand short URIs to their full form, then check them against the blacklist, then either save the long form or reject the edit as appropriate (The error message for a rejected edit should display both short and long URIs so editors know which link tripped the blacklist and why). I guess this is something that would need to happen at the software level though? I'm also wondering if there is some means of automatically detecting which URIs are shortened or if that would have to be curated manually? (e.g. would it know that http://sucs.org/uri/as is a short URI (leading to URL shortening) without being told?) Thryduulf (talk) 03:18, 11 October 2024 (UTC)

. I have not read the meta thread but suspect most cases are good faith and the bad instances are on the edge. It's easy to check, manually go through a couple dozen and see how it looks.

. To expand to long form requires querying a header, external site, or API, might not practical at the MediaWiki interactive layer.

. Shortened URLs could be documented; I made a page that documents archive URLs: Wikipedia:List of web archives on Wikipedia (including some that are shortened URLs). It's easy to start a technical document Wikipedia:List of URL shortening sites on Wikipedia, and hope over time people find it and add knowledge. -- GreenC 04:06, 11 October 2024 (UTC)

We should never rely on anyone to make a future edit. — xaosflux ^Talk 13:56, 17 October 2024 (UTC)

Creating Template:Wikidata Infobox

Hi, I propose to create a template called Template:Wikidata Infobox that creates an infobox from information exists on Wikidata. The same idea is implemented in WikiCommons. See https://commons.wikimedia.org/wiki/Category:Quantum_computing and section infobox which uses {{Wikidata Infobox}}. Cheers. Hooman Mallahzadeh (talk) 13:33, 1 October 2024 (UTC)

Using Wikidata for infoboxes has been discussed many times here and is surprisingly (to me) controversial. Some specific infoboxes do incorporate information from Wikidata (iirc Mike Peel has done some work on this), but I don't think a single generic infobox, whether pulling information from Wikidata or otherwise, will gain consensus. I'll leave a note about this discussion at Wikipedia talk:WikiProject Infoboxes and Wikipedia talk:WikiProject Wikidata. Thryduulf (talk) 13:42, 1 October 2024 (UTC)

Yeah… a LOT of resistance to importing data from Wikidata into our infoboxes. The two main concerns are Verifiability/reliability (although WD has improved on this front, they still are not in line with our policies) and ease of editing (having to go to WD to edit information appearing on WP can be confusing).

I’ll share a personal experience of confusion… the data focused structure of WD is often incomprehensible to me as a text focused editor here at WP. When I try to fix errors on WD, I have extreme difficulty doing so. simply locating the information I need to edit is hard. The way WD pages are organized and structured is alien gobbledygook to me. Blueboar (talk) 14:18, 1 October 2024 (UTC)

Wikidata is a good idea in want of a usable interface. It's use would be massively helped if you could edit data here and it was back flushed to Wikidata. -- LCU ActivelyDisinterested «@» °∆t° 14:39, 2 October 2024 (UTC)

That's how most of the Wikivoyages are set up. It's not automagic (except possibly at German), so you have complete control over which bits you import locally and which bits of your content you push back to Wikidata. The control is important because some content is different. For example, Wikivoyage usually wants to put the lat/long location at the entrance to an location, and Wikidata usually wants the center. There's no need to override each other's locations if these happen to be significantly different (e.g., entrance to Disney World vs center of Disney World). When they're the same, then you can share them back and forth. WhatamIdoing (talk) 04:23, 3 October 2024 (UTC)

Commons only has one kind of infobox. We have a lot of them that have very different data displayed. Personally, I'd love to incorporate wikidata into nearly all infoboxes, but one generic infobox is impossible to suit our needs. Aaron Liu (talk) 14:14, 1 October 2024 (UTC)

I really think that this type of infobox (maybe in collapse form) is the best replacement for infobox of articles that we cannot create any infobox for them like Quantum computing. These data includes links to WikiQuote and its library id etc., which makes them accessible and at hand. I propose to use this kind of infobox in other sections like "See also" section, instead of top, replacing many existing templates like {{Commons category}}. Hooman Mallahzadeh (talk) 14:32, 1 October 2024 (UTC)

Um, what value could be derived from using d:Q17995793 to populate an infobox for the article Quantum computing? Do we really need an infobox for that article to clarify that the subject is an instance of "academic discipline", subclass of "computation", and is the study of "quantum computer" and/or "quantum supremacy"? This is an excellent example of an article that has no need for an infobox. Folly Mox (talk) 02:31, 2 October 2024 (UTC)

Oh I see you want to use the infobox at the bottom of an article like a navbox? That's less objectionable, and also a different kind of box in our jargon. Folly Mox (talk) 02:44, 2 October 2024 (UTC)

@Hooman Mallahzadeh, not every article benefits from an infobox. In my opinion, Quantum computing is one of those. Remsense ‥ 论 06:14, 2 October 2024 (UTC)

@Remsense There are good links to WikiQuote, Wikiversity, Wikidata, WikiCommons, Library of Congress authority ID, that is attached to the main article by simply transcluding this template. This is a big benefit for Wikidata-Infobox of Quantum computing. But collapsing this Wikidata Infobox by default seems reasonable. Hooman Mallahzadeh (talk) 06:26, 2 October 2024 (UTC)

We can already do that by linking to the ones that are important at the bottom—this is extremely common, and is already done on that article! That brings up a key point of resistance to this, though. We don't want to outsource much to Wikidata because we can't as easily decide what not to show, lest we pollute articles with metadata that may be useful in a database but is functionally useless garbage in an encyclopedia article. Fundamentally, we shouldn't ever expect editors to have to use Wikidata also.Remsense ‥ 论 06:28, 2 October 2024 (UTC)

This method is a little hard, ⁣inserting cumulative links «by one template» seems more reasonable to me. Hooman Mallahzadeh (talk) 06:31, 2 October 2024 (UTC)

No, because the decision whether it's worth linking to Wikiquote on a given article should be up to the editors for that article. Remsense ‥ 论 06:32, 2 October 2024 (UTC)

I disagree. According to Zeigarnik effect, placing a Wikiquote link that is empty right now, motivates users to complete that page and put some quote about that concept. Hooman Mallahzadeh (talk) 06:38, 2 October 2024 (UTC)

It's not our job to build Wikiquote, but it is our job not to indiscriminately clutter our encyclopedia articles with useless garbage. Your position would be resented by almost any editor who cares about how the articles they write look and what exactly they present to readers. Remsense ‥ 论 06:59, 2 October 2024 (UTC)

Most readers don't care about those. The small bit that care are satisfied by it being at the bottom. Aaron Liu (talk) 11:34, 2 October 2024 (UTC)

Hooman Mallahzadeh, your idea of a single general use template to build sister project links from the associated Wikidata item for use in the ==External links== section does actually sound like a good idea, but people in this subthread have been confused by your use of the term infobox (which live at the top of the article). However, this sounds identical to the existing template {{Sister project auto}}. How does your idea differ? Folly Mox (talk) 12:35, 2 October 2024 (UTC)

If the notion is to show some source information, it may also be similar in concept to the wider-used template {{Authority control}}. SamuelRiv (talk) 15:07, 2 October 2024 (UTC)

Right, {{authority control}} is already ubiquitous (~2,131,000 transclusions out of 6,931,935 articles including redirects), is even included in the default output of Wikipedia:ArticleWizard, and {{authority control}} already pulls from Wikidata. Folly Mox (talk) 17:54, 2 October 2024 (UTC)

So all that considered, @Hooman Mallahzadeh, I'd suggest just making/forking a template along those lines (or editing the existing template in its sandbox), as this basic concept would be uncontroversial. The rest of the discussion here is hypothetical clutter until people see precisely what you have in mind that may be radically different from what is already being used in these templates above. SamuelRiv (talk) 18:41, 2 October 2024 (UTC)

Something like Category:Earth shows how hard it is to get an automatic, good infobox and not a load of weird entries. "Instance of: "inner planet of the Solar System (Mars, Venus)" And Mercury? if you for some reason list the others here, list them all... "Named after: *soil (Earth) *land (1, 地, 地球) *ball (2, 球, 地球) " Er, what? No idea why the Japanese is shown here. "Inception: 4,541st millennium BC (lead-lead dating, age of the Earth, Young Earth creationism)" Yeah, we sure want a link to Young Earth creationism here... "Dissolved, abolished or demolished date: unknown value (future of Earth)" Not even a link, just the text, as if this is in any way useful. And this is hardly some obscure example. Fram (talk) 14:30, 1 October 2024 (UTC)

{{Infobox person/Wikidata}}, anyone? — Qwerfjkltalk 18:42, 1 October 2024 (UTC)

What a great way to introduce unverified intormation (errors) and nonsense into Wikipedia articles. -- Ssilvers (talk) 20:46, 1 October 2024 (UTC)

Unverified information and incorrect information (what I presume you mean by "errors") are not synonyms. Not everything in WikiData is any of unverified, incorrect or nonsense. Thryduulf (talk) 20:51, 1 October 2024 (UTC)

@Ssilvers There's an easy parameter to only include information with references. Bad references exist and can be easily fixed using the same methods on both sites. Aaron Liu (talk) 21:01, 1 October 2024 (UTC)

Wikidata has different sourcing standards than enwiki - sites that are considered unreliable by enwiki consensus are considered quite fine at Wikidata. Wikidata entries are also left out of existing enwiki cleanup mechanisms. So it's not quite as simple as you're suggesting to apply the same methods to both sites - the two sites have neither the same methods nor a shared understanding of what a "bad reference" even is. Nikkimaria (talk) 00:22, 2 October 2024 (UTC)

Is anyone trying to develop a shared understanding, or does everyone think that it's just OK that the site are disconnected like that, even thought they could help each other much more? If you have links to discussions, I'd be happy to read them. Amir E. Aharoni (talk) 00:46, 2 October 2024 (UTC)

My perception is it's mutual apathy in both userbases: people who write an encyclopedia aren't coterminous with people who want to build a universal database. While the results are unfortunate, it really would be unreasonable to expect one group of volunteers to operate according to the standards of the other. Remsense ‥ 论 06:07, 2 October 2024 (UTC)

Then we should make it known to Wikidata why and which sources are bad and often false. I don't see any source that is considered very valid on Wikidata and removed-on-sight on Wikipedia. Aaron Liu (talk) 01:10, 2 October 2024 (UTC)

[1]. Nikkimaria (talk) 01:25, 2 October 2024 (UTC)

@Nikkimaria, was this a reply to me? Just one comment from one person? Amir E. Aharoni (talk) 02:15, 2 October 2024 (UTC)

It was not a reply to you. This one would be a better place to start for your question, though that perspective is also relevant. Nikkimaria (talk) 02:32, 2 October 2024 (UTC)

Thank you, point taken. In that case, we do need to have parameter-specific overrides at least. Aaron Liu (talk) 02:43, 2 October 2024 (UTC)

Curiously, one of the arguments in that RfC is "FInd a Grave is sourced from gravestones". Wouldn't citing the gravestone be better in that case instead? Veering off-topic here, though. Aaron Liu (talk) 02:46, 2 October 2024 (UTC)

The claim that Wikidata is wrong is just a claim. It's a wiki, it is closely integrated with Wikipedia, and it can be as right or as wrong as Wikipedia itself. Amir E. Aharoni (talk) 21:06, 1 October 2024 (UTC)

Yes, we know Wikipedia is flawed… Which is why we DON’T consider Wikipedia a reliable source, and DON’T use one part of Wikipedia to verify other parts of Wikipedia. Blueboar (talk) 21:18, 1 October 2024 (UTC)

But what's wrong with that if we can machine-verify that the reused part has a source? This isn't about sourcing to Wikidata, it's about reusing sourced content from wikidata, for all the reasons {{excerpt}} is good. Aaron Liu (talk) 21:20, 1 October 2024 (UTC)

Is anyone suggesting to use Wikidata to verify things on Wikipedia? I don't think so. I certainly don't.

People are suggesting to insert things that are written on Wikidata into the English Wikipedia in some cases when it's more efficient to do it. They can be verified just like Wikipedia itself. Amir E. Aharoni (talk) 21:32, 1 October 2024 (UTC)

I don't think Wikipedia should prioritise efficiency. We should prioritise care. If people want to reuse sourced content from Wikidata, they can add the content and add the source. We shouldn't just invite it in uncurated. Folly Mox (talk) 02:41, 2 October 2024 (UTC)

To paraphrase Shrek: But it is curated. What's the difference if it's curated by a Wikipedia editor or by a Wikidata editor? Wikidata it not a completely separate site. The two sites have always been closely related in terms of both community and technology, and with the same ultimate goal of free knowledge edited as a wiki. If there are differences between them in verifiability policy, I'd think about trying to bridge them instead of dismissing the idea of collaboration outright. Amir E. Aharoni (talk) 13:43, 2 October 2024 (UTC)

We shouldn't leave curation of content that is included in articles to people who generally aren't looking at or directly editing said articles. This seems very straightforwardly obvious to me. Remsense ‥ 论 13:51, 2 October 2024 (UTC)

But why do you think that they aren't looking at the articles? I very often edit things on Wikidata because I notice that they appeared incorrectly on Wikipedia, and then I check that the information was updated correctly on both Wikidata and Wikipedia. (It happens more in other languages, such as Hebrew, Spanish, or Russian, because the English Wikipedia uses Wikidata less.)

And what's the difference between changing a thing on Wikidata that will be included in a Wikipedia article and changing a template within Wikipedia that will be included in other pages? (Other than the different URL.) Or changing an image on Commons and Amir E. Aharoni (talk) 14:00, 2 October 2024 (UTC)

The curation processes between the two sites are entirely different. en.wiki is primarily curated at the individual page level, wikidata seems to be primarily curated at cross-sections where lots of pages intersect. (Also worth considering how much of Wikidata is created by bots pulling from other wikis.) CMD (talk) 14:03, 2 October 2024 (UTC)

Because they're editing a database! They might not even speak English, never mind take into account the nuances of how the English Wikipedia (or any other project, to be clear) may be affected by what they are directly doing. Whether you think they aren't totally different websites, they are different websites. I really don't think I have to get sociological to discuss a set of social facts here that are fairly obvious to everyone who contributes to a Wikimedia project. "Wikipedia is one site, Wikidata is another" should also answer your second question: if someone breaks a template, it is not a paradigm shift for me to fix it or ask someone else to, because that occurs on the same website. You're a free spirit and that's fine; any design decision that assumes this quality of contributors in general would be a terrible one. Remsense ‥ 论 14:11, 2 October 2024 (UTC)

Another way to think of it is that they are the same website with different URLs.

And I'm not sure what are you referring to by "quality of contributors" and "free spirit" towards the end. In my understanding, a free spirit in a free encyclopedia is a good thing, but I suspect that you mean something different. Amir E. Aharoni (talk) 19:08, 2 October 2024 (UTC)

Wikidata is definitely quite a different site, ubiquitous in its low standards for inclusion. Aaron Liu (talk) 19:29, 2 October 2024 (UTC)

That, by itself, is not a problem. You can include some things from it on Wikipedia, and exclude others. That's more or less what happens on the English Wikipedia already, but it happens more in some other languages, and it works there fairly well. I don't understand the resistance that some English Wikipedians have to include anything from Wikidata. Amir E. Aharoni (talk) 16:14, 3 October 2024 (UTC)

I agree that it is not a problem. I disagree with what you said about them being the same site and closely related. Aaron Liu (talk) 17:11, 3 October 2024 (UTC)

They:

• Are maintained by the same organization.
• Are built from the outset to be connected, similarly to Commons.
• Share user accounts.
• Share a large part of the editing community. I don't have precise numbers, and of course there are some Wikipedia editors who don't do anything on Wikidata and vice versa, but there is a lot of overlap.

So what makes you think that they are not closely related? Amir E. Aharoni (talk) 22:06, 3 October 2024 (UTC)

Closely related in terms of underlying infrastructure and overlapping goals doesn't equate to being the same website with different URLs. By design, each Wikimedia wiki has its own community, with its own culture, that defines its own guidance and operating procedures. isaacl (talk) 22:14, 3 October 2024 (UTC)

All I'm trying to say is that it's not very different from Commons. It's separate in some ways and the same in some others. Some data from Commons and Wikidata is useful in the English Wikipedia, some isn't. Emphasizing only the differences, as some English Wikipedians do, is neither correct nor useful. Amir E. Aharoni (talk) 22:22, 3 October 2024 (UTC)

All en.wiki takes from Commons is essentially file urls. It's a very different place to en.wiki with very different norms. CMD (talk) 23:24, 3 October 2024 (UTC)

And, um, the files, not just the URLs, right?

Whatever internal norms Commons has, the English Wikipedia takes a lot of files from it. And it can be the same with Wikidata. The English Wikipedia already takes some data from it, and it can take more.

I'm actually not specifically supporting using the generic Wikidata infobox in the English Wikipedia. I'm just trying to understand the resistance that some English Wikipedia editors have to including anything at all from Wikidata. Amir E. Aharoni (talk) 03:12, 4 October 2024 (UTC)

Key elements of the resistance have been explained elsewhere in this thread. They relate to a lower quality of sourcing, higher vulnerability to vandalism, and difficulty in editing Wikidata. Sourcing issues do not apply to commons, and file modifications there are restricted to those with advanced permissions. En.wiki does not import files from Commons, although it does host files separately when they are ineligible for inclusion on Commons. CMD (talk) 04:17, 4 October 2024 (UTC)

Commons benefits from the shared requirements set by the Wikimedia Foundation on media usage, so files are readily usable on any Wikimedia wiki. (Of course, media that embeds information, such as annual stats, suffer from similar problems as Wikidata: its accuracy is dependent on the uploader.) For those concerned about the verification standards on Wikidata, unfortunately I don't really have a good suggestion on creating better processes to validate edits while keeping the "anyone-can-edit, everyone verifies" approach. By its nature, the data is very dense, so I think trying to double check all incoming edits is a more tedious and onerous job than can be expected of volunteers to do. isaacl (talk) 06:35, 4 October 2024 (UTC)

However, the Wikidata editors do not curate content in articles: editors at wikis that use the wiki-data do. Wikidata is the sourced information, and editors decide which ones to include. That is also why I think templates should include per-parameter overrides before switching to wiki-data by default. Aaron Liu (talk) 19:30, 2 October 2024 (UTC)

As far as I know, per-parameter overrides is what is done in Wikipedia in all the languages in which templates pull information from Wikidata, and it makes perfect sense. If this is not the situation anywhere, I'd be very surprised. Amir E. Aharoni (talk) 16:17, 3 October 2024 (UTC)

The Wikidata Infobox on Commons came out of initial work that I was doing here on enwiki - it's the same codebase as {{Infobox person/Wikidata}} and the like, but evolved a lot more due to a lot of constructive input on Commons. It has two views - one for people, one for everything else, and that's been shown to work very well overall, and is a lot more scalable and maintainable than thousands of more specialised templates. Technically, Wikidata infoboxes are quite mature now - and similar ones to this are used a lot on different language Wikipedias. Wikidata is also quite well integrated into different workflows across Wikimedia nowadays (including here with WiR redlists). The main question is a social one: whether the enwp community is interested in pursuing Wikidata infoboxes, and spending the time and energy to constructively contribute to improving data and refining template logic, as the Commons community and other language Wikipedia communities have. Thanks. Mike Peel (talk) 21:00, 1 October 2024 (UTC)

I think that, at this point, we still have too much fear of the unknown and Not invented here syndrome to accept much help from Wikidata. Other wikis will (and do) benefit from it, but we will have to nibble around the edges for another decade.

It's possible that we should be exploring something closer to the bi-directional but manual syncing that the Wikivoyages use (e.g., for official websites and the latitude/longitude of notable locations). If you haven't seen that, then go to your favorite vacation destination at voy: (e.g., voy:en:Paris/7th_arrondissement#See) and scroll down until you find a section that has [add listing] next to the usual section editing buttons. Click that and you'll get a big dialog box. On the right, find the blank for Wikidata and put in a famous landmark (e.g., "Eiffel Tower"; it won't save the edit until you manually tell it to). Click on "quick fetch" to see what Wikidata offers (and then "Cancel" all the way out, so you don't make any changes to the article or to Wikidata). There's also a dialog that lets you choose individual bits (e.g., I want my URL but Wikidata's latitude/longitude, or to remotely replace old information in Wikidata with newer information). Something like this could be done, and could even be set to require sources. Adding sources to Wikidata is usually quite easy, as you just choose the type (e.g., ISBN, DOI, URL...) and add the raw id number/URL directly. WhatamIdoing (talk) 05:49, 2 October 2024 (UTC)

While agree with your assessment of WP decision processes, and I really do think we should ideally be doing more with WD structured data here, WD does not make anything easy (echoing Blueboar above), and frankly (from my attempts to raise this issue with them) does not seem to have any concern with the concept of UX at all. The Commons interface works well, but as soon as it migrates you over to edit Wikidata fields directly, you are forced into their bafflingly perplexing interface, where something that should be as basic as the difference between a "property" and a "reference" (and where to add a citation, which is recommended -- but which is not a "reference") is beneath several minutes of documentation. (The weirdest thing about all of this is that on the WD Discord it seems that people misplacing or not adding data is their #1 maintenance task.)

The project (in its current state) is usable if we can limit as much user interaction as possible to our end, but even Commons has not able been to do this fully. (Note: I feel free to complain about some WD editorship here because I've raised these same issues to WD editors directly many times already, often with poor responses.) SamuelRiv (talk) 15:03, 2 October 2024 (UTC)

The main things I would want to have in place before I could support deeper integration of Wikidata into Wikipedia articles are:

1. an obvious marker in the wikicode for where a parameter is being pulled from Wikidata (like |parameter=:wd or something), with the ability to overwrite it locally just by overwriting it or clearing it
2. clearer error messages that make it obvious to people with no experience in Wikidata which datum the template is missing or unable to understand, which links directly to the property on the Wikidata item that is causing the problem, or if it's not present, one link to the property description and one link to the source item

I've run into a few template errors where the error arose at Wikidata, and in no case was I ever able to resolve these myself. If pro-Wikidata–integration editors here are willing to put in the work to make their templates friendly to Wikipedia editors with moderate and below technical aptitude, then I might consider getting on board, but at present they act like inscrutable black boxes, and their effect in practice is to disempower a large subset of editors in this community. Folly Mox (talk) 12:47, 2 October 2024 (UTC)

I agree with this. I remember the fiascos when {{start date and age}} would scream when improper wiki-data returned multiple dates for just one selector. There should be a way for these templates to error themselves without having the error overwritten by outer, wrapper templates. Aaron Liu (talk) 13:05, 2 October 2024 (UTC)

"Adding sources to Wikidata is usually quite easy, as you just choose the type (e.g., ISBN, DOI, URL...) and add the raw id number/URL directly. "

Random item [2], first unsourced entry: "taxon". "Add reference" opens field "property", which gives a dropdown with 5 possibilities. I want to use a book, so I guess "stated in" would be the right choice. A new dropdown opens, with endless choices, starting with "human"???, "human settlement"???, painting, village, family name... What the bleep is this about??? Perhaps I need to enter "book" and I can continue? Oh, no, it just stops there. Wikidata is extremely non-intuitive and random. Fram (talk) 08:32, 2 October 2024 (UTC)

I want to use a book, so I guess "stated in" would be the right choice.

