Commons:Village pump/Technical

Shortcuts: COM:VP/T • COM:VPT

• Community portal
  • introduction
• Help desk
• Village pump
  • copyright
  • proposals
  • technical
• Administrators' noticeboard
  • vandalism
  • user problems
  • blocks and protections

Welcome to the Village pump technical section

Technical discussion

Village pump/Technical

Bug reports

Code review

Tools

Tools/Directory

Idea Lab

This page is used for technical questions relating to the tools, gadgets, or other technical issues about Commons; it is distinguished from the main Village pump, which handles community-wide discussion of all kinds. The page may also be used to advertise significant discussions taking place elsewhere, such as on the talk page of a Commons policy. Recent sections with no replies for 30 days and sections tagged with {{Section resolved|1=--~~~~}} may be archived; for old discussions, see the archives; recent archives: /Archive/2026/01 /Archive/2026/02.

COMMONS DISCUSSION PAGES

• Commons Help desk
• Village pump (general discussion)
  • Copyright
  • Proposals
  • Technical
• Graphics and photography discussion
  • Photography critiques
  • Image improvement
    • Illustration workshop
    • Map workshop
    • Photography workshop
    • Video and sound workshop
• Categories for discussion
• Undeletion requests
• Deletion requests
• Volunteer Response Team/Noticeboard
• Translators' noticeboard
• Work requests for bots

---

• Contact administrators
  • Vandalism
  • User problems (Dispute resolution)
  • Blocks and protections
• Bureaucrats' noticeboard
• CheckUser requests
• Oversight requests

---

• Telegram
• IRC ^webchat
• Commons mailing list (archive)
• Commons' bugs on Phabricator

Please note

• Feature or bug reports should be filed on Phabricator (see how to report a bug). Bugs with security implications should be reported differently (see how to report security bugs).
• Have you read the FAQ?
  
  

 

Start a new discussion

	SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 1 day and sections whose most recent comment is older than 30 days.

Help with changing links based on display language

On this one File:Great_Mandala_(大曼荼羅)_of_Nichiren_Buddhism.svg I want to have language switching so that in japanese it links to all of the Japanese articles, english to the currently linked articles, etc. How would I do that? Immanuelle ❤️💚💙 (please tag me) 19:12, 20 January 2026 (UTC)

@Immanuelle: To have what link to pages in different languages? Something within the SVG, or something on the file page, or what? - Jmabel ! talk 01:12, 21 January 2026 (UTC)

@Jmabel the clickable links in the svg Immanuelle ❤️💚💙 (please tag me) 01:18, 21 January 2026 (UTC)

@Immanuelle: Those are not in the SVG. Those are just ImageNotes (which contain Wikitext). You can do pretty much anything inside an ImageNote that you can do anywhere else on a page in Commons, and you do it exactly the same way. - Jmabel ! talk 01:22, 21 January 2026 (UTC)

@Jmabel no they are separate from the svg. Go here https://upload.wikimedia.org/wikipedia/commons/4/40/Great_Mandala_%28大曼荼羅%29_of_Nichiren_Buddhism.svg and try clicking on anything in the image. Immanuelle ❤️💚💙 (please tag me) 01:44, 21 January 2026 (UTC)

Oh, interesting, the two ImageNotes (which were what I instantly noticed as clickable) can serve something of the same function, which is why I was confused. I know there are ways to do multilingual SVGs, and I imagine the issues are the same for links in the SVG as for text, but I'm out of my depth there. You might ask one or more of the people who participated in #How to specify a SVG file's default language? above, which is clearly a closely related issue. - Jmabel ! talk 01:50, 21 January 2026 (UTC)

For links directly in the SVG when viewed directly, you can use the <switch> element. However this will use the browser's language which is sometimes different than the language the user expects (And different from whatever site they are viewing). If you want to use this in Wikipedia, I would suggest using an Image map instead. Bawolff (talk) 23:35, 21 January 2026 (UTC)

@Immanuelle:

One option is to link to Wikidata items instead of wiki articles. For example, instead of linking to https://en.wikipedia.org/wiki/R%C4%81gar%C4%81ja, link to its Wikidata item: Rāgarāja (Q1188900). The en.Wiki article will give you the link on the right-hand side. The user can then scroll down to the bottom of the Wikidata page and click a link to his favorite wiki.

A more involved option is a URL that takes the Wikidata item and the HTTP Accept-Language header. The URL would then redirect the user to the user's preferred wiki article if it exists. If it doesn't exist, then the URL would redirect to the Wikidata item.

Glrx (talk) 04:25, 29 January 2026 (UTC)

@Glrx do you know of any files that do that more involved option? Immanuelle ❤️💚💙 (please tag me) 05:40, 29 January 2026 (UTC)

@Immanuelle: No, but the idea has been raised before. The URL needs to be written. It could be on toolforge or a Wikidata endpoint. Glrx (talk) 16:36, 29 January 2026 (UTC)

Blocked for too many requests

Hi, I have a VBA script that I use to download the photos, in practice now usually a recent subset of them, that I have uploaded to Wiki Commons. (I won't go into the reason why I want to download photos that I have already uploaded, but there is a reason.) This used to (say a year or two ago) work perfectly even for hundreds and hundreds of images. Now it struggles to do a dozen or so, before I am blocked, I believe for making too many requests in too short a time. Then eventually it will start working again and let me do a few more, slowing down to a trickle. I suppose this blocking must be a feature recently introduced? I have never been blocked from browser access, however. Even immediately after the script being blocked, I can open pages in the browser (from the same IP address). Anyway, I can try adding pauses between downloads. I've tried five or ten second pauses but it seems to make little difference so far. Anyway, I wouldn't have thought my volume was particularly unreasonable on the scale of server traffic generally. Anyone got any info about what is and isn't permitted, or what I can do to mitigate this issue? Thank you. ITookSomePhotos (talk) 22:02, 21 January 2026 (UTC)

See foundation:Policy:Wikimedia_Foundation_User-Agent_Policy and wikitech:Robot_policy. Unfortunately they have become much more strict with this recently. Bawolff (talk) 23:30, 21 January 2026 (UTC)

Thanks Bawolff, that's useful. ITookSomePhotos (talk) 10:10, 22 January 2026 (UTC)

Do you download the original images or thumbnails? If the latter, you may be affected by mw:Common thumbnail sizes as well. —Tacsipacsi (talk) 21:32, 1 February 2026 (UTC)

Original images. I have found that a thirty second pause between downloads (typically 3-5MB each) helps to prevent rapid blocking, but this does not work indefinitely. Eventually I still get blocked. When this happens, I am now looking at and honouring the "retry after" value, per the documents that Bawolff linked to. I have never seen any value other than 1000 (seconds). Even waiting this long, or even a bit longer for good measure, does not guarantee success. I have seen at least three of these in a row, i.e. total 50 minutes' wait. It seems a bit extreme, given that I am, with 30 second gaps, not putting any more load on the server than normal browsing though a browser. ITookSomePhotos (talk) 18:58, 2 February 2026 (UTC)

How to categorize pages using the OWIDslider gadget?

• Nearly all files in Category:Our World in Data have been meticolously categorized into the cat's subcategories, subcategories of Category:Charts by year of latest data & Category:Maps by year, and Category:Statistics by field / Category:Data graphics by subject (better enabling people to browse, update & find these)
• Since relatively recently, there now is a tool – or more use of that tool – which allows the upload of hundreds of OWID files at once (which is great!)
• There also now is a gadget that uses these hundreds of files to allow semi-interactive viewing, the OWID Gadget (also great!)
• The semi-interactive data visualization uses a page on Commons and all of these pages can be found in Category:Pages using gadget owidslider
• However, categories are missing for both the files and those pages. Now I'd like to also categorize these; I was thinking:
  • The files belonging to one visualization (one topic) should be in one category which can then be categorized as described above and get a charts and a maps subcategory – this would best be done at upload via the OWIDUploader tool. This is especially so because there usually are hundreds of files with one per country which each should go into Category:Charts by country. This is why I'd only lake to do the next things and hope more things get done at User talk:Jmh649/2#OWID uploads with false and missing categories, including for files already uploaded. Any updates on this @Doc James: ? (e.g. I saw Category:Our World in Data graphs by country has been created but I still often come across uncategorized OWID files)
  • The interactive visualizations (example below) should also be categorized into categories like e.g. Category:Meat statistics. Would the best way for that be to categorize pages in Category:Pages using gadget owidslider (indexed also in Commons:List of interactive data graphics)? If so, how could one comprehensively categorize these (e.g. see which of these haven't yet been categorized into a topical cat) and isn't it a problem that the page with so many files loads quite long and incomplete? If not, what would be the better way – what about creating some kind of separate page with just the visualization which is ready to use, loads quickly, and can more easily be categorized (e.g. a page like OWID: Daily meat consumption per person in the meat statistics cat instead of Template:OWID/daily meat consumption per person)?

daily meat consumption per person

Prototyperspective (talk) 18:42, 22 January 2026 (UTC)

Internet is poor currently. Jmh649 is my old old account. Stuff posted there will not get to me. Will look more when internet is better. Doc James (talk · contribs · email) 06:33, 23 January 2026 (UTC)

I posted this on your normal talk page. This is just the name of your archive page which is in your responsibility to rename or keep as is. Okay.

