Wikipedia:Village pump (idea lab)

 Policy Technical Proposals Idea lab WMF Miscellaneous 

• Table of contents
• First discussion
• End of page
• New post

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

• Discussions of technical issues belong at Village pump (technical).
• Discussions of policy belong at Village pump (policy).
• If you're ready to make a concrete proposal and determine whether it has consensus, go to the Village pump (proposals). Proposals worked out here can be brought there.

Before commenting, note:

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

Discussions are automatically archived after remaining inactive for two weeks.

« Archives, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40

Clarifying the number of sources required by GNG

WP:GNG states There is no fixed number of sources required since sources vary in quality and depth of coverage, but multiple sources are generally expected. In theory, this nuanced rationale is useful, but in practice it is rarely applied; editors find their preferred definition of "multiple" (I believe it is typically three, in line with the essay by RoySmith and with most of the remainder defining it as two) and so long as the editor believes each source meet WP:SIGCOV they consider it sufficient.

As such, I believe this nuance doesn't benefit the project, instead making it more difficult for new editors to understand what is required for their article to survive, harder to determine a consensus at AFD, and complicating the work of AFC and NPP. To resolve this, I believe it would be useful to provide a fixed number; as a first thought, I would propose the above quoted text be replaced with At least three sources are required.

Raising this here, to develop the idea and because if an RFC is opened broad input would be desirable due to the significance of GNG. BilledMammal (talk) 16:27, 13 February 2022 (UTC)

We have never spelled out the number because this is far too game-able. We are looking for what is sufficient to show that significant coverage exists. That could come from one really strong comprehensive source like an in-depth biography, or it might require six or more sources that each only have a paragraph or three but combined provide good coverage. Just saying N sources means that editors at AFD will only focus on the number of sources with no consideration of what significant coverage there is. We know this happens. --Masem (t) 16:33, 13 February 2022 (UTC)

I would agree, except editors already focus on finding N sources that individually meet the other requirements of GNG; the only difference is that they use their own definition of N. This idea won't fix the issue, but it won't make it worse and it will fix other issues. BilledMammal (talk) 16:40, 13 February 2022 (UTC)

@BilledMammal, it happens that @SmokeyJoe and I have been discussing this for the better part of the year. I can't say that we've made much progress towards an answer, but we might be getting better about asking the questions.

On the one hand, you have editors like @Masem, who will take a paragraph here and a paragraph there, and so long as the total combination of all independent sources results in being able to write an actual encyclopedia article in the end, then they all count towards notability. One excellent source might demonstrate notability on its own; for another subject, six small sources might be needed.

On the other hand, you have editors, many of whom are focused on (perceived) self-promotionalism, who prefer the WP:SIRS approach, and declare that no source counts towards notability at all unless that single source, evaluated in isolation, would basically be enough to write a whole encyclopedia article on its own. These editors usually require a minimum of two or three excellent sources, and they will not consider less-than-excellent sources at all. A mere paragraph (or three) in a news article is immediately and completely discarded as having no bearing on the question of notability at all; they never reach a point at which one could consider whether the combination of a couple of paragraphs here plus a couple of paragraphs there could produce enough sourced content to make it possible to write an article.

I've never been impressed with the "game-able" claim. We should write down what the actual answer is, because that's consistent with our value of transparency. If we write down "three sources" and people produce three sources containing a single sentence each, then that proves that we wrote down the wrong answer – not that the answer should be kept secret. IMO the actual answer probably sounds more like (a) two different authors writing for two different publishers plus (b) a minimum of 500 words directly about the topic, across all [financially] independent sources, including at least 10 severable facts about the subject and at least one indication of analysis (e.g., comparison to a similar subject or classification as an instance of some larger notable subject). But I'd also ban editors from rejecting sources just because they personally feel that the source's linguistic style is insufficiently scholarly or because the newspaper reporter directly interviewed the subject (and therefore allegedly is no longer "independent"), so you can take my view with an appropriately large grain of salt. WhatamIdoing (talk) 20:01, 22 February 2022 (UTC)

@BilledMammal please read the notes that go along with that essay. -- RoySmith (talk) 16:42, 13 February 2022 (UTC)

Thank you; I originally had both yours and WP:3REFS, but for concision I shortened it to just yours, as the more influential essay in establishing or maintaining the position that N is three, not two or four, even though that wasn't your intent. BilledMammal (talk) 16:50, 13 February 2022 (UTC)

That's still the point: the GNG is 100% centered on showing significant coverage and that can come from 1 high quality source or may require many where the topic itself is not the central topic of each source. Counting sources is not a replacement for reviewing significant coverage, and editors that continue to equate GNG to number of sources using those essays are fundamentally wrong. --Masem (t) 16:59, 13 February 2022 (UTC)

Unfortunately, counting sources (that meet the other criteria of GNG) is what happens, and the fact that there should be more nuance doesn't change that.

On a side note, I disagree that one high quality source is enough, as I don't believe we can comply with NPOV when we only have a single perspective, but that isn't relevant to this discussion. BilledMammal (talk) 17:38, 13 February 2022 (UTC)

What I mean is if I have a book-length biography, high quality, it will likely have its own set of references about that same person. The initial WP article about that person would not need to include those references but that they exist with the biography means more sources exist out there. This doesn't apply to, say, a single long-form obit that is not going to have that type of feature. --Masem (t) 18:05, 13 February 2022 (UTC)

I agree with Masem here. A. C. Santacruz ⁂ Please ping me! 18:18, 13 February 2022 (UTC)

With this clarification, so do I, although I would interpret it as having multiple uncited sources, rather than just one. BilledMammal (talk) 18:22, 13 February 2022 (UTC)