Yeah, so just enter the name of the book... Expecting that it can guess what book you want to cite is quite unreasonable.
You can also just choose "ISBN-13" instead of "stated in". Aaron Liu (talk) 11:38, 2 October 2024 (UTC)

Just like MediaWiki's own "non-intuitive and random" syntax for putting in a reference, Wikidata also has help pages like wikidata:Help:Sources that explain this exact thing to you. Aaron Liu (talk) 11:43, 2 October 2024 (UTC)

Even if it is possible to learn by reading the manual (I would hope so!) I think their challenge to the characterization of it as simply "easy" was fair enough. Remsense ‥ 论 11:54, 2 October 2024 (UTC)

It's at least as easy to add a reference in Wikidata as it is in Wikipedia. That the methods are not the same does not change this. Thryduulf (talk) 11:58, 2 October 2024 (UTC)

I do not dispute this. Remsense ‥ 论 11:59, 2 October 2024 (UTC)

I do. Fram (talk) 12:21, 2 October 2024 (UTC)

One difference worth noting is that on en.wiki someone can search Help:Citation and find something, while in Wikidata most Help: pages are Wikidata items. There is a guide at Wikidata:Help:Sources, but it only guides towards sources that are either Wikidata items or urls. CMD (talk) 12:31, 2 October 2024 (UTC)

Good point on the borked searching, but there is a "Help" button in the main menu.
You seem to have missed the "Different types of sources" section. Aaron Liu (talk) 12:44, 2 October 2024 (UTC)

Not sure how you come to that odd conclusion. The different types of sources section is a guide to creating Wikidata items, not to sourcing them. CMD (talk) 13:05, 2 October 2024 (UTC)

It guides you on how you can cite a database with an item ID (the database is indeed a Wikidata item, but usually, the item for the database you use has already been created), a headstone, Wikisource, and archive records. Aaron Liu (talk) 13:11, 2 October 2024 (UTC)

This threaded conversation is about the assertion of whether adding a Wikidata reference is "usually quite easy", in the context of comparisons to en.wiki. If I stated it was easy to reference something in en.wiki, you just have to create a new article for it, I suspect people would not find that a convincing argument or helpful contribution. CMD (talk) 13:16, 2 October 2024 (UTC)

New items on Wikidata are not articles. Their difficulty is far from that of new articles. That said, auto-generated reference items from e.g. a URL would help a lot. Aaron Liu (talk) 16:21, 2 October 2024 (UTC)

I think you underestimate the difficulty people may have in creating Wikidata items. CMD (talk) 03:02, 3 October 2024 (UTC)

It's genuinely easy. While the "Create a new item" button in the main menu isn't that noticeable, the steps from there are nearly as easy as filling out {{cite book}}. Aaron Liu (talk) 17:12, 3 October 2024 (UTC)

Having created items it is not quite as easy as the nice en.wiki citation creator, but aside from that I don't see what the value is in ignoring the multiple people who have stated they find it difficult to parse Wikidata. CMD (talk) 23:28, 3 October 2024 (UTC)

Auto-generated reference items from URLs are planned for Wikidata, and since Citoid will be used to generate them, they'll be just as erroneous and silly as they are here. Folly Mox (talk) 16:14, 3 October 2024 (UTC)

And that would mean wiki-data included on Wikipedia would be at the same level of quality as the wiki. Aaron Liu (talk) 17:13, 3 October 2024 (UTC)

That doesn't follow at all. Wikidata with a reference generated by the tool would have similar references as those generated by that tool if the editors involved are the same, and if the scrutiny afterwards is the same. But since Wikidata has for example poor bots creating articles or adding problematic references, and doesn't have the same level of vandalism control as Wikipedia (even though here as well too much slips through the cracks), the end result is that Wikidata quality is too often way below Wikipedia quality. Fram (talk) 18:07, 3 October 2024 (UTC)

Items included on wikis will be scrutinized as well as the main wiki is, and I'm not sure what problematic bot contributions you're referring to. Aaron Liu (talk) 20:12, 3 October 2024 (UTC)

As an example, CJMbot[3]. Latest edits include creating an item on a very obscure 17th c. author[4], and then a week later readding the same items and references to it. Another bot then comes along and removes the duplicate items, but adds the duplicate references to the earlier items. Good going... I noticed this bot because at John Cage it added a second date of birth[5] sourced to Museum Dhondt-Dhaenens. Completely unverifiable, the museum exists but what is meant? Some database, a catalogue, information inside the museum? The result is that e.g. the infobox at Commons now since nearly two years shows the correct and an incorrect birthdate. Fram (talk) 07:56, 4 October 2024 (UTC)

It looks as if that bot has wrecked countless Wikidata items, many subtly, some extremely blatant. Some concerns were raiesd on its talk page, but despite reassurances that the bot owner would look into it, nothing was done apparently, and no one noticed the massive amount of disruption. Which yet again illustrates to me that we should use Wikidata as sparsely as possible, and not trust it to be good enough to add content to enwiki. See [6], a chemical compound that thanks to this bot is also a human with a birth date and so on. Not a one-off, see [7][8][9][10]... Fram (talk) 10:12, 4 October 2024 (UTC)

Okay, that is problematic. Looking at its request for bot permissions, it takes CSV databases submitted by museums and attempts to match them for people, hence the comments on "this was indeed a problem in the data". There hasn't been any discussions on the bot linking people to chemical compounds yet, somehow. Hopefully this time I start a discussion at a talk page, someone will notice.
However, when such data is included on Wikipedia, I'm sure people will notice whatever incorrect data there is like you did and fix it. The Recent Changes also has a lot of edits that are shown as manually patrolled, so it's not like their vandalism control is that low. Aaron Liu (talk) 13:54, 4 October 2024 (UTC)

Sorry to disappoint you, but yes, their vandalism control really is way, waaaay too low. I'm still waiting for someone to revert even one of the 5 vandal edits I bookmarked more than 2 days ago, and today it took nearly 4 hours for someone to realise that "Succcca ahahhaha " (English description: "TUA MADRE STR")[11] is not the correct English label of The Lord of the Rings. If even such high profile articles and such blatant vandalism take this long to be reverted... (And yes, this means that the Commons infobox showed this for 4 hours as well). Fram (talk) 15:04, 4 October 2024 (UTC)

Have you considered that everybody who saw that is doing the same as you and timing how long it takes for someone else to fix it? Has it occurred to you to fix things rather than passively disrupting the wiki to make a point? 15:59, 4 October 2024 (UTC) Thryduulf (talk) 15:59, 4 October 2024 (UTC)

Yes, it's lacking. I'm saying that it's not too low. Aaron Liu (talk) 16:29, 4 October 2024 (UTC)

"Yeah, so just enter the name of the book": where? after the "stated in" box? "Stated in" comes with a lengthy explanation, which even on my widescreen ends with "in which a claim is ma..." without any possibility of accessing the remainder of that text. What's the point of the dropdown after you pick "stated in" if none of these make any sense here. But okay, I add a book title after "stated in". Oops, I want to include a book which doesn't have a Wikidata entry: impossible (or what else does the pinkish-red box mean?). Oh right, according to your help page, if you want to use a reference which doesn't already exist as an item on Wikidata, you first have to create an item for it. So easy! If you are truly out of luck, it is a reference with different editions, and then you have to create two new Wikidata items before you can use it as a source. Want to use a newspaper article as a reference? First create a Wikidata itam for the newspaper article. On Wikipedia, I take the dropdown for the type of reference I want to add (only showing things I can actually use, not "human"), it shows me what the standard fields are, and I don't need to create other stuff only to be able to source something. Trying to edit Wikidata is just an endless source of frustration which I happily avoid and don't want to inflict on others at all. Still, it is nice to see how the number of human edits at Wikidata is constantly artificially inflated by e.g. claiming that I have made 951 edits at Wikidata instead of a dozen or so. Oh well, I have bookmarked 5 bits of Wikidata vandalism to see how fast they get reverted, has that visit to Wikidata given me some fun after all. Fram (talk) 12:21, 2 October 2024 (UTC)

where? after the "stated in" box?

Yeah, there's a reason the param is called "stated in". That's more intuitive than the book icon in the source editor's toolbar.

which even on my widescreen ends with "in which a claim is ma..." without any possibility of accessing the remainder of that text

It displays completely on my standard 1080p.

What's the point of the dropdown after you pick "stated in" if none of these make any sense here.

That's just like how for the "author-link" parameter, TemplateWizard—the visual way of inserting templates—automatically suggests every article that begins with whatever you typed. However, I do agree that just like the TemplateWizard, Wikidata should not be automatically suggesting things when you haven't typed anything.

you first have to create an item for it. So easy!

It's genuinely easy. While the "Create a new item" button in the main menu isn't that noticeable, the steps from there are as easy as filling out {{cite book}}.
Fair point on having to fill it in twice for the edition item.

it shows me what the standard fields are

Wikidata also does. Every class of objects (denoted by whatever you put in for "instance of") has a recommended set of statements that you should fill in; these will be automatically suggested when you click on the "add statement" button. Aaron Liu (talk) 12:58, 2 October 2024 (UTC)

You must be looking at a completely different Wikidata (and Wikipedia) than I am than. If Visual Editor is as bad as Wikidata, then that's another good reason not to use it. I don't get a "book" icon, I get a nice textual dropdown with "cite book", "cite web", "cite news" and so on. Standard fields: you claim "Wikidata also does.". No, it doesn't. After I have added a book title after "stated in", nothing happens. I don't get e.g. the pages parameter, or "chapter", or anything. I just have to know which ones are expected. Fram (talk) 13:15, 2 October 2024 (UTC)

If learning how to edit Wikipedia can be a bit steep learning how to use Wikidata is like walking into a brick wall. Especially on mobile I find it simply unusable. I genuinely feel it would be easy to use if you had to write SQL commands. That's not an attack on the concept of Wikidata, rather a criticism of its implementation. -- LCU ActivelyDisinterested «@» °∆t° 14:49, 2 October 2024 (UTC)

You can basically SQL CMD (talk) 14:57, 2 October 2024 (UTC)

But not exactly user friendly. -- LCU ActivelyDisinterested «@» °∆t° 15:25, 6 October 2024 (UTC)

These are pretty good points. Aaron Liu (talk) 16:22, 2 October 2024 (UTC)

I believe the Drag'n'drop gadget is default-on, and this 22-second-long video: https://www.youtube.com/watch?v=jP-qJIkjPf0 shows that it takes only a couple of seconds to import a source directly from Wikipedia into the ref field in Wikidata.

Manually adding a ref such as the one I added here takes very few steps:

1. Click the 'edit' button for that item (if you're not already editing the item).
2. Click '+ add reference'.
3. Choose a reference type from the dropdown list. This may require a little advance knowledge (equivalent to knowing which of the many citation templates to use), but it usually suggests something sensible, like "reference URL".
4. Paste the URL (or other ref information) into the box.
5. Click 'publish'.

I'm certain that every person in this discussion is capable of learning how to do this, and I'm pretty sure that most of us would find that it takes mere seconds after we have learned the simple steps in the process. This takes, at the most, four clicks, pasting the source, plus sometimes typing the name of a suitable reference type (if the type you want isn't already showing in the list).

The drag-n-drop approach doesn't even require that much effort. Just scroll to the list of Wikipedias at the end of the Wikidata item and click [ref] for the language you want to copy the ref from. A copy of the Wikipedia article will open. Find the ref you want to re-use, and drag it over. Script-assisted copying and pasting does not sound like a difficult task to me. WhatamIdoing (talk) 04:18, 3 October 2024 (UTC)

So I first add the source to Wikipedia, and then copy it to Wikidata to be able to, er, use it on Wikipedia? How does this make life easier? The video uses Mabery Gelvin Botanical Garden[12] and adds a second ref to the state it is in, Illinois. So if we had an infobox generated from Wikidata, it would have shows that it was located in Illinois, just like it does now. Apart from the fact that the state is no longer mentioned at Wikidata for some reason... It was added in 2016 (with a retrieval date of 2012, not good) and removed more than a year ago. Good thing we don't rely on that site. Considering that none of the 5 examples of vandalism I yesterday recorded have been reverted, it seems vandal fighting on Wikidata is still problematic anyway. Fram (talk) 07:38, 3 October 2024 (UTC)

The tool is for easily importing existing citations. A tool to help create new citations still needs to be made.
Also, the Wikidata item was changed to be in Champaign county, which seems correct. Aaron Liu (talk) 13:57, 3 October 2024 (UTC)

As said by Chipmunkdavis below, the tool isn't even visible to mere mortals, is it a toggle in the preferences? And I didn't claim that Champaign county is incorrect, but that if the infobox was Wikidata-filled, it would have changed from "Illinois" (good, informative for most people, wanted) to "Champaign county" (not informative for most people, certainly without "Illinois"). Or, to be even more exact, it would have removed "Illinois" from the infobox but not even added "Champaign county" in its place, as on Wikidata, Champaign county is not even referenced... USA wouldn't be shown either, as that is "referenced" to enwiki. The only references in the article are 404 errors. It's not an example I deliberately picked, it's one given by WhatamIdoing (though involuntarily I guess), but it is typical of Wikidata, even after 10+ years. Even when one goes to a basic article, say Illinois[13], it has hardly any reliable sources. Fram (talk) 14:57, 3 October 2024 (UTC)

Yes, the gadget is not enabled by default and needs to be toggled.
Naming locations as within their provinces is a style convention independent of data. Infoboxes could easily make two calls to Wikidata for the location. Aaron Liu (talk) 15:29, 3 October 2024 (UTC)

Doesn't seem to be the default. My Wikidata interface (including for Mabery Gelvin Botanical Garden (Q5477670)) has the various wiki links at the bottom of the page, and without [ref] icons. CMD (talk) 09:30, 3 October 2024 (UTC)

Another key difference (in my experience) between WP and WD: on WP it is acknowledged that there is a learning curve and entry barriers to editing, and in virtually every VP conversation the notion of how to mitigate this is brought up; on WD the conversations I've had seem indicative that editors believe that it is completely intuitive at its core -- either they don't believe they can make it any easier for users, or that users' difficulties are their own fault (this is just the impression I've had from conversations -- admittedly I'm very direct about reporting interface problems). Additionally, when asked, WD editors did not seem interested in running user interaction experiments or examining existing UX data.

Again, this is not to dogpile on WD for no reason. I believe in the project's core purpose, but I feel at this point like I gotta yell at any direction to get them to wake up. SamuelRiv (talk) 14:58, 3 October 2024 (UTC)

This problem definitely exists on Wikidata, but it exists on Wikipedia, too. Amir E. Aharoni (talk) 16:06, 3 October 2024 (UTC)

Regarding the work required to update Wikidata: when I last made a change, which was admittedly a long time ago, I had to go down a long rabbit hole creating Wikidata items for each property value I was added to the citation that didn't already exist, which cascaded to creating more Wikidata items for the property values of the initially created items, and so forth. If this is still the case, then I would be a lot more inclined to update Wikidata if there were a tool to help generate the entire tree of cascaded Wikidata items for me to approve for submission. isaacl (talk) 16:35, 2 October 2024 (UTC)

Continuation

While the initial premise of this thread is challenging. I think it's worth regrouping to deliberate specifically on some other concepts and potentialities vis à vis Wikidata integration others have brought up so far. In particular, I think it's generally agreed that a range of features are potentially viable as long as they're bidirectionally transparent—i.e. Wikipedia users do not have to learn how to use Wikidata or leave Wikipedia in order to pull or update information. We're fairly familiar with this partially being the way short descriptions work, I think? Remsense ‥ 论 03:46, 3 October 2024 (UTC)

I believe Wikidata pull en.wiki short descriptions (and other languages?) for its relevant field only when there is no existing one on Wikidata, and similarly en.wiki only shows Wikidata if there is no en.wiki short description (unless that has already been disabled?). Wikidata will also pull coordinates and other items, but I don't know if much of this comes bidirectionally back to en.wiki. CMD (talk) 14:07, 3 October 2024 (UTC)

You can find more about usage of Wikidata on English Wikipedia at Wikipedia:Wikidata. Over 100 Infoboxes make use of Wikidata in some fashion. See Category:Infobox templates using Wikidata.

The other dimension this conversation neglects, is that English Wikipedia editors would have chance to support other language editions, if we maintain interoperability between Wikidata and Wikipedia; with all other language editions of Wikipedia. There are legitimate concerns about interface complexity (for all Wiki projects) and sourcing standards, but let's tackle them instead of carte-blanche rejecting any synergies. ~ 🦝 Shushugah (he/him • talk) 23:24, 3 October 2024 (UTC)

I'm familiar with our infoboxes using Wikidata. The uses I'm aware of though aren't bidirectionally transparent in the way Remsense mentions, as you very much do have to go into Wikidata to edit them. CMD (talk) 23:31, 3 October 2024 (UTC)

I'll probably again be accused of "passively disrupting the wiki to make a point"[14], but for anyone who believes that Wikidata has a well-functioning vandal patrol system, here are the 5 random bits of vandalism I bookmarked last week, surprise, none of them have been corrected. They are a varied bunch, from someone creating an unsourced BLP violating article on a non-notable person[15] to someone randomly vandalising some labels[16] (which also affects e.g. Commons), from the obscure but obvious[17] to an editor vandalizing 4 different articles without any problems[18], and ending with a Wikipedia editor with an enwiki article, who died fighting for Ukraine against Russia, being blatantly vandalized and insulted[19]. Fram (talk) 09:47, 9 October 2024 (UTC)

To sneak in a very brief aside I meant to post before this spawned a subthread, I do think passive disruption is a bit of a tough sell. Folly Mox (talk) 12:49, 9 October 2024 (UTC)

Putting aside the specifics, would semi-protecting all items used on enwp sufficiently assuage vandalism fears? It would've prevented all of those examples above. It's already the case that high-use items are semi-protected, but I don't really know the culture of semi-protection on Wikidata to predict whether a lower threshold is realistic. Seems worth acknowledging as a possible support condition amid the years-long "it's garbage" vs. "it's useful" debates here. — Rhododendrites ^talk \\ 12:04, 9 October 2024 (UTC)

If we transcluded such content onto Wikidata, the selections of content we actually use would be put under the same, quality countervandal lens that has worked for years. Aaron Liu (talk) 12:10, 9 October 2024 (UTC)

Do you mean "from" Wikidata? Otherwise I don't really understand your comment. And no, in the current system this wouldn't work, vandalism on Wikidata that has an impact here is not detected as quickly as vandalism here (in general, we have vandalism that remains for months or years as well). We can include Wikidata changes in e.g. "recent changes", but then we get things like this included (as "Dm Antón Losada Diéguez (Q3393880); 12:37 . . Estevoaei (talk | contribs) (‎Created claim: Property:P1344: Q12390563)", which has no bearing on Enwiki at all. Which is typical, most of the Wikidata changes we see are either interwiki links, English descriptions (which we don't display anyway), or things like this which will never be shown here. Fram (talk) 12:46, 9 October 2024 (UTC)

Yes, I meant from, sorry. You indeed don't get RC patrolling, but I think the biggest reason WD vandalism doesn't get detected is that it's not prominently displayed. Any vandalism that escapes initial RC patrolling would have little chance of detection unless the vandal has hubris (c.f. anyone who has nuked someone's contributions) or people actually read the article, which they do, and so they fix. If we increase the prominence of such data by actually using it, we will also substantially increase the countervandalism. I hardly see any vandalism on wiki-data we already include on articles. Aaron Liu (talk) 17:23, 9 October 2024 (UTC)

At least some of the examples I gave already impacted e.g. Commons, and no one detected it from there either. When the Wikidata descriptions were displayed on enwiki, vandalism of these descriptions was not significantly or rapidly detected on enwiki and reverted on Wikidata. This RfC discussion from 2017 gives many examples of what goes wrong when enwiki relies on Wikidata for its data, and also gives (near the bottom) examples of prominent Wikidata vandalism lasting for hours and impacting a number of enwiki articles, from during that RfC. It happens all the time. Fram (talk) 18:02, 9 October 2024 (UTC)

Commons isn't very prominent.

lasting for hours

So? That's not abnormal for vandalism, even on Wikipedia. The amount of edits nuking reverts means many cases of vandalism probably are still extant. There's this famous boxer I forgot the name of who had his infobox blatantly vandalized with the nickname changed to something like "imbecile"; it only got reverted after four days. Look at special:Diff/1241738467, which shut off the feedback request service for months. Disregarding the easy protection fix, does that mean all RfCs should be run manually? No. Aaron Liu (talk) 18:35, 9 October 2024 (UTC)

I doubt that vandalism moving Romania to Moldavia would last for hours on enwiki. This is not some sneaky vandalism, this is in-your-face vandalism. As for Wikidata vandalism being perpetuated across infoboxes in many languages (and not being spotted or corrected by the people at these languages either), we have now (since more than an hour) an impossible date of birth for Johan Cruijff, one of the best soccer players ever and thus an article of some importance. The Wikidata vandalism is visible in Catalan[20], Asturian[21], Galician[22], "ha"[23], Norwegian[24], Welsh[25], and so on. Using enwiki as a pool of editors to do vandal control on Wikidata (which is what the above proposal would boil down to) doesn't seem to be a productive way forward for enwiki. Fram (talk) 12:35, 14 October 2024 (UTC)

Vandalism moving Romania to Moldavia wouldn't last for hours on Wikidata either. Johan's DOB was reverted 7 minutes after your comment, and thus the wrong birthdate only stood for 1 hour and 15 minutes. And no, that's not because of your comment, as the reverter is a frequent RC patroller. We can see that Wikidata does have countervandalism. As for the wikis not spotting it, don't forget that it's a workday in Europe, and don't forget the boxer. Aaron Liu (talk) 12:56, 14 October 2024 (UTC)

Found the boxer edit I was talking about: Special:Diff/1249398659. Aaron Liu (talk) 13:04, 14 October 2024 (UTC)

If you start denying reality, it's hardly useful to continue this discussion. And no, I don't believe in such coincidences. When I posted the previous 5 bits of vandalism (one week after they occurred), one was deleted 2 hours later[26], and the others reverted a few minutes after my post (and a repeat occurrence only lasted for 3 hours, hurrah I guess[27]?). Meanwhile, an editor makes just one vandalim revert, which just happens to come directly after my post here, but doesn't care about other blatant vandalism like the 18 edits here[28], which again vandalizes Catalan Wikipedia[29] and even Spanish Wikipedia[30]. Less obscure articles? Stromae, the great Belgian singer: preferred image vandalized on Wikidata, so the infobox on e.g. Catalan Wikipedia or Norwegian[31] shows a wrong image for hours (oh right, during working hours, when no one uses Wikipedia...), as does Commons[32]. Never mind that since May, the Norwegian Wikipedia proudly displays at the top of their infobox, thanks to Wikidata: "Gary Lineker Golden Bollocks".[33] Clearly, using Wikidata to propulate your infoboxes is stupid and reckless, and only gives vandals an extra outlet to play around with. Fram (talk) 14:36, 14 October 2024 (UTC)

Alternatively, vandalism transcluded here will be seen and fixed much quicker meaning that for the exact same amount of effort vandal fighters will be fixing vandalism on multiple languages/projects rather than just one. Thryduulf (talk) 14:41, 14 October 2024 (UTC)

To quote myself from right above: "Using enwiki as a pool of editors to do vandal control on Wikidata (which is what the above proposal would boil down to) doesn't seem to be a productive way forward for enwiki." Wikidata is touted as good for multi-wiki data, some languages fell for this claim, and now enwiki should do the vandal control for them? Everyone here is free to go to Wikidata and do vandal patrolling, no one is stopping you. Some would even argue that if someone like Thryduulf knows that vandalism is affecting Wikidata and other languages Wikipedias, and they make no effort to do any patrolling there, they are actually "passively disrupting" the WMF landscape, no, the knowledge of the world. And as explained above by many editors, no, it's not "for the exact same amount of effort". Never mind e.g. the not yet mentioned issue of duplicated adminning effort, with blocks, protection, ... needed on both sides. Fram (talk) 15:13, 14 October 2024 (UTC)

We're just going round in circles by now, so I'm just gonna address your only new claim:

only gives vandals an extra outlet to play around with

No, it removes outlets. Previously, vandals can vandalize an infobox from an article on any edition of Wikipedia. Now, when they vandalize, they have to vandalize all of them at once and get more quickly reverted. Simplified: Previously, vandals have like 80 places to vandalize. Now, they only have like 20. Common sense shows that vandalism seen on the Spanish Wikipedia is reverted quicker than that of the Catalan Wikipedia. Wikidata centralizes information, and thus vandal control can be less duplicated.

