Jump to content

Wikipedia:Village pump (technical)

From Wikipedia, the free encyclopedia

This is an old revision of this page, as edited by 172.254.222.178 (talk) at 11:56, 20 April 2022 (→How to stop web scraper sites?: Re.). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

 Policy Technical Proposals Idea lab WMF Miscellaneous 
Shortcuts
The technical section of the village pump is used to discuss technical issues about Wikipedia. Bug reports and feature requests should be made in Phabricator (see how to report a bug). Bugs with security implications should be reported differently (see how to report security bugs).

If you want to report a JavaScript error, please follow this guideline. Questions about MediaWiki in general should be posted at the MediaWiki support desk. Discussions are automatically archived after remaining inactive for five days.


Preventing wrapping with references

From the short discussion here. It seems references can wrap in a somewhat unsightly way at line breaks. (Try resizing the window on User:Ovinus/sandbox, and I can provide screenshots if that would be helpful.) I'm on macOS and it occurs in Firefox, but not in Chrome. Is there any way to prevent this without {{nowrap}} and the like? Ovinus (talk) 01:43, 15 April 2022 (UTC)[reply]

@Ovinus: No, and see Wikipedia talk:Manual of Style#Footnotes and line wrapping issue. {{nowrap}} should not be used for this. --Redrose64 🌹 (talk) 22:22, 15 April 2022 (UTC)[reply]

Can we control images shown by Wikipedia app when searching?

I found an iOS Wikipedia app to test a report at Talk:Anal sex#Safe search options. Sure enough, a good way to find Anaesthesia is to type "ana" in the Search Wikipedia box. Doing that shows a list of matches and the second item is Anal sex, helpfully illustrated with the lead image. That is, typing "ana" causes the app to show File:Wiki-analsex.png. Is there a way to mark an image as not suitable for display during a search? Has this been discussed before? I found August 2015. Johnuniq (talk) 01:45, 15 April 2022 (UTC)[reply]

It boggles the mind that MediaWiki:Bad image list isn't supported here. —Cryptic 02:05, 15 April 2022 (UTC)[reply]
Johnuniq, this only work on Minerva I think? The majority of thumbnails seems to be useless anyway due to high compression. Another example WMF developers implementing stuff without giving it second thought. (sadly I've seen this way too often) Do you think "Search ignores MediaWiki:Bad image list, showing NSFW to children looking for Anaheim, California" is a catchy name for a Phabricator task? I thought it was. — Alexis Jazz (talk or ping me) 11:45, 15 April 2022 (UTC)[reply]
Thanks! The original report concerned the Wikipedia App but seeing it on a desktop with the Minerva skin is even more impressive. Johnuniq (talk) 23:17, 15 April 2022 (UTC)[reply]

Recent software upgrade

Resolved

In regard to this recent software upgrade, I've opened an inquiry at mw:Talk:MediaWiki 1.39/wmf.7#Template stopped working on enwiki, because at least one template has stopped working: {{If mobile}} no longer renders text in mobile view. This should be fixed and I have no idea how. Makes me wonder what else has been broken? P.I. Ellsworth - ed. put'r there 02:46, 15 April 2022 (UTC)[reply]

This week's deployment made a change to how we remove some things from content HTML on mobile. Specifically, as it affected this template, we now only outright-remove things with the nomobile class in content namespaces. If_mobile/styles.css used a selector to control the visibility of the mobile-only text that relied on that complete-removal (it basically said "if I'm the second element here, hide me"). I went in and edited the template's CSS so that it does something less fragile, though because of TemplateStyles' restrictions it's not going to be 100% equivalent as there's no perfect proxy available for "the MobileFrontend transforms ran". DLynch (WMF) (talk) 16:25, 16 April 2022 (UTC)[reply]
The {{If mobile}} template has begun to work again, so I will mark this "resolved". P.I. Ellsworth - ed. put'r there 06:27, 17 April 2022 (UTC)[reply]

Janie Fricke

Can someone help me with the {{infobox person}} on Janie Fricke? For some reason, the {{infobox musical artist}} content isn't displaying inside of it and I can't figure out for the life of me what's going on. Ten Pound Hammer • (What did I screw up now?) 04:32, 15 April 2022 (UTC)[reply]

 Fixed The problem was that the {{ubl}} listing her marriages was never properly closed and swallowed the rest of the infobox. * Pppery * it has begun... 04:35, 15 April 2022 (UTC)[reply]

Speedy deletion category counts

Hey, Village Pump Technical folks,

I wish I knew of a better place to request this. But the CSD category counts are off again, at least on 7 or 8 of the categories that indicate that they contain pages but that are actually empty. I'm not sure what to do when this happens, is there a "clear" or "reset" code you can run that would make them accurate again? This has been going on for weeks but now it's happening to more than one category so that has prompted this request for help. Thanks for any tech help you can supply! Liz Read! Talk! 05:10, 15 April 2022 (UTC)[reply]

As of mid-February, purging them is supposed to force a recount. That hasn't fixed any of Category:Candidates for speedy deletion as blatant NOTWEBHOST violations, Category:Candidates for speedy deletion as copyright violations, or Category:Candidates for speedy deletion as vandalism (as displayed on {{CSD-categories}}) for me. —Cryptic 06:19, 15 April 2022 (UTC)[reply]
All three of those had rows in categorylinks without corresponding rows in page (quarry:query/63804). I was able to figure out which page was the problem for CAT:U5 from cl_sortkey_prefix, and restoring and redeleting User:Dronar (Composer)/sandbox fixed it. For the others, well, the only way I know how to get a page title of a deleted page from its old page_id is slow. Working on it. —Cryptic 06:46, 15 April 2022 (UTC)[reply]
It worked for the first three pages I tried. No luck with Draft:The economic state of the PIGS after the financial crisis, previously known as page_id 70437379 - unlike the other three, restoring it gave it a new id. No idea why. I'm not going to try for the seven in Category:Candidates for speedy deletion as spam, lest I break things further. —Cryptic 07:10, 15 April 2022 (UTC)[reply]
Keeping maintainance categories up to date is possible, but the issue is that the developers have thought the issue is about every single category on the wiki. You would have to be an Wikipedian to read between the lines to see that is not the expectation. There is (obviously) more effort to keep every single category updated with an small delay compared to only the maintainance ones. I clarified this in phab:T221795.--Snævar (talk) 07:41, 15 April 2022 (UTC)[reply]

Bad UX at Special:Upload: it allows you to upload badly-licensed files and tags them for speedy, instead of blocking the upload

(I don’t know if that’s the right place to ask, tell me if you have a better idea.)

