Wikidata:Project chat/Archive/2022/08

From Wikidata
Jump to navigation Jump to search


Date of update/check[edit]

In some sources there is an important information: date of updating the source or checking its completeness/correctness/... In the past I used point in time (P585), but now there is property scope constraint (Q53869507) that limits its use to the main value and qualifier. I've changed some value to publication date (P577), but does not seem correct, as in some cases there is both 'publication date' and 'update date'. How this should be done? Is there a more specific property for this? Wostr (talk) 19:24, 4 August 2022 (UTC)[reply]

@Wostr: Would last update (P5017) work? Shinnin (talk) 19:31, 4 August 2022 (UTC)[reply]
@Shinnin: perfect... it's a pity it's not linked from any other property, so I couldn't find it. Thanks. Wostr (talk) 19:59, 4 August 2022 (UTC)[reply]
This section was archived on a request by: Wostr (talk) 19:59, 4 August 2022 (UTC)[reply]

Inconsistency of data for SPARQL result and the webpage?[edit]

Hello. I try to understand the discrepancy I come across.

This property page for creator (wdt:P170) lists 15 related properties (wdt:P1659), but if I make the following SPARQL query, I got 13 related properties: 

SELECT ?item ?itemLabel
{
  ?item wdt:P1659 wd:P170 .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE]" }
}
Try it!

Any idea why there is discrepancy? Cheers! 2001:610:450:40:0:0:6:1032 10:19, 27 July 2022 (UTC)[reply]

Because the query is for other items that link to creator (P170). The question you are asking is this query
SELECT ?item ?itemLabel
{
  wd:P170 wdt:P1659 ?item .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE]" }
}
Try it!
Which does return 15 items. Multichill (talk) 12:17, 27 July 2022 (UTC)[reply]
@Multichill Thank you! It was overlooked! 2001:610:450:40:0:0:0:1035 10:44, 1 August 2022 (UTC)[reply]

Interwiki links for numeric digits and early AD years are completely messed up[edit]

I realised that several Wikidata items like 1 (Q23108) (Year AD1 in Gregorian calendar) and 1 (Q199) (Number 1) are just completely messed up. This continues until at least number/year 100. Each one has hundreds of interwikis, and I don't even understand some languages too. How to fix all of that, I have no idea. But someone needs to fix it somehow. Thanks! CX Zoom (talk) 10:30, 31 July 2022 (UTC)[reply]

The way to fix it would be to use google translate on all the relevant interwiki links and then move them to the correct item. ChristianKl ❪✉❫ 10:34, 1 August 2022 (UTC)[reply]

Wikidata’s 10th birthday in October: call for community events[edit]

Hello all,

As you may know, Wikidata went live on 29 October 2012 ; in 2022, we are celebrating 10 years of Wikidata together! We would like to organize celebration events all around the world, and to create a huge network of decentralized, local and community-led events, that will take place onsite or online, around October 2022.

The goal of these birthday celebrations are to celebrate the achievements of the community, to bring people together, and also to talk about Wikidata to the rest of the world in order to get more people onboard. In various areas of the world, people get together to organize plenty of different birthday events: meetups, workshops, discussions, live streams, editing campaigns… How about you get involved in organizing such an event with your local community?

How to organize an event?

On the Tenth Birthday page, you will find resources to learn how to plan and schedule an event. You will find a few tips to get started, some ideas of events formats and the visual kit to create your own communication material.

If you are interested in organizing an event or an activity for Wikidata’s 10th birthday, we recommend you get in touch with other organizers to get inspired by their experiences and ask for support if needed.

How to get funding for your activities?

WMF and Wikimedia Deutschland are collaborating on a special Rapid Fund campaign for Wikidata’s birthday. If you need financial support to run an event, you can now send an application to receive funding up to 5,000 USD.

Feel free to check the details of the funding campaign, eligibility criteria and supported activities. The program covers a wide range of costs that can be funded: renting space or equipment, internet connection, transport and accommodation, food and refreshments, printing merchandise and communication materials, prizes for contests, and more.

To make sure that your application is processed in time for your event, don’t wait too long to submit your application! The Wikidata birthday events are typically held between mid-September and mid-November, and you should send your application at least 45 days before your activity starts.

If you have questions or need support to apply for funding, feel free to get in touch - we will also organize calls for organizers starting at the end of August.

We will post news about the Wikidata birthday regularly on this talk page and on social channels, as more exciting projects are coming soon: a collaborative celebration video and some blog posts about Wikidata.

If you have any questions, if you want to get involved in Wikidata’s 10th birthday or if you have suggestions, feel free to reach out to me.

Best, Lea Lacroix (WMDE) (talk) 06:50, 1 August 2022 (UTC)[reply]

Property documentation page: show P6607 (constraint clarification) for each property constraint documented[edit]

I was thinking that some property constraints as listed on property documentation pages are usually not obvious to users as to why the constraint exists, and what should be done to resolve the constraint. Constraints listed at country (P17) and YouTube channel ID (P2397) (as examples) are much easier to understand thanks to qualifiers replacement property (P6824), replacement value (P9729) and constraint clarification (P6607) than the generic constraint documentation provided on Property_talk:P17 and Property_talk:P2397 (which don't use any of these helpful qualifiers).

To improve the situation, one or more constraint clarification (P6607) values that may exist in the user's chosen language could be obtained by Module:Constraints and passed as a list to Template:Constraint/text so that property documentation pages can display some help to users on the reason for the constraint occurring. If the resolution is fairly simple (e.g. a case of needing to use an alternative property or value) then Module:Constraints would also pass replacement property (P6824) and/or replacement value (P9729) to Template:Constraint/text so that the constraint documentation can make it obvious to users that constraint violations are easily resolved by using an alternative property or value.

Do you have any thoughts on the above suggestion, or would you be able to assist with implementation of the improvements?

-- Dhx1 (talk) 14:54, 30 July 2022 (UTC)[reply]

I'm certainly in favor of clarifying constraints and how they apply in specific situations, but I don't fully understand what you are suggesting. Could you give more specific or expanded examples of the better situation on country (P17)? JesseW (talk) 23:22, 31 July 2022 (UTC)[reply]
The country (P17) none-of constraint applying to item Grand Principality of Finland (Q62633) is one such example. I hadn't noticed until now, but @Jura1: had already added replacement property (P6824) and replacement value (P9729) to the description of the none-of constraint that appears on the property documentation page (example being country (P17) talk page). This could be extended to the conflicts-with constraint that also uses replacement property (P6824). What would additionally help document all property constraints is adding why the constraint exists. This is done at the moment via the constraint clarification (P6607) qualifier which could be added to the property constraint documentation when the constraint clarification (P6607) language matches the user's language. Dhx1 (talk) 00:48, 1 August 2022 (UTC)[reply]
OK, let me see if I've got it now. For the example you gave; if someone added country (P17) of

Grand Principality of Finland (Q62633) (as I did on the Sandbox item in this edit), they would see a constraint violation message:

none-of constraint
The value for country should not be either Grand Duchy of Finland or Congress Poland.
But they wouldn't be prompted to convert it to the replacements given, or shown the constraint clarification (P6607) message. You are suggesting that the code that shows the constraint violation message be modified to include those? If so, I'm enthusiastically in favor. JesseW (talk) 02:55, 1 August 2022 (UTC)[reply]
Or were you just suggesting that the code for displaying the documentation on the Property_talk page be improved to display the constraint clarification (P6607) values? That also sounds good, if less of a big win than displaying it in the constraint violation messages. JesseW (talk) 02:59, 1 August 2022 (UTC)[reply]
My proposal was just to update the property talk page constraint documentation as this seems fairly simple to do for starters. But I do agree, having more information in the constraint popup messages would be a great help too. It looks like those popup messages are hardcoded into the WikibaseQualityConstraints extension though and would need extension code to be updated to pass extra information to ViolationMessage to print in the case of constraints such as [1]. Dhx1 (talk) 15:57, 1 August 2022 (UTC)[reply]
Great, let's go for updating the talk page templates. I'll see if I can puzzle out how to do that. JesseW (talk) 17:48, 1 August 2022 (UTC)[reply]
I totally agree that it would be good to show constraint clarification (P6607) on the property talk page.
Ideally it would also show whenever constraint violations happen instead of the generic text for the constraint violation. ChristianKl ❪✉❫ 10:59, 1 August 2022 (UTC)[reply]

For implementing the improvement to the talk page descriptions, the relevant function seems to be makeConstraintDocumentation which uses {{Constraint/text}}. I suspect we don't need to modify that template, and can just put the additional info as part of the "description" parameter. I'm somewhat lost, so I'm going to leave it up to others to try further at this point. JesseW (talk) 18:03, 1 August 2022 (UTC)[reply]

Wikidata weekly summary #531[edit]

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week

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.

