The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.A summary of the conclusions reached follows.
Overall consensus is this would be an improvement. As this would be displayed on every page, there were technical concerns about performance. A good suggestion was to take the result of this RFC as an indication of whether or not the community wants this feature. Technical details can be worked out with WMF staff, and performance should be tested. If things do not go well, we would expect this feature not to be deployed and further discussion could happen about whether any required engineering would be worthwhile. (Personal aside: the examples shown do not work correctly in dark mode; presumably this would be fixed before deployment.)
Objections were varied. Some editors were concerned about the quality of the article rating processes, which version was reviewed, and/or the terminology used (especially "good"). Other editors said the fact that the terms-of-art are linked mitigates these concerns to some degree. There was much debate over the timeliness of de-listing and repair of bad edits to good and featured articles. A small number of editors supported removing the tagline completely, or effectively said something like "not broken, don't fix it".
Support was strongly centered on the same themes. Many editors thought it was a good idea or problem-solving to increase reader awareness of these quality ratings. Some said this was an important part of building trust and information literacy, or that this might encourage more editors to improve articles to these high levels, or improve post-promotion quality monitoring.
Mobile: Several editors expressed hope that quality ratings can be somehow displayed on mobile devices in the future. One editor objected to doing this on mobile, apparently misunderstanding that the proposed implementation would not be seen there.
Topicons: The question of whether to keep the same information in topicons was raised; more editors supporting doing so, especially at first so that differential click-through rates could be observed. This can be revisited after more data is collected. It was suggested to incorporate topicons into the tagline immediately, but this was not accepted. This could be revisited if they are eliminated as a standalone thing.
Apologies for the long text to follow but I think a detailed RFCBEFORE and implementation is necessary for such a highly-visible proposal.
There's been perennial proposals for increasing the visibility of page status, with a fair amount of assent but no proposed directions. Many editors in prior discussions have felt the topicon is too small a notice that doesn't accurately reflect the amount of work put into raising articles to featured status. Other editors think the topicons are opaque to readers, and feel that more prominence will draw editors to these backend projects.
I've developed Module:Page assessment raw, a simplified version of Module:Page assessment that uses the newest features in the MediaWiki pageAssessments extension. I wanted to have a duplicate module to reduce the expensive function count (since this will be on every page) and to allow full- or template-protecting the module (for the same reason).
I think the most efficient way to implement this proposal is to fully replace the page MediaWiki:Tagline with a switch-case function to change the tagline based on the output of the Lua module:
{{#switch:{{#invoke:Page assessment raw|get_class|page={{FULLPAGENAME}}|project=Project-independent assessment}}| FA = A ''[[Wikipedia:Featured articles (linked from tagline)|featured article]]'' from
| FL = A ''[[Wikipedia:Featured lists (linked from tagline)|featured list]]'' from
| GA = A ''[[Wikipedia:Good articles (linked from tagline)|good article]]'' from
| From
}} Wikipedia, the free encyclopedia
Support a small change (probably smaller than people expect, considering the banner blindness phenomenon) which could nevertheless increase new editor attraction from people curious enough to click the link. I don't really see any downsides. ~~ AirshipJungleman29 (talk) 20:19, 7 July 2025 (UTC)[reply]
Support. Good articles being invisible on mobile – 65% of our readers – doesn't make sense. Bringing us to some parity with the web version communicates to readers that some verification and vetting effort has been made, especially with the recently increased level of scrutiny required by GAs (and much-discussed at WT:GAN. — ImaginesTigers (talk) 20:23, 7 July 2025 (UTC)[reply]
This sadly won't affect mobile users as the tagline does not appear in Minerva. It would be a good impetus for bugging WMF over at Phabricator to show the tagline though. Dan Leonard (talk • contribs)20:33, 7 July 2025 (UTC)[reply]
@Dan Leonard: Thanks for telling me that, although it is disappointing. I assume this affects web Minerva, too? If so, do we have any statistics on how many users aren't using Minerva at this point? I'd assume the number is relatively low, and largely our most engaged user base. — ImaginesTigers (talk) 22:31, 7 July 2025 (UTC)[reply]
Other respondents mentioned here that mobile topicons don't seem to have any progress and community work on adding a mobile tagline would probably be easier than adding a mobile topicon, not to mention the engineering needed to have parity with tooltips on mobile. Aaron Liu (talk) 16:18, 9 July 2025 (UTC)[reply]
Support a minor change but a positive one. Anything that (1) raises awareness of our quality content and (2) might conceivably encourage readers to contribute is a good thing. Cremastra (talk) 20:48, 7 July 2025 (UTC)[reply]
Support taglines are harder to miss, so this does seem like an improvement. Would this also remove the topicons, or would a highlighted article end up displaying both? Paprikaiser (talk) 20:51, 7 July 2025 (UTC)[reply]
Support per Airship, though I think the current wording is needlessly verbose. I feel it could be more impactful if it said something like "A featured article, meaning it represents the best Wikipedia has to offer" IAWW (talk) 20:55, 7 July 2025 (UTC)[reply]
Sorry, verbose wasn't a good word. I mean the current wording doesn't make the best use of space. Adding "from Wikipedia, the free encyclopedia" is pointless, because the reader obviously already knows this. Those words could be replaced with something different that the reader doesn't already know, like what a "featured article" is. IAWW (talk) 21:23, 7 July 2025 (UTC)[reply]
Support I think this is a better way to make the good and featured stand out. If this proposal passes, I recommend removing the topicons. I would also like to implement this change to all the vital articles as well. For example, George Washington would say A level 3 vital and featured article from Wikipedia, the free encyclopedia while an article that is neither good nor featured, but vital like Christianity would say A level 3 vital article from Wikipedia, the free encyclopedia. Interstellarity (talk) 21:07, 7 July 2025 (UTC)[reply]
Vital article status seems less important, as not only is it less transparent to readers ("is level 1 vital the least or most important?"), but also only relates to the topic itself rather than the quality of the article they are reading. Knowing that what you are reading has been through a formal review process is great to gauge the level of trust you want to give to the article, knowing that the subject has been assessed to be important, less so. Chaotic Enby (talk · contribs) 21:48, 7 July 2025 (UTC)[reply]
For example, a FA/GA badge with enough context could be an interesting and valuable trust signal for a reader, similar to how warning templates are often used as distrust signals when articles are low quality. On the other hand, I assume that a "protected" badge would be of little interest to anyone who isn't logged in and actively considering contributing to the wiki.
Another tension in the current proposal is the form of the indicators themselves. As mentioned in the VP discussions about this, there may be a low awareness of FA/GA for casual readers, and an unlabelled icon might not "onboard" casual readers into explaining what FA/GA are and why they're useful.
While the tagline isn't currently shown in Minerva either, WMF might be more amenable to displaying it as it is likely technically simpler than showing topicons and solves the problem of an unlabelled icon might not 'onboard' casual readers. It's currently tracked, albeit stale, at phabricator:T349117. Dan Leonard (talk • contribs)21:25, 7 July 2025 (UTC)[reply]
Support; I agree with AirshipJungleman that it likely won't be noticed by the vast majority of users, but if one percent are intrigued and find out more about the Good & Featured processes, I think that makes it more worthwhile. I also agree with the takes that it'd be more applicable to mobile than a topicon would be, and I think it adds context to the topicons on desktop if people don't realize you can click on them. Generalissima (talk) (it/she) 22:47, 7 July 2025 (UTC)[reply]
Support as the overdue implementation of repeated consensus for greater visibility. Like Chaotic Enby, I would never want this expanded to other article classes because many (myself included) use Rater for an AI-generated classification that works for our technical needs but communicates little to readers. I would have this appear alongside the topicon, both because the aesthetic badge is a big motivator for article writers and to run Dan's click-through experiment. Hoping that even if the topicons are never added to mobile view, at least this expanded tagline can be. ViridianPenguin🐧 (💬) 22:55, 7 July 2025 (UTC)[reply]
As I said in a previous discussion, I don't agree with the premise. I think the current amount of prominence given to the article rating is appropriate, given the way the rating is determined. Thus I do not support changing the tagline in this manner. isaacl (talk) 22:55, 7 July 2025 (UTC)[reply]
To expand slightly on the determination concerns: I appreciate that the good article/featured article review processes are the only ones we have for this type of article rating. However they do not ensure evaluation by subject matter experts with the background knowledge to best evaluate the comprehensiveness and accuracy of the article. I think giving the rating higher prominence would raise reader expectations that an evaluation has been made by subject matter experts. isaacl (talk) 15:10, 11 July 2025 (UTC)[reply]
Support per above. These two ratings are the ones with the least amount of arbitrariness so I don't think this status should have any less prominence. They're already displayed as topicons everywhere except Minerva, which shows the reader interest. If these statuses are any overprominent, there's already consensus for them being so. Aaron Liu (talk) 23:05, 7 July 2025 (UTC)[reply]
Oppose as the tagline should be kept simple. I agree with isaacl that we don't need to give more prominence to article ratings, which are subjective anyway. Some1 (talk) 23:27, 7 July 2025 (UTC)[reply]
A tagline that changes depending on the article rating isn't simple, and I prefer that the tagline text remains the same for every article, regardless of its rating. Also, it's misleading to state that Example article is "A good article from Wikipedia, the free encyclopedia" when that rating reflects only one editor's review (opinion) of that article. In that instance, the tagline is more misleading than the topicon, especially when most readers won't bother to click on the good article link in the tagline. Many editors in prior discussions have felt the topicon is too small a notice that doesn't accurately reflect the amount of work put into raising articles to featured status Seems like this proposal is more for the editors than for the readers. Some1 (talk) 00:32, 8 July 2025 (UTC)[reply]
What's the impact of this small bit of added complexity?
when that rating reflects only one editor's review (opinion) of that article
I think that's only an argument for renaming GoodArticle. In my opinion, "good article" signifies only as much value as it should. This also means if a reader thinks "hey this is not good-article quality" they can start something on the talk page. Aaron Liu (talk) 18:20, 8 July 2025 (UTC)[reply]
I don't see the harm per se, but more visibility for GA/FA ratings makes more critical the need for participation at WP:FAR and WP:GAR and for those processes to function properly. CMD (talk) 00:38, 8 July 2025 (UTC)[reply]
Support. As I mentioned in the 2023 RFC: I once observed a high school classroom that happened to be teaching research skills on using Wikipedia. The teacher said that the lock icon in the corner meant that it had been reviewed and was safe. Readers have no idea what our esoteric icons mean, so a little explanation of what exactly is verified and what isn't could go a long way towards mutual incentives. This is a smart way to start. czar02:08, 8 July 2025 (UTC)[reply]
Support I don't see the harm in mentioning the fact that the article has reached good/featured status, and it would certainly help reader to become aware of those statuses and as a compliment to the topicons. Speaking of those icons, why don't we include them in the tagline too so the reader will notice it when reading the tagline and be able to recognize it when reading another article on its tagline or on the corner of said artile. So maybe something like "A good article from Wikipedia, the free Encyclopedia." Here I placed the icon after the phrase introducing the article class since it looked akward to have it immediately following the indefinite article "A" when I tried that first. Gramix13 (talk) 02:52, 8 July 2025 (UTC)[reply]
This thought came to mind for me as well. I like the idea of having a more prominent visual, but I feel like it'd be redundant presuming we're keeping the topicon. My preferred approach to accomplish this would be to move the icon from the top right to directly next to the article title, as was previously proposed. Sdkbtalk05:09, 8 July 2025 (UTC)[reply]
Noting that since that 2020 proposal argued for matching the Danish Wikipedia's practice, dawiki has switched to our practice of having the topicon appear in the top-right corner (e.g., da:Israel). ViridianPenguin🐧 (💬) 13:40, 8 July 2025 (UTC)[reply]
Unfortunately not. I checked through their relevant talk pages and Landsbybrønden (village well) archives to no avail in identifying why they switched. ViridianPenguin🐧 (💬) 16:04, 8 July 2025 (UTC)[reply]
Support. An article's good or featured status is a key piece of information that every media-literate reader ought to pay attention to, but our current display is nowhere near prominent enough to make that happen. This is a well-thought-out step in the right direction toward making that happen. I support keeping the topicons, and ultimately moving them next to the article title per the prior proposal. I also continue to hope that phab:T75299 is taken up so that icons display on mobile. Sdkbtalk05:07, 8 July 2025 (UTC)[reply]
Strong support and additional request. I have a script that already does this, but this being an automatic thing for other editors would be extremely useful in helping people maintain articles. Possibly, we could also have 'Currently a featured article candidate', 'Former featured article', 'Currently a good article nominee' too; that might be a bit too technical though. 750h+05:18, 8 July 2025 (UTC)[reply]
Being a candidate or having former status is mainly relevant for editors, not readers, so I would not support this. Sdkbtalk06:03, 8 July 2025 (UTC)[reply]
Support, great idea. I'd also support other measures to improve GA/A/FA visibility, like moving the topicon to be right beside the title. Toadspike[Talk]09:12, 8 July 2025 (UTC)[reply]
My reaction to much of the opposition is that our readers are not stupid. They know what Wikipedia is. They will not see "good article" and think "this must be a 105% perfect article certified by the leading experts in the field and then fully-protected so no-one can edit the page again". They know that Wikipedia is a wiki written by regular people, many of whom are not experts on the topics they write about, and that Wikipedia articles can generally be edited by anyone, anytime. By calling something a "good article" or "featured article", whether in a tagline or with an icon, we simply argue that this article is better than many others. And while were bashing the names as "hokey" or "bizarre", I must point out that "featured" sounds stupid outside the context of being featured on the main page, while "good" is a simple English word that accurately describes the point of the rating. Toadspike[Talk]13:34, 12 July 2025 (UTC)[reply]
Oppose The top of many articles, including GA and FA, are already overly cluttered. I would support the reduction of clutter rather than the adding to it. -- LCU ActivelyDisinterested«@» °∆t°16:01, 8 July 2025 (UTC)[reply]
I understand your point, but this would add practically no clutter, as "From Wikipedia, the free encyclopedia" is still a default tagline. I assume you are generically opposed to the tagline in the first place? Dege31 (talk) 17:17, 8 July 2025 (UTC)[reply]
Support. Awareness of the featured article process is partially what inspired me to begin contributing to Wikipedia, and I feel like increasing this awareness wouldn't just inspire more people to edit, but would also inform readers which articles have been more diligently reviewed to eliminate gaps and verification errors. Sad that this won't be visible on mobile but it's a step in the right direction with no glaring downsides. Fathoms Below(talk)19:50, 8 July 2025 (UTC)[reply]
Support. I can see no downsides to this proposal - anything that helps promote good and featured content, and potentially bring in new editors, is a positive in my book, and this is a pretty nice and yet non-obtrusive way to do so. CoconutOctopustalk20:40, 8 July 2025 (UTC)[reply]
Comment. It sounds like we are proposing adding a module to every single page load on the entire wiki. Has someone who understands mediawiki caching and performance given some thought about if this will cause stress to the servers or performance issues? –Novem Linguae (talk) 22:48, 8 July 2025 (UTC)[reply]
mw.title.new().pageAssessments is an WP:EXPENSIVE function, but since it is only called once it usually shouldn't often be an issue considering the per-page limit of 500. The module does access the class rating via iteration (see the for loop at lines 29–33) rather than via random access, which is admittedly inelegant but probably not too inefficient. Sadly, I got what feels like a WONTFIX for random access at phabricator:T396135. Regardless, whether the community wants something shouldn't be dependent on whether it is currently feasible. I trust the interface admins and WMF will fix things if the community breaks them. This isn't to say I am in any way opposed to a code review of this, which I welcome wholeheartedly. I am sure there is something here that could be made more efficient. Dan Leonard (talk • contribs)23:10, 8 July 2025 (UTC)[reply]
I tend to agree with the link to WP:DWAP here (which has been relevant for parts of the interface like this before, the link to Category in the footer of each page used to be treated as sufficiently expensive for some reason or another that we couldn't let it vary by number of categories or something like that).
The only way to get an authoritative answer to this question would be to ask WMF directly I think.
That aside, I'm actually not really certain this will work the way OP wants it to. Has it been attempted on test Wikipedia? Izno (talk) 23:57, 8 July 2025 (UTC)[reply]
Putting my developer hat on, to my understanding the MediaWiki:tagline is added by a user's skin. As such, running page assessments at that layer is prohibitively expensive since the skin layer is re-rendered every time a user visits a page (as opposed to the parser cache layer where data is computed/rendered every edit and thus can have expensive functions). Unless folks contradict me, to my understanding this will require a significant investment of engineering effort to implement into core-mediawiki (or one of it's extensions) which I'm not sure is worth the outcome.
Putting on my WP:INTADMIN hat on, I'm not sure I'm completely onboard with using JS (or even Lua) to hack and slash at the existing tagline at pages that we as enwiki are wanting folks to visit. (For context, every time such a thing is implemented folks with bad internet connection will see a flash of unstyled content that often makes the navigating/reading experience worse). Sohom (talk) 00:51, 9 July 2025 (UTC)[reply]
I don't think it's prohibitively expensive, just a bit expensive. Performance is not affected for logged-out users as for them the entire page html, including the skin, is CDN-cached. For logged-in users, it does result in re-rendering, however note that the Lua .pageAssessments call is just a single SELECT call in the db. Being marked as WP:EXPENSIVE doesn't give the full picture. – SD0001 (talk) 05:35, 9 July 2025 (UTC)[reply]
It definitely isn't a Lua "hack". Functionality like retrieving page assessments have been exposed to Lua for use cases like this. It doesn't make much sense to implement everything natively in MediaWiki since most MediaWiki installations don't have a concept of FAs or GAs. – SD0001 (talk) 05:37, 9 July 2025 (UTC)[reply]
It's a hack because it hacks a message that shows at a place convenient for desktop users (⅓), but is not shown to mobile web and app users (⅔). It's not a complete solution, it's a hack. Other than that, there's a javascript gadget that also hacks the message, used by some 1000 active users. Why reinvent the wheel?Ponor (talk) 15:37, 10 July 2025 (UTC)[reply]
Changed my mind, it's a complicated js script. I'd still like to see a solution that every reader can see. Indicators are fine. Ponor (talk) 15:48, 10 July 2025 (UTC)[reply]
Using client-side JavaScript to scrape the talk page and modify the DOM after page load is a very different solution than using a MediaWiki extension's intended functionality to modify pages server-side. Mobile users miss out on a lot of things, including navboxes, sidebars, and even categories. Are categories "hacks" because Minerva chooses to hide them? Dan Leonard (talk • contribs)15:45, 10 July 2025 (UTC)[reply]
Agree about the gadget, just checked the code: ugh, that's an ugly one. But here, we're saying "this is important, we want everyone to see it", while at the same time we know that two in three readers will not see it. Categories and navboxes are the things at the bottom. Most users rarely go past the lead. So I'd say there is some importance difference. I'm not strongly opposed to "the hack" – I'd simply like to see a better solution. Use different indicators - find better icons? Use icon+text on desktop, icon alone on mobile? Ponor (talk) 16:10, 10 July 2025 (UTC)[reply]
Topicons aren't shown on mobile because it's a difficult implementation and because they're even more opaque to readers (see phabricator:T75299#10512584): on desktop, {{featured article}} at least benefits its readers with the tooltip "This is a featured article. Click here for more information", but touchscreens don't get tooltips. The tagline isn't shown on Minerva either, but it's probably a much simpler implementation and hasn't been done simply as a screen real estate saving measure. If we can get the tagline shown (phabricator:T349117), it'd serve the same purpose as the tooltip serves for desktop users. I do want mobile users to see this, and I think using the tagline is an important first step to getting WMF to increase visibility. Dan Leonard (talk • contribs)16:19, 10 July 2025 (UTC)[reply]
Oppose What is the point? FAs and GAs are not necessarily better than normal articles, and the reader does not care about who followed internal Wikiprocedures to get something declared FA/GA. Also per Ahecht. Polygnotus (talk) 03:36, 9 July 2025 (UTC)[reply]
While you are technically correct that some articles which haven't been formally reviewed are as good as GAs or FAs, the average GA is certainly better than the average article, which is a stub or not much more. Toadspike[Talk]07:58, 9 July 2025 (UTC)[reply]
They have necessarily underwent quality control, which includes things like "checking whether the sources confirm what the article say"- not exactly obscure, Wikipedia-only procedures. Dege31 (talk) 16:33, 9 July 2025 (UTC)[reply]
Support as an idea. If the current technical implementation is insufficient, then it can still be recognised as something the community supports and perhaps be added at a later point. Even if that is not likely to happen any time soon given other technical priorities, it would be better having than having another RFC every time someone comes up with another way to make FAs/GAs more visible. novovtalkedits08:26, 9 July 2025 (UTC)[reply]
I agree with the idea that consensus and technical details should be separate. The closer should feel free to find clear consensus that the community wants this (because they do). Then a consultation with a WMF dev good at performance and caching (maybe via a Phab ticket tagged #performance_issue and pinging someone like Krinkle?) and/or a trial should probably be encouraged in the close as a next step, but can leave that part vague / not as strong as the community consensus part. –Novem Linguae (talk) 09:10, 9 July 2025 (UTC)[reply]
Oppose hacking a module in to the tagline for every single page is a poor technical implementation, especially for something that is only needed on an extreme minority of pages. — xaosfluxTalk12:29, 9 July 2025 (UTC)[reply]
Also oppose conceptually using the tagline for this purpose; work was already spent on indicators and if wanted I'd prefer improvement to indicators. Indicators are also much more consistent across the Wikipedia's in other languages. — xaosfluxTalk12:32, 9 July 2025 (UTC)[reply]
The French, German, and Russian Wikipedias all use different icons (from us and from each other). Of the Wikipedias I regularly visit only Chinese and Norwegian use the same icons we do. Smaller wikis like Alemannisch don't have quality ratings at all. Toadspike[Talk]13:20, 12 July 2025 (UTC)[reply]
Additional oppose reason, as others have called out already, this will not solve the problem of casual readers (who increasingly use minerva) not seeing the rating - as the tagline element isn't even shown on that skin. I'd rather see indicator support added to minerva. — xaosfluxTalk14:54, 10 July 2025 (UTC)[reply]
Support A tagline seems much more understandable than cryptic and tiny symbols that we use today. If this proposal passes, I would support the removal of original topicons, as they will be made redundant. I don't know how feasible it is, but I would like WMF develop a way to show the taglines in the minerva skin as well. Catalk to me!15:40, 9 July 2025 (UTC)[reply]
Oppose The tagline is there to provide credit to Wikipedia, not provide extra information such as an assessment from a niche rating project that the vast majority of Wikipedia readers and editors don't participate in. I don't think we should be adding overhead to every single page, including non-article pages, just for increasing the visibility of page status. Not to mention that, to those not familiar with our lingo, "A good article from Wikipedia" sounds incredibly hokey, if not conceited.--Ahecht (TALK PAGE)15:53, 9 July 2025 (UTC)[reply]
This is a very controversial opinion. The question of whether cleanup templates' visibility to readers is in conflict with the WP:NODISCLAIMERS philosophy has been a battle on here since the beginning. The official primary purpose, a little fiction we tell ourselves to resolve this, is to categorize articles and flag areas for other editors. Dan Leonard (talk • contribs)02:12, 10 July 2025 (UTC)[reply]
@Dan Leonardto inform that there is (more likely to be) a problem with an extract from an article doesn't specify who gets informed, it might be the readers or editors or both. Polygnotus (talk) 02:26, 10 July 2025 (UTC)[reply]
I'm just explaining Aaron's comment, which I saw as a joke referencing the TfD battles over cleanup templates and the official policy that they aren't for informing readers. Dan Leonard (talk • contribs)02:31, 10 July 2025 (UTC)[reply]
I meant "inform readers" lol. Dan is correct in what I'm talking about, but I did have a point with my joke though in that the tagline would have the inverse purpose. Aaron Liu (talk) 17:39, 10 July 2025 (UTC)[reply]
Readers should absolutely be paying attention to an article's assessed GA/FA status. The fact that something underwent (in the case of FAs) a rigorous review process is a key piece of information as a reader decides to what extent to trust an article (and yes, in an ideal world they'd be verifying everything, but in practice doing that rigorously would take nearly as long as writing the article). It's something we pay attention to even when just reading an article we do not intend to edit. And it's therefore something we should make noticeable enough to readers that they can do the name. Sdkbtalk23:30, 9 July 2025 (UTC)[reply]
@Sdkb I think one of the major weaknesses and strengths on Wikipedia is that the writers are completely clueless about the readers. should absolutely be paying attention to an article's assessed GA/FA status. No, of course not. It means only that someone jumped through some hoops. a reader decides to what extent to trust an article In my experience, people don't work like that. it's therefore something we should make noticeable enough to readers that they can do the name. Why should we tell a reader information that is not helpful to them and that they do not know how to interpret. Doesn't make sense to me. Polygnotus (talk) 00:13, 10 July 2025 (UTC)[reply]
GA/FA statuses only reflect the quality of an article at the time it was reviewed. Some (most?) of the articles with GA/FA status haven't been reassessed in several years or have undergone so many changes since achieving their status that their current quality differs significantly from when they were first reviewed. J.K. Rowling, a "Featured article" (which apparently means that it's "[one] of the best articles Wikipedia has to offer"), currently has two large templates on the article, one of which indicates that it has neutrality issues. So again, these article ratings are subjective, and readers do not need to be paying attention to them any more than they need to. Some1 (talk) 00:31, 10 July 2025 (UTC)[reply]
You're conveniently omitting that there is currently an effort to delist the Rowling article. Which illustrates exactly how it should be working: Quality articles that no longer meet standards should be reassessed. Sdkbtalk02:42, 10 July 2025 (UTC)[reply]
The J.K. Rowling discussion has been opened for almost a month now (and who knows how long it'll take for that discussion to close), and in the meantime, that FA status is there, misleading readers into thinking the article is still "one of Wikipedia's very best works" despite multiple experienced editors arguing that the article should be delisted. Some1 (talk) 03:31, 10 July 2025 (UTC)[reply]
And plenty of others arguing that it should remain listed (and those editors are arguing that the aforementioned maintenance tags were added in an effort to try to get it delisted). I take no position on whether it should or should not remain an FA, but the active discussion around it, as the example you chose, contradicts the notion that FAs are listed and then never looked at/reevaluated again. People who watch them periodically send them to FAR, and WP:URFA/2020 has been going through every single one systemically. Of course, as in most areas of the encyclopedia, we don't have the editor capacity to monitor everything as closely as we wish, but that's no reason to give up on the project or minimize the value it provides to readers — if they know about it. Sdkbtalk17:40, 10 July 2025 (UTC)[reply]
Hence the giant neutrality tag on the top of the article. It's just like a "Disputed" inline tag right after a claim, and here the claim is that this is once of Wikipedia's very best works. Aaron Liu (talk) 17:43, 10 July 2025 (UTC)[reply]
As an editor who has provided a very critical review, I’m not concerned about the time the review has been active. It has been Featured for several years in roughly the same state. A month is pretty normal and I expect it’ll be moved to FARC soon (I’m going to post a follow up review tomorrow, which is how the process works). The moment we start speeding up the process of delisting, you will see (for example) large swathes of gender related content beset by meaningless and inaction able critiques to force delistings. WP:There is no rush. — ImaginesTigers (talk) 09:16, 11 July 2025 (UTC)[reply]
All the common slogans of Wikipedia sound incredibly hokey, if not conceited, to those who are opposed to wikis, the free knowledge movement, ... even the tagline itself, what if someone thinks "no such thing as a free lunch!", we have pages explaining free as in libre vs free as in gratis! This is just a walking on eggshells mentality which is not productive. Dege31 (talk) 16:38, 9 July 2025 (UTC)[reply]
I don't see how either the FA or GA process is niche. If you were talking about WP:ACLASS, then sure, that would be truly niche. But almost 1% of all articles are already either FAs or GAs. As long as it links to the actual WP:FA and WP:GA pages, it isn't any more niche than the topicons already there, which I'd argue are even more cryptic to the casual reader. Epicgenius (talk) 18:16, 9 July 2025 (UTC)[reply]
Less than 1% is pretty niche. How many Wikipedia articles would you have to read to have a 50:50 chance of stumbling on one of the 0.76% that are GA, FA, or FL? NebY (talk) 18:34, 9 July 2025 (UTC)[reply]
But GAs and FAs are some of the most-frequented articles that readers are likely to bump into. They're not randomly distributed. Cremastra (talk) 18:56, 9 July 2025 (UTC)[reply]
Support per above; I have an extension installed which does practically the same thing. To partially address the concerns some users have about the taglines being potentially misleading b/c the GA/FA is super old, the extension says Currently undergoing review of its featured status etc when articles get re-reviewed. ❤HistoryTheorist❤04:21, 10 July 2025 (UTC)[reply]
Support It took me ages to find the tagline. I'd never noticed it, but support anything that lets our readers get more of a sense of relative fidelity of Wikipedia articles. A good thing that the GAR process is alive and kicking: we've been able to remove the icon from the 'worst' articles in recent years. —Femke 🐦 (talk) 07:32, 10 July 2025 (UTC)[reply]
Oppose. We should not be misleading readers into thinking that the article rating is objective or meaningful when it is neither. Thryduulf (talk) 08:19, 10 July 2025 (UTC)[reply]
Are you arguing the average featured article is not better than the average article? Because that's what article ratings having "no meaning" would mean. The reality is the opposite. The average featured article is way way better. IAWW (talk) 15:29, 10 July 2025 (UTC)[reply]
@It is a wonderful world by what objective measure is the comparison being made? What is the definition of an "average article" and an "average featured article"? The median-length featured article is definitely going to be longer than the median-length non-featured article but length alone is not a reliable indicator of anything other than length (e.g. an article that is two thirds reactions from random celebrities on Twitter is worse than an article one third its length that is purely encyclopaedic). Article ratings are meaningless in that they do not reliably convey to the reader any information about the quality of the current state of the article. Even a featured article rating simply says that someone put a lot of effort into satisfying a small number of other people that they put a lot of effort into the article at some point in the past. There is no requirement for writers or reviewers to be subject matter experts so there is no guarantee it is any more or less accurate than any other cited article. Ratings other than featured are worth less than the paper they are written on. Thryduulf (talk) 17:16, 10 July 2025 (UTC)[reply]
If there's a non-featured article whose quality is better than a featured article, that article should be made a featured article. Aaron Liu (talk) 17:45, 10 July 2025 (UTC)[reply]
That assumes that the current version of every featured article is of sufficient standard to be regarded as one of Wikipedia's best, that is simply not true. If you think the effort in getting a rubber stamp of approval from a self-selecting group of non-experts is worth the time it takes, good on you, but that doesn't make the rubber stamp meaningful to readers. Thryduulf (talk) 19:10, 10 July 2025 (UTC)[reply]
If an article does not meet the standards, you can be the one to delist it. The great bulk of articles do not meet Feature quality standards, therefore articles that do are in fact among the best. Aaron Liu (talk) 19:49, 10 July 2025 (UTC)[reply]
It doesn't matter what objective measure you use. Any measure which tries its best to rigorously define the heuristic concept of article quality will show featured articles are better than non-featured articles in 99% of cases. You can see this by picking random featured articles and random non-featured articles – seriously, just go and try it right now! You don't need to rigorously define "average" and "objective measure" or any other words to see this. IAWW (talk) 17:50, 10 July 2025 (UTC)[reply]
Any measure which tries its best to rigorously define the heuristic concept of article quality will show featured articles are better than non-featured articles in 99% of cases. citation needed. Thryduulf (talk) 19:07, 10 July 2025 (UTC)[reply]
The sentence after: "you can see this by picking random featured articles and random non-featured articles" – go try it IAWW (talk) 19:16, 10 July 2025 (UTC)[reply]
https://doi.org/10.1609/icwsm.v3i1.13959 empirically derived features of FAs from FAC discussions, and for two articles I've tried it seems to work. Regardless, Wond's claim did not say the measures have to be objective; that's what you introduced. To see whether an article has better quality, we only need to agree in our judgement, no matter whether it is based on subjective criteria or objective criteria. Aaron Liu (talk) 19:53, 10 July 2025 (UTC)[reply]
Reading the abstract of that 2009 study (I don't have time to read more), it seems to only show that articles that were awarded featured status reliably contained the features the FAC process looked for - which is unsurprising, not really relevant to this discussion, says nothing about whether those features do indicate quality (if the featured article criteria required every featured article to contain a sentence about the colour red and every featured article did contain such a sentence, that would indicate that the FAC process is following its own rules, but wouldn't say anything about the article quality) and may or may not still hold true nearly 15 years later.
To see whether an article has better quality, we only need to agree in our judgement, no matter whether it is based on subjective criteria or objective criteria. true, but no criteria (objective or subjective) were specified let alone agreed - I asked what objective criteria were being used to back up the claim (and still haven't got an answer) because while we could agree to use subjective criteria to verify the claim I do not agree that such would be both relevant and meaningful. Also, it's worth pointing out that putting the article status in the tag line is not agreeing anything with anybody, or telling anybody anything about the average Wikipedia article. It is claiming that this version of this article is an example of Wikipedia's best work. That is not reliably true. Thryduulf (talk) 20:47, 10 July 2025 (UTC)[reply]
Fair point. I still maintain that you should use whatever subjective criteria you should stand behind, but ORES scores also show quality. I strongly disagree that subjective criteria are meaningless. I'm sure we all know the meaning in having "good" article quality, and that is of course subjective. It is meaningless to ask for objective assessments of something being better if we can agree using other criteria that something is better much faster, especially when the selection of such "objective" criteria is, in itself, subjective. All I'm saying is the near-tautology of subjective assessments being able to produce the subjective assessment of "good quality", while you're saying that only objective assessments will suffice, an unfalsifiable claim if you cannot provide an objective measure. The topic of "this version" is being discussed elsewhere in this RfC. Aaron Liu (talk) 21:33, 10 July 2025 (UTC)[reply]
I'm sympathetic to your main argument, that labeling articles at some point in time is incompatible with the wiki model. But I think your comments rubber stamp of approval from a self-selecting group of non-experts, even GA only requires that citations exist, an article that is two thirds reactions from random celebrities on Twitter is worse than an article one third its length that is purely encyclopaedic, and ratings other than featured are worth less than the paper they are written on are very dismissive of the work content reviewers do and I don't think you should be surprised you're getting piled on here over them. Dan Leonard (talk • contribs)19:54, 10 July 2025 (UTC)[reply]
What pile on? It's about even in terms of people who (broadly) agree with me and people who (broadly) disagree. The final statement you quote from me is entirely a matter of opinion, we can agree to disagree regarding that. The quote about reactions relates only to article length not being a reliable metric of quality, I don't understand why you think that is dismissive of the work of content reviewers unless they do regard length as a reliable indicator of article quality? (If they do, that's definitely a black mark for the process).