One uses Special:Upload to upload image and other media files to en-wp without putting them on Commons. This is used mostly for non-free files that comply with WP:NFCC, which have no free license compatible with Commons.

In that interface, there is a long dropdown list of licensing information. Some of the options are "I do not know the license", "The copyright holder gave me permission to use this work only on Wikipedia articles", etc. Those options have the effect of adding a big red template and (I suppose) listing the file for speedy deletion. Here’s an example upload. We had a question recently at the Teahouse of a new user who uploaded a file under one of those bad licensing options, then their file got speedy-deleted (no surprise here).

To me, that seems like bad user experience for the newbies, and more workload for the deleting admins. A license valid "only for Wikipedia articles" is a non-fixable problem so there is little point in posting the image, the probability of deletion after that is 100%. ("I do not know the license" is different, those might be salvageable.) Why not reject the upload with a an information page / popup / whatever telling that the file cannot be uploaded? (The warning template does get previewed in the upload page, but no newbie reads templates anyway.) TigraanClick here for my talk page ("private" contact) 09:54, 15 April 2022 (UTC)[reply]

The problem is that the audience just doesn't understand copyright and we can't explain it to them any further then we already do on that page. But ppl are 'completionists'. They have a goal (I want to use this image) and want to fulfil that goal, irrespective of rules (esp rules that most other platforms do not have). So by giving them an honest answer to fill in, we avoid them lying and simply choosing something that is going to lead to license washing. This makes it easier on admins, as they don't get a ton more images which are incorrectly licensed but harder to filter out from the properly licensed material. The interface for Special:Upload is currently 'dumb', so while preventing the upload is possible, someone would have to write JS to intercept the form submission and explain it to the user (after which the user is still likely to try again and lie btw). —TheDJ (talk • contribs) 10:24, 15 April 2022 (UTC)[reply]
Even "The copyright holder gave me permission to use this work only on Wikipedia articles" might be salvageable. An owner might release a work with a WP-compatible free licence, an editor might ask the owner (through either politeness or ignorance) "is it OK if I use that on Wikipedia?" and select that option, when wider permission already existed. Certes (talk) 11:03, 15 April 2022 (UTC)[reply]

Category:Pages which use score

This seems to be an automated maintenance category about the score extension, which should be bluelinked and turned into a hidden category? Or the software should be changed to give it a better name? Posting here because I lack the technical background knowledge to write a proper category description page. —Kusma (talk) 15:47, 15 April 2022 (UTC)[reply]

Created a basic desc page. – SD0001 (talk) 16:32, 15 April 2022 (UTC)[reply]

Load variable in global.js

Hi, I've come across a strange behavioral difference between common.js on the local wiki and global.js on meta. Let's say I have in my common.js on enwp the following code: 

mw.loader.load('http://127.0.0.1:5500/wikipediatest.js'); // Web server 
var testConfig = 'a';

Then, if the script on the web server has the following code 

console.log(typeof testConfig === 'undefined');

it returns false. This means the variable is loaded from common.js.

But, if the code is in global.js on meta, it returns true. Any idea why this happens and how to successfully load variables from global.js, just like from the local common.js? Any help would be appreciated.--Dragoniez (talk) 19:34, 15 April 2022 (UTC)[reply]

@Dragoniez: You had the variable declared in both your common and global.js at the same time, not? NguoiDungKhongDinhDanh 20:13, 15 April 2022 (UTC)[reply]
It appears it's because global.js is run in its own scope while common.js is run in the global scope. Try window.testConfig instead. Nardog (talk) 22:40, 15 April 2022 (UTC)[reply]
Thank you so much for your comments, both of you two. @Nardog: that was it! I was completely stuck, so you saved me a lot of my time. Much obliged. @NguoiDungKhongDinhDanh: I also appreciate your help. Now my script works, and I know you're a global rollbacker, so you might want to check out my new script for rollback (m:User:Dragoniez/Selective Rollback). --Dragoniez (talk) 03:07, 16 April 2022 (UTC)[reply]
@Dragoniez: I haven't try it yet, but I'm pretty sure that isInGlobalGroup('Global_rollback') will always return false since the group, among others, was renamed in 2014. NguoiDungKhongDinhDanh 03:16, 16 April 2022 (UTC)[reply]
Also, why isInArray()? The built-in method Array.prototype.includes exists for that purpose. NguoiDungKhongDinhDanh 03:21, 16 April 2022 (UTC)[reply]
@NguoiDungKhongDinhDanh: I actually didn't (and in fact couldn't) test that because I don't have the flag but "Global_rollback" seems to be working in m:User:Hoo man/smart rollback.js. I'd have to wait for a feedback from someone. As for Array.prototype.includes, it's just for the sake of browser compatibility (because the method is a ES7 feature). I could also just use $.inArray, but I don't think this is something that must be criticized. --Dragoniez (talk) 03:30, 16 April 2022 (UTC)[reply]

@Dragoniez: That script was mostly written in 2010–2013 and has become outdated. It is still useful, but I'd suggest a complete rewrite. NguoiDungKhongDinhDanh 04:05, 16 April 2022 (UTC)[reply]

A toolforge tool doesn't open for me

Hi all, I have some serious issues trying to open [wikinav.toolforge.org]. It simply shows me a blank white screen with nothing to look at. It's been happening ever since I first came across this site about 3–4 weeks ago. So far, this is the only toolforge site that acts as such. As several editors have pointed out to clickstream data from this site during the course of RMs, I think it'll be very helpful to me to have access to it. Any help to get it opened would be appreciated. Thanks! —CX Zoom[he/him] (let's talk • C • L) 11:16, 16 April 2022 (UTC)[reply]