There are many examples of these files missing categories, here's a recent one I came across which may also elucidate the complexities involved here (e.g. do we want these hundreds of files to show up in deepcategory wall-of-images views or somehow separate them?):

• intitle:Europe -deepcategory:"Our World in Data maps of the world" deepcategory:"Our World in Data maps" -incategory:"SVG maps of Europe by Our World in Data" when filtering for SVG files shows 22,244 results – these are all missing in Category:SVG maps of Europe by Our World in Data which only has about 400

Regarding separation or enabling the user to exclude the hundreds of files from OWID mass-uploads, maybe it would be good to better make sure files are put into Category:Uploaded by OWID importer tool so the user can exclude these and maybe add a note about this at the relevant places. I think views/scans usually should only show one or two page/file per set – a page with the interactive data graphic. Prototyperspective (talk) 12:20, 23 January 2026 (UTC)

All the images we upload with the tool should be in the categroyr:Uploaded by OWID importer tool. We can run bots to add specific images to categories. For example see this https://commons.wikimedia.org/wiki/Category:Our_World_in_Data_graphs_by_country and https://commons.wikimedia.org/wiki/Category:Our_World_in_Data_maps_by_continent.

I am thinking it might be easier to do more categorization after the upload rather than during the upload as these uploads are complicated. Doc James (talk · contribs · email) 07:22, 27 January 2026 (UTC)

With respect to OWID: this is a namespace and would require permission to put content in it. If you can request and get consensus for that would be happy to put these all there by bot.

Per here have uploaded about 211,000 of these images so far. https://commons.wikimedia.org/wiki/Category:Uploaded_by_OWID_importer_tool Doc James (talk · contribs · email) 07:28, 27 January 2026 (UTC)

So that graphs by country cat is populated by a bot; if so that sounds good – then there's less need to set the cat at time of upload and it could explain why you say "it might be easier to do more categorization after the upload rather than during". With respect to OWID: this is a namespace Don't understand what you mean. Basically the main issue of the thread is that the files are missing in the topic categories like Category:Meat statistics or Category:Renewable energy statistics and because it could clutter the pages and the category's search results, that may be good but the interactive visualization should be in these categories – so for example a page OWID: Daily meat consumption per person that loads quickly and contains just the interactive visualization embedded above. The OWID there is not the namespace but the page title (just standardized to have the OWID prefix) so it would be in the "Main"/gallery namespace in this case. If this is a good/best solution to that issue, then it would still need all those visualizations indexed (mostly or entirely?) in Category:Pages using gadget owidslider to get one of such page each. Prototyperspective (talk) 18:47, 27 January 2026 (UTC)

All the pages with the OWID visualizations are listed here Category:Pages_using_gadget_owidslider About 500 of them.

Please read this about naming conventions for namespaces Help:Namespaces Doc James (talk · contribs · email) 23:50, 27 January 2026 (UTC)

Could you please address the questions – I linked that category in the comment you're replying to. Please see the part around and isn't it a problem… above. I thought it would be a good idea to discuss this instead of me figuring out myself what the best solution would be and then unilaterally implementing whatever it is without discussion but this is not much of a discussion. If you were saying you think a new separate namespace would be a good approach, that's not clear and I don't understand why that's needed or would be better than e.g. to create the mentioned kind of mainspace pages. Prototyperspective (talk) 00:09, 28 January 2026 (UTC)

What I am saying is we cannot use the title "OWID:Text" unless we get approval from the community at large as that is a new namespace.

I personally think the Category:Pages_using_gadget_owidslider is sufficient for me to find all these. If you want just the interactive graphs placed on a separate page we could do that, and would just need a naming structure.

What we have currently is here Template:OWID/daily_meat_consumption_per_person and we could have just the interactive graphs at Template:Daily_meat_consumption_per_person but not sure we need this. Doc James (talk · contribs · email) 01:22, 28 January 2026 (UTC)

Thanks for clarifying. Okay so it can't be colons in the prefix as that for pages in the "main"/gallery namespace indicates/requires a new namespace. One could also have the prefix be OWID – or OWID/ for example – the particular suggestion was just an idea/illustration.

I personally think the Category:Pages_using_gadget_owidslider is sufficient for me to find all these as elaborated above, the issue is making these findable for those looking for statistics of a certain topic via either browsing the associated category page or the search. Just having them in that category doesn't achieve that – it's about surfacing these visualizations to people who could be interested in them / to which these are useful.

If you want just the interactive graphs placed on a separate page we could do that do you know if there's a way to do that for all the pages in the Pages_using_gadget_owidslider cat? (maybe using some bot or tool) would just need a naming structure that's what the OWID: visualization name was all about – do you have another suggestion or which prefix(/suffix) would you prefer?

but not sure we need this as explained in the post starting the thread, these pages take long to load and are problematic to edit. It requires downloading lots of data and HotCat does not work on these so one can't edit the usual way and with autocomplete. Opening the wikitext editor takes long and is prone to failing. Also linking to these pages is probably confusing where people wonder what all the many files on the page are about and/or browse these instead of the interactive visualization. An alternative approach would be to remove the files from the template page but I'm not sure if that causes the visualization to break. The pages could link to each other. Prototyperspective (talk) 13:57, 28 January 2026 (UTC)

Okay so like this OWID/daily_meat_consumption_per_person? Doc James (talk · contribs · email) 00:55, 29 January 2026 (UTC)

This is in the gallery namespace which has no prefix. From reading here Commons:Galleries it seems this is okay but might be worth asking a wider audience before we create them all. Doc James (talk · contribs · email) 01:00, 29 January 2026 (UTC)

Asking a wider audience is what I was/am doing here; I thought and think this is the most appropriate large-audience place to discuss this. I pinged you in the comment not because this was addressed to just you. Okay so like this OWID/daily_meat_consumption_per_person? Yes, exactly, thanks. I edited the page to make the visualization larger and center it, what do you think? And also is there a way to create these pages for all the visualizations? Then I'd remove the cats on the large template pages and add them to these pages similar to how files are categorized (nearly all files in Category:Our World in Data except for these OWID Importer mass uploads are by now categorized by subject so this is basically the remaining issue to complete the topical categorization of these data graphics). Prototyperspective (talk) 13:19, 29 January 2026 (UTC)

Yes all these pages can be created by bot. Doc James (talk · contribs · email) 23:40, 29 January 2026 (UTC)

Nice, could you do that if there already is a bot for this and if not create the discussion at Commons:Bots/Requests? Prototyperspective (talk) 11:46, 30 January 2026 (UTC)

Hiding structured data edit links in the printable version

Files like File:St_Peter,_Bounces_Road,_Edmonton_-_geograph.org.uk_-_3802520.jpg and File:The Reverend Clive Foster MBE 1.jpg have pencil-shaped edit links for structured data. Can those icons be hidden in the printable version accessible from Tools, for the same reason why the main "Edit" link is hidden? Maybe the magnifier icons and the question mark icons, too. I'm guessing those are a part of a template somewhere, but I couldn't identify it. whym (talk) 05:01, 25 January 2026 (UTC)

This is implemented by Module:WikidataIB. They can add "noprint" to class surrounding that symbol and it should no longer be there. —TheDJ (talk • contribs) 16:46, 26 January 2026 (UTC)

Thanks, but is that module for SDC in addition to Wikidata? The data values of the 2 files mentioned (at least those in the Summary/Information section) are pulled from SDC, although the properties are from Wikidata. I found core.editAtSDC and core.editAtWikidata in the meantime. Perhaps both modules are to be edited? whym (talk) 12:01, 28 January 2026 (UTC)

Requested now at Module talk:Core. whym (talk) 11:39, 4 February 2026 (UTC)

Is it possible to check file EXIF data with edit filters?