Work together with OpenAI to build free gpt-3 SPARQL query generator into the Query Service[edit]

I was wondering if GPT-3 was capable of generating SPARQL queries compatible with wikidata and turns out it actually works:

https://twitter.com/LukasGalke/status/1552995733527318529

You can try it out for yourself in the gpt-3 playground on their website: https://beta.openai.com/playground/p/QXa2sdTmtHIhdStWtOajx1lA

But obviously that's something that people generally know how to do. So having a text box built into the Query Service that allows anyone to do this, would be very beneficial for actually exploring wikidata for someone that doesn't know how SPARQL works or can't figure out (limited by) the Query Builder.

Since that would need to use the GPT-3 API, it won't be free unless there is a collaboration between Open-AI and Wikidata I guess.

BY NO MEANS are the things it generates perfect. Complex queries might not work even. But having the option to use this alone would be pretty interestingThibaultmol (talk) 13:16, 29 July 2022 (UTC)[reply]

It's a very interesting experiment. Personally I'm a little bit suspicious with GPT-3 since it is a proprietary tool and you need a login even to access to the demo. I think it would be interesting to try the same think with the new BLOOM model which has been released by the Big Science Initiative and is available on the Huggingface platform (https://huggingface.co/bigscience/bloom). Anyway it would be good to have a wikipage relative to this topic. @Thibaultmol: - PAC2 (talk) 06:07, 2 August 2022 (UTC)[reply]

Run Wikidata Query failed on Toolforge server[edit]

Hi. I've set up a K8s job on Toolforge and I got errors when making a query on Wikidata. While the query could be run on my local computer, it could not be on the Toolforge server.

The SPARQL is: 

# Modified SPARQL. Original credit: https://www.wikidata.org/wiki/User:Vojt%C4%9Bch_Dost%C3%A1l
SELECT DISTINCT ?item (SAMPLE(?label) as ?itemLabel) WHERE {
  ?item wdt:P31 wd:Q101352 .
  ?item wdt:P1705 ?label .
  FILTER (lang(?label) = 'en')
} GROUP BY ?item
Try it!

It requires all surnames in English on Wikidata.

The error was:

ERROR: An error occurred for uri https://query.wikidata.org/sparql?query=PREFIX%20wd%3A%20%3Chttp%3A//www.wikidata.org/entity/%3E%0APREFIX%20wdt%3A%20%3Chttp%3A//www.wikidata.org/prop/direct/%3E%0A%0A%23%20Modified%20SPARQL.%20Original%20credit%3A%20https%3A//www.wikidata.org/wiki/User%3AVojt%25C4%259Bch_Dost%25C3%25A1l%0ASELECT%20DISTINCT%20%3Fitem%20%28SAMPLE%28%3Flabel%29%20as%20%3FitemLabel%29%20WHERE%20%7B%0A%0A%20%20%3Fitem%20wdt%3AP31%20wd%3AQ101352%20.%0A%20%20%3Fitem%20wdt%3AP1705%20%3Flabel%20.%0A%20%20FILTER%20%28lang%28%3Flabel%29%20%3D%20%27en%27%29%0A%0A%7D%20GROUP%20BY%20%3Fitem WARNING: Waiting 120 seconds before retrying.

It seems to be a network error to me. Does anyone have clues on this? Feliciss (talk) 13:10, 1 August 2022 (UTC)[reply]

Hi. I've solved this question after I increased CPU to 1 and memory to 4Gi on Toolforge server. Feliciss (talk) 07:26, 2 August 2022 (UTC)[reply]

Anime and Manga project[edit]

I'm planning to create an anime and manga database frontend using wikidata as source data. I'm not a fan of how existing anime and manga databases (like myanimelist, anidb, anime-planet, kitsu etc.) are not public domain. It means every single one is different, and adding or editing any information has to be done separately for each database. This is in addition to their own idiosyncrasies like not including recap episodes and movies for example, or having an unusually stringent add/edit process.

What I want is for the core data to be FOSS/public domain so that anyone can build a service around it. I've looked through a few instances of Anime television series (Q63952888), and it seems like it could use a lot of work.

Here is my plan so far:

  1. Query and store all instances of anime (Q1107) locally.
  2. Build a prototype frontend to display basic data like names, descriptions, genres, cast, voice actors, identifiers, and see how I feel about the project in general.
  3. Make a schema of what additional data I want to pull from wikidata to display.
  4. Try to incorporate those in the frontend.
  5. Use manami's anime-offline-database to create missing anime, and update identifiers for existing anime.
  6. Scrape various sources to compile even more information, and update wikidata accordingly.
  7. ...
  8. Worry about manga, then possibly move on to TV shows and movies.

Context[edit]

I have some experience with making/using a bot to update a wiki. I've done it in the past on the ICotFR Wiki, before the gamepedia/fandom merge. (Contributions)

The tool I built to do it lives at https://git.sr.ht/~txtsd/idlechampwiki. It might be spaghetti, but it did its job well at the time.

I was told on Discord to inform y'all here before I start such a project, so here I am. I have some difficulty organizing and articulating thoughts, but feedback, criticism, and questions are welcome. --txtsd (talk) 07:04, 30 July 2022 (UTC)[reply]

Doing something similar, I found using Mediawiki with the ExternalData extension useful. Mediawiki has lots of display tools, and the ExternalData extension allows you to read directly from Wikidata, or from your own local pages (I used json ones populated with wikidata and then expanded with my extra content). A local cache allows you better control of response time. Mediawiki has lots of display tools like timelines and tabbed pages you might find useful. Vicarage (talk) 08:02, 30 July 2022 (UTC)[reply]
Thanks for this info! The frontend won't be a mediawiki instance, but I can definitely see myself using this elsewhere! --txtsd (talk) 08:11, 30 July 2022 (UTC)[reply]
Notified participants of WikiProject Anime and Manga (this might be of interest to you) - Valentina.Anitnelav (talk) 09:43, 30 July 2022 (UTC)[reply]
Thanks for pinging the group! --txtsd (talk) 12:50, 30 July 2022 (UTC)[reply]
That is very cool, thanks for doing that! As possible inspiration, you can look at vglist (Q96096761) which is a video games database frontend on top of Wikidata.
One likely hurdle you will encounter is that anime and manga items on Wikidata are very often highly conflated − it’s not uncommon to have the manga, TV series, OAV, movie, one or two games etc. all in one single item (see eg A Town Where You Live (Q2819940)). -- Jean-Fred (talk) 11:53, 30 July 2022 (UTC)[reply]
Oooh I'm not sure why I couldn't find vglist when I went looking for a games tracking service several months ago. Thanks for showing me it.
I've noticed the conflation of anime and manga while looking through several items. Is it like that by design? Are they supposed to be together? Because a manga cannot have voice actors, for example, but they'll be listed if the item is an instance of both anime and manga. --txtsd (talk) 12:50, 30 July 2022 (UTC)[reply]
Most Wikipedia only have a page for entire series and that is usually the wikidata item being created.
For the reason, and from the perspective of existing adoption, I think it would be better to base off MALsync tool which offer cross platform data tool. C933103 (talk) 16:44, 30 July 2022 (UTC)[reply]
Ideally each publication (actually, also each edition of a publication) would have separate items, but you'll always have to contend with instances where this isn't (yet) the case. Arlo Barnes (talk) 22:18, 30 July 2022 (UTC)[reply]
I agree with Arlo above about the ideal situation. As well as C933103 about how because many wikipedias just have a page for the series data gets conflated. There has been effort on Wikidata though to separate that conflated data out at minimum into the different medias that the single series page may contain: light novels, manga, animes television series, anime movies, and games. Usually with the originating work linking the others as derivatives. -Jeanjung212 (talk) 09:42, 31 July 2022 (UTC)[reply]
Hmm this seems like a larger effort that needs to be coordinated well, although Nicereddy's script looks promising! --txtsd (talk) 03:36, 3 August 2022 (UTC)[reply]
This is awesome! I'm the developer of vglist, which Jean-Fred kindly mentioned already :)
I actually have a very rough draft of a Ruby script I wanted to use in the near-ish future for splitting anime and manga items that are currently merged into one: https://gist.github.com/connorshea/53d65ae07e912a5904f7767fc3e1ce49
Not sure if that'd be useful for you, but I noticed anime/manga items are a mess right now and have wanted to fix it for a while.
If you want to pick my brain, since I've had to solve similar problems with vglist, feel free! Nicereddy (talk) 01:04, 31 July 2022 (UTC)[reply]
YAS! If we can re-write that script in Python, I'm sure we can coordinate a fair effort to get the items split appropriately in addition to matching them up with their counterparts on Wikipedia and elsewhere. --txtsd (talk) 03:36, 3 August 2022 (UTC)[reply]
Apparently there's also Wikipedia article covering multiple topics (Q21484471), which we could potentially use instead of media franchise for the shared items.
I've updated the GitHub Gist to include the MIT License, so if anyone wants to port this to Python they're more than welcome :) I know very basic Python, but it's definitely not my area of expertise, so I'd need someone else to do most of the work on that 😅 Nicereddy (talk) 14:05, 3 August 2022 (UTC)[reply]