By calling the result a "rubber stamp", the first one is an opinion. For the claim to be factual, you would need to show FAC does not disagree with its insiders, which is verily false. Aaron Liu (talk) 21:34, 10 July 2025 (UTC)[reply]
I'm really disheartened by the comments decrying the content review process here. IAWW's review of an article I wrote was one of the most enjoyable editing experiences I've had on here, and it was very comprehensive and involved a text-source integrity check. Reducing his work to a rubber-stamp without citation checks is insulting. Dan Leonard (talk • contribs)20:30, 11 July 2025 (UTC)[reply]
There are GA reviews like that, but there are also GA reviews like Talk:I-No/GA1. As Stepwise stated below, promoting articles to GA status only requires one person's approval. Some1 (talk) 20:40, 11 July 2025 (UTC)[reply]
Is there a problem with the subject article that makes you think the review was conducted improperly? It obviously has fewer BLP and political considerations than the one I cited so the prose does not need to be checked as closely. Regardless, it still gets a topicon with a "this is a good article" tooltip so I don't see why the modified tagline would be so extreme an addition. Dan Leonard (talk • contribs)20:51, 11 July 2025 (UTC)[reply]
I don't believe the review was conducted "improperly", but find the differences in the lengths and comprehensiveness of the two GA reviews quite jarring. Some1 (talk) 23:21, 11 July 2025 (UTC)[reply]
The first two are factual statements Ha ha. No they aren't. I'm not sure why you're taking this opportunity to complain about editors who want quality articles, but your second statements is clearly false. Perhaps you should look at the GA criteria, which are loose but set a baseline of acceptable quality content, before making clearly and egregiously incorrect statements, which at this point is approaching disinformation. Cremastra (talk) 01:22, 11 July 2025 (UTC)[reply]
Discussion elsewhere has identified that the GA criteria I was reading and the GA criteria that are actually applied are different, so while I thought that was factual it turns out that it wasn't. The first is though.
I'm not complaining about editors wanting quality content - far from it - what I'm saying is that article rating labels are not a reliable guide to the quality of the current version of an article. I'm also saying that the FA and GA criteria used to award those labels are not a guarantee that the version of the article reviewed is one of Wikipedia's best, just that it meets those criteria. If you think that is an attack on editors then you haven't been reading what I've actually written. Thryduulf (talk) 01:35, 11 July 2025 (UTC)[reply]
I mean, the GA criteria excerpts the part from the "how the GA criteria should be applied" guideline on spot checks being minimum, so you could say it is in the GA criteria you were reading. That is an interesting state of affairs indeed. Aaron Liu (talk) 15:01, 11 July 2025 (UTC)[reply]
If the GA criteria as written differ from the GA criteria as applied in practice, then we definitely shouldn't be proclaiming GA status in the tagline. Overloading the ordinary word "good" with an insider meaning is confusing enough. Expecting that people will read a set of criteria and then a further guideline to figure out what "good" is supposed to mean is... not really practical. Stepwise Continuous Dysfunction (talk) 18:41, 11 July 2025 (UTC)[reply]
Maybe you know that that's the only relevant part of that guideline. Maybe someone would figure that out after reading both. But it's not at all clear. Every time a reader has to investigate a behind-the-scenes Wikipedia page to understand what something means, we've failed. Stepwise Continuous Dysfunction (talk) 19:52, 11 July 2025 (UTC)[reply]
We have explanatory essays that go in depth on the specific meanings of everything. That does not undermine the meaning of what is explained at all. The criteria by itself sufficiently explain what is expected of GoodArticles. Aaron Liu (talk) 20:00, 11 July 2025 (UTC)[reply]
That's a pretty serious claim you're making. Do you have any objective evidence that the process is this fundamentally useless? None of the criteria for good and featured articles, after all, are not length, so if they don't measure anything else, that's pretty serious. You think the process is so bad that even checking the sources doesn't increase the average accuracy, that it's just a rubber stamp? This all sounds pretty unbelievable to me, but maybe I'm wrong. If this is all true, why do you think the average participant of this RfC is ignorant of it? Dege31 (talk) 17:57, 10 July 2025 (UTC)[reply]
I do not have evidence for a claim I am not making. Checking the cited sources do verify the content they claim to verify is extremely valuable but only a small part of what is required to get the FA badge, not required for any other rating (even GA only requires that citations exist) [see later discussion, it turns out the documented GA criteria I based that comment on do not match the GA criteria that are actually applied. Thryduulf (talk) 01:42, 11 July 2025 (UTC)], and something that can be done completely independently of the FA process. My claim is that FA status does not reliably communicate anything useful to readers about the current version of the article. The current version of the article might be better than average, even one of Wikipedia's best, but it might even be below average now. Thryduulf (talk) 19:16, 10 July 2025 (UTC)[reply]
That's no longer true. Funnily, the new GA checks are in some sense more strict that the FA criteria on WP:TSI. All GANs require spot checks (since 2023 or so), whereas only newer FA nominators have their articles spot checked. —Femke 🐦 (talk) 20:27, 10 July 2025 (UTC)[reply]
That's good to know, but if we can't even reliably communicate what the GA criteria actually are to editors who know that "good article" is jargon then it is even less useful information for readers than I previously thought it was. Thryduulf (talk) 21:16, 10 July 2025 (UTC)[reply]
The proposed destination for the modified tagline is Wikipedia:Good articles, which states (emphasis mine)
A good article (GA) is a Wikipedia article that meets a core set of editorial standards, the good article criteria, passing through the good article nomination process successfully. They are well-written, contain factually accurate and verifiable information, are broad in coverage, neutral in point of view, stable, and illustrated, where possible, by relevant images with suitable copyright licenses.
Also, I think you may have missed footnote 3 in WP:GACR6, which states "at a minimum, check that the sources used are reliable ... and that those you can access support the content of the article". I think these adequately explain to readers that a good article has had its sources checked for verifiability. Dan Leonard (talk • contribs)21:23, 10 July 2025 (UTC)[reply]
Right, I see what you mean now. I understand the concern, but personally, with the (re)activation of the good & featured article review process, I don't personally think it's as critical. Would you also support removing the good article and featured article topicons, by the same logic? Dege31 (talk) 23:03, 10 July 2025 (UTC)[reply]
It's not something that bothers me enough to propose myself, especially as some people seem rather attached to them, but I would probably support if someone else were to propose it. Thryduulf (talk) 23:08, 10 July 2025 (UTC)[reply]
I've struck with a note as it's slightly more complicated than simply being incorrect, see discussion subsquent to my comment about why I made that statement. Thryduulf (talk) 01:42, 11 July 2025 (UTC)[reply]
There is no requirement for writers or reviewers to be subject matter experts so there is no guarantee it is any more or less accurate than any other cited article. Yes, exactly. Schizophrenia is a Featured Article with almost 4k edits since its FA review back on May 2, 2011 (14 years ago). Who knows if that article is still accurate or up-to-date, but because of its FA status, readers will blindly trust that the article and its content are accurate. Some1 (talk) 18:39, 11 July 2025 (UTC)[reply]
Oppose Per Thryduulf - GA particularly has little to do with quality and is more of an Wikipedia:Esperanza-like process to promote the participants. The GA talkpage currently has an RFC to enforce Quid pro Quo reviewing of articles and to reduce the standard of reviews. GA reviews are in many cases completely subjective and amount to little more than "I like this" or "I don't like this".Nigel Ish (talk) 17:23, 10 July 2025 (UTC)[reply]
It is a bit subjective, yes, but there are detailed criteria on how articles should be evaluated for GA. GAs that don't meet the criteria go through GAR. The linked "reduce the standard of reviews" discussion is not about reducing the standard of reviews but about reviewers who make additional comments beyond the standard of the reviews. And the quid prop quo proposal RfC is currently met with a swarm of opposition. Aaron Liu (talk) 17:48, 10 July 2025 (UTC)[reply]
At this point this discussion discussion has devolved into GA-bashing by people who apparently don't want any kind of quality control and for who any recognition of hard work is evidence of a social clique. Cremastra (talk) 01:18, 11 July 2025 (UTC)[reply]
I don't think I'd ever noticed the tagline before, but I've definitely seen the topicons. The fact that they're coloured images makes them more noticeable than small italics, but having linked text in the tagline would probably cover some of that deficit. Maybe I've overestimating the readers, but because "good article" doesn't have a standard colloquial meaning, I think that if a reader did notice that the tagline said that, there wouldn't be any standard meaning to assume. If they wanted to know what it meant, they might click the link and learn more about the internal processes of the encyclopedia. If they didn't, they wouldn't walk away with any wrong assumptions. "Featured article" is a little different - assumptions could be made - but the meaning could vary wildly. Maybe this so-called "featured" article was chosen at random somehow to feature on the main page at some point in the past. I think the topicons make the meaning clearer, appearing to be badges the article has achieved somehow, but I don't think the additional text in the tagline would harm the project (outside of potential technical burden). 207.11.240.2 (talk) 08:59, 10 July 2025 (UTC)[reply]
Oppose - because as written, the tagline is misleading. For instance, the first example listed: London Beer Flood. When you go to the talk page, it is clearly tagged as: London Beer Flood is a featured article; it (or a previous version of it) has been identified as one of the best articles produced by the Wikipedia community. So it is misleading to imply to our readers that the current version they are reading is the FA version.Isaidnoway(talk)18:28, 10 July 2025 (UTC)[reply]
I think that'd be a good thing. If an article's current state is incongruent with what one expects from being "one of the best articles", readers would be alerted to raise the issues somewhere. This can help ensure quality within FAs. Also, <pedantic>, the banner you quoted says the article or a previous version met the definition, and that the article is a featured article, not that it might be a previous version that is the featured article instead of the current version. Aaron Liu (talk) 20:00, 10 July 2025 (UTC)[reply]
No, what it means is that the FA article could be the current form or an earlier, possibly different, iteration of the article could be the FA. For instance, Michael Jackson was promoted to FA status 17 years ago, and on July 8, 2025, there was a brief edit war over a cleanup tag placed in a section. The article had 16,000+ pageviews that day, so how many readers (hundreds, thousands) read a version (current version) that didn't meet the criteria for a FA, and we shouldn't expect for our readers to try and hunt for a previous version that actually meets the FA criteria. I mean, if we are going to say to our readers with a tagline - the current version you are reading meets the criteria for a FA, but yet editors are squabbling over content that may not reflect what the cited sources actually say, then yeah, go ahead with a misleading tagline.Isaidnoway(talk)23:54, 10 July 2025 (UTC)[reply]
The article was still designated FA. The definition of FA is that the article or a previous version was reviewed and met criteria, not necessarily the current version even if the FA-designation is the status quo. I'm not saying we should expect our readers to hunt for that previous version; I'm saying that increasing this prominence invites readers who realize the incongruence to challenge. I don't see how the cleanup tag changes anything in your argument's favor here. Aaron Liu (talk) 00:47, 11 July 2025 (UTC)[reply]
invites readers who realize the incongruence to challenge the vast majority of readers will not recognise the incongruence, but will blindly trust that an article that proclaims to be top quality is top quality, even if it contains blatant misinformation. Even editors with years of experience working with featured articles will not always see an incongruence for topics (or even topic areas) they are not familiar with. Thryduulf (talk) 00:55, 11 July 2025 (UTC)[reply]
Sorry I missed this reply earlier, but I'm struggling to understand what relevance a banner has to anything in my comment? Thryduulf (talk) 18:58, 22 July 2025 (UTC)[reply]
It's alright. I was referring to maintenance tags. If the maintenance tags that are about eight times larger than the tagline (on desktop; on mobile it's probably gonna be like 3x) and colored in alarming ways say the article contains misinformation, the reader will believe the article contains misinformation over the tagline that says "featured article". Aaron Liu (talk) 19:15, 22 July 2025 (UTC)[reply]
True, but only a very small portion of current revisions (at the time any given reader loads the article) of good or featured articles that are not of that quality (due to vandalism, gradual degradation, changing standards, link rot, POV-pushing, real-world changes, editing disputes, etc, etc, etc) have maintenance tags. Thryduulf (talk) 20:25, 22 July 2025 (UTC)[reply]
I guess that's where we diverge. I have not seen any GA/FAs that have problems untagged for a significant amount of time. Would you give me a post-Coldwell example? Aaron Liu (talk) 20:33, 22 July 2025 (UTC)[reply]
I don't know what you mean by "post-Coldwell" but see everything just before it was sent to FAR/GAR, every vandalised revision, etc. It doesn't matter how long the page is below the standard, it matters that whenever someone views a revision that is below standard (for whatever reason) the tag would be misleading. Sometimes only in a very minor way, other times in extremely major ways and of course everything in between. Thryduulf (talk) 10:09, 23 July 2025 (UTC)[reply]
You have a point. Still, I think this deserves a trial to see if the increased visibility will bring more articles to maintenance categories or GAR and thus counter the issue you describe. Aaron Liu (talk) 17:59, 1 August 2025 (UTC)[reply]
The article had 16,000+ pageviews that day, so how many readers (hundreds, thousands) read a version (current version) that didn't meet the criteria for a FA - Kind of beside the point, but an article doesn't automatically change from "FA quality" to "not FA quality" just because there's a tag (and similarly for GA). Instead it's usually one of two situations:
The tag was justified, and therefore the article was already not up to FA/GA standard beforehand. It can be resolved, or the article brought to WP:FAR/WP:GAR if issues are pervasive enough.
The tag was not justified, nd therefore it changes nothing about the article's rating.
Though, I can't argue with the fact that the current version of an article that previously passed an FAC or a GAN may not necessarily be up to standard. That's why substandard articles are (and should be) listed at FAR or GAR. Epicgenius (talk) 03:43, 11 July 2025 (UTC)[reply]
On MJ's article, if you or I ran across the maintenance tag, sure, we would know it could be either one of the two situations you listed, but would our readers? If this prominent tagline had been in place on MJ, bragging about this is one of our very best articles, and a reader scrolls down to a section that is tagged with - the content you are about to read may not reflect what the cited sources actually say, they are going to walk away scratching their heads, thinking, this is their very best? Of course, this situation would hold true with the FA icon already present, but I just don't see how adding this prominent tagline (more bragging) is a benefit to our readers, when it has the potential to be misleading. It's bad enough these unnecessary icons are already on the page.Isaidnoway(talk)06:50, 11 July 2025 (UTC)[reply]
Oppose, per a combination of Ahecht's arguments (scope creep: rating is not what the tagline is for; and "a good article from Wikipedia" sounds bizarre for everybody not familiar with the technical meaning) and Thryduulf's (rating status is far too unreliable for such a highlighting to be responsible). I may add that both arguments apply particularly strongly to "good articles", for which I would oppose this proposal very strongly. Fut.Perf.☼19:20, 10 July 2025 (UTC)[reply]
Support - clearly sensible, not least as readers on mobiles do not get to see the FA or GA icons. Whatever the merits or demerits of the GAN and FAC procedures, these articles do have a defined level of quality and it's helpful for readers to know that. Chiswick Chap (talk) 08:56, 11 July 2025 (UTC)[reply]
these articles do have a defined level of quality. No they don't. A previous version of the article was assess as having a defined level of quality, but there is no guarantee that the version of the article the tagline is displayed on bears any resemblance to that version. Thryduulf (talk) 09:53, 11 July 2025 (UTC)[reply]
This is just one of the things where Wikipedia works better in practice than in theory. "There is no guarantee that any of this is true" is correct for all of Wikipedia, yet it is highly trusted and extremely widely used. —Kusma (talk) 10:11, 11 July 2025 (UTC)[reply]
Indeed, but that's completely different to a prominent banner saying "this is our best work" with links saying that our best work has been verified etc, being placed on articles that are anything but our best work. While many people seeing a page that has been very obviously vandalised will realise that it has been vandalised, not everybody will and the more subtle the vandalism (or POV pushing, etc) the fewer people will know not to take the statement at face value. Doubly so if the article's POV has been slanted towards a POV the reader happens to share. Thryduulf (talk) 10:26, 11 July 2025 (UTC)[reply]
Thryduulf, I think you overestimate the amount of change that quality articles, especially FAs, see after their review. Ovalipes catharus has some small additions of references and phrasing tweaks ([1]) since it was promoted in January. Malicious edits would be reverted, which leaves potentially problematic edits down to POV-pushing, addition of inaccurate information, and bad writing. These would probably all be caught by the person who brought it to FA in the first place. Cremastra (talk) 15:00, 11 July 2025 (UTC)[reply]
All those mallicious edits, etc were current revisions before reversion. The person who brought it to FA is not watching it 24/7. Also, changes accumulate over time God of War III was promoted to FA in February 2015, it has since undergone significant changes. Is it still FA quality? I have absolutely no idea. Do I trust a 10-year-old rating? if yes, then it's misleading if I happen to have viewed it one of the many reverted revisions was current. If no, then it's pointless. Thryduulf (talk) 15:48, 11 July 2025 (UTC)[reply]
Well, yeah, these intervening 9 years count for almost half of Taylor Swift's career. I wouldn't expect any article, let alone Swift's article, to remain unchanged in that time period.As for the article being bloated and outdated, that is less relevant to the topic currently at hand (mentioning FA status in the tagline) and more like a WP:FAR issue. – Epicgenius (talk) 19:58, 11 July 2025 (UTC)[reply]
The fact that there are complaints about the featured article(s) just illustrates that these article ratings are subjective, provide little meaningful information to readers, and don't need to be given more prominence (and in this case, by modifying taglines, of all things). Some1 (talk) 20:17, 11 July 2025 (UTC)[reply]
these article ratings are subjective, There are literally objective criteria (WP:GACR, WP:FACR) that are used to evaluate articles for GA or FA status. So no, it isn't a subjective rating. The fact that there are complaints about the article just mean that people have opinions. These may indicate that the article doesn't meet the criteria. They may also be unjustified, however, as Dan Leonard indicates below. – Epicgenius (talk) 21:51, 11 July 2025 (UTC)[reply]
I think it's fallacious to imply that just because an article has reached GA or FA status, it shouldn't undergo changes. It's one thing for an article to be modified significantly after its promotion to FA or GA status. Sometimes, this is even required in order for an article to keep its rating, especially for articles about people who are alive or things that still exist.It's another thing entirely for these changes to have significantly degraded the quality of the article, but even a small number of changes by a small number of editors can degrade an article's quality. In short, quantity of changes != quality of changes. – Epicgenius (talk) 16:27, 11 July 2025 (UTC)[reply]
Whatever fallaciousness you might infer, I was not implying that just because an article has reached GA or FA status, it shouldn't undergo changes. I was responding to I think you overestimate the amount of change, began by saying many changes, and didn't discuss their quality or materiality. NebY (talk) 16:50, 11 July 2025 (UTC)[reply]
Whatever fallaciousness you might infer, I was not implying that just because an article has reached GA or FA status, it shouldn't undergo changes. If I was mis-attributing that to you, then I apologize. I was speaking primarily in the context of Thryduulf's comment; they claimed that there is no guarantee that the version of the article the tagline is displayed on bears any resemblance to that version. Which may very well be true, but that comment also implied that articles have to remain more-or-less static after their promotion, which is not the case. I was replying to your comment about the number of changes to selected GAs/FAs because I was trying to convey the fact that a large number of changes may not necessarily be an indicator of an article's decline in quality. It can be an indication of such a deterioration of quality, but this can also be done by one or few editors who remove large parts of an article. – Epicgenius (talk) 17:19, 11 July 2025 (UTC)[reply]
I wasn't saying that articles have to remain more-or-less static after promotion, and I'm not sure how you read that into my comment. I'm simply saying that because articles are not static, a previous version being assigned a quality rating is not a reliable indicator of the current quality of the article. It might be that there have been a thousand changes but no material change, it could be that there have been ten changes and the article is substantially different (which could mean it is worse, better or about the same quality). It is this changing nature that means the assessments are not a reliable indicator of the quality of the version displayed, so we should not be proclaiming that something is an example of our best work when we have absolutely no idea whether it is or isn't. Thryduulf (talk) 17:50, 11 July 2025 (UTC)[reply]
Support mildly increasing the visibility of our assessment processes. We should also strengthen GAR and FAR to ensure the designations remain meaningful. —Kusma (talk) 10:18, 11 July 2025 (UTC)[reply]
I think they are meaningful. I am using the opportunity to assert that FAR and GAR are important in ensuring that the designations are meaningful. We do not have to improve all processes to perfection before considering something like the present proposal. —Kusma (talk) 19:37, 11 July 2025 (UTC)[reply]
Not only is the term "good article" overloading the ordinary word "good", but also, getting that status for an article really only requires the approval of one person. We shouldn't make that look more official than it is.
It also requires the article surviving challenges to the good article status and the reviewer being in good standing.Why not give it a try and see if readers are confused? A lot of people here doubt they will be. Aaron Liu (talk) 19:09, 11 July 2025 (UTC)[reply]
A similar number of people have made very strong arguments that readers will be confused and/or actively mislead. Why should we dismiss those concerns just because some experienced editors with detailed knowledge of the procedures vaugely hope that readers will understand that when shove jargon in their face they will understand both what it means and also that it might not actually mean that? Thryduulf (talk) 19:47, 11 July 2025 (UTC)[reply]
Nobody is claiming readers will know right away what these terms mean, the idea is that they'll be clearer than the status quo of topicons. Czar's story in § Proposal 21: Make GA status more prominent in mainspace shows that readers currently don't understand what these icons mean and actually have serious misunderstandings. A plain-text phrase "good article" or "featured article", with a clear link to these meanings, will hopefully both alleviate confusion and onboard future contributors. Dan Leonard (talk • contribs)20:14, 11 July 2025 (UTC)[reply]
Another comment I somehow missed. If readers have serious misunderstandings about what the jargon means when it is tucked away in a corner and accompanied by a link and/or tooltip to an explanation, why would putting that same jargon front and centre not result in anything other than more readers with serious misunderstandings? Thryduulf (talk) 19:01, 22 July 2025 (UTC)[reply]
My complaint is not about the info being tucked away in a corner but about being obscured by an icon and a tooltip. Tooltip explanations are discouraged by most accessibility guidelines including our own because readers clearly aren’t hovering over them to learn what they mean. Replacing (or here, supplementing) icon-and-tooltip presentation with plaintext and a link is self-evidently clearer and less confusing. Dan Leonard (talk • contribs)19:11, 22 July 2025 (UTC)[reply]
It's not a defined thing but empirical, basically the chances of such reviews ending up at GAR. It's not a real thing besides just having reviews that fit the criteria. Aaron Liu (talk) 20:04, 11 July 2025 (UTC)[reply]
"Good" is a pretty good description, better than "featured" but I think the link will be enough for people to cope. —Kusma (talk) 19:39, 11 July 2025 (UTC)[reply]
If "Good article" actually meant something similar to the non-jargon meaning of "this article is good" then you might have a point. So while I applaud your optimism the evidence of how readers currently interact with Wikipedia suggest to me that it is significantly misplaced. Thryduulf (talk) 19:50, 11 July 2025 (UTC)[reply]
That established Wikipedians can have a good faith disagreement about what the phrase "Good article" means is more than enough evidence that it will mislead some readers. Thryduulf (talk) 22:22, 11 July 2025 (UTC)[reply]
Is there a GoodArticle you think is not good? If so, you should nominate that article for Review after starting a discussion about it, no matter how much the article has changed since it was reviewed. By something approaching induction GoodArticles are therefore good articles. I think the only situation where the GoodArticle process fails to produce good articles is if you believe the criteria are not enough to ensure "good", in which case I'd like to know. Aaron Liu (talk) 22:31, 11 July 2025 (UTC)[reply]
My point is that in Wikipedia jargon "Good article" means that a specific revision of an article was judged by one person to meet a set of very specific criteria (that the current version of the article may or may not now meet). To most readers seeing a tagline saying "good article" would indicate that the version of the article has been assessed to be "good" and thus can be relied upon to be neutral, accurate, and at least reasonably comprehensive and thus by implication articles that are not "good" are "bad" and cannot be said to be poses any of those qualities. While it is true that the current version of many Good Articles is indeed good there are also current versions of Good Articles that are not good. There are also plenty of articles where the current version is accurate, neutral and comprehensive but which are not Good Articles simply because nobody has formally assessed it. Thryduulf (talk) 22:45, 11 July 2025 (UTC)[reply]
there are also current versions of Good Articles that are not good
(besides what I say in the linked reply) Those are also very few.
by implication articles that are not "good" are "bad"
I don't think that's true. The rest are just unreviewed articles, and even not meeting the standard for good doesn't necessarily mean bad. I think this also addresses your last point. Aaron Liu (talk) 01:48, 12 July 2025 (UTC)[reply]
As an editor familiar with the review process you know that "not good" doesn't mean "bad". The same is not true of the average reader who does not know that "Good article" is jargon, let alone what it means.
Re current versions of Good Articles. If you mean stable versions then there probably are relatively few (but still a large number), however when you include every version that is current at some point it is much larger. There is no way for the casual reader to know whether the version they are seeing is the good version or the vandalised version. Thryduulf (talk) 02:41, 12 July 2025 (UTC)[reply]
Sorry, I meant that I doubt that's what readers'll interpret it is. WIthout "good" it's still "articles". I mean at any moment. I don't think aggregating anyone that had any version that was bad is meaningful. And it's not like RC patrol's gone handicapped. Aaron Liu (talk) 02:47, 12 July 2025 (UTC)[reply]
It does mean something similar, and I am indeed optimistic that our readers know that blue text means a link that can be clicked on to obtain clarification. Most of our readers are not using Wikipedia or the WWW for the first time. —Kusma (talk) 23:11, 11 July 2025 (UTC)[reply]
I'd argue that "featured" is a less confusing term than "good" in this context, since it's less generic and actually conveys the connotation that the articles were selected via some process. Stepwise Continuous Dysfunction (talk) 19:55, 11 July 2025 (UTC)[reply]
The proposal has indeed come up frequently. Some editors see increased visibility of article ratings as an improvement to the status quo, but if there is a problem with the status quo it is that it overstates the reliability and importance of article ratings, which is not something making them more prominent can solve (indeed rather the opposite). Thryduulf (talk) 17:26, 12 July 2025 (UTC)[reply]
Support as I like the idea of better signaling which of our articles have met a minimum bar for quality. GAs aren't perfect articles, but we show worse articles on the main page every day. Ed[talk][OMT]20:03, 12 July 2025 (UTC)[reply]
An embellishment is defined as ornamental, or decorative detail, to make something more attractive. Is it to your belief that the good article, and featured article processes are likewise embellishments? Dege31 (talk) 01:54, 17 July 2025 (UTC)[reply]
Support – for the same reasons I proposed more prominent topicons in 2021. As Wikipedia matures, I think it's increasingly important we 1) focus on raising the quality of existing content and 2) help readers learn how to effectively use and understand the varying quality of articles, especially in the current information landscape. Drawing attention to the main ways we review articles, however flawed they are, is a step in the right direction for both these aims, and by raising awareness of peer review processes it might help improve them. I think it would be even better if the "featured article" or "good article" text linked directly to the article's most recent FAC/FAR/GAN, but perhaps that wouldn't be feasible. I understand the valid concerns about the varying quality of FA/GA status articles, but ultimately it is better than no review at all, I trust that most readers understand Wikipedia is not infallible, and the more we focus on peer reviews the better for raising quality and trust in the project. Jr8825 • Talk12:26, 15 July 2025 (UTC)[reply]
Support ways to make GA/FA status more prominently visible, including this proposal. My preferred way of going about it would be to have the icon next to the article title, kind of like how (on desktop) the icon is displayed next to the name of another language in the "Languages" list if that version of the article is good or featured (see e.g. Jupiter). Several of the opposing comments sound more to me like arguments to abolish GA/FA (or at minimum the icons) entirely, which I am fairly certain is a proposition that would be overwhelmingly rejected by the community. TompaDompa (talk) 23:35, 16 July 2025 (UTC)[reply]
Oppose, per several of the arguments made above. If this is intended to bring in more editor participants, we're sending a confusing signal to a group very few of whom are the right target. The terms are opaque and likely to be misleading to our readers, since there are plenty of GAs that are no longer good quality but have not been reassessed yet. Fewer FAs are in that state, but there are some. If the GAR and FAR processes were working as well as we'd like, this would be less of an issue, but we haven't solved that problem yet. Future Perfect at Sunrise puts the case against well. Mike Christie (talk - contribs - library) 00:20, 17 July 2025 (UTC)[reply]
Support trying this out, at least for some period. I believe Wikipedia should try to make GAs/FAs more visible to the common reader, and I think more awareness may also lead to more GAR and FAR helpers. ALittleClass (talk) 02:38, 18 July 2025 (UTC)[reply]
Is there some other language we could rename GAs/FAs to to connote that they have undergone more thorough review and received a higher quality assessment than other articles, without implying that all our other articles are trash? I do think the concept of article quality assessments for articles should be fairly intuitive to most readers, Wikipedia being a work in progress and all that. Sdkbtalk15:25, 21 July 2025 (UTC)[reply]
The problem is that, at least as things currently stand, articles that there is a set of articles that have been assessed as being of a particular quality at some point in their history and a set of articles that are currently that standard of quality. The two sets overlap but are not close to being the same with both high quality articles that have not been assessed as such and articles that were formerly high quality no longer being. Unless and until the significant majority of articles tagged as being of "X" quality currently (at the time any given reader loads the page) are that quality and the significant majority of articles that are that quality are tagged as such any tagline will be inherently misleading in some way. Thryduulf (talk) 15:39, 21 July 2025 (UTC)[reply]
You could get around some of that by something that says something like like "on <date>, a <version> this article was assessed as being <quality standard>" with a link to that quality standard and an explanation that the current version may or may not be of that quality standard, but (a) that isn't a tagline, and (b) is not what is being proposed here. Thryduulf (talk) 15:42, 21 July 2025 (UTC)[reply]
We cannot let the perfect be the enemy of the good. Part of our responsibility of editing the encyclopedia is to ensure that articles that should be FAs/GAs are nominated and that those designated as such but that no longer meet the standards are delisted. It's no different than having an article tagged as needing more citations: That tag was placed at a specific point in time, and represents our judgement at that time, but it doesn't imply that other articles don't also need more citations. And if, as the article evolves, it acquires enough citations, it's our responsibility to remove it. The articles tagged as needing more citations will never correspond precisely to those that actually do need more citations the most, but we still use the notice since it's a good enough (albeit imperfect) indicator. Ditto for quality article assessments. Sdkbtalk16:00, 21 July 2025 (UTC)[reply]
It isn't an integral part of the encyclopaedia, it's an optional status symbol. Just because that status symbol motivates some editors to improve articles doesn't make it integral - look at the countless articles that get improved in other ways and for other reasons. Thryduulf (talk) 17:20, 21 July 2025 (UTC)[reply]
We cannot let the perfect be the enemy of the good. That's true in the abstract, but this proposal is not good - indeed for the reasons explained in detail multiple times it's actually harmfully bad. The comparison to dated tags that specify in detail what the problem is/was and do not proclaim or imply anything about other articles misses the mark in multiple different ways. Thryduulf (talk) 17:16, 21 July 2025 (UTC)[reply]
The German Wikipedia approaches this by putting de:Template:Exzellent at the bottom of pages with a notice "This article was added to the list of excellent articles on day (DD_MM_YYYY) in [permanent_link_to_version this version]. See today's featured article, de:Abreise König Wilhelms I. zur Armee am 31. Juli 1870. That also has serious downsides, for example de:J. R. R. Tolkien links to a version from 2004 that does not have a single inline citation and does not actually meet modern day standards. "The 2004 version of this article met the 2004 version of WP:FA?" is not particularly useful information. —Kusma (talk) 10:22, 31 July 2025 (UTC)[reply]
The latter issue might be mitigated by only displaying the message if the article was promoted or had its status confirmed within the last N years. The German Wikipedia has (or at least had) a very different approach regarding inline citation to the English Wikipedia, so the status of articles is not trivially comparable between the languages. Thryduulf (talk) 10:39, 31 July 2025 (UTC)[reply]
It might not do that to you, but it does to me and Andy and assuredly will to anyone who doesn't know that "good article" is jargon (regardless of whether there is or isn't a link). Thryduulf (talk) 09:54, 22 July 2025 (UTC)[reply]
I really don't think we have enough GAs or FAs that the absence of a "good article" tagline will be so widely noticed. I don't think the proposal will be as impactful as you and Andy seem to think, but we won't find out if we don't try it. —Kusma (talk) 10:01, 22 July 2025 (UTC)[reply]
Whereas I believe that the multiple severe downsides (not just this one) combined with the extremely low benefits that will come even if successful mean that even a trial will come up with a significantly negative benefit:cost analysis. Thryduulf (talk) 12:10, 22 July 2025 (UTC)[reply]
Is there any evidence of severe (!) downsides outside of hypothetical Wikipedia editor discussions? We already have the topicons. Why don't we see any inklings of these severe downsides, if they are a realistic concern? At least I'm not aware of any. Dege31 (talk) 04:08, 23 July 2025 (UTC)[reply]
The top icons are decoration that most people ignore and have to be investigated (but even then others have pointed out that they cause misconceptions). A tagline is extremely prominent and makes a bold statement to everybody reading the article (even more so if it is expanded to mobile) meaning the problems caused by misleading statements will be very significantly amplified. Thryduulf (talk) 10:13, 23 July 2025 (UTC)[reply]
Oppose I'd not noticed the taglines before and suppose that I've been blanking them as fluff. The specific suggestion of calling out FAs and GAs seems quite marginal as we don't have many of them. Why not give an assessment of the article when it has a lesser status too? It might be more useful to warn readers that an article is a stub or just C-class or whatever.
And I don't like the idea that this will be forced onto the mobile interface as space is at a premium in that. See the example of Dopamine which I took a snapshot of for another discussion recently. Notice that this doesn't manage to get all of the first sentence of the article onto the first screenful. Adding a tagline would push it off completely.
So, the idea of giving readers an assessment of the article upfront has merit but the implementation needs work.
