Knowledge:Village pump (proposals)/Archive 211

16024:(quite strongly), because it this would prevent all translation of articles from wikipedias that accept general referencing (which, incidentally, still technically includes our own). (1) Yes, the translator ideally will get the original source, track down the information, and convert general referencing into inline referencing. But often it's impossible to track down all the sources, and we have to assume at least some good faith that the original author did use the sources they provide. A translation is still a starting-point for later editors to convert the general references to inline, and to find new references. Knowledge articles aren't expected to be complete at first appearance. (2) We promise our readers that everything in Knowledge can be traced back to a source. We don't promise how much leg-work it will require. Even an inline reference can require work, if it's just a book with no page number. Or a book that's really hard to track down. Sometimes an article here is quite short, and has a couple of general references that themselves are either quite short, or have a clear section obviously on the subject (e.g. references to Grove in biographies of musicians), and general referencing is straightforward and reasonably helpful to the reader. We are also not obliged to pander to hypothetically incredibly lazy readers who want to be told that the information is in the third word on the fourteenth line on page 37. There are circumstances when general referencing isn't evil, and we should trust AfC and the new page patrol to use their discretion to accept such articles even if not inline-referenced. 9451:. The current central banner system is somewhat sufficient for this purpose. Even if we want to work on the sister projects, one has to be cognizant that every project has different levels of maturity and sets of social norms. We shouldn't be the ones pushing ourselves and sensibilities over to the other projects. If so, can the targeted project take on an influx of editors who are bringing a foreign set of norms and sensibilities over to their projects? As one who occasionally straddles between English and Chinese Wikipedias, as well as commons, I can say that it takes awhile to pick up the norms of the individual projects, an effort that would likely be significantly longer than the run on the banner. If the other projects would like to have us and other projects to contribute, have them to come up with their own campaigns, banners through the central banner system, and processes. If anything, I think improvements in other parts of the interface may be more feasible, for example a better call-out to translate the article if it is known that the viewer is good in a certain language, or call outs somewhere to contribute to wikivoyage if somehow we know if the viewer is actively editing in geographic based articles, Wikisource if one is editing articles related to manuscripts or other types or texts and printed materials. 6939:

information about policy is non-trivial. 3) Because our policies are wiki articles themselves, they too are constantly evolving; without constantly updated training, the AI would forever be operating from an outdated understanding of what our policies actually are. 4) This is not really within the purview of a single project to pursue nor is it likely to gain broad consensus across a wide variety of projects and languages necessary to make it worth the effort and cost. The scenario you're describing is better suited as part of the MediaWiki interface. 5) I suspect there are also non-trivial concerns about license compatibility as well as ensuring an open-source software/tools stack for this. 6) Where is this model going to be hosted? Who is paying for the compute time? The foundation? Some foundation-adjacent entity? Donations? A private research institution? There are too many unanswered questions and this proposal addresses none of them in a way that shows sufficient time and thought was put into making this a realistic suggestion. AI is not a "magic bullet" solution to problems that you haven't validated actually exist; nor that it is a product fit for what this community actually wants.

7259:. Happy to be invited to join this discussion. By way of introduction, I am a new editor who has done some deep diving into AI as well as algorithmic and data biases. As a newcomer, I can say that navigating Knowledge's trove of policies and guidelines can feel very daunting. I’ve read posts in the Teahouse from new editors who say they feel paralyzed due to their fear of being criticized for doing the wrong thing or making mistakes. As such, being able to get easy access to policies would help to reduce policy breaches, support retention of new editors, and help create a “psychologically safer” environment for all editors. In my experience with AI and LLMs such as ChatGPT, they are very efficient when they are dealing with well-defined inputs for a request. A set of Knowledge policies and guidelines would be an example of well-defined inputs. Generally, when LLMs stumble into "hallucinations", it tends to be when they are tasked with open-ended topics such as “blank sheet of paper” requests where they need to create something new with no prior input other than their own information. 8351:

thing in mind is ChatGPT and its fuzzy mistakes and/or the wrong hands that image making 'AI' hallucinates. Artificial Intelligence is much more that this, and I am sure many of you are clear to this. Now what if I assure you that the AI will make equal or lesser percent mistakes than ClueBot NG in terms of overall damage to Knowledge? My point is that I am sure if I had proposed inculsion of bots to Knowledge (assuming they weren't involved yet) I must have met a fierce resistance than I am getting now. By AI, I don't mean hallucinating ChatGPT or an AI which will make Knowledge human editor-less but a algorithm that will better serve the knowledge gap regarding 'policies' and 'guidelines' to new editors, so they can edit without fear of getting 4 warnings quickly and blocked, all because they never knew what they were doing wrong. If you still oppose the idea, please list the reason and I will try to be more objective in answering those.

7768:- First and foremost, I understand and respect your vision to expand Knowledge but I fear that AI (such as ChatGPT) is not to the level that it should be. I understand that AI has been around for quite a long time, but only recently has it been ignited to the public and is used through numerous high-end companies (such as Google, Microsoft, Apple, you name it). I oppose this because Knowledge is about humans checking the article. And yes, it would make it go faster, but it will make mistakes and those mistakes could time consuming for an individual (or a community) to resolve them. I think we should pilot it to a few selective articles rather than making it public to the entire Knowledge articles. I think it's a great idea, no doubt, but we should be cautious when it comes to making it public for all without really piloting it. I have a couple more opinions on it, feel free to reach out if you want to. 4833:- CU and OS are both high-trust roles with real-world legal implications, with access to advanced tools to manage very sensitive information, which can do serious real-world harm to real persons and the project itself if they are misused. They are chosen by a committee which understands the purpose of the roles and evaluates individuals' aptitude and suitability for them, as well as whether or not a candidate fills an identified need. The community elects that committee, Arbcom, and community input is part of the appointment process. Making these specialized positions open to general election puts us at fairly serious risk from motivated malicious groups wanting access to the politically valuable information available; frankly it's already too much of a risk IMO that Arbcom members gain access to these tools through election. 4774:

admins - adminship involves a large number of often very subjective judgments on things like eg. whether someone is likely to continue to be a problem in the future and what sanctions are necessary to prevent this. CU / OS, by comparison, are much more rigid and technical - there is a bit of subjectivity around the edges, but not nearly so much, and it mostly comes down to very simple categorizing using a few straightforward rules. This means that while judgment is important for those roles it doesn't imply as much need for additional community trust. Admins need to understand, and be trusted to accurately enforce, essentially all of Knowledge's policies and practices; CU / OS are each mostly about a much more narrow slice of it, and a slice that generally involves a bit less subjectivity at that. --

7029:

by all of my points -- your idea is incomplete, premature, and lacking sufficient detail or information to be executed on even if there was consensus that it was desirable. You can't simply handwave away critical considerations like "who's going to pay for this" and "is this compatible with our open-licenses and our mission" as trivialities that we'll figure out later. That kind of techbro fastspeak won't fly here. We are the 7th most popular website in the world; even small changes here can have drastic impacts on millions of people. The burden is on you to show that you've actually given the serious consideration and planning due for a change that would affect people in that magnitude. Until then, you don't have a proposal; you have a fantasy.

3332:(other than user pages) require all types of protection to be shown on the top right. The lock icon will display in the same way shown in the protection policy page (linked above). With move protection, the only way to tell that the page is protected is if the move button is not seen (this article for example, doesn't have the move button for me, but it does have subscribe) with the possible exception for administrators as they can move any page even if it's protected. Even if accounts are able to edit semi and EC protected pages, the relevant lock is still there. But like I said, that will not change. Admins will also be able to see if a page is move protected from the green lock icon at protection policy page. 12526:: Sure thing. The contents page is the main gateway for browsing Knowledge. While most people search for a specific article, it would be helpful for people like me that like to browse Knowledge. On mobile, the contents page is not easily accessible compared to desktop so I think in this way, we can make it more accessible in mobile. The contents page has a lot of ways to browse Knowledge by category. Whether people like overview articles by category like looking for some wars going on in the history section or looking at vital, good, and featured articles, it's a great way to be able to browse what Knowledge has to offer. I hope you consider my reasoning and that you understand where I'm coming from. 4407:

fulfill specific purposes. They carry no "social" value (no accountability requirements, etc.), and if someone who hasn't held an NDA role before asks for it, the whole thing basically boils down to "will this person invade upon privacy or leak personal information?". I sure as hell cannot answer that question. Can you?...yeah, you probably can, because you've probably been here for 15 years or something. But you are not the only demographic that goes to RfX, and with this system it won't be any different.If the problem is that you want arbcom out of the way, then sure, I could potentially get behind that. But experienced, non-temperamental people need to make this decision, not the RfA mob.

7457:

queries and responses could we possibly train it on? How will the AI know how to balance essays, guidelines, and policies, or how to recognize an essay that nevertheless has broad buy-in from the community and is a guideline in all but name? How would the AI know how to adapt to IAR, to contradictory essays and guidelines, or to the fact more broadly that our documentation is descriptive rather than prescriptive of Knowledge practices? There is no publicly-available dataset that we could train an AI off of that would even begin to address these use cases. I am raising these concerns as someone with several years of professional experience in AI research.

2361:- FA/GA topicons highlight Knowledge's best content in-place. Readers shouldn't have to go to a talk page or a list somewhere else to see that the content they're reading has been reviewed and is considered Knowledge's finest, and there is benefit to distinguishing it from content that hasn't been through or has failed such a review. Ratings below that level are part of a content rating system that's been dead in the water for 20 years (though projects still use it) and there's no point in highlighting it for readers since there are practically no standards to the ratings, but if you want to see these ratings when reading articles anyway you can enable 5191: 15593:. To be sure, a new, truly unsourced article is very likely to be hit by NPP for draftification, prod, or AFD (perhaps even when it really shouldn't). It's good practice to include at least some references. However, some editors take a narrow view of sourcing and consider implicit sources as "unsourced" when it really isn't (i.e. a newbie editor simply writing out "According to Joe Bloggs, blah blah blah..."). Secondly, the main case where valid "unsourced" articles exist are generally the frontiers that are good to create articles on for 7575:: In editing energy and telecommunications articles, I've noticed that several key editors who made significant contributions have left Knowledge, likely due to criticisms or disputes. These criticisms, not always expressed in a civil tone, have highlighted issues in their newly created articles or edits such as subjective language, the inclusion of original ideas, and potential plagiarism. Implementing a "policy checker" for articles before they "go live" could greatly reduce such friction, helping to retain passionate contributors. 15118:, because this creates a new type of deletion, which makes new page patrol and deletion workflows more complex. I believe that all new deletion proposals should work within our existing types of deletion. I think it would make more sense to expand the scope of BLPPROD, than to add a brand new NOSOURCESPROD that is in addition to the almost obsolete PROD (since folks always just unprod these and they end up at AFD) and BLPPROD. I will also note that many new page patrollers automatically draftify or BLPPROD articles without sources. – 8159:

out of frame, also WMF does not look to have been actively sorting out in something specific before this proposal as seen in CAlbon (WMF)'s reply above. And to the troubles which we may face in this, first we have to start somewhere, second that implementation could be step by step. This cycle is just like an unemployed being denied a job because of lack of experience which he will never get without a job which he is being denied. The ending was not reply to you specifically but rather to the discussion as whole.

14985:, in this highly unlikely scenario, the unsourced article would have saved the second editor a bit of typing. But that time saved typing is cancelled out by the time spent reading the unsourced draft and comparing it to the source material to see if it matches. That's a waste of time. I would never read an unsourced draft -- it doesn't matter what the heck is written there if there is no source listed. I would just gather the sources and write my own summary, completely ignoring the unsourced draft. It's 15674:. I would think that if one was going to go through the expense of yet another RFC on the same theme (the third within six months?), it would have at least been better thought out. As Joe Roe and others have highlighted, clearly it is not. Anyway, we should not introduce this new process, subtly different from existing ones yet also similar and overlapping in purpose. I hesitate to expand our assortment of procedural mechanisms for deletion or quasi-deletion, which is already confusing to editors. 4293:

unsuitable are closed as unsuccessful, community input is sought for all the others. All the comments (arbs, functionaries, community) are then evaluated by arbcom and the successful applicants appointed. An RFA-like process would lead to mostly the same people being appointed as now, just with a greater likelihood of acrimony, incivility, and clearly inappropriate nominations and an increased risk of untrustworthy people being appointed. I'm not clear what problem this is attempting to solve.

3015: 2968: 1287: 2063:(whether in a separate section after the description, or in the nominator's initial !vote) there's expected to be some reasoning for why the heck we're voting on this at all and some sort of case to modify the status quo. The RFC creator gave no such explanation other than what I already cited that "other people mentioned it" and "readers can find out on the talk page" which is not conversant at all with the issues involved. People could propose random changes all day; why is 16186:. An issue for someone who's new to Knowledge, still learning wikitext etc. is that adding a citation in the correct format isn't entirely straightforward. Even using the provided templates, you have to know which detail belongs in which field. It's another technical thing to be learnt. It's quite possible that the reason someone hasn't added any citations yet is that they don't feel confident adding them. "How do I make the references appear in the right place?", "What does 8969:: it's common knowledge that enwiki's love for sister projects is... lackluster at best. Thank god we're not in charge of creating new projects, because otherwise there wouldn't be any. But raising awareness for sister projects raises the probability that we'll be able to help someone find what they're looking for (like, say, a quote or a definition) next time they need something. Maybe they're looking for that information right now, and don't know where to find it! 9221:, good points above on the value of ads and on being cautious with the power of en.wiki within the WMF ecosystem. On a pragmatic note, a general link to go look at WikiSource or WikiData or even WikiCommons is of any use. How to contribute to those is opaque at best. If these resources are promoted, it should be through specific useful ways, such as our practice of including relevant WikiSource pages in See also sections (occasionally they pop up in infoboxes too). 4789:

as can be. CU, on the otherhand, is far from categorizing a few straightforward rules. There is a huge amount of discretion involved and judgement calls are regularly required in terms of whether a check has been justified and if it has whether the results indicate a violation of policies and guidelines. I also think most admins only operate in a narrow slice of the admin universe so they don't actually need to know every policy and guideline. Best,

5219: 1955: 13192: 6557: 6492: 6403: 6353: 6307: 14997:, not writing a summary of secondary sources. Writing off the top of one's head about a topic is not the same thing as summarizing sources. And if somebody's off-the-top-of-their-head blog happens to line up with an NPOV summary of high-quality RS, that's just like a freak coincidence, man. That is not what we should be striving for, or even tolerating, on Knowledge. "But what if the bloggers happens to be right?" is an lol argument. 15631:

topic/subject actually exist. It's not that difficult. If you want new editors to understand the new PROD criteria, why not being more clear in the Article Wizard that Knowledge articles need to have sources? And if my proposal sounds BITEy, then that's because the whole PROD/AfD process itself is BITEy. Please, if that's the reason why you opposed my proposal, please suggest changes to PROD and AfD instead. This is not my fault.

15534:. I also see this as unlikely to succeed, but I'd prefer to see the project move into this direction. I think it would be an improvement to require that article creators include at least one source, and I do not see current procedure as sufficient to deal with the unsourced article problem. I can understand the objections based on grandfathering, but I would prefer the harms of temporary inequity over the harms of doing nothing. 15559:, but geez y'all, it's 2024, citations are expected in everyday culture now, anyone dumping an unsourced article into mainspace (without followup edits) nowadays is willfully ignoring our requirements and clearly has no plans to join the community. We don't need to protect these precious new editors, and we don't need to retain unsupported information that would be deleted if it was in any other format than a standalone page. 15635:

most new editors nowadays don't start out editing in wikicode, they start out with VisualEditor. When they fire up VisualEditor for the first time there is an explicit instruction dialog hovering below the citation button that instructs you how to make an inline citation. It's really not that hard to do for a beginner. Even if VisualEditor somehow cannot process your URL, you can always type that citation in plain text.

5622: 4279:? That wasn't really a direct community process. It started with arbcom being able to veto any candidate and was then advisory in nature only. 15 years is a long time in the project, I'm not sure we should discard an idea based on such an old example. There are certainly things that can be learned from it - along with years of experience in how many other projects have been able to run CUOS elections themselves. — 8570:

digitized public domain texts that anyone can download or distribute; a travel guide; a media repository; and many others. The sister project links are currently buried far down on the Main Page, and are especially distant for mobile viewers who make up an increasing share of our readership. Why would we not want to help readers discover some of the useful resources our sister projects have to offer?

