Changing the font of English Wikipedia to the "Calibri" font

Hi, according to the article Calibri

In 2023, the United States Department of State retired Times New Roman in favor of Calibri for official communications and documents.

In my opinion, Wikipedia can follow the above font and change the font "sans-serif" to "Calibri". Thanks, Hooman Mallahzadeh (talk) 14:58, 27 January 2023 (UTC)

Why? Schazjmd (talk) 15:04, 27 January 2023 (UTC)
@Schazjmd I do not know! But for the same reason that "United States Department of State" uses the font "Calibri" from now. I should do some search for finding its reason. Please you do too. Hooman Mallahzadeh (talk) 15:07, 27 January 2023 (UTC)
Apparently, the reason was accessibility. Enwiki's font is fine as is, being a sans-serif font like Calibri. ~GoatLordServant(Talk) 15:15, 27 January 2023 (UTC)
@GoatLordServant The sans-serif fonts include a variety of fonts, a set containing many fonts, one of them is Calibri font that is a type of sans-serif font. I think the current usage of sans-serif in Wikipedia means:"Change «default font» to «sans-serif type» of it". But the Calibri font is a sans-serif font that adds properties which makes it a unique font. Hooman Mallahzadeh (talk) 15:44, 27 January 2023 (UTC)
I'd oppose this. For one, Calibri is still only shipped with Windows and MS Office - mac users don't necessarily have it. – filelakeshoe (t / c) 🐱 15:17, 27 January 2023 (UTC)
@Filelakeshoe In https://techcrunch.com/2023/01/18/font-furore-as-state-dept-retires-times-new-roman-for-retiring-calibri/ it is written that:

The reason for the change is accessibility and readability

It means that calibri font is much more accessible than "Times New Roman". I believe that the accessibility of this font is great and should be enough so that United States Department of State replaced its font to calibri. Otherwise it does not decide to do this change. Hooman Mallahzadeh (talk) 15:28, 27 January 2023 (UTC)
If Wikipedia forced my browser on Mac OS to render everything in Calibri where it's not installed, it would default to Times New Roman. So much for great accessibility! – filelakeshoe (t / c) 🐱 17:14, 27 January 2023 (UTC)
I think we should spread awareness of the ability of logged in users to set their own font through user CSS. (I use Optima when possible). —Kusma (talk) 15:31, 27 January 2023 (UTC)
  • "sans-serif" is a font-family, meaning we don't force a specific font on our readers or editors at all - and let them use whatever sans-serif font they prefer via their browser settings. Forcing them to a more specific font doesn't seem like a benefit to me. — xaosflux Talk 15:44, 27 January 2023 (UTC)
Perhaps you've misunderstood, @Hooman Mallahzadeh: the reason the US Department of State's decision is in the news is because it is hilariously incompetent, not because it's something to be emulated. It's a myth that serif fonts like Times New Roman are inaccessible and Calibri is a notoriously average font. Wikipedia doesn't actually specify which font to use for body text. It's simply set to your browser's default sans serif font. You are therefore free to change that to Calibri if you wish. – Joe (talk) 15:44, 27 January 2023 (UTC)
@Joe Roe Yes, the current usage means: «change default to sans serif variety of it». But Calibri font is a so much beautiful and readable font, and perhaps a good font for making that a default font. It has benefits for other languages, for example makes Farsi texts (see زبان فارسی) so much beautiful and readable, in a way that the "Times New Roman" font does not. I really persist to make that the default font for Wikipedia. Hooman Mallahzadeh (talk) 15:53, 27 January 2023 (UTC)
@Hooman Mallahzadeh where are you seeing TNR? I opened up a fairly stock Chrome browser and pointed at a random article, the article text is showing in Arial for me (as that is the default sans-serif font in my Chrome instance). — xaosflux Talk 16:16, 27 January 2023 (UTC)
@Xaosflux According to the screenshot of my browser (which is Microsoft Edge):
 
It is TNR for "Standard" and "Serif fonts" and for the "Sans-serif font" the default is Arial. You see that as Arial, because this setting is applied as default for Microsoft Edge and Chrome. But according to the article Arial: "In Office 2007, Arial was replaced by Calibri as the default typeface in PowerPoint, Excel, and Outlook." Hooman Mallahzadeh (talk) 16:32, 27 January 2023 (UTC)
OK, so you can set that to anything that works for you, and so can everyone else. — xaosflux Talk 16:35, 27 January 2023 (UTC)
By setting the font to "sans-serif", Wikipedia is letting each individual user choose whatever font they feel is most beautiful to them. They do this by setting the font they want use for sans-serif in their browser settings. isaacl (talk) 17:03, 27 January 2023 (UTC)
No, this is a pointless nearly 1:1 change. Dronebogus (talk) 19:52, 28 January 2023 (UTC)
I believe that the decision to specify nothing more precise than "sans serif" was taken during mw:Typography Refresh in 2014. This is because it's complicated to support so many languages. Microsoft Office for Windows supports about 75 languages, which is considered quite a lot. MediaWiki supports about 500, and any supported language can be used on this (or any other) wiki. Whatamidoing (WMF) (talk) 19:47, 31 January 2023 (UTC)
IIRC and according to the linked page, they were originally going to go with various non-free fonts that the designers liked on their Macs. Then (IIRC and reading through various pages, but the linked page seems to not have been updated) after a lot of discussion it was found that generic "sans serif" was no worse than various different possibilities across different browsers and operating systems, and avoided having to compromise on principles. Anomie 04:16, 1 February 2023 (UTC)
mw:Typography refresh § Summary of changes says that there were problem with non-Latin characters, even though specifying specific fonts was motivated by trying to improve support for non-Latin characters, so the body font was reverted back to just the generic sans-serif keyword. (I think I remember reading a discussion at the time, but cannot recall the details.) isaacl (talk) 04:31, 1 February 2023 (UTC)
At one point, it was, unfortunately, a matter of "reverting". I heard that the more original, complex plan (which used a mix of free- and non-free fonts) made one of the smaller Wikipedias basically unreadable. They significantly simplified the changes after that discovery. Whatamidoing (WMF) (talk) 16:57, 1 February 2023 (UTC)

New layout: Tools fighting with sidebars

As an avid sidebar maker, I note that the new layout is reducing realestate on the right of the screen for sidebars in preference for "tools". Has there been some discussion of this? Talpedia (talk) 14:38, 28 January 2023 (UTC)

@Talpedia, I see a note on WP:VPT about mw:Reading/Web/Desktop Improvements/Features/Page tools. There should be a [hide] button at the top of the tool. Do you prefer having it collapsed? Whatamidoing (WMF) (talk) 19:55, 31 January 2023 (UTC)
Thanks for the link. Hey again; new year, more wikipedia! I think I probably do prefer it with tools minimized... just felt a bit narrow on a page for pages with sidebars on my laptop (on. Talpedia (talk) 08:49, 1 February 2023 (UTC)
I'm feeling the same way. Speaking of which, I've heard that hiding the TOC will start being "sticky" soon. I think that will be nice for work-me (I spend a lot of time in diffs), but I'm not sure what volunteer-me will end up choosing. Whatamidoing (WMF) (talk) 17:00, 1 February 2023 (UTC)

Sciencepedia

This discussion has been closed. Please do not modify it.
The following discussion has been closed. Please do not modify it.

As I have been working with microalgae non-profit past 8 years as pioner here in Finland, I chose not to patent anything but to publish free "open source patents" so everyone can benefit and develope my inventions/innovations forward.

Our planet will need every effort anyone can give and part of that, science should be free. There is no use of money to anyone if the planet comes impossible for human race to live.

So how about you make sciencepedia, innopedia, inventopedia, you name it? Where anyone can publish idea, prototype, testing etc. and by publishing: you agree everything is open for anyone to benefit also commercially? And everyone could comment, add results of their own testing/prototype and so, articles would expand?

It would be great thing to do as inventors around the world could unite their abilities to create new innovations in order to benefit the survival of the human race. Can you pass this idea forward to someone, who could present it to people high enough on wikipedia authority rank: so maybe they could consider this to be done?

Sincerely, the White Tiger of the West alias Henri Lentonen (talk) 19:49, 1 February 2023 (UTC)

@Henri Lentonen Proposals for new wikimedia projects belong at meta:Proposals for new projects. 192.76.8.64 (talk) 20:12, 1 February 2023 (UTC)

Flag translations whose original get deleted

I have no idea whether this is already done, technically possible, or worth the effort, hence making it as a very tentative suggestion here. The suggestion is prompted by a situation at Wikipedia:Articles for deletion/Lucretia (Baldung) where a hoax article in the German Wikipedia got translated into English, and fortunately someone has brought it to AfD here too.

I was wondering whether it would be possible to create an automatically filled category for "translated articles whose originals have been deleted", so that when an article is deleted on its original language Wikipedia, we get a chance to review its fate here? Of course it might get deleted for a spurious reason there, and we might choose to keep ours, but it's generally a really bad sign if another Wikipedia chooses to delete. The argument for is that AfDs are busy places on every Wikipedia, so to spot a hoax translation like this, we need a regular follower of the German Wikipedia's AfD system to feel motivated to make an English AfD, otherwise the hoax lives on; it's very difficult to find articles that have been deleted! The arguments against are perhaps (1) that it doesn't happen enough to justify the effort, and (2) it wouldn't always work. In this particular example it would have failed because the original translator failed to mark the article as a translation anyway (and, perhaps connected in terms of acknowledgement of sources, got themselves blocked from article- and draft-space for copy-vio). Elemimele (talk) 15:03, 3 January 2023 (UTC)

@Elemimele, I think this is a good idea. I wonder if this task might be best suited for a bot. When people use the Wikipedia:Content translation tool, the first revision will have a standardized edit summary that contains a link to the original. It should be possible to follow that back to the original. Then the question is what to do once you're there. Maybe post something like Template:Copied? Whatamidoing (WMF) (talk) 01:29, 6 January 2023 (UTC)
@Whatamidoing (WMF): thanks for your response. I agree it'd suit a bot, but I don't know anything about how to write or run bots. Yes, even if people make a manual translation, if they follow the en-WP help at Help:Translation the first edit summary should have a link in a fairly standardised way. Also there is a recommendation, but not an obligation, there to add the template {{Translated page|fr|Exact name of the Foreign article}} to the corresponding talk-page. I like your idea of additionally flagging articles with the copied template referring to a non-existent page; that would flag properly-carried-out copies between same-language WP articles, which would also benefit from review if the original article gets deleted. Should I propose this properly somewhere? Elemimele (talk) 06:53, 6 January 2023 (UTC)
@Elemimele, I think the next step is start a discussion at Wikipedia:Bot requests, though it might ultimately require a m:Global bot. Whatamidoing (WMF) (talk) 18:17, 6 January 2023 (UTC)
Thanks for the suggestion; I'll give it a couple of days for discussion as a few editors below have also come up with sensible comments, and then start a proposal. Elemimele (talk) 15:26, 7 January 2023 (UTC)
I think we should more generally have a bot that helps us notice when an article's foreign version is deleted. (And we should not translate articles unless we can check their sources). —Kusma (talk) 18:34, 6 January 2023 (UTC)
Perhaps a bot that flagged on the talk page when any of the articles on other languages linked via Wikidata gets deleted - if that would work. That would remove the need for someone to indicate whether an article had been translated - it would point out that other language versions that may prompt us to review our article.Nigel Ish (talk) 19:01, 6 January 2023 (UTC)
It certainly makes sense to notice all deletions whether they are the original or not. We do, however, still need to indicate that articles have been translated because this is a requirement of copyright. The translation is a derivative work, and even Wikipedia's generous copyright requires attribution (in any case, it's good practice to know where stuff came from). Kusma, I do my best to check sources when translating, but it's often impossible to check all the offline sources. I do try to keep an eye out for additional stuff too; machine translations are now so good that those of us who translate articles manually are wasting our time unless we try to add value in some way or another. The day will probably come, when it's unnecessary to have multiple wikipedias. But not yet! Elemimele (talk) 15:25, 7 January 2023 (UTC)
You would have to handle merge and renames (or does authority control fix that ??)
Does the [| content translation tag] appear on them?
There is also a similar issue when the source article has bee nupdated, but the translated article is frozen at the time of translation., with no local content added. Wakelamp d[@-@]b (talk) 03:25, 2 February 2023 (UTC)

Vector 2022 third question

Hi, I want to put forth a third question to Wikipedia:Requests for comment/Rollback of Vector 2022:

"Should Vector 2022 replace Minerva on the mobile site?"

The reason I want to ask this is because I see great potential for use of Vector 2022 on mobile. This is mainly a mobile-specific question, since Vector 2022 was deployed on desktop at the moment.

I want to know, should I ask the third question, start a new RfC, or what?

P.S. I am using Timeless but I see a lot of potential for Vector 2022. Aasim - Herrscher of Wikis ❄️ 20:10, 2 February 2023 (UTC)

Given how the RfC's spilled out I think it's unique enough to warrant its own RfC, as what you're proposing doesn't have anything to do with potentially rolling back. That and the RfC's big enough as it is right now. —Tenryuu 🐲 ( 💬 • 📝 ) 20:36, 2 February 2023 (UTC)
Aaand... created at Wikipedia:Village pump (proposals)#RFC: Vector 2022 and Mobile. Aasim - Herrscher of Wikis ❄️ 23:02, 2 February 2023 (UTC)
And it got speedy closed for being too early. Is there any prospect for discussing Vector 2022 on mobile or do I have to wait for the discussion to finish? Aasim - Herrscher of Wikis ❄️ 14:41, 3 February 2023 (UTC)
You have to wait for v22 to be actually suitable on mobile. Aaron Liu (talk) 15:23, 3 February 2023 (UTC)

Standardizing Mathematical Notation

Mathematical notation, like computer code, has a lot of room for stylistic liberty. However, many fields of science and mathematics adopt stylistic conventions unique to themselves. As a simple example, there's an entire article just on the subject of Notation for differentiation. The choice of notation is often meant to assist in computational manipulations and/or meant to reduce the cognitive load of complex expressions.

However, differences in notation also act as unnecessary barriers to entry. This only serves to diminish the utility of Wikipedia as a tool for self education. Of course, it is understood that Wikipedia is not a text book, and Wikipedia is not meant to teach. However, practically speaking, and in particular when it comes to topics concerning mathematics, a person visiting Wikipedia is typically looking to gain an understanding of a topic unfamiliar to them. Perhaps, for example, a population geneticist is seeking to leverage equations from statistical mechanics. Even assuming they are a mathematically educated graduate student, they may still struggle to parse the formulae they find during their search purely as a result of the alien notation they find when they go looking.

Therefore, the idea up for discussion is this: Should we develop rules for the standardization of mathematical notation on Wikipedia? What would be the goals of such a standardization? How might the obvious practical concerns be mitigated?

I believe the solution should be a two-fold approach. First, Wikipedia should include an entry for all mathematical notation styles, and preferably in a manner by which alternative notations can be compared or contrasted easily. The page I linked above discussing different notation for differentiation serves as an excellent example. Second, each individual page that primarily concerns a mathematical topic should adopt and reference a single notation style.

The benefits of such a policy are easily understood. The breadth of mathematical notations and the benefits they confer need not be lost or compromised to any great degree. Each page addressing a topic in mathematics (or other topic that involves the use of mathematics) should be sufficiently specific that the application of a single notation should not change their content perceptibly. Furthermore, by choosing a notation, it becomes possible to name and reference that notation style so that readers unfamiliar with it can know how to learn more. This simple change would break down many barriers that currently exist for people looking to extend their knowledge to topics related to those they are already familiar with. VincentRagusa (talk) 21:44, 28 January 2023 (UTC)

Good catch. To give two examples, when should we use an inline <math> vs. a block <math>? When should templates such as {{EquationRef}} and {{Math theorem}} be used? (see also Category:Mathematical formatting templates)? I'll also notify WP:WikiProject Mathematics about this as they have a much better understanding of this topic. CactiStaccingCrane 13:39, 29 January 2023 (UTC)
When there are several different notations for the same topic, this must be quoted in the relevant article (this is generally done). However, this is not the role of Wikipedia to provide a standardization. This would be WP:OR and is clearly forbidden by WP policies. About the style that should be used in Wikipedia, this is the object of the manual of style MOS:MATH, and in particular, MOS:MATH#Notational conventions. Going into further details would need a consensus from the community. In any case, if you have any specific concern, it must be discussed on the talk page of the involved article, or, if several articles are concerned, on WT:WPM, the talk page of the Wikiproject Mathematics. D.Lazard (talk) 14:39, 29 January 2023 (UTC)
I've even seen cases where the notation is not uniform within a single text, e.g., different sign conventions in different chapters, although in every case that I'm aware of the preface mentions it and it and gives the rationale.
My suggestions for mathematical notation in wiki are:
  • Use LaTeX inside <math>...</math> rather than {{math}}
  • Explicitly state the notation used in the article and mention alternative notation, e.g.,
    • (sign convention (+,-,-,-) versus (-,-,-,+) ... (+,+,+,-)
    • Contravariant is   and covariant is   versus covariant is   and contravariant is  .
  • use character names rather than the actual Unicode characters for anything beyond ASCII, e.g., use \gamma rather than  .
Any mathematical style guide should address choice of markup and names of standard entities, e.g.,  ,  ,  ,  , --Shmuel (Seymour J.) Metz Username:Chatul (talk) 15:19, 29 January 2023 (UTC)
Regarding the standard number field entities, for   I have seen <math>\mathbb{R}</math> and more recently <math>\Reals</math>. The latter does not seem standard Latex, but seems to give the same result. Do you know where this would be coming from? (and by the way I agree with your recommendations) PatrickR2 (talk) 16:23, 29 January 2023 (UTC)
Could that be package doublestroke? --Shmuel (Seymour J.) Metz Username:Chatul (talk) 20:58, 29 January 2023 (UTC)
Making people refer to some table of conventions in order to explain a notation is more demanding than just saying what the notation means when the article first uses it. Different pages might write the probability of an event   as   or   or  ; this multiplicity is at worst a mild annoyance as long as each article devotes half a sentence to explaining, as math prose often does anyway. Wikipedia articles on mathematics are difficult and unsatisfying for reasons that run much deeper than bikeshedding over notations can resolve. Some were written 15+ years ago, in the age before Wikipedia asked for sources, and so they are just a random grad student's scattered thoughts on the topic, lightly edited in the time since. Many have been written by people who had no training in explaining things to readers who did not already understand the subject, and it shows. XOR'easter (talk) 16:34, 29 January 2023 (UTC)
I think Notation for differentiation is a quite good example of the challenges. You might say, "We should standardize on Lagrange's prime mark notation," but in physical applications, Newton's dot notation is universal for time derivatives. So you could say, "Then we should standardize on Newton's dot notation," but it's only ever used for time derivatives, never for derivatives with respect to spatial coordinates. Then you might ask, "Why not just use Leibniz's notation?" but it's very awkward to evaluate a derivative at a point in Leibniz's notation, particularly when you want to evaluate at a point with a complicated expression. So then you could say, "How about we go with Euler's notation?" and while it's good notation, it's probably the least-used of any of these. The truth is, there is no standard notation for differentiation. The same is true for many mathematical concepts.
The stumbling block to understanding a mathematics article is rarely notation. Most often it is poor exposition. That is a very different problem and one that is much more difficult to fix. Ozob (talk) 04:24, 30 January 2023 (UTC)
No, this kind of rigid policy does more harm than good. For an essay on this, see User:Jorge Stolfi/DoW/Vogonization, or for a practical example, the "deprecation" of parenthetical references based on a similar over-eager standardization impulse at WP:PARREF has accomplished literally nothing after 2.5 years. If you have an issue with a specific article, take it up on that article's talk page and try to reach some consensus with other editors there. If you have an issue with a specific notation needed in common across several articles, ask a specific question about it on WT:WPM. Trying to make an official wiki-wide pronouncement about this kind of thing is a fool's errand that will accomplish nothing beneficial but will pointlessly antagonize a bunch of editors who never see the discussion and then have their work clobbered by busibodies trying to rigidly enforce standards with reference to faux "consensus" that most editors never considered or discussed. –jacobolus (t) 23:40, 30 January 2023 (UTC)
This gets complicated because while one format might be standard across fields, particular fields can use different notation. There can be a choice be a choice between helping a reader interact with the literature related to an article, or understand you page. An example that comes to mind is machine learning that uses *weird* notation probabalistic expectation but it's everywhere. With all that said, I think given the choice between two notations in a literature, one of which is niche to a field, and the other is more standard across field - we should use the more mainstream one. Talpedia (talk) 09:24, 1 February 2023 (UTC)
The original poster has a good point. Indeed, a textbook or a research monograph often has a table of notations and conventions around the beginning of the book. It makes sense and is certainly beneficial to have similar pages for our readers. The problem seems that of implementation. Like others have pointed out, sometimes there is no standard across the entire mathematics. However, there is often a standard convention within a specific field (e.g., rings are assumed to be commutative in algebraic geometry). This suggests to me that we should have a convention and notation page specific to each field that can be linked from articles in that field. —- Taku (talk) 07:55, 4 February 2023 (UTC)
This seems very difficult to implement in practice for various reasons.
  • Such pages would probably be WP:original synthesis.
  • Linking a symbol in a large formula (written is Latex) is difficult, especially when several symbols should be linked, as it is common, for example in algebraic geometry.
  • Linking a to a notation convention is often much less useful than linking to the underlying concept. For example, if φ(n) appears in an article, a reader who knows Euler's totient function, but do not know which meaning of φ is intended would definitively prefer "φ(n), where φ denotes Euler's totient function". Readers who need the definition, would certainly prefer to get it in one click instead of two, if a notation page is used.
  • For common notations that are commonly used without defining them, we have already Glossary of mathematical symbols. It could be linked in many articles, but I do not see how do this in a way that is useful to beginners, but do not perturb reading for other users.
For coming back to the concerns of the initial post of this thread, I am not sure that the main problem is notation. It appears that, in many cases the main problem of practitioners is to identify the mathematical theory that could help them, and to reformulate their problem to fit the framework of this theory. This requires often hours of discussions between practitioners and mathematicians. D.Lazard (talk) 12:30, 4 February 2023 (UTC)

There is a guideline for standardizing mathematical notation, MOS:MATH. That said, there are technical limitations for mathematical notation being haphazard on Wikipedia, mostly coming down to the Wikimedia developers not being willing to put any effort into fixing longstanding usability problems.

  • LaTeX-math <math> templates are not done as most sites these days have done it for many years using in-browser MathJax rendering; instead, Wikimedia insists on rendering it themselves, leading to formatting that does not match the surrounding text very well.
  • LaTeX-math does not work well on text that might have a link applied to it, because of the above. The link is not colored blue like other links, so you cannot tell that it is linked. For this reason, it is often a bad choice to use LaTeX-math in reference titles that have formulas in them, when the title is also linked.
  • For the same reason LaTeX-math is incompatible with dark-mode skins for Wikipedia. You get black-on-black text.
  • Some basic features of real LaTeX like \strut are missing from Wikipedia's implementation making some formulas impossible to render neatly.
  • I think there are bugs in rendering LaTeX-math in figure captions?
  • On the other hand, mathematics formatted by either basic html formatting (like italics) or by the {{math}} templates is incapable of handling complex formulas, or even simple formulas with square roots. (There is a sqrt template but its use is deprecated by MOS:MATH because it is very ugly.)
  • Basic html formatting is a bad choice for mathematics in any case because of the difficulty of distinguishing | (vertical bar), l (lower case ell), I (upper case vowel i), and 1 (numeral one) in the default sans-serif fonts. In running text this is usually not a big problem because you have enough context to guess what was intended. In mathematics formulas it can lead to changes in meaning.
  • Template-math is unusable in some contexts like the {{unsolved}} template because it uses formatting that toggles between italic and roman, assuming that the surrounding context is roman, and fails when the surrounding context is actually italic.
  • Mixing both kinds of formatting in a single article is unfortunate, because then you get inconsistencies in how variables and formulas look, but it is an unfortunate consequence of the above limitations.