The GNG (and pages related to retaining/deletion of articles due to notability) should be viewed as the potential for articles to be developed further from the likelihood of additional sourcing to be found in the future or that can be added that has been identified but not yet added. EG: if an article goes to AFD and 10 new sources are found and agreed to be sufficient to give that topic significant coverage, then the only thing that should be done is getting those 10 articles added to the article, but we don't have a deadline for that; if years later those 10 sources have yet to be added, it would still be inappropriate to AFD the article because we know those sources exist. Same concept with the "one great source that has a known bibliography/reference section to work from" situation. This comes back again to why we try not to judge just on the number of sources since one source could be a link to potentially multiple additional sources that could build more significant coverage in time. Of course, what more commonly happens today is that editors are just using web-based sourcing that rarely include references to other source, so we judge each source on its own and not what it can led to. --Masem (t) 19:11, 13 February 2022 (UTC)

• Umm, how about 1? A new stub could be one line, sourced, and be GNG OK. — xaosflux ^Talk 17:55, 13 February 2022 (UTC)

42. JoJo Anthrax (talk) 18:15, 16 February 2022 (UTC)

• Just went to newpages, Glaucium calycinum is a brand new article, I see no reason it would be a good deletion candidate for lack of another source yet. — xaosflux ^Talk 17:57, 13 February 2022 (UTC)

  That would fall under WP:OUTCOMES eg WP:NSPECIES. Not so much a GNG issue. --Masem (t) 18:02, 13 February 2022 (UTC)

  I think it would actually fall under WP:COPYVIO; the only difference is that some connecting words have been added, and some technical words have been replaced. BilledMammal (talk) 18:16, 13 February 2022 (UTC)

  Possibly, I'm not saying it was "good" - just that the lack of 2(+) sources are not enough to say it should be deleted for failing GNG. — xaosflux ^Talk 19:06, 13 February 2022 (UTC)

  I think most plants would be better as subsections within their Genus; for example, Aa (plant); all the articles listed there should be merged into it. BilledMammal (talk) 20:13, 13 February 2022 (UTC)

  This is the "notability doesn't depend on what's already cited in the article" problem. It's likely for any species that you could find multiple sources; the fact that there is only {{one source}} indicates a failure of effort, not the non-existence of sources in the real world. WhatamIdoing (talk) 19:44, 22 February 2022 (UTC)

  @BilledMammal, there are only a few standard ways of describing plant species, looking at the same parts in the same order. I wouldn't characterize the article as close paraphrasing, and certainly not of the reference you list as being copied from. That's a very harsh way to treat an editor whose article just came up as an example. A non-templated note about close paraphrasing on the talk page would be enough and have it discussed there. I'm removing the copyvio notice. StarryGrandma (talk) 23:18, 13 February 2022 (UTC)

  I've responded on your talk page. BilledMammal (talk) 23:32, 13 February 2022 (UTC)

  Aforementioned editor here, could we discuss this on the talk page of the relevant article instead of on half a dozen different user pages? If this is an issue, I'd like to be involved and for there to be a central point of public discussion. Thanks y'all. Fritzmann (message me) 23:49, 13 February 2022 (UTC)

The GNG requires two, minimally. This does not mean two is sufficient, but in practice two are sufficient if they are both very good sources.

WP:THREE is about countering WP:Reference bombing. If it’s already been deleted, and you think sources support reversing the decision, then show us your best three, and we will review them very seriously; but DO NOT give us twenty and ask us to spend hours analysis in them all. Also, a rule of thumb that stands up to experience is that if the best few sources are GNG non-compliant, no number of worse sources will be sufficient.

The GNG is just a guideline, and the real decision gets made at AfD. However, when push comes to shove, and we are talking about for-profit company, it’s CEO, or it’s products, the Wikipedia:SIRS approach is the consistent winner. WAID and I have talked at length, without agreement. I believe the crux is that for promotional topics, product reviews are to be presumed non-independent and GNG non-compliant.

Exceptions are traditional encyclopedic topics, like the natural sciences, and distant history. For these, WP:V is enough.

—- SmokeyJoe (talk) 21:50, 22 February 2022 (UTC)

• 5 sources is always sufficient; 4 almost always; 3 usually, 2 often, and 1 never meets GNG but may be sufficient if paired with a notability guideline for special circumstances. Bluerasberry (talk) 14:38, 1 March 2022 (UTC)

  I can view one source as meeting GNG if it's very extensive - most commonly, this is a book dedicated to the article. This is one aspect where NCORP is de jure different and not merely an "enforce more vigorously" Nosebagbear (talk) 13:13, 8 March 2022 (UTC)

Use an edit filter to warn about placing wikilinks next to each other

Hello, I'm new to the whole proposal thing. So, almost all of us has placed wikilinks next to each other, shown in this example. This is to ensure the compliance of MOS:SEAOFBLUE, so the edit filter would most likely have a "warn" setting. I am aware of the pop-up if you link common words like in the example, but it does not address the aforementioned issue. Thoughts? CactiStaccingCrane (talk) 07:03, 4 March 2022 (UTC)

The benefits are fairly clear (warn would indeed the logical setting), but two potential points jump out (and I imagine they did to CSC as well) - the big one being that this could do a lot of undesired warnings. It might be worth running a month of log-only to see what the rate would be. Secondly, would be good to get a specialist saying it'd be fine on a computational front - run-time could be expensive (in comparative terms) for a filter like this. Nosebagbear (talk) 09:48, 7 March 2022 (UTC)

I think maybe placing them client-side would be a better idea then. CactiStaccingCrane (talk) 09:52, 7 March 2022 (UTC)

Help with discussion participation

I propose in the talk page we have an area where users can sign their name to show they are major supporter of an article. This could be beneficial so people in disputes can have editors too easily ping for discussion regarding the article. This will help with participation in the consensus making process. Iamreallygoodatcheckers (talk) 04:09, 6 March 2022 (UTC)

Don't we have the article's page history for that? --Redrose64 🌹 (talk) 08:32, 6 March 2022 (UTC)

• Meh… Any “major supporter” of an article will probably have it on their watchlist. Thus, it is likely that they already know about any disputes that have arisen. Blueboar (talk) 14:32, 6 March 2022 (UTC)