9023: 5161:

likely would not support an application from someone who frequently asked us to suppress things that clearly don't need it, and I likely would support someone whose reports were almost always of things that do need suppression. However, editors who are not oversighters do not have access to this information (indeed when oversight works perfectly almost nobody knows it has happened).

14260:. I don't love grandfather clauses either, but I think this is the least bad of the available solutions (the most bad being "Do nothing; just keep accumulating unsourced material indefinitely".) At some point, we have to turn the water off to the gushing pipe, and then we can focus on cleaning up the remaining mess. We really need to get across the idea of a "reverse BEFORE": 9267:

platform, and so it seem highly inappropriate and immoral and incestuous that we don't allow banners which may improve or even save people's lives, but will allow banners purely to inflate the traffic flow of various WikiMedia projects. If there is an appropriate need to link to another project, that is already done within articles, such as via External links.

5081:

a very well attended RFC for a project this size (where the supporters would likely already be over 100 as well). The best case would be that if successful a graceful transition from the current committee happened. I could see the stewards team stalling any appointment pending additional community deliberations over collisions between multiple policies. —

7405:- The state of the art of AI does not have sufficient quality assurance to be able to operate in such an open-ended domain as Knowledge policy and guidelines. This proposal is begging for trouble, and there is currently no time frame for when AI will reach this level of sophistication beyond a useless guess of "maybe in a few years at best". 4593:

areas like edit-filter work. Non-admins cannot scrutinize most of this activity at all. While trust and judgement are important aspects of being a functionary, there's more to it than that. Community scrutiny is necessary but not sufficient, and the current system of the community being able to provide feedback is exactly what it should be.

1952:- I agree with Levivich that the quality of the GA process can be inconsistent, but in practice I still find that they are almost always better than your average non-good article. Yes there can be stuff like what Doug Coldwell did, but the same goes for every other part and process of Knowledge. Most readers aren't likely to grasp what the 16546:. How many long term unsourced articles are hoaxes? How much misinformation is there sitting around on the project, potentially fooling readers? I think the draftify process works relatively well for new articles, the problem is long-term unsourced articles and unfortunately the community has shown an unwillingness to take it seriously. 8033:. Using ChatGPT to augment the wikipedia search engine isn't a good idea. It hasn't been tested enough, isn't built for searching, and doesn't like certain queries (medical, legal, political, and anything containing the phrase "root access"). If Wikimedia wants to build its own AI to help with searching, however, then I do support this. 7515:

cause a new editor to fix their bad edit before another editor had to deal with it, it could make things better. That said, the model is only a small part of what would be needed to make this work, a bigger aspect would that the UI/UX that makes it seemless for a newer user while not annoying experienced editors (maybe make it opt in?).

13799:(OTOH, the counterargument could well be made that this just shows that the proposal is the best kind of wiki-rule: one that simply codifies existing practice to prevent future confusion. But if that's the argument it would likewise be helpful to have some quantitative details showing how this proposal maps onto existing practice.) -- 9841:, where editors representing different initiatives could host "open house" activities to engage potential volunteers. I didn't pursue the idea, though, since there was almost no interest expressed. However if it were to occur, it would be an opportunity for editors from other projects to set up open houses to publicize their work. 1278:- due to the fact readers have no clue what these mean and the fact the majority no longer meet their FA and GA criteria and 60%+ are mobile viewers don't even see them anyways. I can see how they are an incentive to get editors to improve articles though. That said I would slow down on rfcs that will go nowhere..... Time wasters. 16608:

articles and you make it far more likely they won't develop into a productive editor who improves the encyclopaedia. Short term, arguable improvements should never come at the expense of long term definite improvements. Especially as this proposal would speedy delete articles that have sources which just happen not to be inline.

14298:

appetite to apply this rule to articles that already exist, and without a mechanism to move the grandfather date automatically we'll have to keep having discussions to move it around which wastes time and risks running around consensus by tiring people out with constant discussions. I don't see this going well in practice.

7644:

least a minor change which may make it a process easier (and yet efficient) to edit for Wikipedians. Also I don't think any 'third party tool' or any 'tool' is currently better than Knowledge's own "search engine" which is not very effective, I very first proposed, and even if any exists, they mostly are highly unreliable.

2346:- articles exist for the benefit of the readers, no tenth editors. A reader has no idea how good an article is without reading a significant part of it; how vital it is requires little more than a definition. There is no need to tell a reader how vital an article is; there is much more to tell the reader about the quality. 9160:

not a good way to support sibling projects if that's the goal. It's indiscriminate and for the majority of readers completely useless given what they're here to learn about. Why show a reader a banner about wikispecies if they're here reading about the Andromeda Galaxy? I don't see the benefits outweighing the costs.

13717:, draftifying does the trick already. There do not seem to be many (any?) totally unsourced articles that make it through NPP, so I am not sure there are any articles that this proposal applies to (the category mentioned by the nominator contains mostly old articles that have been recently tagged as unsourced). — 15314:. Articles on notable topics which would be worth having but don't yet meet mainspace standards (lack of sourcing is only one possible reason) should be draftified. That is a step towards eventually having a good article, while deletion is a step away; I think it is obvious which is better for the encyclopedia. 14499:. My reading of the previous discussions was that there is a consensus against (or at least a marked lack of consensus for) automatic deletion of unreferenced articles, old or new. Those previous proposals were weakened by a lack of clarity in scope and terminology, and it seems that this proposal is too: 1681:: GA and FA symbols are no guarantee of any kind of quality, but it is one piece of information that media literate readers should be using to help them assess the reliability of what they read. These topicons may also be motivating to editors who put articles they have improved through GA/FA processes. — 14883:. The first edit to any new page should contain at least one source. One source is the minimum we should require for any page in mainspace. One source is the minimum we should require from anyone creating a new page in mainspace. One source is the minimum required to show that an edit or an article meets 14846:- where the notability is based on the availability of sources, not the current state of the article where it may or may not be sourced. Finally, I think there is no need for this. Sending them to draft is enough for them to fix the article, or if they didn't want to fix it, it will be deleted anyway. 7616:– It'd be a big hassle for Knowledge to implement such a proposal and, at least to me, seems like a pretty unnecessary thing to do, since there already are such third-party tools. If anyone wants to use them, they should feel free to do so (and accept responsibility for any error), like with any tool. 16449:

moving to Draft with a PROD-like tag. Honestly, if the article author can't be bothered to source the article, it can't be that important. We are no longer barn-building here. Virtually every subject of any objkective imortance, and a good many of no objective importance whatsoever, are already here.

16374:

new articles that are completely unsourced or are built on unreliable sources should be draftified as part of the NPP process, even if the NPP reviewer finds that sources exist. The article can then be improved and published to mainspace once sources are added to it (same bar we put on AfC articles).

16241:

As also seen by the recent backlog drive, it's really easy to add a poor source to a shoddily started article and call it a day. It's much easier and better in the long haul if we expect the editor writing the article to establish it based on sources, rather than pretending it's equally worthwhile to

15937:

There are ways other than deletion that can be employed to clear the backlog. I would support a draftification process for unsourced articles to get potentially problematic content out of mainspace, but simply deleting discourages new editors from learning how to write articles and ultimately staying

15346:

to requiring "inline" sources. If there are sources but they aren't inline, that's a case for cleanup, not deletion. Deleting (rather than draftifying) an article only for that reason would be highly counterproductive. New editors often do not know our standards for how to present sources in articles

14976:

Irrelevant? You're the one who brought it up. I agree, it doesn't matter if sources were used or not used in the creation of an article, what's relevant is whether sources are listed in the article or not. If the sources aren't listed in the article, when second editor comes along in order to improve

14961:

Firstly that is irrelevant, secondly it's not true. If someone has written an article about a topic but not included sources then another editor can find sources to verify the claims in the article without needing to know the claims (or even the topic) exist. There is no guarantee of course that such

14292:

mostly because a grandfather clause is generally a bad idea at best: why does this only become a problem on April 1, 2024? As other have pointed out, unsourced articles rarely make it through NPP and AfC, and even if we let PROD be used for this, it can still be contested and forced to go through AFD

14268:

the reference material that verifies what you will write, and cite it. "Write first, hope someone sources it later, in practice let it sit unreferenced or CN tagged for the next ten years" should be an approach that is deprecated and ultimately not permitted (at least not in mainspace; if people want

12571:

is the second link in the sidebar. If that doesn't appear on mobile, that is a problem with the mobile view not the Main Page. We shouldn't add redundant links for desktop users just for the convenience of mobile users, especially for such an obscure and little-used page (the page views tool shows it

12114:

With V22 rolling out, the main page has become quite squashed, with very short line lenghts, tiny images, and a general bloated feel. It would be great to redesign this, probably by stealing ideas from other editing communities. The Spanish Knowledge design would definitely get support from me, but I

9159:

largely per Isaac. If we do every project, we're making a lot of noise for very little benefit. If we select only some, we functionally decide which projects we endorse (and which we don't by omission). Neither is a particularly good outcome. I also think banner campaigns are over-used in general and

8983:

I don't think this is a good idea, beyond directing users to all Wikimedia sites in general. If English Knowledge starts choosing individual sites to promote, and thus selecting ones not to promote, failure to promote a site will be seen as a negative endorsement. This may have an unduly discouraging

8691:

I'd appreciate it if you didn't put words into my mouth. I am under no illusion that Knowledge is 'the best' anything. It is however, the only online encyclopaedia that has any significant readership (thanks in no small part to Google), and is thus worthy of critical scrutiny. And I'm not judging the

8326:

Side note, it's unclear what is meant by "AI" in the proposal. Knowledge has long had AI (it's search bar). I'm assuming the proposal meant generative AI. Strongly against that. Such is a black box with a mind of it's own. We don't need it to be presenting its creations as policies/guidelines or

8258:

As I suggested earlier, I feel continuing to try to discuss your proposal without the involvement of those who are experienced and ready to start planning development isn't the best way forward. At present, this is discussion is about a hypothetical project where there are reasonable doubts about its

8158:

No, no!! Never in my any statement have I ever mentioned that AI will come and interpret policy and suggest editors the next move. Sorry if I am being too expressive as lot of editors here are having same misunderstanding. Ok, so simply as it goes, AI is for new editors, so user script better be kept

7643:

Specifically proposed for new editors (and in extension can be used by others). Some newbies don't even know Wikimedia projects other than Knowledge exist, being aware of third party tools seems improbable. I don't think hassle can be the reason to deny working on (discussion), which may result in at

7103:

I agree, this would be nice. Since training / fine-tuning AIs on volatile user-provided data is a hot area of research and product development at the moment, I don't think it would be a wise use of Knowledge funds to get involved just yet. It feels like we would just be duplicating work that is being

6008:

Ah, yes, anchors, good thinking. Their essential invisibility tends to make me fail to consider them more often than not. In this context, I recken that's probably fine, though, because keeping the thing itself hidden but then implicitly or explicitly drawing attention to it in a visible notice would

5820:

automates the generation of this string directly from the archive config, rendering the four Talk header params unnecessary. Imho, they should be deleted from the template in order to prevent misleading bot notices in the Template header box when the given params get out of sync with the config. More

5617:

Well, amn't i the fool! I can't remember how long we've had TFLs and how long i've been reading them, years and years though, and until this moment i have never realised that they are on a schedule; i have always just scrolled down the Main Page wondering if there'll be a TFL. Boy, do i feel stupid

5535:

Something that I've always been confused about is why TDL has a schedule where one list will appear three days after a list and then the next list will appear 4 days after instead of just three days again. I don't see how this could be for "we could run out of unique lists" purposes because there are

5080:

It would end up being a case if changing the local checkuser policy to allow elections were to be strongly supported by the community, and then there we to be an actual successful election - and how much of a counter argument the committee wanted to put up. This sort of change would certainly require

4788:

Have to be honest, but I think this is backwards in terms of how much trust is required. Admin actions can be and are scrutinized easily and regularly, that can't happen in the same way with CU/OS. OS is maybe straightforward in the way you describe though the information involved can be as sensitive

15895:

can be PRODed. Unsourced articles aren't the same as articles lacking inline citations. He then specifies (3) "Such a PROD can only be revoked after an addition of one inline, reliable, third-party source" – apparently raising the bar even higher, since articles lacking inline, reliable, third-party

15382:

per other guidelines and many other opposers, that sources existing and existing notability is enough to not use an enhanced PROD process. I am not a very big fan of draft space (I think working in articles space where there's many eyes is often better if it seems notable. And user space or off wiki

13899:

citations would result in the deletion of articles on topics whose existence and notability is not only verifiable, but is actually verified with sources actually cited in the article. The proposal is a solution in search of a problem, as there is no problem with unreferenced new articles that could

12936:

As I have already said, independent reliable sources are needed to established notability, so such sources should be made explicit by citing them in any new article. As Levivich says, an article with no sources is a blog. Hell, when I was blogging a few years ago, I always included a list of sources

12770:

Unfortunately (that relatively little used) feature was provided by a 3rd party and that 3rd party wasn't able to keep that integration up with the changes of Mediawiki and Wikimedia for 8 years, after which it broke and there is no easy way to restore or rebuild that without significant investment.

12115:

think we can make this into a more responsive design, so that we can use modern screen sizes more optimally. There are two columns, so the arguments for V22 giving up screen space are not transferable. Easiest is probably taking one small step at a time. Is it possible to widen the main page in V22?

10219:

One idea that might get approval is moving to a single column. The current layout was well designed when pages used the whole screen, but there are very few words on each line now that we have two thin columns shoehorned into a narrow stripe down the middle of the screen with an acre of empty white

9956:

I believe it's time to consider updating the design of the main page. I'm not certain when the current style was implemented, but it seems to date back to 2006 or even earlier. Nowadays, there are numerous modern and colorful box templates available that could give the page a more contemporary look.

9895:

Some years ago Search was changed so as to show matches in some other projects, my edits in those other projects have greatly increased since then. I don't know if search also does this on mobile as I never go to Knowledge on my smartphone. But it would be interesting if we could see stats as to how

9647:

to a lot of relevant articles where we have more images on commons than in the article. I'm not against adding relevant links to our sister projects at the end of articles, but at the top? I think the only time we should do that is where the Knowledge article is a disambiguation page for a word that

9567:

I agree that more should be done within wikipedia to "help readers discover some of the useful resources our sister projects have to offer", but I don't think that banners is a good way to do it. Some user research is needed to figure out productive ways to embed links to other projects in wikipedia

5160:

Another point that I don't think has been made is that, particularly for OS applicants, one thing that is often significant is what the requests they've submitted have been like as this can give an indication of what their judgement will be like in assessing requests from other people. For example I

4002:

If you look at what proposals come out of the committee to increase their productivity, they are almost always about streamlining case management and especially dealing with appeals. That none of them relate to CU or OS appointments should give you a clue that it's really not a significant aspect of

3850:

page when it's such an uncommon occurrence seems to have very little if any benefit. The lock icons not being used often is a non-problem. And most users that want to move a move-protected page are going to click on the button regardless of the color. If you just want to gray it out for yourself, it

3481:

I'm fine with a lock icon for creation-protected pages if technically feasible (obviously not by placing a lock template on the non-existent page). I'm also fine with an upload protection icon although almost all files are at Wikimedia Commons and we don't need to invest a lot of time into improving

2722:

readers don't care about them (which I haven't seen any good evidence of, just assumptions), they're helpful from an editor's point of view. If I see a blatantly terrible article with a little green icon on the top right, that catches my attention and makes me want to either send it to GAR or fix it

2003:

is fine (which is what FA / GA icons do), the problem with VA marking was - too many to list, see previous discussions. This proposal comes across as an extremely weak "well this one proposal was voted down, let's do the same thing in a separate vaguely similar area out of some misguided fairness."

