Wikipedia:Village pump (idea lab)

From Wikipedia, the free encyclopedia
 Policy Technical Proposals Idea lab WMF Miscellaneous 
The idea lab section of the village pump is a place where new ideas or suggestions on general Wikipedia issues can be incubated, for later submission for consensus discussion at Village pump (proposals). Try to be creative and positive when commenting on ideas.
Before creating a new section, please note:

Before commenting, note:

  • This page is not for consensus polling. Stalwart "Oppose" and "Support" comments generally have no place here. Instead, discuss ideas and suggest variations on them.
  • Wondering whether someone already had this idea? Search the archives below, and look through Wikipedia:Perennial proposals.

Discussions are automatically archived after remaining inactive for two weeks.

« Archives, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45

idea: new mainspace entry, to list major or noteworthy articles from magazines, or other periodicals[edit]

Here is a new idea that I would like to present here for consideration, and discussion. As you know, Wikipedia is a resource with sweeping comprehensiveness. We have articles here for every sitcom, every commercially produced record, every cartoon, every notable band, every king or president, every congressman, every legislature, every published book or novel, every amino acid, the list goes on and on.

ok, here is my question. given our comprehensive nature in terms of scope, topics, and content, what would be some possible approaches to create an index of every notable article published? I.e., articles from magazines, or journals, or various reputable periodicals?

to put it another way, if a novel becomes notable just by being published, then why shouldn't the same hold true for notable magazine articles?

for example:

  • if a public figure of major notability publishes their thoughts on their own main field of interest. why shouldn't that be worthy of inclusion in some article that serves as an overall index?
  • as another example, if a major public figure publishes their views on a n issue of major importance, then why shouldn't that be deemed notable to be listed in a major entry on Wikipedia serving as an index of major articles in magazine and periodicals?

some arbitrary examples of articles which clearly should not be included:

  • ordinary news articles on every day events
  • movie reviews of a non-noteworthy nature.

Ok, so clearly, based on some of the above, the standard of whether a magazine/journal article meets WP:Notable would hinge upon some possible parameters, amongst others:

  • if the author of the article is notable in general.
  • if the author has a notable connection to the subject matter of the article, which makes it worthy of inclusion
  • if the article itself is of sufficient noteworthiness to warrant inclusion
    • obviously "sufficient noteworthiness" is a concept that we would need to clearly work out specifically, and to define.

ok, obviously a number of details would need to be worked out. we would need to define clear standards of notability, in terms of what would be included, and what criteria and what basis would be used to determine what could be included. yes, that would all need to be clearly delineated.

Ok, so what are some folk’s thoughts here on this? I would truly welcome any feedback, ideas, input, or comments on this idea. Thanks! Sm8900 (talk) 17:14, 19 December 2022 (UTC)Reply[reply]

Re: if a novel becomes notable just by being published. Just being published isn't sufficient for notability for novels or any other type of book, see WP:NBOOK. Schazjmd (talk) 18:00, 19 December 2022 (UTC)Reply[reply]
'Notability' as the term is used in Wikipedia policy, concerns article topics - things we decide we can adequately write about ourselves, based on available sources. It isn't an assertion about the importance of a subject to the outside world, and it would be improper for Wikipedia to engage in such subjective categorisation of such material. AndyTheGrump (talk) 18:07, 19 December 2022 (UTC)Reply[reply]
I agree with Schaz and Andy's comments, and of course there are the NOTs: not everything, not directory, etc. Nevertheless, I've often thought that what would be useful to our readers is if every citation's author and publisher were blue links, so that readers would always be able to learn more about the sources used in an article. I've thought about having NPROF-style auto-notability for reliable source publishers and authors, but there are alternatives that don't involve changing Wikipedia's notability guidelines, such as using Wikidata to be the index of information about publishers, authors, and works. Levivich (talk) 18:23, 19 December 2022 (UTC)Reply[reply]
I was going to say that Wikidata already has items for cited sources, with corresponding items for all of their associated attributes. (It's what I found makes adding info to Wikidata so time consuming.) isaacl (talk) 23:32, 19 December 2022 (UTC)Reply[reply]
If you have not seen this before, this category with many subcategories may be of interest Category:Works originally published in periodicals (there may be other parent categories/lists that are also useful)g, which I think will include many such notable articles categorized by things like type of publication and/or country and so on for examples of what articles which have had articles. Skynxnex (talk) 18:23, 19 December 2022 (UTC)Reply[reply]
welll..... what about, Yes, Virginia, there is a Santa Claus...? so here is our first contestant. see what I mean?> Face-smile.svg so how about some general guidelines, which might perhaps proceed from the few but genuinine items already fitting within this category? And presumably the implicit guidelines which govern items for inclusion of this type? Sm8900 (talk) 21:55, 19 December 2022 (UTC)Reply[reply]
@Sm8900 in very broad terms, if a topic is notable it may support a standalone article. I can't see us recording for example every tweet someone produces (...publishes their thoughts...); but if someone publishes a statement about something it may already be included in their biography, etc. I think I'm a bit lost on what you want to accomplish; Would this idea result in a new page being produced? If so, would you write (in your sandbox) one draft article/list/etc that you think wouldn't be inclusion-worthy today, but if this idea was adopted would be? Having an example would help me give better feedback. — xaosflux Talk 14:01, 26 December 2022 (UTC)Reply[reply]
those are good ideas, @Xaosflux. for now I'm just brainstorming on how such an entry might be written. if I can find a magazine article that is worth covering, then I will write a draft. So your point is noted. thanks. Sm8900 (talk) 14:28, 3 January 2023 (UTC)Reply[reply]

Section break 1[edit]

ok, how about this as an idea? an article from a magazine or periodical can be considered noteworthy for inclusion, if it has signfiicant published coverage in third-party media outlets; in other words, if it has received coverage from various media outlets elsewhere, in addition to the media outlet that published the article itself.

I think this standard is so simple as to be self-evident or a truism, since it is simply the same standard that is teccnically applied to any current topic, current event, or prominent individual or creative work that is being considered for inclusion as an entry on wikipedia. --Sm8900 (talk) 15:19, 21 December 2022 (UTC)Reply[reply]

  • No, I feel guilty about being overly negative, but most of the possible outcomes undermine what Wikipedia actually is. First problem is how do we define a "major article"? Taking the example of science articles (purely because I need to take a practical example in a field where I feel competent to guess what the proposal might do, in concrete terms), I don't know of any journal that distinguishes between "major" and "minor" articles. Are we going to define everything in Nature as major/notable, but everything in Annual Reviews of Plant Biology as probably not? But Nature publishes novel research, which is a primary source, and Annual Reviews of Plant Biology publishes reviews, which are secondary sources, and therefore highly appropriate as references in articles concerning plant biology. But if we include ARPB, we have to include an enormous multitude of other, equally reputable journals that publish review articles from time to time, and of which most readers will never have heard. How do we compile that list? We could, I suppose, do it by impact factor, but there are big debates in the science world at the moment about how fair impact factor is as an indicator of the true relevance of a journal or an article, and impact factors are specific to academic publication, so they presumably wouldn't extrapolate to deciding what's a "major article" when it comes to military matters, or the film industry. And how do we then deal with the fact that some subjects, like gene editing, or the role of gut microflora in health, will produce truly vast lists of relevant articles if we accept all reputable news-sources and science-journals? And what do we and our users do with this list? This prompts the really big question: if all we're doing is a big, automated search of all published literature, and dumping the result on the user for them to sort out, what makes us better than Google? Our strength is in giving an approachable, well-written overview of a subject based on good secondary sources, with sufficient carefully-chosen references to give our readers a foothold in the broader literature. Our strength is in our human, collaborative editing, and we shouldn't be tempted by the siren of enormous automated data-processing. Google will always be far, far better at it than we are. The true indicator is the fate of the Science Citation Index, which did in the field of the sciences what I think this idea is intended to do more generally. Thirty years ago, I and my colleagues used the science citation index all the time. Now I don't know anyone who uses it, simply because Google does the job much better. Automated literature indexing as a field distinct from broad internet search providers is a comet-watching dinosaur donning woolly underpants and staring extinction in the face. Elemimele (talk) 11:14, 22 December 2022 (UTC)Reply[reply]
@Elemimele, you are making some outstanding points above. However, I actually have changed my own initial proposal somewhat. I agree that creating or compiling an exhaustive list, or index, or summation of all notable articles is simply not needed, and is not feasible or viable at all; actually that was not what I meant to propose, but I see how my proposal above might seem to suggest that somewhat.
here is my current actual suggestion; if a published article has received truly major media coverage, then perhaps we might consider whether it is viable for an entry. such as, if hypothetically, if a leading politician wrote an article, which then subsequently attracted major news coverage and headlines worldwide, over its importance, or dramatic content, etc., then perhaps we might deem it worth of an entry here.
Cover of the September 1917 edition of the magazine featuring Sherlock Holmes
we have numerous articles for short stories, many of which appeared as items in a periodical or magazine when first published. In fact, the categories cited above by another commenter, as being focused on items from a magazine, are mainly populated by published short stories. in the 19th century, many authors like Charles Dickens and Arthur Conan Doyle, published short stories in notable publications. The Strand Magazine is one of the best-known magazine, for fans of Sherlock Homes.
my main thought is that if we did move ahead with this in any way, it would still not precipitate any major change in content, or expansion of content overnight, at all. On the contrary, this idea would be implemented on an extremely narrow basis. We absolutely, would not seek to publish any type of index, that would be exhuastive or comprehensive in regards to all noteworthy articles etc. as you rightly point out above, any such effort or attempt would be highly inadvisable. Your points on this are spot-on.
so therefore, I propose that perhaps we might all keep in mind for the future that perhaps if an article appears on the landscape sometime in the future, perhaps we might simply consider now as to whether such an article could be deemed worthy of its own entry here, and what might be some valid and useful criteria for doing so.
and in addition, here is another thought; a set of regular weekly columns by a columnist, who has had some type of noteworthy role or impact, should be worthy of inclusion here, simply as being on the same level and within the same general category as any other major type of published work. So I just wanted to add that idea and note as well.Ok, so please feel free to reply, for anyone here who may wish to do so. And @Elemimele, I greatly appreciate your thoughtful, helpful, and well-reasoned insights above. Thanks! Sm8900 (talk) 16:41, 22 December 2022 (UTC)Reply[reply]
I fail to see the issue. We already have articles on such noteworthy publications, and no one is suggesting that they should be deleted. Über die von der molekularkinetischen Theorie der Wärme geforderte Bewegung von in ruhenden Flüssigkeiten suspendierten Teilchen is a snappily named article on one such scientific publication, and no special guidelines (whether notability or MOS) seem to be needed. April 26 Editorial is a completely different example, this time of a front page article in China. WP:CREEP: nothing new seems to be needed, or at least no actual problem has been identified so far. Fram (talk) 14:42, 3 January 2023 (UTC)Reply[reply]
@Fram, actually, my only point was whther we can have entries on individual articles within such publications. to recap, we already have numerous articles on numerous short stories, many of which also started out as literary works published in magazines, so that was one large part of the basis for my initial thoughts on this.
yes, the entry that you cite on the 1905 article by Einstein is a perfectly valid example of one such entry. so you have helped me very much to broaden my own knowledge on this. thanks. Sm8900 (talk) 22:56, 3 January 2023 (UTC)Reply[reply]
Sm8900, the answer is here. Does the individual article meet those criteria? If so, it's possible (though not guaranteed) that we should have an "article about the article", as it were. If no (and most articles are a "no", as very few receive such "meta-coverage"), then we should not. So, the answer is "Sometimes we should do that, usually we shouldn't". Seraphimblade Talk to me 01:38, 11 January 2023 (UTC)Reply[reply]
@Seraphimblade, I totally agree with your points above on this topic. that makes sense. thanks. Sm8900 (talk) 15:44, 12 January 2023 (UTC)Reply[reply]

Creating a feedback system (like/dislike) over time for articles of Wikipedia[edit]

Hi, nowadays in the contemporary World Wide Web, feedback system (via like 👍, and dislike 👎) is very common. But for an interactive system (like articles of Wikipedia, that content of an article changes over time) we should show "like/dislike over time", and the mean of "like number on a month" can be shown to the user on the top of an article, plus a diagram for "number of like/dislike in each day" for that month.