Granularity of items / Best to merge or to split ?[edit]

Hello, I've started a discussion whether it is a good idea or not to get complicated things splitted/merged ? I know WD:Bonnie and Clyde and what I understand is that it's better to split+add interlinks than to merge. Querying multiple-topics merged items can be misleading (what it's coordinate location (P625), according to the searched topic, for instance) ? Can infobox handle and show the correct data for a specific topic? Bouzinac 💬●✒️●💛 07:10, 3 August 2022 (UTC)[reply]

It's generally good to err into the direction of splitting as that makes the ontology clearer. ChristianKl ❪✉❫ 14:09, 3 August 2022 (UTC)[reply]

Movement Strategy and Governance News – Issue 7[edit]

Movement Strategy and Governance News
Issue 7, July-September 2022Read the full newsletter

Welcome to the 7th issue of Movement Strategy and Governance News! The newsletter distributes relevant news and events about the implementation of Wikimedia's Movement Strategy recommendations, other relevant topics regarding Movement governance, as well as different projects and activities supported by the Movement Strategy and Governance (MSG) team of the Wikimedia Foundation.

The MSG Newsletter is delivered quarterly, while the more frequent Movement Strategy Weekly will be delivered weekly. Please remember to subscribe here if you would like to receive future issues of this newsletter.

  • Movement sustainability: Wikimedia Foundation's annual sustainability report has been published. (continue reading)
  • Improving user experience: recent improvements on the desktop interface for Wikimedia projects. (continue reading)
  • Safety and inclusion: updates on the revision process of the Universal Code of Conduct Enforcement Guidelines. (continue reading)
  • Equity in decisionmaking: reports from Hubs pilots conversations, recent progress from the Movement Charter Drafting Committee, and a new white paper for futures of participation in the Wikimedia movement. (continue reading)
  • Stakeholders coordination: launch of a helpdesk for Affiliates and volunteer communities working on content partnership. (continue reading)
  • Leadership development: updates on leadership projects by Wikimedia movement organizers in Brazil and Cape Verde. (continue reading)
  • Internal knowledge management: launch of a new portal for technical documentation and community resources. (continue reading)
  • Innovate in free knowledge: high-quality audiovisual resources for scientific experiments and a new toolkit to record oral transcripts. (continue reading)
  • Evaluate, iterate, and adapt: results from the Equity Landscape project pilot (continue reading)
  • Other news and updates: a new forum to discuss Movement Strategy implementation, upcoming Wikimedia Foundation Board of Trustees election, a new podcast to discuss Movement Strategy, and change of personnel for the Foundation's Movement Strategy and Governance team. (continue reading)

MNadzikiewicz (WMF) (talk) 13:40, 3 August 2022 (UTC)[reply]

Autobiography[edit]

When you create an entry for an autobiography should you both author=Joe_Smith and topic=Joe_Smith (already implied)? When I look at entries we have a mixture of the two. RAN (talk) 18:30, 31 July 2022 (UTC)[reply]

The goal is to allow queries based only on main subject (P921) and queries based only on author (P50). See WD:Books for help with bibliographic properties. — Metamorforme42 (talk) 19:41, 31 July 2022 (UTC)[reply]
I guess Richard Arthur Norton (1958- ) is asking about the inherent "duplication" of one information. When something is an autobiography, it is already implied that the main topic is the author (P50). Still, I would insert both. There are definitely cases where a book is written by two authors, one is the book's theme and the other can be a journalist helping on the book. Vojtěch Dostál (talk) 08:47, 3 August 2022 (UTC)[reply]
Note that there could be a actual book written by a real author which is a fictional autobiography of a fictional character - is it qualifying for Q4184? Or where it was authored by a ghost writer. Or where authorship is disputed Mateusz Konieczny (talk) 11:52, 3 August 2022 (UTC)[reply]
  • I added both, I know some purists believe implied information should not be redundantly added, the database should be as streamlined and as terse as possible. --RAN (talk) 15:33, 3 August 2022 (UTC)[reply]

Self-contained export of items[edit]

Hi! The Linked Data Interface provides a way to export entities, and the default rfd option provides a self-contained/complete export of an entity (i.e., all properties, claims, etc. can be reconstructed from this file). However, the rdf format is not always optimal. I am therefore wondering if there are any alternative ways to export a self-contained/complete description of an item? Such as a json file with a main entity and also all entities related to the main entity? -- 45.67.9.165 14:04, 3 August 2022 (UTC)[reply]

The page you linked to shows how to get a JSON file: [2]. As for a "complete description", that's subjective; i.e. I presume you want to get the linked entities as well, but how deep do you want to go? I don't know whether there's an option to request linked entities as well, but you could always do it programmatically. Silver hr (talk) 18:18, 3 August 2022 (UTC)[reply]
Thanky you (it was I who asked); I will explore the different formats more carefully. For example, it seems that the .jsonld contains more information than the .json files. Ajarmund (talk) 19:46, 3 August 2022 (UTC)[reply]

Add category of managers in the team item[edit]

We are adding category of players in the team item with this way: Liverpool F.C. (Q1130849) -> category for members of a team (P6112) -> Category:Liverpool F.C. players (Q7135717)

How we can add the category of managers? Liverpool F.C. (Q1130849) -> ??? -> Category:Liverpool F.C. managers (Q6229420)

Philocypros (talk) 23:02, 3 August 2022 (UTC)[reply]

Q6229420 is already linked to Q1130849. That is amply sufficient. Perhaps let go of the idea that the team item needs to point to ... all other items and accept that other items pointing to the team item will suffice. I'm sorry, but not surprised, to see we have P6112. Wholly unnecessary property. --Tagishsimon (talk) 23:09, 3 August 2022 (UTC)[reply]

You can now check out the current development state of the upcoming REST API for Wikibase[edit]

Hello everyone,

We are working on making it easier to access Wikidata’s data programmatically so that more people can build great applications on top of our data. As part of this we are building a new REST API for Wikibase. A while ago we asked for feedback on the draft specification for that new API (see previous communication). It is in development now. While it is not in the state which we would like to run on Wikidata, you can now have a look at the current development state on Beta Wikidata and follow along as we build out the API. You can find the current state of the specification we are working from at https://doc.wikimedia.org/Wikibase/master/js/rest-api/. (Small caveat: The specification might sometimes be slightly ahead of the code development and contain functionality that still needs to be implemented.)

Please note that the API is still in the experimental development state. There might be changes in the functionality available and API’s behavior that we will not announce separately.

The current state can already do the following:

  • Retrieve the data of an Item with GET /entities/items/{item_id} and filter what fields (i.e. type, labels, descriptions, aliases, statements, sitelinks) are returned when reading the Item data
  • Retrieve all statements of an Item with GET /entities/items/{item_id}/statements
  • Retrieve the data of a single statement of an Item with GET /entities/items/{item_id}/statements/{statement_id} or GET /statements/{statement_id}
  • Conditionally request the data only if it has changed since the specified revision/timestamp (using If-None-Match, If-Modified-Since HTTP headers)
  • Create a statement on an Item with POST /entities/items/{item_id}/statements
  • Authenticate and authorize as a Beta Wikidata user when making edits using the API, as well as provide edit tags and edit comment, and mark an edit as one made by a bot.

If you want to give it a try, here are some REST API calls for an example Item (“house cat”) on Beta Wikidata:

The following things are still missing before it can go live on Wikidata in a first version:

  • Replacing the statement on an Item with PUT /entities/items/{item_id}/statements/{statement_id} or PUT /statements/{statement_id}
  • Editing a statement on an Item with PATCH /entities/items/{item_id}/statements/{statement_id} or PATCH /statements/{statement_id}
  • Removing the statement from an Item with DELETE /entities/items/{item_id}/statements/{statement_id} or DELETE /statements/{statement_id}
  • Automated edit summaries

The following will likely not be available in the first version but follow later:

  • Creating or deleting an Item
  • Getting a statement from an Item based on the Property ID in the statement
  • Any operation (reading, adding, editing, removing) on sitelinks, labels, descriptions and aliases
  • Any operation on entity types other than Items (i.e. Properties, Lexemes, …)
  • Translated error messages

If you have any feedback on what’s there so far please let us know at Wikidata talk:REST API feedback round.