1402:

I think all quality ratings should be shown, not just GA/FA. Analogous to maintenance tags, it's useful for readers to know how much they should trust what they're reading on Knowledge. Moreover, topicons are unobtrusive and might have the benefit of converting a few readers to editors. As an aside,

15070:

What in the world am I doing that would be considered "imposing my opinion on the encyclopedia"? Comments like this is why I get frustrated discussing things with you. Nobody here is imposing anything on anyone, and this IS an RFC about modifying policy. This IS the right place to express the views

14992:

And why do some editors spend so much energy defending unsourced articles? FFS, this is Knowledge, it's coming up on almost 25 years now. The first step in writing an article is to gather not one source, or two, but three high-quality, GNG-compliant sources. If you don't have that, you don't have a

14837:

clearly state that unsourced articles have to be draftified, not PROD-ed. PROD only applies to BLP articles that are unsourced, not other articles. If this is implemented, the procedure of NPP will have to be changed as well which requires further discussion before it can be implemented. We have to

14099:

We need to build this idea that finding and including sources is step #1 of creating an article. And the burden shouldn't fall on overloaded volunteers to "prove a negative" when people don't bother with sources. But are there enough of these to be setting this up? I've done thousands of NPP's and

13904:

that there should no further proposals for the creation of an "unreferenced PROD" for the next five years. I do not believe that there is any chance of consensus for the creation of a new PROD in the near future, and the community does not have time to !vote on what is essentially the same proposal

13172:

The fundamental problem for newcomers is that editing Knowledge is not convenient. It takes a substantial effort to do so. So, one way to make this easier is to improve on the article wizard and ask people to find a few sources before citing them. I imagine that the new article wizard would ask you

12423:

page in mobile view. Instead, I would have to search for in the search bar and go into desktop view to access it. I am proposing that we include this page somewhere on the main page. No opinion on where to put it on the main page, but my first thought would be to put it the other areas of Knowledge

11587:

What's needed is a brainstorming RfC process like is currently being done for RfA. This gives an opportunity for a variety of ideas to be suggested and we can then see what sticks. For example, I'm most interested in structural change -- amending the section order and content so that the featured

11234:

Why? No clear argument for why this would be desirable. Can we be sure it would not lower general accessibility depending on the device being used to access the site? I am not adamantly opposed to sprucing things up a bit. But the oft quoted adage "if it aint broke, don't fix it," definitely comes

10557:

Fucking excuse me? How the heck are the articles that we've vetted low quality? ITN gives us some global perspective and is one way readers could keep themselves up to date in current events. DYK makes everything a bit more fun for everyone and trivia is fun. Both highlight our utility as an online

10457:

I agree that the Slovakian one is too dated. I don't see any problems with the Spanish version under vanilla V22. Maybe swap POTD with On this day for slightly longer line lengths for the latter, make Other projects full-width (and maybe unbox it), and of course adapt ITN to our format, but I don't

9199:

the entity being advertised. We already link to the sister projects in the main page (left panel and bottom) and we routinely mention relevant sister projects explicitly from articles where applicable (many articles include templates linking to specific pages on Commons, Wikivoyage, Wiktionary) and

8852:

for the same reason I don't like banner ads for donations or dishwashing soap. You diminish the encyclopedia by pasting ads for things that are completely unrelated to the topic they are searching. First and foremost, the READERS matter, and this diminishes the encyclopedia by putting information

8300:

I apologize for being unclear: personally, I do not recommend that you expand this discussion further. Instead, I think you should let those who are interested in planning development to drive the conversation, so that they can use their expertise and constraints to help direct further discussions.

8205:

You gave a very specific example of an AI hearing a generic problem and deciding what guideline it thought applied. That is interpreting policy. Your example also specifically includes a suggestion of the editor's next move: "For more information, you may ask the editor who reverted your edit for a

7805:

are active examples. AI has wide applications, and I think that there are lot of applications to it on WP too, to some I oppose just like you, but others, like AI optimized simple search, or beta tested policy chatbots are some I have eyes on, as they rather than replacing existing resources, would

7231:

Sean.hoyland gave one pointer above to a Wikimedia Foundation team that might be able to give some pointers or advice. You can also look at the foundation's pages to find other potential contacts. Think of everyone you know and if they have any related experience or connections to those who do, and

7028:

If we're going to use that analogy, your suggestion is like switching from candles to incandescent bulbs, without having addressed why the lighting difference matters, what type of lamp we're going to use, whether the power grid can support it, and who's going to pay for the electric bill. I stand

6798:

ChatGPT requires email for registration. But it also has that 2021 bias, and couldn't list any policy/ guideline's specific article. ChatGPT is more like proposal 2, but that makes it erroneous. It's also third party, so it may not be fed with new ones and also combine non-wikipedia related data it

6117:

There is a counterargument to be made, using the same scenario: If, after jumping to the top of the article, for whatever reason, one decides one wants to go back to the redirect's target location, again for whatever reason, the proposed change adds a new way to do that - Ctrl+F "redirected". Works

5024:

It could certainly create one of those community consensus crisis sort of situations. Regardless, I certainly would want a wholesale cleanup of everything related, including the arbpol, if this were to gain traction. Realistically, I think this proposal should have gone through more development and

4618:

Arbcom managing this seems to be working well, and producing both enough permission holders to get the work done but not spreading access too widely. Arbcom is exceedingly unlikely to appoint any person the community expresses a decent amount of reservation in, and is aware of things the community

3914:

I propose that we implement a community-based process for appointing CheckUsers and Oversighters similar to how RFAs are run. Because these positions require a high degree of community trust, I believe consensus should be higher like in appointing bureaucrats and interface administrators. I believe

3466:

because participants saw it as unnecessary bloat. Personally I always found them a little useful, even though us unregistered types haven't been able to move pages for a long time since it was still a quick indicator that a title was potentially contentious, but I can understand why it was removed.

1348:

and still in use today needs revision to be practical. GA ranks are varying quality; FA ranks are easier to attain for less popular topics than highly read ones where they are most needed. Removing icons could be the start of recognizing that our quality reporting system is insufficient for meeting

1228:

I'm neutral about removing FA topicons. At least that process can be said to involve multiple editors reviewing it, and the FA folks take their FA reviews pretty seriously AFAIK (unlike GAs, which some take seriously but others don't). I don't think it matters much if the FA topicon is there or not

16537:

I wasn't aware of the RfC on unsourced articles, but the consensus at that RfC is incredibly disappointing. Simply put, we should not have unsourced articles. We are supposed to be an encyclopaedia, and having unsourced articles makes us little more than a social media site or a free webhost. Just

15797:

How/when would that happen? Due to the backlog due to a handful of active NPP'ers being asked to do too much of the million editors' work and otherwise being too difficult and painful, it won't even get looked at at NPP until after the draftifying time limit runs out. Not that I think that there

15634:

You might ask, why do I demand an inline citation in my proposal? This is because an inline citation helps me to verify a specific statement in the article. If that reference is placed below, a reader would have no idea what is the specific statement that we are referring to. Just to mention this,

15630:

that new articles should have an inline citation. Why is it so difficult to codify that into policy? If you translate an article from a foreign language to English, then it is very trivial for them to copy the citations to the English version of the article or find one source that verify that this

13956:

Establishment of notability (in the Knowledge sense) requires the existance of reliable, independent sources, so if we know what those sources are, why not cite them in the article? In other words, if no reliable, independent sources are cited in the article, where is the proof that the subject is

10743:

That does look much better! Combining the search box with the first box seems alright, though maybe I'd use the tagline "Search free knowledge". I'd also recommend making the globe logo stick to the bottom of its box, replacing the whitespace between the first two boxes with a horizontal rule, and

9266:

to all forms of distracting and inappropriate advertising. There will always be sound and moral reasons behind any banner campaign - I'd like us to allow various charitable banners which fund-raise to help starving children, or prevent climate change, etc. But Knowledge was set up not to be such a

8569:

Sister Wikimedia projects have a lot to offer readers, and as one of the most viewed sites on the internet (globally!) we should help introduce readers to these resources. As you know, other Wikimedia projects include a dictionary/thesaurus which includes translations; a travel guide; a library of

8350:

Firstly, AI here does not directly indicate to generative AI, rather it may be complementary. Secondly, the first and foremost thing I propose(d) is AI in search bar, not just as significant as it looks on paper, but practically useful. It is the present 'tune' I say, that as anyone says AI, first

8058:

something similar with their internal content. We may do the same with Knowledge's internal pages and (maybe conversations/ discussions also) for editors (especially new) as I already proposed. AI already serves Knowledge's content for readers, with this editors may benefit from something similar

7514:

I have put some thought into this and I do think there could be merit for a tool to help new editors see their revision draft is violating a WP policy prior to submission. That isn't necessarily a "WP Policy chatbot" but probably something more akin to a spellchecker but for WP policy. If it could

6992:

1) The stance is like 'We have candles, why need bulb?" The very reason I proposed this is that it is really difficult for a non-experienced user (even a mid-experienced user like me having 3000+ edits) to find relevant policy, and sometimes not sure if there exists any for it even with the legacy

5815:

template has no control over Talk page archiving, but it does have four params used to generate a "bot notice" in the header box which says something like: "Archiving: 90 days" (plus a tooltip with more info). It's just a string displayed in the header, which may or may not reflect what the actual

5681:

I wish to seek to change the wikipedia policy WP:SOURCE. Currently this states "Any material lacking an inline citation to a reliable source that directly supports the material may be removed and should not be restored without an inline citation to a reliable source." in the section Responsibility

5406:

Welcome to the sandbox! Anybody can edit this page and it is automatically cleared regularly (anything you write will not remain indefinitely). You can either ] the source code ("''''''" tab above) or use ] ("''''''" tab above). Click the "'''Publish changes'''" button when finished. You can click

5065:

In essence I agree with Xaosflux that the arbitration policy should be amended to align with community desires on appointing checkusers and oversighters. Thus I raised this need in my initial comment. In theory the arbitration committee could just agree to rubberstamp community appointments, but I

5036:

I'm not sure if there would be a crisis. The stewards would have to ignore the English Knowledge arbitration policy and accept community-requested assignments, and it doesn't seem to me that's likely to happen. I agree that it would have been fruitful to have more discussion to develop a rationale

4502:

Previous attempts to try this have descended into a popularity contest and produced no useful feedback about the candidates' suitability for these specific roles. These roles do not necessarily need popular editors who avoid sticking their heads above the parapet for fear of not being elected, but

2808:

addition of sitewide ratings for all mainspace articles and for all devices. The purpose of such a system is to indicate an article's quality to readers in advance, so they could anticipate whether an article is worth reading or not - or at least have a glimpse of what to expect. It's no different

2465:

The topicons are an easy-to-find, yet also unobtrusive, way to indicate that an article has received (and passed) some level of vetting. The fact that they're visible in mainspace also means that they're relatively easy for readers to find and interpret as well, a benefit that would be lost (or at

16715:

per Pppery and Sdkb's criticisms of a grandfather clause. These articles with the unreferenced tag are avoiding draftification in the NPP process by satisfying the relevant notability criteria, so it stands to reason that reliable sources can be found for other editors to create inline citations,

16607:

The issue is that when you speedily delete an article rather than sourcing it, tagging it, prodding it, moving it to draftspace, etc. you don't give any opportunity for the editor to improve it by adding sources, you reduce the opportunity to educate them about the desired standards for Knowledge

16479:

why is there a need for an inline source? It's not a strict requirement for articles in general, nor is necessary to establish notability, and it only creates an additional burden on the editor. I see many "oppose" here because of this requirement, and I was hoping you could explain the rationale

15597:

grounds. These are often translations of other language's wikipedia articles where there very well may be sources, but not ones easily consulted in English. Now, yes, other language Knowledge editions have weaker notability requirements than enwiki, and yes, some of these are just authentically

14918:

A draft with no sources is useless, it doesn't help anyone start an article, because you need at least one source in order to summarize sources. Also, while a petty concern, the editor who started an unsourced draft shouldn't get article creation credit. The first step of writing an article is to

13813:

For what it’s worth, on March 1 and 2 I attempted to patrol the category in question and see if it was possible for one person to keep it down to zero. It sort of is but you run into articles that are unlikely to get deleted but next to impossible to source, and then things get out of hand if you

13789:

So as long as the proposer reasonably believes that the deletion will be uncontroversial there doesn't appear to be any particular procedural barrier to prodding these articles now. Third, it would be helpful to see evidence that there is a problem that needs solving. A PetScan query for articles

12483:

I think there's a larger underlying problem here of it being difficult to access sidebar links in mobile generally. Some links are justifiably not present there, as they relate to tasks better done on a desktop, but the contents page seems useful to everyone. I'd like to see an effort to push the

12218:

Nobody ever seems to mention this, so I think it might just not be generally known -- seriously, go look if you don't believe me -- there is a different Main Page on the app with completely different sections that are not put together by the editing community. It has totally different stuff, some

10695:

Something feels wrong about the concept of giving that much prominence to our sister projects. Knowledge readers are hardly gonna go there and this introduces a lot of colorful icons that clutter up your attention. Previously it'd only take attention when you scroll/look down and want to dedicate

9880:

Whilst I'm not necessarily opposed to focusing on other Wikimedia projects, personally I prefer a venue that is open for any initiative looking for more participation. There are a lot of areas on English Knowledge that could either be helpful to more readers, if they knew about them, or could use

7536:

Maybe we could add that in 'beta features'. Proposal 2 or related would surely be more complex idea than 1, but it will increase the overall efficiency and quality of articles produced/ or edits. However 'AI' is the "popular" term so I used it, but the real solution would be to somehow reduce the

7456:

I still think that this is a solution in search of a problem, and that it will in fact find several new problems if implemented. The same concerns of lack of reliability apply: how do we know that the chatbot will direct people to the correct p&g page given a query at runtime? What dataset of

6993:

tools. To be clear, there are numerous similar policy pages, and it is difficult to find out where the exact guideline is located and it takes lot of time if not hours finding one. Also even, forums avoid/ aren't sure of such questions as they themselves find it hard to locate sometime (not sure).