Since you've been acrimoniously listing vandalism on Wikidata that reflects onto the Catalan edition, why don't you look at this giant list of vandalism on the Catalan WIkipedia instead? Aaron Liu (talk) 15:21, 14 October 2024 (UTC)

That math doesn't math. If there is a version of an article on 80 Wikipedias and they all draw their infoboxes from Wikidata, then there are still 80+1 places to potentially vandalize, because there's still the rest of the article - this proposal doesn't get rid of that. A vandal who wants to write "poop" at the top of the article on Catalan Wikipedia can now do so at either Catalan Wikipedia or Wikidata, and someone who wants to prevent vandalism from appearing in that article at Catalan Wikipedia needs to monitor both. Nikkimaria (talk) 15:36, 14 October 2024 (UTC)

Ah, so that's what it means. Fair point. But the information in the infobox is arguably (one of?) the most important part of an article, and the math would math there. Aaron Liu (talk) 16:44, 14 October 2024 (UTC)

Not if you allow overrides. Nikkimaria (talk) 16:50, 14 October 2024 (UTC)

A workday in Europe? But it's St Angadrisma's Day today. --Redrose64 🌹 (talk) 14:53, 14 October 2024 (UTC)

They take days off Christian feast days in Europe? Aaron Liu (talk) 15:07, 14 October 2024 (UTC)

Have you been to Cardiff on 1 March, or Dublin on 17 March? --Redrose64 🌹 (talk) 17:36, 14 October 2024 (UTC)

I know there are big saints, but I don't think Angadrisma is one that any country takes time off for. Aaron Liu (talk) 17:47, 14 October 2024 (UTC)

It's possible to think of it as vandal control on the English Wikipedia that also happens to be vandal control on Wikidata and Wikipedia in other languages. Amir E. Aharoni (talk) 13:34, 14 October 2024 (UTC)

There are simple implementations, which may perhaps be needed to be coded at WMF level, that we could implement to monitor cross-wiki vandalism from our watchlists here, which seems to be basically what you're talking about -- any change of a wikidata (or commons etc) transcluded material onto a watchlisted article should result in a notification on that editor's WP watchlist (or similarly implemented tool).

Of course, if for example every citation in an article becomes linked to wikidata (and this is beyond the scope of OP), this is a lot of items to place on a watchlist, but I don't know that it matters since the number of expected changes on WD and Commons items is so small (assuming one can watchlist individual properties and not entire items). But automating the process should not be an issue, and notification cross-wiki is already to some extent doable.

Of course it doesn't address what's brought up previously that WD is f-ing awful to edit, which really is something that can be fixed if editors there acknowledge it. SamuelRiv (talk) 15:50, 14 October 2024 (UTC)

Wikidata is WMDE, not WMF. WhatamIdoing (talk) 23:57, 16 October 2024 (UTC)

Really? There's nothing at d:Wikidata:Introduction saying that, indeed at the bottom of every page is the same little "a WIKIMEDIA project" box that we also have, that if clicked takes you to https://wikimediafoundation.org/ - admittedly, this is not the same as wmf: but it's not at all clear what the distinction is. --Redrose64 🌹 (talk) 07:20, 17 October 2024 (UTC)

The WMF hosts it, but WMDE does the software. See mw:Wikibase: "Wikibase is a free, open-source software made and supported by Wikimedia Deutschland and a large community of contributors." The Wikidata Change Dispatching & Watchlists project (the stuff that puts Wikidata changes in your watchlist) is one of WMDE's projects. WhatamIdoing (talk) 19:36, 17 October 2024 (UTC)

New fonts

I always wanted to see other fonts on Wikipedia (except Comic Sans MS, that is so bad). Try adding other fonts. Electrou (formerly Susbush) (talk) 16:09, 11 October 2024 (UTC)

Add this to your Special:MyPage/common.css:

body {
	font-family: "name of font family";
}

, with the part between the double quotes replaced with the actual font family, of course. Aaron Liu (talk) 16:32, 11 October 2024 (UTC)

Ok. Electrou (formerly Susbush) (talk) 16:54, 11 October 2024 (UTC)

I want Ubuntu font, can I get the code? Electrou (formerly Susbush) (talk) 17:04, 11 October 2024 (UTC)

does it support Ubuntu font? Electrou (formerly Susbush) (talk) 17:13, 11 October 2024 (UTC)

It doesn't work (I tried Ubuntu font) Electrou (formerly Susbush) (talk) 17:17, 11 October 2024 (UTC)

@Electrou, what web browser and operating system are you using? WhatamIdoing (talk) 18:56, 11 October 2024 (UTC)

I'm using the latest version of Google Chrome, and I'm on mobile (I use desktop view when needed) Electrou (formerly Susbush) (talk) 20:30, 11 October 2024 (UTC)

As Chaotic Enby suggests, you can't display Wikipedia in any font that isn't on the device you're using. If you open Google Chrome and go into the settings/preferences, there is an item (on my laptop, it's under "Appearance") that says something like "Customize fonts". It should have a drop-down menu that lists all the fonts. If "Ubuntu" isn't in that list, then your device can't display that font. WhatamIdoing (talk) 20:43, 11 October 2024 (UTC)

If you can't install it on that device, then you can still use fallback fonts! They're mostly useful if the same page can be viewed from different devices with different sets of fonts installed, and display if the preferred font isn't install.

body {
	font-family: "preferred font", "fallback font", "fallback fallback font";
}

For example, my userpage is best viewed in Josefin Sans, but also has a series of fallback fonts as it is not available on every device. Chaotic Enby (talk · contribs) 21:38, 11 October 2024 (UTC)

Theoretically, you can also do @import url('https://fonts.googleapis.com/css2?family=Ubuntu:ital,wght@0,300;0,400;0,500;0,700;1,300;1,400;1,500;1,700&display=swap'); to download the fonts automatically. However, I don't think that'll work, since MediaWiki:Common.css is the top stylesheet and I'm not sure if the user's stylesheet is loaded individually. Aaron Liu (talk) 21:47, 11 October 2024 (UTC)

You can load fonts from other servers such as the Google font server, or the toolforge proxy for the Google font server, should you choose. Note page referrer info will get sent to the font server whenever the font is loaded (typically it will get cached so it will load infrequently), so some people are wary of doing this. Also note the toolforge mirror isn't recommended for general use. isaacl (talk) 22:00, 11 October 2024 (UTC)

User's stylesheet is also loaded individually, precisely so that @import works. – SD0001 (talk) 19:42, 17 October 2024 (UTC)

I have installed the font, but it doesn't display it. Electrou (formerly Susbush) (talk) 07:55, 12 October 2024 (UTC)

Have you installed it on your mobile phone? Chaotic Enby (talk · contribs) 20:36, 11 October 2024 (UTC)

Add 'Contribute' links to main page on iPad app

My last period of major editing was in 2017, and at the time the editing interface in the iPad app was pretty barebones, but now the mobile apps have really nice support for basic types of editing. Especially the iPad app, which has a customized interface that isn't quite the same as either the visual or source editors on web. I actually really wish that interface was supported on the iPhone app, that seems to use a less feature full version of the web interface for editing for some reason, but that's besides the point.

I edit about 50% on an iPad, and at those point the app is a better experience than using iPad safari, which has some persistent issues with editing for me. However, it's hard to find information about how to contribute to wikipedia on the iPad app, or to access things like community dashboards/proposals/task lists.

It'd be nice if there was a section of the main screen on the iPad that contained the same links as the "Contribute" section of the menu on the web, or otherwise encouraged and supported new contributors. It would not only support editors like me who just want to get to those things more easily, it might engage current readers who use the iPad app and thus aren't regularly being reminded about chances to contribute on the main screen like the majority of web users are. The iPad app is good enough now that its a really nice editing experience, and I think could work as the the primary interface for new editors. Alternatively, those links could be added to the main menu on the app, which would solve the convenience issue, but not engage/encourage new editors.

I don't know the procedure for changing what shows up on the main screen of the app, or if it's something the community even has influence over. My apologies if this has been brought up already, or there are reasons I don't know about in terms of why this can't work, or I'm discussing it in the wrong place, I just know that mobile editing is less common and less frequently brought up in discussions about supporting new contributors, so as a long time mobile editor I thought it'd be worth mentioning a pathway to better support and encourage mobile editing. penultimate_supper 🚀 (talk) 14:09, 17 October 2024 (UTC)

Pinging ARamadan-WMF, who has worked with the app teams. WhatamIdoing (talk) 19:38, 17 October 2024 (UTC)

Hello @Penultimate supper,

This is Amal Ramadan, Sr. Movement Communications Specialist with the Apps team at the Wikimedia Foundation. Thank you for your thoughtful suggestion and for sharing your editing experience with us. It’s great to hear that you find the iPad app a valuable tool for contributing to Wikipedia!

We appreciate your feedback about adding a 'Contribute' section to the iPad app's main screen or menu. Your suggestion aligns with our goal of improving user engagement and supporting new contributors. We're already working on a navigation refresh for the iOS app, and you can follow the project’s progress here:

https://www.mediawiki.org/wiki/Wikimedia_Apps/Team/iOS/Navigation_Refresh

For iPad-specific improvements, we also have an epic ticket focused on enhancing the iPad experience. You can stay updated through this link:

https://phabricator.wikimedia.org/T377283

I’ve added your suggestions to the ticket to ensure the team considers them during the iPad enhancements.

In the meantime, if you encounter any further issues or have additional suggestions, feel free to reach out through the support email within the app or ping me at any time.

-------------------------------------

Thank you, @WhatamIdoing, for the mention! ARamadan-WMF (talk) 17:16, 18 October 2024 (UTC)

Thank you! And thank you for your work, the apps are such great experiences, honestly my favorite reading experience of the encyclopedia, and a really solid way to edit. Appreciate all the work you do! penultimate_supper 🚀 (talk) 17:32, 18 October 2024 (UTC)

Limited Blackout of Wikipedia in the UK

Overview: This proposal suggests a partial blackout of Wikipedia for users accessing the site from the UK, implemented as a temporary measure to highlight concerns regarding the Online Safety Act 2023. The act requires services to adopt verification or estimation technologies and stricter measures for managing harmful but legal content, especially content accessible to children, with full enforcement expected by OFCOM in 2026.

Why Consider a Blackout? Wikipedia's first pillar is that it is an online encyclopedia with an educational mission. As such, a blackout, even when intended to educate the public about potentially harmful legislation, should only be used in the most serious of circumstances. It must be executed thoughtfully to align with our mission and to avoid damaging our credibility and reputation as a reliable educational resource.

Details of the Proposed Blackout:

• Nag Page Implementation: The blackout will not be a complete shutdown but instead will present a "nag page" when UK users attempt to access Wikipedia. This nag page will:
  • Display information about the Online Safety Act 2023 and its potential implications for free access to information.
  • Collect user information (such as location) to show the scale of concern, but importantly, this data will not be processed or used further, respecting user privacy.
  • Provide an option for users to continue to the encyclopedia content after viewing the information, ensuring they still have access if desired.
• No Set Date: This proposal does not yet suggest a fixed implementation date. It aims to gather community feedback and observe OFCOM's actions as it defines the specific requirements of the act leading up to 2026.

Rationale: Wikipedia’s mission to provide free, reliable information could be significantly impacted if stringent and potentially restrictive measures are enforced without careful consideration. By adopting a limited blackout (nag page) approach, we raise public awareness while ensuring minimal disruption to access. Such a move would emphasize our role as an advocate for open knowledge and user rights while balancing our educational objectives.

Consensus Requirement: Given the potential impact on Wikipedia’s reputation and mission, this proposal seeks to engage the widest consensus possible before any action is taken. A blackout must remain an exceptional tool reserved for critical issues to ensure that it retains its effectiveness and credibility.

---

Comments and Feedback Welcome: Please share your thoughts, ideas, and suggestions to refine this proposal. We aim to ensure that any action taken is well-considered, aligns with our educational mission, and maximizes the positive impact for both Wikipedia and its users.

The true elf (talk) 22:29, 17 October 2024 (UTC)

With all due respect, what do you mean "our"? This is your first edit. If this is an alternate account, sock puppets are not allowed to participate in project space, so if you have a primary account, please log into that instead. Seraphimblade ^{Talk to me} 23:27, 17 October 2024 (UTC)

yeah this is my first edit I do kind of mean hour my bad. Even though I haven't edited anything directly I have made suggestions under

Furthermore, I have made edits to Wiki data, and I have a Wikipedia article that I am sorry about. I am working on it offline and will upload it at some point. 2A00:23C6:C780:4901:D52A:AFC2:3B4C:1F47 (talk) 05:42, 18 October 2024 (UTC)

WP:PROJSOCK is not quite an absolute, bright-line rule per the 2021 RFC. But do please log back in. WhatamIdoing (talk) 02:10, 19 October 2024 (UTC)

Oh, never mind—you wrote this with an AI, now that I read it again. Own words please, not a chatbot. Seraphimblade ^{Talk to me} 23:29, 17 October 2024 (UTC)

sorry about these of AI it was just using as a helpful tool to get my ideas together and formatted well I've got this severe dyslexia and it's a lifesaver sometimes. However I have read for everything and I stand by what it says I wouldn't be editing Wikipedia articles with using AI like this 2A00:23C6:C780:4901:D52A:AFC2:3B4C:1F47 (talk) 05:53, 18 October 2024 (UTC)

Well, don't edit Wikipedia like that, especially if you're making a major proposal. If you have a disability, just say so, like you did. No one would ever fault you for that, but we'd want to hear what you, not a bot, have to say. Seraphimblade ^{Talk to me} 08:02, 18 October 2024 (UTC)

I agree with the sentiment behind that remark, but the problem is that there are some people around who would fault people who have a disability, even though there is nothing they can do about it. Such people may be a small minority, but most disabled people have come into contact with some, especially on the Internet. That makes us a bit wary. Oh, and I'm totally against all blackouts - nobody should presume to speak for everyone else. Phil Bridger (talk) 16:37, 18 October 2024 (UTC)

This doesn't sound like an actual blackout. This sounds like a request for a (large) sitebanner.

(I agree with you that disclosing disabilities can trigger complaints. It can also trigger helpful responses.) WhatamIdoing (talk) 02:13, 19 October 2024 (UTC)

Warning when transcluding images of a certain size or higher

Since this happened to me on accident yesterday, why not bring it here. When you add an image to an article (specifically using source code), you may be unaware of how large that image will be when you click "publish". That's why I'm suggesting that when adding images of a certain size (let's say... 7,000 x 7,000 px or higher), before you are able to publish it will give you a warning that you are uploading images of a size that may not load properly on certain devices, so that you can maybe go back and either change the image, or something else. This diff is a really good example of why this may be needed. SirMemeGod  15:59, 18 October 2024 (UTC)

I can see why that particular image might cause a problem, but how many others would? Do we have figures for how many images would trigger such a warning? These are genuine, not rhetorical, questions. Phil Bridger (talk) 16:20, 18 October 2024 (UTC)

I'm not sure, but I could assume that any image over a certain size could start crashing browsers, like the one I linked. I just need to find out what that size is. SirMemeGod  16:26, 18 October 2024 (UTC)

Pretty much any image that isn't in thumb or frameless has no value. Aaron Liu (talk) 16:29, 18 October 2024 (UTC)

@Sir MemeGod: I don't see why you would want to put the full-size image into an article. You should use the |thumb parameter, and if people want to see a bigger version, they click the thumb image to reach the file description page, which not only shows a larger version, it also offers a selection of sizes to view. --Redrose64 🌹 (talk) 18:07, 18 October 2024 (UTC)

A warning can be added to the commons page, as shown on File:Van Gogh - Starry Night - Google Art Project.jpg. CMD (talk) 18:14, 18 October 2024 (UTC)

I started an edit filter request. Aaron Liu (talk) 04:12, 19 October 2024 (UTC)

Thank you, that is exactly what I was talking about, I probably just worded it badly. SirMemeGod  23:45, 19 October 2024 (UTC)

Wikipedia has a problem with inactive WikiProjects

Hundreds of the over 2000 WikiProjects present on Wikipedia are either inactive or semi-active. Many WikiProjects overlap in scope, or are so niche that very few users have interest in participating. Even for WikiProjects that cast a wide net, some are inactive or semi-active. Historically speaking, WikiProjects were more active when Wikipedia was much younger, and we've seen more WikiProjects become inactive or semi-active.

Why is this a problem?

Having inactive/semi-active WikiProjects that overlap in scope or are incredibly niche causes confusion for editors who seek to collaborate on a topic, or seek help from others familiar with the topic. Additionally, inactive WikiProjects take up valuable shortcut real estate, shortcuts which could be used for other WikiProjects or policies.

An idea for a solution

This is where I'd like the community's feedback. It is unclear whether the outright deletion of WikiProjects is ideal, or if smaller WikiProjects should be merged into larger ones. It's also unclear if a singular solution can be applied to all cases of inactive WikiProjects.

As a personal example, there exists WP:WikiProject Vancouver, WP:WikiProject British Columbia, and WP:WikiProject Canada. Project Vancouver and Project British Columbia are both relatively inactive, with Project Vancouver being relatively narrow in scope. In this case, I believe that merging WikiProject Vancouver into WikiProject British Columbia is ideal, making it a subproject or task force. — Your local Sink Cat (The Sink). 20:09, 20 October 2024 (UTC)

Merging into subprojects/task forces could definitely be an idea – that way, we can keep the established structure while consolidating it into something easier to follow. Regarding members of the merged WikiProjects, should they automatically be added to the wider one, or just given an alert so they can choose whether to join it or not? The second option seems more natural to me.

There is also the broader question of why WikiProject activity is declining. I am far from a specialist on this topic, and wonder if specific things changed that made joining WikiProjects less practical or less useful than it was before? Chaotic Enby (talk · contribs) 20:22, 20 October 2024 (UTC)

Well the proposed parent project participants should decide whether to merge in the niche projects, if the project is completely inactive. At the top of the moribund project page, it could refer people to the relevant active project. And talk page project tagging for dead projects could be removed, or replaced by the parent project. I would say just alert former participants that are not part of the parent already. Graeme Bartlett (talk) 20:26, 20 October 2024 (UTC)

People who are active on pages don't own them. It's a courtesy to invite them to the discussion, but I don't believe they should get any more say than anyone else. Thebiguglyalien (talk) 20:32, 20 October 2024 (UTC)

The inactive projects are usually marked as historical and allowed to stays in place. Ruslik_Zero 20:39, 20 October 2024 (UTC)

A WP:WikiProject is not pages. A WikiProject is people. You can no more vote to make editors join a different group than you could tell your classmates who they were required to be friends with when you were all 10 years old.

The reason merges have to be voluntary is because that's the only way they work. We can merge, redirect, and move pages all day long, but if the individual participants don't voluntarily choose to participate in the new "larger" group, then we won't actually have a larger group: we'll have a bunch of people quitting.

I heard a joke about Church splitting (I didn't expect that to be a red link) many years ago that went like this: The church had a fight over whether the new hymnal books were to have a red cover or a blue cover. In the end, 10% of them split off to form a church with red books, 10% formed a church with blue books, and 80% of them got so disgusted that they quit going to church. Maybe this only seems like a plausible outcome if you're from the South, but we don't want any WikiProject participants to get so disgusted that they quit Wikipedia, or even just quit watching the WikiProject pages.

Consequently, we've established a procedure for voluntary merges. Basically, you make an offer and wait a month to see whether anyone objects. If anyone objects, then stop. If they don't, then merge. I'd guess that we've had objections about 10% of the time, so it rarely interferes with the ultimate outcome. WhatamIdoing (talk) 03:44, 21 October 2024 (UTC)

I think the reason why there are so many inactive Wikiprojects around is simply that nobody has bothered to do anything about them. I certainly, and probably most other editors, see that as a pretty low priority. Phil Bridger (talk) 20:50, 20 October 2024 (UTC)

I've merged a few projects I participated in due to scope overlap, including two recently, it's not even that hard to get consensus, but no one wants to bother. If there's a specific project that bothers you just propose it. PARAKANYAA (talk) 21:46, 20 October 2024 (UTC)

Merging WikiProjects comes up at the WikiProject Council talk page – the most recent discussion can be seen at Wikipedia talk:WikiProject Council/Archive 26 § Systematically merging WikiProjects. Previously. the only set of instructions was for converting a WikiProject into a task force, which might be preserving more infrastructure than necessary if the WikiProject is inactive. WhatamIdoing started Wikipedia:WikiProject Council/Guide/Merging WikiProjects as a first pass at instructions for merging a WikiProject without making it a task force. For better or worse, there is often resistance to implementing merges, as can be seen at the end of the archived discussion to which I linked. (I gave some of my thoughts on why WikiProjects might not be attracting participants in that discussion.) isaacl (talk) 21:55, 20 October 2024 (UTC)

I should get back to that page. @Joe Roe was working on a merge, and we need more template information. The TFD folks have suggested that they'll be generous with their template-replacing bots, but we need to write out a procedure for it, so people know what to do.

@Sink Cat, please consider putting Wikipedia talk:WikiProject Council on your watchlist. That's where a lot of this gets discussed. In terms of the general concept, I think we should aim for a world in which there are many fewer, but much larger (on average) active WikiProjects, but the whole thing happens one baby step at a time. The biggest challenge is that nobody is doing the work. Just reviewing 2,000 groups to produce a recommendation about how to merge them could be a full-time job for a month, and then each group has to be contacted with personalized messages to explain the idea, name the targets, and ask if they're willing. If someone wants to spend, say, 20 hours/week for the next year doing this, then that would be great, but that's the scale we're talking about. WhatamIdoing (talk) 03:55, 21 October 2024 (UTC)

Thanks for the response! I'll get in contact with them to see what can be done. — Your local Sink Cat (The Sink). 04:11, 21 October 2024 (UTC)

Not sure to whom you are referring when you say "them"? Note the word "council" is a bit misleading: it's just essentially the WikiProject wikiproject. isaacl (talk) 05:12, 21 October 2024 (UTC)

Echoing this – it's possible to get consensus for merges but it's really time consuming and slow. I know I've got a few outstanding because you have to wait 30 days after opening a discussion to action it (sensibly enough) and I'll often find that 30 days later I don't really have time to do it. – Joe (talk) 08:22, 21 October 2024 (UTC)

filter for removing closing markdown of comments?