I get the screen but "An error occurred while fetching data for the current title. Try another one.". Trying another one doesn't help. Shame, it's been a very useful tool for me too. Certes (talk) 12:06, 16 April 2022 (UTC)[reply]
Just turn off your adblocker for toolforge.org domain. – SD0001 (talk) 16:11, 16 April 2022 (UTC)[reply]
Thanks but that didn't seem to help, and nor does using wget to avoid browsers completely. Certes (talk) 16:40, 16 April 2022 (UTC)[reply]
I never keep ad blocker on. Also, other toolforge tools seem to be working for me. —CX Zoom[he/him] (let's talk • C • L) 17:05, 16 April 2022 (UTC)[reply]
It works on most pages for me. I'm having problems with https://wikinav.toolforge.org/?language=en&title=Nazir_Ali but this may because the page was moved during the month currently reported (March 2022). https://wikinav.toolforge.org/?language=en&title=Nazir_Afzal and https://wikinav.toolforge.org/?language=en&title=Mercury work correctly for me, so I'm representing the space properly and it's not a problem specific to dabs. (Turning on my adblocker can break either of those, depending how I configure it.) Certes (talk) 17:17, 16 April 2022 (UTC)[reply]
My issues are with the whole tool. All of the above-mentioned links (and each one I come across in RM discussions) are just a white sheet of paper for me. —CX Zoom[he/him] (let's talk • C • L) 17:20, 16 April 2022 (UTC)[reply]
Open your browser console (see Wikipedia:JSERROR #6) and share a screenshot (on imgbb.com or somewhere else). This would help identify if there's any extension that's blocking HTTP requests, or if there is any JS code your browser is unable to parse. – SD0001 (talk) 05:58, 17 April 2022 (UTC)[reply]
Why not share the screenshot here? See third bullet on this page's editnotice, and WP:WPSHOT. --Redrose64 🌹 (talk) 06:53, 17 April 2022 (UTC)[reply]
What is your browser and operating system? Izno (talk) 21:31, 16 April 2022 (UTC)[reply]
Thanks all. Apparently this was an Android-only issue (bulk of my activity is on Android). It works just as described by Certes above on dad's Windows PC that I had access to on weekends. Still, this one tool remains the only toolforge tool that doesn't work out on Android. —CX Zoom[he/him] (let's talk • C • L) 08:37, 17 April 2022 (UTC)[reply]
@CX Zoom it took a little browsing to find, but it looks like you can report issues with that tool here. — xaosflux Talk 10:17, 17 April 2022 (UTC)[reply]
Thank you very much Xaosflux. —CX Zoom[he/him] (let's talk • C • L) 19:48, 17 April 2022 (UTC)[reply]

Hello,

I am having an issue with the efn template. The following code renders text:

{{efn|Also on the [https://digi.kansalliskirjasto.fi/sanomalehti/binding/573741 30 April 1892 program] were, among others, compositions by [[Armas Järnefelt]] (''Lyrical Overture'', 1892), [[Robert Kajanus]] (''Finnish Rhapsody No. 1'', 1881), [[Fredrick Pacius]] (''[[Maamme|Vårt land]]'', 1848), and [[Martin Wegelius]] (''Daniel Hjort Overture'', 1872).}}

The following code, which is preferred because it would take the reader to the proper page of the PDF, does not render text (difference is highlighted):

{{efn|Also on the [https://digi.kansalliskirjasto.fi/sanomalehti/binding/573741?term=Sibelius&term=Kullervo&page=2 30 April 1892 program] were, among others, compositions by [[Armas Järnefelt]] (''Lyrical Overture'', 1892), [[Robert Kajanus]] (''Finnish Rhapsody No. 1'', 1881), [[Fredrick Pacius]] (''[[Maamme|Vårt land]]'', 1848), and [[Martin Wegelius]] (''Daniel Hjort Overture'', 1872).}}

Any ideas on what I am doing incorrectly? Thanks, ~ Silence of Järvenpää (talk) 18:56, 16 April 2022 (UTC)[reply]

The one that doesn't render has an = sign which makes everything ahead of it a parameter name. Try
{{efn|1=Also on the [https://digi.kansalliskirjasto.fi/sanomalehti/binding/573741?term=Sibelius&term=Kullervo&page=2 30 April 1892 program] were, among others, compositions by [[Armas Järnefelt]] (''Lyrical Overture'', 1892), [[Robert Kajanus]] (''Finnish Rhapsody No. 1'', 1881), [[Fredrick Pacius]] (''[[Maamme|Vårt land]]'', 1848), and [[Martin Wegelius]] (''Daniel Hjort Overture'', 1872).}}
—Trappist the monk (talk) 19:05, 16 April 2022 (UTC)[reply]
@Silence of Järvenpää: Since the url contains =, you need to give the first parameter a name, like this: {{efn|1=etc.}}. NguoiDungKhongDinhDanh 19:10, 16 April 2022 (UTC)[reply]
Trappist the monk, NguoiDungKhongDinhDanh ... that solution works. Thanks! ~ Silence of Järvenpää (talk) 19:16, 16 April 2022 (UTC)[reply]

Match Infobox parameter

I want to do something like 

{{#invoke:string|match|s={{:David Brearley}}|pattern={{!}} *successor *= *([^%n]*)}}

but that gives String Module Error: Match not found ({{#invoke:string|match}} seems to substitute all templates recursively). ― Qwerfjkltalk 21:08, 16 April 2022 (UTC)[reply]

When a module meets some transcluded content inside itself, that content is expanded first and then the module finishes running. I would suspect your error is looking for {{!}} instead of its character representation. Izno (talk) 21:33, 16 April 2022 (UTC)[reply]
{{#invoke:string|match|s={{:David Brearley}}|pattern=successor *= *([^%n]*)}} also throws an error. Qwerfjkl's right: All templates are expanded recursively before matching. See ExpandTemplates result for {{#invoke:string|match|s={{:David Brearley}}|pattern=<table class="[%w ]+"}}. NguoiDungKhongDinhDanh 21:40, 16 April 2022 (UTC)[reply]
NguoiDungKhongDinhDanh: substituting it, however, works, which is fine for my purposes. Qwerfjkltalk 21:52, 16 April 2022 (UTC)[reply]
@Qwerfjkl: Not really. It gives [[Robert Morris (judge)|Robert Morris]]\n| pro. NguoiDungKhongDinhDanh 21:57, 16 April 2022 (UTC)[reply]
I suspect that's due to a flaw in my Lua pattern, rather than the principle. Qwerfjkltalk 22:02, 16 April 2022 (UTC)[reply]
Yes, %n isn't a special sequence in Lua (yet) and just matches the letter n in "pronunciation". \n would match a newline. Certes (talk) 22:18, 16 April 2022 (UTC)[reply]
Also, %n in Lua means the nth match, and you have no previous matches isn't special and just matches "n", but (for the reasons Izno just stated) it doesn't work with \n either. An alternative way to do this is with {{Template parameter value|David Brearley|Infobox officeholder|1|successor}} = Robert Morris. Certes (talk) 21:41, 16 April 2022 (UTC)[reply]

Wikipedia enhanced

Hello!

