Community Wishlist/Brainstorming

This page is for brainstorming Community Wishlist wishes before submission.

Digital assistant

This is somewhat covered by the wish Build a Wikipedia search that addresses common queries asked by new volunteers but I think the title "Wikipedia search" could be confused with article search and excludes other Wikimedia projects. Also, there is no need to target this feature at new volunteers.

Problem

It is difficult to find help documentation when contributing to a Wikimedia project. Users turn to mentors and help desks unnecessarily.

Possible solution

We need a search tool that prioitises, in order: FAQs, Help: namespace, Project: namespace, Help desk archives, Village pump archives. I am not tied to the name "Digital assistant".

This is particularly important for new users who wouldn't have to bother mentors and help desks as often, and it always faster to look something up yourself. Experienced users would benefit also.

A working search tool to start with would be great (see example below for why the current search tool is terrible, at least on English Wikipedia). I later envisage a "smart" digital assistant. In the "insert photo" example below it could detect that you recently uploaded a photo and suggest steps to add it to an article using visual editor, or show you the syntax to use in source mode. If you have recently tripped an abuse filter the assistant can offer advice.

I would place the Digital assistant search box at the top of help pages, help desks, in welcome messages and accessible from the Dashboard (basically a link replacing the watchlist that has favourites and a Digital assistant search above the watchlist).

Simulated user experience

I ran through an example on this wish talk but I will repeat it here.

For example say you want to insert a photo into an article. Where do you ask? Somehow you navigate to w:Help:Contents (note there is currently no help link in the mobile interface). Then you use the given search box. "insert photo" restricts search to Wikipedia and Help namespaces. The first five results (after to advice to start an insert photo page) are an RfC on a controversial photo, a University of Dayton Wikiproject, Reference desk archive, Mediation Cabal case and an Articles for Creation Help desk post from 10 years ago . On the first page of results there isn't a single link to a FAQ page or anything in the Help: namespace.

So you find the FAQ, it has a search. One result: answers an "On this day..." event order question. I guess you give up on Wikipedia altogether or go to the w:WP:Help desk.

Google's AI overview for a search of "wikipedia how to insert a photo" is excellent at answering this question by the way. It seems to use an instruction set from a non-Wikimedia website.

Also, for advanced editors, when I want to look something up like blocking policy on English Wikipedia I generally just go to the browser bar and type "WP:BLOCK". With the assistant I imagine even a search for the typo "BLOK" will turn up something useful.

Any input is welcome.--Commander Keane (talk) 05:16, 14 March 2025 (UTC)

DiscussionTools

This could be mobile web specific. When you reply to a comment using DiscussionTools it will have in the grey input box "Replying to User123". I don't actually want to reply to that user, I just want to add a new comment that is indented so it doesn't get confused with the comment directly above. Does the "Replying to User123" serve a notification purpose? Is there a way to add a new comment at the bottom? Help:DiscussionTools doesn't clear this up for me.

Also, is it possible to add a new comment at the bottom of a thread using a dot point or resetting the indent with DiscussionTools. Here I just wanted to add a new dot point at the bottom, but I had to edit the entire section in wikitext mode to do so (thus losing the instant preview of DiscussionTools).

Finally, this may be an editor behavioural issue but linked to the "reply to" button implementation, is there any reason why editors (on enwp and commons at least - and I am guilty) reply mid way up a discussion thread instead of at the bottom. Without being there for every new comment I have to read the timestamps to work out the order of posting and separate threads of discussion can appear.--Commander Keane (talk) 05:16, 14 March 2025 (UTC)

Wikimedia-wide preferences

I would greatly appreciate the ability to set Preferences globally for all Wikimedia projects. This is so that:

• I don't have to switch on "Advanced mode" for every project I visit on mobile. I not sure why "Normal" mode exists for logged in users at all, there is no person icon in the top right to access your talk page, contribs etc. (possibly submit as independent wish)
• I can have a consistent skin when visiting other projects. The default Vector 2022 is bad on mobile so I gave to change it per project. A workaround for this would be specifying a dedicated desktop skin when you are on a mobile device, or improving the mobile interface so that desktop mode is redundant.
• I just found out that while Edit recovery was switched on for English Wikipedia it was off on meta and Commons. A global setting would be handy.
• I would like to switch on email notifications for pings globally for intermittent periods.