Re: the specific suggestion of calling out FAs and GAs seems quite marginal as we don't have many of them: the good and featured systems have broad community support and are currently represented on articles already by topicons with tooltips like "this is a good article". This proposal is intended only as an extension of that as shown above in § Background (tagline), and so any extensions beyond what is already very highly supported would be undue. Also re: it might be more useful to warn readers that an article is a stub or just C-class or whatever a "warning" would violate WP:NODISCLAIMERS. Dan Leonard (talk • contribs)18:30, 22 July 2025 (UTC)[reply]
WP:NODISCLAIMERS is irrelevant. Every page has a general disclaimer per WP:GENDIS. And there's already a well-established set of tags to show that that an article is a stub. The trouble is that these appear at the bottom of articles where the reader won't see them until it's too late. It's better to make the status of an article clear to the reader at the outset, so that they have this context as they are reading it. Andrew🐉(talk) 22:28, 22 July 2025 (UTC)[reply]
You literally used the word "warning", though. That's exactly what NODISCLAIMERS is about. This proposal is about promoting articles that have reached a high standard of quality that the community is proud to present to readers. Giving a "warning" to readers that an article is "just C-class" is the opposite. That could be considered at a later point but should not be part of this proposal. Dan Leonard (talk • contribs)22:32, 22 July 2025 (UTC)[reply]
B-class and lower can be changed by anyone anytime. I'm not confident in the quality assurance for those.Also, if you want more mobile screen space, dismiss that banner. Aaron Liu (talk) 18:42, 22 July 2025 (UTC)[reply]
Re:mobile accessibility- that might be easy for you or me, but the "dismiss banner" button is *really* tiny on mobile. I don't think my mother could reliably click it, for example, she just doesn't have the eyesight anymore. GreenLipstickLesbian💌🦋18:51, 22 July 2025 (UTC)[reply]
Fair point. But that's an issue we could easily solve by enlarging the button; we don't even need to file a task for it as we can simply do it through an interface admin. Aaron Liu (talk) 19:11, 22 July 2025 (UTC)[reply]
Support. This would be a positive for readers and editors alike. Perhaps people opposed should also advocate for removing WP:TMVs from articles—they link to editor-focused stuff as well. We have giant banners when an article has issues, but a small link when it has been determined to meet a set of standards is a bridge too far? Heartfox (talk) 03:15, 23 July 2025 (UTC)[reply]
Weak oppose I like tag lines to be the same for every article. If it is to be tried, I would propose to do it for the Featured articles first. Readers are more likely to have heard from them from the main page. The articles are higher quality and more editors are needed to certify them as featured. The Good article slogan sounds worse. Ideally it should be measured whether readers understanding of the featured article process increases after the change. Rolluik (talk) 09:02, 31 July 2025 (UTC)[reply]
I think it makes sense to do that first. I have not myself participated in a FA or Good article process. I'm aware of the cleanup efforts for older FA and Good articles (and also that it seems to have stalled for the featured articles). I think that new featured and good articles seem mostly similar quality wise for most readers but as an editor I see that the featured ones have mostly better sources. Keep in mind that English is not my native language, so I care less about spelling mistakes and similar issues. I'm sceptical though that changing the tag line will have a big effect on anything (increase of FAC's, Good article nominations, FAR's, readers knowledge of quality assesments...). I don't think it will be detrimental either. Maybe we should give people an easy link to the version of the article that was reviewed in the tag or near the tag (not just on the talk page). I see no real problem with giving newer FA's that were on the main page this tag line (If we were confident enough to put it on the main page...). Rolluik (talk) 20:16, 4 August 2025 (UTC)[reply]
Strong oppose. I'm sorry; I really believe the article status should be more visible, but I can't stand the tagline being modified in such a drastic way. In my opinion, the benefits of keeping the current tagline intact and free of links and other notices outweigh the benefits of letting the user know an article is good/featured. I'm all in for doing the same thing while keeping the precious tagline intact! --FaviFake (talk) 20:54, 31 July 2025 (UTC)[reply]
I'm guessing like the "this article is a stub" messages but at the top, perhaps. It could also include the most recent date it was promoted/survived review as recommended above. Aaron Liu (talk) 21:06, 31 July 2025 (UTC)[reply]
This seems fine to me, I mostly opposed the tag changing due to esthetic reasons. The only downside is banner blindness (and scrolling distance) but most readers seem fine with maintenance templates and similar; and featured/good articles are not supposed to have those. Rolluik (talk) 20:27, 4 August 2025 (UTC)[reply]
Strong oppose to "Good article" tagline, weak support to the rest. Our neutral point of view is one of Wikipedia's core content policies. It's not up to us to declare to the reader our articles are "good". I believe our quality article assessments are quite helpful for Wikipedia's editors and I'm happy with the current topicons (which [being icons] don't declare anything and link to the full explanation), but outright writing at the beginning that this is "A good article from Wikipedia" is not neutral. I'm attracted to the argument that this could increase new editor attraction but I don't see that outweighing the opposers' concerns. I also don't believe making the tagline more verbose in response would be a satisfactory workaround. Sophocrat (talk) 02:19, 1 August 2025 (UTC)[reply]
Note: I'm much more open to doing this only with Featured articles and lists due to the arguments others have made. As a bonus "Featured" is much more neutral than "Good". Sophocrat (talk) 02:24, 1 August 2025 (UTC)[reply]
Could you specify how exactly this proposal violates
All encyclopedic content on Wikipedia must be written from a neutral point of view (NPOV), which means representing fairly, proportionately, and, as far as possible, without editorial bias, all the significant views that have been published by reliable sources on a topic.
i.e. neutral point of view? An article assessment is not "view[s] that have been published... on a topic".
As Aaron Liu says I refer to the principle. We don't declare ourselves to be a reliable source (even though we strive to be!) to the reader. In articles we should avoid stating opinions as facts as our aim is to inform the reader, not make judgements for them. That said, I think I would support doing this only with Featured articles and lists as they have the highest quality standards and because the descriptor "Featured" is just a fact (as those are indeed prominent articles) whereas "Good" is a judgement from us (however well-founded). Sophocrat (talk) 21:36, 1 August 2025 (UTC)[reply]
Support I've been using a userscript that does something similar and it is quite useful. (Wish I could tell you which one, but nothing in my common.js is obviously it.) Loki (talk) 16:20, 4 August 2025 (UTC)[reply]
Weak oppose: On the one hand, it'd be simple, as there's already a gadget for this, it could just require flipping it to default-on. On the other hand, there's already a gadget for this - and more pointedly, we already put an icon in the upper right of the page header for FAs, FLs, and GAs, so having this be by-default seems slightly redundant. - The BushrangerOne ping only05:02, 8 August 2025 (UTC)[reply]
we already put an icon in the upper right of the page this proposal is mostly inspired by Czar's story here, detailing how readers are not just unaware of the meaning of icon-based indicators but actively misguided. I think text and a transparent hyperlink is a clearer presentation. Dan Leonard (talk • contribs)05:22, 8 August 2025 (UTC)[reply]
Strong support – I'll take anything at this point... time and time again we've needed increased visibility for our best articles (most readers don't even know these distinctions exist, let alone the difference between featured vs good)... – Aza24 (talk)21:31, 8 August 2025 (UTC)[reply]
Weak oppose My greatest concern is that most Wiki users have no idea about what this actually means, especially since "good" can be interpreted much differently. In my experience what most readers care about (almost exclusively) is factual correctness, being of course part of being a GA it's not the entire story and (hopefully) most other articles meet that standard too. Highlighting it there makes me feel like it's implying that other articles are inherently "bad" and cannot be trusted. I could however see it work out with FA as not being Featured misses such an implication in my opinion. Squawk7700 (talk) 23:08, 8 August 2025 (UTC)[reply]
Oppose. As much as I'm a fan of the FA and GA processes and like to gain such honours for the articles I've written, it's basically a Wikipedia internal concept. I don't think we should be publishing such things in a single line tag which is used across the Web, and lacks context for readers of they're not familiar with the processes. While in general, on average such articles are likely to be better written and more reliable than others, it's not a universal truth and it would be a mistake to imply in the tagine such articles can be implicitly trusted. — Amakuru (talk) 00:30, 15 August 2025 (UTC)[reply]
Don't FAs convey the quality status in a way that is already explicitly designed for public consumption, being featured on the home page, and indicated with an icon on pages for users that are not logged in? The GA process too implies a basic quality threshold has been reached, and is already displayed publicly. That some GAs or FAs fail to live up to their status is surely more of a clarion call for the review process than it is a reason to view the baby as no better than the bathwater. With the icons already present, the proposed changes aren't making anything visible that isn't already. As always, anyone interested in the context of the terms could navigate through to the explanations, as they can now with the icons. Iskandar323 (talk) 10:43, 15 August 2025 (UTC)[reply]
Amakuru isn't the only one to comment that the f.a. system is an internal process. Another problem is that this isn't available on mobile. Finally: I'm not aware if there was consensus or a discussion to implement the topicons. Logoshimpo (talk) 05:25, 22 August 2025 (UTC)[reply]
I would repeal the decision made for those two discussions since too many arguments here against this selfreference has been brought up. Logoshimpo (talk) 04:13, 23 August 2025 (UTC)[reply]
Support This small change would go a long way to help our readers see when an article is of a higher quality. The icons aren't immediately clear, but text with a link would be much more overt. The more we provide average readers with even a basic understanding of the back-end processes, which is sadly quite uncommon, the better I think they'll be able to engage with our articles. --Grnrchst (talk) 13:05, 20 August 2025 (UTC)[reply]
This small change would go a long way to help our readers see when an article is of a higher quality. no, it will let readers see when some previous version of an article was assessed as being "featured" or "good" and will imply that the revision of the article they are seeing is of that standard (it may or may not be) and that articles not so tagged are inferior (they might be, they might also be superior). I'm also not understanding why a reader should need to know anything about back-end processes in order to engage with articles? Thryduulf (talk) 13:37, 20 August 2025 (UTC)[reply]
I answered your question here — in short, it's helpful info for readers to know to what extent they should trust an article. And in the replies below, myself and others talk about the process by which editors monitor and delist GAs/FAs that no longer meet standards. It's not perfect, of course, but readers know that we are an imperfect work in progress. It's no different than maintenance tags, which also represent a judgement about an article from a particular point in time, which may be outdated but which we try to update if so. This proposal is also no different in kind from our current practice of displaying GA/FA topicons, which we do for the exact same reasons, the only distinction being that a tagline has a slightly higher chance of actually being noticed. Sdkbtalk14:43, 20 August 2025 (UTC)[reply]
This is a great point – the worst possible downside of this process is still no worse than maintenance tags (giant reader-facing banners that serve as an often outdated disclaimer) and is effectively the same as the long-accepted reader-visible icons in the top right. The average case, though, is an improvement that helps readers: re why a reader should need to know anything about back-end processes in order to engage with articles, readers have long shown a vague understanding that Wikipedia might have some kind of content review process, but are still very much unaware of how we make these judgments. See for instance the many first-timer comments like "who wrote this crap it's all wrong" that have plagued article talk pages since the beginning of the project. Linking curious readers directly to Wikipedia:Featured articles tells them exactly how articles are reviewed and provides an overview of our content policies. Dan Leonard (talk • contribs)15:03, 20 August 2025 (UTC)[reply]
Strongly oppose. There are already things like the featured star in the top right that indicate this information. Displaying it so prominently gives an undeserved sense of importance to a rather unimportant classification. In addition, the WIkipedia branding has been on articles for twenty years OmegaAOL (talk) 00:29, 5 September 2025 (UTC)[reply]
Support, conditional on there not being technical issues with the implementation (e.g. I think it would be bad to add a WP:EXPENSIVE call to every single page and then downwardly adjust the limit for actual page content to 499 but this seems surmountable to me) jp×g🗯️23:11, 13 September 2025 (UTC)[reply]
@Dan Leonard: Can we not also have FLC, FAC, GAN? like so...
{{#switch:{{#invoke:Page assessment raw|get_class|page={{FULLPAGENAME}}|project=Project-independent assessment}}| FA = A ''[[Wikipedia:Featured articles (linked from tagline)|featured article]]'' from
| FL = A ''[[Wikipedia:Featured lists (linked from tagline)|featured list]]'' from
| GA = A ''[[Wikipedia:Good articles (linked from tagline)|good article]]'' from
| FAC = A ''[[Wikipedia:Featured article candidates/PAGENAME/archive#|featured article candidate]]'' from
| FLC = A ''[[Wikipedia:Featured list candidates/PAGENAME/archive#|featured list candidate]]'' from
| GAN = A ''[[Talk:PAGENAME/GA#|good article nominee]]'' from
| From
}} Wikipedia, the free encyclopedia
OR, if getting the archive# and GA# might be tedious, we could simply say...
Can we not also have FLC, FAC, GAN? like so...
{{#switch:{{#invoke:Page assessment raw|get_class|page={{FULLPAGENAME}}|project=Project-independent assessment}}| FA = A ''[[Wikipedia:Featured articles (linked from tagline)|featured article]]'' from
| FL = A ''[[Wikipedia:Featured lists (linked from tagline)|featured list]]'' from
| GA = A ''[[Wikipedia:Good articles (linked from tagline)|good article]]'' from
| FAC = A ''[[Wikipedia:Featured article candidates|featured article candidate]]'' from
| FLC = A ''[[Wikipedia:Featured list candidates|featured list candidate]]'' from
| GAN = A ''[[Wikipedia:Good article nominations|good article nominee]]'' from
| From
}} Wikipedia, the free encyclopedia
This RFC proposal is intentionally limited to be an extension to the topicons, which already get enough community support to exist. I disagree with this idea (drive-by junk GANs shouldn't be shown to readers) but also just don't think we should overcomplicate the proposal by going beyond what is already represented by topicons. Dan Leonard (talk • contribs)16:30, 11 July 2025 (UTC)[reply]
Personally, I would better understand those on the oppose side for non-technical reasons, if they could bring in some real data, or real examples, of harm caused by the smaller outreach variants already extant: that is, the topicons, and the talk page assessments. After all, if it is problematic, there should already be evidence. I haven't seen this presented in significant levels. To me, these pitfalls feel remote, and rare. I feel like the potential downsides aren't so big that we can't even do a test run to see how it goes. The large majority of readers read within the confines of the lead paragraphs, so that lessens the potential cumulative impact, too.
Nonetheless, valid concerns. The supporters should also answer: is there will, and capability for scaling up the maintenance of these articles? While this has been ramping up in recent years, this imposes a higher standard.
I've also thought about an idea (this is not a proposal, but fuel for separate discussions, if I, or anyone else, wants to take it further) that maybe takes into account some of the reluctance. It would involve an article losing its reader-facing indicators of GA status or FA status after X years of no review, or Y edits if it's very high activity. That way, there would be a guaranteed minimal level of accountability. Dege31 (talk) 00:38, 12 July 2025 (UTC)[reply]
Eh, it’s a cosmetic change to the website so I think it's fine to balk at it subjectively. It is a shame that my proposal can't have any data or examples of how this would improve reader outreach (although there does seem to be some interesting-looking papers on FAs), as any analysis of such data would only be possible post hoc. If this passes I do hope to do a 30 day postmortem to see how many people click on the statistical redirects and see how many new editors participate in the project pages. Dan Leonard (talk • contribs)00:51, 12 July 2025 (UTC)[reply]
Regarding scaling up the maintenance of these articles, I hope that one side benefit of this passing may be that, if GA/FA status confers additional prominence compared to the status quo, there will be both more incentive for editors to pursue that status for articles that deserve it and more editors noticing/sending to FAR/GAR when an article has that status that does not deserve it. Sdkbtalk04:43, 15 July 2025 (UTC)[reply]
and something similar for FA. Though I do wonder how long after the RfC closers should the discussion be started and whether it should be started even if it succeeds. Aaron Liu (talk) 03:42, 9 August 2025 (UTC)[reply]
I just took a look at de.wiki and their rating system and I'd be more supportive about this proposal if our system would be more similar to theirs because their equivalent to GA "Lesenswerter Artikel" lit. "Article worth reading" reflects the meaning of GA a lot better than "good" in my opinion. But I get that that's unfeasible already because the translation is way too long and sounds a bit awkward. Squawk7700 (talk) 09:26, 9 August 2025 (UTC)[reply]
In addition, quality is not the only factor in what makes an article "worth reading" - a generally terrible article is very much worth reading if it contains a single reliable source that verifies as correct (or not correct) the one claim you are attempting to verify or it unlocks your understanding of whatever it is you are researching. Even a featured article is not worth reading if it doesn't contain the information you are looking for (which could be for many reasons, including not quite being in scope, the article being outdated, a source no longer being available, the relevant portion being removed (with or without consensus) or vandalised, etc). Thryduulf (talk) 16:20, 9 August 2025 (UTC)[reply]
I don't like this as much. While I support any addition of plaintext with wikilinks, the English Wikipedia has long had opposition to emphasizing old revisions. See for instance the massive difference between how pending changes are used between German and English. I think it goes against the wiki model and its core belief that articles always get better over time. Articles should never have prominent links to old revisions. Dan Leonard (talk • contribs)17:18, 9 August 2025 (UTC)[reply]
Could you elaborate on the PendChang difference or the opposition, especially its reason? I think when the wiki model works, it's isn't at all bad to highlight how the article has evolved since it was established to pass a minimum standard. I don't see how the inclusion of the link leads to any default implication that the article got worse. It's just a tool for all the examine, serving the "transparency" part of our ethos. AFAIK even on enwiki, logged-out users see the latest approved revision (an old one) on PendChang-protected pages by default. Aaron Liu (talk) 17:47, 9 August 2025 (UTC)[reply]
The reason we gave GAR and FAR is because classifying an article as GA or FA applies to future revisions, unless and until that status is revoked. It isn't just for the one reviewed version. If it were, we wouldn't still call them GAs and FAs or have processes for removal. Linking to historic revisions can confuse readers, especially when transcluded templates have since been deleted (the deletion of the lang-xx family breaks the first sentence of many historic revisions). Dan Leonard (talk • contribs)22:12, 9 August 2025 (UTC)[reply]
Those are about effectively applying pending changes to every single article. If we can have pending changes protection as it is at enwiki today, I don't see why we can't have a link to an old revision especially when it's labeled as an old revision. I still believe that having a link to the old revision is a useful point of comparison. Anyways, we can drop the revision link if needed. Aaron Liu (talk) 00:37, 10 August 2025 (UTC)[reply]
I referred to enwiki's rejection of dewiki Gesichtete Versionen as an example of the philosophical differences in the wiki model. You can see arguments like incompatible with the basic principles behind wikipedia ... they no longer qualify as a wiki (Erachima), against wiki ethos (IanOfNorwich), creeping implementation of flagged revisions in disguise (S Marshall), 'A rose by any other name?' (OmniArticleEditor). The German Wikipedia's practice of placing a little check mark on specific logged revisions of articles, and presenting them to readers as if they are better than or more reliable than the current one, represents a significant change to the wiki model akin to Citizendium. Articles are assumed to still improve after promotion, and if they decline then they are demoted. Kusma already discussed this above, with the example of a dewiki article that links to an old revision without images or inline citations and has broken template calls (as I mentioned). Dan Leonard (talk • contribs)01:08, 10 August 2025 (UTC)[reply]
Thanks for the ping to this interesting discussion. Although I personally oppose any and all implementation of Pending Changes, vehemently and on a philosophical level, if we are going to have to put up with the awfulness of Pending Changes in our Wiki, then Pending Changes interacts with Good and Featured Articles in complex ways and I'm leery of one-size-fits-all decisions. I think we need to be mindful that Good and Featured Articles are an anomaly -- a holdover from old days of Wikipedia, back when we got to say things like "This is a good article" without having a reliable source for that contention. I think the fact that these assertions are made by Wikipedians rather than by trustworthy sources is highly relevant to the decision of whether to link them.
Some good and featured articles are about things that scholars have largely finished thinking about. If our subject matter is, say, Tropical Depression Ten (2007) then I'm with Aaron Liu. I don't think any massive reevaluation of that topic is likely, so I think we could quite legitimately pick one revision of that article and say, "This! This is the Featured Revision of this Featured Article!" and crystallize it thus for all time.
But other good and featured articles are about things that are still in flux. If our subject matter is, say, DNA nanotechnology then I'm with Dan Leonard. There could be a new discovery that could substantially change the article, at any time, and I would say that to pick one historical revision and imply that it's the platonic ideal of that article isn't the greatest idea I've ever heard.—S MarshallT/C03:08, 10 August 2025 (UTC)[reply]
I remember having discussions about pending changes in the past but with edit filters in place, this wikipedia seems to prefer protection Logoshimpo (talk) 04:20, 10 August 2025 (UTC)[reply]
Thanks. I really don't think having a message like what I mentioned at all implies the old revision being the apex. "became a good article in {revision}" retains the meaning of "everything after this" Dan mentioned.I also find what's extracted from the pending changes discussion weird as applied here. It seems Dan's point is that this is supposed to illustrate opposition to presenting a specific revision with a checkmark, a star, or a plus sign. But that's what we've lived with and supported for a decade and a half, just not at the scale of every single mainspace article. Aaron Liu (talk) 03:46, 11 August 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
RfC: Update messagebox module with new Codex icons
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.A summary of the conclusions reached follows.
The community did not numerically support this change, mostly for aestheic reasons, though some because they did not see a good reason to take any action. Even among those who supported the change, many complained about the broom icon not looking like a broom or otherwise looking bad (and apparently it's not even part of Codex?). The icon colors also did not match the accent colors of amboxes. Ambox.notice option 2 was also disliked, and it is only distinguished from ambox.content by color, which is an accessibility problem.
Arguments in favor cited improvements for dark mode and color-blind users. Perhaps someone will try again with better aestheics and those goals in mind. -- Beland (talk) 07:33, 19 September 2025 (UTC)[reply]
Should the icons in the message box module be updated from the current Ambox ones to the Codex ones? 13:56, 11 August 2025 (UTC)
ambox.style Not Codex, but is the OOUI for style icon...
ambox.notice
Option 2 ambox.notice
The UI on Wikipedia employs Codex components/design tokens/icons (see Special:Version) such as the message system with the Message component. These messages employ CSS and JavaScript, which we can't do yet in wikitext, see T401186. Instead, the icons have to be uploaded to Commons and used from there, as was the case with OOjs/OOUI on other wikis. We missed the 2019 update to OOUI icons unlike MediaWiki's mw:Module:Message box/configuration.
The license for Codex icons is MIT license, the entire package is GNU Public License. I believe the license might be an issue, as I was informed by @Redrose64 that MOS:PDI states one cannot remove image link= for attribution reasons, the only mention of this is this line:
For CC BY-SA, GFDL, or similarly licensed images, blank |alt= and |link= attributes should not be used. It is Wikipedia's policy to link those images for attribution...
Contrary to that point, Codex is already widely implemented on wikipedia.org: Codex icons and its components are used extensively in the Wikimedia ecosystem as its default UI system. So it makes me wonder if GPL is considered similarly licensed images. I don't see us able to maintain a consistent style and appearance (one of Wikimedia's architecture/guiding principles) with the Wikimedia sister projects, if we don't implement this workaround until the Codex-Wikitext extension (T357463) is released. waddie96 ★ (talk)16:39, 10 August 2025 (UTC)[reply]
@Pppery: Style changes are a necessary part of wiki, in order to modernise. For instance, the padlocks for protection changed in order to fit the theme of the wiki better. This is another example of that. —Matrix(!)ping onewhen replying {u - t? - uselessc} 18:33, 10 August 2025 (UTC)[reply]
I note the padlock discussion was about images displayed at 20×20px and significantly depended on changing from using color as the only distinction to including symbols as well for improved accessibility. This discussion relates to images displayed at 40×40px where the existing icons already differ significantly in shape (more so than the proposed replacements!) and are supplemental to the text of the box itself. Anomie⚔16:42, 11 August 2025 (UTC)[reply]
I removed it for your convenience, but I thought it important to point out as at my edit request I would like to make it clear that those icons were to remain the same and not be replaced with any other ones. waddie96 ★ (talk)15:12, 11 August 2025 (UTC)[reply]
Yes, I removed it, as you could see it was in the lower quadrant as simply carried over as it already exists in its latest format. waddie96 ★ (talk)15:11, 11 August 2025 (UTC)[reply]
Copying what I said in the discussion section since its partly the basis of my oppose: The existing designs are clear and easily identifiable, whereas the new Codex designs are thinner and can be confused. Accessibility needs to be taken into account here, not just ILIKEIT and IDONTLIKEIT arguments.Tenshi! (Talk page) 20:59, 14 August 2025 (UTC)[reply]
Strong oppose – it ain't broke. The new icons are also hard to interpret and, well, ugly, as the broom doesn't look like a broom at all, and the other ones follow the hideous trend of having icons look more and more like letters. Let's not make things more complicated for our readers to understand. Cremastra (talk·contribs) 20:05, 10 August 2025 (UTC)[reply]
It looks like the format brush from MS Word, which is just as a good as a broom. In fact I personally abhor that old motherbleeping[Joke] broom. It may have looked good and trendy in 2005 but since my (wiki)birth it's always looked aesthetically absolutely horrible, low-contrast, and at small sizes unsightable. Aaron Liu (talk) 02:50, 11 August 2025 (UTC)[reply]
It looks like the format brush from MS Word thanks, I didn't know that... and I expect many others didn't either. We shouldn't expect our readers to know the icons from a word processor. Also, a brush is not a broom. That looks like a paintbrush. Cremastra (talk·contribs) 14:24, 11 August 2025 (UTC)[reply]
Maybe I shouldn't have mentioned the Word format brush, because the paintbrush has the same although less specific connotations either way. ambox.style is for stylistic issues. Brooms shift that meaning to "cleaning up the lint", while a paintbrush continues the meaning of "lick of paint" or "dressing it up", which is how style tags are resolved. Aaron Liu (talk) 14:29, 11 August 2025 (UTC)[reply]
Oppose Personally, I find these new icons ugly and would rather keep the existing ones. Not going to fight over that though.As far as the license goes, the Expat (MIT) license states The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. For uploaded images, we've interpreted this as meaning that we need the link to the file description page where the notices are "included". The CC BY-SA license has similar language for notices about copyright, license, and attribution. Images included in the software may not be subject to those terms in the same way, since "all copies" might be taken to refer to the distribution of the software as a whole rather than the rendered HTML. Find a lawyer to try to figure that out. Anomie⚔20:58, 10 August 2025 (UTC)[reply]
Although most of those seem like they'd probably be {{PD-simple}} anyway. The stylized paintbrush is the only one that's not a plain shape with a single text character on it. Anomie⚔21:14, 10 August 2025 (UTC)[reply]
After giving it more thought, switching to "oppose" as I find the new icons to be distractingly harsh and soulless. They might be ok at small sizes, but 40×40px in amboxes is not that. Anomie⚔13:44, 12 August 2025 (UTC)[reply]
If feasible, I have Strong Support in Vector 2022 only: V22 uses the Codex design system. I get that a lot of people hate that skin in general too because it wain't broke, but I'm fairly sure these people would not be using Vector 2022. Technical musings: This might be feasible if we output both images and do a conditional-CSS thing to hide one of them, and unlike JS approaches loading CSS will block the display instead of creating a content flash. For when CSS doesn't load, the page would already look broken enough anyways. Otherwise, I support the original proposal We've had a flat site interface design for a very long time; the time to prevent that has long passed. Now, most of our users do see the new flat design. And unlike the last flat-design proposal, there's no possibility for confusion of the content and notice icon among colorblind users (unless we go with Option 2, which I oppose) and the subtle improvements between the OOUI and Codex circle symbols make it actually look good now. Aaron Liu (talk) 02:50, 11 August 2025 (UTC)[reply]
Was there ever really a "time to prevent that"? Paid designers are going to make trendy new designs whether we want them to or not, and their work falls under WP:CONEXCEPT unless they give us the option or we stage a major revolt. Anomie⚔12:19, 11 August 2025 (UTC)[reply]
There's a whole lot more going on there than "flat design". Discussion there seems to have been far more concerned with the functionality changes. Anomie⚔16:32, 11 August 2025 (UTC)[reply]
Support but note that Codex icons are identical to OOUI. There are several reasons behind this including supporting dark mode, ensuring visual consistency and improving accessibility. I came up with this list a while ago to identify icons that can be replaced with flat design with little problem.
I also want to discuss web accessibility and color blindness in this discussion. Here is what the original vs proposed icons potentially look for someone with total color blindness:
Although I'm not sure that just throwing a greyscale filter over the images is really an accurate representation of any real-world color blindness. Anomie⚔22:36, 18 August 2025 (UTC)[reply]
It isn't, but it probably represents the worst case scenario of complete color deficiency. I did put some permalinks to an online non-WMF tool used to check color accessibility below.
Oppose Laughably bad. No reason given for change other than ILIKEIT and a claim that enwiki should change its style to match some other style (why not the reverse?). Johnuniq (talk) 04:41, 11 August 2025 (UTC)[reply]
Support per Aasim. Consistency is good and the new icons look more modern. WP:BROKE is an unconvincing argument given the low effort that is required to make these changes. Sam Walton (talk) 05:41, 11 August 2025 (UTC)[reply]
Oppose the broom change, that new icon is clearly a paintbrush, which brings to mind very different connotations. CMD (talk) 07:55, 11 August 2025 (UTC)[reply]
Oppose per Pppery. We shouldn't radically change appearances without having demonstrated a real benefit. I don't see how icons that blend into text are a benefit -- they are harder to notice and can be skimmed over easier. They're less 'in your face', which is what we want from warning notices. JackFromWisconsin (talk | contribs) 17:22, 11 August 2025 (UTC)[reply]
Support for consistency and style reason. WP:BROKE isn't a good argument if there is a specific issue with the existing icons (in this case, a striking lack of consistency with the surrounding MediaWiki interface). The first option for ambox.notice is preferable as it is closer to the original intent. Chaotic Enby (talk · contribs) 18:37, 11 August 2025 (UTC)[reply]
Regarding the broom being replaced by a paintbrush, I agree that it is a major change in meaning that might not be ideal, and would prefer having a broom icon in a matching style instead. Chaotic Enby (talk · contribs) 18:41, 11 August 2025 (UTC)[reply]
I never noticed the ambiguous meaning. It looked like a broom to me. Must this be another ambiguous image illusion? Aasim (話す) 18:43, 11 August 2025 (UTC)[reply]
Support most, oppose broom. Whatever that is, it's not a broom. But otherwise, consistency is good. More generally, even if nothing is broken, changing icons every 10 years or so helps prevent a website from looking too stale, even if the change is totally cosmetic. The likes of Apple and Google make sure to include some pointless style changes so you can "tell" that this is the latest version of IOS or Android. While we aren't a business, we also don't want to be TOO boringly consistent in style. SnowFire (talk) 19:37, 11 August 2025 (UTC)[reply]
Support. I support switching to Codex when the changes are small, such as in this case, where the icons basically look the same. The whole point of having a design system such as Codex is to get everything on the website to have a unified look, which is more professional and is (usually) easier for maintenance purposes. –Novem Linguae (talk) 21:38, 11 August 2025 (UTC)[reply]
Oppose. We don't want the modernization of icons, which may somewhat be unclear, especially the broom one. We don't need consistencies in icons especially in maintenance templates. Consider use the old icons. Fabvill (Talk to me!)12:07, 12 August 2025 (UTC)[reply]
Support most, oppose broomWP:ILIKEIT arguments seem fine in discussing our aesthetic presentation, but beyond that, Aasim and Aaron Liu have shown that the new icons are clearer than the status quo in dark mode. Disagree with JackFromWisconsin, as the existing icons were not intentionally designed to be garish, nor would the new icons be ignored on a primarily text-based site only using black text and blue hypertext. ViridianPenguin🐧 (💬) 06:39, 13 August 2025 (UTC)[reply]
Support for consistency. Like it or not, Codex is the common design system of all Wikimedia projects so the decision to eventually standardise on it has already been made (and so IMO !votes on the basis of WP:NOTBROKEN are invalid here). Consistent style is an important part of any professional publication and we are no exception. – Joe (talk) 07:05, 13 August 2025 (UTC)[reply]
The mobile website does all sorts of really stupid stuff. Pointing at it hacking in different, often poorer icons doesn't seem like a good argument to me. Anomie⚔10:40, 13 August 2025 (UTC)[reply]
Regardless of how one feels about the new icons on desktop (or the other changes in mobile), the current desktop icons would be rather illegible at the size of the current mobile ones. novovtalkedits04:37, 14 August 2025 (UTC)[reply]
@Graeme Bartlett Have you seen the existing icons in dark mode? Ugly is a understatement for how they look (they have flipped shadows, and are unnecessarily bright). We really should be adopting a stance/culture of "Our interfaces must look good in the light mode, folks who use dark mode will have to just live with it". There are often cases when I do need to switch to the dark mode because of accessibility reasons (eye strain when reading in lower-light conditions) and the iconset that we use has really been a sticking point for me personally when it comes to reading and editing articles. Sohom (talk) 14:32, 13 August 2025 (UTC)[reply]
I'm guessing you mean shouldn't. I really agree that dark mode accessibility is one of the strongest points in favor of the new icons, and something that I'm happy to see the WMF incorporate in their recent design philosophies. That is one of the reasons why "consistency" here is not only about aesthetics, but also about going along with a more widely accessible system. Chaotic Enby (talk · contribs) 16:32, 13 August 2025 (UTC)[reply]
Oppose the replacement for ambox.style and ambox.notice option two, support the rest. Ambox.notice has an "i" for "information", which the first option replicates. The proposed replacement for ambox.style looks more like a paintbrush than a broom, which can be interpreted as "whitewash this" or "cover this up" rather than "clean this up". The first two are fine. I'm actually in the camp of "if it ain't broke," but I also recognize that it appears very unprofessional when there is inconsistency between projects. Since WMF converted to Codex, we should as well. — Jkudlick ⚓ (talk)15:51, 13 August 2025 (UTC)[reply]
Support the red, orange, and the blue i. I agree with the others that the paintbrush is not a broom, and the exclamation point would fit better in other places than a "notice". Izno (talk) 20:55, 14 August 2025 (UTC)[reply]
Support all except for the paintbrush. I also support the "blue i" icon. The current designs are indeed broken in dark mode, which is available to people who are logged out (in other words, the vast majority of readers). Using the Codex designs fixes this actual problem for a large chunk of our target audience. A personal taste for the older icons should not overcome the needs of our readers.
