Wikidata:Project chat/Archive/2022/10

From Wikidata
Jump to navigation Jump to search
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

RIP Ahoerstemeier

Ahoerstemeier was quite a prolific editor here so I thought it best to announce that he has passed away (see his condolence page on the German Wikipedia and obituary on the English Wikipedia). Graham87 (talk) 10:31, 28 September 2022 (UTC)

  • Since they are no longer living, and would not be involved in self promotion, and would have an obituary online, shouldn't they get a Wikidata entry under their real name? --RAN (talk) 16:32, 29 September 2022 (UTC)
Thank you for the notice; I’ve added him to WD:Deceased editors and updated his user page. If someone who knew him better wants to update his obituary, please do. Galaktos (talk) 13:31, 3 October 2022 (UTC)

Purpose of non existent entities?

What is the purpose of non-existent entities such as Q111015181 and Q111023369? I could understand if they were vaporware or proposals for the future, but some of these details seem to come from thin air.

PS: I'm not sure if it's possible to close an RFC, but I goofed by misunderstanding what that meant and initially posting my question there. Is it possible to request the closure of an RFC created in such a way? Thanks. Mbrickn (talk) 04:14, 3 October 2022 (UTC)

They make sense to me. For example it is possible for there to be articles about PlayStation 6 (Q111015181) even though they don't exist. Here is one. They also serve as placeholders for eventual items when they presumably become a thing. BrokenSegue (talk) 05:00, 3 October 2022 (UTC)
It makes sense to have items for things that are likely to exist, I agree. But I'm not sure why PlayStation 6 (Q111015181) has a mass (P2067) statement — surely that's not known yet? Sam Wilson 05:48, 3 October 2022 (UTC)
This is basically what I'm saying. Some of the other details don't make sense either, like the inclusion of the PS3 GPU and CPU in the RSX and Cell, the publication date, and the price. Mbrickn (talk) 11:17, 3 October 2022 (UTC)
I've deleted Q111023369 under WD:N. It had no sitelinks or references of any kind. — Martin (MSGJ · talk) 06:42, 3 October 2022 (UTC)
I have deleted your errant RFC. Another time you might consider using {{Delete}}. Bovlb (talk) 13:08, 3 October 2022 (UTC)

10th Anniversary of Wikidata

I just noticed that Wikidata was launched on 29 October, 2012. Are there any plans to mark the 10th anniversary? Leutha (talk) 10:48, 3 October 2022 (UTC)

Yes! Wikidata:Tenth Birthday is the overview, with links to some of what has been/will be taking place. Vahurzpu (talk) 12:25, 3 October 2022 (UTC)
Cool. I heard my local wiki was promising cake and a snazzy workshop, but as we all know the cake is a lie! Do take them up on the offer on the workshop though. :-) Infrastruktur (talk) 16:49, 3 October 2022 (UTC)

Coolest Tool Award 2022: Call for nominations

The fourth edition of the Coolest Tool Award welcomes your nominations! What is your favorite Wikimedia related software tool? Please submit your favorite tools by October 12, 2022! The awarded projects will be announced and showcased in a virtual ceremony in December.


MediaWiki message delivery 18:30, 3 October 2022 (UTC)

Wikidata weekly summary #540

Here's your quick overview of what has been happening around Wikidata over the last week.
  • 🎊 Events
    • Upcoming
      • Next Linked Data for Libraries LD4 Wikidata Affinity Group call
        • October 4, 2022: Lana Soglasnova and Roman Tashlitskyy will talk about their preliminary work in creating and editing items for Slavic people and some of the complications involved in this work, especially around transliterating names in various languages. Agenda
        • Wikibase Working Hour! 25 October 2022, 2pm Eastern (Time zone converter) Amy Ruskin of Northeastern University Library, will speak on the topic of Wikidata vs. custom Wikibases: Community history case studies, Boston’s Chinatown. The Boston Research Center (BRC) is a digital community history and archives lab based in the Northeastern University Library. One of our current projects involves taking an inventory of historical materials related to Boston’s Chinatown, and we have been using Wikibase to store multilingual data about the linked collections, organizations, and people. In this presentation, we will discuss our experience of getting started with a custom Wikibase and give an overview of our progress so far on the Chinatown Collections project. Amy Ruskin is the Data Engineer in the Digital Scholarship Group in the Northeastern University Library. She has a Master's degree in Information Studies from McGill University and a background in computer science and statistics. Registration link
      • Live editing session on Twitch, in French, by Vigneron, October 4 at 18:00 CEST (UTC+2)
      • [Small wiki toolkits] Upcoming "How to interact with Wikidata via Pywikibot" workshop on Friday, October 7th, 16:00 UTC
      • The State of Wikidata and Cultural Heritage: 10 Years In. Oct 4, 2022 08:00 AM in Pacific Time (US and Canada)
      • Talk to the Search Platform / Query Service Team—October 5th, 2022
      • Wikidata Office Hour GLAM-Hack − online meet-up about the "library world" with Wikidata. October 5th, 12.00 UTC (online via Zoom)
    • Ongoing
      • Wikimedia Österreich, in cooperation with Wikimedia Deutschland, started the DACH Culture Contest as a Börthday present for Wikidata 10th anniversary! It features two categories: "libraries in Austria, Germany and Switzerland" and "Culture in Austria, Germany and Switzerland". So if you would like to improve data about libraries, books, music, art, video games, cinema etc as a börthday treat, join us and get the chance to win some nice prizes! The contest will last until October 16, 2022.
      • Weekly Lexemes Challenge #61, Time
  • 🧑🏿‍💻 Development


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.

Representing how politicians vote

I think the voting record of politicians is an incredibly important point of data.

In NZ Parliament, each bill goes through three readings, where parties vote on bills (or if it's a conscience vote, politicians independently vote). If it passes all three readings, it becomes law.

If I wanted to represent which politicians voted for and against the a particular bill (for example Fair Trading (Soliciting on Behalf of Charities) Amendment Bill (Q113945916) at each of its three readings), is there a way this could be represented? Is there a "voted for" and "voted against" property? I can see votes received (P1111), but this appears to be for counting amount of votes received, rather than individual votes received.

How might this be structured? One vote might be easy enough with simple "voted for/voted against" properties with a value for each politician, but there are three votes for every bill. Would you need an item per reading of the bill? And detail the vote there?

Any experience/examples of recording the voting history of politicians would be appreciated. Supertrinko (talk) 23:10, 2 October 2022 (UTC)

tracking voting is hard (there are lots of votes on lots of things) we should probably start with more modest ambitions. like just tracking all legislators with external IDs to vote tracking websites that already exist. I don't think we currently have the properties you suggest. BrokenSegue (talk) 15:33, 3 October 2022 (UTC)
Unfortunately, the only vote tracking website in NZ is voted.nz, and that is specific to conscience votes only. Realistically, the only true record of votes is in Hansard (Parliamentary verbatim spoken record). So if NZ wants a true record of voting, a database needs to be established. I'm already going to be going through old hansard digitising old content, so I'll be placing voting information in a database anyway, I may as well look towards collecting it in a way that it'll be compatible with wikidata. I think a generic yes/no for/against vote would be easy enough to structure and could be used across a multitude of use cases. Supertrinko (talk) 20:50, 3 October 2022 (UTC)
Ah, I was thinking more U.S. centric where a number of databases exist. If there are no existing databases what sources do you plan to use as references for the statements? Sounds like you plan to create your own database by digitizing content. I'm not opposed to trying to store vote data on wikidata but I think it's really tricky. BrokenSegue (talk) 21:20, 3 October 2022 (UTC)
NZ stores our votes in the official parliamentary record, what we call "Hansard". Essentially detailed minutes of debate in parliament. And at the bottom of each debate, it will record the votes on the bill in plain text format. Not very machine readable at all.
So yeah, it'll definitely be tricky and unfortunately will simply require manually recording votes from those minutes/hansard. Supertrinko (talk) 03:31, 5 October 2022 (UTC)

Heuristic references

I've asked for some thoughts over at Help_talk:Sources#Heuristic_references - briefly, what guidance should we have about when and how to use heuristics like inferred from image (Q105573271) (responses there ideally to keep them together). T.Shafee(evo&evo) (talk) 23:11, 5 October 2022 (UTC)

Modelling of buildings with multiple uses over time

Most heritage buildings have had multiples uses over time. These various uses can be documented in the same Wikidata item, as in Sanderson Centre for the Performing Arts (Q38386517) or in Aeolian Hall (Q4687855). This modelling strategy seems to work relatively well but it requires to qualify most statements with start and end times. At the same time, this may also present challenges with properties such as date of official opening (P1619) (that may apply to the initial opening or to a reopening after a closure and renovations) or architect (P84) (that may apply to the architect of the initial construction or to the renovation). My questions follow... Fjjulien (talk) 02:59, 4 October 2022 (UTC)

Are there circumstances where it would be better to have two distinct items representing different usages of the same building at different points in time? If renovations are made to the exterior of the building and it's no longer the same architectural structure? If the interior of the buildings is profoundly renovated? For example, I stumbled upon The Registry Theatre (Q110374281) and The Registry Theatre (Q56285771), which represent the same building on the exterior, although with renovations in the interior. Is there a justification for keeping these two items separate or should they be merged? Fjjulien (talk) 03:14, 4 October 2022 (UTC)
I think these two can be merged (if they indeed represent the same building). Please be careful not to merge institutions and buildings, but in this case both theatre building (Q24354) and land registry building (Q30310495) are buildings, not institutions/organizations. Vojtěch Dostál (talk) 09:32, 4 October 2022 (UTC)
@Vojtěch Dostál As per your suggestion, I documented both uses of the building within Q110374281, which I then renamed as "The Registry Theatre" to reflect the most common name that the building is known as today. I also kept the other item, Q56285771, which I modelled as a performance hall (Q112688641) (i.e., a performance space within the building) with its own attributes that are distinct from the building's. This dual modelling approach matches OpenStreetMap, where the performance space is represented as a node and the building as an area. Fjjulien (talk) 03:21, 6 October 2022 (UTC)
@Fjjulien Yes. That wasn't exactly my suggestion but it makes sense. Sometimes venues may be considered just a part (eg. a room) of a building. Vojtěch Dostál (talk) 06:00, 6 October 2022 (UTC)
Under which circumstances should structure replaced by (P167) and structure replaces (P1398) be used? The examples provided seem to suggest that these properties are to be used when a building has entirely ceased to exist (it was burned down or demolished) and a new building was erected in its location. Is this indeed the case or could these properties also be used when a building is substantially renovated to serve another purpose?
Beat Estermann (talk) 22:21, 31 March 2017 (UTC) Beireke1 (talk) Beireke1 (talk) 12:07, 2 May 2017 (UTC) - collaborating with Romaine on Belgian data on performing arts venues. Affom (talk) 15:26, 12 May 2017 (UTC) Anvilaquarius (talk) 11:41, 15 September 2017 (UTC) PEAk99(talk) PEAK99 (talk) 20:32, 29 January 2019 (UTC) Boxomi (talk) 22:21, 24 September 2019 (UTC) Antoine2711 (talk) 00:19, 5 December 2019 (UTC) Fjjulien (talk) 21:15, 13 February 2020 (UTC) Vero Marino (talk) 21:15, 13 February 2020 (UTC) Bello Na'im (talk) 08:34, 24 September 2021 (UTC) Titanboo (talk) 17:23, 8 July 2022 (UTC) dlh28 (talk)18:21, 22 September 2022 (UTC)
Notified participants of WikiProject Cultural venues Fjjulien (talk) 03:22, 4 October 2022 (UTC)
I'd rather advocate for a splitting, if there is too many different properties for a stuff at time X and same stuff at time Y Bouzinac 💬●✒️●💛 11:02, 4 October 2022 (UTC)

Global ban for Александр Мотин

Per the Global bans policy, I’m informing the project of this request for comment: RfC/Global ban for Александр Мотин. - Sleeps-Darkly (talk) 03:30, 6 October 2022 (UTC)

Flag replace request (png to svg)

Is there someone that can kindly replace File:Brasschaatvlag.png with File:Brasschaat vlag.svg in flag image (P41) of Brasschaat (Q693513)? Many, many and many thanks in advantage!!! 93.32.64.24 14:03, 6 October 2022 (UTC)

Looks like this has already been updated by Gatto bianco. :) –FlyingAce✈hello 17:00, 6 October 2022 (UTC)

Help with SPARQL query

Hi, I've been using this SPARQL query to get the entities with values for ClassInd audiovisual work ID (P8958) and ClassInd rating (P3216) that don't have content descriptor (P7367) as a qualifier for ClassInd rating (P3216). I got most of it, but the problem I have now with the query is that some of the values it returns have ⧼novalue⧽ as the qualifier value and I want to exclude them from the result too. How can I change the query to remove those entries from the result set? 

SELECT ?item ?value ?id WHERE {
	?item p:P3216 ?statement; wdt:P8958 ?id .
	OPTIONAL { ?statement pq:P7367 ?qualif } .
	FILTER( !BOUND( ?qualif ) ) .
	?statement ps:P3216 ?value
} LIMIT 100

Agabi10 (talk) 22:15, 6 October 2022 (UTC)

@Agabi10: I think the triple you're seeking is ?statement a wdno:P7367 ., so, as below. Here's some documentation. iirc, a is shorthand for rdf:type, and ⧼novalue⧽ for a predicate is represented by making the subject a member of the class of nodes which have a ⧼novalue⧽ for a given property.
SELECT ?item ?value ?id WHERE {
  ?item p:P3216 ?statement; wdt:P8958 ?id .
   FILTER NOT EXISTS { ?statement pq:P7367 ?qualif . } 
   FILTER NOT EXISTS { ?statement a wdno:P7367 . }
   ?statement ps:P3216 ?value
} LIMIT 100
Try it!
--Tagishsimon (talk) 22:24, 6 October 2022 (UTC)
Yes, it seems that now is working as I wanted. Thanks. Agabi10 (talk) 22:32, 6 October 2022 (UTC)

Picasso copyright violation

Q20553609 contains this image which is probably a copyright violation. As said a few paragraphs ago (Landscape at L'Estaque (copyright violation ?)), this may be public domain in the US, but not outside. Laurent.Claessens (talk) 15:44, 11 October 2022 (UTC)

You need to challenge the copyright status of the image at Commons, not here. Bovlb (talk) 16:01, 11 October 2022 (UTC)
This section was archived on a request by: Image was deleted at Commons. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:50, 12 October 2022 (UTC)

internalization problem

Mandariine (talk • contribs • logs) created a fork Q48772092 because the article in italian is not internationalize and removed the italian link from strikebreaker (Q1340502). I tried to explain that although the italien concept is slitly different it is the same concept. could someone fix the problem. thx, Yanik B 13:29, 30 September 2022 (UTC)

pfff... ou comment créer des pb là où y en a pas ! voir aussi ! fr:crumiro est la trad de it:crumiro qui retrace l'histoire et la spécificité des briseurs de grève italiens mais n'est pas un article général sur le concept international ! au même titre que fr:briseur de grève au Québec fr:crumiro est un article détaillé de fr:briseur de grève ! le jour où les copains transalpins écriront un article général il pourra être lié à l'article général en français ! pour le moment ce n'est pas le cas ! si vous voulez que ça le soit : à vos plumes vos encriers ou vos claviers ! alley : courage ! Mandariine (talk) 13:51, 30 September 2022 (UTC)
Slighlty different means that it's different for the purposes of Wikidata. ChristianKl ❪✉❫ 13:17, 7 October 2022 (UTC)

The right level of granularity with P131

For most users of Wikidata, the right level of granularity located in the administrative territorial entity (P131) is at the municipality level. When a dataset is uploaded to Wikidata, the available string of text for P131 is usually a municipality. When data is reused outside of Wikidata, the municipality is also usually the level of granularity that users will ask for. But then municipalities themselves have a need to locate their public buildings and other local assets at a more precise level: they need to sort them by municipal district (Q2198484), administrative sector (Q28203007) or other types of municipal subdivision. How could we best meet the information needs of municipalities and the information needs of other users? Both values would deserve to be ranked "preferred" value: the municipality-level value can be deemed best referenced value (Q98386534); the municipal subdivision value can be considered most precise value (Q71536040). Could we use qualifiers to make it possible to query either the municipality-level value or the subdivision value? Or should we write fancy queries to infer the municipality-level value even if it's not explicitly stated? Fjjulien (talk) 04:00, 4 October 2022 (UTC)

I think it's best to use the most precise value (or multiple with a preferred value). And then infer higher levels of division by walking the graph. BrokenSegue (talk) 04:48, 4 October 2022 (UTC)
This is a very difficult area, and one not made any simpler by making uncited dubious assertions "For most users of Wikidata, the right level of granularity located in the administrative territorial entity (P131) is at the municipality level" nor suggesting palpably wrong solutions "the municipality-level value can be deemed best referenced value (Q98386534)", nor by confusing precision and granularity "the municipal subdivision value can be considered most precise value most precise value (Q71536040)". It's difficult because arrangements and concepts vary from country to country. Difficult because communities are served by multiple administrative entities simultaneously. Difficult because we really have not even got started yet: WD concentrates on - shall we say - municipal authorities and currently tends to ignore domain specific adminstrative territorial entities (flood defence committees, health boards &c); it won't always be this way, these things will be added. Difficult because authorities do not always nest neatly (hello New Zealand). And then there's history; the administrative entities today are not those of 50 or 100 years ago, and there are valid reasons why WD users might want WD to store such data. I suspect the solution(s) will involve all of a) qualifiers b) walking the class tree c) fancy queries predicated in part on country specific knowledge (with the concomitent issues: where do we store country specific information providing the knowledge needed to understand the country's data). How far we are away from producing an accessible dataset is easily illustrated by moving to the top level of country government: it is not possible, afaik, successfully to query for all of the legislative houses of country governments. Much less to successfully query for country-region and municipal-level government across multiple countries. Whereas there is much to be done on a country by country basis to improve WD's dataset, I think we're quite far from being able to apply one-size-fits-all solutions in this area. --Tagishsimon (talk) 07:58, 4 October 2022 (UTC)
To be honest I have seen this parameter (mis) used in such a variety of ways that it could be probably best if we had an RfC and then the result would be written in the property description. Ymblanter (talk) 13:46, 4 October 2022 (UTC)
My usage (fortifications, museums) of located in the administrative territorial entity (P131) is usually to select things for maps and lists when I want "about 50 of them", so it much depends on how common things are. historic county (P7959) is ideal for me, I wish there were a common size I could adopt worldwide, rather picking State in the US, Departement in France etc. But for dockyards I'd want country, churches, then municipality. So what I'd want is a graph walker that stopped at X thousand people, and have located in the administrative territorial entity (P131) at the level that "an educated person in the country might have heard of", so generally colloquial municipality names, and not those horrid invented local authority names. Vicarage (talk) 14:51, 4 October 2022 (UTC)
It might be worth looking whether any standardization body create a concept that could be used and does what you want. ChristianKl ❪✉❫ 12:37, 7 October 2022 (UTC)

Order of "also known" for title article

Hello, I was trying to restore my edits in Q622682, and reorder the title article in English, such as "ζ(3), Zeta(3), Apery's constant". However, the order did not change after I restored it, and it remain display "Zeta(3), Apery's constant, ζ(3)".

P.S.: I was trying to copy from English, but then I accidentally removed ζ(3). So I added it again but the order is not exactly like the previous edit. If I did something wrong, please remind me. Dedhert.Jr (talk) 04:22, 4 October 2022 (UTC)

if I'm understanding you correctly. the order of the aliases does not matter on wikidata. this is also true of the order of statements. BrokenSegue (talk) 04:37, 4 October 2022 (UTC)
The problem is that it matters for usability but the data model itself declares that the order doesn't matter. I think it would make sense to add the order into the data model and be able to manually change it. ChristianKl ❪✉❫ 12:33, 7 October 2022 (UTC)

Aliases and alphabetization

Aliases are not currently alphabetized, which can make it difficult to find a variant if you only know approximately how it's spelled (and for medieval artists, there can be dozens and dozens of variants).

Would it be possible to integrate an "alphabetize aliases" function into the software, in much the same way that properties are clustered regardless of the order in which they're added (assuming that phrasing is sensical)? DS (talk) 15:28, 4 October 2022 (UTC)

this should be doable with a short js function. I don't think everyone would want it alphabetized since sometimes the more common aliases are at top. BrokenSegue (talk) 23:54, 5 October 2022 (UTC)
Given that the data model says that aliases are not stored in a defined order it would make sense to sort them alphabetically when they are displayed. ChristianKl ❪✉❫ 13:03, 7 October 2022 (UTC)

“Art school” ambiguity

I just found out that there are two terms for “art school”, Q383092 and Q1792623, and the two terms are not mergeable because the German label for Q383092 specifically says the term means a school of higher education (other languages that I understand aren’t making this claim), but the description of Q1792623 (used only by German and Polish) specifically says it’s an educational institution “at any level”.

I wonder what’s the best way to resolve this situation. Al12si (talk) 23:46, 5 October 2022 (UTC)

Perhaps one is an "art college" ? DS (talk) 01:13, 6 October 2022 (UTC)
The German does make the distinction that Q383092 is art college (Kunsthochschule). But the Chinese and Japanese Wikipedias uses Q383092 to specifically mean art school at any level (the meaning described by Q1792623)
In North American English the usual term for “art college” (I graduated from one) is “art school”. I’m suspecting speakers of other languages don’t realize this is the case since this isn’t made clear in the description? Al12si (talk) 01:20, 6 October 2022 (UTC)
You start by thinking about the actual concepts that are involved. There are two concepts that are worth separating. One is about art schools on any level and the other is about schools of higher education. Given that both of those concepts are described by Wikipedia's we need concepts for both. You decide about which item is supposed to link to what. It can be helpful to do a reverse lookup before doing that to see how the items are used within Wikidata.
Once that first step is clear, you go through the Wikilinks and move them to the relevant item.
Afterward you think about what the best name of each concept happens to be. The English labels "art school" and "art college" seem good to me. It's good to also add descriptions. In the languages I don't speak I use the Wikilinks to guide what the label happens to be. I personally believe that wrong labels and descriptions are much worse than no descriptions exist, so I see a case for removing any labels and descriptions that have a good chance to be misleading.
It often happens that those Wikilinks exist because people from the individual Wikipedias want Wikilinks even when concepts are a bit different. You can create Wikilinks to redirects on relevant Wikipedias to provide that. At the moment that's quite cumbersome but hopefully soon it will get easier (the new way is finally in testing). ChristianKl ❪✉❫ 10:43, 7 October 2022 (UTC)

Removing unsourced ethnicities

Did we agree to use a bot to remove unsourced ethnicities? I remember agreeing we should not do it for religion. See for example this edit removing ethnicity and this edit. I though we agreed to find sources for unsourced ethnicities, rather than delete them and lose the information. I could see removing the information if you find an obvious contradiction or error, and no source. RAN (talk) 16:03, 29 September 2022 (UTC)

[conflicted many times] I think the bot is acting on the basis of Wikidata:Bot requests#request to depreciated ethnic group only sourced with P143 (2021-10-23) (opened one year ago). Personally I tend to support the removal per the motivations of the proponent, @Fralambert:, who anyway proposed as best solution the deprecation with the plausible reason that we risk a readdition of such removed statements from Wikipedias. At the end of 2021, a discussion similar to this (Wikidata:Project_chat/Archive/2021/12#removing_unreferenced_ethnicity_rather_than_finding_a_reference) led to the revert of a similar mass-removal of unsourced ethnicities. --Epìdosis 16:35, 29 September 2022 (UTC)
@Richard Arthur Norton (1958- ) Please see Wikidata:Project_chat#Ethnicity_statements_again and the linked bot request page. Vojtěch Dostál (talk) 16:23, 29 September 2022 (UTC)
  • Isn't the a minimal number of people involved (a quorum) to pass something so controversial? It look like three people agreeing with each other on something proposed on 24 September 2022, 5 days ago. We have Requests for Deletions that are older. --RAN (talk) 16:29, 29 September 2022 (UTC)
    I am not aware of any such rule. Anyway I see 7-8 people supporting it and 1 against at the bot request page, and one additional supporter here in project chat. Vojtěch Dostál (talk) 16:38, 29 September 2022 (UTC)
Indeed, if it were "so controversial", one would have expected to see more - or any at all - discussion on this board when the deletion proposal was raised here five days ago. --Tagishsimon (talk) 16:47, 29 September 2022 (UTC)
In a previous discussion on Project Chat the last time this was attempted, there was not a consensus for this move. To my knowledge none of the participants in that previous discussion were notified of the Bot chat discussion, and this move should be reversed until there's a clear consensus. Gamaliel (talk) 17:02, 29 September 2022 (UTC)
I don't care too much one way or the other - I am just a bot operator responding to an old request with a clear consensus - but please note that the community was also already notified at Property_talk:P172#Bot_request one year ago. I think I am not responsible for reading all the discussion archives in Wikidata to see if this had ever been discussed somewhere else. Vojtěch Dostál (talk) 17:23, 29 September 2022 (UTC)
I suspect that until users understand that unreferenced ethnicity statements will be removed, ethnicity statements without references will continue to be added. WD tried undeleting such that there would be time to provide references, last time around; but the time was not, it seems, used to provide references. Why would it be different this time around? --Tagishsimon (talk) 17:29, 29 September 2022 (UTC)
Hmm... yeah... if anything, starting a discussion about the "mandatory reference" constraint on this property would be a more logical step for those who have opposed this. Because the constraint makes no sense if we just ignore it anyway. Vojtěch Dostál (talk) 17:36, 29 September 2022
That's not the case, I personally added perhaps thousands of references to ethnicity statements using OpenRefine after the unsourced ethnicity statements were restored. Gamaliel (talk) 19:35, 29 September 2022 (UTC)
No, you definitely shouldn't have to do that! But then involved parties shouldn't have to read every bot request either. Gamaliel (talk) 19:35, 29 September 2022 (UTC)
  • One thing we should do before deletion/deprecation is explore all possible heuristic approaches that infer the property. Some values were imported from Wikipedia infoboxes and categories, but after looking at several deletions, the value can sometimes be found in in the Wikipedia prose text where the word "Jewish" or "Judaism" or "African American" or "Polish" or "Poland" may appear. We just need a new heuristic model: "Inferred from Wikipedia text". We should also have a separate list of instance_of=human" where people are are likely to volunteer to help add in a reference. --RAN (talk) 18:46, 29 September 2022 (UTC)
    That’s a terrible idea. Take Polish: The term can mean holding citizenship in any of the many Polish states there have been over the centuries, it’s often used for people who were born somewhere in modern-day and/or historic Poland or used to live there. Sometimes knowing Polish is enough. Sometimes it’s enough that a name vaguely sounds Polish. (And yes, sometimes ethnicity is meant.) As for “Jewish”: I have seen articles (not only on Wikipedia) where people were described as “Jewish” for all kinds of reasons, sometimes little more than carrying a name that sounds Jewish or being married to a Jew. And those seem to be relatively easy cases compared to other nations. It might be different for “African American” or similar cases but for the (Central) European context, requiring a proper reference is essential. --Emu (talk) 20:18, 29 September 2022 (UTC)
The heuristic model has a built in error rate, yet we already use it in over 10,000 entries. We use it based on surnames and first names. It detected over 500 errors in gender based on given-name. Remember, we are not adding heuristic ethnicity where it does not exist already, we are using it where ethnicity already appears in the Q-entry. Self identification of ethnicity also has a built in error rate, since we do not have access to their DNA. People believing they have Native American ancestry is one of the most common errors disproved by DNA analysis. --RAN (talk) 23:53, 29 September 2022 (UTC)
Then we should consider deleting those statements. Statements have no value just because they are already there.
Again, it seems that you are trying to use a very American concept of ethnicity for the whole world. It doesn’t really work for Europe and it might not work for a lot of different places. Maybe we could find common ground by looking at different regions differently? --Emu (talk) 10:33, 30 September 2022 (UTC)
I would be more offended by having my identity denied and deleted. --RAN (talk) 00:32, 30 September 2022 (UTC)
Native American is not a etnicity, this is more say you are European. Mixing Cherokee, Atikamekw or Slavey seem wrong to me. Fralambert (talk) 01:14, 30 September 2022 (UTC)
ethnic group (P172) is property that may violate privacy (Q44601380) which means "Values for living individuals should generally not be supplied unless they can be considered widespread public knowledge or are openly supplied by the individual themselves (otherwise hidden supporting references are not sufficient). As an example, the fact that someone's address is accessible by looking at a domain name registration doesn't imply that it's considered widespread public knowledge for the sake of this policy."
Inferred from Wikipedia text is not enough to demonstrate that this standard has been reached. Not listing the identity of people is not denying their identity. ChristianKl ❪✉❫ 16:19, 7 October 2022 (UTC):

If something is cited in a Wikipedia article - Wikipedia being the most widely used reference work in the history of humanity - then that certainly meets the definition of widespread public knowledge. Gamaliel (talk) 16:24, 7 October 2022 (UTC)

The fact that there's a moment in time where a Wikipedia article says something isn't proof of anything like that because anyone can edit Wikipedia. In the case of Polish as ethicity, we also need a source that actually claims that the person has that as ethicity and not just as nationality. Jewish is also a term that's often used to speak about the religion of a person. Ashkenazi Jews (Q34069) on the other hand is an ethnic group. ChristianKl ❪✉❫ 16:40, 7 October 2022 (UTC)
  • We use the most precise value available, just like we do for place of birth. For some people, we only know the country, for others we know the hospital. Having no information on ethnicity, is much worse than having too broad an ethnicity. As for "European" being too broad, it at least distinguishes between people that have African and Asian ethnicities. You can always burrow deeper later, and even do a breakdown like in an Ancestry DNA test with percentages, as shown in programs like Finding Your Roots. --RAN (talk) 21:15, 3 October 2022 (UTC)
    I aggrand that for him, we should at least use the Library of Congress, his enthnicity is implied by the Bain News Service. At least for Michelle O'Bonsawin (Q113557494) I use the government biographical note from her nomination. Fralambert (talk) 20:58, 29 September 2022 (UTC)
Since people who have Japanese nationality are Japanese, they are not usually referred to as "ethnically Japanese". Or should it be "Yamato people"? Unlike "Ainu people" or "Ryukyuans," the majority of Japanese are "Yamato people," so they are almost never mentioned in sources.--Afaz (talk) 01:58, 30 September 2022 (UTC)

Call for presentations for WCNA 2022 and Mapping USA

This is a bit of a short notice call for presentations by Sunday, October 15 (soft deadline) for WikiConference North America (Nov 11-13) held jointly with OpenStreetMap US's Mapping USA. We encourage you to please submit your Wikidata and other Wikimedia project proposals at Wikiconference:2022/Submissions and OpenStreetMap US proposals at Call for Proposals: WikiConference North America + Mapping USA. On Friday, Nov 11, we are looking for lightning talks in particular. We welcome workshops, editathons/mapthons, birds of a feathers, and presentations on Saturday and Sunday.

  • It would be great to have someone explain the any data intersections between Wikidata (& structured data on Commons, too!) and OpenStreetMap or OpenHistoricalMapss.
  • If you have been involved in the renaming effort from the US Interior Secretary's Order 3404 to remove Sq__ from geographic names & are active on the OpenStreetMap Slack, please see my comment on the #wikimedia channel.
  • Plus any other topic that you want to present on — we are looking for non-mapping presentations as well, particularly for Sunday, Nov 13.

Peaceray (talk) 16:32, 7 October 2022 (UTC)

How can I state that something can be used for a given purpose?

Here are some examples of what I mean

Note that some of these purposes might be referenced by multiple items (see the example on "store digital data"). Another example would be that some other online dictionaries (e.g. Collins English Dictionary (Q260453)) also show the "IPA transcription of words in English".

Rdrg109 (talk) 01:47, 1 October 2022 (UTC)

I shared this question in Wikidata Telegram group and someone mentioned the following in this message.

has use (P366) seems relevant.

If I were to use has use (P366) to state that Dictionary.com (Q16960686) can be used for "lookup IPA transcription of a word in English", I would need to create a Wikidata item for "lookup IPA transcription of a word in English" and then add the following statement Dictionary.com (Q16960686)has use (P366)lookup IPA transcription of a word in English. Some of the statements I could think of that could then be added to the Wikidata item "lookup IPA transcription of words in English" are
What do you think of this idea? If you think this is not a good idea, could you propose an alternative?
The reason why I'm interested in storing this information: People would then be able to search Wikidata items by their purpose. For example, someone interested in knowing which websites show IPA transcription of words (e.g. a linguist) would be able to generate a list of websites that can be used for "lookup IPA transcription of a word in a language" or any of its subclasses.
Rdrg109 (talk) Rdrg109 (talk) 06:08, 1 October 2022 (UTC)
If you can boil down the syntax to "subject verb object", "toaster cooks bread" I can see it fitting in the WD framework, but if you need more nuance I can see you quickly getting tied in knots creating new specific objects, which I think is unwise. Vicarage (talk) 08:33, 2 October 2022 (UTC)
Theoretically, you could do Dictionary.com (Q16960686)has use (P366)lookup IPA transcription of a word in English through nested qualifiers: Dictionary.com (Q16960686) has use (P366) lookup relative to (P2210) IPA transcription relative to (P2210) English word (Q89663479), however, Wikidata doesn't allow qualifiers to have qualifiers. So I suppose you could break that into a chain of items: lookup of IPA transcription of an English wordsubclass of (P279)lookuprelative to (P2210)IPA transcription of an English word, and IPA transcription of an English wordsubclass of (P279)IPA transcriptionrelative to (P2210)English word (Q89663479). Ditto for any other complex concept. Silver hr (talk) 10:22, 8 October 2022 (UTC)

Statements on subclasses

I am wondering if there is a similar rule to the following that applies to statements other than subclass and instances: "If an item A is an instance of class B, and class B is a subclass of class C, item A is implicitly also an instance of class C. There is no general need to add a statement for the relation A→C to Wikidata." Example: if a class A has the statement X (property studied in (P2579) with the value anatomy (Q514)), does a class B, which is a subclass of A, have to include the same statement X? Sjoel (talk) 09:01, 6 October 2022 (UTC)

Yes, there is. The relationship is that A is a B and B is a C implies that A is also a C is called transitivity. subclass of (P279) subclasses transitive Wikidata property (Q18647515) to indicate that this holds. instance of (P31) on the other hand is not transitive. ChristianKl ❪✉❫ 10:30, 7 October 2022 (UTC)
I thought instance of was transitive because of its value hierarchy property (P6609) setting Vicarage (talk) 22:43, 7 October 2022 (UTC)
I forgot that we had that. I think in this case that means if A instance_of B and B subclass_of C -> A instance_of C. That's also transitivity. On the other hand A instance_of B and B part_of C does not lead to A instance_of C.
Intuitively, no example comes to mind that's would not transitive over subclass_of. ChristianKl ❪✉❫ 22:13, 8 October 2022 (UTC)

how do i search if enwiki article (URL) is attached to any wikidata item

i want to search whether Robert Love (disambiguation) is attached to any item. i want to search using search box on wikidata website. Robert Love (Q16094413) has link to disam page. —Jindam vani (talk) 10:32, 4 October 2022 (UTC)

The search box cannot be used for that. But you can use a query like this: https://w.wiki/5mwT . The title of the page must be verbatim. Infrastruktur (talk) 11:05, 4 October 2022 (UTC)
Actually you can: Special:Search/"Robert Love (disambiguation)" (note the double quote). GZWDer (talk) 11:24, 4 October 2022 (UTC)
See also Special:ItemByTitle/enwiki/Robert Love (disambiguation). Lucas Werkmeister (talk) 12:28, 5 October 2022 (UTC)
Interesting. Although this is undocumented behavior. From the documentation you can basically surmise that labels and aliases are indexed. From that I assumed that descriptions and sitelinks were not, which I guess was not the case. Infrastruktur (talk) 13:35, 6 October 2022 (UTC)
Which documentation are you refering to? ChristianKl ❪✉❫ 20:05, 7 October 2022 (UTC)
Wikibase-specific extensions are documented on this page: https://www.mediawiki.org/wiki/Help:Extension:WikibaseCirrusSearch . Unfortunately the only way to know what is indexed or not is by reading the source code. Infrastruktur (talk) 06:30, 9 October 2022 (UTC)

How to state the role of a company, if not in P31?

Hello,

(I initially posted at Wikidata_talk:WikiProject_Companies#Businesses_and_role, but in the lack of answer I’m bringing this to a wider audience).

Over at Wikidata:WikiProject_Video_games we deal with entities − in general companies − that make (video game developer (Q210167)) and publish (video game publisher (Q1137109)) games. For example, Stray (Q96247255)developer (P178)BlueTwelve Studio (Q113156099)publisher (P123)Annapurna Interactive (Q38805988).

My question is about the instance of (P31) to use on these companies. Historically we have been using video game developer (Q210167) and video game publisher (Q1137109) and they are by far the most common pattern (see queries below [note that plenty of companies do both development and publication]) and we have plenty of constraints to enforce that modelling on external identifiers for such entities (see eg TheGamesDB publisher ID (P7642) or OGDB company ID (P7570))

I have noticed using such P31s runs afoul of the recommendations of WikiProject_Companies, which recommend to simply use, I believe, business (Q4830453) (the difference with enterprise (Q6881511) or company (Q783794) is a bit beyond me ^_^). That’s fine − but how then are we supposed to express that Annapurna Interactive (Q38805988) is a video game publisher (Q1137109)? We already use industry (P452) with video game industry (Q941594) ; product or material produced or service provided (P1056) feels a stretch and could apply to both developers and publishers (depending how you might interpret 'produced') ; occupation (P106) is only scoped to people… Thoughts? Jean-Fred (talk) 20:12, 3 October 2022 (UTC) 

#title: Most common P31 for entities that develop games
SELECT ?type ?typeLabel (COUNT(?game) as ?games) WHERE {
  ?game wdt:P31 wd:Q7889.
  ?game wdt:P178 ?developer.
  ?developer wdt:P31 ?type.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
} GROUP BY ?type ?typeLabel
ORDER BY DESC(?games)
Most common P31 for entities that develop games 
#title: Most common P31 for entities that publish games
SELECT ?type ?typeLabel (COUNT(?game) as ?games) WHERE {
  ?game wdt:P31 wd:Q7889.
  ?game wdt:P123 ?publisher.
  ?publisher wdt:P31 ?type.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
} GROUP BY ?type ?typeLabel
ORDER BY DESC(?games)
Most common P31 for entities that publish games

Jean-Fred (talk) 20:12, 3 October 2022 (UTC)

why not let something be both a video game developer (Q210167) and a business (Q4830453). That makes sense to me since it's possible for a video game developer (Q210167) to be a non-profit or an art-collective. I prefer using instance of (P31) in these cases. BrokenSegue (talk) 21:28, 3 October 2022 (UTC)
Hmm, I guess because my understanding of P31 is that it’s supposed to be about the very core of what an entity is, and video game developer (Q210167) isn’t: there are examples of research labs, political parties, cereal brands etc. who have made or published games − making it their P31 seems off. To me it’s very much like occupation (P106) − we only set instance of (P31)human (Q5). Jean-Fred (talk) 11:50, 5 October 2022 (UTC)
I mean looking at the SPARQL queries you created it seems like already the behaviour I described is most common. BrokenSegue (talk) 23:56, 5 October 2022 (UTC)
@BrokenSegue: What do you mean, to have both as P31? That does not seem to be the case − there are 269 items with P31 set to both video game developer (Q210167) and business (Q4830453) (and 128 to video game publisher (Q1137109) ) − a very far cry from the tens of thousands of developers we have.
If you meant, to have video game developer (Q210167) or business (Q4830453) as P31 − then sure, that’s why I said in my opening post “My question is about the instance of (P31) to use on these companies. Historically we have been using video game developer (Q210167) and video game publisher (Q1137109) and they are by far the most common pattern”.
Jean-Fred (talk) 08:31, 10 October 2022 (UTC)
There is the option of field of work (P101) video game development (Q1061635). P101 oddly missing from Wikidata:WikiProject Companies/Properties. --Tagishsimon (talk) 00:09, 6 October 2022 (UTC)
Hmmmm, interesting. That could work I guess. I suppose this would mean more or less abandoning the video game developer (Q210167) item. Jean-Fred (talk) 08:31, 10 October 2022 (UTC)

Braque "Maison à L'estaque" image

The page Houses at l'Estaque contains an image (P18) of a Cézanne because the correct painting by Braque is not public domain. I think that it would be better to put that image in the property related image instead. Laurent.Claessens (talk) 07:51, 10 October 2022 (UTC)

Better just to remove it altogether - done. It's not helpful to suggest a Cézanne is P6802 related to a Braque; their relation is properly specified using the P180 'depicts' property in Commons Structured Data. --Tagishsimon (talk) 08:10, 10 October 2022 (UTC)

Landscape at L'Estaque (copyright violation ?)

Landscape at L'Estaque contains this image by Braque who died in 1963. The image may be public domain in the USA for the same reason as this one, but it is most probably not public domain in most of the world.

I think we should remove the image from wikidata and update the copyright status on Wikimedia. Laurent.Claessens (talk) 08:50, 10 October 2022 (UTC)

@Laurent.Claessens If you are right, the correct course of action is to remove it from Wikimedia Commons and move it to enwiki, rather than just removing a link from Wikidata. Ping @Fuzheado:, the uploader. Vojtěch Dostál (talk) 09:29, 10 October 2022 (UTC)
I've started a deletion request at Commons, and left a message at en:Talk:Georges Braque. --El Grafo (talk) 10:22, 10 October 2022 (UTC)

Wikidata weekly summary #541

Here's your quick overview of what has been happening around Wikidata over the last week.
  • ✨ Other Noteworthy Stuff
    • The fourth edition of the Coolest Tool Award is looking for nominations. Please submit your favorite tools by October 12, 2022. The awarded projects will be announced and showcased in a virtual ceremony in December.
    • The September 2022 summary for the Wikidata Query Service backend update is out.
  • 🧑🏿‍💻 Development
    • MUL language code: fixed the order of languages shown in the termbox on desktop, especially when mul is used (phab:T311617)
    • Lexicographical data:
      • Made Special:NewLexemeAlpha use the same font as the rest of the wiki (phab:T313166) – this should be the last blocker for replacing the old special page
      • Disallowed *-x-qid with lowercase Q as Lexeme language codes in favor of *-x-Qid (phab:T317863)
    • REST API: worked on handling conditional HTTP request headers
    • Changed the unexpectedUnconnectedPage page prop so that Special:UnconnectedPages can show the latest pages first, which should finally resolve this old security task in production (phab:T300770)
    • Cleaned up the Wikibase and WikibaseLexeme ontology files (phab:T314360)
    • Added a few more globes for geocoordinates (phab:T314611) - Thanks, Mike!

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.

Accidentally created new item - need help merging

I accidentally created a new item for a church building because the original item did not use the full church name. I was unable to merge them, though. The error said something about the item name, so I changed them to match and I am still unable to merge them. See Q7590668 and Q7590668. What am I doing wrong? TamFultz (talk) 22:44, 10 October 2022 (UTC)

Dunno. I've merged St. Michael's Roman Catholic Church (Q114575109) into St. Michael's Roman Catholic Church (Q7590668) without issue. --Tagishsimon (talk) 22:47, 10 October 2022 (UTC)

Ontology test cases

@Multichill: @MisterSynergy: - maybe you are still interested in ontology testcases.

Code is now actually published, https://github.com/matkoniecz/wikibrain should be now installable

Still, plenty of bad design and insane code architecture, but should work at least. If there would be an interest I am likely to improve it or accept PRs.

previously discussed at https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2022/08#Is_there_a_better_place_to_report_blatant_failure_of_Wikidata_classification? --Mateusz Konieczny (talk) 01:29, 5 October 2022 (UTC)

Found this by coincidence, the ping somehow did not work. Maybe because you did not sign within the same section?! Anyways…
I think I have now been able to get a copy of all your packages that are supposed to be necessary for running these tests. Code is a bit mess indeed, but I'll try to figure out what's on here. —MisterSynergy (talk) 20:59, 7 October 2022 (UTC)
Yeah, using filesystem instead of database is quite cingeworthy there to see nothing about config setup I managed to invent. (right now I am refactoring part of code that I use to run wikipedia/wikidata link validation in OpenStreetMap ). @Multichill: as apparently ping has not worked well? Mateusz Konieczny (talk) 08:23, 8 October 2022 (UTC)
Honestly, not sure whether it's worth to put more effort into this thing.
As much as I am aware, the important functionality for Wikidata is this:
  • Determine whether a given item is connected to one of some items (predefined list) via a subclass-of property path
  • Print this path if available
Both of this seems to be doable with much less code, in a way that outsources practically all complex logic to the Wikidata query service (look for SPARQL ASK queries and the RDF GAS API available in Blazegraph).
The current implementation is not in best shape and in fact pretty difficult to work with for a plethora of reasons. I don't know how much you need this thing for other projects (such as OSM etc.), but I'd really suggest to consider a rewrite. Because the general idea is a good one actually, but it deserves a robust implementation. —MisterSynergy (talk) 17:51, 8 October 2022 (UTC)
Yes, the current version should be utterly rewritten. I will put more effort into this if it will be actual limiting factor. Right now it appears that interest in fixing this issues is quite limited and current horrible version is good enough to reveal more than entire Wikidata community can fix. Though maybe I should reach out somewhere else, not only Wikidata talk:WikiProject Ontology - maybe Discord/Telegram channels used by Wikidata community? Mateusz Konieczny (talk) 18:14, 8 October 2022 (UTC)
The best discussion page (to find specialists) is the Ontology WikiProject, the best discussion page (for a large audience) is this page. Discord and Telegram channels are difficult because barely anything that happens there transfers over into this project. —MisterSynergy (talk) 18:29, 8 October 2022 (UTC)
The problem is that I reported some problems there, with only small part getting fixed. This may be indicating that Wikidata community is not seeing a viable path toward Wikidata being able to classify things and being fundamentally unable to distinguish events from physical objects and so on. Mateusz Konieczny (talk) 10:30, 9 October 2022 (UTC)
IMO the process of fixing these issues properly is quite difficult, and not formally in any way defined. Thus only few users really engage in this process. Sure, one could remove a seemingly inappropriate P279 claim somewhere in the chain and make the tests pass, but as you mention by yourself, problems tend to reappear some time later because someone thinks that this problematic claim should indeed be there.
A proper solution of these problems considers at least:
  • Existence and nature of P279 claims in the item in question beyond the problematic one; generally the position of the item in question within the Wikidata ontology/P279 classification tree; an investigation with a graph tool is often helpful, but not simple.
  • Sitelinks to Wikipedias; do all cover the same topic? Do some of them cover a wider or narrower aspect of the concept than others? Language skills are important here. The multipurpose nature of Wikidata sometimes conflict with each other (in particular "provide sitelinks for Wikimedia projects" and "general knowledge base").
  • Backlinks to the item in question; these may require changes when the P279 is altered, but are usually not a blocker for a repair. It often needs advanced query and automation skills to repair backlinks, particularly if there is a larger number of them involved.
  • Constraint violations that appear or disappear, within the item in question as well as in linked ones.
The available tooling to solve these problems is probably not well enough; I also think that these issues are better fixed by a group of editors, rather than by (even skilled) individuals. —MisterSynergy (talk) 20:32, 9 October 2022 (UTC)
"these issues are better fixed by a group of editors, rather than by (even skilled) individuals" @MisterSynergy: What you mean by this? One person making analysis, another person making decision what should be edited and other making edit? I am a bit confused why it would be better Mateusz Konieczny (talk) 11:15, 11 October 2022 (UTC)
Collective analysis and decision-making; if there is more than one edit to make, potentially also the editing. These problems are pretty much the most difficult ones to solve here, in many regards. —MisterSynergy (talk) 14:35, 11 October 2022 (UTC)

(solved) Example of a found problem: Polish-Saxon Post Milestone Nowogrodziec (Q83545869) - milestone is action according to Wikidata

en: Saxon post milestone (a type of milepost in the former Electorate of Saxony) [1]

en: signpost (type of road sign) [2]
en: directional sign (type of road sign) [3]
en: traffic sign (signboard displaying information for road users) [4]
en: traffic control device (general term describing infrastructure deployed for road regulation; includes traffic lights, traffic signs, traffic calming devices, for example) [5]
en: media (storage and delivery agent of information or data) [6]
en: means (means by which an item performs a function) [7]
en: action (something an agent can do or perform) [8]

Mateusz Konieczny (talk) 01:29, 5 October 2022 (UTC)

Media is the culprit, as its a concept, neither an action or physical thing. It keeps cropping up in ontology moans. A split into 'storage medium', and understanding that 'communication' is both a noun and a verb would be a start, but I'm no linguist. I expect WD is full of these problems. Vicarage (talk) 08:42, 5 October 2022 (UTC)
Removed media in https://www.wikidata.org/w/index.php?title=Q17480810&diff=1743617461&oldid=1708214464 - without replacement Mateusz Konieczny (talk) 10:47, 5 October 2022 (UTC)

(two initial comments were duplicated when splitting section)

Example of a found problem: Tulalip Tribes of Washington (Q1516298) - tribe is human according to Wikidata

en: federally recognized Native American tribe in the United States (Native American tribe, band, or native village formally recognized by the American federal government) [9]

en: Native American tribe (group or community of indigenous peoples in the United States) [10]
en: Native Americans in the United States (indigenous peoples of the United States) [11]
en: Americans (citizens or residents of the United States of America) [12]
en: inhabitant (person who lives in a certain place) [13]
en: human (common name of Homo sapiens, unique extant species of the genus Homo) [14] banned as it is a human

Mateusz Konieczny (talk) 18:47, 8 October 2022 (UTC)

Where test cases should be stored in Wikidata

Media is the culprit, as its a concept, neither an action or physical thing. It keeps cropping up in ontology moans. A split into 'storage medium', and understanding that 'communication' is both a noun and a verb would be a start, but I'm no linguist. I expect WD is full of these problems. Vicarage (talk) 08:42, 5 October 2022 (UTC)
Removed media in https://www.wikidata.org/w/index.php?title=Q17480810&diff=1743617461&oldid=1708214464 - without replacement Mateusz Konieczny (talk) 10:47, 5 October 2022 (UTC)
"I expect WD is full of these problems" - and what worse, once fixed they often reappear. Automatic test cases may help detecting reoccurrence of such problems Mateusz Konieczny (talk) 10:47, 5 October 2022 (UTC)
Posted some additional distinct examples to Wikidata talk:WikiProject Ontology Mateusz Konieczny (talk) 08:14, 7 October 2022 (UTC)
@Mateusz Konieczny It's worth thinking about how automated test cases could look like. We could have a property called "test-is-not" that links from "Saxon post milestone" to "action" and than an a page that has an automated list that lists all cases where this is violated. ChristianKl ❪✉❫ 13:01, 7 October 2022 (UTC)
Right now my actual tests are in https://github.com/matkoniecz/wikibrain/blob/master/test_wikidata_structure.py - and they are in form "this is a real physical object at specific location, mappable in OpenStreetMap - not some concept, event, person, species etc". But actual tests would be much better to be in more specific form. Though maybe it would be worth considering attaching them to action (Q4026292) and similar elements? Mateusz Konieczny (talk) 14:34, 7 October 2022 (UTC)
How much test do you think would be good to have for action when we attach the property to action (Q4026292)? ChristianKl ❪✉❫ 20:03, 7 October 2022 (UTC)
How much? @ChristianKl:
When creating an infrastructure for automated tests it's useful to think about how much tests we would ideally have because that's important for the decision whether we want to store the tests at "Saxon post milestone" or at "action". If there would be for example a hundred test for action, storing those tests at action would be bad.
Solving existing problems is harder than prevent new problems from arising. If you have a existing test and someone makes an edit to invalidate the test, it's straightforward to just revert that change. When it comes to solving existing problem we need to think more about where our model should be changed. ChristianKl ❪✉❫ 22:06, 8 October 2022 (UTC)
"If you have a existing test and someone makes an edit to invalidate the test, it's straightforward to just revert that change" - is it even possible to detect specific edit that caused constraint violation in the sublass tree? I see no obvious way to achieve this right now. Though maybe
  1. generate tree
  2. parse history to find dates when each "is subclass of" was added
  3. find newest one
would work well and is not completely horrific? And it would be doable with problems I report here, as nearly all are reoccurrences of the problem. Mateusz Konieczny (talk) 10:27, 9 October 2022 (UTC)
Yes, I would expect to do it that way. I think that with the problem you post here, it's not clear which subclass of (P279) statement is at fault and making the decision is less straightforward than looking at what was newest. ChristianKl ❪✉❫ 13:25, 11 October 2022 (UTC)
After thinking about this: invalid link could be added before missing valid one was added so you need to look at the entire structure anyway Mateusz Konieczny (talk) 21:45, 11 October 2022 (UTC)

changes in search result display

Tracked in Phabricator
Task T306883

There has recently been a fairly radical change to the display of search results and I haven't seen anyone mention it (I'm a dummy who doesn't code or use SPARQL, I just use that box that says "Search Wikidata at the top right of every page). Search results are now accompanied by square thumbnail previews of the item, assuming it has a image (P18) value - the majority have empty white squares looking for all the world like broken links. I can see the images helping somewhat in searches ("John Smith" yields many people with that name, many without images, and how many people known what the most famous John Smith looks like)? Another problem, is that it's much easier to click on the thumbnail image than the item link, which inefficiently brings the user to the Commons image itself, NOT the item they were presumably searching for. A similar change has occurred in search results on Commons (I've raised the issue here). Did any of this get discussed publicly before hand? Has anyone else noticed the changes? Do we like the changes? -Animalparty (talk) 03:55, 7 October 2022 (UTC)

Cosplayers' images now appear in search results for anime characters. It's creepy and I wish it would stop. Afaz (talk) 04:14, 7 October 2022 (UTC)
What is the point of those cosplay images anyway? Is it only because Commons had loads of cosplay on it to begin with? They add nothing to the items in question. —Xezbeth (talk) 10:35, 7 October 2022 (UTC)
It seems to me like this reveals an existing problem, that items get a bad image (P18) value. If an image is creepy when shown next to the item in question, it's likely also going to feel creepy in other use-cases where people depend on good image (P18) values. ChristianKl ❪✉❫ 10:47, 7 October 2022 (UTC)
Those images likely exist because actual images from the anime in question are protected by copyright and thus there's a believe that having the cosplay image is better than not having an image. If we have an anime wikiproject the decision should likely be made there. ChristianKl ❪✉❫ 13:09, 7 October 2022 (UTC)
@Xezbeth: for example Lum Invader (Q2300752).The problem is that the classes are different. A picture of a real human being is shown in the fictitious character item. People will be upset if the "Statue of Liberty" item shows a person in costume instead of a statue. Afaz (talk) 00:47, 11 October 2022 (UTC)
I already marked a bunch of cosplay photos as deprecated with reason for deprecated rank (P2241) does not depict subject of item (Q106626749). And I don't think these are all meaningful images. For example Lum Invader (Q2300752) image was added with https://fist.toolforge.org/wdfist/ - this is basically just a tool that adds the first image from articles as a statement. It is impossible to prevent this on a tool level. It is also useless to delete such statements, because other tool users will just reimport them. The meaningful solution as I see it is to create a new property for cosplay photos (we already have dozens of such properties like image of grave, nighttime view, winter view, etc.) and forcefully set "image" for all anime characters to novalue to prevent tools like wdfist and template harvesters (wdedit/pymediawiki/pltools) from importing cosplayers again. --Lockal (talk) 03:49, 11 October 2022 (UTC)
I noticed maybe half an hour ago, and I'm not sure yet if I like it... At the very least, there should be an option in Preferences to turn it off/on. –FlyingAce✈hello 04:33, 7 October 2022 (UTC)
Clicking on the image should certainly go to the article. Vicarage (talk) 07:10, 7 October 2022 (UTC)
I like the change. To the extend that images show up that are not suitable, I believe that those are generally not suitable for image (P18) and it's good to discover them and remove them.
I agree that clicking on the icon should bring the user to the item.
Instead of the current default I would like that in cases where there's no image for the person, that a truthy small logo or icon (P8972) from its class gets used instead of the current placeholder. https://commons.wikimedia.org/wiki/File:Font_Awesome_5_solid_user.svg looks much nicer than the current placeholder. Icons that work at 16x16 as required by small logo or icon (P8972) generally make good placeholders.
For searches where there are results of multiple different classes but the individual items have no images, seeing the icon for the class makes it easier to know what the item you are searching.
Currently, properties don't show any images. Once, the small logo or icon (P8972) behavior is implemented I think that should change. We at Wikidata can give them proper images. ChristianKl ❪✉❫ 11:20, 7 October 2022 (UTC)
The thumbnail can be hidden via CSS (.searchResultImage-thumbnail {display: none;}). This is only a hack.--GZWDer (talk) 11:32, 7 October 2022 (UTC)
This appears to be related to phab:T245673, but that was resolved two years ago, so presumably someone did something more recent to enable it. Bovlb (talk) 16:27, 7 October 2022 (UTC)
  • But does anyone know where these large-scale change came from? Who made them? A cabal of coders on MediaWiki? An initiative at MetaWiki or the WMF? It appears some powers-that-be, high above lowly day-to-day users and volunteers, decided to make this change. While there are probably benefits as well as drawbacks, I don't like things happening in mysterious echelons with no accountability. -Animalparty (talk) 17:23, 7 October 2022 (UTC)
    It all comes from this phab:T306883 ticket. Nthep (talk) 11:44, 8 October 2022 (UTC)
  • Some CSS to remove the useless placeholder when there is no image and to compact display:
.searchResultImage-thumbnail-placeholder { display: none; } .searchResultImage .searchResultImage-thumbnail { height: auto; } .mw-search-results li { padding-bottom: 12px; }
Cheers, — Envlh (talk) 17:30, 7 October 2022 (UTC)
  • Looking for a way to restore display the complete thumbnails like they should on Commons special search (500 items list). They look very much truncated and 20% don't load at all. My .common.js does not accept the above css, so I can't test it. Please fix this set of obvious display bugs and repair special search on Commons to provide a usable tool again for massive list image reviewing. Pelikana (talk) 11:42, 9 October 2022 (UTC)
    Common.css Bovlb (talk) 14:36, 9 October 2022 (UTC)
  • Is this the first time you all notice such things happen? In my 10+ years experience in RuWiki, Commons and WD, such changes that are sent down to projects without proper notice and/or discussion occur quite regularly. Michgrig (talk) 10:11, 8 October 2022 (UTC)
    Usually, development that creates changes in Wikidata gets listed in the Weekly Summary. ChristianKl ❪✉❫ 22:31, 8 October 2022 (UTC)
Hi everyone, I'm the Community Relations Specialist for the Structured Data Across Wikimedia project, which also is responsible for the deployment of this new feature of article thumbnails. I'm deeply sorry for the lack of proper announcement about the deployment of this feature, and the least I can promise is that this will not happen again for next deployments.
Also, I'm collecting your feedback about bugs and problems you're having with this feature, and I already shared them with the development team. I will keep you posted about them. Meanwhile, feel free to ping me and let me know any problem you're noticing. Thanks a lot and again sorry for the disruption. --Sannita (WMF) (talk) 11:52, 10 October 2022 (UTC)
The feedback from me is simple: it should be possible to hide it completely in the Preferences. I can't see such option and I have to hide it using CSS, after I spent several minutes looking for answers who the hell enabled it by default without any prior notification. That's not how it should work. Wostr (talk) 18:14, 10 October 2022 (UTC)
Hi @Wostr, this has already been filed as a Phabricator ticket (see phab:T320337), and we're looking into this as I speak. I will let you all know as soon as investigation phase is done. Sannita (WMF) (talk) 08:35, 11 October 2022 (UTC)
@Sannita (WMF) Instead of a setting in preferences, I think it might make more sense to show the setting on the search page. That makes it easier to discover and also easier to change if the user wants it to behave differently for some time. ChristianKl ❪✉❫ 13:11, 11 October 2022 (UTC)
@ChristianKl It's one of the proposals made in the ticket, actually. I can't promise anything about it, since the jury dev team is still out on this. Sannita (WMF) (talk) 15:43, 11 October 2022 (UTC)

Help needed to close two requests for comments

Hello

This Wikidata:Requests_for_comment/Gender_neutral_labels_for_occupations_and_positions_in_French request for comments have been closed in August and reopened by a contributor. I think that there is no more contribution. If any uninvolved contributor could close the request for comments, it would be greatly appreciated.

We have exactly the same situation with this one Wikidata:Requests for comment/How to avoid to use male form as a generic form in property labels in French?. PAC2 (talk) 17:47, 8 October 2022 (UTC)

I do think, that RfC should be closed by admins and not just by any contributor who feels that the RfC should be closed. Therefore the admin noticeboard is also the best place to seek someone to close an RfC. ChristianKl ❪✉❫ 21:56, 8 October 2022 (UTC)
Per WD:RFC: "Any uninvolved editor can close a request for comment, usually after about 30 days."
We do not need an admin here. French language skills are more helpful in this particular case. —MisterSynergy (talk) 22:00, 8 October 2022 (UTC)
Yes, any uninvolved contributor can do it. PAC2 (talk) 17:39, 10 October 2022 (UTC)
In the Wikimedia movement, you usually discuss rather than vote. The two "requests for comments" have worked as votes with no prior discussion. GrandEscogriffe (talk) 17:50, 11 October 2022 (UTC)

More specific instances?

Hi everyone. I want to review the state of the municipalities in Spain, while creating their mayors, town halls, etc. When i do the first query it doesn't return all of them because some are in municipality of Catalonia (Q33146843), municipality of Aragon (Q61763947) or similar. In the case of other countries this circumstance does not exist, it will always be commune of France (Q484170) in the case of France. Do I have to change all the municipalities so that they are municipality of Spain (Q2074737) or do I have to put in those that are from Catalonia and Aragón all those that compose it? Because right now some have municipality of Spain (Q2074737) and others municipality of Aragon (Q61763947). The same case occurs to me with other cases such as Catalonia school (Q111236457) where terms from a certain area of ​​the country are being specified. I would like to talk about it and make a decision for the future. Thanks. Vanbasten 23 (talk) 21:02, 10 October 2022 (UTC)

Probably you want ?municipio_de_España wdt:P31/wdt:P279* wd:Q2074737. in your code. municipality of Aragon (Q61763947) is a subclass of municipality of Spain (Q2074737). --Tagishsimon (talk) 21:35, 10 October 2022 (UTC)
@Tagishsimon: i know how do it, but it does not make much sense that several elements have a municipality in Aragon and other municipalities in Spain, when they all belong to the same site, right? --Vanbasten 23 (talk) 07:21, 11 October 2022 (UTC)
It would be ideal if items were modelled in a consistent fashion; the normal course on WD is to employ the more granular value - e.g. municipality of Aragon (Q61763947) - rather than the class value municipality of Spain (Q2074737). --Tagishsimon (talk) 07:55, 11 October 2022 (UTC)
Wikidata has different granuality at different points. Wikidata doesn't have a pro-Nationalism policy that would make us prefer to reference Spain over Catalonia. ChristianKl ❪✉❫ 14:53, 11 October 2022 (UTC)

What exactly is Wikidata about

?? AkJackster (talk) 17:01, 11 October 2022 (UTC)

It's a database of structured data; see Wikidata:Introduction, w:en:Knowledge graph and w:en:Resource Description Framework. It is about making that structured data available for use on wikipedias, and more generally by all-comers. --Tagishsimon (talk) 17:06, 11 October 2022 (UTC)

Roman patrician as an occupation (P106)

Some 555 items have an occupation (P106) of patricii (Q154668). Is this intended? Intuitively, I wouldn’t consider patricians an occupation (and the item has instance of (P31) social category (Q2305862), population group (Q851990), social class in ancient Rome (Q1392538), although at the moment I fail to comprehend whether this is compatible with the value-type constraint (Q21510865) of occupation (P106)). If not, what would be the correct property for declaring someone a patrician? --2A02:8108:50BF:C694:384C:CEE:13AD:CD09 08:36, 12 October 2022 (UTC)

Similar issue to this recent thread on slaveholders. Conclusion was, social classification (P3716) a better choice of property than occupation (P106). --Tagishsimon (talk) 09:11, 12 October 2022 (UTC)
Okay, thanks. Then I’ll go for that. --2A02:8108:50BF:C694:384C:CEE:13AD:CD09 09:42, 12 October 2022 (UTC)

Merge billiard table and carom table?

Should Q3500725 billiard table and Q862731 carom table be merged? I tried to do it manually, but I'm getting errors that it is impossible, even after I migrated all interwikis I could find. This hasn't happened to me before, so I ended up here asking for help. I'm not an expert on the topic, but strictly speaking, carom tables are probably best regarded as a subcategory of billard tables. They are categorised this way on Commons, where c:Category:Carom_tables is in c:Category:Billiard tables. As far as I can tell, however, almost all Wikipedia language versions have a general article on billiard tables that describe all types of tables, including carom tables; these are connected to the Q3500725 billiard table item. The only exception is German Wikipedia, where de:Billardtisch is a disamb page referring to three separate articles: de:Billardtisch (Karambolage), de:Billardtisch (Pool) and de:Snookertisch. The article de:Billardtisch (Karambolage) was linked to Q862731 carom table when I just found it, together with Dutch Wikipedia's nl:Biljarttafel. However, despite the first sentence of the latter Dutch article stating 'Een biljarttafel voor carambolebiljart is rechthoekig van vorm en tweemaal zo lang als breed', thus associating it specifically with carom billiards games, it has no such specification in its title, and there are no other Dutch Wikipedia articles on billiard tables in general or on specific types. What does Wikidata do in such cases? Does it link the German disamb page de:Billardtisch with Q3500725 billiard table, or not? Does it connect nl:Biljarttafel with Q3500725 billiard table, or not? In my opinion, connecting nl:Biljarttafel with Q3500725 billiard table is the easiest solution; the title allows for it to be expanded with information on all other types of non-carom billard tables, and this harmonises the distribution of information across Wikipedia languages versions. But I wanna make sure I do things right and do not create a mess by disregarding the rules. I don't particularly care about this subject, but I hope to use this case to better understand how Wikidata works. Cheers, Nederlandse Leeuw (talk) 15:18, 12 October 2022 (UTC)

Hi Nederlandse Leeuw, the merge probably fails as carom table (Q862731) is declared a subclass of billiard table (Q3500725) [[15]]. This conforms with your statement "carom tables are probably best regarded as a subcategory of billard tables". If carom tables are just a subcategory of billard tables (besides pool and snooker) the items should not be merged. I have no opinion about the Dutch Wikipedia link (if it should link to billiard table (Q3500725) or to carom table (Q862731)) but I think the German article should stay at carom table (Q862731) as it is specifically about the carom table - snooker and pool have their own pages in this Wikipedia edition. - Valentina.Anitnelav (talk) 15:32, 12 October 2022 (UTC)
That makes sense, thanks for explaining it. I'll try to fix it, but if it doesn't work I'll come back here. Cheers, Nederlandse Leeuw (talk) 15:39, 12 October 2022 (UTC)
Ah, I see you already fixed it, thanks! Nederlandse Leeuw (talk) 15:41, 12 October 2022 (UTC)

duplicate painting "Campo Sant'Angelo, Venice"

Q78640027 and Q41597104 are duplicate.

We should keep Q78640027 and remove Q41597104 as the first one contains more information. Especially the location : the painting is in the MET as stated in Q78640027, but not in a private collection as says Q41597104. Laurent.Claessens (talk) 13:28, 11 October 2022 (UTC)

@Laurent.Claessens, have you tried to Merge the items? Michgrig (talk) 18:08, 11 October 2022 (UTC)
Also, paintings can be in different collections at different points in time, which should be recorded by multiple time-qualified collection (P195) values on a single item. The Met provenance record indicates it was owned by Mr. and Mrs. Charles Wrightsman, New York, until 2019. This is presumably the private collection mentioned in the 1989 Canaletto (Q41591006) catalog. -Animalparty (talk) 20:18, 11 October 2022 (UTC)
I checked, both are listed in The Complete Paintings of Canaletto. as 122. Doesn't seem to have an item here yet, just c:Canaletto catalogue raisonné, 1968 on Commons. Merged the two. Multichill (talk) 20:15, 13 October 2022 (UTC)

Start date for locations, what to do about error flags

See Henry K. Lattin (Q20668053) for example, where we get an error flag because the death date of a person is prior to the official start date of locations. Generally the start date is for the current form of government. A village gets reclassified as a town, later a town becomes a city, and the most current date is used. Should we move the birth/death location to the next biggest entity, like the county? Or just ignore the error flags? Or adjust the start/inception dates earlier to when the location was founded/settled? For this person we have obituaries saying he died in Farmingdale, and the entry for the location at Wikipedia has "Postal Address (1845)" and Village (1904). Should we have multiple start dates with some sort of qualifier, and have the earliest preferred_rank, to get rid of the error flags? I am sure others have encountered this problem. RAN (talk) 18:52, 13 October 2022 (UTC)

In order, 1) ignore the flag 2) question what the uncited 1904 date on the settlement means; does it need qualifying, is there a need for another item, am I pointing at the right thing &c&c. Multiple start dates might be a solution, although I'd be loth to rank them b/c it's not clear which if any should be preferred. Distinctly uninclined to use a less granular value because the more granular item has a dubious start date. It's taking on a lot to have to improve, correct or clarify a settlement item when one's mission is a person item, not least because it's unlikely you'll have easy access to sources. --Tagishsimon (talk) 19:18, 13 October 2022 (UTC)

Bogus “No match was found” when adding P7561

I try to fill out category for the interior of the item (P7561) for St Peter's Church of Ireland, Drogheda (Q55028957) with “Interior of St Peter's Church of Ireland Church, Drogheda” but this fails with “No match was found” with no way to add this link. How to fix this? --AFBorchert (talk) 20:56, 13 October 2022 (UTC)

category for the interior of the item (P7561) expects another item as value, not a string. --Ameisenigel (talk) 21:20, 13 October 2022 (UTC)

Should we merge Q21859769 and Q1824337?

Reiche Liesing (Q21859769) https://www.wikidata.org/wiki/Q21859769

Liesing (Q1824337) https://www.wikidata.org/wiki/Q1824337

As far as I can tell, these 2 items refer to the same river. But I can't merge them (I think because they each have a different commons category linked). Can someone help? Ficaia (talk) 06:01, 14 October 2022 (UTC)

To my eye, WD, Commons and OSM tell the same story, that Reiche Liesing and Dürre Liesing meet to form Liesing, which eventually falls into Schwechat. Check out for instance, Q1824337#P885 and Q21859769#P403, or the coord of 48°8'15.18"N, 16°15'47.34"E which is the source coord for one of them, and the mouth coord for the other. --Tagishsimon (talk) 07:15, 14 October 2022 (UTC)
Yes, although for the government agency Umweltbundesamt (see this dataset) the part from Hochroterd (Q113995446) to Pfarrkirche Rodaun (Q1678819) is also called “Liesing (Liesingbach)“ and not “Reiche Liesing”. But I agree, there is no reason to merge, this just seems to be a terminology problem. --Emu (talk) 09:46, 14 October 2022 (UTC)

delete Q114659499

I accidently started is and later noticed that Q21317910 aleady exists Kersti Nebelsiek (talk) 09:04, 14 October 2022 (UTC)

merged --Luckyz (talk) 11:26, 14 October 2022 (UTC)

Book authors and editors

Many books have editors specified instead of authors. How to specify editors as plain text? Currently I use author name string (P2093) property (for example in Seldin and Giebisch's The Kidney: Physiology and pathophysiology (5th Edition) (Q114595502)), but this way I cannot distinguish editors from authors (separate sections of books can have their own authors, and editors compose those sections within a book as I understand). D6194c-1cc (talk) 10:16, 14 October 2022 (UTC)

unknown value Help + object stated in reference as (P5997) can be used with the editor property. ChristianKl ❪✉❫ 16:21, 14 October 2022 (UTC)
Thanks! I'll try it out but with object named as (P1932) property. I think it is more relevant. D6194c-1cc (talk) 20:43, 14 October 2022 (UTC)

Question about what genre to use for a TV show

Hi, I'm really doubtful about what value should I use in the statement "genre" regarding TV shows that only show movies. Kacamata (talk) 16:56, 14 October 2022 (UTC)

I think you need to be more specific if you want a response. "TV shows" and "movies" are normally not the same things. "TV shows" do not normally, in my experience, "show movies". TV channels show TV shows and, sometimes, movies. --Tagishsimon (talk) 17:23, 14 October 2022 (UTC)
@Tagishsimon In Brazil, we have several television programs that are simply movie sessions. They simply show movies. Like Q97378653, Q10370683 or even Q18243965. I'm working to improve these items here, but I'm stuck in the "genre" thing. Kacamata (talk) 19:25, 14 October 2022 (UTC)
you can make a new genre item for this if you can't find one that already exists. then add it to other similar shows. BrokenSegue (talk) 23:22, 14 October 2022 (UTC)

Book editions

Can somebody explain what is the common practice about book editions in Wikidata? Some books have different editions with different editors, ISBNs, and other codes (like Google Books code). I have found that ISBN numbers have some qualifiers to describe an edition. But ISBN-13 (P212) has a very limited number of allowed qualifiers, and also there is another ISBN-10 (P957). Should I create different Q-entity for an edition? How to name it? Should I just add edition to the name? D6194c-1cc (talk) 21:13, 11 October 2022 (UTC)

See, especially, Wikidata:WikiProject Books. Aim is for one item covering the literary work, and one item per edition. --Tagishsimon (talk) 21:47, 11 October 2022 (UTC)
This aim might make sense in theory and robot logic, but in practice it is often impractical and confusing for humans. A "fully Wikidata-fied" work like Frankenstein; or, The Modern Prometheus (Q150827) may have thousands of editions, translations, reprints, and versions. Text books and other reference books may have several to dozens of editions (some with miniscule differences but different identifiers). Should someone wish to use a book to cite a claim on Wikidata, they may need to sort through dozens of nearly identical items to find the appropriate item (e.g. maybe only volume 2 of the 10th (deluxe) edition of an encyclopedia contains the birthdate of a person). As Wikidata become ever larger and more data-dense, actually finding meaningful data and items may become increasingly difficult. I realize this comment strays more into the futurist side of Wikidata than the immediate question of the OP, but it's an issue that I think will become ever more pressing. -Animalparty (talk) 00:03, 12 October 2022 (UTC)
I submit that being practical and non-confusing for humans isn't, and shouldn't be, the primary aim of Wikidata. Wikipedia is the general knowledge project aimed at direct human consumption. On the other hand, AFAIK, it was always the goal of Wikidata to be machine-readable. Thus, the structure of knowledge on Wikidata should reflect the structure of the real-world things which it describes, however complex it is. If that leads to Wikidata sometimes being impractical to be used by humans directly, then humans should use tools to make that job easier. Silver hr (talk) 00:53, 12 October 2022 (UTC)
And indeed managing references is a task very suitable for a tool, which could handle the search for existing, and creation of new, items transparently in the background. --2A02:8108:50BF:C694:384C:CEE:13AD:CD09 08:40, 12 October 2022 (UTC)
I'm bound to ask @Animalparty: how else would you structure the data, if we take as fact that there are thousands of editions of Frankenstein; or, The Modern Prometheus (Q150827)? Options seem to be a) a single work item, no edition items, which frustrates anyone interested in the editions b) work item and edition items, which seems a fairly ideal way of representing the real world situation c) a single item holding data concerned with the work and its editions, which seems impractical. I'm not sure what the referencing ask/issue here is: you seem to be suggesting that in order to make it easier to cite volume 2 of the 10th (deluxe) edition, it would be better if WD did not have an item for volume 2 of the 10th (deluxe) edition? --Tagishsimon (talk) 09:07, 12 October 2022 (UTC)
Thanks for the explanation. I've followed your advise and created 5 entities about Donald W. Seldin and Gerhard H. Giebisch The Kidney: Physiology and pathophysiology 5 editions (The Kidney: Physiology and pathophysiology (1st Edition) (Q114616452), The Kidney: Physiology and pathophysiology (2nd Edition) (Q114656525), etc.). Since each next edition has some added information, those books are different with different contributors. D6194c-1cc (talk) 09:13, 14 October 2022 (UTC)
But I doubt about The Demon-Haunted World (Q2482106). It has many editions and translations into other languages. Do they need separate entities? Those editions and translations seem to be the same book. Currently editions are covered by ISBN section. If I create separate entity for a translation, how do I specify in the main entity list of different editions? D6194c-1cc (talk) 09:19, 14 October 2022 (UTC)
Use has edition or translation (P747) on the original and edition or translation of (P629) on the derived edition (as listed in the table on Wikidata:WikiProject Books#Edition item properties) --M2Ys4U (talk) 17:20, 14 October 2022 (UTC)
Thanks for the help! D6194c-1cc (talk) 07:53, 15 October 2022 (UTC)

Should central bank require currency property?

I believe central bank should require a currency property, and that the currency property should be allowed on entities that are an instance of central bank.

Currency, a currency can have a central bank/issuer and I think an inverse property such as this should exist bank of somaliland currently is using this property but the reserve bank of australia is not.

Constraints require currency to only be applied to a fictional country or administrative region - so the bank of Somaliland currently has a constraint violation. Currently there is nothing linking a central bank to the currency it issues, and I think it makes sense to use the property currency for this. I am going to be bold and take the initiative to do this DominusVilicus (talk) 04:12, 12 October 2022 (UTC)

Sounds appropriate. People can look up the country of the central bank, and from that infer its currency, but i suspect there are exceptions to this. Again, I'm too lazy to actually check. On the topic of reverse properties, these are intentionally avoided most of the time as they lead to redundant data which will inevitably have synchronization issues. Infrastruktur (talk) 08:04, 12 October 2022 (UTC)
In Wikidata, constraints are something that properties have. central bank (Q66344) is an item and not a property and thus can't have constraints. properties for this type (P1963) is the more appropriate way to make that central banks usually do have currencies. ChristianKl ❪✉❫ 16:07, 14 October 2022 (UTC)
I would just like to note that central bank (P1304) is also a property, used in instances of countries generally DominusVilicus (talk) 07:54, 15 October 2022 (UTC)

Think big! Open letter about Wikimedia Commons

Dear friends of free knowledge,

Wikimedia Commons is in crisis. There are numerous concerns and complaints about our central media platform, for many years.

Therefore, this open letter asks the Wikimedia Foundation to Think big! about the future of Wikimedia Commons.

In late August 2022, we at the Commons Photographers User Group talked about Wikimedia Commons. The result of these and other talks is this open letter.

We invite everyone to sign this open letter to show how important Wikimedia Commons is to you. You may be a regular Commons contributor, a Wikidatan, an editor of Wiktionary or Wikivoyage, or maybe you represent an affiliation. We also strongly invite other people who are involved with Commons directly or indirectly, maybe in the context of a GLAM.

Please inform others about this open letter.

Kind regards, Ziko (talk) 19:53, 14 October 2022 (UTC)

Wikimedia Commons is not in crisis. It's fine to wish Commons to be different, not fine to wave hyperbolic shrouds around. --Tagishsimon (talk) 13:42, 15 October 2022 (UTC)

Archives at (P485) appropriate use

I'm wondering if it is appropriate to use the 'Archives at' property (P485) when an institution holds only a portion of an individual's archives. Our library collection includes correspondence from numerous individuals, but for many of the correspondents, the majority of their archives are held at other institutions. We'd like to use Wikidata to indicate that we hold a set of correspondence, but I'm uncertain if P485 is the correct property to use. One example: we have in our collection a set of correspondence from Spencer Fullerton Baird to William Brewster, an ornithologist. The majority of Baird's archives are at the Smithsonian NMNH. Is it appropriate for our institution (Harvard Museum of Comparative Zoology) to add a P485 statement to Baird's Wikidata entry to indicate what we have? Thanks in advance for any insight. Jmdeveer (talk) 21:50, 14 October 2022 (UTC)

The property can have multiple values, so if you have a substantial partial collection, go ahead Vicarage (talk) 04:35, 15 October 2022 (UTC)
applies to part (P518) and collection or exhibition size (P1436) can also be useful qualifiers. Jheald (talk) 14:43, 15 October 2022 (UTC)

New Wikidata game matching new Wikipedia articles and categories

Hi all. I've set up a new Wikidata game that matches new Wikipedia articles and categories to Wikidata items. This is to try to reduce duplicate items created by Pi bot. Please give it a go! Thanks. Mike Peel (talk) 11:54, 15 October 2022 (UTC)

Merge Ben Bolt (Q59291518) and Ben Bolt (Q56407662)

Hi all, I'm not having much luck with merging these two items. They both refer to the same poem / song but the former has Wikisource associated with it and the other has Wikipedia -- can't figure out how to get the "merge two items" tool to work. Would someone kindly help me merge them? RexSueciae (talk) 13:14, 15 October 2022 (UTC)

They're distinct things; a poem, and a ballard derived from the poem. They will not merge b/c both point at eachother - Q59291518#P4969 & Q56407662#p144. --Tagishsimon (talk) 13:40, 15 October 2022 (UTC)
Oh, I see now. It doesn't make much sense to distinguish them -- at least, I feel like Wikisource should be connected to the latter because that's the one that's currently linked to Wikipedia. But that's a simple enough fix, hang on a second. RexSueciae (talk) 14:56, 15 October 2022 (UTC)

occupation (P106)deity (Q178885)

Currently there seem to be seven items with occupation (P106)deity (Q178885). While this in itself already seems dubious, when I have a look at the first entry the query returned (Juliana Paes (Q466954)), I feel inclined to assume some kind of hoax or at least grave error:

…on an item claiming to deal with a female (Q6581072) human (Q5) born 1979 speaking Portuguese (Q5146); description: Brazilian actress. Asking for a second pair of eyes here. --2A02:8108:50BF:C694:882C:E109:BAF0:30CE 13:48, 15 October 2022 (UTC)

It's fairly standard vandalism, from February; hopefully now reverted. --Tagishsimon (talk) 13:56, 15 October 2022 (UTC)
OK. I’ll have a look at the others; deity (Q178885) is probably not an occupation. --2A02:8108:50BF:C694:882C:E109:BAF0:30CE 14:07, 15 October 2022 (UTC)

“Faculty” an occupation?

Bill Thomas (Q132302)occupation (P106)faculty (Q180958) doesn’t make much sense, but there is a reference for that claim: http://bioguide.congress.gov/scripts/biodisplay.pl?index=T000188 – which provides the highly incomplete sentence “faculty, Bakersfield Community College, Bakersfield, Calif., 1965-1974”, from which the statement is probably derived. I’m not sure about this word usage – does that mean he was a university teacher (Q1622272) or (just) a faculty member (Q5428874)? He probably wasn’t the entire faculty (Q180958), which isn’t an occupation in any case. --2A02:8108:50BF:C694:882C:E109:BAF0:30CE 15:02, 15 October 2022 (UTC)

Seems like this data would be more appropriate to employer (P108). --Tagishsimon (talk) 15:06, 15 October 2022 (UTC)
Yes, possibly. What I’m also unsure about, though, is the organization directed by the office or position (P2389)Bakersfield College (Q4849288) qualifier. Bakersfield College (Q4849288) should probably be the employer (P108) (wherever faculty (Q180958) would go then). --2A02:8108:50BF:C694:882C:E109:BAF0:30CE 15:16, 15 October 2022 (UTC)
Q132302#P108. --Tagishsimon (talk) 15:21, 15 October 2022 (UTC)

Same item with different code

Hi all. I've tried to add the portuguese Wikipedia article tittled "Abobadilha" (Q9561041) to the proper concept wikidata page (Q21573989) but I can't do it. Can anybody fix it? Thanks. Jabalcón (talk) 15:26, 15 October 2022 (UTC)

In this case a merge seems the simplest route. Had the sitelink been on an inappropriate item, then removing it first, and adding it second, would work. Now all at tile vault (Q9561041) --Tagishsimon (talk) 15:32, 15 October 2022 (UTC)

Horn-satisfiability vs. Markierungsalgorithmus

Horn-satisfiability (Q1239194) “is the problem of deciding whether a given set of propositional Horn clauses is satisfiable or not” (en:Horn-satisfiability). The German sitelink de:Markierungsalgorithmus, however, is an article about a specific algorithm for solving this problem. I would expect a computational problem and the algorithm(s) for solving it to be separate entities, hence I propose splitting Horn-satisfiability (Q1239194) into two items, one for the problem and the other for the algorithm.

Thanks in advance for your suggestions! --2A02:8108:50BF:C694:882C:E109:BAF0:30CE 18:52, 14 October 2022 (UTC)

splitting them makes sense. only suggestion for a property would be facet of (P1269). BrokenSegue (talk) 19:48, 14 October 2022 (UTC)
For the records: Bellman–Ford algorithm (Q816022)computes solution to (P2159)shortest path problem (Q1058754). I haven’t found an inverse property, though (cycle detection (Q60614789)solved by (P1136)Floyd Cycle Detection Algorithm (Q1588200), but solved by (P1136) states to be for people who solved a problem, not algorithms; I’ll change that statement soon). --2A02:8108:50BF:C694:146A:852:36CA:CE6E 20:17, 16 October 2022 (UTC)
The item has now been split: Horn-satisfiability (Q1239194) and Q114708728, with Q114708728computes solution to (P2159)Horn-satisfiability (Q1239194). I’ve requested an inverse property for that. --2A02:8108:50BF:C694:146A:852:36CA:CE6E 20:45, 16 October 2022 (UTC)
PS. By the way, does anyone know what that algorithm is called in English? It’s the algorithm described here, but I doubt that “Greedy-Horn” is the common name. --2A02:8108:50BF:C694:146A:852:36CA:CE6E 20:51, 16 October 2022 (UTC)

Protected page edit request

Please add id:Bata jemuran to adobe (Q183496). Thanks in advance V-Aids (talk) 10:04, 15 October 2022 (UTC)

✓ Done -wd-Ryan (Talk/Edits) 03:29, 16 October 2022 (UTC)

Cannot find statement

SELECT ?item ?problem
WHERE
{
  ?item wdt:P2159 ?problem.
  ?problem wdt:P1343 ?item.
}
Try it!

claims there is a statement End Software Patents (Q109381800)computes solution to (P2159)software patent debate (Q2996086), but I cannot find that statement in End Software Patents (Q109381800)… where is it? It’s probably questionable (the End Software Patents project does, probably, not “compute” a solution to the software patent debate). --2A02:8108:50BF:C694:146A:852:36CA:CE6E 20:38, 16 October 2022 (UTC)

It has been removed, but for whatever reason, that removal was not reflected on WDQS. I've made a trivial edit, which has updated WDQS properly. --Tagishsimon (talk) 02:35, 17 October 2022 (UTC)

Should Currency require ISO 4217 code?

Currently, using the currency property requires the value to have an ISO 4217 code. An example of this might be in used in El Salvador for the currency of the country, however this presents a problem as not every currency has a ISO 4217 code. Since El Salvador uses bitcoin as their national currency, this creates a constraint violation

This same violation might exist across historical, fictional and proposed countries and entities.

How should this be addressed? I don't think it makes sense for every use of currency as a property to require the currency value to have an ISO 4217 code. Thoughts? DominusVilicus (talk) 13:45, 11 October 2022 (UTC)

I think we should either remove the constraint or add exceptions to the constraint BrokenSegue (talk) 15:58, 11 October 2022 (UTC)
I am going to assume most of the 723 constraint violations are false positives (historic currencies etc.), which speaks for removing the constraint. Infrastruktur (talk) 16:15, 11 October 2022 (UTC)
I think exceptions should be added if there is a finite and manageable number of them. That, however, depends on the domain of currency (P38). If the domain is countries, then the number of exceptions would be manageable. If, however, the domain includes fictional countries or entities such as exchanges, then anything, including the ever-growing list of cryptocurrencies, can be a currency, and I assume most of those will not have an ISO 4217 code, thus making the exception list infinite and unmanageable. Silver hr (talk) 00:40, 12 October 2022 (UTC)
The same applies to Central Bank/Issuer. So I think this constraint should be removed from both of them. Not every currency will have such a code, and it does not make sense to force a statement for it. Neither does it make sense to add an exception for every historical, proposed or fictional currency, country or bank.
Is there anything we can do to remove it as a constraint, but suggest it as a recommended property? @BrokenSegue @Infrastruktur @Silver hr DominusVilicus (talk) 03:15, 12 October 2022 (UTC)
@Silver hr: I took your advice and had a proper look at where it was used. https://w.wiki/5ojN . Turns out it was used mostly on historic countries, and there are too many cases to be covered by constraint exceptions. Infrastruktur (talk) 07:19, 12 October 2022 (UTC)
By saying there are too many cases to be covered by constraint exceptions, do you mean that there would be a technical or performance problem or that you wouldn't be willing to do the work of adding them?
This discussion touches upon the very definition of currency (P38). If the constraint is to be useful in warning about range errors for the property, the domain should also be constrained to countries (and the few other entities ISO 4217 covers). And if that would be the case, I don't see this finite number of exceptions as unmanageable. Otherwise, if the domain isn't constrained to countries, the ISO 4217 constraint is useless as well. Silver hr (talk) 07:56, 12 October 2022 (UTC)
I have no idea about the performance. By the looks of it ISO 4217 has a table for historic currencies but this is limited to relatively recent ones, which implies most of the 660 historic countries will have to be added as constraint exceptions. And while you could add them in bulk with a tool like Wikibase-CLI, I suspect people will be unwilling to manually maintain such a big list. Infrastruktur (talk) 08:27, 12 October 2022 (UTC)
I don't see what maintenance would be necessary; what is historic cannot change, so once it's put in, it's done. Silver hr (talk) 09:13, 12 October 2022 (UTC)
eh, I think maintaining a list so large would be cumbersome, especially since exceptions are supposed to be, well, exceptions, however in this case they historical states, and proposed states constitute a large sum of the total.
I think it's obvious that not every currency requires an ISO 4217 code (historical, digital, cryptocurrencies, fictional, etc), so therefore I believe it should be removed DominusVilicus (talk) 13:43, 12 October 2022 (UTC)
As I've asked already, what maintenance? Silver hr (talk) 03:00, 13 October 2022 (UTC)
I have removed ISO 4217 as a constraint, and instead have opted to use properties for type on the wikidata item, to suggest this on instances of currency instead of the property because it makes the most sense. DominusVilicus (talk) 13:23, 17 October 2022 (UTC)
Sorry I'm late to the party. But shouldn't historic currencies have "end date" and maybe a special class "Historic Currency"? Then we don't need to remove the constraint, just apply it to contemporary (hm, "current") currecnies Vladimir Alexiev (talk) 15:02, 17 October 2022 (UTC)

Post stamps, money - which property for end of validity?

I am working on list of czechoslovak post stamps. Every stamp have publication date (P577) and also date, when they were withdrawn from circulation (maybe significant event (P793)) and also end of validity - after this date they cannot be used. use of significant event (P793) for this is problematic because of using qualifier. I found service retirement (P730) and discontinued date (P2669) but according name and description they are not the right one - they still exist, but are not valid. JAn Dudík (talk) 10:28, 15 October 2022 (UTC)

You could create an analog of circulation (Q5121688) for postage stamps (and a superclass for both), and then use end time (P582)Xrelative to (P2210)postage circulation? and end time (P582)Xrelative to (P2210)postage validity (Q13637596). Silver hr (talk) 13:56, 17 October 2022 (UTC)
Use of significant event (P793) does not seem to be problematic. Significant event: end of validity, qualifier point in time. --Tagishsimon (talk) 13:59, 17 October 2022 (UTC)

Fandom articles to Wikimedia special items

I haven't gotten a response on Property_talk:P6262, so I'll ask here.

For Fandom article ID (P6262), should we or should we not add IDs to list articles, disambiguation pages, and categories? Examples:

I think we should because it would be similar to adding interwiki links to connect the sites. It would allow us to connect more from Wikipedia to various Fandom wikis.

Another possibility could be, for example, minecraft:Category:Blocks on block (Q91074594) with qualifier object has role (P3831)—>Wikimedia category (Q4167836).

​ -wd-Ryan (Talk/Edits) 20:48, 16 October 2022 (UTC)

This can also apply to other wiki sites like Liquipedia ID (P10918). -wd-Ryan (Talk/Edits) 02:27, 17 October 2022 (UTC)
list articles makes sense to me. disambiguation pages are a little weird but I suppose there's no harm. BrokenSegue (talk) 17:35, 17 October 2022 (UTC)

standards importer (ISO and maybe others)

Wikidata:Bot_requests#standards importer (ISO and maybe others) (2022-10-16): please share what you think Vladimir Alexiev (talk) 15:11, 17 October 2022 (UTC)

Wikidata weekly summary #542

Here's your quick overview of what has been happening around Wikidata over the last week.
  • 🧑🏿‍💻 Development
    • REST API:
      • Finished work on handling conditional HTTP request headers (If-Match, If-None-Match, etc)
      • Working on providing more clear error reports to PATCH route errors
    • Mismatch Finder: Implementing a new random mismatch feature (phab:T302289)
    • The text in the namespace tabs for Lexeme: and EntitySchema: pages can now be easily translated (phab:T316770) - thanks Amir!
    • Made the Wikibase error message "Malformed input" more meaningful (phab:T304943)
    • Continued working on the deployment of a patch to improve the Query Service lag integration in MediaWiki API's maxlag (phab:T315423)
    • Fixed a bug where the Kartographer map preview is erroneously closed, when clicking on zoom actions (+/-) (phab:T303103)
    • Still working on fixing a mobile termbox issue related to the "mul" language code (phab:T318137)
    • Added ircs as an allowed protocol for the URL datatype (phab:T320643) - thanks Lectrician1!

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.

editing helper

I could imagine having a JS script installed that pops up on a key and presents a menu with a few Q-items that fill the currently opened input field would be quite helpful. Has someone already written such a script? SCIdude (talk) 16:03, 17 October 2022 (UTC)

how would it decide what few Q items to show? BrokenSegue (talk) 17:36, 17 October 2022 (UTC)
These would be completely determined by editing of the script. Nothing fancy is needed. --SCIdude (talk) 07:22, 19 October 2022 (UTC)
But of course, if you want to add some design, the script could read the list of Qs from a specific user page. Anyway, I personally don't need that. SCIdude (talk) 07:27, 19 October 2022 (UTC)
Maybe some own modification of WUS might be what you are looking for. JAn Dudík (talk) 12:12, 19 October 2022 (UTC)
Interesting. I can see how it can be tweaked for my purpose. Thank you. SCIdude (talk) 17:49, 19 October 2022 (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. SCIdude (talk) 08:30, 23 October 2022 (UTC)

https://www.wikidata.org/wiki/Q109108991

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.

Also requested at Wikidata:Administrators' noticeboard#Entity_approval — Martin (MSGJ · talk) 20:06, 18 October 2022 (UTC)

Hi, I would like to restore this entity, as it is listed on VIAF database. http://viaf.org/viaf/34164419218120232996 2405:204:3312:9941:0:0:E16:18A5 15:30, 18 October 2022 (UTC)

The above discussion is preserved as an archive. Please do not modify it. Subsequent comments should be made in a new section.
This section was archived on a request by: —‍Mdaniels5757 (talk • contribs) 14:03, 23 October 2022 (UTC)

Explain my edit

Jar'Edo Wens was a longstanding hoax on the English Wikipedia that was deleted in 2015. It claimed that "Jar'Edo Wens" was an Australian aboriginal god. Because it was translated into some other languages before deletion, it was given a data page here, Q20191952. Someone later created an article about the hoax and turned "Jar'Edo Wens" into a redirect to the article about the hoax, and Q20191952 now references the hoax and has data about hoaxes. However, Q20191952 also has a "main subject" property saying that it's an instance of "god - of - Indigenous Australians". Since the hoax obviously isn't a god of Indigenous Australians, I removed the property, but when I clicked "remove", it removed it instantly without giving me the chance to leave an edit summary. How can I leave an edit summary, so this isn't seen as vandalism? Is there a way to do a dummy edit here? 123.51.107.94 01:24, 20 October 2022 (UTC)

@123.51.107.94: If you Undo a previous edit here then you can provide an edit summary, but in general the Wikidata UI does not allow for edit summaries for most simple changes like this. ArthurPSmith (talk) 01:36, 20 October 2022 (UTC)
You can also use the gadget EditSum, but this requires an account. — Metamorforme42 (talk) 08:41, 21 October 2022 (UTC)
This section was archived on a request by: —‍Mdaniels5757 (talk • contribs) 14:03, 23 October 2022 (UTC)

Describing outdoor performing arts venues

The WikiProject Cultural venues has meticulously documented how to represent indoor performing arts venues, such as performing arts buildings (Q57660343) and performance halls (Q112688641). However, there are no clear guidelines on how to describe outdoor venues dedicated to the performing arts.

There are no subclasses of event venue (Q18674739) and only a handful of subclasses of open-air venue (Q99692455) that refer specifically to outdoor performing arts venues. These include outdoor concert venue (Q50418254) and open-air theatre (Q11183017). However, both of these items are too narrow in scope to include venues such as Mississauga Celebration Square (Q107071940) that present more than just concerts and have a permanent stage, but no permanent seating (as open-air theatre, described as "a theatre without a roof", would imply).

Would it be helpful to propose a new subclass of open-air venue to describe all outdoor performing arts venues? And to group outdoor concert venue and open-air theatre as subclasses of this? Dlh28 (talk) 15:31, 14 October 2022 (UTC)

@Dlh28 These are interesting questions. As you pointed out, none of outdoor concert venue (Q50418254), open-air theatre (Q11183017), and their superclass open-air venue (Q99692455) are currently subclasses of event venue (Q18674739). Creating a performing arts specific superclass for them would seem a little "conceptually dissonant" to me, because members of the WikiProject Cultural venues have so far used event venue (Q18674739) as a superclass for all performance spaces (as illustrated in this class diagram). Therefore, whether or not we choose to make outdoor concert venue (Q50418254) and open-air theatre (Q11183017) subclasses of some "open-air performance space" class (or whatever the name we would give it), we should also consider making each one of a subclass of event venue (Q18674739). This would have the merit of producing a cleaner, more coherent class hierarchy for all performance spaces.
Beat Estermann (talk) 22:21, 31 March 2017 (UTC) Beireke1 (talk) Beireke1 (talk) 12:07, 2 May 2017 (UTC) - collaborating with Romaine on Belgian data on performing arts venues. Affom (talk) 15:26, 12 May 2017 (UTC) Anvilaquarius (talk) 11:41, 15 September 2017 (UTC) PEAk99(talk) PEAK99 (talk) 20:32, 29 January 2019 (UTC) Boxomi (talk) 22:21, 24 September 2019 (UTC) Antoine2711 (talk) 00:19, 5 December 2019 (UTC) Fjjulien (talk) 21:15, 13 February 2020 (UTC) Vero Marino (talk) 21:15, 13 February 2020 (UTC) Bello Na'im (talk) 08:34, 24 September 2021 (UTC) Titanboo (talk) 17:23, 8 July 2022 (UTC) dlh28 (talk)18:21, 22 September 2022 (UTC)
Notified participants of WikiProject Cultural venues Fjjulien (talk) 01:22, 19 October 2022 (UTC)
My previous proposed solution would actually only solve half of the problem, as we would still have to clarify the relationship between the new "open-air performance space" class and with the parent class open-air venue (Q99692455). If "open-air performance space" is a subclass open-air venue (Q99692455), which is subclass of venue (Q17350442), then "open-air performance space" would also indirectly be a subclass of venue (Q17350442) (and a subclass of event venue (Q18674739), as suggested above). Would this be a significant problem? Probably not. (This being said, I find this to raise another question: are "venue" and "event venue" truly conceptually different or are they the same? But let's not diverge). Fjjulien (talk) 01:38, 19 October 2022 (UTC)
I would also like to point out that performing arts performances are regularly presented in places that were not built primarily to serve this function: a church, a community centre, a shopping mall, a long-term care facility, a barn, an abandoned industrial building, etc. While some of these places may be characterized as an event venue (Q18674739), they could not be described as subclasses of a performing arts building (Q57660343) because that is not what they were designed for (although they may be occasionally repurposed for this kind of use). As we gather more and more information about actual performing arts events in databases such as Artsdata, we may realize that a significant proportion of these events are presented in places that were not purposely designed for the presentation of live performances. This may lead us to make (and accept) the distinction between the design and intended use of a place vs the actual use. Fjjulien (talk) 02:00, 19 October 2022 (UTC)
Circling back to Mississauga Celebration Square (Q107071940), the case that prompted your questions... Should the community decide not to create an "open-air performance space" class, you could still describe Mississauga Celebration Square (Q107071940) as a subclass of both an event venue (Q18674739) and open-air venue (Q99692455) (in addition to fixed construction (Q811430) and possibly public space (Q294440). All of these statements can be true and coexist at the same time. Fjjulien (talk) 02:10, 19 October 2022 (UTC)

Edits erroneously detected as potential vandalism

One of my recent edits was detected as potential vandalism. May you please elucidate me why? This could seriously deter unexperienced users when shown falsely.-- Kop Buster (talk) 14:37, 17 October 2022 (UTC)

You were adding values for DDB ID (P4948) that contained a series of consecutive consonants. The latter can potentially indicate vandalism, in this case it wasn't. I excluded the property from the filter that warned you. Lymantria (talk) 17:03, 17 October 2022 (UTC)
On a related topic, I am curious as to why are additions of LyricsTranslate ID (P7212) tagged as possible vandalism (see this diff for example). I have been patrolling this tag for the last few days, and this particular IP user keeps coming up because they do a lot of similar edits; as far as I can tell, all their edits are constructive. –FlyingAce✈hello 02:29, 19 October 2022 (UTC)

Брага: mashing or mash?

I've got the question whether ru:Брага / uk:Брага are rather the process mashing (Q1276781) or the product mash (Q114743955)?-- Noraberto (talk) 19:19, 19 October 2022 (UTC)

The product. Ymblanter (talk) 20:10, 19 October 2022 (UTC)

Equivalent of occupation (P106) for companies

Property occupation (P106) is to be used for people (and some similar entities such as fictional characters), but apparently not for companies, so I wanted to replace it in TopFoto (Q110975090) – but which property to use instead? --2A02:8108:50BF:C694:D010:1DEA:F053:C8EC 16:14, 9 October 2022 (UTC)

industry (P452) seems like the most relevant property to me. --M2Ys4U (talk) 16:25, 9 October 2022 (UTC)
And in addition, field of work (P101) to allow for greater specificity - what role is played by the company in the industry. --Tagishsimon (talk) 17:13, 9 October 2022 (UTC)
Like this? Or should photography (Q11633) be field of work (P101)? --2A02:8108:50BF:C694:D010:1DEA:F053:C8EC 17:24, 9 October 2022 (UTC)
Yes, possibly more like this ... photography probably isn;t an industry. --Tagishsimon (talk) 18:20, 9 October 2022 (UTC)
there is photo industry (Q1439700), though ... El Grafo (talk) 08:59, 20 October 2022 (UTC)

There is another non-person item with occupation (P106), a municipality this time: Baseri Kundal, Rajasthan, India (Q72926832)… It has a number of properties possibly not intended to be used there, such as native language (P103). --2A02:8108:50BF:C694:D010:1DEA:F053:C8EC 20:26, 9 October 2022 (UTC)

I've removed some of the errant statements; can see what the user was seeking to do - represent the languages, industries, ethnic groups &c associated with the village ... not sure whether & how WD models such things. --Tagishsimon (talk) 20:33, 9 October 2022 (UTC)
Yes. If someone familiar with modelling this kind of information happens to come by: Feel free to re-add the data. --2A02:8108:50BF:C694:D010:1DEA:F053:C8EC 20:39, 9 October 2022 (UTC)
(Funnily enough, I raised pretty much the same question above at #How_to_state_the_role_of_a_company,_if_not_in_P31? Jean-Fred (talk) 08:41, 10 October 2022 (UTC))
Maybe there should be a note in the description of occupation (P106) which property/ies to use instead for non-human entities? --2A02:8108:50BF:C694:C65:464B:9B56:426B 13:10, 10 October 2022 (UTC)
I have added such note. --Wiltbeider Thierry Dubois 39 (talk) 22:18, 16 October 2022 (UTC)
@Wiltbeider Thierry Dubois 39: I've reverted that edit. The incidence of business &c having a P106 statement is very small; IMO way too small to justify the property description being used to fight an insignificant problem. --Tagishsimon (talk) 22:47, 16 October 2022 (UTC)
There are at least 1000 of them for business (Q4830453) and company (Q783794), so at least not very small in absolute terms (maybe small compared with the total number of 217973 business (Q4830453)/company (Q783794) items). --2A02:8108:50BF:C694:BDFB:F1DB:F27F:16E6 09:55, 20 October 2022 (UTC)

100 million data items on Wikidata

Tracked in Phabricator
Task T321282

Wikidata surpassed 100 million data items!

--PantheraLeo1359531 (talk) 11:47, 19 October 2022 (UTC)

Determined roughly by myself. Probably Q114758237. --Wolverène (talk) 12:12, 19 October 2022 (UTC)
Actually we have 108545108 items. the number display in main page is number of items+properties+lexemes. GZWDer (talk) 12:43, 19 October 2022 (UTC)
As much as I remember, property pages are for some reason not considered "content pages". Thus the counter on the main page and on Special:Statistics includes item (99.303M) + lexeme pages (700k). —MisterSynergy (talk) 13:07, 19 October 2022 (UTC)
This is my understanding as well - it is lexemes plus main items. Properties are not considered content pages under mw:Manual:$wgContentNamespaces but strangely EntitySchema are, per the site configuration - these have `content: ""` set which I think indicates a content namespace. This query will count all items in all namespaces if you want to check the totals... Andrew Gray (talk) 17:13, 19 October 2022 (UTC)
I always thought it comes from here which indicates that main namespace (by default) and Lexeme namespace (Wikidata-specific) are content namespaces in this project. Not sure whether additional content namespaces can be defined elsewhere. See also: mw:API:Siteinfo#Namespaces and mw:Manual:$wgContentNamespaces. —MisterSynergy (talk) 17:33, 19 October 2022 (UTC)
I'm a bit puzzled to be honest, as I'd looked there too, but my best guess was that the one reported by the API must be closer to how the system is counting it? Either way it does not make a great deal of difference (there are only a handful of ES) but would be nice to know for sure... Andrew Gray (talk) 18:47, 19 October 2022 (UTC)
This looks like the "content" attribute for namespace 640 is hard-coded into the EntitySchema extension. Not sure whether this is a good idea. (I have very limited understanding of MediaWiki internals, so I might be wrong…) —MisterSynergy (talk) 19:09, 19 October 2022 (UTC)
We are counting Entity Schemas as content pages, but Properties are not being counted. We have created a ticket to include Properties in the count, but please let us know if there are any potential negative side-effects that we should be aware of. -Mohammed Sadat (WMDE) (talk) 13:03, 20 October 2022 (UTC)
I think that one was too early. It went fast, but my guess it is one of Q114758251, Q114758252 or Q114758253. Ainali (talk) 14:43, 19 October 2022 (UTC)

Year of birth and death in English description

Greetings! Is there any rule or discussion about how the year of birth and death should be indicated in the English description of a person's element?

I mean something like this Q3182477: English botanist, first curator of Kew gardens (1798–1888)

Should the data in brackets be placed at the beginning or at the end of the description? Do you need a long hyphen or a short one?

Is there an agreement on this? Pallor (talk) 14:44, 19 October 2022 (UTC)

All the model item (P5869) for human (Q5) put it at the end of the description. There's a mix between long hyphen or a short hyphen in the model items. I don't think we need uniformity about long hyphen and short hyphens. ChristianKl ❪✉❫ 15:03, 19 October 2022 (UTC)
It should not be indicated at all unless there are special reasons. See Wikidata:Requests for comment/Use of dates in the descriptions of items regarding humans for an introduction into this debate. --Emu (talk) 15:06, 19 October 2022 (UTC)
It should be indicated because it is, generally speaking, very useful to have it in the description. --Tagishsimon (talk) 15:20, 19 October 2022 (UTC)
We have discussed this many, many times. Everything can be “useful” for some workflows and applications. Indeed descriptions like “Occupation: Interior Designer Date of Birth: January 01, 1930 Date of Death: January 01, 1994 Birth City: Minneapolis Birth State/Province: Minnesota Birth Country: United States Resident City: Minneapolis Resident Country: United States Archive/Repo” (Q76124272) can be useful. Why not copy all statements to the description? That would be über useful! --Emu (talk) 15:47, 19 October 2022 (UTC)
It is conventional to include some combination of citizenship, occupation and dob/dod. We do so because these are the most useful disambiguators of human name labels, quite apart from being immediately informative. I take note that you do not have a better argument than the above straw man. I live in hope that one day you will listen to users who value dob/dod in descriptions and put your peculiar personal belief aside. --Tagishsimon (talk) 18:47, 19 October 2022 (UTC)
It is conventional to use citizenship + occupation without DOB/DOD. Of the first 100 human search results for “American writer” only 9 have some sort of DOB/DOD. Help:Description doesn’t mention any form of DOB/DOD. It’s a pet project of some users to add DOB/DOD (some even employed large bot jobs in the past), not any form of consensus. But sure, frame it as some pet peeve of mine and ignore all arguments that Epìdosis so diligently collected on the page I linked, I’m sure that this will further your noble cause. --Emu (talk) 20:30, 19 October 2022 (UTC)
Any additional information can be useful for disambiguating (there is no real use of descriptions beyond that, eventually), but obviously it’s not desirable to have the desciption overloaded with unneeded pieces of information (not to mention the character limit). Ideally, descriptions would contain the minimum number of statements needed for full disambiguation. Of course, a strict interpretation of this would lead to items where the label is already unique having no description at all. And this would (sometimes, perhaps frequently) require changing the descriptions of all items with the same label when a new item is created. There has to be some compromise. In my perception, in the case of humans, nationality and occupation are already sufficient most of the time (occupation alone would probably do in many cases), but are DOB/DOD harmful? However, if they are included, they should use an EN dash (–) for typographic reasons. (Though I am not sure whether there are different opinions on this depending on style guide.) --2A02:8108:50BF:C694:BDFB:F1DB:F27F:16E6 09:31, 20 October 2022 (UTC)

Thank you all for your answers, especially the linked May discussion, because I was not familiar with it. In order to avoid misunderstandings: the question was not whether it is necessary to indicate these data, but in what format should we enter the DOB/DOD that has already been indicated and will occur in the future in the English description? It is obvious that - whether we want it or not - this data ARE ALREADY INDICATED in the description, it would be a shame to pretend that it does not exist or that it is only a temporary phenomenon. It is also not worth assuming that anyone would take the trouble to delete this data from all the elements within a day or two, i.e. after a few days the question would become outdated.

There are indeed edit wars because of this, which is unfortunate: it distracts from meaningful edits, it captures the attention of administrators, all because of an otherwise insignificant issue. I have the impression that there are two types of views: one does not want this data to be included at all, precisely because of this it does not have a position on what format it should be in if it is included. The other side - regardless of whether it considers the indication necessary or not - approaches the issue in a practical way.

I await the opinions of the latter in the future

  • is it better to do it at the beginning or at the end? In German and Hungarian descriptions, it often happens that the date appears at the beginning, in the others - e.g. Dutch - always at the end, in English 99 percent at the end, but even today there are editors who always put it at the beginning.
  • do you need a long or short hyphen? (There is a special spelling rule for this in Hungarian, it is an error for us if there is not a long hyphen, but in English I don't know if there is any rule for this or, for example, library practice...) Pallor (talk) 10:18, 20 October 2022 (UTC)
Leave it as it is: style, typography, positioning, complete absense of DOB/DOD in descriptions — if the description is not incorrect, don't touch it. Uniformity of descriptions is not a goal and does not yield any benefits, so we do not need to strive for it. —MisterSynergy (talk) 10:27, 20 October 2022 (UTC)

Question about Stoomoven [Dutch]

steam oven (Q2114471)
food steamer (Q1158805)

Is steam oven (Q2114471) identical to food steamer (Q1158805)? Dixiparty (talk) 11:44, 20 October 2022 (UTC)

As you can see in the sitelinks: nl:Stoomoven is different than nl:Stoomkoker. Multichill (talk) 12:05, 20 October 2022 (UTC)
@Multichill: I'm not an impaired person. What is the exact difference between those?--Dixiparty (talk) 12:06, 20 October 2022 (UTC)
I see you're new here, please be polite. Uncivil comments are not appreciated.
You are asking if the two are identical and the answer is no. The first one is a type of oven the second one a cooking appliance where you boil water and put whatever to steam on top of it. Multichill (talk) 12:23, 20 October 2022 (UTC)

Attempting to merge two items, I seem to have been flagged as a vandal

I am attempting to merge the lesser item Q11205631 into the older, better item Q819545 using the "Merge two Items"-tool.

I kept getting "Error: Conflicting descriptions for language fa." for different languages so I began deleting entries from the lesser item that already existed in the greater item. I don't know all of these languages so I don't know if they are equal or not. And the "Merge two Items"-tool only gives me one conflict at a time.

Could someone designate me as "not a vandal" so I can complete the removal edit I have in my browser and finally merge these items? Or am I doing this all wrong? BucketOfSquirrels (talk) 18:11, 20 October 2022 (UTC)

Merged. Maybe use the merge gadget next time. Help:Merge --Tagishsimon (talk) 18:26, 20 October 2022 (UTC)
Thanks! Yeah, I think I'll do that next time. The "Merge two Items"-tool seemed like the quickest way to do it. And then it became a sunk cost. It was a frustrating experience. BucketOfSquirrels (talk) 20:59, 20 October 2022 (UTC)

Mass shooting vs. Mass murder

Should items use both of these as P31 or just only one of them? Trade (talk) 22:48, 20 October 2022 (UTC)

Mass shooting is a subclass of mass murder. Presuming shooting was the only thing going on, only the first value is necessary. --Tagishsimon (talk) 23:30, 20 October 2022 (UTC)
Both should probably be used. There are cases like Q33000144. It was a mass shooting. However, there were no deaths so it wouldn't be a mass murder. DoublePendulumAttractor (talk) 01:59, 21 October 2022 (UTC)
There's clearly a problem here. Little Rock nightclub shooting (Q33000144) is defined as a mass shooting (Q21480300), which is defined as a subclass of mass murder (Q750215). --Tagishsimon (talk) 02:06, 21 October 2022 (UTC)
You can have lots of shooting without lots of deaths. Shouldn't take rocket science to realize not all shootings (mass or single) are murders. And mass murder can be committed without shooting, e.g. mass stabbing, mass burning, mass poisoning, etc. -Animalparty (talk) 02:51, 21 October 2022 (UTC)
Shouldn't be rocket science to do the minimum of background reading - such as the en wiki article - to come to understand that some authorities define mass shooting in terms of fatalities, some do not. Right now, the mass shooting item on WD is consistent with an 'involves fatalities' definition, but is used erroneously on Q33000144. Ideally WD reflects sources, not the blazing intuition of its users. So, not for the first time, this corner of WD is a mess, should anyone wish to get involved in cleaning it up. --Tagishsimon (talk) 10:36, 21 October 2022 (UTC)

New Lexeme creation page will be live on Wikidata on November 2nd

Hi everyone,

Among our development goals this year is to make the lexicographical data part of Wikidata easier to understand for people not familiar with lexicography. This included reworking the Lexeme creation page to improve the editing experience of users. We plan to replace the Special:NewLexeme page with the new one on November 2nd!

As you may recall, we made a number of tweaks to the old page and asked you to test it and give feedback (see the previous announcement). We addressed the issues the community raised, and we would like to thank everyone who participated in the testing and provided feedback.

While the new Special:NewLexeme is already scheduled to be deployed, we would still like to hear what you think. If you have any questions or suggestions please let us know on the Wikidata:Lexicographical data talk page.

Cheers, -Mohammed Sadat (WMDE) (talk) 09:31, 21 October 2022 (UTC)

Data cancellation

I have not given any authorization to publish my profile and my date of birth. I therefore ask that the data concerning me be immediately deleted. Thank you Elena Gremigni (talk) 10:03, 21 October 2022 (UTC)

Loophole?

I’ve discovered something that seems like a loophole in the instance of (P31)/subclass of (P279) system to me. In my understanding, A instance of (P31) B and B subclass of (P279) C implies A instance of (P31) C. However, there is (for example) Cocopah people (Q165320)instance of (P31)Indigenous peoples of the Americas (Q36747) and Indigenous peoples of the Americas (Q36747)subclass of (P279)Americans (Q2384959), which would imply Cocopah people (Q165320)instance of (P31)Americans (Q2384959), then Americans (Q2384959)subclass of (P279)inhabitant (Q22947), which would imply Cocopah people (Q165320)instance of (P31)inhabitant (Q22947), and finally inhabitant (Q22947)subclass of (P279)human (Q5), which would ultimately imply Cocopah people (Q165320)instance of (P31)human (Q5). Now Cocopah people (Q165320) is not a human (Q5), but rather a group/tribe of them… Is my understanding wrong and this an intentional modelling? Or is this an unintended side-effect of one statement in the chain? (I would suspect Indigenous peoples of the Americas (Q36747)subclass of (P279)Americans (Q2384959), since an indigeneous people of the Americas is already not an American, but a group of them.) Or do qualifiers have to be respected here? (Though it is not obvious to me how.) --2A02:8108:50BF:C694:146A:852:36CA:CE6E 21:21, 16 October 2022 (UTC)

That seems to be a modelling error. See Wikidata talk:WikiProject Ontology and Wikidata:Project_chat#Ontology_test_cases for more cases like this Mateusz Konieczny (talk) 22:44, 16 October 2022 (UTC)
Cocopah people (Q165320) being an instance of Indigenous peoples of the Americas (Q36747) is saying that it's a member of a group of humans, which is a single human. Ghouston (talk) 11:11, 17 October 2022 (UTC)
If this is the case there are many such instances that must be fixed (for this item alone). Query -wd-Ryan (Talk/Edits) 17:44, 17 October 2022 (UTC)
Putting concerns about correct labelling (singular/plural) aside, a specific indigenous people being instance of (P31) (i.e. is-a) indigenous people of somewhere does not look wrong to me, but that indigenous people of somewhere, which should thus be a class of indigenous people, being instance of (P31) (i.e. is-a) group of humans does. It should be subclass of (P279) group of humans. In any case, the problem here is that at some point in the hierarchy a class of groups of humans (indigenous peoples are groups of humans) is mis-classified as a group of humans (whose members are then humans instead of groups of humans). --2A02:8108:50BF:C694:F1BE:DAFE:689A:1C4F 20:28, 17 October 2022 (UTC)
There's a similar example in a thread above, too. --M2Ys4U (talk) 12:37, 17 October 2022 (UTC)
Now archived to https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2022/10#Example_of_a_found_problem:_Tulalip_Tribes_of_Washington_(Q1516298)_-_tribe_is_human_according_to_Wikidata without fix Mateusz Konieczny (talk) 06:55, 22 October 2022 (UTC)
Posted in https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Ontology#tribe_vs_human to increase chance it will be spotted and processed. Mateusz Konieczny (talk) 06:55, 22 October 2022 (UTC)

Daniel Orme item

The British engraver Daniel Orme (1766-1832) is Daniel Orme (Q5218343). However, Daniel Orme (Q94914398) appears to be a duplicate item for the same person with a mistaken date of death of 1802. The VIAF identity https://viaf.org/viaf/56880092 lists only a small handful of works that are all by the "real" Orme. What should be done here? Merge?

Further more, there is a very confused item D. Orme (Q108213698) which appears to have taken metadata from some database relating to an item (Orme's trade card) and used it to construct a Wikidata entity for the person. Inductiveload (talk) 16:59, 22 October 2022 (UTC)

Merge them, deprecate incorrect statements, maybe move the trade card record's P31. --Tagishsimon (talk) 18:39, 22 October 2022 (UTC)
Thanks! I merged the first two, and also found another VIAF cluster for him! I think I'll pass on the card one actually, as there are nearly 500 of them (https://w.wiki/5rBn) all with the same issues! Inductiveload (talk) Inductiveload (talk) 19:53, 22 October 2022 (UTC)

list of highly corrected claims?

Is there a page that lists properties / prop+object pairs that had the most (bot or manual) corrections? My interest starts with but is not limited to main subject (P921) statements of scholarly articles that were created by naive heuristics. I have now corrected so many of them that I'm starting to believe I'm not alone, and some stat could be worthwhile. SCIdude (talk) 08:23, 23 October 2022 (UTC)

Withdrawal from a treaty

France (along with others) has announced withdrawal from Energy Charter Treaty (Q2001857). All participating countries are currently marked as signatory (P1891), but that does not allow end time (P582). How to represent a country withdrawing from a treaty? Pauljmackay (talk) 10:12, 23 October 2022 (UTC)

Put an end time and an end cause on the statement. If the constraint whinges, the constraint is wrong and needs to be changed. --Tagishsimon (talk) 14:11, 23 October 2022 (UTC)

Mass murder

Could someone please make the two queries? --Trade (talk) 16:29, 23 October 2022 (UTC)

https://w.wiki/5rLp
https://w.wiki/5rLr -wd-Ryan (Talk/Edits) 16:39, 23 October 2022 (UTC)

Occupation: wife

Continuing the occupation (P106) questions, some 96 items have occupation (P106)wife (Q188830). I acknowledge that in many countries married women were (occasionally still are) expected to dedicate themselves to being a wife, occupationally, so to speak… but I dare to doubt that this makes occupation (P106)wife (Q188830) an appropriate modelling for Wikidata. What would be a suitable alternative property? position held (P39)? social classification (P3716)? --2A02:8108:50BF:C694:3C5C:2B2:84A2:6682 20:43, 26 October 2022 (UTC)

In the absence of objections I will now change occupation (P106)wife (Q188830) to social classification (P3716)wife (Q188830). --2A02:8108:50BF:C694:DC3D:88D9:AC3D:EDB4 21:12, 29 October 2022 (UTC)

This section was archived on a request by: --2A02:8108:50BF:C694:DC3D:88D9:AC3D:EDB4 21:26, 29 October 2022 (UTC)

Edits reverted

User:Dorades reverted my edits on font (Q4868296), Lego (Q170484), emoji (Q1049294) and ringtone (Q690856). Why? They are useless? 2001:B07:6442:8903:DD4B:6839:FDBD:2576 15:34, 20 October 2022 (UTC)

Hello! I'm glad you want to contribute to Wikidata. The property reference URL (P854) is used underneath statements to record where information was imported from. Also, the websites you added aren't really appropriate for these items. Instead, you can edit the items for the websites, such as Emojipedia (Q22908129). See the statement main subject (P921) = emoji (Q1049294). -wd-Ryan (Talk/Edits) 16:15, 20 October 2022 (UTC)
CC @Dorades: -wd-Ryan (Talk/Edits) 16:16, 20 October 2022 (UTC)
Thank you, wd-Ryan, for explaining, I have nothing to add. --Dorades (talk) 09:39, 24 October 2022 (UTC)

How do I add multiple statements from the same reference, without re-entering the full reference every time

I tried to read Help:Sources#Different types of sources#Web page and Help:Statements#Adding_statements.

I don't understand the instruction "Source your statement with reference URL (P854) and link it to the URL of the webpage".

On Wikipedia I would simply use the "Edit source"-feature and paste a reference name everywhere I needed to. BucketOfSquirrels (talk) 20:38, 21 October 2022 (UTC)

@BucketOfSquirrels: A reference can be copied from a statement X to statements Y, Z, etc. using the gadget DuplicateReferences, that you can turn on in Special:Preferences#mw-prefsection-gadgets. --Epìdosis 21:00, 21 October 2022 (UTC)
I have to question why some of these gadgets aren't enabled by default. People not having them enabled seems to cause new users a lot frustration @Epìdosis:--Trade (talk) 21:47, 21 October 2022 (UTC)
I would personally enable by default at least Merge and DuplicateReferences, I think it would have just positive effects. --Epìdosis 21:52, 21 October 2022 (UTC)
Note that DuplicateReferences has a bug where you can only copy references that already existed. You need to refresh to copy references that you just made. -wd-Ryan (Talk/Edits) 02:23, 22 October 2022 (UTC)
Thanks all. I found this old topic on the guide's talk page. I hope I'm not violating a policy on rallying/picketing, (I don't remember the term and I couldn't find the policy), but could some experienced editors have a look at it? BucketOfSquirrels (talk) 17:01, 22 October 2022 (UTC)
Honestly, I really wonder why wikidata can't adopt something more like wikipedia's reference system, rather than having every statement have a separate reference entry. This REALLY overwhelms some items that have thousands of statements - when every statement has the same reference, that duplication can increase the size of the item drastically. Instead why not have references attached to the item, and then each statement could refer back to a particular reference by its id within the item (ref. 1, ref 9, etc.)? Has there been any discussion on here or in phabricator about switching to a better reference system like this? The RDF representation would be simplified too I think with this method. ArthurPSmith (talk) 12:27, 24 October 2022 (UTC)
 Support I’m sure it would, but I’ve no idea whether there has been any discussion… --2A02:8108:50BF:C694:D412:B8C4:744F:D6DD 13:24, 24 October 2022 (UTC)

Goalkeeper an occupation?

It’s an occupation (P106) question again…

Some 19 items have occupation (P106)goalkeeper (Q172964). Given that relatively small number (compared to the multitude of soccer etc. teams), I suspect this is due to most goalkeepers not having such a statement – inconsistent modelling, in other words. What I’m not sure about is what the desirable modelling should be in this case. Goalkeeper is probably not an occupation (that would be player of whatever sport the person in question plays), so position played on team / speciality (P413) should be used?

As an aside, goalkeeper (Q172964) has both instance of (P31)position (Q1781513) and subclass of (P279)position (Q1781513), which does’t look good in terms of ontology @Mateusz Konieczny. --2A02:8108:50BF:C694:D412:B8C4:744F:D6DD 09:04, 24 October 2022 (UTC)

"instance of (P31)position (Q1781513) and subclass of (P279)position (Q1781513)" - edited, maybe it is better now and humans no longer end with weird classification due to that subclass Mateusz Konieczny (talk) 11:50, 24 October 2022 (UTC)
Thanks! In the absence of comments or objections regarding the question, I will now replace occupation (P106) with position played on team / speciality (P413) for goalkeeper (Q172964). --2A02:8108:50BF:C694:D412:B8C4:744F:D6DD 13:47, 24 October 2022 (UTC)

Single-value constraints

Is there a way to exclude from constraint reports, items who list more than one value for a property, but where the other values have deprecated rank? Seems to me it should only report when there are more than one value having normal or preferred rank. You can see an example here: Wikidata:Database reports/Constraint violations/P5080 on the item for Ole Gunnar Solskjær (Q18976). It does not show up on this page however: Special:ConstraintReport/Q18976. Infrastruktur (talk) 16:07, 22 October 2022 (UTC)

It would be ideal if separator (P4155) accepted a rank as a value. I don't think it does. Possibly reason for deprecated rank (P2241) could be used as a proxy P4155 value; deprecated statements should carry this qualifier (although even then, iirc, the separator qualifier values need to be distinct if there is more than one instance of the same separator qualifier on a set of property statemements for an item.) --Tagishsimon (talk) 16:17, 22 October 2022 (UTC)
Good to know. I'm trying this out on the aforementioned property. But I wonder if the bot operator would consider making this the new default? The Wikibase software does not consider statements with deprecated rank to be constraint violations either. @Ivan A. Krestinin: Infrastruktur (talk) 13:28, 23 October 2022 (UTC)
Krbot2 just ran on the P5080. The separator I added seems to have made no difference to the report, so I am going to remove it. Infrastruktur (talk) 14:56, 24 October 2022 (UTC)
Yes, I believe this request appears on User_talk:Ivan_A._Krestinin#Deprecated_rank_not_taken_into_account_for_P648_"single_value"_report? ~every month :) Vojtěch Dostál (talk) 15:03, 24 October 2022 (UTC)

Catholic or not?

I’ve stumbled across religious (Q2566598), which, according to its English description, is just “member of a religious order” without mentioning a denomination, but both its German and its French label declare it to be member of a Catholic order. The English Wikipedia article explicitly calls it a term from the “terminology of many Western Christian denominations, such as the Catholic Church, Lutheran Churches, and Anglican Communion” and there are no German or French Wikipedia articles linked. Nor has Catholicism been linked by a property. Does being a religious (Q2566598) imply being Catholic (probably not)? Should the item be split or just the descriptions corrected? (I have no idea whether this only applies to French and German.) --2A02:8108:50BF:C694:D412:B8C4:744F:D6DD 09:11, 24 October 2022 (UTC)

This item is explicitly catholic since 2014. So please create a new one if you want to have a more generic one. Ayack (talk) 09:49, 24 October 2022 (UTC)
That statements has been removed – but thanks, then I’ll re-add it and split the item. --2A02:8108:50BF:C694:D412:B8C4:744F:D6DD 09:54, 24 October 2022 (UTC)
@Ayack: The newly created item is religious (Western Christianity) (Q114836842), but unfortunately religious (Q2566598) is protected and I cannot edit it. I would suggest:
Also, I have not checked whether the values of equivalent class (P1709) and the external identifiers for religious (Q2566598) are correct. But religious (Q2566598)Commons category (P373)Category:Religious occupations looks fishy. --2A02:8108:50BF:C694:D412:B8C4:744F:D6DD 10:02, 24 October 2022 (UTC)
Done. --Wiltbeider Thierry Dubois 39 (talk) 20:51, 24 October 2022 (UTC)

Importing a purely numerical value from a Wikidata value that includes a unit of measurement.

I want to make this work on Wikipedia:

import syntax (the four pipes aren't really there) = {{||||#statements:P2043|from=Q48815436}}

{{convert|import syntax|m|ftin}}

To do this I need the numerical value imported without the unit of measurement. How do I do this? Where can I find a list of flags for the "Parser function"?BucketOfSquirrels (talk) 20:24, 24 October 2022 (UTC)

On English Wikipedia you can do the following:
{{convert|input=P2043|qid=Q48815436|ftin}}
I'm not sure if this helps you? — Martin (MSGJ · talk) 22:06, 24 October 2022 (UTC)
Thanks! I just found this on Wikipedia;
{{convert|{{wikidata|property|raw|Q48815436|P2043}}|m|ftin}}
which is what I technically asked for, but I like yours better. BucketOfSquirrels (talk) 22:36, 24 October 2022 (UTC)

ページ名の付け方

くろとんこだ 2400:2411:AAE2:B700:D8E6:2A64:3F4E:F26E 00:46, 25 October 2022 (UTC)

Wikidata weekly summary #543

Here's your quick overview of what has been happening around Wikidata over the last week.
Wikidata passed another milestone on October 19.

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.

emerging patterns on Music from Wikidata: evaluation

Hi, I’m Valentina Carriero, a phd student at the University of Bologna working mainly on ontologies and knowledge graphs.

In a recent work, that I presented yesterday at the Wikidata Workshop (https://wikidataworkshop.github.io/2022/papers/Wikidata_Workshop_2022_paper_1411.pdf), we developed a method for extracting what we call "emerging patterns" from a Wikidata subgraph on a specific domain (in our case, music!). We identify the most relevant (i.e. more instantiated) classes of that domain, and suggest a set of properties, and the appropriate range(s) for those properties, to be used with instances of that class, based on the actual usage in the Wikidata sub-KG on music. We randomly (for now) compare our results with the current support in Wikidata (including property constraints and your WikiProject Music), in order to show how we can support the reuse of the Wikidata ontology. In addition to reflecting the actual usage, our patterns integrate current suggestions by identifying the "correct" (based on the usage) pairs of domain (type constraint) and range (value-type constraint) for a property.

We would be very interested to perform a user-based evaluation of our method and our patterns on the music domain, in order to understand how they can actually support the WikiProject Music, and all editors that happen to edit some music-related items or have some knowledge about music, and how we can actually impact on the Wikidata community work and Wikidata KG completeness/correctness. For now, our patterns are basically .tsv files with all "triplets" above a threshold (that can be customised by the user) (see an example here https://github.com/valecarriero/wikidata-emerging-patterns/blob/main/results/patterns/Q482994/Q482994-dr-pairs-frequent-properties_85_50_18.tsv for the album pattern), but, if possible, we could think together if there is a way to integrate our suggestions of properties and ranges with tools/schemas/constraints currently used by the Wikidata community when new data is injected or when a project like yours is built, in order to perform this evaluation! Here's the link to the GitHub repo with a brief documentation https://github.com/valecarriero/wikidata-emerging-patterns

Let me know if any of you is interested to talk more about how we can make these patterns actually useful for you and to participate in the user-based evaluation (that shouldn't take long at all)!

Thanks a lot!

Valentina

email: valentina.carriero3@unibo.it Vale Carriero (talk) 07:53, 25 October 2022 (UTC)

What is the difference between Q1410110, Q1880737 and Q658255?

What is the difference between

  • Q1410110 (branch = branch office)
  • Q1880737 (branch office = branch outlet)
  • Q658255 (subsidiary company = daughter company)

To me they look all the same. Can at least two be merged (and preferably all three into one)? JopkeB (talk) 08:42, 25 October 2022 (UTC)

subsidiary (Q658255) is an independent legal entity (as explained in the English description at branch office (Q1880737)). branch office (Q1880737) is a little confusing, many sitelinks go to "filiale" or similar concepts but the English "branch office" may have a slightly different meaning. Actually I'd me more worried about the difference between branch (Q232846) and some of the three you mentioned. Vojtěch Dostál (talk) 08:55, 25 October 2022 (UTC)

P5606’s P2559 “Wikidata usage instructions” should be P2378 “issued by”

There is a typo of P2738, which should be P2378 JuguangXiao (talk) 22:37, 24 October 2022 (UTC)

Property:P5606 JuguangXiao (talk) 22:38, 24 October 2022 (UTC)
Sorted. Easiest thing would have been for you to make the change. --Tagishsimon (talk) 22:46, 24 October 2022 (UTC)
As far as I know, not everyone can edit Property, no? I have edited few Items, but Property page shows no edit UI to me. JuguangXiao (talk) 21:53, 25 October 2022 (UTC)

Noun Project Attributions on this site

Hi there,

I'm the Marketing Communications Manager at Noun Project and want to thank you for using some of our icons on this site, especially this page: https://www.wikidata.org/wiki/Wikidata:Data_Import_Guide

Can you please include the hyperlink below for attribution purposes?

Please use the following attribution including the hyperlink: Presentation icon from the Noun Project

Thank you! 2603:8000:A603:C961:A861:F2F7:83B:F486 21:05, 25 October 2022 (UTC)

CC3 unported - presuming this is the licence under which the images were made available - requires "You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use." Appropriate credit is given on the wikicommons page you arrive at if you click on any of the images - e.g. at https://commons.wikimedia.org/wiki/File:Noun_Project_start_icon_69466.svg . --Tagishsimon (talk) 21:19, 25 October 2022 (UTC)

How to tag livery of aircraft/hot air balloons?

I'd like to link aircraft/hot air balloons with the item of their livery. Example: HB-QMU (Q106043344) -> Raiffeisen Switzerland (Q681189). Which property is best suited for this relation? Should main subject (P921), depicts (P180) or dedicated to (P825) be used for this purpose?

Joshbaumgartner (talk) Milad A380 talk? Danrok (talk) 15:06, 29 August 2013 (UTC) Fale (talk) Jura Novarupta (talk) -- focusing on airports MB-one (talk) 15:06, 12 August 2017 (UTC) Fuzheado (talk) 15:31, 19 September 2017 (UTC) econterms (talk) 01:42, 17 April 2018 (UTC) Bouzinac (talk) Bingobro (Chat) 13:16, 1 July 2020 (UTC) Malvinero10 (talk) 22:22, 2 August 2020 (UTC) Respublik (talk) 15:37, 26 September 2022 (UTC)

Notified participants of WikiProject Aviation 1-Byte (talk) 11:15, 23 October 2022 (UTC)

This is an excellent question. The same would apply for sponsors on race cars or even sports uniforms, I imagine. main subject (P921) probably does not work as a vehicle is not really a work about the displayed sponsor. depicts (P180) might work but not with the sponsor exactly (since the vehicle is not depicting the company as a whole) but maybe an item for the particular logo or company branding that is actually depicted. dedicated to (P825) is a little different as dedication and sponsorship are different kinds of relationships. sponsor (P859) also may apply, but this would only be in case the bank really does sponsor the balloon in question, and would not even require the balloon to display the sponsor name to be valid. Of course there are many cases of vehicle with company names and logos on them not as a result of sponsorship but just because their owners like the look or because they are replicating historical looks and dedicated to (P825) would not really apply for those cases. Ideally, a property along the lines of 'advertises' would be better, but in lieu of such, I'd recommend using depicts (P180) to a new item for the displayed logo/branding (different from the organization item) for what is visibly displayed on the vehicle as well as dedicated to (P825) to the organization when actual sponsorship is involved. Josh Baumgartner (talk) 04:03, 26 October 2022 (UTC)

relative (P1038) and sibling (P3373)

I have correctly added relative (P1038) and sibling (P3373) on Giorgio Pustetto (Q69154523), Carlo D’Alpaos (Q69152719) and Andrea D'Alpaos (Q3615652)? 2001:B07:6442:8903:693C:3B49:53C4:C006 08:43, 26 October 2022 (UTC)

Setting head of government (P6) or similar qualifier for non-notable persons

In our local Wikipedia we regularly had to change the head of government (P6) for cities, towns, villages, etc. Unfortunately can't create an item for all the persons, because almost all of them do not meet the Notability criteria. Is there other way to add such data? I think that this data is important for these entities and I think that its place is in Wikidata, but I'm not sure in what form we can add them. Do you have any ideas? --StanProg (talk) 13:42, 25 October 2022 (UTC)

 Question Are we absolutely positive that these items don’t meet the notability criteria of the project? I would imagine that if someone is the head of government (P6) of a bigger municipality they would be fairly easily describable with publicly available sources. Furthermore, we may also take into account the third point of the policy; if there is a real structural need for the item, it should be acceptable. --everyone's favorite Blua lago(let's have a chat y'all) 14:43, 25 October 2022 (UTC)
Almost all "mayors" do not have article in Wikipedia or on any other project so I think they do not meet the Notability criteria. I'm not talking just "mayors" of bigger municipalities or big cities, but ones on smaller ones, towns and even villages with 100 citizens for example. I can assume that the ones of the bigger settlements could meet the criteria, but the ones of the smaller ones definitely do not. Maybe there should be some kind of limit for the size of the settlement? I'm not sure if being a mayor of small village can fit into the "real structural need" criteria. Any thoughts? StanProg (talk) 14:56, 25 October 2022 (UTC)
@StanProg In my opinion, article in Wikipedia is not a real argument for inclusion or non-inclusion of an item into Wikidata. Mayors do deserve items (and some countries have created items for all or most of their mayors). There is a bigger problem - with sourcing the data in countries which do not have a public database of mayors. Vojtěch Dostál (talk) 15:03, 25 October 2022 (UTC)
At least we don't have such problem, because our mayors database is publicly available. I will wait for some more opinions. It's better to be sure that this is OK, before starting adding them as it's a lot of work. StanProg (talk) 15:28, 25 October 2022 (UTC)
mayors are much more notable than a lot of people already in wikidata. if you have strong references for this data then I say include it. BrokenSegue (talk) 17:39, 25 October 2022 (UTC)
I agree with Vojtěch Dostál - these people should have their own items. We have items for most sports people whether or not they have wikipedia articles about them. And for lots of scientists that don't have wikipedia articles. ArthurPSmith (talk) 15:44, 25 October 2022 (UTC)
I agree with Vojtěch Dostál too – both in the assessment of notability and with regard to the problems in sourcing. For example, it would be tough for Austria because there isn’t any public database and in many cases hardly anything is known about the mayor apart from the name and maybe the party they ran for. --Emu (talk) 15:47, 25 October 2022 (UTC)
A person doesn't need to have extensive biographic information published to satisfy WD:N. If there is even a single credible list of mayors with nothing more than names and dates served, that would probably be sufficient. As future research unfolds, and/or more databases, books, and newspapers become digitized, items can be expanded or merged as needed (some public officials might be more prominent in business or other areas: it may not be apparent until later than J. Schmidt the farmer was also J. Schmidt the mayor). I think Wikidata really shines when disparate databases can be connected. As an example, this database entry contains very little info besides name and dates served, even though it's an official government source. But cross-referencing it with books and other resources, it can be identified as Robert Hairston (Q20989132) -Animalparty (talk) 18:58, 25 October 2022 (UTC)
Yes – and no. The items should also have enough information for future generations to know who they are supposed to be about – without that, cross-referencing can’t work. To give an example, it gets tricky when different people with the same name serve as mayor – and that’s pretty normal in some regions as there is a limited set of local family names and an even more limited set of popular first names. Depending on the heuristic, you are bound to create a lot of duplicates or conflations. --Emu (talk) 19:49, 25 October 2022 (UTC)
I see no difference between an item that says nothing more than " J. Schmidt, mayor of town X in 1865" and a scientist item that can say no more than "14th co-author of a single paper". There are undoubtedly tons of duplicate entities already on Wikidata awaiting identification and merging, and more will certainly arise in the future. As more information is revealed about "J. Schmidt who was mayor of town X in 1865", add it! If the next mayor is also a J. Schmidt, but there is doubt or ambiguity as to whether it's the same person as, don't merge! But if nothing is ever verified beyond name and mayorship, it's still a notable entity. -Animalparty (talk) 20:30, 25 October 2022 (UTC)
Thank you all for the responses. They were very helpful. --StanProg (talk) 09:05, 26 October 2022 (UTC)
@StanProg: If you really don't want to create an item (or if you don't want to create an item right now), a standard idiom would be to set P6 = somevalue, with qualifier object named as (P1932) = <name>. ("Object" because they are the object of the statement). But as others have said above, do feel free to create an item. Jheald (talk) 13:08, 26 October 2022 (UTC)
  • I have been working on lists of mayors for over a decade, most of those lists have been deleted from Wikipedia. I have been trying to recreate them here. See for instance Talk:Q106370185 where you can create a master list for the position. You can indicate interim holders of the office, so they appear in italics, because they do not get a number. --RAN (talk) 21:32, 26 October 2022 (UTC)

Yemeni/Qatari cyclist Ahmed Al-Bardiny / Ahmed Albourdainy

There's something weird going on at Ahmed Al-Bardiny (Q21004777) and Ahmed Al-Bardiny (Q114868038). There's an edit war going on here and on ENWP, but I think there may actually be a conflation between two (or more) people. Could someone with more subject-matter expertise please take a look? CC @GAN, YoaR, Hddjjd, Seacactus 13, Hasan muntaseer. Bovlb (talk) 16:20, 26 October 2022 (UTC)

  • @Bovlb: Now in the element Q21004777 in the identifiers section there are two links to velostatistics sites. They show a rider who was born on 04/04/1990 and has Qatari citizenship. — GAN (talk) 16:38, 26 October 2022 (UTC)
    Does the Yemeni cyclist exist and pass notability criteria? Bovlb (talk) 16:40, 26 October 2022 (UTC)
    @Bovlb: On en-wiki, Yemeni is now listed as "Ahmed Abdulhakim Ahmed Al-Hemyari (born May 15, 2002)". In his achievements, the race of 2013 is also indicated there --- at 11 years old ??? We managed to find a maximum of three riders with the name "Ahmed" and Yemeni citizenship. But I did not find any significance in cycling among them, even taking into account the use of Module:Cycling race.
    There is one link on the en-wiki in the "References" section, but it leads to "youtube.fandom.com/wiki/Ahmed_Al-hemyari".
    — GAN (talk) 17:15, 26 October 2022 (UTC)
    Can protection against unregistered put? — GAN (talk) 17:19, 26 October 2022 (UTC)
    I have added two weeks of semi-protection on both items and the ENWP article. Bovlb (talk) 17:30, 26 October 2022 (UTC)
    I am the creator of this article. This article is for a Qatari cyclist born in 1990, not a Yemeni cyclist born in 2002. These accounts have never made any previous contributions and are clearly vandals. I have reverted it back to its original state. If there are actually any others of a similar name out there, then overwriting the original article is not at all the proper way to go about creating a page for them. Seacactus 13 (talk) 18:01, 26 October 2022 (UTC)

Guidance on Notability / Inclusion

Hello...

Frequent contributor, first time poster. I was looking for some guidance on "notability."

We use Wikidata for a number of things, but for certain key topics (public companies, government agencies) and people (elected officials, public company C-level executives), if we don't find a wikidata ID after checking twice, we add it with appropriate links.

We're very comfortable that elected officials, candidates, public companies, federal and top administrative-level agencies meet criteria for inclusion. But we were looking for guidance on two questions we have:


1. Finance / Wall Street. Would a stock analyst -- someone who writes reports and asks questions that has influence on the value of a public company -- be acceptable for inclusion in Wikidata?

2. Similarly, is there a "floor" for elected officials / public figures? Would a city council member in Sacramento be notable? What about an elected official in Tyler, Texas? In finance, a CEO / CFO of a public company would meet the criteria, but what about the head of investor relations? What about financial reporters?


We've reviewed the guidelines here: https://www.wikidata.org/wiki/Wikidata:Notability. We can make the case, say, a stock analyst fulfills:


It refers to an instance of a clearly identifiable conceptual or material entity. The entity must be notable, in the sense that it can be described using serious and publicly available references.

It fulfills a structural need, for example: it is needed to make statements made in other items more useful.


...in terms of providing greater depth in Wall Street knowledge graphs, or for newspapers / news shows, or municipal coverage. But we didn't want to be flagged for spamming for adding folks like this if it was outside the mission, or if the data was too granular. We wouldn't add anything that couldn't be backed up with source links and material.


Guidance would be appreciated. On these two points, we're asking before doing to be safe. Thank you. Unclejohann (talk) 14:06, 26 October 2022 (UTC)

I think we need to accept that everybody is notable. Not only are people famous for 15 minutes, they are related to famous people, or are notable within their communities. I am conscious when adding guest of honour lists for science fiction conventions or award winners that while authors are well covered, as are some scientists and fans, others are missing, and their omission is confusing and disrespectful of the status of a guest of honour or award winner. But while I cannot justify a new entry just for that item, they might well appear at other events or be councillors or stock analysts too. And like you cannot have a non-interesting number, you could always extend the notability threshold a little further. Vicarage (talk) 14:30, 26 October 2022 (UTC)
@Vicarage Everybody is indeed notable, if they can be described using serious and publicly available references. It’s the wrong question to ask whether somebody deserves an item – of course they do, we are all human beings with our unique stories that should be respected and cherished. But this is somehow beside the point: It’s not enough for some user to claim something, we need independent proof. That (and the necessary user base to keep all those items free of bias and vandalism) seems to be the main point why we care about notability after all. --Emu (talk) 16:44, 26 October 2022 (UTC)
I've always been suspicious about wikipedia's desire for references. All sorts of untruths are in published material, and most facts can be supported somewhere. But notability is about importance, not verification, and while a parish councillor or employee could be confirmed in role with references to organograms, its a question of whether this is interesting or clutter. The constant stream of deletion requests can't just be because people puff themselves up in the wrong way. The issue is how would the WD dataset be degraded if there were 10,000 Michael Jacksons in it and you couldn't find the right singer, coder or beer critic in a dropdown. If notability was a ranking, so the famous rose to the top, and a hundredfold increase in entities didn't phase the software, I'd be very happy to add in everybody, but I'd need help to know that the secretary of a village council worked as a stock analyst and was a SF fan in their spare time. It seems awkward that I'd be happy to document people with unusual names, but not common ones. Can anyone familiar with the database comment on the practicality of inclusiveness? Vicarage (talk) 17:21, 26 October 2022 (UTC)
I’m not exactly sure what you are trying to say, to be honest. --Emu (talk) 20:19, 26 October 2022 (UTC)
In general, if there's a public interest for having data about a person and there's already data out there that allows us to source items, than the person is notable for the purposes of Wikidata. ChristianKl ❪✉❫ 15:16, 26 October 2022 (UTC)
  • I agree that eventually the most famous or well known people will get diluted as the number of entries grow. We have considered adding all of Findagrave and other large databases in the past and the worry was about dilution. Currently the lowest Q-numbers appear first in our index and eventually we will figure out other ways to make, newer-famous people rise to the top of the index when a search is done. That shouldn't stop us from adding new entries. --RAN (talk) 20:33, 26 October 2022 (UTC)

Notability shouldn’t be a question of the availability of references… but inclusion into Wikidata shouldn’t be a question (solely) of notability… --2A02:8108:50BF:C694:3C5C:2B2:84A2:6682 20:45, 26 October 2022 (UTC)

  • To quote User:Emu: "I’m not exactly sure what you are trying to say, to be honest." --RAN (talk) 21:34, 26 October 2022 (UTC)
    Whether someone/something is notable should not depend on whether there are references available to verify claims about that subject; but if such references are not available, then inclusion of the subject into Wikidata should not happen, even if it is notable. (Otherwise we are collecting claims that may as well be entirely made-up.) --2A02:8108:50BF:C694:3C5C:2B2:84A2:6682 21:42, 26 October 2022 (UTC)
    Within Wikidata notable means what our notability policy means, regardles of people somewhere else consider to be notable. ChristianKl ❪✉❫ 14:58, 27 October 2022 (UTC)
    That’s why “should”. The entire comment was opinion-based. It further wasn’t meant to endorse any specific notion of notability, but rather make a point about the principles by which a specific notion of notability could/should be constructed (no idea whether Wikidata’s follows them). --2A02:8108:50BF:C694:79FB:19DA:5E58:94A6 20:08, 27 October 2022 (UTC)

Description

Is "none" used as a description for a Wikidata item, like this? Saroj Uprety (talk) 18:20, 28 October 2022 (UTC)

No. --Emu (talk) 18:49, 28 October 2022 (UTC)

has written for (P6872)

Are there best practices for how to model and format has written for (P6872)? If it's an online periodical, and the author has a dedicated author feed/profile, I generally qualify the statement with URL (P2699) or an equivalent property such as:

Thus, the URL (or property) is both a self-evident reference and a visible, useful link to the subject's writing for that publication. A reference using the same URL might be added (even if redundant).

When author feeds are not available, I simply use a third party reference verifying the claim, such as Paul Sexton (Q114835138)has written for (P6872)The Sunday Times (Q221986)Rock's Backpages author ID (P6462)paul-sexton.

What I'm basically asking is whether using URL (P2699) as a qualifier (rather than, or in addition to, a reference) might be problematic somehow, e.g. from a querying/coding/data management aspect, or be considered spammy or promotional. Personally I appreciate having the link more visible (not hidden as a reference), as being a human (not a machine), convenience is nice. Thoughts? -Animalparty (talk) 20:03, 24 October 2022 (UTC)

I support the URL as a qualifier in such a circumstance, don't see it as problematic, do see the benefit users get from it. Qualifiers add context and value to statements, and this is a good example. --Tagishsimon (talk) 20:13, 24 October 2022 (UTC)
On the contrary, I personally don't think these URLs should be included as qualifiers. They're not really qualifying the statement. The URL may instead belong to independent described by source (P1343) statements, or, in the case of the New York Times, there is already an identifier than can be included in the item as an independent P8297 statement. Vojtěch Dostál (talk) 11:04, 29 October 2022 (UTC)

Unanswered question

I have posted the question in wrong page? 2001:B07:6442:8903:3895:A2FF:B208:B1A3 15:13, 28 October 2022 (UTC)

It's probably just that nobody had time to answer. Please don't take in personally, we're all volunteers. Vojtěch Dostál (talk) 15:16, 28 October 2022 (UTC)
Some people don't bother creating an account and some people don't bother replying. Two might be related or not. Multichill (talk) 10:12, 29 October 2022 (UTC)

Happy 10th birthday, Wikidata! 🎉

Hello all,

As you may know, Wikidata went live for the first time on October 29, 2012, allowing thousands of people to create items, populate the wiki and create the amazing free knowledge base that we know today. On Wikidata's 10th birthday, we would like to address many thanks to everyone who has been involved with Wikidata over the past years: long time editors and fresh new contributors, people carefully curating the data and the ontologies, people creating tools, people reusing the data, people adding content in many different languages, people advocating for Wikidata in various places, and many others whose contribution made Wikidata what it is today.

Now is time to celebrate! Here's what you can do to participate in the celebration:

  • Write a message on the birthday presents & messages page, where you will also find the gifts that some people prepared for the Wikidata community and a message from the development team
  • Join one of the many events organized by local Wikidata groups, onsite all around the world or online. You can for example join one of the Wikidata online meetups taking place at 08:00 and 17:00 UTC on Jitsi.
  • Read some articles that were published for the occasion, and watch the collaborative birthday video showcasing Wikidata community members from all around the world
  • Join the discussions on social media using the hashtags #WikidataBirthday and #Wikidata10.

Happy 10th birthday Wikidata, and thanks for being part of this amazing project! Cheers, Lea Lacroix (WMDE) (talk) 06:59, 29 October 2022 (UTC)

Mix'n'Match for P3124

Some months ago one of our Polish collegues imported to Mix'n'Match part of the Polish Science (Q97032597) database, for which we have our Polish scientist ID (deprecated) (P3124) property. To be precise, he imported only the female scientists. Could someone please modify or otherwise extend this import to include all scientists from that source, which is considered highly reliable here in Poland? I don't feel technically competent enough to do it on my own. Thank you in advance. Powerek38 (talk) 09:47, 29 October 2022 (UTC)

Ship prefix templates

My English Wikipedia day began by observing Ginkos520isthebest's actions in creating bare item Template:TSS (Q114901211) from an (English?) Wikipedia Template bad delivered me a notice. While I am indef block'ed on that place I did investigate and memories flooded back of the Maud and I wondered why that was used on the template and then I remembered it was on that famous ship on which I once travelled over 50 years ago to be sick after an enjoyable cod and chips meal. Back to this place I observed the item did not have an instance of or description. I checked out Template:SS (Q13447040) and was able to gather what to use as an instance of (P31) there, albeit I personally think the TSS can stand for Turbine Steam Ship but en:Ship prefix does not confirm this. Anyway due to indef block I cannot ask at the :en WikiProject for help.-- User:Djm-leighpark(a)talk 09:27, 29 October 2022 (UTC)

Anyway to get the point I am minded while I am still waking up before doing other stuff that there would be some usefulness of link ship templates such as Q114901211 and Q13447040 together in this place and wondered if and how there is a way or best practice for doing this. Thankyou. -- User:Djm-leighpark(a)talk 09:27, 29 October 2022 (UTC)
✓ Done: No answers so I've had another look at my own stupid question. There may be some traction around template or module that populates category (P4329) or its inverse, but I'm now old and its too late and everyone's partying & best wishes to them. I'm just festering in isolation like a leper. I've probably worked it out ... if I didn't ask someone my stupid question I'd probably not have worked it out for myself. Might need to consult a singer or a snake to get some tooling. -- 23:19, 29 October 2022 (UTC) User:Djm-leighpark(a)talk 23:19, 29 October 2022 (UTC)

review score (P444) and point in time (P585)

Should P585 be the time the last review was posted or the time the review score were updated on WD?--Trade (talk) 21:32, 30 October 2022 (UTC)

The former. Date retrieved in the references is best for the updted on WD date. --Tagishsimon (talk) 21:50, 30 October 2022 (UTC)

cannot update photo

I am trying to update the photo of Canadian Foreign Minister Joly (using File:Secretary Blinken Meets with Canadian Foreign Minister Joly and Mexican Foreign Minister Ebrard (52138406616).jpg) but get the following error msg:

Could not save due to an error.File names are not allowed to contain characters like colons or slashes

Thanks in advance Ottawahitech (talk) 20:05, 27 October 2022 (UTC)

Ottawahitech If you are trying to update the value image (P18) for the item Mélanie Joly (Q15054353) you simply need to omit the "File:" prefix. So in this case simply "Secretary Blinken Meets with Canadian Foreign Minister Joly and Mexican Foreign Minister Ebrard (52138406616).jpg". Also, in general, consider that there may not be a compelling reason to add or substitute new image, so long as any existing image is a high quality, representative image of the subject. Newer doesn't always mean better. -Animalparty (talk) 00:32, 28 October 2022 (UTC)
@Animalparty, Thank you for answering my question AND pinging me.
This is probably not the right place to discuss if photo's should be updated? Ottawahitech (talk) 20:20, 31 October 2022 (UTC)

Wikidata weekly summary #544

Here's your quick overview of what has been happening around Wikidata over the last week.
  • 🧰 Tool of the week
    • sparklis (source) : query builder in natural language that allows people to explore and query SPARQL endpoints with all the power of SPARQL and without any knowledge of SPARQL, nor of the endpoint vocabulary.
  • ✨ Other Noteworthy Stuff
    • Word Finder is a tool based on Wikidata Lexemes that solves anagrams and find missing letters in crosswords.

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.

best practices for adding/changing images

I think the image for video game (Q7889) could be improved, as it features people playing video games rather than a video game itself. What do I need to consider before changing it? Nivekuil (talk) 18:20, 31 October 2022 (UTC)

there can be more than one image. also the images need to be properly licensed so most video game screen shots aren't acceptable. BrokenSegue (talk) 22:41, 31 October 2022 (UTC)
Ah, but should there ever be more than one image (P18)? WD is not Commons, P18 not a dumping ground, and other image properties exist. (As to the instant issue, perhaps bold, revert, discuss. [16]) --Tagishsimon (talk) 22:51, 31 October 2022 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2022/10&oldid=1767066833"