A list of people to WP:CANVAS? CambridgeBayWeather, Uqaqtuq (talk), Huliva 19:09, 7 March 2022 (UTC)

WP:AALERTS more or less covers this. Headbomb {t · c · p · b} 23:01, 18 March 2022 (UTC)

NSONG for covers

From WP:NSONG: Songs with notable cover versions are normally covered in one common article about the song and the cover versions. This rule... makes no sense to me. Since I've ranted about this at length already, I'm just going to copy my reasoning for bringing this up here:

... a good song article is going to have big sections about production, release, reception, and themes. Those are going to be very different in cover songs that have taken on a life of their own. Off the top of my head, I'm thinking of Marvin Gaye's I Heard It Through The Grapevine (as well as Creedence Clearwater Revival's), The Beatles' Twist and Shout, The Animals' The House of the Rising Sun, Simon & Garfunkel's Scarborough Fair, Gillian Welch and Allison Krauss's I'll Fly Away, and They Might Be Giants' Istanbul (Not Constantinople), but there are many more. If a Wikipedia article on a song solely focuses on the song's lyrics and tune, that's a pretty terrible article—but that's the only thing a notable cover has in common with its original. If a cover has independent notability, it should have a separate article. Pigeonholing a cover into the original's article fundamentally misrepresents what a song's Wikipedia article should focus on, and leaves out a lot of notable information that could very well have its own article.

So, I want to brainstorm on language that can replace that line. I haven't been able to get it past the rules-lawyer in my head, though, so I'm putting out my best guess and I want to hear your thoughts on how it can be improved:

Songs with notable cover versions are normally covered in one common article about the song and the cover versions, except where a comprehensive overview of a cover in the original article cannot reasonably constitute due weight and must be spun off into a separate article.

It's... pretty conservative at the moment. Basically, covers (even notable covers) are merged to the original wherever possible, but if a cover is just too big or renowned or has a life of its own, it'll get its own article. Any ideas for this? theleekycauldron (talk • contribs) (she/they) 11:10, 9 March 2022 (UTC)

For the record, the original discussion is at Wikipedia talk:Notability (music)#Really, a cover can never have an independent article? – I don't want to relitigate whether or not we should be drafting this rule at all at the moment, since we're not consensus polling. theleekycauldron (talk • contribs) (she/they) 11:11, 9 March 2022 (UTC)

Let the community decide about major new features?

Now maybe I'm wrong and I just always miss it, but my impression is that the WMF would often develop features with minimal community input, minimal community testing, doesn't give the community any real chance to consider possible alternatives and proceeds to roll out the new feature. For example mw:Flow which was quite the faceplant, cross-wiki uploads plague Commons with numerous copyvios to this day due to a flawed design, DiscussionTools which takes up bandwidth even if you've disabled it, the VisualEditor which isn't universally loved and I don't expect a poll before the Vector 2022 skin which (in its current state at least) has upsides and downsides over Vector classic. Damn, almost forgot about IP masking! (are we still in the dark about what that's actually going to be like?) What I'm thinking is that the community should be informed well in advance and there should be some mechanism to ensure major new features will be backed by the community, or at least not hated. The community should have a chance to prepare for it (like updating tools/bots/help pages) and suggest changes or alternatives. For example with Flow, the community could have rejected it before it made a faceplant. While IP masking can't be rejected as the motive for that is legal, the community could propose implementation details/alternatives which do exist. And brace for its introduction, of course. I'm just brainstorming here, am I making any sense? — Alexis Jazz (talk or ping me) 15:41, 10 March 2022 (UTC)

To shoot from the hip, whatever my thoughts on the rest of it, DiscussionTools is one of the single biggest improvements to the editor experience in the history of the project and the associated community outreach has been superb. Anyway, there's some information on ongoing projects on mediawikiwiki. I would say it could be updated/communicated more frequently, but perhaps it's already current. Enterprisey (talk!) 18:03, 10 March 2022 (UTC)

I don't know.. maybe like apply to work for the foundation ? —TheDJ (talk • contribs) 20:58, 10 March 2022 (UTC)

also, which part of the community ? different parts always show up to different steps and different projects and always have different opinoins and generally ppl leave out those without 'community' voice —TheDJ (talk • contribs) 21:01, 10 March 2022 (UTC)

TheDJ, work for the foundation, that sounds like wonderful idea. As for which part, well I was just brainstorming. I suppose if the WMF was more responsive to questions that community members are already asking and if developers didn't spend 90% of their community time on Phabricator (which is largely disconnected from the community at large) and in Flow discussions, we'd be halfway already. — Alexis Jazz (talk or ping me) 03:46, 11 March 2022 (UTC)

As TheDJ alludes to, the community doesn't speak with one voice, plus design by large group conversation is often ineffective. On user-interface related features such as skins, there needs to be some liberty for new designs to be tried and tested, as learning from failure is important. I agree there ought to be a feedback loop, and in some cases, trials are appropriate. (I've not followed the progress, but I do know that suggestions were sought from the onset regarding IP masking, and most recently feedback was requested on different approaches.) Although focus groups shouldn't be the ultimate last word on design choices, as they're a small sample of opinion, it could be helpful for the WMF and the community to work together, at the start of the design cycle of a feature, on getting a suitable set of volunteers who could provide prolonged engagement during the development process. isaacl (talk) 21:36, 10 March 2022 (UTC)

Great idea. Like FRS but for WMF tech. Enterprisey (talk!) 00:40, 11 March 2022 (UTC)

I wasn't thinking of a notification service. I think for a given feature, the developers and community should agree upon types of users who could give useful feedback, and find volunteers who are willing to have ongoing discussions throughout the development process. This would allow the development team to have fast feedback cycles, enabling them to go through different options more rapidly. isaacl (talk) 00:49, 11 March 2022 (UTC)

