Wikidata:Project chat/Archive/2023/06

From Wikidata
Jump to navigation Jump to search


Merge candidate?

The English Wikipedia commonly circumlocutes in its article titles to work around the BLP1E policy, which says that people notable for a singular event do not deserve a biography article. Therefore, the article is scoped to the event itself. Mortara case (Q108126419) and Edgardo Mortara (Q438681) are today's example. The former, 16 articles correspond to the person Edgardo Mortara, and in the latter, 7 articles about the event itself. eswiki alone is included in both lists.

Could these items be merged? Is there a normal treatment of such dichotomies when enwiki creates them? If I could convince the eswiki community to merge the two articles, would that clear the way for a merge here? Elizium23 (talk) 23:18, 3 June 2023 (UTC)

No, Wikidata an item about a criminal case is separate from that of a person so they don't belong merged on Wikidata. On Wikidata we only merge items that are about the same thing. ChristianKl ❪✉❫ 00:11, 4 June 2023 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. --Elizium23 (talk) 16:00, 4 June 2023 (UTC)

Proposed config change: Add pagelang right to wikidata-staff group

Tracked in Phabricator
Task T337760

Hi all,

the Wikidata staff group is supposed to have all the rights that administrators, bureaucrats and translation administrators have (according to this 2014 RFC), but I noticed that it’s currently missing the “Change page language” right, which was introduced two years after the group was added. I suggest that we fix this by adding the right to the user group, so that it’s properly redundant with translation admins again like it’s expected to be. Do you agree? Lucas Werkmeister (WMDE) (talk) 16:01, 30 May 2023 (UTC)

seems reasonable BrokenSegue (talk) 16:05, 30 May 2023 (UTC)
+1 --Ameisenigel (talk) 19:48, 30 May 2023 (UTC)
+1 —MisterSynergy (talk) 19:49, 30 May 2023 (UTC)
+1 --Matěj Suchánek (talk) 20:33, 31 May 2023 (UTC)
I instead propose that the staff group will inherit all rights from admin and translation admin via mw:Manual:$wgGroupInheritsPermissions.--GZWDer (talk) 14:51, 1 June 2023 (UTC)

Qualifier editing

The qualifers are presented in vertical columns each character on a different line. I cannot find any way of entering the data for any new qualifier or for correcting the existing data if needed. This applies also for references. What am I doing wrong? Afil (talk) 04:02, 1 June 2023 (UTC)

It's the same edit link of the statement. Bean49 (talk) 09:19, 1 June 2023 (UTC)
What is your browser width? This sounds like you are using a browser that is under 1000 pixels wide so all the text ends
u
p
l
i
k
e
t
h
i
s
I think it starts going wrong around 1000 pixels wide, here's an example at 890 pixels
http://whatismyscreenresolution.net/multi-screen-test?site-url=https://www.wikidata.org/wiki/Q42&w=890&h=600 Piecesofuk (talk) 09:22, 1 June 2023 (UTC)
Might be related to an issue some mobile users are having at the moment? https://www.wikidata.org/wiki/Wikidata:Report_a_technical_problem#Page_jumbled_up_in_mobile_view Infrastruktur (talk) 13:58, 1 June 2023 (UTC)
Suggest you raise it on Wikidata:Report a technical problem. The previous issue was allegedly fixed & the phab ticket closed - https://phabricator.wikimedia.org/T336956 --Tagishsimon (talk) 14:45, 1 June 2023 (UTC)
Thank you. I am using a PC and usually use the Microsoft Edge browser. For some reason the width is different if I use Chrome, where the problem dissapears. I will use Chrome when I need to access qualifiers. Thank you for the clue Afil (talk) 00:50, 2 June 2023 (UTC)

Meta (name)

There are two objects called "Meta" that refer to the name of a person. One (Q1437426) is used by the English Wikipedia article "Meta (name)" and the other (Q12796400) is used by several Wikipedia languages (de, et, nds, no, sl, sv). However, these objects should probably not be merged, since the first one refers to "given name and surname" and the other to "female given name". How should this be resolved? Right now, the English Wikipedia article "Meta (name)" does not have any links to the others, but that is very confusing. Should the Wikipedia articles be renamed? How should the Wikidata objects be linked? Ahltorp (talk) 12:18, 1 June 2023 (UTC)

There's also the family name (Meta (Q57167572)) and a Japanese family name (Meta (Q26000745)). I moved the Commons link from Q26000745 to Q57167572, but only 3 of the 5 people in the Commons category are Albanian so I don't know if Q57167572 should just have Albanian as the language or should be changed to multiple languages. I have seen combined items linked to the given name and family name with has part(s) (P527), and with part of (P361) on the separate items linking back, for example Stuart (Q26902542) "given name and surname", Stuart (Q7626248) "male given name" and Stuart (Q23041619) "family name". Peter James (talk) 19:05, 1 June 2023 (UTC)

Technical issue

Please have a look at this topic. I would like to import a certain number of IDs, but until this isn't solved, I can't... 2A01:CB1D:8CC3:6500:2D8E:560:2681:4BAD 12:53, 2 June 2023 (UTC)

How to sort values within a property?

How can I get multiple values of a Wikidata property to sort in a desired order?

In particular, how can I get multiple values within total revenue (P2139) to sort in date sequence in Pregnancy Justice (Q30288051)? When I use it in Wikipedia, the last "Revenue" entered, in this case for 2017 is displayed. I want the most recent (i.e., 2019), but that's not what I'm getting -- and w:Pregnancy Justice shows the value for 2017 NOT 2019.

?? Thanks, DavidMCEddy (talk) 16:51, 2 June 2023 (UTC)

I set the most recent data to preferred and now it works. I don't know how the Wikipedia-Wikidata integration for revenue works but seems it doesn't account for the date qualifier. BrokenSegue (talk) 17:05, 2 June 2023 (UTC)
Thanks. DavidMCEddy (talk) 17:28, 2 June 2023 (UTC)
@DavidMCEddy: ff the rank is not set (it is now) but you know what the latest date is, you can use calls to module Wd like
{{#invoke:Wd|property|eid=Q30288051|P2139|P585=2019}}→2,593,376 United States dollar
{{#invoke:Wd|property|eid=Q30288051|P2139|P585=2017}}→1,993,978 United States dollar
I don't think that's possible with Module:WikidataIB though. Ponor (talk) 17:43, 2 June 2023 (UTC)

Etiquetes de propietats

Hello,

We are analyzing the Wikidata ontology from the user's perspective. One of the areas of evaluation is the functioning of the Wikidata Query Builder. We would like to know that the categories or descriptors for the values ​​that can be selected within fields Property and Value correspond to the labels of these elements. If this is the case, we understand that if tags have not been assigned to properties and items, the categories or descriptors will not appear in Wikidata Query Builder queries. We couldn't confirm whether in the case of properties (which accept a single tag per language) the corresponding tag may not be available in all languages. In the case of items, we have the page User:Mr. Ibrahem/Language statistics for items https://www.wikidata.org/wiki/User:Mr._Ibrahem/Language_statistics_for_items detailing the number of untagged items for each language: There is a similar page for properties.

Thank you so much.

Miquel Centelles Miquel.centelles2023 (talk) 10:26, 2 June 2023 (UTC)

If you want to understand Wikidata's ontology it would be useful to learn the terms of the elements of the ontology. "Tag" is not a word that refers to anything in the Wikidata ontology.
Also, who's "we"? ChristianKl ❪✉❫ 17:21, 2 June 2023 (UTC)
https://fima.ub.edu/directori/ficha17 --Arnoseven (talk) 21:53, 3 June 2023 (UTC)
Potser en Miquel seria més benvingut si no se'l ensenyava per un error de traducció automàtica. "Etiquetes" in Catalan can mean "tag" or also "label" so I think this editor has a fairly good grasp of terminology already. Elizium23 (talk) 22:14, 3 June 2023 (UTC)
@Miquel.centelles2023 Yes, Query Builder will autocomplete based on labels of existing items and properties. If a property or an item has not been assigned a label, it will not be discoverable using that specific text string. Vojtěch Dostál (talk) 18:41, 4 June 2023 (UTC)

Problems with two publishing houses

I have some problems with two publishing houses: Electa (Q17538695) and Edizioni Polistampa (Q65945053). The problem is about WorldCat Identies and Entities. Could you help me? Thanks.-- Carnby (talk) 08:44, 4 June 2023 (UTC)

WorldCat Entities property says it's an "identifier for a person, work, or place", and I've noticed that identifiers that referred to groups of people or organisations that had a WorldCat Identities ID were not included in WorldCat Entities. As there is no identifier in Entities, the Identities URLs redirect to VIAF or return "not found". I added a Library of Congress authority ID (P244) to Q65945053 as that was the source of the WorldCat Identities ID (superseded) (P7859). Peter James (talk) 16:53, 4 June 2023 (UTC)

Empty disambiguation page items

Listed one of these for deletion after fixing a page move today, and on examination turned up a hundred or so - https://w.wiki/6nnH - should they all just be deleted? I'm guessing there is no real value in keeping these around.

(I tried running a similar query on categories but it timed out; templates gets 12) Andrew Gray (talk) 11:19, 4 June 2023 (UTC)

Most of them are Japanese names, and never had any sitelinks but they are linked from other items. Peter James (talk) 14:52, 4 June 2023 (UTC)
There is a related report at Wikidata:Database reports/to delete/empty disambiguation items. These items are difficult to delete because many of them are still linked from somewhere else, and often their sitelinks have been inappropriately moved to another (name) item. Since nobody is really trying to fix this, many of these items qualify for deletion as they do not have any use in this form. ---MisterSynergy (talk) 15:24, 4 June 2023 (UTC)

Incorrect merge new user

I am no Wikidata expert but got a notification for a German Wikipedia article that was connected to this Wikidata entity.

I am pretty sure this was merged incorrectly: Fonthip Watcharatrakul (Q3747508). I can’t do anything about it but notify here. Frupa (talk) 19:21, 4 June 2023 (UTC)

@Frupa Danke für den Hinweis, ich habe den Merge korrigiert! --Emu (talk) 19:35, 4 June 2023 (UTC)

Questions regarding images at Wikimedia Commons

I have uploaded an image to Commons.[1]

In that image at Commons, I have stated that it depicts a merchant vessel in the section titled "Items portrayed in this file." I have two questions:

  1. Should I create a unique Wikidata Item for that image?
  2. How is the wikidata I already added to the image in Commons used if I haven't created an item for the image?

Thanks in advance...

Michael.C.Wright (talk) 16:32, 5 June 2023 (UTC)

Hi. As I understand it Wikidata generally links to category pages on Commons, not to individual files, have a look at Ever Given (Q104378621). If you know the identity of the ship and want to use it for a 'depicts' statement on Commons, it's possible to add a new Wikidata item for that ship, but new items should comply with Wikidata's notability criterias, which basically means the ship needs to be covered on another Wikimedia project site (and this includes Commons, so if Commons have a category page and images for the ship in question, you're good to go).
Adding structured data at Commons is valuable in that it greatly aids searching for photos, more so than simple categorization can, if you're only interested in pictures of turtles that are taken at some location in Australia, queries can do that sort of thing. If combined with the data available in Wikidata this allows for some interesting and powerful use-cases. More information here: commons:Commons:Structured_data . Infrastruktur (talk) 17:37, 5 June 2023 (UTC)
Thank you. I don't know the identity of the ship and the image is now locked so I couldn't edit it anyway. Michael.C.Wright (talk) 20:31, 5 June 2023 (UTC)

Paucabot

Elizium23 (talk) 21:58, 2 June 2023 (UTC)

Background: User talk:Paucabot#Deletion of media legend (P2096).
I didn't notice this thread in which I was being treated as a vandal.
I have stopped from doing this kind of editions and I will not do any more editions like that before we reach a consensus, but I would need the community to understand what's I was trying to achieve:
In 2020, a bot (JoRobot) uploaded legends of picturesfrom catalan wikipedia, but moved everything to media legend (P2096) because the majority were legends. But there was a minority that should have gone to point in time (P585) and that are the ones I am trying to fix now.
I was not deleting information. Take a look at on reverted edit: Special:diff/1906689365. It says, in catalan, John Lydon el 2010 (John Lydon in 2010). So, I have put the only information that has this sentence in the right place (point in time (P585)). There is no loss of information.
I was doing that because in infoboxes in catalan wikipedia can import point in time (P585) as a qualifier of image (P18). If it's not done that way, the date would be redundant as you can see, for example, in ca:Beatrix Potter.
The issues raised by Termininja and Koavf were solved. The latter was a problem of date precision (which I have already corrected) and the former seemed solved as the user accepted my explanation.
Now I have two more objections by Elizium23 anf Tagishsimon that I don't agree with and I would like to hear other people. I really think I'm improving WD in doing so and I don't think I am breaking any rule.
Thanks, Paucabot (talk) 13:59, 4 June 2023 (UTC)
You are not improving WD. You are removing valid statements - media legends - for no good reason. Point in time is not a substitute for media legend, fullstop. Media legend is a string which can be used without further computation. To achieve a media legend from a point in time statement involves fetching additional info from elsewhere. They are not the same. There was nothing wrong with the media legend. Imposing your personal preference and deleting the valid work of other people is sociopathic and wrong. Please stop. --Tagishsimon (talk) 14:05, 4 June 2023 (UTC)
As I told you before, we can disagree on things, but there is no need to be impolite. Paucabot (talk) 14:08, 4 June 2023 (UTC)
You are vandalising WD. What's polite about that?
Be clear: it is fine to add point in time statements if you wish. Fine to create captions on Catalan WP using Lua, if that floats your boat. It is not fine to trash other people's valid work on WD just because you feel like doing so. No-one us going rock up to this discussion to say 'yes, it's fine, trash valid stateents b/c you think your way of doing things is superior'. The only question is how long it will take you to get the message. --Tagishsimon (talk) 14:14, 4 June 2023 (UTC)
I stopped doing this edits when you wrote me on my discussion page and you keep telling me to stop ...
Believe it or not, I was thinking I was doing a good job because these media legend (P2096) are in catalan and, I thought, that way could be multilingual and used by other wikipedias, but I have come to realise that I lost some time ago the presumption of innocence. Paucabot (talk) 14:27, 4 June 2023 (UTC)
You were not doing a good job. You were destroying the valid work of other people because you thought you knew better. You did not know better. Ideally you will now reinstate the media legends you deleted. --Tagishsimon (talk) 14:29, 4 June 2023 (UTC)
Would it be ok for you if the owner of the bot that made these editions (Joutbis) agreed with those changes? Paucabot (talk) 14:31, 4 June 2023 (UTC)
You say you're "moving" data from one property to another. We'd first prefer that you copy data instead, so that the original property is not lost, doesn't that make sense?
For me, keeping media legend (P2096) is absolutely redundant as you can clearly see here. Paucabot (talk) 16:37, 4 June 2023 (UTC)
ffs, you are so so wrong. Yes, one can create the legend from a kit of parts. But that is not the same as being able to use the pre-made string. Stop thinking that Catalan WP is the only client for media legends. Start thinking that any user should be able to access a media legend without having to write code to construct the string. --Tagishsimon (talk) 16:41, 4 June 2023 (UTC)
Stop telling me I'm wrong. We have a disagreement, and I haven't told you once you're wrong (you're not, you only have a different opinion).
I understand that WD is not only for Wikipedias use. Having said that, for third parties to use WD, it would be more useful to have the date of the picture separated from the legend than written in catalan. The catalan community is willing to fix this (have you seen it?) and you're pressing against it for a thing that would improve third parties using WD in all languages. Paucabot (talk) 16:50, 4 June 2023 (UTC)
Secondly, there are objections to the precision of the dates you're adding and to the methods you're using to determine them. It's going to be fraught with inaccuracy to determine the date of a photo based on the data available to us. There is the image description and structured data on Commons, there is EXIF data, and there are the media legends, and these may differ. Does your bot use any heuristics, draw from more than one source, or alert you in case of any conflict? At any rate, I think it would be best when adding a point in time, that you also add a note explaining how you derived it, so there is a trail to follow.
Does this sound reasonable? Elizium23 (talk) 15:52, 4 June 2023 (UTC)
That I could do (add that the date comes from legends from catalan wikipedia, I mean). The dates are as reliable (or not) as the media legend (P2096) are. So, they could have problems, but if it is the case, the problems are already there. If it helps, I promise to double check every point in time (P585) added with commons info. Paucabot (talk) 16:43, 4 June 2023 (UTC)
For the record, I think I am the "other people" whose work is Paucabot apparently vandalizing. I am the operator of JoRobot, and the work I did was entirely on the suggestion of Paucabot. I consider it a common project. Since bots do bulk work, I find it reassuring that a human reviews the work. Let me make it clear that Paucabot is actually a real person, despite the "bot" ending. All the work done on legends by JoRobot and Paucabot should be considered as a whole project. The bot does the brute force, and the human does the refining. And if the human is Pau, then rest assured that the result will be top class.--Joutbis (talk) 16:28, 4 June 2023 (UTC)
Thank you, and pardon my confusion regarding the bot status of Pau. I am still familiarizing myself with policy here, and I had assumed that an account ending in "bot" was designated and flagged, but this account is not. (Is "Cabot" a surname?) Perhaps, also, bots are not the only accounts authorized to make rapid or automated mass changes to Wikidata. At any rate, I hope that we can establish a consensus and we can all take responsibility for our edits insofar as they improve the project according to that consensus. Elizium23 (talk) 16:45, 4 June 2023 (UTC)
@Elizium23: Yes, Cabot is my surname. Thanks for your words. Paucabot (talk) 17:08, 4 June 2023 (UTC)
For the record: The legends Joutbis uploaded with his bot were agreed and discussed with me in this page (in catalan). The bot was approved in Wikidata here. Paucabot (talk) 08:18, 5 June 2023 (UTC)
Vandalism requires intent. You may not like the edits but if they are made with good intent you shouldn't call it that. In this case everyone seems well intentioned and I suggest you be less aggressive. The media legends being removed in general seem very low quality. BrokenSegue (talk) 17:21, 6 June 2023 (UTC)
media legend (P2096) is evil, it is not structured (people put there wikisyntax 🤦‍♂️), captions already live on Commons (although they are not accessible)... So it isn't really bad idea to prune some usage in favor of properties like point in time (P585), depicts (P180), etc. --Matěj Suchánek (talk) 19:08, 4 June 2023 (UTC)
I agree with Matěj Suchánek. Actually I also have been changing media legend (P2096) for depicts (P180) like in these cases: Special:diff/1907176488, Special:diff/1907177678 or Special:diff/1906698235. I think those changes should never be considered vandalism, nor the others we're talking here. Paucabot (talk) 08:10, 5 June 2023 (UTC)

Wikidata weekly summary #579

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Events
    • Introduction workshop about Wikidata for employees of the Koninklijke Bibliotheek (KB, national library of the Netherlands), to make them more familiar and self-reliant with Wikidata - 6 June 2023, Koninklijke Bibliotheek, The Hague, The Netherlands - Slides (in Dutch) on Commons.
    • Next Linked Data for Libraries LD4 Wikidata Affinity Group Working Hour June 5, 2023: In collaboration with the Black Bibliography Project, we will be adding metadata for undescribed or under-described African-American women authors listed in The Pen Is Ours: a Listing of Writings by and about African-American Women before 1910, compiled by Jean Fagan Yellin and Cynthia D. Bond.Event page
    • Registration has opened for the workshop on mathematical research data taking place on 9-11 October 2023 in Leipzig. Talks, demos and lightning talks can be suggested, including Wikidata-related ones.
  • Tool of the week
    • IdentifierInput makes it easier to add identifier properties by letting you put in the full url and then stripping it down to just the identifier for you.
  • Development
    • Wikibase REST API: We finished the new endpoint for patching labels (phab:T332739). We also started work on the new endpoint for retrieving Property data (phab:T337720)
    • Client wikis: We drastically reduced the size of a ResourceLoader module that’s loaded on most Wikipedia page views (phab:334682)
    • Language code: With awesome help from Nikki, we improved documentation for how to add language codes to Wikibase/Wikidata (phab:T335857)
    • EntitySchemas: We are getting the new datatype ready for testing.

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

The curious case of Carl Auböck (or too many Carls)

Carl Auböck (Q117450107) is an Austrian company that was founded by Carl Auböck (Q117447888). This Carl had a son (Q1036805), grandson (Q17310710), and great-grandson (Q117448324) – all of whom were also named Carl Auböck (at this point, I might add that this is not a joke). There is already a Wikidata item for a "Wikimedia set index article" (Q85750661) which references the en:wp article (see above). My question is: How can the set index item be improved so that it includes better cross-linking with each of the individuals as well as the company itself. Cheers, Cl3phact0 (talk) 09:13, 6 June 2023 (UTC)
PS: (See also: Reasonator and Commons)

@Cl3phact0: I usually use has part(s) (P527) to list the the elements of the set index; I suppose one could also use does not have part (P3113) if needed. You should anyway use different from (P1889) to link with the disambiguation page. --Jahl de Vautban (talk) 12:00, 6 June 2023 (UTC)
Thank you, Jahl de Vautban. Perfect! -- Cl3phact0 (talk) 16:48, 6 June 2023 (UTC)

Unregisterd user keeps reverting my chances, no idea how to make it stop

A unregistered user keeps reverting my chances for Q2419294 as they are convinced the institution is part of the government Jhowie Nitnek 18:03, 6 June 2023 (UTC)

have you tried having a conversation with them on their talk page or the item talk page? BrokenSegue (talk) 18:04, 6 June 2023 (UTC)
Bonjour, je lui expliqué sur sa page de discussion la raison pour laquelle je n'accepte pas son changement, mais il refuse obstinément de comprendre. --94.109.120.176 18:06, 6 June 2023 (UTC)
Could you please look at their articles of association before making conclusions, I read it while adding info.
http://www.ejustice.just.fgov.be/cgi_tsv/tsv_rech.pl?language=nl&btw=0406613706&liste=Liste Jhowie Nitnek 18:13, 6 June 2023 (UTC)
They posted something on my talk page and I tried to explain them but they still do it and ignore what I'm saying. Jhowie Nitnek 18:07, 6 June 2023 (UTC)
Comme je vous l'ai signalé plusieurs fois, il ne s'agit pas d'une cinémathèque locale ou régionale, mais de la Cinémathèque fédérale belge, à distinguer de la Cinémathèque de la Communauté française Wallonie-Bruxelles, qui est également, aussi située à Bruxelles.
The revert war seems to be contered on the item description. There does not seem to be much lack of clarity about what the institution is: https://www.cinematek.be/en/about-us states it is a "Public Utility Foundation", which is defined elsewhere [2] as a " a foundation which has as its disinterested/charitable purpose the realization of a work of a philanthropical, ideological, religious, scientific, artistic, pedagogical or cultural nature." https://www.cinematek.be/en/about-us further states it is "a bi-communal institution, which is subsidized by the Federal Ministry of Science Policy and also enjoys permanent support from the National Lottery." Given all of that, amending the description to "cinémathèque de l'État fédéral belge" ("cinematheque of the Belgian Federal State") seems a little misleading. Why not call it what it itself calls itself? --Tagishsimon (talk) 18:26, 6 June 2023 (UTC)
That is what I have trying to explain to them but they don't want to try to understand Jhowie Nitnek 18:32, 6 June 2023 (UTC)
The edit war for the English description is even more incomprehensible to me. “Belgian cinematheque” and “cinematheque in Brussels” seem to both be technically correct. What exactly seems to be the problem? --Emu (talk) 18:33, 6 June 2023 (UTC)
Il y a deux cinémathèque à Bruxelles : la Cinémathèque de la Communauté française Wallonie-Bruxelles et la Cinémathèque royale, qui est belge fédérale. Cette dernière a notamment organisé le réputé Festival du cinéma expérimental à Knokke, en Flandre, diffusé les films classiques en Flandre et en Wallonie, récolte les copies en pellicule et les objets liés au cinéma dans toute la Belgique. Elle est financée par l'État fédéral. Il ne s'agit donc pas simplement d'une cinémathèque bruxelloise, mais c'est le dépôt belge de plus de 70 000 films venant de toute la Belgique entière. En conséquence, « cinémathèque à Bruxelles » ne représente pas l'institution et porte à confondre avec l'autre cinémathèque située aussi à Bruxelles. --94.109.120.176 18:39, 6 June 2023 (UTC)
Maybe we can get to a compromise and call it the "national cinematheque of Belgium"? Jhowie Nitnek 18:42, 6 June 2023 (UTC)
But is that even true? And if so, what does “national” mean in this context? w:en:Cinematek does not mention this epithet at all … --Emu (talk) 18:46, 6 June 2023 (UTC)
"Nationale" signifie belge fédérale car c'est la « Cinémathèque royale de Belgique ». Elle n'est ni régionale, ni communautaire, ni communale, ni provinciale. 94.109.120.176 18:52, 6 June 2023 (UTC)
That the scope is the whole of Belgium and not just one language community (Belgian politics is complicated) Jhowie Nitnek 18:59, 6 June 2023 (UTC)
Okay. I’m not quite sure if using words in descriptions that need to be explained by explaining the current state of Belgium is the best idea :-) --Emu (talk) 19:38, 6 June 2023 (UTC)
We don’t have an item with the label “Cinémathèque de la Communauté française Wallonie-Bruxelles”, do you mean Cinémathèque de la Fédération Wallonie-Bruxelles (Q2973243)? If so, the two items have totally different labels, I find it hard to imagine that there is much room for confusion. --Emu (talk) 18:43, 6 June 2023 (UTC)
Je précise sur la forme que ce n'est pas moi qui modifie : https://www.wikidata.org/w/index.php?title=Q2419294&diff=prev&oldid=1267921963
L'autre cinémathèque à Bruxelles : https://fr.wikipedia.org/wiki/Cin%C3%A9math%C3%A8que_du_minist%C3%A8re_de_la_communaut%C3%A9_fran%C3%A7aise_de_Belgique
Yeah, that’s linked to Wikidata item Q2973243. If you are really concerned about a possible confusion between those two institutions, you might be better off working on that item which can use some work. --Emu (talk) 18:50, 6 June 2023 (UTC)

Incontinent pandas

In an attempt to throw even more data at my pandas to analyze, I ran into an annoying problem, and I was wondering if this is something others have experienced as well.

I'm doing a groupby on two of the columns of a dataframe with 500 000 rows. Looking at the resulting dataframe shows there is now 4.3 million rows, apparently containing index-columns I didn't ask for as well. Changing pandas version from 1.3.3 to 1.5.3 doesn't help. Maybe a data corruption issue? Or am I missing something here? Increasing the number of rows eventually triggers an OOM condition. 

df_title_noagg = pd.concat({'Namespace':df['Namespace'], 'Title':df['Title'], 'Edits': pd.Series(np.ones(df.shape[0], dtype=np.int64))}, axis=1, sort=False)

df_title = df_title_noagg.groupby(['Title', 'Namespace']).count()

df_title_noagg.info(memory_usage='deep')

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 500000 entries, 0 to 499999
Data columns (total 3 columns):
 #   Column     Non-Null Count   Dtype   
---  ------     --------------   -----   
 0   Namespace  500000 non-null  category
 1   Title      500000 non-null  object  
 2   Edits      500000 non-null  int64   
dtypes: category(1), int64(1), object(1)
memory usage: 35.8 MB

df_title.info(memory_usage='deep')

<class 'pandas.core.frame.DataFrame'>
MultiIndex: 4384110 entries, ('109.229.96.232', '(Main)') to ('तुलसीदास', 'Wikidata talk')
Data columns (total 1 columns):
 #   Column  Dtype
---  ------  -----
 0   Edits   int64
dtypes: int64(1)
memory usage: 80.8 MB

By the way, the usernames are from the sanitized version of the database, so there shouldn't be any privacy concerns. Infrastruktur (talk) 20:05, 6 June 2023 (UTC)

It is kinda difficult to guess based on the information you have provided here. Are you working on PAWS, Toolforge, or elsewhere? How much memory do you have? What exactly goes wrong?
I have been running into memory issues when using fairly large pandas DataFrames on PAWS and Toolforge, but 5M rows should still be pretty managable in most environments. However, I have not seen "data corruption" issues with pandas yet. Re. version: it might be an idea to try the recently released v2 version of pandas with the new pyarrow backend (instead of numpy), but I would be surprised if this would actually solve your problems. v1.3 and v1.5 should behave identically as much as I am aware. —MisterSynergy (talk) 20:36, 6 June 2023 (UTC)
Turned out I did miss something. Every page title has an associated namespace. When I asked it to group by namepaces it created 15 times as many buckets as was needed for the operation. (-.-;) Initially I suspected memory issues since pandas often operates on references to the original data, but any merge or concat operation does a deep copy.
It's very nice to have all the raw data loaded into memory. But in this case it was barely able to handle 4M rows running on PAWS. There is no chance it will be able to handle the 18M rows of Wikidata's recentchanges. I will probably have to completely rewrite this particular script to rely on SQL to filter and aggregate all the data that is needed. Jupyter Hub on PAWS lets you use up to 3 GiB of memory, a pure python script might be able to use more, haven't tried yet. Infrastruktur (talk) 06:04, 7 June 2023 (UTC)

Third-party resources policy

At meta:Third-party resources policy you'll find a proposal that will affect quite a few of our gadgets. Please have a look and comment on the talk page. Multichill (talk) 21:33, 7 June 2023 (UTC)

Discussion on connecting Wikidata reference properties to CS1 on Wikipedia

 You are invited to join the discussion at Wikipedia:Module talk:Wd § References mapping. {{u|Sdkb}} talk 20:51, 7 June 2023 (UTC)

Father or Uncle/Mather or Aunt

Suppose that one person can be the uncle/the aunt or the father/the mather of another. It is not possible to determine what the relationship is. How could that be expressed in Wikidata? Thank you. --Fantastoria (talk) 10:22, 6 June 2023 (UTC)

As far as I remember Hawaiian kinship terms don't distinguish between uncle and father. You could create an item "father or uncle" that matches the Hawaiian category and use that with relative (P1038)/kinship to subject (P1039). You however should have a source that's pretty explicit about the relationship being either father or uncle. ChristianKl ❪✉❫ 11:17, 6 June 2023 (UTC)
Yes. Wikidata should model the exact relationship whenever possible and only resort to underspecified statements (“X is the father or the uncle of Y”) if an exact one is really not possible. In particular, there is no point in making an underspecified statement just because there is ambiguity in one language (unless, of course, the only sources available are in that language and there is no way of telling the possible meanings apart). --Data Consolidation Officer (talk) 15:13, 7 June 2023 (UTC)
Ideally, WD has respect for non-Western societal structures. "there is no point in making an underspecified statement just because there is ambiguity in one language" is as offensive as it is crashingly ignorant. w:en:Hawaiian kinship. --Tagishsimon (talk) 15:31, 7 June 2023 (UTC)
@Data_Consolidation_Officer: On behalf of the Wikidata community, I'd like to apologize that you were addressed in this way. I have blocked Tagishsimon for 3 days. Bovlb (talk) 16:53, 7 June 2023 (UTC)
while Tagishsimon was wrong to be so curt/rude I do think they are correct. We should track kinship in terms that are culturally meaningful. Consider that we do not usually distinguish between "grandfather" and "maternal grandfather". This might seem to be a "underspecified statement just because there is ambiguity in [English]". Other languages have more precise terms for this kind of kinship and yet we often use the non-specific English terms. It's even ok to list multiple kinship relations for the same pair of people (if different sources provide different terms for the relationship). BrokenSegue (talk) 17:13, 7 June 2023 (UTC)
@Bovlb: I acknowledge, though, that my wording may have been ambiguous, too, provoking that reaction. Anyway, @BrokenSegue: But do we directly model grandparents, as opposed to length-2 property paths of father (P22) and mother (P25)? The latter variant does distinguish between materal and paternal grandfathers and grandmothers. And I have seen qualifiers being used to distinguish between biological and non-biological (i.e. adoptive, step-)parents. We can of course discuss whether we want different properties for biological and non-biological parents, or whether we want only one property “parent” (with mother/father indicated by qualifiers as well) or maybe even only the property “relative”, but I do not think it would be a good idea to model relationships in an international (!) ontology ambiguously just because the relationship is expressed ambiguously in (some) language(s). More broadly speaking, in my opinion Wikidata modelling should not be based on how relationships are expressed in some language, but on the relationships themselves. That said, I do acknowledge that there may be relationships where it is hard to identify “the relationship itself” (kinship is not a case of that) and I do acknowledge, and this I probably failed to properly convey earlier, that some relationships in the real world may have only ever been expressed in an ambiguous way. This can obviously be the case with kinship, even in English, e.g. if I say “he is a relative of mine”. This has nothing to do with respect for non-Western societal structures. In non-Western societal structures, too, every human has a biological mother and a biological father, regardless of whether that society makes the same conceptual distinctions Western societies do. I have no problem with Wikidata modelling concepts from non-Western societies, and if “father or uncle” is a societally relevant connection (like “godfather” or “legal guardian”, whether or not we want such relationships to be modelled in Wikidata), why not? But if it is just “in language X, we do not distinguish fathers and uncles”, I would not see valid grounds for modelling “father or uncle”. To give a possibly more convincing example, we do not want to have an item for “bank or bench”, even though both are referred to by the same word in German, and we would hopefully think twice before creating an item for something for which, due to ambiguity in German, it is not even clear whether it is a bench or a bank. --Data Consolidation Officer (talk) 20:23, 7 June 2023 (UTC)
@Data_Consolidation_Officer: there are definitely cases where we model things using grandfather (Q9238344) see for example Q76#P1038. In wikidata what matters is what the source says. If the reference says "paternal grandfather" then use paternal grandfather (Q19682162). It's ok to have both that and grandfather (Q9238344) with the more specific one being preferred (and different sources for both). When modeling data I also suggest you consider use-cases. If a particular relationship has meaning to some community then it makes sense to model it for that community.
More broadly speaking, in my opinion Wikidata modelling should not be based on how relationships are expressed in some language, but on the relationships themselves
Wikidata is based on what references say not on what is "true" (we don't have access to what is true) and those references are written in some language. So how the relationships are expressed in language/cultures matters. BrokenSegue (talk) 21:00, 7 June 2023 (UTC)
@BrokenSegue: I’ve always thought that it’s Wikidata’s aim to abstract away from language, providing data about the world in machine-readable form, and now you tell me that Wikidata is actually all about modelling what is expressed in some language? Of course references are written in some language and sometimes the descriptions there are ambiguous, but I’d thought that we are modelling facts, not the way they are expressed. And while there are obviously many areas where the knowledge to be modelled is a culture-dependent one, kinship is certainly none of these, at least not biological kinship (but anything beyond biology would probably open up an entirely new can of worms here…). It’s not a question of culture whose genes you have inherited, and the people you’ve inherited them from are your parents, regardless of what a specific language calls them, whether it distinguished the male and the female one. Their parents are grandparents, their siblings are aunts/uncles (there isn’t a collective word for both in English, but there may be one in other languages) and so on. Obviously sometimes the exact relationship is not known, but that’s something Wikidata has to be able to deal with in any area. It’s not the case that the relationship is inherently unspecified; either A is the biological father of B or not, regardless of whether we know whether this is the case or not. A may be the father of B, but not the uncle, or the uncle, but not the father, or even both the uncle and the father (in case of incest), but it’s not possible for the statements “A is the father of B” and “A is an uncle of B” to be partially true. If we do not know whether they are true or not, then, yes, we may need a means to model this lack of knowledge, which may require something like “father or uncle” (of which both “father” and “uncle” must be subclasses!). But that’s a matter of whether and how we model lack of knowledge, not a matter of culture or of non-Western societal structures.
In all cases, our modelling should use a consistent property inventory. Being able to express kinship with father (P22) and mother (P25) or alternatively with kinship to subject (P1039)-qualified relative (P1038) adds complexity, creates potential redundancy and opens up the possibility of inconsistencies. For example, when writing a SPARQL query to get the siblings of ?someone, I’d have to know that ?someone wdt:P22/^wdt:P22 ?sibling; wdt:P25/^wdt:P25 ?sibling. FILTER(?someone != ?sibling) is insufficient because some siblings may only be reachable via property paths involving p:1038 and pq:P1039. The latter may actually be preferrable because it allows us to model underspecifications (the above “father or uncle” or something like a sex-agnostic “parent”) without the need for a complex hierarchy of specialized properties, but this shouldn’t depend on how the relationship is expressed in some natural language. --Data Consolidation Officer (talk) 08:41, 8 June 2023 (UTC)
It's just not feasible to always express kinship with specific properties like father (P22). If the relationship is complicated we don't want to have to create all the intermediate people (and we may not know all the intermediate people) to establish the lineage through direct relationship. I agree it's more complicated to have both. But unfortunately the ontology on Wikidata already requires you to know complicated things like that and in many cases we choose to be redundant rather than require our users to be more savvy (for example sometimes we qualify place of birth (P19) with a country (P17) even though in theory you should be able to determine the country based on date of birth and walking the graph). Sometimes compromises need to be made to serve our users. BrokenSegue (talk) 00:05, 9 June 2023 (UTC)

Tattoos？

wd doesnt have a property to record whether a person has tattoos and where the tattoos are? 🤔 RZuo (talk) 07:59, 8 June 2023 (UTC) also what the tattoos are.--RZuo (talk) 11:07, 8 June 2023 (UTC)

Maybe there has been no need to model this until now? You may propose a new property. --Data Consolidation Officer (talk) 08:43, 8 June 2023 (UTC)
Many modern-time people has tattoo(s) or other kinds of body modification (up to wearing the earrings), often it's not even a person's 'trademark'. In short, we don't have a property for this or a similar thing. --Wolverène (talk) 10:02, 8 June 2023 (UTC)
i think body modification is certainly a "property" (characteristic) of a person. it's a common metric on model websites for example. wd has medical condition (P1050) which every person would live thru multiple times.--RZuo (talk) 11:07, 8 June 2023 (UTC)
an old discussion which briefly mentioned tattoos: Wikidata:Property_proposal/Archive/17#medical_condition. RZuo (talk) 11:13, 8 June 2023 (UTC)

“refers to”, “means” or similar property?

Infovarius has added the statement Q117560305said to be the same as (P460)male human (Q84048850), for which the software flags a “potential issue” (“male human (Q84048850) should also have the symmetric statement said to be the same as (P460)Q117560305.”) What is probably meant is that the Russian term “чувак” refers to a male human. But is said to be the same as (P460) really the property to use here? To resolve the “potential issue” (given that we generally want to resolve potential issues), one would have to add the symmetric statement to male human (Q84048850), cluttering that item with “said to be the same as” statements for possibly a lot of terms in a lot of languages. What I would like to have here is a property along the lines of “refers to” (“чувак refers to male human (Q84048850)”) or “means” (“чувак means male human (Q84048850)”), but is there any such property? --Data Consolidation Officer (talk) 15:08, 7 June 2023 (UTC)

We have lexemes and senses to store information about what terms refer to what. Items ideally are about concepts and not the names of concepts. While unfortunately, there are some items that are expections, I would not want to have extra properties for those that encourage people to create more of those items. ChristianKl ❪✉❫ 08:21, 9 June 2023 (UTC)

Merge candidate

The Japanese company Tendo Mokko (Q7699648) has a duplicate entity Tendo Mokko, Inc. (Q117115795). As the former contains more information than the latter (and the "Inc." designation may be erroneous as well – the information on the company's own site refers to "TENDO CO.,LTD."), the latter would seem to be the redundant entry. [NB: If I had the technical skills to properly execute merge/redirect operations on WD, I would gladly take care of this myself (I just read Help:Merge and don't feel that I do).] Cheers, Cl3phact0 (talk) 17:02, 8 June 2023 (UTC)

@Cl3phact0, no technical skills are required to run the Merge gadget that you can enable in your preferences. Michgrig (talk) 21:43, 8 June 2023 (UTC)
Thank you, Michgrig, I'll have another look. -- Cl3phact0 (talk) 06:21, 9 June 2023 (UTC)

Dernière relecture avant impression du livre Le mouvement Wikimédia

Bonjour. Un livre au sujet du mouvement Wikimédia sera très prochainement publié au format papier par une maison d'édition. Avant cela, il fut préalablement publié et longuement retravaillé sur Wikilivres à la page Le mouvement Wikimédia. Ceux d'entre vous qui auraient l'envie de relire la version web, avant qu'elle ne soit imprimée sur papier sont les bienvenus. Et il est ensuite possible, comme de coutume, de laisser un message sur la page de discussion de l'ouvrage. En vous remerciant par avance et vous souhaitant une belle fin de journée à tous, Lionel Scheepmans ✉ Contact Désolé pour ma dysorthographie, dyslexie et "dys"traction. 22:16, 8 June 2023 (UTC)

J'ai pris la liberté de corriger vos liens pour qu'ils pointent vers la version française de Wikibooks. Bovlb (talk) 19:02, 9 June 2023 (UTC)
C'est très aimable ! Merci beaucoup @Bovlb. Lionel Scheepmans ✉ Contact Désolé pour ma dysorthographie, dyslexie et "dys"traction. 14:06, 10 June 2023 (UTC)

http://aa.xtraz.net

A series of dates of birth and death were added using this website, but the matching is off, resulting in a bad date import. We now have a list pf people who died before they were born: Wikidata:Database_reports/items_with_P569_greater_than_P570 caused by the addition. Is anyone tracking the bot that is adding the data? I only noticed because of the error flag from people dying before they were born, there will be more errors. RAN (talk) 19:21, 7 June 2023 (UTC)

I believe @Mahir256 might have some insights into this? --Emu (talk) 08:20, 8 June 2023 (UTC)
  • @Nuria2701: The errors are still there, three days later. We need to fix the errors where a person died before they were born, and more importantly detect other improper additions of a death/birth date that was caused by poor matching using http://aa.xtraz.net as a source. --RAN (talk) 01:00, 10 June 2023 (UTC)
I have reposted this to the Administrators' noticeboard. --RAN (talk) 03:59, 11 June 2023 (UTC)

Throttle item creation by anonymous users

User:Bovlb/Throttle item creation by anonymous users

I have a policy proposal to restrict item creation by anonymous users. I plan to post this as an RFC in the near future, so I am not looking for !votes at this time, just feedback on how to improve the proposal, or whether I'm even on the right track. Bovlb (talk) 01:54, 11 June 2023 (UTC)

Might be able to provide a better basis for setting the rate-limit. Last 30 day period there were 6019 item creations by 2334 users. These numbers count all item creations by IPs in addition to unpatrolled item creations by registered users. The breakdown is as follows (the count here is the number of affected users. press 'expand' to see table):
Creations Count
1 1702
2 261
3 112
4 61
5 41
6 23
7 19
8 14
9 11
10 12
11 5
12 5
13 6
14 4
15 4
16 5
18 5
19 1
20 1
21 1
22 2
23 3
24 1
25 2
28 3
29 2
30 1
31 1
32 2
33 1
35 1
37 2
38 3
39 1
40 1
42 1
45 1
48 1
49 1
50 1
52 1
54 2
56 1
75 1
84 1
91 1
93 1
95 1
192 1
As you can see there is a sharp drop-off in the amount of users who create more than 10 new items per month and another drop-off after 18 item creations. Setting the limit at 18 item creations will affect 44 users and 2290 users will be unaffected. What the filter period should be will affect how bursty the new item creation is allowed to be, so IMO a one month period is too much and a one week period is too inflexible, Setting the period to 2 weeks will allow creating 9 items in one go, but they will then have to wait another 2 weeks to create 9 more items. Infrastruktur (talk) 08:09, 11 June 2023 (UTC)
I would consider extending it to item edits as well, but with a higher threshold than for creations. Vicarage (talk) 09:25, 11 June 2023 (UTC)
I'm fine with throtteling for anonymous users. I think it's important that there are clear messages that tell the user that they can create more items if they sign up for an account.
As far as numbers go five new items within a seven day window seems reasonble to me. ChristianKl ❪✉❫ 14:06, 11 June 2023 (UTC)

La Ilustración de la Mujer

Bonjour, Je cherche à fusionner deux éléments wikidata qui traitent exactement du même sujet et qui ont des pages wikipedia: La Ilustración de la Mujer (Q17423135) et La Ilustración de la Mujer (Q18620725). Merci de votre aide! Cahtls (talk) 19:38, 10 June 2023 (UTC)

On a dis que l'une etait publie a Madrid, et l'autre, a Barcelona. On a deux categories differentes a Commons. Ymblanter (talk) 18:56, 11 June 2023 (UTC)

Many Considerations on France

There are currently Considerations sur la France (Q42188185) (en-wiki and en-quote), Considerations on France (Q19148997) (fr-source) and Considérations sur la France (Q115865002) (archive.org full text). They are currently linked with edition or translation of (P629) but should they be merged? MKFI (talk) 11:44, 12 June 2023 (UTC)

No, they should not. Considerations sur la France (Q42188185) covers the concept of the work itself. They other two items seem to cover the 1814 and the 1821 print edition, respectively. --Emu (talk) 11:59, 12 June 2023 (UTC)

Wikidata weekly summary #580

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Events
    • Next Linked Data for Libraries LD4 Wikidata Affinity Group Call June 13, 2023: Helen Williams will discuss her work with Wikidata at the London School of Economics Agenda
    • Wikidata Workshop 2023 will be co-located with the 22nd International Conference on Semantic Web (ISWC) in November 2023 in Athens. The submission deadline is Thursday, 20 July 2023. More details and call for papers
  • Tool of the week
    • narrow.css attempts to make the Wikidata desktop interface work better on narrower screens.

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

Linking Commons cat & Wikipedia article

I am not sure what exactly should be the relation of Category:Canadian National No 6 (tugboat, 1948) (Q116550802), Canadian National Tug no. 6 (Q17986243), and Canadian National Tug no. 6 (Q116550789), but surely they should be stitched together in a way that connects Commons:Category:Canadian National No 6 (tugboat, 1948) to en:Canadian National Tug no. 6 (and fr:Canadian National Tug no. 6), which is not currently the case. Could someone who has more idea than I how this is normally done please fix this? Thanks. (Ping me if you need me, I don't maintain a watchlist on Wikidata.) - Jmabel (talk) 03:33, 10 June 2023 (UTC)

Moved from the talk page--Ymblanter (talk) 19:00, 11 June 2023 (UTC)
https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Ships#Same_ship_by_different_names and https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Ships#Are_Wikidata_items_for_ships-by-different-names_to_be_merged_into_one_Wikidata_ship-by-name_item? discuss the problem of hulls with unique IMOs but different names. As this ship only ever had one name, and the IMO Wikimedia category has no content other than the ship name, its a good candidate for merging, and I've done so and tidied up. Vicarage (talk) 19:59, 11 June 2023 (UTC)
@Jmabel: Ymblanter (talk) 16:20, 12 June 2023 (UTC)

New EntitySchema data type is ready for testing on Test Wikidata

Hello,

As many of you may already know, we have been working on introducing a new Wikidata data type that will make it easier to find EntitySchemas and use them to connect to other Wikibase Entities. This will allow editors to refer to existing EntitySchemas in statements to indicate what class of Items, Lexemes etc. are governed by an EntitySchema. This new EntitySchema datatype is now live on Test Wikidata for testing and your feedback.

This is an important step to more consistent and reusable data. It’s now live on Test Wikidata for more information, or to give give feedback see Wikidata talk:Schemas

Cheers, Arian Bozorg (WMDE) (talk) 15:50, 13 June 2023 (UTC)

Thank you! I’m happy to finally see some progress related to schemas. —MisterSynergy (talk) 20:59, 13 June 2023 (UTC)

Odd presentation of "different from" data

In the "Wikidata Infobox" at https://commons.wikimedia.org/wiki/Category:Godshill,_Isle_of_Wight, it says "different from Godshill", which reads oddly. I have noticed in other places that this "different from" field produces odd-looking results in Wikimedia Commons. What is needed for place names is a list "Other places called ~", or something of that nature, where all the places are properly disambiguated, not just the base name repeated, or even repeated multiple times, e.g. as at https://commons.wikimedia.org/wiki/Category:Newport, which is said to be "different from Newport" and "different from Newport"! I understand that the links go to different places, but still it looks weird. How can this be improved?

18:09, 13 June 2023 (UTC) ITookSomePhotos (talk) 18:09, 13 June 2023 (UTC)

A hard problem as inevitably if 2 things share the same label, Newport will always say "see also Newport". One solution is to use the WD descriptions (which we require to be unique when taken in combination with the label), so Newport, Wales will say "see also Newport (town on the Isle of Wight)". This is rather long for an infobox though. I'd not go for an explicit solution for places, as you'd need a separate one for people, films, projects etc. Vicarage (talk) 18:22, 13 June 2023 (UTC)
Right, generally speaking it needs to be a list along the lines of "other Xs called Y", or however it can be made most generic. Though the disambiguated names inevitably would be longer than the base names, at least the presentation would make sense, which is not the case right now. ITookSomePhotos (talk) 18:43, 13 June 2023 (UTC)
And in fact at e.g. https://commons.wikimedia.org/wiki/Category:Newport it doesn't seem necessary to include this "other places" list at all in the "wikidata" box, since the whole point of the page itself is to list all the places of that name. But am I right in thinking that the behaviour of the "Wikidata" box at Wikimedia Commons is controlled at Wikimedia Commons, so it could pick up any data available at "Wikidata" and present it any way it chooses? I don't think I totally appreciated that earlier, so perhaps these issues would be better raised at Wikimedia Commons, so that anyone interested can make the sensible changes to how "Wikidata Infobox" works? Is that right? Or would it entail changes at Wikidata too? ITookSomePhotos (talk) 19:33, 13 June 2023 (UTC)
Its the Commons people who choose how to use our data. We do need to consult if there isn't a clear one-to-one mapping of concepts, but that's not the case here. Vicarage (talk) 19:43, 13 June 2023 (UTC)
I see, thanks. ITookSomePhotos (talk) 21:08, 13 June 2023 (UTC)

Updated picture not showing up in Wiki

old image
new image

Hi,

I changed the image at https://www.wikidata.org/wiki/Q2034581 to a brighter version, as shown on the right, but this is not being picked up in the "Wikidata Infobox" at https://commons.wikimedia.org/wiki/Category:Godshill,_Isle_of_Wight, despite the system having had ample time. What do I need to do to make the new image display?

ITookSomePhotos (talk) 18:06, 13 June 2023 (UTC)

I purged the page by going to https://commons.wikimedia.org/w/index.php?title=Category:Godshill,_Isle_of_Wight&action=purge . That purged the page cache and the template rebuild itself. --William Graham (talk) 18:31, 13 June 2023 (UTC)
Thanks, it's showing correctly for me now. Is a separate "purge" action supposed to be required, or was it just an isolated glitch that caused the update not to happen automatically, do you know?
After a certain amount of time the cache will expire and the page will rebuild. Off the top of my head I don't know the cache expiration time for Commons. I believe it's no longer than 24 hours. Probably best to ask over at Commons for the exact number. --William Graham (talk) 18:50, 13 June 2023 (UTC)
Thanks, I left it at least 36 hours. Anyway, I guess either it would have happened eventually or it was an isolated glitch. Thanks for your help with this. ITookSomePhotos (talk) 19:24, 13 June 2023 (UTC)
Usually changes like this are dispatched to other wikis fairly quickly (within minutes, I believe), but unfortunately Commons is excluded from this mechanism due to its size (T188730 and/or T179010, not sure which, but more likely the former). Due to this, without the purge I think the change would have only shown up on Commons on the first visit after 21 days ($wgParserCacheExpireTime). Lucas Werkmeister (WMDE) (talk) 15:17, 14 June 2023 (UTC)
Could you explain what you mean by "on the first visit"? ITookSomePhotos (talk) 21:43, 14 June 2023 (UTC)
Oh, sorry, you mean for 21 days the change would not show (for anyone), then any time anyone visits after 21 days it would show for them, is that right? ITookSomePhotos (talk) 21:46, 14 June 2023 (UTC)

Amsterdam Q9899

Can someone please add logo Logo of Gemeente Amsterdam.svg? Page is semi-protected. 147.161.249.77 06:50, 14 June 2023 (UTC)

✓ Done FYI for future reference, if you sign up, you can add it yourself. RVA2869 (talk) 07:42, 14 June 2023 (UTC)

order (Q193622)

Hi, This issue is kind of related to a topic started earlier today.(https://www.wikidata.org/wiki/Wikidata:Project_chat#Can_we_PLEASE_put_some_limits_on_the_use_of_English_Wikipedia_as_a_source_of_sen_BLPs?)

An order (Q193622) is not an award in itself. They are usually divided into different ranks. (grade of an order (Q60754876)) Most statements are added from Russian Wikipedia (Q206855), I guess with Infobox export gadget. Any suggestions because this is getting ridiculous... example: https://www.wikidata.org/wiki/Q151301#P166 RVA2869 (talk) 19:38, 14 June 2023 (UTC)

Kentucky Colonel

At Kentucky Colonel (Q632482) are all the instances_of correct? RAN (talk) 21:23, 14 June 2023 (UTC)

Most of "subclass of" look incorrect and probably can be removed. For the instances of, I think "job title", "personal jurisdiction", "natural person", "people", "position", and "commission" can be removed. Although some or all of those might be moved to subclass of. -- William Graham (talk) 22:59, 14 June 2023 (UTC)
To be fair, their own website suggests that commission (Q60918) is appropriate here (or is at least the closest thing I could find). — Huntster (t @ c) 23:19, 14 June 2023 (UTC)
I would argue that at least natural person (Q154954) and people (Q2472587) are inappropriate, because the item is about the award, not the awardees. "Kentucky Colonel" is not in and of itself a person. personal jurisdiction (Q7170598) doesn't make any sense. occupation (Q12737077) seems somewhat redundant to position (Q4164871). It feels like a lot of these could be trimmed down. — Huntster (t @ c) 23:16, 14 June 2023 (UTC)
I took a look at the linked items and it seems that one or more booster organizations, membership organizations that admit people of this title, have packed this item and related items with their point of view, bordering on spam. Based on this page on one organization's website https://www.kycolonelcy.us/kentucky-colonels-on-wikipedia it looks like they feel spurned by deletions on Wikipedia and this may be a new outlet. -- William Graham (talk) 23:28, 14 June 2023 (UTC)

Tool to report and notify

Is there any tools/script that make it easy to report and notify users? Kacamata (talk) 23:59, 13 June 2023 (UTC)

To notify a user (make a user get an alert): {{ping|Username}} or [[User:Username|Username]], or leave a message on his/hers Talk page. What do you mean by "report"? JopkeB (talk) 09:48, 14 June 2023 (UTC)
@JopkeB: I believe what Kacamata wants is something like WP:Twinkle, to make it easier to add warning templates to user talk pages and to create vandalism reports at the admin noticeboard. AFAIK there is no such tool for reports, but there is a user script for warnings at User:Bene*/userwarn.js (it adds a "Warn this user" link to the Tools menu, which allows you to choose the warning template from a drop-down menu). –FlyingAce✈hello 12:57, 14 June 2023 (UTC)
@FlyingAce That what I wanted. Thanks, I'll try this script. Also, thanks @JopkeB for responding. Kacamata (talk) 20:47, 14 June 2023 (UTC)
You may also try User:Ahmad252/scripts/UserWarning.js or my fork. Jklamo (talk) 08:08, 15 June 2023 (UTC)

Backwards number

When @OBCPO1 added the ISNI (P213) number to Andrew Agwunobi (Q106910762) it came out backwards and right-aligned (see previous version) The only way I could fix was to add the same value again and then remove the backwards one. Has anyone seen this before? — Martin (MSGJ · talk) 15:28, 14 June 2023 (UTC)

Apparently the value ended with a right-to-left mark (U+200F, Q1017375), which was enough to cause the whole thing to be displayed right-to-left. Lucas Werkmeister (WMDE) (talk) 15:48, 14 June 2023 (UTC)
Thanks - seems you are right. Just fixed another one of these [https://www.wikidata.org/w/index.php?title=Q7183484&diff=prev&oldid=1915647872 here] — Martin (MSGJ · talk) 07:47, 15 June 2023 (UTC)

Can we PLEASE put some limits on the use of English Wikipedia as a source of sen BLPs?

It completely defeats the purpose of BLP if people just lists English Wikipedia as the source of controversial and privacy-violating statements

I'm inclined to remove these type of unsourced statements on BLPs because finding proper citations is really tedious. Especially when the statements are blindly spammed with little regard for modeling

/end of rant--Trade (talk) 23:02, 12 June 2023 (UTC)

Citing a Wiki article defeats the purpose of citations. I don't understand why anyone does it other than being an easy cop-out to beat the "requires citation" warnings. Banning such data dumps would be a net positive. — Huntster (t @ c) 23:14, 12 June 2023 (UTC)
@Huntster: Because this kind of referencing is promoted by Wikidata through their little helpers. --Gymnicus (talk) 08:52, 13 June 2023 (UTC)
When I find an item that cites a Wikipedia in references but has no history of a corresponding sitelink, I usually end up deleting it, citing "fake references" as one of the reasons.
I had assumed that someone somewhere was advising spammers to add such references. @Gymnicus What helper tool do you think encourages such references? Bovlb (talk) 16:53, 13 June 2023 (UTC)
Wwwyzzerdd for Wikidata Trade (talk) 22:27, 13 June 2023 (UTC)
OK. So far as I can see, that would only work for items that did have a sitelink to the relevant project. I was talking about a completely different scenario, where there are claims referenced to Wikipedia but no sign of a sitelink. Bovlb (talk) 00:24, 14 June 2023 (UTC)
One of the problems is that far too often these references to Wikipedia lead to articles where the data being referred to have no actual references. We cannot know where the data came from, or even worse if it is some kind of circular reference where the Wikipedia article writers found the information on WikiData (unreferenced of course), added it to Wikipedia, and then someone comes along and states the WD data is a reference to WP. It's simply bad all around to allow WP being stated as a source on WD. — Huntster (t @ c) 00:35, 14 June 2023 (UTC)
Right. The imported from Wikimedia project (P143) reference is really intended just to record why the claim was added, not to justify the claim. Bovlb (talk) 00:38, 14 June 2023 (UTC)
If we want our data to actually have value and integrity, how about a filter that stops the action and warns the user to use a third party source rather than a Wikipedia article? — Huntster (t @ c) 14:01, 14 June 2023 (UTC)
we don't even trigger warnings for no sources on most statements much less "good sources". BrokenSegue (talk) 17:15, 14 June 2023 (UTC)
Not all statements are equally contentious. Trade (talk) 20:54, 15 June 2023 (UTC)
We currently have https://www.wikidata.org/wiki/Wikidata:Living_people . Do you feel that the existing policy isn't sufficiently enforced? Or do you feel it should change? If you want it to change, how would you want it to change? ChristianKl ❪✉❫ 15:07, 14 June 2023 (UTC)
For starter, by not having P813, P143 and P4656 count towards the citation-needed constraint (Q54554025) constraint--Trade (talk) 15:23, 14 June 2023 (UTC)
I'd be ok with this. I don't know if it's feasible though. BrokenSegue (talk) 17:39, 14 June 2023 (UTC)
phab:T195052 apparently anticipated that it "Would start with just binary in the first version and maybe do a second version that looks for real references (and not imported from)." Bovlb (talk) 17:42, 14 June 2023 (UTC)

c:Category:Uses of Wikidata Infobox with no image - can dates of months be excluded?

c:Category:Uses of Wikidata Infobox with no image has now more than 800.000 subcategories and it is growing. To reduce that number and because I think it is good to have images in Wikidata items (a picture may say a thousands words, especially when you are not familiar with the Latin script), I am adding images to Wikidata items, and set image (P18) of Wikidata items with no appropriate image at no value (for instance of people who died before ±1900 without a portrait).
There are also Wikidata items of dates of months, like c:Category:1 April and c:Category:16 August. There are images in the corresponding Commons categories, but they are not representative for all these dates between ± 2.000 years ago and now (perhaps I might find images for 21 June and 21 December, but that's it). So I think it is rather nonsensical to have images in Wikidata infoboxes of these categories. And therefor the corresponding Commons categories might as well disappear from c:Category:Uses of Wikidata Infobox with no image.
My question is: Do you know a solution for these kinds of abstract ideas? Can they for instance also have image (P18) set at no value? Or is there a better solution? JopkeB (talk) 08:18, 15 June 2023 (UTC)

I think these questions are better addressed downstream by the authors of the template on commons and Commons itself. I am not a heavy commons user, but I don't see a reason to feet over a hidden maintenance category. Not all maintenance categories require action, likely including this one.
Regarding Wikidata, I do not think it necessary or even appropriate to create statements of image (P18): no value.in most circumstances. William Graham (talk) 13:00, 15 June 2023 (UTC)
There is a problem with having a mechanism for WD contributors working through a search space and wanting to record their null findings. But I don't think that's appropriate for image (P18) and people, as images can always be discovered for anyone of any era, and their absence is not notable. I do wonder why we are recording in WD the results of a Commons infobox design. Surely the uses WD is put to are independent of its structure, and anyone wanting to list items without images would be better to do a SPARQL query. Vicarage (talk) 13:51, 15 June 2023 (UTC)
May 20 - calendar sheet
I found the solution myself. We can use icons of calender sheets for each day. It is a lot of work, but it is not nonsensical to have them in Wikidata items for dates of months. They can be found in c:Category:Calendar icons (all icons of dates there now have been better categorized and used in a Wikidata item).
For over 300 dates there is not yet an icon. Who wants to help, can do so! You may choose yourself the color, layout, language (only Latin script and Arabic numerals) and extension. Take care that in Commons you give credit to the maker. Choose as Commons categories: (1) the date (a grandchild of c:Category:Days by month) and (2) c:Category:Calendar icons or a subcategory.
I withdraw this question.
Thanks for your answers and thoughts. JopkeB (talk) 14:35, 15 June 2023 (UTC)
This looks a bit like forum shopping because it was already brought up at c:Template_talk:Wikidata_Infobox#Category:Uses_of_Wikidata_Infobox_with_no_image_-_please_exclude_dates_of_months. Multichill (talk) 20:42, 15 June 2023 (UTC)
Your assumption of bad faith is once again misplaced. It was suggested in that discussion to ask here, and the OP already notified that discussion of this one. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:58, 15 June 2023 (UTC)

Redirection

Hiː

This item Q28060193 (Governmental weather service) used in a French Infobox redirect to Météorologie (e.g. Palestinian Meteorological Department) while there is no redirection in English. Can anyone explain how the redirection is done? Pierre cb (talk) 03:30, 15 June 2023 (UTC)

Q87766926#P31 links to Q28060193, and the link in the infobox goes to the item in Q28060193#P101; I don't know exactly how these templates and modules work, but lines 98-103 of fr:Module:Infobox/Organisation contain code related to "Type" in the infobox, and "defaultquery" in line 101 refers to line 60 of the same module, which specifies the properties to use when making the link ('P101', 'P366', 'P361', 'P279'). Peter James (talk) 14:25, 15 June 2023 (UTC)
Thanks Pierre cb (talk) 12:16, 16 June 2023 (UTC)

Any interest in collections of civic data?

I’m a lawyer at the intersection of law and technology, and a number of projects related to legal and political data have been on my mind for some time. In my pursuits, I’ve developed, come across, and contributed to a number of data projects in the space. A major problem is the lack of a central home for these data projects, and issues with availability of one database or another as funding runs out or maintenance devolves. Do you all think there is a home for that kind of data here?

Examples of data:

• A database of court case data, including case events, parties, judges, locations, jurisdictions, and more.

• A database of laws

• A database of court precedents

• A database of representative’s votes on particular bills, over the lifetime of a bill

This is something I would pitch to each of the relevant parties, but I wanted to first check that it’s an appropriate endeavor.

Some use-cases I imagine include fact checking claims about jurisprudence, fact-checking claims on political stances, and a repository of accurate statements of the state of the law. The benefit to the legal tech research community would be a centralized home for the analysis of data, and a simplified way to engage others in development of this data with reduced barriers to entry. Currently, the cost of accessing federal court case “metadata” is $225,000 per district court, per year of court activity. There are 94 district courts in the US, and that figure does not include the text of actual filings.

There is a dearth of knowledge about the courts and the law which results in or, at a minimum, contributes to access to justice issues.

Let me know if there is an interest in housing such data here, or if there would be similar kinds of data that would be helpful for the community.

Thanks everyone! Shalalalaw (talk) 13:37, 16 June 2023 (UTC)

The projects you theorize seem very expansive and very loosely defined. There are an extraordinarily large number of court cases, and laws across all countries and jurisdictions worldwide. I think starting with a much smaller more manageable project would be a good starting point. Also, the idea about tracking bills and votes seems to be met by GovTrack (Q5588553) for the United States congress.
Some concepts and resources that may be helpful:
Finally, not to discourage you, but some new users see Wikidata as merely a free data host to house their personal or organization project data. Consider if your idea may be better placed in a relational database system that you (or someone you employ/contract to) host and manage.
-- William Graham (talk) 14:56, 16 June 2023 (UTC)
Also, I have found the Legisworks Historical Statutes at Large Data dataset to be useful as a resource. -- William Graham (talk) 15:01, 16 June 2023 (UTC)

WikidataCon 2023: Call for Proposals

Hi everyone,

Later this year WikidataCon is happening in Taipei and online. The call for proposals is now open and I'd love to see many of you submit talks about cool stuff you're working on as well as important topics we should discuss as a community. You can submit proposals at https://pretalx.com/wikidatacon2023/. If you need any help with your submission please feel free to reach out.

Cheers Lydia Pintscher (WMDE) (talk) 18:32, 17 June 2023 (UTC)

Help with SPARQL query

Hi, I'm having problems trying to get a SPARQL query to get entities that are instance of (P31) of Wikimedia category (Q4167836) that have a sitelink with eswiki but no spanish label that doesn't time out. This is what I've got, but I don't know how to edit it to avoid the timeout: 

SELECT ?i ?article ?label_es WHERE {
  ?i wdt:P31 wd:Q4167836;
    ^schema:about ?article.
  ?article schema:isPartOf <https://es.wikipedia.org/>.
  FILTER(NOT EXISTS {
    ?i rdfs:label ?label_es.
    FILTER((LANG(?label_es)) = "es")
  })
}
Try it!

Agabi10 (talk) 13:42, 17 June 2023 (UTC)

You can also use the Query Builder to create your queries without having to do SPARQL editing. CrystalLemonade (talk) 15:54, 17 June 2023 (UTC)

Hello @Agabi10: I have found the following similar recipies in my FAQ:

In your case, it might be something like: 

SELECT ?item
WHERE
{
  ?sitelink schema:about ?item .
  ?sitelink schema:isPartOf <https://es.wikipedia.org/> .
  SERVICE wikibase:mwapi
  {
    bd:serviceParam wikibase:api "Search" .
    bd:serviceParam wikibase:endpoint "www.wikidata.org" .
    bd:serviceParam mwapi:srsearch "haswbstatement:P31=Q4167836 -haslabel:es" .
    ?item wikibase:apiOutputItem mwapi:title.
  }
}
Try it!

https://w.wiki/6qmT

Instead of haslabel is would also possible to use hasdescription for example. --M2k~dewiki (talk) 16:05, 17 June 2023 (UTC)

As far as I know, there is a bot, which is copying sitelinks to labels, if there is no label yet, so the list usually might be quite short for any language at least for descriptions (maybe not for descriptions, due to multiple identical labels + additional text in brackets for the sitelink, where then a identical combination of label + description ist not possible, since the combination has to be unique) M2k~dewiki (talk) 16:10, 17 June 2023 (UTC)
@M2k~dewiki: the query doesn't seem to be working. Most of these entities have eswiki sitelink, they still don't have a spanish label and they're not appearing in the results. Agabi10 (talk) 16:34, 17 June 2023 (UTC)
@Agabi10: maybe this is due to common problems with the text search. For the last few hours I get the error message: Bei der Suche ist ein Fehler aufgetreten: Deine Suche konnte aufgrund eines vorübergehenden Problems nicht abgeschlossen werden. Bitte später erneut versuchen. when using the full text search:
M2k~dewiki (talk) 16:37, 17 June 2023 (UTC)
General Search is now working again, the query above still does not find these items.
According to
missing labels are copied by User:MatSuBot from sitelink, bot operator is User:Matěj Suchánek. I would assume, that the queries are split into smaller parts by the bot to avoid the timeout. M2k~dewiki (talk) 11:11, 18 June 2023 (UTC)
@Agabi10: a similar discussion can be found at
M2k~dewiki (talk) 13:11, 18 June 2023 (UTC)
From discussions like these, a general approach to avoid timeouts is to split the result set into smaller pieces by adding additional filter criteria. M2k~dewiki (talk) 13:31, 18 June 2023 (UTC)

Search broken

Search appears to be broken. Abductive (talk) 18:29, 17 June 2023 (UTC)

For me too. A Phabricator task has been opened at phab:T339811. Tol (talk | contribs) @ 01:53, 18 June 2023 (UTC)
Same here. Vojtěch Dostál (talk) 06:12, 18 June 2023 (UTC)
Since yesterday...--Kresspahl (talk) 07:44, 18 June 2023 (UTC)
General Search is now working again
M2k~dewiki (talk) 11:12, 18 June 2023 (UTC)

Change the languages on the login page

I think it would be better to not just use Western European languages (and Esperanto) on MediaWiki:Loginlanguagelinks. So I am proposing that this be changed to the language set that is used at simple.wikipedia.org: Arabic, German, English, Spanish, French, Hindi, Indonesian, Italian, Dutch, Portuguese, Russian, Chinese. Let's start this discussion. Lights and freedom (talk) 16:50, 16 June 2023 (UTC)

I'm unfamiliar with what that mediawiki page controls? Can you explain? I'm confused why Esperanto would be on any such list. BrokenSegue (talk) 01:07, 17 June 2023 (UTC)
@BrokenSegue When you login, the instructions ("Username", "Password", "Help with logging in", "Forgot your password", "Keep me logged in", etc.) are available in several languages which can be customized by that page. I don't know why Esperanto is there - I assume one of the early developers was into international auxiliary languages. Meanwhile, Wikimedia Commons' login is in 37 languages, including Galician, Low German, and Lojban, because admins keep adding new languages whenever someone asks. Lights and freedom (talk) 01:48, 17 June 2023 (UTC)
yeah your list seems good. i'd suggest a data driven approach but it probably doesn't matter BrokenSegue (talk) 03:48, 17 June 2023 (UTC)
@BrokenSegue fwiw, while I don't think we can count what interface language users have in their settings, the category counts at category:Babel - Users by language are probably a decent approximation - there are six over 1k (ru, en, de, es, fr, it), three 600+ (ja, plus zh/pt if combined), two 500+ (nl, la!), two 400+ (hi, ar). So other than the appearance of Latin and the omission of Indonesian (315x; ml and pl also in the 300+ range), pretty close to the suggested list. Andrew Gray (talk) 11:48, 17 June 2023 (UTC)
When looking at the list, it seems to me like Japanese should belong there. It's the country with the second most Wikipedia pageviews and there are more speakers in it than Dutch and Italian. Japanese also has four times the number of active editors as Dutch on Wikipedia. Persian also seems to me like a language that might reasonably belong. ChristianKl ❪✉❫ 11:37, 17 June 2023 (UTC)
I agree with adding Japanese, and maybe removing Dutch as it probably doesn't bring much added reach. The other languages, I'm not sure about. Lights and freedom (talk) 17:31, 17 June 2023 (UTC)
I would add that this is particularly important for Wikidata, as it's not supposed to be restricted by language. Lights and freedom (talk) 21:58, 17 June 2023 (UTC)
I added the list to the one used by simple.wikipedia and added Japanese. I think most uses of Latin in Babel are for people using it for Latin names for species, anatomical concepts and other scientific concepts where the user would not use it as their main display language. ChristianKl ❪✉❫ 22:57, 19 June 2023 (UTC)

Label and title

Hi y'all,

Could someone have a look at list of The Young and the Restless characters (Q3375960) and tell who is right between Quesotiotyo and me? For me, this item is about a list (per instance of (P31) and is a list of (P360)) but Quesotiotyo thinks this "represents an article" and even add the label as title (P1476).

Cheers, VIGNERON (talk) 07:09, 19 June 2023 (UTC)

It should be lower case per the thousands of other list items. English Wikipedia would not capitalize list if it were technically possible since it's not a proper noun. I don't actually know why Template:Lowercase title (Q4282320) isn't used on lists, but it shouldn't matter for Wikidata's purposes. It's definitely not a title. —Xezbeth (talk) 08:16, 19 June 2023 (UTC)
The only reason that we have these items is because they exist on Wikipedia and other Wikimedia sites, so yes, it does matter for our purposes how those sites choose to format the article titles. Wikidata should fully match the corresponding sitelinks on these Wikimedia internal items. That is why we also include parentheticals for lists where they would otherwise be omitted from our labels (such as list of English writers (A-C) (Q42326746)). It is the same way that we treat other item types such as disambiguation pages, categories, and templates.
--Quesotiotyo (talk) 18:05, 19 June 2023 (UTC)
Wikimedia list items are not about a particular article on a particular Wikipedia but they are generally about the concept of the list that's then represented by the article of individual Wikipedia. As such I don't think title (P1476) is appropriate here and the item name should not be capitalized. ChristianKl ❪✉❫ 12:59, 19 June 2023 (UTC)
That is a good point, that these kinds of items are not about any one particular sitelink, but should encompass any and all of them. Title (P1476) doesn't quite work that way when there are multiple sitelinks and so I will remove it (plus the sitelink titles are already stored along with the item, which I had forgotten about).
Thanks,
--Quesotiotyo (talk) 18:30, 19 June 2023 (UTC)
correct BrokenSegue (talk) 18:46, 19 June 2023 (UTC)

Wikidata weekly summary #581

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Discussions
    • Open request for adminship: Vargenau (RfP scheduled to end after 25 June 2023 07:51 UTC)
    • Closed request for adminship: Infrastruktur (closed on May 12th without prior mention)
    • Closed request for permissions/Bot:
  • Tool of the week
    • Mapping Diversity is a platform that utilizes Wikidata to discover key facts about diversity and representation in street names across Europe. E.g. 59% of the streets and squares in Paris are named after people. Of those, 8.9% are named after women.
  • Development
    • REST API:
      • We finished the endpoint for getting data from Properties and it is now available on Wikidata.
      • We started expanding that endpoint with filter options and conditional http headers.
    • EntitySchemas: The new datatype is now available for testing and feedback.
    • mul language code: We've continued working on making the mobile termbox (labels, descriptions, aliases) ready for the new language code. (phab:T329644, phab:T338302, phab:T329655, phab:T316767)
    • Vector 2022 skin: We are fixing an issue with the bolding of results in the main search suggester. (phab:T327510)


You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.


MediaWiki:Wikibase-SortedProperties underwatched?

Requested edits going as far back as a year have gone unresolved. The order in which statements are displayed is an important aspect of user experience. How can we get more admin responsiveness here? Swpb (talk) 14:56, 14 June 2023 (UTC)

what about just not protecting it? is it really a magnet for vandalism? BrokenSegue (talk) 19:16, 14 June 2023 (UTC)
The protection message says the page "forms part of the MediaWiki interface", which suggests that breaking it (e.g. blanking it) would break a lot, e.g., basic membership statements wouldn't be at the tops of items, like properties wouldn't appear together, etc. So protection is probably justified. Swpb (talk) 19:40, 14 June 2023 (UTC)
It's in MediaWiki namespace, so only admins and interface-admins (and wmde-staffers) can edit it; there is no custom protection configured for that page. The issue is that it is fundamentally a flawed approach to use this page for sorting properties, and each requested change may upset other editors. —MisterSynergy (talk) 19:42, 14 June 2023 (UTC)
Correct me if I'm wrong, but isn't it currently the only available approach for sorting properties? Swpb (talk) 19:55, 14 June 2023 (UTC)
@MisterSynergy: ? Swpb (talk) 14:50, 15 June 2023 (UTC)
No, there is nothing else really available. —MisterSynergy (talk) 15:23, 15 June 2023 (UTC)
If the person who used to update that list got tired of the bike-shedding I'm not surprised. It should be possible for confirmed users to edit the order, possibly following a discussion. We could designate a new property for property sort order, for simplicity's sake I suggest a plain 1-5 scale to determine ordering of identifiers and non-identifier statements. The list can then simply be generated at regular intervals and uploaded with an admin-bot. The tool could disregard any edits done the last hour (?) or so to give time for reverting any vandalism. Infrastruktur (talk) 16:16, 15 June 2023 (UTC)
I don't think it's possible for regular users to edit the page without a software change. A 1-5 property would offer a lot less precision than the interface page. This page is the MediaWiki solution for every wiki that uses structured data; why re-invent the wheel but make it pentagonal? If one person found the task too tedious, that's fine, but there are lots of admins; someone should be able to adopt such an important page. Swpb (talk) 15:13, 16 June 2023 (UTC)
"A 1-5 property would offer a lot less precision than the interface page." This is a good thing as it would prevent pointless minor changes to the ordering. That was the idea behind suggesting an intentionally constrained range. It was just an example. If you pick 4 for normal rank and 5 for unimportant rank that would leave 1-3 ranks for very important, important and somewhat important ranking, IMO plenty. Infrastruktur (talk) 05:32, 17 June 2023 (UTC)
Property sorting is about much more than relative importance. Swpb (talk) 19:34, 20 June 2023 (UTC)

Label, description, sitelinks as Properties?

Tools like Wikidata:Tools/OpenRefine can pull properties from reconciled data, as well as update them on WD. But I can't think of a way to do the same with labels, descriptions and sitelinks, because (afaik) they don't have their Ps and Qs. I'm new to all this and I find it a little strange that they're treated differently — and for what reason? Are there any secret properties (or virtual properties that APIs can decode properly) to set these values programatically? Something like "Spanish description (Pxxx)" for Q13 is "miedo al número 13"; "frwiki title (Pyyy)" for Q13 is "Triskaïdékaphobie"? Ponor (talk) 18:04, 20 June 2023 (UTC)

@Ponor Most tools can handle these in a similar way they handle properties. You can use "Len" or "Den" to get the English label/description in the corresponding OpenRefine dialog window. Vojtěch Dostál (talk) 18:17, 20 June 2023 (UTC)
@Vojtěch Dostál: awesome, thanks! Sounds like the best kept secret (until today). OpenRefine adds SPARQL: in front of Lxx/Dxx, but it does the trick. As Sxxwiki does for the sitelinks. Ponor (talk) 19:39, 20 June 2023 (UTC)

Creating items for internet archive collections

The Internet Archive organises their material into collections (eg https://archive.org/details/ship-design-drawings for ship plans) which I'd like to reference using described by source (P1343) qualified by URL (P2699). These collections seem substantial enough to have their own entry here, but what instance of (P31) should they have, as they are not really a web archive (Q30047053) which is for archiving web sites. Do we have many internet archive collections on WD already? Vicarage (talk) 06:44, 20 June 2023 (UTC)

Suggest you create a new item for "Internet Archive collection" which is part of "Internet Archive" and a possible subclass of "web archive". — Martin (MSGJ · talk) 13:40, 20 June 2023 (UTC)
Sounds sensible. I've created Ship Design Drawings archive (Q119789127) which is an Internet Archive collection (Q119789115) which is in turn a subclass of the broader information repository (Q6031177), and used it with HMS Wolfhound (Q5635005) Vicarage (talk) 14:44, 20 June 2023 (UTC)
@MSGJ You changed Ship Design Drawings archive (Q119789127) to be a subclass. I think it should be an instance because its the end of an inference chain. Vicarage (talk) 13:55, 21 June 2023 (UTC)

Results of the survey on ontology issues reusers are facing

Hi everyone,

A while ago we did a survey among reusers about the different types of ontology issues they are facing when building applications and more using data from Wikidata. The results are available now. More details at Wikidata talk:Ontology issues prioritization.

Cheers Lydia Pintscher (WMDE) (talk) 09:03, 21 June 2023 (UTC)

scolarité - Property:P69 invisible

Hi. I would like to see Property:P69 in the Infobox of many people. The information John Muir High School is registered (exemple Ken Whittingham (Q1986502) but not published one the page Ken Whittingham. It happens many times and I don't understand why. Thank you for your help. Laugriville (talk) 14:55, 21 June 2023 (UTC)

I believe this is a local problem on the French Wikipedia. The template fr:Modèle:Infobox Cinéma (personnalité) does not seem have have an "educated at" field to map educated at (P69) into. You might propose adding it (see fr:Discussion_modèle:Infobox_Cinéma_(personnalité)#Formation_/_éducation) or switch the article to using a different infobox. Bovlb (talk) 17:53, 21 June 2023 (UTC)
@Laugriville Bovlb (talk) 00:50, 22 June 2023 (UTC)

Delete Property JSTOR topic ID (P3827) - JSTOR Topic ID

The JSTOR Topic ID property is no longer in use and all associated links in Wikidata are permanently broken. The JSTOR Topic ID property was not created by someone at JSTOR and we do not know its origins. This property is causing issues with Google's crawl process and with images for JSTOR.

To give some background, the JSTOR Topic ID was used to map Wikidata items to topics in JSTOR. However, JSTOR has ceased using these topic IDs, which has led to a significant number of broken links in Wikidata. This not only disrupts the user experience but also affects the search engine's ability to accurately index and retrieve relevant information.

Given the long-term discontinuation of the use of topic IDs by JSTOR, the related Wikidata property (P3827) has lost its functional relevance. Thus, I propose that this property should be deleted to maintain the integrity of data in Wikidata and prevent any potential misinformation or misdirection. Adding a note from our internal search team:

"We noticed a recent increase in the number of pages that are indexed but blocked by robots.txt. This is not likely related to the usage issues we are seeing, but something to look into anyway. We noticed topic pages were one of the main culprits, which is kind of weird because we thought were were handling them appropriately through the 404 redirect. It looks like google may be trying to crawl and index them because of external references via wikidata." Nabilamari12 (talk) 16:32, 21 June 2023 (UTC)

CC original property supporters: @ Pigsonthewing, Magnus Manske , Sadads , Runner1928, Strakhov, YULdigitalpreservation. Bovlb (talk) 17:59, 21 June 2023 (UTC)
We don't delete properties because external bodies break their inbound links (Q119417469 refers). Where are these "significant number of broken links in Wikidata"? As can be seen, for example, at Q431#P3827, our values for this property are not even linked (and have not been since September 2020). I would also draw the attention of User:Nabilamari12 to WD:PAID. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:43, 21 June 2023 (UTC)
P.S. I also note discussion on the property talk page, where it is reported that we were told by a JSTOR representative in September 2020: We're looking to reactivate our topics functionality sometime near the end of the month; I would expect that then these links should function for you as they did before. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:57, 21 June 2023 (UTC)
We do not usually delete dead identifiers because there may be archives of the data. For example at [3]. Deleting the data would hurt the integrity of Wikidata. Who are you quoting in that last paragraph? BrokenSegue (talk) 04:58, 22 June 2023 (UTC)

Variable railway gauge

There are train types with variable gauge. It is a special property of a railway vehicle/train to be able to do this. There is variable gauge (Q1133339) but this general, not a specific train type. For example I have two values for the property track gauge (P1064) in RENFE Class 121 (Q1244119). But it would be preferable to specify this is a variable gauge train with two values. By the way: I could not find the Spanish gauge of 1668 mm. source: openrailwaymap I created Commons:Category:Convertible gauge trains. Smiley.toerist (talk) 08:59, 22 June 2023 (UTC)

People with different names

Next issue with Japanese names (though I’m positive this one does not apply to Japanese ones only): Rin Nonomiya (Q11644898) appears to have two different names, Rin Nonomiya (野々宮りん) and Mei Hasegawa (長谷川めい). How to deal with this as far as given name (P735) and family name (P734) are concerned? In particular I’m wondering how to make clear which given name belongs to which family name (family name (P734) doesn’t seem to be a valid qualifier for given name (P735)). (At least one of the two names is obviously a nickname/pseudonym but I have no idea which one.) --Data Consolidation Officer (talk) 20:00, 21 June 2023 (UTC)

I think you just list all their forenames in given name (P735) and all their surnames in family name (P734). In general, it's not possible to construct a full name from these fields. There are other name fields that could possibly be used. Ghouston (talk) 00:36, 22 June 2023 (UTC)
That’s a pity but I suppose it’s on purpose. By the way, is there any semantic difference between name (P2561) and name in native language (P1559) when the language of the former is the native language of the subject, and should name (P2561) be used if name in native language (P1559) is already present? And, since we are already talking about naming properties, which one to use in the case of Rin Nonomiya (Q11644898), where at least one of the names is a pseudonym but it is unclear which one? Two values for name in native language (P1559)? Or alternative name (P4970)…? --Data Consolidation Officer (talk) 21:00, 23 June 2023 (UTC)

Templates that exist in most projects

Hi, I'm looking for a list of templates that are available in most wikis (list of Q items that are Wikimedia templates sorted by number of projects that have them). Is there a way to do it? Thank you. Bennylin (talk) 12:02, 22 June 2023 (UTC)

In principle something like this:
SELECT ?template ?templateLabel ?count
WHERE 
{
  {
    SELECT ?template (COUNT(?sitelink) AS ?count)
    WHERE
    {
      ?template wdt:P31 wd:Q11266439;
                ^schema:about ?sitelink.
    } GROUP BY ?template
  }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
} ORDER BY DESC(?count)
Try it!
The query times out, however. Maybe someone (at Wikidata:Request a query?) has an idea how to optimize it. --Data Consolidation Officer (talk) 15:54, 22 June 2023 (UTC)
The results are basically the same as just all the items with most sitelinks, by the way.
SELECT ?template ?templateLabel ?count WITH {
  SELECT ?template ?count WHERE {
    ?template wikibase:sitelinks ?count. hint:Prior hint:rangeSafe true.
    FILTER(?count >= 250)
  }
  ORDER BY DESC(?count)
  LIMIT 100
} AS %results WHERE {
  INCLUDE %results.
  ?template wdt:P31/wdt:P279* wd:Q11266439.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
ORDER BY DESC(?count)
Try it!
--TweetsFactsAndQueries ([[User talk:|talk]]) 16:25, 22 June 2023 (UTC)
Ah, yes, I wasn’t aware of wikibase:sitelinks. This and the FILTER(?count >= 250) does the trick:
SELECT ?template ?templateLabel ?count
WHERE 
{
  ?template wdt:P31/wdt:P279* wd:Q11266439;
            wikibase:sitelinks ?count.
  FILTER(?count >= 250)
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
} ORDER BY DESC(?count)
Try it!
--Data Consolidation Officer (talk) 12:06, 23 June 2023 (UTC)

"Wikidata property for items unsuitable as sources"

Would this be a bad idea? There are definitely some identifier that i think we should avoid to use as references (unless the source is about the site itself obviously) Trade (talk) 21:23, 15 June 2023 (UTC)

Obviously no identifier should be treated as unsuitable unless there is a high degree of consensus from the community before someone brings that up--Trade (talk) 21:28, 15 June 2023 (UTC)
I don't think I understand how you imagine using such a property. Can you give an example usage? BrokenSegue (talk) 01:09, 17 June 2023 (UTC)
Everipedia would be one example Trade (talk) 21:00, 17 June 2023 (UTC)
Is that because Everipedia has a lot of crap data (I've never seen this site)? Vladimir Alexiev (talk) 14:55, 21 June 2023 (UTC)
No I don't mean an example of such a source. I mean an example of how the property would be used. Also, I was not aware Everipedia was problematic (other than its association with blockchain)? BrokenSegue (talk) 16:02, 21 June 2023 (UTC)
Everipdia have a history of stealing content from Wikipedia. Everything else is mostly spam and a lot of unsourced stuff in addition to offering paid editing. --Trade (talk) 14:01, 24 June 2023 (UTC)
Oh wait you aren't suggesting a new property. I think I understand now. BrokenSegue (talk) 16:02, 21 June 2023 (UTC)
Yes i am suggesting an item to use with P31 Trade (talk) 13:53, 24 June 2023 (UTC)

Merge Wizzard error

Unable to merge (apparent) duplicate items Q4717091 and Q4713774. The conflict seems to be duplicate entries in en:wp and pt:wp. -- Cl3phact0 (talk) 16:29, 22 June 2023 (UTC)

@Cl3phact0: if there are duplicate articles in a language we can not merge items. That said it looks like the two rivers are different (comparing the "duplicate" articles). BrokenSegue (talk) 17:24, 22 June 2023 (UTC)
@BrokenSegue: Thank you, I'll have a closer look. It does seem like something is amiss as both items have the same name (though I have no particular knowledge of the geography of the region – I came across this while trawling Special:Random). The item that was suggested for improvement was: Gălășeni River (Q4717091). I simply tried to add the Description "river in Romania" when I became aware of the conflict. (Admittedly not likely to change the the course of history.) -- Cl3phact0 (talk) 19:03, 22 June 2023 (UTC)
yeah it would require someone with knowledge of the area to solve this conflict. you could possibly mark it as said to be the same as (P460) to highlight the issue BrokenSegue (talk) 19:07, 22 June 2023 (UTC)
Fixed by adding the county name to each description (just like how the villages of village : small clustered human settlement smaller than a townGălășeni (Q724030): village in Sălaj County, Romania and village : small clustered human settlement smaller than a townGălășeni (Q726035): village in Bihor County, Romania were done).
--Quesotiotyo (talk) 21:59, 23 June 2023 (UTC)
Thank you. Gălășeni River (Q4717091): river in Bihor County, Romania; Gălășeni River (Q4713774): river in Sălaj County, Romania – simple solution! Also, thanks for the example using {{Q*|n}} nomenclature. Quite a useful trick that. -- Cl3phact0 (talk) 06:36, 24 June 2023 (UTC)

Merge assistance required - person used two names

I have merged two items that both used the name Robert Arthur Buddicom (Q21537396) but when he moved from England to South Australia with a woman other than his first wife, he mostly used the name Robert Arthur Buddicom (Q94657179). Both had Wikidata items, and some of the identifiers exist under both names. I've created and intend to expand further the English Wikipedia article, initially based on the Australian Dictionary of Biography article under the Buddicom name even though he mostly used Bedford in Australia. I need help from someone with more expertise than I have to properly combine these entities please. -- ScottDavis (talk) 13:31, 24 June 2023 (UTC)

@ScottDavis: ✓ Done — Huntster (t @ c) 13:07, 25 June 2023 (UTC)

Getting started

Hello everyone, I am new to wikidata and has anyone really been as far as to even want to decide to try and even go as far as to use more like? Josivise (talk) 22:08, 24 June 2023 (UTC)

What exactly do you want to find out? --Wolverène (talk) 22:16, 24 June 2023 (UTC)
they are just trolling. see [4]. BrokenSegue (talk) 06:27, 25 June 2023 (UTC)
lol --Wolverène (talk) 06:30, 25 June 2023 (UTC)

duplicate pages (IMO 9495753) & (IMO 8129656)

Q84258388 and Q56372055 refer to the same ship, IMO 9495753. These need to be merged. The former name is "KS Titan 2" (2008-2013) the current name appears to be "Bull Ray" (since 2013) -- 64.229.90.172 20:46, 25 June 2023 (UTC)

Done Vicarage (talk) 21:23, 25 June 2023 (UTC)

Q101805659 and Q110155918 refer to the same ship. IMO 8129656, "Titan 2". These need to be merged. There also exists Q101805681 which could also be merged into the same, as most Commons categories exists as the multilingual link on the topic Q-page, instead of as a separate category Q-page -- 64.229.90.172 20:50, 25 June 2023 (UTC)

Create an account and you can do it yourself. Multichill (talk) 20:54, 25 June 2023 (UTC)
Done and tidied. You might want to read https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Ships#Are_Wikidata_items_for_ships-by-different-names_to_be_merged_into_one_Wikidata_ship-by-name_item? to see how the interaction of multiple ship names here and on Commons is contentious. Personally I think we should have one item per hull, but don't want to upset wikimedians. Vicarage (talk) 21:27, 25 June 2023 (UTC)

How do I export structured data added to a file at commons?

I would like to download the structured data added to a file. I understand it will be a JSON file, but I do not know how to do it. Something like https://commons.wikimedia.org/w/index.php?title=File:Star-background.png&action=raw&slot=data gives a maintenance error. Not sure if the data is the name of the slot. Just a guess. Thanks for your help.

PS I also asked here, but I am unsure if this is the correct community to ask. (https://commons.wikimedia.org/wiki/Commons:Help_desk#How_do_I_export_structured_data_added_to_a_file?) I will share the info. [[kgh]] (talk) 08:04, 25 June 2023 (UTC)

@Kghbln: POTD as json, rdf & and more. Multichill (talk) 20:52, 25 June 2023 (UTC)
@Multichill Thanks a bunch!! Indeed, this is the way! The next question would have been how I may figure out the entity ID. However, I got this info via &action=info. Clicking the link to "Special:EntityPage" results in a bad request since neither .json nor .rdf is appended, but once you do it manually, it works. I believe this is a bug. Anyhow, great to know. I will self-answer on commons to get interested people to your answer here. [[kgh]] (talk) 15:00, 26 June 2023 (UTC)
@Kghbln:The entity id is just the page ID prefixed with "M". On every image you can find a link in the toolbar on the left labeled "Concept URI". I don't think it's a bug, you're probably not using it correctly. Multichill (talk) 15:28, 26 June 2023 (UTC)
@Multichill Ah, well, this is a bit faster and easier. Anyhow, try to click on the first link, "A tinted ..." at the very bottom on https://commons.wikimedia.org/w/index.php?title=File:Balaklava_sick_2.jpg&action=info. I understand that no format is preset. However, just clicking on a link should not result in a bad request - invalid ID. I still think it is a bug, but it does not worry me. Clicking on Concept-URL redirects me to the file page, which is expected but does not help people trying to get hold of the structured data manually. [[kgh]] (talk) 16:02, 26 June 2023 (UTC)

Q9738728

Please check what happened in the history of this Wikidata element as of 5:21 pm on 25 June 2023‎. First the editor unmerged another Wikidata element, and then the same editor does precisely the same thing? What's the justification? Thank you for your analysis and eventually send it to the rightful person. Thanks. GualdimG (talk) 12:47, 26 June 2023 (UTC)

The user already responded to you on Talk:Q9738728 with an explanation. Why don't you follow up there? Bovlb (talk) 14:38, 26 June 2023 (UTC)
You think, really, that was a acceptable explanation? Please check what happened. GualdimG (talk) 15:26, 26 June 2023 (UTC)
I don't think it's a perfect explanation. Undoing and redoing merges as a way to review differences is very much not an ideal process. Perhaps there is a tool improvement to be suggested here.
Nevertheless, the user did respond to you with an explanation, and you have not engaged with it. The ball is in your court here. In most cases, we expect editors to make an attempt to sort things out with each other before coming here.
CC @Tm. Bovlb (talk) 15:39, 26 June 2023 (UTC)
I think the editor was very aware of what was doing. Is not an inexperienced editor. "Undoing and redoing merges as a way to review differences is very much not an acceptable process". But, and now? May I reset wat was done? It will began an editing war? If the historic of the element stays like that, "the crime does pay". Is not the first time he does this, and will do it again. GualdimG (talk) 16:05, 26 June 2023 (UTC)

Wikidata weekly summary #582

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Discussions
    • Open request for adminship: Wolverène (RfP scheduled to end after 2 July 2023 18:46 UTC)
    • Closed request for adminship: Vargenau (successful) Welcome onboard \o/
    • New requests for permissions/Bot: RomesfulImporterBot (Task: Creating and updating Russian Lexemes by transferring information from the Russian Wiktionary)
  • Tool of the week
  • User:Nikki/NarrowUI.css is a User script that tries to make Wikidata work better on narrower screens, including mobile devices.
  • Other Noteworthy Stuff
    • TU Berlin research group "The Restitution of Knowledge" are exploring the use of Wikidata to build a Linked Open Data (LOD) project that maps German colonial military expeditions into Africa. Interested in getting involved or learning more about their project? Reach out on the project's talk page: Wikidata:WikiProject TheRoK/Data Models.
  • Development
    • Wikibase REST API: We are working on making it possible to get statements from a Property. (phab:T338383)
    • EntitySchema: We're looking into the feedback we received from the testing of the new datatype.
    • Ontology issues: We have released the survey results. The next step is looking into solutions.
    • mul language code: We are continuing to adapt the desktop UI to make it ready for the new special language code. (phab:T329644, phab:T338302, phab:T339103, phab:T329655, phab:T329655)
    • Vector 2022: We are changing the style for search suggestions as the default of the skin isn't working well for Wikidata. (phab:T327510)

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

Emoji

Hi, can you please confirm if emojis are allowed to use in aliases? Thanks. Madamebiblio (talk) 23:08, 18 June 2023 (UTC)

it seems to be pretty common practice. e.g. heart (Q826930). so i assume it's defacto accepted. BrokenSegue (talk) 23:54, 18 June 2023 (UTC)
Could this be a usecase for a (new?) property along the lines of “associated emoji/emoticon/pictogram”? --Data Consolidation Officer (talk) 19:10, 21 June 2023 (UTC)
Doesn't Unicode character (P487) do exactly that already? El Grafo (talk) 12:21, 26 June 2023 (UTC)
Er, yes, that’s why I put “new” in brackets with a question mark. Didn’t know whether there was already a property for that and was too lazy to search. --Data Consolidation Officer (talk) 09:34, 27 June 2023 (UTC)

Merge candidate

Grillo telephone (Q91940918): 1960s flip phone telephone from Italy and Grillo telephone (Q5609175): 1960s flip phone telephone from Italy appear to be redundant. The former seems to be a subset of the latter. Should they be merged? -- Cl3phact0 (talk) 08:32, 27 June 2023 (UTC)

They should be merged if the concepts are exactly the same. If one is a subset of the other, then this means the concepts are not exactly the same. — Martin (MSGJ · talk) 13:43, 27 June 2023 (UTC)
To be more precise: the first item is one iteration of the Q5609175#P195 claim of the second. The label and description are confusingly similar. Something about this seems incorrect as is. -- Cl3phact0 (talk) 14:40, 27 June 2023 (UTC)
It appears the first item was created as a scrape of the Israeli museum website. So, while theoretically Grillo telephone (Q91940918) could be changed to instance of (P31)exhibit (Q9252000)/subclass of (P279)Grillo telephone (Q5609175), there's no data that's unique to the museum entry, so merging the two seems the best course of action. — Huntster (t @ c) 14:53, 27 June 2023 (UTC)

Batch remove spam'd references

Any suggestions for a way to batch remove spam'd URLs from wikidata. Looks like a coordinated campaign tried to spam links to globalzonetoday.com (which is just a low quality site) using lots of random IPs. BrokenSegue (talk) 18:23, 27 June 2023 (UTC)

Only by bot I am afraid. Ymblanter (talk) 18:53, 27 June 2023 (UTC)
Not what you asked for, but maybe add the site to the spam blacklist? Infrastruktur (talk) 20:01, 27 June 2023 (UTC)

Items for Japanese surnames: dealing with spelling variations

When I logged in today, I was notified that a link had been made from Mika Nonomiya (Q27918071) to Nonomiya (Q116878297), an item I had created to be able to add a family name (P734) to Nonomiya Ikuko (Q110733705).

The issue is that Mika Nonomiya (Q27918071)’s surname is spelled 野々宮 in kanji (Q82772) while Nonomiya Ikuko (Q110733705)’s is spelled 野宮. (Both are spelled identically when using kana (Q187659).)

Should this be one item (with more than one native label (P1705)) or should there be two items for the surname “Nonomiya” (one for 野々宮 and one for 野宮)? Apparently this depends, at least partly, on whether it’s “just” two different spellings of the same surname or two different (but homophonous) surnames, but I’m not knowledgeable enough about the etymology of Japanese surnames to answer that question. (I do know that said etymology is often not reflected in the kanji (Q82772) spelling, which is one reason why I bring up this topic here.) --Data Consolidation Officer (talk) 21:07, 18 June 2023 (UTC)

Even if the reading of the kana is the same, if the kanji is different, it constitutes a different surname. For instance, the surname Saitō (Q42306803)(Saito) has nine different homonymous surnames associated with it, yet each of them is distinct. These variations have diverse origins, but Japan does not consider them and treats them all as separate surnames. Therefore, 野々宮 and 野宮 are different surnames. Afaz (talk) 23:53, 18 June 2023 (UTC)
The kanji used is the same in this example though. Does the use of an iteration mark (々) make it a separate surname? —Xezbeth (talk) 04:35, 19 June 2023 (UTC)
Nonomiya (野々宮) and Nonomiya (野宮) are distinct family names, despite the only difference being the presence or absence of the repetition symbol "々". They are strictly differentiated in official documents. However, in postcard addresses, it is often considered a common mistake. Each name is derived from a different place name and has a distinct geographical distribution[5][6]. Afaz (talk) 07:35, 19 June 2023 (UTC)
Reading "野宮" as Nonomiya is a less common case. It is typically read as Nomiya. It may be the surname of a noble or priest. Afaz (talk) 14:02, 19 June 2023 (UTC)
In the case of Nonomiya Ikuko (Q110733705) it is most probably the “noble” variant. Thank you all very much for your answers. I’ll create a new item for 野々宮. --Data Consolidation Officer (talk) 19:06, 21 June 2023 (UTC)
There is now Nonomiya (Q116878297) for 野宮「ののみや」, Nonomiya (Q119822679) for 野々宮 and Nomiya (Q119822783) for 野宮「のみや」. --Data Consolidation Officer (talk) 19:27, 21 June 2023 (UTC)
…and does someone happen to know which one to use for Q110949125 (野宮定逸), Q110950836 (野宮定俊) and Q110948999 (野宮定縁)? --Data Consolidation Officer (talk) 19:33, 21 June 2023 (UTC)
@Afaz: Judging from Q110949125relative (P1038)Q110948999, Q110948999relative (P1038)Nonomiya Sadamoto (Q6537242) and Q110950836relative (P1038)Nonomiya Sadamoto (Q6537242), all three items should probably have family name (P734)Nonomiya (Q116878297) as well? --Data Consolidation Officer (talk) 09:39, 27 June 2023 (UTC)
That's right. They are the nobles of the Nonomiya family (Q11645309). Afaz (talk) 19:50, 27 June 2023 (UTC)
OK, I’ve added family name (P734)Nonomiya (Q116878297) to all three items, and family (P53)Nonomiya family (Q11645309) as well. --Data Consolidation Officer (talk) 08:18, 28 June 2023 (UTC)

Legion of Honour

human : any member of Homo sapiens, unique extant species of the genus Homo, from embryo to adultPatricia Davies (Q119890155): British World War II code breaker received a Legion of Honour (Q163700): highest French order of merit. The award received (P166) property returns an error when this is added. What property should be used? -- Cl3phact0 (talk) 20:10, 27 June 2023 (UTC)

Since a lot of other items are using it that way, and there's nothing on the Talk page suggesting any alternative usage, just add it and ignore the warning. Presumably its status as state order (Q56291528) isn't sufficient for the purposes of award received (P166), and somebody will figure out how to fix it, some day. Ghouston (talk) 00:36, 28 June 2023 (UTC)
Strictly you should say which class of award it is, Knight etc, but still the connection to award is not present. I agree, someone will reconcile orders with awards one day. Its a classic rabbithole. Vicarage (talk) 05:55, 28 June 2023 (UTC)
@Cl3phact0 Given the pictures of the french embassy i'm pretty sure it's a Knight of the Legion of Honour (Q10855271). RVA2869 (talk) 06:29, 28 June 2023 (UTC)

Carlo Rovelli (Q1042470)

In Carlo Rovelli (Q1042470) I can't find the start times about being a professor at Western University and at Perimeter Institute for Theoretical Physics. Also his main work at Aix-Maresille University should be checked. This is his latest CV I could find. Thanks in advance.-- Carnby (talk) 05:40, 28 June 2023 (UTC)

A script to lowercase current language label

A few months ago I have created a a script that allows to easily lowercase the first letter of the label in the current language (by either double-clicking on a button, or tapping and confirming on touch devices). A little bit of advertisement for the sake of the few folks who might save some keystrokes by using it. Improvement suggestions are welcome. --Base (talk) 04:46, 22 June 2023 (UTC)

@Base: That looks to be quite helpful :) I tried it on mobile, but the confirmation pop-up is in Ukrainian I think – any chance it can be set to the browser default, or maybe to English?
One other thing I'm wondering is that Wikidata displays the labels for all the languages I have in the Babel template in my user page. In that case, would it change all of them to lowercase, or just the one I've set up in my preferences? I can see cases where I'd want to leave the capitalization in English or German (say, "Mexican footballer") and change it to lowercase only in Spanish ("futbolista mexicano"). –FlyingAce✈hello 21:33, 23 June 2023 (UTC)
Wait – never mind, it's the label rather than the description. Oops! I still could see a case for leaving the capitalization in German (since nouns are capitalized) while wanting to change other languages to lowercase. –FlyingAce✈hello 21:35, 23 June 2023 (UTC)
@FlyingAce, yes, it is only about labels (with the caveat that it will also remove an alias if that alias matches the lowercased label) and as it is currently set up it works for the current interface language. You can of course switch to another language per page via ?uselang=languagecode but of course it would be faster to just edit the label in that particular language a traditional way. I can add the possibility to edit other languages too, but I am a bit unsure where shall I place the button for it as not to make the interface too cumbersome, perhaps you have some suggestions on that? As to the pop up in Ukrainian, yes, I guess I shall add some l10n :) --Base (talk) 21:55, 23 June 2023 (UTC)
Is it like User:Ricordisamoa/LowercaseDescription.js but not for descriptions? Also there exists cool User:Nikki/LowercaseLabels.js for changing any language label to lowercase. --Infovarius (talk) 19:23, 28 June 2023 (UTC)

Best way to handle items that conflate projects with their products?

E.g. California High-Speed Rail (Q782093), Wendelstein 7-X (Q315738). The first is a rail system, the second a fusion plant. However, it's common in casual idiom to refer to such items and the projects that create them with the same term, and so they often end up as a single Q-item. Obviously, a Q-item should only refer to a project or its product, not both. But suppose I want to query, say, a list of all transport megaproject (Q59264914), including California High-Speed Rail (Q782093). What's the best way to link the latter to the former, as the basis for such a query? P31 is out, since that creates a disjointness violation. You could create a separate item, e.g. California High-Speed Rail project (creator (P170)? operator (P137)?), but I'm not sure it's reasonable to have separate "project" and "product" items in many cases. I've done California High-Speed Rail (Q782093)has cause (P828)transport megaproject (Q59264914), but I wonder if there is a better general solution. Swpb (talk) 18:02, 27 June 2023 (UTC)

If an organisation is set up to deliver a single product, I see little merit in splitting it into 2 terms. I don't see why a Q item can't be both. Some organisations are created with different names, like TransManche Link building the Channel Tunnel, and then the organisation is complicated and interesting in its own right, so deserves a separate entry. Films are often made by production companies set up specially for the purpose, you'd not want those to be separate, as they are a legal fiction within the bigger film company. Vicarage (talk) 18:59, 27 June 2023 (UTC)
I guess it depends on how ontologically consistent we're trying to be. entity (Q35120) is a disjoint union of (P2738) object (Q488383), property (Q937228), and process (Q3249551); every entity is an instance of exactly one of those. If Q782093 is both a rail system (object) and a project (process), then it violates that disjointness, and User:TomT0m/classification.js flags it. I don't think that causes any downstream inference issues though. @TomT0m:, what do you think? Swpb (talk) 19:44, 27 June 2023 (UTC)
"I don't think that causes any downstream inference issues though" - it causes problems in my project where I want to identify for example events/human behaviour events and ones that exist as physical entities (see say Wendelstein 7-X (Q315738)) Mateusz Konieczny (talk) 12:15, 28 June 2023 (UTC)
Can you give some more detail on the problem this causes for you? Swpb (talk) 14:23, 28 June 2023 (UTC)
It causes false positives in https://matkoniecz.github.io/OSM-wikipedia-tag-validator-reports/ when something actually physically existing is also merged with event/abstract project. See also https://github.com/matkoniecz/OSM-wikipedia-tag-validator#story-behind-this-tool Mateusz Konieczny (talk) 16:27, 28 June 2023 (UTC)

Merge Q212431 (Installation Art) and Q20437094 (installation artwork)

Would someone who is more versed with Wikidata than me please consider merging Q212431 (Installation Art) and Q20437094 (installation artwork)?

The English Wikipedia page already redirects one to the other (https://en.wikipedia.org/wiki/Installation_art). Some authority references are identical, like the Art & Architecture Thesaurus ID. Graefestrasse (talk) 14:57, 28 June 2023 (UTC)

"Art" and "work of art" are different concepts for sure. So merge would be wrong. --Infovarius (talk) 22:13, 28 June 2023 (UTC)

Announcing the new Elections Committee members

You can find this message translated into additional languages on Meta-wiki.
More languages • Please help translate to your language

Hello there,

We are glad to announce the new members and advisors of the Elections Committee. The Elections Committee assists with the design and implementation of the process to select Community- and Affiliate-Selected trustees for the Wikimedia Foundation Board of Trustees. After an open nomination process, the strongest candidates spoke with the Board and four candidates were asked to join the Elections Committee. Four other candidates were asked to participate as advisors.

Thank you to all the community members who submitted their names for consideration. We look forward to working with the Elections Committee in the near future.

On behalf of the Wikimedia Foundation Board of Trustees,

RamzyM (WMF) 17:59, 28 June 2023 (UTC)

Animated television series

animated series (Q581714) has existed for many years and has always been about "animated series" generally, to match the various Wikipedia articles that are not about a particular medium. Two months ago animated television series (Q117467246) was created, an item more specifically about animated television series. The animated series item was recently changed from a subclass of television series (Q5398426) to a subclass of series of creative works (Q7725310), throwing up thousands of constraint errors across thousands of items and making animated television series (Q117467246) the one that should ideally be used on items. Having gone through hundreds of items, most were an instance of either animated series (Q581714), television series (Q5398426) or both, which suggests that users didn't have a clear idea of what they should be doing.

My question: is this split acceptable? A similar thing was done with anime (Q1107) and anime television series (Q63952888) in 2019, so the change made sense to me. —Xezbeth (talk) 04:27, 24 June 2023 (UTC)

The item in question has meant TV series specifically and that is how it has been used. Its recent repurposing is breaking thousands of items. It is better to use a new superitem. Máté (talk) 06:49, 24 June 2023 (UTC)
Split is necessary. Máté, it's not "repurposing". There was no original "television" animated series, just look at some early versions as e.g. this - most labels and all sitelinks don't include "television", only one superclass does, but it was changed many times after. --Infovarius (talk) 19:08, 28 June 2023 (UTC)
I wasn't saying not to have two items. I was suggesting to do it the other way around. We have one that has been used on animated TV series so let's keep using it that way. – Máté (talk) 08:19, 29 June 2023 (UTC)
Another issue that I see: how to distinguis "television" from "not television" animated series? What this even means - "television"? Why animated series should be defined by television at all? It is sometimes aired on TV, sometimes several times, sometimes not - it's not a defining property imho. --Infovarius (talk) 19:11, 28 June 2023 (UTC)
animated series (Q581714) would also include animated film series (Q56884562), animated web series (Q117467240) like YouTube videos or flash animations, animated short film series (Q113791292) and series of cartoon shorts that do not air on TV. Granted the majority of items are television series. Category:Animated television series (Q7164609) is used on 50+ Wikipedias so subdividing it the same way here shouldn't be controversial. —Xezbeth (talk) 06:42, 29 June 2023 (UTC)

Given a coordinate, how to easily identify wikidata items near it？

sometimes i find a target on google maps, so i get its coords. the method i'm using now to find nearby wikidata items is to copy paste the coords into wikishootme url, e.g. https://wikishootme.toolforge.org/#lat=48.8&lng=2.3 for (48.8 N, 2.3 E). this method is still a bit cumbersome because i need to copy paste lat and lng separately. just wondering if there are better tools/methods for doing this? RZuo (talk) 10:41, 24 June 2023 (UTC)

Hello @RZuo: in articles (e.g. en:Canton of Bagneux) you can click on the coordinates in the right upper corner and you will get to
In the section "Wikipedia articles" you get various links with these coordinates
for example:
M2k~dewiki (talk) 12:47, 24 June 2023 (UTC)
Special:Nearby / Geodata API currently only supports one coordinate per wikidata object, while rivers might have two coordinates (water spring + river mouth) for example:
M2k~dewiki (talk) 12:56, 24 June 2023 (UTC)
Might also work querying with SPARQL
#defaultView:Map
SELECT ?place ?placeLabel ?location ?instanceLabel ?dist WHERE {
  SERVICE wikibase:around {     ?place wdt:P625 ?location .
 bd:serviceParam wikibase:center "Point(2.3 48.8)"^^geo:wktLiteral .
      bd:serviceParam wikibase:radius 2. # in kilometers
}  OPTIONAL {     ?place wdt:P31 ?instance .  }
  SERVICE wikibase:label { bd:serviceParam wikibase:language 'en' }
  BIND(geof:distance(?loc, ?location) AS ?dist) .

} ORDER BY ASC(?dist)
Try it!
Bouzinac 💬●✒️●💛 17:07, 24 June 2023 (UTC)
@M2k~dewiki, Bouzinac: thx a lot.--RZuo (talk) 10:18, 29 June 2023 (UTC)

How to present the relations between a lake, a dam and a hydropower station？

Dongjiang Lake (Q10870562) is formed because of Dongjiang Dam (Q1241395), which hosts Dongjiang Hydropower Station (Q84764639). is there a usual/standard way of presenting these relations? RZuo (talk) 13:40, 24 June 2023 (UTC)

I think the current standard is at Wikidata:WikiProject Energy/Review-1#Hydroelectricity ontology. --ScottDavis (talk) 14:40, 24 June 2023 (UTC)
Yes those are the main ones. In addition I would use Dongjiang Dam (Q1241395) part of (P361) Dongjiang Hydropower Station (Q84764639) — Martin (MSGJ · talk) 14:01, 27 June 2023 (UTC)
@ScottDavis, MSGJ: thx a lot.--RZuo (talk) 10:18, 29 June 2023 (UTC)

Lemmy instance communities as properties

I'm wondering if URLs of communities hosted on Lemmy instance (Q120061681) servers could be added as a new property you could link to a certain item, just like you can link subreddit (P3984) into items. CrystalLemonade (talk) 20:29, 28 June 2023 (UTC)

@CrystalLemonade: yes that could be done. see Wikidata:Property proposal. BrokenSegue (talk) 21:39, 28 June 2023 (UTC)
Okay, I made a proposal here, but it seems like the proposal page for its category seems to have stopped including them after the "Unicode character (item)" one. CrystalLemonade (talk) 22:52, 28 June 2023 (UTC)
@CrystalLemonade: yeah there is something of a backlog. sorry. BrokenSegue (talk) 05:15, 29 June 2023 (UTC)

Survey of Scottish Witchcraft project

Hi, I'm Ruby and I'm a student working on the Survey of Scottish Witchcraft project at the University of Edinburgh and we are quality checking our data after 5 years work adding Scottish witchcraft in Wikidata. We've created a project page here.

Part of what I'm checking is where we have information on the guilty, half guilty, or innocent verdicts of the Scottish witch trials but can't see a property for verdicts in the Propbrowse tool. Is there a reason why trial verdicts don't seem to be recorded or a different way of modelling this I'm not aware of? Would convicted of (P1399) work and exonerated of (P7781) for items of people as to their guilt or innocence? Exonerated seems to have a different meaning however and implies a correction of a guilty verdict. I still think there might need to be verdict property for items about trials to cover the guilty verdict, not proven verdict, innocent/acquitted verdicts. I am also looking at the information on their sentences (e.g. banishment, excommunication, execution) and what we can do with this information. Currently we are using penalty penalty (P1596) as it seemed the best fit of existing properties but are wondering if any other properties would work better. I note there has been previous discussion about whether verdict needed a property required. Any help or advice would be gratefully received. Delane13 (talk) 15:44, 27 June 2023 (UTC)

I agree that a new property might be good. Maybe as a qualifier for charge (P1595) because a person might be found guilty on some charges and not others? ChristianKl ❪✉❫ 15:59, 29 June 2023 (UTC)

Proper way to link to a reference that was from one date with information stated at another date?

Hi. I was looking at Mario (Q12379) and added a source for Miyamoto's claim that his surname is Mario, in 2015. I then added a statement for his claim that he has no last name; however while the second article was published in 2015 the interview in it is from 2012 and I want to make that clear in the statements. Is there a way to properly distinguish between the date an article was published, and when the thing referenced in it was said? DemonDays64 T | C 22:22, 26 June 2023 (UTC)

@DemonDays64, someone else may know of some random property that will work, but one option would be to wrap the quote in quotation marks, and append –Miyamoto, 2012. — Huntster (t @ c) 04:21, 27 June 2023 (UTC)
Use a "point in time" (or start/end date) qualifier on the statement. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:27, 30 June 2023 (UTC)

Jewish people languages

On Jewish people (Q7325), I see many languages listed as native language (P103), and many others listed as languages spoken, written or signed (P1412). Some are listed in both categories, others in only one. The selection seems very arbitrary. Is there an explanation for this? Lights and freedom (talk) 02:29, 28 June 2023 (UTC)

Frankly, I don’t see a point in having native language (P103) or languages spoken, written or signed (P1412) statements for Jewish people (Q7325) at all. They may make sense for items about ethnic groups closely tied to a language (e.g. indigenous people), but not so much for Jewish people (Q7325), a group essentially defined by the religion of its members (though I do know that the issue is more complex than that). There is simply no such thing as “the language of the Jewish people” (and if there were, it would probably be Hebrew (Q9288)). The statements should be deleted. In general, I’d recommend using native language (P103) and languages spoken, written or signed (P1412) for items about humans only. --Data Consolidation Officer (talk) 08:28, 28 June 2023 (UTC)
There should be a way to designate the languages traditionally associated with a nation or an ethnicity. Should be strange if the item for the Jewish people has no connection to items for Hebrew, Yiddish, Ladino, Italkian, etc. If P103 and P1412 aren't good for this (they surely aren't), which property may fit here? --Wolverène (talk) 08:56, 28 June 2023 (UTC)
(Deleted my own comment because it went a bit too much off-topic.) I’ve posed a question at Wikidata talk:WikiProject Languages#Languages associated with groups of people. --Data Consolidation Officer (talk) 09:07, 30 June 2023 (UTC)
By the way, the country (P17) statements on Jewish people (Q7325) suffer from the same problem. There are probably some Jews in any country (not just the ones listed there) and apart from Israel (Q801), none of them has a compelling reason for being considered specifically associated with the Jewish people. --Data Consolidation Officer (talk) 09:34, 28 June 2023 (UTC)

Request for help with formatting the examples in a property proposal

Hi all

I've just done my first property proposal in maybe two years Wikidata:Property proposal/World Geographical Scheme for Recording Plant Distributions ID. I feel confident the property will be useful and used, but I'm not sure I've done the examples correctlyand can't find any instructions for what is 'correct'. What I'm trying to say in example 1 is that the city of Buenos Aires (the capital of Argentina) uses the code AGE-BA in this database. If anyone knows a clearer way of saying this in the proposal please feel free to correct it.

Thanks

John Cummings (talk) 12:11, 30 June 2023 (UTC)

@John Cummings: Template:Property proposal has detailed instructions on how to do this - did you use the template? In particular your examples should be reversed and should provide a Qid for the item you want the property to apply to. Also if there is a formatter URL it is helpful to link the ID using the formatter in the examples you provide. ArthurPSmith (talk) 14:39, 30 June 2023 (UTC)
Thanks, I just pressed the button to start a new property proposal and used the fields that were provided. John Cummings (talk) 18:17, 30 June 2023 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2023/06&oldid=1930954208"