By chance while doing some other work on pre 2006 pages, I noticed Bernard Lee, which, due to an IP editor, looked like a stub, despite being listed as a GA. I'm not really sure what the intention of their edit was but they removed part of the comment markdown, causing most of the article to be hidden. Is there some way of tagging instances of this happening via Edit filter or such? I could imagine someone maybe more malicious intentionally doing this that would otherwise probably just blank the page or remove large parts, which is already more or less mitigated by current measures. Akaibu (talk) 22:04, 20 October 2024 (UTC)

I think there are two separate things here - an edit filter to detect new edits that result in an unclosed HTML comment. I don't know whether one exists, but Wikipedia:Edit filter/Requested is the place to request this if it doesn't.

The second thing is to check for existing pages with unclosed comments, which sounds like the sort of thing a Wikipedia:Database report would be useful for, Wikipedia talk:Database reports is the place to request one.

In both cases a correctly closed comment after an unclosed one could cause false negatives, but I guess detecting <!-- […] <!-- […] --> would catch those? Thryduulf (talk) 22:30, 20 October 2024 (UTC)

It may also be worthwhile to detect very big comments. But there are likely many of these deliberately inserted, so Thryduulf's idea sounds better. Graeme Bartlett (talk) 09:52, 21 October 2024 (UTC)

Related topic in phab:T304222. — xaosflux ^Talk 15:06, 21 October 2024 (UTC)

By the way, Markdown is a formatting language that we do not use. We use Wikitext. Aaron Liu (talk) 16:30, 21 October 2024 (UTC)

Ai chat bot

I have an idea of building an open source ai model trained on wikipedia available free of charge on wikipedia platform as new generations have less attention span and need to get things done fast. I want it open source so that it can be for the community and get inspired by contributors Yassinsamirhegazy (talk) 13:22, 21 October 2024 (UTC)

Hi, the contents of Wikipedia are CC BY-SA 4.0, so you can use the corpus to train an AI model if you wish to, assuming you release it under a compatible license. Regarding having it available on Wikipedia, this might not necessarily be a good idea, as AI models are prone to hallucinations, and a model trained on Wikipedia contents at a given time will not be editable like the encyclopedia itself. Chaotic Enby (talk · contribs) 13:48, 21 October 2024 (UTC)

I understand your point but this ai model can be just a start to cope up with the new generation of audience 41.46.109.7 (talk) 14:06, 21 October 2024 (UTC)

Have you considered that all of the major chatbot LLMs have already scraped Wikipedia into their corpus? Remsense ‥ 论 14:38, 21 October 2024 (UTC)

Please be aware that many Wikipedians more or less distrust AI, and some have been advocating for a ban on any use of AI in Wikipedia. While I consider a ban to be unenforceable, I would want editors to avoid using any AI without very carefully checking and correcting all output from the AI, which might well be more work than just writing content without using an AI. Donald Albury 14:44, 21 October 2024 (UTC)

Yeah, i've considered that but I want to add this feature some that people spend more time on wikipedia Yassinsamirhegazy (talk) 17:23, 21 October 2024 (UTC)

We want usefulness to readers, not engagement. Aaron Liu (talk) 17:31, 21 October 2024 (UTC)

ChatGPT, Ollama, and pretty much every major platform already did that. Aaron Liu (talk) 14:44, 21 October 2024 (UTC)

<- For interest, FutureHouse's PaperQA2 model is already matching or outperforming subject matter experts in article creation summarizing scientific topics in a very specific domain. See Language Agents Achieve Superhuman Synthesis of Scientific Knowledge. Sean.hoyland (talk) 14:53, 21 October 2024 (UTC)

Also, people (with less attention span) can already interact with Wikipedia content using the labs.google NotebookLM experiment. Sean.hoyland (talk) 15:10, 21 October 2024 (UTC)

Consensus Required Restriction and NPOV tags

I know that editors should seek to remain unbiased, but it seems on divisive topics we can end up with one side who manages to tilt the article toward their POV. We can then end up with half of the editors saying "This article is perfectly fine" and the other half of the editors saying "There are big POV issues, here they are..."

The side who are happy with the bias can actively work to prevent any fixes to the page to address the bias, while simultaneously blocking the addition of a NPOV tag to the page.

It seems that if half of the editors are saying "it's fine" and the other half are saying "there are big issues" this is extremely indicative of a POV problem even if there is not consensus that one exists.

So I'm wondering if there should be an exemption to the Consensus Required Restriction and if some sort of critical mass short of consensus should be enough to allow for NPOV tag.

Thoughts? Bob drobbs (talk) 18:48, 22 October 2024 (UTC)

Putting the {{POV}} should never be an end goal. The goal should be to fix the problem, and adding the banner frequently provides no practical benefit at all. Imagine a wildly non-neutral article about an article under the Wikipedia:Consensus required restriction. Maybe it's an article called 2+2. The article says that 2+2 is generally understood to equal four, but there is a significant minority of respectable mathematicians says that 2+2=5. Here's the story you seem to want:

• Alice adds a {{POV}} tag.
• Eve removes it because she disagrees.
• The discussion on the talk page about the tag ends in a stalemate. Because of the 'consensus required' rules, the tag would normally not be added, but because of the newly carved-out exception, the tag can be added.
• End result: The article is tagged, but it's still wildly unbalanced.

Here's the story we need:

• Alice adds a {{POV}} tag.
• Eve removes it because she disagrees.
• The discussion on the talk page about the tag ends in a stalemate. Because of the 'consensus required' rules, the tag is not added.
• Alice decides to quit worrying about the tag and start worrying about the content of the article. The regulars on the talk page can't reach a satisfactory agreement, so she takes the dispute to a relevant noticeboard or starts an RFC.

Remember: Maintenance tags are not badges of shame. They do not exist to 'warn the reader' or to formally express your disagreement with the article. They exist in a (mostly vain) hope that editors will fix the article's content. If you can fix the article's content without a tag, then everyone wins. If you can't fix an article (tagged or otherwise), then read up on Wikipedia:Dispute resolution. WhatamIdoing (talk) 19:05, 22 October 2024 (UTC)

Depends… how many are on each “side” of the debate?

If there are multiple editors on each “side”, then I don’t think we can say that a consensus actually exists (in either direction). However, if it’s just one or two disgruntled editors against many, then we can say there is consensus, and that consensus does not require unanimity. Blueboar (talk) 19:08, 22 October 2024 (UTC)

WP:BITE rewrite 2

Previous discussions

• Wikipedia:Village pump (idea lab)/Archive 60 § Rewriting WP:BITE
• Wikipedia talk:Please do not bite the newcomers § RfC: Is this rewrite ready to replace the current page?, which only got 8 responses in 3 months, with 4-4 deadlock. In particular, two editors have expressed concern that the rewrite was not persuasive enough, which I attempted to mitigate by expanding the lead with more emotionally-driven prose. I have therefore withdrawn the RfC. There seems to be a consensus that an rewrite is needed, however. While many of the suggestions provided on the RfC and the previous idea lab were added to the rewrite, I am thinking of advertising the next RfC on WP:Centralised discussion, so I would like further feedback for the proposed rewrite: Wikipedia:Please do not bite the newcomers/rewrite.

How do you feel about this updated rewrite? Ca ^{talk to me!} 14:27, 25 October 2024 (UTC)

@Ca, without looking at your rewrite, I often feel like baby steps are the best way to get changes made. WhatamIdoing (talk) 16:37, 25 October 2024 (UTC)

Auto-amending bare URL tags

Is there any reason we aren't or can't scrape and extract basic cite template information (webpage title, domain, visited datetime) at submit-time from bare URL <ref>s? Personally, I use bare URLs all the time as I consider filling out the cite template the most emphatically tedious part of editing WP (as it is when writing scholarly manuscripts) and know that some robot editor will just come along and fix it for me anyway. Since the code already exists and could be done more efficiently server-side, why don't we just pull it into the WP core? Just a small and probably extant idea I had while feeling a little guilty for adding a bare <ref> to a nicely-formatted article with proper tag attributes and template use. Cheers, fellows. Elliott-AtomicInfinity (talk) 01:38, 24 October 2024 (UTC)

This is what mw:Edit check is getting to AFAIK. Aaron Liu (talk) 02:18, 24 October 2024 (UTC)

@Trizek (WMF) will know if they're going beyond "prompt to add a ref" to "try to format the ref". WhatamIdoing (talk) 20:03, 24 October 2024 (UTC)

Not quite what you are looking for, but there is reFill for bare URLs, and Wikipedia:Citation expander for when the URL is within <ref> ... /ref>. You can save your edit, and then immediately run those tools. As always, the output of any such tool must be reviewed and verified. Donald Albury 14:32, 24 October 2024 (UTC)

You might be looking for this.

Or use the visual editor, or use the ref filler in the 2010 wikitext editor's toolbar. Or maybe even embrace the idea that everyone contributes in different ways, and that WP:CITE means what it says about doing your best to accurately communicate what your source is, and that editors can, do, and should work together on the formatting. WhatamIdoing (talk) 20:01, 24 October 2024 (UTC)

It took me far to long to realise you can auto-fill citations from the toolbar, I fear I'm not the only one as it's not well advertised. -- LCU ActivelyDisinterested «@» °∆t° 20:12, 24 October 2024 (UTC)

Nice, but it doesn't always work. It is particularly annoying when I'm trying to cite something that I have accessed through the Wikilibrary (with a long list of authors) and it doesn't work, and I am unable to go to the non-Wikilibrary page for the article. Donald Albury 20:26, 24 October 2024 (UTC)

It is indeed unfortunate that many websites do not properly mark up their pages so that the automated tools and Zotero etc are unable to extract the appropriate information from webpages. But that is a problem that we cannot really solve. —TheDJ (talk • contribs) 20:48, 24 October 2024 (UTC)

Recently (1-2 months back?) I saw a project to improve the automated tools (or maybe just one tool), iirc by adding local code for specific commonly cited websites it consistently gets wrong. Unfortunately I can't find the discussion now, but if someone remembers it (user:WhatamIdoing?) you may be interested. Thryduulf (talk) 22:48, 24 October 2024 (UTC)

Thryduulf, are you thinking of Wikipedia:Village pump (proposals)/Archive 213 § Deploying Edit Check on this wiki (August 2024)? Folly Mox (talk) 23:35, 24 October 2024 (UTC)

No, this was about better (more complete, more correct) automated filling of source metadata (author, publisher, title, etc) when a reference is added. Thryduulf (talk) 00:01, 25 October 2024 (UTC)

m:Web2cit? * Pppery * _{it has begun...} 00:41, 25 October 2024 (UTC)

I think that was the tool, but I'm sure the discussion was on en rather than meta (the canonical capitalisation btw is Web2Cit). Thryduulf (talk) 01:25, 25 October 2024 (UTC)

I really need to be better at checking my links. * Pppery * _{it has begun...} 03:09, 25 October 2024 (UTC)

The other main enwiki tool that formats citations is User:Citation bot. * Pppery * _{it has begun...} 03:09, 25 October 2024 (UTC)

What about Wikipedia talk:Citing sources § citation generator? (September 2024)? Folly Mox (talk) 11:51, 25 October 2024 (UTC)

Oh right, and the "Edit Check" thread I guessed first had a tangential subthread about better metadata, including adding a local translation layer to Citoid as invoked from the Visual Editor interface, but no one gave the subthread a heading for me to link. I think it starts around here. Folly Mox (talk) 11:57, 25 October 2024 (UTC)

It wasn't the citing sources discussion. Might have been the subthread but I'm not certain. Thryduulf (talk) 12:24, 25 October 2024 (UTC)

Or they do mark them up, but they block our servers. That's probably happening with The New York Times. Properly formatted refs are in their interest, too, but it's likely that all they see is some automated thing or another and automatically block it. WhatamIdoing (talk) 23:41, 24 October 2024 (UTC)

Server side automation would have the same issues, and I'd rather editors checked what the automated tools outputted as they sometimes produce nonsense. -- LCU ActivelyDisinterested «@» °∆t° 20:50, 24 October 2024 (UTC)

@Donald Albury, I've had some success with using a DOI in such cases, assuming it has one. WhatamIdoing (talk) 23:42, 24 October 2024 (UTC)

I think I've also had the problem using DOI. It looks like WikiLibrary sets cookies on my device (when I go back to a source a day or two after I looked it up with WL, WL immediately jumps in. I saw this happening even when I was logged in with my alternate account, which is not eligible for WL.), so any attempt to reach a link outside of WL gets diverted back to WL. Donald Albury 18:15, 25 October 2024 (UTC)

Author Profile Page

One should exist for Authors to personally fill out a profile. Many book readers would like to know the "about" information about writers that have books published. 91.242.149.121 (talk) 09:15, 29 October 2024 (UTC)

If you're referring to the authors/editors of an article, editors already have their own user pages, which are accessible from an article edit history. If you're referring to the authors of books that merit Wikipedia articles, this isn't a place for people to tell about themselves(see WP:AUTO). 331dot (talk) 09:18, 29 October 2024 (UTC)

Fix Draftification with a new template

Draftification has long been criticized as a backdoor to deletion. In New Pages Patrol (NPP), it is common to move new articles that are not ready for mainspace to draftspace. This way, articles that could potentially be suitable for Wikipedia, but are not yet, are preserved. The article creator then gets a chance to improve their article without NPPers breathing down their necks or having it taken to Articles for Deletion. If anyone, including the article creator, objects to draftification, the article should be moved back to mainspace (draftification should be reversed). This is explained by DRAFTNO #6 and #7. No reason is required for the objection.

Problem: However, we also have a rule that drafts that haven't been edited for six months get automatically deleted, under Criterion for Speedy Deletion G13. So, well-meaning New Page Patrollers will unilaterally draftify new articles that are not yet ready for the encyclopedia. The new editors who created the article may disagree with the move, without knowing that they can object. The new users can get discouraged and leave Wikipedia altogether, and after six months the draft is deleted under CSD G13. As this process happens without community discussions, it results in draftification being called a "backdoor to deletion".

Solution: This problem can be solved without changing policy or current practice. We just need to make it very obvious to new users that they can object to draftification. We can also make it easy to reverse the draftication (assuming the new user is autoconfirmed). I suggest we do this by adding a template to all draftified articles. The template would include a big blue button, similar to the "Submit the draft for review!" button at Template:AfC submission/draft, which says "Object to this move". Clicking this button either: 1. Leaves a message on the talk page of the editor who draftified, notifying them that there has been an objection to the move and requesting that it be immediately reversed. 2. Moves the page back to mainspace automatically, or if the editor's account is unable to perform this task, creates an entry at Requested moves/Technical moves to that effect. The latter is better, but also more technically complex. Adding a similar button to Template:Uw-articletodraft, the warning typically given upon draftification, would also be helpful.

Implementation: Once the new template is ready, it can be added to MPGuy's MoveToDraft userscript, which is the most common way for NPPers to draftify articles. It should be placed above the AfC template on all draftified articles.

I would appreciate comments from technically skilled editors, who could create this template (or tell me that it's impossible), from NPPers who draftify articles, and from uninvolved editors who have opinions on the draftification process. Toadspike [Talk] 10:37, 26 September 2024 (UTC)

This idea isn't really my own, it was obviously sparked by the most recent RfA. A similar idea was previously discussed here, but that discussion proposed a requirement that all editors have to follow (policy), not a technical solution, and turned into a trainwreck. To prevent something similar, I ask all participants to please focus on improving the current situation instead of debating the morality of draftification as a whole. Toadspike [Talk] 11:03, 26 September 2024 (UTC)

Notifying the users who commented most directly on this topic at the RfA: @Alalch E.@User:Onel5969@User:Hobit@User:Fangz@User:Nsk92. I have also notified the NPP Talk page and posted a message on Discord. I am not sure how to notifying all participants of the previous discussion (aside from doing it manually) and I am not sure that is productive considering how many people were involved and how offtopic it got, so I won't do that for now. Toadspike [Talk] 11:07, 26 September 2024 (UTC)

Are you sure you want to make this an RfC? Is there a BEFORE somewhere? Aaron Liu (talk) 11:32, 26 September 2024 (UTC)

Good point, I am not sure if the RfC label applies, so I'll remove the templates. I was looking for ways to notify people and misread RFCBEFORE. Toadspike [Talk] 11:34, 26 September 2024 (UTC)

• Oppose: The draftification message could be tweaked, but a big button to reverse the move will lead to more AfDs, higher strain on NPP, more BITEY behaviour, and worse editor retention. Draft space is incredibly valuable, and people have some incredibly warped views about the space. If we did something like this then we'd end up chasing away new editors because learning how to make your article meet our complicated guidelines in under 7 days (AfD tag) is not easy for a lot of folks. Draft space gives them the opportunity to work on the content, to receive advise, and to make articles that will actually survive at AfD and allow them to stick around. Really we need to draftify more, and I've taken it upon myself to begin to do so again and encourage others to do. I'm big on editor retention. This is not the way to do it. Hey man im josh (talk) 12:15, 26 September 2024 (UTC)

  The problem with unilateral draftification is that it can also be incredibly bitey, especially when done for arbitrary reasons that have nothing to do with any of the reasons why something might be deleted at AfD (although this is less prevalent than the trivial reasons things are rejected at AfC). We should be draftifying fewer articles and not sending them to AfD either but rather leaving them in the mainspace (With appropriate tags where justified) so that they can be found and improved rather than pretending that they don't exist for six months and then deleting them. Thryduulf (talk) 12:34, 26 September 2024 (UTC)

  I'm not really convinced draftification is any worse than the alternatives - tagging is *also* bitey as well, and one user tagging an article and leaving it in mainspace could lead to another user seeing it and deciding to AfD. Draftification could be a way to protect an article until it enters a better state. But I think the other part I have an issue with is the lack of clear guidelines. Clearly some people have an issue with draftification and others do not, and people have different ideas what it is for. That needs to be made more concrete. Otherwise just saying "we should use draftification less" isn't going to lead to any positive changes. Fangz (talk) 12:43, 26 September 2024 (UTC)

  I agree with the general sentiment – arguing for more or less draftification does not solve the problem that new users basically can't object to it. Toadspike [Talk] 12:47, 26 September 2024 (UTC)

  I envision a template (possibly one specific for relatively new users?) being something like:

  1. Hi, this article has been moved to a draft form because another user thinks it has potential but is not ready for the encyclopedia just yet. REASON:

  2. You can continue to work on it while it's not published, though note that if not editted for 6 months it will be deleted. Here are some useful resources.

  3. When you think the article is ready you can submit the article to a review, which can give useful feedback. []

  4. Alternatively you may return the article to the main encyclopedia at any time and have it be editted while part of the main encyclopedia. See WP: Draft Object. Note however that if other users think there are unfixable issues with the article it may be put forward as a candidate for deletion. Fangz (talk) 12:54, 26 September 2024 (UTC)

  I like the idea for the user warning. Aaron Liu (talk) 12:59, 26 September 2024 (UTC)

  Tagging never leads to an article being automatically deleted. – Joe (talk) 18:43, 26 September 2024 (UTC)

  In my view draftified articles should (semi?) automatically return to the mainspace after timeout instead of be deleted. Or at least be re-evaluated for notability. I do not really see the reason for automatic speedy deletion, except as backdoor deletion. Fangz (talk) 18:59, 26 September 2024 (UTC)

  I like that idea. They don't, though, so it's a bit of a moot point in terms of current policy. – Joe (talk) 06:16, 27 September 2024 (UTC)

  Why can't they just improve it in mainspace, without the sting on of an initial rejection and a six month deletion countdown hanging over them? I don't get why you keep presenting this as a choice between draftspace and AfD. – Joe (talk) 18:46, 26 September 2024 (UTC)

  The reason is that "improving it in mainspace" has its own issues. An article in mainspace has to juggle being of service to the reader to being of service to the editor. This implies formal processes and wikijargon for consistency, unified templates for issues in the article, clear and ruthless labelling of problems and so on. There's a strong tendency for the first experience of an editor to be a very public and humiliating fight against established editors who have a better understanding of wikipedia processes, quickly driving the editor away or getting them blocked. It is also very difficult to improve on this experience as it would imply fundamental changes affecting all sorts of things. Meanwhile improving an article in draft mode allows for a more informal process of communication to shepherd an article towards an acceptable state. Fangz (talk) 19:10, 26 September 2024 (UTC)

  I did a little work on page view statistics recently. The median article gets about one page view per week. So if the new article is typical, then it doesn't have to "be of service to the reader", because there aren't really any readers. Editors (especially NPP and RecentChanges folks) may look at a brand-new article a few dozen times on the first day, but once the reviewers leave it alone, most articles just don't have much traffic.