The paintbrush is neither a broom icon nor actually part of Codex, so I guess I am a weak oppose. I think a broom icon would be a great addition to Codex (the ethos of a wiki is to make mistakes easy to fix, rather than hard to make, so a broom is clearly helpful). Or even if there is some reason not to add to Codex, perhaps we could request a broom in Codex style from the relevant WMF team and/or volunteers. If we find a Codex-style, dark-mode-compatible broom, I support using that. HouseBlaster (talk • he/they)01:43, 15 August 2025 (UTC)[reply]
Oppose as I believe the current icons are easier to recognize. I also believe they are satisfactorily consistent: they are all clean SVGs with pleasant gradients and thin contours. They don't benefit much from being more consistent than that. And that's without considering that thing about I believe the license might be an issue. Sophocrat (talk) 01:47, 15 August 2025 (UTC)[reply]
I do not consider the current icons consistent. The only similarity is all being skeuomorphic/frutiger aero, which is as much as saying Monet and Van Gogh have a consistent artstyle. The speedy-deletion sign and the broom are clearly from very different styles of design compared to the other two signs. (Also, the bigger benefit is consistency with the interface.) Aaron Liu (talk) 16:44, 19 August 2025 (UTC)[reply]
Oppose per 2A0E:1D47:9085:D200:EDFB:7835:FC79:242D. It will create more inconsistencies as some templates (I'm thinking of maintenance templates on articles, which will be seen by readers) will have flat icons while others will have non-flat icons. If someone finds suitable alternatives for the most commonly used ones (such as the book icon in {{More citations needed}}), support all per Joe but weak oppose the broom as it barely looks like a broom at all. OutsideNormality (talk) 01:55, 15 August 2025 (UTC)[reply]
For {{More citations needed}}, would be the most visually similar one, although or convey better the idea of looking for references in my opinion. Any of those would be fine to me.The icon in {{Unreliable sources}} also matches well with (or , if we take the rectangle being searched to represent the article rather than the sources). is closer visually, but could be more confusing as "question mark in rectangle" is a very common symbol and makes it less clear that the rectangle represents the article.{{Disputed}} easily gets .For {{Speculation}}, Codex sadly doesn't have a crystal ball yet (although I could make one in the Codex style if needed). The closest "magic woo" icons (and that's a stretch) are or .{{Current}} and {{Recentism}} can get different colors of .{{Globalize}} gets . Chaotic Enby (talk · contribs) 18:19, 15 August 2025 (UTC)[reply]
Support, those icon suggestions are nice. Sometimes simplicity is better, especially when they're paired with text-heavy boxes. The flat style, though "in vogue" at the moment, is definately easier on the eyes. qcne(talk)20:07, 18 August 2025 (UTC)[reply]
Support Aesthetics are important for webpage design. The Wikipedia logo isn't usually seen on mobile, so it doesn't clash with the codex icons. –LaundryPizza03 (dc̄) 16:44, 21 August 2025 (UTC)[reply]
if the old icons don't work in dark mode Define "work". Except for the broom's shadow, they're all perfectly visible, some just think they're ugly or maybe not dark enough in dark mode. As for accessibility, I note the icons are mainly decorative anyway, the real meaning is communicated by the text of the message box. Anomie⚔13:00, 24 August 2025 (UTC)[reply]
The seven types of ambox (four of which would have their icons changed) are not just decorative. The type is chosen not on aesthetics but is based on the type of issue that the template describes. As with what I mentioned for the unblock icons below, these icons are very functional for a quick glean of this information on the type of ambox, so one can decide faster whether to ignore it, for one. Aaron Liu (talk) 22:31, 25 August 2025 (UTC)[reply]
Oppose I'm unable to easily read the symbols in the new icons due to my astigmatism; they look like plain vertical lines (|) inside the shapes. On the current icons, it's very easy for me to tell that these are exclamation points and a letter i. On dark mode, I'm unable to see the paintbrush handle without squinting whereas the current icon is easily identifiable as a broom. 3df (talk) 05:49, 2 September 2025 (UTC)[reply]
Strongly oppose. 1: I like skeuomorphism and think the older icons look better. 2: Ignoring my aesthetic preferences, the older icons are easier to distinguish. 3: Your icon change will ruin the messagebox for everybody who uses MonoBook or Vector 2010. OmegaAOL (talk) 00:37, 5 September 2025 (UTC)[reply]
As with OOUI buttons, the (upcoming) Codex-Wikitext feature mentioned should or should be easily modifiable to allow alternative icon sets per skin. Aaron Liu (talk) 00:52, 5 September 2025 (UTC)[reply]
Support Important part of UI design is consistency. Although if we wait long enough, I guess skeumorphism will come back so we will avoid having to do anything.. Galobtter (talk) 23:28, 7 September 2025 (UTC)[reply]
Partial oppose. An even more important part of design is making sure that images look like what they're intended to represent. As someone noted above, the new broom icon doesn't look like a broom: it makes me think of a backscratcher or paintbrush, but not a broom. I like all the older images better, but aside from the broom, I don't have a strong opinion; since most of the non-broom icons use exclamation points, the exception looks like ¡ rather than "i", but that's still not a huge problem. Regarding the colourblindness issue — I'm a bit red-green colourblind, but I don't have any difficulty with the current images. Why would this matter anyway? Each icon has a different character or background shape, so even if you're a total monochromat, you can easily distinguish them. Nyttend (talk) 08:03, 11 September 2025 (UTC)[reply]
Strong support for all suggested (including the Qs below) per nom, Aasim, Chaotic Enby, Waddie. as a Vector 2020 and dark mode user, these are a nice change. the whole skin has been very visually appealing and accessible for a user with bad vision, and these would add to it. the (constructive, see below) criticisms for some of the design icons should be taken into account for reworking them as needed. I would also say that the boxes themselves would warrant a bit of change. Juwan (talk) 00:08, 13 September 2025 (UTC)[reply]
I think ILIKEIT and IDONTLIKEIT are going to be themes in this discussion. WP:ILIKEIT and WP:IDONTLIKEIT really apply to deletion discussions mostly but they are also bad arguments for other kinds of discussions. I brought up my reasons for preferring it including dark mode compatibility, consistency with other aspects of Wikipedia (including the protection icons), and accessibility (although I did suggest one change). Arguments on principles on visual design are probably going to be themes similar to the Vector 2022 RfCs. Aasim (話す) 05:37, 11 August 2025 (UTC)[reply]
@Cremastra, Pppery, and Tenshi Hinanawi: my strongest critique would be that much of the discussion relies on subjective perception and analogy rather than user research or accessibility testing. I'm also just going to point out the fallacies in the invalid construction of your argument as comments expressing support or opposition if they are going to represent the community’s position should surely be reasonably substantiated with relevant reasoning, evidence, or examples.
Status quo bias: assumes that because something currently works (or isn’t perceived as broken), it shouldn’t be changed — without addressing whether improvement might be possible or beneficial.
Subjective aesthetic judgment without objective criteria: While personal preferences are valid in casual conversation, as arguments they are weak unless tied to objective usability, accessibility, or design standards.
Personal dislike: This is an ad hominem toward the design, not the proposal. The focus shifts to personal feelings rather than the proposal's functional merits.
False analogy: like the format brush from MS Word comparison assumes visual similarity equals functional equivalence, without establishing that readers will interpret the icon the same way
Overgeneralisation: We shouldn’t expect our readers to know the icons from a word processor. This may or may not be true, but it’s stated as a certainty without evidence of actual reader familiarity. waddie96 ★ (talk)15:34, 11 August 2025 (UTC)[reply]
The existing designs are clear and easily identifiable, whereas the new Codex designs are thinner and can be confused. Accessibility needs to be taken into account here, not just ILIKEIT and IDONTLIKEIT arguments. Tenshi! (Talk page) 15:45, 11 August 2025 (UTC)[reply]
Yes, this. The new icons are less accessible; it is your argument that amounts to ILIKEIT. By definition flat designs with less information are more likely to be confusing, and they should be avoided. Why not have the cleanup icon be File:Broom (PSF).jpg, perhaps rotated 45°? Cremastra (talk·contribs) 16:40, 11 August 2025 (UTC)[reply]
I don't see how they are more easily confused with each other. The key to distinguishing between icons is their differences, not their total size. The added skeumorphic lighting in the Tango icons cancel out as that's something all of them have.I don't see how .content and .notice are any less distinguishable; they have the same amount of differences: color, "direction" of symbol, and an alteration to make it clear the i is a letter. You could I guess say the serif is bolder than just shortening the height but they're already far past the differentiation-baseline for me to dent my subjective differentiation index. Same thing goes for the speedy icon's white inside vs no white inside. Aaron Liu (talk) 03:01, 12 August 2025 (UTC)[reply]
I don't think the new icons are less accessible. What I do know is that the original icons don't look good in dark mode. See this as an example:
Weird, that doesn't happen even logged-out for me in the light/dark comparisons; they only happen in the gallery from Waddie for me. If you click on Extension:Graph and enable dark mode, does the same thing happen? Aaron Liu (talk) 02:56, 12 August 2025 (UTC)[reply]
I'm using the default dark mode gadget. In light mode there is no square around the Dark mode column, but it's there in actual dark mode. CMD (talk) 07:09, 12 August 2025 (UTC)[reply]
Not sure what you mean by "default gadget". I was talking about Vector 2022's built-in dark mode from the appearance menu, which looks like an incognito icon when hidden. Aaron Liu (talk) 14:14, 12 August 2025 (UTC)[reply]
They don't flip, even though IMHO they should. This is because MW renders SVG icons in the backend and then throws out a simple PNG image that is supposed to be appropriately scaled. It would be better to instead have the SVG code spat out to render in the frontend, then stuff like changing the color based on the current theme can work. Aasim (話す) 06:48, 12 August 2025 (UTC)[reply]
Why not have the cleanup icon be: @Cremastra As always, your argument synthesis is derogoratory and your 'discussions' are 'win' or 'lose' mentality—not constructive. waddie96 ★ (talk)02:17, 12 August 2025 (UTC)[reply]
I wanted to add on to the icon you recommended for the "broom".
First, it is grayscale, not colored in. Second, it becomes invisible in dark mode. I am pretty sure the second one can be fixed by inverting the broom, but it does not solve the first problem. This is not being displayed on a grayscale CRT, it is being displayed on a variety of screens from television sets (because yes some people hook their computer to their 4K OLED TV) to HDR monitors to smartphone screens. It should look good on almost all of them. And all the problems as well with the image not being an SVG.
Here is how that icon looks like for the record: . Not good. Even rotating it 45 degrees doesn't really fix the issue:
A few people above are mentioning "consistency with the MediaWiki interface", by which I'm guessing they specifically mean Vector2022. Personally, I really don't see it. Looking at some random pages in a private-browsing window, I see a whole lot of nothing to be consistent with other than the Wikipedia logo in the corner, which these new icons don't match, and whatever we have in the articles themselves, which these new icons don't really match either. I guess "whole lot of nothing" kind of goes with "harsh and soulless", which is the vibe I get from the new codex icons. Anomie⚔22:27, 11 August 2025 (UTC)[reply]
Having a triple-dot menu, the entire top bar, the sticky top bar and user dropdown (only accessible when logged-in for some reason), everything to do with discussion subscriptions, everything to do with notifications, Recent Changes/the Watchlist, the graphic for empty talk pages and editing onboarding, the mentor dashboard, newcomer homepage.... There's also built-in dialog boxes somewhere, beyond the reference previews gadget we have. Aaron Liu (talk) 00:56, 12 August 2025 (UTC)[reply]
So, a bunch of stuff that only shows up when logged in, and changing icons to match logged-in Vector2022 would cause them to mismatch for logged-in users of classic Vector, Monobook, and other skins. I can't say I find that terribly convincing. Anomie⚔13:39, 12 August 2025 (UTC)[reply]
Nope, as a classic Vector user, I also have Codex/OOUI icons in the interface, like the notification icons. If anything, that change is also an improvement for us. Chaotic Enby (talk · contribs) 13:57, 12 August 2025 (UTC)[reply]
Firstly, I'm surprised that no one has commented on what I said about making them only show on V22. Secondly, a quarter of this shows when logged out too. In fact, the Codex warning sign for "You are not logged in" only shows up for logged out users. It's also in all of the editing interfaces. Aaron Liu (talk) 14:18, 12 August 2025 (UTC)[reply]
IMO embedding both and hiding one or the other with CSS is not a very clean solution. Looking at various other places you mention, I again find that 20×20px or smaller icons have a different impact compared to 40×40px ambox icons. I also find the Codex warning sign for "You are not logged in" is oddly truncated, FYI. Anomie⚔14:40, 12 August 2025 (UTC)[reply]
Instead of a brush or a broom, maybe we could go for a color version of the more inviting (File:Codex icon edit.svg), to indicate issues that editors can solve? Other ones I like are (a generic tag icon) or (yes, it's officially the recent changes icon, but simple icons like this are very polysemic). Chaotic Enby (talk · contribs) 14:12, 12 August 2025 (UTC)[reply]
The flat colors make it obvious that these icons don’t match the existing accent colors for the amboxes. I also feel like the orange and red colors are too similar. Amazing how we got it right years ago and yet the WMF tries to reinvent the wheel and makes a simple color-clash mistake like this. —pythoncoder (talk | contribs)17:05, 13 August 2025 (UTC)[reply]
Yeah sorry I don't have time to read all 56 comments since I was last here but:
If we use the icons for accessibility we should ideally not use the current yellow ones, we should use the prescribed yellow warning color at Codex color palette for consistency and for sufficient W2A contrast. See Codex color tokens and Color palette.
Current yellow icon on yellow background would be: #FFCC33 bg/ #FDF2D5 (1.352:1) Fail for large and regular text (AA) Fail for large and regular text (AAA) Fail for UI components and graphical objects (non-text)
The @color-icon-warning on @background-color-warning is: var(--color-icon-warning) bg/ var(--background-color-warning)
(Pretty much trying to match .mbox CSS with @cdx-message CSS
I wonder if a manual of style for icons and colors in templates would be a good idea. It probably can detail stuff like accessibility and icon color, and how maybe we should not solely use color to distinguish function of an icon, as people can have varying degrees of color blindness. WP:DONTFIXIT seems to be one theme in this discussion, as is WP:ILIKEIT and WP:IDONTLIKEIT. But I am not seeing that many arguments based on web accessibility. Sure some icons are supplementary, but even supplementary icons and shapes help people who have dyslexia or do not speak English (like in this Exit sign), people who are color blind (like this traffic light), people who are skimming and want a general idea of what is happening on a specific page, and people in all of these mentioned categories and more. We do have MOS:COLOR and MOS:ICON, but they largely pertain to the use of colors and icons in the prose of articles and not in maintenance templates, user interface, etc. Aasim (話す) 01:48, 20 August 2025 (UTC)[reply]
I think there would be a significant amount of clashing/bikeshedding within the context of such a discussion and a high amount of community inertia (not to mention that we would clash with the WMF Codex/Design team as and when such a style guide becomes out of date). Sohom (talk) 02:46, 20 August 2025 (UTC)[reply]
Yeah the law of triviality is definitely a problem, and I think what kind of icons we are using (flat or skeumorphic or nuvola or old) aren't really worth debating unless if there is an accessibility or readability concern. When I first suggested converting to flat icons all the way back in 2020, the RfC was ended early for not being prime for an RfC, and the only few other times I even entertained the idea in the idea lab, they never went to fruition as RfCs (most recent idea lab post I can find is here). The only things I really do think would be worth including somewhere either on this new style page or in the existing MoS is:
When icons are used as indicators in templates, template colors should match the color of such icons
For accessibility reasons other properties such as icon shape should be used to distinguish between sets of icons part of template series such as the user warning and block templates.
Otherwise, I don't think there is a good reason to revert changes to icons. I think this issue will become less contentious if and when there is an in-built extension tag or parser function that automatically loads the SVG code for any one of the growing icons in the Codex library (and they can potentially be expanded and customized for each design system). Aasim (話す) 04:18, 20 August 2025 (UTC)[reply]
FYI this section did not actually ping anyone because you have to sign your post in the same edit as you added the user names, and you can't ping more than 50 users in the same edit. I really don't think pinging 76 people was necessary here, though. * Pppery *it has begun...14:51, 11 August 2025 (UTC)[reply]
The plan with ambox is to eventually somehow replace the dark-mode background color with --background-color-neutral-subtle (#202122) while fixing the ribbon colors to the left. Aaron Liu (talk) 17:00, 21 August 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
@Beland: Hi Beland, I'd like to revisit the outcome you summarised for the first part of the discussion, as I have concerns about its accuracy.
One significant issue is the phrasing did not numerically support this change. That raises red flags, since consensus on Wikipedia is not determined by counting heads or votes, but by evaluating the strength and policy-basis of arguments. Using numerical language gives the impression of a vote tally rather than a consensus assessment.
Another point is that substantial weight was given to opinions that did not fully engage with the matter at hand. The issue wasn't solely about the paintbrush icon; it was about updating the maintenance template icons to the Codex set as a whole. While one particular icon drew attention, that should not overshadow the broader question of whether the set itself is suitable.
In addition, the arguments summarised in your close seem too narrow compared to the range that was actually presented. To restate more fully:
Arguments in favour highlighted alignment with MediaWiki’s Codex system (already used in Vector 2022 and mobile), improved accessibility (especially in dark mode), and the benefits of a modern, consistent, and professional appearance.
Arguments against emphasised that the current icons are already clear, recognisable, and functional, while some of the Codex icons were seen as thinner, harder to distinguish, or less meaningful (e.g., the broom vs. paintbrush concern). Opponents also cited potential confusion, lack of demonstrated functional gain, and inconsistency with templates not yet updated.
Intermediate views generally supported Codex adoption in principle, but suggested deferring until more appropriate icons (such as a Codex broom) are available, or until all maintenance templates can be updated together for consistency.
From my reading, the discussion pointed to no clear consensus. There is interest in Codex alignment for long-term consistency, but significant opposition remains regarding specific icon choices and the timing of implementation.
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.A summary of the conclusions reached follows.
→ (but modified to have an X inside rather than an !)
→ (but modified to have an X inside rather than an !)
Lv. 4im:
Light mode
Dark mode
→
→
Lv. 4:
Light mode
Dark mode
→
→
Lv. 3:
Light mode
Dark mode
→
→
Lv. 2:
Light mode
Dark mode
→
→
Lv. 1:
Light mode
Dark mode
→
→
(note there is no noticeable difference between the appearance Codex and OOUI icons, and Codex icons already exist for this or can be created easily for this) Aasim (話す) 19:08, 18 August 2025 (UTC)[reply]
As proposer: support, but it would be preferable to wait until it is possible to use Codex icons inline with wikitext before updating. The old templates have this problem of file links to files that can no longer be deleted or updated because they are used so much. There definitely are major accessibility changes to this, including different shapes for Lv. 1 and Lv. 2 warnings, as well as better appearance on dark mode. It's also useful to consider what this looks like with a total color blindness filter I found online.Aasim (話す) 19:08, 18 August 2025 (UTC)[reply]
Okay, I tried creating a template to illustrate the problem clearer. Web accessibility is not an option:
As above, I don't know why you're showing the icons at a quarter the size (by area) that they're actually used at. Anomie⚔22:40, 18 August 2025 (UTC)[reply]
Support. I feel like the new icons are better in dark mode – while flat design and contrast are indeed something to consider, the lighting and white text on the old ones are more problematic in my opinion. I can easily make the cross-inside-octogon since we already have . Chaotic Enby (talk · contribs) 19:34, 18 August 2025 (UTC)[reply]
Oppose for the same reasons I oppose the main proposal: these are harder to recognize (particularly in dark mode) and I haven't seen any convincing arguments why more consistency with the surrounding UI is a good thing. Plus judging by the monochromacy table someone made they seem less accessible. Sophocrat (talk) 20:33, 18 August 2025 (UTC)[reply]
Oppose Same reasoning as above. I note that any argument of "accessibility" is going to be a red herring, since the icons are decorative anyway and the real message of any box is communicated by the text, not the icon. Anomie⚔22:40, 18 August 2025 (UTC)[reply]
Can you clarify which icons you are having trouble distinguishing? For the old icons having an i for both level 1 and level 2 warnings with the only difference being color is problematic from an accessibility standpoint. Aasim (話す) 22:14, 19 August 2025 (UTC)[reply]
Can you maybe reference which icons you are having trouble distinguishing? It would be helpful maybe to help improve this proposal and reach an adequate compromise. Aasim (話す) 03:11, 20 August 2025 (UTC)[reply]
I assume CMD is talking about the solid red clock, the solid red hand, the solid red stop sign, and the dark orange triangle. I agree they are more difficult to distinguish. Cremastra (talk·contribs) 03:27, 20 August 2025 (UTC)[reply]
I think the issue is at the size I am showing them in my half-baked attempt at showing what it might look like for color blind users, even I am struggling to make out the details. But then I am not hunched against my laptop, I have a proper mechanical keyboard in front of my laptop keyboard on my desk that essentially doubles the minimum viewing distance when on a docking station.
Maybe the solution would be to make the icons bigger, so they are 50px rather than 25px, but then that interferes with line placement, necessitating table substitution rather than inline substitution. Aasim (話す) 03:33, 20 August 2025 (UTC)[reply]
A bit ago I was experimenting with unifying all the uw templates in Template:Uw/sandbox and I realized in one of my iterations I used both and in combination with . The icons do look nicer on colored backgrounds that match them, but otherwise when not on a matching color it can look a little hideous. But a hideous color that is still legible is better than un-dark-modeable colors like the white clock that cannot be changed to black as inverting would mess up the color of the hands and the stop X. And on Vector 2022 and Minerva it seems broken, I don't know why. Aasim (話す) 03:27, 20 August 2025 (UTC)[reply]
I've boldly replaced the lv 1 and 3 OOUI icons with the Codex icons, since there is in fact meaningful difference for those. Also, for an ambitious change, one might consider {{cdx-message}}. Aaron Liu (talk) 18:37, 20 August 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Question 3: Should we update the unblock templates to use Codex/OOUI icons?
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.A summary of the conclusions reached follows.
Yes. 2:1 in favor. Supporters found not distinguishing based on color alone is an accessibility improvement, and the new shapes more clearly communicate their meanings. -- Beland (talk) 07:44, 19 September 2025 (UTC)[reply]
Unblock:
Light mode
Dark mode
→
→
Unblock on hold:
Light mode
Dark mode
→
→
Unblock declined:
Light mode
Dark mode
→
→
Unblock accepted:
Light mode
Dark mode
→
→
(note there is no noticeable difference between the appearance Codex and OOUI icons, and Codex icons already exist for this or can be created easily for this) Aasim (話す) 19:08, 18 August 2025 (UTC)[reply]
As proposer: support. The unblock templates have some of the most inaccessible icons IMHO, with only a recolored clock used for all of them. It's also useful to consider what this looks like with a total color blindness filter I found online. Aasim (話す) 19:08, 18 August 2025 (UTC)[reply]
Okay, I tried creating a template to illustrate the problem clearer. Web accessibility is not an option:
As above, I don't know why you're showing the icons at a quarter the size (by area) that they're actually used at. Anomie⚔22:39, 18 August 2025 (UTC)[reply]
The key aspect here is color in this case. If I were to show the icons any bigger I am afraid they would overflow off the screen especially on mobile devices. Feel free to adjust the size of the icons if you think it is necessary for this discussion. Aasim (話す) 23:32, 18 August 2025 (UTC)[reply]
Size is relevant for at least some contrast ratios – for example, WCAG AA and AAA are not the same for regular-sized and large-sized text. However, in this specific case, it is clear that the useful information in the current icons only depends on the colors, which is problematic from an accessibility standpoint. Chaotic Enby (talk · contribs) 23:53, 18 August 2025 (UTC)[reply]
Oppose Same reasoning as above. I note that any argument of "accessibility" is going to be a red herring, since the icons are decorative anyway and the real message of any box is communicated by the text, not the icon. Anomie⚔22:39, 18 August 2025 (UTC)[reply]
Umm... Isn't dyslexia definitely going to cause problems for some editors? Colored icons are much faster to identify than walls of text. Also not everyone who gets blocked here speaks English and an X implies denied and a check implies accepted (in the Western world at least). Aasim (話す) 23:41, 18 August 2025 (UTC)[reply]
Further, if this passes only because people don't like the colored clocks being the only difference, I'd rather see them changed to something like the existing clocks with symbols superimposed in the lower-right corner, like we do for various other things, than the brutalist Codex style icons. Anomie⚔12:26, 25 August 2025 (UTC)[reply]
Support This one has a far better rationale since the original icons are not accessibly differentiable from each other. Sure, there's always other text that tells you that status, but the icon is not purely decorative at all. As a colorsighted person, it serves as a great, fast way to see the status of an unblock request that you can quickly scroll by. On the other hand, you'd have to slow down and read text. There also isn't an easy way to guess that yellow means on hold and blue means it's new without knowing that beforehand. Aaron Liu (talk) 03:12, 19 August 2025 (UTC)[reply]
Not a fan of a big red X for unblock declined, it feels a lot more like rubbing it in than a palette-swapped clock. CMD (talk) 04:31, 19 August 2025 (UTC)[reply]
Maybe check/thumbs down would be more appropriate. I haven't made any changes because I mostly abandoned development of the idea until someone else brought it up. My opinions have not changed that much in the past seven years, but I did see that since we are discussing some of the icons right now we can discuss the others as well.
It really shouldn't matter too much what icons we use and what icons we don't use (apart from accessibility and UI/UX testing), we are an encyclopedia after all, and icons are not part of encyclopedia proper. I don't even know how consensus was achieved for the nuvola icons in the first place, nor the information icons that have not changed since 2005 at the earliest. Aasim (話す) 05:48, 19 August 2025 (UTC)[reply]
I know Phabricator uses ✅ and 👎. That is where I got the idea from. But yeah I see how it can be debatable whether it is appropriate. Aasim (話す) 00:41, 20 August 2025 (UTC)[reply]
Support per Aaron Liu and also common sense: a / makes more sense than coloured clocks for declined/accepted requests respectively, and a pause icon is also more symbolic of "on hold" than a random yellow clock. OutsideNormality (talk) 16:30, 19 August 2025 (UTC)[reply]
However, if it's related to a block, I think there should be a central image to keep the theme (like in this case it was a clock), and then in the bottom right have a changing image based on 'status'. For example: (illustrative example only not a suggestion!).
Support Different-colored clocks don't convey much on their own. For warnings and blocks, there is also the necessity of communicating with users with limited English fluency. –LaundryPizza03 (dc̄) 16:57, 21 August 2025 (UTC)[reply]
Support as of now the icons are basically identical for someone who hasn't spent looking at them. Change in the symbol will be beneficial. (please ping on reply) ~/Bunnypranav:<ping>11:43, 31 August 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
I still think it would be worth considering whether people really want to change to these icons, or if an alternative that keeps a similar style to the old ones while being different in more than just color would be more accepted. Anomie⚔11:19, 19 September 2025 (UTC)[reply]
Question 4: Should we replace the icon for current events?
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.A summary of the conclusions reached follows.
Not with this replacement. The rationale "bikeshedding" is kind of hard to interpret. The only way to make decisions about whether a given change is an improvement without consulting the community is for people to just decide what to do on their own and do it. If the objection is to having the discussion in the first place, it would be helpful to say what the alternative mechanism should be for the next time this comes up. I'm not sure "never make any minor aesthetic changes" is a long-term viable strategy.
Anyway, there were plenty of other editors opposed for more normal reasons. Some provided encouragement to keep looking for a better alternative, saying the existing icon is unsatisfactory. -- Beland (talk) 08:17, 19 September 2025 (UTC)[reply]
Oppose no significant improvement. The current icon has a globe that looks like the Earth and a clock that looks like a clock. The proposed change looks way too cartoony, has an Earth apparently drawn by an anti-Mediterranean conspiracy theorist, and an oversimplified clock. Let's have a nice icon instead of the simplest one available.
No it's not, I was just making the proposition since I thought the above was also derailing to other icons but I see it's simply going to change other icons to Codex. :-) waddie96 ★ (talk)20:44, 20 August 2025 (UTC)[reply]
Just a joke, because the second globe has North America, Greenland, and South America rendered in good detail but doesn't distinguish between Europe and Africa: hence, no Mediterranean Sea. Cremastra (talk·contribs) 21:12, 20 August 2025 (UTC)[reply]
Oppose, no improvement beyond making it flat, and too busy (in terms of colors) to fit with the Codex design language. To fit the theme, a much better choice would be , , or a combination of the two. Chaotic Enby (talk · contribs) 21:24, 20 August 2025 (UTC)[reply]
I think Codex is better suited to smaller icons, like the buttons in the source editor tab in the reply tool. Larger icons like this benefit from a bit more detail and colour, but if we were to move to flat, I don't think Codex would be the right answer, because it would be at this size:
The first thing you notice about that icon is that it is very ugly, because I have no design sense, but I think if we want flat icons that's the level of detail to maintain. Cremastra (talk·contribs) 00:35, 21 August 2025 (UTC)[reply]
Strong oppose: I have seen three topics created by you so far concerning the corporatization of various Wikipedia images and icons. I am opposed to all of those changes. OmegaAOL (talk) 00:40, 5 September 2025 (UTC)[reply]
Please don't discourage people from expressing their opinions. The tradeoff in participating or not is up for each person to weigh. isaacl (talk) 14:58, 10 September 2025 (UTC)[reply]
Support No strong preference between the Icons8 composite and Cremastra's globe-patterned clock (as an SVG). The Icons8 globe by itself () meshes well with the Codex signs and the proposed broom icon; it could be used for Template:Globalize if the Codex-style symbols are adopted for generic use. –LaundryPizza03 (dc̄) 17:06, 21 August 2025 (UTC)[reply]
I concur. There's plenty of nice flaticons similar. A lot of noun project, SVGrepo etc. You just have to dig through 1000+ to get one. But also Microsoft Fluent UI, Antu SVG, Breeze, Icons8, Iconoir, Material Design 3 waddie96 ★ (talk)11:17, 10 September 2025 (UTC)[reply]
Neutral This feels like a trivial change like the color of templates. This icon is flat and consistent with the other Codex icons, but the icon doesn't look that good in dark mode (especially with the white clock). There may be a better icon that is suited for this that can be added as an option B or C. Aasim (話す) 16:13, 22 August 2025 (UTC)[reply]
Strong oppose, we should not be replacing a decently accurate coastline with some really poor blobs. We're trying to be an encyclopaedia, let's try to get some small details right. CMD (talk) 11:28, 10 September 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Just an FYI to those who are watching this discussion: waddie96 is actively changing mboxes and also submitting mbox template edit requests that include these new, flat icons, despite what looks to me like significant opposition in the discussions above. I think they may also be removing linking to the icons, possibly failing to comply with licensing for the icons (I am not a licensing expert, so I could be wrong on this one). You can review their contributions in Template space if you want to see the extent of their recent editing. – Jonesey95 (talk) 04:33, 30 August 2025 (UTC)[reply]
Hey, sorry I didn't think those template edits were controversial or considered as part of this discussion because they're not Codex icons. I will revert the edits. waddie96 ★ (talk) 12:51, 30 August 2025 (UTC) (scratch that, I see one did use a Codex icon waddie96 ★ (talk)13:00, 30 August 2025 (UTC))[reply]
Anomie, I think it's far from a block, when this is the first mention of it. I'm a regular editor of enwiki yes, but to be honest didn't think those template changes were applicable to this discussion. As an administrator you should know that typically, before seeking a block, a final warning in an escalating series should have been posted to the user's talk page. waddie96 ★ (talk)12:58, 30 August 2025 (UTC)[reply]
Thank you for responding appropriately to the notification above! 🙂If behavior is disruptive and continuing, then a block would be appropriate and I don't see a problem in noting that a block is a possible outcome when discussing behavior that may be disruptive if continued. I also don't see the statement you quote in any policy or guideline; I do see something similar in Template:Report vandal, but that wouldn't apply here anyway since WP:Fait accompli is not vandalism. Anomie⚔13:21, 30 August 2025 (UTC)[reply]
Look, I appreciate your thank you. And what I mean to say is, a block is obviously a very serious action. And it will only incite an unfavourable response in a discussion. I believe what makes most users leave enwiki is how we speak to one another. We're generally not kind, or understanding, nor do we AGF as often as we should. I am just as guilty. But I do appreciate bringing up the issue here before reverting my edits. And giving me an oppurtunity to fix up the issue I created. 😄
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.A summary of the conclusions reached follows.
Consensus is against Wikipedia officially endorsing this company or making their files "an inherent part of article's structure". If their content is licensed appropriately, they're welcome to upload it to Commons, and people without a COI may use the content here if they find it useful.Debates on what "licensed appropriately" means when it comes to the training data behind AI text-to-speech aren't really in scope for this page, and at this point the only discussion is people arguing over that. There's probably a good place over at Commons for that sort of discussion. And as a bonus, over there it's more likely to wind up actually affecting anything. Anomie⚔21:54, 14 September 2025 (UTC)[reply]
Dear Community, My name is Amiel.
At TTSREADER we have undertaken to convert a large amount of Wikipedia articles to synthesized speech with the latest and most advanced AI voices available. We currently have some 50,000 articles synthesized both in Male and Female voices and have created an environment that allows one to navigate through different articles, listen and follow the article content. https://ttsreader.com/wiki/
We currently hold also some 50,000 articles in Spanish. We are looking to expand this resource all the time.
I want to bring this project of ours to the attention of the community with the hopes of turning it into an inherent part of article's structure. We are looking to share/give the content that we have, both from our site as well as integrate the audio files in existing Wikipedia articles.
I have been in touch with a number of Authors and Editors as well as reviewed some past discussions on this matter and have received the impression that there is no desire to integrate such audio files in articles because of lack of human authenticity. I disagree with this stance, and would like this to be discussed once again.
Important to point out, the voices we employ have very realistic humanlike characteristics. They incorporate, emphasis, intonation, pauses and tone that are very engaging. All this leads to improved delivery and comprehension. The articles themselves are word for word the content of the original articles derived from the date of capture. The articles have been sanitized to ensure only free-flowing comprehensible content is conveyed. Thus, content of this nature goes a long way to make Wikipedia articles more accessible to the general population and simply accessible to those with disabilities such as the vision impaired or those with language and reading impairment such as Dyslexia and ADHD.
Has any of the audio material the AI voice model has been trained with been used without the permission of their copyright holders? Cortador (talk) 14:41, 17 August 2025 (UTC)[reply]
The AI voice models we work with, are commercial models provided by leading vendors such as Azure, Google and OpenAI. There privacy policies are very stringent and from the research that we have done their model development is without copyright infringements.
The content of the output voice is purely controlled by us and how we apply these models, and the rights to the content produced by the models are of those who synthesize the voice.
I dont see, or at least not aware of any copyrights problems. @Cortador Do you suspect a problem?
OpenAI alone is currently involved in at least a dozen lawsuits (see also here). I don't know whatever research you have done, but there's clearly some disagreement on whether or not these models have been built "without copyright infringements". Cortador (talk) 14:58, 17 August 2025 (UTC)[reply]
Seconding @Cortador's comment. In my opinion, it's best for WP to remain aloof from models with copyright infringement lawsuits against them. There do exist TTS models without any such issues, but it does not appear that the project in question has taken care to select only such fairly trained models for its use. Aurodea108 (talk) 03:47, 11 September 2025 (UTC)[reply]
There are also a bunch of people suing Wikipedia, at this very moment, nearly all of which for inane reasons -- what's the point? jp×g🗯️17:30, 14 September 2025 (UTC)[reply]
That people are already suing Wikipedia does not support the idea that Wikipedia should start getting involved in other matters of questionable legality, especially since there are other perfectly good alternatives available. Aurodea108 (talk) 19:57, 14 September 2025 (UTC)[reply]
You're saying words, but did you first obtain permission from the copyright holder of every piece of audio you've ever heard in your lifetime? You're a human, but have you made sure that no human is currently the subject of a lawsuit? jp×g🗯️17:00, 14 September 2025 (UTC)[reply]
While this looks very helpful as a separate service, I do not think an external tool such as this one should be integrated as an inherent part of article's structure, as it would make Wikipedia reliant on this external provider. If the tool was entirely free and open-source, integrating it to MediaWiki could be a possibility, but it isn't clear to what extent this is the case, and whether these latest and most advanced AI voices available are under a compatible license (or, like Cortador points out, were even trained with permission).The promotional language in the proposal also bugs me a bit, but I believe that this tool should be analyzed on its merits. Sadly, there is little that I can see beyond a user interface, so it is hard to find more about the tool's capabilities or its technical details. Chaotic Enby (talk · contribs) 15:05, 17 August 2025 (UTC)[reply]
Edit: seeing the replies above, I am sad to say that this simply cannot be possible. Wikipedia will not integrate inside our articles a tool that relies on commercial models, especially to that extent. Chaotic Enby (talk · contribs) 15:08, 17 August 2025 (UTC)[reply]
Let me clarify what I meant by "inherent part of article's structure". I mean to say that the audio outputs/audio readings that we have generated I would be happy to give them/ provide them for use as part of Wikipedia articles. In this way, they can provide an added media that contributes to the content's accessibility. I dont want to turn Wikipedia to become reliant on us, that defeats my intention completely. What I have I want to share.