So I propose to add an option for 👍 and 👎 that after a click of a user, its color changes from gray to green/red, and showing the mean like/dislike number per last month at the top of that article, and additionally create a "like/dislike per day" diagram. Thanks, Hooman Mallahzadeh (talk) 12:49, 26 December 2022 (UTC)Reply[reply]

That seems like a great way to make off-wiki brigading more effective. Platforms like YouTube and Netflix have removed similar displays because of brigading issues. ScottishFinnishRadish (talk) 13:04, 26 December 2022 (UTC)Reply[reply]
@ScottishFinnishRadish and Hooman Mallahzadeh: No it's not. The platforms you referring to—their reason for removing it was moot and completely trash—but that's another discussion. Anyway I do agree that a like—disklike feedback does not belong on Wikipedia where content rapidly changes. I mean, numbers don't really mean anything on here. We'd be better off with users expressing what they like or dislike about an article and editors having access to this info. But then, I don't think this would be problem-free either. — Python Drink (talk) 09:29, 15 January 2023 (UTC)Reply[reply]
@ScottishFinnishRadish We can propose solutions for the brigading issue, IP and username supervision is one technique.
One of the benefits of feedback is that finding bad/good articles in the opinion of users becomes more convenient, and we can make a "bad article" more user-friendly and more fluent to become a good article. Hooman Mallahzadeh (talk) 13:12, 26 December 2022 (UTC)Reply[reply]
I think this would attract more feedback about if someone likes or dislikes the topic of an article, as opposed to the coverage of the topic. — xaosflux Talk 13:50, 26 December 2022 (UTC)Reply[reply]
@Xaosflux I don't think so! I think "like number" attracts how well that topic/concept is described. If one finds and understoods his intended and required information in a fast and fluent way, then he/she likes that article, otherwise dislike.
We can show this sentence in the feedbacking system: "Was this article helpful?" and then ask him to feedback the article, to avoid voting about topic instead of coverage of that topic. Hooman Mallahzadeh (talk) 14:02, 26 December 2022 (UTC)Reply[reply]
@Hooman Mallahzadeh thanks for the note, please see Wikipedia:Article Feedback Tool, where a previous experiment in this was done. In 2014 it was disabled following this discussion: Wikipedia:Requests for comment/Article feedback. — xaosflux Talk 14:05, 26 December 2022 (UTC)Reply[reply]
@Xaosflux Wikipedia follows 1% rule, i.e., only 1% of users (or readers) are the actual writers of articles. So I propose history of feedback (past like and comments) should be kept hidden for 99% and only other 1% can view history of like number and comments. This 1% should do something to improve that article and remove defects of article. One way to find this 1% is to check and inspect edit number of a user, if edit number is greater than for example 2000 edits, then this feedback history become apparent for that user. Hooman Mallahzadeh (talk) 14:33, 26 December 2022 (UTC)Reply[reply]
I like the idea, especially to do with bias and uneditorly behaviour, but I agree on the brigading, and that if there are problems the reader should fix. Wakelamp d[@[email protected]]b (talk) 10:55, 27 December 2022 (UTC)Reply[reply]
These systems have had a massive negative impact on society as they drive content towards extremism ie. the more emotional the reaction content the more likes/dislikes it gets, the more attention it gets. I would be very opposed to it. -- GreenC 15:00, 26 December 2022 (UTC)Reply[reply]
@GreenC The question of feedback is "Was this article helpful?" or "Did you get the answer of your question?" Yes(like)/No(dislike). In my opinion this is not emotional at all. Hooman Mallahzadeh (talk) 15:08, 26 December 2022 (UTC)Reply[reply]
Sorry my typo I should have said 'emotional content'. If people are motivated to edit based on number of likes they get (and people do get motivated this way) they will make content they discover gets more likes. There's a ton of academic research on this topic I'm probably not explaining it correctly. -- GreenC 15:15, 26 December 2022 (UTC)Reply[reply]
@GreenC Suppose some one wants a recent and clear photo of Tim Berners-Lee to use in his presentation. He opens Wikipedia and founds a picture of him at 2014. Now is 2022, and he can feedback about this article that:

Not facebook not like thumbs down.pngDislike : I did not satisfy my request, because the picture that is present is for 8 years ago.

This way, writers of this article is alarmed to upload a newer version of Tim Berners-Lee's picture.
I should note again that these comments are visible for only 1% (see 1% rule) but this 1% should do something to improve the article. For example this 1% somehow find a picture, or create a discussion at talk and wants a newer version. Hooman Mallahzadeh (talk) 15:38, 26 December 2022 (UTC)Reply[reply]
@GreenC@ScottishFinnishRadish@Xaosflux Briefly speaking:

99% + 1% of participators of Wikipedia can write feedback, but 1% of participators can view feedbacks and then modify the article.

This is the most important application of feedback. Hooman Mallahzadeh (talk) 16:26, 26 December 2022 (UTC)Reply[reply]
Wikipedia should be an encyclopedia where anyone can edit. It looks like the feedback system that you are proposing actually makes it harder for people to join the project and make edits, since they would just write the problems they had with the article in a feedback box instead. Knowing that the encyclopedia is not perfect and has many issues to be fixed is how I was introduced into editing Wikipedia.
Another big question is how to ensure that we can respond to the feedback efficiently. Having a text box soliciting feedback can invite a lot of writing since we have a lot of readers. 0xDeadbeef→∞ (talk to me) 16:47, 26 December 2022 (UTC)Reply[reply]
Above, the proposer notes IP and username supervision is one technique, but for this idea to work, we'd need to understand how this could be done without costing more effort in review than we'd gain. Nosebagbear (talk) 17:46, 26 December 2022 (UTC)Reply[reply]
I don't think this is a good idea. The basic premise here seems wrong to me - this is a site dedicated to writing an academic encyclopaedia, it is not a social media site designed to feed people content they want to see to keep them engaged - "like" and "dislike" are completely the wrong things to be asking about because people's emotional response to an article is not the kind of feedback that is useful in an academic context. You don't see academic journals asking their readers to react to the articles they publish.
Using this as a metric of how "good" an article is also strikes me as a poor idea - regardless of what you say a significant number of people are going to "like" or "dislike" and article on the basis of its topic, regardless of the actual content of the page. It is also quite possible to write an article that is a good read for a causal observer that is factually inaccurate, uses poor sources, misses major aspects of a topic, is biased etc. Structured processes like featured article reviews seem like a much better way of determining which articles are well written.
Asking people to leave comments on articles has been tried multiple times before and has always ended poorly - the final report on the v5 article feedback tool is worth reading mw:Article feedback/Version 5/Report. only 10% of the comments left were actually reviewed by an editor, and only 12% of the comments reviewed were actually marked as being useful. Consider the example you give above, is "the photo is old" a useful bit of feedback? I would argue that it isn't - the age of the photograph is already obvious to anyone looking at the page, and there are a multitude of reasons why that picture is being used that may not be obvious to a reader - it is entirely possible that there are no high quality, recent photos available under a free licence, for example. (talk) 14:16, 27 December 2022 (UTC)Reply[reply]
Adding on to what other users have said about how terrible like/dislike systems are in general, I think this is contrary to the purpose of Wikipedia in another way. The whole point of this project is that if people don't like an article, they can fix it themselves or express their concerns on the talk page. Thebiguglyalien (talk) 16:43, 27 December 2022 (UTC)Reply[reply]
I agree that the concept of like/dislike is not consistent with our mission. If the aim is to have readers feed back on missing, incorrect or outdated info, the talk page serves that purpose. However, what might be useful is to make that clearer for the reader and provide a one-click method of adding material to the talk page from the main space. As it is now, I believe, there's a realization needed that the talk page is the right channel, then a click to get to the talk page, then a click to start a contribution, then a click to publish the contribution. This could be streamlined, I believe. User:Ceyockey (talk to me) 17:06, 27 December 2022 (UTC)Reply[reply]
We could make a form/inputbox/whatnot to solict and streamline generic feedback on articles, but it comes with a workflow issue: there's a good chance no one will do anything about the feedback. — xaosflux Talk 00:00, 28 December 2022 (UTC)Reply[reply]
I'm more concerned that the only use will be for causing disruption of the sort seen at Talk:Adam's Bridge [1][2][3][4]. ScottishFinnishRadish (talk) 00:09, 28 December 2022 (UTC)Reply[reply]
Far better that the vandal focuses on the talk page than the main space page, but I agree it is nonetheless ANNOYING! User:Ceyockey (talk to me) 00:45, 28 December 2022 (UTC)Reply[reply]
@Ceyockey @ScottishFinnishRadish Wow on Adam's Bridge. I quite like stack exchange's system that you need x reputation to comment on controversial topics. Wakelamp d[@[email protected]]b (talk) 13:28, 28 December 2022 (UTC)Reply[reply]
@Xaosflux@ScottishFinnishRadish@Wakelamp In my opinion, a very brief and small feedbacking system, via 🤍 and ❤️, something like what already exists in the feedbacking system of Instagram, is very useful both for original writers of articles, and for all readers because makes Wikipedia more user-friendly. Hooman Mallahzadeh (talk) 06:39, 12 January 2023 (UTC)Reply[reply]
What is the supposed benefit of such a patently terrible and useless proposal? Besides allowing even more trolling and baseless whining from drivebys to waste editors’ time and prevent work on actual content issues? Dronebogus (talk) 07:13, 12 January 2023 (UTC)Reply[reply]
@Dronebogus Feedback is both for readers and writers and for analyzers of articles, and is different from "Talking about an article" that is already implemented in Wikipedia. Readers that do not have time to modify, do the "minimum action" that they can do for that article (e.g., changing 🤍 to ❤️ is the least modification that a human can do for an article). This least modification has many signs for original writers. Original writers (that are 1% of readers according to 1% rule) alert that "This article was not good and should be modified". The other benefit is to collect statistics and categorizing articles and finding articles that should be rewritten and/or deleted.
See and read the topics:
  • Identify barriers to conversion
  • Find out more about your users
  • Improve UX
  • Improve accessibility
  • Get a heads-up on technical issues affecting customers
  • Connect “what” with “why”
Nowadays, feedbacking system and the method of creating and analyzing them (for example via Natural language processing) is one of the most well-studied areas of World Wide Web. There are many types of feedback, but changing 🤍 to ❤️ is the least feedback.
Do you really think that feedbacking system in Instagram is "patently terrible and useless proposal"? and "waste editors’ time"? If you think so, please tell them to remove that from Instagram.
Feedbacking system is different from "talk page" and has many benefits for readers/writers/analyzers, in a way that existing "Talk pages" do not have. Hooman Mallahzadeh (talk) 07:47, 12 January 2023 (UTC)Reply[reply]
I’m not impressed. Wikipedia is not Instagram. Dronebogus (talk) 07:50, 12 January 2023 (UTC)Reply[reply]
@Dronebogus Wikipedia is not Instagram is true, but Instagram is newer and has understand that this system is useful, and profits from many good benefits of this system.
Lack of feedback is a defect of Wikipedia and makes Wikipedia less useful for readers/writers/analysts. Existing convention for lack of feedback in WP should sometimes be broken. In the first glance it seems useless, but actually has many benefits that later becomes apparent. Hooman Mallahzadeh (talk) 08:02, 12 January 2023 (UTC)Reply[reply]
Feedback could be good for bias - but it could go to a hidden field so that there is less incentive for dog piling,
Being honest, if an article is biased for/against, then it is often impossible for a new editor to do much as they will be reverted or bamboozled with spurious arguments. Even as an experienced editor is it worth getting involved?
For everything else, what about having using proxies? Readability, class, importance, and percentage of good faith reverts. Wakelamp d[@[email protected]]b (talk) 08:40, 14 January 2023 (UTC)Reply[reply]
I strongly disapprove of adding a like/dislike system, for all the reasons stated by others. However, I do believe that Wikipedia could go further in some other form of encouraging feedback from readers, such as a comments section which is more obviously accessible than the talk page, and segregated such so as to not obstruct editor discussion on the talk page. However I have no idea how to do this in a way that doesn't violate WP:NOTSOCIAL or otherwise proves unhelpful (I most certainly do not want Wikipedia to incorporate anything akin to Fandom's useless comments sections). silvia (User:BlankpopsiclesilviaASHs4) (inquire within) 09:48, 12 January 2023 (UTC)Reply[reply]
@BlankpopsiclesilviaASHs4 Here you say

Wikipedia could go further in some other form of encouraging feedback from readers

"Encouraging feedback" is the convention of contemporary World Wide Web in the year 2023, and Wikipedia should not find an exceptional way for collecting feedback via "Talk page".
Why Instagram does not create a talk page for its posts? It is because "creating talk page" is much more expensive way for expressing opinions and requires much more mental attempt from readers, than using feedback.
In my opinion we should not make Wikipedia an exceptional website that uses "Talk page" to express opinions. It should be noted that new sections in "Talk page" is very low growing, and this expensive way for expressing ideas prohibits users from expressing many ideas that only can be expressed in feedback. No implementation of feedback means: losing many comments that can only be expressed via feedback, and cannot be expressed via "Talk pages".
"Feedbacking system" has advantages and disadvantages for Wikipedia, but in my opinion its benefits is much more than its disadvantages.
Additionally, techniques exists to overcome disadvantages. For example, we can make feedbacks hidden for new users and only confirmed users can view them. Other technique may exist to solve some disadvantages of implementing feedbacks. Hooman Mallahzadeh (talk) 10:31, 12 January 2023 (UTC)Reply[reply]
There's a case for allowing thanks from unregistered readers, and allowing praise for an article rather than having to pick an individual edit. As for criticism, a simple "I don't like it" is rarely helpful. What we really need is "the explanation of Bloggs' theorem is unclear", or "please mention its importance to Fijian flautists", which is better done on the existing (or easily created) talk page. Perhaps we should make that more accessible. Certes (talk) 12:18, 12 January 2023 (UTC)Reply[reply]

Category for articles without a talk page[edit]

Every now and then I come across articles without talk pages. Most recently this one. I just go ahead and quickly create the talk page most of the time. However wouldn't it be nice to have a way of grouping articles without talk pages? Ideally this category should be empty. FacetsOfNonStickPans (talk) 19:23, 27 December 2022 (UTC)Reply[reply]