6871:

I assumed that this shall help, or else I wouldn't have proposed it. We may not need it, but it's worth considering the potential advantages such an AI could bring, taking into account factors such as efficiency, accuracy, accessibility, and standardization all throughout Knowledge it could bring

5995:

I think this is a good question. The notice linking to the redirect is useful as a way to get to the redirect page. and it informs the reader why they are at a place they would not expect to be, but it should be displayed where it can be seen, and preferably where it is most relevant, which would

4773:

RfA, while it might be the only method we've come up with for selecting administrators, is not such a glorious success that we ought to be using it as a model for anything else, at least not when the current system is perfectly functional. And there are substantial differences between CU / OS and

4292:

The current process is that applications, which are accepted at any time, are scrutinised by ARBCOM, those that cannot be dismissed out of hand are passed to the functionaries to for scrutiny. Based on the feedback from functionaries and arbitrators' own discussions, applications that are clearly

3622:

That might be alright but then thing is, if a page is (edit) protected, then the lock is still there, even if they are able to edit it themselves. So that is why I proposed a plan to introduce the green move icon lock in the top right to be shown to all. Again, like with edit locks, the move lock

3285:

topicons. Anecdotally, some years ago, in a former life, I was observing a high school social studies class in the American Midwest and the teacher, as part of a lesson on research methods, told the class that the lock icon in the upper corner meant that it was vetted. We should be discussing how

3082:

I've nothing against the topicons; I'd be happy to keep them just because they're useful to experienced editors, a nice reward for hard work, and traditional for the site. However, I don't think we should be making this decision on the grounds that we speculate that they're useful to non-editing

1440:

By that logic, I think one could say: Why should a reader care about that one editor came to the conclusion that an article is GA quality and 5-10ish editors—none of whom are necessarily subject matter experts—came to the conclusion that an article is FA quality? I'll make a note to return to the

1425:

Why should a reader care about how vital wiki editors think an article is? The reader has their own ideas about how vital that article is to them in that moment and in general. The idea of showing all rating indicators is an interesting one in terms of helping readers clue into what they are. I'd

16190:

mean on the form?", "How do I know if I'm choosing the right template?", "What is a template anyway?" People don't automatically know this stuff, and they're already having to read up on how everything else works. I think a friendly offer of help with inserting them would be more useful. "Find a

15504:

Draftification is done at the discretion of editors and new page patrollers, who like most Knowledge editors are generally expected to have coherent reason for things that they do. Making it a binding policy would remove this discretion. Either way, it is not good: if it's meant to substantively

15325:

PROD would force that article to be improved or to be moved to draftspace. And frankly, why is it so difficult to cite one source in an article about a notable topic? I don't get the opposers' reasoning here. If there is no source in an article, how could we know that this article is not a total

14419:

In practice, this is needed because it is very difficult to verify which statements belong to which source (with an exception of very short stubs). I would say that all sentences in such an article are likely to be challenged because they establish that "this topic exists in real life" and "this

13477:

The categories track the tagging date. And iirc some automated tools update tag dates when making unrelated edits. So, it's likely the new categories are just populating with old articles. Autopatrolled editors wouldn't remain autopatrolled if they created unsourced articles nowadays. I actually

13346:

Moving to Draft with a x day prod-like notice. I'm impressed by the relentless work done by a small number of overwhelmed volunteers trying to address this problem, and I support their efforts. I disagreed with the former proposal, but this one is acceptable, grandfather the existing article and

11449:

I want to avoid doing more with this, since I think that's where previous proposals have failed - people want some of the changes but not others so it ends up a whole mess. And since at least some people seem to like this, I might try to put it up for RfC soon, but want to make sure there aren't

10597:

might be getting at is that simply undergoing a vetting process is insufficient; it is the quality of the vetting that is important, and so by extension, the quality of those doing the vetting. If, for example, ITN and DYN undergo a lightweight review which is perhaps keener on filling slots and

10248:

From a purely selfish viewpoint (which is allowed, as this is an aesthetic matter), I want the main page to remain unchanged. I use Vector 2010, and everything looks just fine to me. However, the vast majority of readers don't have the luxury of setting that preference, and are stuck with wide

10157:

You could probably get consensus for the general idea that the MP should be changed, but the consensus would break down over what specifically to change it to, as everyone has their own idea about that. As noted, this is a constantly made proposal. The best chance of success would probably be to

8008:

User scripts are add-ons that are commonly used by experienced users, while new editors are often unaware of them. Even mid-level editors may naturally find them uncomfortable to use. While volunteer-created scripts or tools are valuable, official integration ensures widespread accessibility and

7480:

Your have very valid points of concern. I think that this could be implemented step by step, whenever the community and foundation feel ready. There are lot of proposals in this discussion itself: First two are the proposals I presented, then a new one is too just let the already existing search

7428:

No, I never deemed the proposal to be the way you are thinking, I myself may have opposed such proposal. I proposed for an AI optimized search for Knowledge's policy and guidelines. There are lot of chatbots out there (including ones made on small budget) which could answer question based on web

6820:

Also, new data like various discussions going on ANI, Village pump, IANB etc. could be fed to AI so as to decrease discussions on already discussed topic, a user is not aware/ couldn't find about. Currently only search method is using Knowledge search within specific category which could produce

6237:

The aim of this proposal is to add some technical clarification to the template but also simplify its use. I find the current template clumsy looking and over-engineered in its 3-in-1 approach (i.e. certain templates can be transcluded if certain parameters are used. Why not just transclude them

5864:

The GeoHack system offers a choice of about two dozen map applications. I always choose the same application, but I have to make that choice every time I use GeoHack. It would be nice if my choice could be saved in a preference setting so that whenever I click on a coordinate I go directly to my

5688:

I believe that requiring sources on every page brings a number of problems: 1) it is onerous and inefficient and discourages linking relevant articles to pages, especially for new editors: 2) the relevant article may include more sources, mentions of the article might only include one, so anyone

4592:

Not only is the current system not broken, the community as a whole is very poorly placed to judge an admin's skill at handling those areas that are mostly directly relevant to CUOS skills; revision deletion, SPI activity, deletion-related activity at RC patrol, VTRS activity, and more technical

4107:

But I'm in favor of trying a community based system because it might be better than the current system. In general, I think community-based systems work better than representative-based systems on Knowledge. ANI works better than AE. RFA for all its problems yields better results on average than

3331:

Currently, articles that are protected from editing have the relevant lock in the top right hand corner. Userpages may not have the relevant edit lock on the top right. With regarding edit protection, it will stay as it is but why don't we do the same for move, create and upload so that articles

2784:

be hijacked by sockpuppeting, they present no evidence that this is widespread to convincingly argue the review process' collaborative editing is a net waste. While the GA and FA criteria are no different from the normal aims of article editing, the pursuit of specific quality metrics encourages

1307:

As others have ably noted, these icons provide no discernible benefit to the reader. We would therefore better serve the reader by eliminating this visual noise. Frankly, we would probably do well to reconsider the centering of FA and GA work in general. The hundreds of hours it takes to nurse a

13798:

with listed references (so tagged incorrectly but still subject to this proposal) and one that is currently up for speedy deletion. The remaining six definitely have some issues, but I'm not sure we need this level of policy change to fix what seems to be a couple-of-articles-per-month problem.

11794:

Most editors are not allowed to edit the main page and so we have a frozen format which is alienating too. What's needed is a process for making changes which allows the main page to evolve. Perhaps there could be an annual update. During the year, there would be a beta test version to trial

7063:

Also I believe that just because I did not elaborate every problem we may face with this, doesn't make it useless, and it could further be developed. However, I assume your point to make sense, and would like to further detail my proposals. And yes, we are one of the most visited website, so we

4744:

The process often requires considerable privacy. ArbCom may have to learn, or deal with, very personal information of potential CUOS. An RfA like process is incapable of handling such sensitive issues. We already ask for community feedback for candidates. Why would we want more RfA, when RfA is

3264:

If we are going to remove an icon, I would hide the protection icons from logged out users (I don't see the benefit - they are anyways shown a "view source" button instead of "edit source" if they can't edit the article). That truly is inside baseball unless you actively try to edit an article.

14297:

makes it likely to be kept regardless of the number of citations in the article. The main practical effect I see is the grandfather clause, and it only makes sense if the goal of slowly removing the protection is bought into. The problem is that consensus from about a month ago doesn't show an

4406:

I'm uneasy about making people go through the fiery pits of hell thrice total if they want CUOS. That aside, my bigger concern is the wider community's ability to judge whether someone is trustworthy enough to handle private information. CUOS isn't admin or crat, they're specific roles used to

15602:

notable people who don't have articles yet, and deleting the articles out of misguided "consistency" doesn't make sense. Better to keep the article simple and non-controversial and wait on someone familiar with the foreign language sources, instead. Keeping this situation a valid option for

11779:

The most popular editing view for Knowledge is desktop, not mobile. Alienating the site’s most essential users is a serious risk to take. If there had been no way to revert the Vector 2022 and recent line height changes, I would have been strongly alienated from Wiki and contributed way less.

6938:

for several reasons. 1) We already have a search capability, as well as numerous other features like lists, disambiguation pages and redirect shortcuts to help guide users to find what they're looking for. 2) The risks of AI misinterpreting, hallucinating, or otherwise giving users inaccurate

16735:: This proposal risks discouraging new contributors by setting overly stringent barriers to entry. Instead of penalising new, unsourced articles, we should guide and educate new editors on the importance of sourcing. Let's not make participation in Knowledge more daunting than it already is. 14598:, what information actually needs to be in it? Or should it verify some of the article content? If so, how much? If not, where are you supposed to put the citation, if it doesn't actually relate to any actual article text? What about articles that are obviously not hoaxes (i.e. most of them)? 8227:

How can any AI/ algorithm work without interpreting its query and relate it to the search results. What you mean by 'interpretation' is the vary backbone of any AI that I could propose. Otherwise what is difference between current search bar which just cares to match the words you enter? The

7319:

Proposal 1 is AI- optimized search. Proposal 2 may better suit your oppose. Also the proposal 1 is not asking for issues as it is already evident from my and others(response above yours) experience that new wikipedians do face lot of issues when it comes to make even the slightest edits, and

6451:

Each category page should go into further detail about its purpose. On core and extension tracking category pages the information would most likely be included on the pages themselves but in the case of template/module tracking categories the information would most likely be transcluded from

4633:

RfA is a broken process as it is, leading in part to a shortage of admins. Doing the same for CUOS without improving the RfA process is asking for trouble, not in the least because CUOS deals with way more sensitive data and the community as a whole is ill equipped to assess trustworthiness.

14438:

unsourced statements as likely to be challenged because the person doing the challenging must assess whether the claim is a "sky is blue" type claim (and thus doesn't need a source), plausibly true (in which case they should attempt to find a source, and add it if found or tag it if not) or

9344:

as conceptually wrong. Very, very, VERY few people, even linguaphiles, are hardcore "let me sit down and read the dictionary some" types. Similarly, even people who love reading books aren't going to be enticed by "oh hey, let me go to the local library and pick some books at random" from

16294:

This is basically de facto practice anyway. An article without sources is going to get draftified pronto. Then, either the author will clean it up, or it gets deleted in six months. I don't know that we need a dedicated kind of PROD, I think regular PROD could work if it isn't suitable for

15000:

If an editor happened to use sources in drafting the unsourced article and just forgot to put them in there, that's easily fixed. When the article is prodded, tagged for CSD, or AFD'd, the editor can add the sources. Heck, even after it's deleted, the editor can get it REFUNDed and add the

8477:

Please note that this is a discussion regarding whether Knowledge should do this in the broad sense; detailed arguments like "I don't like the suggested banners" or "but we shouldn't promote " should be saved for later discussion. Please respond to the idea in the RfC statement. Thank-you,

7166:

This proposal is in essence about creating a better search tool. It could be a good project for a university computer science department. If you know anyone with appropriate connections, or any developers with available resources who might be interested, perhaps you can suggest it to them.

2863:

I don't feel strong enough to sit on either side of this discussion. There are genuine grounds for concern about quality control and the ad hoc nature of the review process. It does seem that the topicons are very much inside baseball, so to speak. Nevertheless, despite the extreme example

15857:"All of the objections raised to the two similar proposals in the last few months still apply. I really can't be bothered to repeat them again. We're supposed to be building an encyclopedia here. Clearing a backlog is only incidental to this. Phil Bridger (talk) 08:52, 17 March 2024 (UTC)" 12721:

Thanks for the reply and the advice. But it used to be that you could aggregate topics/chapters and download all of them in PDF in the format of a PDF book, just by clicking download. It would download it and create the index, with page numbers and chapters. All of it in one unified PDF

7084:

An AI that can instantly forward you to the correct venue for a dispute, find help pages for any question, or search the archives of boards, and other capabilities mentioned above would be a massive benefit to editor QoL. A proof of concept might be possible by creating a custom GPT on

1295:

Our loyalty is to the WP:READER, to whom these icons—where they're even noticed, that is (on mobile view they don't exist!)—are merely inside baseball. Their real purpose to allow editors a chance to gussify their user, talk pages etc. Stick em a table on a subpage if you have to :)

15167:

appears to be a list of current prods. Got any reports that list the outcomes of prods, such as deprod vs delete? My point is that many prods are de-prodded, which then requires the patroller to follow up with an AFD. This is my anecdotal experience with using prod during NPP. Thanks.

12937:

at the bottom of a post, so that readers could read more about the subject, if they wanted to. Not providing sources when an article is created is just being rude to other editors who are somehow expected to do the work sourcing the article that the creator could not be bothered with.

4099:

I never said I think an rfa-like process would produce a better result. The proposal is for a community based process, not an rfa-like one. I think rfa is terrible, hence my voting in favor of various reforms, consistently for years. I said I thought rfa produced fewer false positives

11748:

The most popular view or Knowledge is the mobile one and that's already one column. We should therefore be designing and optimising for that as the primary interface. The balance issue is absurd in that it often causes ITN admins to remove news entries -- form rather than function.

14164:. Beside it is very frustrating to see in AfDs the comments "there are sources available" and "normal editing can solve this" after which absolutely nothing happens. Not even by the editors that state that there are sources available (when you know that, why do you not add them???). 14016:

Based on the number of comments in opposition here, I don't think it's obvious at all. However, if it is obvious then you will have no problem gaining consensus for the amendments. Just don't claim the consensus exists until you can provide a citation proving it does actually exist.

8692:

other projects by 'Wikipedian' standards, I'm judging them by the standards of someone who considers Knowledge structurally flawed, even if its objectives are worthy in the abstract. The projects I refer to are in my opinion worse, in several ways, but mostly of little significance.

6970:

Beyond a shadow of a doubt, people have trouble searching for policies and boards. Half the time I use Google Search. Overall, there seems to be unharnessed potential, which could make newbies stick around or make learning and navigating between policies, past and present, easier. -

6965:

Is this referring to an AI generating freeform text, or referring a user to a certain page? I think that this program should probably be restricted to answering in links. I don't see a reason to believe the AI would consistently get that much wrong, or get it wrong more than a user.

16095:

Yes, that's precisely what I meant. I don't condone the German Knowledge's approach; it would be better if they used more inline citations, but unfortunately they often don't, and yet nevertheless they have useful, well-written articles on subjects we don't cover, and the articles

8280:

I have indeed invited dozens of editors shown to have expertise/ special interest in AI or its branches since you last suggested. We even have the attention of director of machine learning at WMF. I will try to expand this over now, streamlining its focus as you said. Thank you.

7001:

3) That's the other primary reason I am proposing this. Assuming that a constantly evolving policy may make the "AI" old, seems like saying editors have super power of going though entire policy updates within less time AI is updated. AI may help editors know what updates are in

2864:

highlighted above (for which I have direct experience) I still think indicating some kind of independent review has taken place is worthwhile, no matter how limited in nature...I might have an easier time supporting this proposal if an alternative was being proposed. Regards, --

1770:

of the GA and FA processes back when I was an unregistered lurker. I can imagine other unregistered users falling down a rabbit hole and potentially getting involved with curating content after seeing one of them. Also Barkeep49 presents another good rationale for keeping them.

11043:

I think people will support a new main page design as soon as they're shown a new main page design that they like. I would encourage people who are so inclined to create and share mockups of new main page designs. Eventually someone will make something that enough people like.

15040:

I wasn't the one to bring this up, you brought it up in response to Zero. I'm not saying that it's not better to have sources in a draft (obviously it is), so most of your arguments are refuting something I'm not claiming. My only point here was that an unsourced draft can be

11368:

I like a lot of that, but (at least on Desktop on MonoBook) the colored section headers only seem to have a thin border on the bottom and nothing on the other 3 sides. Much better than the es.wiki one linked above, which is visually clean but seems to lack any sort of flavor.

16273:—the majority of outcomes are either unsourced articles, or backfilled articles that will likely need to be largely rewritten anyway because the original author wasn't actually working from any sources. Better to do the rewriting first before the article actually goes live. 15505:

change the way that article creation works on Knowledge, it is for the worse, and if it isn't meant to substantively change anything, that is a great reason to avoid making giant disruptions in the public-facing process of a thing that has worked fine for the last 23 years.

14811:

So our current system works, no? None of the articles have been prodded, but those which were problematic have been moved to Draft space as appropriate. Some of those may be moved back into main space after they have been worked on. Better to go to Draft than to be deleted.

14349:. Is there something preventing us moving such articles to draft space that necessitates creating a new hammer? Also, requiring inline citations is a step too far and seems contrary to policy. I hope this isn't really what was intended and the proposal is just badly worded. 13777:. Three concerns: First, the proposal and discussion seem to be conflating unsourced articles with articles lacking inline citations. Requiring all articles to have inline citations regardless of whether any content has been or is likely to be challenged is out of step with 4128:

if this is your opinion, then I think I understand why we feel so differently about this issue. From my perspective, AE is a very significantly better board (for the situations where they overlap) than is ANI, which vies with RFA for the title of worst venue on the project.

15383:

is better at the stage where notability is unclear or if someone wants to draft an article solo.). However, I could see supporting a properly crafted proposal that articles less than 90 days old and entirely unsourced when draftified with care (so maybe a very weak form of

14132:

The greater fool theory, that someone else will do it, is not accurate: see the backlog. See the stats that show how few experienced regular editors. If we can't support our maintainers who are in the trenches doing this work, then you are right, they should not do it. --

2848:

If it helps 1% of the readers, that should be a good enough reason to retain them, unless they actually harm the site or reduce it's quality, at least for some readers. If it's 100% neutral for 99% of the readers, this should not weaken the fact that it helps the other 1%.