Is it possible create an edit filter that would check file EXIF "user comment" field? Many copyvios have user comment "Screenshot" and it would be useful to automatically tag such uploads. Preventing is not necessary as there are legitimate files as well. MKFI (talk) 12:21, 25 January 2026 (UTC)

Let me hasard a guess: it's not possible, unfortunately (at least right now). Perusing Commons:Abuse filter and mediawikiwiki:Extension:AbuseFilter/Rules format, I didn't see any mention of (pre-made) EXIF-related variables. If, and that's a large if, the software could allow for arbitrary user-defined variables, then it's conceivable that such an EXIF-specific rule may be developed, but that's way beyond my own abilities. Pinging @Lustiger seth: whom I know as being knowledgeable about the edit filter technique, he may be able to provide a technical insight. Regards, Grand-Duc (talk) 13:26, 25 January 2026 (UTC)

Currently i don't see any possibility for using exif variables.

There are file attributes such as file_mime, file_size, file_width, but nothing for file content (incl. exif data). So i guess you should create a feature requesting ticket at https://phabricator.wikimedia.org. The only related ticket i found was https://phabricator.wikimedia.org/T170251, but that's only related vaguely.

-- seth (talk) 13:52, 25 January 2026 (UTC)

This would also be an extremely useful feature for identifying likely AI uploads and Facebook copyvios. Omphalographer (talk) 05:07, 26 January 2026 (UTC)

The only way to do that right now, is to use quarry. —TheDJ (talk • contribs) 16:42, 26 January 2026 (UTC)

As an aside, there's a slightly neater way of doing that which you might be interested in: quarry:query/101452. Omphalographer (talk) 18:42, 26 January 2026 (UTC)

I often take screenshots of pictures I have taken to make sure that the geolocation data is removed. Doc James (talk · contribs · email) 08:04, 27 January 2026 (UTC)

@Doc James: that's a really dirty way of removing metadata and, forcibly, reducing the quality. May I suggest that you edit the EXIF directly? There are hints for that in COM:Exif, I'd say "Use EXIFTool and EXIFToolGUI" (as long as you're on Windows), handy pieces of software for that. Regards, Grand-Duc (talk) 10:56, 27 January 2026 (UTC)

On Windows, you can edit EXIF via the file properties, which is even easier. - Jmabel ! talk 18:32, 27 January 2026 (UTC)

Tech News: 2026-05

Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.

Updates for editors

• Wikimedia Foundation invites comments on proposed future of the Product and Technology Advisory Council until 28 February.
• All users with registered accounts can now use passkeys for two-factor authentication (2FA). Passkeys are a simple way to log in without using a second device. They verify the user's identity using a fingerprint, face scan, or a PIN code. To set up a passkey, first set up a regular 2FA method. Currently, to log in with a passkey, users must also use a password. Later this quarter, passwordless login will allow users to log in with a single click and a passkey. Users with advanced rights will also be required to have 2FA enabled. This is part of the Account Security project.
• Unregistered contributors on blocked IPs or blocked IP ranges can now interact on-wiki to appeal a block by creating a temporary account to appeal a block on the user talk page, unless the "prevent this user from editing their own talk page" is enabled. This solves the problem of logged-out users unable to use the default unblock process via user talk page. [1]
• View all 20 community-submitted tasks that were resolved last week. For example, the Two-Factor Authentication (2FA) methods description on the management page has been updated. It is now clearer and easier for users to understand and make use of. [2]

Updates for technical contributors

• A new AbuseFilter variable, account_type, has been added to provide a reliable way to determine the account type being created in the createaccount and autocreateaccount actions. As part of this change, the variable accountname has been renamed to account_name, and accountname is now deprecated. Edit filter managers should update any filters that use hardcoded account type checks or the deprecated variable. [3]
• Image thumbnails that are requested in non-standard sizes, and using non-standard methods such as direct requests to upload.wikimedia.org/… will stop working in the near future. This change is to prevent ongoing external abuse by web-scrapers and bots. Some users with custom CSS/JS, Interface Admins who can fix gadgets and local skins, and Tool-authors, will need to update their code to use standard thumbnail sizes. Details, search-links, and examples of how to fix them, are available in the task.
• Detailed code updates later this week: MediaWiki

Tech news prepared by Tech News writers and posted by bot • Contribute • Translate • Get help • Give feedback • Subscribe or unsubscribe.

MediaWiki message delivery 21:14, 26 January 2026 (UTC)

Extra line break in PD-Canada-Crown template

Moved to Commons talk:WikiProject Templates § Extra line break in PD-Canada-Crown template?: -Consigned (talk) 13:32, 31 January 2026 (UTC)

My Image Annotation button is gone

Is it only me, or are annotations broken for anybody else as well? I neither see any existing annotations nor does the button "Add note" appear. And no, I didn't switch it off in my settings. Chianti (talk) 22:20, 28 January 2026 (UTC)

Works for me. That's about the last thing to load, via JavaScript, after a whole bunch of other JavaScript, so if any other JS is broken, or if your JS times out, it could be missed. If you are enough of a techie to use the browser console, you may find a coherent error message, possibly very tangentially related to the symptom. - Jmabel ! talk 05:49, 29 January 2026 (UTC)

Works for me in Chrome, Safari, and Firefox, but not Vivaldi, and it doesn't give any errors. While testing, I found that annotations won't show if the window is too narrow (less than about 830 pixels wide), so if your browser is blocking access to the window size (or you're using it on a phone in portrait orientation), that might be the problem. --Carnildo (talk) 22:17, 29 January 2026 (UTC)

Ongoing WMF-sponsored work on video2commons

As you probably know, there is not a lot of WMF developer support for Commons these days. The one thing I'm aware of that is moving forward is that they have a contractor working on video2commons. At meta:Product and Technology Advisory Council/Unsupported Tools Working Group#January 2026 they report some recent work; what is probably of most general interest is support for playlists and user-library uploads, better subtitle extraction, and several aspects of support for importing from YouTube.

From what I can tell, the one contractor currently working directly on this is doing good work; still, I continue to believe that Commons could benefit greatly from far more WMF dev support. It is good that they've shown that certain work can be successfully contracted out, but it is equally clear that most cannot, and that we need people at least a handful of developers who bring or, well, develop an understanding of Commons, not just of some individual tool. (With reference to "equally clear that most cannot", I was party to some of the discussions of where to focus this resource, and several higher priorities were rejected because handing them to a contractor would set that contractor up to fail.) - Jmabel ! talk 18:01, 29 January 2026 (UTC)

Out of curiosity, what were the higher priority things that a contractor would be setup to fail if given? Bawolff (talk) 21:18, 2 February 2026 (UTC)

@Bawolff: I don't remember everything that went by, but I believe improvements to Video2Commons were the other leading candidate. I remember there was discussion of the Upload Wizard (apparently a bit of a mess internally, which—being a developer myself—amazes me: it is not doing anything I think should be complicated); CropTool was discussed, but it looks like a new group of volunteers have taken that on successfully; there was some discussion of MediaViewer, which I pretty much don't use so I had and have nothing intelligent to say on that front; there was definitely talk about getting seriously behind one of the batch uploading tools, probably PattyPan, and I think that was one that was rejected for this go-around on the basis of too large (there was only funding for three person-months); I also remember there was some discussion of rewriting Cat-a-lot (another where I personally have little to say: I'm pretty fine with it as it is).

I could be missing something here. When I was brought into this, I was promised it would be at most a few hours of my time in any given month, and I was in a reactive/reviewing mode and did not take copious notes. The most proactive thing I did here was to put about 8-10 tools on their radar that had not been on their initial list.

FWIW, my own ideas as to where I'd put resources are very different. My own view is that the single most valuable thing we could have is a full-time combination PM and volunteer coordinator to help coordinate volunteer-based work on tools, almost the opposite use of resources from a short-term, outside, contract-based developer. My overall take: not nearly enough WMF resources are devoted to tool support or to Commons; given that limitation, what they are doing is sane, but probably not optimal; "sane" is a big improvement on where we were a few years back. - Jmabel ! talk 18:23, 3 February 2026 (UTC)

not nearly enough WMF resources are devoted to tool support or to Commons agree if by "tool support" you mean tool development and I note that I can't think of many things that WMF could do that would be effective and relates to Commons that isn't technical development and doesn't involve it as the main part I remember there was discussion of the Upload Wizard would be great if development on it continues – see the ideas and requests at Commons_talk:WMF support for Commons/Upload Wizard Improvements (eg the recent threads near bottom).

Commons could benefit greatly from far more WMF dev support Very much agree! For example continued work on MediaSearch (e.g. see these issues for it here) or the mobile view of category pages or many of the wishes about Commons in the technical Wishlist: https://meta.wikimedia.org/wiki/Community_Wishlist/Wishes?tags=multimedia (voting open!)

.