David Eppstein (talk) 18:01, 4 February 2023 (UTC)

I think there are bugs in rendering LaTeX-math in figure captions? Yeah, when you click on an image to expand it, the caption of the expanded image doesn't display LaTeX-math. --JBL (talk) 19:46, 4 February 2023 (UTC)

I always hated how there would be a MOS:SECTLINK to a specific section of an article and said section would be removed or have it's name altered and the link would be effectively worthless. Examples include the Natalie Biden redirect and this link on the Template:COI on Wikipedia. I think we should develop some method to attempt to combat this, however, I don't really know what. Knightoftheswords281 (talk) 06:20, 23 January 2023 (UTC)

Cewbot fixes some such cases, and notifies via the talk page when it detects a problem it can't fix. There is also a bot (I forget its name) which modifies links to discussion page sections by adding /Archive123 etc. where appropriate. Certes (talk) 09:56, 23 January 2023 (UTC)
ClueBot III does that when it does the archiving. I don't know if there's a bot that does it generally for manual archives or archives by other bots. Anomie 14:23, 23 January 2023 (UTC)
Would be nice for a bot to fix these automatically, instead of making alerting us to fix it ourselves. DFlhb (talk) 00:40, 5 February 2023 (UTC)

The "Up, Down" buttons in the Teahouse

 
The buttons in the Teahouse

In the Teahouse, there are buttons on the bottom right corner that allow you to automatically go to the top of the page or the bottom. However, I have only ever seen these buttons in the Teahouse. On some pages, it takes a while to scroll to the top or bottom of a page. I propose (not formally, yet) that on any page that contains a lot of text, that these buttons are on that page. How much text, we'll have to figure out. ‍ ‍ Helloheart ‍ 01:50, 20 January 2023 (UTC)

It would probably be better to make it into a gadget for people who want it. For people on desktops who have Home and End keys, the big arrows would just be in the way. Anomie 02:26, 20 January 2023 (UTC)
Maybe it could be in "Preferences". ‍ ‍ Helloheart ‍ 02:39, 20 January 2023 (UTC)
Yes, a preference would be great. My tablet doesn't have Home and End keys... David10244 (talk) 10:09, 21 January 2023 (UTC)
On your tablet, if the screen is long enough to be scrollable, most displays will include a scroll bar - usually on the right side of the window/screen. Does tapping the top or bottom of that bar scroll you to the top or bottom? — xaosflux Talk 16:16, 23 January 2023 (UTC)
Those big, clunky buttons are very much in the way, and very annoying when I go to the Tea house to help out. So much so, that I took the trouble to block them from my view. You can, too; just add the following to your common.css:
#skip-to-top-button, #skip-to-bottom-button {display: none;} /* hide nav buttons */
Otoh, there are several templates that allow you to jump to the top and bottom, are small and unobtrusive, and appear only there, so never cover any text. See, for example, {{skip to top}}, {{skip to bottom}}. There are a few others; see the "See also" section at those templates. Anomie's opt-in suggestion is a better approach. Mathglot (talk) 19:39, 5 February 2023 (UTC)

Project-independent quality assessments

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


Quality assessments are recorded on an article's talk page in wikiproject banners. E.g.

The importance assessment is project-specific, but the quality assessment will usually be the same for all wikiprojects. It evaluates how well the article covers the subject in terms of completeness, prose quality, wikilinks, citations and so on, regardless of which projects are interested in the subject. Rather than repeat the quality assessment in each wikiproject banner, it would be convenient if it could be placed in {{WikiProject banner shell}}, and dropped from the individual wikiproject templates, unless the project has unique rules for assessing quality. E.g. rather than

{{WikiProject banner shell |1=
{{WikiProject Biography|core=yes|living=n|listas=Churchill, Winston |class=GA}}
{{WikiProject British Empire |class=GA|importance=high}}
{{WikiProject Conservatism |class=GA|importance=top}}
}}

It should be coded as

{{WikiProject banner shell |class=GA |1=
{{WikiProject Biography|core=yes|living=n|listas=Churchill, Winston}}
{{WikiProject British Empire|importance=high}}
{{WikiProject Conservatism|importance=top}}
}}

However, the article should still be assigned to project/quality categories, e.g. Category:GA-Class biography articles, Category:GA-Class British Empire articles, Category:GA-Class Conservatism articles.

I do not see an easy way to pass a "global" quality assessment to each of the wikiproject templates. Perhaps there are problems with the general idea of project-independent quality assessment. Comments? Aymatth2 (talk) 15:46, 22 January 2023 (UTC)

  • Easy support. Should have done this years ago! The Wikipedia-wide quality scale is held at Wikipedia:Content assessment and most projects use this (or a very close variant) already. Happy to advise and explore implementation options if this passes. I suspect the syntax will need to be something like {{WikiProject banner shell |class=GA |projects=biography, british empire, conservatism}} — Martin (MSGJ · talk) 20:42, 22 January 2023 (UTC)
  • Oppose The road projects often have assessments differing from what other WikiProjects use. --Rschen7754 20:50, 22 January 2023 (UTC)
    Are these useful in practice? Most assessments I see are outdated and not very accurate; is this better for road projects? —Kusma (talk) 11:08, 23 January 2023 (UTC)
    Why? They n eed to be standardized. Can you show us an example? One minor exception should not stop us from reforming the creaking system in a way that's strictly superior. Piotr Konieczny aka Prokonsul Piotrus| reply here 04:43, 25 January 2023 (UTC)
    PS. I checked Wikipedia:WikiProject Highways and I don't see any differences. If you meant a different project, please link it. Piotr Konieczny aka Prokonsul Piotrus| reply here 10:32, 26 January 2023 (UTC)
    As I linked below, WP:HWY/A. We look for certain sections to be present and up to a certain standard for even awarding Start class. --Rschen7754 05:48, 31 January 2023 (UTC)
    I think you should re-think that. It doesn't make sense for a WikiProject to insist that something is "stub-class" when it's definitely not a Wikipedia:Stub. WhatamIdoing (talk) 16:49, 1 February 2023 (UTC)
  • Ugh. I started out writing an oppose, citing Rs's point that some projects have their own criteria. But then I went looking to find some examples. I looked at Military History and Medicine, since I know those are two of the bigger and better organized projects. I was expecting to find lists of things each project was specifically looking for. I had in mind that Military History might require for a battle things like "Date and length of engagement", "Commanding officers on both sides", "Number of troops involved" "Number of casualties on both sides", that sort of thing. But I couldn't find that. What I did find was Wikipedia:WikiProject Military history/Assessment/A-Class, which is really just a rehash of the generic criteria. Wikipedia:WikiProject Military history/Assessment is more of the same. And likewise with Wikipedia:WikiProject Medicine/Assessment#Quality scale. So, can somebody point me to an example of a project which has assessment criteria that aren't just the generic criteria copied-and-pasted into a project page? -- RoySmith (talk) 22:49, 22 January 2023 (UTC)
  • Strong Support – As a newbie, I have largely refrained from all quality assessment changes precisely because of this fragmentation of quality criteria. Such granular criteria is sometimes hard to find or even understand and may be out of date on a Wikiproject's page/portal (inactive projects?). How's a newbie to know, really? But shouldn't we all roughly agree what makes "bad", "average", and "great" encyclopedia articles? Further, the idea of changing all linked project scores at the same time might be intimidating to new editors (will 5+ Wikiprojects come at me if I change them all?) and also avoidable make-work (in my humble view). Not an expert, but this seems like an elegant implementation solution as well. I'm certainly curious to see if this plays out. — LumonRedacts 04:35, 23 January 2023 (UTC)
  • Support, the vast majority of WikiProjects do not have any in-house assessments, and the few that do should just use additional code and categories, no need to make things more complicated for everyone else. —Kusma (talk) 11:06, 23 January 2023 (UTC)
    Sorry, just noticed we're on the Idea Lab pump here, so perhaps shouldn't have voted. Instead of the "put the class into Banner Shell" idea, we could also have a general "assessment" template at the top and the WikiProject banners could read that via the kind of magic used to make citation templates respect date formatting. WikiProjects could then easily opt out of that if they really want. —Kusma (talk) 11:25, 23 January 2023 (UTC)
    Yes, that could work as well. I guess the banner shell template could be made to act as the general assessment template that you mention, so wouldn't then require a separate template. In addition it would be handy to have the class displayed on the shell rather than tucked inside it? — Martin (MSGJ · talk) 12:30, 23 January 2023 (UTC)
    "This C-Class article is of interest to the following WikiProjects" would work for the shell if we have so many projects that we want to hide their templates. But for a one-project article, perhaps a "This article is C-class" followed by a "This article is covered by the Underwater Knitting task force of WikiProject Textiles" project template works better. —Kusma (talk) 12:53, 23 January 2023 (UTC)
    Mocked up example below. I don't think that takes up any more space than what you are suggesting? — Martin (MSGJ · talk) 15:09, 23 January 2023 (UTC)
    I think my concern is that I'd like to have the ability to have uncollapsed WikiProject templates, especially if there is just one, and I'm not sure that is possible with the banner shell setup. Other than that I am happy with what you propose. —Kusma (talk) 22:04, 24 January 2023 (UTC)
    My fault for starting the bold voting, sorry! — Martin (MSGJ · talk) 12:31, 23 January 2023 (UTC)
  • Support, but... you should be able to override if the need arises, such as one project having A-Class and the others don't (rather than going through and creating A-Class for each project). Using the example from above:
{{WikiProject banner shell|class=GA|1=
{{WikiProject Biography|core=yes|living=n|listas=Churchill, Winston}}
{{WikiProject British Empire|class=A|importance=high}}
{{WikiProject Conservatism|importance=top}}
}}
Regarding what Rschen said about road articles, it really only affects the low end of the scale, and even then only a handful of articles. It's really not a huge deal, but being able to override would allow for some flexibility. –Fredddie 08:38, 2 February 2023 (UTC)
The A Class category exists for almost all wikiprojects, often empty or with very few pages, but it is supported and I'm unaware of any local consensus that have deprecated its usage. It is still a part of the more general WP:Content assessment in use by the vast majority of wikiprojects. That said, I don't want A Class to be deprecated, but if something does meet the global A Class criteria why not use it? CX Zoom[he/him] (let's talk • {CX}) 15:18, 2 February 2023 (UTC)
A-class makes no sense for most project. I have proposed that we remove it from the global scale. You might like to comment at Wikipedia talk:Content assessment#Remove A-class? — Martin (MSGJ · talk) 15:37, 2 February 2023 (UTC)
I was just using it as a hypothetical. I'm not proposing anything with regards to A-Class. –Fredddie 16:29, 2 February 2023 (UTC)

Project-independent quality assessments break1

We could continue to let projects have their own quality assessment criteria, overriding the generic criteria, if the project members feel passionate about it. But I think the great majority of projects are happy to go along with the standard definitions at Wikipedia:Content assessment#Grades.

I still do not see how an assessment can be entered once on the talk page and propagated to all the project templates so they can add it to their quality categories. The parameter |1= in {{WikiProject banner shell}} lists the project assessment templates, and these have been expanded before {{WikiProject banner shell}} gets to look at the parameter value, too late to pass the assessment to the project templates. I may be missing some obvious workaround. Aymatth2 (talk) 14:51, 23 January 2023 (UTC)

Yes there are ways to do this, as Kusma alluded to. The template can actually scan the wikitext of the talk page and extract the class=C rating. — Martin (MSGJ · talk) 15:01, 23 January 2023 (UTC)
@MSGJ: I did not think of that. I find that {{str sub old|{{ {{FULLPAGENAME}} }}|{{#expr:{{Str find|<noinclude>{{ {{FULLPAGENAME}} }}</noinclude>|-Class article is of interest}}-2}}|8}} extracts .. from the template above, which proves it can be done. Seems a bit clumsy, but it does the job. Aymatth2 (talk) 16:35, 25 January 2023 (UTC)
@Aymatth2, @Kusma, @MSGJ, @Piotrus: I was working on a WP:Lua-based solution to global quality assessments. And hopefully, I'll be providing a working demonstration by the end of this week. CX Zoom[he/him] (let's talk • {CX}) 18:22, 25 January 2023 (UTC)
That sounds great, @CX Zoom.
@Aymatth2, I think the main problem will be B-class. Everyone uses the same standard for Stub, which means everyone uses the same standard for Start-class. When C-class was introduced, a few WikiProjects decided to opt-out, but that was so long ago that most of them have probably gone inactive (=non-existent groups can't have an opinion) or changed their minds since then. GA and FA are external ratings, so there's no possibility of diverging on those. But B-class means two different things:
  • In the WPMED-style of rating: A long-ish article that isn't obviously missing major information (e.g., an article about a disease should have a section on symptoms and treatment) and has at least one citation in each ==main section==.
  • In the MILHIST approach: An article that meets the five named and separately rated criteria, following the standards laid out in the group's B-class FAQ.
An article such as Battlefield medicine could end up with different ratings. WPMED rated it B-class, ORES rates it as likely (if barely) B-class, and MILHIST marked it as Start-class last time (they might choose C-class if it were updated). WhatamIdoing (talk) 21:06, 25 January 2023 (UTC)
Wikipedia:Content assessment currently uses the 6-point checklist, so I guess that will be the default criteria unless someone wants to propose a change (presumably at Wikipedia talk:Content assessment. It's not chalk and cheese though. Point 2 is "The article reasonably covers the topic, and does not contain obvious omissions or inaccuracies" and point 1 is "The article is suitably referenced, with inline citations" which more or less covers WPMED's criteria. — Martin (MSGJ · talk) 21:15, 25 January 2023 (UTC)
We can default to the highest or lowest or most common rating. I know some projects have B-class reviews, and don't auto-upgrade when other prohects do a more cursory B-class assessment. This shouldn't be a big problem, but we should decide how the template will handle cases where assessments varies? Piotr Konieczny aka Prokonsul Piotrus| reply here 10:31, 26 January 2023 (UTC)
I think if we do a round of merging assessments, another possibility is to leave all articles with diverging assessments as "unassessed" to invite human review. I'm sure there are some pages rated as A-Class and as Start-Class, and it is probably best to just put them back on the unassessed pile. —Kusma (talk) 11:03, 26 January 2023 (UTC)
  Like — Martin (MSGJ · talk) 12:19, 26 January 2023 (UTC)
A variant:
  • Allow projects to follow a non-standard approach, giving a different project-level assessment from the article-level assessment. This is probably unusual, but let's not stall overall progress by trying to force conformity
  • Scan the talk pages, and when all the project-level assessments are the same, drop them from the wikiproject templates and put them in the article-level template
  • If there are diverging assessments, mark the article "divergent" at the article level, leaving the project-level assessments in place.
    But after human review, give these articles an assessment at the article level, which may differ from one of the project-level assessments if the project follows a non-standard approach.
Aymatth2 (talk) 14:35, 26 January 2023 (UTC)
Great news, please do keep us updated and let us know what kind of help you need, if any. Piotr Konieczny aka Prokonsul Piotrus| reply here 10:25, 26 January 2023 (UTC)

Project-independent quality assessments break2

  • This sounds like a good idea, and would simplify the assessment process. I know this is early, but I'd be willing to help with the implementation of this. — Qwerfjkltalk 21:11, 25 January 2023 (UTC)
    Thank youfor your offer, let's convene a working group (if this passes) — Martin (MSGJ · talk) 21:16, 25 January 2023 (UTC)
    @MSGJ, okay, do you think a proposal should be made somewhere?
    Currently, these are the details:
    • The class should be stored in the WikiProject Banner template
    • WikiProjects can modify the class if they wish
    • There are currently two suggestions for how to implement this:
    • {{WikiProject banner shell|class=GA|1={{WikiProject Biography|core=yes|living=n|listas=Churchill, Winston}}
    {{WikiProject British Empire|importance=high}}
    {{WikiProject Conservatism|importance=top}}
    }}
    Personally, I'd prefer the former, since it is closer to the current syntax, and otherwise passing parameters for specific WPs will be difficult. — Qwerfjkltalk 17:37, 26 January 2023 (UTC)
    Yes, the former does look more straightforward, and I think each separate template should be able to reliably obtain the class in order to do the relevant categorisation. A formal proposal will be needed, and at that stage the main active WikiProjects will need to be informed too. — Martin (MSGJ · talk) 18:05, 26 January 2023 (UTC)

This looks great to me --Guerillero Parlez Moi 22:25, 26 January 2023 (UTC)

Technical implementation

@Aymatth2, @Kusma, @MSGJ, @Piotrus, @WhatamIdoing: As I expected above, I've prepared a working demonstration at Module:Sandbox/CX Zoom/WPBS. Find its usage at User talk:CX Zoom Alt/Subpage1. Notice how the parameters of the various WikiProjects continue work as earlier, except that the "class" parameter of each banner is overridden by the one in WPBS module. If the class parameter of the module is not filled, individual banner class parameter will be passed. Also notice the red-linked maintenance categories invoked by the module on that talk page. In my opinion, we may take this module a step further to include automatic triggering of "needs-infobox", "needs-image", "listas" parameters, instead of having to repeat these in each banner. We can project-wise categorise all infobox-lacking articles with a single parameter. However, there's one caveat, that every project banner will need to have a subpage similar to {{X21}}, {{X22}}, {{X23}}, {{X24}} that would be invoked from within the modular WPBS. However, when taking into account how many talk pages need to be edited anyway in order to pass class parameter into WPBS (preferably by a bot), this seems minuscule. Ideas and opinion welcome. CX Zoom[he/him] (let's talk • {CX}) 10:41, 28 January 2023 (UTC)

@CX Zoom: You are amazing. This will make things much smoother -- Guerillero Parlez Moi 12:40, 28 January 2023 (UTC)
What he said :D Piotr Konieczny aka Prokonsul Piotrus| reply here 14:55, 28 January 2023 (UTC)
@CX Zoom, I think a better implementation is needed. For example, it would be simpler to wrap the content in nowiki tags and then unstrip them in the module. I'll see if I can write a demo. — Qwerfjkltalk 13:30, 28 January 2023 (UTC)
I thought of that first, and made it somewhere in en.beta.wmflabs. But, then it occurred to me that the average assessor might not know what the nowiki tags are for and leave them out, wrongly believing that it is a result of the visual editor nowiki problem. To them, this one looks more intuitive? CX Zoom[he/him] (let's talk • {CX}) 13:47, 28 January 2023 (UTC)
@CX Zoom, in any case, it'd be easier to just move the WP templates to a subtemplate so the wikicode doesn't have to change (as much).
I'll think about a better solution.
I might start writing some code for parsing the talk pages, in preparation. — Qwerfjkltalk 14:04, 28 January 2023 (UTC)
That would work too, but then historical revisions would show up something that looks like gibberish. Imo, something like [[Template:WikiProject <name>/banner]] can be used inside WPBS module, keeping the parent template intact, so that old versions also show what was actually intended at that point of time. CX Zoom[he/him] (let's talk • {CX}) 14:23, 28 January 2023 (UTC)
I am very uncomfortable with forcing about 1,000 projects to use an awkward new syntax like {{X21}}. I would much prefer to hide the changes in {{WikiProject banner shell}} and {{WPBannerMeta}} (or modules invoked by these templates) with backward compatibility. We should be able to first implement the new versions of {{WikiProject banner shell}} and {{WPBannerMeta}}, then progressively convert different sets of talk pages to exploit the new functionality. A big bang approach will not work.
I am in favor of parsing the talk pages if there is no other way of passing the article quality assessment from {{WikiProject banner shell}} to {{WPBannerMeta}}. A crude approach is:
{{str sub old|<noinclude>{{ {{FULLPAGENAME}} }}</noinclude>|{{#expr:{{Str find|<noinclude>{{ {{FULLPAGENAME}} }}</noinclude>|-Class article is of interest}}-7}}|13}}
which on this page yields:
...
I am sure a more robust approach could be developed. Aymatth2 (talk) 19:37, 28 January 2023 (UTC)
I don't think the new syntax causes any harm. For one, this syntax is completely hidden from sight on talk pages. The actual implementation would consist of a corresponding documentation, so anyone bumping into the template page would know how to deal with it. The main reason the syntax is new and awkward is to completely eliminate the possibility of false positives. My original version was close to normal template syntax, but it had the potential of causing false positives due to interference from other templates or misplaced braces or stray vertical bars. And pages that aren't switched to the new system would not fall apart due to the backward compatibility which was the main reason I did not attempt to edit the banner to meet the needs of module but created a separate template keeping existing banners intact. CX Zoom[he/him] (let's talk • {CX}) 20:10, 28 January 2023 (UTC)
Please keep in mind that, if you intend to use this approach, you have to be careful, as the way you have it set up causes an infinite loop (when it transcludes itself to get the string to check, it involves transcluding itself, which leads to a loop), and this populates Category:Pages with template loops. I've fixed this for now in your example by simply adding a noinclude to the first transclusion aswell. Aidan9382 (talk) 20:53, 28 January 2023 (UTC)
@Aymatth2, @Aidan9382, if this method were used it should be done via a module (::getContent() ), as then the page content can be cached, and it avoids a template loop. — Qwerfjkltalk 23:01, 28 January 2023 (UTC)
Also, I don't think backwards compatibility should be a concern. — Qwerfjkltalk 23:02, 28 January 2023 (UTC)
I tried the :getContent() type, and it ended up in error, and I couldn't find an another way to make it work. Maybe you could think of something that can fix the error that I couldn't. CX Zoom[he/him] (let's talk • {CX}) 23:48, 28 January 2023 (UTC)

Most projects follow the standard approach in Wikipedia:Content assessment, but we will not get acceptance unless we allow some of the larger projects to continue to implement their own quality rating approaches. I was thinking more of something like this:

That is, the article class is displayed on the overall banner, and by default is used by all projects, but a project may override the article class

  • If there is no project class, or it is the same as the article class, the article class is used to create categories like Category:C-class Ruritania articles. The article/project class is not shown on the project banner
  • If there is a project class that differs from the article class, the project class is used to create categories like Category:Start-class Socialism articles. The project class is shown on the project banner

The changes to {{WikiProject banner shell}} to accept an article class, and {{WPBannerMeta}} to use the article class if no project class is specified, could be done at once. They would have no effect until people started adding the |class= value to {{WikiProject banner shell}} and removing the |class= values from the project templates. We might want to introduce that quietly and take a few weeks to test it on a wide range of articles and tweak if needed before making the new approach more visible to the broader community of editors. Then we would need a mechanized conversion, or series of conversions, to work through the article talk pages and

  • Add the {{WikiProject banner shell}} where not already present, including articles with no project banners or just one project banner
  • Move |class= values up from the project templates to the overall banner template, if they are all the same
  • Flag articles with multiple |class= values for manual review.