@Chaotic Enby Excuse my promotional language, im not yet attuned to the community's sensitivities. I believe I understand your concerns. I am trying my best to share resources that we have that I believe can contribute to Wikipedia. I ask that you judge what I share and the content based on merits. Ill go even further, if you believe there is a better way/fashion to improve the merits of what I have shared I would be the happiest to hear.
@Cortador thanks for sharing this. Could you please elaborate where there could be a problem with this matter? The Engines I mentioned above are used across the world across different platforms and fields of application all the time and are ever expanding their use and applicability. (Lawsuits take place all the time, does that stop the activity?, most often not.)
Why is this of concern to you? If the materials are developed lawfully and shared legitimately why should this limit the possibility of using these materials? I may be naive but if the outcome is of benefit then surely that should be the metric? Would love to hear. AmielRieger (talk) 15:55, 17 August 2025 (UTC)[reply]
Wikipedia is distributed under a free license, and avoids using third-party services that do not have a compatible license. Even if you decide to provide these recordings for Wikipedia, it needs to be under a similar license (which allows for commercial re-use), which I am not sure is compatible with the terms of the commercial models you are using. Chaotic Enby (talk · contribs) 16:03, 17 August 2025 (UTC)[reply]
Thank you for the feedback so far. To clarify:
No reliance on external services – I am not proposing embedding Azure/Google services into Wikipedia. Instead, I propose to provide the finished audio files, which can be uploaded to Wikimedia Commons and hosted like any other media file. Once uploaded, Wikipedia would not depend on us or any commercial provider.
License compatibility – We already do and will release the audio under CC BY-SA 4.0, the same license as the article text. The models we use (Azure, Google) explicitly grant users rights over outputs, and we have full rights to release them under a free license.
No training data risk – We do not train AI models ourselves. We only apply existing licensed tools to freely licensed Wikipedia text. The audio is a direct transformation of CC-BY-SA content, not a reuse of copyrighted material.
Accessibility benefit – The purpose is to make Wikipedia more accessible: helping vision-impaired users, language learners, and readers with dyslexia or ADHD engage with articles more easily
I fully understand the importance of licensing clarity, and I am happy to provide concrete proof of licensing rights if needed. My goal is to share this work in a way that strengthens Wikipedia’s mission of free, accessible knowledge for all.
Who would be an authority in this matter that will assist me in clarifying any issues? Is there someone I could turn to? Im convinced that the issue of licensing is not a problem. AmielRieger (talk) 17:58, 17 August 2025 (UTC)[reply]
Yes. I believe we have effectively already "released" these recordings under license. You can review them on our site https://ttsreader.com/wiki/.
On our site that we setup we already make these licensing declarations in accordance with the licensing required (as mentioned above). Please see a template for example: https://ttsreader.com/wiki/?article=Albert_Einstein (If you are already gonna visit the site then I would be very happy to hear your feedback)
Furthermore, as a test a couple of months back I uploaded an ogg file of the United States article. @Tbhotch at the time temporarily uploaded it after reviewing the recording in its entirety. It was also taken down by him a few minutes later.
Don't Creative Commons licenses automatically apply to derivative works? If so, they are necessarily released under compatible free licenses, and it is impossible for them to be released otherwise. jp×g🗯️17:03, 14 September 2025 (UTC)[reply]
Oppose Their answer about the source of their model's data is unsatisfactory. In the absence of an explicit statement that they personally sourced all of the inputs from freely licensed sources (i.e. they did not outsource the collection and vetting of inputs to another organization), and that they have and always will follow all of the requirements of those licenses, it is safe to assume that they, like every other AI model, are engaged in wholesale theft. Putting aside any legal arguments, as those are still working their way through the courts, it is unethical for a project built on and around a free license, and which has extensive policies about respecting copyright, to turn a blind eye to how the inputs of AI engines are collected and pretend that it's not a violation of our own community pillars to accept the outputs. The Squirrel Conspiracy (talk) 21:07, 17 August 2025 (UTC)[reply]
I also want to add that the accessibility angle rings hollow to me. People that need text to speech already have tools that are purpose-built for that function, and will do a better job of it then an AI transcript, with the added benefit that what they hear will always reflect the current version of the article. The Squirrel Conspiracy (talk) 21:14, 17 August 2025 (UTC)[reply]
Dear @The Squirrel Conspiracy. I have high regard for your concern for ethics, but I respecfully disagree with you. Furthermore to falsely accuse me and my organization of "wholesale theft" before honestly assessing, discussing and researching us and our content is unethical in its own right. ( I am somewhat taken aback by your smearing) So please deal with content of the discussion and refrain from judging or smearing us publicly. and now to respond to the points you are making:
I recognize concerns about how commercial AI models are trained, but this is not relevant to the licensing of the outputs. The recordings we provide are derived solely from Wikipedia content (freely licensed text) and are released with the same free license. By ensuring outputs are CC BY-SA 4.0, we comply fully with Wikimedia’s requirements and contribute material that enhances accessibility without introducing copyright conflicts. As simple as that.
Any further sophistry is not relevant.
Im surprised that the argument to make Wikipedia even more accessible is a red blanket to you, and the solution to that is not within the realms of Wikipedia. That seems not in line with Wikipedia's ethics. And if we are speaking about ethics, then the ethics of providing such a service (sooner rather than later) and accessibility to those with real limitations far outweigh the counter ethics of legally developed and licensed content based on AI engines. Important to emphasize, the textual content and thus the vocal output is pure Wikipedia text and as mentioned in my previous response all of our WIKI media is available to be used free, in line with Wikipedia licensing requirements.
@Sdkb The intent is good the implementation is not as I would have imagined the standard to be. We already hold some 50k articles in English and some 50k articles in Spanish. Each of these articles has both Male and Female voices. These voices are the leading voices available in terms of their humanlike nature, including intonation, emphasis, pauses and tones. All Im offering is to provide our existing content to Wikipedia.
this is not relevant to the licensing of the outputs This is where we fundamentally disagree. If your approach to concerns about the inputs is "it doesn't matter, just look at the outputs", our understanding of Wikipedia's goals and underpinning philosophy are wholly incompatible. The Squirrel Conspiracy (talk) 22:27, 17 August 2025 (UTC)[reply]
There do already exist TTS models that are based on properly licensed training data. I don't agree with the statement "it is safe to assume that they, like every other AI model, are engaged in wholesale theft." However, from what the project in question has already stated, they didn't select those fairly trained TTS models. Aurodea108 (talk) 20:22, 14 September 2025 (UTC)[reply]
People can run text-to-speech software on articles at any time, locally and with flexibility. What's the point of taking specific versions of articles and making them into inflexible audio files that go out of date almost instantly? Dan Leonard (talk • contribs)22:22, 17 August 2025 (UTC)[reply]
The point is that a for-profit company that no one's ever heard of gets to say that they have a partnership with Wikipedia - one of the most prominent brands on the internet - when they go to investors for their next round of funding. Let's not pretend for a moment that this effort is anything but self-serving. The Squirrel Conspiracy (talk) 22:29, 17 August 2025 (UTC)[reply]
I appreciate the thoughtful concerns raised here and want to address them directly.
First, regarding our company:
I am proud to say that TTSReader and the overriding company Well Source Ltd, is a completely self-funded company. We have never undergone a round of funding, we have no investors, and we are entirely bootstrapped and self-sustaining. We have been in the market for over 10 years, and today we serve more than two million people around the world, many of whom rely on our completely free services. So, while we may not be a household name, it would not be accurate to say that “no one has ever heard of us.”
Second, on the question of being a for-profit company:
Yes, we are a for-profit company. We are not ashamed of that, and we have no intention of hiding it. That said, our interest in contributing to Wikipedia is not at odds with that reality. We believe we have built one of the best TTS products and we have developed relevant content, and so we see alignment between what we’ve achieved and Wikipedia’s broader goals of accessibility and knowledge sharing. We have invested real money to create this content, and want to offer it for free to the Wikipedia community as it is freely available now on our website.
To be clear:
The content itself remains pure Wikipedia text.
The outputs (our audio files) are released under the appropriate free license and fully reusable under Wikipedia’s requirements.
Our intent is not to extract value from Wikipedia, but to add accessibility for audiences and yes to be proudly associated with contributing content to Wikipedia.
I understand the skepticism, particularly around the role of commercial actors. But I would argue that being a for-profit does not automatically make our contributions misaligned with Wikipedia’s mission. If anything, our long-standing sustainability without outside investors demonstrates independence, stability, and seriousness of purpose.
The audio file for an article becomes outdated as soon as the article is edited, which is one of the reasons why these text-to-speech projects aren't really popular on this site. Some1 (talk) 23:19, 17 August 2025 (UTC)[reply]
Thats a valid point. There are many points to ponder regarding this vision.
Our considerations are that, many many articles are not that dynamic in nature and thus the audio files remain relevant over a significant period of time. Furthermore, in most cases the vast majority of content remains the same and so the overall value of the audio file remains high. On our website we "freeze" the articles in time.
We are thinking of a mechanism where the audio files get updated (rerendered) when a certain threshold of change occurs. But this would require development and operating costs that at least for now we don't have those resources.
An alternative would be to integrate a player of sufficient quality into all articles and render them dynamically. This has its own complications both technological and potential overheads. Both I believe beyond the scope of Wikipedia. We for example can provide such integrations, and even provide it for free, but in light of the earlier discussions and perhaps because of association concerns this may not be viable. If anyone in Wikipedia is open to consider this avenue, I would be happy to discuss it.
Why are we even having this discussion? If you want to "provide the finished audio files, which can be uploaded to Wikimedia Commons and hosted like any other media file", then nobody here can stop you from doing that. WhatamIdoing (talk) 01:48, 18 August 2025 (UTC)[reply]
Thank you @WhatamIdoing, I am new to this world of content contribution and I made an initial test with the United States article. I linked it earlier: https://commons.wikimedia.org/wiki/File:United_States.ogg but a couple of days after I posted It was put forward for deletion by @The Squirrel Conspiracy and I see now that @Krd has removed it as of a couple of hours ago. Could you please guide me how to upload content while making sure that it wont be deleted? What should I do different in order to avoid this the next time?
Why we are having this discussion? Well, cause there is content available that I believe adds value to Wikipedia but I sense otherwise from certain players in the community. I have not been convinced. The content remains true to the original text and does not infringe on licensing guidelines etc. What if I load all 50k articles in Male and Female voice ie 100k files and because it bothers some they all get deleted, this is not clever on my part. Lets not forget the 50k Spanish articles we have too. Thus I preferred to raise this issue in this forum as I was recommended by @ChildrenWillListen and hopefully through open and honest dialogue I would be able to progress this topic. I was very encouraged initially when @Tbhotch had listened through the entire content and saw no flaw with the actual content. AmielRieger (talk) 08:02, 18 August 2025 (UTC)[reply]
As was mentioned above, you are free to upload your properly licensed audio files to Commons. Any uninvolved editor can link content from Commons in a Wikipedia article, but whether that link will be left in the article is subject to community consensus. I think it is clear that there is no appetite here for any formal connection with or endorsement from Wikipedia of your project. I suggest reading the essay at Wikipedia:Don't bludgeon the process before posting here again. Donald Albury13:36, 18 August 2025 (UTC)[reply]
A terrible American accent. Then it read "14 March" as "March fourteenth"! Sorry, but that's just wrong. You've lost me! HiLo48 (talk) 00:26, 18 August 2025 (UTC)[reply]
If you prefer a British accent that can also be synthesized ;-). Either way Im glad you listened to the content. Much more than many others. AmielRieger (talk) 07:29, 18 August 2025 (UTC)[reply]
You lost me at “recording”. Our articles change frequently (as editors add, subtract and reword information). So any recording would quickly become out of date. Non-starter right there. Blueboar (talk) 14:07, 18 August 2025 (UTC)[reply]
Oppose any commercial integration with or commercial endorsement by Wikipedia. Wikipedia is not a place to show off your new product nor a place to experiment with novel technologies. Our license permits the company to host a spoken Wikipedia mirror if they so choose, but there are serious unresolved moral and ethical problems with the technology itself, as well as inherent conflicts of interest with commercial entities. AmielRieger's dismissive rebuttal to The Squirrel Conspiracy above suggests that they already lean into the "ends justifies the means" ethical fallacy so common with large language model techbros that Wikipedia has always rebuked, and should continue to rebuke. We should not do this unless it's developed in-house, despite the difficulties that the m:Wikispeech project seems to be highlighting. Ivanvector (Talk/Edits) 14:18, 18 August 2025 (UTC)[reply]
Oppose for three reasons: the audio files would be outtdated the moment someone makes an edit, on-the-fly text-to-speech software is likely more useful for people with accessibility issues, and Wikipedia is not a platform to showcase technology based on large-scale theft. Cortador (talk) 15:23, 18 August 2025 (UTC)[reply]
I have ... received the impression that there is no desire to integrate such audio files in articles because of lack of human authenticity. I disagree with this stance, and would like this to be discussed once again.
So despite the fact that people already explained to you that we don't want it, you are asking the same question again? We still don't want it. And the next time you ask the same question we still don't want it. And if you keep asking the answer will still be 'no'. Hope that helps. Polygnotus (talk) 09:59, 19 August 2025 (UTC)[reply]
Well, irrelevant to Commons and Wikipedia, which are invested in copyright and not in the contractual agreements between two third parties. ꧁Zanahary꧂14:16, 21 August 2025 (UTC)[reply]
Oppose for the reasons explained by Cortador, Polygnotus, and others above, but also and at least as importantly at the moment, i do not believe that any AI production currently sounds even close to natural and will therefore, at least misrepresent our content and, possibly be inaccurate in its mispronunciations and mis-emphases. While this last reason may, just possibly, one day be eradicated, there isn't any chance that currently such "recordings" would be worth integrating into the encyclopaedia ~ or even linking to ~ LindsayHello16:15, 19 August 2025 (UTC)[reply]
Hi Amiel. I'm sorry that people are being hostile towards you. If your audio is already freely licensed, and Wikimedia Commons allows AI-generated material (it does, last I checked), then you should feel free to upload it to Commons. But you should probably check in at the Commons village pump first. Off the top of my head, it may be Commons's position that AI output may not be under a CC license, and is instead presumed to be ineligible for copyright protection and thus necessarily in the public domain. There is a template on English Wikipedia that embeds audio recordings at the top of an article. People could use that template to embed your company's outputs. ꧁Zanahary꧂20:52, 19 August 2025 (UTC)[reply]
We already have WikiProject Spoken Wikipedia for providing audible versions of articles, and I think that project needs more support before we even think of an automated solution. Issues with an automated TTS, like pronunciation problems and an audible version becoming outdated, are directly addressed by the project, which encourages its readers to look up pronunciation guides and focus on featured articles (which are not liable to greatly change over time). --Grnrchst (talk) 13:12, 20 August 2025 (UTC)[reply]
No thanks. Even if only freely licensed models were used, it would still be necessary for every output to be manually reviewed for mistakes and then rectified as needed. And as can be easily gauged from the sentiment expressed at WP:VPW recently about the WMF's AI summaries project, the community isn't interested in undertaking such a large task. Not worth the effort and time spent. JavaHurricane17:17, 21 August 2025 (UTC)[reply]
While I agree with the other concerns, I don't get this. TTS is not generative and does not hallucinate factual errors. Any minor audio glitches should be fixed with eyes wiki-style. Aaron Liu (talk) 17:36, 21 August 2025 (UTC)[reply]
This doesn't make sense. There is no reason to believe that TTS is more error-prone than humans at reading text aloud. ꧁Zanahary꧂17:59, 21 August 2025 (UTC)[reply]
does not hallucinate factual errors is untrue, at least in my experience with AI TTS tools. And this doesn't account for issues such as mispronounciations (especially with technical terms, for instance) which would still need correcting.
Any minor audio glitches should be fixed with eyes wiki-style: wouldn't be an issue if the load of articles for which audio files are being generated was not too large (as would be the case with audio files generated by humans under, say, WP:SPOKEN). Maybe I'm a cynic, but I don't see the load not being large with an AI TTS service being used. JavaHurricane12:36, 22 August 2025 (UTC)[reply]
Oppose. Who knows but that we might (accidentally) end up with another Bhutanese passport situation, for one thing. Also, don't most people with Internet access already have access to screen reader software? Your operating system should already have some feature that reads aloud the current version of a webpage (at least basic text, which we use), which avoids the aforementioned problem of our content constantly changing. Nyttend (talk) 09:58, 31 August 2025 (UTC)[reply]
Oppose many great opinions above to back this one up, but simply put enwiki as a project cannot be incorporating content that is from propriety software. Open-source software that gives public domain content should be a pre-requisite for our project. Besides we have an in-house extension in development that is opensource and freely licensed called meta:Wikispeech.
I don't really understand why people are bolding their responses to this, since it doesn't really seem like the kind of thing that is contingent on our approval or disapproval, but I guess that if that's what we are doing I may as well support it. As far as I can tell, the proposal is identical to WP:Spoken Wikipedia -- recordings being made of static versions of articles and then made available to readers with appropriate provisos. Any objection made on the basis that this is bad or dangerous, or provides some other detriment, would also apply to Spoken Wikipedia recordings. There are also a bunch of objections made on the basis of people disliking specific software companies on account of their being evil. This is true -- most software companies are, shall we say, ethically challenged -- but I think you will agree that a similarly puritanical proposal to require all Wikipedia users to run Linux (or even to use Firefox) would be an obvious nonstarter. There remains the completely unprecedented idea, lately in vogue, that anybody who speaks must do so with the express approval of the holding companies which own every copyrighted work he has ever heard in his entire life, which I think fundamentally does not make sense, although IP trolls on Bluesky have been saying it over and over for years; the idea that the companies are "brothers", which makes their products bad(?); altogether I do not really see an objection to this, except that it makes us as human editors feel bad to have our "brand" diluted. jp×g🗯️17:21, 14 September 2025 (UTC)[reply]
(Sorry to repeat what I said above; I guess I should've read all the new replies before replying.) Spoken Wikipedia involves human readers, which necessarily has some degree of more accurate pronunciation and intonation. Having TTS everywhere provides no advantage to what can already be done with the user's browser or operating system. Aaron Liu (talk) 17:46, 14 September 2025 (UTC)[reply]
I have never before heard the idea that Wikipedia should omit features/content on the basis that some readers probably have the technical resources to do it themselves. For example, we have dark mode — despite the fact that some people (like myself) can write their own stylesheets — I've heard nobody say dark mode was therefore a waste of time, certainly never that we shouldn't allow people to implement it on their own [t/d]ime. I think this argument would, in any other circumstance, be considered a total non sequitur; the style of reasoning is used nowhere else on the project.
It is true that some people have browser TTS; I don't have any easy way to do it on my desktop. I do on my cell phone, which was a fairly expensive high-end one when I bought it. I do not really agree with the idea that we should not allow people to contribute content for free because people with expensive cell phones can in theory set up a way to recreate that same content with their own hardware. jp×g🗯️19:18, 14 September 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Do you think we should change the village pump's header color? Personally, I dislike this yellow-brownish color; it feels outdated and isn't in the usual Wikipedia style. We have so many better colour options over at pages like:
and many others. I suggest we pick a replacement for the current color. The colors below are just some examples; the first three are used on the Main Page.
Update 16:08, 21 August 2025 (UTC): Zanahary has created actual mockups; you can see them in the § Mockups section. (I also collapsed the other color choices below, which weren't being voted on.)
Option A
this is an example of black text with color applied to its background and border
Option B
this is an example of black text with color applied to its background and border
Option C
this is an example of black text with color applied to its background and border
Option D
this is an example of black text with color applied to its background and border
Option F
this is an example of black text with color applied to its background and border
LMFAO. Chaos! Hyperbole much? Encountering an unexpected color change, editors will be beside themselves and won't know what to do! If I didn't know you, I'd suspect trolling. ―Mandruss☎ IMO. 07:10, 11 September 2025 (UTC)[reply]
The current dehydrated beige-yellow-ish looks fine as-is, but D also looks great. Everything else is too light, imo. EF500:29, 21 August 2025 (UTC)[reply]
Support for decorative reasons. This will give visual impact and this will be decoratively clear for colorblind users. Additionally, this will make the Village Pump creatively nice and beautiful. Fabvill (Talk to me!)01:09, 21 August 2025 (UTC)[reply]
Whatever you do, don't forget about dark mode whenever you pick colors nowadays. it has to be legible in both, or you need additional colors for dark mode. —TheDJ (talk • contribs) 11:48, 21 August 2025 (UTC)[reply]
All of the options look ok to me in dark mode. A and C are pretty dark, but I don't see how that would be a problem in dark mode. You want dark? You got dark. Mustn't wake up the wife. Text-background contrast is fine for all. ―Mandruss☎ IMO. 07:59, 11 September 2025 (UTC)[reply]
Support D as proposer; I like purple in general and D is also used on the Main Page, so I agree with the other comments. The village pump should be welcoming and purple is perfect for that. FaviFake (talk) 12:01, 21 August 2025 (UTC)[reply]
Support, the color is indeed ugly right now. I have no problem with any of the mockups made by the proposal nor by myself below. ꧁Zanahary꧂14:34, 21 August 2025 (UTC)[reply]
Sure, no harm, purple would probably look nicer than the current option. But this really isn't the most pressing issue facing us at the moment; bold editing might have been better than launching a giant survey. I'd suggest this be moved to the talk page. Sdkbtalk15:01, 21 August 2025 (UTC)[reply]
But this really isn't the most pressing issue facing us at the moment - Oh I wasn't aware we were supposed to limit discussion to the most pressing issue (wouldn't that mean one thread on the page at a time?). bold editing might have been better than launching a giant survey. Bold editing would have been promptly reverted by one of the Opposers in this discussion. Then we'd be here anyway. We just skipped two needless steps in the process, B and R, and led with D. No, it doesn't need to go to talk. ―Mandruss☎ IMO. 07:28, 11 September 2025 (UTC)[reply]
OpposeWP:DONTFIXIT, there's nothing wrong with the current color and there's nothing better about any of the suggestions below. All this will come down to is which color happens to get the most "I like it" votes from people who bother to comment at all. Anomie⚔16:11, 21 August 2025 (UTC)[reply]
Not so much. We're looking at actual real tangible things like contrast ratios. Besides, I have no problem with popularity for something like this. ―Mandruss☎ IMO. 11:56, 11 September 2025 (UTC)[reply]
The only discussion of contrast ratios is making sure all the options have good ratios. Your own chart below shows that the current one has a better ratio that many of the proposed options, but even the worst is well above the minimum. So, no, that doesn't really seem to be part of anyone's decision here. Anomie⚔12:06, 11 September 2025 (UTC)[reply]
Fine. In that case, I have no problem with popularity for something like this. Also, one of my greatest Wikipedia pet peeves is resistance to change. "What we have is fine" is a horrible argument, just horrible. ―Mandruss☎ IMO. 12:32, 11 September 2025 (UTC)[reply]
Did the current color set even get so much as "a popularity contest vote"? The colors for talk page templates did (Wikipedia:Talk page templates/vote). It would be unfortunate if we rejected the results of current discussions as a mere vote, if the current state didn't even get a discussion in the first place. WhatamIdoing (talk) 16:08, 11 September 2025 (UTC)[reply]
Support D (or option 6 below, with option 5 as a second choice). I don't think WP:DONTFIXIT applies here as there is an identifiable issue, even though it is of an aesthetic nature. I also wouldn't be opposed to revamping the tab design entirely, as the border still gives it a very dated feel. Chaotic Enby (talk · contribs) 17:30, 21 August 2025 (UTC)[reply]
After trying the mockups below, also support option F as a third choice. My only worry is that it might be a bit too close to the WP:CENT color. Although Option D would likely look better with a lighter border like F's. Chaotic Enby (talk · contribs) 12:54, 22 August 2025 (UTC)[reply]
One thing I forgot to consider: on blue and purple skins, blue and purple links will naturally be less readable. I added colored links to my mockups – looking at the contrasts, I now have a preference for option F over option D. Chaotic Enby (talk · contribs) 18:19, 22 August 2025 (UTC)[reply]
Looks like that would only work in Vector 2022 and Minerva. Those CSS variables don't seem to be defined in Vector, Monobook, or Timeless. Anomie⚔18:20, 31 August 2025 (UTC)[reply]
Yeah apologies that's why you define them like this: var( --background-color-success-subtle, #dff2eb );. thanks for letting me know, I didn't actually know that :) waddie96 ★ (talk)22:59, 31 August 2025 (UTC)[reply]
Option A. Tastefully understated, calming, not one of the overused blues. Like the paint in a doctor's office. Brown looks like dirt; if I see that on the walls of my doctor's office, I'm finding a different doctor. And that's how to put together a proposal, by the way. ―Mandruss☎ IMO. 07:13, 11 September 2025 (UTC)[reply]
Comment. I used Snook's Colour Contrast Check. For the uninitiated, higher numbers are better, and all values exceed WCAG 2.0 Level AA (4.50) and WCAG 2.0 Level AAA (7.00) by a wide margin. Black-on-white and white-on-black are 21.00. I'm assuming #000000 for black and #FFFFFF for white, though I can't get Firefox's color picker to verify that for text.
The examples here (as well as the current header) all specifically set color:black, overriding the Vector 2022 (and Minerva) default text color of #202122. When I toggle Vector 2022's dark mode, the current header's colors are overridden entirely (we wind up with #eaecf0 text on a transparent background (showing through black), with the colored border remaining unchanged) while the examples here are not changed for dark mode at all. Anomie⚔12:00, 11 September 2025 (UTC)[reply]
There shouldn't be a problem if just the colors are switched in the existing {{Start tab}} invocation. You said I'm assuming #000000 for black and #FFFFFF for white, though I can't get Firefox's color picker to verify that for text, so I was providing more information on the text color used for the header and why it's actual black rather than Vector's normal dark-grey. Anomie⚔13:57, 12 September 2025 (UTC)[reply]
Comment. Revised table. For the purposes of contrast checking, I think we need to look at the paler of the two shades in each option sample. For the previous table, I was looking at the darker shades. I'm just correcting the record; the conclusion is the same: Contrast is not an issue.
Comment. This is not a suggestion for proposal expansion (I know better). But I note that there's a lot of brown on the site; just look at the top of an article talk page, for starters. I can imagine the output of this proposal becoming the new en-wiki color, thereby conveying a certain site-wide cohesion—like there is actually somebody in charge of the whole site; like there is some coordination happening. If it's good at village pump, it's good anywhere. One could argue that the brown does exactly that; problem is, it's a terrible color choice. And the browns aren't the same, anyway.As I understand it, this is what CSS is for. With a virtual flick of a switch, we should be able to change en-wiki's color site-wide. Anything less is 20th century technology. The uses for that switch are yet to be imagined, but I'm certain they would exist. ―Mandruss☎ IMO. 02:41, 13 September 2025 (UTC)[reply]
I think it would be nice if the Villages Pump had different color schemes to more easily distinguish them, but for the record, the heretofore-besmirched "yellow-brownish color" is actually a quite venerable piece of Wikipedia design history: the palette used by talk page headers and message boxes is called "ClockworkSoul's Coffee Roll" and it dates to a big RfC from some twenty or so years ago (prior to that, there was no unified formatting for the headers at all, and they were just all totally different styles and it looked like puke). jp×g🗯️19:20, 14 September 2025 (UTC)[reply]
I note none of the examples given here actually show what the header would look like. Even the one that claims to be the current color isn't, nothing in the current header uses this color. If we want to make a choice, we should probably have accurate mockups to choose from. Here's some wikitext to mockup the current coloring:
Selected tab
Other tab
Other tab
Other tab
Other tab
Other tab
This is the actual current village pump color. It uses the 50° row from Help:Using colours: "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
Unfortunately I can't guess at what was intended for any of FaviFake's suggestions here to mock them up similarly. Anomie⚔12:09, 21 August 2025 (UTC)[reply]
You're right. I only copy-pasted these "coloured boxes" from the pages i linked to to get a general sense of which one people would prefer. If a colour is more liked than others, say, green, we could then figure out all the specific shades of green for the border, inactive state, background, etc. I could work on creating more accurate mockups, but I think other people would do a much better job than me. I'm no colour scientist. (However, I've changed the incorrect colour you pointed out; it was intended to be more visually pleasing, but I agree it should match the current colours.)
The following discussion has been closed. Please do not modify it.
If you don't know what you're doing, perhaps you should take a step back and ask people for help before making half-baked proposals. Anomie⚔12:50, 21 August 2025 (UTC)[reply]
Excuse me? Again, my goal wasn't to propose specific colors. I wanted to see if there was consensus for changing the current colour. As I stated in my proposal, the examples were just that, examples of colours I thought looked better. FaviFake (talk) 14:10, 21 August 2025 (UTC)[reply]
An initially inaccurate summary along with suggestions that don't correspond to how the colors would actually be used doesn't sound fully baked to me. FaviFake has been doing a lot of this sort of thing lately. Anomie⚔14:33, 21 August 2025 (UTC)[reply]
I do agree there is something wanting in the current color choice. However, I'm not sure whether to support, due to a lack of discussion on which color, if any, is best accessibility-wise. A second concern is how the new header would appear in light vs dark mode, and on mobile vs. desktop. Dege31 (talk) 00:41, 3 September 2025 (UTC)[reply]
This is the village pump header using the 50° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
This is the village pump header using the 40° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
Option 2
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the 90° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
Option 3
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the 140° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
Option 4
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the 190° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
Option 5
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the 240° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
Option 6
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the 290° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
Option 7
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the 340° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs.
Options A, B and D derive their color palettes from the 150°, 210° and 270° rows respectively, with the only difference being the use of "main background" for the lightest color instead of "accent color", while options C and F have unique color palettes. Here is what the header would look like using these color palettes directly (as well as the 150° row for comparison).
Option A
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the Option A palette. That's "Title background" as the background, "Title border" as the border, and "Light Box background" for the inactive tabs. On Vector 2022, blue links look like this and purple links look like this. On other skins, blue links look like this and purple links look like this.
150° row
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the 150° row. That's "header background" as the background, "header border only" as the border, and "accent color" for the inactive tabs. On Vector 2022, blue links look like this and purple links look like this. On other skins, blue links look like this and purple links look like this.
Option B
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the Option B palette. That's "Title background" as the background, "Title border" as the border, and "Light Box background" for the inactive tabs. On Vector 2022, blue links look like this and purple links look like this. On other skins, blue links look like this and purple links look like this.
Option C
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the Option C palette. That's "Title background" as the background, "Title border" as the border, and "Light Box background" for the inactive tabs. On Vector 2022, blue links look like this and purple links look like this. On other skins, blue links look like this and purple links look like this.
Option D
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the Option D palette. That's "Title background" as the background, "Title border" as the border, and "Light Box background" for the inactive tabs. On Vector 2022, blue links look like this and purple links look like this. On other skins, blue links look like this and purple links look like this.
Option F
Other tab
Other tab
Other tab
Other tab
Other tab
This is the village pump header using the Option F palette. That's "Title background" as the background, "Title border" as the border, and "Light Box background" for the inactive tabs. On Vector 2022, blue links look like this and purple links look like this. On other skins, blue links look like this and purple links look like this.
Contrary to the popular opinion, Palette 'Blue' and 'Purple' (Options B, D & F) are clearly mixing-up with the Blue & Purple links, which will make it very hard for the color blind to distinguish.