@FacetsOfNonStickPans I really don't find anything wrong with a non-created talk page; simply existing without having anything useful on a talk page isn't a benefit, for example look at some recent new talk pages: Talk:Zootah, Talk:Jetcost, Talk:Jesus Loves the Little Children, Talk:Itha. In every one of those cases it would satisfy the idea above of an article "with a talk page", but most every one of those can just be speedy deleted as useless. I don't think a "category" would be a good way to deal with these either, if someone wanted to do something constructive they could use a query, such as this one. — xaosflux Talk 20:59, 27 December 2022 (UTC)Reply[reply]
Yes, there are number of reasons why a mainspace article may not have a talk page. And then that talk page may have no useful information such as WikiProject templates. I am an average non-technical editor; the link largely makes sense and those redlinks within the limit have indeed been identified. I will explore this more myself. Thank you! FacetsOfNonStickPans (talk) 22:54, 27 December 2022 (UTC)Reply[reply]
I remember the good ol' days before WikiProject tagging became widespread when most articles didn't have talk pages. For example, my first edit in the talk namespace was to create the talk page for the Dionne Warwick article. I think it's about time for you whippersnappers to get off my lawn! :-) Graham87 09:45, 28 December 2022 (UTC)Reply[reply]
I used to be quite conscientious about creating talk pages for articles I had started, including adding what I thought were appropriate project banners, but I have pretty much given up on that. Of the last 26 articles I have started, going back 2 1/2 years, ten still do not have a talk page. These days, I am more concerned about how an article fits into the hyper-space of related topics (i.e., how many rabbit holes are available from links in the article) than I am about whether a talk page exists to link to more or less quiescent projects. - Donald Albury 18:44, 28 December 2022 (UTC)Reply[reply]
My preference is to create a talk page if and only if there's something to talk about, whether that's discussion on how to treat some subtopic in the article or a banner assessing importance and quality. Certes (talk) 19:10, 28 December 2022 (UTC)Reply[reply]
Automatically deleting dormant/unactive talk pages, provided certain conditions are met, is also not possible. At the other extreme, identifying the most active talk pages does exist. FacetsOfNonStickPans (talk) 10:04, 29 December 2022 (UTC)Reply[reply]
I vaguely remember a proposal to automatically create talk pages whenever an article is created. It would have addresses edge cases, such as a user being unable to create pages due to permissions, but also unable to communicate on talk page due to being unable to create it. It would make it easier for people to contribute to a talk page, without the heightened complexity of creating one too. ~ 🦝 Shushugah (he/him • talk) 15:15, 31 December 2022 (UTC)Reply[reply]
If I understand the table of permissions at Wikipedia:User access levels correctly, everyone not subject to a block (including non-logged in users) is able to create an article talk page. isaacl (talk) 17:06, 31 December 2022 (UTC)Reply[reply]
Using the current talk page to express like/dislike has a few issues;
  • the existence of comments is not visible to readers of the article What about having the talk tab show the number of comments?
  • Many comments on articles are never acted on/seen. It would be nice if all the low importance, stud, start, B article comments that have not been acted on could be in one place for a project/rather than spread on thousands of pages. Wakelamp d[@[email protected]]b (talk)
Wakelamp d[@[email protected]]b (talk) 09:03, 14 January 2023 (UTC)Reply[reply]
@FacetsOfNonStickPans, why do you want talk pages to exist? Whatamidoing (WMF) (talk) 00:00, 4 January 2023 (UTC)Reply[reply]
Whatamidoing (WMF), sorry for the longish explanation. The reasons for why they already exist have been developed over time through consensus. Talk page messages like, "This is the talk page for discussing improvements to the ... article. This is not a forum for general discussion of the article's subject." explain this. Even where there is no discussion, there are a number of navigation links. WikiProject tags help in populating the respective WikiProject assessment statistics, the grouping which provides its own set of usefulness.
If I argue for the opposite, that I don't want talk pages to exist where there is no discussion, then I must also move the extra navigation tags elsewhere or do without them. If I do not want navigation tags, say for the WikiProjects, then I must accept that assessment statistics are not accurate quantitively. However pages above a certain quality rating all have talk pages; so this leads to the conclusion that only stub or start level articles are likely not to have a talk page as compared to b level or higher quality assessments. (I am not talking about talk pages for disambiguation articles.) On the other hand some WikiProjects have regular tagging competitions. There have been many discussions for improving talk pages, only very recently I remember one which was even displayed as a banner.
I started this village pump section with the aim of identifying how many articles do not have a talk page. xaosflux was kind enough to create an sql query on I soon realized that this query did not get rid of disambiguation articles. Also, I did not want to increase the limit to something like 10,000 just to see how many pages were returned since again, I didn't want disambiguation pages. This was the limit of my own understanding of the sql query. I realized that a quick search of past queries or google search would not help me adapt the query to my search needs, so I have placed learning sql in my to-do list.
These past few days I've been giving more thought to navigation links, say the difference or similarities between Categories and WikiProjects. However, as I think about this more, increasing the quantitative accuracy of WikiProject assessment statistics seems unnecessary since qualitatively they may still lack, and as I wrote, only pages which are rarely visited are likely not to have a talk page. Even if you want to search for edits made by an editor according to WikiProject, these missing talk pages shouldn't really make much of a difference.
So you've asked me why I want talk pages to exist. No, rather, as I initially started out with, I just wanted to know how many articles do not have talk pages and I thought a good way for that would be to create an automated category. I wrote "Ideally this category should be empty," ideal isn't always possible or needed. Now that I know sql queries can also get this answer, I've got my answer, irrespective of whether I can actually run a Quarry query or not as yet. FacetsOfNonStickPans (talk) 12:01, 5 January 2023 (UTC)Reply[reply]
There are several threads in this. First, we've no evidence that {{talk header}} provides a practical benefit to anyone. With the recent Wikipedia:Talk pages project, which brought us the auto-signing Reply button, it's probably less useful than it ever was before. If it were unambiguously useful, it should probably be part of the software for the page, instead of an optional template. The facility already exists in the software; it'd just be a matter of putting a message in MediaWiki:Talkpageheader (and maybe flipping a config switch somewhere), and then it would automatically appear on all talk pages. My guess is that this isn't used because it isn't as useful as it might seem at first glance.
I have a soft spot for article ratings, as volunteer-me has done well above 10,000 of them for Wikipedia:WikiProject Medicine. Keeping in mind that I'm therefore biased, I don't think that the current system is very good. People were talking a decade ago about storing that information in a separate, structured, dedicated space. This could be something vaguely like the way a File: page works: You see it all on the same page, but in the database, the image and the text on the page are stored separately, and clicking the Edit button at the top of an image on Commons opens the wikitext, not graphics editing software. Imagine a world in which information like quality ratings, categories, and associations with WikiProjects were stored separately, so WikiProjects could track an article, and ratings could be assigned to an article, without technically starting the talk page itself.
Because of the way we record quality assessments, there are no pages with any assessment rating that doesn't have a talk page. However, there could be articles that would qualify for relatively high assessment ratings and haven't yet been tagged. The mw:ORES system should be able to identify this (perhaps in conjunction with a SQL query in Quarry).
Pages like Wikipedia:WikiProject Directory/Description/WikiProject Medicine list editors who contribute to articles within a given area; however, Reports bot hasn't updated them for six months ("Task 2") and getting that active again would require someone with technical skills. I'm not aware of any way to get a reliable list of untagged articles for a WikiProject; keyword-based efforts such as User:AlexNewArtBot/MedicineSearchResult have, at least for medicine-related articles, tended to have a lot of false positives. The ORES work looked at identifying the main subject of an article, which I think would probably work better, but I don't know if they reached a useful point before they stopped. Even then it would be difficult: if you have a BLP physician who is notable for being an athlete, then identifying a single WikiProject (or general subject area) might not correctly identify the group of editors who would value the article the most. Whatamidoing (WMF) (talk) 01:25, 6 January 2023 (UTC)Reply[reply]
Just with regard to {{talk header}}
Whatamidoing (WMF), one less thing to be considered sacred despite the unanimous speedy keep in 2017; rather than just the entire template {{talk header}}, I specifically meant the line quoted, but as you said either it unambiguously stays or why at all. Maybe that option of adding the template or not is what is also important. It is only used on 700k pages and doesn't display on the mobile version unless switched. So then rather than merely a navigation message to new readers, a reminder to older ones, it is used as a categorizer as well. Some Wikipedians add it to every talk page they create, or maybe some sort of bandwagon effect for editors who predict edit warring.
Some description in the template documentation or a talk page summary (/FAQ style but no one has asked the questions as yet) on how the template originated would be interesting, such as a major discussion surrounding it, relevant research, other than just saying it originated with an edit in 2005. Then again, richer histories may also have no practical benefit, and the history is all there if someone is really interested. FacetsOfNonStickPans (talk) 17:14, 9 January 2023 (UTC)Reply[reply]
There were almost certainly no RFCs (here's what the RFC page looked like a week later; RFC wasn't a normal way to have such discussions back in 2005), but there were multiple attempts to delete it, which could give you some insight into what people were thinking back then. Whatamidoing (WMF) (talk) 22:05, 9 January 2023 (UTC)Reply[reply]
Whatamidoing (WMF), on summarizing TfDi Sept 2005, TfDi Oct 2005, TfDe Feb 2006, we get groups related to location, visibility, text, readers and aesthetics. Data usage was an initial concern, "huge performance hit". Reminding users to sign their comments has gone.
  • Location - where are the entry points for new editors and how does this template exist/coordinate with similar intent templates
  • Usage - universal/default versus the existing Template:Talk header#usage, "it's used on over 1500 talk pages ... speaks volumes for the fact that people think it's useful." Now it is 600k+.
  • Visibility - according to the number of edits, according to traffic on a talk page, according to most watchlisted pages, users can be given a option of hiding the message, shrinking the message (which expands every month)
  • Text - How many words are really needed? Why not add the guideline fixit as well.
  • Readers - How many readers actually read their welcome message, do users with more than 50 edits need reminding
  • Aesthetics - The recent changes to User Talk Page show that aesthetics are a concern, "just plain unsightly"
Merges include a longer version, the now deleted template:talkheaderlong, into the current one. An important merge happened in TfDi Oct 2020 related to reference search tools. Higher level long duration research could answer "who knows how much worse things might be without them?". (Quoted text is from the linked pages.) FacetsOfNonStickPans (talk) 18:51, 11 January 2023 (UTC)Reply[reply]
There are a lot of things that could be done here, if someone wanted to refine that template. For example, I believe that there is CSS code at the English Wikipedia that would let you hide the entire box for anyone who is autoconfirmed or extended-confirmed.
Not everyone receives a welcome message at this wiki, and I've long believed that we should assume that WP:Nobody reads the directions (this applies equally to talkheader and the welcome messages).
I think that the biggest barrier to removing it would be that we're all used to it now. Humans tend to believe One True™ Way for something to work is whatever it was on the day I originally learned it.
@Czar was asking about how to measure the use a while ago. I don't think you can measure how many people read it, but I understand that it's possible to track how many people click on a link in a template. See wikitech:Provenance for a description. After setting it up, you'd just replace the current links (e.g., ) with the full URL, including the tracking code, and wrap them in Template:Plain link so that it looks the same as always. Whatamidoing (WMF) (talk) 21:59, 11 January 2023 (UTC)Reply[reply]
What I wouldn't give to have a UX team at the WMF run some user interviews for us on banner blindness and clickthroughs on this template. I'd happily give up this real estate to the Growth team and have them figure out how best to use the real estate for new users because clearly we have no idea as a community if this template is the best we've got! In my opinion, clicking the talk page is a better indicator of curiosity and interest than clicking the edit tab. czar 03:40, 12 January 2023 (UTC)Reply[reply]
I'm not sure if it requires a dev to set it up, but once it's been created, I think we could add it to any page/template we wanted. Neil might know how this system works.
Volunteer-me has wished to be able to run it on some navboxes. Navboxes are the sort of thing that looks useful, but they're invisible on the mobile site (more than half our page views), and nobody's ever complained about their absence. So maybe they're not as useful as they look? Counting clickthroughs would help answer that question.
Tangentially, if you want to know about article-to-article clickthroughs, then take a look at This won't help with the talk page banner, because it only counts in the mainspace, but it can be interesting for any article that you're working on. Whatamidoing (WMF) (talk) 17:44, 12 January 2023 (UTC)Reply[reply]
@Whatamidoing (WMF) @Czar hi! I'm back from offsite and vacation now.
You don't actually need to do any set-up to use the provenance system. Just add ?wprov=mytag to the link, and it will automatically be captured in the webrequest data stream.
However, the other end of the process, querying the data, will be troublesome. It can only be done by someone with confidential data access. That someone also needs to write SQL and deal with querying one of our largest data sets.
So, practically speaking, you wouldn't be able to use the data you can easily generate unless you have the support of an engineering team which can direct an analyst to do the querying. (Or, theoretically, if you can convince someone with the appropriate acccess to do it as a volunteer.)
Sorry for the bad news! Neil Shah-Quinn (WMF) (talk) 21:34, 23 January 2023 (UTC)Reply[reply]

Idea for small topics and article series[edit]

Hi everyone. I'm currently working on the article series about Japanese figure skater Yuzuru Hanyu. The latest experiences lead to a new idea that might be fruitful for smaller topics and series with close interlinking (scope of ca. 3 to 6 articles). This concerns especially long biographies that had to be split into multiple sub-articles.

I noticed that the current navigation sidebar for our article series has the following issues:

  • It is a powerful tool if placed at the top of the sub-articles, but it gets lost in the mass of surrounding text on the main biography. This is not ideal, since the biography is the most visited page of the series and hence the most crucial one for linking and traffic.
  • The sidebar works properly in desktop mode but is not available in mobile view (similar issue with navigation boxes at the bottom of the page). To make mobile readers aware of the series as a whole, we can use the "See also" section, but that one is visually not very striking. Many readers didn't know that there is a sub-article about Hanyu's Olympic seasons, until it was featured on Wikipedia's main page this month.
  • While the current sidebar of the series is nice in optics, it covers a lot of space with little content, which is not economic.