3630:

The reason I propose those changes is because they are shown in the relevant links but the non edit protection locks are barely used. It also tells users that the page is protected (the move option should be visible to admins if protected) With regarding salted pages (see

13235:

tools and wording changes for newcomers are instated. The risk of this detracting newcomers is high enough that I think the community should have a consensus that the new wording/templates etc. alleviate harm, and they should be ready to go before any changes are made. ―

16212:, many unsourced articles have sourcing available and it just takes an editor to spend 10 mins to find them. The problem is that this rule would be deleting potentially encyclopedic subjects written by editors that lack the expertise to add sources. Also, not all pages 4648:

In discussions about the low number of applicants for functionary permissions in the 2023 round it was noted that the shortage of new administrators was likely an influential factor. RFA's brokenness is the most significant reason for the shortage of new administrators.

3243:

Many editors have stated that finding out about the GA/FA process has made them motivated to contribute to Knowledge. I think the article ratings should be made more explicit to the reader, but if we can't bother to improve the top-icons, then we should get rid of it.

13731:

The NPP process does not allow unsourced articles to be draftified. It maybe happens often, but it's not part of the policy. An NPP reviewer is required to look for sources, and if they find sufficient ones to support notability, they are supposed to tag the page as

13616:

per the most recent discussion and disallow any new discussions on this for a few months. I've already seen several notable topics without sources or with only one source being moved to draftspace and then deleted after six months. A PROD would make this even worse.

12259:

So is someone going to let the people at DYK and FLC know that they were deemed unimportant and removed from the main page, then? If this is just being done unilaterally by the fiat of app developers, does it really matter what the community decides should be there?

13523:

In general, I expect almost all the articles in those categories are old articles. Notability would have to be completely obvious and there would need to be no BLP element in the article for a reviewer to pass an unsourced article in 2024, if someone were to start

6229:

I believe that the third group is the largest of the three. Such tracking categories are used to track missing or invalid parameters in template transclusions amongst other things. They are not integrated into the MediaWiki instance and are therefore not listed on

11062:: a minimalist interface, with a prominent search bar, to which I'd add prominent display of TFA and FP. Like maybe FP centered at the top, search bar below that, and TFA below that. But I'm no web designer though so I'm not sure exactly how that should all look. 9326:

project. I'm generally opposed to taking up screen real estate to try and recruit readers, let the readers read in peace without distraction. BTW, the way to let readers know about sister projects they may find useful is "inline," the way we do it now, with those

7489:), then we have something like Knowledge Library type, difference that it provides access to some third party model like ChatGPT fed with p&g which could be used in a while, etc. I have even posted an example of basic AI which we could think of in first step. 4051:

That is an inappropriate question and you're old enough to know that Thryd. No, I'm not going to name the CUs/OSs who I think shouldn't have been appointed over the years. But for starters, the ones who were suspended or removed from those positions would be good

16220:

lacking sources do little harm to Knowledge. What about disambiguation pages, obviously those shouldn't have sources. At what point does a list page become a disambiguation page, and vice versa? And pages with valid general references should never be deleted. —

9361:

having its first EL being to Wikisource? Great. The reader is interested in the topic, here's a link to where you can read the story on Wikisource. Or disambig pages including a prominent Wiktionary link, especially for less-used words or foreign terms like

15421:

I know this is not going to succeed, but ultimately only this approach is consistent with the widespread (and beneficial) practice of improving verifiability by removing unsourced content from articles—which can only be restored if there is a reliable source.

9321:

due to banner blindness, and due to the fact that just because a project is a sister project doesn't mean it's a good project. I just don't see the value in telling readers about other projects. BTW, I also would oppose a banner that encouraged people to edit

15638:

And why do I ask that citation must come from reliable source? Because we don't want people to write an article about a notable topic with absolutely shitty sources, like Twitter posts, gossip websites, forums, etc. If the whole article is talking about the

5689:

looking for useful information might not see it; 3) in a rapidly moving field sources may be updated in an article but that might be missed on linked pages. In any case it is easy for anyone to click on the link to see the article with all relevant sources.

4147:: I'm sorry for personalizing this discussion with my earlier comments. I thought by personalizing this discussion I could make a point about not naming names and personalizing the discussion, but that was a dumb idea on my part in retrospect, my apologies. 9327:

boxes/links/hatnotes/whatever-you-call-them that say, e.g. "full text is available at WikiSource," or "there is an entry at WikiData," or "there is media on Commons," etc. It's better to put those more-targeted notices in the place where they'll matter.

15774:

newbies worse and give less feedback on how to write an acceptable article, without any change to the amount of unsourced content in mainspace. The backlog of unsourced articles is completely unconnected to the new articles that this proposal is about.

15443:

I think it's premature to predict the outcome. A straight vote count is running a bit under 40:60 against the proposal, but a number of objections (including my own) are to specific details (e.g., specifically requiring an inline citation instead of a

15643:

and every citation in that article refers to a self-published Blogger page, then that article is just as useless as it is uncited, because we cannot verify whether that article is spewing out myths or not. Many editors here objected my proposal due to

5009:

leaves the question open. However by ratifying the arbitration policy, the community agreed on giving this responsibility to the arbitration committee, as well as agreeing that this scope has to be altered via the arbitration policy amendment process.

4095:

You know what I meant by "false positive": that people were appointed who shouldn't have been. You're not confused about that. What you are asking is for me to substantiate the claim, which is an inappropriate question, because it would require naming

13867:– Let's suggest something new. Before PRODDING, one should check for sources. If it turns out that the unsourced article has no reliable sources to verify, then just PROD it; otherwise, it should be draftified as an easy and less bitey than prodding. 13328:

was rejected. Let's not make this another topic area where people keep pushing essentially the same proposal with slightly different wording until, through tendentiousness and exhaustion, they manage to get something in (and then ratchet and repeat).

3412:

already supports move protection, it just needs admins to apply the template when they move protect a page. It cannot be used for create and upload protection as there is no page to put the template on. It might be possible to add the padlock icon to

11249:

I think where previous proposals have failed before is in trying to do many changes at the same time (e.g. changing the emphasis of different parts of the main page while also restyling it). Anyways I have a modest proposal, under the principal that

12035:

require the article to be created, expanded fivefold or made GA this week. That's basically articles which are either new or newly worth featuring. A stub which occupied the title of this week's good work for a decade shouldn't block it from DYK.

11403:

is intentional - I wanted to have a simple header style rather than having an unnecessary box around the headers. Yeah, I liked ptwiki's a lot more than eswiki's, which is probably technically the "best" in terms of modernization but is too bland.

6917:. While practicing editing is important, AI tools can greatly enhance the editing process on Knowledge by providing valuable assistance, improving accuracy, and streamlining search. They are designed to complement human editors, not replace them. 1179:. Some editors in that discussion said that they would support removing the good and featured article topicons as well. If either readers or editors are interested in whether they are good or featured articles, they can be found on the talk page. 9838: 13391:

Did you find many such new articles? New page reviewers can mark unsourced articles reviewed if the topic is notable but I doubt that happens very often. Someone would have to sort those articles by creation/expansion-from-redirect date to find

2554:, below, as to the "randomness" of GA reviews. Any bad GAN review can be deleted, and any bad FAC/FLC review can be stricken through. My point is that the existence of bad reviews is not a good argument, in my view, for removing the topicons. – 12242:

Not really? The sections are Featured article, On this day, Featured image, Most visited articles, and Random article. It's not totally different by any means, and I don't think it needs consideration when we think about a main page redesign.

8602:

less substantive WP page. If not for public figures, for intellectuals, especially politically controversial ones, promoting some of their own - such as, of course, their best critics' - quintessential paragraphs should really be the standard

1213:

Strongly support removing GA topicons. GA reviews are essentially random. It's just one editor reviewing it and there is literally no training or qualifications that the reviewing editor needs to have. To suggest a GA article has gone through

14772:. I'm not seeing those articles as any more problematic than other similar unsourced articles. I think it is better that we make a judgement about which articles we delete rather than simply sweep away everything that meets a broad criteria. 13695:

That's why there is a moving deadline to the left to clear out the backlog. Stage one of this proposal is to prevent further growth of the backlog, stage two of this proposal is to start tackling the remaining articles from newest to oldest.

13216:

All of the objections raised to the two similar proposals in the last few months still apply. I really can't be bothered to repeat them again. We're supposed to be building an encyclopedia here. Clearing a backlog is only incidental to this.

8080:

Using ai to help users find policy pages seems fine, but asking it for advice when it comes to editing conflicts seems like a poor idea. We already have the teahouse and help desk, and the likeliness of ai giving an incorrect answer is high.

2938:

We having no evidence for something doesn't necessarily mean it isn't true. Readers curious to find out the meaning of the icon would hover over it - it says "This is a featured article. Click here for more information". The click leads to

8236:

is just a simple advice which may be made compulsory for AI as a disclaimer for issues related to 'warnings for unknown reason'. If you still oppose the idea, please list the reason and I will try to be more objective in answering those.

4523:

A very sensitive position (including being able to do real life harm to people) which has very specific proven-trust requirements. I'm more comfortable with some experienced (elected) people making that analysis and decision. Sincerely,

3845:

I think graying out the Move button is a bit dubious from a UX perspective. The warning on the Move page after clicking on the button is already quite clear. Trying to illustrate that a page is move protected prior to an attempt to move a

4971:

I didn't want to repeat the process since I linked to it, but mostly yes. An amendment can be also be proposed for ratification if it is approved by the committee. The majority vote to ratify must have at least 100 supporting statements.

12410: 10424:

Yes, I do agree with that, especially the Slovakian one. However, you know we still need ideas from other contributors, which can be challenging. Nevertheless, I will make an effort to advocate the idea of changing the main page design.

8136:

An AI interpreting the policies and guidelines is simply asking for trouble. Nor is there a need, as pointed out above, to force anything official right now given that user scripts exist and the WMF may be carrying out its own research.

7429:

results in real time by 'searching' on internet, but the Knowledge(AI-opt-search) work would be simpler even in term that it is just based on internal pages of just Knowledge. In other words, right now Knowledge's traditional search is '

9957:

What are your thoughts on starting this initiative? After all, the main page represents our entire community. I understand that changing a familiar style can be challenging for many users, but it's part of the natural cycle of updates.

3772:

I proposed it so that users are able to see if the page is move protected or not, just like edit protected. With regarding salted pages, I meant to add, they would also be on the top right if at all possible, otherwise I suggested the

15489:

This is not a "very costly move" as you have said, as multiple editors has pointed out that this is not that much different to how modern Knowledge processes new unsourced articles. We just don't allow them to come to the mainspace.

14327:

There have been multiple recent consensuses that explicitly opposed to using prod or speedy deletion to clear the backlog, so this is clearly something the community does not want and continuing to propose it is getting tendentious.

12358:

When it comes to modern design, the Wikimedia Foundation understands that initiating random discussions here on Knowledge can be challenging, as it is difficult to satisfy everyone, and people often resist adapting to new changes.

4307:

If that summary is correct (and I have no reason to doubt it) then it sounds like an ideal way to appoint people to those positions. Certainly better than any RFA-like process. Let's remember that ARBCOM is elected by popular vote.

3226:

On one hand, it's definitely an improvement over nothing, and should minimize confusion. On the other hand, topicons look better, and nobody might care enough anyways so that these who care would click on the icons to learn more...

12479:

The portals link was added to that box as a small concession after we decided to remove the more prominent portal links at the top of the Main Page. I'm not saying that's a reason to keep it, but just historical context you should

5484:

please remove this redirect and instead create a disambiguation page of “People’s Publishing House”, so that other pages with similar names, for example “People’s Publishing House (India) can be listed on that disambiguationmpage.

1223: 16146:, PROD is the prescribed process for completely unsourced BLPs (which seems to be the majority of articles that NPP deals with). Do you feel that BLPPROD is also BITEy? Do you use draftification to get around the BLPPROD rules? 6603: 7792:

First of all I appreciate your civil tone. I understand the lack of confidence on currently available AI platforms, and that implementing it site wide may bring issues, increasing the overall time to solve. But for that you said

2779:

Opponents of the GA and FA topicons claim they are visual noise, yet they criticize the choice to not include them in mobile viewing as proof the review process is wasteful. While Levivich is correct that the peer review process

9485:. First of all, banners are bad, and encyclopedia shouldn't advertise anything. Second, because there are fewer readers and fewer editors, some projects are not really useful, and it's not en wiki aim to improve other projects. 9284:

per SilkTork. Besides, a lot of people, both editors and readers, are already frustrated by the endless, huge banner ads that get foisted on articles every December, grubbing for money for the WMF. We don't need to add to that.

5407:"'''Show preview'''" to see a preview of your edits, or "'''Show changes'''" to see what you have changed. this page is cleared regularly, feel free to try your editing skills below. <span class="plainlinks clickbutton": --> 12499:

I don't use either Contents/Portals or Contents regularly, but it seems that they are fairly equivalent, and the Portals look much nicer and have more links. I don't see the point of replacing the Portals with Contents alone –

14273:

in a draft or userspace, that's up to them.) An alternative might be to draftify unsourced new articles, and forbid returning them to mainspace until at least one source is added, but when you're already dealing with a flood,

7556:

Yeah, I do think helping new editors overcome the hurdle of their (and everyone's) first edits often being pretty rough. Structured Tasks can help with this, but the "policychecker" idea has merit, at least as an exploration.

5796: 14518:, cited or not? Those are not synonyms, and anyone who has got stuck into trying to write or salvage articles will know that there is a big difference between an article that lacks citations and an article that lacks sources. 8660:

edia, not Knowledge. This is the exact Knowledge-is-the-best-and-most-important-project conceit I've been talking about. And I wonder which project you object to— Wikidata doesn't follow our notability guidelines, Wiktionary

4108:

arbcom CU/OS appointments (calculate the #-removed-per-#-selected ratio and you'll see a much higher % of CU/OS appointments are reversed than the % of RFA passes who are later desysoped, even though both are extremely rare).

3870: 14081:

sourced, not whether any sources are currently listed in the article. We should be making it easier for editors (new and old) to contribute notable articles to the encyclopaedia, not putting even more barriers in their way.

10393:

I kind of like the Spanish version, but Vector 2022 forces the text to wrap in a weird way because it's so narrow. The Turkish version is pretty similar to en.wiki's, and wouldn't be much of an improvement. I also looked at

7904:

Note that the new editor does not know about script and rather says 'Hindi', a language. The editor need not be expert in linguistic field, but this initial gf edits should not discourage them from editing constructively in

4003:

their workload. Over the last five years, functionaries have been asked by arbitrators to evaluate 26 candidates for CU and 16 for OS (people applying for both tools are included in both counts), This is not a large burden.

16592:

We currently have many thousands of unsourced articles, the backlog is enormous. Having unsourced articles is a credibility problem for a start. With no sources, how can we verify we're not spreading hoaxes to our readers?

3290:. I've also been a big proponent of requesting Foundation staff to run user testing on some of our community design norms, i.e., banner blindness or in this case topicon literacy. I can track down those requests if useful. 13923:

that "the burden to demonstrate verifiability lies with the editor who adds or restores material." I do agree that retroactively apply this proposal to very old articles is a very dumb idea, but I do think that we need to

6669:

I am busy right now to handle any opposes which I surely have answers to. Next time I will objectify this more and come with clearer proposal and beforehand answer common opposes. Anyone in support may start discussion in

11588:

picture and ITN swap places so that ITN can expand coverage of its recent death entries. Issues like this can't be resolved by the editors who maintain the individual sections and so an overall mainpage forum is needed.

8228:'interpretation' I consider wrong is that WP AI is given specific cases by user, say related to NPV, and the AI presents a 'solution' rather than simple 'AI optimized search' that too based on its own 'interpretation' of 5092:

Sure, that would be a reasonable approach. We might be thinking of different connotations for the word crisis. I agree that most likely there would be sufficient support to amend the arbitration policy at the same time.

2322:

The icons are an extremely powerful motivator to get involved and improve Knowledge. It only works because they are shown by default. I'm sure it would be trivial to squelch display with a script for an opt-out option.

9001:

There are far fewer reasons to do this than not to do this; some of the reasons that I can think of off the top of my head include NPOV & not wanting to have to look into the editorial practices of other projects.