Yesterday we had a meeting with high school students in Albania discussing about Wikipedia in general, its usage by them, and its pros and cons. There were different interesting remarks that were made by the students which gave their opinions on SqWiki and EnWiki but there were 2 proposals which interested me personally from a technical point of view.

A lot of students said that they struggled a lot identifying the needed information on Wikipedia. The overall idea is that they had different questions and they wanted answers and their sources fast, not to learn about the topic as a whole. Quite some times Encyclopaedia Britannica was brought as an example for text formatting with their 2-min summaries and the top questions some articles had. They referenced WW1 as a historical article often (they generally said that they used Wikipedia mostly for its history articles) so I believe what they had in mind was this: WW1 - Britannica. In it you can notice all the aforementioned elements, the fast questions/fast answers part, the possibility for short summaries and even articles only loading the next section once you've scrolled down enough. If you compare it with our WW1 article here you'll start to see what they're talking about when you see the length of its sections (or the total number of its sections). Some of them did say that the way Wikipedia was organized wasn't disorienting because the table of contents and its sections were enough to not get lost in text but many said that they basically wanted sublimed information according to their needs.

As mentioned, there were 2 proposed solutions for this by the students themselves:

  1. Create a technical way that can highlight text according to what you search for. Technically speaking we're talking about an enhanced version of the find tool most browsers have. Its function would be to find where the part of information you're looking for is located in the text by what keyword you provided on it and autofocus on that paragraph or couple of paragraphs highlighted in yellow.
  2. Create different versions of Wikipedia according to the users needs/preferences, the same as we do with different languages. In this scenario we would have different versions for the cat article. One would be, let's say, for children in primary schools, one for students in high school and one for students studying zoology. Even more versions can be implemented.

What are everyone's thoughts on these 2 proposals? How feasible would be to implement any of those or variations of them in practice? - Klein Muçi (talk) 12:33, 17 April 2022 (UTC)[reply]