I experimented with some alternatives and developed a simple navigation bar that could be placed at the top of the article:

Extended content
Yuzuru Hanyu article series


  • This navigation bar should be accessible in mobile view as well. I am not an expert for technical stuff and formatting, so there might be better solutions than this one, but it should give a rough idea how the template could look like.
  • It is much more compact than the current sidebar and covers less space in the article. It is also more visible and accessible, and fits the rather simple formatting of Wikipedia better than the colorful box.
  • Readers know from the beginning that the article is part of a series consisting of multiple, closely interrelated pages. For example, if readers want to look up the list of Hanyu's awards, they don't have to scroll down the full biography to find the link to the respective sub-article but can click on it directly at the top of the page.
  • A navigation bar like this may invite more readers to visit the sub-articles, and it would be a nice opportunity to display good and featured topics in the article mainspace. Personally, I didn't know that Wikipedia has FTs and GTs, until I noticed one through its talk page by coincidence. This is sad because they represent Wikipedia's best works and deserve recognition. A possible labelling could look like this:

There are many topics and article series, especially large ones with 15–20 articles or more, where the current sidebars and navigation boxes are more suitable. Those can definitely stay. However, for small series, a simple horizontal navigation bar like the two examples above, might be a better alternative, especially if it's accessible in mobile view.

I am really interested in your opinions and suggestions. Thank you very much in advance, and I wish you all a happy new year! Henni147 (talk) 14:13, 30 December 2022 (UTC)Reply[reply]

@Henni147:, maybe the Horizontal TOC can accomplish what you are asking for. I added
{{horizontal TOC|nonum=yes|align=center|limit=3}}
to Yuzuru Hanyu article for you to view. It is positioned at end of lead & before first section title. Regards, JoeNMLC (talk) 19:45, 3 January 2023 (UTC)Reply[reply]
@JoeNMLC: Thank you very much for your response. My idea does not really depend on the table of contents. The issue is that the current article series sidebar kind of gets lost between the surrounding text. It is placed below the very long infobox, and most readers don't notice it there at all. In my opinion, it is important that readers are aware that this biography article is part of a larger series that consists of multiple, closely interlinked articles that build a unit in content.
Update: I think, I have found a proper solution for our issue now. I embedded the article series as a module into the infobox (see this draft), which has the following advantages:
  1. It can be displayed in mobile view.
  2. The links are placed high enough in the biography article that readers don't have to scroll down much.
  3. It doesn't require the creation of a new template.
  4. Bonus: FA/GA icons can be nicely added.
On the sub-pages, the article series template can stay. Henni147 (talk) 08:28, 5 January 2023 (UTC)Reply[reply]

Flag translations whose original get deleted[edit]

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)Reply[reply]

@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)Reply[reply]
@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)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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[@[email protected]]b (talk) 03:25, 2 February 2023 (UTC)Reply[reply]

On the Attitude of Administrators and Seasoned Editors[edit]

Lately it has come to my attention the rude and unwelcoming manners of a certain number of administrators and seasoned editors, either in the teahouse or in the interactions on talk pages on different articles, even though I understand that disruptive/vandalizing behaviour is harmful for the project and must be addressed swiftly and that the administrator positions entails certain rights and prerogatives , i extend a cordial invitation on everyone here to adjust the attitude to reflect the expectations of welcoming, politeness and patience and even more keenly to those who have spent more time and effort on wikipedia to follow the guidelines and policies that they're so eager to enforce, lets continue building a better and more inclusive and welcoming project. Juanriveranava (talk) 21:10, 3 January 2023 (UTC)Reply[reply]

I presume that you are talking about Wikipedia:Teahouse#Extended confirmed pages. We should make vandals feel unwelcome, because they are unwelcome. This is a workplace to build an encyclopedia, not a social club. Phil Bridger (talk) 21:26, 3 January 2023 (UTC)Reply[reply]
not only, don't jump to conclusions, i'm contributing with an alternative pov that have been shared by others, the resistance to admit that this is becoming an issue is astounding, it seems to me that policies are enforced when suited to those who incur in said harsh behaviour, but those policies that are created to counter that rude manners and not. like the guidelines to become a teahouse host, don't bite newcomers, be bold and so forth and so are disregarded in a very selective and unfair manner. — Preceding unsigned comment added by Juanriveranava (talkcontribs) 22:39, 3 January 2023 (UTC)Reply[reply]

Now you have made me feel unwelcomed, I guess I'm a vandal in your consideration, but hey more power to you. I understand that policies are not laws, and my comment was directed towards the improvement of the project, now I understand how unwelcomed I am. have a nice day mate. Juanriveranava (talk) 22:39, 3 January 2023 (UTC)Reply[reply]

The other day, I saw a warning about vandalism on an editor's talk page. I clicked through to the edit, and found what's probably a bug in the software. I've asked one of the devs to look into it. And now what do I do? Let it pass, or go back to the patroller and say "Hey, you know, if you see a scrambled mess with multiple Special:Tags on it, you should be leaving a note at Wikipedia:Village pump (technical) about software bugs, instead of leaving escalating warnings on the editor's talk page"? I don't know that it would solve any problems (one reviewer, out of hundreds or more, would be more aware of one uncommon situation, out of hundreds or more), and it might result in the reviewer claiming editors should magically know that the software was going silently post something different from what showed on the screen at the time they tried to post their edit. If we approached editors with "Hey, that was a mess. Do you need help?", it'd be much more pleasant for everyone. Whatamidoing (WMF) (talk) 00:28, 4 January 2023 (UTC)Reply[reply]
Yeah, exactly my point, all this welcoming policies like, be bold and so forth seem on the light of this rude behaviour empty wording, not put into practice or put at the sole discretion of the administrators or "seasoned bigger than life editors" who bash on newcomers or novice editors, all this very discouraging for anyone trying to contribute.
Then when I make an honest conciliatory call, I got comments like the one [email protected]Phil Bridger in the fashion of, this is not a social club suck it up, and leave this to the grown ups.
to top it I just got a comment on my talk page about how unintelligible my writing is, and how low my english proficiency is, I was suggested to use google translation instead, first time in my life and interactions someone has told me that (not saying im a Shakespeare of the sorts but come on), I've read several comments on talk pages where this practice is not recommended, and in any professional situation using google translation copy paste is frowned upon. So contradictory messages, in the end it seems that one should not even try to contribute because if you do and make a mistake you can get a comment in the fashion of, REMIND THIS I'M AN ADMINISTRATOR AND I WILL BLOCK YOU, IS THAT CLEAR?
But as my mother says, sadly it is what it is, thank you for your contribution @Whatamidoing (WMF) Juanriveranava (talk) 03:10, 4 January 2023 (UTC)Reply[reply]
Yes, it's hard. It's especially hard to be generous to newcomers over the long term. At least for myself, I can become frustrated if I have explained this a hundred times. I have to carefully remember that the previous 100 explanations went to 100 other people, and it's all completely new to the 101st person. Whatamidoing (WMF) (talk) 01:45, 6 January 2023 (UTC)Reply[reply]
@Whatamidoing (WMF) while providing some advice to go to VPT with an issue sounds good, I suggest WMF staff or contractors refrain from suggesting that editors not make certain edits. If you were to suggest this in your capacity as a volunteer it would be more approriate. — xaosflux Talk 17:01, 4 January 2023 (UTC)Reply[reply]
I found the problem in my capacity as a volunteer, and I will be filing the Phab ticket in my capacity as a staff member, because it involved mw:Mobile visual editor.
But while we're on this subject, the idea that the "role" could matter more than the "whole person" does not seem to be shared by most of the world. There are discussions that I don't feel comfortable joining any longer, because someone may misunderstand my own personal opinion as an Official Declaration™. Whatamidoing (WMF) (talk) 01:40, 6 January 2023 (UTC)Reply[reply]
To me, at least, the distinction is clear. When you make an edit as WhatamIdoing you are editing as an individual person but when you edit as Whatamidoing (WMF) you are representing the WMF. Is that the way that you, and others, see it? Phil Bridger (talk) 18:42, 10 January 2023 (UTC)Reply[reply]
That's the way I see it, but the fact that I see it this way probably tells you more about my culture than about the One True™ Way of understanding the world and the people in it.
I've been fairly lucky with this division so far. There have been a few discussions that I've decided to stay away from out of fear that someone might misunderstand my personal opinion as the WMF's orders (most of y'all have known me for years, but there are new folks every day), and some years back, I had to correct one or two people who tried to claim that a WMF employee endorsed one side of a content dispute, but it's generally been okay. Other staff have been less lucky. One extreme example: A long-time editor went on an AWB run to correct a common punctuation error (in his volunteer account, of course), and found that social media was talking about how "the WMF" was editing articles about politicians and obviously endorsing m:the Wrong Version of an article about a controversial candidate because he fixed a comma without fixing everything else. Whatamidoing (WMF) (talk) 22:09, 11 January 2023 (UTC)Reply[reply]

Avoiding driving off experts from Wikipedia[edit]

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)Reply[reply]

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)Reply[reply]
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)Reply[reply]
I sometimes point to Wikipedia:Expert editors. Gråbergs Gråa Sång (talk) 11:41, 4 January 2023 (UTC)Reply[reply]
(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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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] and see how well it is doing these days. Donald Albury 01:48, 5 January 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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[@[email protected]]b (talk) 03:30, 2 February 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]

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)Reply[reply]

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)Reply[reply]

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)Reply[reply]

Expand notability criteria.[edit]

In the notes section of WP:NOTE note number 1 it states "... directories and databases, advertisements, announcements columns, and minor news stories are all examples of coverage that may not actually support notability when examined, despite their existence as reliable sources." I propose that we should add a another bullet point under "General notability guideline" section "Significant coverage" to say "coverage in databases alone, does not qualify as significant coverage for the purposes of establishing notability" to make it clear that presence in a database is not enough to establish notability. The problem to be solved is the idea that every database entry deserves its own article just because a database and nowhere but databases include a particular species/other. EvilxFish (talk) 05:14, 13 January 2023 (UTC)Reply[reply]