3623:

will be there for all, including unregistered users and admins. If a page is not move protected, then the lock will not show even to unregistered users despite the fact that unregistered users cannot move pages anyway.

13117: 10265:

I use the full width view of new Vector, but when checking it on the narrow width, it looks fine to me. A little narrow perhaps, but not nearly constrained enough IMO to obviate the advantages of a two-column view. ―

4322:

Mostly - except that that is the general "routine" process, arbcom also can just appoint anyone they want if they pass with a 50%+1 committee vote (we see this the most when there is a request from a former funct.) —

3037:

Related to this: I wonder what a reasonable comparison would be. (If it's 1,000 clicks, is that a lot or a little?) I wonder whether it could be compared against the Portal: links that used to be at the top of the

15467:; not sure there is much more to be said about why this would not be a great idea. I think this would go against the spirit of the project, and that it would be a very costly move in return for not much improvement. 7343:

Sorry, got trigger-happy when I saw 'AI.' I could see it working if it was used as an addition to the search. Something similar to how Bing uses Copilot to summarize search results. We would just have to be careful.

7195:

I think you need to first find people with appropriate experience and resources who can work on this type of project. This will allow them to engage in any discussions and thus guide them down more productive paths.

4072:

OK, that wasn't a well phrased question. I was trying to understand what you meant by "false positive" (which you've now made clear means people who were appointed who you think should not have been) and importantly

3510:

Alternate proposal: What if we grayed out the move button and put a lock next to it? I agree with Daniel and ToBeFree that locking the edit button would be too confusing, but this could work if technically possible.

15837:

The last proposal said that we should PROD all articles right now that does not have a source. That's absolute lunacy. What I'm suggesting here is what the community has effectively done to new articles since 2020.

11570:

When you try to get a lot of people to make a decision together, they often spend more time attacking each other than they do collectively moving in any productive direction, if I understood the metaphor correctly.

2247:

I'm piling on to say that FA and GA are very important tools for developing quality articles and qualifying articles should be highlighted. The icons are also a learning experience for anyone wanting to click them.

2105:

If my GA/FA articles didn't get the topicon, I rather doubt I'd have even known they existed, or felt compelled to get my articles to that status. The badge is useful to readers, and a great inducement for editors.

16295:

draftifying. But I do agree with the sentiment that we are no longer in the wild west days of Knowledge. Sourcing exists, can be found, and should be found before a totally unsourced article gets thrown out there.

14193:

years, and the standards have dramatically increased since then. Back then, this might've been acceptable, but it is no longer, hasn't been for quite a while, and this just codifies that. This is also in line with

13900:

possibly make it expedient to create a new sticky PROD similar to the BLP PROD. We are not being swamped with unreferenced new articles, and it is very easy, and takes very little time, to do a WP:BEFORE search. I

8259:

viability, which means there isn't any impetus to work towards refining the scope and objectives. I think the proposal needs to have more concrete resources and expertise behind it in order to get better feedback.

7537:

pressure of getting 'edit wrong due to policy/ guideline' fear from new editors, because those who have such worries are the ones who are most dedicated. And it may surely ease the navigation process like gps did.

1958:

means anyway, in my mind it serves as more of a motivational little trinket for editors, and I'm concerned that removing the most prominent area where it's featured could decrease the impetus to improve content. ―

6020: 5983: 10687:

The third column has line lengths that are way too short in the limited width. That makes the excessive space to the right of the descriptions all the more jarring. And that is in my version of V22 enhanced with

6129: 7279:

Feel like this is an unnecessary way to teach policy (and it's just asking for issues). We should avoid AI in general, and instead we should be improving new user onboarding so that we don't need this solution.

16703: 16543: 13137:

to aid completely new editors to create a new article. I think that the wizard should be shown more prominently in "Knowledge does not have an article with this exact name" notification box, as well as making

11912:

Not necessarily, there's quite a few people in this thread who would support that idea. That's also what the purpose of an RfC is: gather consensus, and if it is a quick-fail, so be it. What do others think?

4929: 12879:

inline, reliable, third-party source. That source does not need to completely establish the topic's notability (because that will be decided in AfD); its only job is to verify that this topic is not a hoax.

7868: 1308:

single article through FAC would surely be better spent improving some of the millions of articles that have been substantially untouched for a decade or more. But in any event this would be a good step. --

15732:

There is no excuse for creating an unsourced article in 2024. Unsourced articles cannot be repaired by adding sources; they must be rewritten because without the sources we cannot tell if they are COPYVIO.

3169:

without the colored title text would make the most sense here. It shows the article assessment in an intuitive way and clearly indicate links for curious readers to read more about our assessment criteria.

11258:. It could do more work from someone who actually knows how to design (I like ptwiki's main page header a lot and that could be incorporated), but all I did was basically remove unnecessary (imo) styling. 4888:

for the nth number of good reasons above, plus ArbCom just started essentially open instead of timed applications - which was the big step forward needed in this process. Lets give it a chance to work. --

3482:

the display of the few remaining local image pages. A move protection icon, however, is an unneeded gimmick that is more likely to confuse readers than to help the few editors who intend to move the page.

15712:

cleanup, in that a PROD indicates that "this article that does not have inline citations does not belong to Knowledge". And it is not wrong to do so, because it corresponds to our current best practices.

11078:

This, but for WikiProjects, which are largely bland and uninviting. This is something I've felt for a long time, but have just now worked up the courage to say... so be nice. heh. (but isn't it obvious?)

10811:

I think the portal links should be omitted, especially since previous consensus is to do so. Also, I don't think that the Knowledge Ad should be present, and I'm not sure about how the logo is presented.

6726:

Proposal 2: Proposal 1 + Giving AI more AI feel, by letting it become more than better search engine, by becoming suggestion judge for small situations. Example: You editors are creative, you may add one.

6121:

Ah, but now that I imagine myself doing that, selecting "redirected" in the notice at the top, then Ctrl+F, then just hitting Enter once or twice may be even more convenient than typing it in. Never mind!

5987: 3915:

in this way, the community has more of a say of who gets to be a CU and OS other than our current arbitrators who currently hold those positions. I would also be open to saying with arbitration approval.

10480:

I would concur that it's a good idea. IMO, some of the other Wikipedias have Main Pages that put enWP's one to shame. But as others have stated, good luck finding something that everyone can agree on. ―

5527: 1813:

per Barkeep and Bilorv. Moreover, just because mobile readers are unable to see them it does not mean we should remove them from PC readers (if anything it means that we should add them to mobile view).

16571:

have unsourced articles. Opposition centres around how to balance our goal of creating a verifiable encyclopaedia against our equally-important goals of being an encyclopaedia that anyone can edit that

13988:. If you want to change "credible claim of significance" to "credible claim of significance supported by an inline citation to a reliable source" then you need to explicitly propose amending that page. 12433: 12213: 11986:. I do not know how many new articles are created on Knowledge each day, but it could be dozens, and we only need to list a handful of them to make the Main Page show Knowledge is a developing project. 5961:

For most of the many redirects that target sections of articles, instead of entire articles, that means that this line isn't in view, which makes little sense to me. After all, if the page includes any

4904: 3929: 3924: 3626:

In addition, the pending changes protection (check mark) and semi protection (user icon) are both grey with the semi protect lock being slightly darker so all could be grey, but I still prefer it to be

4180:

I see no current problem with ArbCom selecting them, and this has the slightly greater risk of granting the power to the wrong people. Moreover, it could lead to a flood of premature self-nominations.

3287: 15365:; unclear that this necessarily resolves a problem - as the February unreferenced backlog drive showed, even experienced Wikipedians were making mistakes and incorrectly sourcing articles. Regards, -- 10642:, indeed, I know no German or Dutch! But yes, a "huh" because I wasn't speaking for myself on the quality of the vetting, merely that it's a view (among others, of course, as you point out.) Cheers, 4174: 1662: 7806:

only complement them. And as they say it, something is better than nothing (being added), especially when it is promising to look at. I would also like to know 'more opinions' you mentioned to have.

6653: 3312: 13828:

This is exactly what I meant. This PROD would essentially be a formal enactment of our "unspoken rule", which is that new articles on Knowledge must have sources. A lot of people here don't get it.

7692:

a GPT with access to all Knowledge namespace articles concatenated into a single file. Go ahead and ask it about Knowledge rules and guidelines and maybe it will actually give an accurate answer. –

2879: 9063:. The purpose of Knowledge is to be an encyclopaedia, not to recruit users for other projects. If the Wikimedia foundation thinks they need more users elsewhere, they can do their own advertising. 7059:

I presented my ideas here (on proposal) as conjectures, indicating that they are open to improvement and further development. This allows for collaboration and the exchange of ideas among editors.

6234:. In theory, it is possible to have tracking categories which are manually included on pages the same way content articles are manually categorised, but I am not aware of such tracking categories. 4276: 13028:

We should also find ways to retain new editors if this proposal is enforced, because that would set an even higher barrier for entry for new editors to Knowledge. This is the reason I why invited

7006:

4) This is just a tool. I may be wrong, but it is not even a major update like enforcing new vector design on IP editors/ readers. Just a sidebar for easy searching, and policy updates will do it.

4503:

trustworthy, active admins who know how to use the tools and deal with abuse within the bounds of policy. ArbCom appointments with community consultation strikes the right balance in my opinion.

3904: 2123:, mostly per voorts. We should show more of this, not less; readers deserve to know if an article has passed some quality gate. On the topic of substandard GA/FAs, yes they exist, no we shouldn't 1473:

This is how I think about the distinctions: start is longer than a stub, but either not very well-cited or written, C is a start class article that's starting to meet some core PAGs, and for B we

12841: 5834: 4805:

RfA has serious problems and I don't think we want to replicate them to other processes. The current system for CU and OS appointments seems to work well, and allows adequate community feedback.

16392:

I wanted to correct one erroneous item which is that draftifying articles is typically an available option. Due to the backlog at NPP they are all generally past the time limit for doing that.

12018:

are new articles. I was hoping some recent change patrollers could enlighten me as to how many articles are likely to be created on Knowledge each day, to know whether my suggestion is tenable.

5146:

gets here (not in the next few months, but could possibly be later this calendar year), that we might need more CUs than we have now, particularly during the first few months of the transition.

3949:(Moral Support) I'd love arbcom to be out of the CUOS business, primarily so that we can focus on putting people on the committee primarily on the strength of their dispute management skills. — 6708:

Create AI for Knowledge by taking existing model and feeding it the guidelines and policies. This will make it easier to find relevant policy. Example: I ask AI for any policy regarding 'using

1461:

I've had a similar thought before myself, but I've come to believe that in practice, the distinction between Start/C/B-class is often too poorly defined to be useful information to the reader.

12590: 12373:

And thereby decides it has the authority to unilaterally override the community? While I don't really care what the app does since I don't use it the principles there are fundamentally wrong.

10790: 10619:

I also wonder what you mean by "hein". A dictionary search shows that it is 1. surprisingly not German 2. Dutch for death 3. French and Portuguese for "huh?". Is that last one what you meant?

5548: 2824: 1193: 16494:

Because this is de facto required for all new articles on Knowledge. Inline source helps the person who actually use the reference section have an easier time verifying facts in the article.

10234:

The white space is probably from the skin you are using. It looks fine in legacy vector. Easier to change the skin to something that you like rather than change the main page for everyone

14895:, our core content policies. One source is not too much to ask. And for this reason, there is little reason to save an unsourced draft. Without a source, a Knowledge article is meaningless. 3378: 3356: 1382: 11201:'s for Vector 2022 users? This way we could keep up the "modern" look for Vector 2022, but keep the "legacy" style for Vector legacy. I just don't know if this is technologically feasible. 7832:: For those who oppose major change related to AI, the ai need not take a very inclusive part in Knowledge's working for now, but a simple AI search like this would really ease the problem: 6165:

I will just mention that it works completely differently on mobile. Here, no matter where it sends you, a notice appears for a few seconds at the bottom of the screen, and then disappears.

15004:

Much more likely, the unsourced article is unsourced because the author didn't summarize sources, they wrote off the top of their head. This is not worth saving, nor is it worth defending.

14977:

the unsourced article, the second editor has to start by looking at sources and coming up with a summary of those sources. Then the second editor can look at the unsourced draft and yeah,

3366:. I am not ECP myself (have ~260 edits). This is just an example of what it looks like for non ECP users to try and (re)create a salted page. You may wish to use an incognito tab to do so. 8906:

I could just as easily say that proposals like this one are detrimental to the English Knowledge because they consume editor time that might otherwise have been spent improving articles.

6799:

is fed on with its response, which will make it even more erroneous. Proposal is to simply feed an AI on Wikipedian data only and program it to link specific policy it found the info on.

14591:

the deletion is expected to be uncontroversial. Does this remove the 'uncontroversial' requirement? Does that mean that this new type of PROD can be used multiple times? Or after an AfD?

3653:

Actually, edit locks do not show up for users who can edit the article anyway, at least not on mobile. I have not seen them for EC-protected articles since I reached extended confirmed.

1388:

The topicons are useful for readers as indicators that an article has gone through some kind of review process. This is in contrast to vital articles which is editor facing information.

15188:

Nice to meet you too. That link points to many articles where the PROD tag has been there for nearly a week and nobody has removed it, making your claim that that always happens false.

4957:

So that would be 100 petition signatures and a majority on an up and down vote? I believe that is what it was on the vote that removed any remaining provision for an appeal to Jimbo.--

3326: 2516:

per the commenters above. These are not mere decorations; the icons signify that the GAs and FAs have undergone their respective quality control processes, and are thus helpful to the

2305:

The proposal might be an argument for a double-check system for GAN's, but it isn't a reason to remove the topicons. They at least show articles that have been through some vetting.

12387:

It's not possible because the app exposes all the versions, so, how will you decide which style to use? If you ask each community separately, it will take years to reach a consensus.

12324: 11688:

You don't think that a larger image and prose that isn't chopped into tiny sections that resemble the old "See spot. See Spot run. Run, Spot, run." children's books would be useful?--

8635:. Several of the 'sister projects' are worse than useless. At least one is run in a manner entirely contrary to the stated objectives of Knowledge. We should not be encouraging them. 3496:

Displaying a lock for move-protected articles would confuse far more people than it would help, I agree with ToBeFree. I'm ambivalent about displaying a lock for the other two cases.

2746:

As primarily a reader of Knowledge, I find the GA/FA top icons to be useful information. Removing them would do some readers a disservice, and there is no advantage to removing them.

1559:

Per Barkeep. I'm not sure how valuable fronting the lower ratings would be (they're often out of step with the quality of the article) but I'm not immediately opposed to adding them.

11708:

We do not have an overabundance of blurbs, blurb words, recent deaths or DYK hooks. Besides size limitations, there's a reason newspapers separate every single article into columns.

2625:: The usefulness of these icons is different than the possible usefulness of a vital articles topicon. Also not seeing an actual reason listed for why we would want to remove these. 1188: 9670: 1730:

in this discussion or the previous one. The process basically works, and judging from the silence, it doesn't seem to be contentious these days. Congratulations all around. - Dank (

13814:

miss a day. It’s a tough project. I just want to note that this is a burden on editors to fix, and a lot of times adding a inline source to a plainly bad article doesn’t do a lot.

10158:

propose incremental changes, one at a time- a wholesale redesign would never gain consensus. Even a small change would take much work to convince others to support and implement.

1740:

Featured lists should undoubtedly follow featured articles in this regard. If any decision is ever made to change or remove the FA icon, featured lists should get the same result.

13167: 13121: 10960:

Honestly I quite like eswiki's, but I would only change the style, and leave how ITN and etc are formatted as is. But, strongly support changing the "Welcome to Knowledge" header

7014:

6) I am not considering myself eligible to answer questions related to financial side, as it dependent on what the community and foundation decides and how this proposal evolves.

5958:

When one arrives at an article via a redirect, the page is rendered with an additional line saying "(Redirected from )", directly (?) below the "" byline below the article title.

14313:

This is why I suggested that "In the future, this "imaginary" deadline could be gradually push backwards to tackle older and older articles, until the backlog is fully cleared."

9432:

Rather than asking readers to take an action to see banners about other Wikimedia sites, we may as well just ask them to visit another page to learn about other Wikimedia sties.