The problem is that these people and more so their availability is highly unreliable. You can't plan around it, they don't answer fast enough, and they too are heavily biased. There have been plenty of WMF projects where WMF was nudged into a certain direction only to be recalled later by other volunteers. —TheDJ (talk • contribs) 09:09, 11 March 2022 (UTC)

Yes, I was going to say more on that aspect but omitted it for conciseness. There is of course no guarantee that any selected group would remain engaged; it can only be hoped that at any given opportunity, someone in the group will respond. As I mentioned, the received feedback should be considered carefully as one source of comments and not treated as absolute truth. isaacl (talk) 15:25, 11 March 2022 (UTC)

To echo what Enterprisey said, at least for DiscussionTools, the WMF has been soliciting feedback on the tool here on the various Village Pumps, at mw:Talk pages project/Replying, and in Tech News since early 2020. --Ahecht (TALK
PAGE) 22:34, 10 March 2022 (UTC)

Ahecht, if DiscussionTools is the result of working closely with the community I may have been wrong. DiscussionTools is nice on the surface, if you don't care about features, customizability or resources. It's more convenient than source editing, but in some ways (philosophically, no idea if any code is shared) it's Flow in disguise, just not exposing the "structured" part of "structured discussions" to the end users (us), which some may actually consider to be a good thing. My main concerns with recent developments: DiscussionTools can't be disabled. (you can hide the links but not actually get rid of them) For IP masking several users asked on Wikipedia:Village pump (WMF)#How we will see unregistered users what we will see instead of IPs and similar and related questions on Wikipedia:Village pump (WMF)/Archive 4#New IP Masking implementation updates available also remained unanswered. — Alexis Jazz (talk or ping me) 03:24, 11 March 2022 (UTC)

"you can hide the links but not actually get rid of them" This has more to do with the fact that we are going to more consistent Parser output for all users (less variance, hopefully one day, no variance) that with DiscussionTools. "what we will see instead of IPs" As far as I know that's exactly what they are figuring out right now, right ? —TheDJ (talk • contribs) 09:19, 11 March 2022 (UTC)

TheDJ, the kind of features that DiscussionTools adds can be accomplished with some very tiny extra HTML parser output instead of the massive bloat they implemented. From DiscussionTools that would be the data-mw-comment-end span tag. Their IDs are suboptimal but that's another matter, an ID is all you need in the HTML. IMHO the better solution is to actually fix signatures to become more machine readable, both in wikitext and when parsed, which would cause all required data to become part of the wikitext, and no parser adjustments would be needed. Somehow either nobody suggested this or they were ignored, either way, seems like a communication problem to me. And about IP masking, well, I don't know because the WMF developers don't have 2 minutes to state something like "the exact form of IDs for anons is not yet set in stone, the leading contender is X, we are also considering Y and Z" in a relevant WP:VPWMF discussion. @Certes: you keep better track of this than I do I think, please prove me wrong and tell me the WMF made a more informative statement since I last checked. — Alexis Jazz (talk or ping me) 15:50, 11 March 2022 (UTC)

I've seen nothing but only really watch Wikipedia; there may be updates on Meta or MediaWiki. Masking doesn't seem to be happening quickly, so I'm hoping that it will prove infeasible and go away. Certes (talk) 18:31, 11 March 2022 (UTC)

Certes, you have (possibly unintentionally) demonstrated my point about the disconnect: I [...] only really watch Wikipedia. You are invested in Wikimedia, even knowledgeable about technical details and clearly interested in them, you're just not tuned into the channels the developers primarily use. You made three posts on Phabricator this year. Friction between developers and community would be expected when communication channels are not effective. — Alexis Jazz (talk or ping me) 19:50, 11 March 2022 (UTC)

@Alexis Jazz: As you may have read in Tech News, the WMF just announced its intention to impose a session-based system. I suspect that it will be trivial to discard session information and start again with a clean slate, forcing us to ban unregistered editors and cut off our source of new contributors. Certes (talk) 22:44, 14 March 2022 (UTC)

I see they say "We do acknowledge that deleting a cookie is easier than switching an IP, of course, and do respect the effects it would have." I'd like to know how they respect the problems this will cause. Doug Weller talk 12:50, 15 March 2022 (UTC)

you can hide the links but not actually get rid of them – afaik, that's because at Wikipedia scale, "don't split the cache" is like "don't cross the streams". ⁓ Pelagic ( messages ) 06:38, 17 March 2022 (UTC)

The reply tool is fantastic and was designed with lots of feedback from the community, I think it's a really good way of improving talk pages without changing the underlying wikitext structure. "It increases the size of the page slightly" is probably the most nit-picky feedback possible, anyone here could make a much larger change to the page size by adding a few images.

I don't think that the issue here is exclusively "the WMF doesn't get feedback", "the WMF gets the wrong feedback" is just as big of an issue. If you go and look at the original flow prototyping pages on meta/mediawiki I think the original focus group used to create the overall design consisted of 5 people who had never used a wiki before. Asking people who are (for want of a better word) completely ignorant about how consensus based wiki projects work to design one of it's most essential features is a recipe for disaster. The WMF seems to be so obsessed with recruiting new editors that they are willing to completely ignore the people who run the site on a day to day basis.