Makes sense to me. I'll note that coverage in database is enough to meet wikidata:Wikidata:Notability, which is more broader than Wikipedia's one. Piotr Konieczny aka Prokonsul Piotrus| reply here 05:51, 13 January 2023 (UTC)Reply[reply]
Makes sense to me too. Therefore, I expect there to be fervent opposition imminent... JoelleJay (talk) 06:39, 13 January 2023 (UTC)Reply[reply]
ok reading with abject horror the can of worms I am potentially opening, I would like to point out a few things:
1. I do not believe every database entry, by virtue of being in a database, be it an obscure species, chemical etc, should have an article.
2. If you do believe this, perhaps you would support the creation of a bot that makes millions of Wikipedia pages by scraping databases...
3. This does not mean lists of things are not notable even if this is adopted, because of WP:NLISTITEM, which means that even if individually species/chemicals/asteroids/etc are not notable a list of them may be notable.
4. This is not an issue of reliability of databases, only if they are alone enough to establish notability. If yes you may as well agree everything is notable because all one needs to do is find anything in a database and it becomes notable.
EvilxFish (talk) 07:03, 13 January 2023 (UTC)Reply[reply]
The plot thickens as I read more into things like WP:SPECIESOUTCOMES (which is not policy but is treated as one) I am convinced what we actually need is a new policy in WP:NOT called WP:NOTADATABASE or something similar. This would override the unofficial rules various wikiprojects have employed. EvilxFish (talk) 11:09, 13 January 2023 (UTC)Reply[reply]
We already have WP:NOTDATABASE. – Joe (talk) 17:18, 13 January 2023 (UTC)Reply[reply]
I believe it needs clarification then, because it is clearly being disregarded in favor of large volumes of articles which for example invoke (the none policy of) WP:SPECIES. EvilxFish (talk) 02:04, 14 January 2023 (UTC)Reply[reply]
  • It does not make any sense to me, not because it's wrong, but because it's obvious and is likely to confuse the readers. Has anyone ever argued that being listed in a database qualifies as significant coverage? And even if there could probably be discussions in which some users advocate the notability of a subject because of its inclusion in a specific database, they for sure are invoking a different argument/reason of notability, certainly not that "listing in a database = significant coverage". A mere database listing is a BLATANT example of trivial mention, and then it does not require any additional sentence in our guideline. Otherwise it's like saying "Meat is prohibited, only vegetables are allowed. But be careful, steak is not allowed!", to which I would react wondering if other meat exceptions are instead allowed. --Cavarrone 11:32, 13 January 2023 (UTC)Reply[reply]
    @Cavarrone:Although it may seem obvious it is not, consider for example this deletion discussion here in which it is basically argued and it is the precedent that anything part of a database in relation to a species was a paper, and inclusion in a database is enough for notability for species specifically. I then made the same argument about having a bot above and they said this was actually something put into practice. It appears that although we have a notability criteria for most things, certain projects choose to ignore them and instead substitute their own, where even a trivial mention appears to be enough for them. Therefore this clarification becomes necessary to iron out Wikipedia's position. EvilxFish (talk) 13:05, 13 January 2023 (UTC)Reply[reply]
    As far as I can see there the argument that the subject passes GNG because of databates is not explicitly stated, more simply it is claimed that the subject is inherently notable because of their inclusion on certain databates, which is a different argument we can agree or not, but has not really to do with the notion of "significant coverage". Notability is not only based on the GNG guideline, people can support the notability of a subject through different guidelines and arguments, up to the point of ignoring entirely rules and guidelines, obviously the closer will weight valid and less valid arguments and consensus, but adding your proposed addendum (which sounds to me like "and beware, a trivial mention is not significant coverage!!!") will not prevent anyone from using the same exact argument you are trying to invalidate. --Cavarrone 15:55, 13 January 2023 (UTC)Reply[reply]
    The general irony of course of the ignore all rules policy is rules are definitely not ignored in AfD discussions for example, for most classes of article. What would you suggest that helps? The problem ideally I want to avoid, is mass creation of articles by bots to turn Wikipedia from a encyclopedia that covers topics that have received significant coverage in news/review articles/books/etc, to naught more than a collection of stubs derived from databases of species/chemicals/proteins etc. EvilxFish (talk) 02:17, 14 January 2023 (UTC)Reply[reply]
  • hmmm… what about the case of a building being listed on the National Register of Historic Places, and similar landmark preservation databases? Not all databases are the same… and a few are prestigious enough that I think they do impart notability. Blueboar (talk) 12:41, 13 January 2023 (UTC)Reply[reply]
You cold consider it as a reason/hint/sign of notability, or maybe not, but certainly it does not qualify as "significant coverage" (and it is already like this, we don't need any addendum or guideline modification to state this). Cavarrone 12:58, 13 January 2023 (UTC)Reply[reply]
Just to be clear, for most things it is like this but not everything, as noted in my reply above EvilxFish (talk) 13:18, 13 January 2023 (UTC)Reply[reply]
There are about 20 new places, on average, added to the National Register of Historic Places every week. I find it hard to accept that a given historic place is "notable" in the Wikipedia sense without reliable sources (beyond the NRHP listing) providing significant coverage of that place. - Donald Albury 15:23, 13 January 2023 (UTC)Reply[reply]
@Donald Albury seriously?! Why on earth is it considered an automatic GEO pass with such a ridiculously low inclusion threshold? Are the blurbs on each one really that significant and independent (I would guess they are absolutely not from a quick skim)? If my podunk 18k-person town can have 11 listings there is no way NRHP entries are distinguishing sites of encyclopedic importance from those of hyperlocal mild interest. This seems like an enormous contributor to systemic bias... JoelleJay (talk) 07:12, 15 January 2023 (UTC)Reply[reply]
@JoelleJay: Remember that the Geoland assumption of automatic notability is for populated places. That does not apply to almost all of the places on the NRHP. Now, I have expanded several articles about places on the NRHP, sometimes with little more than the nominating document available. I do think relying only on the nominating document for an article is, at best, marginal, but I suspect any attempt to delete an article about a place on the NRHP that has no independent reliable sources cited would fail at AfD. There are also articles in WP about places that are on local lists of historic buildings, which may be less notable that places on the NRHP. I don't like the many stubs about places on the NRHP that lack indpendent reliable sources, or even a reference to the nominating document, but I am not willing to take on that fight at this time. Donald Albury 14:37, 15 January 2023 (UTC)Reply[reply]
@Blueboar: I would have to reject that premise, as all NRHP listings are accompanied by a separate document nominating the property for listing in the register. The nominating document is not a database, and is its own source separate from the database itself (obviously, because the document still exists for a property that is nominated but rejected for listing). BD2412 T 19:46, 14 January 2023 (UTC)Reply[reply]
It is not always obvious where that nominating document is to be found. Even when I can find the nominating document, the resulting WP article may be skimpy if I don't find other reliable sources. See Warm Mineral Springs Building Complex. Donald Albury 20:38, 14 January 2023 (UTC)Reply[reply]
The nomination form I'm seeing appears to be a single page of administrative items and some box checking -- it contains orders of magnitude less information than literally any job application does and obviously would not count towards GNG as it is primary, not independent, and contains zero SIGCOV. JoelleJay (talk) 07:21, 15 January 2023 (UTC)Reply[reply]
The "Historical Statement of Significance", such as this, can be useful, although it is not independent. However, I have only found such statements when a local government posts them, as the City of North Port did in this case. I am not aware of how to find those statements for other places on the NRHP. Donald Albury 15:05, 15 January 2023 (UTC)Reply[reply]
I support this proposal, although I think it might be better to call it a "clarification" of the notability criteria, rather than an "expansion". - Donald Albury 15:23, 13 January 2023 (UTC) Edited 15:26, 13 January 2023 (UTC)Reply[reply]
  • Notability's primary use case is to keep people from using Wikipedia as a promotional vehicle for their whatever; company, organization, product, etc. How does the proposal above help with that? --Jayron32 15:31, 13 January 2023 (UTC)Reply[reply]
    It states "On Wikipedia, notability is a test used by editors to decide whether a given topic warrants its own article." on WP:NOTE first line. I think the key is what is wikipedia? Is it a database of every species, chemical, protein etc or is it an encyclopedia that discusses a wide range of notable topics. I'd argue the latter. EvilxFish (talk) 01:47, 14 January 2023 (UTC)Reply[reply]
    I know what WP:N states; but the main use of it is to keep spam away. For things which lack that element, it's a fairly low priority to worry about notability. I'm not really much concerned with a short article about some species of roundworm, as I am about some biography of an "up and coming singer" or about some company trying to increase its SEO results. --Jayron32 02:24, 17 January 2023 (UTC)Reply[reply]
    I'd agree the other example described are worse, but I still see this as spam. There are literally millions of known species with more to be discovered. If we maintain the status quo one can expect Wikipedia to become mostly a collection of stubs from scraped species databases rather than an encyclopedia. EvilxFish (talk) 04:07, 17 January 2023 (UTC)Reply[reply]
    Except, since no one stands to gain financially nor improve their reputation nor increase their visibility by abusing Wikipedia, these sorts of things aren't really a problem. --Jayron32 12:04, 17 January 2023 (UTC)Reply[reply]
    Then why do we have policy stating wikipedia is not a database or directory? How can we expect editors to watch and maintain millions of separate articles on low-information topics no one is interested in -- wouldn't that make it easier for vandals to insert promo/spam? And anyway people absolutely do benefit financially from creating articles on individual proteins or chemical reactions or whatever: refbombing with (their own reviews of) primary papers they wrote, introducing self-serving eponyms, putting their names in the body of the article ("research by Dr. X et al revealed..."); having a wiki article on a discovery you made is huge for career progression (which is exactly why a colleague of mine created an article on a topic our lab published--he even went through AfC, where reviewers failed to catch that all his sources were primary and non-independent). Do you know what venture capitalists like to see when investing in a startup? Evidence the scientific basis of a new technology is reliable and accepted. Pharma articles are already notorious for attracting COI editors; how much more difficult would it be to monitor UPE contributions scattered across millions of pages as opposed to a few thousand? JoelleJay (talk) 19:08, 17 January 2023 (UTC)Reply[reply]
    It is an incorrect assumption that the use of "scraped species databases" are the usual sources for Wikipedia's species articles. Instead, most information comes from manually curated databases that are created by editorial teams of professional taxonomists representing some of the world's leading authorities on the subject. Even when not stated explicitly, these database entries include assessments on the validity of the species descriptions while also taking into account previous reviews and diverging taxonomic opinions (example here). These databases can consist of secondary or even tertiary source material that would be a gold standard to use in any Wikipedia article.
    Also, note that a comprehensive collection of (stub) articles is a definition of an encyclopedia, something we are all striving to produce. As stubs are the foundational building blocks of articles, they should not be considered a problem simply because they are stubs per se, as their addition supports the role of the project. To now describe these efforts by so many Wikipedia editors as "spam" is an inaccurate use of the term. Loopy30 (talk) 13:57, 17 January 2023 (UTC)Reply[reply]
    In fact, Wikipedia is a database. A database of text, which is displayed to you as articles, but underneath that user interface it's a database. See for instance MW:Manual:Database access for descriptions of how to access it at the database level rather than at the user interface level. —David Eppstein (talk) 01:55, 14 January 2023 (UTC)Reply[reply]
    I added a comment at the end, because I agree this obviously needs further clarification and not all databases are equal. EvilxFish (talk) 02:10, 14 January 2023 (UTC)Reply[reply]
  • I'll point out that as an attempt to get your hook into a few tens of thousands of taxon articles, as indicated above, this is going to fail, because a taxon's presence in (multiple) databases is not considered an indication of notability, but of validity (i.e., the classification has been accepted by the field). The presumption of notability rests on an unrelated consensus based on the importance of taxonomic units in the life sciences. - Which, I admit, I probably presented in a misleading manner here by leaving out the middle man. --Elmidae (talk · contribs) 16:27, 13 January 2023 (UTC)Reply[reply]
    I will not dispute the validity of databases as, they are a valid source, as is the paper from which they are derived, but as you rightfully point out, it is not an indication of notability. If we accept that taxonomic units are important and should be presented somewhere, would you agree, due to the nature of the coverage and the general notability criteria for every other article on Wikipedia that is used to determine inclusion or exclusion, this data is best presented in a project such as Wikidata instead? EvilxFish (talk) 01:55, 14 January 2023 (UTC)Reply[reply]
  • There are a few things wrong with this proposal. One is that it assume that no databases contain significant coverage. Most don't, but some, such as the NRHP, do for some subjects. Another is that it assumes that the general notability guideline is the only possible route to notability. It is not. Notability guidelines are guidelines as to whether it is possible to write a policy-compliant article on a subject. If that has already been done then there is no need to invoke any guideline. Phil Bridger (talk) 17:07, 13 January 2023 (UTC)Reply[reply]
    • I agree completely with this comment. BD2412 T 19:47, 14 January 2023 (UTC)Reply[reply]
  • A very similar proposal was discussed just a few months ago and failed to get consensus in a sitewide RfC. As discussed there, databases are like any other type of source: some contain significant coverage, some don't. A blanket rule like this seems like a non-starter to me. – Joe (talk) 17:15, 13 January 2023 (UTC)Reply[reply]
  • I'll repeat a comment I already left at Wikipedia talk:What Wikipedia is not: There is no valid distinction between information in databases and information in other kinds of sources. Being in a database is a matter of data storage and access technology, not of what type or depth of information is being stored. For instance, all Wikipedia articles are stored in a database; that's how the Wikipedia servers keep track of article content. Probably most newspaper articles online these days are stored in a database. It would be stupid to say "articles sourced from nothing but online newspaper articles are not notable", but that's what this proposal would amount to, because those newspaper articles are database entries. —David Eppstein (talk) 23:38, 13 January 2023 (UTC)Reply[reply]
  • Agree with the above comment, some databases have significant content, others do not so a catch as catch can approach is not helpful imv Atlantic306 (talk) 00:21, 14 January 2023 (UTC)Reply[reply]
  • based on the above comments, would a further clarification such as "coverage in a database that describes the subject in minimal detail, usually derived from a single source." That way a database say of historical sights, that is constructed from multiple sources would constitute a good source, but a database of chemical species created by literature scraping from the papers that first synthesized them and report properties such as melting points, would not be a good source for determining notability? I agree based on the above comments one cannot tar all databases with the same brush and, well, this is why we are discussing this to come up with a sensible proposal. EvilxFish (talk) 02:00, 14 January 2023 (UTC)Reply[reply]
    The requirement of depth of coverage is already in GNG, for a very long time. So if you are talking about amending GNG to add that requirement, you need to adjust your time machine. If you are talking about somehow imposing a new rule that every SNG must be subservient to GNG, so that the depth of coverage rule from GNG is imposed on them too, then your proposal goes well beyond being a minor change to GNG and is I think a total nonstarter. If you are demanding that past AfD participants change all their past opinions so that the record of typical past AfD opinions in OUTCOMES can change to reflect your new rule, then you need a much stronger time machine. —David Eppstein (talk) 03:00, 14 January 2023 (UTC)Reply[reply]
    Please refrain from being condescending as per WP:CIVIL. I suppose I am arguing that some subject specific guidelines need to be brought closer to the general ones, on the basis that those subject guidelines yield what I have argued are undesirable outcomes which I have mentioned. And working with you and the rest of the community I would like to work towards deriving such a proposal that achieves an appropriate level of compromise. I would also point out that WP:OUTCOMES are not rules, just general outcomes and because something happened in the past that does not mean it is invariable correct. Finally I would point out WP:5P5, which makes it clear that just because something was accepted policy in the past does not mean it should be policy indefinitely into the future and discussions like this should focus on the merits and flaws of the proposed changes and avoid status quo bias. EvilxFish (talk) 03:22, 14 January 2023 (UTC)Reply[reply]
    I am adamantly opposed to spreading the infection of hype-based notability (that is, GNG) to subject notability guidelines that are based on accomplishments rather than hype (example: NPOL, where you can become notable by winning a major office but not by pouring lots of money into publicizing a losing campaign) and are working perfectly well as is. —David Eppstein (talk) 05:53, 14 January 2023 (UTC)Reply[reply]
    This is why I had the "prose coverage by a human" requirement in the RfC proposal linked somewhere above; like you, I do not think physical/chemical/etc. parameters (e.g. melting point) scraped from primary literature with no secondary analysis can ever be considered SIGCOV. Somehow this take was controversial... JoelleJay (talk) 21:53, 14 January 2023 (UTC)Reply[reply]
  • Databases, per se, are not a problem. As others here, and before have said, 1) a database is just a way of storing and accessing records and 2) the quality and reliability of a database varies. This community has a process for evaluating whether a source is reliable and an entire process to gather community input on the reliability of a source. --Enos733 (talk) 05:29, 14 January 2023 (UTC)Reply[reply]
    @Enos733: I agree with this statement and it is the reason why I attempted to make an amendment in the previous bullet point. The issue I would like to resolve is that a mention in a database alone (with the primary source it is derived from) is grounds for notability in some subject areas. The reason I feel strongly about this, is if this is accepted policy we see bots that create a vast number of stub articles for certain areas, which for most other subjects would fail notability criteria. I then argue if for example all species are worth being documented, the more appropriate project to do so would be wikidata and policy should change to reflect that. EvilxFish (talk) 06:05, 14 January 2023 (UTC)Reply[reply]
    This is where we differ. I have no issues with stub articles. I think that a stub can be helpful to many viewers, who just want some basic information about a subject. The problem with bots is different, but again, changing how some databases are used is a blunt instrument that does not solve your concerns. - Enos733 (talk) 07:49, 14 January 2023 (UTC)Reply[reply]
    I would argue it does serve to address them. Currently it is accepted that if even a trivial mention of a topic in a database is present (for some subject areas) therefore that topic is notable, despite WP:GNG. This facilitated the use of bots to generate tens of thousands of stub articles on topics which fail WP:GNG but because they are noted in a database which, in the case of species, lists basic attributes like size family etc, they are treated as notable. This is not to argue that stubs are a bad thing when they are created for topics with a range of coverage that satisfy WP:GNG only the existence of 10's of thousands of them for insignificant topics on the basis that if they are in a database => they are notable. Perhaps I am attacking this from the wrong angle though and I instead need to focus on addressing commonly accepted practices (not policies) such as WP:SPECIES. What do you think? EvilxFish (talk) 08:10, 14 January 2023 (UTC)Reply[reply]
I think if you wanted overturn WP:SPECIESOUTCOMES, it is probably best litigated as its own issue. I don't feel that it would be overturned by the suggestion you are offering. NeverRainsButPours (talk) 10:56, 14 January 2023 (UTC)Reply[reply]
  • I'd like to thank everyone for the comments and feedback and helping me to think this through a bit better, I think I agree with @NeverRainsButPours: that my focus should be on changing policies such as WP:SPECIESOUTCOMES and maybe one or two similar ones. I will work out the best venue to do that. If it's here I am not sure if it is best to start a new discussion as it will be substantially different from what this started with. What do you guys think? EvilxFish (talk) 11:08, 14 January 2023 (UTC)Reply[reply]
    I think it would be best to start by engaging with the relevant WikiProjects (e.g., Wikipedia:WikiProject Tree of Life and some of the top-level biology projects if you're proposing to change SPECIESOUTCOMES). That said, I think you would do well to pause for a while and do some subject-specific background research before plunging ahead with your next proposal. (e.g., the fact that all valid species must have a description or diagnosis is significant in explaining why SPECIESOUTCOMES is true). It might save you some trouble. Choess (talk) 17:31, 14 January 2023 (UTC)Reply[reply]
    thank you, I will bear that in mind and try look for historical discussions pertaining to this. EvilxFish (talk) 02:26, 15 January 2023 (UTC)Reply[reply]
  • Sorry, but this makes no sense to me at all. Databases can be good or bad, primary or secondary, superficial or in-depth, etc. This proposal would eliminate a wide array of perfectly acceptable sources for no good reason. Honestly, this looks like what happens when a rule-of-thumb approach to the intrinsically difficult question "What deserves its own page in an encyclopedia?" gets taken as the ultimate answer. My advice is to spend a month not speaking in all-caps shortcuts, discover how strange they sound, and contemplate the headline "Wikipedia deletes thousands of endangered species: Not notable enough, say editors". XOR'easter (talk) 14:17, 14 January 2023 (UTC)Reply[reply]