Cheers, - Mohammed Sadat (WMDE) (talk) 08:43, 1 August 2022 (UTC)[reply]

What is GET /entities/items even supposed to do? It says "Retrieve a list of Wikibase Items", but it has a "_fields" value indicating it also returns data? Yet it doesn't allow you to specify the QIDs to retrieve data from. The spec for this call makes zero sense. If it is a lister it should probably get a different name in the API to avoid confusion with the data retrieval calls. If it is a lister I would also make the following changes:
  • Allow for specifying sort orders: none (database order) and by QID.
  • Allow to specify QID as offset.
  • Add limit parameter and allow specifying QID as end of list, as well as a numeric position. (based on whatever the sort order is)
Infrastruktur (talk) 22:19, 1 August 2022 (UTC)[reply]
Thanks for your feedback. GET /entities/items is not actually part of the specification (which can be found here), nor of the first version. We are still looking into how useful having the list of all Items would be as well as how feasible it would be both from the technical as well as client experience perspective. It seems like this is something other access methods like the dumps are better suited for.
Let’s continue further discussions on Wikidata talk:REST API feedback round to keep it all in one place. - Mohammed Sadat (WMDE) (talk) 11:04, 4 August 2022 (UTC)[reply]

Wikidata’s 10th birthday: contribute to the celebration video[edit]

Hello all,

As you may know, Wikidata’s 10th birthday is around the corner, with plenty of local and online events taking place in October 2022 to gather the Wikidata community.

On top of that, we are working on a collaborative video to celebrate the 10th anniversary of Wikidata and the achievements of the community. In this short clip, we would like to integrate footage from many people from all around the world sharing birthday wishes, saying a few words about Wikidata, and showing places around them.

This is where you come in! You can contribute directly to this project by recording and sending us a video before September 18th.

You can do it on your own or together with your local Wikidata community. You can take the opportunity of an upcoming onsite event (for example, a Wikimania gathering) to record something with more people.

You don’t need a lot of equipment: videos filmed on smartphone are very welcome. You will find all the details and requirements on this page.

On top of that, if you need support to record a video, or simply if you want to take the opportunity to gain new skills, we are happy to offer an online workshop “how to film with your smartphone”, you will find more details about the workshop on this page.

Feel free to share this message with your local community and anyone interested. If you have any questions or need support, please check the documentation page, and feel free to write on the talk page. If you have technical questions, you can also contact Anne, our video producer, directly by email.

We’re very excited to receive your videos and to create this celebration content with you! Best, Lea Lacroix (WMDE) (talk) 13:28, 4 August 2022 (UTC)[reply]

Make {{Item classification}} include Q numbers[edit]

It's very frustrating (at least to me) that the default superclass hierarchy displayed by {{Item classification}} uses a different format that {{Q}} (i.e. Universe (Q1)). Specifically, the format lacks copyable Q numbers. Is there consensus for these to be added? I don't want to work on doing so before getting at least some encouragement. JesseW (talk) 17:16, 31 July 2022 (UTC)[reply]

OK, I decided to work on it anyway. :-) Here (in the sandbox) is the feature, if we decide we want it. JesseW (talk) 00:12, 1 August 2022 (UTC)[reply]
A more central question would be whether should make the item id part of the linked in the first place. I would expect that in most cases the ability to copy the id is more valuable than having a larger click area. ChristianKl ❪✉❫ 10:43, 1 August 2022 (UTC)[reply]
I think we can decide that separately, after getting a consensus (hopefully) for adding Q numbers to {{Item classification}} at all. I tend to agree with you about making the Q number un-clickable, but even having it part of the link makes it easier to copy than if it's only in the link target. Anyone else willing to speak up -- maybe we can go ahead after a week without objections? JesseW (talk) 13:02, 1 August 2022 (UTC)[reply]

Also opened a discussion on the talk page, Template_talk:Item_classification#Show_IDs. JesseW (talk) 13:08, 1 August 2022 (UTC)[reply]

{{Item classification}} uses {{SuperclassTree}} which uses {{Tree}} which uses Module:Tree. Template:Tree/sandbox use Module:Tree/sandbox but I don't see any of your changes. PAC2 (talk) 05:38, 2 August 2022 (UTC)[reply]
@PAC2: My change (linked above) is to Module:Tree/sandbox, adding a show_id boolean parameter. There aren't any relevant changes to make to Template:Tree/sandbox, but I've now changed Template:Tree/sandbox/doc to demonstrate the feature. Thanks for asking about it, do you see it now? JesseW (talk) 12:18, 2 August 2022 (UTC)[reply]

That's perfect PAC2 (talk) 15:17, 2 August 2022 (UTC)[reply]

  1.  Support I think it's an improvement. PAC2 (talk) 21:26, 3 August 2022 (UTC)[reply]

After nearly a week, and comments from two users (one explicitly supporting it), and no objections, I'm going to go ahead and make the change. It's easy to revert if people object. JesseW (talk) 20:12, 4 August 2022 (UTC)[reply]

Wikidata:List of properties/e-commerce[edit]

I think this list is incomplete. Somehow PLbot not add all the properties @Pasleim:. Eurohunter (talk) 16:15, 4 August 2022 (UTC)[reply]

Subclass trees[edit]

Clustered them here as these discussions are related. Multichill (talk) 16:33, 25 July 2022 (UTC)[reply]

Specific church listed as object that exists outside physical reality[edit]

https://www.wikidata.org/wiki/Q2064095

type https://www.wikidata.org/wiki/Q2977 en: cathedral (Christian church, which is seat of a bishop) [3]

en: church building (building constructed for Christian worship) [4]
en: geographic object (a representation of a geographic entity) [5]
en: representation (role, function or property of an abstract or real object, relation or changes) [6]
en: relation (general relation between different objects or individuals) [7]
en: abstract object (object with no physical referents) [8]
en: non-physical entity (object that exists outside physical reality) [9]

(I am happy to report more cases like this if community here is interested in fixing them and such report are welcome, let me know if I should stop or pause posting about such issues)

Mateusz Konieczny (talk) 22:31, 13 July 2022 (UTC)[reply]

San Francisco botanical garden is classified by Wikidata as a skill[edit]

https://www.wikidata.org/wiki/Q589884

type https://www.wikidata.org/wiki/Q167346 en: botanical garden (well-tended area displaying a wide range of plants labelled with their botanical names) [10]

en: show garden (garden for the mediation of certain topics) [11]
en: garden (planned space set aside for the display, cultivation, and enjoyment of plants) [12]
en: land use (characterization of land based on what can be built on it and what the land can be used for) [13]
en: spatial planning (technique for physical organisation of space) [14]
en: planning (process of determining the activities required to achieve a desired goal) [15]
en: skill (learned ability to carry out a task) [16]

(I am happy to report more cases like this if community here is interested in fixing them and such report are welcome - let me know if that is not useful. @Pigsonthewing: who recommended to report such cases here.) Mateusz Konieczny (talk) 15:31, 25 July 2022 (UTC)[reply]

Would this be improved by defining garden (Q1107656) as being a subclass of geographical feature (Q618123)? That might also help with some of the other hierarchies called out here. Pauljmackay (talk) 11:52, 31 July 2022 (UTC)[reply]
garden (Q1107656) has way too many superclasses; I'm working on trying to straighten it out. JesseW (talk) 17:03, 31 July 2022 (UTC)[reply]
Big thanks! Let me know if at some point you would be interested in more reports like this one (I am unsure should I add more after listed right now are fixed or should I let this topic to die in archive) Mateusz Konieczny (talk) 21:35, 31 July 2022 (UTC)[reply]
I think the right place to report these is on Wikidata_talk:WikiProject_Ontology -- that will get the attention of the people interested in fixing these sort of messes, as opposed to here that gets everyone's attention. JesseW (talk) 23:04, 31 July 2022 (UTC)[reply]

All coffeehouses are listed as non-physical entities[edit]

https://www.wikidata.org/wiki/Q672804

type https://www.wikidata.org/wiki/Q30022 en: coffeehouse (establishment that serves coffee and tea) [17]

en: drinking establishment (business whose primary function is the serving of beverages for consumption on the premises, e.g. cafés, tea houses, inns) [18]
en: eating or drinking establishment [19]
en: business (activity of making one's living or making money by producing or buying and selling products) [20]
en: economic unit (economics term; economically independent decision maker, such as a private household or a company) [21]
en: unit (entity regarded or used as an elementary structural or functional constituent to measure, analyse or describe another entity) [22]
en: abstract object (object with no physical referents) [23]
en: non-physical entity (object that exists outside physical reality) [24] banned as it is an object that exists outside physical reality !!!!!!!!

(I am happy to report more cases like this if community here is interested in fixing them and such report are welcome - let me know if that is not useful. @Pigsonthewing: who recommended to report such cases here.)

Mateusz Konieczny (talk) 15:32, 25 July 2022 (UTC)[reply]

Ceramic picture classified by Wikidata as a non-physical entity[edit]

https://www.wikidata.org/wiki/Q75320653

type https://www.wikidata.org/wiki/Q45621 en: ceramic (inorganic, nonmetallic solid prepared by the action of heat) [25]

en: solid (solid matter; matter which does not deform easily under pressure) [26]
en: physical state [27]
en: attribute entity [28]
en: abstract object (object with no physical referents) [29]
en: non-physical entity (object that exists outside physical reality) [30] banned as it is an object that exists outside physical reality !!!!!!!!!!