5253: 4869:

per Thryduulf finding only 42 CU/OS evaluations over the past five years, suggesting that the current system poses a minimal burden on ArbCom while avoiding the toxicity of RfA.

15015:

Knowledge articles, even if they're hosted at wikipedia.org. The starting point for every article and every editor is sources. Anyone who starts anywhere else is doing it wrong.

14765: 8984:

effect on the expansion of the related communities. I am wary of putting English Knowledge in a position where it can determine the success or failure of other Wikimedia sites.

8610:

for this precise purposes that e.g. hovers under an infobox closer to the top of the corresponding WP article rather than at the very bottom, would really do WQ some justice. (*

5938: 5051:

I'm inclined to trust the steward when he says such a situation with the stewards would create a crisis. Perhaps it would be resolved the way you suspect but perhaps not. Best,

2551: 1133: 13891:. The topic of an unsourced article is often notable. The content of unsourced articles is often accurate and verifiable. An article is not "unsourced" merely because lacks of 9468:

Banner blindness is bad and getting worse. Also, I cannot say that I endorse all of our sister projects; forcing the community to vote on who we endorse only ends in disaster.

9345:

Wikisource or Wikibooks. And yet that's precisely what these suggested banners provide: essentially a link to the front door of the library, or to a full dictionary. What is

7324:

finally. We are not a company, but encyclopedia's community which should be not conservative in approach of helping Knowledge's editor base growth in fears which make no sense.

3463: 16707: 2903: 16316:. However, there must be safeguards. 1. Draftifying is better than PROD. It preserves good text, even if unsourced. 2. The PRODing editor and the deleting/draftifying editor 5001:

I think the wording in the current arbitration policy has given sole responsibility to the arbitration committee for appointment of checkusers and oversighters. I agree that

1169: 15640: 15404:

I think that any articles without citations should either provide a source or be deleted.But there should be a ~month long period to provide sources before deletion occurs.

13525: 13113: 2834:

pay absolutely no attention to article ratings and their associated topicons. However, because I don’t think they harm the project in any way… I am neutral on the question.

13984:

says "verified with an inline reliable source from the first revision of the article" not "verifiable". If you want to change that, you need to explicitly propose amending

6279:. It should be obvious to any administrator not to delete tracking categories, but it can be stated that tracking categories may at times be empty or even most of the time. 6076: 6012:

That said, I suppose a generic phrasing like "(Redirected from to this location)" would work for both, or even all three, cases. That'd leave the space issue to be solved.

5297: 1982:

Not clear what the harm is, and if the only benefit is making people feel good about getting an article to GA/FA (by having a shiny icon on the article) that seems enough.

14100:

don't think that I've ever seen one with zero sources and just a few with zero in-line sources. The widespread / pervasive volunteer-crushing problem is zero GNG sources.

13347:

raise the quality bar for new articles by a very small degree. We already do this with AfC, where quality standards are much higher, it is nothing we don't already do. --

13173:

for URLs/book titles (and pages), and once you create a draft it would show you how to expand these fragments of info into fully-fledged "wikipedia-compliant" citations.

7437:' but much less sophisticated and easy to build. Also this would complement the search and may not be major change, it is proposal 2 which may suit your oppose. Regards, 6595: 6024: 6003: 4055:

Arbcom elections and RFAs have also produced false positives, in fairness, but I think RFA has produced the fewest false positives (proportionately) of the three systems.

1176: 15655:

I'm just gonna sum up my proposal as follows. Missing content on Knowledge is terrible. Having incorrect content masquerading as facts on Knowledge is much, much worse.

15274: 13325: 12837: 9519:~ considering how blind banners make people, how useless (and even antagonising) much online advertising is, there is no way to consider this a good idea. Happy days, ~ 7981:. Wrong page? This seems like a request for software rather than something that needs consensus. Any volunteer may create a user script or Toolforge tool that uses AI. – 6016: 5979: 2718:

Oppose per Barkeep, Bilorv, Epicgenius, Aaron Liu, and others. I don't see any point in removing topicons; I find them useful and evidently a lot of others do too. Even

7797:, I would also point that we have bots, which do not directly edit and cross check the articles (as a dystopian editor-less Knowledge may have, and I oppose this), the 6526: 6464: 6174: 6040: 5661: 13928:

communicate with new editors that new articles require at least 1 cited source. We can help them to find the sources, but ultimately the burden is on them to find it.

10774:

I don't think WPAds should be isolated. Maybe it could replace the search box.Also, any ideas for how to eradicate that awkward gap below the image for the first box?

10052: 10044: 6536: 6125: 6112: 5184: 15959:- (has anyone else mentioned this? sorry if so) this looks as though it is about to deal with most of the issues raised here, to the liking of some if not of others. 9820: 6160: 6133: 5206: 2713: 16699: 11830: 11483: 11425: 10526: 6192: 6094: 4452:

What is the problem that has to be solved? The present RfA-process is quite a bit of mud slinging, so it only decreases the number of able & willing candidates.

14386:

So, contrary to policy, you are actually seeking to bypass the challenge part (for the contentious material) and jump straight to deletion (of the entire article)?

10681:

I don't know any good ways to fix this, but flushing the Welcome banner's text to the left leaves a lot of blank space in the rest of the box, making it feel weird.

8939:

No, a proposal that seeks to improve the encyclopedia is not a detriment. This, however, is a proposal that diminishes the encyclopedia, which is why I opposed it.

2755: 12733: 12626: 11207: 9969: 9794: 9698: 6613: 6067: 5698: 4852:

These roles have real-world legal effects and are required to sign an NDA. We definitely shouldn't elect them by popular vote. The current system works just fine.

4334:

This is equivalent to former admins regaining the tools on request as long as at least one 'crat is happy to flip the bit. The only difference is that there is no

1302: 13942:

This conflates notability with content; we can establish notability in the absence of sourcing in an aricle, verification of content requires sourcing. Regards,

13742:

and mark it as reviewed, not to draftify it. I would support allowing NPP reviewers to directly draftify unsourced pages, but this is not the case at the moment.

3590:

Erm, I said one comment above in support of just greying it out. I suppose this proposal would be a strong recommendation to the WMF to implement such a feature.

2647:

GA reviews are essentially random. It's just one editor reviewing it and there is literally no training or qualifications that the reviewing editor needs to have.

1376: 16753: 11982:

ways in which s/he would like the Main Page changed. I do have a suggestion, and that is we could include a box listing some of the new articles in Knowledge in

9850: 9238:, as leeky put it. Other sister projects are very well what readers may be looking for, like WikiVoyage or Commons. There should be an option to opt-out though. 6535:

Re your idea to combine "tracking category" and "possibly empty category" based on the idea that tracking categories shouldn't be deleted, dated categories like

2479: 1859: 50: 16236: 12236: 9664: 8463: 6733:

seem to help, so will this which may be easier to build. But couldn't find discussion for one like this. The proposed proposals may be amended for better use.

6268:

That one template be used for each purpose instead of combination templates or a mix of combination templates and individual templates per the current practice.

5917:. He might know whether a bit of .js or something could solve this problem. If he isn't on wiki in the next couple of days, I suggest taking this question to 4104:), but that's not the same thing as saying an rfa-like process would produce fewer false positives for CU/OS appointments. I don't know if it would or wouldn't. 16038:(1): Is it that difficult to ask for one source? Our built-in translation tool also translate the references for you, so I don't see why this is so hard to do. 15547: 12984: 12889: 12819: 12298:

If consensus determines that there are specific objections to the landing page in the mobile app, that discussion can be brought to developers for changing it.

8978: 7010:

5) Knowledge itself would not have got the success, if it was not launched due to such considerations. Solving the problems is necessary to establish anything.

5470: 3206:. Nobody except the tiny minority of high-volume Knowledge editors cares about how we rate articles. Like: there are 8,000,000,000 people in the world, and 2896: 2818: 2353: 1761: 1208: 16168:

guideline. This doesn't mean we should expand the pattern we use to deal with unsourced BLPs (which exists for a specific reason) to the more general case. ~

15770:

be raised. The proposal is about new unsourced articles and suggests to delete them instead of draftifying as we currently do. This would make our problem of

14429: 13205: 13182: 13041: 12926: 12908: 10535:

Redesign it, but in a way that removes ITN and DYK. Both are massive investments for very little return, and much of the content they display is low quality.

9917:

It probably does on the website since mobilefrontend is mostly a different skin that uses the same engines. It definitely doesn't do that on the app, though.

9413: 9151: 9011: 7820: 7787: 3635:

for an example), the lock could be displayed within the infobox or even the top right, or somewhere before the 'This page is protected from creation, so only

3197: 2741: 2634: 1788: 1604: 1576:: It's good to show readers in an unobtrusive way that the article they're reading has been reviewed and is of quality. There is no benefit to removing them. 9187: 8554: 5637: 5463: 5438: 5272: 4111:

As for you personally, not really the place to discuss that nor a topic I care to spend time on, but for example this line of questioning was a poor choice.

2704: 2508: 2097: 15823:- the suggestion has already been turned down, apparently twice - no justification for bringing it up again so soon. For the rest, as per James500 and Joe. 15617:

Do I feel comfortable to say that we tolerate unsourced articles because someday, somebody will magically put a citation into it? No. Without work from the

14638:

Maybe not this specific discussion, which is already quite long. If you want to develop this idea further, I'd suggest going back to basics and asking what

12899:

backlog. In the future, this "imaginary" deadline could be gradually push backwards to tackle older and older articles, until the backlog is fully cleared.

9230: 9071: 2796: 2424: 2300: 2239: 1540: 1362: 16200: 15984: 14761: 14212: 12895:

If this proposal is accepted by the community, it would greatly streamline our efforts to cleanup uncited articles and prevent the growth of the cancerous

10729:: using Galobtter's design for the main portion. I think a large prominent search box on the main page is a good idea, though, as opposed to a redundancy. 10598:

reducing backlogs than ensuring the integrity of the main page, then it would be unsurprising, in some eyes, if these processes came under extra scrutiny,

8865: 8825:

Banners steal our readers' attention. By distracting them and taking them off task, we slow down them learning what they're here to learn. Ads make people

8644: 5423: 2771: 2205: 16744: 16576:

instead of removing them. I don't think it's fair to conclude that, if others see that balance differently from you, they're not "taking it seriously". –

15026:

statements in mainspace. It's rather an indictment of Knowledge that this was not the first policy, and that it's still not policy almost 25 years later.

9866:, had a paragraph overview of each project, links to any welcome portals on that project, and links to any coordination or other related groups on en.wp. 9295: 9055: 8887: 8124: 8090: 5450: 5345: 5325: 3944: 3741:

If a page is not move protected, then the lock will not show even to unregistered users despite the fact that unregistered users cannot move pages anyway.

2873: 2441: 1805: 1273: 16727: 16366: 15722: 15703: 15133: 15044:

You are entitled to your opinion about how other people's workflows for writing articles is "wrong", but unless and until there is a consensus to modify

14284: 13975: 13654: 13636: 12941: 12744: 12729: 12696: 12666: 12637: 12630: 12622: 11438:

I added 1px of margin, which was missing since I removed the 1px border. It looks ok to me now, although you might be right that more margin is needed. @

11286:

I like that! It's simpler but clearer. Although something seems a little off with the header box, the lack of border makes it feel poorly-defined to me.

9546: 9477: 9213: 8844: 8799:). I think that such a campaign would be best served by a focused group of 3 or 4 of them that is more able to effectively direct attention their way. ― 5211: 4643: 2591: 2545: 2378: 2314: 2271: 2192: 2115: 1554: 16555: 16090: 15947: 15929: 15683: 13875: 13308: 13226: 13112:

If you think the new editor article-writing experience is so demoralizing, maybe we should just not let new editors create articles in the first place?

13001: 9629: 9577: 9460: 9178:

per leeky. General support for the idea, without committing to doing all of them (after all, we can't say too much about the Incubator, Meta, etc.). —

9170: 8817: 8582: 8370:"Now what if I assure you that the AI will make equal or lesser percent mistakes than ClueBot NG in terms of overall damage to Knowledge?" That's a big 8073: 7927: 7375:- I'd find it so useful to find policies and guidelines, as long as the AI is only used for that and is not involved in anything else around Knowledge. 6837: 6815: 6786: 6771: 6749: 6688: 6546: 6481: 4385:

Thryduulf gives a good summary. I don't currently suspect that the CU/OS system we have is appointing the wrong people, and RfA is exceedingly toxic. —

3794:

Why should someone who can't move any page anyways see if a page is move-protected? They can edit most of them, so edit padlocks are definitely needed.

3219: 3179: 2843: 2673: 2397: 2166:. I would prefer we go in the opposite direction and make the icons more visible and more intuitive. Readers will not know which one is more "trusted". 1991: 1977: 1824: 1585: 1568: 1397: 1238: 15413: 9525: 9379: 9336: 9276: 8386: 7971: 7945: 4880: 4613: 4563: 4497: 4428: 3974:

ARBCOM were elected on the basis that this would be one of their responsibilities. They can tell us if they think they've got too much on their plate.

3632: 2582: 2257: 2013: 1878: 1653: 1317: 16304: 16259: 15374: 14962:

claims will present a comprehensive and neutral view of the topic, but that's true regardless of whether all, none or some of the claims are sourced.

14491: 14177: 13951: 13905:

again and again and again in quick succession. It is not particularly easy to check the large number of noticeboards for perennial proposals, either.

13837: 12584: 11171: 11146: 10648: 10630: 10612: 9494: 9258: 8411: 8251: 8215: 8200: 8173: 8146: 8042: 8023: 7996: 7160: 6926: 6902: 4739: 4444: 3744:

Then what purpose would the lock serve? Why would we want anons to see them? I don't think this or "it's not used much" are valid reasons to use them.

3120: 2690: 2230:

The great majority of Knowledge's readers are unaware that there is a quality scale. Indicators of article quality need to be more visible, not less.

16765: 16403: 16345: 16050: 16033: 15568: 14448: 14414: 14394: 14379: 14357: 14337: 14322: 14308: 14111: 14091: 14026: 14011: 13997: 13454: 13293: 13268: 13254: 12474: 12460: 10617:

Thanks for the detailed explanation, though I don't think the vetting that we do is low-quality. At least at ITN we do a pretty big referencing pass.

9948: 9590: 9396: 8993: 7566: 7551: 7524: 7133: 7113: 7098: 4602: 4515: 3253: 3238: 2446:

Adding: I fully acknowledge that both review processes have serious issues, but the average quality of a GA/FA far outstrips that of other articles.

2338: 1843: 1694: 1420: 1339: 16645: 16286: 15968: 15871: 15847: 15832: 15744: 15664: 15612: 15585: 15396: 15335: 15318: 15242: 14904: 14821: 14806: 14781: 13937: 13914: 13081: 13063: 12771:(i'm not blaming the 3rd party, I'm pretty sure it wasn't exactly sustainable on their end either, they haven't even updated their website since). — 9928: 9511: 5494: 4951: 4844: 4825: 4754: 4628: 4465: 3505: 3476: 2175: 1290: 16101: 16025: 15908: 15197: 15183: 15154: 14873: 14859: 14740: 14671: 14657: 14633: 14616: 14252: 13853: 13808: 12737: 12716: 10552: 9559: 8377:

On a side note, if somebody's getting four warnings, they're obviously not somebody who's going to go to the effort to search for policies at all.

8009:

consistency across the platform. So I propose a direct integration of this feature in Knowledge's interface, and so I started the discussion here.

7268: 7073: 7047: 7023: 6957: 4861: 4783: 4725: 4535: 4302: 4191: 4032: 3955: 2222: 2158: 2133: 1735: 1673: 1622: 1513: 16012: 15998: 15761: 15438: 13673: 13362: 13336: 13105: 13019: 12557: 12535: 11820: 11774: 11615: 10569: 9427: 9034: 8365: 8338: 7879:

This could be the possible reason that your edit was removed. For more information, you may ask the editor who reverted your edit for a response.

6888: 6859: 4898: 4700: 4686: 4672: 4658: 3134: 3055: 2856: 1938: 16423: 15387:

is done) cannot be moved back to article space without adding at least one source that has a credible claim of at minimum verifying the article.