Write an essay?[edit]

Based on the above comments, it seems clear that not all databases are the same. Some do establish notability while others do not. Further guidance on all of this would be helpful. So… I would suggest that we start by crafting an essay to outline it. Blueboar (talk) 15:24, 14 January 2023 (UTC)Reply[reply]

I don't see the need for any further clarification myself. As already said in the general notability guideline, for a topic to pass that particular notability guideline coverage in reliable sources is required to be significant. It makes no difference whether the coverage is stored in a database or not, just as it makes no difference what type of paper it may be written on. Phil Bridger (talk) 16:27, 14 January 2023 (UTC)Reply[reply]
I don't think there's a need for a general essay about databases and notability. Per others, whether or not you can call a source a "database" doesn't really shed any further light on whether the individual topics in it are notable. Attempting to subdivide and classify databases for purposes of notability will just wind up restating GNG, and because the restatement won't perfectly parallel GNG, will encourage more vexatious litigation. If there are issues with a particular database or set of databases, that should be addressed in appropriate SNGs; major aspects of notability, such as SIGCOV and RS, require subject-specific context to interpret, which is why it's important to have SNGs (rather than attempting pure, abstract reasoning from the GNG alone). Choess (talk) 16:42, 14 January 2023 (UTC)Reply[reply]
Attempting to subdivide and classify databases for purposes of notability will just wind up restating GNG -- the issue is that some people actually do interpret physical parameters scraped from primary studies as "SIGCOV" when they're collected in a single database (which of course would mean the types of entries that happen to have lots of recorded physical parameters, like all the billions of objects in some astro databases, automatically "have more SIGCOV"). A simple reminder at NOTDATABASE that SIGCOV requires direct, secondary prose analysis by a human would clarify things. JoelleJay (talk) 22:07, 14 January 2023 (UTC)Reply[reply]
I think what I need to do is as stated above, address specific policies such as WP:SPECIESOUTCOMES. Editing general policies will not solve the issue I am attempting to tackle. EvilxFish (talk) 17:08, 14 January 2023 (UTC)Reply[reply]


I have reconsidered this proposal based on the above comments and agree with the assessment provided by multiple users as to why this is a bad idea and will not achieve what I would like. I will give this some more though and focus instead on things such as WP:SPECIESOUTCOMES more specifically. I think for the sake of clarity I will start a new section for this once I have formulated the idea somewhat. EvilxFish (talk) 01:57, 15 January 2023 (UTC)Reply[reply]

@EvilxFish, I don't know how familiar you are with the concept of a Valid name (zoology), but if you want that essay to be useful, then you really do need to understand it. I'm going to use your nom at Wikipedia:Articles for deletion/Harpalus numidicus as an example. At the time of your nom, it contained one ref, to a WP:DEADURL. Most editors don't know much about species, and a quick search on your favorite web search engine isn't likely to be productive. This isn't exactly an easy task.
But: We have to avoid the mistake of assuming that the citations currently present in the article are the only sources in the world, and in this subject area, we have one indisputable fact: it is impossible to have a Valid name (zoology) for any animal or a Valid name (botany) for any plant unless and until someone has published a description in an acceptable scientific work, that is comprehensive enough and detailed enough that other researchers can reliably distinguish the species you've found from very similar species.
So, while you are looking at a confusing mess on the search page, and a dead link in the article, people who happen to be familiar with the subject area are saying: The database says that this has a valid name, and therefore I can guarantee that there is a detailed description of the species somewhere. In this case, someone has recently very kindly added a ref to the original three-page-long scientific journal article to the article. It's understandable that you couldn't find it in a quick WP:BEFORE, and it's even understandable that you wouldn't already know that such papers (or books) exist for every single validly named plant and animal species in the entire world. This IMO is what editors need to be told in the essay you're writing: even if it looks like the article is just repeating a paltry database entry, if the subject of the article is a validly named plant or animal, we can guarantee that WP:NEXIST is fulfilled. You just have to have the subject-specific knowledge about how to find the sources. This is easy in some cases (e.g., Fishbase lists them) and harder in others, but valid name == paper published. WhatamIdoing (talk) 20:44, 17 January 2023 (UTC)Reply[reply]
But does it mean multiple secondary, independent papers with SIGCOV have been published? JoelleJay (talk) 03:44, 18 January 2023 (UTC)Reply[reply]
For the overwhelming majority of them? Yes. I work in the Lepidoptera area, which is one of the largest amassments of species (both in regards to valid names and in regards to species that have a Wikipedia article). Are the articles generally in a terrible state? Well, yes. Between the multiple iterations of species-creation-bots and some other since-indeffed problematic article creators, we've got a massive backlog of articles that need to be whipped into shape and only a small number of editors willing to work on them. Are they generally under-sourced? A-yup. But do those sources exist? Almost without exception.
It's actually finding and accessing them that's the problem, and for somewhat obscure species, Google nor Google Scholar are all that helpful: a lot of the information is in paywalled articles, in non-searchable scans of old articles, or in not-yet-digitized articles and books. Lots of the relevant coverage is in old, dead trees.
Throw in species with a bunch of genus transferrals or junior synonyms in their history and a good portion of the relevant scientific works will be found under those names, which you need to actually be aware of to even look for those in the first place (to then promptly run into the same Google-ain't-helpful issue as above, and then once you finally find a place a particular book or paper has been scanned, you're presented with the joy of, say, working with a blurry scan of early 19th century German riddled with scientific jargon and abbreviations, in an old, barely legible type with ink that has started bleeding a bit and paper has gone well past turning yellow.)
Are there times where I can't find multiple such papers? Sure. The more obscure a species, the older its original publishing date, the smaller its range, and the further it occurs from the Anglosphere, the harder it becomes. (A good part of that is because digitizing efforts so far overwhelmingly focus on material from Europe and the Anglosphere) In almost all of those cases, though, it's a safe assumption that if the species is still considered valid, such sources do indeed exist.
The issue is figuring out "okay, is this still considered an actual, valid species", and for that you don't need SIGCOV. Reliable, up-to-date databases listing the species as valid, or recent scientific works including the species in a list of valid species are enough for that. AddWittyNameHere 20:14, 18 January 2023 (UTC)Reply[reply]
@AddWittyNameHere: From what I have understood so far is the databases are constructed based on a report from the primary literature that initially described the species. I'm curious what secondary sources will usually exist for these species and how does one go about finding them given your above mentioned difficulties in searching using things like google? Is it things like textbooks that reference the original paper when discussing the species, maybe in the context of other similar ones (good secondary source coverage)?
With regards to don't need SIGCOV, isn't this equating validity and notability? There are many things which are valid but don't have articles because they are not notable based on either the general criteria or specific criteria of particular rule subsets. If the argument is that any valid species is also notable, why should this be the case? Thank you for helping me understand this area better! EvilxFish (talk) 01:44, 19 January 2023 (UTC)Reply[reply]
Apologies up ahead: this will be one long response. I've summarized and simplified where possible, but the questions you're asking mostly aren't ones that have short, simple answers.
Regarding the databases, kind of but not quite. To majorly simplify it (let's just say that there's a reason physical copies of the International Code of Zoological Nomenclature take the shape of a book and not a flyer), a name can end up being not valid for a variety of reasons and in a variety of ways, and reliable, up-to-date databases update their information once that happens. So, an initial description is a necessary part of it, but not the whole of it.
Onto secondary sources (or rather, secondary coverage, because a lot of taxonomic literature is a mix of primary in some regards, secondary in others.)
Possible sources of significant secondary coverage include but are certainly not limited to:
Monographs and other taxonomic revisions; identification guides; new records/data/additional description (e.g. distribution, host plants, seasonality, parasitoids--yes, the new data is primary, but these frequently include a summarization or overview of information from previous literature, which is not); articles on closely related taxa; relevant checklists, overviews and standard works; reviews of specimen collections hosted by museums; factsheets and scientific literature outside the taxonomic corner for agriculturally or medically significant species as well as invasive exotics and deliberately-introduced species; works dealing with biodiversity, climate change and/or protection of endangered species; works on species that are food source and/or host to the species in question, or to which the species in question is food source and/or host.
Finding them: first step is identifying a source. Databases, reference lists in related articles and books, tracing back taxonomic changes, looking for works published while the species was known by a different name if you know of any such names, figuring out the title of the work in which the species was first described and looking for citations thereof, monographs and other taxonomic revisions of relevant higher taxa, checking papers of the most relevant authorities in that particular taxonomic corner to see if there's anything relevant, check if checklists for the region in which the species occurs mention any refs, keyword searches in journal archives and Wikipedia Library, and so on. If you own actual subject-relevant paper-and-ink resources, check those.
Then once you know the name of a work in which the species is, or might be, covered, the next step is actually accessing that work. If copyright-expired, check archives like the Biodiversity Heritage Library to see if it's been digitized. If not that old, check Wikipedia Library, ResearchGate, journal's site for archives, Google Scholar (sometimes it doesn't find things on keyword which it does find on author and title) and Google Books, author's site because some upload scans or pdfs, university site because same, check for access through whatever institutes you have access through, check physical university library if you have access to one, if you've got a partial copy, search for some random sentences for it, sometimes Google finds them then after all, try some different search engines, ask on Wikipedia's Resource Request, ask on subject-relevant communities whether someone has it, possibly even e-mail the author to try and see if they'd be willing to provide you with a copy.
Validity is not, by itself, notability, that is true. Plenty of things are valid in the sense that they exist, yet are not notable by Wikipedia standards, because there isn't (enough) actual coverage in appropriate sources. But validity in the specific context of zoology has a meaning well beyond merely conveying "this exists". It means "this taxon exists; it is not the same as another taxon in use; it has been described in a way sufficient to distinguish it from other taxa in use; and this is the correct name to use for it." When all of those criteria are fulfilled—which they have to be in order for something to be an actual valid species—that is a near-guarantee that significant-enough coverage does exist, a high chance that one or more editors familiar with the subject area are able to find and access some of it or at least be able to point out where such sources exist, and for sufficiently obscure species, a pretty low chance for editors without that familiarity to do the same.
Which means that, for editors without a familiarity with the subject, it's going to be just about impossible to tell apart the identical-looking cases of "obscure species article has only a database as source in spite of existing SIGCOV" and "obscure species article has only a database as source because no SIGCOV actually exists", while the former is quite common and the latter is not. As the notability of an article's subject is not determined based on the presence of sources in an article, but on the existence of such sources (WP:NEXIST), using validity of the species as the benchmark to decide the likelihood of notability-conveying sources existing therefore makes sense: it is generally fairly easy to prove or disprove validity (and can thus be checked by people with significantly less familiarity with the subject than checking for the existence of sources would require), it genuinely does have a very close connection with the likelihood of there being coverage, and it avoids AFD being flooded with species articles which are then a. kept upon someone with familiarity with the subject digging up some sources, b. kept upon someone with familiarity with the subject explaining that the validity of the species means such sources almost invariably exist or c. deleted only to be recreated down the line as soon as someone with said familiarity has the time to dig up said sources. AddWittyNameHere 07:43, 19 January 2023 (UTC)Reply[reply]
I really appreciate your detailed and thoughtful responses! I am approaching this from A) a background in molecular biology where Wikipedia's pages on individual proteins are often similarly sourced to databases but secondary SIGCOV cannot be presumed; B) familiarity with the NASTRO guideline where even though all objects in certain databases correspond to numerous descriptions, measurements, and other parameters that are often verified/repeated across multiple studies, appearance in those databases is explicitly not allowed as evidence of notability precisely because it would result in millions of articles; and C) our recent tightening of athlete notability, where for criteria to even have a presumption of SIGCOV there must be at least one GNG source present in the article, and for new criteria proposals there must be evidence-based consensus that 90-95% of all subjects meeting the criterion also meet GNG. So to me it is rather incongruous to have on the one hand the requirement by the notability guideline that (presumption of) notability must be achieved through either the GNG or the dozenish SNGs listed at N (the only places where NEXIST can be automatically generalized to all qualifying topics), and on the other hand have tens of millions of subjects that can bypass those expectations based on a non-guideline that AFAIK has not been examined on a community-wide scale. I also strongly prescribe to the WP:NOPAGE perspective; my feeling is that if all the DUE secondary SIGCOV on something can be summarized in an infobox and/or within another article, we do not need a standalone on it.
That said, I am glad there are editors like you who actually put in the effort to source species articles appropriately and have the patience to explain how/why they can be expected to comply with N. The issue (and I've made exactly this same argument regarding athlete bios) is with all the other editors/bots creating thousands of microstubs that serve as essentially worse versions of their originating dedicated databases and appear at the top of Google search results despite being unmaintained and barely monitored. This just isn't sustainable. JoelleJay (talk) 22:40, 19 January 2023 (UTC)Reply[reply]
You're welcome. :) For discussions on a community-wide scale, the subject pops up every few years in places like this and some other community-wide noticeboards. So far, the outcome has always been roughly a. "if it's verifiably a valid species, it's notable" or b. "no consensus to change (therefore defaulting to a)". How many of these discussions got enough attention to truly consider them community-wide, as opposed to simply appearing in a community-wide rather than WikiProject-specific location, I don't quite know from top of my head, however.
I also strongly prescribe to the WP:NOPAGE perspective; my feeling is that if all the DUE secondary SIGCOV on something can be summarized in an infobox and/or within another article, we do not need a standalone on it.
Sure, I can agree with that in theory. Thing is, for species that generally isn't an actual option once all due secondary SIGCOV is actually taken into account, rather than simply what is currently in the Wikipedia article.
But even for species where such information could genuinely be summarized in another article—its genus. That might work for small genera. Say, ten or so species at most.
But many, many genera don't have ten species. They have sixty. Or three hundred. Or well over a thousand. Articles on such genera simply listing all existing valid species in it already become fairly unwieldy (see e.g. Eupithecia, one of those massive 1000+ species genera which I've recently been working on).
Now imagine that list, but with author and year of description after each name, plus some amount of distribution data ranging from type location to full known distribution. Add the synonyms with their authors and years—and many of them will have a synonym or two, and some have dozens. Some description. Then depending on what information is known, vernacular names, seasonality, whether it overwinters as egg or larva or pupa or imago, whether it's got multiple forms, whether it's been assessed as vulnerable or endangered, any medical or environmental or agricultural significance, and so on. That's all DUE information when it comes to the species. A good bit of it is hardly DUE on the genus article, though. So you'd either end up with a bunch of genus articles that are an unwieldy amassment of bits of species-specific knowledge that aren't necessarily DUE for the genus, or you'd end up providing even less information than the species articles do.
Not to mention that all those species articles? The pages would still have to exist, just as redirect to the genus page. All those current redirects to the species articles? They'd also still have to exist, just re-pointed to the genus page in order to avoid double-redirs (and then you need to either create a massive amount of anchors in the genus article so people are led to the relevant bit of information). Some of the categories? They'd still need to be in place, but on the redirs because they apply to the species not the genus. So the likely net result would imo be "equal number of mainspace pages even if not articles and replacing one area of undersourcing and possible incongruence with another, while not actually improving navigation, readability, or information provided, and potentially reducing all three of them". While introducing a whole heap of work to change everything around, which would mean even less time and effort spent on actually improving the sourcing and information provided on species as long as it's ongoing (read:several years). AddWittyNameHere 18:11, 20 January 2023 (UTC)Reply[reply]
@AddWittyNameHere: I too would like to thank you for this very detailed response and I am starting to get a clearer picture of how ideas such as WP:SPECIESOUTCOMES came about, I will definitely be including this in the essay I am planning on writing maybe this weekend about the topic. I now have doubts about whether anything should/needs to change, but I feel as if writing all this down will help others at least understand the defacto policy. EvilxFish (talk) 07:58, 20 January 2023 (UTC)Reply[reply]
You're quite welcome! AddWittyNameHere 18:14, 20 January 2023 (UTC)Reply[reply]
With regards to don't need SIGCOV: As a side note, no subject actually requires a citation to a notability-proving source in the article. If you blank all the sources from, (e.g.,) Cancer or Queen Victoria, those subjects still 100% meet GNG, including the SIGCOV component. It's convenient if you add sources that make it easy for other, less-informed editors to figure out that the notability criteria have been met, but it's the existence of the sources in the real world, not the citation of them in the article, that actually matters. WhatamIdoing (talk) 18:50, 23 January 2023 (UTC)Reply[reply]
Thank you for raising that point and when I come to write this up eventually, I will be sure to take it into consideration and make a mention of this important point! Just through this discussion here I have learned a lot more about the area, thank you! EvilxFish (talk) 01:34, 19 January 2023 (UTC)Reply[reply]
@EvilxFish:, there's another point regarding SPECIESOUTCOMES that hasn't really been addressed yet in this thread. The number of species names ever published is several times the number regarded as valid. Most species names are synonyms. Nobody is mass-creating articles for synonyms from databases (some people do create a lot of redirects for synonyms); however sometimes people create an article using outdated sources for a name that is considered a synonym. One of the reasons species articles don't get deleted, is that if an article for a synonym is created, it can just be redirected to an article for a valid name. Plantdrew (talk) 22:31, 20 January 2023 (UTC)Reply[reply]