It is good that they've shown that certain work can be successfully contracted out, but it is equally clear that most cannot probably there is still lots of things that can be contracted out and so far I've not seen explanations for which things can't be (the reasons for why).
Other ideas are hiring more devs (remote, in chapters, and locally in the US), m:Wish bounties, and facilitating & aiding more open source volunteer development, e.g. via the concrete feasible ideas that I've outlined here.
Prototyperspective (talk) 18:50, 3 February 2026 (UTC)

A couple of issues with bringing in [relatively short-term] contractors for s/w development generally, nothing here specific to WMF:

• In general, it is easiest to use contractors for very well-defined tasks that can reasonably be expected to be completed in a specific time frame, easily be tested to determine task completion, and not need a great deal of ongoing maintenance after that. The farther you get from that in any respect, the harder to use contractors.
• Using contractors neither leverages nor builds up organizational memory. Unless a contractor is very fast on the uptake, it is far more crucial than for hiring a long-term employee to hire someone who is already strong in all of the relevant technologies, so in making a hire you may have to trade that off against overall skill level. Also, the lower the overall quality of any existing body of code, the harder it is for a contractor to take on doing almost anything with it, compared to someone who is in it for the long term.

Jmabel ! talk 23:25, 3 February 2026 (UTC)

Re wish bounties - bug bounty (not the security type but the type where you pay people to work on specific feature requests) has failed in basically every open source project it has been tried. I would suggest avoiding that unless people look carefully into why it failed in other projects and avoid falling into similar traps. Bawolff (talk) 09:24, 4 February 2026 (UTC)

Interesting; could you provide some useful link(s) regarding this here or at the talk page of the page linked? (examples where it failed, info on why, etc) Haven't heard that and thought I heard of some quite successful cases. There's probably many ways this can be done – for example an org could offer bounties for implementation of any of a large set of wishes or Wikipedians could crowdfund development and specify ranked wishes etc. I'd love to donate to get tangible wiki wishes implemented under certain circumstances. Prototyperspective (talk) 12:07, 4 February 2026 (UTC)

After googling, maybe its not as clear cut as i thought. I did find https://ziglang.org/news/bounties-damage-open-source-projects/ and https://www.dokuwiki.org/bounties . That said, it is an idea people try to push in the open source world a lot (mostly companies who are hoping to get a cut) but there are very limited success stories. Its important to keep in mind that the going rate for mediawiki consulting is somewhere in the range of $50-$150 USD/hr (or higher), so usually bounties don't make financial sense for the people doing them as they are usually not high enough. Of course often people have other additional motivations in addition to or instead of $$$, but keep in mind adding low amounts of money can sonetimes back fire; many people who would do things for free would refuse to do things for small and medium amounts of money. The other problem with bounties is sometimes there is misalignment - people try and do the bare minimum and you end up with unusable crap or worse the person who created the bounty doesn't understand the problem and makes it for the wrong thing. In many situations the technical part is only a small part of the work, and getting someone to do it does not really move the needle on actually making it happen (I agree 100% with Jmabel on the benefits of a good PM). As an aside, i have a vauge memory of User:Eloquence trying to setup a bounty for DPL or RSS mediawiki features a long time ago. However i can't find any reference to it online so maybe that is only in my imagination. Bawolff (talk) 14:16, 4 February 2026 (UTC)

P.s. Don't let me disuade you though, even though i have doubts about the value of monentary rewards, getting a list of feature requests to a state where you could run a succesful bounty program (i.e. you have a ranked list of features where the requirements are well documented and all appropriate stakeholders have bought in to the design choices, plus dedicated CR resources) would be hugely valuable. People would probably start doing things off that list without any monentary compensation. The primary blockers for volunteer devs is that nobody knows what the community actually wants, nobody really wants to put in the social effort to get consensus on the technical requirements, nobody really wants to do things if its a gamble whether you will get timely code review. Fix all that, and you wont need to pay people. Bawolff (talk) 14:43, 4 February 2026 (UTC)

Interesting feedback, thanks. I don't see the point with the bounties not making financial sense; it's some monetary incentive vs no such incentive. I was well aware of the potential backfiring issue but I'm not sure if that's what you meant: I was thinking of if some people get money for their development, the volunteers who do it without compensation would feel less motivated and/or would implement the issues with bounties instead of the other things (latter is not necessarily all bad). I don't think people doing just the bare minimum would be a substantial problem as it would still have to pass review and validation and just doing the minimal implementation is fine. Do the more advanced stuff later and just build the minimum thing first of all is a great principle. This also makes it much less likely there's work on things that will later be scrapped. There are lot of open bugs and wishes that are such a list of feature requests already, such as many of the good-first-bug things or issues with lots of subscribers who wait for it to finally be implemented etc. I don't see many devs working on many of these; sth that could be done is to make such lists more visible; e.g. via the aforementioned banner that links to a landing page where such wishes and issues could be listed where devs can see how some of them are interesting to them and sth that would be worth implementing. That nobody knows what the community actually wants is false and probably most issues don't need some "consensus on the technical requirements". Timely code review is something the WMF can and should improve, for example by hiring more devs that do these or again facilitating more volunteers to join these efforts. I don't say wish bounties are needed; it's one option and it's especially an option since the WMF is still not showing much interest in actually increasing development. I again would donate (under certain circumstances) to a non-profit organization that uses >95% of its donations for actual tangible software development of concrete issues & wishes and such could be achieved via wish bounties so I think there may well be some potential there. Prototyperspective (talk) 16:27, 4 February 2026 (UTC)

By backfiring I mean psychologically. There is a weird thing for some people, where once you put a dollar number on something people start to value it very differently. If its free, they center the work on more intangible values, but once a dollar value is assigned they stop thinking about the intangible values and start viewing it solely through a money lens. Admittedly this is a bit just my personal opinion - it could be wrong or maybe only apply to a small minority of contributors. Re, contributors getting jealous of those getting money - while I agree, i think that ship has long sailed with all the WMF paid people. By minimum, i don't mean Minimum-viable-product (Which i agree is a good thing), I mean code that is at minimum level of quality which might be difficult to review and cause manitenance problems (or other externalities later). Re issues with lots of subscribers. Often these have hidden issues or reasons for not doing them that need to be resolved. For a bounty program to be successful, hidden issues need to be surfaced, since the bounty-completer is likely to be an outsider and won't have the unspoken context. Regardless, I would encourage people interested in this idea to keep a top 5 list of what they think the most viable & valuable tasks to turn into bounties would be. It would give us something concrete to talk about even if nobody has ponied up any cash yet. Bawolff (talk) 22:26, 4 February 2026 (UTC)

See m:Talk:Wikimedia Foundation Annual Plan/2026-2027#The problem that underlies most issues and challenges noted here and elsewhere.
Prototyperspective (talk) 00:47, 5 February 2026 (UTC)

"Hist" becoming "History"

Tracked in Phabricator
Task T244411

I noticed that the "hist" link on Special:Watchlist and Special:Contributions now says "history". Is there a reason this was changed? I'm not sure it's necessary and the fact it's wider than "diff" makes it look very awkward. - The Bushranger (talk) 06:05, 31 January 2026 (UTC)

I found that weird, too. But my main reaction is: I'm sorry anyone put any time into changing that, and hope they don't then waste more time changing it back. - Jmabel ! talk 23:47, 31 January 2026 (UTC)

See phab:T244411. The change has been reverted. NguoiDungKhongDinhDanh 00:05, 1 February 2026 (UTC)

Still seeing "history" on my watchlist. I assume the reversion is still percolating through the system? - The Bushranger (talk) 04:51, 1 February 2026 (UTC)

That's correct. It's not deployed yet, at least not here. NguoiDungKhongDinhDanh 04:53, 1 February 2026 (UTC)

Changes take 0 to 14 days to be deployed, depending on day of the week and any possible unforeseen problems that cause temporary holds —TheDJ (talk • contribs) 16:01, 2 February 2026 (UTC)

Tech News: 2026-06

Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.

Updates for editors

• The "Page information" feature, which gives validating information about a page (example), now automatically includes a table of contents. If there is a local MediaWiki:Pageinfo-header page created by individual users, it can now be removed. [4]
• View all 21 community-submitted tasks that were resolved last week. For example, VisualEditor previously added bold or italic formatting inside link descriptions, making the wikicode complex. This has now been fixed. [5]

Updates for technical contributors

• There was no XML dump on 20 January. Additionally, from now on, dumps will be generated once per month only. [6]
• The MediaWiki Interfaces team removed support for all transform endpoints containing a trailing slash from the MediaWiki REST API. All API users currently calling those endpoints are encouraged to transition to the non-trailing slash versions. If you have questions or encounter any problems, please file a ticket in phabricator to the #MW-Interfaces-Team board.
• Detailed code updates later this week: MediaWiki

Weekly highlight

• Users are reminded that the Wikimedia Foundation has shared some guiding questions for the July 2026–June 2027 Annual Plan on Meta and Diff. These focus on global trends, faster and healthier experimentation, better support for newcomers, strengthening editors and advanced users, improving collaboration across projects, and growing and retaining readership. Feedback and ideas are welcome on the talk page.