That only leaves us with 'Green' (A) and the current 'Yellow' (C). Yellow is going to be replaced, so I would prefer Green - Light Green (Option 2 in this list) since not only is it different from the existing Green 'tq' function, but also it has a tint of legacy of its 'soon to be predecessor' Yellow. Cdr. Erwin Smith (talk) 08:21, 4 September 2025 (UTC)[reply]
That is actually genius! But we might need to create another VP topic to gather consensus for this, I suspect it'd be much more controversial. FaviFake (talk) 17:49, 13 September 2025 (UTC)[reply]
Numerically, there were 7 supports and 7 opposes. However, Wikipedia discussions are not a vote but a discussion. On the "for" side, people pointed out that the page was not well formatted, and cluttered up space. However on the "against" side, people pointed out that the page should be improved, not deleted, and in the standard Vector 2022 view that most people use, it didn't clutter much space. Therefore, I see no consensus to remove the contents link. —Matrixping mewhen u reply (t? - c) 16:14, 18 September 2025 (UTC)[reply]
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Yes as proposer. I'm not convinced that this page is really helpful for browsing Wikipedia, so I think would be helpful to have a discussion on removing the link. I initiated a discussion regarding whether to remove the link, but the discussion got stale, and it was eventually archived. I hope to achieve this by doing an RfC regarding this. Interstellarity (talk) 13:20, 22 August 2025 (UTC)[reply]
I do not support the proposal. I think it is reasonable to have a starting framework for browsing an encyclopedia, and that a contents link in the side bar fits this purpose. The rationale expressed in the previous proposal was more suitable as motivation to improve the lists of content and their presentation. isaacl (talk) 15:34, 22 August 2025 (UTC)[reply]
Oppose per Isaacl. It's worth noting that the way most readers see Wikipedia (open an article in a logged out private browsing window if you haven't in a while), the main menu is hidden behind a hamburger (three dashes stacked on top of each other) dropdown option in the upper left. So it's not as if it's a terribly prominent link that needs to justify being shown to everyone all the time. Sdkbtalk15:42, 22 August 2025 (UTC)[reply]
Oppose per Sdkb and isaacl. "Contents" is a useful link and shouldn't just be removed. If you don't like clicking on the link, either don't, or suggest improvements to the article itself. I'm sure there is a better contents page design out there if we're motivated enough to find one. --📎 JackFromWisconsin (talk | contribs) 15:50, 22 August 2025 (UTC)[reply]
Support – This is not how most users browse Wikipedia, and it only clutters the sidebar. I wished they'd remove it with Vector 2022, but alas it's still here. --FaviFake (talk) 16:12, 22 August 2025 (UTC)[reply]
As I alluded to in my !vote, what Vector 2022 did was essentially remove all the sidebar links by placing them under a hamburger menu (rather than directly visible on the page) for most users. I think that was the right call. Since while the Contents link isn't particularly useful, nor are the others. The actual links in the sidebar are under community control. Sdkbtalk16:25, 22 August 2025 (UTC)[reply]
My thinking was (which I agree that I should have explained), that people these days are far more likely to use search to find the content they are looking for rather than trying to navigate through a top-down list, and hence the link isn't useful. * Pppery *it has begun...18:15, 22 August 2025 (UTC)[reply]
Oppose. If you just cut off the (subjectively) least-useful link you gain nothing but a different (subjectively) least-useful link until you have no links at all. We have nothing to suggest that those who use it don't find it useful, and we have no suggestions for something more useful to take its place. It's not causing any harm, and nobody is forcing you to click the link - heck someone with more technical know-how than me could probably give you some fancy custom CSS or js to hide the link if it bothers you. Thryduulf (talk) 16:56, 22 August 2025 (UTC)[reply]
I'm looking for something that is basically a list of articles rather than a list of lists. Something like an A-Z list, but Wikipedia has too many articles to list alphabetically. If we can find a page where similar articles are grouped together similar to an outline. Example: History of the United States goes under United States, similar to Outline of the United States, but for all Wikipedia articles. Not sure if we have a page for that, but I'd be willing to create one if there's a need for it. Interstellarity (talk) 17:49, 22 August 2025 (UTC)[reply]
I was being tongue in cheek, but actually using this is probably a bad idea as not every article has a corresponding list (cf. WP:NLIST). Though I wouldn't be surprised if nevertheless it's still more useful for your average Wikipedia reader. novovtalkedits09:43, 24 August 2025 (UTC)[reply]
Oppose per the arguments above. If you think it's subpar, you are welcome to improve it. In my opinion, it makes sense for an encyclopedia to have a contents listing like Wikipedia:Contents. Consider also that, as mentioned above, Vector 2022 hides the entire sidebar in a dropdown by default, so it's not like the link clutters the page at all. (If you want to hide it, you can use this CSS: #n-contents {display: none;}) OutsideNormality (talk) 17:46, 23 August 2025 (UTC)[reply]
The page views for this page are kind of weird. ~450K per month until May 2020, when it halves. (What did we do that month, rearrange the Main Page?) Then it almost halves again, starting in late January 2023, when WP:Vector 2022 was deployed. The question that I'd like to be able to answer is: When readers get to that page, are they satisfied with it? What happens next, if an ordinary reader ends up on that page? WhatamIdoing (talk) 21:59, 23 August 2025 (UTC)[reply]
This seems like a good thing for the WMF to do. They are good at making a/b tests and running experiments. Maybe a technical team get can involved a new approach could use categories and something a little more dynamic. Definitely dreaming big here. 📎 JackFromWisconsin (talk | contribs) 02:55, 24 August 2025 (UTC)[reply]
Trying to formulate some sort of vote that runs along the lines of Keep but delete the most prominent part, which is the "Browse by subject area". Firstly, "Wikipedia's content is divided into broad subject areas" does not reflect how our content is divided, to the extent that it is this is through the category system which functions as a series of overlapping tags rather than discrete silos. Secondly, the specific subject pages like Wikipedia:Contents/Culture and the arts seem like weird duplications of outline articles such as Outline of culture. Thirdly, we already have a system that divides important content into subject, Wikipedia:Vital articles, which is included as part of Wikipedia:Contents so there's some weird nesting dolls of content division. And of course, the lists don't seem maintained (possibly because they duplicate the article space as well as Vital articles). Getting rid of these is a much better option than investing time into overhauls (of content that exists elsewhere!) as some have suggested above.However, while we don't divide out content into broad subject areas, it does seem useful to have a space that explains how we do divide our content. "Browse by format" seems useful to readers, give or take how much each format is used (hopefully it would be adjusted if those change). Introducing readers to the Featured and Good quality content systems would help them understand at least to some extent how content is developed. If readers do want subject lists, Vital articles provides an already curated list (although it should not be in the "Articles by quality" section). There are probably other tweaks that can be made, but repurposing the page to talk about Contents in the system sense rather than the Subject sense seems like it would be more helpful and duplicate the search bar less. CMD (talk) 04:14, 24 August 2025 (UTC)[reply]
I'm concerned that this is more of a pointy deletion....user for years has been unable to implement they preferred layout of vital articles being seen first. My concern is the vital article project is trying to take this over. I believe the vast majority of editors here see the vital article project as extremely unstable and not representative of our worldwide viewership. Moxy🍁20:07, 24 August 2025 (UTC)[reply]
I didn't really even know about this page. From the UseModWiki dumps, it dates back to January 2001. It has me thinking, how would you design a useful contents page for this whole project? It seems like something that ought to be in the sidebar, if it could be made into something nicer. Makes me want to draw concept art. 3df (talk) 07:02, 10 September 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
E.g. I am viewing Maheboob Khan, and click the link "Maula Bakhsh" which bring me to Dutch Wikipedia article. Similarly, the link "Maritim Hotelgesellschaft" in Maritim Travemünde links to German Wikipedia. This has two disadvantages: (1) It clearly violates WP:ASTONISH; (2) This does not encourage creation of articles in English Wikipedia. So I propose we should only link to foreign Wikipedia via {{Interlanguage link}} and create an AbuseFilter to prevent users from adding direct links to non-English Wikipedia in articles. GZWDer (talk) 01:56, 31 August 2025 (UTC)[reply]
Comment - This is already deprecated per H:FOREIGNLINK, which explains that you should use the inter-language link template so that the English Wikipedia article for it shows up as a redlink which is accompanied by a smaller bracketed link to a foreign language article. I've gone ahead and fixed the issue with the article.--JasonMacker (talk) 03:13, 31 August 2025 (UTC)[reply]
I was surprised to see this, because it's been officially discouraged for years. @GZWDer, do you really mean to be asking if someone would figure out how many of these errors exist and fix them for you, or were you just trying to get a rule written down, even though nobody will read it? WhatamIdoing (talk) 04:53, 31 August 2025 (UTC)[reply]
WP:AFD is neither a policy nor a guideline. WP:RFC is neither a policy nor a guideline. WP:BRD is neither a policy nor a guideline. WP:5P is neither a policy nor a guideline. WP:TE is neither a policy nor a guideline. Editors want to do the right thing, and they don't require good advice to say "policy" or "guideline" at the top of the page to do the right thing anyway. WhatamIdoing (talk) 20:13, 31 August 2025 (UTC)[reply]
Not sure if interwiki links are explicitly "deprecated" by H:FOREIGNLINK, but the ill template is the best practice and avoids the issues raised. What gets tricky is edits such as Special:diff/1308072402, which places the link somewhere a template may break things. CMD (talk) 05:21, 31 August 2025 (UTC)[reply]
Oppose per Thryduulf. We shouldn't get in the way of adding links to other Wikipedias when relevant. Of course, I don't have any objection to banning links like the one you propose (I agree with your references to ASTONISH and RED), but if we go creating an abuse filter, it won't be able to distinguish between good and bad. Nyttend (talk) 09:51, 31 August 2025 (UTC)[reply]
Oppose. This does not seem to be a big enough problem to justify an edit filter, which can only be edited by a few people and needs testing. The use of {{Interlanguage link}} is already listed as best practice by H:FOREIGNLINK and it doesn't matter a bit whether it is called a policy or a guideline. What matters is that it's a good idea. Phil Bridger (talk) 10:25, 31 August 2025 (UTC)[reply]
Comment I like this idea but I'm not exactly sure what is being proposed. My complaint is that articles should be internally linked to internal wikipedia articles. What I mean is that English wikipedia articles should only link to other English wikipedia articles. Logoshimpo (talk) 11:05, 31 August 2025 (UTC)[reply]
We also link to Wiktionary and Wikisource. We don't do this very often, so you might not have seen it, but one hardly wants to dumb down our writing ("but readers won't know what that word means!"), and sometimes a link to the dictionary definition makes a good compromise between brilliant prose and helping readers understand it. Wikisource links tend to be for non-notable historical documents ("issued the 1789 proclamation for the election") or in a list of works (* "The Bluebell" by Anne Brontë). WhatamIdoing (talk) 20:22, 31 August 2025 (UTC)[reply]
I'd support upgrading a slimmed-down version of H:FOREIGNLINK to guideline status or adding it to the MOS, but I strongly oppose an addition to the AbuseFilter as overkill for a really minor problem. Toadspike[Talk]17:57, 31 August 2025 (UTC)[reply]
Thanks, I missed that. We should consider upgrading "may be helpful" to "should be used" based on the strong preference for the Ill template here and elsewhere. Toadspike[Talk]14:07, 9 September 2025 (UTC)[reply]
Oppose articles in other language Wikis are useful in the absence of an English article since translation programs, however flawed they are, are available and better than nothing.--Wehwalt (talk) 19:56, 31 August 2025 (UTC)[reply]
That's exactly the role of {{ill}}: it does give a bluelink to the existing article in the other language, explicitly identified by what language it is, but it also has a redlink to the enwiki article. That means readers can immediately read (and decide if they are able to read) but there's also a tracked inbound link and an easy route to creating the article (for example, de novo or by translation). DMacks (talk) 02:14, 1 September 2025 (UTC)[reply]
Oppose, seems to already be covered by the MOS and Help. No need for more instruction creep, and there are potential reasons to allow such links. —Locke Cole • t • c00:26, 2 September 2025 (UTC)[reply]
Oppose. there are many times when linking to a non-English Wikipedia entry is actually the best option, as we don't always have corresponding articles. The template is the best solution, but adding an abuse filter seems unnecessary. Firsfron of Ronchester00:45, 2 September 2025 (UTC)[reply]
Oppose abuse filter; per Toadspike, I would support strengthening of language at MOS:IWL with more precise directions for the use of {{ill}}. In my opinion, {{ill}} should always be used when linking to a non-English wiki. Wrackingtalk!01:59, 2 September 2025 (UTC)[reply]
Except in rare cases, it's my impression that if {{ill}} isn't used, it's not for a lack of directions. It's because the editor didn't know that the option existed, and they'd be happy to have someone add it for them. If we could find a way to locate 'missing' uses of the template, and had a volunteer to fix them, then I would expect that to be welcome assistance. WhatamIdoing (talk) 02:24, 2 September 2025 (UTC)[reply]
Fair enough. I agree that editors not knowing about {{ill}} is probably a primary cause of this, and a way to locate 'missing' uses of the template would be great. Wrackingtalk!02:37, 2 September 2025 (UTC)[reply]
{{ill}} is also a real pain to use. Highly non-intuitive. I have to consult the documentation every time I have to use it. Usually takes a couple of goes to get it right. Hawkeye7(discuss)02:43, 2 September 2025 (UTC)[reply]
+1 all of this. I think something like {{ill|langcode:article}} would rock, exactly matching the style of [[:langcode:article]] interwiki links (and the natural extention of piping, as {{ill|langcode:article|displaytext}} for [[:langcode:article|displaytext]]). Or {{ill|:langcode:article}}, which probably makes it trivial to program (a "if first parameter has leading colon, parse off the first colon-delimited string and push it into the second parameter" wrapper converts it directly to the currently-handled syntax). DMacks (talk) 02:57, 2 September 2025 (UTC)[reply]
I would object to introducing such complexity across every use of this template when most of the usages probably won't use this complex feature (complex as it involves string operations). How about instead introducing a new template that uses this syntax? Aaron Liu (talk) 01:05, 3 September 2025 (UTC)[reply]
Nearly all uses of the template link to a single language code, but the template is optimised for multiple languages. So we can have things like: {{ill|Charles Darwin (botanist)|lt=Charles Darwin|fr|Charles Darwin|de|Charles Darwin|es|Charles Darwin}}. This illustrates the difficulty remembering how it works. The first guess of most users would be {{ill|fr|Charles Darwin}}. That is wrong; the English text has to come first. And while the editor might expect the next parameter to default to the display text, this actually requires |lt=, an abbreviation I'm not aware of being used anywhere else. Hawkeye7(discuss)01:43, 3 September 2025 (UTC)[reply]
If I were writing the parameters from scratch, I'd have chosen something like {{ill|target=local link|text=display text|lang=code|article=other wiki article|lang2=code|article2=other other wiki article}} (e.g. {{ill|target=London|text=Capital of the United Kingdom|lang=cy|article=Llundain|lang2=fr|article2=Londres}}) as those are names and ordering that intuitively make sense to me. Whether they made sense to anybody else I'm not sure but the ordering seems to match what others are suggesting. I don't have a clue how easy this would be to program. Thryduulf (talk) 03:41, 3 September 2025 (UTC)[reply]
I don't see why that means we can't have a lang alias for just the first language.I agree that "lt" is weird. I'd use display, text, or something. Aaron Liu (talk) 16:25, 3 September 2025 (UTC)[reply]
Can't an edit filter not just disallow an edit, but also tag an edit? Not sure how hard/easy it is to add a custom tag, but if there were a way to filter recent changes to show edits that added an interwiki, it might make what you're thinking of easier to track. =) —Locke Cole • t • c04:31, 2 September 2025 (UTC)[reply]
Answering my own question, but yes, from WP:EFBASICS, denying altogether is an option, but so is warning, tagging, or simply logging. Warning might not be a bad option, if the warning is customized to let the editor know about the interlanguage templates, it could be both educational and a way to remind editors in case they forgot. —Locke Cole • t • c04:44, 2 September 2025 (UTC)[reply]
I did a search and found that this was an edit filter before (780); it was disabled in 2016 by MusikAnimal (Too infrequent and for the good-faith edits I'm not too happy with even throwing a warning) Also, I left a note at WP:EFN to alert watchers of this discussion.Wrackingtalk!05:27, 2 September 2025 (UTC)[reply]
Correcting myself. It seems 780 (hist·log) was meant to catch attempts at connecting different-language, same-topic articles via a link at the bottom of the article (also called interlanguage linking); this is now handled via Wikidata. The issue we're discussing here is the in-text linking of non-English wiki articles instead of using {{ill}} template (per MOS:IWL, WP:ASTONISH). Wrackingtalk!05:42, 2 September 2025 (UTC)[reply]
Oppose per WP:CREEP and WP:IAR. Note that there are Wikipedias for Simple English and Scots which would not be especially astonishing. Fixing up such usage with templates or whatever is routine work for gnomes. Andrew🐉(talk) 05:42, 2 September 2025 (UTC)[reply]
I disagree - anytime a link sends the user outside en.wikipedia.org this should be clearly messaged before the click is made. There's no reason to have exceptions to this principle. (I for one would find it very astonishing to find myself at the Scots(?) wikipedia w/o prior heads-up). Regards CapnZapp (talk) 08:12, 2 September 2025 (UTC)[reply]
Comment What does "deprecate" mean? I thought this was already discouraged. I don't think trying to actively prevent users from direct links is useful. Generally, spending effort making it hard or impossible to do wrong is futile; we should instead trust users to do the right thing. We can always revert/admonish/ban mistakes and errors after the fact. Discouragement should be sufficient. Does this mean I'm opposing? Then so be it. If this proposal's "deprecation" includes a suggestion to make sure our discouragement is clearly communicated, however, I'm in support. CapnZapp (talk) 08:08, 2 September 2025 (UTC)[reply]
Even if you considered it "already discouraged", there are still many such links existing and should potentially be replaced with templates. GZWDer (talk) 12:59, 2 September 2025 (UTC)[reply]
GZWDer That's not the proposal discussed here? Certainly not ALL such links should be replaced. When context makes the out-of-English-Wiki link obvious, {{ill}} is less necessary. {{ill}} is less desirable on pages with very many links. So if you want to discuss how to find them so you can manually convert them, go ahead, nothing is stopping you. Cheers CapnZapp (talk) 11:25, 3 September 2025 (UTC)[reply]
Declaring something to be unwanted doesn't help us find and fix the problems. I realize that since I spend so much time working on policies and guidelines, it surprises people to hear me say it, but Wikipedia:Nobody reads the directions. This is just the reality we work in. It often takes two years for editors to notice changes even to the highest-traffic policy pages. Writing down the answer is IMO good, and it's helpful to the rare person who goes looking, but seriously: Policy and guideline pages are not magic pixie dust. Not one of us has actually read them all. Need proof? This whole discussion began because someone thought we needed a rule that has already been documented in at least two guidelines and one help page for years.
WhatamIdoing First off, maybe you're responding to GZWDer and not me? (Your first sentence appears directed at me; your last at them) Anyway, of course nobody always read the rules. That doesn't mean we should always (or even often) back up the rules with things like edit filters that physically stop editors from breaking them. Is this a case where strong measures are warranted? I say no. Yes, I've found cases where an "invisible" off-en-wiki link was inappropriate, but... then I fixed it. This is not on par with, say, how we ask IP editors to solve a captcha before they can add external links to articles. CapnZapp (talk) 11:30, 3 September 2025 (UTC)[reply]
I would be happy to have stronger wording that discourages linking to other Wikipedias by means other than {{ill}} and would support some project to systematically search and eliminate such links (for example, insource:/\[\[\:de\:/ helps search for German Wikipedia links; I also use CSS to underline these links so I notice them more easily), but I can imagine some places (like tables where the linking is explained outside of the table) where they are appropriate, so I hesitate to support this as written. —Kusma (talk) 10:21, 2 September 2025 (UTC)[reply]
Support in principle. This should not be blocked by an edit filter, but if an EF can flag and/or log such additions so they can by fixed by the gnomishly inclined, that would be great. older ≠ wiser11:46, 2 September 2025 (UTC)[reply]
Playing around with the insource: search code above, it looks like this might affect something on the order of 1% of articles (~70K articles). If we assume (for nice round numbers) an average of 10 edits per article, that suggests that it might tag one edit per day. Would 1–10 edits per day be worth flagging? Is anyone here committing to watching for and fixing the flagged edits? WhatamIdoing (talk) 17:53, 2 September 2025 (UTC)[reply]
This is definitely something I would watch for. I already use {{ill}} quite a bit in normal copy-editing when I come across underlinked articles or articles with existing non-enwiki links. Wrackingtalk!20:29, 2 September 2025 (UTC)[reply]
If an article change on my watchlist was flagged, I'd likely check it out. I'm not sure I'd going hunting them down in other articles. older ≠ wiser10:19, 3 September 2025 (UTC)[reply]
Opppose, looks like we have enough guidance already, and to physically block with edit filters is overkill. I'd rather see a bot that converts bare links to {{ill}} as a solution. -- GreenC18:22, 2 September 2025 (UTC)[reply]
Would you be OK with the an edit filter for the purposes of warning (to let the editor know, or at least remind them, of the template method of linking in article-space) or tagging the edit? —Locke Cole • t • c18:33, 2 September 2025 (UTC)[reply]
Oppose Mostly a non-problem; such links are very rare. They appear in a different color from en links, which mitigates the astonishment problem. No need to add more rules to ban a rare but occasionally useful tool. --Trovatore (talk) 18:46, 2 September 2025 (UTC)[reply]
I think this may be due to a skin. Non-enwiki links appear normal blue to me, so opening them and landing on a non-English page is astonishing. Wrackingtalk!20:10, 2 September 2025 (UTC)[reply]
Hmm. I use Monobook with a couple of customizations, but I don't think either of them affects this behavior. Maybe we could make the different color the default? --Trovatore (talk) 20:48, 2 September 2025 (UTC)[reply]
Comment on {{ill}}. I wasn't really familiar with this template, but I just looked it up, and to be honest I'm not convinced it's a good solution. Apparently it auto-converts the link to an en.wiki link when an article of that title is created. The problem I see with that is that there's no way of predicting in advance what the en.wiki article will be about. If, say, the title is a false friend, or just a different meaning, then all of a sudden you have a link that goes to an inappropriate article, and there's no human intervention to notice it. In general I think foreign-language links are not such a bad thing as some people seem to think. We should give readers credit for being able to deal with seeing stuff they can't read. It's not like they'll be in a situation they don't know how to deal with. Either they can read it, or they can't read it but they know they can't. In the latter case, they either try to find some way of understanding it, or they understand that this link is not useful to them, and no big deal. --Trovatore (talk) 18:54, 2 September 2025 (UTC)[reply]
I am a frequent user of {{ill}} and I love it. Indeed it requires some care to prevent wrong links, but so does just wikilinking a word in the article. The invisible automatic changing to an enwiki link is followed up by a bot edit that will show up on the watchlist.
The main downsides of direct interwiki links are that they are almost the same colour as enwiki links so there is no warning where you'll end up (I fix this in my user CSS by adding underlines), and that the links do not turn from red to blue when a suitable enwiki article is created, so they tend to stay in articles despite better targets being available. —Kusma (talk) 20:07, 2 September 2025 (UTC)[reply]
It really seems like this is more a problem with the default skin than it is a problem with direct interlang links. Could we get in a request to make them more distinct in the next update? --Trovatore (talk) 20:58, 2 September 2025 (UTC)[reply]
I don't know what interwiki links look like in the default skin (I use Monobook). Changing the link colour won't make people notice when an article matching the foreign one has been created though. —Kusma (talk) 21:09, 2 September 2025 (UTC)[reply]
I also use monobook and it comes out in light blue, whereas en.wiki links come out in dark blue.
True about the article-creation issue, but again, I don't see the foreign links as so bad, so this doesn't seem like a deal-breaker to me. The link can be upgraded in the normal review process. On the other hand links to the wrong article are bad; those are actual errors, not just a missed opportunity to link an article in English. --Trovatore (talk) 22:05, 2 September 2025 (UTC)[reply]
Depending on lighting conditions, I don't see the difference between the two shades of blue without effort. The"wrong link" issue you mention is something that can happen with every red link on Wikipedia and is not related to interlanguage linking. —Kusma (talk) 22:27, 2 September 2025 (UTC)[reply]
(edit conflict) There's already T333577, which is likely to continue being ignored. As for colors, looks like the default colors for unvisited links are
The problem I see with [auto-conversion of ill's by the bot] is that there's no way of predicting in advance what the en.wiki article will be about... all of a sudden you have a link that goes to an inappropriate article
You say you weren't familiar with the template, so as someone who has added hundreds of interlanguage links and seen many thousands more, I can say that this has not been a problem, or at least, no worse than people adding direct wikilinks to the wrong article. Both occur (rarely); both are annoyances; neither is a valid reason to deprecate ill's or autoconvert wikilinks. If there is a real problem with interlanguage links, it is that since by definition the English article does not exist yet, the ill creator has to invent an English title for the red link, and it can happen that two people each create an ill about the same thing with two different red link titles for the future English article. That does happen sometimes, but it is more of an inconvenience than a problem, analogous to two editors creating direct red links to the same topic and calling it two different things. If and when an English article is eventually created from one of the two ill red links, it usually gets shaken out eventually, either by the other one getting fixed, or becoming a redirect, and then eventually Cewbot comes along and culls it. So, you needn't worry too much about this case, as it is not a serious problem, and we needn't consider it as part of this Rfc. Mathglot (talk) 23:10, 7 September 2025 (UTC)[reply]
Comment - So if this is enforced by an abuse filter, will and/or how will this affect edit summaries that use those links? Usually for translation attribution these links are needed, so if this is disallowed how will this be done? The {{Ill}} template does not work in edit summaries (or any template I'm pretty sure). This could also very well be fixed in the abuse filter itself and I just don't know it, but I'm curious regardless. Sophisticatedevening(talk)19:36, 2 September 2025 (UTC)[reply]
Oppose – This is basically a 'best practices' issue, but forbidding one practice is not the best way to encourage a better one. I share your concerns upon clicking a blue link that takes you unexpectedly to nl-wiki—I don't like that, either. I am also a huge supporter of {{interlanguage link}}, so we see eye-to-eye there as well. Nevertheless, I must oppose your proposal; it is not necessary, the remedies are draconian, and there are better ways to deal with this. Additionally, you haven't considered some of the possible reasons for keeping them around, or alternative remedies. I also oppose it on the grounds of instruction creep; existing guidelines suffice for this. The fact is, that interlanguage links are part of Wikipedia, and there are legitimate uses of them. You did not mention their use in citations, for example. There are other situations where they might be useful, for example when the link is called out in running text as a foreign link and therefore the ASTONISHMENT goes away; e.g., at Kemperplatz#Sources or any of these articles.[slow link] Presumably you would make an exception for those, and there may be other cases like that as well. As has been said, many users of direct interlanguage links may be unaware that template {{ill}} exists, and forbidding use of the former is not the best way to promote the use of the template; that should be achieved with the carrot, not the stick. Some here have mentioned difficulty in using the {{ill}} template, and to the extent possible, that should be handled by improving the template doc. One positive step that could be taken that I have not seen mentioned thus far, is to trap new interlanguage links as they are typed and pop up a dialog box, similar to the disambig popup dialog that appears in edit mode when you type a link that is a disambig page. A solution like that would probably go a long way to minimizing growth of this problem, and WP:AWB could find and assist in repair of such links. Forbidding them is overkill, and not necessary. Mathglot (talk) 22:49, 2 September 2025 (UTC)[reply]
I don't want to click on your expensive link, but I'm curious about how many articles were found in that search.
Could we do a bot run to tag the links (skipping anything inside a template or otherwise feeling like it might break something)? It wouldn't necessarily have to be visible, but if we could get a group together to sort through these, like Wikipedia:WikiProject Disambiguation handles dab links, then we might be able to solve the problem. WhatamIdoing (talk) 00:35, 3 September 2025 (UTC)[reply]
There were 144 results but it is a lower bound, and probably only a small minority of the total, as I was looking for a particular pattern likely to yield on-page translation attributions similar to the Kemperplatz case. That said, every little bit helps, I suppose. (P.S., that search can be slow, but I don't believe it counts as expensive, as it includes a double-quoted insource term, which works off an index and is described as "ideal" here. That might be a good question for VPT, though.) Mathglot (talk) 01:07, 3 September 2025 (UTC)[reply]
Then I think it's missing most of the articles, because a simple search on insource:/\[\[\:de\:/ turns up about 20K articles, just for the one language. WhatamIdoing (talk) 01:38, 3 September 2025 (UTC)[reply]
It's only part of Mathglot's argument, but yes, I don't think there are many articles that link to other-language Wikipedia articles qua articles. And many that do should not – Kemperplatz § Sources treats a German Wikipedia article as a source, in violation of WP:CIRCULAR. Such uses should be replaced with {{translated page}} on the talk page and possibly a maintenance tag like {{expand language}} on the article. jlwoodwa (talk) 04:16, 3 September 2025 (UTC)[reply]
No, this discussion is not related to only those ~144 articles; the 20K for dewiki (and other non-enwiki) is what we're talking about. If you click the "slow link", it shows articles that link to non-enwiki articles andself-reference those articles, e.g., The German Wikipedia article [[:de:Konkordienbuch]] (at Book of Concord).
For Spanish, insource:/\[\[\:es\:/ returns 13K results. Not all are especially problematic—I wouldn't consider eswiki links in citation templates to be high-priority, e.g., |publisher = [[:es:Universidad Empresarial Siglo 21|Universidad Empresarial Siglo 21]] (at Jimmy Wales). Here is an example of a cleanup I did at Mexico after finding eswiki links via that search pattern. Wrackingtalk!05:36, 3 September 2025 (UTC)[reply]
Comment. I believe {{ill}} is expensive, and so one use case for other solutions I would bring up is possibly when the sheer number of links rule out using ill? CapnZapp (talk) 11:35, 3 September 2025 (UTC)[reply]
If seeing something written in a foreign language without being warned in advance is astonishing then I don't know what is non-astonishing. Some people must go through their whole lives being astonished. Phil Bridger (talk) 11:59, 3 September 2025 (UTC)[reply]
I don't think anyone here is saying seeing something written in a foreign language is astonishing. What many might find astonishing (or confusing) is to click on a wikilink that looks identical or nearly so to any other ordinary wikilink and being taken to an article in another language (and with the concomitant navigational framework in the other language as well). older ≠ wiser14:43, 3 September 2025 (UTC)[reply]
Comment I see several comments above discussing how to find the problematic links. Here are some possibilities, in approximate order of escalation:
People manually do searches like insource:/\[\[\:es\:/ for every language code.
Someone could re-enable CW Error #68 for enwiki (cf. the report for eswiki), which interested editors could use to find pages to clean up.
A log-only edit filter could be created, and people could use that log to find edits to follow up on.
An edit filter could add a tag to edits, without warning or preventing them. Interested editors could watch for edits with the tag for review.
An edit filter could warn users against adding such links, possibly teaching them about {{ill}}.
An edit filter could prevents edits that would add such links.
Seems like there's not much support for the last, and probably not a whole lot for the second-to-last either. But some arguing against seem to be overlooking the earlier options. Anomie⚔14:06, 3 September 2025 (UTC)[reply]
You missed out the most likely scenario - that people spend far more time than it would take to enact any of those proposals on discussing the "issue" here and end up being too exhausted to do anything about it. Phil Bridger (talk) 14:23, 3 September 2025 (UTC)[reply]
Good list @Anomie. Another possibility is to more clearly differentiate links to other language wikis. Unfortunately, as you've said, T333577 doesn't appear to be going anywhere. But perhaps, to also address concerns about {{ill}} being 'expensive' as well as cases where the link shouldn't change when an English article is created, there might be another template specifically for creating stable links to other language articles with color coding or other visual indicators. Perhaps include mention of the language, such as (German article: Bundestag) somewhat similar to how {{langx}} displays the language label. older ≠ wiser15:13, 3 September 2025 (UTC)[reply]
Having the link show a different color when you happen to read the article doesn't help an editor find articles that have this problem. I like the idea of re-enabling the WP:CHECKWIKI report. It'd probably be a good idea to find a couple of editors who are willing to work on the list before running it frequently. WhatamIdoing (talk) 16:58, 3 September 2025 (UTC)[reply]
Maybe if the color were more distinct it wouldn't be so much of a "problem". I'm still not convinced there's anything wrong with interlang links per se. --Trovatore (talk) 18:59, 3 September 2025 (UTC)[reply]
I'm not sure that any amount of fiddling with the color will work. On the one side, you have people saying "It must be different – really, really different, so it's super obvious that it's weird and different and not normal". When we do that, then other people appear and say "Now it doesn't look like a link! Nobody knows they can click on it. It just looks like someone randomly decided to put {green|purple|orange} text in the middle of an article for no reason at all!" (And most compromises are met with "It's basically unreadable in dark mode".) WhatamIdoing (talk) 19:03, 3 September 2025 (UTC)[reply]
Looking at Anomie's table, I think the Monobook dark contrasted with the Timeless light would work pretty well. Dark mode looks atrocious to me but I don't see that this would be any less readable in dark mode than dark mode is anyway. --Trovatore (talk) 22:00, 3 September 2025 (UTC)[reply]
Again: no amount of fiddling with the colors helps editors who are starting at the Main Page figure out which articles contain potentially unwanted links to non-English Wikipedia articles.
What's wanted – assuming we're trying to fix this problem – is "Here is the list of articles that need {{ill}} added to them". This is not a common enough problem that we can change the color and then tell people to click Special:Random a few hundred times and manually scan the page for the color change. WhatamIdoing (talk) 17:45, 14 September 2025 (UTC)[reply]
Oppose an edit filter (way too blunt), but support codifying {{ill}} as best practice, which is already de facto the case but isn't strongly codified yet. Indeed, guidelines should follow and codify community practices. It still leaves leeway for edge cases or people not knowing the guideline, as it would only mean that regular interlanguage links can be converted to {{ill}} (when reasonable) by later editors. Changing link colors to better distinguish interlanguage links is also a good thing, although such a visible change should probably be a proposal of its own. Chaotic Enby (talk · contribs) 14:08, 7 September 2025 (UTC)[reply]
Oppose all. The English Wikipedia already has a severe lack of information about non-anglophone countries, and any form of this proposal would just make matters worse. James500 (talk) 05:14, 10 September 2025 (UTC)[reply]
Bot to make list-defined references editable with the VisualEditor
Proposal: a bot that replaces {{reflist|refs=...}} with <references>...</references>
Rationale: The reason is that there are issues with the list-defined references that are based on the template {{reflist}}. The VisualEditor can't parse references that are inside templates. This comes from a design decision, it has been like this for over 10 years and developers apparently don't want to change it (see T52896). It means that in the VisualEditor, list-defined references that are within a reflist template can't be modified, and are not displayed (you instead get the message "This reference is defined in a template or other generated block, and for now can only be previewed in source mode"). So when using the VisualEditor, you can't visualize these references and you have to switch to the source editor in order to manipulate these references. However, the parsing works with list-defined references that use the <references> template. Here is an example of an edit that makes the replacement.
One more important benefit I just discovered: references within <references> can also be parsed by Wikipedia's translation tool, unlike those inside reflist. Which means that list-defined references within <references> are not removed by Wikipedia's article translation tool, and are instead properly converted to the corresponding reference format in the target language.
Technical details: This proposal is not about replacing all instances of {{reflist}}, only some of those that have the parameter "refs", which represents more than 55,000 articles. Replacing {{reflist|refs=...}} with <references>...</references> does not affect the rendering of the references (outside the VisualEditor) if the reflist only has a parameter |refs=. The reflist template can have some other parameters, like |colwidth=. One possibility is to apply the change if |refs= is one and the only parameter in the reflist template. This is safer and should still cover most instances of the problem. Another possibility would be to have a more sophisticated bot that can handle reflists with the parameter |refs= and some other additional parameters, which would allow for more replacements.
Previous discussions:
There was a long discussion in Wikipedia talk:Citing sources on this a few months ago. The discussion was initially about deprecating list-defined references (which didn't get consensus), and then switched to replacing {{reflist|refs=...}} with <references>, here of one of the paragraphs of the closing comment:
"There was 2:1 support in favor of deprecating {{reflist|refs=}} and replacing existing instances. I updated the linked documentation pages to do so. Someone will need to write a bot and follow the procedure at Wikipedia:Bots/Requests for approval. At least one editor had concerns about bots making incorrect edits. There was also discussion of whether or not such changes should be bot-flagged so they don't show up on watchlists, and whether it should be required that other changes be made at the same time. The bot approval process is designed to take these concerns into account and balance them against the proposed benefits; that would be the place to raise them. (It might be helpful if whoever makes the requests notifies the editors who participated in this discussion.)"
Support. This maintains list-defined references in that article while solving other problems. I think that this discussion is overkill, but BOTREQ frequently wants to make double-extra-certain that a bot task won't need to be reverted later. WhatamIdoing (talk) 22:16, 3 September 2025 (UTC)[reply]
For clarity: This will affect less than 1% of articles, and it will change this:
I have two questions: (a) Is the rendered result is perfectly identical what's being replaced? (b) Will it interfere with anything else contained in {{reflist|refs=}}? I employ HTML comments amongst list-defined sources for myself and others and want to be sure any bot changes won't strip or modify them. Thanks, all. — Fourthords | =Λ= |23:01, 3 September 2025 (UTC)[reply]
(a) Yes, it's exactly the same. AIUI the reflist template is just calling the <references>...</references> tags. (b) It shouldn't touch anything else except the exact characters planned to be replaced. (Why would it?) WhatamIdoing (talk) 23:12, 3 September 2025 (UTC)[reply]
Comment. If we are going to do something like this with the rationale that we cannot have refs inside templates, then we should more consistently examine contexts where refs are inside templates and eliminate all of them rather than piecemeal gutting our useful templates. Other templates that often have refs inside them include all of our infoboxes, {{efn}}, {{blockquote}}, and {{block indent}}. Should we perhaps discuss substing all of those? If not, why should this one context be treated exceptionally from others? —David Eppstein (talk) 23:17, 3 September 2025 (UTC)[reply]
In {{efn}}, {{blockquote}}, and {{block indent}} you generally have a text field in which you can see and edit the source code of references. It may not be ideal, but unlike for list-defined references, I don't see good alternatives and at least you don't need to switch to source editor. Alenoach (talk) 23:42, 3 September 2025 (UTC)[reply]
{{reflist}} has a field in which the source code of references appears. If VE is able to edit the source code of references in {{efn}} but not {{reflist}}, then the whole premise of this RfC (that VE is unable to edit refs in templates) would appear to be false. Are you suggesting that merely changing the type of the "refs" parameter of {{reflist}}, in the template parameter block of {{reflist/doc}}, from "String" to "Content", would allow VE to see and edit the source code of references in reflists? —David Eppstein (talk) 00:43, 4 September 2025 (UTC)[reply]
It might be able to see and edit them, but it still won't parse them so they won't show up on the "Re-use" tab when you insert a reference. --Ahecht (TALK PAGE)00:54, 4 September 2025 (UTC)[reply]
Does when I try it. Note there needs to be more than 10 references and your browser window needs to be wide enough. It also looks like Vector 2022's default settings (Medium text and "Standard" width) don't meet the "wide enough" criterion; I have to switch to Small text or "Wide" width to get it to wrap. Anomie⚔23:57, 3 September 2025 (UTC)[reply]
Hmm. Looks like the font size of the references themselves winds up the same with both {{reflist}} and <references />, but that's achieved with different CSS. The difference means that {{reflist}} applies the column-width of 30em based on the reduced font size, while <references /> applies the same 30em with the unreduced font size. Looks like T334941 exists already, although there they don't seem to have found this difference. Anomie⚔11:47, 4 September 2025 (UTC)[reply]
Is it possible to get tag-references to be responsive to more dense columns? I've found 30em far too wide in cases where short references are used. Ifly6 (talk) 00:25, 7 September 2025 (UTC)[reply]
Yeah, looks like the {{reflist}}30em is equivalent to telling .mw-references-columns to wrap on 27em because of that font-size:90% that reflist does.