Making a "free" ChatGPT for use on Wikipedia[edit]

I've been playing around with ChatGPT for some time by experimenting and seeing how can a similar model be used on Wikipedia, and I'd had lots of fun doing it. I found that it is the most effective at performing tasks that is not original research: paraphrasing text, brainstorming, debloat a section, etc. Of course, using a closed-source proprietary AI will not go well with Wikipedia's goals, but I wonder have there been any development from the WMF to make a "free" libre large language models? It seems that Wikifunctions and Abstract Wikipedia can integrate such an AI or someone can make a frontend gadget that makes calls to the AI at wikitech:Toolforge, but I haven't seen any activity about that. CactiStaccingCrane 14:34, 19 January 2023 (UTC)Reply[reply]

@CactiStaccingCrane: there is a bit of an academic community surrounding Natural language processing (the topic in which ChatGPT lives), where people share details about models. One interesting feature is that often the authors work for the big firms, so for example meta will publish a paper that gets used by Google, and then extended and used by OpenAi.
OpenAI's earlier model, GPT3, can do many of the things that ChatGPT can do, but ChatGPT, based on InstructGPT, is a bit better-suited for dialogue tasks. There is a paper describing what InstructGPT does.
Based on what's available at the moment, things like GPT-J and GPT-neo have freely available weights and are similar to GPT-3. It'd be quite easy to wire these up somewhere, potentially together with Wikipedia:User scripts, for them to be used by individuals. The issue to scaling this beyond yourself would be GPU time and WMF and Wikipedia's ethical concerns. Chris Albon heads up an ML team at WMF - so it might be interesting to see what they are upto. Talpedia (talk) 14:35, 28 January 2023 (UTC)Reply[reply]
It might be worth checking out Galactica as well - whose weights are freely available and was trained on papers. Talpedia (talk) 14:36, 28 January 2023 (UTC)Reply[reply]

The "Up, Down" buttons in the Teahouse[edit]

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)Reply[reply]

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)Reply[reply]
Maybe it could be in "Preferences". ‍ ‍ Helloheart ‍ 02:39, 20 January 2023 (UTC)Reply[reply]
Yes, a preference would be great. My tablet doesn't have Home and End keys... David10244 (talk) 10:09, 21 January 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]

Changing Wikipedia's media player's audio volume[edit]

Always when there is a sound or video file you can click on on a Wikipedia page, the audio is by default set to the loudest volume. This is very annoying. One must remember to mute the computer before clicking on the media player and lower the volume. Countless times I have forgotten that and the loud audio has made me jump, and probably woken up my neighbours. If there is a way for Wikipedia's developers to change the default volume to "medium" I, and probably many others, would be grateful. 2A07:A880:4601:1051:B6C5:AC44:87CB:4D90 (talk) 08:00, 20 January 2023 (UTC)Reply[reply]

There could be a little volume selector next to the audio. ‍ ‍ Helloheart ‍ 03:22, 22 January 2023 (UTC)Reply[reply]
There is, that is not their complaint. —TheDJ (talkcontribs) 12:35, 23 January 2023 (UTC)Reply[reply]

Project-independent quality assessments[edit]

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)Reply[reply]

  • 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)Reply[reply]
  • Oppose The road projects often have assessments differing from what other WikiProjects use. --Rschen7754 20:50, 22 January 2023 (UTC)Reply[reply]
    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)Reply[reply]
    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)Reply[reply]
    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)Reply[reply]
    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)Reply[reply]
    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)Reply[reply]
  • 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)Reply[reply]
  • 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)Reply[reply]
  • 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)Reply[reply]
    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)Reply[reply]
    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)Reply[reply]
    "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)Reply[reply]
    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)Reply[reply]
    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)Reply[reply]
    My fault for starting the bold voting, sorry! — Martin (MSGJ · talk) 12:31, 23 January 2023 (UTC)Reply[reply]
  • 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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]

Project-independent quality assessments break1[edit]

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)Reply[reply]

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)Reply[reply]
@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)Reply[reply]
  • Support. This is a partial solution to the problem discussed at Wikipedia:Village_pump_(miscellaneous)#Improper_handling_of_assessment_for_inactive_WikiProjects. Please note this is also similar/related to Wikipedia:Village_pump_(proposals)#Display_assessments_on_inactive_wikiproject_banners (ping User:Aymatth2, I hope there's no conflict between these proposals; maybe they should be merged somehow to prevent to separate discussions?).--Piotr Konieczny aka Prokonsul Piotrus| reply here 04:46, 25 January 2023 (UTC)Reply[reply]
    Piotrus: this is funny. When I tried to discuss this idea at the original discussion you said it was "out of scope". And now Aymatth2 has set up separate discussions you are suggesting they should be merged?! — Martin (MSGJ · talk) 12:40, 25 January 2023 (UTC)Reply[reply]
    I see these as two separate discussions, which can be resolved independently. Restoring assessments for inactive projects is fairly simple and low impact. Most editors would not notice the change. Starting to capture quality assessments at the article level rather than the project level has a broader impact, potentially involving mass conversion of many talk pages. Aymatth2 (talk) 16:35, 25 January 2023 (UTC)Reply[reply]
    I hardly insist on merging discussions if others don't see the need. My point was that your proposal was about replacing the existing system of WikiProject grading, which I said would not be accepted by some WikiProjects. We don't need to replace stuff, we just need to create a working backup, since the problem to be solved is simply that when a WikiProject is marked as inactive, it's assessment scheme is pointlessy disconnected from the larger system. Piotr Konieczny aka Prokonsul Piotrus| reply here 10:29, 26 January 2023 (UTC)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
👍 Like — Martin (MSGJ · talk) 12:19, 26 January 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]

Project-independent quality assessments break2[edit]

  • 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)Reply[reply]
    Thank youfor your offer, let's convene a working group (if this passes) — Martin (MSGJ · talk) 21:16, 25 January 2023 (UTC)Reply[reply]
    @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 paramters for specific WPs will be difficult. — Qwerfjkltalk 17:37, 26 January 2023 (UTC)Reply[reply]
    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)Reply[reply]

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

  • Support I have on occasion used a different quality assessment per wikiproject (strong article, but weak for climate change and visa versa), but the increased administrative burden is not worth it. —Femke 🐦 (talk) 17:59, 27 January 2023 (UTC)Reply[reply]
  • Looks good, don't think this will be too controversial so long as it's made clear that individual wikiprojects can still assess quality separately if preferred as for example MILHIST will want to do with lists. (talk) 03:18, 29 January 2023 (UTC)Reply[reply]
  • Support, Please yes. Abductive (reasoning) 11:10, 29 January 2023 (UTC)Reply[reply]

Technical implementation[edit]

@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)Reply[reply]

@CX Zoom: You are amazing. This will make things much smoother -- Guerillero Parlez Moi 12:40, 28 January 2023 (UTC)Reply[reply]
What he said :D Piotr Konieczny aka Prokonsul Piotrus| reply here 14:55, 28 January 2023 (UTC)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]
Also, I don't think backwards compatibility should be a concern. — Qwerfjkltalk 23:02, 28 January 2023 (UTC)Reply[reply]
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)Reply[reply]