Final thought - they do seem to have gotten a lot better over the last few years (at least in terms of software development) and do seem to be making efforts to listen to the community, most of the big disasters were 5+ years ago. 192.76.8.70 (talk) 09:49, 11 March 2022 (UTC)

Images one could disable or load on click/hover/whatever. I know many people here probably don't care, but shouldn't we take those who don't have unlimited data and 10mbps+ broadband into consideration? Some people are still stuck on 1mbit/s ADSL (or worse) or pay per megabyte, even in developed countries. (granted there are people who have no access to the internet at all, some not even to clean drinking water, but that's another issue) Now if the additional HTML was actually necessary to achieve the goal, that would be a fair argument. But it isn't. The HTML (just the HTML, don't get me started on the API call) that gets added currently is about 95% bloat. Actually I'm rounding down from 95.3%, this is not a number I'm pulling out of thin air. (if anyone has doubts I can share my homework) So if this is the result of close collaboration with the community, I was dead wrong thinking community involvement would improve things. (either that or it would have been even worse without the community involvement) And while I don't know about all the disasters from 5+ years ago, I know that structured data and crosswiki uploads are an ongoing train wreck. And here's another thing that worries me: mw:Talk pages project states Some features may involve introducing new wikitext. Although, any changes to wikitext will be limited to those that enable new features that benefit contributors. Features like replying to specific comments or watchlisting particular discussions. This could align with what I'm doing, but given the current state and direction of DiscussionTools I'm sure they'll find a way to mess it up, maybe even break my efforts in the process. And this Google doc from phab:T273341#7539540 worries me even more. There's the wikitext, which is easily publicly accessible and can easily be copied/forked/analyzed/etc and there'll be a shadow ledger which won't be so easy to copy/fork but the dependency on it will just keep increasing. The WMF developers still want wikitext to be something it's not. They still want Flow. Maybe that's what the community told them during the collab sessions, in which case I guess I'd have to eat my humble pie. — Alexis Jazz (talk or ping me) 19:33, 11 March 2022 (UTC)

"It increases the size of the page slightly" is probably the most nit-picky feedback possible, It certainly does. Examining the HTML that is served, I find that every single post has gained several hundred extra bytes, in the form of five extra <span>...</span> elements (one at the start of the post, the rest after the timestamp) and one <a>...</a> element (the reply link itself). This is present even if you have disabled the feature. --Redrose64 🌹 (talk) 22:10, 11 March 2022 (UTC)

As a point of fact, the [reply] tool was created as the result of the mw:Talk pages consultation 2019. During that five-month, multilingual, multi-wiki consultation, hundreds of editors shared their opinions about what should be built (and not built). @Alexis Jazz, I believe you were mostly editing at Commons then: c:Commons:Talk pages project 2019 was not especially well-attended despite multiple announcements and people linking to it in other discussions (e.g., [1][2][3]). Commons contributors mostly seemed to post at the central discussion rather than the project-specific page. The English Wikipedia's local page is at Wikipedia:Talk pages consultation 2019/Phase 1 and Wikipedia:Talk pages consultation 2019/Phase 2. Between the two discussions here at the English Wikipedia, I think we had something like 150 editors posting 900 comments – just in this one community. Overall, there were thousands of comments from at least 600 editors in at least 20 languages at more than 20 wikis. Since then, there have been dozens of discussions about specific details, at this wiki and also at other wikis. Multiple features inside the [reply] tool exist because volunteers asked for them after the project started, including the ability to add a custom edit summary and an extra button for adding the page to your watchlist. As for its success, I notice that SineBot has made fewer edits since this tool was deployed last Monday.

The visual editor, too, was requested by volunteers. The first request that I'm personally aware of was made in 2004 (i.e., before most of us here created our accounts). The decision to create a visual editor was taken during the original 2009–2010 strategy: process, which lasted more than one year and involved editors from around the globe. The idea of creating a visual editor was one of the most strongly supported proposals during those discussions. The visual editor was IMO deployed too soon, and they did not take my advice to deploy it at another wiki first (the English Wikipedia's articles have the most complex formatting, and therefore is IMO not the right place to begin deployment of any new article-editing tools), but these subsequent mistakes do not change the fact that this idea originated from volunteers. At this point, the visual editor is relatively popular, with half of newly registered editors choosing the visual editor for their first edits at this wiki (there's a far higher percentage at some other Wikipedias) and the visual editor being used approximately once for every two edits made in the 2010 wikitext editor here at the English Wikipedia. Even people who dislike the concept of a visual editor on principle would rather add, delete, or rearrange columns in a table by clicking three times in the visual editor than manually typing the wikitext code on each line of the table.

The original name for Flow was "LiquidThreads version 3". LiquidThreads was originally proposed by a volunteer, in response to the English Wikipedia's WikiProject Aircraft voting to hold their discussions off wiki. At that time, the WMF had zero staff. The first version was written by a volunteer. The conceptual shift from LiquidThreads version 2 (which is still used at Wikinews) to "Flow" (so named because it was meant to support a Workflow pattern built by local admins and tech volunteers, rather than being restricted to simple discussions) was the request to make it capable of handling Wikipedia:Articles for deletion and Wikipedia:Arbitration/Current (imagine a world in which ArbCom clerks didn't need to manually count the votes). IMO it never got far enough to show its real promise for managing workflows, but the goal was to extend existing, volunteer-conceived software to meet the needs of existing volunteers. Even in its limited discussion-only state, Flow has been popular in some communities (not, however, this one).

I would like to particularly encourage everyone here to offer the kind of "practical" feedback that Alexis suggests above. Yes, we're apparently stuck with some things, and other things are built because other communities requests them (or other parts of this community – the newcomers at the Teahouse probably benefits more from the auto-signing and auto-indenting [reply] tool than the regulars at VPT, after all). But I'm certain that, no matter which product is being discussed, that a clear articulation of needs and goals would be helpful and very much appreciated.

If you have any experience with writing a User story, then I've found that to be an effective model when talking to product managers. The idea is to say what you want to accomplish rather than how you want to accomplish it. To give an extreme, but hopefully illustrative, example, you want to write something like As a RecentChanges patroller, I want to know whether this new person is likely to speak English well, so I can share relevant advice (e.g., a link to WP:EMBASSY or a link to a help page written in the dominant language of their country) or As an admin calculating a block range, I need to know what IP range this vandal is using, so that I can block the correct range. What's not useful is As a RecentChanges patroller, I need to know the editor's IP address because that's how we've always done it. Goal-oriented stories sometimes result in magic buttons that say things like "☑︎ Block the nearby IPs, too". Whatamidoing (WMF) (talk) 19:08, 14 March 2022 (UTC)

I think the IP-masking reps have been making significant efforts to engage, but the change they're tasked with is fundamentally more distasteful to the communities than DiscussionTools, so they are stuck pushing it uphill. ⁓ Pelagic ( messages ) 06:49, 17 March 2022 (UTC)

Pelagic, they tried to engage, but when questions are asked (like "what's it gonna look like?") in response to that the WMF goes quiet. Even if they have no answers yet, just saying that would help a lot. — Alexis Jazz (talk or ping me) 22:23, 18 March 2022 (UTC)

Updating the protected page symbols?