  I think the reason we are unwilling to "improve it in mainspace" is because we're scared that we'll forget that it was there, and years later, someone will be embarrassed to discover that an WP:UGLY article has been neglected ever since. We are using draftification and other threats as a way to make other WP:VOLUNTEERS improve the article to our idea of acceptable quality. WhatamIdoing (talk) 20:40, 26 September 2024 (UTC)

  I don't know if we're looking at different draft namespaces, but an "informal process of communication to shepherd an article towards an acceptable state" sounds like the precise opposite of our current AfC process. – Joe (talk) 06:19, 27 September 2024 (UTC)

I don't like the idea of a button but I do think the template should be changed. I think having a button suggests it's a default option, but I think a link is okay. Fangz (talk) 12:37, 26 September 2024 (UTC)

• This is pretty much what {{subst:draft article}} does but MoveToDraft uses {{subst:AfC submission/draft}} for whatever reason. C F A 💬 12:41, 26 September 2024 (UTC)

  For the record, I would support using either {{draft article}} or creating a new template with your proposed changes. C F A 💬 12:44, 26 September 2024 (UTC)

  All AfC drafts are supposed to use the AfC template. If you're talking about the "publish now" button, that's for autoconfirmed only. Aaron Liu (talk) 12:45, 26 September 2024 (UTC)

  But articles aren't inherently "AfC drafts" once draftified. People are allowed to move them back to mainspace themselves, no questions asked, or optionally submit them to AfC. The AfC draft template makes no mention of the first option. C F A 💬 12:48, 26 September 2024 (UTC)

  BTW, the MTD script was changed yesterday to use {{draft article}}. -MPGuy2824 (talk) 12:48, 26 September 2024 (UTC)

  Oh, I didn't notice. That's better, at least. C F A 💬 12:49, 26 September 2024 (UTC)

  I wonder if we will ever quit using the inappropriate and legally unsound "publication" language in that template. On this wiki, it's the author, not the AFC reviewer, who "publishes" it. WhatamIdoing (talk) 20:43, 26 September 2024 (UTC)

  @WhatamIdoing: Can you think of a better word? (Seriously, I've been trying for ages and can't come up with anything.) – Joe (talk) 07:33, 27 September 2024 (UTC)

  Yes: WP:MOVE. WhatamIdoing (talk) 20:28, 27 September 2024 (UTC)

  Yeah but then you have to specify "move to mainspace" or something and that feels really jargony. "Move out of draft", maybe? – Joe (talk) 04:49, 1 October 2024 (UTC)

  That would work. WhatamIdoing (talk) 05:19, 1 October 2024 (UTC)

  Well, that might solve the problem. I doesn't explain to new users how their article turned into a draft, but at least there's an easy way to reverse it. I won't withdraw this for now, maybe others have thoughts. (P.S. you might want to update User:MPGuy2824/MoveToDraft) Toadspike [Talk] 12:51, 26 September 2024 (UTC)

• This is the idea lab so no bolded comment from me, but I have mixed feelings. I am in favour of softening the experience for newcomers, but I'm opposed to the concept of draftification being automatically reversible. If a new page patroller reviews a new article and moves it to draft because it's clearly unsuitable for mainspace, the creator should need to do more than just say "I object" in order to move their clearly unsuitable article back again. I've recently proposed that all of draftspace should be move-protected at the semi level (the proposal was not well received - fair enough). This is probably the rule I ignore more than any other on Wikipedia, mostly dealing with spam sockfarms that try to abuse the rule to promote their garbage. Besides, a new user whose submission is quarantined to draft space and they're left with instructions and a list of suggestions with helpful links is already getting better treatment than most editors ever have or will, and if their reaction to that is to rage-quit then they're probably not a good fit for the collaborative environment of Wikipedia anyway. Ivanvector (^Talk/_Edits) 12:57, 26 September 2024 (UTC)

  @Ivanvector, you know the joke about "If you ask three people, you'll get four opinions"? I wonder if we ask three NPPers what "ready for mainspace" means, if we'd get four opinions. AFAICT, "ready for mainspace" most often means "contains at least as many refs as the median article, but higher quality ones". All the children in Lake Wobegon are above average, and all the new Wikipedia articles must be, too. WhatamIdoing (talk) 20:12, 26 September 2024 (UTC)

  I feel like I might vaguely recall a discussion on that topic sometime in the not too distant past. Folly Mox (talk) 22:55, 26 September 2024 (UTC)

  176 comments from 22 editors, and I probably had 22 opinions all by myself. ;-) WhatamIdoing (talk) 22:23, 27 September 2024 (UTC)

  All pages are effectively move-protected at the semi level already. Moving requires an (auto)confirmed account. SilverLocust 💬 07:17, 5 October 2024 (UTC)

• As far as I see it draftification should never be used for subjects which pass GNG, and it should only be standard for things like films/TV series/games which are in the works but have not yet begun production. Subjects with debatable notability should be sent to AFD to the issue can be resolved.★Trekker (talk) 13:00, 26 September 2024 (UTC)

  Subjects that pass WP:GNG should never be draftified at all, instead they should be tagged and dealt with using normal community procedures. I agree that films/TV series/games/political events probably best fit the bill for draftifications, but so do potentially notable but underdeveloped articles. Sohom (talk) 13:33, 26 September 2024 (UTC)

  This is out of step with the present form of Wikipedia:DRAFTIFY, and I don't think it makes sense anyway. Articles that fail GNG should not be draftified, they should be AfDed. Films etc that are in the works should not be draftified merely because they aren't in production, and it's not really a great use for draft space because there's no guarantee that there would be a change of situation to establish notability within 6 months. Articles should be draftified only if the reviewer believes the article can be editted into an acceptable state within the time window. This implies a pass of GNG - i.e. a belief that reliable sources are potentially out there. Remember that GNG is about the *subject*, not about the state of the article. Fangz (talk) 14:09, 26 September 2024 (UTC)

  In my view the correct use of draftification is sort of as an alternative version of the WIP template. An acknowledgement that the article is not ready and should be being worked on and will likely have multiple issues, but in a protected sandboxed environment to avoid overly zealous moderation and promotion of misunderstanding for casual readers, and without implying the original editor is the one working on it. For new users it should offer a less formal and jargony process to learning how to improve an article than tagging based methods, because the latter has to balance the need to inform *readers* as well as editors. Fangz (talk) 14:23, 26 September 2024 (UTC)

  If you evaluate that a article passes WP:GNG, then there is not point in draftifying it, you could just add a {{sources exist}} template, patrol and move on. Alternatively, if you evaluate that a article fails WP:GNG, there is no point in wasting the article creator's time and you should WP:AFD/PROD it.

  The only case where you would draftify a article is if you saw a article that a) had a credible claim to significance/notability b) does not meet/prove notability in it's current state c) has been created in the last week or so by a inexperienced article creator. Sohom (talk) 14:43, 26 September 2024 (UTC)

  Not sure if we're disagreeing or we're having some semantics thing about what "passes GNG" means.

  But anyway there's issues beyond notability, in my view that's probably more useful. Fangz (talk) 15:08, 26 September 2024 (UTC)

  If an article has a credible chance of being kept or merged at AfD then it should not be draftified.

  If an article would definitely fail AfD and there is no editing that can fix that it should be sent to AfD. Thryduulf (talk) 15:57, 26 September 2024 (UTC)

  I think then you're pretty much arguing that the draftification process should be removed entirely, and I don't agree with that. It has its advantages. It should not be made a mandatory process by any means but just as some users prefer to work on articles as a draft and then push to the public wiki, it can be a better resolution to certain issues than the alternatives. Fangz (talk) 17:44, 26 September 2024 (UTC)

  I'm not sure that the Draft: namespace has any advantages over a user sandbox, and m:Research:Wikipedia article creation and m:Research:AfC processes and productivity says that the Draft: namespace is where articles go to die.

  I do think that we've fallen into a false binary here. The options are not "garbage in the mainspace" vs "auto-deleted as in the draftspace". There are other options (e.g., sticky prods for uncited articles, userification, bold stubbification, bold merging, developing a more consistent and predictable standard for evaluating articles, etc.). WhatamIdoing (talk) 20:20, 26 September 2024 (UTC)

  I think there is a argument to be made that this landscape might have changed a fair bit since this research was done. The latest data that these projects consider is from 2014-2017. WP:ACTRIAL happened after that research was done, and Wikipedia's policies have changed since those times. Sohom (talk) 20:48, 26 September 2024 (UTC)

  It's possible that things have changed, and I'm never one to turn down a new research project if you happen to be volunteering to do it (I believe that all the necessary data is public), but looking at the overall deletion rate in that namespace, it seems unlikely that the result will be materially different. WhatamIdoing (talk) 20:31, 27 September 2024 (UTC)

  I think then you're pretty much arguing that the draftification process should be removed entirely, and I don't agree with that. I'm sorry to pick on you but this is the clearest example yet of the circular reasoning that has got us into this mess: draftification must be good because we do it, so we must keep doing it because it's good. From literally the moment draftspace was created and people started doing this (before that, the equivalent process of userfication was expressly forbidden without prior discussion), others have been pointing out that the underlying logic makes no sense. Draftification is only for articles that shouldn't be deleted, but it's also only for articles that can't be in mainspace. But since fix good content in place is a part of the editing policy and almost all the community accepted reasons for deletion involve the potential of the article, not it's current state, the intersection of those two sets is functionally zero (apart from some consensus-established edge cases like paid creations or upcoming films).

  This is why attempts to clarify and improve policy around draftification—and I've been closely involved in many of them—keep failing. You try to find a solid basis for guidelines and there just isn't one. We really need to stop trying to square the circle of justifying draftification as it is practiced now, and start asking what we the community actually wants to achieve with it and whether what we're doing now fulfils that aim. So far it's not looking good for the send-them-all-to-draftspace-and-the-god-of-notability-will-recognise-his-own camp, because there's not a shred of evidence that it helps improve content, retain editors or manage the NPP workload, and as WAID says above the empirical studies we do have concluded the precise opposite. But that picture could change with more research – somebody just needs to step up and do it! – Joe (talk) 07:01, 27 September 2024 (UTC)

  @Joe Roe Draftification is only for articles that shouldn't be deleted, but it's also only for articles that can't be in mainspace. That is the exact reason why draftspace exists in the first place. Imagine you see a article with the following content: Nicholas Carlini is an amazing researcher at Google working on adversarial machine learning. created in the last week or so and sourced to a person's personal web-page. On doing a quick google search, you see that the person exists and is a researcher at said company, however, due to your unfamiliarity with adversarial machine learning topic-area you are not able to immediately identify the person's impact on the field. Do you 1) WP:BITEly nominate the article for deletion 2) leave the content up for somebody to deal with it (and hope that the other somebody will not choose option 1) or 3) draftify the article with a note that more sources are required to prove notability? Sohom (talk) 11:25, 27 September 2024 (UTC)

  @Sohom Datta None of them. What you do is add a template to the article noting the lack of sources, leave a friendly message on the creator's talk page explaining the issues in plain English, and leave a note about it at Wikipedia talk:WikiProject Computer science. Depending on what your research found you could add more information, add some sources that might or might not demonstrate notability, remove the peacock terms, etc. Yes, this is more effort than blinding draftifying or AfDing but it is far more important that things get done well than things get done quickly. Thryduulf (talk) 12:37, 27 September 2024 (UTC)

  @Sohom Datta, thanks for creating Nicholas Carlini, whose first version does not contain the hypothetical sentence you gave in your comment above. In your example above, why can't that stay in the mainspace? I frankly don't love it, and I'd immediately pull the word "amazing" out, but what's the policy basis for saying "that article truly can't be in mainspace"? WhatamIdoing (talk) 21:10, 27 September 2024 (UTC)

  @Fangz I'm not arguing for the elimination of draftspace, it has it's uses as an optional space where articles can be developed over time so they don't have to meet all the relevant content policies from the very first edit. I'm also not arguing for the elimination of all draftifcation, just the majority of unilateral draftification because, as Joe has put better than I can, it is not a net benefit to the project as currently practised. Thryduulf (talk) 12:46, 27 September 2024 (UTC)

  There's a middle ground between meets-GNG-mark-as-reviewed and fails-GNG-send-to-AfD: recently created articles where the sources in the article do not validate GNG, but where the new page reviewer hasn't done a BEFORE search. I think it's perfectly fair (and permissioned within the current draftification process) to say "this recently created article doesn't demonstrate GNG yet, but I'll kick it back to the creator in draft form to put in some more sources." Dclemens1971 (talk) 04:43, 29 September 2024 (UTC)

  Punting it to draftspace without doing a BEFORE is definitely not something we should be tolerating. Thryduulf (talk) 10:21, 29 September 2024 (UTC)

  This would mean we're either leaving these articles unpatrolled (which is obviously undesirable), or giving new page patrollers the job of finding sources on every article where the original author hasn't, which would be ideal in, well, ideal conditions, but puts the burden of actually sourcing the encyclopedia on a very small group of editors. In my opinion, there should be a way to ask the original author to add sources to show it meets GNG, beyond just putting a "notability" tag and being done with it. Chaotic Enby (talk · contribs) 19:53, 1 October 2024 (UTC)

  I agree with Chaotic Enby. Drafification is a good solution because it strongly encourages the author to improve the article, and, most importantly, gets it out of mainspace so that it isn't a problem for innocent readers – without forcing NPPers to clean up other peoples' messes. Cremastra (talk) 20:06, 1 October 2024 (UTC)

  Drafticiation [...] strongly encourages the author to improve the article. That's the theory but the evidence is that in practice it very rarely does this. There is also little to no evidence that most pages moved to draftspace are actually a problem for innocent readers rather than being a problem for those who want immediate perfection. Thryduulf (talk) 20:47, 1 October 2024 (UTC)

  About wanting to ask the original author to add sources to show it meets GNG, beyond just putting a "notability" tag and being done with it, I wonder if it's actually possible to do this in a non-coercive way. The options right now are:

  • Just ask (what the {{notability}} tag does).
  • Ask under threat of deletion (WP:BLPPROD and WP:PROD).
  • Move article to Draft: space (essentially holding the article hostage, to be deleted if you give up or can't figure out how to do it).
  • Send to AFD today.

  AFAICT a method for "force another WP:VOLUNTEER to improve the article to my standards" option has proven pretty elusive. But if you want to reach that point, I suggest that you take a baby step towards it in the form of getting a policy (any policy, really) to actually, directly, unambiguously say that every article must cite at least one source. Until the community agrees that this actually is a requirement, then we have no hope of getting them to increase the requirement all the way up to "show it meets GNG". WhatamIdoing (talk) 00:20, 2 October 2024 (UTC)

If a new editor thinks their article is ready for mainspace, they will put it there. They will also happily revert the move. If a new editor is unsure, they will probably ask for help first or use draftspace. Cremastra (talk) 19:35, 26 September 2024 (UTC)

I think the concern expressed by Joe and others who support the "backdoor" theory is that new users do not know how to revert the move to draftspace. Do you disagree with that assumption? Toadspike [Talk] 19:53, 26 September 2024 (UTC)

I think most users do not know how to revert the move, yes. I also think we shouldn't hand it to them on a silver platter, because that likely largely annuls the whole point of draftification. What is the solution to this? I couldn't tell you. Cremastra (talk) 20:09, 26 September 2024 (UTC)

Is "the whole point of draftification" to make my view of the subject's value more powerful than the newbies' view? Security through obscurity kind of works for that, but not reliably. WhatamIdoing (talk) 20:22, 26 September 2024 (UTC)

They don't know how, maybe, but more importantly that they don't know that they're allowed to. We have to remember how very unusual our collaborative process is. If an inexperienced editor contributes an article to Wikipedia and then it is swiftly unpublished with a message that there's something wrong with it, they won't think, hmm, I'm not sure if I agree with that, I'm going to revert and/or discuss this with my peer-editors to find a consensus. They'll think that with someone the authority to decide what happens to articles has rejected my contribution, and I'm a mere newbie. At that point they will either give up (the majority) or they'll persevere and get into cycle of trying to satisfy first the NPP reviewer and then a succession of AfC reviewers until they finally give up or manage to write a GA, which seems to be roughly the standard AfC is applying these days Even very experience editors fall into this trap because even though the templated messages try to communicate the full range of options the user has (now at least, after I and others have spent several years fighting for it), it's really hard to communicate that we're all equal and all have a say here within a draft–review structure that implicitly elevates the opinions of reviewers over others. – Joe (talk) 07:31, 27 September 2024 (UTC)

I've pulled the most recent 10 articles moved to mainspace with the AFCH script. They are:

• 607th Air Control Squadron – 80 words per Wikipedia:Prosesize, 2 refs: Stub per mw:ORES article quality rating.
• Will Putnam – 296 words, 15 refs: Start.
• Coat of arms of Saskatoon – 380 words, 4 refs: Start.
• Urban Realism – 304 words, 4 refs: Start.
• Devara: Part 1 (soundtrack) – 547 words, 29 refs: C class.
• Honor Thy Father and Mother: The True Story of the Menendez Murders – 562 words, 8 refs: C class.
• XP-21279 – 200 words, 11 refs: Start.
• Crew Dragon Launch Abort System – 641 words, 23 refs: B class.
• Vince Zampella – 536 words, 18 refs: C class.
• Jeffrey Kahn – 179 words, 14 refs: Start class.

That's an average of 372.5 words and 12.6 refs. The median article has 338 words and 4 refs. Compared to existing articles, 53% of our existing articles have fewer than 372.5 words, and 83% have fewer fewer than 12.6 refs. One in six articles has fewer words than the shortest in this list. One of three articles is shorter than the second-shortest in this list.

I think it is clear from these numbers that AFC is expecting more refs than existing community practice, and that they are trying to accept only articles that are already as long as ones that editors have been working on in the mainspace for years.

BTW, during the same span of time, more than 100 pages were deleted from the Draft: namespace. You shouldn't assume this means that more than 90% of drafts get deleted, because deletions are bursty and this is a relatively small sample size, but that's about what I expected. WhatamIdoing (talk) 20:56, 27 September 2024 (UTC)

• A Conclusion: I am sadly not surprised at the current state of this discussion. Some of the heated off-topic arguments verge on NOTHERE behavior. I am very disappointed to see this from experienced editors. To those of you who simply commented on the proposal: I appreciate you a lot.

Since the default NPP draft template was changed to Template:Draft article a day before this discussion began, I think my proposal is moot. I don't see how we could improve that template much, but I may raise some minor wording changes on the Template Talk. If someone wants to close this discussion, that's fine; if others wish to continue discussing other things here, I wish you the all best. Toadspike [Talk] 21:16, 27 September 2024 (UTC)

Worth also talking about the usertalk notification MTD leaves, which only provides one option: submit for review. Agree in principle we shouldn't trick people into thinking draftification/AfC is mandatory for a typical article creator. — Rhododendrites ^talk \\ 13:29, 28 September 2024 (UTC)

• Strong Oppose All it will do is destroy the draft system as it stands and eventualy destroy Wikipedia. This almost happened between 2008 and 2012, before the draft process was available, when Wikipedia was flooded with paid/coi editors and there was no effective system to deal with them. Do folk not understand what draftification is. Every publisher has draft process. It is NOT a route to deletion. That is what the detractors of the system say, many of them who are paid to oppose it and destroy it. It is the one of the core safeguards we have against the complete destruction of Wikipedia. scope_creep^Talk 11:46, 29 September 2024 (UTC)

  That comment is almost entirely evidence free assumptions of bad faith. Please try engaging with the discussion rather than just knee-jerking oppose to changing the status quo because it would change the status quo. Thryduulf (talk) 12:56, 29 September 2024 (UTC)

  Its not evidence free and I resent the fact that you have said my comment bad faith. Why would I make the comment if I didn't know what I was talking about. I've worked in NPP/AFC since it was created and was involved in some of the early discussions. I now how exactly how UPE/paid editors behave. It would lead to an exodus of editors after the place gets flooded with adverts. It would be free-for-all. The reality is that the editor who posted hasn't thought it through and hasn't looked in the archives to see what the situation was like then. scope_creep^Talk 16:05, 29 September 2024 (UTC)

  "Trust me, I was there" is not evidence. Your comment assumes bad faith from those disagreeing with you, and of everybody submitting new articles. Not every editor is paid (and disclosed paid editing is explicitly allowed), not every paid edit (disclosed or otherwise) is bad, not every paid editor (disclosed or otherwise) is attempting to harm the encyclopaedia, not every paid edit (even undisclosed ones) does harm the encyclopaedia. Thryduulf (talk) 16:19, 29 September 2024 (UTC)

  That is true to certain extent, but the majority of editors who create modern biographical, organisational and product articles which make up the majority are undeclared paid editors. They do not have our best interests at heart and never have done. scope_creep^Talk 16:53, 29 September 2024 (UTC)

  Even if that is true (and you haven't provided any evidence, of either your assertion or the implications of it that these articles harm Wikipedia and/or that draftification as currently implemented and practice prevents that harm), that doesn't mean that draftification as implemented currently can't be improved and that any changes to the status quo will mean the death of Wikipedia. Thryduulf (talk) 17:02, 29 September 2024 (UTC)

  @Scope creep, what percentage of articles in the draft space do you think get deleted? WhatamIdoing (talk) 18:51, 29 September 2024 (UTC)

  If drafts get deleted, that's because their creators have abandoned them. That's what G13 is. Perhaps more effort should be spent encouraging article writers to improve their articles after they got moved to draft (where they can be improved without interference), but draftification is not deliberate, malicious backdoor deletion, and I resent it being characterized as such. Cremastra (talk) 19:47, 29 September 2024 (UTC)

  Is this a Double-barreled question? The comment you're replying to said only "route to deletion", and you've turned it into four separate parts:

  • deliberate
  • malicious
  • backdoor
  • deletion.

  I wouldn't personally characterize any of them as malicious, but I think a fraction of them are deliberate. IMO claiming that nobody ever sent a borderline subject to AFC instead of AFD (which has lower standards in practice) would be rather extraordinary. I frankly don't think we're all so stupid that we can't figure out which route is most likely to end up with the result we prefer.

  If we characterize AFD as the "front door" for deletion, then it seems fair to describe letting articles expire in the Draft: space as the "back door".

  But the original comment is merely that it's not a route to deletion. But if 90–95% all of the articles put on that path actually do end up getting deleted, then is it not basically fair to say that it is one of our routes to deletion? WhatamIdoing (talk) 20:22, 29 September 2024 (UTC)

• Oppose. The current verbiage of the tag makes it clear to anyone with a lick of common sense, that the article has potential, but in its current form it is not ready for mainspace. Some of the comments here from folks clearly indicate a lack of understanding of what the draftification process is for. If an article, in its current form, passes GNG, then there are only certain circumstances where it should be draftified (e.g. paid editing), but if an article probably would pass GNG, but does not in its current form (e.g. there are not enough in-depth sources from independent, reliable sources to meet the standard), than that is a poster child for draftification. When I was more active in reviewing articles, I created several custom responses, which took the standard message and massaged it a bit depending on the reason for draftification (e.g. UPE, lack of GNG) or a specific topic (e.g. NFOOTY, Populated places). In some instances those messages contained an offer to ping me directly when they felt the article was ready for mainspace. I am all for article creation, but I also care about the quality and reputation of Wikipedia, which is often seen as the punchline for jokes regarding garbage information on the internet. And I would completely disagree with those who say that draftification is not a net benefit. In fact, I think it is one of the most useful tools to helping improve the quality on WP. Is it always used correctly? No. But that's an education problem with individual users, not an overriding issue with the process itself.Onel5969 ^{TT me} 14:34, 29 September 2024 (UTC)

  I agree with Onel5969. (But also remember to not leave !votes as this is the idea lab, not a formal proposal). Cremastra (talk) 14:36, 29 September 2024 (UTC)

  Apologies, Cremastra. Onel5969 ^{TT me} 19:38, 29 September 2024 (UTC)

  That might be fine in theory, but it doesn't match the what is happening in practice. Especially given that articles are being moved to draftspace for not being of sufficient quality that are C or even B class. If an article is neutrally written and meets the GNG then there is no justification for moving it to draftspace just because someone might (or might not) have been paid to edit it. Thryduulf (talk) 15:21, 29 September 2024 (UTC)

  @Thryduulf: Do you have a specific example in mind when you mention C or B class articles? scope_creep^Talk 16:13, 29 September 2024 (UTC)

  See @WhatamIdoing's comment in this discussion. Thryduulf (talk) 16:15, 29 September 2024 (UTC)

  This list is the state of articles when they come out of the Draft: space. For articles going in to the Draft: space, here's a current list:

  • Draft:Sargodha Satellite Fields – 376 words, 5 refs (reason: "more sources needed")
  • Draft:Näsir Zaman – 20 words, 10 refs (reason: "more sources needed")
  • Draft:Ahsan Khalil – 48 words, 15 refs (reason: "more sources needed")
  • Draft:Martin-J. Sepulveda – 232 words, 9 refs (reason: "likely covert advertising")
  • Draft:Gregory Bortz – 270 words, 22 refs (reason: "likely covert advertising")
  • Draft:Yury Yuriyovich Shikula – 40 words, 3 refs (really ~150 words; reason: "more sources needed")
  • Draft:Dwight Carlton Harris – 151 words, 3 refs (moved by the author for having no sources; the same editor added sources half an hour later)
  • Draft:Rhodesia Railways 19th class – 4 words (it's a list), 0 <ref>s, but two sources named (reason: "no sources, custom reason")
  • Draft:De Zandbak – 107 words, 5 refs (reason: "more sources needed")
  • Draft:Kian Breckin (footballer) – 316 words, 15 refs (moved by the author with no reason given, but it was deleted at AFD last year)

  I have skipped redirects, some round-robin page swaps, and a couple of editors moving AFC submissions from User: space to the Draft: space, and tried to include only articles being moved from the mainspace to the Draft: space. I can't get the ORES ratings for these articles, but at a glance, I think that Start and C-class is not an unreasonable description. WhatamIdoing (talk) 19:31, 29 September 2024 (UTC)

  First, thanks for providing the list. The issue is, in reviewing those drafts, most are solid drafts, and not " Especially given that articles are being moved to draftspace for not being of sufficient quality that are C or even B class." Although I think a more careful explanation could have been made. For example, the first one would have been better with a "more in-depth references from independent, reliable sources" needed, rather than simply saying "more sources needed", as there isn't a single, in-depth reference from an independent, reliable source in the draft. The second and third examples are the exact same issue. The 4th and 5th examples are properly labeled as covert advertising (both editors have been blocked for it - in addition, the 4th one didn't have a single in-depth reference from an independent source, either). The 6th example, while having 3 sources, none are in-depth, and while it might be a spelling difference on the translations of the 2nd and 3rd refs, it does not appear that the article's subject is mentioned in any of them. The 7th article is not a true example of draftification, as it was moved by the author. The 8th and 9th article have zero independent reliable sources (for the 9th, the newspaper referenced does not have a page number, and the link does not appear to bring up anything in depth about the hack lab). Not sure about the 10th, for the history is a bit wacky, but again, does not look like an example of draftification.