Tech news prepared by Tech News writers and posted by bot • Contribute • Translate • Get help • Give feedback • Subscribe or unsubscribe.

MediaWiki message delivery 17:40, 2 February 2026 (UTC)

Not mentioned here, as it isn't really relevant for most projects, but I fixed a number of video pipeline problems.

• Transcode pipeline had files that had crashed but for which the crash had not been registered properly. phab:T385270
• Some Opus files were not properly recognized phab:T414643
• Some Ogg files had duration 0. After this week that should be fixed phab:T414348

Additonally, a few things are currently in progress. They are either fixed or about to be fixed in getID3, which should trickle down to Commons somewhere in the next few weeks, including:

• very small files would crash
• midi files without tempo events had no duration phab:T414645
• flac files could have no duration phab:T414641
• chunked uploads of webm sometimes crashed phab:T403213
• streamed (youtube) webm does not have duration phab:T357035

And of longer term interest in the image area

• MediaWiki will now directly render SVGs by default. This is not active for WMF, but the hope is that by moving this forward we will someday get there.
• JPEG XL files can soon be recognized, which is a prerequisite for parsing and thumbnailing them at some point phab:T270855

I hope that makes people happy. And if you are interested in working on problems similar to do, i encourage you to do so and to tag me as reviewer of your work on gerrit. —TheDJ (talk • contribs) 11:36, 4 February 2026 (UTC)

Thanks, great efforts and results! Sjoerd de Bruin (talk) 21:01, 5 February 2026 (UTC)

TheDJ makes important statements here. I think Com:Textured 3D is watched with high anticipation, too --PantheraLeo1359531 😺 (talk) 10:01, 6 February 2026 (UTC)

{{int:Talkpagelinktext}} getting substituted on custom signatures

{{int:Talkpagelinktext}} is a template for translating "talk", but when using a custom signature with templates, it automatically substitutes every template, even the translation template, thus losing the translation. Is there a workaround for this? HyperAnd [talk] 13:18, 5 February 2026 (UTC)

Creative Commons Search Portal now uses MediaSearch by default

A few weeks ago, I discovered that the search.creativecommons.org uses a regular text search instead of the media search interface. The text search is not well suited for finding media files, it only searches for a keyword and it does not show any media files. Well, the good news is that this is now fixed. I also corrected the project name there, which was "Wikipedia Commons". Nemoralis (talk) 16:55, 6 February 2026 (UTC)

Protection question

This page currently says it's cascade-protected, but I was able to edit it as a non-admin. What's happening to cause the discrepancy? Sdkb ^talk 20:03, 6 February 2026 (UTC)

@Sdkb Per point 3 of Commons:Village pump/Technical/Archive/2025/03#Tech News: 2025-10, since March 2025, the cascading protection in this case only includes upload protection, and doesn't include edit protection. Thanks. Tvpuppy (talk) 20:22, 6 February 2026 (UTC)

Ah, thanks! That's a nice improvement! We need to update the notice in that case. Sdkb ^talk 21:32, 6 February 2026 (UTC)

Thumbnail problem

Anyone know why no thumbnail shows for File:View west on Boyer Avenue, Seattle, December 15, 1911 (MOHAI 13058).jpg, and/or can fix it? - Jmabel ! talk 04:01, 7 February 2026 (UTC)

@Jmabel: I can see the various thumbnail sizes (330px, 960px, etc.) fine on my end. Are you still experiencing the issue? ~Kevin Payravi (talk) 08:03, 7 February 2026 (UTC)

@Kevin Payravi: I can see several sizes of thumbnail, but on the file page itself the preview is blank. If that's just me and not a more general problem, then it's odd but not important. - Jmabel ! talk 19:55, 7 February 2026 (UTC)

I see both thumbnails in Firefox. There have been multiple threads about the thumbnails missing problem here, two or one in I think the latest archive. Prototyperspective (talk) 22:04, 7 February 2026 (UTC)

Problem of accessing Commons

Since 2025, when accessing Commons without using a VPN, it has sometimes been inaccessible for a day, sometimes for several days, and occasionally, clicking "Publish changes" after editing fails to go through. May I ask what is causing this? Huangdan2060 (talk) 14:16, 7 February 2026 (UTC)

If even regular reading access is limited your IP might be on an AI scraper list. Maybe you can get help at ca@wikimedia.org. GPSLeo (talk) 14:50, 7 February 2026 (UTC)

Thank you, but the page you provided cannot be accessed in China. Huangdan2060 (talk) 01:09, 8 February 2026 (UTC)

That's not a page, it's an email address. Jmabel ! talk 05:12, 8 February 2026 (UTC)

Irritating preview for grayscale GeoTIFFs

Hi!

Tracked in Phabricator
Task T416775

I filed a task. When special GeoTIFFs like "digitale Geländemodelle" (digital terrain models) are uploaded, we see a plain white preview. This is irritating and may lead to premature deletion requests. When I put this file in QGIS, I see graphical information. This task may be useful, as Commons acquires more and more GeoData. --PantheraLeo1359531 😺 (talk) 20:56, 7 February 2026 (UTC)

Help downloading full-resolution images

Could anyone please help obtain the full-resolution images found here? I'm able to upload the as-displayed images, but—if you click on the images—you're able to zoom in further. I also tried dezoomify (website & extension), but it did not work. There are for use on the article Serpent labret with articulated tongue and, as works by the U.S. federal government, should be in the public domain. Thanks, --Usernameunique (talk) 17:16, 8 February 2026 (UTC)

I read a bit in the source code, hinting at that the source is using Open Seadragon. Using that info to google, I landed upon https://hackernoon.com/how-to-download-hi-res-images-from-museum-websites-b4446387e75d That may help other people investigate further, I stopped at that point... Regards, Grand-Duc (talk) 20:04, 8 February 2026 (UTC)

Perhaps I'm missing something, but are you referring to this image version (2000x1321)? I just copied the image URL from the page you linked and remove the size parameter (&max=980) to get the image in its actual size. Thanks. Tvpuppy (talk) 23:02, 8 February 2026 (UTC)

Thanks, both. Tvpuppy, I tried that too, but it's still smaller than the zoomed-in version—so unless you can zoom in beyond full resolution (i.e., you zoom in more, but it just gets fuzzier), it seems there's a higher-resolution version. --Usernameunique (talk) 01:02, 9 February 2026 (UTC)

I've uploaded these at the available resolution for now (here, here, and here), but if anyone is able to download the full resolution, please let me know. Thanks, --Usernameunique (talk) 20:04, 16 February 2026 (UTC)

Tech News: 2026-07

Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.

Updates for editors

• Logged-in contributors who manage large or complex watchlists can now organise and filter watched pages in ways that improve their workflows with the new Watchlist labels feature. By adding custom labels (for example: pages you created, pages being monitored for vandalism, or discussion pages) users can more quickly identify what needs attention, reduce cognitive load, and respond more efficiently. This improves watchlist usability, especially for highly active editors.
• A new feature available on Special:Contributions shows temporary accounts that are likely operated by the same person, and so makes patrolling less time-consuming. Upon checking contributions of a temporary account, users with access to temporary account IP addresses can now see a view of contributions from the related temporary accounts. The feature looks up all the IPs associated with a given temporary account within the data retention period and shows all the contributions of all temporary accounts that have used these IPs. Learn more. [7]
• When editors preview a wikitext edit, the reminder box that they are only seeing a preview (which is shown at the top), now has a grey/neutral background instead of a yellow/warning background. This makes it easier to distinguish preview notes from actual warnings (for example, edit conflicts or problematic redirect targets), which will now be shown in separate warning or error boxes. [8]
• The Global Watchlist lets you view your watchlists from multiple wikis on one page. The extension continues to improve — it now properly supports more than one Wikibase site, for example both Wikidata and testwikidata. In addition, issues regarding text direction have been fixed for users who prefer Wikidata or other Wikibase sites in right-to-left (RTL) languages. [9][10]
• The automatic "magic links" for ISBN, RFC, and PMID numbers have been deprecated in wikitext since 2021 due to inflexibility and difficulties with localization. Several wikis have successfully replaced RFC and PMID magic links with equivalent external links, but a template was often required to replace the functionality of the ISBN magic link. There is now a new built-in parser function {{#isbn}} available to replace the basic functionality of the ISBN magic link. This makes it easier for wikis who wish to migrate off of the deprecated magic link functionality to do so. [11]
• Two new wikis have been created:
  • a Wikipedia in Jju (w:kaj:) [12]
  • a Wikipedia in Nawat (w:ppl:) [13]
• View all 23 community-submitted tasks that were resolved last week.