(I am happy to report more cases like this if community here is interested in fixing them and such report are welcome - let me know if that is not useful. @Pigsonthewing: who recommended to report such cases here.)

Mateusz Konieczny (talk) 15:34, 25 July 2022 (UTC)[reply]

@Mateusz Konieczny: Please don't ping me each time you post one of these.
Instead of https://www.wikidata.org/wiki/Q75320653 or [https://www.wikidata.org/wiki/Q75320653], please use [[Q75320653]] (or better still {{Q|Q75320653}}, which will provide the label, so you don't have to type it). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:36, 27 July 2022 (UTC)[reply]
OK, if I post more I will not ping you and will use this formatting. I pinged you as you were the person claiming that it is a good idea to report such issues here. Though likely I will not post more for now as it appears that it is not useful, as I expected. If anyone wants such report, let me know how many such errors you would want to get for processing. (note: for my rant about how wikidata classification system is unusable, see https://lists.openstreetmap.org/pipermail/talk/2022-July/087654.html ) Mateusz Konieczny (talk) 14:11, 27 July 2022 (UTC)[reply]
There is also the issue that many items have multiple entities under subclass of (P279). If you traverse up the tree, these conflicting classifications can be found. But is that true for all the traverses? E.g. is ceramic (Q45621) being a subclass of material (Q214609) more valid? I would also suggest that untitled (Q75320653) should be made from ceramic (Q45621), not a subclass of it. And possibly that ceramic (Q45621) should have a "state" of solid (Q11438), not a subclass of it (not exactly sure what property to use there). Pauljmackay (talk) 14:29, 31 July 2022 (UTC)[reply]
@Mateusz Konieczny, Pauljmackay: Oh, you ran into the problem that the material and the artwork made from this material are mixed up. You probably want to use ceramic art (Q13464614). See Wikidata:WikiProject Visual arts/Item structure for more information. Multichill (talk) 17:49, 31 July 2022 (UTC)[reply]

Is there a better place to report blatant failure of Wikidata classification?[edit]

Is there a better place to report utterly invalid classifications in Wikidata which are revealed by subclass trees?

In my view situation is as follows:

  • Wikidata has very low quality classification system
    • Wikidata classification system is not allowing to reliably answer questions such as "is this an event" or "is it a physical object" or "is it ship or group of humans" or "is it physical or non-physical entity".
  • If someone wants to fix individual problems they can do this - it takes some time but is not really hard
  • If you are not willing to fix on your own then noone else will do, as critical failures in classification schemes are completely normal, common, and not treated as an emergency
    • Is there a better place to report such issues than here?
  • If you are willing to fix classification issues you should expect that some other editor will break them again sooner or later
    • Caching Wikidata and updating only updated items may be a solution if fixing problems again is too irritating
  • Overall, using Wikidata classification system as of 2022 is likely a mistake and if anyone plans to use it they should reconsider it
    • At least I am consider significant time spend on it by myself as a mistake

Have I missed something? Especially where I am claiming that for something there is no sane way to achieve it?

Mateusz Konieczny (talk) 07:42, 30 July 2022 (UTC)[reply]

You seem to assume the goal of Wikidata is to have a perfect model of everything. That's impossible. We live in an imperfect world that's not black and white. The broader the scope of the model, the more inconsistencies it will have. Some things can be improved, others can't without breaking different parts of the model.
Generally the model should be locally consistent. So don't try to traverse the whole subclass tree, but limit the number of steps. Multichill (talk) 10:30, 31 July 2022 (UTC)[reply]
I tend to agree, it will not be a perfect model at any time.
However, I think there is a general consensus that the subclass tree—sometimes referred to as “the Wikidata ontology”—is not good enough based on what’s probably possible. The issue has for instance been addressed at the “2021 Data Quality Days” event (see File:DataQualityDays 2021 - Ontology-issues talk.webm) and there are phab tickets regarding potential improvements such as phab:T295275. None of that immediately improves the ontology, but I think there is long-term hope for a better result. —MisterSynergy (talk) 18:17, 31 July 2022 (UTC)[reply]
I am not expecting it to be perfect. The problem with Wikidata is that after fixing entries they kept being broken again and again. I have test set of about 150 cases that were fixed at some point. If I run it based on latest Wikidata data, not cached entries - over 100 are broken again. Also, classifications issues are quite bad, routinely ending with utter nonsense. It is not just some isolated problem, it is that my validator of OSM data turned into a validator of Wikidata structure. Maybe I was misusing Wikidata and should not expect it to work in the first place, but it was not indicated or caveated anywhere at all. I think I started from examples in https://query.wikidata.org/ - and either I missed such warning or it is missing - but there is no hint that such classification queries are not reliable at all. Also, some people advocating Wikidata use are denying that problem exists or claim that it is really limited (that is what triggered my recent postings here). Every single of my examples posted so far is from list of this test cases that kept being broken again and again. When I asked for help - many people helped me with specific technical details but noone warned me that entire attempt is doomed and answering such questions as "it is an event or place" with Wikidata is impossible to do reliably Mateusz Konieczny (talk) 21:26, 31 July 2022 (UTC)[reply]
Would you be willing to publish your test set of 150 cases? I’d be interested to look into it, in order to see what’s going on. You’re reporting plenty of individual cases here which makes it difficult to really get an overview. —MisterSynergy (talk) 21:34, 31 July 2022 (UTC)[reply]
Outputs are at User:Mateusz_Konieczny/failing_testcases - and number of failures is noticeably lower right now, thanks to all people who fixed something (I also has run it with set of patches and short-circuiting that avoid triggering of some issues). Code itself is at https://github.com/matkoniecz/wikibrain/blob/master/test_wikidata_structure.py Note that some failures may be caused by my misuse or poor use of Wikidata structure but many are clear examples of broken Wikidata ontology Mateusz Konieczny (talk) 12:05, 1 August 2022 (UTC)[reply]
I can't find the packages wikimedia_connection and osm_handling_config. Can you help me here? wikibrain seems to be your own package, available on Github—right? —MisterSynergy (talk) 13:40, 1 August 2022 (UTC)[reply]
Oh, that is entirely my fault - I have never published this packages (partially because I mostly gave up on this project) and osm_handling_config is a horrifying hack and/or old mistake. I will try to make it runnable. Do you now what is the proper method of passing config to python script? In this case about location of cache. Config file in some predetermined location? Environment variable? Config file in place where script runs? Because I am pretty sure that "build a package where function returns config values" is not preferred solution. Passing as script parameter will not work for example with tests. Mateusz Konieczny (talk) 21:19, 1 August 2022 (UTC)[reply]
I don't care if it has poor quality. If it runs, it runs—and otherwise I'll try to figure out what's wrong.
Same holds for the config for which there are several methods how you can do this. The only thing you don't want to do is to publish private config information into your public repository. If there is such information in your config, consider publishing a config-sample file that mocks what config is expected to be there and how it should look like, without containing any private information. —MisterSynergy (talk) 22:02, 1 August 2022 (UTC)[reply]
I made first steps toward making this code usable, but have not yet tested it. I will ping you once I will reproduce install on a clean OS. Mateusz Konieczny (talk) 13:04, 6 August 2022 (UTC)[reply]
I entirely agree with you about the lack of warning regarding high-level classification problems at Wikidata. They aren't reliable or stable (yet), and that's not made clear enough. I think a helpful way to reduce the problem is, as User:Multichill suggested, is to kinda chop off your interpretation before you get into the problematic high-level classes, by hard-coding properties like "is it an event or a place" at mid-level (like geographical feature (Q618123)) which (I hope?) are more stable. Also, if/when we can come to consensus (over at Wikidata:WikiProject_Ontology) on parts of the high-level hierarchy, it can be documented there, and enforced with a bot. But that requires actually coming up with agreed-on standards. JesseW (talk) 23:18, 31 July 2022 (UTC)[reply]
These is a Wikidata:Classification noticeboard, which was active in 2015-2016.--GZWDer (talk) 21:45, 2 August 2022 (UTC)[reply]
The problem you describe is a consequence of Wikidata being a wiki that anyone can participate in. Semantics and ontology building are non-trivial, and as long as people without the requisite knowledge edit the wiki, the classification tree will be unreliable; in other words, it will likely never be reliable. Silver hr (talk) 13:39, 3 August 2022 (UTC)[reply]
@Silver hr: not buying that. Just replace "Wikidata" with "Wikipedia" in your statement. People have been saying that forever about Wikipedia and that turned out pretty good. But I do agree that modeling this is not easy at all just like writing a featured article. I guess it's made harder because edits all over the place instead of one article to put on your watchlist.
@Mateusz Konieczny: you currently seem to have paths of items that shouldn't be connected as a test. Do you also keep paths of items that should be connected? I've noticed that people trying to fix these issues often end up breaking different chains that shouldn't be broken. I wonder if we can make some sort of recent changes for subclass of (P279) changes so it's easier to track modifications that might cause problems.
When it comes to solving these issues we can probably describe different strategies how to do it. For example add the relevant subclasses to a (relatively) well defined concept lower in the tree and worry a bit less about the redundancy. That reduces the risk of things getting broken when people are changing things higher up in the tree. I'm sure more and better strategies exist. Multichill (talk) 12:58, 6 August 2022 (UTC)[reply]
@Multichill: I keep also some to prevent false negatives, see for example https://github.com/matkoniecz/wikibrain/blob/master/test_wikidata_structure.py#L34 and other "assert_unlinkability" tests (11 in my test set). Though such test starting to fail was typically (always?) triggered by bugs in my code. And in general false negative is typically caused by items being simply not classified at all, not by classification system being not inclusive enough Mateusz Konieczny (talk) 13:04, 6 August 2022 (UTC)[reply]
@Mateusz Konieczny: that made me think of something. Lua supports testcases, see examples here and many more on Commons.
Maybe we should set up LUA test cases for these so it's easy for anyone to see if it breaks and we can work together on adding more cases? Multichill (talk) 13:31, 6 August 2022 (UTC)[reply]
@Multichill: That would be great if possible to do! I am not promising help on LUA itself, but would be happy to provide more test cases (in addition to this here) and help a bit with fixing revealed issues Mateusz Konieczny (talk) 14:17, 6 August 2022 (UTC)[reply]
The content of Wikipedia is free-form text in a natural language. Almost every person can communicate in a natural language. Far, far less people can express themselves in a formal language, and that is pretty much what we're doing on Wikidata. But I don't particularly wish to get into this topic deeply as I don't believe it would be productive. Let's just say I admire your optimism. Silver hr (talk) 16:51, 6 August 2022 (UTC)[reply]