1. In the code editor you can highlight a word and it highlights all other times that word shows up.
2. An kid version has been rejected several times on meta, for example meta:Wikipedia Kids. There is an non-WMF kids article website at vikidia.org. An NSFW (Not safe for work) filter would be more feasible, and there is some filtering in place on this wiki for that purpose.--Snævar (talk) 13:03, 17 April 2022 (UTC)[reply]
@Snævar, thank you for your insight! In regard to the first proposal, the idea was for the tool to be intelligent enough as to understand the meaning of the provided word and autofocus on the part of the text that gives the best answer in aspect to that, a part which should be well outlined where it starts and where it ends. This mechanism is already being applied by Google in its queries, mostly with Wikipedia and YouTube. If you search for something, you don't only get a link where the key word you provided is located. You also get a passage from Wikipedia highlighted in bold which Google thinks is the most suitable part of information of what you're looking for according to the keyword provided. The same happens with videos where Google offers a timed YouTube video with a well-defined start and end period in which it thinks what you're looking for is located. They would just like to have these tools inside Wikipedia to query information from within it, without having to utilize Google or other search platforms for those. Google also offers the ability to "answer" directly asked questions to it, something Britannica also provided in a sort of way. Again, something they would like to be implemented here. So, better said, they would like to make querying information from within Wikipedia better with ways such as those mentioned above.
As for the "kid version", thank you for the link as I hadn't seen it before. Also I didn't know about the NSFW filtering so I'd be happy to know more about it if you can provide more information. How does it work exactly? What are you referring to? The main idea in this aspect can be thought in 2 ways I believe:
  1. Judging by the arguments provided in your meta link, if we are to roll only with 2 versions, language and simple language (English and simple English), then, if we want to scale that globally we would need to also make way for simple Albanian, simple Greek, simple Chinese, etc.
  2. More elegant but way harder to implement: Create a tool that can pre-format existing articles in different modes according to one's needs. No new wiki needs to be created, you just choose what version want to have your article provided to you. With my limited technical knowledge I believe such a system could be created by setting up a lot of hooks/anchors/tags in articles according to the different versions we want to implement similar as to how multi-language projects such as Mediawiki and Commons apply translation tags to their pages. There, of course, may be more elegant ways to approach this. - Klein Muçi (talk) 15:13, 17 April 2022 (UTC)[reply]
    Writing for different levels of experience/language comprehension is more than a language pass over individual sentences; completely different article structures may be appropriate, and different breakdowns of topics into different sets of articles. For this and scalability reasons, I think separating the source for each version is a better approach, including giving each version its own edit history. Thus at present I think separate wikis would be preferable. isaacl (talk) 15:57, 17 April 2022 (UTC)[reply]
    I was actually picturing my scenario with hooks/anchors/tags to be what you actually describe as completely different article structures may be appropriate, and different breakdowns of topics into different sets of articles. It would be hard to acquire 3 or more different "working staffs" for 1 language (or get 3 different wikis out of the incubator) while on the other hand, it seems easier in my eyes to have 1 staff write the same thing in 3 or more different way.
    But maybe my hooks/anchors/tags scenario can't actually provide what I say it can so maybe the multiple-wikis approach is better. - Klein Muçi (talk) 16:36, 17 April 2022 (UTC)[reply]
    For example, if article A-grade 3 has one set of headings, and article A-grade 8 has a completely different set of headings, then I think the two articles would essentially have to be in separate parts within the same source file. (There might be some re-use of citations.) Topic X might have a more rich breakdown into articles X1, X2, and X3 at a grade 8 level, but be rolled up as a section of article W at a grade 3 level. Putting everything into one source file imposes a tax on everyone to maintain all language versions, whether or not they are interested in contributing towards this goal. The combined edits across variants would make diffs harder to read, and would make it harder to remove and re-apply changes for a specific variant. I can't envision how to make this work in practice in a scalable manner. isaacl (talk) 20:23, 17 April 2022 (UTC)[reply]
    Okay but practically speaking how much luck would someone have in persuading the global community in creating a dozens of versions of Wikipedia in each language all of a sudden? So far we haven't had even a simple-another-language beside simple-en, let alone have multiple versions of each language. If we are to make this idea be just "create new wikis from scratch" I don't believe it will go far ahead, especially after considering how many times the wikikids ideas have been rejected. - Klein Muçi (talk) 01:49, 18 April 2022 (UTC)[reply]
    @Klein Muçi: Actually, Basque Wikipedia do have a special namespace where articles are simplified for kids to read: eu:Txikipedia:Azala. NguoiDungKhongDinhDanh 02:52, 18 April 2022 (UTC)[reply]
    It's similar to when someone says "I want to start a WikiProject to do X": the first question is "do you have a community of editors who want to do X?" So first we should look for editors interested in creating a Wikipedia variant for a specific audience. Once a community is identified, ways to facilitate the work can be planned (which might not necessarily be hosted by Wikimedia; if there is a community of interested Wikipedians, though, it should be easier to get a Wikimedia project approved). But I don't think it would be effective to have articles that are a union of all variants for different reading levels and different audiences (of which there are infinite combinations: zoology students, sociology students, historians, artists, legislative staff who are trying to write laws and so need broad understanding of many areas, etc.) into one source file. Separate files with separate edit histories for each is more manageable. isaacl (talk) 04:15, 18 April 2022 (UTC)[reply]
    @Isaacl, you're right on that. This was discussed together with concepts such as m:Education so maybe it would be better if I further discussed it in its corresponding talk page. This would also help with the "find a community part". - Klein Muçi (talk) 09:39, 18 April 2022 (UTC)[reply]
    @NguoiDungKhongDinhDanh, interesting concept. I hadn't thought of utilizing namespaces up until this point. But if you want to have different namespaces for different kinds of publics, not only for kids, the more you create, the more it starts looking like you're building either portals or wikiprojects. - Klein Muçi (talk) 09:48, 18 April 2022 (UTC)[reply]
  • Not sure if this is "technical" (yet) - but perhaps some sort of "reference desk" integration would help with part of this? — xaosflux Talk 14:48, 17 April 2022 (UTC)[reply]
    @Xaosflux, the reference desk was what I was thinking about as well when they were entertaining the questions and answers idea. That of course would help, the more subtle it was, the better. Something similar to how the mentorship system works: You press the ? symbol on a page and get a writing square where you can write your question and then hopefully you get the answer soon without having time to realize that you're asking real humans. Maybe you can also get pre-written answers automatically if the said question was already asked and answered before. Eventually a lot of answers would be provided automatically and that space could also have a an auto-updating FAQ section in regard to that.
    I emphasize the subtleness part because people usually want to be discreet when asking questions.
    Unfortunately, all I described above wouldn't work at all in small wikis like SqWiki which can't sustain a reference desk considering their very limited human resources but having such a system implemented on EnWiki could be a good starting point. - Klein Muçi (talk) 15:25, 17 April 2022 (UTC)[reply]
  • The first proposal sounds like Google's featured snippets feature (which I see you mentioned in a followup reply). Are you suggesting that Wikipedia's in-house search capability be enhanced to be able to understand posed questions and figure out snippets to highlight? Given that this is a tricky problem that big companies like Google, Amazon, Apple, and Microsoft have invested considerable resources on solving, it might not be the most effective area for Wikipedia to focus on. The second proposal needs interested communities to support each separate version, just like the different language versions. isaacl (talk) 15:37, 17 April 2022 (UTC)[reply]
    @Isaacl, apparently, yes. Thank you for the link! I didn't remember its name and I didn't know about the yellow highlight it provided on-page once you followed that snippet. I usually get all my information from the snippet provided on Google and I have rarely followed through. That's where the students I mentioned must have gotten the idea also. Having such a function would basically cut down the work for the reference desk in here and make it possible for use even on small wikis.
    As for the second part, is that all it is needed though? Would I be able to just go to the incubator right now and start a new project (following its guidelines) that could potentially grow to be "Wikipedia high school"? (Of course not with that name. :P ) Do you think it could be more feasible if we had an extension written for that that utilizes tags to auto-format article to the user-chosen version? - Klein Muçi (talk) 15:57, 17 April 2022 (UTC)[reply]
    Unless Wikipedia can do in-house semantic search better than what external search engines do, I don't see a strong motivation to do it, and I don't even think it can get close with the disparity in resources invested.
    As I responded to another one of your comments, I think separate wikis are best. isaacl (talk) 16:02, 17 April 2022 (UTC)[reply]
    @Isaacl, there are again 2 implementation plans here I can think of:
    1. Make Wikipedia behave the same as Google. Ideally speaking, you search in the search box, you get shown the snippet of information you want to read about or better yet you are sent directly to it, with the snippet part autofocused and highlighted. (Same thought can be extended for multimedia smart querying because there can be miracles when you believe... ♪♫♪♫)
    2. A minor version of the one above. The snippet function only works after you have already chosen an article. You have a built-in search function within its article which works the same as above, autofocusing and highlighting text according to your query.
    This of course depends on if Wikipedia can actually provide that functionality, as you said. Any idea who can provide further insight from a tech-viewpoint of the current search function?
    Sorry for that. I hadn't seen that comment. Will reply there. - Klein Muçi (talk) 16:23, 17 April 2022 (UTC)[reply]
    The key question is why not just use external search engines, which have already done (and continue to do) extensive research and AI model training for turning your query into target pages and snippets within them? This is not a trivial problem, and the search engines collect much more data that they can use to refine their models. (No apologies necessary; I was just cross-referencing my response to avoid duplication.) isaacl (talk) 20:11, 17 April 2022 (UTC)[reply]
    @Isaacl, I guess it depends on the specific workflow. From what it was said to me I believe they searched for WW1 on Google, came here, saw an enormous article, got scared by it and said "if only there was a way to quickly find the answer to my question about WW1 without having to skim through all this text". Some thought "if only Wikipedia could have summaries or FAQs about articles like Britannica" and some others said "if only we could have an enhanced way to search and find the right text snippets we want".
    Of course they can get similar results with their wishes by better utilizing search engines. Or by being better at choosing key words and just CTRL-F-ing the article. And I'm sure many people already do precisely those things. But apparently, there are other people who intuitively chose this other workflow and asked if it could be given assistance in their way. If that's impossible I guess we'll have to live with the current version. Klein Muçi (talk) 01:24, 18 April 2022 (UTC)[reply]
    I'm pretty sure the vast majority of users ask their questions on external search engines (whether directly through a web/app interface, or through a voice interface), rather than Wikipedia's. There's no way for Wikipedia to build a similar semantic search system on its own, without getting all that search query data to improve its algorithm. It would be more effective to integrate external search providers into Wikipedia. isaacl (talk) 03:58, 18 April 2022 (UTC)[reply]
    @Isaacl, any ideas how such an integration would look like in practice? It has worked very well with Content Translation Tool so maybe it would be a great idea if it can be achieved even on this aspect. - Klein Muçi (talk) 09:51, 18 April 2022 (UTC)[reply]
    I'm not thinking of anything at all like Content Translation Tool so we may be talking about different things. I'm thinking of a feature (perhaps a gadget) that would let you configure a supported external search provider and then provide an easy way to access it. It could completely replace existing search boxes or add an option to let you switch between the internal engines and external ones. isaacl (talk) 15:09, 18 April 2022 (UTC)[reply]
    No, I believe we are on the same waters. I mentioned CTT because it provides its function by standing on the shoulders of Google Translate and Yandex. At the same way we could provide a better search provider by utilizing the help of major search engines like Google, etc. What you describe looks like a very nice idea. Ideally speaking, I believe such a feature/gadget would have a nice interface that feels organic with Wikipedia (not clunky) and, maybe, the results that it brings are also brought back here again and are filtered to only be related to Wikipedia. Does that seem like an accomplishable deed to you? - Klein Muçi (talk) 15:52, 18 April 2022 (UTC)[reply]
    In terms of user interface, everything is possible given enough effort. However all projects also have opportunity cost. So I don't personally have a feel of the relative tradeoffs. isaacl (talk) 16:02, 18 April 2022 (UTC)[reply]