Updates for technical contributors

• A new global user group has been created: Local bots. It will be used internally by the software to allow community bots to bypass rate limits that are applied to abusive web scrapers. Accounts that are approved as bots on at least one Wikimedia wiki will be automatically added to this group. It will not change what user permissions the bot has. [14]
• Detailed code updates later this week: MediaWiki

Meetings and events

• The MediaWiki Users and Developers Conference, Spring 2026 will be held March 25–27 in Salt Lake City, USA. This event is organized by and for the third-party MediaWiki community. You can propose sessions and register to attend. [15]

Tech news prepared by Tech News writers and posted by bot • Contribute • Translate • Get help • Give feedback • Subscribe or unsubscribe.

MediaWiki message delivery 23:27, 9 February 2026 (UTC)

Process Duplicates

Can somebody explain what will happen if I will press "Process Duplicates". I found it on File:2022 California Proposition 27 by County.svg and a duplicate. I do not understand how to handled the two content pages. I did search for a Help: page, but did not find any. Thanks, Ellywa (talk) 22:39, 10 February 2026 (UTC)

Looks like you get a tool intended to let you merge any content from the two different file pages and then to turn one file into a redirect to the other file. However, it doesn't look to me like it's a very clear UI (e.g. both file descriptions are editable; which one will be used once the merge takes place?)

Anyone know who is in charge of this thing? - Jmabel ! talk 00:43, 11 February 2026 (UTC)

FastCCI dysfunctional again?

I was completing the overly incomplete Category:Videos of Trochilidae after creating a redirect to it from the easier-to-find (in HotCat & UploadWizard autocomplete, Commons search, and Web search engines) Category:Videos of hummingbirds. To do so, I used deepcategory:"Trochilidae" -incategory:"Videos of Trochilidae". There, I noticed lots of videos of helicopters.

(All of this is a normal method to complete a category which could be applied in many cases, maybe at some point more systematically or semi-automatically or routinely.)

There is just one way to see why a given file is underneath a category branch. That is going to the respective category -> Page information -> copy page ID -> append ?fastcci_from=id with the id (29437) to the URL of the file one would like to check that one has to open in a new tab. For example:
https://commons.wikimedia.org/wiki/File:Escuadrilla_Hist%C3%B3rica_Mentor_y_Patrulla_Aspa_-_BugWarp_408.webm?fastcci_from=29437

However, it doesn't work. There just recently was this thread about the topic of category<->file path but FastCCI is already dysfunctional again.

@Dschwen: Could you restart it please and if possible look into why that tool is going down all the time and fix that (or create an issue about how it could be fixed)? Another approach to this would be for somebody to implement m:Community Wishlist/Wishes/A way to see why a file is somewhere underneath a specific category (tool to show cat-path) (voting open) where this functionality would be in a separate tool that doesn't go down all the time and where one can display the cat-path within the deepcategory view (linked above) without having to open a new tab an all the time-intensive manual things. Thanks, Prototyperspective (talk) 13:04, 11 February 2026 (UTC)

 Comment, about the helicopters videos, the reason they are under Category:Trochilidae is because "Trochilidae --> Things named after hummingbirds --> Airbus Helicopters H120 Colibri". Thanks. Tvpuppy (talk) 17:37, 11 February 2026 (UTC)

Ah thanks so it's another case of Things named after… categories introducing unexpected/offtopic files; see Commons:Village pump#Moving Things named after xyz categories out of xyz cats. Nevertheless, it would be nice if the tool could be restarted. Prototyperspective (talk) 17:46, 11 February 2026 (UTC)

Download ZIP contents through OpenRefine

I would like to upload a batch of orthophotos of Thuringia. Each tile comes as ZIP file with TIF, meta and tfw file. Is there a way to let the Wikimedia servers extract only the TIF to be uploaded? Having them down- and reuploaded on my PC probably takes some time. Thanks! --PantheraLeo1359531 😺 (talk) 16:15, 12 February 2026 (UTC)

Probably a good candidate for Commons:Bots/Work requests. - Jmabel ! talk 18:27, 12 February 2026 (UTC)

CommonsDelinker is dead

CommonsDelinker (talk · contributions (views) · deleted user contributions · deleted uploads · recent activity (talk · project · deletion requests) · logs · block log · global contribs · CentralAuth) Still processes deletions and cat moves, but no file moves since 2026-01-22. There are ca 1750 accumulated unprocessed requests. The page toolsadmin lists 7 administrators: @User:AntiCompositeNumber @User:Grin, @User:Magnus Manske, @User:Mdaniels5757, @User:Peter Gervai, @User:Steinsplitter, @User:Zhuyifei1999. -- Taylor 49 (talk) 02:02, 14 February 2026 (UTC)

Looks like it's working again. AntiCompositeNumber (they/them) (talk) 22:26, 16 February 2026 (UTC)

@AntiCompositeNumber: Hi, I don't see it replacing files. It only removes files, and renames catgories. Yann (talk) 09:48, 17 February 2026 (UTC)

Bot was crashing because it was failing to log in to one wiki. Should be working now. AntiCompositeNumber (they/them) (talk) 02:56, 18 February 2026 (UTC)

Thanks a lot! Yann (talk) 16:14, 18 February 2026 (UTC)

There was an ongoing problem (for many years now) that due some mediawiki... whatever, let's call it bug, some projects fail login and respond unholy errors instead which make the bot getting stuck after a while.

I have implemented a workaround, so theoretically this login reason will not get the bot stuck anymore, hope dies last. Thanks for the ping. -- grin ✎ 02:51, 22 February 2026 (UTC)

Sidenote: do not ping other people (old developers) in, unless you deliberately plan to annoy them. It is not an accident they have not wanted to maintain it anymore. (Only exception is Steinsplitter, who operates the Category Mover "bot subtask".) -- grin ✎ 02:54, 22 February 2026 (UTC)

How to create barnstar?

Anyone know how to create such barnstar and icon in Gimp? I have seen this video but it's less advanced. Eurohunter (talk) 16:33, 16 February 2026 (UTC)

@Eurohunter, you might try Commons:Graphic Lab. Nemoralis (talk) 21:21, 18 February 2026 (UTC)

Tech News: 2026-08

Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.

Weekly highlight

• The SRE Team will be performing a cleanup of Wikimedia's Etherpad instance, the web-based editor for real-time collaborative document editing. All pads will be permanently deleted after 30 April, 2026 – if there are still migration projects in progress at that point the team can revisit the date on a case by case basis. Please create local backups of any content you wish to keep, as deleted data cannot be recovered. This cleanup helps reduce database size and minimize infrastructure footprint. Etherpad will continue to support real-time collaboration, but long-term storage should not be expected. Additional cleanups may occur in the future without prior notice. [16]

Updates for editors

• The Information Retrieval team will be launching an Android mobile app experiment that tests hybrid search capabilities which can handle both semantic and keyword queries. The improvement of on-platform search will enable readers to find what they’re looking for directly on Wikipedia more easily. The experiment will first be launched on Greek Wikipedia in late February, followed by English, French, and Portuguese in March. Read more on Diff blog. [17]
• The Reader Growth team will run an experiment for mobile web users, that adds a table of contents and automatically expands all article sections, to learn more about navigation issues they face. The test will be available on Arabic, Chinese, English, French, Indonesian, and Vietnamese Wikipedias.
• Previously, site notices (MediaWiki:Sitenotice and MediaWiki:Anonnotice) would only render on the desktop site. Now, they will render on all platforms. Users on mobile web will now see these notices and be informed. Site administrators should be prepared to test and fix notices on mobile devices to avoid interference with articles. To opt out, interface admins can add #siteNotice { display: none; } to MediaWiki:Minerva.css. [18][19]
• View all 19 community-submitted tasks that were resolved last week. For example, an issue on Special:RecentChanges has been fixed. Previously, clicking hide in the active filters caused the "view new changes since…" button to disappear, though it should have remained visible. The button now behaves as expected. [20]

Updates for technical contributors

• New documentation is now available to help editors debug on-site search features. It supports troubleshooting when pages do not appear in results, when ranking seems unexpected, and when you need to inspect what content is being indexed, helping make search behavior easier to understand and analyze. Learn more. [21]
• Detailed code updates later this week: MediaWiki

Tech news prepared by Tech News writers and posted by bot • Contribute • Translate • Get help • Give feedback • Subscribe or unsubscribe.