How to translate names of languages[edit]

Viewing pages like United States of America (Q30) in Korean shows many names of languages that have not been translated to Korean [for example, Tunisian Arabic (Arabic script)]. How can users translate these names to Korean? Intolerable situation (talk) 01:03, 6 August 2022 (UTC)[reply]

Are you asking where to go to enter these translations, or how to actually translate them? Middle river exports (talk) 14:54, 6 August 2022 (UTC)[reply]
@Middle river exports: He probably mean that name "Tunisian Arabic (Arabic script)" remains untranslated and will show in English independently from language version of Wikidata you use. Eurohunter (talk) 15:28, 6 August 2022 (UTC)[reply]
@Intolerable situation: https://www.wikidata.org/w/index.php?title=Q30&uselang=ko#P1549 looks mostly Korean to me.
Translations are from CLDR, see https://translatewiki.net/wiki/CLDR for more information.
In this file you can find the English keys. For example pt-br is in that file, but missing in the Korean file so that shows as "Brazilian Portuguese". Multichill (talk) 17:44, 6 August 2022 (UTC)[reply]

Twinkle[edit]

Does Twinkle is deployed at Wikidata? I do not find nothing at Wikidata:Twinkle. --NGC 54 (talk | contribs) 18:50, 6 August 2022 (UTC)[reply]

@NGC 54: TwinkleGlobal works here. Bencemac (talk) 19:04, 6 August 2022 (UTC)[reply]
@Bencemac: But there is no local Twinkle version? --NGC 54 (talk | contribs) 19:14, 6 August 2022 (UTC)[reply]
As far as I know, there is no. Bencemac (talk) 19:22, 6 August 2022 (UTC)[reply]

Should billionaire (Q1062083) be used as a value of instance of (P31)?[edit]

I discover that some items use billionaire (Q1062083) as a value of instance of (P31). See the list here : https://w.wiki/5WWb

I find it weird for many reasons :

  • first, there are many uncertainties in the way we state that someone is billionaire. Should we compute the capital in euros, dollars or any other local currency? what do you count in the capital?, etc
  • second, billionnaire is a very volatile status. You may be billionnaire in 2020 and lose this status the year after, etc. How do we deal with temporal variations?
  • last but not least, there is no reason to say that billionnaires are different from other humans. Putting this information in the P31 property gives to the statement a kind of ontological statement.

For all those reasons, I think that we should find another solution for encoding this information.

Is there any previous decision, request for comments or debate about this issue? WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. --PAC2 (talk) 06:19, 28 July 2022 (UTC)[reply]

Having lots of money is a notable characteristic, but it would be better captured with net worth (P2218) with qualifiers of date, currency and a reference of Forbes, Sunday Times etc Vicarage (talk) 06:31, 28 July 2022 (UTC)[reply]
@PAC2: no, and the person who added it (user:TiagoLubiana) has reverted part of it. See Wikidata:Project_chat/Archive/2022/03#How_to_represent_billionaires_on_Wikidata?.
@TiagoLubiana: can you please finish your clean up? Multichill (talk) 08:36, 28 July 2022 (UTC)[reply]

Thanks for your answers. PAC2 (talk) 17:50, 28 July 2022 (UTC)[reply]

  • It should not happen. All items for individual human beings should use human (Q5) and not use other Belgium (Q31) items like this. ChristianKl ❪✉❫ 10:54, 1 August 2022 (UTC)[reply]
    I believe "instance of" in the sense of "is a" would be a better property to use. One is not often noted as having the characteristics or qualities of a billionaire but instead as being one.
    • Thousandaires get spell checked because no one uses the term.
    • Millionaire/s, Billionaire/s, and the like could be considered both an instance and a class. An instance of an individual who is of the class of one with a net worth of greater than a million or a billion units of currency. An instance of billionaire would inherit the property of having the net worth of greater than one billion.
    • Start and end time could be used as qualifiers of course.
    • Associated cause qualifier could also be linked to the net worth of greater than $1 billion.
    • The currency used would likely be the one primarily used by the individual but I see no reason others could not be used if used in references.
    • A script could be applied to all humans with existing properties of net worth greater than $1 million (or the pertinent awards) and apply instance of millionaire of billionaire accordingly. Terrickisaiah555 (talk) 04:39, 4 August 2022 (UTC)[reply]
Sigh. Why would that be better than the current method of using has characteristic (P1552)? It really is as if a bunch of people who have not even done the most basic research and are completely and utterly unaware that by convention WD does not use P31 to characterise humans, have parachuted in here to talk through their hats. See https://w.wiki/5Y37 --Tagishsimon (talk) 04:43, 4 August 2022 (UTC)[reply]
That seems a bit judgmental. I would say that seems to be an instance of a condescending reply XD. It also seems to have the quality of discouraging collaboration.
I am aware that the idea is to prevent labeling people however it is important to remember that there is a difference between identifiers and properties.
Being able to identify people who have the properties that are considered to make one a millionaire or billionaire would be helpful and it seems like there some fear of identifying humans in any sort of way other than humans at the loss of the use of the tools. Using the quality requires searching for both humans and either billionaire or net worth greater than a certain amount. I see no reason to not just utilize class inheritance. Having the quality of billionaire would typically mean you have characteristics that make you likely to be or to have been an instance of billionaire at some point which is not necessarily the same as actually having been a billionaire at some point. What do ya think @Tagishsimon ? Terrickisaiah555 (talk) 05:27, 4 August 2022 (UTC)[reply]
I commend you on your excellent reply, Terrickisaiah555. Now, what part of the established method of denoting that a person is a billionaire, by attaching 'billionaire' as a value to a has characteristic (P1552) statement, requires us instead to consider using P31 for the purpose? And if we are to use P31 for this purpose, what other characteristics or qualities of humans should we also attach using P31? --Tagishsimon (talk) 05:49, 4 August 2022 (UTC)[reply]
Well thank you, I appreciate your feedback. You seem to have the qualities of a billionaire. You wouldn't happen to be one would by chance would you? :) I would say that a billionaire is likely to have certain qualities that increase their chances of continuing to be a billionaire. This, however, is not easily defined. I would keep only the properties of billionaires that define it as such; An instance of human who at some point had a net worth of over 1 billion units of currency. @Tagishsimon Limiting billionaire to a value instead of an identifier of an inherited set of properties, greatly limits and confuses things unnecessarily. Instance of billionaire is much easier to query than multiple properties and values/ranges/etc. I hesitate to get into other useful identifiers of humans at the moment. Maybe this could serve as a use case for other examples. Terrickisaiah555 (talk) 06:11, 4 August 2022 (UTC)[reply]
Ontologically, there's no reason we couldn't have subclasses of the form human (Q5) AND [some constraint]. However, this would produce a potentially infinite number of human (Q5) subclasses because you can always make up and combine constraints. For example: with sex you get male human and female human; two subclasses. With year of birth you get human born in [year]; thousands of subclasses. Combine that with sex, you've doubled the number. With country of birth you get human born in [country]; hundreds of subclasses. Combine that with the previous, you multiply the previous number by hundreds. It would be a combinatorial explosion.
You are right in that if a small number such subclasses existed, it would be easier to write some queries because you'd only need to specify the subclass instead of the constraints that define it. But how long would it take you to find the subclass you're interested in if there are trillions of direct subclasses of human (Q5)? I think a better solution lies on the user-facing side of WD: have a wiki-repository of queries that return classes of interest, such as billionaire (or possibly use Schemas, not sure). Silver hr (talk) 14:19, 4 August 2022 (UTC)[reply]
" But how long would it take you to find the subclass you're interested in if there are trillions of direct subclasses of human (Q5)?"
If you know what you are interested in, you could either pick any subclass with those specific properties (since it would just be instances with those criteria), or search for subclasses with the properties and choose those that best describe the case you are looking for.
You would be able to use as specific a class as needed or possible. This could always be changed to something more specific as more info becomes available. Derivatives of super-classes would allow one to find what they are looking for without needing to be too specific. @Silver hr
"With country of birth you get human born in [country];"
In this case, it is usually better to use instance of human with properties indicating their birthplace as there is not a currently a term, but rather a description (combination of terms) of a subject and it's properties (birth place). This also allows more refinement and specificity. If a term can be found that is specific to an individual born in a country rather than just citizenship than that would likely be a good use of a class however there are not really any terms such as that that I know of. Terrickisaiah555 (talk) 19:32, 4 August 2022 (UTC)[reply]
So, basically, you're saying it would be easier to input or query instance of (P31)billionaire (Q1062083) rather than net worth (P2218)1000000000 or net worth (P2218)unknown valuelower limit (P5447)1000000000, and that we wouldn't have a potential combinatorial explosion of direct subclasses because this would only be limited to subclasses with established names such that we'd have, for example instance of (P31)billionaire (Q1062083) and instance of (P31)president (Q30461) rather than instance of (P31)billionaire president? On the face of it, I could agree, but there is a problem: you can't have qualifiers for qualifiers. So, the qualifiers on a net worth (P2218) statement would be lost if the net worth (P2218) statement itself were a qualifier for instance of (P31)billionaire (Q1062083). (And I don't think merely transferring the net worth (P2218) qualifiers onto instance of (P31)billionaire (Q1062083) would be semantically correct, at least not in all cases.) Silver hr (talk) 18:39, 6 August 2022 (UTC)[reply]
Correct, it would be easier to query and would make more sense in my opinion.
If it is useful to have a billionaire president as a class that describes individuals than I see no reason it would not be something we should create and utilize. This would increase the number of classes, however it would mitigate the need to have multiple properties allowing a clearer presentation of the data specific to the item. In the case of billionaire presidents, it does not appear that this has often been the case thus far and would not likely provide any benefit by have it as a classification. I would say the limiting factor would be the usefulness in providing structure. A few instances do not really give a need for a way of categorizing.
The other reason this would be beneficial is in the standardization. By determining and suggesting alternative classifications it will provide a more efficient, standardized, intelligent and adaptable system in the long run.
As far as the net worth being lost, I would just have it as a property of the individual rather than the class of billionaire. One could be an instance of billionaire with an unknown specific net worth (other than the implied range.) The billionaire class could have a cause of net worth being above the amount associated with it that should not conflict with the existence of the individuals net worth. Terrickisaiah555 (talk) 20:07, 6 August 2022 (UTC)[reply]