Although I have no real need for project independent preferences, some advanced users will, so the existing behaviour will have to remain. I also think global preferences make more sense for new editors engaging in multi-wiki activities. Commander Keane (talk) 07:09, 14 March 2025 (UTC)

Isn't there already Special:GlobalPreferences? Prototyperspective (talk) 21:43, 16 March 2025 (UTC)

Thanks that is fantastic, I wasn't aware of that special page. A note at the top of each individual Preferences page would be handy to point this feature out. Special:MobileOptions needs a Special:GlobalMobileOptions for the advanced mode, expanding of sections etc though, unless I missing that page also. Really I am not sure why Mobile has its own options, they could be incorporated into Preferences. Commander Keane (talk) 23:02, 16 March 2025 (UTC)

I would have suggested to create a phab issue about displaying a note at the top but apparently there already is "Global preferences: (set your global preferences)" in the User profile preferences page (the initial preferences page) so that may suffice. Prototyperspective (talk) 09:40, 17 March 2025 (UTC)

Ahh ok. On mobile the first preferences page is a list of buttons to User profile, Appearance, Editing etc. The Global preferences button is in the User profile section which I have never visited (my email and signature are fine as is). I can see how new users would gravitate to the User profile section, so it may not be a big issue. More knowledge is better though if we want people to have a pleasant cross-wiki experience. I can imagine a Wikipedia user being upset that their Commons upload was deleted and they thought they had turned on email watchlist and talk page notifications. Thanks for investigating :-). Commander Keane (talk) 11:05, 17 March 2025 (UTC)

An awesome categorisation system for Commons

Problems with current categorisation system on Commons

• Language support. It is exclusively in English, which includes navigation and inserting/modifying everying in English. There are disputes over category names.
• Structure. Due to subcats everything gets tucked away so when you visit a broad subject you have to dig into layers of categories to find something useful. Wall of images view for category pages including images in subcats was a highly popular feature request.
• Constant tedious restructuring of category trees. When a category grows too large it has to be split: geographically, by date, by filetype etc. This is done by changing each file page.

There are numerous other downfalls (please suggest them if you have time).

Solutions

I think Structured Data on Commons (SDC) could be utilized to solve many of these problems and supersede the current categories system. It uses Wikidata's internationalisation (and same WikiBase software).

However, some of the current issues will be transferred to SDC. I am unsure about the relationship between search and categorisation for discovery of files, or how the implementation of human defined category structures will interact with SDC.

Note that current wishes Add machine translated category titles on WMC (a patch to the current category system) and Add a date range filter to Special:MediaSearch (which mentions the inadequacy of the Query Service that will surely need to be improved for SDC to succeed) are related to this wish. Commander Keane (talk) 07:46, 14 March 2025 (UTC)

SD is inferior to categories in many ways. Maybe 0.2% of files have some SD set and when it's set, it's usually not all that is displayed in the file and likely just got copied from the categories which are obviously far better-maintained and populated. SD also don't have a category page and duplicate work already done. Categories are great and should be developed further. SD again is a pipedream ideology that never worked well and it if it ever works then just for things that bots write to the files. Prototyperspective (talk) 21:40, 16 March 2025 (UTC)

Categories are useful for people from within Wikimedia (maybe just Commons), SDC would make searching useful for people outside, and consequently would encourage image re-use.

I am working in computer vision (AI) research, and there was numerous times already when Wikimedia Commons would be the first place to go to... if it offered a good search engine / query service. This is something SDC was designed for. Unfortunately, the adoption is poor, and some WMF experiments with edit microtasks surely did not contribute in favor of it.