I think this illustrates some of the misunderstanding that folks who don't like draftification make. You look at the list provided, and you go, wow, lots of references, most not stubs or micro-stubs, why in the hell were they draftified? Hell, I did that myself, wondering if all 10 were done by a single editor, who perhaps did not have a firm grasp of draftification. But then you dive into the merits of the sourcing, or the upe issues, and it appears all 8 of the draftifications appear justified.Onel5969 ^{TT me} 20:32, 29 September 2024 (UTC)

@Onel5969, I wonder if you could explain "the newspaper" in the 9th article a little better. You say that the article has "zero independent reliable sources", but traditional print newspapers are independent reliable sources. Then you say it doesn't have a page number, but the link takes you directly to a scanned copy of the correct page; the cited article [title given in the citation] is in the last two columns. None of that makes the newspaper less independent. Is your concern that the article appears to predate the use of the name in the article title ("De Zanbak" means "The Sandbox")? WhatamIdoing (talk) 20:51, 29 September 2024 (UTC)

Could be the translation, but there does not appear to be anything connecting the group mentioned in that article, with De Zanbak. But even if there is, agf, that still is the only in-depth independent source. Onel5969 ^{TT me} 01:15, 30 September 2024 (UTC)

I'm glad we agree that a newspaper is an independent source. WhatamIdoing (talk) 01:20, 30 September 2024 (UTC)

If a new editor includes 5 sources in their submission and it gets moved to [somewhere I didn't put it] because "more sources needed" or "no sources" how many of them are going to take the time to learn that the experienced editor actually meant none of these sources contain what I think is significant, in-depth coverage in independent reliable sources and then have the confidence to say "actually, this experienced person with the power to remove my article from Wikipedia is wrong and I'm right, I'll learn how to challenge them and how and where to express my view in a way that the powerful people will listen to me" rather than just give up at some point along that path? And before anyone says it, no, just because a few bad faith editors might be among the dissuaded does not justify the loss of good faith editors. Thryduulf (talk) 21:29, 29 September 2024 (UTC)

I guess that's the difference between editors who care about quality on WP, and those who care about quantity. But that's why I said that the rationale given could have been better. Onel5969 ^{TT me} 01:17, 30 September 2024 (UTC)

Is it really a quality vs quantity question?

Or is this the difference between editors who would rather see a page run through Wikipedia:Articles for deletion or Wikipedia:Proposed article mergers instead of being unilaterally hidden until it gets deleted without the level of community oversight that we expect from AFD? For example, I'm not convinced that "De Zanbak" is a viable subject for an article, but I think there are several ways that we could address that concern, and I don't see the Draft: space helping. In fact, the only thing that moving that page to the Draft: space does that's different from moving that page to the User: space is: It's far more likely to get deleted during the next year if it's in Draft: space than if it's in User: space. WhatamIdoing (talk) 01:26, 30 September 2024 (UTC)

Well, since draftification isn't a "backdoor to deletion", nor is it "hiding an article", it's definitely a question of quality vs. quantity. Draftification, in short, is a quality control measure. These are articles that might be notable enough for mainspace, but simply aren't in good enough shape to be there. But, like other vehicles in WP, good faith editors might disagree on an article's notability, so for example in the De Zandbak articlem, Jay8g (who tagged it for notability), and Jonathan Deamer (who draftified it) might deem it potentially notable, while you, WhatamIdoing, might have simply sent it to AfD, because you do not feel it notable. But that doesn't mean the system isn't working. Perhaps we can tweak the current verbiage in the template to include where resources about where an editor can reach out for help might be added (e.g. AFC or Teahouse)?Onel5969 ^{TT me} 09:20, 30 September 2024 (UTC)

Well, since draftification isn't a "backdoor to deletion", nor is it "hiding an article", you say that as if there is no possible way good faith editors could disagree, but that simply isn't true. Whether either of those things is true is a matter of opinion (and, in my opinion, one that is consistent with the evidence presented). Thryduulf (talk) 10:07, 30 September 2024 (UTC)

Hi. No, editors can certainly have different interpretations and disagree on issues. However, in this instance, it is not a matter of disagreement. In order to hold those views indicates a lack of understanding of what the draftification process is. That's not what draftification is, it is, as I've said, simply a quality control measure. It would be like saying, it's a matter of opinion whether or not this person wrote an article about themselves, that can be interpreted as not being COI editing. Of if a an article simply cut and paste the info from Encyclopedia Brittanica, you cannot say it's your opinion that that isn't a copyvio. I mean, I have the utmost respect for you, Thryduulf, and you do a great job on WP. There are things on WP which are subjective (e.g. exactly what constitutes SIGCOV), while others are objective, (e.g. UPE/COI editing, copyvio). What draftification is falls into the latter category. All that being said, we can disagree on whether or not an individual article should or should not have been draftified. You say the evidence presented shows that it was not warranted that those articles be sent to draft. Going through the sources, however, it looks like draftification was justified. That is a difference of opinion. Onel5969 ^{TT me} 14:20, 30 September 2024 (UTC)

It's your opinion that moving an article to the Draft: space is simply a quality control measure.

It's my opinion that moving an article to the Draft: space is also simply a quality control measure that, compared to the available alternatives of leaving it in the mainspace, sending it to AFD, or moving it to User: space, substantially decreases the likelihood of the quality being improved and substantially increases the likelihood of the article being deleted.

Oh, right: Those last two points ("substantially decreases the likelihood of the quality being improved" and "substantially increases the likelihood of the article being deleted") aren't "opinions". They're objective facts. WhatamIdoing (talk) 22:17, 30 September 2024 (UTC)

Rhodesia Railways 19th class is not a list; it's a train that was in operation for multiple ranges of time. Even if it were a list, the empty headings and only content being a table is nowhere near start-class, maybe even substub. Aaron Liu (talk) 20:47, 29 September 2024 (UTC)

The existing content in the article is an infobox and a table. Tables are the format preferred by Wikipedia:Featured lists. Empty sections aren't banned, and ratings are based on what is already there. I'd rate it as |class=List today. WhatamIdoing (talk) 20:53, 29 September 2024 (UTC)

only content being a table have you actually read the page? That infobox is full of content, there are two apparently reliable sources and the table itself has about 20 rows of content. Thryduulf (talk) 21:33, 29 September 2024 (UTC)

Sorry, yes the infobox as well. I still wouldn't call it a start, though. Aaron Liu (talk) 22:03, 29 September 2024 (UTC)

Wikipedia Main Page Proposal

Moved from WP:VPT

I think that the Wikipedia main page would be more educative and with a section riddles, proverbs, idioms, wise saying. You know, a collection from many languages around, their origins, past meanings, reforms, present meanings, examples of their usage in history (past & present), their literal meanings, word for word rendering in english, etc. I don't know, who has better ideas? Let Wikipedia be a fun place too for visitors and readers to always learn more. I'm looking forward to seeing this by the start of next year and in other language wikis. Any and all contributions are accepted. elias_fdafs (talk) 20:02, 13 October 2024 (UTC)

@Elías Fortaleza de la Fuerza Sánchez: I moved your idea to the idea lab here, it was not a technical issue. — xaosflux ^Talk 20:09, 13 October 2024 (UTC)

While this does sound more like a Wiktionary or Wikiquote thing, I feel like there might be fruitful discussion to be had about showcasing featured content from sister projects in the general case. Folly Mox (talk) 20:32, 13 October 2024 (UTC)

Oh dear, another Main Page redesign suggestion. Good luck with that. --Redrose64 🌹 (talk) 20:51, 13 October 2024 (UTC)

The thing is, one person's "wise saying" is another person's "deepity". I don't think having these on the Main Page, especially in a dedicated section, would actually be very encyclopedic. However, like Folly Mox says, a more general concept of showcasing sister project content (a word etymology, a quote, etc.) could be interesting! Chaotic Enby (talk · contribs) 21:03, 13 October 2024 (UTC)

How about a small section with whatever the sister project featured thingy is? It could cycle daily. Cremastra (talk) 22:57, 13 October 2024 (UTC)

That could very well work! Chaotic Enby (talk · contribs) 17:22, 14 October 2024 (UTC)

The tricky thing is actually transcluding something from another project, which I don't think is possible without mw:Scary transclusion. Cremastra (talk) 17:25, 14 October 2024 (UTC)

I assume you mean without scary transclusion? jlwoodwa (talk) 18:54, 14 October 2024 (UTC)

Fixed. Thanks, Cremastra (talk) 18:58, 14 October 2024 (UTC)

The problem with idioms and proverbs is that, usually, they are regional. They are widely used at some area, but hardly mentioned or even unknown in others. For each user that see such a section and says "oh, that's the origin of that proverb" we'll have several who will say "what, was that a proverb? Never heard about it". Besides, explaining their background is just impossible with the limited text in main page boxes. Perhaps DYK may be a better venue to show those articles in the main page. Cambalachero (talk) 13:18, 15 October 2024 (UTC)

Articles on idioms would be helpful, especially if they mentioned pitfalls when translating between languages. However, I don't believe that the main page is an appropriate venue. -- Shmuel (Seymour J.) Metz Username:Chatul (talk) 13:58, 15 October 2024 (UTC)

As I discussed at Wikipedia:Village pump (proposals)/Archive 213 § Proposal: Create quizzes on Wikipedia, I suggest finding people interested in creating that type of content, creating a project page, and producing the content regularly on whatever schedule you can manage. From that experience, you can try to figure out how to make the process sustainable. isaacl (talk) 15:59, 15 October 2024 (UTC)

I think that it could be put in Portal:Current events, on the side under the "2024 at Wikipedia's sister projects" box. There's plenty of room, and a "____ of the Day" could be fun. WhatamIdoing (talk) 00:03, 17 October 2024 (UTC)

The main page is already overcrowded. Tinynanorobots (talk) 11:21, 29 October 2024 (UTC)

I agree with that 41.114.177.180 (talk) 11:14, 31 October 2024 (UTC)

Describing Notability in plain English

The name we use for Wikipedia:Notability has long been a source of confusion. People can guess the basic concept of many policies and guidelines from the plain-English meaning of the title (e.g., Wikipedia:Reliable sources are sources you rely on; Wikipedia:Copyright violations is about violations of copyright etc.), but this one has quite a different meaning. You can have several sources that directly say ____ is a notable musician, and we'll reply that he's not WP:Notable.

I'd like to brainstorm some alternative phrases that could be used instead of "notability", or as a supplement to it, that would be less confusing to people unfamiliar with our internal jargon.

Background

From Wikipedia:Glossary:

NN, non-notable

Abbreviation found in comments at Wikipedia:Articles for deletion and in edit summaries, indicating that the article's subject is not notable enough for a Wikipedia entry. A subject is non-notable if editors agree not to have an article about this subject. Their decision is usually based on things like not finding enough reliable sources to write a decent encyclopedia article, but it can also be based on things like a desire to present a small subject as part of a larger one.

Notability, Notable

A characteristic held by article subjects that qualify for separate, stand-alone articles. A notable topic is one that "is suitable for a stand-alone article or list when it has received significant coverage in reliable sources that are independent of the subject." Note that "notability" is a property of a topic, and has nothing to do with the quality of an article, or whether an article exists for the topic.

From a recent discussion:

• Words for the concept: criteria, concept, test, quality, qualities, requirements, notedness, guideline, threshold, when to create an article
• Words for the result: separate article, stand-alone article, separate page, stand-alone topic, new topic, own page, article creation, article suitability, inclusion
• Some specific ideas:

Collapsed list of prior ideas
The following discussion has been closed. Please do not modify it.
Article concept guideline Article creation criteria Article creation guideline Article sourcing test Article suitability criteria Article test guideline Article threshold Criteria for article creation Guide to which topics should be included as articles on Wikipedia Guideline for when a topic should have its own article Inclusion criteria Is the subject written about in reliable sources? New topic test Notedness Own page threshold test Page sourcing guide Primary notability criterion Qualifying for a separate article Separate article criteria Source availability Source depth Stand-alone concept Stand-alone article criteria Stand-alone topic criteria Stand-alone topic criterion #1 (#2, #3, etc.) When to create an article

Feel free to expand the box if you want to see some of the prior ideas. It's collapsed because some research on brainstorming ideas suggests that looking at other people's ideas can reduce the total number of ideas shared. Duplicates are fine!

Your ideas (“notability”)

Please share your ideas here. Even a 'bad' idea might inspire the next person to think of another option. WhatamIdoing (talk) 21:04, 22 October 2024 (UTC)

I had also been thinking about this. The core issue, in my opinion, is that we're trying to describe as "notability" something that is closer to "for someone/something to have an article, they should be well-documented in reliable sources". At its core, the term "notability" carries more of a connotation of relative importance, leading to a lot of newcomers, and sometimes even other users, being misled as to what makes a topic notable. On the other hand, the actual guidelines describing it focus on the existence of reliable sources about the topic, with importance only being used by some guidelines as a proxy for these sources being likely to exist.

A word like well-sourced or well-documented would carry this idea better, without the a priori of "importance/fame is what matters" that "notability" carries. Chaotic Enby (talk · contribs) 21:13, 22 October 2024 (UTC)

Lots of topics are documented by primarily primary sources (eg like many news event), but we require coverage by secondary sources, so those would worsen the situation. — Masem (t) 21:20, 22 October 2024 (UTC)

Good point, although it could still be a first start, and no word can fully convey our notability guidelines either. Maybe encyclopedically sourced could help convey the fact that not any source works? Or, as you employ the term "coverage", we could make the difference between (primary) documentation and (secondary) coverage and call it well-covered? Chaotic Enby (talk · contribs) 21:27, 22 October 2024 (UTC)

Chaotic Enby: I am still chewing over all this. I fear that encyclopedically sourced could be easily misunderstood and potentially lead to more newbies trying to cite Wikipedia itself. I really don't have a sense of how well people outside Wikipedia and academia understand the distinctions between primary, secondary, and tertiary sources. I think well-covered does capture the essence of what we're trying to convey without causing more confusion. — ClaudineChionh (she/her · talk · contribs · email) 23:42, 22 October 2024 (UTC)

I'd say that people outside the wiki-verse and academia understand primary/secondary/tertiary "not at all", and I'd say that people in the wiki-verse understand the distinction "poorly". Editors struggle with WP:PRIMARYNEWS, especially when it comes to the question of notability ("But event is obviously important, so obviously this breaking news article is secondary"), and most of them could not explain why Wikipedia:Secondary does not mean independent. WhatamIdoing (talk) 00:17, 23 October 2024 (UTC)

Masem, the GNG requires secondary sources, but NPROF does not. Both are still wiki-notable. We therefore need a handle for this concept that does not assume the GNG approach. WhatamIdoing (talk) 00:13, 23 October 2024 (UTC)

I like “When to create an article”. SmokeyJoe (talk) 21:22, 22 October 2024 (UTC)

I caution against this because as it is used now, notability is not used as a check at new page review, and primarily is a method used for evaluating whether to delete an article at AFD. We should have an advice page with that title about how to make sure you have a good topic and reviewing the notability of the topic is a good starting point as part of it. — Masem (t) 21:35, 22 October 2024 (UTC)

So "When to have an article"? WhatamIdoing (talk) 00:18, 23 October 2024 (UTC)

“When to have a separate article”?

A nod to WP:Structurism as an existing concept. Also a hint that newcomers should add content to existing article, and no be trying to add a new orphan topic as their first contribution.

- SmokeyJoe (talk) 02:12, 23 October 2024 (UTC)

• I'm inclined to something like "stand-alone topic criteria" (to clarify that the point is determining whether a topic warrants a stand-alone article) or, in a similar vein, "when to create an article" (or possibly "when it can be appropriate to create an article", since in occasional cases it can be appropriate to cover a small number of closely related topics that could be notable in the same article rather than separately). I do agree with the notion that notability isn't the best name for this guideline and that having a term that's more like plain English. Hydrangeans (she/her | talk | edits) 21:26, 22 October 2024 (UTC)
• I like "stand-alone article criteria" as it is focused on when to create an article. --Enos733 (talk) 21:28, 22 October 2024 (UTC)
• Maybe something like suitable topics or suitable article topics would work. That implies the existence of editorial judgement and that some topics simply aren't suitable. User:Chaotic Enby, this was inspired by your idea of "well-sourced", which has the potential to be confused with well-cited (i.e., the number of refs in the article right now, rather than the number of sources in the real world). WhatamIdoing (talk) 00:21, 23 October 2024 (UTC)

  From all the suggestions so far, this seems the most understandable so far. It's short, it communicates that some things are excluded and that there is 'judgement' involved. —TheDJ (talk • contribs) 09:46, 23 October 2024 (UTC)

  I really like this, because (1) it removes the emotional pain of being deemed "not notable", and (2) it avoids the confusion we currently have of subjects that are notable but about which we cannot have a meaningful article. To clarify: (1) AfD and Teahouse waste a lot of time trying to explain to upset people that Wikipedian notability has nothing to do with importance, creativity, or future value to humanity. We could save a lot of trouble by focusing on the objective need for sources rather than the subjective view of importance. (2) Notability is currently only half of the two tests we need to pass: We can't have an article unless the subject is notable and someone's written something about it for us to summarise. We have a lot of guidelines that say "XXX is generally considered notable", which result in unexpandable stubs because although it can be demonstrated from primary sources that two old ladies and a chicken once lived near a railway siding in Ohio (making it a genuine inhabited settlement), there is now nothing there, and no one has ever written anything about it, or ever will. Focusing on what is necessary to create an article would cut to what actually matters, practically, rather than getting tied up in legalistic debates about what constitutes a notable thing. Elemimele (talk) 16:14, 23 October 2024 (UTC)

  But all too often, we get the argument that we should have an article on XXX because sources exist, even though no one has demonstrated the existence of such sources. Attempts to add the requirement that reliable sources must be cited to create or keep an article have been repeatedly rejected. I would support replacing "notability" with something like "specific topics are suitable for articles if they are well-sourced, NPOV, and meet certain broad topic requirements (i.e., replacements for SNGs)". Donald Albury 17:32, 23 October 2024 (UTC)

  @Donald Albury, do you mean that topics are suitable/notable if they're neutral, or do you mean that articles are suitable/notable if they're neutral? I'm not sure if you mean that you want to repeal WP:NRVE and expand the deletion policy to say that a (deserved) {{POV}} tag is grounds for deletion, or if you mean that citing sources in a neutrally written article must be possible, even if it hasn't happened yet (including for many years). WhatamIdoing (talk) 17:44, 23 October 2024 (UTC)

  I was just trying to express that availability of reliable sources is required, but not sufficient, and that other policies and guidelines also must be met for an article to be suitable for inclusion in WP. Donald Albury 18:30, 23 October 2024 (UTC)

  I do like suitable much better than notable (and came here to suggest it). It's more accurate, but still gives some flexibility in definition, where something like well documented might be open to misinterpretation / lawyering. Folly Mox (talk) 17:50, 23 October 2024 (UTC)

  Suitable / Suitability is the first suggest I like, it helps show that this is a Wikipedia stand not a general idea. -- LCU ActivelyDisinterested «@» °∆t° 20:23, 23 October 2024 (UTC)

  Thanks a lot! I do very much like suitable topics (suitability?) too, as it is broad and flexible enough to cover our various policies on the topic. Chaotic Enby (talk · contribs) 19:26, 23 October 2024 (UTC)

  @Seraphimblade also deserves credit for this idea, since a comment from him is the source of "article suitability" in the list above.