As a comment, there are lists of billionaires across Wikipedias, and this grouping seems relevant. We almost never have a precise net worth (P2218) (this changes literally every minute) , and the property is way way less queriable than a straightforward "billionaire" statement. We have no ways to query for billionaires on Wikidata. Not even an approximate query.By removing this P31 statement without providing a reasonable alternative, we are making Wikidata less useful. Anyways, started removing what is left of it (https://quickstatements.toolforge.org/#/batch/94128). TiagoLubiana (talk) 19:58, 1 August 2022 (UTC)[reply]

Since the definition of a billionaire is "a person with a net worth of at least 1 billion", if the net worth of someone isn't known, this could be expressed with net worth (P2218) and a lower bound of 1000000000. The Quantity data type supports a lower and upper bound, although it doesn't look like it's technically possible to have only a lower bound on a specific value, or any bound on an unknown value. It is, however, possible to use lower limit (P5447) as a qualifier, so the billionaire statement could be net worth (P2218)unknown valuelower limit (P5447)1000000000. Silver hr (talk) 15:47, 3 August 2022 (UTC)[reply]
I think it's worth to launch a Request for comments to discuss about this question and find the best solution. PAC2 (talk) 21:28, 3 August 2022 (UTC)[reply]
RfCs are not normally employed as a method of settling trivial modelling questions. RfCs expect that non-trivial effort has gone into solving the issue before taking the RfC step, something that has not happened in this case.
TiagoLubiana's desire to use P31 to represent the quality of being a billionaire is completely unsupported; by convention, P31 is not used to represent qualities of humans. Their "we are making Wikidata less useful" reflects their lack of knowledge of WD practice in this area, which amply provides a model to store the information - has characteristic (P1552) ... see https://w.wiki/5Y37 --Tagishsimon (talk) 04:35, 4 August 2022 (UTC)[reply]
instance of (P31) does not exist to store as much information as possible. In most cases it's best if it has exactly one value. One of the sources for who happens to be a billionaire happens to be the list from Forbes. We can easily store information from it via net worth (P2218).
Apart from that I think the batchjob by TiagoLubiana (talk • contribs • logs) violated our living people policy. net worth (P2218) is property that may violate privacy (Q44601380). Adding such information without external sources about the net worth is very questionable especially when the information is added in an automated way to a lot of items. ChristianKl ❪✉❫ 08:42, 6 August 2022 (UTC)[reply]
Okay, you are really enthusiastic about this issue, and are making personal accusations that border on troll behavior.
I wont' feed it. I've gotten too many enraged comments and personal attacks when trying new things with P31 statements. Seriously, let's be more careful with our language and value the "assume good faith" and "no personal attacks" policies.
Explicitly saying "User X has no idea of A,B,C", "User X desire is completely unsupported" or "User X don't know what they are doing" and even "User X has violated a serious policy" is, at the very least, aggressive wordings and may be considered personal attacks.
As per tagging billionaires: I give up, I won't touch it. If anyone wants to pull that info in the future, just use the code here: https://github.com/lubianat/upwork_billionaires TiagoLubiana (talk) 16:22, 6 August 2022 (UTC)[reply]
The policy does say "should be supported by a reliable public source" and the data you added isn't. Pointing out that you violated the policy isn't a matter of assuming that you violated it in bad faith or a personal attack. It's just pointing out that we have a policy that exists to prevent what you did. ChristianKl ❪✉❫ 08:18, 7 August 2022 (UTC)[reply]

By the way: I realize I was grumpy and passive-aggressive sometimes in this discussion. I am sorry, and I apologize. TiagoLubiana (talk) 16:56, 6 August 2022 (UTC)[reply]

People's Guerrilla Army[edit]

I wish there was an English language link for the "Ejército Guerrillero Popular" page of the Spanish Wikipedia. A redirection that is not as such to Shining Path, but to People's Guerrilla Army but the page does not let me do that since it says that it is already in use, what should I do? Cesar David MP (talk) 02:41, 7 August 2022 (UTC)[reply]

Possibly turn the redirect page into a disambiguation page, presuming there is another page than Shining Path that "Ejército Guerrillero Popular" might resolve to. --Tagishsimon (talk) 13:24, 7 August 2022 (UTC)[reply]

Wikidata weekly summary #532[edit]

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Events
    • Upcoming
      • Wikimania 2022, August 11 to 14, online event. The Hackathon will take place August 12-14. On this page you can find a summary of sessions and community gatherings related to Wikidata and Wikibase.
      • Next Linked Data for Libraries LD4 Wikidata Affinity Group call August 9, 2022: Pieter Vander Vennet on MapComplete, a thematic OpenStreetMap viewer and editor which uses species, language, and image data from Wikidata. Agenda.
      • The Wikimania Hackathon starts next Friday, August 16-22! There are still lots of spots in the schedule to add your Wikidata related sessions or project ideas (anyone can present a session)
    • Ongoing
      • Wikimedia Indonesia's Wikidata edit-a-thon (datathon) for the 77th anniversary of the Indnesian Independence Day started on 5th August and will be held until 12th August. Participants are instructed to edit items containing the statement country of origin (P495): Indonesia (Q252).
      • Toolhub is a catalog of 1500+ tools used every day in a wide variety of workflows across many Wiki projects. We are currently improving the search functionality and need your input – whether you are already familiar with Toolhub or not. Please take 5-10 minutes to leave feedback.
    • Ongoing
      • Weekly Lexemes Challenge #53, Sheep
    • Past
      • First online meet-up fully organized by volunteers of the Indonesian Wikidata Community has been held on 30th July where we edited items on Indonesian ethnic groups.
  • Tool of the week
    • Wiki tools - adds dozens of Wikipedia and Wikidata functions to your Google sheets.
  • Development
    • Lexicographical data:
      • Continuing to address feedback from the testing (e.g. phab:T312292, phab:T313113, phab:T313466)
      • We have pushed back replacing Special:NewLexeme with the new Special:NewLexemeAlpha a bit to address more of the testing feedback.
    • Continuing to tackle allowing sitelinks to redirects under some circumstances (phab:T278962)
    • REST API:
      • Finishing up the endpoints for removing and replacing statements and adding authentication and authorization to them
      • Looking into feedback from first testing