Still, sometimes it feels like the WMC people completely forgot they are maintaining repository of files (mostly images), not concepts. On one hand, they are overburdened with licensing/legal stuff, and at some point the database was getting closer to a crash. On the other, they spend resources on providing me with information that this signature origins from someone who died from pneumonia in 1938 and his first names included Karel. Even though it is unrelated to the image(s). Simple "is: signature (Q188675), made by: Karel Čapek (Q155855), text: 'Karel Čapek'" is enough. Structured data at its finest.

Or how do I extract all photographs of Karel Čapek himself from c:Category:Karel Čapek? It's full of artworks, recordings, signatures, etc. I could heuristically filter out extensions like .svg, but I will still be left with JPEG's that need to be assessed by a sophisticated method or manually. Why should I bother?

Sad to see the wasted potential... --Matěj Suchánek (talk) 13:04, 17 March 2025 (UTC)

Categories are useful for people from within Wikimedia (maybe just Commons), Don't know what you mean – categories are there for everybody to use. SDC would make searching useful for people outside, and consequently would encourage image re-use As do categories....but what do you mean with that, SDC doesn't do that. if it offered a good search engine / query service Commons categories can be queried. Secondly, the MediaSearch is great and I've suggested a set of improvements for it on the Commons talk page about it. Simple "is: signature (Q188675), made by: Karel Čapek (Q155855), text: 'Karel Čapek'" is enough. Structured data at its finest. Nobody sets such SD, unlike categories. In contrast, you can use petscan or deepcategory:"Signatures of people by country" to achieve exactly what you like. Once again SD is shown to be a redundant, misleading, time-sink. Or how do I extract all photographs of Karel Čapek himself from c:Category:Karel Čapek? Using deepcategory (or using categories if you're querying). Secondly, by making excluding these simple by implementing what I suggested here: phab:T376440#10354943. And third, most people categories that contain many pics have subcategories that contain just the photos such as c:Category:Donald Trump by year. I hope I addressed all points you raised and cleared things up. Prototyperspective (talk) 15:19, 17 March 2025 (UTC)

Thanks for trying to clear things up, but I am well aware of the current options.

what do you mean with that, SDC doesn't do that Wikimedia Commons Query Service is there, although still in beta.

Nobody sets such SD, unlike categories. And this is my point. See what Wikidata has achieved without categories. And it's supposed to cover the whole universe. Commons needs to deal "only" with its images, and everything entered into Wikidata is available for Commons, too.

you can use petscan or deepcategory:"Signatures of people by country" to achieve exactly what you like. deepcategory has suboptimal performance and needs artificial limits because of the risk of users inducing cycles and potential overflow during the tree search. [A] good search engine should be scalable and performant. From an "outsider's" POV, deepcategory:"Signatures of people by country" is obscure. Why not just deepcategory:"Signatures"? On the other hand, "depicts: a signature" – simple, understandable.

most people categories that contain many pics have subcategories See what effort all the categorization takes? And I wanted it for Karel Čapek, I shouldn't need to care if it contain[s] many pics.

By the way, MediaSearch is just as useless as deepcategory:"Karel Čapek". I wanted photographs of Karel Čapek himself, that is, depicts (P180): Karel Čapek (Q155855). Everything else is basically DIY.