MediaWiki message delivery 19:14, 16 February 2026 (UTC)

Issue uploading Inkscape SVG

I attempted to upload an updated version of File:Preferred_PM_polling_Canada_46th_federal_election.svg today. The graph is created in LibreOffice, exported as an SVG, and then inserted into Inkscape where the legend is edited, before being saved as an SVG. I have never had a problem uploading such SVGs before. However, today, I was prevented from uploading the update by the following error, which I have collapsed for convenience as it is daunitingly long. How can I fix this? Cremastra (talk) 22:19, 17 February 2026 (UTC)

Isn't this all code like if you open image in Notepad? I would like to know what is this code. Eurohunter (talk) 22:40, 17 February 2026 (UTC)

You have included a non-SVG image inside the SVG. That’s not allowed. —TheDJ (talk • contribs) 22:56, 17 February 2026 (UTC)

Okay, I suspected that. But how did I do that in the first place? Cremastra (talk) 23:28, 17 February 2026 (UTC)

Its actually kind of the opposite. They embedded an svg inside an svg, which isn't allowed. If it was a non-svg image it would be allowed. Basically, the svg has to be directly included instead of being indirectly embedded. Probably the issue is that you "inserted" the file into inkscape instead of opening the file with inkscape. Bawolff (talk) 19:43, 22 February 2026 (UTC)

Note: I removed the file content you pasted here and copied it into Phabricator Pastebin. It is available here: https://phabricator.wikimedia.org/P88886. Nemoralis (talk) 21:18, 18 February 2026 (UTC)

Wrong coordinates can't be removed

At File:Edinburgh Airport.jpg I can't remove the coordinates of the point of view. Is this a known problem? (I'm not asking about workarounds.) Prototyperspective (talk) 11:42, 19 February 2026 (UTC)

@Prototyperspective: you removed the template but not structured data; they both need to be removed. I have done that. However since the coordinates are present in the file EXIF, it is possible that a bot might add them again. MKFI (talk) 15:38, 19 February 2026 (UTC)

you removed the template but not structured data this is what this thread is about. it is possible that a bot might add them again does somebody know whether it will? Prototyperspective (talk) 16:03, 19 February 2026 (UTC)

I have done that How? It did not work when I tried so at least it's unintuitive or there's a bug. Prototyperspective (talk) 16:27, 19 February 2026 (UTC)

You can always override it by entering the correct coords or tag the problem using {{Fact disputed}}. RadioKAOS / Talk to me, Billy / Transmissions 15:48, 19 February 2026 (UTC)

Those are two workarounds. Also both have issues. I'm asking how to remove the geoinfo from the SD, it doesn't seem to be possible when I tried. {{Fact disputed}} is for file contents, not the metadata. Prototyperspective (talk) 16:04, 19 February 2026 (UTC)

@Prototyperspective: I don't know why it didn't work for you, I simply went to the structured data, clicked edit on coordinates and then the trashcan icon. For suppressing the wrong coordinates there is also {{Location withheld}}. MKFI (talk) 16:46, 19 February 2026 (UTC)

I tried it on a different file and there it removed the coordinates. It only doesn't remove them when clicking "Remove all" (there is no blue publish icon when pressing that and confirming the popup that shows then). Prototyperspective (talk) 16:54, 19 February 2026 (UTC)

It depends, I also can not remove the coordinates from the interface. When I need to remove them I just revert the bot edit. Ymblanter (talk) 17:37, 19 February 2026 (UTC)

Might be related: phab:T313638. Nemoralis (talk) 20:58, 19 February 2026 (UTC)

Commons contributors by number of uploaded files (in use)

Commons:Commons contributions achievements/Commons contributors by number of uploads is a relatively new meta page showing a table of Commons users with a column for the number of files upload.

• I just wanted to check which quarry I used to create it but the server times out with error [65d16284-b14b-40ad-aff6-d0fa12d44efb] 2026-02-19 18:17:35: Fatal exception of type "Wikimedia\RequestTimeout\RequestTimeoutException" – the table can still be accessed in the archives here. Maybe somebody knows how this can be fixed (the points below require either that or creating a new page).
• Does somebody know how to add a column for a) rank b) whether or not the user is a bot c) number of files in use in any Wikimedia project d) maybe number of files in use in Wikipedia mainspace separately.
• Maybe people have further ideas for columns – e.g. maybe it's possible to have sortable column to also show number of files by filetype, and/or fraction of files in mainspace use, and/or files that are own work...

I think these would be interesting stats. Prototyperspective (talk) 18:26, 19 February 2026 (UTC)

For the timeout, splitting this across multiple pages might help. Bawolff (talk) 21:04, 22 February 2026 (UTC)

What is interesting is the percentage of files in use. Yann (talk) 21:35, 22 February 2026 (UTC)

Does somebody know how to add a column for … c) number of files in use in any Wikimedia project d) maybe number of files in use in Wikipedia mainspace the percentage could be calculated from that but I'm not sure it's more meaningful than number of files in use; percentage would imply one barely ever uploads other kinds of files. Prototyperspective (talk) 21:42, 22 February 2026 (UTC)

Well, having 1,000 files in use is not the same after uploading 1,000,000 files or after uploading 2,000 files. Yann (talk) 21:49, 22 February 2026 (UTC)

This could however be read by also looking at the other column that has that number. Anyway, I wouldn't have an issue with the percentage which btw is also shown in the Commons app. Prototyperspective (talk) 22:14, 22 February 2026 (UTC)

https://quarry.wmcloud.org/query/102335 (adding in use slowed down the query significantly. Percentage, and in use on main ns on a wikipedia are probably also possible but would slow it down more) Bawolff (talk) 06:22, 23 February 2026 (UTC)

Amazing, thank you! Since this would be run only rarely I don't see much of a problem with it taking long to load (in this case ~13 minutes). Regarding percentage, I think this can be seen in the columns and if one added it, one can't use it for sorting because at the top would be lots of users who uploaded just 1 or so file(s) which is in use. I would have no issue with it being added but don't think it's needed or very useful.

Do you know how one could show the number of files in use on Wikipedias mainspace? This is useful because for example some files are used only in sandboxes, talk pages or are pronunciation audios used in lots of wiktionary and Wikidata pages. Prototyperspective (talk) 14:11, 23 February 2026 (UTC)

Query updated to show that. Note this only includes actual main space, will not include non-main content namespaces like portal. It determines if the site is a wikipedia by if the wiki name ends in "wiki" and os less than 9 letters. I think that should include all wikipedias and exclude other sites, but there could be edge cases possibly. Bawolff (talk) 18:12, 23 February 2026 (UTC)

Aside: while percentage of files that are in use can be interesting, it should not be seen as a metric of who is a great contributor of content to Commons. A very high percentage usually means someone whose uploads have been driven precisely by what is need it in some particular sister project or projects, and who rarely or never uploads more than one image related to a particular subject. Example to the contrary: I just uploaded several dozen photos I took of contributing properties of NRHP-recognized historic districts in Binghamton, New York. Most of these buildings previously had no image on Commons. Few, if any, are likely to end up "used" in the near future because the en-wiki articles on these topics are stubs or near-stubs, the Wikivoyage article has exactly one image of the city's historic center, and I can't think where else they would show up on a sister project. Nonetheless, they are clearly well within Commons scope, and I can't imagine an argument against us actively wanting such images. - Jmabel ! talk 20:50, 23 February 2026 (UTC)

Super interesting, thanks! I find the use in WP-mainspace column the most interesting – since there is no way afaik to enumerate the table based on whatever column is sorted by, could you add a column for the # of number in use on WPmNS? Moreover, maybe a separate list without the bots would be interesting too but it seems like the human/bot boolean is totally false. Many bots, incl the user with most uploads do not have the bot flag but https://commons.wikiscan.org/ does show the correct bot flag – do you know why that is or how to show the correct value in that column? Maybe the accounts' bot flag was removed from somewhere where it shouldn't have been removed or it's not set but sth else is set that could be shown. Minor correction: it should be 'Files uploaded', not 'Images uploaded'. Prototyperspective (talk) 00:54, 24 February 2026 (UTC)

Sorted on the other column https://quarry.wmcloud.org/query/102363 Bawolff (talk) 02:19, 24 February 2026 (UTC)

Thanks! I've now put it on a page where links are clickable and which can be found from within Commons: Commons:Commons contributors by number of uploads and Wikipedia uses thereof (1-5000)

I've put further ideas and remaining issues regarding the list(s) into section "Development". For example, it would be nice if the usernames could be linked to their Uploads page (and I think linking there instead of their userpage also avoids them getting pinged). Prototyperspective (talk) 18:02, 24 February 2026 (UTC)

Those are files uploaded and files in use right, not just images? Because the tables are named as if these were just images so I'm confused whether to rename the column. Prototyperspective (talk) 22:39, 24 February 2026 (UTC)

Yes these are files of all types. The term image is used in the database for historical reasons. Bawolff (talk) 06:21, 25 February 2026 (UTC)

Idea: see top domains of source links & top uploaders to category

This could be useful to identify more Commons:Free media resources for example, interesting statistics (imagine seeing pie charts for categories with a click), better partnership & requests & recognition potentials, and broadly to better see where files are coming from.
For example, in categories relating to data graphics probably, ourworldindata.org would be a top domain and in some other cat it could be some academic journal.
There may be many applications for this and probably many are found only once this is possible over time. Maybe this is already possible via some Quarry queries. A difficulty is that one would have to search the {{Information}} template's source field and that currently only seems possible via the insource search oeprator workaround (see thread).