On testing, even just knocking the default column width for <references/> down to 29em is enough to get us columns in Vector 2022 at standard width with the sidebars shown. Though obviously going a bit narrower makes it more likely that people with windows less-wide than the full "standard" width would still get the columns... DLynch (WMF) (talk) 01:46, 7 September 2025 (UTC)[reply]
Different order. reflist sets it on the wrapper-div that it creates, while core-references sets it on the ol that's inside the other wrapper that core uses for responsive columns. They both wind up reducing the font size that the actual references display at, but one affects the column widths and the other doesn't.
I would be fine with deprecation if the tag were sufficiently responsive for shortened footnotes. Most of the time it just picks the equivalent of |colwidth=30em when the output would save a lot of whitespace if it picked |colwidth=20em (or even |colwidth=15em) instead. Ifly6 (talk) 05:35, 5 September 2025 (UTC)[reply]
I have no idea. Replacing {{reflist}} with the references tag universally entirely would, I am under the impression, break three-column reference list formats. Ifly6 (talk) 20:25, 6 September 2025 (UTC)[reply]
Since this appears to be where the columns issue is actually discussed I'll repeat what I say below: I just implemented the {{refwidth}} thing I mentioned in the last discussion. It currently has three presets that correspond to 20em, 24em, and 27em respectively. An example of this in action is at User:Aaron Liu/sandbox#References.I had the idea after @Rjjiii also mentioned this concern in the previous discussion. For some reason I didn't see their reply, so I'll address it here: The refbegin thing is irrelevant as that isn't list-defined references. LDRs are defined in reference lists—the ones that are usually numbered and list uses of <ref></ref>—and the example based on the OSS page does not use them. See Help:List-defined references. Aaron Liu (talk) 20:40, 6 September 2025 (UTC)[reply]
Support: Although @David Eppstein is correct this doesn't solve every problem, that doesn't mean it doesn't solve a problem, and a pretty major one at that. This is a problem I run into often with the visual editor; I usually just have to change the reference format to move them out of refs to effectively edit the article. This would solve that, and only a modest # of articles would be affected, so I support it. Mrfoogles (talk) 00:01, 4 September 2025 (UTC)[reply]
Oppose. VE is broken and we should not extend the damage to the rest of Wikipedia by making our templates broken and clogging up our watchlists with bot edits as well. —David Eppstein (talk) 00:44, 4 September 2025 (UTC)[reply]
This it not completely specific to VE though: Wikipedia's translation tool somehow has a similar issue where it can't parse refs inside "reflist" and will just remove them. Alenoach (talk) 01:02, 4 September 2025 (UTC)[reply]
"VE is broken" doesn't really capture the issue here. The built-in MediaWiki references tag is how we used to do things, and the {{reflist}} wrapper is a hack workaround for a now-irrelevant issue. Templated wikitext is almost inherently impossible to parse, so mishandling of an expensive template is not a bug in VE but an inherent part of wikitext. Deprecating a redundant and useless wrapper for a built-in MediaWiki function is useful cleanup, even if VE handled it fine. Dan Leonard (talk • contribs)02:19, 4 September 2025 (UTC)[reply]
Support I'm of the opinion that we should deprecate {{reflist}} altogether for all instances where we're not using |group=, since that's the only thing where the bare tag doesn't have feature parity yet. Yes, the template allows you to specify a fixed column width, but we shouldn't be doing that anyway as the default responsive option is a more flexible and modern approach. Most of the other uses of reflist date back a decade or more to when the raw tag was much less capable. This is a good first step, and would be 55,000 more articles where editors can see an example of using the raw tag instead of the pointless wrapper. --Ahecht (TALK PAGE)00:47, 4 September 2025 (UTC)[reply]
@Dan Leonard{{reflist|group=lower-alpha}} actually formats the reference numbers as lowercase letters, but to do that with bare tags you need to do <divclass="reflist-lower-alpha"><referencesgroup="lower-alpha"/></div>. --Ahecht (TALK PAGE)18:06, 8 September 2025 (UTC)[reply]
This is something that's going to (sort of) solve itself as parsoid rendering rolls out, because parsoid is going to expose data-mw-group="lower-alpha" on the <ol>. Then wikis will be able to add their own site CSS on that attribute if they want. DLynch (WMF) (talk) 21:27, 8 September 2025 (UTC)[reply]
There are other problems with {{reflist}} than just VE compatibility, such as hiding all references if the page exceeds the WP:PEIS limit (the bare tag isn't affected by template limits, although some individual references may not render if they use a CS1 template after the limit is reached). --Ahecht (TALK PAGE)00:57, 4 September 2025 (UTC)[reply]
Support per above. Note that this proposal is only about replacing instances that define new references within the template and not all invocations of it (and even then, the fixed column width feature can be implemented through a separate templatestyles template that selects from one out of three widths). This template is being singled out among the templates that define refs because it's the most common one that would break the new subreferencing feature and changing invocations should not be very intrusive. Aaron Liu (talk) 01:39, 4 September 2025 (UTC)[reply]
Support. Seems like a sensible idea. I would strongly prefer that whatever tool or bot is used to do this should not make these edits in isolation, to avoid clogging watchlists. If this can be turned into something that's done as part of other edits, at least until the number of remaining articles to modify is much smaller, that would be best. Re the "fix VE" argument; sure, if that looked likely, but it doesn't, and there are more new VE users every day. There's no reason to make their lives harder for the sake of a principle. Mike Christie (talk - contribs - library) 02:16, 4 September 2025 (UTC)[reply]
Support deprecating {{reflist}} altogether except for instances where specific functions unique to the template are required for some reason. A lot of people, in this discussion and many others involving VisualEditor, love to use the refrain "our practices are fine, VisualEditor is broken". Even this proposal seems to think this is the VisualEditor team being lazy with this comes from a design decision, it has been like this for over 10 years and developers apparently don't want to change it. While the editor might have some bugs (the :0 ref naming scheme comes to mind), sometimes its limitations demonstrate an issue with our practices. Wikitext is essentially an unparsable language,[1][2][3] so it's almost inconceivably difficult for any WYSIWYG editor to handle the complexity of parsing <ref> calls hidden inside of a template call that is filled with conditional expressions. Not to mention how to handle that for a multilingual project where hundreds of language editions and sister projects have copied or adapted a similar wrapper. Sometimes, it's a good idea to deprecate templates that just wrap simpler basic wikitext functions. The community has shown distaste for templates that exist only to wrap basic wikitext syntax like boldface (Wikipedia:Templates for discussion/Log/2012 April 13 § Template:Bold) or italics (§ Template:Italic) and would certainly benefit from limiting use of a template that in 99% of cases just directly calls <references>. The bare Cite extension tag was how we used to use references, and {{reflist}} was only invented to wrap the tag in a <div> with columns. That became redundant in 2017 when the Cite extension got responsive design. Since then, there's almost never been a use for the template and its eleven-year reign should have ended. Instead it lives on as a zombie, mentioned in too many documentation pages and the muscle-memory of editors. As I mentioned at Wikipedia talk:Citing sources/Archive 58 § Sister projects have managed this, the Polish Wikipedia got rid of their wrapper template via a bot, and it has worked well for them. The VisualEditor is how people are learning to edit, for better or for worse, and it is crucial for this project's survival to meet our new editors with collegiality and help them productively edit the encyclopedia. Dan Leonard (talk • contribs)04:05, 4 September 2025 (UTC)[reply]
Italian Wikipedia also did a bot run to replace their {{reflist}} version [9] once there was no more need for it due to the introduction of <references responsive> in mw:Contributors (2015-2021)/Projects/Columns for references. My team (WMDE Technical Wishes) did an investigation last year (phab:T377043) to check which features of {{reflist}} equivalents across major wikis are still missing for <references>. Our main take away is basically that nowadays most features (except fore some rarely used ones and some right-to-left language specific features) are already part of <references> which many community members might not know. Johannes Richter (WMDE) (talk) 08:53, 4 September 2025 (UTC)[reply]
Thanks for the clarification! I would also support deprecating reflist, given your explanations. There could be a spin-off voting on this particular question to see if there is a sufficiently large consensus. But since it's less surgical, I expect that getting consensus would be harder. Alenoach (talk) 22:46, 5 September 2025 (UTC)[reply]
Comment, reposted from the Bot requests discussion: The monthly parameter usage report for Template:Reflist suggests that there are 183,000 articles using |refs=. It seems like any sort of replacement would need to start with a well-advertised RFC that successfully deprecated |refs=. This proposal (not an RFC) appears to be deprecating that parameter de facto, possibly without stating it explicitly. Maybe I missed it. – Jonesey95 (talk) 05:02, 4 September 2025 (UTC)[reply]
If this bot gets approval, I would just take it as given that the parameter should be deprecated, and then removed once all the substitutions are done. Otherwise we'll have to do ongoing work to remove instances that use it. This discussion is pretty visible, RFC or no. -- Beland (talk) 05:18, 4 September 2025 (UTC)[reply]
Suppport as proposed. I don't think it's a good idea to try to add unrelated tasks into the same bot, as Mike Christie has proposed. If the edits need to be rolled back due to malfunction, that will undo other useful edits, and the more tasks the higher the chances of malfunction. It would also delay and complicate the bot approval process unnecessarily. Part of the point of having a bot flag is so people can filter them out of their watchlists, to address complaints about cluttering those. So many bots are already running and making small changes, not to mention editors making small changes, the incremental annoyance caused by this run is pretty much negligible, especially considering what a tiny percentage of articles (0.78%) are affected. Also, long-term ease of use and its effect on editor retention are more important than avoiding a small one-time annoyance. I support letting the bot handle any additional parameters if they are present in, say, 100+ articles. My bias would be toward removing any customizations and going with the default rendering. For any low-frequency parameters we can review those articles manually and make sure the customizations aren't needed. I'd be happy to help with that. -- Beland (talk) 05:13, 4 September 2025 (UTC)[reply]
Supportprovided that no changes are made if {{reflist}} is invoked with |colwidth=. (Unless something is done to make <references> automatically select whatever column format that minimises its vertical space in 2022 Vector or something like that.) Ifly6 (talk) 05:18, 4 September 2025 (UTC)[reply]
I think it's better for Reflist to always be replaced as long as the |ref= parameter is present. On column widths, I just implemented the {{refwidth}} thing I mentioned in the last discussion. It currently has three presets that correspond to 20em, 24em, and 27em respectively. An example of this in action is at User:Aaron Liu/sandbox#References. Aaron Liu (talk) 15:50, 6 September 2025 (UTC)[reply]
I guess you mean specifying an arbitrary column size like |14em. To my knowledge, no, unless we implement inline styles for the <references> tag. Aaron Liu (talk) 20:55, 6 September 2025 (UTC)[reply]
Support limited to where |refs= is the only parameter in the reflist template. Like Mike Christie, I see no reason to hope for WMF improvement of this VisualEditor fault after a decade of inaction. ViridianPenguin🐧 (💬) 06:23, 4 September 2025 (UTC)[reply]
Support per above, and also support combining this work other task, if possible. A large portion of editors regularly use VE, especially new editors. This will help a lot IMO. ~/Bunnypranav:<ping>09:48, 4 September 2025 (UTC)[reply]
Support (weak). Many newer editors would benefit, and they are likely unaware of this discussion and the technical issue. This change would allow VE editors to see, reuse, and modify list-defined references. VE still would not support removing list-defined references. Also, the {{reflist}} template likely would never have been made for the current version of the "references" tag. Rjjiii (talk) 03:30, 5 September 2025 (UTC)[reply]
For what it's worth, this change happening would make it significantly more likely that we'd put time into fixing things like removing a list-defined reference. It's been low-priority because we knew that a large number of places that use list-defined references wouldn't be affected, but... DLynch (WMF) (talk) 20:48, 5 September 2025 (UTC)[reply]
This is tracked by issue T356471. I have pinged a developer to try to get some attention to it. I expect that fixing the bug for ref removal shouldn't be particularly hard, but they need to prioritize it. Alenoach (talk) 21:13, 5 September 2025 (UTC)[reply]
Support per Alenoach and Dan Leonard. Compelling proposal. It seems like ideally we should get rid of {{reflist}} altogether in favour of the native syntax, but this proposal is a good first step. – SD0001 (talk) 14:08, 5 September 2025 (UTC)[reply]
Support{{reflist}} being so widespread has been a major factor in us not investing time in improving VE's support for list-defined references in general. Ideally it'd be nice to remove the template altogether, because we'd still be left with needing a complex edit when someone wants to add the very first list-defined reference to an article in VE. It also won't help with one of our most common cases of uneditable template-defined references, which is those defined inside infobox parameters, but it'd be a step towards a simpler future. DLynch (WMF) (talk) 20:57, 5 September 2025 (UTC)[reply]
(I'd specifically throw a +1 behind all of Dan Leonard's comments above, which seem spot on to me about the general difficulties of dealing with wikitext-in-template-parameters.) DLynch (WMF) (talk) 21:20, 5 September 2025 (UTC)[reply]
Support. Making it possible to edit something is always a worthwhile goal, and worth the "downside" of bot spam. (A bot is not just improving a few articles, it is improving lots of articles!) And to be clear, the |refs= parameter should be eliminated so we don't end up in this situation again. I'd even go so far as to say {{reflist}} should be a subst-only template. My complaints about {{reflist}} can be saved for another day, but for now let's solve one problem at a time. HouseBlaster (talk • he/they)22:05, 7 September 2025 (UTC)[reply]
Oppose. Fix the actual problem where it lies, don't ask someone else to put up with a workaround.
*: [WMF, stage left] — Oh, Wikipedia, dear! We have this great VE tool, but we fucked up; it doesn't play nice with LDRs. Can you pretty please change the way you have always done it with {{Reflist}}, and pull our ass out of the fire on this one? Would really appreciate it; ta very much! *: [Wikipedia] — grumble... er, well, I dunno; sometimes I think I'm just too accommodating. All right, I guess... Just this once, then. *: [WMF exits stage right, smiling slightly] *: [later...] [Wikipedia] — Oh, WMF, there you are! You know, I really hate to pester you about this all the time, but there's this little thing about VE making inscrutable numeric citation names like <ref name=":17"> instead of something mnemonic like, say, <ref name="Einstein-1905">. I know it's been barely ten years since we first mentioned it, but, um, do you think... *: [WMF] — [coughs] Er, actually, I've been meaning to talk to yo about that, but I'm late, I'm late, for a very important date... [scurries off, stage right]
I've been editing for about 20 years and have created over a thousand articles. And I like list-defined references and have been using them since I discovered them comparatively recently. But the technical basics still seem obscure and unclear. What I'm still not fully understanding is the difference between:
Having the same keyword interpreted in different ways depending on the position of a slash seems quite arcane. And I'm not aware of any way of adding the latter pair using the Visual Editor. There still seems to be a long way to go to make this simple and easy for newbies.
In 20 years of editing Wikipedia, I have never used or even noticed <cite> and am not familiar with what it might do. So, I don't understand it at all. But I guess that it's a raw markup tag which is commonly buried in templates such as {{citation}} or {{cite book}}. I suppose that you have to be a real old-timer to have been exposed to Wikipedia's fundamental primitives.
More generally, I've not done much html or xml raw coding and so the slash syntax is not familiar. But now I check on it, it appears that closing slashes are somewhat deprecated in HTML5 in void tags, especially self-closing tags.
Editors writing articles shouldn't have to understand such technical tag issues. The problem in making life simple for them is that there are far too many ways of citing references – each with their own arcane and complex syntax. This violates the KISS principle.
I see. Well, I've used those various forms of <ref> but didn't understand that there was a general pattern to the syntax – I just copied what worked elsewhere. And with list-defined references, I've been using the {{r}} template rather than <ref> so that the inline citations are succinct. For example, rather than <ref name=BBC/><ref name=NYT/>, I put {{r|BBC|NYT}}.
As others have said, this is just regular HTML tag syntax that's already used similarly in a bunch of places. If you want, you could think of it as the difference between reflist and refbegin/refend.
The latter would just mean having a way to create list-defined references inside VE, since that's the only reason to have the open/close tags, so there's room to put something inside it -- and which is used is basically an implementation detail that VE should be able to switch between automatically depending on whether there's anything inside the tag.
This would be technically trivial, but there's UX complexities around asking a user whether the reference should live inside the list -- it's not a question that really makes sense to someone who's not editing the source, after all. We should probably work out what the community of a given wiki thinks is the "correct" place for references to live, and default to doing that. DLynch (WMF) (talk) 14:22, 8 September 2025 (UTC)[reply]
The correct place for references is where they appear in the article – down at the bottom in a section called References. Full inline references are nightmarish in the source code and that's why I quickly switched to using list-defined references when I discovered them. The short footnote ({{sfn}}) found in many featured articles is similar – listing the main sources in a section called Sources and then using short tags to cite them in the text.
But getting the community to agree on a simple uniform scheme won't happen easily because of WP:CITEVAR and the hard-core vested interests. The Visual Editor should focus on one good scheme and encourage that so that the new generation of editors grows up with a more uniform standard.
While some do find reading source code tricky, that isn't really an issue for visual editor editors. The big stumbling block for list-defined references is likely that such a reference system usually requires two edits instead of one, something visual editor also might be able to smooth out if it had some way to automatically shift the reference as part of the edit saving. CMD (talk) 16:14, 8 September 2025 (UTC)[reply]
VE already does some amount of automatic shifting-around, in the form of making sure that the definition of a reused inline ref is always attached to its first usage in the article. Attaching some similar behavior to add to attach the definition into the reference list wouldn't be a stretch. DLynch (WMF) (talk) 22:04, 8 September 2025 (UTC)[reply]
List-defined references are a minority on the site; I find it easier to have single-use references inline in the text so that nothing gets out of sync if they are removed or replaced. -- Beland (talk) 21:46, 8 September 2025 (UTC)[reply]
I'm personally a fan of list-defined references, particularly for anything that's being reused. If I was making completely arbitrary changes to the wikis to enforce my own aesthetic preferences, I'd probably make it so that <ref name="foo">only worked to reuse a list-defined reference. Alas, I'm not getting consensus on that one. ;-P
More-seriously, I think that VE has been working as it has for so long that suddenly switching the default behavior would require some thought. A starting point of working out an acceptable heuristic for VE that'd make it go with the flow to match whatever style is in use in an article, rather than always doing inline refs would probably not be too controversial, though. DLynch (WMF) (talk) 22:01, 8 September 2025 (UTC)[reply]
Support. This is a really minor change that would produce major benefits with no downsides. The opposition seems to consist of polemics against VE and misunderstandings of the technical purpose of the template being replaced. Toadspike[Talk]00:47, 10 September 2025 (UTC)[reply]
Support - I'd prefer if VE would fix their own bugs, but as the devs have been pretty consistent for a decade about not supporting use cases that they don't like and there isn't a visual difference, we might as well. Several hundred of those 55000 articles are from me, and I guess I'd rather have the page more fully work in VE then keep html syntax out of the wikicode. --PresN14:57, 10 September 2025 (UTC)[reply]
Support. Yes, VE is much better than it was 10 years ago, but it still has some bugs when it comes to things embedded in other templates. Since the developers aren't fixing the LDR issue, this should be a workaround. Not an ideal workaround, of course, but one that should be effective. Plus, there is already consensus to deprecate the |refs= parameter in {{reflist}}. Epicgenius (talk) 13:17, 17 September 2025 (UTC)[reply]
^Siebenmann, Chris (2011-11-12). "(Not) parsing wikitext". Wandering Thoughts. University of Toronto. Archived from the original on 2025-05-03. I don't know how to write a formal parser for it, something based on parsing theory in order to have all of the advantages of real parsers ... unlike a traditional language, wikitext doesn't have any errors
^Dohrn, Hannes; Riehle, Dirk (2011). "Design and implementation of the Sweble Wikitext parser: Unlocking the structured data of Wikipedia". Proceedings of the 7th International Symposium on Wikis and Open Collaboration. WikiSym '11. Mountain View, California. p. 73. doi:10.1145/2038558.2038571. ISBN978-1-4503-0909-7. Moreover, the generated HTML can be invalid. The complexity of the software also prohibits the further development of the parser and maintenance has become difficult. Many have tried to implement a parser for Wikitext to obtain a machine-accessible representation of the information inside an article. However, we don't know about a project that succeeded so far. The main reason we see for this is the complexity of the Wikitext grammar, which does not fall in any of the well-understood categories LALR(1) or LL(k).
^Wicke, Gabriel (2013-03-04). "Parsoid: How Wikipedia catches up with the web". Diff. Wikimedia Foundation. Archived from the original on 2025-08-13. There is no guarantee that the expansion of a template will parse to a self-contained DOM structure. In fact, there are many templates that only produce a table start tag (<table>), a table row (<tr>...</tr>) or a table end tag (</table>). They can even only produce the first half of an HTML tag or Wikitext element (e.g. ...</tabl), which is practically impossible to represent in HTML.
Match icons with Wikipedia skin (Vector 2010, Monobook)
For people using the Vector Legacy or MonoBook skins, there are quite obvious design inconsistencies between the rest of the skin and the new "flat" icons. In addition, many of these new icons were not designed for the scale of old Wikipedia skins and can look blurry.
The old icons are still hosted on Wikimedia.org at their old URLs; I propose that we match the old icons to these old skins one of these ways:
1) Check for and redirect users of a legacy (V2010, MonoBook) skin to the older icon URLs, or
2) Rehost the old icons under a new schema: wikipedia/en/ is already used for new icons, so have wikipedia/en/legacy for the older icons. Have the icon refs point to that directory instead for users of older skins. OmegaAOL (talk) 04:05, 5 September 2025 (UTC)[reply]
As a user of Vector 2010, I don't necessarily agree with this proposal. A lot of the reasons for using Vector 2010, such as the page layout, are independent of (or even despite) the skin's older style, and some users like myself do prefer the more modern look of flat icons. An alternate option would be to make the choice of icons separate from the choice of skin, which would suit both the users preferring V10 with flat icons and those preferring the old icons. Chaotic Enby (talk · contribs) 20:01, 8 September 2025 (UTC)[reply]
Change editnotice appearance for mobile contributors
Hi all. I've started editing on my phone recently, and I've experienced the annoyance of editnotices for editors on mobile devices:
They're disproportionate to the rest of the UI and imposing
Their icons cause the text to stack to sometimes a screen's length high
So I propose improving their appearance for mobile editors with responsive web design, particularly since mobile devices constitute more than 40% of all Wikipedia readers, and growing.
The proposed changes' visual effects can be seen in the before and after screenshots below, additions to the CSS can also be seen below, and the actual changes viewed on Template:Editnotice/testcases on your desktop and mobile device by enabling the mobile sidebar gadget in your preferences (or use Chrome DevTools to simulate it on your device); it can also be perused in its entirety with the Template:Fmbox CSS at Template:Editnotice/styles.css too. My changes also include updating the colors used from static Hex to global variables, so that they automatically change with dark mode.
Instead of making the table cells in the Editnotice display: block; which is a 'quick fix', it would be wiser to convert the HTML <table> layout used in Template:Fmbox (as well as all the other mboxes) to use <div> instead, then use flexboxes (display: flex;) to get the horizontal and vertical layouts as the layout would be more responsive.
Edit notice before adaptive CSS on iPhone 380px ✕ 844px.
Edit notice after adaptive CSS on iPhone 380px ✕ 844px.
@mediascreenand(max-width:639px){/* Target smartphones */.fmbox.mbox-image,/* Stack icon and text vertically */.fmbox.mbox-text,/* Make the font size normal */.fmbox.mbox-imageright,.fmbox.editnotice-header,.fmbox.editnotice-headerspan{display:block;font-size:100%!important;width:90%;}.fmbox.mbox-imageimg.mw-file-element,.fmbox.mbox-imagerightimg.mw-file-element{height:auto;width:auto;max-height:40px;text-align:center;}}@mediascreenand(max-width:300px){/* Target tiny screens */.fmbox.mbox-image{/* Hide the icon */display:none;}}
Instead of making the table cells in the Editnotice display: block; which is a 'quick fix', it would be wiser to convert the HTML <table> layout used in Template:Fmbox (as well as all the other mboxes) to use <div> instead, then use flexboxes (display: flex;) - agree, and IMO this should be done to all other message boxes too. Sapphaline (talk) 14:58, 6 September 2025 (UTC)[reply]
@Gnomingstuff: I hear you, and I agree that editnotices are important tools. The tricky part is weighing up the cost vs. benefit. Large, heavy editnotices might reduce vandalism, but they also risk discouraging good-faith editors (lowering editor retention) — and at some point the effect on vandalism bottoms out anyway. I don't think bumping the text size or icon slightly really changes that dynamic, since the notice still takes up a lot of space. We could always experiment with making the icon a bit larger (say 60×60px or more) and nudging the text size to 105%.
Your example I must disagree with though as it's a false argument, showing one example as the status quo is a hasty generalization and false equivalence. waddie96 ★ (talk)22:26, 6 September 2025 (UTC)[reply]
@Waddie96, I wonder if you could produce a mockup image that would help people understand what you're talking about. I doubt that anybody looks at the above screenshot and thinks "You know what's really great about edit notices? Having 25% of the screen be empty white space, with a blue dot most of the way down the screen, and all my important written instructions getting smushed over to the side in small text".
I don't understand this opposition. THEYCANTHEARYOU describes situations where editors can't see or understand things like editnotices. This proposal intends to repair serious HTML accessibility issues present in editnotices, thereby making them more readable to editors who might otherwise not understand them. Dan Leonard (talk • contribs)01:16, 8 September 2025 (UTC)[reply]
The original proposal took issue with edit notices being disproportionate to the rest of the UI and imposing. That's the entire point.
The upstream systems that care about message boxes only care about ambox, hence why Module:Message box/div/doc details a plan that makes ambox (and tmbox) last in the rollout. fmbox and cmbox should be pretty safe, but either way, that's why there's a staged rollout. Izno (talk) 21:13, 8 September 2025 (UTC)[reply]
It looks to me like the stop sign is slightly bigger (taller) than the others, but I believe these details can be changed. Also, I'm looking at this on a laptop, and the appearance on mobile (=66% of traffic) is going to be more important. WhatamIdoing (talk) 17:54, 14 September 2025 (UTC)[reply]
Yes, icon size and line-height we can adjust (I believe it may be line height I did not adjust, but sized down the text. Can size it down proportionally. But main thing is the concept for now.
What I think is most important is either coming across a community developer/contributor who knows if MediaWiki core has implemented some mobile adaptations for web (I know they have an entire interface etc. for mobile apps, that's fine)? Or anyone know of someone they can ask. Otherwise last resort, I can log a Phabricator ticket, asking if this is feasible, and ensuring it won't interfere with the mobile app implementation. Let me know what you think. waddie96 ★ (talk)00:34, 15 September 2025 (UTC)[reply]
Will the icons' appearance on Wikipedia Mobile (Web) be different from Wikipedia Mobile (App)?
Asking this because I viewed those icons on the web version, and wouldn't wanna give false inputs which might confuse you.
And we're discussing the web browser mobile app version yes, (I have used the app just not enough to confidently declare "I've used the app"), because the apps parse/process/sanitize the wikitext beforehand in some way I'm sure, for example the infobox is brought straight to the top just after one paragraph of lede (same as MobileFrontend), but also it doesn't go to the Main Page, it has it's own custom main page.
Hello, I was it would be really cool if entering the Konami Code did something crazy! I'm not exactly sure what, but it should be epic. Maybe on April 1st, it could make you an admin. 70.244.42.80 (talk) 00:27, 7 September 2025 (UTC)[reply]
Do people actually use game controllers to read Wikipedia? Where would they enter the Konami Code?
There would have to be a place for them to type the code. Think about it: if you are reading a Wikipedia article, and you accidentally press a few keys on your keyboard, nothing happens, right? It doesn't type those letters on the screen anywhere. Something would have to be 'listening' for the code. WhatamIdoing (talk) 01:45, 7 September 2025 (UTC)[reply]
It's also not 'typing the code'. If you actually push ↑+↑+↓+↓+←+→+←+→+B+A and then click the Search button, you'll end up at BA, because all the arrows get ignored by the search box. WhatamIdoing (talk) 16:53, 7 September 2025 (UTC)[reply]
Yeah, if you push the arrow keys that will happen but if you type the characters (which is easy for anyone with meta keys set up (e.g meta-> produces →) you get taken to the article. I presume getting mediawiki to interpret arrow keys as anything other than scrolling/navigating input boxes would require a very significant change to the code? Thryduulf (talk) 19:13, 7 September 2025 (UTC)[reply]
It would have to be done in Javascript, as that's the only way to capture keyboard events in a browser. The primary impact would be client-side, so for part of the same reasons that easter eggs became forbidden at Microsoft, I think we need to be cautious about deploying code for which most people won't see any net positive, particularly for those with slower devices. isaacl (talk) 21:53, 7 September 2025 (UTC)[reply]
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
This doesn't seem suitable for RM, even if I could get past some technical obstacles.
Oppose, the name "idea lab" is perfectly descriptive. I like its slightly whimsical title, which pairs well with the likewise fun "village pump" (we aren't a village and there isn't a pump). "Proposals development and filtering" is a mouthful that is also confusingly similar to this page. It also isn't even correct, the idea lab is a lab, not a filtering process. Anyone can still post here without having gone through the "filter" of the idea lab. Dan Leonard (talk • contribs)01:07, 8 September 2025 (UTC)[reply]
It also isn't even correct, the idea lab is a lab, not a filtering process. Recent discussion seemed to disagree. I'm too old, tired, and semi-retired to go find that. And development = lab. Labs develop things. I understand the appeal of "friendly" names as a design principle, but we're generally not targeting an audience who need friendliness. If one really does, they probably lack the experience to be making site proposals anyway. anyway. This is not WP:TEAHOUSE. This is where big boys and girls play, that's just a fact of life. We shouldn't discourage low-experience editors here, but we needn't encourage it, either. ―Mandruss☎ IMO. 01:16, 8 September 2025 (UTC)Edited after reply 07:19, 8 September 2025 (UTC)[reply]
Recent discussion seemed to disagree. I'm too old, tired, and semi-retired to go find that. If it was a filtering process, then wouldn't you have posted this proposal there first? Dan Leonard (talk • contribs)01:19, 8 September 2025 (UTC)[reply]
LOL. Foxes guarding henhouse. I suspect that would have been summarily executed without a last cigarette. Not even worth much discussion, they would say. Go away, essentially. Three editors in firm opposition, zero in support, nobody else sees a need to even comment unless the situation changes before archival. I've been there a number of times and speak from some experience. ―Mandruss☎ IMO. 01:25, 8 September 2025 (UTC)[reply]
And even if it isn't, why would friendliness be bad? Unnecessary, maybe, but all the alternatives have far more wordword. Just why? And what's wrong with "Idea"? Proposals are when you develop and write an idea out, so the new title is just wrong. Aaron Liu (talk) 18:12, 8 September 2025 (UTC)[reply]
Anyone can still post here without having gone through the "filter" of the idea lab. And nothing in this proposal suggests that they can't. It does suggest that, if they go to VPPDF first, it will be to (1) develop the idea (potential proposal), and (2) decide whether the idea is proposal-worthy (filtering). This could easily be explained at the top of VPPDF (and/or this page), and I'm volunteering to write that. ―Mandruss☎ IMO. 07:50, 8 September 2025 (UTC)[reply]
"Proposals development and filtering" is a mouthful - Not a problem. So don't put it in your mouth. Just refer to VPPDF or WP:VPPDF, as the situation requires. that is also confusingly similar to this page. - Disagree. Unconfusingly similar to this page. Part of the rationale is to show that the two pages are (somehow) associated, which today is completely invisible at first glance. ―Mandruss☎ IMO. 10:44, 8 September 2025 (UTC)[reply]
Oppose, since a concise title is best. This would also likely cause confusion with the title of this page: Village pump (proposals) and Village pump (proposals development and filtering) both have "Village pump (proposals)" as part of their name. "Idea lab" is more accurate and makes a clear distinction between the two sections. SuperPianoMan9167 (talk) 01:13, 8 September 2025 (UTC)[reply]
'Comment. I will take this opportunity to preemptively respond to an argument that has not been made, but surely will be before this is over.
Argument: But what about the other village pump page titles? They're friendly. Why should this one stick out??
Response: This has been near the top of my user page for years: "Perfect is the enemy of good. Don't oppose small improvements because they fail to solve the problem 100%; the result is usually no improvement at all. Small improvements can be followed by other small improvements." Sure, we could expand the proposal to "un-friendly" all pages for consistency with VPPDF. I can't think of a better way to commit proposal suicide. That consistency is not critical. As a factor in this, it might have a little weight. Key word little. ―Mandruss☎ IMO. 07:34, 8 September 2025 (UTC)[reply]
That's a pretty dismissive response. Doesn't even respond to many of my points, suggesting that you can't counter them. "Non-problematic" - how so? Etc. Unhelpful, unconstructive, almost a democratic vote, a not-not-vote. We can do better. ―Mandruss☎ IMO. 13:37, 8 September 2025 (UTC)[reply]
Well, we could start with the proposal's rationale: More accurate and more descriptive. Call it what it is. Less accurate and less descriptive is problematic IMO. One can disagree, but they can't claim I haven't explained something. ―Mandruss☎ IMO. 21:48, 8 September 2025 (UTC)[reply]
Oppose new name is unnecessarily long and hard to remember. I fail to see how idea lab is bad; this is a place for developing on ideas which will eventually become proposals; while also filtering out ideas which may not work well the community and the project. The underlined part is my description of the current name, using most words from the proposed name which, I believe, conveys that the current name makes sense and is sufficient. (please ping on reply) ~/Bunnypranav:<ping>13:23, 8 September 2025 (UTC)[reply]
unnecessarily long - Feel free to suggest a way to shorten it while achieving the goals of this proposal. I'm at a loss. (This is called "collaboration", which is what Wikipedia editors are supposed to do as I understand it.) hard to remember - Thankfully, nobody needs to remember it. I suggest using the shortcut name. It's easier to remember in two simple steps:
Type "VP" for village pump.