Main ideas after the discussion

  1. Reference desk integration with articles;
  2. External search provider on-wiki;

If anyone can help by providing the first steps on how such ideas can be implemented practically it'd be welcome. - Klein Muçi (talk) 01:48, 19 April 2022 (UTC)[reply]

Password reset

Hi! I have no plans to password reset, I'm just confused how can we change the password of our Wikipedia account if we don't specify our email address. In Special:PasswordReset, you must provide either a Username or Email address to reset your password, what happens if you only provide a username, where you can recieved the emails? —Princess Faye (my talk) 01:27, 18 April 2022 (UTC)[reply]

You can change your password at anytime if you know your password at Special:ChangePassword. If you forgot your password, the only supported method for reset is email. — xaosflux Talk 01:40, 18 April 2022 (UTC)[reply]
To directly answer the question, you specify the email address at Special:Preferences#mw-prefsection-personal-email or when you create your account. When you go to Special:PasswordReset and specify a username, it looks up the e-mail address listed there and sends an email to it. If you specify an e-mail address, it looks up the username that matches email address and sends a password reset email for it (I'm not sure what happens if there is more than one such username). * Pppery * it has begun... 01:45, 18 April 2022 (UTC)[reply]
@Xaosflux and Pppery: If your Wikipedia account does not have an email address, and you forget the password, and you go to Special:PasswordReset to change the password, and since you don't have an email address on your account, you just entered the username… you won't receive an email and you still can't change the password, right? All I know is that if you enter either of the two in Special:PasswordReset, username or email address, you will still receive an email. —Princess Faye (my talk) 02:19, 18 April 2022 (UTC)[reply]
Because I'm wondering if your Wikipedia account doesn't have an email address, anyone can change its password because there is no email address. —Princess Faye (my talk) 02:31, 18 April 2022 (UTC)[reply]
Or if you forget your Wikipedia account password, and there is no email address attached to that account, and you can't provide an email address in Special:PasswordReset, the results, you can no longer change your password, correct? —Princess Faye (my talk) 02:31, 18 April 2022 (UTC)[reply]
Well, yes, but actually no. Sysadmins can change it for you if you ask. NguoiDungKhongDinhDanh 03:05, 18 April 2022 (UTC)[reply]
Only if they have definite, absolute proof that the person asking for password change is really who they claim to be. --Redrose64 🌹 (talk) 06:19, 18 April 2022 (UTC)[reply]
Like a committed identity, I assume, although Help:Logging in#What if I forget my password? says "(Rare) If you have a long-established account and have previously set up a 'committed identity', ... it may be possible to recover the account" (emphasis added). Nardog (talk) 06:35, 18 April 2022 (UTC)[reply]
@Nardog and Redrose64: If you are logged out of all Wikipedia projects and literally have no email address on your Wikipedia account, and you also don't have an email address to enter in case you go to Special:PasswordReset, but you just put your username on Special:PasswordReset. My question is, if Wikipedia will send an email for the Password reset, where it will be received if what you provided is just your username? —Princess Faye (my talk) 07:49, 18 April 2022 (UTC)[reply]
If there is no email address set, then it doesn't send any email. The user will still see the same confirmation message as if an email had been sent, this is deliberate to prevent bad faith actors being able to get information about an account. See phabricator:T238961 and the code. the wub "?!" 09:29, 18 April 2022 (UTC)[reply]
@The wub: Is an account safe if it doesn't have an email address?? and what else is the disadvantage of not having an email address on the Wikipedia account other than not being able to reset the account password? —Princess Faye (my talk) 09:45, 18 April 2022 (UTC)[reply]
Without an email address, there is no way to reset a lost password (other than sysadmin intervention already mentioned, and that shouldn't be relied upon). Without providing an email address you also won't be able to send or receive emails through the Email this user function. the wub "?!" 12:49, 18 April 2022 (UTC)[reply]
As long as you don't forget your password, your account is relatively safe (2FA is recommended however) even without email. NguoiDungKhongDinhDanh 13:08, 18 April 2022 (UTC)[reply]
2FA is not recommended for standard users (nor is it even available) - as there is no supported 2FA reset process for broken authenticators and lost scratch codes. — xaosflux Talk 13:40, 18 April 2022 (UTC)[reply]
┌───────────────────────────┘
It should be. The risk of losing scratch codes and the authenticator is trivial comparing to that of being compromised by third parties. NguoiDungKhongDinhDanh 14:53, 18 April 2022 (UTC)[reply]
Yes, a sysadmin has the technical capability to reset (which they usually would do by (re)setting your email address and then having you do a reset), but that is not supported in our production environment. That doesn't mean it won't or has never happened - but no one should ever rely on it. — xaosflux Talk 09:50, 18 April 2022 (UTC)[reply]
Perhaps this summary will help:
  • If you know your password,
    • and you have e-mail set up, you can use or change your password whenever you want.
    • and you don't have e-mail set up, you can still use or change your password whenever you want.
  • If you lost your password,
    • and you have e-mail set up, you can get a new password.
    • and you don't have e-mail set up, you cannot get a new password, unless some very rare circumstances apply.
      • As an example of those rare circumstances, if a WMF employee lost account access, they could get the WMF office to reset their password. If this happened to me, it would probably require a video call with someone in the office who already knows me plus a note from my manager.
      • For most people, this generally requires being personal friends for years with someone who has the necessary user rights. It is not at all something you should be relying on. If you don't already have an established way to contact one of those people off wiki (e.g., via that e-mail address you haven't set up...), then it won't work.
For the most part, unless you have a valid, working e-mail address set up, if you lose your password, then you should assume that you have lost that account and go to Special:CreateAccount. Your first edit can be setting up a userpage that explains that you lost your password, just like all these editors did. Whatamidoing (WMF) (talk) 16:54, 18 April 2022 (UTC)[reply]
Tracked in Phabricator
Task T109909

Edit summary

At the bottom of the edit window in the text editor, there are some edit summary choices that can be selected instead of typing a custom summary. These pre-defined choices are namespace dependent. For article, there are things like "add/improve refs" or "removing unsourced content". For Talk pages, it's "Comment", "Reply", "Suggestion". For Drafts, it's the same choices as on Talk pages. I think editing a Draft is more like editing an article and the article choices should be available. Can this be changed? MB 02:25, 18 April 2022 (UTC)[reply]

Yes, by tweaking the DefaultSummaries gadget. Feel free to request on its talk page. NguoiDungKhongDinhDanh 03:01, 18 April 2022 (UTC)[reply]
MediaWiki talk:Gadget-defaultsummaries.js#Interface-protected edit request on 18 April 2022 --Ahecht (TALK
PAGE) 15:18, 18 April 2022 (UTC)[reply]

strange issue at WP:UAA/Bot

See my comments at Wikipedia talk:Username policy#weird bug Beeblebrox (talk) 18:40, 18 April 2022 (UTC)[reply]

Transcluding <noinclude>

On my creative-venture wiki (and its Referata forebears), I've long had trouble with the <noinclude></noinclude> wraparound on a preloaded template of mine. What I call "Template starter" sets up the bare bones for creating a template, adding a doc-page section that is always hidden from view outside a template's page à la Wikimedia.

With my latest attempt yesterday, I followed the instructions provided at mw:Manual:Creating pages with preloaded text#Loading the preload file (<no<includeonly></includeonly>include></noinclude>), but instead of the expected: 

<!-- Please replace this with template content --><noinclude>{{doc tag|categories=<includeonly>Templates</includeonly>}}</noinclude>

all I'm getting is just the comment after loading: 

<!-- Please replace this with template content -->

I've tried getting this to work time and again, to the point where spaces before the tag name (< noinclude></ noinclude>) must do as a compromise in our current setup. Of course, that's not really what I even wanted to deal with; I simply want to get the proper <noinclude>s on the page without any hassle. How else to go about it?

(MW 1.37.2; code loaded via Extension:Preloader)

(Crossposted from mw:Topic:Wtumv01tz8qmndq8 [Support desk thread], which has seen no replies/further activity at press time)

--Slgrandson (How's my egg-throwing coleslaw?) 19:12, 18 April 2022 (UTC)[reply]

@Slgrandson, this works: 
<!-- Please replace this with template content --><<noinclude />noinclude>{{doc tag|categories=<<noinclude />includeonly>Templates<<noinclude />/includeonly>}}<<noinclude />/noinclude>
There are <noinclude /> tags (equivalent to <noinclude></noinclude>) inside each tag to prevent the other tags from being used, until transclusion (or preload). Qwerfjkltalk 20:51, 18 April 2022 (UTC)[reply]
@Qwerfjkl: Even with that improvement, it still doesn't take effect via Preloader--all of the raw code makes it onto the target page, spaces or not. Upstream bug on Preloader's part? (Keep in mind it's marked as "unmaintained", the last update having occurred in June 2018. Original co-developer Robchurch (talk · contribs) hasn't been seen on Wikimedia since 2019 [his last edit coming after a decade-long hiatus (!)], so to partner @Troyengel: Can you please look into this issue/investigate once you're available?) --Slgrandson (How's my egg-throwing coleslaw?) 23:44, 18 April 2022 (UTC)[reply]
As it turns out, my suspicions were justified: it is an upstream problem dating back to at least April 2021. Per word from @Troyengel:
I have no way to test this (this project is looking for a new maintainer, I haven't used it in years) and this change is significant enough in design/technique that it needs someone to test it before we merge. If you're aware of someone else who could help validate, send them this way!
Maybe Phabricator can take up the case from here as well?
--Slgrandson (How's my egg-throwing coleslaw?) 23:55, 18 April 2022 (UTC)[reply]
Or better yet, someone from GitLab who can finish this merge/fix sooner rather than later. --Slgrandson (How's my egg-throwing coleslaw?) 00:14, 19 April 2022 (UTC)[reply]

Tech News: 2022-16

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

Changes later this week

  • The new version of MediaWiki will be on test wikis and MediaWiki.org from 19 April. It will be on non-Wikipedia wikis and some Wikipedias from 20 April. It will be on all wikis from 21 April (calendar).
  • Some wikis will be in read-only for a few minutes because of a switch of their main database. It will be performed on 19 April at 07:00 UTC (targeted wikis) and on 21 April at 7:00 UTC (targeted wikis).
  • Administrators will now have the option to delete/undelete the associated "Talk" page when they are deleting a given page. An API endpoint with this option is also available. This concludes the 11th wish of the 2021 Community Wishlist Survey.
  • On selected wikis, 50% of logged-in users will see the new table of contents. When scrolling up and down the page, the table of contents will stay in the same place on the screen. This is part of the Desktop Improvements project. [1]
  • Message boxes produced by MediaWiki code will no longer have these CSS classes: successbox, errorbox, warningbox. The styles for those classes and messagebox will be removed from MediaWiki core. This only affects wikis that use these classes in wikitext, or change their appearance within site-wide CSS. Please review any local usage and definitions for these classes you may have. This was previously announced in the 28 February issue of Tech News.

Future changes

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

23:10, 18 April 2022 (UTC)

50% of logged-in users will see the new table of contents for all skins? — GhostInTheMachine talk to me 09:29, 19 April 2022 (UTC)[reply]
That means 50% of will see the new style, and 50% will see the old style. However, I suspect that it only affects people using New Vector. SGrabarczuk (WMF) would know. Whatamidoing (WMF) (talk) 20:06, 19 April 2022 (UTC)[reply]
Thanks @GhostInTheMachine. This only applies to selected wikis and those who use Vector 2022. Desktop Improvements are Vector 2022 and nothing else. SGrabarczuk (WMF) (talk) 22:38, 19 April 2022 (UTC)[reply]

Lots of tools down

If you see any tools broken over the next hour or so, it's probably due to some database maintenance that's going on right now. -- RoySmith (talk) 15:16, 19 April 2022 (UTC)[reply]

Why does the first link work correctly (outside of ref tags) but not the second one? RandomCanadian (talk / contribs) 17:13, 19 April 2022 (UTC)[reply]

@RandomCanadian: I'm really not sure about the why but I've made an edit to User:RandomCanadian/sandbox6 that demonstrates the workaround. -- John of Reading (talk) 17:32, 19 April 2022 (UTC)[reply]
For those curious, the problem is that the ref was emitting the HTML 
<span class="reference-text"><a rel="nofollow" class="external text" href="http://www.twse.com.tw/pdf/en/{{{1}}}_en.pdf">Consolidated Information of Listed Company {{{1}}}</a>, Taiwan Stock Exchange Corporation</span>
instead of the intended 
<span class="reference-text"><a rel="nofollow" class="external text" href="http://www.twse.com.tw/pdf/en/2603_en.pdf">Consolidated Information of Listed Company 2603</a>, Taiwan Stock Exchange Corporation</span>
notice how the two {{{1}}} are left literally instead of being replaced by the parameter value.
This is phab:T4700, some of which is covered by Help:Footnotes#Formatting ref tags and mw:Help:Cite#Substitution and embedded parser functions. --Redrose64 🌹 (talk) 19:47, 19 April 2022 (UTC)[reply]

Petscan

Why this link stopped working and other categories still works? Eurohunter (talk) 20:46, 19 April 2022 (UTC)[reply]

The link seems to work for me. I get 57 results starting with Madagascar–Mauritius relations. Database maintenance affected tools earlier; perhaps PetScan broke temporarily. Certes (talk) 22:49, 19 April 2022 (UTC)[reply]

Adding external font-face in template css

Feel free to move my question if it isn`t the right place.

How can I add an external font-face in template css, for example for Google fonts like this: <link rel="stylesheet" type="text/css" href="//fonts.googleapis.com/css?family=Open+Sans" />. If I add this link, the page won`t let me save. רונאלדיניו המלך (talk) 21:48, 19 April 2022 (UTC)[reply]

You can't, it's not allowed. --Redrose64 🌹 (talk) 22:17, 19 April 2022 (UTC)[reply]
Not sure what file you're trying to modify—CSS files don't have elements denoted by tags (and thus no <link> element). If you are trying to modify a CSS file being used by a template, note as per mw:Manual:CSS#Using the url() function, you can't load fonts from non-Wikimedia servers. For fonts available from Google Fonts, though I suspect it won't work, you can try using the Toolforge font server proxy to load the desired font file. (I haven't tried it other than within my personal CSS file, where it can be used.) isaacl (talk) 22:21, 19 April 2022 (UTC)[reply]
Using a <link /> element in this manner is valid in HTML (it doesn't go in a stylesheet), it loads an external font that may then be referred to by a declaration for a font: or font-family: property that is present in a stylesheet that is itself loaded from the same HTML document. But the tag is not whitelisted, and so the MediaWiki software does not permit such use. --Redrose64 🌹 (talk) 22:38, 19 April 2022 (UTC)[reply]
Yes, I know it's valid HTML and that it's not listed as something the MediaWiki software supports. I didn't think the software would block a wikitext-based page save with a <link /> element, though, so I wasn't sure if the original poster was referring to a CSS file that they were trying to modify. isaacl (talk) 22:46, 19 April 2022 (UTC)[reply]
@Isaacl: The whitelist for font contexts is empty, so that won't work. Anomie⚔ 22:42, 19 April 2022 (UTC)[reply]
Hmm, so does that mean the described method in mw:Manual:CSS#Using the url() function to use font families provided through the Universal Language Selector extension won't work either? (I realize those instructions aren't specifically for TemplateStyles and so more restrictions can apply.) isaacl (talk) 22:50, 19 April 2022 (UTC)[reply]
Indeed more restrictions apply to TemplateStyles than to site CSS. That's one of them. But you could try requesting a configuration change to add ULS font URLs to the whitelist I mentioned if you want. Although better in that case might be to get ULS to detect the need to supply the font itself rather than bypassing it. Anomie⚔ 11:39, 20 April 2022 (UTC)[reply]

Font size too big in geonotices

There is a huge geonotice at the top of my watchlist. Anybody willing to look into reducing the font size of geonotices? There is a <style> tag somewhere that has the CSS #watchlist-message .geonotice span {font-size: 144.5%;} that I suspect is the cause. Screenshot of CSS. Thanks. –Novem Linguae (talk) 22:07, 19 April 2022 (UTC)[reply]

This has been discussed before. Basically, when set to normal font size, people were simply not noticing the notices. Also, please don't send me off to imgur (my PC slows to a crawl, my virus scanner complains, and I need to reboot), we have a perfectly satisfactory system ourselves. --Redrose64 🌹 (talk) 22:21, 19 April 2022 (UTC)[reply]
@Redrose64 sorry, but our screen shot system onwiki sucks for this sort of purpose. It is very understandable why, but that doesn't make it not suck. This is not an endorsement for imgur, just that for tech troubleshooting, dealing with our necessarily complex file hosting rules are a pain for many contributors. The way it works on phab is immensely better. — xaosflux Talk 10:17, 20 April 2022 (UTC)[reply]

How to stop web scraper sites?

I brought this up a few times last year, though nothing really materialized, but am hoping to raise this issue once again. I was just wondering if there is anything other than the spam blacklist or XLinkerBot to prevent citations to scraper sites from being used here? It seems that the area of recent deaths is where these are most prevalent, as there is just so much crap out there now that seems to scrape certain terms from Twitter (and probably Wikipedia) into cartoonishly amateur bot-generated "news articles" to generate clicks. Here are some previous discussions: 1 2; any further thoughts are greatly appreciated. Connormah (talk) 05:22, 20 April 2022 (UTC)[reply]

Lookup the dns names, find their owners, blast owners on twitter ? —TheDJ (talk • contribs) 08:40, 20 April 2022 (UTC)[reply]
Technically, not much can be done. The normal countermeasures such as captcha or passwords etc. obviously are not feasible. Refusing access based on IPs or other reverse DNS information is complex, resource-intensive and temporary. Non-technical solutions such as TheDJ suggested are probably the only readily available defense. 172.254.222.178 (talk) 11:56, 20 April 2022 (UTC)[reply]

Template R to diacritic

Can someone experienced with redirect templates take a look at this discussion here? - Klein Muçi (talk) 11:51, 20 April 2022 (UTC)[reply]

Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(technical)&oldid=1083738342"