Just thinking outside the box... --Matěj Suchánek (talk) 17:20, 17 March 2025 (UTC)

• First sentence was not addressed, just for those reading
• Wikimedia Commons Query Service is there, although still in beta I know and have used it. Still, regarding "SDC would make searching useful for people outside, and consequently would encourage image re-use", SD doesn't do that.
• See what Wikidata has achieved without categories Wikidata has so far achieved not much. I'm trying to change it a bit but so far nearly nobody knows of the site and nearly nobody uses it. (This is in large part due to a lack of bulk imports of data that people actually use.) Unlike Commons, which is used at least somewhat and has clear value in terms of applications/usefulness.
• And it's supposed to cover the whole universe. I understand you apply some Wikidataish kind of thinking to everything, being very active on that site. Law of the instrument: "it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail"; this also applies to the way of thinking and activity one deploys but problems Commons are no nails and developing ways to improve it also are not nails.
• You did not address the point "Nobody sets such SD, unlike categories", just for those reading
• deepcategory has suboptimal performance and needs artificial limits Then it should be improved. However, I think those limits or some of them have been removed by now. In any case you could use this for your example where it works perfectly.
• From an "outsider's" POV, deepcategory:"Signatures of people by country" is obscure Couldn't say it any better – once again, this should be made more accessible.
• Why not just deepcategory:"Signatures"? You could also use that.
• On the other hand, "depicts: a signature" – simple, understandable. One can also use "category: a signature" which is just as simple; it's just the interface. Moreover, other kinds of SD is more complex and the users don't know how something particular is specified while for categories it's uniform and always the same via the category.
• See what effort all the categorization takes? Please now take a moment to consider the time and effort required to duplicate what has been laboriously set in countless hours of volunteer contributions via categories already by redundantly also setting things in SD. Please really take a moment to think about this given you are working toward dumping this mountain of time-expenditure on scarce volunteers.
• I shouldn't need to care if it contain[s] many pics. Then instead of complaining, create the subcategory. Alternatively, realize that this is actually not an important issue – people can use deepcategory on that cat to just scroll through all pics in the category or with 2 or use 2 or so clicks to go through the few subcats. The few seconds extra time spent to find a photo are a triviality and if there is any reason for why a photos subcat is needed, it would get created.
• Several people think SD is some sort of panacea but ultimately it just impedes practical realistic solutions.

Prototyperspective (talk) 18:29, 17 March 2025 (UTC)

Several people think SD is some sort of panacea but ultimately it just impedes practical realistic solutions.

I do not say that, and I disagree about it impeding anything, even in the least. I try to be as realistic as possible. But I also like to present ideas which challenge current habits.

Anyways, the topic we are discussing is named "An awesome categorisation system for Commons". Commander Keane outlines well the problems, I add a user story. And claim that SDC is actually awesome and its proper adoption is a solution.

As for me complaining, the whole wishlist thing is about complaining. I'm cannot do something (or I can but with much effort), and I wish I could. In particular, I wish something like [1] was actually useful because this is a real query service. (And yes, now it isn't useful, but I ask for proper adoption after all.) Furthermore, I want this to work instantly for many concepts, hence any additional effort spent will always be an important issue.

I would never participate in what you claim pushing SDC forward is. (Quite recently, I was called "sentimental" because I opposed careless mass deletion with no attempt to re-use.) In fact, I dedicate most time here not to making content, but to making sure others' work does not degrade, their time is not wasted and is used as efficiently as possible. Better structured content will improve image search, SEO and perhaps finally make Commons the number one place-to-go for images (like Wikipedia for facts), increasing image re-use and giving the effort spent by the scarce volunteers wider recognition. Hopefully.

In the end, it's up to the Commons community if they consider this a real problem. Perhaps not now. By the way, thank you very much for taking your time and replying.

--Matěj Suchánek (talk) 22:09, 17 March 2025 (UTC)

Sure, challenging present ideas and thinking outside the box is good but this isn't the silver bullet that a few people portray it to be. Maybe SD has some use for meta-data like whether or not the image is black-and-white, whether it has audio, etc but for the contents I think I explained most already. I should repeat that it would duplicate work and that SD does not have category page. And also I should add that depicts (P180): Karel Čapek (Q155855) wouldn't even actually work (unlike deepcategory) because depicts like that are also added to artworks, not just photos. The part about complaining was not meant this literal, it's more like this problem can be fixed in a straightforward easy way with no need for super-laborious infeasible remakings reforms. Good that you linked the query which should show you that the SD query result isn't an actual improvement over visiting the category (quite the opposite). Thanks nevertheless for taking the time to think about how to improve the current state of things. Prototyperspective (talk) 22:34, 17 March 2025 (UTC)