Type the widely-used acronym used for that widely-used Adobe file format for documents. You know the one, it's virtually a household word these days.
Oppose Unconvinced by the rationale - the proposed title is a mouthful, doesn't accurately describe what the Idea Lab is for, and I see nothing wrong with the current name. I also agree with SuperPianoMan that having two village pumps whose subtitles start with 'proposals' is undesirable. Sam Walton (talk) 14:03, 8 September 2025 (UTC)[reply]
First, you're making a point or two that has already been countered, suggesting tha you haven't taken the time to read everything and fairly consider it. Second, does "idea lab" accurately describe what the Idea Lab is for? I don't see it. But if we can't improve our game substantially in a short time (two days?), I'll go ahead and pull the proposal and save myself a lot of time. This is what a 12-year editor should know to expect, but I'm somewhat thick at times. ―Mandruss☎ IMO. 14:10, 8 September 2025 (UTC)[reply]
(edit conflict) Oppose. Even after reading all of the comments in this section, I'm still at a loss as to how the proposed change is an improvement or quite why the supposed problem is actually something that needs improvement. The current title (well, the "idea lab" part at least) is succinct, descriptive and directly relevant to the page's content and purpose without be fussy or prescriptive. Not everything in the idea lab is, or is meant to be, a developing proposal - somethings are literally "I've had this idea, what do you think of it?". Thryduulf (talk) 14:10, 8 September 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
The proposal (or at least this formulation of it) does not have consensus, and the proposer has indicated a desire to withdraw it from consideration (as well as linked to a VPIL thread for further brainstorming). jp×g🗯️09:22, 14 September 2025 (UTC)[reply]
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Hello all. My proposal is that all so-called "series" templates which do not indicate an order of the contained articles to be read should be deleted. Such templates clog up space in the article, and are merely glorified navboxes. A series box should present a set of articles to be read in (ideally) a particular order.
Here are examples of those series boxes which should be deleted:
Support, these duplicate the purpose of navboxes and additionally cause issues with MOS:SANDWICH. Having a chronological order (such as in campaignboxes) is a justifiable reason to keep a sidebar, as they add additional information. However, given the wide scope of this change, I wonder if it could be better to structure this proposal as a RfC. Chaotic Enby (talk · contribs) 13:15, 10 September 2025 (UTC)[reply]
To clarify, I placed the discussion there given its very wide scope (affecting thousands of templates, each of them in use on many content pages), rather than the current size of the discussion itself. Chaotic Enby (talk · contribs) 16:14, 10 September 2025 (UTC)[reply]
Oppose I don't think that the arguments in favor of removal are strong enough. To me, the situation is not anywhere near bad enough to justify this. JuxtaposedJacob (talk) | :) | he/him | 14:24, 10 September 2025 (UTC)[reply]
The argument is simple: these series boxes are unnecessarily taking up space in the article (usually at the lead paragraphs!) as they are mere duplicates of navboxes. ―Howard • 🌽3314:33, 10 September 2025 (UTC)[reply]
Strong Oppose - you are in effect suggesting the deletion of almost all sidebars on Wikipedia articles unless they happen to be a historical chronology.Bottom navboxes and sidebars can and do coexist, often times sidebars are a subset of bottom nav boxes that tend to be more complete, but also overwhelming to many readers.It would also force all readers to have to scroll to the bottom of articles instead of being able to open a main topic and being able to jump to related main topics right from the prominent sidebar.Like you are suggesting deletion of {{discrimination sidebar}} or {{Democracy sidebar}}, {{Psychology sidebar}}, {{LGBTQ sidebar}}, {{Research sidebar}} or {{Geography sidebar}}, or just about almost all sidebars outside of ordinal history. Likely snow close and most definitely needed CENT listing. The proposal should have been a lot clearer about the impact this would have, listing a lot more prominent examples as I just did. Raladic (talk) 15:56, 10 September 2025 (UTC)[reply]
I fully support the deletion of those templates also. These can all and should be formatted into navboxes. Take for example the Democracy, LGBTQ, Discrimination, and Geography sidebars which are filled to the brim with countless articles exactly as overwhelming a navbox would be. I oppose a snow close and prefer a full discussion into the matter. ―Howard • 🌽3316:06, 10 September 2025 (UTC)[reply]
These sidebars are how most readers navigate major topics. Template:Sidebar is currently transcluded 320,000 times Toolforge: Sidebar (~320,000). Likely the majority of such sidebars is not chronological or otherwise ordinal. We should get a quarry count to have a full understanding of the scope this would impact, which is going to be huge and will nto serve our readers by taking away the primary way how many readers navigate between articles.
If we have a sidebar with the same amount of content as a navbox, what is the purpose of the sidebar? I am not against all sideboxes, but I am against those sideboxes which claim to present a "series" but are overwhelming with content quite obviously not presented as a series. I believe ordinal sidebars are much more helpful as they have defined limits for inclusion and ordering. A navbox can serve as a more indiscriminate collection of topics. ―Howard • 🌽3316:48, 10 September 2025 (UTC)[reply]
Q: Why have a sidebar template at the top of an article, when you could have a navbox with the same links at the bottom of the article?
Strong Oppose Like Raladic, I do not understand how "chronology improves organization" can be used to conclude that non-ordinal series templates are so difficult to organize that they deserve across-the-board deletion. {{Donald Trump series}} is obviously an outlier given that Donald Trump is the most heavily covered individual on Wikipedia by many metrics. {{Progressivism sidebar}} and similar templates for political philosophies are highly useful in my reading experience to relate people of an ideology by their occupation. Having removed portals from the Main Page in 2022, we rely on such sidebars as a better maintained form of organization. ViridianPenguin🐧 (💬) 16:49, 10 September 2025 (UTC)[reply]
I’ll also note that if the OP has a problem with the Trump template - part of that is likely due to the fact that the template is poorly written right now and should use collapsible subgroups so that it’s transcluded use can have everything be default collapsed and only the relevant to the page section is expanded upon transclusion using section state parameters like we do on other more organized sidebars. Raladic (talk) 17:00, 10 September 2025 (UTC)[reply]
Strong Oppose This is a huge ask for an arbitrary reason. The OP is effectively requesting the mass-removal of sidebars for a reason not specified at the WP:SIDEBAR guideline and does not state any rationale apart from "I don't like it". The guideline addresses that sidebars can take up a lot of "real estate" but explains what should go into a sidebar vs. what should go into a navbox. When done well, sidebars (such as {{Joe Biden series}}) can aid readers in finding relevant articles to the topic at hand. If an article has both a sidebar and a navbox, the sidebar should either serve as a complement to the navbox or contain only a closely-related subset of articles in the navbox. IFF they are identical lists, then the sidebar can be removed from that article, but this wholesale approach is effectively using a wrecking ball instead of a ball peen hammer. — Jkudlick ⚓ (talk)17:20, 10 September 2025 (UTC)[reply]
Oppose. I cannot tell the difference between your examples of ordered and unordered sidebars. They all just look like sidebars to me. and I think removing all sidebars would be pretty drastic. do you have some diffs or evidence of complaints about sidebars? Is this maybe a solution looking for a problem? –Novem Linguae (talk) 18:19, 10 September 2025 (UTC)[reply]
From what I understand, ordered sidebars (like the example {{2023 United States banking crisis sidebar}}) show the articles in a specific, meaningful order (usually chronological or logical), which adds utility, although navboxes can also do that. I don't have specific diffs in mind (and it would have been good to see a WP:RFCBEFORE equivalent before jumping into a major discussion like this), but one aspect is that having more floating content can easily amplify MOS:SANDWICH issues. Chaotic Enby (talk · contribs) 18:25, 10 September 2025 (UTC)[reply]
I don't use these sidebars, so I do not think they are particularly useful (like most navboxes are not very useful) and they take up space that could be used for more images. However, per WP:LEOPARD, I oppose deprecating them via the Village Pump; if you want to delete them, just use WP:TFD. —Kusma (talk) 18:24, 10 September 2025 (UTC)[reply]
If WP:LEOPARD's guidance to tag each affected template were actually followed for this proposal, it would be interesting (purely academically) to see how many TfD's that would be, given that the {{TfD}} template is limited to a max of 50 at a time per WP:TFDHOW. Raladic (talk) 18:35, 10 September 2025 (UTC)[reply]
Yeah, no. Removing some class of templates en masse is precisely a case where you should use the RFC process. TFD is 1) not equipped to handle that kind of volume, and 2) is not the kind of place to establish a precedent for sending pages for deletion. Izno (talk) 21:06, 10 September 2025 (UTC)[reply]
My point is that TfD has proper mechanisms that ensure that interested parties get notified (tagging the templates). The Village Pump does not. I would be happy with any other notification mechanism that actually works. —Kusma (talk) 07:59, 11 September 2025 (UTC)[reply]
Oppose sidebars aren't redundant to navboxes, they're an alternative. In fact, I prefer sidebars, because unlike navboxes they look nice and high enough in the article content that readers are liable to actually use them. Cremastra (talk·contribs) 20:23, 10 September 2025 (UTC)[reply]
I am increasingly of the opinion that we should delete the use of {{sidebar}} in mainspace, but this proposal proposes only some half step that honestly doesn't make any sense. Sidebars like navboxes have for a long time not been required to have some order (lol), and changing what is effectively an inclusion criterion for links in a sidebar is nitpicking at what these do. Izno (talk) 21:11, 10 September 2025 (UTC)[reply]
Further comment: I will admit this is my first time starting a village pump discussion and was not expecting this level of attendance. I don't want to make this a full-blown RFC either, at least not at this point. It is perhaps my solution which is a bit too far (deleting all non-ordinal sidebars outright), but this is something I am willing to scale back if a different solution solves the following problem:
Some sidebars are just too long and indiscriminate. There are too many cases of them resembling or duplicating navboxes outright. These kinds of sidebars are taking up space in the first paragraphs and this can cause sandwiching.
If a consensus exists that this is a made-up problem, or merely opinion-based, then I will drop the entire discussion and move on. If there is a consensus that this is a genuine problem affecting the layout of pages but my solution is far-fetched, then I believe there should still be some kind of style guide to prevent these excessive side-bars. ―Howard • 🌽3322:11, 10 September 2025 (UTC)[reply]
Some sidebars are just too long and indiscriminate. There are too many cases of them resembling or duplicating navboxes outright. These kinds of sidebars are taking up space in the first paragraphs and this can cause sandwiching. - I'd say that's primarily opinion based.
While, as I mentioned above, some sidebars may not be well-designed (e.g. being default-expanded in all sections), that is an entirely differently fixable problem.
We have many very well-designed sidebars, that take up minimal space and can be controlled through parameters to only expand the section that is relevant to the page where it's transcluded, and they are serving many users to jump to related articles instead of having to scroll all the way down to a navbar at the bottom.
So I'd suggest if you have a particular template that you have an issue with, assuming it's not protected above your permissions, go, be WP:BOLD, or raise an edit-request at the talk page of the template if it might be a more contentious change, and improve it, give it collapsed sections, make unnecessarily big images a bit smaller. The world is your oyster.
I will say that of course you are not alone in that some other editors also may not like sidebars (as this thread has shown with some agreeing with you), but I think on-average, they serve more readers than they may annoy (some) editors and we do have to remind ourselves that we're writing this think primarily for our readers first and the average reader may not be as tech-literate as many editors are and appreciate a simple prominent box to find related articles that they may otherwise miss as statistically speaking, many readers don't go past the lead as WP:SUMMARIZE reminds us. Raladic (talk) 01:41, 11 September 2025 (UTC)[reply]
The LGBTQ and Discrimination sidebars are extremely cluttered (Psychology less so), so I would actually consider them bad examples. Even though the sections are collapsed, the content within them is still filled with a navbox-level amount of words. For example, when expanding the "manifestations" section of the Discrimination sidebar, one finds 211 words, and making the sidebar fully expanded gives 765 words. At what point can it be considered excessive? ―Howard • 🌽3312:53, 11 September 2025 (UTC)[reply]
I support the deprecation of all sidebars. They are fairly useless and take up a huge amount of horizontal space, which is in short supply for readers since we switched to Vector 2022. They also result in far too much strife to be worthwhile, like the recent nightmare over the Conservatism and Liberalism in China sidebars. Relegating this content to navboxes solves most of these issues, while leaving them intact for any readers who may still wish to use them. Toadspike[Talk]07:47, 11 September 2025 (UTC)[reply]
I must support starting to phase out sidebars. They are so cursed; infoboxes are already CVS receipts, and now these are tacked on underneath. If we want to sandwich a picture into the first section, in some cases we need to move the sidebar out of the lead for technical reasons. The navboxes at the bottom are much better. I would also give the (crackpot) idea that sidebars could work if they were instead placed in the vector 22 right side toolbox, but that idea might be too radical. 3df (talk) 08:22, 11 September 2025 (UTC)[reply]
Oppose as grossly disproportionate to the (mostly trivially) fixable problem of some sidebar templates being bloated and/or poorly organised. Thryduulf (talk) 13:10, 11 September 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
I suspect that some of these are content disputes (see WP:FETCH), others should be {{Better source needed}}, and that others are simply sources added by less experienced editors, who are expecting a "moderator" to check their work and remove the tag. But all of them are requests for a citation, when a citation is already present, which is an oxymoron.
I don't have an exact count, but I think it happens in a few thousand articles. I think we should have a bot remove these redundant tags. What do you think? WhatamIdoing (talk) 16:40, 11 September 2025 (UTC)[reply]
I've come across a few of these, it's definitely jarring, but I would be against mass removed by a bot. Some could just be removed, but other might highlight actual issues just with the wrong template. This search for {{citation needed}} times out, but returns 884 articles[11], {{fact}} - 74[12], {{CN}} - 411[13]. {{Citation needed}} has 81 redirects so there probably more, but these are all small enough that they could be cleaned up manually. -- LCU ActivelyDisinterested«@» °∆t°18:44, 11 September 2025 (UTC)[reply]
I did a search for >{{citation n and got three thousand articles before it timed out. My search also picked up non-refs (e.g., <!-- hidden comment -->{{citation needed so I tried again and got 2,300 for only </ref> tags. This misses all re-used refs like <ref name="Smith" /> and of course any redirect like {{fact}}.
There are many maintenance categories with tens of thousands of entries, and some of them have been cleared down. Even if it is a few thousand entries that is well in the realm of manual clearance. -- LCU ActivelyDisinterested«@» °∆t°19:18, 11 September 2025 (UTC)[reply]
I would be against removal by bot, because it is unlikely that a bot would be able to distinguish whether to remove the citation or the "citation needed" tag, and there are several other possibilities such as changing "citation needed" to something else. And human editors would relish having several thousand articles to work on. Phil Bridger (talk) 19:40, 11 September 2025 (UTC)[reply]
One possible case where this is not strictly an error per se (but is clearly far from optimal) is where where the citation supports only part of the associated content. Simply removing the CN template in such circumstances would be incorrect. What might work is for a bot to replace the citation needed tag with a new tag that states something like "The adjacent reference needs to be checked." and puts the article in a maintenance category. Thryduulf (talk) 23:39, 11 September 2025 (UTC)[reply]
That's only ever so slightly better than {{Citation needed}} but it's still not a great fit as that is for situations where someone has tried but failed to confirm the source verifies the article but has been able to for some reason (typically being unable to access the source). That's a specific issue that needs addressing, the problem here is that the combination of tags is not specific, it could mean multiple different things:
The citation needed tag is outdated
The reference failed verification
Someone was unable to verify whether the reference is correct or not
A better source is needed
The reference verifies only part of the associated content
What we want people reviewing this maintenance category to do is: identify what the issue is, fix it (if they can) or adjust the tagging to make it clear to other editors what the problem is. Thryduulf (talk) 10:11, 12 September 2025 (UTC)[reply]
remove {{citation needed}} (e.g., this is a reliable source and it supports the material as written; this is a reliable source, and I've corrected the Wikipedia article to match the cited source)
Why there are some undo/rollback edit such as edits 10 years ago like this or this did not contain Undo and Rollback tags? Can someone should tag Undo and Rollback for edits that were 10 years ago or more? Fabvill (Talk to me!)02:23, 12 September 2025 (UTC)[reply]
It's technically possible, at least in principle - tags are stored in a different database table than article text, or edit summaries, or other revision metadata, and there's nothing forcing it to only be updated at the time of an edit - but it's not practical to look at old edits and determine if they were actually rollbacks or undos. Edit summary doesn't help, since you can write whatever you want in there. I used a manually-constructed rollback-looking edit summary a couple times before registering, because I didn't know rollback existed and thought that was just the accepted etiquette when reverting someone, and looking back I'm sort of surprised nobody ever hollered at me for impersonating an admin or whatever. I suppose you could look for a combination of a proper rollback edit summary, contemporary presence of the rollbacker or administrator user group (which is itself difficult to prove programmatically), and an edit that restored a previous revision's text, but it would be difficult to program, likely slow to execute, and not be of much benefit: after all, you were already able to look at those old edits and determine they were undos and rollbacks despite them not being tagged. Proving an undo, when the edit being undo wasn't the most recent (and so not equivalent to a rollback) would be even worse. —Cryptic11:50, 12 September 2025 (UTC)[reply]
@Fabvill: The simple answer is that it isn't possible to do this, since mw-rollback and mw-undo tags are automatically applied by the software. The only way to apply a rollback-like tag or undo-like tag, would be to create a whole new tag just to deal with old edits and identify them as using rollback and undo, but that is of dubious benefit to the page history, honestly. If you have a tangible benefit to doing so, I suppose it's not impossible to add a new "manually-applied rollback tag" or similar for undo, but I'm not seeing a reason justifying that. EggRoll97(talk) 06:09, 14 September 2025 (UTC)[reply]
Proposal: Remove the Contents link from the sidebar
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.A summary of the conclusions reached follows.
No consensus to remove. Editors who expressed a clear opinion were evenly split between removing and keeping. Editors in favor of removal pointed to low click rate and said other navigation systems, like portals were better, and this link was a distraction. Many participants did not express a clear opinion, so it seems the need to remove this high-profile link is not obvious. Editors suggested improving the page or improving the links from the main page; this link was hard to find for some. This question might be worth revisiting if those efforts are made and do not result in a perceived improvement in usefulness or engagement. Proposing page deletion might be an appropriate forum next time, as some editors said this page is useless if it is not linked from the main menu. -- Beland (talk) 14:30, 12 September 2025 (UTC)[reply]
Considering the amount of pageviews this page gets, I just don't see the value in keeping this page. When looking at the page, it disappoints me that while it does cover most types of articles, it doesn't cover biographies pretty well. Biographies make up a huge portion of our Wikipedia pages, so unless the page is improved to include them, I think it would better to just remove the link. Interstellarity (talk) 12:28, 7 July 2025 (UTC)[reply]
The standard for appearing in the sidebar is much higher than "not causing harm." If it's not pulling its weight, it should be removed. SnowFire (talk) 15:53, 8 July 2025 (UTC)[reply]
The number of views is about 3800/day, essentially all of them from desktop, almost none from mobile. WP:Contents and its subpages are an attempt at providing alternative ways to discovering Wikipedia content (not through direct search), a time-honoured traditional approach. I am not sure it is working well for many people and I do not know how well-maintained it is, but if there is no link to this from sidebar or at least the Main Page, it won't work at all. —Kusma (talk) 16:11, 8 July 2025 (UTC)[reply]
Contents is indirectly linked on the Main Page, via Contents/Portals. Maybe replace that with a straight link to Contents? Dege31 (talk) 17:30, 8 July 2025 (UTC)[reply]
De-cluttering sidebars/menus generally has one underlying goal: to quit 'distracting' people with something that's not useful/helpful, so that they will be more likely to find/click on something that is useful/helpful. WhatamIdoing (talk) 07:10, 13 July 2025 (UTC)[reply]
Support removing the page, as it does not help navigation, nor does it make sense to have 7 million articles condensed into a scrappy and incomplete list in the sidebar. Portals probably do a better job. Pksois23 (talk) 07:31, 18 July 2025 (UTC)[reply]
Just to check, the sidebar and its links are only relevant for logged in users right? Taking action/no action would not affect most readers, so the links are mostly there for newer editors (or to try and beguile readers creating accounts into becoming editors)? CMD (talk) 07:53, 18 July 2025 (UTC)[reply]
As I said, there is no Main menu when I am logged out. I don't think logged out users can turn settings on and off. CMD (talk) 12:38, 18 July 2025 (UTC)[reply]
It's hidden as the three bars left of the Wikipedia logo in the top left. When you click it it should have an option that says move to sidebar. At least on vector 22 it's like this Pksois23 (talk) 12:45, 18 July 2025 (UTC)[reply]
I really don't understand the issue. How should we cover biographies on the Contents page? Or if you don't know (which is quite reasonable) what benefit would you expect to get that you don't now? All the best: RichFarmbrough21:45, 30 July 2025 (UTC).[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
RfC: merging Wikipedia:Requests for comment/User names with AN(I)
Support as proposer. There have been three (3) archived requests in 2025, and there were six (6) in 2024. We have too many noticeboards, discussion venues, and processes; dealing with this one seems like some low-hanging fruit. We previously shut down WP:EAR and WP:N/N for disuse. At the VPI discussion, it was brought up that this might slightly discourage filing reports. I see that as a feature, rather than a bug: only two of those nine reports found consensus that the username was inappropriate, most recently over a year ago, so slightly increasing the threshold for a filing would cut down on unnecessary discussions. HouseBlaster (talk • he/they)16:16, 14 September 2025 (UTC)[reply]
Support per nom. While ANI is definitely getting a bit large, the very low volume of requests this would add isn't really an issue compared to the advantage of simplifying a whole noticeboard away. If we want to limit ANI bloat, we could move something like TPA requests away to AIV instead. Chaotic Enby (talk · contribs) 16:21, 14 September 2025 (UTC)[reply]
Oppose This will just create confusion I think. The current page is being used for when UAA has declined to block a username, but the person reporting would like a community opinion. Bringing it to AN or ANI will result in overblocking, I suspect as blocking admins may not be familiar with the detailed rules we have. Secretlondon (talk) 16:43, 14 September 2025 (UTC)[reply]
SupportWP:AN or the suggestion in the discussion below for WP:UAA. OpposeWP:ANI, for reasons similar to Secretlondon's (though I'd say the bigger problem will be the dramaboard mob not reading the rules, rather than admins). WhatamIdoing (talk) 17:39, 14 September 2025 (UTC)[reply]
Oppose. RFC/N is a specialized venue for a specialized set of cases—notably, applying a policy that many users, even experienced users, even admins, often misunderstand. The proposer hasn't presented any evidence that RFC/Ns are resulting in unjust outcomes or otherwise hindering the smooth running of the encyclopedia, just that there isn't much volume of cases and that some can be resolved through normal admin actions, neither of which is inherently a problem. Nor have they presented any evidence that AN or AN/I is better-equipped to handle such cases. If it ain't broke, don't fix it, and it ain't broke. Instead, I'd rather we better advertise the existence of RFC/N, particularly to UAA filers (who frequently report users on bases that do not justify a summary block, but might be disallowed at RFC/N if anyone bothered to file), and consider expanding RFC/N's scope to also include signature issues, which are related to username issues and are not well-suited to AN(I). -- Tamzin[cetacean needed] (they|xe|🤷) 17:46, 14 September 2025 (UTC)[reply]
Moving that "specialized set of cases" to a more widely used venue might result in more people understanding that policy better. That could result in fewer unjust accusations being made in the first place, which would support the smooth running of the encyclopedia. WhatamIdoing (talk) 18:10, 14 September 2025 (UTC)[reply]
Oppose (summoned by bot). ANI every time I drive by or am summoned to it is definitely too large, and having a specialized venue would enable admins to get to specific username requests quickly, and sort them out. I would be more open to making them a dedicated section, but I still feel like that one page for everything could have a load time impact, as well as increase the chance of edit conflicts. InvadingInvader (userpage, talk) 14:16, 16 September 2025 (UTC)[reply]
Discussion (merging Wikipedia:Requests for comment/User names with AN(I))
No - it's used when the submitter thinks it should be blocked but the admins don't agree. They've all gone through UAA and been found as borderline. Secretlondon (talk) 16:40, 14 September 2025 (UTC)[reply]
I see it as a separate section of the page. Given the very few username discussions it shouldn't be a problem. The other alternative would be to create a template to use and hold the discussion on the talk page of the user. Just spitballing here. 331dot (talk) 17:34, 14 September 2025 (UTC)[reply]
Honestly, a dedicated section for deeper discussions would be quite helpful. Avoids sending editors flying from one noticeboard to another, and especially avoids the risk of ANI drama, while keeping the number of noticeboards low. I'm inclined to support that as my first choice. Chaotic Enby (talk · contribs) 17:48, 14 September 2025 (UTC)[reply]
That might necessitate renaming UAA to something else; Usernames for attention; Username reporting and discussion; something else..... 331dot (talk) 18:25, 14 September 2025 (UTC)[reply]
Yet another perfect place to change "administrator" to "administration", just like with AN/ANI. They're not pages for administrators, they're pages for the administration of the encyclopedia. ScottishFinnishRadish (talk) 14:23, 16 September 2025 (UTC)[reply]
Every once in a while, we have a subject where the article is being heavily edited by editors with competing views from all sides, such that the article ends up being fully protected either due to addition of strongly contested content or edit warring or both. I propose that whenever an article is fully protected while being heavily edited in this manner, a subpage draft of the article should automatically be created, where editors can continue their efforts unimpeded by the page protection. That way, good additions can be made and sources added within the context of the appearance of the article without the actual article being touched. If editors want continue edit warring over the substance of the article, letting them do so on a draft subpage allows the dispute to play out and possibly even result in workable compromise language without either disrupting the article, which remains locked, or requiring administrators to fix every typo and add every legitimate point and reliable source proposed on the talk page. BD2412T22:59, 16 September 2025 (UTC)[reply]
And who gets to decide which of the edits on the sub-page get moved to the article? That's if any editing actually happens on the sub-page.
I can perhaps see the merits in a page that exists for the sole purpose of collecting additional sources (plus the minimum context needed to identify what they relate to). Only an uninvolved (admin? extended-confirmed?) editor would be allowed to remove citations from that page that had not been added to the main article with consensus. Anything else, discuss it on the talk page. Thryduulf (talk) 23:57, 16 September 2025 (UTC)[reply]
@Donald Albury: If editors are directed to make their proposed edits to the subpage, I expect a reasonable number will do that. As for who decides which edits get moved, that would probably still be the admins (until page protection expires), but they need not do it right away, so disputes will play out on the subpage. I think sometimes editors can't come to an agreement on the talk page because they can't envision what it will look like in the article, and can't edit the article. BD2412T00:00, 17 September 2025 (UTC)[reply]
I just have trouble envisioning that editors who edit war in main space are going to play nice on a sub-page. I suspect many won't bother because it isn't real. Donald Albury00:57, 17 September 2025 (UTC)[reply]
I've seen one or two varieties of this on specific contentious pages a long time ago. It might sometimes be helpful but in general we should not encourage the idea that people can continue to treat pages like a forum to erect their favorite theory. Johnuniq (talk) 02:52, 17 September 2025 (UTC)[reply]
The smaller action of locking the lead within a subpage would alleviate a large amount of potential conflict while still allowing work on the substance of the article. CMD (talk) 03:20, 17 September 2025 (UTC)[reply]
I actually haven’t seen a fully protected article before, and seeing now how rare it is, that makes sense. But in clicking through to Spinosaurus—an article I would presume is uncontentious—it is not at all obvious what has gone wrong to the point that experienced editors are unable to edit it. Clearly there’s some dispute, but there is no notice nor section on the talk page. Is it too much to ask that perhaps fully protected pages should have some section created, linking recent edits and or editors involved? Someone decided that Erika Kirk needed a banner, and while I disagree with this as a general solution (the talk page is a much better place for it), it speaks to a shared desire on the part of editors to understand what is going on. — HTGS (talk)09:17, 17 September 2025 (UTC)[reply]
I fully protected United States Department of Defense last week because of an edit war involving multiple users, including some with edit counts in the thousands. I modified that 12 hours later to ECP protection with BRD required. Most of the edit warriors have stayed away from both the article and the talk page since then. Editing the talk page doesn't appear to be attractive to those edit warriors, and I doubt a sub-page would be any more attractive. I think full protection of a page should be kept as short as possible, and, at least with articles that fall in a contentious topic area, other editing restrictions should be used to prevent edit warring. Donald Albury12:39, 17 September 2025 (UTC)[reply]
Apparently there's a debate over just how big that dinosaur is. There are three discussion on the talk page, and two of them are about it.
In re "Is it too much to ask that perhaps fully protected pages should have some section created, linking recent edits and or editors involved?": I think one of the purposes of protecting a page is to incentivize the edit warring factions to find the talk page. WhatamIdoing (talk) 21:12, 17 September 2025 (UTC)[reply]
Is there really a problem with creating this list on Wikipedia? Then users, on a volunteer basis, could start adding everyone who might have been on that island.
You could create a Red List and a Yellow List — for those who were definitely there, and those who possibly were.
(BLP violation removed), for example, would go on the Red List — that's pretty obvious. (BLP violation removed) would go on the Yellow List, since he said that "this list doesn’t exist" — maybe he’s on it himself.
This list could then be treated as real and used to damage the reputation of those pedophiles.
Of course, all accusations must be backed by links to the sources or information the claims are based on. Ivapol (talk) 09:51, 17 September 2025 (UTC)[reply]
While a list of people who actually were on the island might have legitimacy, Then users, on a volunteer basis, could start adding everyone who might have been on that island. is way too broad. You might have been on the island. We (TINW) should be careful to not defame innocent people. -- Shmuel (Seymour J.) Metz Username:Chatul (talk) 09:59, 17 September 2025 (UTC)[reply]
This template uses the description list markup to display the "hatnote", which is bad for accessibility and also not compliant with the manual of style (Wikipedia:Manual of Style#Indentation: Do not use : (description list markup) to indent text in articles).
If you use a <div> rather than a <span>, it should avoid the manual notes will be on the same line as notes generated by the template difference. Anomie⚔13:18, 17 September 2025 (UTC)[reply]
🤷 If you want some things but not others, I guess you're out of luck to reproduce the current behavior without wikitext list markup of some sort. Anomie⚔13:53, 17 September 2025 (UTC)[reply]
If the implementation of the {{Updated}} template is modified to pass its message (including the reference) to {{Hatnote}}, then the message should be in a <div>...</div> element, with the reference contained within, and following text will be on a new line. However, any uses in the form of the testcase at Template:Updated/testcases § Reference following, with text on a new line after the reference would have to be changed so the reference is passed to the template. This does seem to be the originally intended way to use the template, but if it's true that a trailing reference is frequently used in football articles, then there may be a lot of instances of this alternate usage. isaacl (talk) 15:50, 17 September 2025 (UTC)[reply]
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
I think this could be really useful as you have a choice to change your Wikipedia time that you see on your computer (a local option just for the computer itself)
How I think this would work is that you would go into preferences > appearance > time offset, and it would give you a list of options to change your "server time" to.
Your signature's time being your local time is just inconvenient to people whose signature's time is UTC. Imagine that your signature's timezone was UTC+3 at the moment you made this comment I'm replying to; it would appear as if you replied to me 3 hours after I left my comment. Sapphaline (talk) 12:39, 18 September 2025 (UTC)[reply]
The time after the signature is in UTC so that everyone sees the same time, which when referring to a post (e.g. "the post made by EditorShane3456 at 12:28 today") is a good thing. Phil Bridger (talk) 12:37, 18 September 2025 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
from Wikipedia, the free Encyclopedia." Here I placed the icon after the phrase introducing the article class since it looked akward to have it immediately following the indefinite article "A" when I tried that first. Gramix13 (talk) 02:52, 8 July 2025 (UTC)
) next to the article title? TompaDompa (talk) 18:12, 22 July 2025 (UTC)
Display an assessment of an article's quality in its page header (documentation). Dan Leonard (talk • contribs) 16:37, 4 August 2025 (UTC)
ambox.speedy .delete
ambox.content
ambox.style Not Codex, but is the OOUI for style icon...
ambox.notice
Option 2 ambox.notice
with 
but otherwise everything else looks good. Aasim (話す) 04:14, 11 August 2025 (UTC)
would be the most visually similar one, although 
or 
convey better the idea of looking for references in my opinion. Any of those would be fine to me.The icon in {{Unreliable sources}} also matches well with 
, if we take the rectangle being searched to represent the article rather than the sources). 
is closer visually, but could be more confusing as "question mark in rectangle" is a very common symbol and makes it less clear that the rectangle represents the article.{{Disputed}} easily gets 
.For {{Speculation}}, Codex sadly doesn't have a crystal ball yet (although I could make one in the Codex style if needed). The closest "magic woo" icons (and that's a stretch) are 
or 
.{{Current}} and {{Recentism}} can get different colors of 
.{{Globalize}} gets 
. Chaotic Enby (talk · contribs) 18:19, 15 August 2025 (UTC)
.jpg/20px-Broom_(PSF).jpg)
. Not good. Even rotating it 45 degrees doesn't really fix the issue: Aasim (話す) 07:00, 12 August 2025 (UTC)
(File:Codex icon edit.svg), to indicate issues that editors can solve?
(a generic tag icon) or 
(yes, it's officially the recent changes icon, but simple icons like this are very polysemic). Chaotic Enby (talk · contribs) 14:12, 12 August 2025 (UTC)
, but it goes a bit too far in the other direction. Surprised we don't have amber in the Codex or OOUI color palettes. Chaotic Enby (talk · contribs) 17:36, 13 August 2025 (UTC)
? Aaron Liu (talk) 18:28, 20 August 2025 (UTC)
→ 
→ 
(but modified to have an X inside rather than an !)
→ 
→ 
→ 
→ 
. Chaotic Enby (talk · contribs) 19:34, 18 August 2025 (UTC)
as well. –LaundryPizza03 (dc̄) 16:54, 21 August 2025 (UTC)
. The icons do look nicer on colored backgrounds that match them, but otherwise when not on a matching color it can look a little hideous. But a hideous color that is still legible is better than un-dark-modeable colors like the white clock that cannot be changed to black as inverting would mess up the color of the hands and the stop X. And on Vector 2022 and Minerva it seems broken, I don't know why. Aasim (話す) 03:27, 20 August 2025 (UTC)
, which has a white interior and is currently in use at kowiki's equivalent of Template:Uw-ublock. –LaundryPizza03 (dc̄) 16:55, 21 August 2025 (UTC)
→ 
→ 
→ 
→ 
(illustrative example only not a suggestion!).
Previous
New
This article documents a current event. Information may change rapidly and initial news reports may be unreliable. The latest updates to this article may not reflect the most current information. Feel free to improve this article or discuss changes on the talk page, but please note that updates without valid and reliable references will be removed.
This article documents a current event. Information may change rapidly and initial news reports may be unreliable. The latest updates to this article may not reflect the most current information. Feel free to improve this article or discuss changes on the talk page, but please note that updates without valid and reliable references will be removed.
) meshes well with the Codex signs and the proposed broom icon; it could be used for Template:Globalize if the Codex-style symbols are adopted for generic use. –LaundryPizza03 (dc̄) 17:06, 21 August 2025 (UTC)