Protected pages on wp currently use lock symbols to indicate restricted editing. However, in the open access / open knowledge movement more generally closed vs open locks are used to indicate closed vs open access resources (indeed we also use {{Closed_access}} to indicate paywalled resources). Because of this, it's left our symbology out of synch with the wider use in online resources. If we were creating the icons today from scratch, what would we use? I wonder whether something more like etc might work better. What do people think? T.Shafee(Evo&Evo)^talk 05:24, 11 March 2022 (UTC)

I know you're probably just looking for general ideas and this isn't a formal proposal or anything, but for anyone looking to make this a thing I would suggest checking out the 2018 RFC that led to our current images being adopted for a good format to ask the question. Primefac (talk) 07:20, 11 March 2022 (UTC)

Oh crap, is it that time of year again? --Redrose64 🌹 (talk) 21:28, 11 March 2022 (UTC)

@Evolution and evolvability, if we were creating our symbology from scratch, I think you're right that shields might've been preferable to locks. But locks aren't terrible, and at this point, the community is so used to locks that it's unfortunately not even worth talking about; it just isn't going to happen.

The only big icon update that I think it's really worth our time to push for is the FA and GA symbols. There's also currently an open discussion on the taxonomy box pencil icon where I'd love to have your and others' input. {{u|Sdkb}} ^talk 05:25, 13 March 2022 (UTC)

Community statement

I've drafted a potential community statement on the arrest of a Ukrainian Wikimedian and the Russian government's threats to censor Wikipedia at meta:2022 Russian invasion of Ukraine/Community statement. Before opening it up to signatories and promoting it widely, I'd appreciate feedback. Should anything about it be changed? {{u|Sdkb}} ^talk 00:54, 15 March 2022 (UTC)

If you want to add a little extra to it, it could mention Wikipedia as being a source of credible info for the Ukrainian and Russian people.North8000 (talk) 02:55, 15 March 2022 (UTC)

Article we talk about today at WikiCanada chat Wikimedia says it ‘will not back down’ after Russia threatens Wikipedia block.Moxy- 03:08, 15 March 2022 (UTC)

Good article; already linked! And yeah, that could be a nice thing to mention if we can find a place to put it. {{u|Sdkb}} ^talk 03:25, 15 March 2022 (UTC)

@Sdkb: - I really like the letter content. Perhaps the only thing that comes to mind is that the header (currently "Russian invasion of Ukraine") suggests a more "support Ukraine against the Russian invaders" focus, whereas the letter is (rightly (for Wikimedia)) more on the censorship by Russia & Belarus. Without the title becoming too long, perhaps a rewording to make the title more directly clear to meaning? Nosebagbear (talk) 15:33, 15 March 2022 (UTC)

Give WMF feedback on model cards

The WMF Privacy and Machine Learning Platform teams are developing model cards to increase visibility, transparency, and accountability of algorithmic decision-making on WMF platforms. The broad goal is for every ML model hosted by WMF to have a model card for the community and public to understand, discuss, and govern that model. As part of rolling this out to more models, we want to get feedback from as many people as possible.

To that end, we would love for you to give some feedback on the talk page of our prototype! Of course, please feel free to ask any questions/give any general comments as a thread to this conversation.

Thanks :)

- Htriedman (talk) 19:00, 17 March 2022 (UTC)

What is a "model card"? Jo-Jo Eumerus (talk) 19:03, 17 March 2022 (UTC)

Sorry, I'm sure you are trying your best, but the only feedback that I can give is that the WMF should sack the 90%+ of its staff who work on such irrelevant projects, stop trying to raise more money than it needs, and concentrate on providing the hardware and software infrastructure that we need. Phil Bridger (talk) 19:08, 17 March 2022 (UTC)

Harsh words, but fair. Certes (talk) 20:32, 17 March 2022 (UTC)

You shouldn't really hijack any thread made by WMF staff with off topic comments about this sort of thing. I'm sure you have valid concerns but it seems quite rude, especially as this *is* software related and just because you might not understand it doesn't mean its not useful to other members of the community. ✨ Ed talk! ✨ 20:53, 17 March 2022 (UTC)

Agree completely. It's almost disruptive to reply to any WMF thread as if it's a WMF soapbox. A. C. Santacruz ⁂ Please ping me! 21:02, 17 March 2022 (UTC)

These model cards seem very useful! Machine learning can get quite mysterious, especially when the data starts getting and I feel like this openness definitely helps.

Whist a push, deeper accessibility to run the models through APIs, similar to what ORES has, would be great too for userscript and tool makers, but I'd strongly recommend some sort of authentication check be done unlike ORES to make sure quotas are adhered to. Maybe if this content was merged onto the new API Wiki (api.wikimedia.org, I know it isn't quite an API but a central spot for all technical docs would be helpful) or Wikitech that would be nice. ✨ Ed talk! ✨ 20:58, 17 March 2022 (UTC)

• Jo-Jo asked a pertinent question above… please do not assume that people know what “model cards” are and what they are supposed to do. Explain! Blueboar (talk) 21:26, 17 March 2022 (UTC)

Thanks, all, for the helpful comments. Model cards are a kind of standardized documentation about machine learning models — something akin to a nutritional label. Ideally, they should be standardized in their look and content, understandable by people who aren't domain experts, and provide clear answers to basic questions (who, what, why, how, etc.). Although the paper I've linked provides a specific schema for model cards, I've been figuring out what that kind of documentation should look like on various wikipedias. If we are going to host models that have a very large impact on what wikipedia looks like (e.g. the ML models that power various patrolling tools), and scale up their utility in the future, this kind of documentation is useful — and in the medium term it may be required by law in the EU for models that have an impact on basic human rights, like online speech. Htriedman (talk) 22:50, 17 March 2022 (UTC)

LOL... I still have no clue what you are talking about... but thanks for at least trying to explain. Blueboar (talk) 18:39, 21 March 2022 (UTC)

Give page movers the right to move past move protection?