  I'm leaning a bit towards "topics" (or "subjects"?), because "articles" could be argued to exclude lists, and because of the endless problem of "it's not notable because the article's quality is currently too low". WhatamIdoing (talk) 02:13, 24 October 2024 (UTC)

  I like that a lot. On Wikipedia, a suitable topic for an article is one where either sufficient reliable, independent sources can be found or which meets the criteria outlined in a subject-specific suitability guideline (SSG)... – sounds both more understandable and closer to how we actually decide whether or not to keep articles, in practice. – Joe (talk) 08:02, 25 October 2024 (UTC)

• Just to throw out an idea that I think can minimize disruption, is to consider a comparable situation to the relatively recent rename of of the old Naming Convention page to WP:Article Titles while specific advice of naming for various fields are still at "Naming Convention". In that same vein, if the GNG was moved to its own page (thus sitting alongside the sepearate SNG pages) and what's left at WP:N left there, then renaming that leftover to some of the suggestions above would still allow us to keep the principle of notability via the GNG and SNG while having a better landing page at a more familiar term and to explain the GNG and SNG functions within that. It would minimize a mass edit on p&g pages. The GNG and SNGs can be described as tests used on Wikipedia to measure how notable (real world definition) a topic is within the suitability on an encyclopedia. Masem (t) 18:11, 23 October 2024 (UTC)

  There might be some advantages to splitting the GNG out onto its own page, but I think that might need to be a separate discussion. WhatamIdoing (talk) 18:36, 23 October 2024 (UTC)

  Well, in terms of providing g a word that is closer to the real definition for our practice related to when a topic is suitable for it's own article, treating the existing idea of notability through the GNG and SNGs as is and focusing on a clear word for the broad concept is a clean solution. Masem (t) 19:39, 23 October 2024 (UTC)

  The issue is that GNG is about how well-documented the topic is in independent secondary sources, which doesn't necessarily map to real-world notability, and using the latter word for it has been just as much source of confusion. Chaotic Enby (talk · contribs) 19:45, 23 October 2024 (UTC)

  We can frame the GNG and SNGs as semi objective, source based tests to evaluate real-world notability, and establih in the top level guideline that one reason to allow a topic to have an article is via demonstrating real world notability using the GNG and SNGs tests. That moves us away from having notability take the wiki definition. We still need a clear understandable title for the top level guideline, and that would also discuss more that the GNG and SNGs, such as the current NLIST advice. — Masem (t) 16:55, 24 October 2024 (UTC)

  The issue is, basing article suitability on real-world notability is a very iffy definition. GNG (and to a lesser extent the SNGs) provide us with a better foundation for defining what is suitable for our encyclopedia, which is quality of independent secondary sourcing. "This person is important" is ultimately a less relevant criterion than "this person has enough secondary sources to write an article about them", if our goal is to write an encyclopedia (tertiary source, i.e. relying on secondary sourcing). Chaotic Enby (talk · contribs) 17:02, 24 October 2024 (UTC)

  Consider that several of the SNG do give measure of real world notability based on claims of importance which can be given by a single reliable primary source, the expectation they can be expanded. The key is that with a tiny bit of rewording of the GNG and SNGs are set as the evaluation of real world notability with the expectation of sourcing and coverage required for an encyclopedia, either which establishes one way a topic is suitable for a stand alone article. Masem (t) 17:35, 24 October 2024 (UTC)

The reality is that wp:notability is the operative word for "allowed to have a separate article" and the talisman for the fuzzy ecosystem/process which decides that. It incorporates with wp:notability guidelines, degree of compliance with WP:not (a measure of the degree of enclyclopedicness of the article) and a bit of influences from real world notability/importance. Any term needs to acknowledge this. If one tries to base it on summarizing just what the wp:notability guidelines say, IMO it won't work. Sincerely, North8000 (talk) 02:00, 24 October 2024 (UTC)

So WP:What topics are allowed to have a separate article? WhatamIdoing (talk) 02:10, 24 October 2024 (UTC)

Yes. And the strongest consideration in it would be wp:notability per the notability guidelines, but the above other factors described above are also a part of that consideration. North8000 (talk) 15:31, 24 October 2024 (UTC)

I prefer to see a more obvious name for the guideline for article creation. Something such as like "Guideline for article creation" would be more obvious. TFD (talk) 16:21, 24 October 2024 (UTC)

Agree. My point was agreeing with the structure/content. If we want this to have legs, we'll need something with an even shorter with a good acronym for it. North8000 (talk) 17:08, 24 October 2024 (UTC)

My concern about WP:Guideline for article creation is that I would expect the advice on a page with that name to overlap considerably with Help:Your first article. There's more to article creation than identifying whether this is a suitable/acceptable/appropriate/notable topic for an article. WhatamIdoing (talk) 17:17, 24 October 2024 (UTC)

• Procedural request I don't mean to gum up the process. But there's a risk of having 100 different ideas from 99 different editors, which can make it difficult to reach a consensus. Whenever this brainstorming step is over, I want to recommend compiling them into a list, sorted by some type of subcategory. That way we can slowly funnel our way towards something that can earn a consensus. I believe you'd probably find (at least) three or four types of names:

• Non-descriptive (compare WP:NOT or WP:SIZE): inclusion criteria, inclusion test, article creation threshold, etc.
• Type of outcome (compare WP:DISAMBIG or WP:STANDALONELIST): separate article, stand-alone article, separate page
• Standard of sources (compare WP:RELIABLE or WP:VERIFIABLE): independent sources, third party sources
• Standard of coverage (compare WP:OR or WP:COPYRIGHT): significant coverage, minimum coverage, coverage threshold

I lean towards something more descriptive, because "inclusion criteria" just shifts the complaints from "Wikipedia has an arbitrary definition of notability!" to "Wikipedia has an arbitrary list of inclusion criteria!" Newcomers and outsiders notoriously don't read passed the headline, or even twist ambiguity in bad faith to attack Wikipedia with misinformation campaigns. It would help the project much more if the guideline title summarized an uncontroversial standard for our encyclopedia. (Currently: if no reliable, independent sources can be found on a topic, then it should not have a separate article. or A topic is presumed to be suitable for a stand-alone article or list when it has received significant coverage in reliable sources that are independent of the subject.) Shooterwalker (talk) 17:24, 24 October 2024 (UTC)

I agree that "inclusion criteria" could shift the complaints from "Wikipedia has an arbitrary definition of notability!" to "Wikipedia has an arbitrary list of inclusion criteria!" However, the current name has an additional problem, namely "I have three Wikipedia:Reliable sources that WP:Directly support a claim that this subject is 'notable', so why are you claiming that it's non-notable?" That problem would go away with a name like "inclusion criteria". WhatamIdoing (talk) 19:57, 24 October 2024 (UTC)

We do have arbitrary criteria, though. All the WP:SNGs are arbitrary. The WP:GNG is arbitrary in what it considers "significant", "reliable", and "independent". Individual decisions on articles are arbitrary in how these guidelines are interpreted and how strictly they are applied. It's better to be open about that than pretend, like too many 'notability theorists' do, that we've come up with a 350 word rubric that objectively divides all of human knowledge into worthy and unworthy. I think most people can understand that to make a large project like this manageable, you have to agree on some boundaries – and respect them, even if they don't agree with them. – Joe (talk) 08:18, 25 October 2024 (UTC)

Except we do also have rubrics, and we also have to work together, so we have found it necessary to define together. There's boundaries, you say? What are they? We have to go about answering that question together. -- Alanscottwalker (talk) 11:20, 25 October 2024 (UTC)

Yes, that's what I'm saying. WP:GNG and the WP:SNGs are the boundaries we've arrived upon. They aren't objective, they're arbitrary and subjective, but that's okay. – Joe (talk) 11:31, 25 October 2024 (UTC)

Perhaps it would be more precise to say that there is an element of arbitrariness and subjectivity to decisions, especially for borderline subjects. WhatamIdoing (talk) 16:35, 25 October 2024 (UTC)

But, they are not just arbitrary. They describe real world things (sources) and using them for coming to decision (measure), and even more importantly, the rationale for doing so (writing based on what reliable others do). Arbitrary would be no definition, no rubric at all among us. We may suck, but we don't usually just rely on throwing darts or dice to delete articles. Alanscottwalker (talk) 18:39, 25 October 2024 (UTC)

Do they really describe real-world sources? Consider "The person has had a substantial impact outside academia in their academic capacity." The basis for deciding this is: Wikipedia editors say so.

Or "The person's academic work has made a significant impact in the area of higher education, affecting a substantial number of academic institutions." I tried adding a requirement for Wikipedia:Independent sources to that, and it got reverted 75 minutes later. WhatamIdoing (talk) 19:21, 25 October 2024 (UTC)

Yes, they do, those impacts are sourced, somehow -- you actually do have to prove it to each other. (Besides, you already know, this "independence" is a both matter of degree, and not strictly necessary to be in the definition for all real world sources -- and as a matter of various qualities a real world source might have, we are generally more concerned with trustworthy). Alanscottwalker (talk) 20:56, 25 October 2024 (UTC)

I've been told that the usual way of qualifying under the second one ("significant impact") is to write a book that is used in (undergraduate?) classes at multiple universities. But some classes, and some universities, appear to be more equal than others for this determination, which is arbitrary, using the definition as a decision made according to individual personal preference rather than by its intrinsic qualities.

I agree with you that supporters of these two criteria use sources whose independence can often be most politely described as a "matter of degree", and they appear to agree with you that independent sources are "not strictly necessary". (For example, I have seen sources accepted by other editors that were just a few links to class syllabi, saying that the text for the class would be Big Textbook by Alice Expert. In GNG terms, these are mere passing mentions in self-published sources, and would not be accepted for any other subject at all.) WhatamIdoing (talk) 21:36, 25 October 2024 (UTC)

I strongly disagree with the statement that the WP:GNG is arbitrary. (Or even the additional requirements of the WP:SNGs.) At worst, the application of WP:N requires some level of judgment, based on a consensus of editors applying the principle. But the evidentiary standard for writing an article is based on real, practical, and empirical experience. And it helps our project when the world understands that we write articles based on evidence. Shooterwalker (talk) 01:41, 29 October 2024 (UTC)

It's not completely random, but it is arbitrary, in the sense that nobody on the face of the earth except for English Wikipedia editors uses this definition of "notable". For example, I think virtually all people would agree that a YouTube celebrity with twenty million fans was "a famous or important person" -- it is only Wikipedians who have a secret alternate definition where it means "has had three newspaper articles written about them". jp×g🗯️ 12:11, 29 October 2024 (UTC)

That's a different point, and the real point of this brainstorm. Asking for sources isn't an arbitrary standard, but in hindsight, the word "notability" is an odd choice of words to describe it. Shooterwalker (talk) 22:04, 29 October 2024 (UTC)

• Frankly, I think that it is an obvious error that we incorrectly refer to our guideline as "notability", and that we call things that meet this guideline "notable". It is not arbitrary -- there are rules -- but they aren't about notability. No normal speaker of the English language, when they say "notability", means what that guideline says.

If I'm going to be totally honest, it feels like -- whether designed intentionally or not -- the guideline's name is designed to make sure that newbies give invalid arguments during deletion debates, thus ensuring that their autobiographies/advertisements/etc are deleted and they are dismissed, because they stupidly assume that the word means the thing it means in 99.999% of its usage in the English language. For example, the obvious direct interpretation of "Smith is not notable" is:

• The speaker's subjective opinion, which you can argue against by saying "Yes he is".
• A claim that he is not very famous, which you can argue against by saying X million people listen to his podcast
• A claim that he is not very successful, which you can argue against by saying he made X million dollars or has Y thousand clients or employs Z hundred people.
• A claim that he is not very unique, which you can argue against by saying that he's the first X to ever Y, or the only Z who's ever Qd while Zing.

We do not accept any of these arguments. If you make any of these arguments, we sneer and ridicule you for being an idiot.

I would propose that the "notability" guideline be called something that does not, in any way, create "two-tier" sentences (e.g. ones where there's an obvious plain English interpretation, and then a second Wikipedian English interpretation where it means something else). For this reason I think stuff like "impactfulness topic criteria" might be helpful, but would not fully solve the issue, as people still know what the word "impactful" means, and would argue that things were impactful, when what we actually meant was impactful, and only a moron would think that meant impactful. It should be something that nobody would ever think to define in terms other than looking up the Wikipedia policy for it. For example: "includability" or "florfbap". jp×g🗯️ 12:07, 29 October 2024 (UTC)

Maybe "sourceability"? It brings the idea of having quality sources, while not being an already existing word, to make it clear that it's a unique Wikipedia concept. Chaotic Enby (talk · contribs) 16:01, 29 October 2024 (UTC)

I was about to suggest this, one advantage is "sourcable" and "sourcability" having the same grammatical categories as "notable" and "notability", easing rewrites.--LaukkuTheGreit (Talk•Contribs) 10:39, 1 November 2024 (UTC)

How do you apply that to WP:NPROF, which doesn't really care about sources? WhatamIdoing (talk) 15:42, 1 November 2024 (UTC)

Honest answer, you deprecate NPROF. I don't know why this one guideline has been repeatedly giving exemptions to sourcing requirements in an encyclopedia that should rely on secondary sources. Chaotic Enby (talk · contribs) 15:53, 1 November 2024 (UTC)

Until that magical future appears, I think we need a name that encompasses SNG criteria that are not directly dependent on sourcing. WhatamIdoing (talk) 16:57, 1 November 2024 (UTC)

"Sourcability" is far too close in meaning to WP:Verifiability and implies a weaker aspect when in fact what notability currently is is more complex than just WP:V itself. (that's one reason why notability remains a guideline rather than a policy, because of how complex it is) Masem (t) 17:08, 1 November 2024 (UTC)

• Binominal expression or complete sentence This is a great discussion and thank you to OP for starting it. I think all the suggestions presented so far are great, though, they each introduce some of the ambiguity that already exists with Notability. I guess my comment is that we seem to be caught up on coining an all-encapsulating single word that could replace Notability. Maybe this is a case where a binominal expression ("Nice and Plenty") or even a complete sentence would be more appropriate to communicate the complexity that WP:NOTABILITY houses? Chetsford (talk) 18:10, 1 November 2024 (UTC)

Determining who should be an electionadmin

Following Wikipedia:Village pump (proposals) § Enabling SecurePoll elections with the electionadmin right (permalink), I am starting a discussion on who should actually get the electionadmin permission. (The permission is necessary for scrutineering the results.) I see a couple of potential options:

1. Bundling the permission with CheckUser
2. Creating a separate user group which simply contains the electionadmin permission, which is assignable...
   1. ...by the Arbitration Committee, or
   2. ...by community consensus at (either a new WP:Requests for election administrator or some existing page, such as WP:AN)

I would lean towards option 1, with option 2a as a second choice. The less bureaucracy the better, and CUs are trusted enough to use the permission responsibly. They have also already signed the NDA. If we are going to create a separate right, ArbCom is the body which is best equipped to assign (and importantly, monitor the use of) tools which give access to non-public information. Are there other options I am not thinking of? Reasons to pick a particular option? Other comments? HouseBlaster (talk • he/they) 02:47, 28 October 2024 (UTC)

Pinging participants at the VPR discussion: @ActivelyDisinterested, Bluerasberry, Bunnypranav, Chaotic Enby, EggRoll97, Isaacl, JSutherland (WMF), Just Step Sideways, Levivich, Novem Linguae, Pinguinn, Pppery, SD0001, Sdkb, Sohom Datta, Thryduulf, and Xaosflux. Thanks, HouseBlaster (talk • he/they) 02:47, 28 October 2024 (UTC)

Option 1. At least until T377531 is done, in which case I would give securepoll-edit-poll to admin (or maybe crat) while leaving securepoll-view-voter-pii only for checkusers. * Pppery * _{it has begun...} 02:51, 28 October 2024 (UTC)

Option 1. If scruntinising for socks, then going with CUs group – robertsky (talk) 02:55, 28 October 2024 (UTC)

Option 2a with the right being given initially to any current CU or former CU in good standing who asks for it. That way it can be added/removed independently of the CU rights if there is any reason to do so, and allows for any changes in the future. Thryduulf (talk) 03:34, 28 October 2024 (UTC)

Given that ACE scrutineers are routinely granted local CU for the purpose of properly scrutinizing the election, I do not think an electionadmin without CU makes any sense. (I can see the argument for having CUs without electionadmin, however.) In other words, I do not think it should be added independently of the CU right, but removing it would be acceptable. HouseBlaster (talk • he/they) 03:40, 28 October 2024 (UTC)

I think the ability to assign/remove it independently of CU is more important than whether it ever is given to a non-CU or removed from someone without removing CU. Thryduulf (talk) 03:45, 28 October 2024 (UTC)

Option 2a and personally, I don't think it necessarily needs to be restricted to CheckUsers. As long as the appointment comes with a vote of ArbCom and has community consultation, it satisfies the WMF's criteria for access, assuming the recipient is identified as well. I see no reason to lock it behind the CheckUser right, though I do think ArbCom is the right choice for who should be assigning it. EggRoll97 ^(talk) 04:38, 28 October 2024 (UTC)

We should postpone considering this until phab:T377531 is resolved. Communication from WMF regarding SecurePoll related groups/rights has not been technically accurate. In particular I want to note the following points:

• electionadmin group gives access to sensitive data as of today, but this is actually because of a bug. If/when the bug fix is merged (gerrit:1083337), no PII would be leaked – and the ability to setup and configure polls becomes quite low-risk and can be bundled into the sysop toolkit.
• The user right which actually does expose PII, securepoll-view-voter-pii, is extremely sensitive! It allows viewing the IPs and UAs of all the voters in a single page. This is a much higher level of access than CheckUser which only allows viewing the data for a user individually, and the use of such access along with the given reason go into logs which are audited by ArbCom/Ombuds for compliance with local and global CheckUser policy. Compare that with securepoll-view-voter-pii which allows viewing PII en masse without any audit trail (phab:T356442).

– SD0001 (talk) 07:25, 28 October 2024 (UTC)

This idea starts with the presupposition that PII must be collected on these. That is something that can be discussed. Perhaps it doesn't need to be, especially if we are going to keep using manual whitelists for the electoral rolls. — xaosflux ^Talk 09:48, 28 October 2024 (UTC)

Per SD, I think that all admins should have electionadmin right, since scheduling elections seems like great mop work. However, I disagree that the right to view voter private information shouldn’t be given to all CheckUsers. I think that it should, since the users that have accessed the information are already logged (T271276), so there is an audit trail. CheckUsers are also the ones who know about socks, so they seem like a great fit.

Also, guys, this is the idea lab, not VPR. Don’t pile on! Aaron Liu (talk) 11:28, 28 October 2024 (UTC)

• Unless we start doing an election a week we'll ever need more than a handful of election admins, so I'm not sure giving it to every admin is necessary. It sounds more like interfaceadmin (of which we have 10) or bureaucrat in that regard (15, and they aren't busy). – Joe (talk) 11:32, 28 October 2024 (UTC)

  We don't "need" it, but I see only benefits in having all admins able to do it. Aaron Liu (talk) 12:55, 28 October 2024 (UTC)

  What benefits would there be in having way (as in perhaps 100x) more people than needed? The drawbacks I can see immediately are: increased risk/attack surface (we generally follow the principle of least privilege for even the most minor rights); increased chance of misunderstandings arising from the lack of clarity over who's responsible for elections (look at what's happened already); increased chance of conflicts when too many people are trying to coordinate one election; increased expectations for new admins from adding yet another bundled responsibility. – Joe (talk) 18:01, 28 October 2024 (UTC)

  The arbitration committee election is run by volunteers who co-ordinate to avoid overlapping work. I'm confident that Wikipedia's collaborative community will continue in this tradition with administrator elections.

  Bundling privileges together is a tradeoff: sure, election admins could be a separate user group, with the overhead cost of the processes to add or remove members. That has to be weighed against the risk of someone setting up votes without community approval. I think I agree with SD0001 that it's not a high-risk scenario. There are much better ways for someone who obtained access to an admin's account to disrupt Wikipedia than trying to setup clandestine polls. isaacl (talk) 18:34, 28 October 2024 (UTC)

  That's what I'm basing my numbers off. The ArbCom election is run by three coordinators, three scrutineers, and a couple of people setting up the pre-RfC. So eight people, and none of them seem particularly run off their feet. We have over eight hundred admins.

  I imagine the potential for abuse would not be in setting up bogus elections but in manipulating or sabotaging real ones or (if CU ends up being included in this) doxxing voters en masse. Admin accounts are scarce and valuable enough commodity for it to be worth the effort for some people. But again, we apply the principle of least privilege to rollback and page mover. Why wouldn't we do it here? – Joe (talk) 19:54, 28 October 2024 (UTC)

  Regarding the arbitration committee election: anyone is eligible to co-ordinate the election RfC and the non-votewiki portion of the election itself. They don't all rush in and try to perform tasks without co-ordinating with each other.

  SD0001 suggested that the electionadmin right be bundled with the admin toolset since, after the bug is fixed, it does not require CheckUser privileges. I appreciate that, in your view, the risk is sufficiently high to warrant the overhead of managing the electionadmin right with a separate process.

  Regarding the page mover and rollback rights, they remain bundled with the sysop group, though. Since those rights are granted separately, if the principle of least privilege were followed, admins should have to apply for them separately rather than getting them bundled. I've previously stated my support for tailoring permissions with matching roles (see this comment thread on protection for Did You Know queues for example). But I recognize that there is an overhead cost, and there have to be willing volunteers to pay it. isaacl (talk) 22:09, 28 October 2024 (UTC)

Who will have access to the vote details? Unrelated to the other private details who voted, and how they voted should be encrypted and access extremely limited. If that's not the case these won't be 'secure' elections, and voters need to be very visibly warned that their votes aren't private -- LCU ActivelyDisinterested «@» °∆t° 12:51, 28 October 2024 (UTC)

The creation and setup can be done by 'crats as Joe mentioned. For the PII exposing rights, I do agree with SD's views, I also have another option, this right can be asked only by existing Check Users, and the ArbCom can grant it to the ones who have a respectable experience using CU tools. ~/Bunnypranav:<ping> 12:55, 28 October 2024 (UTC)

One thing to keep in mind, to see SP PII you have to be listed on the specific poll (and you have to be in the electionadmin group to be eligible to be listed on the specific poll). So if we decide to collect the enhanced PII in SP, then it is still limited to only the users registered to the specific poll. — xaosflux ^Talk 13:21, 28 October 2024 (UTC)

Regarding the attack vector for users with the electionadmin, a compromised account could add more scrutineers to the poll. Thus the risk would be that everyone in the group with the securepoll-view-voter-pii right could gain access. So for simplicity, I think we should be prepared to accept that everyone with the securepoll-view-voter-pii right might have access. If want to be able to move users in and out of this group on a per-poll basis, then there should be a separate user group for it. isaacl (talk) 18:45, 28 October 2024 (UTC)

Note: It is certainly possible we could have multiple, overlapping secure polls at the same time on different topics with different admins, etc. — xaosflux ^Talk 18:52, 28 October 2024 (UTC)

Sure. Everyone who has the securepoll-view-voter-pii right should be trusted for all polls currently in progress. isaacl (talk) 18:59, 28 October 2024 (UTC)

• Community consensus, as we do for scrutineers. I really don't like the idea that arbcom would be involved in any way in the admin election process, and they don't need more workload anyway. Just Step Sideways ^{from this world ..... today} 17:24, 28 October 2024 (UTC)

  So one thing that needs to be considered is, first assuming we collect the PII in SP, is that if a scrutineer wants to be able to further investigate a voter using other on-wiki data, they will also need to be a local checkuser to be able to check those logs. So it needs to be established that those checks are appropriate, and additionally we currently give arbcom exclusive decision making authority on who may be a local checkuser. — xaosflux ^Talk 18:50, 28 October 2024 (UTC)

• Echoing Xaosflux above, I believe electionadmin / securepoll-view-voter-pii does not grant every electionadmin the ability to see every voter's data ever. I think it just grants the ability to be added to a poll during setup. There is compartmentalization. So how this would work in practice if we gave the right to all the checkusers is that when we have an election, we'd probably want to pre-pick 3 checkusers to scrutineer the election, add them to the poll, then only those 3 would have access to the data.
  Right now the way I am asking WMF to set up SecurePoll for us is to let any admin create a poll, and then only that poll's designated electionadmins can edit a poll or scrutineer. phab:T378287.
  There's 3 pending patches related to phab:T377531 that may change some details of how SecurePoll permissions work. We may have some additional clarity after those patches are merged and have been in production for a few weeks. –Novem Linguae (talk) 19:08, 28 October 2024 (UTC)

  @Novem Linguae: Where is the consensus for giving all admins the ability to create a poll? It seems contrary to Levivich's close of Wikipedia:Village_pump_(proposals)#Enabling_SecurePoll_elections_with_the_electionadmin_right: An RFC to determine how the new right should be distributed can be launched at any time; it may be advisable to advertise that RFC on WP:CENT. – Joe (talk) 22:05, 28 October 2024 (UTC)

  It's the suggested setting in the extension documentation at mw:Extension:SecurePoll, and exposes no PII. However if folks object we can change the ticket. –Novem Linguae (talk) 22:29, 28 October 2024 (UTC)

  I think we need a discussion of who should have it. The RfC linked above gave this meta page as an explanation of what an 'electionadmin' is, and it says (perhaps contrary to the default in the documentation) that it's a right that allows users to set up elections with SecurePoll. – Joe (talk) 07:38, 29 October 2024 (UTC)

  Roger. Will move it to electionadmin for now. –Novem Linguae (talk) 08:45, 29 October 2024 (UTC)

• I'm opposed to local CUs doing scrutineering. For electionadmin (excl PII), I don't see why we don't give it to the Electoral Commission only? They're elected to manage the ACE process. This is distinct to option 2 in my eyes; it would be assumed consensus upon election of an Electoral Commission (perhaps the crats could action the userrights changes). For AELECT, it's trickier but could be discussed in post-trial discussions. For PII scrutineering, I think we should continue using stewards for ACE; for AELECT it's up in the air (and subject to them willing to do it, and the frequency of admin elections being reasonable for them, I'd prefer stewards handling it). ProcrastinatingReader (talk) 22:11, 29 October 2024 (UTC)

  The stewards have already indicated that they are not willing to continue handling enwiki admin elections. Thryduulf (talk) 22:20, 29 October 2024 (UTC)

  Looks opposite to me? See: [34] and [35] - seems like their initial comment was misinterpreted? ProcrastinatingReader (talk) 22:21, 29 October 2024 (UTC)