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)Reply[reply]

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)Reply[reply]
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)Reply[reply]
We will get fewer complaints if we let different groups have the ability to opt out. WhatamIdoing (talk) 18:46, 28 January 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
  • "This page does not require a rating on the project's quality scale": You're correct, it is given by {{WPBannerMeta}} and would need to be edited to fix this. CX Zoom[he/him] (let's talk • {CX}) 15:46, 29 January 2023 (UTC)Reply[reply]
  • 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.: Doable, but again, needs editing {{WikiProject banner shell}}, {{WPBannerMeta}}. CX Zoom[he/him] (let's talk • {CX}) 15:46, 29 January 2023 (UTC)Reply[reply]
  • I can foresee problems with new projects and changes to project parameter: Can you please explain? CX Zoom[he/him] (let's talk • {CX}) 15:46, 29 January 2023 (UTC)Reply[reply]
    Once we have created the 1,000-odd {{X23}}-type templates, we have to create new ones for new projects, and change them when projects want new parameters passed to their templates. Aymatth2 (talk) 16:28, 29 January 2023 (UTC)Reply[reply]
    Honestly, is that that big of a deal? Most WikiProject templates are template-protected already, it is not like a random editor would edit the template without reading documentation and the parameters fails to pass through. The syntax is pretty easy to understand if you read the source code. CX Zoom[he/him] (let's talk • {CX}) 16:38, 29 January 2023 (UTC)Reply[reply]
  • Parsing the article-level class out of the page seems much simpler.: Probably not possible, or it may be. Refer to Aidan9382's comment above and Wikipedia:Village pump (technical)#Get un-nowikied parameter through module. The problem is you never know what part of the page someone puts the banner, at the top, below talk header or between ITN and AFD note or something else. You need the parse the entire page while ensuring that you don't fall into an infinite loop. CX Zoom[he/him] (let's talk • {CX}) 15:46, 29 January 2023 (UTC)Reply[reply]
    I think Aidan9382 solved the recursive call problem by adding <noinclude>...</noinclude> around the page transclusion. The position of the banner on the page is not important, although it should be just before the project templates. The page is already in memory and {{Str find}} and {{str sub old}} are efficient. There may well be a better way to parse the page, but this way works. @MSGJ: any comments? Aymatth2 (talk) 16:28, 29 January 2023 (UTC)Reply[reply]
    A cute way to simplify the logic would be to have {{WPBannerMeta}} wrap the article class value in <onlyinclude>...</onlyinclude>. That way {{ {{FULLPAGENAME}} }} would return the class value. See User:Aymatth2/Sandbox2. But that means the talk page could not be transcluded by anyone else, so I think is ruled out. Probably not much more efficient anyway. Aymatth2 (talk) 17:26, 29 January 2023 (UTC)Reply[reply]
@Aymatth2, perhaps section transclusion? — Qwerfjkltalk 16:25, 30 January 2023 (UTC)Reply[reply]
In fact, I might start working on this soon. — Qwerfjkltalk 19:50, 29 January 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]

Support French design Andre🚐 22:19, 4 February 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]

Moving forward[edit]

  • @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)Reply[reply]
    User:Evad37 may want to know about this for the User:Evad37/rater.js. WhatamIdoing (talk) 02:59, 31 January 2023 (UTC)Reply[reply]
    Looks great. CX Zoom[he/him] (let's talk • {CX}) 09:01, 31 January 2023 (UTC)Reply[reply]
    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)Reply[reply]
    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)Reply[reply]
    Absolutely - there will be a significant changes to a few templates required — Martin (MSGJ · talk) 12:49, 1 February 2023 (UTC)Reply[reply]
  • 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)Reply[reply]
    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)Reply[reply]
    @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)Reply[reply]
    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)Reply[reply]
    @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)Reply[reply]
    @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)Reply[reply]
    Fair enough. — Qwerfjkltalk 17:24, 31 January 2023 (UTC)Reply[reply]
    Although a clearer name than "special-criteria" e.g. project-assessment-criteria might be better. — Qwerfjkltalk 20:27, 31 January 2023 (UTC)Reply[reply]
    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)Reply[reply]
    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 assesment in the banner shell rather than an assessment in project banner. Donald Albury 16:40, 1 February 2023 (UTC)Reply[reply]

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)Reply[reply]

Draft proposal: Support for article-level quality assessments[edit]

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:

Article-level assessment example.jpg
  • {{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)Reply[reply]

Very clear, and hopefully uncontroversial. Thanks for putting this together — Martin (MSGJ · talk) 13:08, 3 February 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
Sorry, missed the point. Yes, importance is clearly project-specific. Aymatth2 (talk) 15:44, 5 February 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
I am reluctant to introduce yet another parameter to {{WPBannerMeta}}, but perhaps that is the only solution. Two options:
  • 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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
Enthusiastically support this proposal, both in spirit and in practice; it's long overdue. DFlhb (talk) 23:30, 4 February 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]

Combating piped link rot[edit]

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)Reply[reply]

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)Reply[reply]
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)Reply[reply]
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)Reply[reply]

Bring back the classic Table of Contents?[edit]

Hi everyone, Wikipedia has rolled out a new interface/design of Wikipedia. I appreciate the effort and while I am not sure about whether a redesign was necessary I think I can mostly defer to those who put more thought into it.

There is one thing that I feel more strongly about: the new design of the Table of Contents. The original TOC was a great way to get a quick glimpse at the contents of the whole page. The new format is harder to read and use because:

  1. its is design as a narrow column,
  2. subsection items are collapsed and not visible unless you uncollapse the list, and
  3. when there TOC is long (as is in this page), you have to scroll through it to be able to read all the contents. In other words, the ability to get a sense for the whole page by glancing at an easy-to-read TOC, is gone. Its advantage of being visible no matter where you scroll, and that it highlights what part of the TOC you are at, is nice. So without

While the new TOC displayed as column on the side could continue to exist, I wonder if other wiki editors would favor also bringing back the classic TOC, placed below the lead section.

I am not making a formal proposal yet. I am just trying to get a sense for whether this is something that only I care about or if the larger community has feels similarly. Oh, and if this is already being discussed somewhere else, please let me know and I can join that conversation. Thank you. Al83tito (talk) 13:57, 24 January 2023 (UTC)Reply[reply]

@Al83tito - There is loads and loads of discussion at WP:V22RFC2, where your voice will most certainly be heard. casualdejekyll 14:31, 24 January 2023 (UTC)Reply[reply]
@Casualdejekyll, great, thank you for pointing me there. I've added there my support for the rollback. Al83tito (talk) 15:54, 24 January 2023 (UTC)Reply[reply]
@Al83tito: please share your views at Wikipedia:Requests for comment/Rollback of Vector 2022#Bring back the TOC. Æo (talk) 16:23, 24 January 2023 (UTC)Reply[reply]
@ÆoThank you for the invitation! I just did. I am glad this is a shared concern. Al83tito (talk) 17:11, 24 January 2023 (UTC)Reply[reply]
I have a problem with the new style table of contents, but only as regards to high-traffic Wikipedia-space pages, such as WP:ANI and the Village Pump pages. People who watch or regularly edit those pages are mostly signed-on, so can change the skin if necessary. I don't think that there is any great problem with articles, which are what most readers read. Phil Bridger (talk) 19:52, 24 January 2023 (UTC)Reply[reply]
@Phil Bridger: I think the same; feel free to express your own views at Wikipedia:Requests for comment/Rollback of Vector 2022#Bring back the TOC. Æo (talk) 15:25, 26 January 2023 (UTC)Reply[reply]

Changing the font of English Wikipedia to the "Calibri" font[edit]

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)Reply[reply]

Why? Schazjmd (talk) 15:04, 27 January 2023 (UTC)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]
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)Reply[reply]
@Filelakeshoe In 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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
  • "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)Reply[reply]
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)Reply[reply]
@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)Reply[reply]
@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)Reply[reply]
@Xaosflux According to the screenshot of my browser (which is Microsoft Edge):
Screenshot for fonts of Microsoft Edge.png
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
No, this is a pointless nearly 1:1 change. Dronebogus (talk) 19:52, 28 January 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]

Pale blue top area and pale green left side[edit]

Hello Wikipedia. A tasteful easy way to make the site look smart would be to have the top inch of the site very pale blue and have the left strip very pale green. It wouldn’t be overwhelming. — Preceding unsigned comment added by (talk) 01:55, 28 January 2023 (UTC)Reply[reply]

Good idea. CactiStaccingCrane 03:25, 28 January 2023 (UTC)Reply[reply]

Skins Lite with user CSS[edit]

Could we provide more support and publicity for locally stored CSS to customise Wikipedia's appearance without affecting other users or burdening our servers? This might take the form of browser add-ons (probably recommending existing ones rather than reinventing the wheel) and a library of CSS tweaks to download. This method couldn't make any changes as drastic as Vector 2022, but potential applications include changing the font to Calibri or putting a blue bar across the top of the screen. We could even extend the principle to JavaScript, but that brings a greater risk of malware than CSS alone. Certes (talk) 09:44, 28 January 2023 (UTC)Reply[reply]

@Certes we do have a couple of places, see: Wikipedia:User scripts#Where can I find user scripts? to start. — xaosflux Talk 12:53, 28 January 2023 (UTC)Reply[reply]
That's great for JavaScript; I'm an appreciative regular user of many scripts found via there. I was thinking of something simpler (and safer) for CSS to change fonts, background colours and possibly even element widths. Certes (talk) 19:07, 28 January 2023 (UTC)Reply[reply]
The common.css file in your user space, also accessible from Special:MyPage/common.css, can be used for personal CSS rules. Help:User style has some examples and links to other examples. Are you looking for something more than that? An index of enhancements might be nice; someone would have to volunteer to take on that task (it's possible one exists and I just haven't found it). isaacl (talk) 22:06, 28 January 2023 (UTC)Reply[reply]
Yes; again I'm using common.css to collect various snippets that I've copied or made up. Perhaps Help:User style just needs a few more incoming links. Certes (talk) 22:41, 28 January 2023 (UTC)Reply[reply]
Not sure if you're looking for more automation in applying snippets? That could be written as a user script or gadget, similar to the script installer. (CSS files can be loaded via Javascript, and so CSS snippets can be managed by the script installer (and I do this myself for one snippet I wrote), though this isn't great from a performance perspective as each snippet would be loaded separately.) Or perhaps more cookbook-like documentation is what you would like? isaacl (talk) 23:11, 28 January 2023 (UTC)Reply[reply]
It's an open-ended suggestion, but I find the [Java]script installer easy to use, so something similar for CSS might help people who would like to change Wikipedia's appearance, even if those changes aren't ideal for use in the default appearance. Certes (talk) 23:56, 28 January 2023 (UTC)Reply[reply]
Sure, just trying to explore what types of options you are considering. A tricky part with UI customization is working out conflicts. Some snippets might not work together unless their rules are adjusted, or unless they are placed in a specific order in the CSS file. (Although new UI widgets added with Javascript can also conflict, I think CSS problems are generally more subtle, as interactions between style rules aren't always readily apparent.) It would be a nice collaborative project to get a tool working that could enable different customizations from different authors. isaacl (talk) 00:16, 29 January 2023 (UTC)Reply[reply]

New layout: Tools fighting with sidebars[edit]

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)Reply[reply]

@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)Reply[reply]
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)Reply[reply]
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)Reply[reply]

Standardizing Mathematical Notation[edit]

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)Reply[reply]

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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
Could that be package doublestroke? --Shmuel (Seymour J.) Metz Username:Chatul (talk) 20:58, 29 January 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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)Reply[reply]

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)Reply[reply]

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)Reply[reply]

A Project Review Page (= Shared Watchlist)[edit]

Suggestion :

  • Create a Project Review Page (PRP) (modified RSS feed)
  • Editors can then add as the project watch page as a single item to their watch list

Why? :

  • Editors are swamped by large watchlists, or are not aware of important issues on a project
  • Sometimes just want to know an experienced editor has reviewed a change
  • Projects were intended to be busier and are not acting as the first step in Wikipedia:Requests_for_comment#Before_starting_the_process
  • May reduce some negative behaviours by making them more visible

Features :

  • Can now ping a project, which adds to the PRP
  • Comments on the PRP are magic-ed/synched to appear on the article comment pages they refer to
  • Projects can specify what appears on their PRP
  • PRP shows human readable summary with content folded to make it easier to read
  • Ability to flag item quickly that they have been reviewed, or more eyes/editors are needed Wakelamp d[@[email protected]]b (talk) 13:43, 30 January 2023 (UTC)Reply[reply]
It's not quite what you asked for, but Special:RecentChangesLinked/(list name) shows recent changes to pages linked in a list. I use it daily on a few lists, though such changes don't pop up automatically on my watchlist. Certes (talk) 16:08, 30 January 2023 (UTC)Reply[reply]


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)Reply[reply]

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

Vector 2022 third question[edit]

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)Reply[reply]

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)Reply[reply]
Aaand... created at Wikipedia:Village pump (proposals)#RFC: Vector 2022 and Mobile. Aasim - Herrscher of Wikis ❄️ 23:02, 2 February 2023 (UTC)Reply[reply]
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)Reply[reply]
You have to wait for v22 to be actually suitable on mobile. Aaron Liu (talk) 15:23, 3 February 2023 (UTC)Reply[reply]

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

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)Reply[reply]

Re-ping, due to typo: User:Widr. Mathglot (talk) 00:21, 5 February 2023 (UTC)Reply[reply]
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)Reply[reply]
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)Reply[reply]
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 (