I want to know how people feel about this.. So, there's this right called Page mover, and it allows users to do stuff related to moving pages (suppressing redirects, overriding title blacklist.) Separately, there's also this thing called move protection, which makes it so only administrators can move the protected pages.

So, uh, if there's a right that specifically makes it so one can move more pages, then why is there a protection on moving pages that doesn't let them, the people trusted to move pages correctly, through?

(I've never really done a proposal before so I wanted to hit by here first to check if I'll get snowed out in an actual proposal of this) casualdejekyll 01:39, 18 March 2022 (UTC)

@Casualdejekyll please first read through similar discussions from August 2018 and April 2021. — xaosflux ^Talk 01:48, 18 March 2022 (UTC)

Well, @Xaosflux, the April 2021 proposal appears to have a lot of votes that oppose but would be neutral on PMR-level move protection existing. That doesn't make much sense to me - the proposal is to replace admin move protection with PMR-level move protection, so they're saying the oppose and are neutral at the same time? casualdejekyll 01:59, 18 March 2022 (UTC)

@Casualdejekyll I think one of the technical challenges is that to move a page you have to be able to edit the page, so this would require a lot of technical work for something that doesn't occur that often: pages that are full-protected against "move", but not full-protected against "edit" - that actually need to be moved - and that the best solution isn't to reduce the protection. You would need to come up with some good statistics to show that the scope of the problem is large enough to build a new technical system. — xaosflux ^Talk 09:25, 18 March 2022 (UTC)

A quick glance through Special:ProtectedPages shows that about 80% of the first 1,000 move-protected non-redirect articles are also full-protected. @Casualdejekyll, it looks like you don't have this user right. How often has this situation affected your own editing? Whatamidoing (WMF) (talk) 18:37, 21 March 2022 (UTC)

Comment section after citations

It should be possible to add a little snippet of text after a citation that says what the source is/what it says. This would be especially helpful for long sources like PDFs where people could say what page they found the info in question on or just put a quote of what it says. — Preceding unsigned comment added by Ranicher (talk • contribs) 17:24, 19 March 2022 (UTC)

That is possible. If using free-form citations then anything you like can be entered, and when using citation templates such as {{cite book}} or {{cite web}} the page= and quote= parameters can be used. Phil Bridger (talk) 18:06, 19 March 2022 (UTC)

For works without page numbers, you can use the parameter "at=" (instead of "page=" or "pages=") (see Template:Citation#In-source locations) in cite templates to indicate section/chapter headings or other identifiable markers in the document. - Donald Albury 21:01, 19 March 2022 (UTC)

Even when using a citation template instead of free-form text, if one of its parameters doesn't quite fit your need, you can add a snippet of text after it. isaacl (talk) 21:13, 20 March 2022 (UTC)

Proposed additions to the Main Page

The purpose of this RFC is determine what sections should be added and removed from the Main Page. I have identified a few features missing from the main page which would make Wikipedia more interactive. Other editors are welcome to add proposals here. Interstellarity (talk) 20:57, 19 March 2022 (UTC)

Proposal 1: Add links to the level 1 vital articles to replace portal links on the top right

This change would show that the purpose of the vital articles project is for readers and not editors.

Support

1. Interstellarity (talk) 20:57, 19 March 2022 (UTC)

Oppose

1. Oppose the WP:VA is more of a WikiProject than anything else. Whilst they are some of the most important articles, the VA list shouldn't be commented on from the main page. Best Wishes, Lee Vilenski ^{(talk • contribs)} 23:25, 19 March 2022 (UTC)

Neutral

Discussion

Proposal 2: Add a list of the most popular articles of the week

I'm always interested in what articles are popular every week and it might incentivize experienced editors to improve the articles.

Support

1. Interstellarity (talk) 20:57, 19 March 2022 (UTC)

Oppose

Neutral

Discussion

• I believe the mobile/app version of Wikipedia has this already, or something similar; it gets discussed occasionally at WP:ITNC by some who want to see the In The News box be more like a most-read article list. 331dot (talk) 00:06, 20 March 2022 (UTC)
• I'm not sure about this one. Don't articles linked from the Main Page get a large increase in page views? If so, then anything on the Main Page would then tend to end up in the "most popular" box, leading to those articles getting more page views and staying as the "most popular". Can you clarify more what would count as "most popular" and how it would account for increased page views from the Main Page? RudolfRed (talk) 20:26, 20 March 2022 (UTC)

  What about the most popular pages last week? Enterprisey (talk!) 22:23, 20 March 2022 (UTC)

  @RudolfRed and Enterprisey: To be clear, I am talking about the most viewed articles in a given week. We could possibly list the ten most viewed articles this week as well as the ten most viewed articles of last week. Interstellarity (talk) 22:50, 20 March 2022 (UTC)

  Won't that cause a positive feedback? I believe it would tend to keep an article high on the list for the next week. Donald Albury 01:07, 21 March 2022 (UTC)

• You guys are aware of the existence of the Top 25 Report, right? casualdejekyll 18:49, 21 March 2022 (UTC)

Proposal 3: Add quizzes throughout the Main page

I think creating quizzes on Today's featured article, In the news, as well as Today's featured picture would be great because I like to test my knowledge to see what I know about the article.

Support

1. Interstellarity (talk) 20:57, 19 March 2022 (UTC)
2. YTKJ (talk) 18:45, 20 March 2022 (UTC) Yes, this sounds like a fun addition to Wikipedia.