  But is it really good to require relying on Stewards for local stuff? It just seems more logical to me that we make it all an on-wiki process. I just don't see why not local CU. Aaron Liu (talk) 22:22, 29 October 2024 (UTC)

  I think so, at least for scrutineering. Stewards are happy to deal with it AFAICT and it's been the way it's done historically. Aside from independence reasons, I'd be concerned with local CUs doing election scrutineering with PII because that would give local CUs a dual-purpose. Their purpose atm is tackling abuse, and they're restricted by the Wikipedia:CheckUser policy. In particular, they need cause, and their checks are logged. But scrutineers see all voters', which is a large portion of the active editor userbase, and their checks aren't logged. I find it hard to reconcile this in my head with the restrictions placed upon CUs in our local policy, which become meaningless to my mind if CUs can see most active editors' IPs come ACE/AELECT anyway. ProcrastinatingReader (talk) 22:29, 29 October 2024 (UTC)

  Not exactly sure what @Xaosflux meant above, but the task he mentioned only said that it was a bit too easy (as in friction, not permission) to access the PII, not that it's not logged; in fact, the task links to phab:T271276, which made access logged.
  Correct me if I'm wrong, but don't the list of users who can access PII have to be whitelisted by the electionadmin for every election? Aaron Liu (talk) 23:42, 29 October 2024 (UTC)

  Yes, to see securepoll data you have to be an explicitly listed admin for the specific poll you want to view; to be eligible to be listed you have to currently be in the electionadmin group. Viewing the securepoll data is logged in securepoll. It is also possible to configure a poll to not collect personal data at all, so just the public list of usernames would be available. — xaosflux ^Talk 23:53, 29 October 2024 (UTC)

  Oops, I pinged the wrong person, sorry. I meant @SD0001 lol Aaron Liu (talk) 01:30, 30 October 2024 (UTC)

  I didn't know about the feature mentioned in phab:T271276 and didn't see it in localhost, probably because of some missing setup. It does seem to be logged after all. Nevertheless, it just records that someone saw the full list – there's obviously no way of logging whose PII they happened to look at or take note of. That's more of what I meant. – SD0001 (talk) 22:19, 30 October 2024 (UTC)

  Since scrutinners need to be whitelisted, I don't really see a problem. Aaron Liu (talk) 23:23, 30 October 2024 (UTC)

  Maybe $wgSecurePollUseLogging = true; and then visiting Special:SecurePollLog will turn it on. –Novem Linguae (talk) 23:35, 30 October 2024 (UTC)

  A possible solution CUs having a lot of instant access, if a CU wants to become a PII scurtineer, they would have to apply to it, and a group of trusted users (crats/ArbCom/anyone else that the community deems trusted) can approve it. This perm can be granted/requested per election, and they would be added into the securepoll view PII whitelist. ~/Bunnypranav:<ping> 10:35, 30 October 2024 (UTC)

  In summary, you're proposing we ask bureaucrats to give people the electadmin right. Honestly, since we (probs. justifiably) don't appear to want SecurePoll to be frictionless, I think that's a good-enough idea. Aaron Liu (talk) 22:21, 29 October 2024 (UTC)

  The original purpose of the arbitration committee election commission was to be able to decide on questions that needed to be settled quickly to be effective, and thus a community RfC discussion wouldn't be feasible. Community expectations has shifted the role to include more management of community comments. Most of the election management continues to be done by other volunteers. Personally, I don't like continuing the trend of making the commission more central to the arbitration committee election process as I'd prefer that it remain a largely hands-off role. I like how the arbitration committee elections are run through a grassroots effort. isaacl (talk) 22:38, 29 October 2024 (UTC)

  How are election admins on votewiki currently decided? It looks like Cyberpower and Xaosflux have had electionadmin for various periods.[36][37] Is consensus needed, or do WMF just appoint people who express interest in helping manage the process? ProcrastinatingReader (talk) 22:46, 29 October 2024 (UTC)

  When an election is being configured on votewiki the coordinators have accounts made and are added to the group, then are added to the election, as are the scrutineers. In elections such as ACE, the coordinators are removed when voting begins so they can't access the PII that gets collected. — xaosflux ^Talk 23:55, 29 October 2024 (UTC)

  When there are non-technical coordinators, they may not get added as they wouldn't have anything to do - in which case the WMF resource generally does all the work. — xaosflux ^Talk 23:56, 29 October 2024 (UTC)

  Can something similar to this be continued, then? Coordinators are added to electionadmin, and either local crats or admins flip the bits each ACE? ProcrastinatingReader (talk) 07:10, 30 October 2024 (UTC)

• I'm going to split a couple of ideas out in to subsections below. There are a few complicated things that need to be considered. Let's assume this isn't for ACE right now, but for any other use case we decide to use securepoll for. — Preceding unsigned comment added by Xaosflux (talk • contribs)
• create new separate userright Setting up elections is a specialized, new, sensitive, very social process. Although elections require checkuser services and support, there is little overlap in the roles. I presume how this is going to work is that we have a noticeboard for requesting elections, a form to complete in that noticeboard, and then the electionadmin will use the tool which generates the election instance. When the election is complete, then I think the electionadmin should turn the results over to a checkuser for scrutiny. The major English Wikipedia election is Wikipedia:Arbitration Committee Elections December 2023/Coordination, but I think that whomever is in this role for English Wikipedia should be open and eager to collaborate across wikis with elections including Picture of the Year, WMF board elections, special elections like the Movement Charter, and Community Wishlist. All of these elections have had very serious social challenges which are beyond the role of technical functionary. It may not fall to the role of electionadmin to resolve all social issues, but the electionadmin certainly should not create election instances carelessly or without confirming that community support for an election is in place. The results of Wikimedia elections direct investments at least in the tens of millions of dollars. This election committee should consider the possibility of requesting a budget from the Wikimedia Foundation to communicate the elections, train election coordinators, discuss election policy and best practices across languages and wikiprojects, and try to establish some social and ethical norms that apply through Wikimedia projects. I would like for people to trust our elections and believe that Wikipedia is democratic. Activities like "promoting democracy" are not in the scope of checkuser duties. If we assign this userright to checkusers, then I think that will restrict elections to what checkusers currently do, rather than allow us to design elections to meet community needs. And yes of course - people with electionadmin rights should not get checkuser access to personal data. Bluerasberry (talk) 21:11, 1 November 2024 (UTC)

  While I have many words to say against your argument on electionadmins, this is a workshop, so at this point I think we should accept "create new electionadmin user group" and "add electionadmin to all admins" as separate options.

  If we assign this userright to checkusers, then I think that will restrict elections to what checkusers currently do, rather than allow us to design elections to meet community needs.

  What else would scrutineers need to do besides inspecting election PII and checkusering to ensure democracy? Aaron Liu (talk) 22:20, 1 November 2024 (UTC)

@Aaron Liu: Scrutineering and election administration are non-overlapping workflows. Part of scrutineering is checkusering. Another part of scrutineering could include confirming that someone is eligible to vote by non-standard, non-machine readable criteria, which Wikimedia elections often include. For example, elections over off-wiki processes often give voting rights to people who contribute to Wikipedia outside the Wikimedia platform, such as by processing Wikimedia Commons content for upload or similar. Bluerasberry (talk) 00:31, 2 November 2024 (UTC)

I'm fairly sure election administrators configure the list of eligible voters, not scrutineers. I have trust in administrators to conduct diplomacy and even more trust in a potential group of "specialized, sensitive, and very social" users. Aaron Liu (talk) 00:39, 2 November 2024 (UTC)

Scrutineering and election administration are non-overlapping workflows. I agee. I envision splitting electionadmin into a user right that adds and edits polls, and a user right that scrutineers polls. They are currently kind of combined. phab:T377531. –Novem Linguae (talk) 05:55, 2 November 2024 (UTC)

You lost me a bit in the second half of your post, because you switched to talking about global elections and WMF budgets. I think that would be unrelated to what we're talking about here, which is developing the ability for enwiki to host its own non-global elections, with the goal of 1) not depending on and using the resources of global partners such as stewards and WMF Trust & Safety, 2) developing the technical and social ability to hold many more elections than we do currently, and 3) increasing autonomy. Our use cases are things like WP:ACE and WP:AELECT. By the way, global elections are their own special beast, and are much more technically challenging than local elections (phab:T355594, wikitech:SecurePoll#How to run a board election), and basically require WMF Trust & Safety and WMF software engineer support no matter what, unlike local elections which will run completely self-sufficiently once we have a system in place. –Novem Linguae (talk) 05:53, 2 November 2024 (UTC)

Do we need PII in SP for local elections?

So this is something that is going to be important as to who is going to be allowed to do what, and how they will be allowed to do that. Polls don't have to collect PII. If they don't, they will still collect usernames. PRO's are that if we don't collect PII in the vote action, then the bar of who can administer elections is much lower. The con is that checkuser data of the vote-action isn't collected. Keep in mind the username is still collected - and checkuser investigations of everything that person has done on-wiki can continue as per normal. This is very close to how it is in RFA now, if the only edit/action that wasn't checkuser recorded for someone was their "vote". — xaosflux ^Talk 14:26, 30 October 2024 (UTC)

That is actually is pretty good option. As there is a suffrage requirement, the chances of abuse are a lot lower. ~/Bunnypranav:<ping> 15:25, 30 October 2024 (UTC)

I like the idea of collecting the PII from the voting data, as it's a good way of excluding some socks and catching others. However that's not the really the purpose of having a poll, so I don't see why its collected. Maybe I'm missing something, is there some other reason to collect the data? -- LCU ActivelyDisinterested «@» °∆t° 15:33, 30 October 2024 (UTC)

The purpose of collecting the PII is to detect attempts to circumvent the one person, one vote requirement by one person voting from multiple accounts. This is done by analysing the public and non-public information in the same way that checkusers at SPI do. At arbcom elections struck votes fall into five categories:

• Editors in good standing voting twice from the same account. This is permitted and should just automatically invalidate the older vote but occasionally it doesn't happen. Scrutineers strike the earlier vote.
• Editors in good standing voting twice with different accounts in good faith. e.g. someone with a valid alt account wanting to change their vote but forgetting which account they used first time, or forgetting that they'd already voted. Scrutineers strike the earlier vote.
• Editors discovered to be sockpuppetteers by normal means after they have cast votes. If only one account has been used to vote the most recent vote by that account is allowed to stand, if multiple accounts have been used to cast votes then all the votes are struck.
• Known sockpuppetteers voting with one or more accounts discovered to be sockpuppets by the scrutineers. Scrutineers strike all votes.
• Editors, not previously known to be sockpuppetteers, discovered by scrutineers to be voting with multiple accounts in bad faith. Scrutineers strike all votes.

Without PII being collected I believe that the first three types of multiple voting would still be detectable, the rest would not. Sockpuppetteers who vote using only one account will not be detected by scrutineers regardless of whether PII is collected or not. Thryduulf (talk) 16:43, 30 October 2024 (UTC)

The last two fall into the using the poll to catch editors who are socking that I mentioned, but should this be something that's part of the poll? I guess the main issues would be a sockpuppetiers setting up accounts that allow for voting, and then never using them again. That would be near impossible to catch unless they slipped up before hand.

Say the was a EC requirement to vote, a malicious actor could setup multiple accounts, use them to make good edits in completely separate areas to avoid scrutiny, and only bring them together within a vote in an attempt to sway the outcome. The normal methods for catching sockpuppets would be ineffectual in stopping that.

I was thinking this might be overkill, but now I'm starting to think I was wrong. -- LCU ActivelyDisinterested «@» °∆t° 17:14, 30 October 2024 (UTC)

At least the CUs can detect same IPs/same ballot + proxy. Aaron Liu (talk) 20:13, 30 October 2024 (UTC)

CUs who aren't election admins can't see anything more about the vote/voter that a normal admin can, even if they have a reason to look. Thryduulf (talk) 20:20, 30 October 2024 (UTC)

Sorry, I meant the scrutineers, not either of the above. I know that this is somewhat CU procedure and typed my thoughts out wrong :) thanks for the correction Aaron Liu (talk) 20:32, 30 October 2024 (UTC)

I think scrutineers can see the IP address associated with a ballot only if the poll is set to collect PII? Thryduulf (talk) 20:39, 30 October 2024 (UTC)

Looking back on the discussion, I read Disint's comment wrong. Aaron Liu (talk) 21:01, 30 October 2024 (UTC)

only if the poll is set to collect PII. I don't see any option to turn off PII collection on the page Special:SecurePoll/create. One way to turn that off would be to not grant the user right securepoll-view-voter-pii to anyone. Unless I am missing an option somewhere. –Novem Linguae (talk) 21:13, 30 October 2024 (UTC)

I'm not sure if creating a poll works on enwiki yet. Aaron Liu (talk) 23:24, 30 October 2024 (UTC)

This was in my localhost testing environment. –Novem Linguae (talk) 23:36, 30 October 2024 (UTC)

Hmmm, seems like this feature didn't get off the ground (only hide the voter list from other voters did). Sort of why we are in idea lab though -- if this is useful we could put in a feature request do "disable PII collection". — xaosflux ^Talk 22:02, 30 October 2024 (UTC)

• Electionadmins do not need access to personally identifying information Someone should be able to scrutineer election data. Right now checkusers do that. I do not think electionadmins should have access to personally identifying information, but they should be able to consult with checkusers or have some way to confirm election validity. Bluerasberry (talk) 21:15, 1 November 2024 (UTC)

  This is about whether scrutineers should have access to PII, not about electionadmins. Scrutineers are the people whitelisted for each election to view a list of the browser-used and IP-used for every vote. Aaron Liu (talk) 22:21, 1 November 2024 (UTC)

  ?? I must be misunderstanding. Who gets whitelisted to scrutineer, and on what basis? As I understood, checkusers can already do this, and the discussion is about whether users with the electionadmin right could additionally scrutineer. Is "whitelist to scrutineer" an additional class of users? Bluerasberry (talk) 00:36, 2 November 2024 (UTC)

  No, this subheading is about whether scrutineers should be able to access PII, not electamins.
  For each election, highly trusted users (so far, with the votewiki elections, those users have been stewards) are asked if any of them would like to volunteer to scrutinize the election (and just that election, though they can also volunteer to scrutinize futre elections separately).
  After that, when setting up the election, two lists have to be added by an electamin: 1. a list of all users who may vote; and 2. a list of users who may view the PII-containing logs of voting.A list from which a software may "bouncer" to deny everybody not on the list is called a whitelist. Aaron Liu (talk) 00:49, 2 November 2024 (UTC)

  Who gets whitelisted to scrutineer, and on what basis? Depends on the election. One common format is to pick 3 stewards to scrutineer an election. Then WMF T&S gives them electionadmin permissions on votewiki, and they are added to just the election they'll be scrutineering. As I understood, checkusers can already [scrutineer]. I don't think this is correct. The checkuser group does not have any SecurePoll related permissions by default. We would have to change the #Wikimedia-site-config via a Phabricator ticket. However giving checkusers these permissions is a logical idea since checkusers have already signed the NDA and are already trusted to handle the kind of voter data that SecurePoll collects. –Novem Linguae (talk) 06:03, 2 November 2024 (UTC)

  Checkusers do not currently scrutineer our elections and never have. Stewards, those who are not from enwiki, do it. Nobody aside from the three designated stewards who are scrutineering (which are any three stewards who volunteer) are the only ones who see the PII of voters in elections. ProcrastinatingReader (talk) 09:55, 3 November 2024 (UTC)

• No PII means no scrutineering. Are we comfortable with having 600 vote WP:AELECTs or 1,600 vote WP:ACEs without anyone double checking IPs and user agents for obvious socks? I'm leaning towards yes collect PII. Also SecurePoll does not currently appear to have an option to turn off PII collection. –Novem Linguae (talk) 06:22, 2 November 2024 (UTC)

  No PII means no scrutineering. I wouldn't say so. It would mean elections have the same level of (sockpuppetry) scrutineering as RfAs, where only usernames are visible. I don't really think a sock is going to change the outcome of an ACE election. At the same time, it may well help ensure the integrity of elections, if even through deterrence, thus I'm ambivalent on whether to collect PII.

  I think either stewards scrutineering with PII or no PII scrutineering are OK with me. I'd prefer either of those options to local CUs scrutineering though, which I find a bit discomforting. ProcrastinatingReader (talk) 09:54, 3 November 2024 (UTC)

  It would mean elections have the same level of (sockpuppetry) scrutineering as RfAs At the moment, not exactly – an RfA vote is an edit, so its PII is available to CUs, whereas a SecurePoll vote is not logged to Special:Log, so the CU tool has no access to its PII. I've filed phab:T378892 regarding this.

  I concur that local CUs should not get access to the wholesale PII of all voters. Scrutineering should be either done by local CUs using the CU tool only (assuming the ticket is resolved), or by external stewards like currently. – SD0001 (talk) 10:48, 3 November 2024 (UTC)

Do we need to encrypt the backend poll data?

How people voted isn't available through the securepoll system, but when setting up a poll you can optionally choose the configure encryption. This will prevent vote data from being able to be accessed by system administrators who read the datastores. This provides quite strong voter secrecy. The downside is that cryptography is hard, and will require election administrators to understand these aspects, develop and strictly adhere to secure processes for key management. As this larger idea is about who can be an election admin, if we need this component we will need a way to ensure that such admins are not only trustworthy, but that they are also technically competent. — xaosflux ^Talk 14:26, 30 October 2024 (UTC)

I'm confused, you say How people voted isn't available through the securepoll system and encryption will prevent vote data from being able to be accessed by system administrators. So to clarify without encryption can system admins see how people voted, or is that information store elsewhere? -- LCU ActivelyDisinterested «@» °∆t° 14:35, 30 October 2024 (UTC)

My understanding is they can't unless election admins give them the key (basically a very strong password) Aaron Liu (talk) 14:48, 30 October 2024 (UTC)

If encryption prevents them from accessing the datastore, can they access the unencrypted datastore without need of a key? -- LCU ActivelyDisinterested «@» °∆t° 14:55, 30 October 2024 (UTC)

No, that's the core concept of encryption. Aaron Liu (talk) 14:58, 30 October 2024 (UTC)

When you create a poll, you can choose to optionally encrypt the poll data. This can be done with SSL or GPG keys. If you encrypt the poll, the stored data can't be read by system admins (note, this is not a wikipedia admin, or 'election admin', but the back-end server administrators). Finalizing the poll requires loading the decryption key in to the tallying mechanism. If the poll isn't encrypted it is possible the vote data could be accessed by system admins accessing the raw stored data. In either case, the software doesn't ever produce a voter:choice output. — xaosflux ^Talk 15:01, 30 October 2024 (UTC)

Ok so the choice that voters make isn't accessible even without encryption, which would suggest encryption isn't needed. What general type of information about the vote data is accessible without encryption? -- LCU ActivelyDisinterested «@» °∆t° 15:05, 30 October 2024 (UTC)

It's not available through the poll system (the confidentially risk is only of stored raw data for server admins). The public data is what you can already see on all elections: date of vote, name of voter, and if the vote has been replaced or stuck. — xaosflux ^Talk 15:19, 30 October 2024 (UTC)

Sorry this doesn't make it clearer. Yes or no, can a sysadmin see how people have voted by accessing the database if it's not encrypted? -- LCU ActivelyDisinterested «@» °∆t° 15:27, 30 October 2024 (UTC)

Yes. Aaron Liu (talk) 15:30, 30 October 2024 (UTC)

Thanks, and sorry for being slow. -- LCU ActivelyDisinterested «@» °∆t° 15:34, 30 October 2024 (UTC)

Note that the current poll encryption feature still doesn't entirely prevent an actively malicious sysadmin from, say, modifying the software to do something with the unencrypted data either before it's encrypted or after it's been decrypted to be counted. Of course that's much harder to pull off than just reading the unencrypted database (especially if you don't want to leave any traces) and requires a bit more server-side access, but not impossible. Taavi (talk!) 15:14, 30 October 2024 (UTC)

Yes this is true of all safety measures, but not an argument against them. -- LCU ActivelyDisinterested «@» °∆t° 15:27, 30 October 2024 (UTC)

I don't think we need to safeguard anything from the WMF. Aaron Liu (talk) 14:58, 30 October 2024 (UTC)

(no opinion in whether it should be encrypted, just stating some facts) Not all sysadmins are WMF staff. And there are a total of 192 sysadmins, which is much more than you might expect. * Pppery * _{it has begun...} 17:34, 30 October 2024 (UTC)

Out of curiosity, where's the 192 number coming from? Folks in the ops LDAP group would definitely have enough database access to modify votes in the SQL database, but that's only 65 people I think. Who'm I missing? –Novem Linguae (talk) 21:26, 30 October 2024 (UTC)

deployment and restricted also have those permissions as I understand it. * Pppery * _{it has begun...} 22:52, 30 October 2024 (UTC)

Yes. The first two can modify the running code, and analytics-privatedata-users can also read things from the database (in addition to the restricted group). Taavi (talk!) 04:20, 31 October 2024 (UTC)

analytics-privatedata-users contains 270 people, 128 of whom are already in one of the other groups, making 334 people total. No, I don't expect any of them to go snooping, but it is what it is ... * Pppery * _{it has begun...} 21:53, 31 October 2024 (UTC)

Hmm, do they know they're operating WMF services? Aaron Liu (talk) 23:25, 30 October 2024 (UTC)

Obviously I can't read the minds of all of them, but probably, given that one of the requirements is signing L3. * Pppery * _{it has begun...} 01:09, 31 October 2024 (UTC)

Yes. Taavi (talk!) 04:34, 31 October 2024 (UTC)

Given that voting choices could be accessed I would say it needs to be encrypted. Obviously this only makes it harder to access the information not impossible, but that is true of all such measures. Account passwords are required even though as a security measure they can be overcome.

Voters would expect that their votes are secure, or if not they need to be well informed that their votes are not. -- LCU ActivelyDisinterested «@» °∆t° 15:41, 30 October 2024 (UTC)

This should be decided on a case-by-case basis. If a WikiProject is using SecurePoll to elect its coordinators, using encryption seems like overkill. For ArbCom elections, on the other hand, I see no reason not to encrypt. For such significant elections, there would be no shortage of volunteers who can handle OpenSSL keys. – SD0001 (talk) 17:59, 30 October 2024 (UTC)

Leaning no to encryption. Seems like overkill. "Make the workflow more complicated for every electionadmin in every election" vs "make it harder for a rogue sysadmin to tamper with an election one time or a couple times until they get caught/fired/access removed" seems to be the tradeoff to weigh here. –Novem Linguae (talk) 21:28, 30 October 2024 (UTC)

It's not tampering with the result that is the problem, and reading the vote choices is unlikely to get caught. I wouldn't vote if I knew my vote was so easily accessible. -- LCU ActivelyDisinterested «@» °∆t° 22:01, 30 October 2024 (UTC)

"... in every election" Encryption can be configured differently for each election. – SD0001 (talk) 22:02, 30 October 2024 (UTC)

I'll have to try out encryption in a SecurePoll test wiki sometime. I should probably also take a peek at the database and see exactly what it encrypts. But my impression is it increases complexity for the electionadmin, who has to do stuff like generate encryption keys, then make sure the encryption key doesn't get lost else the entire election's results are lost. This will reduce the pool of folks that can easily administrate an election, limiting it to a small pool of technical users who are familiar with this encryption workflow. –Novem Linguae (talk) 06:08, 2 November 2024 (UTC)