13961: 13726: 12168:

Would creating styles in the template namespace just for a sandbox in userspace to include templatestyles be an abuse of the template namespace?

11804: 11789: 11758: 10530: 10036: 9912: 9137:. leeky puts it perfectly. Most readers don't know what on Earth a Wikivoyage is, when in reality, it could be exactly what they're looking for. 8623: 7676: 7658: 7631: 7503: 7468: 7451: 7416: 6471: 4923: 4814: 4479: 3969: 3860: 16525: 16503: 16489: 16384: 16137: 15522: 15499: 15484: 15302: 14482:

methods are better ways to address this issue than quickly deleting new articles. Such as a more leisurely deletion process or drafification.

13769: 13705: 13386: 11889: 11850: 11834: 11521: 11507: 11487: 11429: 11413: 11395: 11377: 9769: 9755: 9123: 9109: 9096: 8445: 7756: 7731: 7701: 7598: 7584: 7395: 6980: 4768: 4398: 4219: 4205: 2455: 1490: 1468: 16617: 16467: 16109: 14462: 12765: 12690: 12517: 12129:

Since I've noticed that many people prefer the Spanish version, I believe we should focus on structuring this discussion more effectively now.

11244: 11093: 9856:

Maybe rather than banners for individual projects, there could be a banner about sister projects in general linking to a page that transcluded

9727: 9441: 8771: 8757: 8701: 8686: 8487: 8310: 8295: 8268: 7241: 7226: 7205: 7190: 7176: 6342: 4587: 4317: 4037:

Please can you give some examples of these "many false positives" and explain how a different process would have produced a different outcome.

3838: 3805: 3789: 3767: 3662: 3648: 3520: 3398: 3274: 3106: 3077: 1917: 1127: 1123: 1115: 1111: 1107: 1103: 1099: 1095: 1091: 1087: 1083: 1079: 1075: 1071: 1067: 1063: 1059: 1055: 1051: 1047: 1043: 1039: 1035: 1031: 1027: 1023: 1019: 1015: 1011: 1007: 1003: 999: 995: 991: 987: 983: 979: 975: 971: 967: 963: 959: 955: 951: 947: 943: 939: 935: 931: 927: 923: 919: 915: 911: 907: 903: 899: 895: 891: 887: 883: 879: 875: 871: 867: 863: 859: 855: 851: 847: 843: 839: 835: 831: 827: 823: 819: 815: 811: 807: 803: 799: 795: 791: 787: 783: 779: 775: 771: 767: 763: 759: 755: 751: 747: 743: 739: 735: 731: 727: 723: 719: 715: 711: 707: 703: 699: 695: 691: 687: 683: 679: 675: 13539: 13518: 13489: 13472: 13443: 13424: 12027: 12009: 11597: 11128: 11109: 10308: 10056: 9311: 8953: 8934: 8920: 8901: 7363: 7338: 7307: 5598: 4270: 4236: 3997: 3491: 2563: 2528:

are that way. The reader may not know what they mean initially, but they'll certainly know if they click on them - the solution is to make it

2466:

least reduced) if the indicators were kept confined to talk pages. (I also second the comments made by Barkeep, Bilorv, and The Night Watch.)

1747: 16602: 16587: 14717: 14701: 12660: 12492: 12368: 12342:

That's because the app doesn't use the main page as its landing page, and that's what JPxG's talking about, though I consider it irrelevant.

11532:

Probably, yes it is time, however the problem is not to change the design, it is how to herd the cats without them canabalizing themselves. —

11267: 11224: 11187: 10955: 10806: 10692:. Under the normal limited width all the columns have line lengths that are too short and the sister projects' descriptions run off the page. 10673: 10499: 10434: 10415: 10388: 10324: 10073: 10007: 9882: 9842: 9743: 9616:

the sister projects, and picking which ones to advertise seems wrong (and will probably get bogged down in a huge no-consensus RfC anyways).

9602: 9433: 9370:

more aggressively, sure. That might be a good way to spend time. But these banners don't have any relevance, and therefore aren't useful.

9195:- Advertising (especially in banner form) is annoying and intrusive, distracting and wasting of screen space. It doesn't matter how noble or 9045: 8985: 8728: 8302: 8275: 8260: 7860:

WP AI: I found the following guideline regarding the usage of Indic script (which the Hindi language is predominantly written in) in infobox:

7233: 7212: 7197: 7168: 5768: 5732: 5690: 5558:, it's because TFL appears Mondays and Fridays, aligning with days of the week. That seems to work fine, so I don't see a need for a change. 5521: 5500:

If they really are of equal importance then a two-entry disambiguation page may be the best solution. However, the new article you created,

5286: 5094: 5067: 5038: 5011: 4973: 4943: 4058:

But you really should be ashamed for asking for names. I'll give you one name: yours. That's not nice but neither is what you just asked me.

3717:

Oh, I see what you are talking about now. I think that the padlock would be a type of topicon, and those really should be visible on mobile.

671: 667: 663: 659: 655: 651: 647: 643: 639: 635: 631: 627: 623: 619: 615: 611: 607: 603: 599: 595: 591: 587: 583: 579: 575: 571: 567: 563: 559: 555: 551: 547: 543: 539: 535: 531: 527: 523: 519: 515: 511: 507: 503: 499: 495: 491: 487: 483: 479: 475: 471: 467: 463: 459: 455: 451: 447: 443: 439: 435: 431: 427: 423: 419: 415: 411: 407: 403: 399: 395: 391: 387: 383: 379: 375: 371: 367: 363: 359: 355: 351: 347: 343: 339: 335: 331: 327: 323: 319: 315: 273: 269: 265: 261: 257: 253: 249: 245: 241: 237: 233: 229: 225: 221: 217: 213: 209: 205: 15457: 15264: 12856:

unsourced articles, because currently there are still new articles that does not have sources attached to it (see the 2023/2024 category at

12608: 12199: 11963: 11949: 11733: 11473: 11459: 11359: 11343: 11329: 11295: 11053: 10983: 10919: 10899: 10881: 10867: 10853: 10839: 10821: 10785: 10769: 10755: 10738: 10721: 10469: 8546: 5155: 4798: 4663:

I don’t disagree in the slightest. I was just hedging to avoid nitpicking about all the possible factors that lead to a shortage of admins.

3983: 3726: 3712: 3696: 3678: 3615: 3601: 3585: 3550: 3536: 2954: 2615: 1403:

I think the distinction that's been drawn between quality ratings and vital article status is strained. Both are done through the work of a

1156:

topicons for good and featured articles. I find it very unlikely that consensus will change from here given the overwhelming !vote to keep.

16417: 15080: 15065: 15035: 14971: 14956: 14942: 14689: 14063: 13823: 12788: 12353: 12163: 11719: 11701: 11683: 11667: 11649: 11633: 11580: 11281: 11008: 10969: 10516: 10210: 10149: 9999:

It still looks good on Desktop and even on my phone, in Monobook and Minerva. Everything is in one column and the content is very readable.

9809: 9713: 7104:

done elsewhere. At the current pace of development, I would expect that production-ready open source systems that do this are not far off.

5612: 5580: 5359: 4372: 4358: 4347: 4329: 4261:

The last year in which CUOS were selected by the community (it has been done before) was a disaster for providing the relevant privileges.

4156: 4138: 4120: 4090: 4067: 4046: 4012: 3336: 3322: 3156: 2931: 2141:. This does not really seem like a good idea to me; if there's a bunch of crappy GAs we do not need to address topicons to deal with them. 311: 307: 303: 299: 295: 291: 287: 283: 279: 201: 197: 193: 189: 185: 181: 177: 173: 169: 165: 161: 157: 153: 149: 145: 141: 137: 133: 129: 125: 121: 117: 113: 109: 105: 101: 16174: 16155: 15896:

citations aren't the same as articles merely lacking inline citations. There must be clarity on what exactly is proposed to be changed. –

15351: 15108: 14928: 14913: 14127: 13403: 12396: 12382: 12179: 11933: 11903: 11071: 10229: 10105: 9984: 6777:

Re proposal 1, ChatGPT can already answer this question decently. Presumably its training data includes pages in the Knowledge namespace.

5930: 5853: 5566: 5316: 5195: 3298: 3032: 2914: 1253: 16437: 15618: 12914: 12138: 12124: 11549: 11313: 10452: 10284: 10258: 10243: 10189: 9890: 9875: 6618:

Summary: Renamed user g5s6n3yi8z7g08cs made unapproved automated edits performing various tasks over the course of a couple months using

5996:

usually be at the redirect target. At the section header would be appropriate for R to section. Not sure about R to anchor. Cheers · · ·

5170: 5075: 5060: 3430: 1999:. The provided rationale is extremely unconvincing - "some people mentioned it?" Lay out the case yourself if you believe it. Marking 1454: 1435: 16639:

section, I'm not too worried that the references aren't "inline" and require some syntax formatting to not have it summarily deleted. —

15809: 11995: 11565: 11036: 10087: 8436:

Despite some support for the idea, clear consensus holds that banners advertising sister projects should not be run on the front page. —

8186:

be related to editor's problem. Reading the policy/ guideline page and deciding next move is solely editor's responsibility and choice.

8110:

be related to editor's problem. Reading the policy/ guideline page and deciding next move is solely editor's responsibility and choice.

5113: 5102: 5087: 5046: 5031: 5019: 4996: 2197:

also noting that I will support introducing FA and GA topicons to mobile and apps, but not supporting the entire removal of the badges.

2076: 2039: 1896: 16232: 14148: 13572: 12293: 12277: 12254: 12104: 12088: 10167: 5822: 5802: 5776: 5762: 5740: 5726: 5143: 4986:

Maybe ... while changing the arbcom policy requires those hoops, changing the cu/os polices don't really. This would really need to be

4981: 4966: 3753:

With regarding salted pages (see red link example for an example), the lock could be displayed within the infobox or even the top right

16682: 16663: 12333: 9831: 5905: 4939: 4935: 4285: 4254: 16539: 15784: 15543: 14405:, and nowhere does it say that deletion of the article is the appropriate course of action for verifiable but not verified material. 13761:. From above, it appears that newly created unsourced articles are already sufficiently handled by our existing rules and processes. 13606: 12309: 12074: 9823:

about the appropriateness of displaying the banner for Ukraine's Cultural Diplomacy Month globally without any geographic targeting.

6099:

Definitely "in addition", rather than "instead of". At the very least, this would help editors whose muscle memories have them press

5882: 13690: 12045: 11478:

Another idea would be to use light gray rounded rectangles, because rounded rectangles are a very common part of modern web design.

10521:

I think that it’s probably best to make changes one by one, so that consensus would be more likely. Like one change per discussion.

10095: 5336:

I think it is a good idea, and it seems more user-friendly. It draws more attention to itself, but I don't think that is a problem.

3001: 2971: 1716: 14039:

I agree with DonaldAlbury above: the point of having one source, no matter if it's not reliable, or trivial, is to ensure that the

13781:

and quite a jump from settled practice. Second, is is not clear what it means for the affected articles to be "eligible" for PROD.

13751: 13126: 9586: 6371: 6198: 5138:

This is a tangent, but if any of you are willing to take on these roles, please consider volunteering. I am worried that whenever

2475: 1367:

The top-icons lack context. I'm in favor of removing them or improving the top-icons to something that make more sense to readers.

9357:

that the reader has already shown themselves interested in, hence them being at an article in the first place. So something like

8762:

Is it your intention to get into long-winded discussions with everyone who participates here, or just the ones you disagree with?

3018: 2487:

per Voorts. Acknowledging the lack of professionalism at WikiProject Good Articles, the icons at top mean something. I'd suggest

15852:

My bad - I should have implied "substantially similar", not "identical". On which, I'll just quote a brief comment from up above:

14947:

It doesn't matter if they were used or not, if the sources are not listed in the draft, the draft is not helpful to anyone else.

13158:

easier to understand for new editors. But we need new editors and we MUST NOT make Knowledge harder for newcomers to contribute.

9800:

One could look at siteviews of projects, number of edits made, number of active users... and see if there was noticeable change.

7781: 7663:

I don't think that things shouldn't be done because they are hard to do. I just personally think that it's not worth it. I'd say

6591: 6522: 6460: 5585:

I think the "every Monday and Friday" pattern makes it easier for people to know when to look for it than your proposal would. --

2988:

and changed the links in the template accordingly, so after a month or so we should have good data on how many clicks these get.

1226:

who had hundreds of GAs that turned out to be full of misinformation (and copyvio and other stuff, the editor ended up indef'd).

16081:

I wouldn't personally want to emulate this approach, but it is something to think about as a potential source of complications.

15648:, but why is it so hard to ask people just cite one of those excellent sources to the first sentence, to verify that this topic 14909:

Isn't the reason to draftify an article exactly so that it can be brought up to mainspace standards before living in mainspace?

13787:

the page is not a redirect, never previously proposed for deletion, never undeleted, and never subject to a deletion discussion.

12992:

Number two is extremely difficult. Anyone who knows English could be a new editor. There are many hundreds of millions of them.

11348:

Can't we add padding to the bottom of TFA, like ptwiki seems to do? Also the manual alignment doesn't seem to be working for me

10558:

encyclopedia and good work on our behalf. Removing it would make no sense at all. It has worked, it is working, and it returns.

3335:

The only way to see the green (move) and blue (create/also known as salting) lock is if the URL was typed for some reason, like

3113:

I have occasionally told people who have questions about whether/when Knowledge articles are reliable to look for the topicons.

2213:. Do not see any positive from this action. On the contrary, will be some amount of value subtraction (if that is even a term). 1609:

Why? I think it's a convenient way to show that articles are of a decent quality to our readers. It's not causing any problems.

13029: 10315:

As Joe Roe said, Wikipedians are usually a little less open to change than a cube of iron. But do you have any specific ideas?

4077:

you think an RFA-like process would have produced a different result. I will also follow-up your answer regarding me by asking

67: 16408:

Do we need to increase the time limit that articles can sit in draftspace … so NPP can work through their perennial backlogs?

14842:

where alternatives before deletion have to be considered before moving on to the deletion. This proposal also didn't consider

14662:

I wish that the opposers would detail more about what this proposal is missing rather than just regurgitating talking points.

11604:

We're currently kinda designing some layouts, so a more formal workshopping process should begin when we have more proposals.

11602:

I don't think expanding coverage of RD entries is needed much, and the POTD often needs a dedicated row due to its image size.

7181:

Yes, AI optimized search tool is the first and foremost thing I proposed. Other things can be discussed within the community.

5682:

for providing citations. I propose amending this with the additional sentence "Sources may be contained in a linked article."

5005:

states that local elections can still be preferred over appointment by the arbitration committee, and it could be argued that

14208: 13791: 13595:

and move on. Others might choose to bat it back and forth to draftspace a few times first, but the end result is the same. –

12798: 12725:

It really was a great tool for creating, in seconds, a PDF book of topics you liked including an index, page numbers and all.

12616: 10439:

With all due respect, the gradients in the Slovakian one make it look very dated, like an early-2000s PowerPoint template. --

9860: 9718:

Adding 'sister project' banners is inevitably going to make other content less likely to be seen, given finite screen sizes.

9147: 8853:

in the way of what they came for: verifiable facts about a topic. We aren't here to promote anything, including ourselves.

6963:

The risks of AI misinterpreting, hallucinating, or otherwise giving users inaccurate information about policy is non-trivial.

6392: 5501: 1913: 16426:, but I noted at the time that the specific figure of 90 days was just a starting point that can and should be revisited. – 15798:

are very many completely unsourced new articles. The pervasive problem is lack of GNG sources for articles that need those.

12572:

is the least used link in that part of the sidebar). Incidentally, the correct place to suggest changes to the Main Page is

11197:

Is it possible to have, say, the current main page visible to users with the Vector legacy skin, but a main page similar to

8614:, where a special "featured articles" modality was, however, discerned as a likely prerequisite for all of this though...) 4484:

The last attempt at community-based CUOS appointments was a total failure, and the current system is more than good enough.

87: 14624:, I think all of these concerns are very valid and thus this proposal is not complete. Should I move this to the idea lab? 14366:: "All quotations, and any material whose verifiability has been challenged or is likely to be challenged, must include an 12831: