Wikipedia:Village pump (technical)

From Wikipedia, the free encyclopedia
  (Redirected from Wikipedia:Vpt)
Jump to navigation Jump to search
 Policy Technical Proposals Idea lab WMF Miscellaneous 
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).

Newcomers to the technical village pump are encouraged to read these guidelines prior to posting here. 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.

Frequently asked questions (FAQ) (see also: Wikipedia:FAQ/Technical)
Click "[show]" next to each point to see more details.
If something looks wrong, purge the server's cache, then bypass your browser's cache.
This tends to solve most issues, including improper display of images, user-preferences not loading, and old versions of pages being shown.
No, we will not use JavaScript to set focus on the search box.
This would interfere with usability, accessibility, keyboard navigation and standard forms. See task 3864. There is an accesskey property on it (default to accesskey="f" in English). Logged-in users can set a gadget in their preferences.
No, we will not add a spell-checker, or spell-checking bot.
You can use a web browser such as Firefox, which has a spell checker.
If you have problems making your fancy signature work, check Wikipedia:How to fix your signature.
If you changed to another skin and cannot change back, use this link.
Alternatively, you can press Tab until the "Save" button is highlighted, and press Enter. Using Mozilla Firefox also seems to solve the problem.
If an image thumbnail is not showing, try purging its image description page.
If the image is from Wikimedia Commons, you might have to purge there too. If it doesn't work, try again before doing anything else. Some ad blockers, proxies, or firewalls block URLs containing /ad/ or ending in common executable suffixes. This can cause some images or articles to not appear.
For server or network status, please see Wikimedia Metrics.
« Archives, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187

Skip to top
Skip to bottom

Stub creator tool[edit]

Sometimes I come across a notable topic that we don't have an article for, and I want to quickly create a stub and seed it with some references, but it's a topic that I don't have experience creating articles about, so I don't know how these articles are "supposed" to look or what they're "supposed" to have.

I'm dreaming of a tool that will let me quickly create a stub by having templates for various topics, where I can choose the topic, fill out some basic information (like infobox parameters and a paragraph of prose) and references, and it will create the page with the correct infobox, navboxes, categories, create the talk page with the right WikiProject tags, etc. Basically, select the article topic (e.g. athlete, scientist, album, book, car, town, school, etc.), fill out a form, click, and it makes a stub with all the right parts. Editors could create new templates for new topics, or custom/personalized/alternate templates, kind of like custom warnings in Twinkle.

Does anything like this exist? Levivich harass/hound 08:05, 17 February 2021 (UTC)

Levivich, don't think such a tool exists, but would be cool. Copy this post to WP:SCRIPTREQ unless you get a response here. (There unanswered requests don't usually get archived, so someone could see it and write a script some time.) – SD0001 (talk) 13:28, 18 February 2021 (UTC)

Levivich Are you looking for something like Template:Biography? Slywriter (talk) 17:28, 21 February 2021 (UTC)

@Slywriter: Yes! Thank you! That's basically what I was looking for... except one of those for various different topic areas. Do you know if we have other templates like that for other topics? I poked around the category tree a bit and can't seem to find a category for that particular kind of template (I don't even know what you call it... an "article template"? An "article generator template"?). What I was looking for was a tool that had a bunch of those kinds of templates so I can select the topic I'm creating an article about and it'll load the appropriate template (or form or whatever) for me to fill out. Knowing a template already exists is a huge help (especially if/when I take up SD0001's advice about posting at SCRIPTREQ)... thanks for pointing me to it! Levivich harass/hound 17:34, 21 February 2021 (UTC)
Levivich,(sorry if ping not needed), so a search for exact text "sample layout" lead me to WP:MOS/Layout, which is handy guide but no samples and another for US Fed legislation Wikipedia:WikiProject_United_States_Federal_Government_Legislative_Data/Proposed_layout. Beyond that, I only found a handful of images showing layouts. Also a review of the history of the user who created the biography (15 years ago) doesn't show them creating/editing any others beyond infoboxes. Slywriter (talk) 18:04, 21 February 2021 (UTC)
@Slywriter: (pings are great on busy pages!) Thanks for looking. (I'm pretty sure there's also a song, album, or musician template in one of the wikiprojects, to add to the list.) I guess the script writing might be the easy part; the hard part is writing all the templates, which requires knowing the best practices and MOSes for all the topics. I guess it wouldn't be so tough to find a recent FA for each topic and convert it into a template like {{biography}} (and then run the template by the appropriate WikiProject); maybe that's something I'll chip away at. Levivich harass/hound 18:14, 21 February 2021 (UTC)
Levivich, this may or may not be helpful,
My quick review (only went through A-C) show as some examples aviation has detailed guides but no templates, baseball has a fork of biography, and chemistry has simpler guides, no templates.
{{Aerostart}} {{Aerobiostart}} {{Airlinestart}} ? — GhostInTheMachine talk to me 20:22, 21 February 2021 (UTC)
Perhaps straying too far from your goal but does seem like a missed opportunity not to have layouts readily available as part of the MOS/layout article and then branching from there into the various categories (Arts, Biographies, History, Science etc) and subcategories(painting/book, actor/football).
Anyway, seems like much of the information is somewhere, just not well-organized to direct a new editor towards nor for an experienced editor to find without significant work. Slywriter (talk) 18:51, 21 February 2021 (UTC)
@Slywriter: Not really straying too far from the goal; having the model layouts just as MOS sub-pages (and listed somewhere in the MOS) would be helpful just by itself, and is probably more than half the battle... from there, creating subst-able templates, or a script, is really just incremental convenience. Levivich harass/hound 19:01, 21 February 2021 (UTC)
WikiProject style advice is one place, we do also have specific subMOSs for some topics with varying detail. Example MOS:VG has a short version 'here's what this should look like' and longer form 'here's what goes in each section'. --Izno (talk) 19:52, 21 February 2021 (UTC)

Visual editor copies ref to top of page[edit]

Hi all, does anyone know why visual editor seems to do this from time to time? I didn't click anywhere near the top of the article and just was expanding the ref. I don't have any other examples off hand, but I have noticed this happening every now and then. Rather odd. --TheSandDoctor Talk 16:36, 17 February 2021 (UTC)

TheSandDoctor, Wow, that's weird. I've never seen it do that myself. Do you have some other examples? -- RoySmith (talk) 17:04, 17 February 2021 (UTC)
@RoySmith: Definitely a weird one. Unfortunately not off hand or that I could find in around a 30 minute search. I know that I have seen it before (typically IPs notice and correct lol), but it seems to happen every few hundred to few thousand edits. Special:Diff/1002506672 is the closest I could find, where it inputted random text at the start of the article. --TheSandDoctor Talk 18:24, 17 February 2021 (UTC)
TheSandDoctor, You've got under 500 edits that have been reverted. Maybe look over that list and see if any ring a bell? -- RoySmith (talk) 18:37, 17 February 2021 (UTC)
@RoySmith: Good idea. Found one on Taylor Swift Special:Diff/988510759 --TheSandDoctor Talk 18:52, 17 February 2021 (UTC)
TheSandDoctor, Thanks. Were these both instances of you having copied the completed cite template from somewhere and pasted it into the article? The next time this happens, could you please set your time machine to t minus 5 minutes and video record your earlier editing session? -- RoySmith (talk) 19:02, 17 February 2021 (UTC)
@TheSandDoctor: What's your web browser/operating system, and do you find that you "lose focus" after some things? (Like you know where the cursor is supposed to be, and then you click on something, and the cursor seems to be nowhere)? Whatamidoing (WMF) (talk) 00:32, 23 February 2021 (UTC)
@Whatamidoing (WMF): I was using Google Chrome on Windows 10. In the Twitter article case, all I did was update the ref and then click apply & then save the edit. As for the Swift article, it was too long ago for me to remember. --TheSandDoctor Talk 03:42, 23 February 2021 (UTC)
@TheSandDoctor, thanks for your quick response. I realize this is a tiny detail, but do you happen to remember if you needed to click the button twice to get the ref, or to generate it twice? Whatamidoing (WMF) (talk) 18:01, 23 February 2021 (UTC)
@Whatamidoing (WMF): Could you please clarify the question? Do you mean whether it took two clicks to have the "edit" button show up? If so, that sometimes happens but do not know if it happened in this case, unfortunately. If not, please elaborate and I will answer as best I can. --TheSandDoctor Talk 18:14, 23 February 2021 (UTC)
No, I was wondering if you needed to click one of the "Cite" buttons twice. But I think it will be a wild goose chase. Whatamidoing (WMF) (talk) 16:48, 24 February 2021 (UTC)
Potential example from my watchlist. CMD (talk) 06:11, 24 February 2021 (UTC)
It happened to me last night in Firefox, and an editor fortunately undid the whole thing instead of correcting it, which meant that I saw it. (It's not a page that I routinely check the history for). The key detail in both my and @Chipmunkdavis's diff is that it's replacing the first template at the top of the wikitext with the citation template.
I think we have enough information to file a bug report. I'll post the bug number here when I've got it. Whatamidoing (WMF) (talk) 16:50, 24 February 2021 (UTC)


I'm randomly logged out since few deys (monday?). What is the problem this time? Why this problem keeps coming back from time to time? Like atleast 1-2 periods a year. Eurohunter (talk) 18:26, 17 February 2021 (UTC)

@Eurohunter: Sometimes that is due to security and everyone is logged out when a bug is found (that was once or twice last year), other times it could be cookie related. Did you clear your cookies recently? — Preceding unsigned comment added by TheSandDoctor (talkcontribs) 18:55, 17 February 2021 (UTC)
@TheSandDoctor: No it's already like I'm log out few times in hour now. It looks like it log out me but when I refresh I'm loged in again. I never clear my cookies when it's happening. It's not the first time. Eurohunter (talk) 21:08, 17 February 2021 (UTC)
Eurohunter, We've heard this a couple of times over the last few days here.. the reports and the fact that they happen very incidentally, but still somewhat 'often' to some people, make it sound like maybe in some regions some people are accidentally hitting a server which is having problems with login sessions.. .. I advice filing a report in Phabricator. —TheDJ (talkcontribs) 09:10, 18 February 2021 (UTC)
@TheDJ: Interesting. This problem has been seen at least several times in the last few years. Do you have idea what could cause this has returned? Eurohunter (talk) 16:51, 18 February 2021 (UTC)
Eurohunter, there are a 1000 different potential causes for login problems. It keeps coming back because it is a problem that people actually recognise, unlike many other problems that happen each and every day with similar causes. Its recognition bias. —TheDJ (talkcontribs) 08:40, 19 February 2021 (UTC)
I am going through my watchlist, and just now I clicked a diff link on that and found I was shown the page with "Not logged in" and a "Log in" link at the top; inside one second the "Central login" popup appeared. Checking my system, I find that I presently have a flaky connection to my router, so it might have dropped a packet at some point resulting in an incomplete cookie. returning to my watchlist, it displayed as normal and other diff links also worked fine. --Redrose64 🌹 (talk) 16:48, 18 February 2021 (UTC)
For what it's worth, I've been having similar experiences repeatedly over the past several days. I log in, it shows that I've logged in, and then when I go to another page, I get logged out. After I log back in, I seem to stay logged in for the rest of the session. It looks as though when I first log in, the log in doesn't fully "stick". I'm definitely not clearing anything from cookies or browsing history to cause it. --Tryptofish (talk) 19:25, 20 February 2021 (UTC)
It just happened to me again, so I paid close attention to what happened. I logged in successfully. I then went to my watchlist and clicked to see the combined new changes on one of the pages that I watch, and when I got there, it said that I was logged in centrally, but that I needed to re-log in for en-wiki. --Tryptofish (talk) 20:19, 22 February 2021 (UTC)
It shouldn't say that you "needed to re-log in for en-wiki". This is what it should say:
Central login
You are centrally logged in as Tryptofish. Reload the page to apply your user settings.
which should be in a little box at upper right, just below the search box, which itself is just below the "Not logged in / Talk / Contributions / Create account /Log in" links. Usually, I find that reloading the page stubbornly shows me as not logged in; but following a link from the page to a different page works. --Redrose64 🌹 (talk) 21:19, 22 February 2021 (UTC)
Thanks for the reply. Yes, you are exactly correct about how it said it. Obviously, this isn't the worst thing in the world, but it's been happening to me a lot lately, and seems more like a bug than a feature. --Tryptofish (talk) 22:48, 22 February 2021 (UTC)

The French Wikipedia's birthdate[edit]

Extract from a web page showing a list of dates, each associated with a web link and a numeric value.
First Wikipedia revisions history.

the 15th of January 2001 is widely recognized as the Wikipedia birthdate. What about the French Wikipedia's birthdate ?

On the 11th of May 2001, posted on the Wikipedia mailing list, a message advertised the creation of But in the introduction of the article French Wikipedia, 23 March 2001 is mentioned as the starting date of the french edition of Wikipedia (until the 20th of July 2006, the introduction displayed August 2001 as the birthdate. On the 28th of January 2019, the 23 March 2001 date appeared with a citation (24th of September 2010) from ZDNet France). The article's french version does mention the same date, but with a "citation needed" warning.
The oldest version of the french home page is available on Internet Archive (IA) and dates back to the 19 of may 2001. Unfortunately, the archive only shows the 4th revision. On, the oldest revision of the home page dates back to the 11th of October 2002.
I've wondered whether frWiki could have been mistaken for the french version of Nupedia. According to IA, a Nupedia French Language Translation Project was open since at least the 9 April 2001. A thread of messages (IA archive) from Francais-L mailing list, hosted on, shows however that the french Nupedia was not online before the 1st of April 2001 (see messages' content on IA).
On Wikipedia-l mailing list, the 16th of March 2001, Jimmy Wales wrote:
"I want to set up some alternative language wikipedias. French and German would be good [...] I intend to setup the following domain names and wikis:, (both pointing to the same thing)".
On the 16th of April 2001, on the Francais-L mailing list, Larry Sanger posted the following message (IA archive):
"Another thing to consider is to encourage Jimmy Wales [...] to create a French Wikipedia, under the title I've already asked him to do this for all the foreign languages we're translating into on Nupedia, but I think Jimmy is busy these days--so if you encourage him, he'll create it soon, I am guessing. (If you were to tell him that you have an article, in French, that you want to add, that would probably persuade him.)"
This confirms that was not available before April 2001.

On, the first revision of the known oldest article, Paul Héroult, dates back to the 4th of August 2001. On IA, a 19th of May 2001 version is available. Some revisions are apparently missing on

The to change took place on the 2nd of November 2002.

Questions: what is the accurate French Wikipedia's birthdate ? Does a reliable source exist about it ? Could relevant Wikipedia revisions be retrieved ?

Note: this message may not be considered a technical issue, so, please, feel free to move it into a Village pump section you think is more appropriate.

--ContributorQ (talk) 18:55, 18 February 2021 (UTC)

@ContributorQ: it might help to ask them, they have an page for non-french questions here: w:fr:Wikipédia:Bistro_des_non-francophones/enxaosflux Talk 19:03, 18 February 2021 (UTC)
Hello @Xaosflux, I am a regular french Wikipedia editor. On Le Bistro (17 January 2021) (the french Village pump), the question has been discussed, but no one has provided a decisive convincing proof. --ContributorQ (talk) 19:41, 18 February 2021 (UTC)
This is a good question that has come up on occasion (not only French). A common table of established dates would be very useful indeed, it would then be possible to do things like {{NUMBEROF|fr|birthdate}}, which could be added to a new column in List of Wikipedias and many other places ({{NUMBEROF}} has been exported to 70+ wiki langs). The date table should be open to change and discussion, I suspect Commons Tabular is better than Wikidata for this purpose, possibly a page on Wherever easily machine readable and commonly available. -- GreenC 19:15, 18 February 2021 (UTC)
@ContributorQ: I did a couple of database queries and the earliest edit in the current French Wikipedia database is this one to Utilisateur:BernardVatant and dates to 1 June 2001; the [[earliest deleted edit is ironically to Wikipédia:Nupedia Translation Project and dates from 8 June of that year. Graham87 11:16, 19 February 2021 (UTC)

@ContributorQ: Double-pinging you because I think I've struck gold here! This directory of database dumps contains a file called wiki-fr.tar.gz. Inside that gz archive under the "/wiki-fr/lib-http/db/wiki" directory, the first line says "990174692³3HomePage³3*³30³3217.14.192.xxx³30³3id³2111". 990174692 is a Unix timestamp and is equivalent to 18 May 2001, 08:31:32 (UTC/GMT). The line also indicates that the edit was made by the IP address (the last octet of IP addresses was always obscured by UseModWiki). The corresponding entry in the difflog file does indeed sound like it could be the French Wikipedia's very first homepage edit. I will place the original text at fr:Utilisateur:Graham87/HomePage, for want of a better location, because (a) it relates to the French Wikipedia and (b) I don't know how to format it within this message. Frwiki editors can do whatever they like with my subpage.

This directory contains a script by Tim Starling for porting the very early English Wikipedia dumps to a more modern format. Adapting the script to the French Wikipedia database dump would probably take a lot of work and is beyond my skill level. Graham87 11:57, 19 February 2021 (UTC)
Also see this message on Wikipedia-l, which I found from following footnotes 45–46 at History of Wikipedia. Graham87 12:17, 19 February 2021 (UTC)
There is a lot of relevant history in the August 2001 English Wikipedia database dump at the title "International Wikipedia". Despite the gaps in history, I have imported it all to its current location, Wikipedia:Multilingual coordination; the relevant edit about the French Wikipedia and several others is this one. Also note that the Wikipedia-l thread above begins on 11 may 2001, not 18 May. I can't explain the discrepancy; perhaps the subdomains existed but had no content before 18 May. Graham87 12:44, 19 February 2021 (UTC)
First Wikipedia logo (2001).

@Graham87: Excellent! Thank you very much for your time and the job done. It helps a lot. With the informations you provided, I could retrace a more precise sequence of events.

In the March 2001 Wikipedia-l archive there is no message announcing that a french version of Wikipedia has been created. On the 16th of March 2001, Jimmy Wales posted his wish to create one, a Catalan wikipedia and a German Wikipedia. A few minutes later, he also announced the online availability of A message by JimboWales, on Wikipedia ("Multilingual coordination" meta page), confirms that:
"I have set up Catalan and Deutsch (or Deutsche, that's still up in the air!), and I anxiously expect there to be some problems or complaints as people start to use them[...]
I think that we should have French and Spanish next, but I don't know how to write 'French' and 'Spanish' correctly. If someone will tell me, I will set them up."

On the 19th of March 2001, JimboWales announced the creation of ("Multilingual coordination" meta page). The day after, a confirmation of the online availabality of, and has been published.
On the 11th of May 2001, on Wikipedia-l, Jason Richey announced the creation of the subdomain and eight other new wikis or subdomains (, xx = chinese, esperanto, hebrew, italian, japanese, portuguese, spanish or russian).
On the 13th of May 2001, someone asked for a french version ("Multilingual coordination" meta page).
On the 16th of April 2001, on the Francais-L mailing list, hosted on, Larry Sanger posted the following message (IA archive):
"Another thing to consider is to encourage Jimmy Wales [...] to create a French Wikipedia, under the title I've already asked him to do this for all the foreign languages we're translating into on Nupedia, but I think Jimmy is busy these days--so if you encourage him, he'll create it soon, I am guessing. (If you were to tell him that you have an article, in French, that you want to add, that would probably persuade him.)"
On the 18th of May 2001, in a reply to a request made on the 15th of May 2001 by Jimmy Wales, Jason Richey implicitly announced, on Wikipedia-l, the creation of the alias which was confirmed by Larry Sanger, the same day ("Multilingual coordination" meta page). Three days later, on the "Multilingual coordination" meta page, JasonR announced the creation of

The table below shows data extracted from the frWiki dump (wiki-fr.tar.gz). The timestamps and other data are infos retrieved from revisions logged in the "recent changes" file (rclog) and "diff log" file (diff_log).

From 2002 frWiki dump (
Revision # Timestamp Date Page Content
1 990174692 2001-05-18 8:31:32 Home page This is the home of the new French Wikipedia. Please replace this text with something appropriate en francais, and start writing!
4 990259515 2001-05-19 8:05:15 Home page (see record on IA) This is the home of the new French Wikipedia. Please replace this text with something appropriate en francais, and start writing!
Please see our instructions (sorry, in English) on how to edit a page. (Perhaps these could be translated...
Paul (Louis-Toussaint) Héroult
1 990259529 2001-05-19 8:05:29 Paul Heroult (see history page on IA) [...]
5 991380665 2001-06-01 7:31:05 Home page Cette page est la racine de Wikipedia en Français, que vous êtes invités à faire grandir ...
Please see our instructions (sorry, in English) on how to edit a page. (Perhaps these could be translated...
Voirour instructions
(sorry, in English) pour le mode d'emploi : comment éditer une page.
A suivre ...
1 991380815 2001-06-01 7:33:35 BernardVatant (see on frWiki) Page contact pour Bernard Vatant
Très intéressé par ce projet collaboratif.
J'espère pouvoir y consacrer un peu de temps.
Mes passions : l'Astronomie et le partage des connaissances.
Pour en savoir plus :
Me contacter : bernard@[...]

All collected infos are consistent with the May 2001 birthdate (11th, or 18th); none does tally with the 23 March 2001.

The tables below show data extracted from some pther wikis dumps. The timestamps and other data are infos retrieved from the first revision logged in the "recent changes" file (rclog) and "diff log" file (diff_log).

From 2001 wiki dump (
Timestamp Date Home page's content
979586833 2001-01-15 19:27:13 This is the new WikiPedia!
From 2002 xxWiki dumps (
Wiki Timestamp Date Home page's content
de 984703365 2001-03-16 00:42:45 This is the new German language wikipedia. I obviously need to translate all of the error messages, page text, etc
Advice solicited! Just write here on the homepage for a couple of days, and then we'll get started next week.
I do not speak German. Larry Sanger speaks some. So only English speaking people (who also speak German) are likely to be able to effectively communicate with me.
ca 984776841 2001-03-16 21:07:21 (see history page on IA) This is for the Catalan language wiki project.
ja 985041045 2001-03-19 22:30:45 irrashaimase!
kore wa nihongo no wikipedia desu.
kakite kudasai.
eo 990174584 2001-05-18 8:29:44 This is the home of the new Esperanto Wikipedia. Please replace this text with something appropriate in Esperanto, and get to work!
fr 990174692 2001-05-18 8:31:32 This is the home of the new French Wikipedia. Please replace this text with something appropriate en francais, and start writing!
it 990174783 2001-05-18 8:33:03 This is the home of the new Italian Wikipedia. Please replace this text with something appropriate in Italian, and start writing!
pt 990175390 2001-05-18 8:43:10 This is the home of the new Portuguese Wikipedia. Please replace this text with something appropriate in Portugues, and start writing!
Please see our instructions (sorry, in English) on how to edit a page. (Perhaps these could be translated...)
sv 990611216 2001-05-23 9:46:56 Välkomna till Svenska Wikipedia!
Denna sidan är startsidan (HomePage) för den svenska wikipedian. Målet med Wikipedia är att skapa ett uppslagsverk med uppslagsord av alla möjliga typer där allt innehåll är fritt att använda enligt GNU Free Documentation License.[...]
nl 992991055 2001-06-19 22:50:55 Welkom bij de nieuwe (19/06/2001) Nederlandstalige Wikipedia.
Wikipedia is een gemeenschapsproject met het doel vanaf nul een complete encyclopdie te creëren. De Engelstalige versie is in januari 2001 opgestart en bevat al meer dan 9000 pagina's. We zouden graag voor de Nederlandstalige pagina's een vergelijkbaar succes zien[...]
pl 1001517728 2001-09-26 15:22:08 = Polskia Edycjia Wikipedii =
26 Wrzesnia 2001 postanowilismy wystartowac polska edycje Wikipedii
Zapraszamy wszystkich do tworzenia nowych artykulów i tlumaczen tekstow z wikipedii.

@GreenC: FYI.

PS: It would be nice to import in each current xxWiki database the very first revision of the homepage and the first article.

--ContributorQ (talk) 17:12, 20 February 2021 (UTC)

@ContributorQ: No worries ... glad to help. Nice analysis! The only major things that I would add/modify are that the page histories in the 2002 dumps are probably not complete (see below) and that the edits to the multilingual coordination page were all made not to Meta (which didn't exist in early 2001) but to the actual English Wikipedia. Re the deletion of the history: see Wikipedia:Usemod article histories and an explanation at MeatballWiki; for instance in the French dump, the earliest surviving edit to the Paris article is a very well-developed page dated 26 January 2002 (UTC) (so it is obviously not the actual first edit to that page). The reason the HomePages probably survived so well is that they were relatively lightly edited. There may also be cases where history of a page from, say, May and November 2001 might have survived but the intervening edits in August may not have. The August 2001 English Wikipedia dump is complete but the 2002 dumps for the other languages are almost definitely not. Graham87 18:35, 20 February 2021 (UTC)
@Graham87: You're right, Multilingual coordination was indeed not a meta page. I corrected that.
I think what makes sense here is the coherence between the data extracted from the dumps and the Wikipedia and mailing lists' announcements.
Paul Heroult, the first frWiki's article, has been imported from the french Nupedia version as emphasized by its content (see, for comparison, the 16 Apr 2001 messages from Francais-L, on IA), the timestamp and the uploader's pseudo (an IP: Therefore the first revision of the french HP and the first article seem pretty reliable.
It's a pity however that article histories contain some missing or dubious entries. --ContributorQ (talk) 07:09, 21 February 2021 (UTC)

Source has no page number on crucial page[edit]

I want to use a reference in a source that has page numbers for virtually all the text - but not for the "Authors note" at the beginning, which has some useful information. If I were referring to any other page in the book, the ref would look like[1]:99 (for page 99). I am sure that there must be a way of stating "author's note" instead of a page number, but cannot find it. Any ideas? ThoughtIdRetired (talk) 21:12, 18 February 2021 (UTC)

You can use Template:Cite book#In-source locations and say |at=Authors note. PrimeHunter (talk) 21:56, 18 February 2021 (UTC)
Thanks, that would seem to work, but has a problem if the same source is needed elsewhere in the article (which is highly likely). That means I would have to use the full reference a second time (giving it a slightly different refname) and use the {{rp|99}} code for the page numbers. The problem is that it would be a mess for the reader, as they cannot track all the parts of the article that rely on the one book.
I have just tried again with [2]:Author's note and it seems to give exactly what I want. I have no idea why it didn't work the first time I tried it.ThoughtIdRetired (talk) 22:45, 18 February 2021 (UTC)
@ThoughtIdRetired: This is one of the typical reasons for the use of shortened footnotes. Rp is sometimes the other mechanism used, but I find it ugly. --Izno (talk) 06:23, 19 February 2021 (UTC)
@Izno:I feel that shortened footnotes make it more tedious to work out where the text of an article comes from. You might find the reference more easily, but it is difficult to get a view of how much of the article relies on that work, and it is slower to get back to the point you had read to in the article. I have done quite a bit of trying to work out where some articles are sourced and therefore this is the voice of experience. It's a particular issue if several authors have written more than one work on the subject - I have to make written notes in complex cases. So, the citation method I have chosen may not be pretty, but it scores better on functionality. Obviously, just a point of view.... ThoughtIdRetired (talk) 08:10, 19 February 2021 (UTC)
@ThoughtIdRetired: Typing <ref name="Greenhill 1988">{{cite book |last1=Greenhill |first1=Basil |title=The Merchant Schooners |date=1951 |publisher=Conway Maritime Press |location=London |isbn=0 85177 475 X |edition=1988}}</ref>{{rp|author's note}} gives[1]:author's note. I find shortened refs abominable, I find articles strewn with errors because people get them wrong. DuncanHill (talk) 02:25, 20 February 2021 (UTC)


  1. ^ a b Greenhill, Basil (1951). The Merchant Schooners (1988 ed.). London: Conway Maritime Press. ISBN 0 85177 475 X.
  2. ^ Greenhill, Basil (1951). The Merchant Schooners (1988 ed.). London: Conway Maritime Press. ISBN 0 85177 475 X.

@Izno, ThoughtIdRetired, and DuncanHill: Just to point out the opposite point of view and what I consider to be the advantages of {{sfn}}s. Using {{sfn|Greenhill|1988|p=47}} adds reusable refs to the reflist without having to resort to the <ref name="Greenhill 1988 p47"> malarkey :) which one often comes across. All refs to the same page are automatically grouped together, and you only have to change the page number(s) in the {sfn} to achieve this.[1] Plus, you only have to hover your mouse over the {sfn} to see the basic ref, and hovering again or clicking on it takes you to the {cite book} in the Bibliography, and the back button returns you to where you were.[1] With the even easier <ref>Greenhill 1988 p. 47</ref>, just as commonly found, you have to physically scroll down to the Bibliography and scroll all the way back up again to find where you were.
I agree that screwed-up {sfn}s are annoying: often that's because the editor didn't include |ref=harv in {{cite book}} etc. The few examples I have come across have mostly been of this type. However, you now don't need it any more: CS1/2 copes with it automatically, and a bot has been removing |ref=harv from {cites} for a while.[2] Since I use {sfn}s all the time I find fixing them only takes a few moments, and I haven't come across a bad one for some time. I also agree that {sfn}s and {cite book} can take longer to learn than <ref name="" />, and more slightly more time-consuming to use in an article. Last time I looked the relevant information was scattered over several Help pages with not much attempt to construct a single comprehensive instructive page about using just {sfn}s and {cite book} etc. Things may have improved. Would anyone know how many articles actually use {sfn}s? I know it's a minority.
In order to use sfns properly, I believe the {cite book} params have to be correct as well: the actual edition cited should be in (parentheses), and the original edition in [square brackets]. I would tend to use |year= and |orig-year= which displays slightly differently: {{cite book |last1=Greenhill |first1=Basil |title=The Merchant Schooners |year=1988 |orig-year=1951 |publisher=Conway Maritime Press |location=London |isbn=0 85177 475 X}}.[3] You can use |ref={{harvid|Greenhill|1988a}} in {cite book} and {{sfn|Greenhill|1988a|p=}} to distinguish different books published in the same year.[4] You can use |loc= as in {{sfn|Greenhill|1988a|loc=Author's note}}[5] Obviously reffing is very personal and can be contentious. I started a section on my talk page with a view to gathering people's ideas and possibly writing an Essay. If anyone would like to contribute, let loose with their gripes etc. with reasons, please feel free. MinorProphet (talk) 05:39, 20 February 2021 (UTC)


  1. ^ a b Greenhill 1988, p. 47.
  2. ^ Greenhill 1988, p. 49.
  3. ^ Greenhill 1988, p. 52.
  4. ^ Greenhill 1988a, p. 296.
  5. ^ Greenhill 1988a, Author's note.
  • Greenhill, Basil (1988) [1951]. The Merchant Schooners. London: Conway Maritime Press. ISBN 0 85177 475 X.
  • Greenhill, Basil (1988). Pleasure Schooners. London: Conway Maritime Press. ISBN 0 85188 478 X
As a note, you do not need to use |ref={{harvid|Greenhill|1988a}} in {cite book}, just set year=1988a: {{cite book |last1=Greenhill |first1=Basil |title=Pleasure Schooners |year=1988a |publisher=Conway Maritime Press |location=London}}, which outputs the necessarily disambiguated citeref: CITEREFGreenhill1988a. --Izno (talk) 07:40, 20 February 2021 (UTC)
Thanks, learn something every day. >MinorProphet (talk) 10:01, 20 February 2021 (UTC)
@MinorProphet: When building a {{cite book}} (or similar) with a view to using {{sfn}}, the only parameters that matter for the {{cite book}} are: |last1= (or |last= if there is only one author), |last2=, |last3=, |last4= (as many as are applicable) and |date=; other parameters such as |edition= and |orig-date= do not contribute to the sfn anchor. As regards suffix letters for years, see refs 25, 50, 51, 55 at Reading Southern railway station. --Redrose64 🌹 (talk) 11:50, 20 February 2021 (UTC)
@Redrose64: Hi, I think I was aware that {{sfn}} only needs |last=etc. and |date=, but my wording was clumsily expressed. What I was trying to say was that since you can use {{cite}} params in various ways not necessarily envisaged by their authors, are |edition=1988 and |year=1988 equally valid methods for displaying what amounts to the same concept? In other words, is there (or should there be) one and only one way to use eg {{cite book}}? There will probably be both purists and hackers. Also, it's very much a personal choice since everyone seems to have developed their own favourite style of cite, regardless of the many very specific Chicago/Harvard/Oxford/CS1|2 styles. Thanks also for the helpful pointers to Reading SR station - coincidentally I was at school there in the 1970s, and remember well the old GWR station with the original entrance building. >MinorProphet (talk) 13:06, 20 February 2021 (UTC)
At present, if you are not specifying the day and month you can use |year= as an alternative to |date=; but |year= is not guaranteed to work forever - there are those who would seek to eradicate all parameter alternatives for the cite templates. If you don't stick to what's in the documentation, it may not work and you will probably get a bot trampling all over your edit. --Redrose64 🌹 (talk) 13:31, 20 February 2021 (UTC)

I seem to remember a very long time ago that I encountered some difficulty with using |date= with only a year. It would have been when I first started trying to understand {{cite book}} coupled with {{sfn}} etc.; perhaps merely an initial misunderstanding. Anyway I've used |year= ever since with books, and |date= with {{cite journal}} and {{cite news}} but I really don't care what the param is, even |go-forth-and-multiply=...(lol?) If |date= takes any combination of DDMMYYYY etc., I'm quite happy to use it.
@ThoughtIdRetired and DuncanHill: I have been slow to understand your preference for a {{cite book}} in the reflist, with perhaps no bibliography. As was pointed out it, is indeed much easier to locate every reference to a book etc. by using <ref name="" /> in the reflist along with {{rp|36}}:36 etc.: all refs are linked to the source on one line. My personal feeling (along with Izno, I think) is that it somehow slows down the whole flow of the article, as if the reference itself were more important than the information being imparted. It obviously depends on a number of factors: many social science journals use Chicago-style (Greenhill 1988, 38) as a matter of course which also (imo) breaks up the flow. Mind you, many scientific articles are simply not concerned with forging a literary style, but merely presenting information in the plainest and simplest way possible. Again, it's a matter of preference. I am not being aggressive or dismissive, merely trying to understand why people prefer certain reffing styles.
So, why might someone want to locate every cite of a particular source in the whole of an article? As a general reader, when coming across an article that interests me, I might well attempt to hunt down the source, but I wouldn't necessarily want to pinpoint every ref in a specific book, even as an editor or creator of an article. Thoughts? MinorProphet (talk) 15:31, 20 February 2021 (UTC)

Most of the sfn errors I come across (and I have fixed hundreds) are undefined works, or multiple undifferentiated works, or not matching the name in the sfn brackets with the names in the defined citation. Sometimes they are because of editors copying lumps of text from other articles and not bothering to check the refs afterwards, or editors "cleaning up" bibliographies and further reading sections without realising that some of the works are being called by refs. I've also seen them broken because someone has, cleverly and helpfully, defined refs in a template which happen to have the same name as other refs used on the page calling the template. As a reader I want to be able to check a citation quickly and easily, and for me short citations make that harder. They don't break up the flow of the text because if I just want to read the text I can choose not to look at the refs. When it comes to "literary style" in article space I'm with Beckett, who tried to write without it, and Q, who told us to kill our darlings. "Just the facts, ma'am", as Joe Friday didn't quite say. DuncanHill (talk) 15:58, 20 February 2021 (UTC)
I am caught by surprise by the volume of comments here - I will have to study them in more detail later as a bit rushed right now.
An immediate answer to So, why might someone want to locate every cite of a particular source in the whole of an article? As a general reader,.... - in 2 parts. (1) General reader: when I first started using wikipedia - before doing any editing - if I wanted to find out more on a subject it seemed sensible to pick the major sources for an article and read them. Rightly or wrongly, I feel my preferred method of referencing allows you to quickly identify where the bulk of the article is sourced. (I note that "further reading" sections are an unreliable source of helpful reading material - if the books are not used as references in the article, why are they important?) (2) As an editor, more by accident than design, I seem to spend a lot of my rather limited editing time on articles that I feel are something of a disaster (in terms of content). The first job is to work out where the article is sourced. (That often answers the question: why is this so bad?) So often there is an impressive list of references, but when you get into it, you realise that most of the meat of an article is sourced to something one would struggle to define as an RS. The weirdest that I have found is a ref that appeared to be based on Fast Sailing Ships by David MacGregor - it's a pivotal work and cited in Clipper - where you would expect it to be used often. It then turned out to be an on-line link to just the index of this book, was used once, and supporting text that suggested the editor had never actually read the source. (Fixing this is a work in progress.) So - working out the overall reference base for an article seems key to sorting out any content problems.ThoughtIdRetired (talk) 17:20, 20 February 2021 (UTC)
Thanks all for your carefully and generously-expressed views. I feel that the original thread has shifted far enough to be restarted in the following sub-section. MinorProphet (talk) 21:40, 20 February 2021 (UTC)
DuncanHill, if you're looking to fix sfn templates without matching full references in articles, have a crack at Category:Harv and Sfn no-target errors, current population 26,000 articles. A limited fraction of them are false positives, but most need fixing. Category:Harv and Sfn multiple-target errors also has about 3,600 articles at the moment. – Jonesey95 (talk) 22:52, 20 February 2021 (UTC)
@Jonesey95: - I'm not looking to fix sfn errors, I just keep finding them! I've got a script that highlights them on the page. DuncanHill (talk) 00:20, 21 February 2021 (UTC)

Further thoughts on refs[edit]

It has been increasingly obvious to me for some time that there are a number of deep-seated problems with "WP" and its overall attitude to sourcing and referencing. The comments here exemplify what I have began to call WP's psychosis in regard to reffing. Maybe my impending essay could be titled "Why is this so bad?"™

1) @DuncanHill: I sympathise completely with your distress about failed {{sfn}} refs. I haven't experienced a tenth of your discomfort. I regularly click on 'Random article' to get a feel of what's around, but it's mostly Ukrainian 3rd-division footballers, Iranian villages pop. 270 in 2004, Japanese railway stations and species of minute sea snails in Patagonia with approx. 1 ref per article, or 2 if you're lucky. I get the feeling that you are a fan of {{sfn}}, Duncan, but the attempts of ill-informed editors have left you in a state of disrepair. Why do they get it so wrong? I refer to my previous post, which points towards a complete lack of comprehensive and authoritative help in this specific area. The choices of reffing presented to a noob editor are mind-bendingly multifarious. The possibilities presented in the various reffing "Help" pages are utterly contradictory and are displayed as a smorgasbord of equal-valued options, none of which can possibly be "better" than any other, since "consent" is apparently the prime factor.
Why is it so bad? Because although sfns and cite books are one of the best ways to approach referencing (imo), they are also the hardest possible way on WP, and quite frankly the "Help" is derisory. It's no accident that there are so many sfn fails: you might as well ask "Why do so many people die on Mt. Everest?" PS Why do you refer to William Faulkner, 'darling killer' as Q?
2) @ThoughtIdRetired: Thanks for your early reactions: I look forward to some further thoughts. I feel that you perhaps approach almost the same problem from almost 180°: the ability to pinpoint the significant sources (RS) of any article (if any) seems indeed crucial to working out whether it's even worth fixing (WP:Blow it up and start over), or could could just do with a little smartening up around the edges. If you would like some assistance with 'Clipper' I'd be happy to have a look at least. Slocum, Alain Gerbault, Dana, my seafaring literary heroes. I seem to remember having bought a copy of The Log of the Cutty Sark or something similar, a few years ago. I'll be getting the opportunity to dig out some boxes soon.

We all know when we come across a well-written, welcoming, intelligent and cogently-argued article, with enlightening, well-chosen sources and an over-arching sense of style and purpose, understandable by experts and newcomers alike: but these are, alas, most rare. Even the best-contrived, illuminating text in an encyclopedia article must fail unless it is backed up by solid references. Surely the quality of the refs underpin and define the entire edifice of our amazing enterprise: yet the reffing here is an utterly contradictory, gloomy and psychotically destructive enterprise as long as any editor can do what they damn well choose. There are far too many reffing options, and none can be be allowed a priority under the current system. The best reffing takes time and effort to learn, and the Help is pitiful: you can do this, or that, or the other, you can achieve it in 47 other ways which we will mix up and intersperse with other pointless and dead-end options, or you can devise your own insane system: but if you create an article with reffing system X, that's how it will be for the rest of time unless some sort of "consensus" is reached, which usually consists of two equally intellectually unattractive argumentative saddos engaging in an edit war before subjecting themselves to Arbitration presided over by sleeper fifth columnists and QAnon New Page Patrollers who by their billion edits have wormed their way into Adminship. Hmm, perhaps David Foster Wallace might have begun his career with a similar sentence. Sleep beckons. MinorProphet (talk) 21:40, 20 February 2021 (UTC)

@MinorProphet: No, I'm not a fan of sfn, even when done perfectly. DuncanHill (talk) 00:20, 21 February 2021 (UTC)
@MinorProphet: Like Faulkner, I was misremembering Q - "Whenever you feel an impulse to perpetrate a piece of exceptionally fine writing, obey it—whole-heartedly—and delete it before sending your manuscript to press. Murder your darlings". DuncanHill (talk) 15:36, 21 February 2021 (UTC)
Seem to be 3 headings for issues here.
Getting citation information into an article accurately and efficiently. My preference for this is the {cite book} template. I have used {sfn} but find the technical requirements tedious and/or difficult to learn (i.e. I have missed the easy way to use it). The multiple problems with short references in Wikipedia suggest I am not alone in this. One problem with {cite book} is that you then have to use an extra template to put a page number in the first time. The second problem with {cite book} is that on the second use of a reference, you have to remember the name you gave the ref. Solutions of this include having a second copy of the whole article open and using "edit find" to get the right one. This is relevant if there is a prolific author on the article subject who has written several useful refs. References I use a lot are stored on my user page (so that gives consistency between articles on similar subjects.)
Then there are problems getting citation information out of an article. Short citations require some adept handling of the mouse to show the full reference details (you already have the page number at this point) - but it works. However, you cannot go to the reference section and make a quick assessment of the number of times a source is cited. (Unless I am missing something.) {cite book} has the problem of "ugly" inclusion of the page number in the text of the article. It is pretty pointless to show that then, because the reader does not know which ref it applies to without hovering the cursor over the reference's number. Playing (amateur) systems analyst for a moment, it is surely possible to improve the display of {cite book} references by only displaying the page number on user request (either by cursor hover, as part of seeing the whole reference, or an on/off preference switch for the page).
The last problem with references is source selection. This is the first and major fundamental role of an editor - decide which sources are important and authoritative, providing the basis for the article. (The second fundamental role is putting all that material into an intelligible and readable form.) To put this trivially, I am sure I could find a source that said the moon was made of cheese (in the children's book section), but not appropriate for Moon.
More seriously, I don't know if all subject areas have this problem, but I have been editing in areas where there are a huge number of books that are totally inadequate. For Highland Clearances there are many books written for the "tourist bookshop" that simply pander to the misconceptions of the people who buy them. This is even addressed by academic historians (Tom Devine covers it quite extensively: The Scottish Clearances: A History of the Dispossessed, 1600–1900. London: Allen Lane. ISBN 978-0241304105 pgs 9-11).
In maritime history, particularly on subjects like Clippers, there are many "coffee table" books (similar criticism to the Scottish tourist bookshop fare), but also authors who appear authoritative, but have big failings. Eric Kentley (16 years as a curator at the National Maritime Museum) wrote Cutty Sark, the Last of the Tea Clippers, published by Conway Maritime (a very good quality publisher for this subject) (2014). His lack of grasp of the subject is shown by some of the reviews of this book on Google (not the reviews that say "present for...." etc.) (Not sure I want to risk anything legal for Wikipedia by listing all the stuff Kentley got wrong.) Interestingly the worst of these were edited out in the second edition. (I note Cutty Sark does not use this work.) Basil Lubbock is a fantastically prolific writer - and often the only source on some sectors of the subject - but even his Wikipedia page warns that he get things muddled up a lot.
In short, you need an understanding of the subject to choose the sources.
Another danger is the book that is readily available online. If google books let you see all the text of a book that is not totally ancient, that surely tells you that there is no value left in that work for the publisher. The book of a similar age that you cannot see on google books, but would have to buy or find in the library, is almost certainly more authoritative. The partial views that you can get on google books means that an editor is seeing things out of context - and can misunderstand. Some authors are a real risk for this - Eric Richards (Scottish historian) often lays out the argument he wants to demolish in some detail, then says what he thinks. The snapshot views offered by google books would completely misrepresent his output.
Then we have the source that is authoritative on some aspects, but not others. For instance [1], where a philosopher (appropriate in discussing the morality/legality of the bombing of Dresden) was cited for the death toll of the bombing of Hamburg. He simply had the wrong number - and any editor on top of the subject should have spotted this, or if they didn't know, checked it. My thinking here possibly goes further than WP:CONTEXTMATTERS warning about "Information provided in passing by an otherwise reliable source".
Overall, Wikipedia editors need higher standards on source selection. A bit more time reading and a bit less time editing.
I should add - I have, at times, made some real howlers on Wikipedia - but am fortunate that other editors have politely corrected them. ThoughtIdRetired (talk) 14:55, 21 February 2021 (UTC)
I've found a few howlers caused by poor source selection. A couple of favourites - a misattribution of a famous quote to the wrong speech, because one historian blindly copied another's error, another a totally spurious quotation ultimately sourced to a satirical work. Both of these would (or should) be obvious to anyone familiar with the subject. I'm sure I must have made a few too. Hopefully someone picks them up! DuncanHill (talk) 15:50, 21 February 2021 (UTC)
Perhaps we need a how-to-write-an-article guide that starts with references, as in "First, gather your sources and make sure they are acceptable, reliable and properly referenced." Then it becomes easier to build an article with a well-stocked reference section, on which every subsequently written section and paragraph will depend. I am only a newby in Wikipedia, but this is based on my experience writing other essays in an academic context, where the first thing I did with a new book was to write its reference down (in the locally relevant style) before starting to make notes from it. It saved me from thrashing around later trying to build a ref for a book that I had handed back long before. This might even <pious wish>discourage some of the more fantastical Draft pages from getting as far as being declined for lack of references</pious wish>. Writing in this way leads me a to use a hand-built reference list (all nicely in alphabetical order), with lots of sfn's picking out the relevant page numbers.--Verbarson (talk) 14:05, 22 February 2021 (UTC)

(edit conflict)

@ThoughtIdRetired:, thanks for your comprehensive comments. 1) Getting cite info into an article. I am a firm believer in {cite book} because it always displays the information in a completely consistent way. It's orientated towards scientific journals with volume, issue and page number as simple numbers, rather than the literary "Oxford" style with 'vol. III' and date at the end. This can be worked around with |others= and |series=. Although you are correct in saying that only <ref name="" /> shows up all the references to a single cite compared to {sfn}, it's not particularly difficult to difficult to scroll through the reflist and make a rough guess as to its relative importance. Do you or anyone really need an exact count?
You correctly identify the many problems of using <ref name="" /> with {cite book}, which is why I started using {sfn}s. I started an instructional article with examples, called The joy of sfn, but I haven't done much work on it recently. Maybe I should have another look. As always, it's very much a personal thing, and you are right that getting to grips with {sfn} takes quite a lot of work. I reckon that most of the effort is fact understanding {cite book}, and when you know it, {sfn} is a mere doddle, although DuncanHill knows how much it can go wrong. I've made a quick list of what can go wrong further down. In fact the whole mechanism of how any combination of [ref]...[/ref] and [sfn}s actually fit together is nowhere explained on an official Help page (most people probably wouldn't need it); although it the basics were kindly laid out for me here, especially the second answer 2.

As I see it, there are various ways of making a ref.

  1. The very easiest for the editor: <ref>Greenhill, p. 36</ref>, and the reader has to go looking in the bibliography section for the relevant work, which can be formatted either manually or by {cite book}.
  2. <ref name=>"Greenhill 1988"</ref> plus {rp|36} (I think your preferred method) which as you say involves "ugly" inclusions in the body of the text. Note that it only works if you include the entire cited source in the (first) instance of |ref name=, whether you use {cite book} or do it by hand.
  3. The get-around is very time-consuming for the editor, which involves creating a separate ref for every page used: <ref name=>"Greenhill 1988 p36"</ref>. There is no intrusion, but the cite then needs to be in the Bibliography section, again either manual or with {cite book}. Note that in all these three cases {cite book} takes no dynamic part in the referencing system, it merely displays information, and the MediaWiki software (I think) just puts them together when it compiles the {reflist}.
  4. You can make an inline {harv} eg {{harv|Greenhill|1988|p=36}} (Greenhill 1988, p. 36) but I think this is now deprecated.
  5. In my opinion, {sfn} completely overcomes all these difficulties. {cite book} automatically creates a {harv} which an {sfn} can link to via a CITEREF. {sfn} is actually just a {harv} wrapped in a <ref></ref>, When I create any {cite book} I put in a hidden {sfn} eg <!--{{sfn|Greenhill|1988|p=}}--> All I need to do is remember the author and date of the book (not hard), or it's in the bibliography section hidden in the cite. The entry in the reflist automatically links to the cite in the Bibliography, which appears if you hover over it, or jumps to it if you click on it. What could be easier? (cue distant hollow laughter...)

2) Getting cite info out of the article. I agree with all your points. As to the hovering the mouse, there may well be adjustments in Control Panel → Mouse which might improve things. I agree that the output of {cite book} could be tweaked quite easily, but I think only a very few people have the knowledge and ability to do it, and it would need a concerted effort to bring about any change.
3) Source selection. This is obviously a separate issue to the reffing systems discussed above. It would be somewhat unfeasible to compile a list of "approved" reliable sources, and since anyone can edit WP, you are going to get a wide range of sources chosen for various reasons which you discuss. "Anyone can edit" is both one of the great strengths and weaknesses of WP. Very often editors are not specialists or experts: many of us are merely informed readers with an interest in one or many subjects. But as long as coffee-table books (or worse) are used as sources, Wikipedia will remain a "coffee-table encyclopaedia". I agree with everything you say about Google books. Google search seems to deliberately omit from its results, and almost everything out of copyright is there in full with pdfs and plain OCR text. Even some contemporary books are there in full for short-term loan online like a proper library.

@DuncanHill: Just to recap on your reasons why {sfn}s get broken:

  • The {sfn} will appear to be good, but there's nothing it can link to because {cite book} is completely missing
    • Copying lumps of text from other articles and not copying the associated {cite book}, and not bothering to check the refs afterwards.
    • Removing works from Bibliographies without checking that there is a corresponding sfn in the text.
  • Duplicate CITEREF: (I think?) Editor has defined refs in a template (eg sfn, harv) which happen to have the same name as other refs used on the page calling the template.

Other common reasons:

  • Mismatch between params of {sfn} and {cite book} where both exist:
    • In {cite book}, any or all of |last1= (etc.), |date= or |year= is either missing or incorrectly defined. Even if the {sfn} is completely correct, the link will fail.
    • As I pointed out earlier, you must definitely use the correct date of publication using |date=, and |orig-date= and NOT eg |edition=1988, since CITEREF cannot extract information from |edition=.
    • In the {sfn}, any of the first two params may be incorrect, eg {sfn|Greenwood|1998|p=36} instead of {sfn|Greenhill|1988|p=36}
    • In the {sfn} if there is more than one authors (etc.) one or more may be missing or incorrectly spelled, even if the {cite book} is completely correct.
    • An extraneous pipe sign {sfn|Greenwood||1998|p=36} will result in an empty ampersand (Greenwood & 1998, p. 36)
    • If you define two {sfn}s with different pagination schemes eg {sfn|Greenwood|1998|p=36} and {sfn|Greenwood|1998|pp=36}, you will get an error when previewing.

@Jonesey95: I installed Trappist the monk/HarvErrors.js and fixed Koornwinder polynomials as a test - seems to have worked a treat. Btw, there seemed to be no 'Install' button as suggested in the installation instructions. It also pointed out in James Douglas (English Army officer) the problem of using {{sfn|Greenhill|1988|p=36}} along with |edition=1988 in {cite book} as pointed out in the list above, giving Harv error: link from CITEREFKenyon1993 doesn't point to any citation. and Harv warning: There is no link pointing to this citation. The anchor is named CITEREFKenyon1986. (Haven't fixed it yet.) This is exactly what ThoughtIdRetired is using in his OP in the section above, and I'm afraid it is simply wrong and leads into error. It may be why you're not a fan of {sfn} since they would never work if you use {cite book} in this way.

Here's what you posted in your OP: <ref name="Greenhill 1988">{{cite book |last1=Greenhill |first1=Basil |title=The Merchant Schooners |date=1951 |edition=1988 |publisher=Conway Maritime Press |location=London |isbn=0 85177 475 X}}</ref>{{rp|99}}, which gives you this:[1]:99 which happens to work, although cite book is wrongly formatted.

{sfn} can only ever work with this if you use {{sfn|Greenhill|1951|p=36}}, resulting in (Greenhill 1951, p. 36), which is obviously not what you need.

Compare with {{sfn|Greenhill|1988|p=36}} along with this [2] (Greenhill 1988, p. 36)

So, TIR, that may be one of the reasons why you don't get along with {sfn}. MinorProphet (talk) 17:42, 22 February 2021 (UTC)

Thanks, User:MinorProphet. I may be working from an insufficient knowledge base, but here are some additional thoughts.
(i) The refname I have given: "Greenhill 1988" is intended for the {citebook} template. I have found that it can be confused by some editors to be a Harvard reference, but I could have instead given the refname "strawberry jam" - but then I would have had to remember which one was which.
(ii) In some cases it is absolutely crucial to identify which edition a reference is from. Over and above the risk that page numbers do not align, some editions have substantially different content - sometimes even contradicting what was said in an earlier version. I see that I am using year, orig-year and edition differently from you. It would help enormously if {citebook} had a direct link to an easy to read brief set of instructions on what should be in each field - hover over a field to see what information should go there? I have just revisited [[2]] and even knowing the usage preferred by MinorProphet, it is a little challenging to find that interpretation in the guidance as written. Perhaps I am getting a little impatient with instructions that require a bit of work, but can't some of the CE enthusiasts on Wikipedia make this less impenetrable? ThoughtIdRetired (talk) 20:46, 22 February 2021 (UTC)
...and WP:CITE seems to be a little light on "so how do I actually do that?" (though better than the help text given above). Writing instructions is a special skill - there must be someone in the Wikipedia community who can cover this fundamental part of editing. ThoughtIdRetired (talk) 21:21, 22 February 2021 (UTC)
@ThoughtIdRetired:Thanks for your reply. I also had a look at the Date section you referred to. Many things have changed or been updated since I last looked at it, and I would agree that what I said earlier isn't reflected in the Help page. In my opinion, the parameters for {cite book} should be filled out by copying the info on the physical title page and the next page (eg edition, isbn etc.) This means, I feel, that |date= should be the date of publication (whatever edition it is) and not the date of the original edition. You have also highlighted the grim state of actual practical help available in the documentation pages, including Template:Cite book. I absolutely agree that there should be one and only one style of reffing. WP:CITESTYLE says that "A number of citation styles exist" but doesn't tell you that none of these are available through {cite book}, a subset of CS1|2 which is a hodge-podge of Chicago and APA invented by WP:en coding editors and tweaked as they saw fit over the years. You are quite right about writing instructions, it's a particularly specific skill, but the documentation is usually written by the guys who write the templates, and unfortunately the two capabilities are almost never present in the one head. Just to check, |date= can take an {sfn} style disambig for two books from the same year, although the documentation at Date says that |year= fulfils this function. {{sfn|Smith|1988a|p=304}} [3]
  • Smith, John (1988a) [1903]. John Smith's Big Book of Fun (42nd ed.). Big Books Corp. MinorProphet (talk) 03:15, 23 February 2021 (UTC)


  1. ^ Greenhill, Basil (1951). The Merchant Schooners (1988 ed.). London: Conway Maritime Press. ISBN 0 85177 475 X.
  2. ^ Greenhill, Basil (1988) [1951]. The Merchant Schooners. London: Conway Maritime Press. ISBN 0 85177 475 X.
  3. ^ Smith 1988a, p. 304.
...and on source selection - it's largely a matter of self-discipline by editors - "do I understand this enough to chip in?" That does not require you to be a total expert, and the co-operative nature of Wikipedia is the safety device for an editor getting it all wrong. But conversely, editors who go for high edit counts worry me - how on earth do they ever find time to research all those edits? More rules, I guess, would be a problem. Better compliance with the rules we have would help.ThoughtIdRetired (talk) 21:04, 22 February 2021 (UTC)

@Verbarson: Such things do exist, such as Your First Article and Help:Referencing for beginners, and Wikipedia:Reliable sources (RS), plus all the associated 'See also's at the end of the articles. On the other hand, because WP allows almost any referencing style, it is left up to the individual to make a choice with no advice given (or even allowed) as to what is "The Best Way"™. It really does take a fair amount of reading/researching experience to be able to recognise an RS, and as shown above there are very many ways in which any reffing method can go wrong. I've been on WP for around 12 years, but I continue to make basic slips such as missing off the / in a closing [ref]. I remember the confusion when it happened the first time, and how long it took to find the solution. Now I can recognize it instantly, but sometimes I can make multiple mistakes, and it takes five or more 'Previews' to get a single extended {ref} or {efn} right. Many intelligent people have never had the opportunity to write anything at university level (essentially, I feel, the general standard of writing, sources and reffing which WP is aiming for), and I also don't think that the sort of all-in-one guide you suggest exists. Maybe it would be far too long.

  • Notability, COI, Self-Promotion, WP:An article about yourself isn't necessarily a good thing, Advertising, etc.
  • Sources - Bibliography, {cite book} etc.
  • Reffing types - {ref}...{/ref} and the rest, discussed above, vs. templates ({sfn}, {harv} etc.
  • Planning - General layout - Bullet-point outline of sections
  • Content - style - WP:MOS which I call Ministry of Style, guidelines for GAs and FAs as ideals.

I agree that far too many people have no appreciation of these things, perhaps they just have to learn the hard way which takes time and effort, with no short cuts I'm afraid. MinorProphet (talk) 18:48, 22 February 2021 (UTC)

@DuncanHill: - I wonder if you could say why you don't like {sfn}s at all, even when properly used? I am genuinely interested, and would like to hear your thoughts. I realise that they are quite lot of extra trouble - much of the effort is having to get {cite book} right - and easy to get screw up as 26,000 articles show: instead a manual bibliography and a ref - /ref of some sort achieves much the same thing, if perhaps slightly less elegantly imo but with much less margin for error. MinorProphet (talk) 19:09, 22 February 2021 (UTC)
@MinorProphet: It's not just sfn, it's the whole Harvard style. It seems totally unsuited to Wikipedia. Look at David Lloyd George. Look at refs 1 & 3. Ref 1 I can point my mouse at the little 1 and see the book and author in preview. I click on the 1 and go straight to the details. Click "back" and I'm back where I was in the text. Ref 3 I point my mouse at and it just says "Harnden 2011". I click on the 3 and go to where it says "Harnden 2011", then I click on that and end up at the details. I then have to click "back" twice to return to my point in the text. Yes, I do know I can click on the preview, but that's far less intuitive than clicking on the number. As a reader who looks refs up it's a pain. As an editor who checks refs it's a pain. We should make it as easy as possible for people - readers and editors to look up refs. DuncanHill (talk) 20:08, 22 February 2021 (UTC)
Thanks User:DuncanHill - I am puzzled why a computer-based encyclopedia needs to use a referencing system developed for the paper-based world. One single referencing system, with all the problems sorted out, will surely achieve the highest level of reader-understanding and editor-usability. My vote would be {citebook} with better help text (hover over for help on what should go in each field - click for fullest detail?) and the page numbers of refs only showing on hovering over the ref number, to help the flow of readability of the article (with option to make all ref page numbers visible in the article?). Of course, there would be an absolutely mountainous barrier to consensus forming on such a proposal, but I don't think that should kill a reasonable aspiration.ThoughtIdRetired (talk) 20:55, 22 February 2021 (UTC)

@DuncanHill: Thanks very much for your views. The term "Harvard style" has been badly abused on WP, since it's generally mixed up with {sfn} and {harv} via CS1|2, and in fact Parenthetical referencing or Harvard-style inline Author-date referencing eg {Greenbrook 1988, p.46) which {harv} implements is now officially discouraged since October 2020. If I understand you correctly, are you disagreeing with the whole system of using {sfn} and {harv}, coupled to {cite book} in a separate bibliography? It seems to me that some editors/readers are quite happy with a References section populated with full cites, created by {cite book} using {ref name=""}, although {cite book} now creates its own CITEREF based on author and date anyway, ready to be linked to by [sfn}.

It's interesting that you picked David Lloyd George, which is an appalling mish-mash of reffing styles with many tens of editors in the last year. Having installed the Trappist the monk/HarvErrors.js as recommended by Jonesey95 as above, I can now see all the {cite book}s in the Bibliography with no ref pointing to them (ie roughly 50%).

Ref [1] uses the {ref name=""} style, but the whole {cite book} which appears in the Bibliography section is ignored and duplicated in the ref itself, ie briefly {ref name=""}(cite book |last= etc.}}{/ref}{rp|13} This, in my opinion, is completely wrong in this particular article, since the {cite book} entry in the reflist cancels out the one in the bibliography. You are right that hovering over the [1] shows the complete source, but that's because the entire {cite book} is included in the ref and every subsequent {ref name ="" /}. This is ThoughtIdRetired's favoured approach, I think, and it does indeed allow you to see every link to any source in the reflist. It is also true that you cannot link directly with this this reffing style to a {cite book} in the Biblio. I think may be possible in some other way using {ref}s, but I have never bothered to find out since {sfn} does it for you. As far as I know (and I may well be wrong) the only other widely-used way is a plain {ref name=""}Greenhill, p.37{/ref}, which means that editor and reader alike have to scroll down to the Biblio to find out. In other words, {cite book} should be either exclusively in the notes OR the bibliography, but not both.

Ref [3] uses {sfn} in the completely correct way. As you say, if you just click on the [3] that takes you to the reflist, and and clicking on that takes you to the Biblio, and you have to click back twice. But are you aware that when you point (hover) your mouse over the blue[3] so that it shows "Harnden 2011, p. 11", and then move the mouse slightly so it points to the blue link Harnden 2011, that it then shows the {cite book} entry in the Biblio? You can go there directly by clicking on it. The whole chain is displayed by two very slight movements of the mouse (admittedly two rather than one). If the cite book contains a url, you can even click on that and be taken there direct with an even smaller movement of the mouse.

Not every article needs a Bibliography section, especially short-ish ones, but a major article like this one most certainly does, and it has been seriously abused. In fact, one of the more concerned editors (see Talk:David Lloyd George#More problems with refs - Grigg 2002 vol3.) has been making just these changes, combining {ref name} with {cite book}, for this very reason that {sfn} "takes more clicks" which I feel is simply untrue. Has no-one ever tried hovering their mouse? You don't have to actually click on anything with {sfn}! Looking at the whole hodge-podge has made me feel slightly nauseous.
I imagine that an article in this appalling state would take an entire week to fix, since 50% of the refs would have to be changed into either {ref}s or {sfn}s, including getting "consensus" (lol).

At some point this thread is going to get archived (five days of inactivity), so might I suggest if that happens we could carry on this most interesting and fruitful discussion we could do so on my talk page under User talk:MinorProphet#Further thoughts on referencing. Thanks anyway for all your most generous comments so far, it has been really stimulating. >MinorProphet (talk) 03:15, 23 February 2021 (UTC)

Quick reply. "But are you aware that when you point (hover) your mouse over the blue[3] so that it shows "Harnden 2011, p. 11", and then move the mouse slightly so it points to the blue link Harnden 2011, that it then shows the {cite book} entry in the Biblio? " - no, it doesn't - do you have some kind of script which makes it do that? Even if it did, it's still more moves than necessary. The bibliography on Lloyd George is a mess I agree, it has been used to list just about every book or article that a particular editor has found. I've tried trimming it occasionally, and it needs re-ordering. I am the "more concerned" editor in the talk page thread you linked to. I found a mass of duff refs (wrong book) and turned them into correct ones. It is infinitely more important to have accurate refs than ones which please your aesthetic sensibilities. I do not see any reason not to list all books cited after their appearance in the refs section. A "works cited" followed by a "further reading" section is what DLlG needs. The errors which matter are the no pages, no works, wrong authors, wrong dates, which are, in my experience, far more common using sfn or harv. Having a citebook not called by a reference isn't, to my mind, an error at all, just an artefact of the gadget. DuncanHill (talk) 03:47, 23 February 2021 (UTC)
@DuncanHill: Hi, thanks for your swift reply. I'm really sorry not to have recognised your name. I did look through the article history to check if anyone here was involved, but I haven't had had much sleep recently (although not under the influence) and I simply didn't make the connection. I hope you aren't too offended. As far as I know I have no script installed which allows this hovering effect. I'm using an ancient version of Firefox (v. 47) on XP SP3, which can sometimes give strange effects. As far as I remember it just happened one day and it has continued consistently for at least five years. By your description of the {sfn}, you have no other choice except to click directly on the [3] or the Harnden 2011 when you point at /hover over it: either of which takes you to the reflist, from where another click takes you to the {cite book} in the Bibliography section. Can you not point to/hover over the Harnden 2011 link even in the reflist to see the complete cite? What browser/OS are you using? I can certainly see that it would make using {sfn}s a completely and frustratingly tedious bore, which is the exact opposite of my experience and why I'm a fan. I wonder whether @ThoughtIdRetired: has the same experience as you. I think will start a new thread to enquire.
On the subject of DLlG, skipping through the talk page I became aware of the mass of screwed-up refs of what ever type. You are quite right that "Having a citebook not called by a reference" isn't a real error, I was only made aware of it having installed the HarvError script I mentioned. I think this has only come about relatively recently, since changes to CS1|2 meant that |ref=harv was no longer needed in {{cite book}} to allow {sfn} to link to it (a major source of sfn fails), but was able to generate its own CITEREF automatically. Thus the 26,000 articles that use {cite book}, {citation} etc. with no associated ref. In the old days plenty of articles used {cite book} without |ref=harv, simply as a consistent way of formatting the cite, but it has turned into something else, perhaps an error-generating beast whose 'errors' are not necessarily errors at all. I might have a look at some of my own articles...
As I have said, and your experience shows, {sfn} etc. is essentially more complex than almost any {ref}...{/ref} or {ref name=""}, coupled or not with {cite book}, perhaps even by an order of magnitude: but I still maintain that the unless the params of {cite book} are used consistently and 'correctly' (whatever that means), there will still be room for errors even if {sfn} is not being used.
My personal approach is: since {cite book} generates its own CITEREF, almost exactly the same as an intelligently-chosen {ref name=""} why not use it? Again it's personal.
It seems there are two approaches to the appearance of the reflist. One is to have everything in it in long format, regardless. The other is for it to contain only short footnotes, generated either by a very simple {ref}Greenbridge, p. 204{/ref} [or (Greenbridge 1998, p. 204) if needed for disambig], or {sfn}s. This keeps the reflist very compact, with only {cite web}s making a difference. Plain {ref}s require everyone to scroll down to the Bibliography, {sfn}s with their hoverability (for me at least) don't. Again, I find {rp|204} intrusive as it breaks up the flow of the text, although this is obviously not a difficulty for everyone. Sorry, I've been rather long-winded. I appreciate your taking the time to reply. Cheers, >MinorProphet (talk) 05:17, 23 February 2021 (UTC)
Well, well, well (three holes in the ground). Check out #Displayed output of sfn below. Izno kindly pinpointed the source of the behaviour. It's to do with Preferences → Gadgets → Browsing and a combination of Navigation popups and Reference Tooltips. For me, when Nav. popups is disabled and Ref. Tooltips is checked, all I have been describing bursts into life. MinorProphet (talk) 08:41, 23 February 2021 (UTC)

Force template to break onto new line[edit]

How do I force two consecutive {{rquote}} templates to break onto new lines? Such as the two quotes here. ➧datumizer  ☎  10:38, 19 February 2021 (UTC)

I made a hacky solution with an enclosing div.[3] Maybe somebody can do better. PrimeHunter (talk) 11:49, 19 February 2021 (UTC)
40%, no div? --mfb (talk) 18:10, 19 February 2021 (UTC)
That places them side by side again for me unless the window is narrow. Using Firefox, desktop, Vector. PrimeHunter (talk) 18:20, 19 February 2021 (UTC)
Ah... I tested it to the full width (and everything smaller than that) but my screen here isn't that big. If I zoom out then I can reproduce the problem. It's strange that a relative width leads to this behavior. Keep the div solution? Make the width 51% to ensure the boxes can't fit next to each other? --mfb (talk) 18:33, 19 February 2021 (UTC)
I ended up switching to a different template. Thanks though. ➧datumizer  ☎  07:36, 22 February 2021 (UTC)

Pending Changes again[edit]

After this archived thread was resolved, I am once again unable to set pending changes on articles. @Xaosflux:. -- Jezebel's Ponyobons mots 17:00, 19 February 2021 (UTC)

phab:T273317 may be still a problem, could you see if you can add/remove users from the new page patroller and autopatrolled groups? You can test with Special:UserRights/Xaosflux_ep (just set it to expire in a day). — xaosflux Talk 17:16, 19 February 2021 (UTC)
Done. No problem there.-- Jezebel's Ponyobons mots 17:44, 19 February 2021 (UTC)
Added to phab:T275017. — xaosflux Talk 19:20, 19 February 2021 (UTC)
hi @Ponyo, thanks for the report and @Xaosflux thanks for putting it on Phabricator
As a possible workaround, can you try to add yourself to "pending changes reviewers" group to see if that helps you to workaround the issue? Martin Urbanec (talk) 02:09, 20 February 2021 (UTC)
(I know it can sound unrelated, but when debugging the previous occurance of the same issue, it sometimes worked with reviewer, but not without for some weird reason, so that's why I'm suggesting it) Martin Urbanec (talk) 02:10, 20 February 2021 (UTC)
For information, the same happened with my last edits. --Delfield (talk) 15:18, 20 February 2021 (UTC)

Pending changes auto-accept error again[edit]

In this this thread last month, several users (including myself) mentioned an error where our edits are not being auto-accepted on semi-protected articles. This techincal issue was eventually resolved. Now I'm having the same issue again, as seen here (same article as last time). Can this technical issue be fixed again? Thanks. Maestro2016 (talk) 02:13, 20 February 2021 (UTC)

Suspect this is same problem as Wikipedia:Village_pump_(technical)#Pending_Changes_again above. — xaosflux Talk 02:19, 20 February 2021 (UTC)
I'm still having the same issue, as you can see from recent edits to the same article. I'm still getting the "pending changes" thing again. Maestro2016 (talk) 19:15, 21 February 2021 (UTC)
I also just experienced this issue on Scratch_(programming_language) [4] - the diff shows the version as current, but the Scratch page shows 1 pending revision, mine. Might it be because I undid an accepted revision? Dialectric (talk) 03:07, 25 February 2021 (UTC)

Hello, I don't know where to put this, but my edits to pending changes pages have recently not been accepted automatically. This is weird because I've been able to make many edits to pending changes pages without any problem earlier this month. My account is almost a year old and I have over 500 edits, I've been able to edit extend-confirmed and semi-protected pages without any issues. Clear Looking Glass (talk) 05:47, 25 February 2021 (UTC)

Watched articles[edit]

Is there any way to mark articles I'm watching from templates or categories? Eurohunter (talk) 18:00, 20 February 2021 (UTC)

What do you mean? I think the answer is 'no'. Certainly we shouldn't 'mark articles'. --Izno (talk) 18:33, 20 February 2021 (UTC)
I'm not sure what you mean but Wikipedia:User scripts/List#Watchlist shows:
  • Watchlist mark[1] (source) – bolds watched pages in Category and "User Contributions" listings. Also adds a "Show watchlist controls" link to enable watching and unwatching directly from these listings.
I haven't tried it and the script hasn't been edited since 2015. PrimeHunter (talk) 20:43, 20 February 2021 (UTC)


  1. ^ Copy the following code, click here, then paste:
@PrimeHunter: @Izno: I mean how do you know then which articles from certain category are you watching? It doesn't works. There is other script on PLWP pl:Wikipedysta:Kaligula/js/watchCat.js. Eurohunter (talk) 17:45, 21 February 2021 (UTC)

Range block[edit]

My ignorance of technical matters is complete. My IP address is currently subject to a range block. (1) The instructions given to IPs are confusing and misleading, and I think they ought to be changed. Is this the right place to raise this issue? (2) Up until today, this has not been a problem for me, since I can simply log in to my account. Today, the system kept telling me that I was not allowed to edit Wikipedia because my account had been blocked – it was logging me out in the middle of a simple edit. Sweet6970 (talk) 12:40, 21 February 2021 (UTC)

If this is not the correct place to raise my query, I would be grateful if someone would tell me where I should do so. Sweet6970 (talk) 10:42, 22 February 2021 (UTC)

Depends on whether it’s a global block on Meta or a local one on English Wikipedia. Can you share part of the message you receive, Sweet6970? (No need to divulge the IP address if you’re not comfortable with that.) — Pelagicmessages ) – (08:15 Fri 26, AEDT) 21:15, 25 February 2021 (UTC)

Global user page[edit]

As I understand it, it is possible to configure one's preferences at meta to create a global user page that will be transcluded from meta to all wikis, MediaWiki:Help:Extension:GlobalUserPage. When a user does that, a user page is no longer editable on en-wiki and its history log is not viewable here either. Most of the time this is fine, but I can see some potential problems, e.g. if somebody starts putting some inapprorpriate material (spam, using the userpage as a webhost, personal attacks etc) on their global user page, and we are not able to address the issue here. Another situation concerns users who are banned or blocked (e.g. for sockpuppetry) on en-wiki. Often in such cases we tag their userpages accordingly but it would seem that for a user with a globally transcluded user page we don't have this capability. Or do we? I saw a user who got indef blocked at ANI yesterday and they seemed to have implemented a global user page option via meta today. Is there anything that can be done in such situations? Nsk92 (talk) 13:36, 21 February 2021 (UTC)

The easiest way is to create a user page locally. All local user pages override global ones. Note that meta has also policies against inappropriate material like spam, webhosting, attacking others, so for some pages requesting deletion there is an option too. Majavah (talk!) 13:43, 21 February 2021 (UTC)
If there is blatant inappropriate material on a meta-wiki user page, you may tag it for speedy deletion on Meta, if you aren't sure you can ask at meta:Meta:Requests for help from a sysop or bureaucrat. You should not put project-local scarlet letters on someone else's meta-wiki page though. — xaosflux Talk 15:41, 21 February 2021 (UTC)
Nsk92: the local user page can still be created if there is a transcluded meta page. Try it on your example. If someone is using their meta global page for abuse, a global lock is probably in order also. –xenotalk 13:46, 21 February 2021 (UTC)
The blocked user in question, User:Tisquesusa, did have a local user page before. Then it suddenly disappeared and got replaced by a global user page transcluded from meta. The globally transcluded page does not contain any abusive or improper material at the moment (there was a G11 user subpage that I CSD tagged today and it got deleted). But the situation still somewhat concerns me. In fact I don't understand how I can try to create/edit a local user page for this user now (assuming I wanted to do that). I can't create a red link. I can't access the history log for the user page. The global user page just sits there. Nsk92 (talk) 13:59, 21 February 2021 (UTC)
The user page was deleted as WP:CSD#U5 by Deb today. I don't see how it qualifies as U5, mind you. An admin can recreate it if necessary. Jo-Jo Eumerus (talk) 14:18, 21 February 2021 (UTC)
I can restore it if you think it's worth the effort - it looks pretty awful, mind you. Let me know. Deb (talk) 14:26, 21 February 2021 (UTC)
The page that was deleted (as both G11 and U5) was a subpage User:Tisquesusa/Más Muisca, not the parent user page. I don't remember what the user page itself contained, but the subpage was an advertisement page for a guided tour/adventure operation run by the user. That's U5 in my book. Nsk92 (talk) 14:30, 21 February 2021 (UTC)
@Deb and Nsk92: U5 only applies when the owner has made few or no edits outside of user pages. Tisquesusa has made lots of non-userspace edits, so the criterion cannot possible apply. * Pppery * it has begun... 15:03, 21 February 2021 (UTC)
Ah, yes, I see you are right. I should have only tagged it as G11. Nsk92 (talk) 15:11, 21 February 2021 (UTC)
Nsk92: what do you mean "I can’t create a red link"? Unless there was a change I didn’t hear a about, any user can create another user’s user page. –xenotalk 14:34, 21 February 2021 (UTC)
Only if it doesn't exist already. When I type 'User:Tisquesusa' in the search window and press 'Go', I am taken to the global user page, User:Tisquesusa. If I press 'Search' instead, I get a line 'There is a page named "User:Tisquesusa" on Wikipedia.' Nsk92 (talk) 14:43, 21 February 2021 (UTC)
When visited locally, it can be edited locally, despite the global presence. Generally if there is abusive content it would simply be turned into a redirect to the user talk page. –xenotalk 14:46, 21 February 2021 (UTC)
But how, exactly? How can it be turned into a redirect or edited locally, if necessary? I can't access the local history log for this user page now. And the edit button for it is not available either. Nsk92 (talk) 14:49, 21 February 2021 (UTC)
Hmm, try clicking here. Maybe it is limited to administrators? –xenotalk 14:57, 21 February 2021 (UTC)
Works for me, even though I'm not an admin. * Pppery * it has begun... 15:03, 21 February 2021 (UTC)
Your link works, and it allows me to create a local user page (it does look like it has been deleted or displaced somehow, presumably by the global user page being activated). But I have no idea how you created this link. Did you just manually type an entire http address? Interesting and strange ... Nsk92 (talk) 15:08, 21 February 2021 (UTC)
I’m using the desktop monobook responsive view; I have some custom scripts but I don’t think they are what’s adding the edit button for me. –xenotalk 15:09, 21 February 2021 (UTC)
Weird. I am also using the desktop monobook view, but I don't have an edit button for that page. Nsk92 (talk) 15:17, 21 February 2021 (UTC)
Nsk92, the link you are probably looking for is "Add local description" which takes you to the edit page. It is confusingly named, but I think it is because it uses the same system as images on commons use (i.e. random file File:Rueda de prensa sobre la sentencia del tribunal europeo acerca de los desahucios en España (8558751810).jpg when viewed on enwiki has the "Add local description" link instead of "Edit"). The "Add local description" link makes more sense for images, as you are adding a local description of the image. However, it doesn't make as much sense for global userpages. Perhaps this should be renamed for global userpages to something like "Create local userpage"? Dreamy Jazz talk to me | my contributions 15:29, 21 February 2021 (UTC)
(edit conflict) Depending on preferences there may be one or more of the tabs "Add local description", "Create", "Create source", "Edit", "Edit source", or MonoBook variants. Don't you have any of them? PrimeHunter (talk) 15:31, 21 February 2021 (UTC)
You are right, there is an "Add local description" button. I didn't realize that it acts as a local edit button (which would presumably override a global user page?) Nsk92 (talk) 15:55, 21 February 2021 (UTC)
← Yes it would Nsk92. Dreamy Jazz, good suggestion for the interface change. The responsive view was giving me an intuitive pencil but in landscape, it is indeed Add local description. –xenotalk 16:07, 21 February 2021 (UTC)
Xeno, if I'm not wrong, would that be achieved by placing the text needed in MediaWiki:Create-local? I also presume that mediawiki pages can contain parser functions and magic words which will work per page (so that the page knows where it is being used and then can modify the wording based on the namespace). It could, if I assume correctly for those both, use a parser function to check if the namespace number is 2 (i.e. userspace) and then output "Create local userpage" instead. Dreamy Jazz talk to me | my contributions 18:38, 21 February 2021 (UTC)
I've created a sandbox for this and when adding "?uselang=sandbox" to the URL it seems to work as intended. So my assumptions were correct. Dreamy Jazz talk to me | my contributions 18:50, 21 February 2021 (UTC)
I think this change is pretty minor, so I'm inclined to sync the sandbox (what is currently in MediaWiki:Create-local/sandbox) to MediaWiki:Create-local. However, I'll wait for a bit in case there are objections. Interestingly this page had no history until I edited it. That seems to be the same with MediaWiki:Create-local. Dreamy Jazz talk to me | my contributions 19:00, 21 February 2021 (UTC)
There are around 26,000 MediaWiki messages and only around 2,000 pages in the namespace. The rest display a MediaWiki default. PrimeHunter (talk) 18:21, 22 February 2021 (UTC)
I've gone and merged the sandbox to the main page so that it shows for everyone with their language set to en. Dreamy Jazz talk to me | my contributions 13:52, 25 February 2021 (UTC)


On, we have

MLA Style Manual

  • Wikipedia contributors. "Julius Lekakeny Sunkuli." Wikipedia, The Free Encyclopedia. Wikipedia, The Free Encyclopedia, 27 Aug. 2018. Web. 27 Aug. 2018.

Shouldn't the second one be Wikimedia Foundation

Also on, we have

  • "Plagiarism." Wikipedia: The Free Encyclopedia. Wikimedia Foundation, Inc. 22 July 2004. Web. 10 Aug. 2004,

There's a contradiction here. Headbomb {t · c · p · b} 17:12, 21 February 2021 (UTC)

"Wikimedia Foundation, Inc." was removed in 2008.[5] "Wikipedia, The Free Encyclopedia" was repetead in 2010.[6] PrimeHunter (talk) 17:35, 21 February 2021 (UTC)

Stop Chrome from scrolling the editing box when I press enter[edit]

I asked this already at the reference desk but was unable to arrive at a diagnosis or solution. When I am editing Wikipedia, Chrome has some weird annoying feature that is horrendously annoying. Usually, but not always, when I press the enter key, the editing textbox will scroll so that the new line is at the very top of the text box, rather than the text box staying where it is and the new line pushing the text down under it (as is the expected behaviour). This only happens on Wikipedia (although I have not found any sites that use plain multiline text boxes to test... everything is script-based these days), and it happens on every computer I try it on. It happens logged out as well as in incognito mode, so this is the behaviour that IPs would be experiencing unless they use the visual editor.

I've used Chrome for years, but this behaviour has only been happening to my knowledge for the past several months. How do I stop text boxes from scrolling when I press enter? I'm up to date with v88.0.4324.150. - Floydian τ ¢ 17:46, 21 February 2021 (UTC)

This happens to me when I paste in one or more lines of text, but it is not consistent. Subsequent pastes do not always make it happen. Using Chrome on Lubuntu.--Verbarson (talk) 15:20, 22 February 2021 (UTC)
He's already tried mw:safemode, and that didn't help. I'm stumped. Anyone else have any ideas? Whatamidoing (WMF) (talk) 00:34, 23 February 2021 (UTC)

Image resolution on mobile[edit]

What is the highest resolution image that will display on mobile? I have noticed that if I go to Commons and click on a high resolution version in some cases it will not load. Is this the operating system or Mediawiki or something else limiting download size? Cheers, · · · Peter Southwood (talk): 18:38, 21 February 2021 (UTC)

Pbsouthwood, Could you be looking at what I described in T270209? -- RoySmith (talk) 18:49, 21 February 2021 (UTC)
RoySmith, My experience has only been with .png files, Android OS, Samsung tablet and a few phones. I haven't done much experimentation because why bother if it is a known limitation. · · · Peter Southwood (talk): 04:33, 22 February 2021 (UTC)

I have had a few friends test. iPhone 11 and 12 seem to manage, but Android phones tested so far fail on the original of File:Dive_sites_of_the_Whittle_Rock_Reef_high_resolution.png ‎(14,040 × 9,930 pixels, file size: 8.71 MB, MIME type: image/png) The next highest resolution available by default is not legible for the small print. Is it likely to be OS, browser, or timeout for some reason? · · · Peter Southwood (talk): 08:50, 22 February 2021 (UTC)

Maybe memory? · · · Peter Southwood (talk): 10:48, 22 February 2021 (UTC)

Pbsouthwood, definitely memory. That’s a gigantic image. It either runs out of memory, or isn’t even able to address images of such a size in some level of graphics libraries it uses. —TheDJ (talkcontribs) 08:32, 23 February 2021 (UTC)
Fair enough TheDJ, the problem is then that the highest resolution available from the default set of pngs is too low to read the text. Is there a way to work around this? · · · Peter Southwood (talk): 12:19, 23 February 2021 (UTC)
Perhaps something more than 1240 and less than 14K? Can confirm Android does not like this file. Slywriter (talk) 12:24, 23 February 2021 (UTC)
Pbsouthwood, with default pngs, do you mean the links in the file page with preset sizes ? In that case, open any of them and modify the url to retrieve a width that is somewhere in between (hmm, just tried that, seems that the thumbnail server is having trouble even scaling something to those in between sizes...It just throws errors) In general I would say.. this is not a suitable file format to store that amount of information... SVG or something might be better. —TheDJ (talkcontribs) 13:35, 23 February 2021 (UTC)
TheDJ, That is what I meant by default pngs. I uploaded a svg of the same map. It is a little better in that it loads on my tablet, but crashes when I try to zoom in to read the text. See File:Map of Whittle Rock dive sites 2021.svg. The svg is also large, and the map will get more detail as it becomes available. Cheers, · · · Peter Southwood (talk): 16:56, 23 February 2021 (UTC)
Correction: The 6,771 × 4,789 pixels version crashes in zoom, but the original svg seems to zoom in just fine. However tapping the image from the article does not load the svg, it loads a bitmap rendering, presumably a png. · · · Peter Southwood (talk): 17:07, 23 February 2021 (UTC)
Pbsouthwood, That's a VINO (Vector In Name Only). Sure, it's packaged in an SVG format, but it's mostly just a bunch of encoded PNG bitmaps. That doesn't solve the problem that you're dragging a huge amount of data around.
I know what I'm suggesting is a lot of work, but I think the only way to make this work is to actually trace the contour intervals so you can generate a real vector version of this. -- RoySmith (talk) 17:48, 23 February 2021 (UTC)
RoySmith, That is the original Inkscape svg from which the png was rendered, so the work has already been done. It is probably possible to trim it down a bit, but not sure how much. · · · Peter Southwood (talk): 18:02, 23 February 2021 (UTC)
Pbsouthwood, I'm not an expert either Inkscape or SVG, but when I look at the data in the file, I see 70 lines that start with
I assume each of those is a bitmap image that you imported into Inkscape. Pulling those lines out, they add up to 78% of the bulk of the original file:
$ du -h x Map_of_Whittle_Rock_dive_sites_2021.svg
7.8M x
10M Map_of_Whittle_Rock_dive_sites_2021.svg
so that's clearly the place to start. Maybe you could down-convert the individual images to something much lower resolution and re-import? PNG uses lossless compression. My hunch is that JPG's algorithms would do a much better job of compression than PNG's on these kinds of images. I would start by converting the original png files to jpg, and experiment with how low you can crank jpg's quality parameter and still get a reasonable rendition.
I'd also guess that you'll find people at Wikipedia:Graphics Lab who are much more clueful about this stuff than I am. -- RoySmith (talk) 18:22, 23 February 2021 (UTC)
[ec]It turns out I could clean it up a lot. Will upload cleaned version and see how it goes. Cheers, · · · Peter Southwood (talk): 18:25, 23 February 2021 (UTC)
RoySmith, I found the same thing by deleting all the construction files I used. I did not think they would be such a large part of the file. Live and learn! Cheers, · · · Peter Southwood (talk): 18:30, 23 February 2021 (UTC)
The svg file is now down to 2.48MB, now how do I persuade the mobile unit to upload the svg not a low res rendering of it? · · · Peter Southwood (talk): 18:58, 23 February 2021 (UTC)
Pbsouthwood, I don't know the answer to that one, sorry. -- RoySmith (talk) 19:15, 23 February 2021 (UTC)
No problem RoySmith, You have been very helpful, and I appreciate your help. Maybe TheDJ has some ideas? · · · Peter Southwood (talk): 19:24, 23 February 2021 (UTC)

Old dumps?[edit]

Would somebody be able to pull me a dump from sometime in the week of February 7 or 14, and throw me a list of the contents of Category:Canadian films as of that date in my sandbox at User:Bearcat/Temp? The category currently has 4,627 articles in it, but I know for a fact (I check it on a regular basis, what with that being my primary editing domain these days) that there were 4,655 just a couple of days ago — and since this is after I added three new films today, that means 31 films have disappeared from the category within the past day or two. One or two, I could handwave away as either an article getting deleted or a technical counting error, but not 31, so I need to figure out what's been disappearing and why. Thanks. Bearcat (talk) 01:45, 22 February 2021 (UTC)

@Bearcat: I've saved the last 130 changes to the category on that page. User:Nardog/CatChangesViewer is good for this sort of thing. — JJMC89(T·C) 02:26, 22 February 2021 (UTC)
Okay, thanks. That definitely helps, as I see a significant cluster of IP edits that entail changing Category:Canadian films to Category:Canadian direct-to-video films. Bingo. I'll definitely bookmark that page for future reference. (Also, by the way, even the 4,655 of a couple of days ago also felt like the category had been shrinking, but I wasn't able to recall an exact prior number to prove it — but it turns out I was definitely right, because I've already got the category back up to 4,659 and still have about 60 more pages to correct.) Bearcat (talk) 02:37, 22 February 2021 (UTC)
Bearcat, is that not overcategorising? · · · Peter Southwood (talk): 04:44, 22 February 2021 (UTC)
The film project has a longstanding consensus that "[Year] films" and "[Country] films" catgories are supposed to be all-inclusive (hence they're tagged with the {{allincluded}} template). Yes, there are some other contexts where that would be considered duplicate categorization — but there are also situations where a WikiProject is allowed to establish a consensus that "all-inclusive" categorization is warranted in certain categories if there are compelling reasons for it. So if you'd like to try to convince WikiProject Film to overturn its prior consensus on that, feel free to take a shot — but the existing consensus is that Category:2003 films is supposed to directly contain all films released in 2003 regardless of whether they're also in subcategories like Category:2003 animated films or not, and Category:Canadian films is supposed to directly contain all Canadian films regardless of whether they're also in subcatgories like Category:Canadian direct-to-video films or not, so people shouldn't be removing those categories without a consensus to change the current practice. (I should also note that when said consensus was established, I was opposed to it on the grounds that film wasn't a special case that needed to be handled differently than novels or TV shows or music where all-inclusive categorization isn't done — but I lost that battle.) Bearcat (talk) 04:54, 22 February 2021 (UTC)
Bearcat thanks for the explanation, I had not encountered this exception before, but I usually only categorise articles that I edit significantly. I guess the IPs hadn't either. This is one of those cases where an exception is causing a lot of extra work for no obvious good reason. Cheers, · · · Peter Southwood (talk): 05:14, 22 February 2021 (UTC)


Is there a way to see WikiProjects by article count similar to Wikipedia:List of Wikipedians by article count. ~~ CAPTAIN MEDUSAtalk 13:08, 22 February 2021 (UTC)

BamBots has a list of 1998 WikiProjects with an indication of activity. Projects generally state a count of different classes of "their" articles — for example WikiProject Books/Assessment statistics. These are created by User:WP 1.0 bot. The bot page tracks 2560 WikiProjects, but I could not find a full list — just thisGhostInTheMachine talk to me 13:39, 22 February 2021 (UTC)
Here's a query that will show all the Template:WikiProject* usage on the article talk pages. -- WOSlinker (talk) 13:42, 22 February 2021 (UTC)
Awesome! Is that query persistent? — GhostInTheMachine talk to me 14:00, 22 February 2021 (UTC)
If you want to run it yourself, just use the Fork button to create your own copy. And then you can run whenever you want. -- WOSlinker (talk) 15:08, 22 February 2021 (UTC)
Thanks WOSlinker. Did that and linked to it in the WikiProject Directory. CAPTAIN MEDUSA That directory seems to be the main location for this sort of report — GhostInTheMachine talk to me 10:49, 23 February 2021 (UTC)

Unsuccessful Login Attempts[edit]

A few editors including myself have been notified of multiple unsuccessful attempts to login using the user's account and a wrong password. It appears that there is someone whom I will call a whacker, a poor imitation of a hacker, trying to log in using various user accounts. I have checked my Contribution History, and all of the edits are edits that I made. I have no reason to think that any of the login attempts succeeded. I have no intention of changing my password, because it isn't compromised, and is working as it should. So I have a two-part question. First, does anyone have a guess as to what the whacker is doing or trying to do? It probably doesn't matter.

Second, I have a technical question. When does the system lock out an IP address? Is there a rule so that after 3 or 5 unsuccessful attempts to log in from an IP address, the device is locked for a reasonable length of time, such as an hour? It isn't so much a matter of protecting the system, since strong passwords appear to be a sufficient protection. It is more a matter of maybe discouraging the whacker.

Third, and this is just a wild guess. I wonder whether the whacker is hoping that one of the humans behind the accounts will panic and will change their password to something easily guessed. Don't do that, then. Robert McClenon (talk) 19:58, 22 February 2021 (UTC)

Simplify marking WikiProject status on talk page[edit]

See discussion Template talk:WikiProject banner shell#Add class/importance attribute on shell and please advise where else is right forum to discuss. I know it's potentially a major change. Shushugah (talk) 20:15, 22 February 2021 (UTC)

Emptying a category[edit]

Is there an easier way to empty a deleted category than manually removing the category from each article? Fences&Windows 23:40, 22 February 2021 (UTC)

That sounds like a bot-like task. Could a bot be written to remove the deleted categories from the articles after the categories are deleted? That isn't easier for the bot coder, but it sounds as though it would make it easier for the closer of the CFD. Robert McClenon (talk) 23:54, 22 February 2021 (UTC)
I know little about the details of categories. If my comment above is completely off the mark, then I know even little less than I thought I do. Robert McClenon (talk) 23:54, 22 February 2021 (UTC)
There's already a bot to implement discussions at CfD, which reads its instructions from Wikipedia:Categories for discussion/Working. * Pppery * it has begun... 23:57, 22 February 2021 (UTC)
@Fences and windows: Quite so; if a category is deleted legitimately, i.e. via a WP:CFD, the normal procedures of that forum will pass the actual empty-and-delete task to an adminbot. On occasion, a category is populated by a template - for these, an amendment to the template code should be sufficient. There is normally no need to empty a category manually. --Redrose64 🌹 (talk) 10:43, 23 February 2021 (UTC)
Aha, thanks Pppery and Redrose64; I haven't done CfD for a long time. The context is categories created by a banned user, which have been tagged for speedy deletion - should those then go via CfD? Specific example: Category:Native American superheroes tagged by MarnetteD. Fences&Windows 11:41, 23 February 2021 (UTC)
If the category page has not been deleted, file a WP:CFD pointing out that WP:CSD#G5 applies. If it's already deleted, it's probably best to start a thread at WT:CFD/W asking for assistance. --Redrose64 🌹 (talk) 21:47, 23 February 2021 (UTC)

Tech News: 2021-08[edit]

00:16, 23 February 2021 (UTC)

Reply tool[edit]

Just checking in to see if anyone else has information for me about mw:Talk pages project/Replying. I'm talking to the devs about offering this as a default-off Beta Feature here. There are lots of comments at Wikipedia talk:Talk pages project#Experiences from people who've been using it, but we might be biased. ;-) So if you're aware of any problems that make you nervous about having more people try it out, or if you've been secretly cleaning up messes every time I use it and you just never wanted to mention it, please ping me. (Try it out here: and look for the [reply] buttons. If you switch to the visual mode, it'll make it really easy to @-mention editors. As long as nobody breaks the page with, say, an unescaped half of a parser code, an unclosed div tag, or a broken wikitext table, then it should work even on a page as big as this one.) Whatamidoing (WMF) (talk) 00:45, 23 February 2021 (UTC)

Displayed output of {{tl|sfn}}[edit]

Hi all, I've just learned in another lengthy discussion above, (latest posts in #Further thoughts on refs), that my experience of {{sfn}} is somewhat different to at least one other editor. When I hover over a ref generated by an {{sfn}}, it shows me the blue-linked short footnote found in the {{reflist}}, eg Harnden 2011, p. 206. Remaining in the main body of text, If I move my mouse a fraction and hover over Harnden 2011, it displays a second box (not sure of the exact name) which contains the usual output of {{cite book}} which generated the CITEREF in the first place. Clicking on that takes me straight to the Bibliography (or wherever the {{cite book}} is located}. Is this the default behaviour, or is it just me? I'm using ancient FF v47 and XP SP3, which shows certain quirks (eg recently all lists have one column only, whatever the width eg {{reflist|20em}} NB That's not the problem here.) The other editor says he has to click on the ref to get to the reflist, and then again to get to the Biblio, and has to click back twice to get back to where he started. So, if my experience is the proper sequence, why might other users not be able to see the second box through a hover? As far as I know I have no script installed which might do this, but I am very un-technical in this respect. Cheers, >MinorProphet (talk) 05:52, 23 February 2021 (UTC)

When I hover over a shortref, it shows me no additional pop-up/second box. I have to click on that reference to see the whole citation. The whole citation then of course doesn't have a link back to either its use in the text or its shortref. The lack of immediate information in the pop-up, and the lack of the ability to right click a source link from that pop-up without darting around other areas of the page, makes me prefer longform citations. CMD (talk) 06:57, 23 February 2021 (UTC)
@Chipmunkdavis: Hi, I'm not surprised you prefer longform, I'm sorry that your experience seems so poor compared to mine. What browser and OS are you using, to help narrow down where exactly this behaviour occurs? (Could be just me, only 1% of users worldwide use XP, max. possible version of FF is 52.9, and is now at v.85...) MinorProphet (talk) 07:28, 23 February 2021 (UTC)
(I'm to blame for your reduced viewing experience of columns, but you're now one of the <2% using a browser that can't handle them with standard CSS, and most who can't are in fact stuck on Chrome mobile from the same time period... The use case for which will almost always be a single column. I judged that it would be reasonable to stop sending 3x the CSS just for that 2%.)
Out of curiosity, what color is the second box? Secondly, when you go to Gadgets in your preferences, do you have Popups enabled? Third, what options are checked on the Beta page in your preferences? --Izno (talk) 07:35, 23 February 2021 (UTC)
Aha. I didn't even know a second pop-up was possible before this post. Following this tip, I disabled pop-ups in the Wikipedia gadgets, and now I get a second pop-up. How counterintuitive. CMD (talk) 07:44, 23 February 2021 (UTC)
The two kinds of pop ups not playong each other nicely is a known issue, largely caused by the fact that the nav popups gadget was first written some 15 years ago and received little in the way of updates. The other is more modern... --Izno (talk) 09:06, 23 February 2021 (UTC)
(edit conflict) Hey, Izno, I really can't complain if I use software invented before they burnt down the Great Library of Alexandria.
  1. All the popups are b/g plain white, f/g black text, links in blue.
  2. If you mean Navigation popups: article previews and editing functions pop up when hovering over links, then no, it's not checked. On the other hand, I do get article previews.
  3. No Beta features enabled at all.
  4. Aha! How about Browsing → Reference Tooltips, which is checked. Just disabled it, and all the incredibly useful info disappears and a vast sadness rolls over me. Restored Tooltips, and all is light and joy. Why on earth is this not default behaviour? All the stunning usefulness of sfn and cite book is otherwise completely negated. >MinorProphet (talk) 08:06, 23 February 2021 (UTC)
Obviously respect and thanks for pinpointing the source of the behaviour so quickly. MinorProphet (talk) 08:43, 23 February 2021 (UTC)
My memory says it is the default behavior for accounts registered after a certain date, and I think there was some discussion about making it the default viewing experience also for unregistered editors. But I am unsure. That all said, I've been watching the discussion above and would like to point out the books referencing project, which, whenever WMDE gets off its butt to finish it, will allow one to get both the lesser and fuller citation in one popup (or at least that is in the current backlog), at least for articles which convert to that style. --Izno (talk) 09:06, 23 February 2021 (UTC)
Reference tooltips works for me when logged out and on other browsers, so it might be the default in newer browsers. CMD (talk) 09:03, 23 February 2021 (UTC)
The browser should have nothing to do with it. Whether you get ReferenceTooltips or one of the others is down to your Wikipedia user settings. ReferenceTooltips first became available as a gadget on 4 April 2012, and was initially opt-in for all logged-in users. At 00:26, 17 July 2012 (UTC) it became a default gadget, which means that it was enabled for logged-out users, and was also enabled (on an opt-out basis) for logged-in users registered after that time. Logged in users who registered before 17 July 2012 did not have their setting altered, so unless they had already enabled the gadget, it remained switched off for those users. --Redrose64 🌹 (talk) 23:25, 23 February 2021 (UTC)
If I turn off popups then the ref tooltips thing seems to work, but I don't get page previews. Even if I did I understand that page previews lack the edit/history/talk etc links that popups have and which make them so invaluable. DuncanHill (talk) 10:11, 23 February 2021 (UTC)
There are at least three different kinds of popups, which are not compatible - if you turn one of them on, you also need to turn off all of the others. At Preferences → Gadgets we find:
  • Navigation popups: article previews and editing functions pop up when hovering over links
  • Reference Tooltips: hover over inline citations to see reference information without moving away from the article text (does not work if "Navigation popups" is enabled above)
and at Preferences → Beta features there is Reference Previews which shows
  • Show a preview of a reference by hovering over its footnote marker.
    Please note: If you’re using the Navigation popups gadget or the Reference Tooltips gadget, you won’t see Reference Previews.
The default is that Reference Tooltips is enabled, the other two disabled. Perhaps the compatibility warnings can be strengthened. --Redrose64 🌹 (talk) 10:56, 23 February 2021 (UTC)
Would it be technically possible to unbundle the reference tooltips from the rest of Navigation_popups, if that is the only redundant aspect to ReferenceTooltips? Seems odd there are two gadgets doing the same job with one being worse at it. CMD (talk) 11:07, 23 February 2021 (UTC)
Do you mean stop nav popups displaying the ref when you point at it? That would just make things worse. DuncanHill (talk) 11:23, 23 February 2021 (UTC)
Ber-limey, what an utter and seemingly irreconcilable mess! Welcome to It_Wasn', aka the daily refuge of maintenance coders the world over. Although I laud everyone's technical and/or coding ability; also people's apparent understanding of the technical depth of what is going on; and our exceptionally refined ability to discuss both technical and emotional issues in the same breath, it appears we may have arrived at a pretty pass. There seem to be a number of previous issues which have caught up with us, and a number of current issues which seem unresolvable unless certain decisions are taken. And I suggest that the responsibility rests with us, here, now. "What, then (sang Plato's ghost), what then?"[10] >MinorProphet (talk) 12:37, 23 February 2021 (UTC)

Best way to intra wiki link multi language edit-a-thon?[edit]

I hosted in the past, an edit-a-thon at Wikipedia:WikiProject Organized Labour/Online edit-a-thon Tech February 2021, however there are no corresponding Labor WikiProjects in German or Italian Wikipedia, so I hosted the pages in my personal username space, e.g. see de:Benutzer:Shushugah/Tech-Arbeiter*innen_Edit-a-thon, and manually inter wiki linked them, since they're ineligible for a WikiData item due to being in username space. Another option would be to...boldly create a WikiProject in German wikipedia, but Italian and other language wikipedias don't have such option.

I considered hosting on Meta Wikimedia, however that complicates linking to Wikipedia articles, which is a deal breaker for welcoming new editors. Shushugah (talk) 13:47, 23 February 2021 (UTC)

@Shushugah: what is preventing you from linking to articles from meta-wiki? Here is an example: meta:Special:PermaLink/21143516. — xaosflux Talk 15:17, 23 February 2021 (UTC)
I regularly link across Wikipedia languages and Wikipedia projects, but requiring new users to prefix every article with the Language AND Project is a high technical burden. Having a Meta place holder to encourage new Wiki platforms does make sense though. Shushugah (talk) 15:45, 23 February 2021 (UTC)

Watchlist parameter[edit]

Hi all,

my question is when I enable the HIDE->bots parameter - in order to reduce the size of my watchlist - the problem is the hidden articles may contain as well non-bot new edits before the latest/recent bot edit, that have been made earlier (but was not present in my earlier listing, since those edits did not have been made then, but between my two listings). With other words, in case the last=(recent) edit has been made by a bot, then that article will not displayed on the actual listing of the watchlist, and it will remain until a non-bot edit will be performed, however, if again a bot edit would follow (and meanhile I would not make a fresh listing by any purpose), again it would be/remain hidden, and I would not know anything about any intermediary non-bot edits, like that. This is the problem. I would like just to hide those articles in my actual listing, which had just and only both edits from a given time, but not the non-bot edits have been done after the given time, but before the recent/latest bot edit. I hope I formulated my problem in a way to be correctly understood. Thank You.(KIENGIR (talk) 22:34, 24 February 2021 (UTC))

@KIENGIR: this is known issue phab:T11790, not something we can fix directly here on the English Wikipedia. You can follow that ticket for more details. — xaosflux Talk 00:40, 25 February 2021 (UTC)
Thank, I looked through the thread, but as I see the outcome tended the opposite direction I wish to achieve...or I would be wrong?(KIENGIR (talk) 00:55, 25 February 2021 (UTC))
That problem is still "open" meaning that no one has implemented a fix for it, yet. — xaosflux Talk 01:11, 25 February 2021 (UTC)
I have not tried this but elsewhere, others have suggested this:
Trappist the monk (talk) 01:41, 25 February 2021 (UTC)

Edit filter to prevent http: //%5B(and cousins) making good links bad[edit]

So while playing on toolforge and running a search on %.%[11](which terrified me that I would blow up a server), I stumbled onto the fact that potentially hundreds of links/refs are broken because http://%5b (as well as %5b%20 , %5E and a few others) are appended to the beginning of the url.

Working my way through cleaning up what I can but wondering if an edit filter would be viable to that warns an editor adding a double http:// in an edit summary (I say warn because possible their are valid reasons - Interner archive comes to mind) Slywriter (talk) 02:31, 25 February 2021 (UTC)

Here is an example one added. No apparent cause. 10 years ago. Unclear why it's happening. %5B = open square-bracket. Old MediaWiki bug? -- GreenC 02:53, 25 February 2021 (UTC)
Wow, sorry, totally forgot to add examples. I've already cleaned up a bunch but let me go back through them and see if there is any similarity in when the edits were added. Slywriter (talk) 03:05, 25 February 2021 (UTC)
2012 Special:Diff/472050129
2018 Special:Diff/822925977
For it to be a bug, it would be a long standing one. Not seeing any commonality in tags either (web, mobile, iOS, visual)Slywriter (talk) 03:26, 25 February 2021 (UTC)
Yeah hard to say. BTW notice some of the URLs have trailing %5D that also should be removed. It may be pretty difficult to automate a fix given how many forms it takes. -- GreenC 04:18, 25 February 2021 (UTC)
Like this URL http://%5B in Drugyel higher secondary school should be [ Drukgyel school]
5B & 5D there. Cut and Paste of a wiki link gone wrong?
Tested what happens when you manually type/cut and paste [] into the edit screen(&visual editor manual cite wizard) as [http://[]] but that didn't trigger a change to unicode, the brackets remained.
Doubt I'll have much luck figuring out the how they do it but the brackets showing up give me the idea to search against other special wiki characters (<>{}~) unicode equivalents. Might be some other unicode characters that will cause obvious problems at beginning or end of a url.
If the double html remains constant, may be able to at least automate the generation of a list that can be manually checked. Slywriter (talk) 05:31, 25 February 2021 (UTC)

With some sleep and a new look at the data, assuming I am correct in my assumption that my search string is returning an alphanumeric sorted list of ALL websites links found in article space, then the problem is less than a 100 links. Only reason I question my assumption is the surprisingly low number of IP only web addresses that are used for links. Slywriter (talk) 14:49, 25 February 2021 (UTC)

It can happen if you try to use the toolbar link feature on code which already has external link syntax. An article may say [ Drukgyel school]. Click the chain icon in the toolbar and paste the code in the top field. You get a warning and cannot save. If you select "To an external web page" then you can save but it becomes [http://%5B [ Drukgyel school<nowiki>]</nowiki>]. I don't know whether nowiki was always added. PrimeHunter (talk) 21:45, 25 February 2021 (UTC)
Thanks for that explanation. The sequence also explains why it's mostly, though not exclusively, low quality pages that its occurring. Situations where it seems the editor was determined to get the link in.
I've found a few other bad link styles like web addresses starting with a period . ::For now, I'll work on cleaning up through toolforge but once I have a good list of the errors is here the best place to see if someone with better backend knowledge can setup a query that I can run on my own or updates a page to show likely bad URLs? Would be better for long term maintenance and I suspect lower overhead than me running constant toolforge searches. Slywriter (talk) 22:08, 25 February 2021 (UTC)

How do we disable magic links?[edit]

This 2017 en.WP RFC determined that once bots and scripts have replaced magic links (ISBN et al.), magic links should be disabled. As far as I know, bots and scripts have done this work (except for new untemplated and nowiki-wrapped ISBNs that are added continually by manual editing and buggy VE copy-paste editing), and the ability to disable magic linking locally has been provided.

This MW page appears to explain how to disable magic linking on a per-wiki basis. It is past time to do it here, if we have this local control and if there are no show-stopping feature requests that depend on magic links. There are a bunch of magic-link-related bugs and feature requests linked at T145589, but I am unable to determine whether any of them will be affected if we disable magic linking locally.

Is there anyone here with the sysadmin-level knowledge to figure out whether we can finally disable magic linking here at en.WP? – Jonesey95 (talk) 03:18, 25 February 2021 (UTC)

servers crashing[edit]

An error occurred while attempting to preview your changes. The server did not respond within the expected time.

11 times...... 0mtwb9gd5wx (talk) 08:56, 25 February 2021 (UTC)

Latest version of Firefox appears to break Wikimedia single sign-on[edit]

Firefox 86.0 comes with "total cookie protection" (see ) that sandboxes cookie storage to prevent cross-site tracking by third party cookies. Unfortunately, it also seems to break the Wikimedia single-sign-on mechanism. See for a description of how it works.

This clearly either needs WMF liason with Mozilla to whitelist the single sign on mechansim, or to use the Storage Access API to request it be permitted on a site-by-site basis. Where is the best place to report this? -- The Anome (talk) 09:18, 25 February 2021 (UTC)

The Anome, welcome to the club. My Safari ticket is here phab:T226797TheDJ (talkcontribs) 09:42, 25 February 2021 (UTC)
This clearly needs to be fixed ASAP -- how do we get the WMF's attention on this? -- The Anome (talk) 10:03, 25 February 2021 (UTC)
But does it need to be fixed if it isn't broken? Using firefox as always, version 86, has continued to give great service, and hasn't broken yet. just an observation. -Roxy the grumpy dog. wooF 11:03, 25 February 2021 (UTC)
@The Anome: to be clear, simply having FF v86 doesn't appear to be breaking logon, does it? Is this an opt-in feature that users have to select? (I think the option is labeled All third party cookies (may cause websites to break)). — xaosflux Talk 12:06, 25 February 2021 (UTC)
Logon is fine, cross-site logon between sites ending in "" is also fine. What doesn't work is SSO between top-level domains, such as Commons or Wikidata, if you log in from a site. -- The Anome (talk) 12:13, 25 February 2021 (UTC)
@The Anome: is this only in "strict mode" ? — xaosflux Talk 12:23, 25 February 2021 (UTC)
I believe so. But the system sould support strict mode, as this is what any privacy-conscious user should be using already. -- The Anome (talk) 13:46, 25 February 2021 (UTC)
In the mean time, can anyone verify if someone enables this strict mode (that warns it may be breaking) if they can still exempt site-by-site as talked about in this article? — xaosflux Talk 14:06, 25 February 2021 (UTC)

Ugly workaround[edit]

I threw together a hacky little user script at meta:User:Suffusion of Yellow/central.js. Works for me with FF86/Linux with all third party cookies blocked. Install in your global.js, then click "Central login" under "tools" on any site you are currently logged in to, and follow the directions. You will be logged in to all the other wikis, too. Suffusion of Yellow (talk) 21:26, 25 February 2021 (UTC)

Technical help needed on the helpdesk.[edit]

Technical minded people may want to take a look at Wikipedia:Help_desk#Series_of_failed_pings. I can't figure out what's going wrong with this person's pings. ~ ONUnicorn(Talk|Contribs)problem solving 20:19, 25 February 2021 (UTC)