Oppose

1. User:力 (powera, π, ν) 23:18, 19 March 2022 (UTC)
2. completely against the point of Wikipedia. Best Wishes, Lee Vilenski ^{(talk • contribs)} 23:21, 19 March 2022 (UTC)
3. Oppose, per Lee. RudolfRed (talk) 20:24, 20 March 2022 (UTC)
4. Oppose, per Lee.--User:Khajidha (talk) (contributions) 20:34, 20 March 2022 (UTC)
5. Nah, but why nah? One, this would require some software work, and software developers are in short supply. If someone really wants to write a quiz extension lets see what it looks like first. If someone wants to write this as a javascript - lets see the example again. This would require volunteers to write and maintain questions and answers. Now, if some wikiproject really really wants to do this, and do it as a script, then maybe I'd be ok with it being an opt-in / onclick load gadget. But would want to see some proof of concepts on both the software and the process first. — xaosflux ^Talk 21:17, 20 March 2022 (UTC)

   A first iteration doesn't even need to be interactive: just have the questions and answers on separate pages. I agree before considering it for the main page, there should be a track record of the quiz creation process running regularly and smoothly. (I don't know if the Signpost is interested in a column like this; it could be a place to try it out.) isaacl (talk) 21:34, 20 March 2022 (UTC)

   The software already exists; see mw:Extension:Quiz. It's used at both Wikibooks and Wikiversity, and maybe other places. For the record, just because the software exists does not mean that it can be used on the largest wiki in the world. But if you wanted it (for any reason, not necessarily for this proposal), we could ask. Whatamidoing (WMF) (talk) 18:43, 21 March 2022 (UTC)

Neutral

1. I like the idea in general. Quizzes are fun, and add interest. Even the NY Times has a weekly news quiz. But before we can consider putting it on the front page, the idea should be tested somewhere else to see how well it works. -- RoySmith (talk) 18:54, 21 March 2022 (UTC)

Discussion

Proposal 4: Add Browse by category at the bottom of the Main Page so readers can look for a specific article without using the search engine

There will always be readers that are not quite sure what article they are looking for. Having basic categories like People, Science, Technology, and Society would make the user interface more usable rather than using the old-fashioned categories.

Support

1. Interstellarity (talk) 21:02, 19 March 2022 (UTC)

Oppose

1. I don't think throwing a reader at Category:People is going to be very useful. The category system is decent for finding related low-level topics, but not so much for browsing such broad categories such as those proposed. — xaosflux ^Talk 21:20, 20 March 2022 (UTC)

Neutral

Discussion

• Wikipedia already has this, they are called Portals, but they have kind of diminished in visibility and importance over the years. I'm also not really sure it is easier than simply typing in the article or subject one is looking for in the search bar. 331dot (talk) 00:01, 20 March 2022 (UTC)
• In addition to the portal links (which oddly enough you suggested removing), for desktop readers there is a "Contents" link in the sidebar that provides some navigational guidance. Perhaps this link should duplicated in a more prominent location. isaacl (talk) 21:28, 20 March 2022 (UTC)

General discussion

• Interstellarity There are countless attempts to change the main page, but never consensus for them. I would urge you to better hash out ideas at the Village Pump first. I would say that quizzes are probably a nonstarter, this is an encyclopedia, not a gaming site. 331dot (talk) 23:18, 19 March 2022 (UTC)

  I don't mean to shoot down your ideas. But changing the Main Page of one of the top used websites in the world requires a lot of time and effort, usually from multiple people and more than an RFC. I would gently suggest that you withdraw these proposals from this page and discuss them at the Village Pump. 331dot (talk) 23:30, 19 March 2022 (UTC)

  @331dot I have moved the proposals from the main page talk to the idea lab so that they can be developed further. I have converted them to informal proposals so that they can be discussed. BTW, the reason why I suggested quizzes on Wikipedia is because Britannica, another encyclopedia has quizzes on the site. Why would people be against quizzes? Interstellarity (talk) 23:51, 19 March 2022 (UTC)

  Interstellarity Wikipedia is a user-built collection of human knowledge. It is not a place to test the knowledge of its users. I would not be surprised if there were outside websites that use Wikipedia information as a basis for quizzes. Built-in quizzes are a distraction and outside our mission. 331dot (talk) 23:57, 19 March 2022 (UTC)

  OK, I see now. It seems like Wikipedia leans toward providing the information and just an encyclopedia while Britannica is more of a multi-purpose website that's more than an encylopedia. Interstellarity (talk) 00:00, 20 March 2022 (UTC)

  Yes, and that works for Britannica, because their mission is broader than Wikipedia's. One is not worse than the other, it's just different. 331dot (talk) 00:04, 20 March 2022 (UTC)

  And maybe Britannica has to do those other things to survive because we have outcompeted them in the niche of online encyclopedia.--User:Khajidha (talk) (contributions) 20:55, 20 March 2022 (UTC)

  I don't think anyone would stop you from creating quizzes in your own user space, if that's what you're interested in doing. Before it can be even considered for a prominent spot for readers, though, there should be a strong track record of new quizzes being regularly created, and a few people regularly assuming this duty. isaacl (talk) 21:24, 20 March 2022 (UTC)

• We aren't an "interactive" website though. It would have been more prudent to get some input before creating an RfC. Best Wishes, Lee Vilenski ^{(talk • contribs)} 23:23, 19 March 2022 (UTC)
• Why does this have Support, Oppose and Neutral sections? Per the box at the top, this page is not for consensus polling. Stalwart "Oppose" and "Support" comments generally have no place here. Instead, discuss ideas and suggest variations on them. --Redrose64 🌹 (talk) 07:56, 20 March 2022 (UTC)

  Probably because it was originally on talk:main page and got moved here.Nigel Ish (talk) 10:11, 20 March 2022 (UTC)

• We already have a Browse by category function. It appears top right but is labelled All portals. It is currently proposed for removal. Certes (talk) 12:11, 21 March 2022 (UTC)