Given new AI coding tools, maybe implementing this via some new gadget and/or query wouldn't be super difficult so I thought I'd just put it here even though the potential applications aren't yet very clear and probably not nearly as large as those of many other things talked about in the archives of this discussion page. Prototyperspective (talk) 14:27, 21 February 2026 (UTC)

You can do external links with quarry and just assume that every external link is a source, but you would have some false positives. Alternatively you can use SDC, but i dont know how complete SDC is when it comes to sources, and linking SDC to categories is hard. For small categories you can use the api (extimageinfo) but that doesn't scale. Doing top uploader of a category is easy. Bawolff (talk) 06:31, 23 February 2026 (UTC)

For example, https://commonswalkabout.org/?lang=en&f.P180=Q22667&cf=P7482 gives a list of sources of railway images. The results aren't super interesting which suggests to me that the dataset is fairly incomplete. Bawolff (talk) 06:41, 23 February 2026 (UTC)

Oh, i guess sdc is also hard to query due to the complex (and quite frankly inconsistent) schema with wide usage of qualifiers. Anyways here is a slightly better SDC query for sources of train images https://w.wiki/HxSN Bawolff (talk) 07:01, 23 February 2026 (UTC)

That query is interesting, thanks, but it only displays sources of a small fraction of files which were imported by bots (and shows these bots' source sites). I just checked a bunch of files and they nearly all did not have the SD set. Moreover, this really is about scanning files in a category and while this is possible with the Commons query service (I've used this to set the language of files on Wikidata using the Commons categories about language), afaik one can't scan also subcategories of the selected category (see also).

So I think the approach with Quarry has more potential / is better suited for this. Scanning external links is not necessarily worse than checking just the source field. However, a big problem there is that it would be biased by files that link to many URLs. Maybe this could be mitigated by counting only one link per file. However, isn't it possible to search for URLS between |source= and the next | (probably via regex)? Prototyperspective (talk) 14:28, 23 February 2026 (UTC)

Quarry does not have access to article texts so can't do anything with them. Commons blazegraph can do subcategories (via federated query to the wikidata category server), but i think it can only get access to which category an image is in via federated query to mediawiki api, which makes things more difficult, especially for large categories. Bawolff (talk) 16:03, 23 February 2026 (UTC)

Attempt using sparql federation [22]. Seems to work - can find sources and do subcategories, but i suspect it will have issues on latge categories. Bawolff (talk) 18:07, 23 February 2026 (UTC)

example with WDQS aggergating into bar chart. Only works on small-ish categories. Bawolff (talk) 19:31, 23 February 2026 (UTC)

Woa, very interesting – so it seems there is a way. I thought one would need a second tool, maybe used by some gadget, that takes the data and extracts the URLs per item which can then be analyzed and be returned as chart...but it seems the second query already does that too.

That this only works in very small categories of course more or less negates its usefulness however, but there must be some way around that...maybe running it somehow iteratively/batch-wise (possibly like these 1000 batches) or by making some exception/permission where this query can run for longer or something. Studies could be based on this query and we could have lots of interesting stats and insights or use-cases for this but it would first need to work on larger categories that have more files and/or subcategories. Prototyperspective (talk) 00:40, 24 February 2026 (UTC)

Tech News: 2026-09

Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.

Weekly highlight

• Reference Check has been deployed to English Wikipedia, completing its rollout across all Wikipedias. The feature prompts newcomers to add a citation before publishing new content, helping reduce common citation-related reverts and improve verifiability. In A/B testing, the impact was substantial: newcomers shown Reference Check were approximately 2.2 times more likely to include a reference on desktop and about 17.5 times more likely on mobile web. [23]

Updates for editors

• The InterwikiSorting extension, which allowed for the sorting of interwiki links, has been undeployed from Wikipedia. As a result, editors who had enabled interwiki link sorting in non-compact mode (full list format) will now see links reordered. The links moving forward will be listed in the alphabetical order of language code. [24]
• Later this week, people who are editing a page-section using the mobile visual editor, will notice a new "Edit full page" button. When tapped, you will be able to edit the entire article. This helps when the change you want to make is outside the section you initially opened. [25][26]
• The Reader Experience team is inviting editors to assess whether dark mode should still be considered "beta" on their wiki, based on their experience of how well it functions on desktop and mobile. If the feature is deemed mature, editors can update the interface messages in MediaWiki:skin-theme-description and MediaWiki:Vector-night-mode-beta-tag to indicate that dark mode is ready and no longer considered beta.
• The improved Activity tab which displays user-insights is now available to all users of the Wikipedia iOS app (version 7.9.0 and later). Following earlier A/B testing that showed higher account creation among users with access to the feature, it has been rolled out to 100% of users along with some updates. The Activity tab now shows your edited articles in the timeline, offers editing impact insights like contribution counts and article view trends, and customization options to improve in-app experience for users.
• View all 21 community-submitted tasks that were resolved last week. For example, a bug that prevented DiscussionTools from working on mobile has now been fixed, restoring full functionality. [27]

Updates for technical contributors

• The Global Watchlist lets you view your watchlists from multiple wikis on one page. The extension that makes this possible continues to improve. The latest upgrade is the inclusion of a new hook, ext.globalwatchlist.rebuild, which fires after each watchlist rebuild. This allows you to run gadgets and user scripts for the Special page. [28]
• Detailed code updates later this week: MediaWiki

Tech news prepared by Tech News writers and posted by bot • Contribute • Translate • Get help • Give feedback • Subscribe or unsubscribe.

MediaWiki message delivery 19:00, 23 February 2026 (UTC)

Sub-sub-category is somehow appearing on parent-parent category

If you look at Category:Hilary Knight, you can see Category:Hilary Knight in 2010 as a direct subcategory. If you go to Category:Hilary Knight in 2010, you will not find category:Hilary Knight. I don't understand how this is possible. This isn't a caching issue, as Category:Hilary Knight in 2010 was made months ago. How is this happening? —Justin (koavf)❤T☮C☺M☯ 15:59, 24 February 2026 (UTC)

@Koavf: I did a null edit on Category:Hilary Knight in 2010, and at least for me it appears to be fixed. - Jmabel ! talk 17:58, 24 February 2026 (UTC)

Fixed it is. Wild since I purged both of those categories twice. So odd... —Justin (koavf)❤T☮C☺M☯ 18:22, 24 February 2026 (UTC)

This section is resolved and can be archived. If you disagree, replace this template with your comment. —Justin (koavf)❤T☮C☺M☯ 18:22, 24 February 2026 (UTC)

—Justin (koavf)❤T☮C☺M☯ 18:22, 24 February 2026 (UTC)

How to develop a tool like Flickr2Commons?

I'm thinking of maybe giving development of a Study2Commons tool a try (upload images from an individual study) based on the code of Commons:Flickr2Commons.

So I downloaded the code from bitbucket but what now? How can I launch and test things – is there a simple beginner guide for this kind of development somewhere? I don't have my own server and it would eventually be deployed on toolforge I guess. Prototyperspective (talk) 18:50, 24 February 2026 (UTC)

EXIF date import conditional?

Hey, not sure if this is a bug or intentional so thought I'd ask first (and couldn't find this asked previously). Is it intentional that the "Date work was created or first published" field is auto-filled by the Upload Wizard if at the start the file was marked as mine for copyright purposes, vs. it is not auto-filled if it's marked as "I have permission from my employer to upload"? I would think it would scrape the metadata either way. Thanks. ErinAtTCDDigitalCollections (talk) 09:54, 25 February 2026 (UTC)