In this regard, I recommend reading w:Wikipedia:Wikipedia Signpost/2025-03-22/News and notes#Interview with WMF's CPTO, especially her thoughts on Wikimedia Commons. --Matěj Suchánek (talk) 08:22, 22 March 2025 (UTC)

Interesting link but I think it's something of general relevance, not specific/related to the subject of this thread.
She wrote There's not as many videos honestly, but a lot of images, really incredible ones that's just false, there are also many videos including pretty incredible ones. Over 100 k videos from YouTube alone, and Google & DuckDuckGo apparently censor them from their Videos tab. One example category of high-quality videos is this but there's also many other ones (btw video2commons used to import nearly all of these seems currently broken and WMF is not taking care of it). the primary focus of the Commons community is the collecting of free content, rather than its dissemination. Yes, there is not much consideration so far about whether the platform is actually used, useful and known – a top issue in that regard is I think how it's indexed in Web search engines (see the proposal I just linked).
I think there's a lot that could be done if we were to think about this system holistically and the ways in which they, the different projects complement or don't complement each other. and How much more powerful and effective would it be if we were really thinking hard about the ways in which we could showcase all of the incredible images Agree so I recommend really looking into Community Wishlist/Wishes/Suggest media set in Wikidata items for their Wikipedia articles.

• And please also see the discussion at c:Commons talk:Media knowledge beyond Wikipedia#Response from the Selena Deckelmann, CPTO at the Wikimedia Foundation

Thanks for bringing this up. Prototyperspective (talk) 16:16, 22 March 2025 (UTC)

Thank you for having me look at that. It's a very interesting reading as I see some of "my" ideas surfacing there. --Matěj Suchánek (talk) 20:51, 22 March 2025 (UTC)

"Current files in use" link

A bit from left field, but I think the addition of a MediaWiki page that lists all files in use on a page would be useful.

Currently at the bottom of English Wikipedia articles on mobile it says "Content is available under CC BY-SA 4.0 unless otherwise noted" but the licensing of the files (some may be full copyright, some CC BY, some FAL etc) is not actually noted. This may not be meeting legal obligations and it certainly isn't ethical to treat reusers and uploaders like this.

On desktop the message says "Text is available under the Creative Commons Attribution-ShareAlike 4.0 License; additional terms may apply" which is more accurate but leaves ambiguity over the licensing of media.

I propose to change the text to include a link to a page displaying the files with the licence and author alongside. Something like "Text is available under CC BY-SA 4.0, click here for media details".

I would like the page to include the captions used but that is just for editorial benefits.

This page would also list files used in site banners, templates etc. It would allow wikis to use non-clickable media while still meeting licence terms (eg the annual Picture of the Year banner on Commons has an awkward info link on it). Commander Keane (talk) 00:29, 15 March 2025 (UTC)

Mobile first

This is a collection of experiences on Mobile web that are deficient to Desktop. Individual wishes could be filed if needed.

• Ability to easily check the status of a protected page (the edit button is an unclickable pencil with padlock symbol, and a popup says "this page is protected...[view source]"), a link to the log would be good
• phab:T358733. A working button to expand all sections. On mobile, all sections are collapsed by default and while you can expand all using preferences it would be better to have an expand/collaspe button
• nowiki tag insert button, under bold/italic for DiscussionTools
• ...

Commander Keane (talk) 11:16, 19 March 2025 (UTC)

• Categories can't be seen (and visited) – they're still hidden on mobile phab:T24660
• No way to reach for example the pageviews of a page or its Wikidata item since the Tools button isn't present on mobile

Prototyperspective (talk) 19:21, 19 March 2025 (UTC)