Is that the right general direction? I can write up a proposal, and if it is accepted we can follow User:MSGJ's suggestion and start a working group. Aymatth2 (talk) 15:58, 28 January 2023 (UTC)

That would be very easy to implement by changing lines 39-45 of the module. I will update the module as per wherever the consensus of the eventual discussion takes us to. However, in my personal opinion, we should enact a blanket quality rating, the way we already do for FA/GA. CX Zoom[he/him] (let's talk • {CX}) 16:06, 28 January 2023 (UTC)
I agree in theory. The quality rating shows the extent to which the article is well-written, complete, sourced, wikilinked etc., which has nothing to do with projects. But some projects have specific views on what information must be given for an article to be considered complete, and how the information should be arranged. If we try to force through a blanket rating we will not get consensus. Maybe a year or two after this change has settled down we could revisit the idea of eliminating project quality guidelines. Aymatth2 (talk) 16:46, 28 January 2023 (UTC)
We will get fewer complaints if we let different groups have the ability to opt out. WhatamIdoing (talk) 18:46, 28 January 2023 (UTC)
FWIW @Aymatth2, @WhatamIdoing: I've created Module:Sandbox/CX Zoom/WPBS 2, which causes the internal banner class parameter (if it is supplied) to override the shell parameter. See its usage at User talk:CX Zoom Alt/Subpage2. The page has same wikitext as User talk:CX Zoom Alt/Subpage1, but results in different class for one of the banners owing to differences in how conflicts are handled. Let me know if that is what was intended, or something else. Thanks! CX Zoom[he/him] (let's talk • {CX}) 21:02, 28 January 2023 (UTC)
I agree we should allow WPs to use their own class if they want to, at least for now.
I don't have any.particular opinion on moving the article rating to the WP banner shell, though it might be a bit easier to miss.
We might need to consider non-mainspace talks as well, as some of those use classes. — Qwerfjkltalk 09:30, 29 January 2023 (UTC)
I agree that this will be the path of least resistance. I am also a firm believer in "baby steps" (or "change by stealth" to give it another name, i.e. reform without people noticing any drastic changes). The firm support for this idea gives a clear direction of where this is heading in the future — Martin (MSGJ · talk) 12:54, 1 February 2023 (UTC)
Perhaps we should start with an opt-out fir Military history A and B class. I think implementing the B-class checklist everywhere could be beneficial (it is the only part of the lower level assessment process giving any feedback) but let's leave that for the next round. —Kusma (talk) 10:57, 29 January 2023 (UTC)
Or a phase-in. Hardly anybody does assessing anymore, so an override should be available to WikiProjects that care (it will be as rare as rocking horse manure). Abductive (reasoning) 11:10, 29 January 2023 (UTC)
Since, MILHIST seems to the only major WikiProject doing quality ratings as part of its WikiProject services, why do we not ask them to rate in the shell instead, that other projects will also follow. Also, I'm new enough that I lack the knowledge of any of the background history of GA/FA classes. How did they get consensus to be put on articles on for-all basis, but not Start or Stub? Quality should be universal, unlike importance rating which may vary project to project. As for B-class checklist, that is a great tool, if it were a part of shell, we could utilise that information to create additional categories like, WP Germany articles without sources, etc. CX Zoom[he/him] (let's talk • {CX}) 11:54, 29 January 2023 (UTC)
There is a lot of conservatism in the editor community. Editors who have learned a syntax that works will expect it to continue to work. Any hint that functionality will be lost (like project-specific assessments) may hit fierce opposition. It is much easier to get acceptance for small, incremental changes than for larger, higher-impact changes. In this case, the cleanest solution might be to drop support for project-specific quirks and implement a single "ratings" template taking parameters like |class=, |project1=, |importance1=, |project2=, |importance2= ... but that will never be accepted.
The lowest impact approach, so the least likely to hit roadblocks, is to
  • Implement the new functionality in {{WikiProject banner shell}} and {{WPBannerMeta}}, without requiring any new project files
  • Change the documentation to reflect the new article-level |class= parameter, and let this settle down for a few months. The new parameter will be gradually introduced on a range of articles as editors become familiar with it.
  • Then start a new project to perform bulk-conversion of talk pages where all the projects have the same quality assessment. This should be fairly uncontroversial, but we might be surprised...
  • After that, start a discussion on how to deal with the talk pages that have more than one quality assessment.
It takes patience, and we may never get to a point where quality assessments are always project-independent (which I would prefer), but it would be a big improvement and is doable. Aymatth2 (talk) 15:01, 29 January 2023 (UTC)
Alright, I agree with your idea. I made a second demonstration (linked above) where individual project rating (if one exists) overrides the shell rating. Let me know if that is what you were wanting, or something else. Thanks! CX Zoom[he/him] (let's talk • {CX}) 15:09, 29 January 2023 (UTC)
The demo at User talk:CX Zoom Alt/Subpage2 is closer. There should be a banner at the top giving the article-level class, and the class should not be displayed by the project templates unless it differs from the article-level class. I assume these would be handled by tweaks to {{WikiProject banner shell}} and {{WPBannerMeta}}. I assume the message "This page does not require a rating on the project's quality scale." is because the templates are not in the article talk namespace. But this approach still requires the {{X21}}-type templates. We would have to create about 1,000 of them before implementing the change, and then I can foresee problems with new projects and changes to project parameters. ¨Parsing the article-level class out of the page seems much simpler. Aymatth2 (talk) 15:33, 29 January 2023 (UTC)
┌─────────────────────────────────┘
@Aymatth2, perhaps section transclusion? — Qwerfjkltalk 16:25, 30 January 2023 (UTC)
┌─────────────────────────────────┘
In fact, I might start working on this soon. — Qwerfjkltalk 19:50, 29 January 2023 (UTC)
┌─────────────────────────────────┘
Demo at User:Qwerfjkl/sandbox/WPclass. I've called the class manually, but it should be built into the templates (or WPBannerMeta, preferably). — Qwerfjkltalk 16:29, 30 January 2023 (UTC)
It appears inexpensive:
Parser profiling data

DiscussionTools time usage 0 008 seconds
CPU time usage 0 223 seconds
Real time usage 0 288 seconds
Preprocessor visited node count 7,168/1,000,000
Post-expand include size 75,370/2,097,152 bytes
Template argument size 35,091/2,097,152 bytes
Highest expansion depth 35/100
Expensive parser function count 13/500
Unstrip recursion depth 0/20
Unstrip post-expand size 17,226/5,000,000 bytes
Lua time usage 0.051/10.000 seconds
Lua memory usage 2,481,610/52,428,800 bytes
Number of Wikibase entities loaded 0/400
— Qwerfjkltalk 16:55, 31 January 2023 (UTC)
 
French-inspired design
French Wikipedia has done this for more than a decade. One article assessment for all projects, with individual importance ratings. See fr:Modèle:Wikiprojet for the template, and see fr:Discussion:Allemagne for how it looks. I'm placing it here for technical inspiration.

We could even adopt the design of the French template, with only two changes: allow projects to override the general assessment (the French template doesn't, but we should, for MILHIST's sake; and it could look like the Importance ratings do on the French template); and change #2, replace the row of links with the current "[show]" link in our WikiProject banner shell, so projects still have the ability to put whatever they want in their templates. That way, we can adopt the French design, with no loss in functionality. DFlhb (talk) 22:01, 4 February 2023 (UTC)

Support French design Andre🚐 22:19, 4 February 2023 (UTC)
The principle is good, but I think there could be more efficient designs. Not sure I like how much space is taken up by the class rating. Happy to participate in follow-up discussions at Template talk:WikiProject banner shell at a later stage. — Martin (MSGJ · talk) 22:33, 4 February 2023 (UTC)
Any design changes does indeed merit further consensus, and I don't want to complicate this great proposal by pushing through another big change. I'm posting this mainly for any technical inspiration (if that's needed; but we're in good hands here), and in terms of design, I'm really just throwing out a wild idea, for inspiration's sake; our proposal should likely be much more conservative. I've added a pic to illustrate how it would look with the "two changes" I talk about above. DFlhb (talk) 22:51, 4 February 2023 (UTC)

Moving forward

  • @Qwerfjkl, Certes, CX Zoom, Aidan9382, and MSGJ: The demo at User:Qwerfjkl/sandbox/WPclass looks good to me. It solves the main technical issue. The logic should be in {{WPBannerMeta}}, so the wikiproject template can pass {{WPBannerMeta}} a project-level |class= value if one has been specified, and otherwise {{WPBannerMeta}} uses {{Template parameter value}} to look for an article-level.|class= value. And {{WikiProject banner shell}} should be adjusted so it displays the article-level |class=value, and formats categories when there are no wikiproject templates in the 1= parameter. These are straightforward changes. I think this is ready for proposal as a project, if nobody objects. Aymatth2 (talk) 20:44, 30 January 2023 (UTC)
    User:Evad37 may want to know about this for the User:Evad37/rater.js. WhatamIdoing (talk) 02:59, 31 January 2023 (UTC)
    Looks great. CX Zoom[he/him] (let's talk • {CX}) 09:01, 31 January 2023 (UTC)
    I think it would be nicer if the class was displayed in the banner shell rather than in each separate template. That is the way I assumed it would be implemented. It would also strongly hint to editors that they should change the parameter in the shell template rather than trying to change it in each separate banner. — Martin (MSGJ · talk) 12:44, 1 February 2023 (UTC)
    I believe that is plan going forward. But it would require editing the wpbs template. CX Zoom[he/him] (let's talk • {CX}) 12:46, 1 February 2023 (UTC)
    Absolutely - there will be a significant changes to a few templates required — Martin (MSGJ · talk) 12:49, 1 February 2023 (UTC)
  • I am concerned about the tone this discussion is taking. It seems that a group of a few editors have decided that they want to make assessments universal, and that WikiProjects (not only MILHIST, but WP:HWY and all subprojects) who disagree with these universal assessments will be steamrolled into submission, now or in the future. --Rschen7754 06:20, 31 January 2023 (UTC)
    While that is something I'd personally like to see, don't worry because it won't be happening without an RfC. Also, when a project like MILHIST or HWY has sensible and workable assessment criterion that is being followed already, they are more likely to become the new norm than some other random criterion. At the end of the day, most assessments then will be performed by the same editors who perform most assessments now. CX Zoom[he/him] (let's talk • {CX}) 09:09, 31 January 2023 (UTC)
    @Rschen7754: that is a valid concern. Most projects use the general quality criteria at Wikipedia:Content assessment, but some important projects have more specific quality criteria. It would far beyond the scope of this proposal to eliminate project-specific assessment criteria. Any attempt to do so would almost certainly fail.
    Next step is a formal project proposal to allow editors to add a general quality assessment for the article using a new |class= parameter in {{WikiProject banner shell}}, and let the wikiproject banners inherit this assessment. But wikiprojects may continue to give different quality assessments using their specific criteria.
    Once the new functionality has been added to {{WikiProject banner shell}} and {{WPBannerMeta}} and is settled down, we can consider a follow-up project to run through article talk pages converting to the new parameter format when all the wikiproject banners have the same quality assessment.
    The follow-up project may explicitly exclude wikiprojects like Wikipedia:WikiProject Military history, leaving the quality assessment on the project banner even when it is the same as the assessment on the {{WikiProject banner shell}}. That is, for example, the article is C class under the general quality criteria, and is also C class under the project-specific criteria.
    It may be useful to add an "opt out" parameter to {{WPBannerMeta}}, e.g. |special-criteria=y, for projects that have their own quality assessment criteria. Aymatth2 (talk) 14:11, 31 January 2023 (UTC)
    Agreed, wikiprojects by local consensus, may opt-out. Those which do not have specific criteria will take in the universal rating. CX Zoom[he/him] (let's talk • {CX}) 14:34, 31 January 2023 (UTC)
    @Aymatth2, I thought that specifying a parameter would be enough. Then, any projects that don't opt-out can have their WP banners modified so that the class is moved to the shell, with the projects remaining passing on their class to meta banner. — Qwerfjkltalk 16:52, 31 January 2023 (UTC)
    @Qwerfjkl: Editors will keep doing things the old way for some time. They will add an opt-in project banner holding the quality assessment to a new article talk page, and will not add a {{WikiProject banner shell}}. Once things have settled down, perhaps after bulk conversion, we could use tracking categories to find and fix cases like this. Or we could just leave them. on the basis that skipping {{WikiProject banner shell}} and repeating the assessment on one or two project banners is fine, but with three or more project banners it is better to wrap them in {{WikiProject banner shell}}, with |class= given there. Initially, anyway, the opt-out parameter would be more to force display of the |class= value even when the same as the article-level |class= value. Aymatth2 (talk) 17:15, 31 January 2023 (UTC)
    Fair enough. — Qwerfjkltalk 17:24, 31 January 2023 (UTC)
    Although a clearer name than "special-criteria" e.g. project-assessment-criteria might be better. — Qwerfjkltalk 20:27, 31 January 2023 (UTC)
    I wouldn't say that this is a "group of a few editors". It now seems to be widely supported by a great number of editors, and I have seen very little opposition so far. We can (and will) give it more exposure in future proposals, but I suspect this consensus will be reflected. It goes without saying that WikiProjects will have to accept the will of the general community and adapt accordingly. But I really hope that most will see this as a positive development which will lead to more robust assessments and clearer criteria. — Martin (MSGJ · talk) 12:48, 1 February 2023 (UTC)
    Having been lurking around this discussion since it started, and having worked extensively in the past on Wikiprojects that are now more or less dormant, I guess it is time to voice my support for having one quality assessment in the banner shell rather than an assessment in project banner. Donald Albury 16:40, 1 February 2023 (UTC)

I have drafted a proposal below. It has to be as clear and uncontroversial as possible. Any suggestions for improvement are welcome. Aymatth2 (talk) 01:07, 3 February 2023 (UTC)

Draft proposal: Support for article-level quality assessments

Quality assessments define how close we are to a distribution-quality article in terms of completeness, organization, prose quality, sourcing, wikilinks etc. Most projects follow the general guidelines at Wikipedia:Content assessment, but some large projects have specialized assessment guidelines. This is to propose adding a |class= parameter to {{WikiProject banner shell}}, which can display a general quality assessment, and letting project banner templates "inherit" this assessment. {{WPBannerMeta}} will look after the details, so the project banner templates will not have to change.

Projects with specialized quality assessment approaches, which will be recognized by {{WPBannerMeta}} using a new |QUALITY_CRITERIA=special parameter, can continue to record these assessments on their project banners and link to their specialized quality assessment scales.

Importance assessments are project-specific, showing how important the article is in providing complete coverage of the project subject area. An article may be high importance for one project, low importance for another, and irrelevant to most projects. This proposal does not affect importance assessments.

Banners using article-level general quality assessment are illustrated below:

 
  • {{WikiProject banner shell}} may now accept, validate and display an optional |class= parameter as shown above.
  • {{WikiProject banner shell}} may be added to an article talk page with no wikiproject banners, in which case it will populate a general category like Category:C-Class articles
  • If a new parameter |QUALITY_CRITERIA= has the value "special", the project class will be displayed and used to create categories as at present. The project class will be displayed even if it is the same as the article class. Projects will be canvassed to set this parameter if they want to use special quality assessment criteria.
  • Otherwise, the project is assumed to follow the general assessment approach, which is true of most projects today
    • {{WPBannerMeta}} will retrieve the article-level |class= value (if present) using:
      {{Template parameter value|{{FULLPAGENAME}}|WikiProject banner shell||class}}}}
    • If no article-level class value is found, the wikiproject banner will be processed as at present
    • If the wikiproject banner does not supply a |class= value to {{WPBannerMeta}}, or if it supplies a value the same as the (non-blank) article-level class, the class will not be shown in the project template, since that would be redundant. The article class will be used to form categories like Category:C-Class Linguistics articles
    • If the wikiproject banner supplies a class value that differs from the (non-blank) article class value, the talk page will be placed in a tracking category and the project class will be used to form categories like Category:C-Class Linguistics articles
  • A future project may consider bulk change to remove |class= values from wikiproject banners where the value is the same as the article level class, and where the wikiproject uses the general Wikipedia:Content assessment approach. That is outside the scope of this proposal.

Comments? Aymatth2 (talk) 01:07, 3 February 2023 (UTC)

Very clear, and hopefully uncontroversial. Thanks for putting this together — Martin (MSGJ · talk) 13:08, 3 February 2023 (UTC)
I've been following this as it has been something I've advocated for awhile, but somewhat vaguely as I'm not that good wiki templates. In the third bullet point, what does it mean to say "a wikiproject banner has not passed an |ASSESSMENT_LINK = value"? Does this mean Wikiprojects with their own ratings need to modify their project banners to note this? Or does |class need to change, to differentiate it from banners which just have a different value through entropy? Thanks again for putting this together. CMD (talk) 13:23, 3 February 2023 (UTC)
The dentistry project currently has its own quality scale, described at Wikipedia:WikiProject Dentistry/Article rating. This is linked in the template via the parameter |ASSESSMENT__LINK=Wikipedia:WikiProject Dentistry/Article rating. So as things stand, dentistry would keep its own scale and would continue to display the article ratings on its banner. If in the future, members of dentistry project decided that there scale was not really different from the global scale, they could remove the |ASSESSMENT__LINK= parameter — Martin (MSGJ · talk) 13:30, 3 February 2023 (UTC)
@Chipmunkdavis: Wikiprojects with their own rating scales will already be passing an |ASSESSMENT_LINK = value, pointing to their scale. They do not have to change anything, I will try to clarify this. Aymatth2 (talk) 14:50, 3 February 2023 (UTC)
@MSGJ: I missed the Dentistry scale. Bad example. Will fix it. But yes, if they no longer want to use a special scale, they would drop this link. Aymatth2 (talk) 14:50, 3 February 2023 (UTC)
Most projects use the ASSESSMENT_LINK to point to a project page about assessment. It doesn't meet they specifically want to use a custom scale, so these templates may need looking at in the future. If the project is inactive or barely active, it would probably be safe to switch them over to the global scale. And active projects can be consulted on their preference. — Martin (MSGJ · talk) 14:57, 3 February 2023 (UTC)
Oh, I see! Thanks both, yes, I've figured it out. I think MSGJ is right, for example, Template:WikiProject Countries has an ASSESSMENT_LINK, but it does not run its own assessments. Template:WPBannerMeta/doc currently says ASSESSMENT_LINK is filled in by default, will this behaviour need to change? CMD (talk) 15:08, 3 February 2023 (UTC)
I have tweaked the wording about |ASSESSMENT_LINK= to show that "no" or a link to Wikipedia:Content assessment#Quality_scale means the project uses the general approach"- Have also added a note that projects may want to review whether they want a special approach or the general approach. I did not realize how many use a "special" approach that is in fact identical to the general approach. Aymatth2 (talk) 15:49, 3 February 2023 (UTC)
What about projects that adopt the standard quality scale, but gave project-specific importance examples/criteria? (who therefore use |ASSESSMENT_LINK)? (I should note that this isn't a blocker for me at all.) DFlhb (talk) 23:26, 4 February 2023 (UTC)
@DFlhb: To {{WPBannerMeta}} the |QUALITY_SCALE is just a list of values, like Stub, Start, C, B, A, GA, FA, List etc. Most but not all projects use the standard / extended scale. A project could follow the standard assessment approach, but have project-specific examples. It is clearest in that case if the template links to Wikipedia:Content assessment#Quality_scale, although it will also link to the wikiproject page, and from there to the examples. If there are project-specific criteria though, the project does not follow the standard assessment approach. For example, Wikiproject Ruritanians may require that an article give the subject's date of birth to be rated above C class. A GA-class article for an early Ruritanian may be rated C-class by Wikiproject Ruritanians because the date of birth is unknown. Aymatth2 (talk) 15:07, 5 February 2023 (UTC)
Just to make sure we're talking about the same thing: I'm talking about specific importance examples, like Wikipedia:WikiProject_Apple_Inc./Assessment#Importance_scale. We copied the quality scale verbatim, but provide Apple-specific criteria for importance ratings. DFlhb (talk) 15:20, 5 February 2023 (UTC)
Yeah I was just going to say that Aymatth2 hadn't understood the point you were making! — Martin (MSGJ · talk) 15:24, 5 February 2023 (UTC)
Sorry, missed the point. Yes, importance is clearly project-specific. Aymatth2 (talk) 15:44, 5 February 2023 (UTC)
WPBannerMeta currently does not distinguish the assessment link for the quality scale and the assessment link for the importance scale. It perhaps should be reiterated that we are only proposing to merge quality ratings, so therefore we should probably retain the ASSESSMENT_LINK parameter for the purpose of the project's importance scale. Therefore we need to look for another method to determine whether a project has opted in/out of the combined quality scale assessments. Here is a suggestion. If a project banner is using |QUALITY_SCALE=standard (or have not set this parameter which means it defaults to the standard scale) then it is 95% sure that they have not customised their quality scale, so it should be safe to assume they have not opted out. — Martin (MSGJ · talk) 15:32, 5 February 2023 (UTC)
I do not think we can use |QUALITY_SCALE=standard / extended for much. It just means the project uses Stub, Start, C, B, A, GA, FA, List etc. values, which is true of most projects. Aymatth2 (talk) 15:48, 5 February 2023 (UTC)
I am reluctant to introduce yet another parameter to {{WPBannerMeta}}, but perhaps that is the only solution. Two options:
  • Split ASSESSMENT_LINK into QUALITY-LINK and IMPORTANCE-LINK
  • Add a new QUALITY-CRITERIA=general/special parameter. If omitted, and |ASSESSMENT_LINK has been resolved by {{WPBannerMeta}} to Wikipedia:Content assessment#Quality_scale, then assume "general", else assume "special".
I do not much like either... Aymatth2 (talk) 15:44, 5 February 2023 (UTC)
I think it depends on whether projects will be opting in to the combined assessments or opting out. For example we could drop a note on the talk page of every WikiProject and say what is happening, and what they need to do in order to opt out of it. This could be something like adding the |QUALITY_CRITERIA= parameter. And then after some time, we can assume that any project that has not set this parameter in their banner template is happy to use combined quality assessments. — Martin (MSGJ · talk) 20:31, 5 February 2023 (UTC)
I like that. Opt out by setting |QUALITY_CRITERIA=special. Default opt in. Will adjust the write-up. Nothing is going to happen until editors start populating the article-level |class= and blanking out the project-level |class= when the same, so the functionality can be implemented and the projects given time to respond. I have adjusted the write-up. Aymatth2 (talk) 21:17, 5 February 2023 (UTC)
Default opt-in is the way to go, especially with so many inactive WikiProjects (who wants to fix a thousand templates?) DFlhb (talk) 21:36, 5 February 2023 (UTC)
Enthusiastically support this proposal, both in spirit and in practice; it's long overdue. DFlhb (talk) 23:30, 4 February 2023 (UTC)
Support This makes sense and looks good. I think it's an improvement on the current design. Schazjmd (talk) 01:06, 5 February 2023 (UTC)
Comment Not sure where to place this, but somewhere above Qwerfjkl (I think?) was talking about the Talk page content parsing issue, and I just wanted to point out a case where this was done successfully. The {{Talk header}} which appears a the top of some Talk pages includes a set of "find sources" links that are influenced by what WikiProjects the article belongs to. To see it in action, contrast the "find sources" links viewable at:
That conversation was some days ago, so this may be passé by now, but thought I'd mention it just in case, so you can see how we did it. Mathglot (talk) 20:13, 5 February 2023 (UTC)
@Mathglot, yes, I ended up using {{Template parameter value}} as suggested by Certes. Note that this does not entirely work; it doesn't account for redirects, which is probably not that hard to fix. — Qwerfjkltalk 21:43, 5 February 2023 (UTC)
We faced the same issue of redirects, and handled it in subtemplates to avoid muddying the logic of the main. See e.g., {{Find sources/proj/is med}}. If you took that approach, given the 80/20 rule, you might have to create a dozen or two such (or one, long, #switch) to handle 80% of the WikiProjects with redirects, but that doesn't seem overly onerous. Mathglot (talk) 22:12, 5 February 2023 (UTC)
@Mathglot, it's just the redirects of WPBANNER we have to worry about. An easy solution would be {{for loop}}. — Qwerfjkltalk 22:16, 5 February 2023 (UTC)
We could change wikiproject banners that call redirects such as {{WikiProjectBannerShell}} so they call {{WikiProject banner shell}} directly. There are not many of them. Then {{Template parameter value}} seems good for extracting the article class value. Aymatth2 (talk) 22:43, 5 February 2023 (UTC)
We took a hybrid approach in order to keep the regex logic as simple as possible, altering pages using redirects with a reasonably small number of transcluders, while retaining unaltered the high-use redirects because of the excessive effort required to change them, and aiming the regex at those. Mathglot (talk) 23:04, 5 February 2023 (UTC)
Would it not be easier to just look for the first occurence of class= on the page and use that? — Martin (MSGJ · talk) 23:18, 5 February 2023 (UTC)
@Mathglot: Your approach may be useful in {{WikiProject banner shell}}, which will need to check if there are any wikiproject banners, and if not add the article to general categories like Category:C-Class articles. {{Find page text|%{%{wikiproject }} seems to do the job. Aymatth2 (talk) 22:43, 5 February 2023 (UTC)

@Abductive, Aidan9382, Andrevan, CX Zoom, Certes, Chipmunkdavis, DFlhb, Evad37, Femke, Fredddie, Guerillero, Kusma, LumonRedacts, MSGJ, Mathglot, Piotrus, Qwerfjkl, RoySmith, Rschen7754, Schazjmd, and WhatamIdoing: This has been a very productive discussion. Thanks to all for their ideas and solutions. I have opened a formal proposal at Wikipedia:Village pump (proposals)#Project-independent quality assessments. Any further input on the proposal is welcome there. Thanks, Aymatth2 (talk) 18:03, 6 February 2023 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

question re discussions of articles

hi. where is a good place to discuss individual new navboxes, articles, and other regular items that I have created? i would value some forum to discuss and to hear other's ideas. is the village pump a good place for this? if not.... then might it be possible, maybe, that perhaps we should either expand an existing forum, or else perhaps devise an entirely new forum, where such discussions could occur? open to any ideas or comments. thanks. Sm8900 (talk) 14:38, 7 February 2023 (UTC)

I think that depends on the nature of the discussion. Generally speaking, the talk page of the article, navbox, etc. is the preferred location, but discussions about things also take place in various fora like WP:AFD, WP:TFD, noticeboards, WP:DYK if you're nominating something for that, WP:Peer Review, etc. You really haven't given enough info in your question to be able to give a satisfactory answer. ~ ONUnicorn(Talk|Contribs)problem solving 14:46, 7 February 2023 (UTC)
And projects have talk pages that are good for discussing more general things, and also templates used in multiple articles. Graeme Bartlett (talk) 22:30, 7 February 2023 (UTC)

Search engines are encouraging article splits

If I search "Windows 365" on Google, I don't get a Wikipedia link in their "knowledge panel". That's because (until recently), Windows 365 was just a section of Microsoft Windows, rather than a dedicated article. That discourages mergers, even in cases where they make sense (for example, minor software companies, whose applications are just subsections of the company's Wikipedia article; we want the Wikipedia page to show up when people search the applications' names too, not just the company's name).

Is there a way we can fix that? Should we get the WMF to contact Google, to ask them to take into account redirect printworthiness for knowledge panels? DFlhb (talk) 17:06, 5 February 2023 (UTC)

How does this discourage mergers? If people use SEO based arguments during merge/split discussions, we should discourage the use of SEO based arguments. —Kusma (talk) 17:33, 5 February 2023 (UTC)
Nobody uses these arguments; but I suspect this is a factor in the tendency of some editors to prefer splitting articles, even when a subject could be better treated with a single article. DFlhb (talk) 17:38, 5 February 2023 (UTC)
@DFlhb, there are some old essays at m:Separatism and m:Mergism that may interest you. In the real world, the two views may be called Lumpers and splitters. Whatamidoing (WMF) (talk) 00:02, 8 February 2023 (UTC)

Kiwik - limited 2 way comms between talk pages

Kiwik is a palindrome, so it can be read in two ways - so it sort of suits a 2 way comm. I would appreciate peoples thoughts on this wish as I had a brainwave on the wishlist due date. It is an attempt to manage editor workload by being able to effectively ping a project rather than a person, and to create the basis for sharing watchlists and creating workflows. A kiwik template is just a wikilink pointing towards another talk or user page, that automatically creates a matching Kiwik link pointing towards the first page. The payload/message is in the wiki description. Wakelamp d[@-@]b (talk) 10:30, 9 February 2023 (UTC)

Update re WikiProject History

Just an FYI and PSA: WikiProject History is currently utilized mainly as a general noticeboard for history-related inquries, topics, requests, etc. anyone and everyone here is welcome to drop by with any editing initiatives, ideas, questions, requests, efforts, projects, or activities that they might like to suggest or propose there. we can use your input.

please note that this wikiproject is largely dormant, mainly because most history-minded editors tend to be more drawn to wikiprojects that pertain to more specific historical topics, eras, and regions, etc. if you want to see a truly overarching wikiproject for history that is fully active, check out WikiProject Military History. they are truly phenomenal. we try to do our bit for the war effort as well.   Sm8900 (talk) 20:27, 6 February 2023 (UTC)

Thanks for the reminder about this WikiProject's existence. Just joined! I'd love to see some sort of organized activity or initiative in the near future that could attract new editors and get it fully functioning. Thebiguglyalien (talk) 23:58, 6 February 2023 (UTC)
that's great. thanks for joining. feel free to post an intro on the talk page, just to say hello, or for any other reason. thanks! Sm8900 (talk) 14:59, 9 February 2023 (UTC)

Workshopping some guidance for content of Wikipedia articles about years/dates

Scope of the discussion

There's been some contentious discussions lately about the way to handle the content of Wikipedia articles about years; specifically about what kinds of content should or should not be included in said articles. Some of these have devolved into discussions about user behavior; this is NOT about that, please continue those discussions in other forums. When there is a conflict of this nature, it is a clear symptom that something is wrong with the system as a whole, and we need to correct that. This is designed to be an open space to propose solutions to the problem. The goal of this discussion should be to develop a clear set of guidelines dealing with calendar-type articles (specifically year articles like 2022 and day articles like January 1). To be clear, we are not yet voting on any such guidelines (that's not what the idea lab is for), rather this is a place to workshop some ideas for a potential future RFC, to be posted somewhere like WP:VPP, to establish and/or update existing guidance on how to manage the content of calendar-type articles.--Jayron32 17:40, 10 February 2023 (UTC)

Statement of the problem

Near as I can see, the main issue is that, at an article like 2022 or like January 1, we cannot include everything that happened in those time periods. The locus of the dispute has been on the birth and death sections of those articles. The article 2022 cannot possibly list all of the people with Wikipedia articles who died in 2022. It's too massive a number, Category:2022 deaths shows just how big the scope of the problem is. We can't just put all of those people in "deaths" section of the 2022 article. The problem is how to decide how to manage that section appropriately.--Jayron32 17:40, 10 February 2023 (UTC)

Proposed solutions

Ideally, this section is for proposing, commenting on, workshoping, tweaking, and discussion ways we can solve the problem. Good guidance allows us to move forward as an encyclopedia and reduces friction between editors. Lets figure out what that guidance should be. Near as I can tell, the way forward involves two possible types solutions; though these are not mutually exclusive and likely whatever we decide will involve some amount of both, and quite possibly something novel that I haven't even thought of, but will turn out to be a really good idea (I'm not that smart, and I'm sure someone else will cut this Gordian knot in this thread). Here are, what I see, are the two ways to fix this problem.

  • When a year article like 2022 is showing signs of becoming too large, we split out the "Births" or "Deaths" sections into separate articles on their own. This allows each of them to grow to a larger size, relieving the pressure to add more people where it's already too big. If those articles become too big, then we can further split them down by months. This would not need to be done for all year articles, just in situations where WP:SIZE is becoming a problem, so years like 1175, which is not overwhelmed by the number of people in those sections, would not need its own article. The benefit of this advice is that it pretty much exactly mirrors the existing guidance at places like WP:SUMMARY, which recommends splitting larger articles into smaller ones as necessary. The downside is that, in some cases, breaking this down into "Deaths in XXXX" or even "Deaths in January XXXX" might still be too large, and at some point we run into some of the WP:CLN issues where the lists start to mimic categories.
  • We develop some minimum standards for inclusion, as WP:N just produces a full list of all people who died in such a year with Wikipedia articles. Having some standards of inclusion more stringent than WP:N helps cut down on size, and having these standards be clear, written down explicitly, and verifiable by anyone heads off debate as to who is, or is not, "important" enough for the "Deaths" or "Births" section of a given year. There's some guidance started at Wikipedia:Days of the year#Births and deaths, which expresses that there needs to be a standard higher than "Has a Wikipedia article", however it doesn't seem to be adequate to indicate what that standard is.

My idea is that we ultimately use a little of both, but I want to stop talking now and let the discussion develop and see what ideas others have. That's all I have to say for now. Please feel free to comment on my ideas, add your own, or just discuss the issue in general below. --Jayron32 17:40, 10 February 2023 (UTC) Jayron32 17:40, 10 February 2023 (UTC)

Just a random thought, but when it comes to "size", are we worried from a technical perspective of loading too-large a page to edit? Or is it a reader concern for these specific articles? If it's the former, wouldn't one possible solution be to not break them into separate articles, but to break them into sub-pages that are then transcluded into the main page similar to templates? Or actually I guess you could do both; separate articles that are transcluded into the main article? To be clear I'm not against developing minimum standards, just trying to think outside the box a little especially if it's an editing concern.
In so far as listings go, I know obviously the types of devices used to view articles varies wildly (and mobile is taking up a huge chunk of that), but since births/deaths are just lists, perhaps we could use columns to make it take up less physical space on mediums that allow that (desktop/print)? —Locke Coletc 19:06, 10 February 2023 (UTC)
I think it's great that this is finally getting sitewide attention so a valid consensus can be developed. I'll briefly recap the things I've said on the matter to this point:
  • A year article should give coverage to subjects per WP:WEIGHT (and specifically WP:PROPORTION) based on WP:Reliable sources. There's been talk of using "year in review" articles from major news outlets to determine what should get the most weight in a year article.
  • A year article should briefly touch upon each topic of the year per WP:SUMMARY with links to further Year in Topic articles. It should broadly cover all aspects of the year, including info on major political developments, armed conflicts, and scientific advances, among other things. Essentially all of the things that are most heavily covered in reference to a given year. There should also be a discussion about how much culture should be covered, because the cultural elements of a year certainly have at least some weight in the year's coverage.
  • WP:PROSE favors prose over lists whenever possible. It's not a requirement, but I think there's a strong argument that year articles should be properly developed just like any other article. A few months ago, one other user and I wrote 2001 to demo this concept, and I think it provides better encyclopedic coverage of the subject. Separate Timeline of Year articles could exist if there's value in keeping the timeline format.
  • Regarding births and deaths, there is an ongoing RfC here to decide whether WP:SPLIT should be invoked for the births and deaths sections, and the consensus appears to be leaning toward support. WP:PROPORTION also applies, as "celebrity deaths" is not a major aspect when considering a year from an encyclopedic standpoint and should not receive undue coverage in an article about a year. I personally consider them to be WP:TRIVIA. Deaths that are major events, such as Assassination of Shinzo Abe, would still be included in main year articles.
  • Any lists split into their own articles, whether timelines or birth/death lists, would have to comply with WP:STANDALONE, which sets a basic sitewide standard.
There's a lot to consider regarding year articles, as their format really haven't been updated since before Wikipedia introduced its modern content P&G. The most important thing, however, is that a sitewide consensus is made before any particular standard or format is enforced across articles. I'm hoping that this discussion can form the WP:RFCBEFORE to make that happen. Thebiguglyalien (talk) 19:32, 10 February 2023 (UTC)
There is 'one' problem though, with getting as many editors as possible involved. The bigger the number involved & the bigger the number of options? the less likely a consensus on anything will be achieved. GoodDay (talk) 20:34, 10 February 2023 (UTC)
It would be more work, for sure. But Wikipedia is designed to build on input from many people, and it's not really a consensus if only a few people give input. My hope (and I believe Jayron32's hope) is that this discussion is open-ended so we see what gains traction, which can narrow down options and inform more specific decisions on implementation going forward. If anything, WP:CENT should probably be involved at some point, though I'll leave that decision to others. Thebiguglyalien (talk) 00:25, 11 February 2023 (UTC)
I don't want to be repetitive, so +1 to biguglyalien's post. At bottom, I think these year articles can be developed same as any other article: starting with a handful of sources ("XXXX in review") and following NPOV to determine what to include. And writing in prose. I think the best approach is to write one or two or three like that, and then see what we learn from that, and then we'll end up with some models to follow. There are also "Deaths of XXXX" sources that can be used for death lists (whether or not they are split). That'll probably still leave us with a few areas where we need to invent an inclusion criteria: (1) current year articles (no review sources yet), (2) list of births (which I think should be: anyone on a list of deaths also gets on a list of births, thus no living people on lists of births, but that probably needs an RfC). Levivich (talk) 22:43, 10 February 2023 (UTC)

Best thing to do, is open an RFC covering all concerns about the status of the Year pages. At the moment there's (IMHO) too many concurrent discussions, in different places. GoodDay (talk) 19:12, 10 February 2023 (UTC)

To be clear, I don't want to limit this to the Births/Deaths issue (not that I can; this is an open forum. I'm not a particularly important part of this process). It's just that the Births/Deaths issue was the main source of contention recently. We should feel free to opening this to the whole page standards, if that's what's on people's hearts to do. --Jayron32 19:17, 10 February 2023 (UTC)
Oh, just one more point. I do agree that focusing on Births and Deaths would be a mistake. The kind of controversies that have arisen over individual births and deaths have also arisen over individual events, but it's less noticeable because contributors are less likely to make changes to event entries in retrospect. Unfortunately, every time someone remotely noteworthy dies, his/her fans immediately leap to record it in the Year article, often not even being aware of the existence of Year in Topic articles where the entry belongs. They often proceed to include them in their year of birth as well, even if they don't have a reference for that. So we need to extend the discussion. Deb (talk) 14:38, 12 February 2023 (UTC)

My suggestion is to turn years into "portals" of events, ideally with these suggested sections: Politics, Economics, Culture, Science, Disasters, Sports, Culture, and a Timeline of widely covered events.

  • Politics covers topics like international relations, elections, wars, and NGOs.
  • Economics covers topics like the stock market, finance, interest rates, crypto, and money
  • Culture covers topics like entertainment, movements, changes to human society, and for recent years, notable trends on the internet.
  • Science is self explanatory: major scientific advancements and notable innovations which happened within the year.
  • Disasters is also self explanatory: wars, earthquakes and other natural disasters, ongoing genocides, and mass tragedies.
  • Sports covers FIFA, FIBA, the Olympics, athletic records, etc.
  • Timeline is like the "events" section from pre-prosification year articles. The proposed inclusion criteria is based on media coverage from multiple countries and if the event in question meets our other policies.

Anyways, these are just my thoughts. I would also encourage everyone to give consideration to the idea of putting the most notable deaths in a year a photo in the Deaths articles. InvadingInvader (userpage, talk) 01:54, 11 February 2023 (UTC)

I quite like this suggestion. My concern with the proposal of splitting the Year articles is that we may end up back at square one, with squabbles over which events and births/deaths should be included and still no firm consensus as to what the criteria should be (look back through the Talk pages and you'll find that I spent many years discussing these with other contributors and at the end of it very little was agreed). Deb (talk) 16:47, 11 February 2023 (UTC)
I'll try to get a draft out of this idea in a bit. A little busy IRL so I can't constantly write and improve, but I'll try to speed it up. InvadingInvader (userpage, talk) 00:41, 12 February 2023 (UTC)
Update: y'all can view my progress so far at User:InvadingInvader/Years draft. Nowhere near complete, but I thought I would share this now, and y'all are welcome to help. InvadingInvader (userpage, talk) 01:15, 12 February 2023 (UTC)
I suggest you take a look at 2001 a lot of what you're trying to do has already been done there. JeffUK 11:13, 12 February 2023 (UTC)
I'm familiar with it. InvadingInvader (userpage, talk) 04:47, 13 February 2023 (UTC)

Talk page template to record nominator and nomination timestamp for some article events

At the moment there is no reliable way to algorthmically determine the nominator and nomination timestamp for a historical good article nomination (failed or passed). The {{Article history}} template records almost everything about a good article, but not those two fields. Searching page history for a {{GA nominee}} template will find the answer in many cases, but not all GA nominations used that, and those that did don't always have a nominator parameter. There are other impediments -- vandalism to the talk page that removes the nominee template, for example, will make it appear that the subsequent reversion is the edit that added the template; that makes it look as if the reverting editor is the nominator. I posted a query to the article history template talk page a day or two ago, but got no response, and I think this might be the better page anyway.

I'm only talking about GAs back to about March 2008 -- before that date there were no GA subpages, and often a review would happen with no nomination request. There's no way to capture that information, of course. For GAs since March 2008, what's the best way to record nominator and nomination timestamp? I'm not familiar with DYK and ITN, but perhaps they have the same issue. Would extra fields in article history be the best way? Or a separate GA nomination template? Mike Christie (talk - contribs - library) 15:43, 14 February 2023 (UTC)

Is it really important for us to know that? --Jayron32 15:45, 14 February 2023 (UTC)
I think it can be useful. Both FAC and GAN (disclosure: this is partly through my own efforts) pay attention to the number of promoted articles and the number of reviews an editor has performed. Those statistics are used by some editors to make decisions about what articles to review -- see recent discussions on both WT:FAC and WT:GAN. For example, some editors prefer to review nominations by editors who have reviewed more articles than they have nominated -- reviewers are a scarce resource and these numbers help motivate nominators to review. The reviewing number has its own inaccuracies, which I'm working on correcting, but it's meaningless without an accurate nomination number. Mike Christie (talk - contribs - library) 15:50, 14 February 2023 (UTC)
And for a very concrete reason why these things are important to track, see Wikipedia:Good_article_reassessment/February_2023. This effort to clean up old GAs by a now-banned editor would be impossible without attribution of nominators. Mike Christie (talk - contribs - library) 17:15, 14 February 2023 (UTC)
But would we have even got as far as making that reason necessary if it wasn't for the general habit of thinking that someone who has created a few good articles must not be needed to be monitored when creating more? We need less looking at who has nominated, not more. Phil Bridger (talk) 19:16, 14 February 2023 (UTC)
Ideally, yes, but not having the ability to track this information wouldn't have stopped Doug, and it isn't stopping anyone else currently star-collecting. It just stops our ability to accurately report on what cleanup is needed. For example, that cleanup effort is focusing on Doug's promoted GAs. I don't know how many failed GAs he has nominated that aren't going to be PDELed because we don't know about them. I know of exactly one, because I failed it. We ought to be able to look up his failed GAs, which he would no doubt have filled with content of the same quality as his promoted GAs, and clean those up too. I don't know a reliable way to find those. Do you? That's what I'm asking for. Mike Christie (talk - contribs - library) 19:35, 14 February 2023 (UTC)
Without commenting on the general question, I think we should carefully consider using rare cases as a reason to build tracking mechanisms. New scenarios will pop up, but that doesn't necessary mean the best approach is to pay an ongoing cost. Sometimes building special-purpose tools/queries to deal with each new infrequent case is better.
Regarding this specific case, I think all content submissions by the editor in question are now suspect, whether or not they were in articles submitted for good article status. Giving priority to articles at GA status is just trying to keep the status meaningful. I'm not sure those that failed to achieve GA status are more important than those that weren't nominated and yet still could have problematic content. isaacl (talk) 22:42, 14 February 2023 (UTC)
Yes, that's a fair point with regard to that editor (or any similar editor) and their submissions. I would still like to see the data available for the first reason I gave. Mike Christie (talk - contribs - library) 22:47, 14 February 2023 (UTC)

search algorithm

Create a smarter search algorithm or improve the display of the output from the present search algorithm.

This is a problem that Google and other search engines have which make searching much more difficult by including items completely unrelated to what is desired.


I recently searched Wikipedia for "Vines of Belize".

I was amazed that the "found" items were anything that included the word "of".  Hence it included topics like Coinage "of" the world, Inflation, Deflation, Stagflation, and other topics worthless to my search.


In normal English usage, the phrase "of Belize" is a prepositional qualifier that limits the noun that it is qualifying.


Similarly, if I had search for "Belize vine", the word "Belize" is used as an adjective qualifying the noun that follows it.  I didn't phrase my search as "Belizean vine" because the word "Belizean" is unlikely to be used in the article if the article merely lists some of the countries in which the vine is found.


A smarter search algorithm would first search for the main noun (vine) and then eliminate those entries that don't include the word "Belize".


Instead, Wikipedia's search algorithm returns any entry that contains "Vines" or "of" or "Belize".


Although the above suggestion would require that your search algorithm be smart enough to understand word usage, there is a simpler implementation that eliminates having to be this smart about a given language.


Although not as smart as the change I'm suggesting, a simpler search would return just those entries that contain all 3 words, or at least order the list so that entries that contain all 3 words are listed first, followed by entries that contain just 2 of the search words, followed by entries that contain just 1 of the search words.


Another improvement could be made by eliminating all words that are articles (such as "a", "an", and "the"), eliminating connectors (like "and" and "or"), and common prepositions (like "of", "on", "in", "for", etc.). These words could be language dependent.


Wikipedia does not do this, and thus wastes their users' time by mixing it all up so people have to wade thru hundreds of irrelevant articles.

It just occurred to me that there is another change you could make that would not require that you change your current search algorithm !

Add an option to the displayed list of found entries, perhaps call it "Smart ordering" that takes the output from your search algorithm and reorders them to list those containing all of the search words first.

Personally, I wouldn't care how the remaining are ordered if the top entries contain ALL of my search terms, but the result would be much more valuable to the average Wikipedia user if the remaining were ordered by the number of search words actually found.

You might even include a check box by the original search to exclude words that are articles ("a", "an", and "the"), connectors ("and", "or"), and prepositions ("of", "for", etc.).

Or you could include this check box beside the "Smart ordering" button on the displayed results page.

-- Darrel Joy -- a frequent Wikipedia user 190.197.122.145 (talk) 19:56, 16 February 2023 (UTC)

Poll Readers/IP editors on Vector change

A few logistic questions first, since I know we can post banners, but don't know the limits.

  • Can we anonymously track response by users to a binary question? Anonymous being the same extent they are granted now as a passive reader.
  • Can we target IP editors specifically? Is there a part of the process they see but public does not? And can a banner/link/something be inserted there?

Proposal:

1. Put up for 72 hours a banner asking "Do you prefer the old look or new look"?

2. Should the banner be shown to all users or only ip editors?

Editors have other ways to change the look and give feedback. IP editors may not avail themselves and readers are not expected to. This seems a reasonable way to judge whether the dislike felt internally is equally as passionate outside. Slywriter (talk) 05:53, 17 February 2023 (UTC)

The Vector22 team are already planning on doing surveys of readers - they may already be active, come to think of it, given the timeline. I know that they held off doing it in the first week because they wanted to avoid change shock affecting the outcome in favour of V10 - I don't know why they couldn't do a survey in both 1st and 3rd weeks. Nosebagbear (talk) 09:52, 20 February 2023 (UTC)

Improved template messages

Hi there, I think that it would be a good idea to make the template messages able to link to specific paragraphs or text.


Thanks,


Blutankalpha Blutankalpha (talk) 12:20, 10 February 2023 (UTC)

Could you please provide some specific examples illustrating your concern? Template messages do often include links to shortcuts, which I would think is usually good enough. DonIago (talk) 14:07, 10 February 2023 (UTC)
Templates can only point to things a wikilink can point to, so they can point to sections (e.g. Saturn#Orbit_and_rotation) but not to a line of text on another page (which could easily be edited away). — xaosflux Talk 14:13, 10 February 2023 (UTC)
Ok, That is a good point.
This matter is closed
Blutankalpha Blutankalpha (talk) 10:27, 20 February 2023 (UTC)

Wikipedia:Notability (politics)

I think that this proposed guideline is ready for being a notability guideline. Can someone start discussion for making this proposed guideline into notability guideline? Thanks. ​​​​​​​𝐋𝐨𝐫𝐝𝐕𝐨𝐥𝐝𝐞𝐦𝐨𝐫𝐭𝟕𝟐𝟖🧙‍♂️Let's Talk ! 03:45, 18 February 2023 (UTC)

WP:PROPOSAL outlines the instructions. Curbon7 (talk) 04:14, 18 February 2023 (UTC)
... Didn't we just go through this? --WaltClipper -(talk) 17:13, 19 February 2023 (UTC)
Can we just not. --Jayron32 14:03, 20 February 2023 (UTC)

Avoiding driving off experts from Wikipedia

I fear experts may be disincentivized from contributing, or driven off by WP:FETCH-like behavior.

One problem, is that even with WP:V, nothing can substitute true expertise. I can read research papers, academic books, watch physics/math lectures, I'm still not going to be able to contribute to these topics with the level of understanding that an expert has. So when experts come here, contribute, and are reverted because of lack of sourcing, not because of any specific content objection, I think that's excessively burdensome. Why not just add a {{cn}} tag?

See this discussion, which seems representative of a more widespread problem: it focuses on procedural issues, instead of content (the only commenter who has discussed the content is the expert whose edits were reverted).

Interested in people's thoughts. Amusingly, WP:BURDEN and WP:PRESERVE were discussed at length recently (see here and here); but those discussions only barely addressed technical articles; most of ours are languishing, due to lack of expert attention, and I think we need to discuss that as its own subject.

How do we make it easier for experts to improve these articles, while minimizing the risk of unverifiable content being added? Surely, we have a ways to go. DFlhb (talk) 00:36, 4 January 2023 (UTC)

Are there any unique factors when working with an expert that would warrant its own policy considerations? In the meantime, I wonder if our policies need to be updated to more clearly state what is considered an inappropriate removal or restoration. And as a behavioral guideline, I think WP:DONTBITE should be applied a little more firmly. Thebiguglyalien (talk) 07:02, 4 January 2023 (UTC)
I agree that WP:BITE policy regardless whether of who. That said, WP:ACADEME is a nice essay that I share with academics who struggle with Wiki style of editing versus research paper. ~ 🦝 Shushugah (he/him • talk) 08:39, 4 January 2023 (UTC)
I sometimes point to Wikipedia:Expert editors. Gråbergs Gråa Sång (talk) 11:41, 4 January 2023 (UTC)
(1) Those are both very good links. I've been thinking about this question. I have given up editing WP in my own professional field following two events. The first was an attempt to improve an article on a method, where I encountered a bigger expert coming the other way, who politely refused all possible changes to the article. Basically they were editing from the university that developed the method in the first place, and were unable to accept the concept that anyone outside their own colleagues has made any significant contribution since, or that the method has in any way been improved or enlarged upon. I find this rather a pity, so the article remains as it was: good, but basically a fossil, remembering a happy past. The second was an attempt at cleaning up another article, which is full of references to minor primary research papers containing good(ish) ideas that went nowhere, while missing quite important concepts now used in commercial scientific instruments. I became frustrated because it was clearly a target for semi-knowledgeable grad students and early-stage professionals, using a bit of Wikipedia editing as part of their training, and unable to realise that every research paper on the subject claims that it's an important advance. Meanwhile I couldn't add the things that are really being used, because the best sources for them are heavily linked to the manufacturers, so they're not seen as independent (Wikipedia has a very strong nose for even the slightest commercial COI). After a while, I realised that as an educational resource, Wikipedia is pretty rubbish: it's better to devote one's time to institutional and personal web-resources. Which brings us to point 2:
(2) Few experts understand that Wikipedia is an encyclopaedia. It's supposed to be readable by any person of normal educational and academic background. It's not supposed to be a specialist reference resource for use by professionals and experts. There are big divides on this by subject area. For example, look at Mensural notation, a very technical and complex subject in music. The article is readable, and makes sense, requiring little previous knowledge. Now look at Integral, which is a very basic concept in maths. The article kicks off with the first sentence of the lead: "In mathematics, an integral assigns numbers to functions in a way that describes displacement, area, volume, and other concepts that arise by combining infinitesimal data." Basically unless you know what an integral is, that sentence might as well be written in Chinese, but if you do, it makes perfect sense. The whole article is a subtle fight between those who wanted to explain integrals in a way that an average shop assistant could understand if they wanted, versus those who wanted to make sure that every sentence conforms to a mathematician's 100% rigorous approach, and the result is an article that's worthless: it doesn't tell an expert anything they didn't know already, and it's a very poor way for anyone else to find out what an integral is. In many ways, I wish experts would stay away unless they are both experts in the subject matter, and in writing about it in a non-technical way. I suspect that many experts see the subject from too close, and those who are good at writing encyclopaedically about their subject don't do so here, not only because they're fed up of being bitten and challenged by non-experts, but also because they've realised there are better places to do it.
(3) The OP also touched on the matter of citation, and suggested we should be a bit gentle with experts who are writing from knowledge, writing correct information, but not citing. I'd advise avoiding discussing this at the same time as discussing how to include experts anyway. Unfortunately WP has two groups of editors: the delete-uncited-on-sight, and the don't-disrupt-the-flow-by-deleting-correct-statements people. Both sides are utterly convinced that they are Totally Right. Any attempt at discussing the subject merely ends with both sides declaring there is no point in discussing it because not only are they Right, but everyone knows they're Right. It's a doomed discussion, and poisons discussion of anything else. I'd keep it separate! Elemimele (talk) 13:21, 4 January 2023 (UTC)
Touching on part of your theme, after first trying to edit in the area in which I received my PhD (50 years ago, but never worked in it), I decided I couldn't keep up with the grad students in the field. Similarly, I have stayed away from the field I worked in for 20 years before retiring. Now, I edit what interests me, and don't recall running into any editors who claimed to be experts in the various areas I work in, although experts would be welcome to help sort out competing and ambiguous sources. - Donald Albury 20:28, 4 January 2023 (UTC)
Personally, I'd hesitate to extrapolate from the sampling of experts who have edited Wikipedia in the past to generalize that very few experts understand that Wikipedia is an encyclopedia targeted at the general public. There are plenty of experts who understand the need to tailor messages for the intended audience. For those making a living in their field of expertise, I agree that that in many cases, there is limited upside in contributing to Wikipedia, versus finding other outlets for public education and potentially becoming a source cited by Wikipedia. isaacl (talk) 22:18, 4 January 2023 (UTC)
I don't think interaction with experts is the main problem; it's more about reducing the risk of their edits getting lost in the fray. Probably the biggest disincentive for contributions is the idea that "it won't matter, 'cause it won't last" (some likely even would add: "and a nonexpert would eventually destroy my improvements anyway"). Maybe having a guideline recommending that these reverted edits be placed in a talk page header, so they can get reviewed, even a year, 3 years, 5 years later? (better late than never). Or, adding a link to the talk page header with recommendations specific to expert contributors: "state you're an expert, say what's wrong, and give a source that could be used to fix it", to let them know that they can help us a lot with just a few minutes of their time, without them feeling they need to learn how Wikipedia works and just not bothering.
Or having some way for experts to contribute to articles in their userspace, not have these drafts deleted after some arbirtary time, and similarly link them in a talk page header so they can be reviewed at some point? These types of more m:Eventualist approaches seem like they would be most fruitful, without treating expert to lower contribution standards than other editors. DFlhb (talk) 20:36, 4 January 2023 (UTC)
But how would you deal with the fact that we have no mechanism for verifying who is, indeed, an expert? You cannot label an edit as being made by an expert unless you have verified that the user making the edit is indeed an expert. After the Essjay controversy, Jimmy Wales proposed that Wikipedia adopt a system for verifying experts, but the community said no. While that was 15 years ago, I would be surprised if the community is ready to officially recognize experts. If it is ready, there is the question of the bureaucracy that would be needed to administer it. Donald Albury 21:04, 4 January 2023 (UTC)
I love Jimbo's idea. One mechanism might be the VRT. Also just learned there's a relevant draft proposal on verification; posting in case others haven't seen. DFlhb (talk) 21:24, 4 January 2023 (UTC)
But there are so many huge problems. (1) lots of experts won't want to be identified or identifiable, or 'outed', and will not get involved; (2) if experts can be validated by staffers in private, how do we have the transparency to know they're really experts, or do we have an unknown clique with special editing privileges; (3) how do we retain casual-experts who happen to spot an error while drifting past as readers, and correct it (often as an IP editor); and (4) do we actually want an encyclopaedia operating in Britannica mode, written by experts rather than everyone who can find a source? It's a good beast, but a different beast. You might end up with a not-very-good encyclopaedia written by the sort of ex-expert who has time on their hands and nothing much else to do because they're not actually all that good at what they do. Elemimele (talk) 21:48, 4 January 2023 (UTC)
The problems are so huge that they would completely change the nature of Wikipedia, and undo over two decades of work that has pretty much put traditional general encyclopedias out of business. For verification to work it would need a huge bureaucracy to support it, and you would still have the problem that extremist POV-pushers would constitute the majority of "experts", as others would not be prepared to go through this process in order to do voluntary work. Self-certification would be even worse, as many people over-estimate their expertise. Phil Bridger (talk) 22:18, 4 January 2023 (UTC)
This seems like a pretty strong misrepresentation of the proposed idea. You're implying that it's some overhaul of editing when it would just be a way for editors to verify their credentials if they liked. Also, there are already systems in place for verifying private information to the WMF. Thebiguglyalien (talk) 22:46, 4 January 2023 (UTC)
Just want to add to this: we're going back into the "it's an interaction problem" territory, the opposite of what I support. The larger goal is to fix the issue of good ideas being almost "lost to time" in talk page archives or revision histories. The credential verification isn't even a requirement for my ideas; I'm brainstorming ways to make m:Eventualism work better; that's all. DFlhb (talk) 23:27, 4 January 2023 (UTC)
The problem is that the only decision-making mechanism on English Wikipedia is based on consensus. If some edit is identified as a "good idea", then editors will work at putting it into the article now. It's operationally difficult to maintain a list of ideas that aren't determined to be good now, but that a different set of editors might think are good in future, because any edit can meet that definition. isaacl (talk) 00:02, 5 January 2023 (UTC)
Citizendium is an object lesson in recognizing experts. Articles could be written by "authors", but only "editors" could approve an article. In academic fields, editors had to hold a PhD in the field, and be working in the field they were an editor for. In non-academic fields they let someone who had published articles on the subject be an editor. Even though I have an earned PhD, I could not have been an editor for that field because I had not held a position in the field. Check out [citizendium.org Citizendium] and see how well it is doing these days. Donald Albury 01:48, 5 January 2023 (UTC)
The Essjay rule still applies though. If you edit an article and are not an expert, then you are a sock, and will be dealt with accordingly. The level at which articles need to be pitched is always problematic; there is no micro-Wikipedia. Nor are we in the business of lies-to-children. We try to pitch the article at the general reader, but we also know that the more complicated a subject, the more likely it is that the reader has expertise. Anyone looking at an article on integration will be at least a high schooler, for that is when the subject is taught. My third grade math text said: "a circle is a set of points". What a mind-blowing concept! So if it is good enough for the third grade, the rest should have no problems. Hawkeye7 (discuss) 03:20, 5 January 2023 (UTC)
The reader I have in my head when I look at Integral is a high-school kid's grandmother: the kid has come home talking about integrals and she wants to know what the kid is going on about. She's intelligent, but her maths education happened 60 years ago (and at a time when girls were expected to cook, not integrate), so she turns to the world's best general reference book for help. It'd be disgraceful of us not to do our best! But what's she going to make of the first sentence of the section on Lebesque integrals: "It is often of interest, both in theory and applications, to be able to pass to the limit under the integral"? Lebesque did a better job of explaining in terms of loose change in his pocket. The diagrams in the article are much, much better than the text. To write about integrals in an encyclopaedic way that is useful to an intelligent grandmother you need someone who's an expert not only at maths, but also at little old ladies. Elemimele (talk) 14:26, 5 January 2023 (UTC)
There are multiple issues raised here. The first is that while it'd be disgraceful of us not to do our best!, our top level articles are seldom examples of our best work. They are very hard to write! So the experts prefer writing up more specific but more manageable topics. (I intended to rewrite one within my own field of expertise over the holiday, but found it more congenial to write about the guy who built Disneyworld.) The second issue is how we can cater for the level of background knowledge of the reader, which determines what information they are seeking. There are three cases in your example: the grandmother (who sounds very much like my own, who attended a domestic arts school back in the 1950s), who has little background; the high schooler, who would be in year 10 or 11, when the topic of integration is introduced; and the college student, who will be encountering the Lebesque integral. (A crucial concept, as noted above, was slyly slipped into the third grader's text, but this was part of the New Math movement of the 1960s.) The Lebesque integral subarticle can assume that level of knowledge; in most cases, the more specific an article, the more we can infer about the reader. But what about this article? We Wikipedians know that "Formal definitions" means "college level math in this section" but most readers don't know that. We now have a tug of war among editors that is common to many mathematical articles. What is the logical ordering? I would argue for pushing that section down the article, and bringing the section on the Fundamental Theorem of Calculus (which the high schooler will encounter) up. But other editors will argue that the ordering is more logical the way it is: with the detailed proofs and concepts coming first. In other words, the issue is pedagogical, not mathematical. Hawkeye7 (discuss) 20:41, 5 January 2023 (UTC)
That last sentence is spot on. The most valuable experts here are those who consider the pedagogical side. Elemimele (talk) 06:57, 6 January 2023 (UTC)
I think user boxes were supposed to help with encouraging multiple people with experience and expertise to commen, and know they will be supported. But if you ask for help on a user talk then you could be accused of spamming. Wakelamp d[@-@]b (talk) 03:30, 2 February 2023 (UTC)
As a PhD-possessing expert who edits almost exclusively in my field of work and study, I'm not really sure that this is really a problem of Wikipedia editors driving off experts. I think these things are true but I'll only speak for myself: This is a really weird place to write and it's different from how I write in almost any other context. Moreover, because most of my interactions are with people with whom I have no relationship and have much less familiarity with the subject, I find myself explaining things over and over again, sometimes things that are glaringly obvious to me and my colleagues but unknown outside of those circles. For example, a few days ago an editor was asking me why we rely on the Carnegie Classification of Institutions of Higher Education to label some U.S. universities as "research universities;" it's a very legitimate question and one that I should be able to answer but it's questioning such a basic, common practice in my discipline that it was equal parts frustrating and amusing (amused at myself for struggling to answer such a basic, reasonable question - not amusement at another editor's ignorance!). Similarly, I've reverted many edits made by editors who confuse a capital campaign with an endowment; another very obvious distinction to me but clearly not obvious for many other people.
In my discipline, I think that we bear most of the fault for not wanting to engage here and contribute to this public good. This is a weird place with a community and practices unlike any other so it takes a lot to stick with it and learn your way around. There are certainly things we can do to make it easier for new editors. But I haven't experienced much that is specific for experts who are new editors that must be changed. Deferring to someone else's expertise without evidence is definitely not the way to go. It would certainly make my editing here easier in some instances but it would be a massive change in our foundational culture and practices, a change that I would not support and I doubt would garner significant support project-wide. ElKevbo (talk) 04:12, 6 January 2023 (UTC)
Some of you may also be interested in Wikipedia:Teahouse#Non-expert review guild? by GuineaPigC77. Whatamidoing (WMF) (talk) 18:13, 6 January 2023 (UTC)
I think the focus on "real" experts, credentials, whether they can talk to non-experts, etc. is perhaps a bit beside the problem exposed in the original post. Those of us who remember the Web before it had version numbers (through, say, the early 2000s?) will recall that a significant proportion of it used to be individually-written web pages on the author's little niche interest. (A representative sample for the youngsters.) Sometimes the people who wrote these sorts of things were Genuine Certified Experts as regarded the subject they were writing about; more often, perhaps, they were adjacent to the subject. e.g., the person writing lucidly about the Lebesque integral and its applications might be a programmer implementing a math library rather than a professor of mathematics. It might be someone with no formal credentials at all about the subject, but a passionate amateur student; superficially interested people generally couldn't be bothered to engage deeply enough and for long enough to do this sort of thing. I think Wikipedia absorbed a lot of that passionate amateurism, and rechanneling it was responsible for a lot of our early growth.
Unfortunately, as Wikipedia has become a load-bearing part of the noosphere, we've had to face an increasingly complex threat model. Many more people now edit Wikipedia, not out of a sort of naive enthusiasm for knowledge, but because of a desire to promote (or suppress) some person, organization, or ideology; and the widespread consumption of our information means that errors, even those made by a well-meaning but ignorant enthusiast, can have a great deal of impact. As a result, the way we interpret content policy has become increasingly rigid and compulsive, and focused on protecting us from the lowest common denominator editor. Even if the modal editor is a crook or an idiot, firm application of policy will (we hope) result in them creating accurate articles, will they, nill they. The problem is that this general trend in policy and the interpretation of policy is paid for by the slow, gradual immiseration of editors who are knowledgeable about a particular topic. When trying to make a specialized topic intelligible to a lay reader, you will almost always find that certain pieces of disciplinary knowledge, like those ElKevbo mentions above, are assumed to be understood by the reader of the reliable sources you are using to write the article, and you will not be able to cite a clear, explicit statement of that piece of knowledge from the source. The conscientious editor will find that the particular statements drawn from the reliable sources are intimately intermixed with background knowledge, and is faced with an extended hunt through peripherally relevant sources to gain an explicit warrant for those pieces of background knowledge. It is, frankly, exhausting, a strain on working memory, and deters sustained contribution.
I don't know what to suggest as a solution. These changes to policy happened organically, and for a reason. But I think our current approach will tend to be self-perpetuating; having adapted our policies to editors who can't be trusted to know what they're doing or to act honestly, we will select for an editor base of that type. Choess (talk) 03:50, 7 January 2023 (UTC)
I understand your comment about filling in the background. Much of my editing is about history and/or archaeology. As I dig through sources I run into events, places, and concepts I have not heard of, and which are mentioned only in passing in a source, and for which no article yet exists in enWP. So I try to fill in those gaps. Such attempts all too often turn into a descent into a multi-branching rabbit hole. Donald Albury 23:26, 7 January 2023 (UTC)
I was just checking through my notes for MSc lectures I'll be giving shortly, and note that I explicitly warned last year's students not to read any of the relevant articles about the subject on Wikipedia, as the articles are riddled with errors, out of date, and full of trivialities that went nowhere; instead I furnished the students with a list of mainstream textbooks and review articles, and links to generous professors in the US who've put good teaching materials on their own websites. It's a bit depressing reading my own opinion. But there's no way I'm going to try to clear up that mess. I hope (and genuinely believe) that my subject is particularly badly covered, and that I'm not misinforming myself when I use Wikipedia as a reader on other topics. The trouble is, situations like this make me wonder whether other experts feel the same about their fields, and undermine my faith. Elemimele (talk) 13:29, 9 January 2023 (UTC)
I'm not an expert in the areas that I mainly edit in, so I rely on the reliable sources I can find (including the books from academic and other reputable presses I have accumlated in the last 15 years). I do find when working on existing articles that much of the content is not supported by citations, or is supported by citations to blogs, promotional sites, well-meaning but ill-informed "official" sites, or off-line sources that I am not familiar with, and cannot find coverage of on-line. All too often, the cited sources do not support, in part or at all, the content preceding the citation. I also look back at my early work on WP and cringe. In one early article I got a city name wrong. I saw the mistake seven years later, and after searching to see which editor had introduced the error, was embarrassed to discover that I had made the error when I wrote the article. Yet, I use WP all the time to look up something I know little or nothing about. I will also continue to do what I can to improve the content of WP, however little that may be in the grand scheme of things. Donald Albury 17:43, 9 January 2023 (UTC)
You've put your finger on one of the problems that some technical experts have here: that they encounter non-experts relying on reliable sources. But the non-experts don't realise that you sometimes have to have some expertise in the subject to recognise a reliable source. The articles in my field, an instrumentation-related branch of science, are riddled with trivia fished out of journals that might as well be re-titled "Annual Reviews of Whacky Ideas that Went Nowhere", but because they're peer-reviewed review articles, they're automatically deemed Reliable. Some of these ideas are 10, 20 years old but no one has ever developed them any further because they don't work. No one ever publishes a subsequent article saying "we read this idea and tried to do it, but it failed", that's not how publishing works. Instead, the fringe ideas just fade away. If you're writing a review for professionals, people expect to be told a few whacky things that they don't know; professionals know this, and don't expect the ideas to be mainstream. But non-experts don't know, so all this stuff gets trotted out in a Wikipedia article as though it were the bread-and-butter of the subject, with not the slightest attempt at distinguishing between what's done and what someone once briefly thought might be worth a try. Meanwhile, I found when I first tried to add some information about what people actually do, the real stuff, it would get reverted because the sources that best support this are often produced by the producers of the instruments, and so it's deemed non-independent stuff, tainted by commercialism. Or it's teaching information produced by labs that do it, in which case it'd be reverted as a "blog". But very often there are five manufacturers all writing more-or-less the same thing because they're describing, accurately, and up-to-date, what is actually done (i.e. it's not really non-independent, because since they all write the same thing, they might as well be writing about one another's instruments, and not their own), or multiple labs, of very high-quality output, are producing similar teaching material, so their "blogs" are mutually supported by the fact everyone who knows anything is saying the same thing. Wikipedia is like a bunch of people who want to know how a back-hoe works, but who refuse to listen to JCB, or a group of back-hoe operators, because they are utterly convinced that a guy called Bert who loves going to truck shows and wrote a book about it is somehow more reliable than a team of experts who actually build the things. And that's a problem. But I do think this depends enormously on the field concerned. There are almost certainly some fields where genuine experts are rare, and well-informed amateurs armed with good sources might be better. But again, it's really hard to assess your own ability to edit in a field without having the expertise. In a sense, the whole of Wikipedia is founded on ignoring the Dunning Kruger effect, and bits of it seem to get away with it quite well! And sorry, I'm rather changing views on experts here, having previously complained that experts are sometimes rubbish at explaining their subject. Elemimele (talk) 23:13, 9 January 2023 (UTC)
Is it because the experts get outnumbered? It is such a pity that contacting editors that have expertise listed in their infobox is spamming :-( Wakelamp d[@-@]b (talk) 12:21, 24 February 2023 (UTC)

This is an excellent way of putting peer-reviewed content over here, "Category:Wikipedia articles published in peer-reviewed literature", and then that little open book sign appears on the right where the GA and FA symbols are; this somehow seems to connect to this thread on experts. Though this does seem to address expert-level-content more than experts-as-individuals. FacetsOfNonStickPans (talk) 11:43, 10 January 2023 (UTC)

How long does it take them to figure out I'm not actually an expert in their field? Hawkeye7 (discuss) 17:43, 12 January 2023 (UTC)

One other factor that drives experts from Wikipedia -- & IMHO prevents more than a limited amount of serious development on any subject -- is the lack of any real payback for the work. Outside Wikipedia if one writes an essay or monograph on a topic, the writer expects to receive something in return: money, or credit, or simply ownership. Instead a Wikipedian donates their time, expertise, & incidental costs of writing an article to the project, after which it becomes the property of everyone, & (as the slogan reads) "anyone can edit it". Yes, we contribute to Wikipedia out of love, but unless a contributor gets something tangible from the contribution this work is in effect unrequited love. The result is that only a fraction of the already small group of contributors will doggedly fight to keep an article usable & the information correct, & even then (as pointed out above) those few may have a subtly incorrect or out of date understanding of the topic.

While this does not discourage any contributions, it has a dampening effect: one is going to limit ones editing time, research time, & enthusiasm for a given topic if they are fighting ignorance without any recognition of their efforts. Articles will reach a point of improvement, then stay at that point -- or degrade if the original author has moved on -- due to this lack of recognition & the policy about WP:OWNERSHIP. Now I'm not saying we should throw out those policies, after all this concerns a central part of Wikipedia culture, & we can all point to instances where this required radical altruism has helped Wikipedia. However, having been made aware of the negative impact this required altruism has, perhaps we should think about loosening this requirement. Or accept Wikipedia is doomed to being only so useful. -- llywrch (talk) 19:48, 20 January 2023 (UTC)

I am joining this conversation a little to late, but as an expert that left wikipedia, so I thought I would share my experience. In my case, a fellow expert who wanted influence articles from a particular perspective. Eventually things escalated to the point that editor decided to publish papers, in tangentially related journals not really in the subject area. For example something sociology or philosophy of mathematics than rather than mathematical statements. Friends of his would then change articles to the desired POV and with an citation that said exactly what would dovetail into the WP article. As is pointed out above it takes "some expertise in the subject to recognise a reliable source". Editors who were not experts, understandably felt that the point of view was verifiable and notable since it was published, and even I have to admit seemed custom written for the wikipedia article. Since it grew out of the dispute this made sense. I felt my choices were to go down the path of writing my own crazy articles, or just accept this.
As a side note, a lot of the disagreement was about when/where/how do discuss infinitesimal ideas. I am personally horrified to see it pointed out that the jargon into the lead sentence of integral. But in terms of full disclosure, I have edited that sentence (see here if your curious) so I shouldn't be considered an unbiased observer. And I am not interested in reopening old grievances. Thenub314 (talk) 15:58, 9 February 2023 (UTC)
Because so many editors would rather revert than add, e.g., a {{cn}}, I have decided that WP:BOLD is often not tenable and use the talk pages heavily prior to putting in the time to do major edits. I really wish that MOS would discourage reverts for reasons where adding a template is more appropriate.
Also, there are subjects will I will no longer contribute because some editors have created a hostile work environment, A more robust dispute resolution mechanism, not requiring consent by all editors involved, would help. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 01:34, 13 February 2023 (UTC)

I don't think this is going to actually happen, but I have had the following idea for a while: the Wikimedia foundation should hire a small group of academics as paid editors, who I call chief editors. In the very early days of Wikipedia, Larry Sanger was a paid editor who was in the position of editor-in-chief. The chief editors would kind of play a similar role. Their primary role would be in making editorial judgement of on which reliable sources should an article be based on. As pointed out above, consensus is not a good way to select the best sources for articles; especially for scholarly articles, the approach based on the credentials is better. The chief editors would be hired by the foundation, perhaps based on the nominations/recommendations from the Wikipedia community. We the community should nominate those who have a good of understanding of NPOV and have some academic degrees as well. Getting paid from the foundation should also help ensure avoiding the COI concern. Anyway, just some idea. —- Taku (talk) 13:35, 13 February 2023 (UTC)

AGREED - DFlhb, I would never recommend anyone to this site. It doesn't matter what you know or what you can prove, the only thing that really matters is whether you have more privileges than another user or have a friend who does. The person with the most power almost always wins at the end of the day. I won't name names, but I know some users here who have retired more people than Mike Tyson did by coming up with endless challenges and objections to the point of making users quit. The site is about 20% of people who are interested in helping other people learn about subjects and 80% of people who use this site for power trips. KatoKungLee (talk) 02:41, 24 February 2023 (UTC)

Concern about biases towards older, foreign and obscure topics

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


I like to write a lot about older and foreign things. However, this nearly always ends up leading to a few problems:

- The internet wasn't really used widely until the last 20 years. Therefore, there's not a lot of sources for various things like there are now. Take a sporting event for example. I can go to multiple local news sites or national sports sites and find results of various games. I can get takes on how everyone performed, along with videos and interviews to further prove this. That's not really possible when you go into things from prior to the 1990s because....

- If you want to find out about things that weren't major general events prior to the 1990s, you are going to have to get a book, magazine or newspaper. Finding any of these things from prior to the 1990s is going to cost you money unless you get lucky and it's in a library or online for free. Worse, when it comes to newspapers, you likely cannot even find copies of various newspapers and such anywhere due to people not saving them or not releasing them publicly.

- Now these two things are problems, but then we dip into problem #3 - foreign sources. Finding the book/newspaper and hoping it covered what you wanted (which is an expensive guessing game, because you don't know what's in the book until you have read it) is hard enough, but then you need to be able to read it. Sure, if it's in German or Spanish, we can probably use google translate, but if the source is in a language like Chinese or Japanese, you can't type in Chinese/Japanese on an American keyboard without knowledge of the language. And in general, a lot of the world is still not online which limits possible sources. The vast majority of internet articles online are coming from richer countries where more people have computers and the internet.

Overall, I think Wikipedia needs to change how they view articles on older and foreign topics. It's discriminatory on money, location and language and for article writers, you are already kind of facing the gauntlet to begin with against people with more power than you. KatoKungLee (talk) 19:02, 22 February 2023 (UTC)

  • In what way does "Wikipedia needs to change how they [sic] view articles on older and foreign topics"? We aleady have the policies WP:NONENG and WP:SOURCEACCESS. Phil Bridger (talk) 19:36, 22 February 2023 (UTC)
    • Phil Bridger - I think having less scrutiny on older and foreign topics would be helpful since so much of it requires spending money or being in various locations, which are both discriminatory. Many books are expensive. Many books are out of print and difficult if impossible to find. Some publishers don't want to ship to other countries either and many types of media were never made available outside of their country. It would be easier and cheaper for me in many situations to write my own book based on sources that would not be allowed here than to actually track down the original books that they came from. It sounds odd, but it would work.
While the resource request section is cool, I would guess that 99.9% of resource requests are never granted, making it unreliable as a tool to get resources from.KatoKungLee (talk) 20:00, 22 February 2023 (UTC)
I've always been successful when using Resource Request, and looking at the latest archive nearly ever request is resolved, so I don't know where you're getting the 99.9% figure from. -- LCU ActivelyDisinterested transmissions °co-ords° 15:16, 24 February 2023 (UTC)
  • I would prefer if editors used books and newspapers far more than they currently do, regardless of the topic. There are many places to get them for free. We even have WP:The Wikipedia Library that provides open access to a ridiculous number of articles and books for regular editors. Beyond that, Archive.org is an invaluable resource for books, especially older books. Thebiguglyalien (talk) 20:07, 22 February 2023 (UTC)
    • Thebiguglyalien - Newspapers are tough. You have to be in the area to get access to a lot of them and they often don't have full collections. And who knows what the situation is like in non-US countries, since you would have to know the language to get anywhere with it. You also have to hope it has what you are looking for, since there's no guarantee they covered it. While archive.org and other sites are great, the selection is also not complete. KatoKungLee (talk) 20:39, 22 February 2023 (UTC)
      newspapers.com has absolutely extensive records of newspapers in the US, and I know there are similar repositories for the UK and Aus. Curbon7 (talk) 19:57, 24 February 2023 (UTC)
  • It would be a deep disservice to older, "foreign", and obscure topics if we ran a policy of allowing their articles to be inherently less reliable or inherently lower quality. CMD (talk) 02:56, 23 February 2023 (UTC)
  • Indeed, I raised this subject here. And this could also happen to plot summaries in TV/movie articles older than around 2000 pretty easily. I'm not saying to throw out our sourcing guidelines, but WP:NODEADLINE and use common sense. --Rschen7754 04:11, 23 February 2023 (UTC)
    • TV plots are a good example. The only way you could prove that the episode was about something was if you had a TV Guide. Collecting those to prove a summary is an expensive task when anyone who has seen the episode would know what it's generally about. The people who are going to challenge it though would rather raise questions about it then spend 10 minutes and find out first hand.
    • A general common sense rule is also badly needed here as some of the most vocal people here seem to have a total lack of it.KatoKungLee (talk) 14:18, 24 February 2023 (UTC)
      Summaries can be referenced to the work, this is common practice. As to this idea it has been floated in one form or another multiple times. If anyone had any common sense they would let it go. -- LCU ActivelyDisinterested transmissions °co-ords° 14:21, 24 February 2023 (UTC)
  • WP:V is non-negotiable, regardless of the subject. Using non-English or difficult to access sources is fine, but sources must exist. If no sources exist then the content is WP:OR and should be removed. -- LCU ActivelyDisinterested transmissions °co-ords° 20:41, 23 February 2023 (UTC)
    Agree. This proposal is basically saying “researching obscure topics is hard, can we pretty please cut corners?” Which the answer to is obviously no. Dronebogus (talk) 20:50, 23 February 2023 (UTC)
    At this point I've likely spent 40+ hours searching out sources for obscure language orthographies, finding such works isn't easy but referencing isn't optional. -- LCU ActivelyDisinterested transmissions °co-ords° 21:02, 23 February 2023 (UTC)
  • I don't think KatoKungLee is suggesting to ignore the verifiability guidelines, but rather, to lesson the notability restrictions on people with significant accomplishments from foreign countries in the pre-internet era, for which finding sources is a very difficult task. BeanieFan11 (talk) 21:55, 23 February 2023 (UTC)
    • BeanieFan11 - This. I can go to the local library and pull up dozens of sources on a local figure from the 1930s to make them seem notable, because local libraries have the old newspapers, books and magazines that would probably have that information. If I went to the same library to get that info on a local figure in Uzbekistan in the 1930s, I'm not going to find anything. Getting the books or sources needed for that person would require me to either spend a ton of money or to be in Uzbekistan to do research, and I don't have the resources to do something like that. And even if I did, I'd still have to be able to read Uzbeki to a degree to find information. Essentially asking someone to visit a foreign country, purchase expensive books or know a foreign language to submit an article is insanity. KatoKungLee (talk) 14:30, 24 February 2023 (UTC)
      • On the contrary, I would consider it much more insane to accept an article written by someone who doesn't have access to reliable sources on the topic and wouldn't understand them if they did. Caeciliusinhorto-public (talk) 14:41, 24 February 2023 (UTC)
        • Caeciliusinhorto-public - How do you get the local sources if you don't speak the language or aren't in the area? You have to hope someone else writes about it or does the work for you. Relying on secondhand information isn't that good either. Original research isn't allowed here and publishing first hand accounts isn't allowed here either.KatoKungLee (talk) 14:57, 24 February 2023 (UTC)
          • You've essentially answered your own question. If there's a subject you can't write about, then you let someone else do it. If no one else does, then it doesn't get covered. It's an unfortunate truth, but that doesn't make it less true. There are things we can try to do to mitigate this, such as recruiting more bilingual editors, but non-Anglosphere topics will always be slightly more difficult to write about than Anglosphere topics, and that's something we have to acknowledge and work with. Thebiguglyalien (talk) 15:59, 24 February 2023 (UTC)
            • I think having some information about a subject is better than having no information at all. You can always add more to an article and just having it invites people to improve it. Nobody can improve an article that doesn't exist and with increased scrutiny on here, less articles are getting published period.KatoKungLee (talk) 16:59, 24 February 2023 (UTC)
          • KatoKungLee exactly my point. If you don't have access to reliable sources you shouldn't be creating an article. Caeciliusinhorto-public (talk) 16:00, 24 February 2023 (UTC)
            • Not only do people's ideas of reliable sources differ on the site, but people's knowledge of reliable sources also differ on the site. An article can always be improved, and the articles existence usually invites improvement. An article that doesn't exist doesn't really encourage much group research.KatoKungLee (talk) 16:59, 24 February 2023 (UTC)
              • Existing articles can also be made worse. In fact, my experience suggests that most articles on obscure topics don't change very much at all for long periods of time until either (a) someone comes along with access to actual reliable sources and makes a concerted effort to make the article better or (b) someone comes and inserts misinformation, making the article worse. The first case is not made easier by weakening our requirements for reliable sources, but the second is. I would argue that an article which misinforms readers is worse than no article at all, so any proposed reform which makes it easier to add misinformation is one which makes Wikipedia worse, not better. Caeciliusinhorto (talk) 17:37, 24 February 2023 (UTC)
@KatoKungLee The solution to your issue is actually very simple. If you don't have access to the sources, then you don't write the article. Leave it for other editors who do understand the language and /or do have access the sources. Roger (Dodger67) (talk) 20:11, 24 February 2023 (UTC)
I can get dozens of references on local people where I live at the library that other people can't who aren't local. I can't do the same for people from foreign, poorer and smaller countries. It doesn't mean they aren't important, it's just unreasonable to expect people to have to spend large amounts of money and time to hunt down sources that are not available in their country.KatoKungLee (talk) 16:47, 25 February 2023 (UTC)

There is also a "bias" that works in the reverse direction. When somebody cite a less-available source, it is harder to scrutinize whether or not the source supports what was written. North8000 (talk) 16:56, 24 February 2023 (UTC)

  • The unfortunate thing is that lots of people want to use Wikipedia for promotion, whether for financial gain or simply to promote what they are interested in as a hobby or obsession. That means that an established, trusted editor has to have access to the sources, whether that is physical or linguistic access. Many editors seem to be frightened of anything that is not on the Internet or is not in English, even more so if it is not in the Roman alphabet. I don't think there's an awful lot we can do about that. What would you suggest, more specifically than "having less scrutiny"? Phil Bridger (talk) 18:19, 24 February 2023 (UTC)
I was just noting that the "bias" works both ways. I can't think of any change to make with regard to wp:verifiability, not any big problem there. The bigger problem is with a combination if wp:notability and wp:before. We need to establish that including and identifying GNG sources is a key part of developing articles. Not the job of somebody else to prove a negative amongst a pile of non-english or inaccessible sources. In other words, get rid of or modify wp:before. Sincerely, North8000 (talk) 19:28, 24 February 2023 (UTC)
Some of you might be interested in this opportunity:
The mw:Editing team is starting a new project, mw:Edit Check. The basic story is: Imagine a world in which the visual editor prompts editors to add inline citations. The problem is: How much is too much? It'd be annoying if you get interrupted after each character, but pointless if you never see it. If you're looking at a diff in your watchlist or Special:RecentChanges, what are you looking for (that a computer might be able to recognize)?
They're hosting a meeting in Google Meet this coming Friday, 3 March 2023. More information is available at mw:Editing team/Community Conversations#3 March 2023. I hope that many of you will be able to attend, but if you aren't, please consider leaving your advice to the team on the talk page.
Thanks, Whatamidoing (WMF) (talk) 20:33, 24 February 2023 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

A semi-automated tool for IP vandal-fighting based on revert statistics

I'd like to discuss an approach to IP vandal-fighting, that would introduce a semi-automated tool to lighten the load on admins, while hopefully improving our catch rate. Currently, semi-protection is used on article (and other) pages, which protects them from editing by unregistered users, when disruption becomes excessive. A weakness of this approach currently, is that it is a manual process, requiring use of the {{Edit semi-protected}} template at the Talk page, or a request at Wikipedia:Requests for page protection.

My proposal has two parts. One addresses "identification", i.e., the "when disruption becomes excessive" part, and attempts to automate that. The second part, would use the result of the identification process, to either flag/report the page, or semi-protect the page automatically, under some conditions. The proposal does not address, and brings no improvement to vandalism from registered users.

Part 1 — Identification

The identification process is based on the theory that the proportion of reverts on a page tells us something about vandalism. This is a very rough association, as contentious topics are also likely to have a higher proportion of reverts which are not due to vandalism. The process proposed here attempts to tease out the difference.

The proposal depends on the gathering of some statistics about reverts on an article, which might be something like:

  1. ip-edits: number of IP edits in last time-period (a series; week, month, quarter, year?)
  2. ip-reverted: number of IP edits in last time-period that were reverted
  3. reg-edits: number of registered user (non-bot?) edits in last time-period
  4. reg-reverted: number of registered user edits in last time-period that were reverted

(maybe also ip-reverts: number of IP edits in last time-period that are reverts of another user edit)

We would then derive:

  • ip-reverted-pct: percent of IP edits reverted in last time period (w/m/q/y)
  • reg-reverted-pct: percent of registered user edits reverted in last time period
  • reverted-index: calculate ip-reverted-pct / reg-reverted-pct

Over time, we would also have data available to calculate mean values over a large number of articles for the derived statistics, which would yield:

  • mean-ip-reverted-pct: percent of IP edits reverted in last time period (w/m/q/y)
  • mean-reg-reverted-pct: percent of registered user edits reverted in last time period
  • mean-reverted-index: calculate mean-ip-reverted-pct / mean-reg-reverted-pct

Possibly we could have dual, or multiple sets of these values, depending on "contentioius topic" status, or other factors. Possibly some appropriate db query could provide a rough initial approximation for these values for starters, until we have more data.

Part II — response

The meat of the proposal, involves comparing the derived reverted-index to mean-reverted-index, and depending on some conditions (minimum thresholds of accumulated data, etc.) and some configurable parameters (one standard deviation from the mean?) would then do something, which could be:

  • generate a human-readable report, for review by semi-protection team (or pending changes?) and further action
  • semi-protect the article on the fly, if conditions and thresholds are met (maybe stricter/higher threshold, 2 SDs?, than just for reporting?)

Note: about 5% of edits to a page are vandalism, per this study (2006 data).

Note that any on-the-fly protection could have a built-in throttle almost for free; that is, once protected, the ip-reverted numbers on articles having a lot of vandalism would start to fall, and soon show up in the weekly/monthly figures as drops in values. A counterpart to on-the-fly semi-protection could examine just those articles and monitor against some other (configurable) param, and automatically remove semi-protection when that "drop" threshold was reached. This would mitigate any collateral damage caused by semi-protecting an article based on high revert-ratios that in fact was not related to vandalism. (I would argue that any such collateral damage would represent far less damage to the project in the aggregate, than letting all of it through and requiring someone to notice and a manual process it; I'd call it simply, "an abundance of caution", but that's a separate discussion.)

This proposal makes some key assumptions about the meaning of these statistics, and whether they correlate well, or at all, with vandalism, and for that we would probably need a small starter set of data from a handful of articles which could then be assessed by humans to see whether there's a fit. Someone like the Quarry folks (pinged below) might know how to gather a starter set like this for a few articles, so we can test the theory and see if it holds up.

The hope is, that by providing some automation of the detection of possible IP vandalism, with the concomitant possibility of semi-automated or automated semi-protection, we can greatly improve the detection of IP vandalism, somewhat lighten the load on users, who have to stop what they are doing when they notice suspected cases in order to find out how to report it and then report it, and vastly lighten the load of admins having to deal with semi-protection requests, which are probably only a small fraction of the known IP vandalism, and an even smaller fraction of all IP vandalism, with a possible submerged iceberg out there, that no one even notices.

Adding possibly interested users; ClueBot operators: @Cobi, Rich Smith, and DamianZaremba:; Xtools revert stats: MusikAnimal; db query folks: @Cryptic, Novem Linguae, Certes, and Joe Roe:; WP:AIV folks: @Yamla, ToBeFree, Daniel Case, Wldr, Zzuuzz, Bbb23, Materialscientist, Hut 8.5, HJ Mitchell, and KrakatoaKatie:; recent related WP:ANI discussion participants: @Boing! said Zebedee, NinjaRobotPirate, and Kusma:, and pt-wiki users @Renato de Carvalho Ferreira, JMagalhães, MisterSanderson, Érico, and PauloMSimoes: (who took part in the August 2020 IP-banning discussion[English] on pt-wiki.) Thanks, Mathglot (talk) 00:07, 5 February 2023 (UTC)

Re-ping, due to typo: User:Widr. Mathglot (talk) 00:21, 5 February 2023 (UTC)
You provided a thorough and thoughtful discussion but I don't see where you've provided a specific proposal. North8000 (talk) 00:42, 5 February 2023 (UTC)
I guess I implied a proposal, without really stating it:
Let's gather the required revert stats (maybe we have them already?) and create the infrastructure necessary to calculate the needed values, generate reports on them, and maybe auto-semiprotect those pages for which it exceeds a certain threshold.
But as this is still in brainstorming stage, I wasn't quite ready to make a specific proposal, since I'm sure I'll benefit from feedback and new information, which likely will change the proposal. But still, clarity is always better, so this is my first take at a specific proposal. Thanks, Mathglot (talk) 00:56, 5 February 2023 (UTC)
Thanks for the ping; I'll wait for data before judging this. My initial thought was "this can't work without a distinction between vandalism and good-faith edits in the data", but then again, page protection is applied to prevent disruptive editing in general. ~ ToBeFree (talk) 01:01, 5 February 2023 (UTC)
(edit conflict) ToBeFree, the proxy for that distinction, is the comparison of article revert-index to the mean; i.e., how the ratio of reverted IP edits to reverted user edit compares to articles in the encyclopedia more generally? If it's twice or ten times as much, why is that? Still, this calculation is a blunt instrument, and as I envisage it, just a starting point. Sharpening it might involve factoring in contentiousness of the article, and other features of the reverted IPs, or the reverting users. I know when I look at my long Watchlist, that I skip or pay less attention to reverts by editors I trust, than ones I don't recognize, and when I see the edit summary "Fix typo" on an edit by an IPv6 which adds +338 bytes to the article, I examine it very carefully. I'm sure we all have an instinct about what to look for in a revert, and this is just a very rough first cut to try to translate that instinct into something quantifiable based on actual data we can deal with and base calculations upon. Necessarily, the first version won't be a very good approximation, but if it's feasible at all, we have to start somewhere. That' why I pinged some of the ClueBot folks, because they may be doing something like this, though for a different task. Mathglot (talk) 01:23, 5 February 2023 (UTC)
That comparison is surely useful, but does not provide the distinction I was referring to, as there is no way to see from the data whether a specific article has a high amount of good-faith disruptive IP editing or bad-faith disruptive IP editing. Fortunately, this distinction is probably not needed as both should lead to protection. ~ ToBeFree (talk) 01:26, 5 February 2023 (UTC)
ToBeFree, I see, thanks for that clarification. Certes has already weighed in with some data below, which is exciting to see; I'm still looking at it, to see what it can tell us. Rather than misinterpret you again, can you describe the data you would like to see, or think of a query (in English) that would provide the data you are waiting for? Mathglot (talk) 01:38, 5 February 2023 (UTC)
That's actually already what I had been waiting for; I'll reply below. ~ ToBeFree (talk) 01:43, 5 February 2023 (UTC)
I've done a few preliminary counts for the week 00:00 28 Jan to 00:00 4 Feb (excluding 4 Feb, as its bad edits may not yet be reverted). Registered editors made 2,253,389 edits of which 35,146 (1.56%) were reverted. IPs made 220,447 edits of which 41,862 (18.99%) were reverted, a ratio of 12.18. I don't have the standard deviation but I've listed the articles with a ratio over 25, limiting it to those with more than 10 reverted IP edits (to eliminate articles with one IP vandal and no registered vandals). Results are here. Most of them have an infinite ratio because no registered editors made a reverted change last week. Of course, some reverted edits may be a good-faith vandal fighter having their work undone by a persistent vandal. Certes (talk) 01:08, 5 February 2023 (UTC)
Excellent; thanks very much for this. This is just what we need, some hard data we can begin to look at, so we can have a discussion based on the data, and not speculation. Mathglot (talk) 01:28, 5 February 2023 (UTC)
Thank you very much! Looking at the evaluated revision history of "1939_Japanese_expedition_to_Tibet", which is displayed as "11 ip_total, 11 ip_reverted, 1 reg_total, 0 reg_reverted", the data quality could probably be improved by counting successive edits from the same IP address as only one contribution, and counting it as reverted if at least one of the edits (alternatively, any of them) has a mw-reverted tag. ~ ToBeFree (talk) 01:47, 5 February 2023 (UTC)
Quarry 71078: Articles where edits are much more likely to be reverted if from IPs
Quarry 71078: Articles where edits are much more likely to be reverted if from IPs
ns rc_title ip
total
ip
reverted
reg
total
reg
reverted
0 1939 Japanese expedition to Tibet (edit | talk | history | links | watch | logs) 11.0 11.0 1.0 0.0
0 2011 United States debt-ceiling crisis (edit | talk | history | links | watch | logs) 22.0 21.0 5.0 0.0
0 2017 Snooker Shoot Out (edit | talk | history | links | watch | logs) 13.0 13.0 1.0 0.0
0 2022 European heat waves (edit | talk | history | links | watch | logs) 40.0 11.0 21.0 0.0
0 2023 German Masters (edit | talk | history | links | watch | logs) 42.0 25.0 92.0 1.0
0 2023 North Island floods (edit | talk | history | links | watch | logs) 24.0 16.0 349.0 4.0
0 2023 Pakistan Super League squads (edit | talk | history | links | watch | logs) 22.0 22.0 2.0 0.0
0 2023 Pakistani general election (edit | talk | history | links | watch | logs) 13.0 12.0 11.0 0.0
0 Afra Saraçoğlu (edit | talk | history | links | watch | logs) 32.0 16.0 6.0 0.0
0 America's Next Top Model (season 7) (edit | talk | history | links | watch | logs) 15.0 14.0 1.0 0.0
0 America's Next Top Model (season 9) (edit | talk | history | links | watch | logs) 34.0 13.0 8.0 0.0
0 American football positions (edit | talk | history | links | watch | logs) 11.0 11.0 4.0 0.0
0 Barbapapa: One Big Happy Family! (edit | talk | history | links | watch | logs) 12.0 11.0 4.0 0.0
0 Benigembla (edit | talk | history | links | watch | logs) 66.0 11.0 3.0 0.0
0 Brahui language (edit | talk | history | links | watch | logs) 14.0 13.0 2.0 0.0
0 Cascadia movement (edit | talk | history | links | watch | logs) 13.0 13.0 5.0 0.0
0 Criminal stereotype of African Americans (edit | talk | history | links | watch | logs) 11.0 11.0 9.0 0.0
0 Delicious Party Pretty Cure (edit | talk | history | links | watch | logs) 33.0 16.0 11.0 0.0
0 Disney Princess (edit | talk | history | links | watch | logs) 11.0 11.0 2.0 0.0
0 Elimination Chamber (2023) (edit | talk | history | links | watch | logs) 27.0 23.0 76.0 1.0
0 Fateh Burj (edit | talk | history | links | watch | logs) 13.0 11.0 1.0 0.0
0 Food 4 Less (edit | talk | history | links | watch | logs) 18.0 13.0 9.0 0.0
0 Georges Méliès (edit | talk | history | links | watch | logs) 13.0 13.0 1.0 0.0
0 Gum (edit | talk | history | links | watch | logs) 12.0 12.0 3.0 0.0
0 Harshvardhan Rane (edit | talk | history | links | watch | logs) 33.0 12.0 3.0 0.0
0 Hero: 108 (edit | talk | history | links | watch | logs) 18.0 16.0 3.0 0.0
0 Hurlock, Maryland (edit | talk | history | links | watch | logs) 16.0 16.0 4.0 0.0
0 Indonesia's Next Top Model (season 3) (edit | talk | history | links | watch | logs) 79.0 15.0 32.0 0.0
0 Intimidation (edit | talk | history | links | watch | logs) 11.0 11.0 4.0 0.0
0 Janet Nguyen (edit | talk | history | links | watch | logs) 89.0 16.0 1.0 0.0
0 Jerry Lawson (engineer) (edit | talk | history | links | watch | logs) 12.0 12.0 10.0 0.0
0 Jesse Lee Soffer (edit | talk | history | links | watch | logs) 16.0 13.0 5.0 0.0
0 JimJam (edit | talk | history | links | watch | logs) 15.0 13.0 7.0 0.0
0 Jon Snow (character) (edit | talk | history | links | watch | logs) 12.0 12.0 7.0 0.0
0 Joseph Barboza (edit | talk | history | links | watch | logs) 16.0 15.0 9.0 0.0
0 Justin Jefferson (edit | talk | history | links | watch | logs) 14.0 14.0 11.0 0.0
0 Lam Kor-wan (edit | talk | history | links | watch | logs) 41.0 14.0 8.0 0.0
0 Lara Secondary College (edit | talk | history | links | watch | logs) 19.0 19.0 3.0 0.0
0 Larry Page (edit | talk | history | links | watch | logs) 16.0 14.0 7.0 0.0
0 List of Asian stadiums by capacity (edit | talk | history | links | watch | logs) 52.0 15.0 4.0 0.0
0 List of Fuller House episodes (edit | talk | history | links | watch | logs) 12.0 11.0 2.0 0.0
0 List of Go, Diego, Go! episodes (edit | talk | history | links | watch | logs) 18.0 18.0 4.0 0.0
0 List of Pretty Cure films (edit | talk | history | links | watch | logs) 24.0 15.0 1.0 0.0
0 List of current automobile manufacturers by country (edit | talk | history | links | watch | logs) 13.0 12.0 3.0 0.0
0 List of equipment of the South African Army (edit | talk | history | links | watch | logs) 12.0 12.0 4.0 0.0
0 List of former TV channels in the United Kingdom (edit | talk | history | links | watch | logs) 15.0 12.0 3.0 0.0
0 List of militaries by country (edit | talk | history | links | watch | logs) 20.0 18.0 10.0 0.0
0 List of programs broadcast by Treehouse TV (edit | talk | history | links | watch | logs) 62.0 12.0 9.0 0.0
0 Lucas Merolla (edit | talk | history | links | watch | logs) 30.0 14.0 9.0 0.0
0 Maybe (Machine Gun Kelly song) (edit | talk | history | links | watch | logs) 13.0 13.0 4.0 0.0
0 Melinda Dillon (edit | talk | history | links | watch | logs) 12.0 11.0 7.0 0.0
0 Mike Little (edit | talk | history | links | watch | logs) 12.0 12.0 4.0 0.0
0 Oberholzer murder (edit | talk | history | links | watch | logs) 11.0 11.0 1.0 0.0
0 Pedro Porro (edit | talk | history | links | watch | logs) 15.0 12.0 32.0 1.0
0 Philippine peso (edit | talk | history | links | watch | logs) 14.0 11.0 3.0 0.0
0 Quicksilver (wrestler) (edit | talk | history | links | watch | logs) 18.0 18.0 4.0 0.0
0 Real World/Road Rules Challenge: Battle of the Sexes 2 (edit | talk | history | links | watch | logs) 13.0 13.0 2.0 0.0
0 Roberrt (edit | talk | history | links | watch | logs) 12.0 11.0 6.0 0.0
0 Roger Tuivasa-Sheck (edit | talk | history | links | watch | logs) 15.0 15.0 3.0 0.0
0 Sara Corrales (edit | talk | history | links | watch | logs) 11.0 11.0 1.0 0.0
0 Sony Pictures Imageworks (edit | talk | history | links | watch | logs) 11.0 11.0 3.0 0.0
0 St Mary's Rochfortbridge GAA (edit | talk | history | links | watch | logs) 17.0 13.0 1.0 0.0
0 Statement (computer science) (edit | talk | history | links | watch | logs) 12.0 12.0 3.0 0.0
0 Stetson University College of Law (edit | talk | history | links | watch | logs) 14.0 14.0 1.0 0.0
0 Superbook (edit | talk | history | links | watch | logs) 13.0 12.0 6.0 0.0
0 Swashbuckle (TV series) (edit | talk | history | links | watch | logs) 27.0 27.0 10.0 0.0
0 Syria Palaestina (edit | talk | history | links | watch | logs) 14.0 14.0 9.0 0.0
0 Tejasswi Prakash (edit | talk | history | links | watch | logs) 13.0 11.0 3.0 0.0
0 The Friendly Type (edit | talk | history | links | watch | logs) 16.0 14.0 1.0 0.0
0 The Theory and Practice of Oligarchical Collectivism (edit | talk | history | links | watch | logs) 12.0 12.0 2.0 0.0
0 Third Servile War (edit | talk | history | links | watch | logs) 13.0 13.0 1.0 0.0
0 UKTV (edit | talk | history | links | watch | logs) 18.0 18.0 3.0 0.0
0 VRT NWS Journaal (edit | talk | history | links | watch | logs) 13.0 13.0 5.0 0.0
0 Varma (surname) (edit | talk | history | links | watch | logs) 11.0 11.0 3.0 0.0
0 Verona Villafranca Airport (edit | talk | history | links | watch | logs) 11.0 11.0 6.0 0.0
0 Walter Reed (edit | talk | history | links | watch | logs) 14.0 14.0 3.0 0.0
0 Williams Street (edit | talk | history | links | watch | logs) 65.0 30.0 1.0 0.0
0 XEFE-AM (edit | talk | history | links | watch | logs) 12.0 12.0 8.0 0.0
0 Čačak (edit | talk | history | links | watch | logs) 11.0 11.0 3.0 0.0
1 Talk:Ce Acatl Topiltzin (edit | subject | history | links | watch | logs) 11.0 11.0 2.0 0.0
1 Talk:Isla Bryson case (edit | subject | history | links | watch | logs) 14.0 13.0 130.0 3.0
1 Talk:Metohija (edit | subject | history | links | watch | logs) 12.0 12.0 18.0 0.0
1 Talk:Numbers (TV series) (edit | subject | history | links | watch | logs) 12.0 12.0 3.0 0.0
118 Draft:The Bad Batch (2016 film) (edit | talk | history | links | watch | logs) 68.0 19.0 2.0 0.0

SQL query:

SELECT rc_namespace, rc_title,
 SUM(CASE WHEN actor_user IS NULL THEN 1 ELSE 0 END) AS ip_total,
 SUM(CASE WHEN actor_user IS NULL AND ct_tag_id IS NOT NULL THEN 1 ELSE 0 END) AS ip_reverted,
 SUM(CASE WHEN actor_user IS NOT NULL THEN 1 ELSE 0 END) AS reg_total,
 SUM(CASE WHEN actor_user IS NOT NULL AND ct_tag_id IS NOT NULL THEN 1 ELSE 0 END) AS reg_reverted
FROM recentchanges
JOIN actor ON actor_id = rc_actor
LEFT JOIN change_tag ON ct_rc_id = rc_id AND ct_tag_id = 590 /* mw-reverted */
WHERE rc_timestamp BETWEEN '20230128' AND '20230204'
GROUP BY rc_title
HAVING ip_reverted > 10 AND ip_reverted/ip_total * reg_total/(CASE WHEN reg_reverted = 0 THEN 0.0001 ELSE reg_reverted END) > 25
ORDER BY rc_namespace, rc_title
Certes' query as Wikitable. Mathglot (talk) 01:53, 5 February 2023 (UTC)
Links and current SQL source added ~ ToBeFree (talk) 01:59, 5 February 2023 (UTC)
Hi, Certes; I forked to 71079 trying to add percent columns, but I'm guessing you can't re-use previously established 'AS' labels in subsequent parts of the statement. You'll see what I tried to do in creating cols ip_rvt_pct and reg_rvt_pct. Or, does every col have to be a SUM because of the GROUP BY? Is that fixable? Thanks, Mathglot (talk) 02:52, 5 February 2023 (UTC)
The column alias doesn't become usable during the main SELECT, only at the GROUP BY stage. There are workarounds but basically you need to repeat the expression, e.g. (SUM(CASE WHEN actor_user IS NULL AND ct_tag_id IS NOT NULL THEN 1 ELSE 0 END) * 100) / SUM(CASE WHEN actor_user IS NULL THEN 1 ELSE 0 END). Certes (talk) 12:15, 5 February 2023 (UTC)
I think a bot which tries to automatically identify pages with high rates of IP vandalism and either protect the page or file a request somewhere (maybe WP:RFPP) might well be workable, but I'd suggest:
  1. Semiprotection is much more likely to be appropriate if lots of edits are being reverted in a short space of time. If a page gets 10 IP edits in a year and all are reverted then it probably isn't appropriate to semiprotect the page. If it gets 10 reverted IP edits in a day then semiprotection is a lot more likely to be appropriate.
  2. The number of reverts might be more useful than the number of edits. For example Disney Princess is included in the above query output because an IP made 10 edits in a row which were reverted in one edit. This isn't appropriate for semiprotection. If the IP was reverted 10 times then semiprotection is more likely to be useful.
  3. You could try to determine whether the edits are likely to be vandalism by looking at the edit summary. If an edit is reverted with a standard rollback message, or if the edit summary mentions vandalism or spam then semiprotection is more likely to be appropriate. If people are writing custom edit summaries for the reverts then it's less likely.
  4. The number of IPs making the edits is also important. If one IP is making all the edits then protection probably isn't a good idea and it would make sense to just block that IP (or report them to AIV) instead. If edits are coming from multiple IPs or one vandal on a range then semiprotection is more likely to be useful.
  5. You could also look at the people doing the reverts. If IPs are being reverted by experienced editors (especially multiple experienced editors) then protection is more likely to be appropriate. If on the other hand an IP is reverting another IP, or an IP is reverting a new user, then automatic protection is probably a bad idea (you might protect the page in a vandalised version and then stop the person reverting the vandalism from fixing it).
Hut 8.5 11:09, 5 February 2023 (UTC)
We might also look at the long-term revert % of regular editors rather than this week's (because most articles have no such reverts this week). For efficiency, that suggests a two-stage process:
  1. Find pages where many different IPs have been reverted recently (that's quick)
  2. For those pages only, calculate the long-term reversion % of registered editors, to filter out contentious topics
In practice, it might be worth not doing the second step: if IPs keep arguing over whether India is better than Pakistan, we may still want to protect.
quarry:query/71090 lists pages where more than 50% of IP edits have been reverted, and at least five different IPs have been reverted, over the previous two days, Certes (talk) 13:09, 5 February 2023 (UTC)
It might be desirable to calculate statistics on the source of IP vandalism by ASN and restrict anonymous edits from problematic networks. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 15:47, 5 February 2023 (UTC)
@Certes, I think you could restrict the query to the mainspace, as I understood the proposal to be primarily aimed at the mainspace. You could probably also remove bots and similar edits (e.g., AWB).
@Mathglot, without saying anything at all about the merits of your proposal, I wonder if the upcoming m:IP Masking work will introduce a complication. About a year or so from now, there won't be "unregistered editors"; there will instead by automatically created "temporary accounts". User:172.0.0.1 will become User:*2022-1234 instead (or something like that). This is, in the backend, a third account type. There are advantages (privacy for the user; better communication options for the rest of us), but whatever you build now might break when this happens. I would expect it to be fixable, but I would also expect pretty significant hiccups in the transition process. Whatamidoing (WMF) (talk) 00:00, 8 February 2023 (UTC)
We only need to know whether each edit came from an account or an IP, and which IP edits come from the same IP. That data should remain available after IP masking, though it may move to different database tables. The main risk may be that this tool will become redundant because IP masking makes vandal-fighting impractical and we reluctantly have to turn away unregistered editors altogether, but that's a separate debate. Certes (talk) 00:11, 8 February 2023 (UTC)
Right now, we have two categories of accounts:
  • Registered editors
  • Unregistered editors
After IP masking, we will (probably) have three categories of accounts:
  • Registered editors
  • Unregistered editors
  • Temporary editors
with the expectation that there won't actually be anyone in the "Unregistered editors" list (though it will continue to exist in theory). I don't know what kind of update will be necessary to the code (maybe it will be as simple as finding all the bits that say "unregistered" and replacing them with "temporary"), but if you're calculating this dynamically, then there will be a transition period. If the switch happens on February 1 (and assuming a 30-day look-back period) then on February 1, you'll have 30 days' worth of unregistered editors and 0 days' worth of tempoary editors – a problem, if you want to use only temporary editors to calculate which articles are at risk from temporary editors today. Halfway through the month, you'll have 15 days' of unregistered editors and 15 days of temporary editors. This could give you weird numbers. Once the time period is over, then all should be well, but during the transition, it might require some extra effort. Whatamidoing (WMF) (talk) 01:36, 8 February 2023 (UTC)
I think unregistered and temporary editors are the same thing for this purpose: they're not logged in. The only problem, if and when we have a transition, is that someone who was 123.45.67.89 yesterday might be Hidden-IP-246810 tomorrow and we're deliberately prevented from connecting the two. It will be as if everyone changes their IP address at once (possibly renewing their licence to vandalise, Eternal September-style). But we still don't know the details of IP masking, and some of us still have hope that it will never be achieved, so let's not let that tail wag the dog. Certes (talk) 01:57, 8 February 2023 (UTC)
@Certes, I believe the plan is for admins and other trusted users (think Wikipedia:Requests for permissions, not an RFA-like process) to have access to the IP for all temporary users. It just won't be visible to anyone with internet access.
IP masking will happen; there is no longer any doubt about that. If you want to follow it, then I suggest watching m:IP Editing: Privacy Enhancement and Abuse Mitigation#Statements from the Wikimedia Foundation Legal department. Whatamidoing (WMF) (talk) 22:25, 21 February 2023 (UTC)
I'm hoping that we can just look at IP edits and not have to consider edits by registered accounts. If so then we can assume edits are not by AWB or bot (or that if they are, we certainly want to know about them!). Certes (talk) 00:16, 8 February 2023 (UTC)
Mainspace is certainly the main focus but if IPs are being reverted in other namespaces then something is amiss and needs attention. If they're vandalising a template or WP:something then we want to know, and you have to be a real nuisance to get reverted repeatedly on a Talk: page. It doesn't cost much to check all the namespaces while we're there. Certes (talk) 00:19, 8 February 2023 (UTC)
(edit conflict)@Whatamidoing (WMF): I am aware of the IP masking project (didn't know it was called that) and have had it in mind since the beginning, but since I don't have sufficient details about it, and the basic idea here seemed intricate enough to begin with without adding the complication of new types of account (yet), I wanted to expose the outlines of the idea here to see how it might evolve, and to get some good folks thinking about it. I think we could still evolve the basic idea a bit more, thinking about different forms of queries to test, gathering more suggestions (as you just did to Certes) and more data to prove (or refute) the principles behind it, and then at some point, if it's worth considering a transition from brainstorming level (VPI) to proposal level (VPR), we could discuss at that point whether it's worth building a simple prototype or report based on these queries (or others) which might be the germ of a useful additional tool for vandal fighting. If all of those "if's" line up as I hope they will, then, perhaps, it would be worth asking ourselves whether we should go ahead and build something simple (a 1x/week bot report?) with the foreknowledge that it will have to change once the new account type arrives, or do we just shelve it for the time being and wait for the change? It's hard to answer that now with little data to go on, but I'm biased to doing something sooner rather than later, partly because it might expose the idea to more eyeballs (and therefore more, and better evolution), partly because I don't know if I really trust the timeline I'm hearing about the IP account transition, given some other expected dates that later changed (not blaming, just trying to be realistic), and partly because if the "old way" (IP accounts) actually works, then for 12, or 18, or 24 months, we've got one more tool in our toolshed for vandal fighting for that period of time, and I'm all in favor of that. (Maybe also because I'm eager to see if this could actually be of benefit, and if so, I'd like to try and get something concrete out there to look at, which might help bootstrap it to the next level. Certainly if you (or anyone from WMF) has any inside knowledge about IP account transition that would doom this or make it unworkable, the sooner we find out about that, the better; likewise if we know that there will be hiccups (how big?) that would be useful knowledge, too. Thank you for mentioning this; it was inevitable that this would be raised at some point, and it's just as well that you did so now. Mathglot (talk) 00:43, 8 February 2023 (UTC)
I think that a bot report has a lot of potential, though the exercise in general reminds me of Regression toward the mean. In healthcare, there's a drive to address high-need patients based on the medical care they needed last year, and it often ends with researchers noting that high-need patients in year 1 are often not high-need patients in year 2, regardless of whether you do anything. It could be that using "yesterday's" unwanted edits to predict where "tomorrow's" unwanted edits will happen is just not very effective. For example, looking at the first article in the table above, an IP was reverted last week, but that was the first time an IP was reverted on that article in about a dozen years. It could be another dozen years before it happens again. Whatamidoing (WMF) (talk) 01:43, 8 February 2023 (UTC)
Collecting statistics by ASN would definitely require support on the IP masking side, and privacy concerns would limit who could run the bot(s). --Shmuel (Seymour J.) Metz Username:Chatul (talk) 15:10, 8 February 2023 (UTC)
@Chatul, are you aware of any work currently being done that depends on the Autonomous System Number? Whatamidoing (WMF) (talk) 22:27, 21 February 2023 (UTC)
No there would have to be new code to look up the ASN for the IP address, and there would have to be some analysis of privacy concerns to ensure that the new code would not allow unauthorized users to identify anonymous posters. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 13:59, 26 February 2023 (UTC)

Idea: wiki file format

Here I go, hoping I'm not trying to reinvent the wheel

I think it would be a great idea if wiki could create a couple standard file format, to group hyperlink text into a single file (therefore replacing the file/folder thing enforced by HTML...

My idea would be a single archive fail containing:

  • the text content in wiki code, with each page as a file
  • a folder containing the necessary media associated with pages

It would be a great tool, for a lot of stuff:

  • I'm always wishing I had a wiki dedicated for this or that base of knowledge (in an association, to group all acronyms, and other interconnected notions; in development, to share knowledge on a project, and, as myself, as a )TTRPG player, to keep lore on our world updated and easy to browse)
  • one can imagine publishing "books" in that format, effectively replacing pdf/doc files, a bit in the same spirit as latex files but more in a CBR/ebook spirit
  • it would be awesome to be able to "extract" a collection of Wikipedia pages (determining which and suggesting things to add with the see also feature and links in the already selected page) for offline use, effectively creating a sharable, savable "reading list" from the android app.
  • it would reduce bandwidth use in that respect, and therefore, maintenance cost and environmental impact.
  • it also would mean that Wikipedia would have the first "interpreter" of that file type, but that would/could rapidly change: people could want different skin, style, etc. for their wiki file readers.

On top of that, I said typeS of file, one can imagine variations of this:

  • one with only the text content
  • one with the media attached
  • one with the modification history
  • etc.

In any case, if the android application can be "split" between a sort of reader and a sort of browser/research query manager, it would be a perfect implementation of the kiss principle and UNIX philosophy. Alefith (talk) 03:07, 21 February 2023 (UTC)

@Alefith maybe a partial answer, there is a standard export format for wikipages, in XML. Here is an example of what it produces: Special:Export/Stigmella_corylifoliella. That is for the current version, using Special:Export you can also get the full page history with all attribution. Notably, it does not contain images. — xaosflux Talk 19:22, 22 February 2023 (UTC)
We also have a PDF renderer, see this example for that same article. — xaosflux Talk 19:24, 22 February 2023 (UTC)
Yes these are parts of the functionality I'm looking for, but it's impossible to do what I want to do.
Let's take an use case, so that it's more clear:
Let's consider three pages:
what I want is:
  • to be able to compare the two type of moth (say in different tab, or better yet in a split window)
  • to add one type of moth, or insect, or whatever, or edit something (without having to type HTML/XML, or anything more than wikitext)
  • that the links towards any of the pages encapsulated in the file (any of the three mentionned, or any other I created afterwards) are pointing towards the page and not towards the wikipedia (or source wiki) website.
Basically, a file would be the data of a "mini-wiki". And you could consult or modify the pages with a wiki website, or with other tools (say, a different wiki engine, or one can imagine offline apps, etc.)
As far as I understand, the only two things I can do with the tools you proposed are:
  • having one pdf / XML file with every page concatenated (that probably won't even work because of table of content conflicts) and then relink every link in the pages that point towards an incapsulated page (ie moth, attacus atlas, and stigmella corylifofiella) so that they point to the right section of the file, and not towards the source wiki.
  • having one XML file for each wiki page, and manually switch any time I want to consult a linked page
Basically, going from hypertext data to text data.
You could say what I'm looking for is HTML with the markup limited to wikitext (and therefore the syntax as simple as wikitext as well), links always relative and always towards included pages, no scripting, and the styling handled externally.
Honestly, if that existed, I'm pretty sure at some point it could be used a lot, say, for documentation of IT project, etc. Basically, it would be a new *.txt . Literally a hypertext file to replace a text file, hey how about *.htxt as an extension? Alefith (talk) 21:47, 22 February 2023 (UTC)
I don't think this is what you're looking for, but there is also the "parse" endpoint, example output. — xaosflux Talk 23:30, 22 February 2023 (UTC)
Also, the rvprop output, example here. — xaosflux Talk 23:31, 22 February 2023 (UTC)
No it isn't but thanks! Alefith (talk) 13:06, 27 February 2023 (UTC)

Trivia in year articles and Template:Year article header

All the year articles have an intro generated by Template:Year article header which is mostly trivia. For example, the intro for 1754 is:

1754 (MDCCLIV) was a common year starting on Tuesday of the Gregorian calendar and a common year starting on Saturday of the Julian calendar, the 1754th year of the Common Era (CE) and Anno Domini (AD) designations, the 754th year of the 2nd millennium, the 54th year of the 18th century, and the 5th year of the 1750s decade. As of the start of 1754, the Gregorian calendar was 11 days ahead of the Julian calendar, which remained in localized use until 1923.

I think a lot can be removed. No need to say the day of the week for 1st January. I would also remove the whole "the 754th year of the 2nd millennium, the 54th year of the 18th century, and the 5th year of the 1750s decade", just keeping the century and millennium.

Any thoughts? Vpab15 (talk) 18:34, 26 February 2023 (UTC)

It's useful information, and I like the idea of a standard paragraph like this for all year articles. I agree about trimming the detail per Vpab except I'd keep day of the week (where else would you find that information?). I think we could also ditch the Roman numerals, as I do not think English Wikipedia has many Roman readers. I'd move the whole paragraph further down in the lead, perhaps it could be the standard last paragraph of the lead. Then there is the question of what do we write for the first paragraph to replace it, though? Levivich (talk) 18:58, 26 February 2023 (UTC)
I always imagined the Roman numerals might be a useful search term to include for readers who have seen them in the credits of movies and TV shows.
Overall I think including these kind of calendrical factoids is in scope for main year articles, but I agree that "the 754th year of the 2nd millennium, the 54th year of the 18th century, and the 5th year of the 1750s decade" serves little purpose, and it duplicates what’s in the infobox immediately to its right. Barnards.tar.gz (talk) 15:22, 1 March 2023 (UTC)

Users should try to attempt to improve articles first and mark them for deletion second

I don't want to name any specific articles here in the interest of keeping the peace.

In the last week, I've personally edited at least 7 different articles to make them more eligible for being on wikipedia. These articles had notices on them and I went in and fixed them. I didn't know anything about a lot of these subjects and I don't even know how to do a lot of things on wikipedia. The fixes probably took me less than 5 minutes.

While I agree that people should do the best they can with articles and should put effort into finding sources, I think it's wrong to mark articles for deletion without any attempts to improve the articles themselves. In many situations, fixes seem to take even less time than marking them for deletion does.

I'd like to suggest that a rule be made where people should first attempt to fix articles before deleting them. Or, if perhaps something could be done involving users whose first move is to delete an article without ever thinking of improving them. No, not every article can be fixed or is worthy, but the mindset of immediately going for the deletion is very concerning.

KatoKungLee (talk) 21:11, 23 February 2023 (UTC)

Based on some comments you've posted in a couple of AfDs, I suspect you're thinking of my edits. Please feel free to look through my edit history, but I can assure you that I spend plenty of time improving articles that have problems. I also nominate a lot of articles for deletion, but only after looking to see if they can be improved first. Since I work on footballer biographies most of the time, it is only natural that nominating articles for deletion is a large part of what I do - for example, we have over 50,000 football articles assessed as stubs with low importance; many of those are simply not ever going to meet our notability policies and guidelines. Jogurney (talk) 21:23, 23 February 2023 (UTC)
I did not mention anyone specifically by name. Please don't assume.KatoKungLee (talk) 21:36, 23 February 2023 (UTC)
Absolutely editors should be fixing articles that they believe should be part of Wikipedia, but fixing an article you don't believe should be part of Wikipedia doesn't really make any sense. -- LCU ActivelyDisinterested transmissions °co-ords° 21:25, 23 February 2023 (UTC)
I suggest a better idea would be that editors creating articles should have to add sources proving the articles notability, and so don't waste the time of other editors who have to check if the articles are notable or not. -- LCU ActivelyDisinterested transmissions °co-ords° 21:52, 23 February 2023 (UTC)

I'd like to suggest that a rule be made where people should first attempt to fix articles before deleting them

The problem from my perspective is the amount of work it could take to improve an article and dig up sources, and how the people who don't want the article to exist are often not motivated to work on it. Placing burdens like that on people who are already in "good deed mood" rather than "this is important modd" and "this is interesting mode" seems unworkable. The cases that come to mind are things like niche psychological or medical diagnoses with with missing or poor sources which when they exists are old and tangetial. The burdens I would place on the deleter would be more along the lines of "don't delete an article that contains a version of the article you would be happy with if you removed material" and "actually look at the sources" Talpedia (talk) 21:35, 23 February 2023 (UTC)
  • This "rule" you are speaking of already exists! One of the "requirements" before nominating an article for deletion is to follow WP:BEFORE. Of course, some nominators do not do this, or do not spend enough time on this, but it is already part of the process of AfD, and editors have been "warned" or chastised for not doing this. Natg 19 (talk) 21:53, 23 February 2023 (UTC)
    Greater and more explicit enforcement of Wikipedia:Deletion is not cleanup might help. Imagine if selected editors were told, "Look, you keep nominating articles for deletion because the current version is ugly, but that's not how notability works. AFD is not a place for demanding that other editors clean up an article and add sources. If you do this again, you'll be TBANned from nominating articles for deletion." WhatamIdoing (talk) 22:04, 23 February 2023 (UTC)
    Only if we equally BAN editors from article creation who consistently don't reference they articles properly. Wasting other editors time by not fulfilling BURDEN is disruptive. -- LCU ActivelyDisinterested transmissions °co-ords° 00:11, 24 February 2023 (UTC)
    @ActivelyDisinterested, please see my note about the meeting above. What's a "proper" level of citations? How could that be calculated in software? Whatamidoing (WMF) (talk) 20:36, 24 February 2023 (UTC)
    I would have thought a good rule of thumb is each block of text for additions, if someone has pressed return they have completed some specific point (you can see this behaviour in a lot of Wikipedias articles). That's not the "correct" answer, but based on a design stand point. I don't think an absolutely correct answer is possible, but each sentence would be aggravating while waiting an entire edit wouldn't encourage the desired behaviour (note the block of text could be one sentence, but could be several). How to deal with edits that change details is harder, maybe some form of lesser nudge asking if the changes require a new reference. -- LCU ActivelyDisinterested transmissions °co-ords° 21:26, 24 February 2023 (UTC)
    Whether that rule is a rule is debated; some editors do not consider it to be required and there is no consensus on whether it is. For example, when nominating mass created articles many editors think it is reasonable to put in the same level of effort that the creator put in, which precludes a WP:BEFORE. BilledMammal (talk) 23:26, 23 February 2023 (UTC)
    One of the questions at the, sooner to be defunct, AfD RFC was going to be whether BEFORE should be obligatory or not. As it stands it's an unanswered question. -- LCU ActivelyDisinterested transmissions °co-ords° 00:09, 24 February 2023 (UTC)
  • It is important to remember that a nomination for deletion is just that - a nomination. It is the beginning of a discussion, not the end of it. Lots of articles get nominated and end up being kept. That said, see WP:BURDON... it is up to those who wish to keep the article to bring it up to minimum standards. We are all volunteers, and you can not make someone else do the work - you can only make yourself do it. Blueboar (talk) 22:01, 23 February 2023 (UTC)
    OTOH, the minimum standard is "it would be possible to find sources, if someone spent enough time and effort to do so". There is no minimum standard for what the current version of the article must look like. A 100% unsourced substub saying nothing more than "Cancer is a kind of disease" is not eligible for deletion. WhatamIdoing (talk) 22:06, 23 February 2023 (UTC)
    I think we should explore making unsourced articles eligible for draftification that can only be contested by adding a source. BilledMammal (talk) 23:26, 23 February 2023 (UTC)
    "A 100% unsourced substub saying nothing more than "Cancer is a kind of disease" is not eligible for deletion. " That's a problem. Such an "article" is worse than useless and should be eradicated on sight with no discussion needed.--User:Khajidha (talk) (contributions) 04:07, 25 February 2023 (UTC)
Users should try to ensure that articles they create will meet notability criteria first, and move them into article space second... AndyTheGrump (talk) 22:07, 23 February 2023 (UTC)
  • The debate between “inclusionists” and “deletionists” is as old as Wikipedia. It isn’t going to end soon. But we can strive for a middle ground. I have long felt that if we are going to ask deletionists to do a BEFORE search prior to nomination, we should also ask inclusionists to do an AFTER fix of the article should the result be “keep”. While it is very frustrating when deletionists nominate an article that can easily be sourced… it is equally frustrating when inclusionists say that they found sources, but never bother to actually cite them IN the article. Blueboar (talk) 02:03, 24 February 2023 (UTC)
    There's a very big difference between finding enough sources online to convince yourself that the topic is notable (typically takes less than 10 minutes) on the one hand and, on the other, writing reliable encyclopedic content (which will involve hunting for more sources than the bare minimum required for notability, reading those sources, and then carefully writing up the text in a way that doesn't accidentally misinterpret those sources: this takes at least hours). – Uanfala (talk) 14:19, 25 February 2023 (UTC)
    @Blueboar I totally agree. We've got WP:BEFORE. We need WP:AFTER. Deletion decisions should be made based entirely on what's in the article. If people find better sources during a discussion, they should update the article to include them. Otherwise, all we end up with is an article that's still not sourced, and that's a disservice to our readers. I've largely stopped nominating articles for deletion because the arguments at AfD drive me nuts. -- RoySmith (talk) 18:51, 24 February 2023 (UTC)
    Why? Did someone impose a WP:DEADLINE that can be triggered by an AFD nomination, and forget to tell the rest of us? Maybe the rule should be that if you nominate an article for deletion, and your BEFORE search proved inadequate, then it's your job, as the sloppy nominator, to take any sources I provided to you, on the silver platter of the AFD page, and stick them in the article yourself. Why should I have to do any extra work at all, just because some editor was so stupid as to think a modern, thousand-bed teaching hospital probably never had anything written about it? WhatamIdoing (talk) 20:40, 24 February 2023 (UTC)
    Maybe the solution is if the AfD closes as keep the nominator should remove all unreferenced information from the article. Then per WP:BURDEN anyone wanting to return the removed text would have to reference it correctly. -- LCU ActivelyDisinterested transmissions °co-ords° 21:33, 24 February 2023 (UTC)
    How does blindly removing content, especially after I've handed you the sources needed for the article, help any readers? Why should an editor being stupid (every basically functional adult knows that it's not possible to build a huge teaching hospital without both government agencies and newspapers taking notice of the event) result in an article being gutted by the stupid editor? WhatamIdoing (talk) 00:36, 25 February 2023 (UTC)
    Why would an editor stupidly waste their's and other editors time finding good sources that they then don't add to an article, why? Improve the article, everyone wants good articles. Maybe stupid ideas are required to combat stupidity. -- LCU ActivelyDisinterested transmissions °co-ords° 01:10, 25 February 2023 (UTC)
    Perhaps everyone can avoid calling editors stupid? For better or worse, different editors have different expectations about how to work collaboratively towards building better articles. Preferring one method or another doesn't mean one editor is less intelligent than another. isaacl (talk) 22:17, 25 February 2023 (UTC)
You're not going to have any luck with this, unfortunately. The trend for what people (or at least the subset of people who participate in the discussions on this topic) want to happen is actually in the opposite direction.
In lieu of that, the best way to accomplish what you are suggesting is to proactively source and improve articles before they are nominated. On that note, there are about 129,000 candidates. Gnomingstuff (talk) 21:25, 25 February 2023 (UTC)


We shouldn't try to force this (nor bad-idea-wp:before) onto the ham-handed scale of inclusionist vs. deletionist. With some exceptions (e.g. wp:not and wp:speedy type) which are generally not disputed where article should not clearly not even exist, AFD is about wp:notability. A key part of the job of building the article is establishing wp:notability, which generally means finding and including a couple of GNG sources. Without that, in Wiki terms, you haven't really created anything. Like me handing you a windshield wiper and saying "here's the car I built, it just needs somebody to complete it" :-) Once that is done, it doesn't go to AFD. AFD is about articles that shouldn't exist, not about articles that need improvement. Sincerely, North8000 (talk) 19:49, 24 February 2023 (UTC)

I don't agree that "establishing notability", by which you seem to mean "proving notability to the satisfaction of editors unfamiliar with the subject" is a key requirement for an article. A 100% unsourced substub saying nothing more than "Cancer is a kind of disease" is not eligible for deletion because Cancer is an incontestably notable subject. Proving that will happen, automatically and incidentally, when the article is expanded (because WP:MEDRS), but notability depends on the real world, not on the current state of the article, even if the current state of the article is little more than a maker's nameplate and a windshield wiper. WhatamIdoing (talk) 20:47, 24 February 2023 (UTC)
An article containing only "Cancer is a kind of disease" should be pushed to draft space, and continually creating unsourced articles is disruptive behaviour as it's a massive time sink for other editors. -- LCU ActivelyDisinterested transmissions °co-ords° 21:30, 24 February 2023 (UTC)
Why should an article about an obviously notable subject be hidden in draft space? The one thing that we know about draftspace is that it will get less attention from other editors. Is that what you really want to accomplish with an article on a clearly notable subject? Do you want to establish a rule that effectively says "If you don't want to spend several hours working on this yourself, then don't create it at all"? WhatamIdoing (talk) 00:15, 25 February 2023 (UTC)
Because it's obviously not ready for main space. -- LCU ActivelyDisinterested transmissions °co-ords° 01:12, 25 February 2023 (UTC)
@WhatamIdoing: Of course there is a range of situations. I think that your Cancer example is so non-typical (new article= has not had an article in Wikipedia with absolutely obvious wp:notability) that I don't think that it is useful for the discussion other than to illustrate the point that wp:notability goes with the subject not the content of the article. Also proving is a pretty extreme degree, beyond what I was talking about which is that if you don't have some sources, you haven't done the basic job of starting an article. But my main point to the OP was: put in a couple GNG sources, and you don't need to worry about AFD, even if the article quality is low. And my main vague idea that I'm promoting is that step #1 of writing an article is finding sources. North8000 (talk) 21:50, 24 February 2023 (UTC)
Yeah, creating an article based on sources that don't establish notability, or especially based on no sources at all, is functionally equivalent to adding the subject's name to a categorized list of "maybe-notable" topics. Either the material in the article is unencyclopedic fluff sourced to trivial mentions/non-independent bodies, or it's not sourced and should be deleted; regardless, it doesn't serve readers as a trustable encyclopedia entry. JoelleJay (talk) 23:50, 24 February 2023 (UTC)
Does it? Do you really think that having little blue clicky numbers is what causes readers to trust an article? I don't. There are multiple factors that affect readers' willingness to trust a page, and I doubt that little blue clicky numbers even make the top 10. Readers barely notice them and almost never click on them. The biggest factor seems to be that the page aligns with their expectations.
I'm also not talking about "maybe-notable" topics. First of all, readers don't care whether the topic is Wikipedia:Notable. They care whether the topic contains the information they're looking for. Think about it: Have you ever heard anyone complain that Wikipedia had the information they were searching for, and they thought this was a bad thing? Can you imagine what that would even sound like? "Yeah, @JoelleJay, you people over at Wikipedia are totally screwing up. I put <niche subject> in Google, and Wikipedia had a short little article on it that answered my question. What a bunch of garbage. Why did you all even bother writing the thing I needed?" I've never heard anything even remotely like that, and I'll bet nobody reading this page has, either.
Second, I find that these are still red links:
Do you doubt that any of these are notable subjects? I don't, and I doubt that you have any qualms about those subjects either.
I believe that a unsourced substub that says "French Renaissance sculpture is sculpture produced during the French Renaissance", followed by a ==See also== list of relevant sculptors would be better for readers than nothing at all, which is what we have now. Do you really believe that nothing at all would be more informative/educational/helpful to readers? WhatamIdoing (talk) 00:30, 25 February 2023 (UTC)
Ok, I have sneakily shot all these foxes by redirecting them to the relevant section of Sculpture in the Renaissance period. This is a recent machine-translated version of the Spanish article, and has many, many faults, but its sections are still much better than nothing. Btw, we didn't even have Italian Renaissance sculpture until I did it for last year's Core Contest. Johnbod (talk) 17:15, 25 February 2023 (UTC)
People complain all the time that the wikipedia entry for X topic is totally useless and/or has outdated, incomplete, or incorrect information. They will be and are rightfully pissed when what they're expecting is an encyclopedia article and what they get is a stub with three sentences and a database ref, or 30 sentences of unsourced material they have no way of easily verifying, or 30 sentences of pure trivia sourced to an interview. An article that is not based on notability-granting sources cannot provide an encyclopedic summary of the topic even if it really is notable, so what questions could it be answering anyway? Of course your Renaissance examples shouldn't exist in mainspace! Having some See Also links makes it even worse since they're inevitably providing an unbalanced view of the topic to boot, so yes readers are better off not getting a shitty info-sparse stub SEO'd to the top of Google results. A dedicated resource where that info is actively curated is going to be orders of magnitude more helpful than a tautology and random context-free links. JoelleJay (talk) 01:39, 25 February 2023 (UTC)
Pages can always be edited to be made more up to date even by people who don't know much about the subject. Pages that don't exist can't be. There's also little guarantee that the sources will always be available as books go out of print and website links get broken.KatoKungLee (talk) 16:52, 25 February 2023 (UTC)
JoelleJay, to say these Renaissance sculpture articles should not exist in mainspace seems to me in contradiction to the way this encyclopedia evolved. Originally every article was like that. The goal was to make progress, not write only articles that had some minimum number of references. There are far fewer articles that are in that state (clearly notable, completely unreferenced) any more, but to say they are a bad thing is to say that the community's attitude to article creation has changed, and I don't think that's true. WP:REDYES is similar: it exists because incompleteness is seen as not inherently bad -- the missing material encourages others to write to fill the gap. Perhaps your take on this is becoming more common, but I would be surprised if it has become the majority view. Mike Christie (talk - contribs - library) 12:26, 25 February 2023 (UTC)
If it has no content it should just be redirected, for instance we don't have an article on "German Renaissance sculpture" but we have a very informative article on German Renaissance or even Sculpture in the Renaissance period#German sculpture. Why should we give readers a poorly authored article when we have a fairly informative article on the greater subject or a well written section covering the topic. If someone wants to create an article that actually offers the reader more value they can then remove the redirect. This isn't the early days, we already have many good articles that are more informative than some stub that blocks readers from finding what they are looking for. -- LCU ActivelyDisinterested transmissions °co-ords° 13:16, 25 February 2023 (UTC)
I agree that if there's an appropriate redirect target that's better than an unreferenced stub. That wouldn't be the case for many articles related to recent events, though. If something like the Staffordshire Hoard were to be discovered today, an article titled Worcestershire Hoard with one sentence and no references would be fine, as far as I can see. It's there to be improved and it's better than a red link, which in turn is better than no link. If the point is these situations would be rare, then I agree with you, but when these rare situations happen I don't think there's anything wrong with these stubs. Mike Christie (talk - contribs - library) 13:24, 25 February 2023 (UTC)
The fringes of all arguments is why there is WP:IAR, if something new is emerging then other editors will shortly expand the article. It shouldn't though be standard practice for topics that may or may not be notable, that is the area where it wastes other editors time. -- LCU ActivelyDisinterested transmissions °co-ords° 16:59, 25 February 2023 (UTC)
An article that gives so little information is of no value to readers, even if it is on a notable topic that can be expanded by other editors. I don't see how this is debatable. If the article is taking up a title that would be much better served as a redirect to a more comprehensive section on another page, it is actively doing a disservice to readers. We aren't in the early days of Wikipedia where people visiting had no preconceptions of accuracy or completion or encyclopedicity; nowadays readers are arriving from Google search results anticipating an encyclopedia article, and especially when the topic is something "obviously notable" they will expect it to be much more informative than a couple sentences. JoelleJay (talk) 23:30, 26 February 2023 (UTC)
An article that gives so little information is of no value to readers is an opinion, not a fact. It is also an opinion that is not shared by everyone. I have personally found an apparently abandoned substub to be very useful on occasion. (I have also found several long articles to be useless, including material that was cited. Does anyone happen to know what it means to provide log management and analytics services that leverage machine-generated big data to deliver real-time IT insights? I know what all the individual words mean, but I still have no idea what this company does.)
I wonder how you know that nowadays readers are arriving from Google search results anticipating an encyclopedia article, and especially when the topic is something "obviously notable" they will expect it to be much more informative than a couple sentences. Have you talked to any non-editors about their expectations? Have you asked them if they would rather find nothing at all, than to find only a basic definition or substub, like "Alice Athlete (1899–1962) competed in the Olympics for Country"? WhatamIdoing (talk) 00:12, 28 February 2023 (UTC)
Have you talked to any non-editors about their expectations? Of course I have? People in my lab and my social community constantly use Wikipedia as an overview source and importantly as a place to find comprehensive refs to read for themselves. As soon as they learn I edit Wikipedia they have complaints about how uninformative and incomplete particular pages are (usually this is for individual genes/proteins; some of the sporadic biomed edits I make are following up on such comments). Most of them get there from the top results on Google, which they expect to contain some information. JoelleJay (talk) 04:17, 28 February 2023 (UTC)
But do they say that they would rather have nothing at all? BeanieFan11 (talk) 16:04, 28 February 2023 (UTC)
Well, they regard clicking on the wikipedia link a complete waste of time, so I would say yes. JoelleJay (talk) 23:33, 1 March 2023 (UTC)
Irrespective of whether a lay reader's trust is informed by the presence of blue clicky numbers, the blue clicky numbers are the epistemic cornerstone of everything we do here. signed, Rosguill talk 07:07, 25 February 2023 (UTC)
I believe that the little blue clicky numbers are important to editors, but they do not seem to be important to readers. WhatamIdoing (talk) 00:13, 28 February 2023 (UTC)
They're important to building a reliable encyclopedia; readers indirectly care about them, even if they don't directly do so. BilledMammal (talk) 01:32, 28 February 2023 (UTC)
I am a reader before I am an editor, and I find myself quite annoyed when I click into a Wikipedia article only to find it is essentially nothing. For me this is most common for species articles, which are often a single sentence or two. Maybe it has a link to IUCN page. In that case, I have wasted my time and could have gone directly there from the google search. An unsourced sentence and a list of see alsos sounds even worse than that. CMD (talk) 13:17, 25 February 2023 (UTC)

Rather than mass get rid of stubs because they're stubs (like some are suggesting), why not improve them?

An idea: At User:BilledMammal/Lugnuts Olympian stubs, there's discussion on whether to propose mass getting rid of Lugnuts' stubs by the thousands because they're stubs. I have what I think would be a much better plan: start a project called something like Wikipedia:WikiProject Fix Lugnuts' stubs where we give rewards for those who can do so, e.g. I'd be willing to give a barnstar to someone who can expand two of his so-called "permastubs" into something that would pass the criteria at WP:DYK; one for three further; and one more for each five additional. I think something like this would be much more productive for the encyclopedia than mass getting rid of stubs (and something like this where we give rewards for stub article improvers would work for any topic area, I think). Thoughts on something like this? BeanieFan11 (talk) 02:21, 1 March 2023 (UTC)

We should keep and work to improve all sports biography stubs that include at least one reference to a source providing significant coverage of the subject, excluding database sources, as is required by Wikipedia:Notability (sports). We should delete or redirect all such stubs that do not meet that minimum requirement after an AfD debate of a week or more. Information about such non-notable athletes can be presented in list articles, or team/season articles. Such freestanding biography articles can easily be recreated if the necessary significant coverage is discovered and cited. Find the significant coverage first, and then write the article. Cullen328 (talk) 02:38, 1 March 2023 (UTC)
That proposal isn't to get rid of them, it is to temporarily move them to draft space unless adopted by a wikiproject or user, in which case the articles would be moved to project or draft space. It also isn't because they are stubs; it is because of WP:GNG and WP:SPORTSCRIT #5 issues and because the quantity makes it impossible to deal with through our normal processes. BilledMammal (talk) 02:41, 1 March 2023 (UTC)
Both of you are missing my main point: should we create a project where we give rewards for stub improvers? BeanieFan11 (talk) 13:51, 1 March 2023 (UTC)
I don't think the point was missed. WikiProjects are just groups of editors sharing common goals who choose to collaborate together on a WikiProject page. If you can find other people interested in participating in a project that works together on improving stubs, great! The key first step is finding interested editors; after that, the group can decide how it should organize (within of course bounds of general community consensus). isaacl (talk) 17:41, 1 March 2023 (UTC)
AGREE - First, the 24/7 scheming against Lugnuts is harassment and it shouldn't be allowed here. It does not matter whether Charles Ponzi or Jesus wrote an article. All that matters is the article's content. We do not want a wikipedia where articles get changed and deleted over the latest hottest trend of the week.
Mass deletions of articles or mass drafting of articles is completely against everything this site should be. Massive red flags should be raised about what's really going on at this site if that is ever seriously considered. Willie Magee (cyclist) has no involvement in anything related to Albert Johnstone. These two articles should not be compared in any way.
Articles should also not have to meet standards that do not exist yet. If a law is made tomorrow that every wikipedia article has to mention the word "jello" in it or be deleted, 99% of wikipedia articles would have to be deleted. That would never be possible even if everyone on the site worked on it and tons of good articles would be gone for silly reasons.
BeanieFan11 is 100% correct. Wikipedia users should always try to improve an article over going for the speedy deletion. Users wanting to get articles deleted should have to show proof that they performed a basic Google search and checked online newspaper databases, Google Books and other recommended free wikipedia sourcing options first. They should also have to open up a help topic on one of the sections here looking for more sources. There's plenty of people who would be interested in helping, just we have to know about it. KatoKungLee (talk) 17:36, 1 March 2023 (UTC)
It sounds like DYK already does this? If anyone fixes up a stub into something more (and that article has at least one item of interest), they can get a shiny DYK question mark. CMD (talk) 17:52, 1 March 2023 (UTC)
What I am suggesting is having a project which gives out barnstars to those who can expand stubs – I think a barnstar would be more motivating than a small  . BeanieFan11 (talk) 18:02, 1 March 2023 (UTC)
I would have thought it the other way around, although I'm not sure how this could be studied. Wikipedia:De-stub-athon may interest you. CMD (talk) 18:08, 1 March 2023 (UTC)
It doesn't really matter—if there are people interested in doing one, the other, or both, they can just do it! isaacl (talk) 23:50, 1 March 2023 (UTC)
  • No to both - No, we shouldn't delete (or delete-via-drafticiation) those stubs. The stubs do not have much positive or negative effect on the project or anyone else. Mass creation of so many stubs was undesirable, but also not disallowed, and the creator has since been indeffed. We also shouldn't be celebrating those creations by creating a special event just for them. We have millions of stubs that need attention, with many subjects a whole lot more pressing than an Olympian. I cannot imagine there would be consensus for mass-draftify or mass-delete, and while anyone can start an article improvement drive if they wish, I'll register my opinion that it's not a good idea, either [to focus on Lugnuts' articles]. Deal with them on a case-by-case basis if it's an area you care about, knowing that the creator won't be doing it any more, and that they're not doing any harm or any good if you choose to ignore them. — Rhododendrites talk \\ 18:16, 1 March 2023 (UTC)
  • The correct solution is two parts: 1) fix the stubs that you can fix (if you want to) and just do it yourself, without asking anyone to help at all (because this is a volunteer organization and no one can make anyone else help with anything) 2) If you find a stub that you believe cannot be fixed (and, no one else can tell you that you don't believe it, it's your belief) then tag it for deletion. That's how we proceed. We neither delete them all (because some may have reliable sources we can use to improve them) nor do we demand that they are all improved (because some of them may never be able to be improved). Instead, we simply take on whatever work we feel like doing to make Wikipedia better, and we don't tell anyone else they need to do it also. That means we don't tell anyone they have to improve the articles along with us (we just do it) and we don't tell anyone they have to tag the articles for deletion (we just do it). In the entire history of Wikipedia, among all editors that have ever interacted with the encyclopedia, nothing has ever gotten done because we demanded that others do it. Instead of starting threads telling people what they should do, just do it yourself. If you don't feel like doing it yourself, don't. --Jayron32 19:03, 1 March 2023 (UTC)
    I'm not sure how proposing a WikiProject constitutes "demanding" anything of anybody, and I'm not sure how anyone could draw that conclusion in good faith. The entire purpose of the village pump is to bring things to a wider audience. Gnomingstuff (talk) 18:21, 2 March 2023 (UTC)
  • No need to delete or immediately expand them all. Just let them be stubs until interested editors expand them, or nominate them for deletion, like any other stubs on Wikipedia... ---Another Believer (Talk) 19:06, 1 March 2023 (UTC)
If someone wants is put forward a suggestion on something let's wait until they do, rather than going over this now and if and when anything is suggested. -- LCU ActivelyDisinterested transmissions °co-ords° 19:46, 1 March 2023 (UTC)
@ActivelyDisinterested: Well, the proposal has been made: Wikipedia:Village pump (proposals)#RfC on draftifying a subset of mass-created Olympian microstubs. BeanieFan11 (talk) 15:15, 2 March 2023 (UTC)
Thanks for alerting me, I'll comment there. -- LCU ActivelyDisinterested transmissions °co-ords° 15:18, 2 March 2023 (UTC)
It would be great if you can find a group of editors interested in improving the articles in question. Ultimately, everyone wants more comprehensive articles. isaacl (talk) 00:08, 3 March 2023 (UTC)