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

Stale import from wikipedia?[edit]

https://m.wikidata.org/wiki/Q16972633 claims it's named after a town in Japan but the cited section claims otherwise. Was this imported automatically at some point and is there some facility to quality check this? Nivekuil (talk) 22:59, 9 August 2022 (UTC)[reply]

The imported claim value had been changed by an IP user here. Now undone. —MisterSynergy (talk) 23:05, 9 August 2022 (UTC)[reply]

How to express that an entity (technology, language etc. etc. ) was used during a period of the history?[edit]

I would like to express that an entity was used during a period of the history. For instance, that lead tin yellow pigment was used from the 13th century until 18th century. I could use https://www.wikidata.org/wiki/Property:P2754 but for the end of the period I could only find discontinued https://www.wikidata.org/wiki/Property:P2669 I was wondering if I could use something more general like inception for the start, so I can use it also for other entities like for instance uncial script https://www.wikidata.org/wiki/Q784235. MRCGCM53 (talk) 11:55, 10 August 2022 (UTC)[reply]

I think start time (P580) and end time (P582) would be appropriate. Silver hr (talk) 16:57, 10 August 2022 (UTC)[reply]

Wikidata integration in OpenRefine: could this be your project?[edit]

Hello Wikidata community!

For the past five years I have been maintaining the Wikidata integration in OpenRefine. I am really grateful for the amazing things the community has done with it. Still, there are plenty of things to improve in this integration.

After all these years, I now want to focus my attention in other building sites of OpenRefine. The reconciliation functionality is calling for many improvements. I also want to continue my work on the scalability and reproducibility of OpenRefine. Those are features which will of course benefit the Wikimedia community, but also users coming from other horizons.

Therefore I am looking for someone else to maintain the Wikibase integration in OpenRefine. This means that I am willing to really let go of it and let someone else be in charge. Of course I would still help onboard new developers and keep reviewing pull requests in this area for a while, but I will soon stop developing new features and fixing bugs there myself. This applies to the following projects:

I honestly think this is an enjoyable thing to be in charge of! Among the perks:

  • You get to work on a project that is used by a lot of people, and that's really gratifying!
  • The OpenRefine team is small, so expect low bureaucracy and a lot of freedom. It is however full of passionate Wikimedians who will support your work: Spinster, Antoine2711, Loz.ross, Thadguidry, Owenpatel, Abbe98, Ainali… and myself!
  • I claim we are now pretty good at onboarding people, thanks to various rounds of Outreachy and Google Summer of Code internships, among others
  • There are opportunities to get your work funded, for instance from the Wikimedia Foundation, Wikimedia chapters, or other Wikibase stakeholders.

So the floor is yours! Do you dream about a five-star lexeme support in OpenRefine? Do you think a ShEx integration is overdue? Or maybe you just want to move that crucial button somewhere more accessible and also fix the layout of this dialog? Be my guest! You don't need to be a Java wizard or to write Javascript like you breathe - if you already have programming experience, you will be able to pick this up as you go.

Feel free to respond here, on OpenRefine's dev mailing list, or even by opening issues and pull requests already! Also, Antoine2711 and I are hosting a session at the Wikimania hackathon to give you an overview of the situation of Wikidata integration in OpenRefine - join us there! − Pintoch (talk) 19:22, 10 August 2022 (UTC)[reply]

forced pregnancy (Q5467540)[edit]

Do we really want to list every death from lack of access to abortion on Wikidata? Trade (talk) 12:49, 25 July 2022 (UTC)[reply]

Before looking at the history I was wondering if LAP959 found a new subject to work on. Sure looks like it. This user has a pattern of using significant event (P793), notable work (P800), depicts (P180) in the wrong way, see User talk:LAP959 for more examples. Just remove it. Multichill (talk) 14:03, 25 July 2022 (UTC)[reply]
To Trade (talk) and Multichill (talk) : These are women whose deaths caused a public outcry and, in some cases, resulted in a major change in a country's legislation, for example Ireland, which changed its laws due the scandal.
Each cases is thoroughly documented in newspaper articles. I think that rather than simply eliminating the information, a better, more inclusive approach might be to add sources that make it clear why these particular deaths of girls and women are historically significant.
In 2012, Savita Halappanavar, age 31 and 17 weeks pregnant, went to a hospital in Galway, Ireland. Doctors there determined that she was having a miscarriage. However, because the fetus still had a detectable heartbeat, it was protected by the Eighth Amendment. Doctors could not intervene – in legal terms, ending its life – even to save the mother. So she was admitted to the hospital for pain management while awaiting the miscarriage to progress naturally. Over the course of three days, as her pain increased and signs of infection grew, she and her husband pleaded with hospital officialsto terminate the pregnancy because of the health risk. The request was denied because the fetus still had a heartbeat. By the time the fetal heartbeat could no longer be detected, Halappanavar had developed a massive infection in her uterus, which spread to her blood. After suffering organ failure and four days in intensive care, she died. This was likely not the only time someone had suffered, or even died, as a result of being denied abortion in Ireland. But the publicity surrounding the case prompted a new wave of activism aimed at repealing the Eighth Amendment. In 2013, the Protection of Life During Pregnancy Act was signed into law, which did not fully repeal the Eighth Amendment but legalized abortions that would protect the mother’s life. - https://www.pbs.org/newshour/health/what-irelands-history-with-abortion-might-teach-us-about-a-post-roe-america
Please see:
Wikipedia's Hostility to Women (The Atlantic)
"Some female editors have been the target of harassment from their male colleagues—and the gender bias has spilled over into the site’s content, too."
https://www.theatlantic.com/technology/archive/2015/10/how-wikipedia-is-hostile-to-women/411619/ LAP959 (talk) 16:25, 25 July 2022 (UTC)[reply]
How is this related to the Atlantic article? --Trade (talk) 16:38, 25 July 2022 (UTC)[reply]
@LAP959 I reverted your edits. This is not the right way of using this item. Ayack (talk) 16:45, 25 July 2022 (UTC)[reply]
*** Ayack (talk) , please suggest the proper way to link the significant changes in Irish law (in 2013 and in 2018) to the death of Savita Halappanavar, which was caused by forced pregnancy in 2012. In 2018 the Eighth Amendment was repealed. There is a Wikipedia article about it. (https://en.wikipedia.org/wiki/Death_of_Savita_Halappanavar)
"On 25 May 2018, the Irish people voted by 66.4% to 33.6% in a referendum to repeal the Eighth Amendment. They approved the Thirty-sixth Amendment of the Constitution Bill 2018 to delete the current provisions of Article 40.3."
Death and Suffering: The Story Behind Ireland’s Abortion Ban and its Reversal
"The death of Savita Halappanavar in an Irish hospital in 2012 after she was denied an abortion during a miscarriage caused outrage across Ireland."
Thank you. *** LAP959 (talk) 16:54, 25 July 2022 (UTC)[reply]
One could to model it using cause of death (P509) at Savita Halappanavar (Q5247527). But this gives a constraint violation, currently. You could also use significant event (P793) at Savita Halappanavar (Q5247527), I think. - Valentina.Anitnelav (talk) 17:38, 25 July 2022 (UTC)[reply]
Thank you. Should be better now. LAP959 (talk) 05:35, 26 July 2022 (UTC)[reply]
I think manner of death (P1196) would be the right property. ChristianKl ❪✉❫ 10:27, 1 August 2022 (UTC)[reply]
@ChristianKl . Following up on your suggestion, I tried to use manner of death (forced pregnancy) but it returns a constraint error:
one-of constraintHelp Discuss
The value for manner of death should be one of the following:
Could manner of death constraints be updated to include the consequences of a forced pregnancy? Thank you. LAP959 (talk) 07:32, 3 August 2022 (UTC)[reply]
I added it to the list of values. I can't see a good reason why it shouldn't belong there. ChristianKl ❪✉❫ 08:12, 6 August 2022 (UTC)[reply]
Thank you! LAP959 (talk) 16:52, 11 August 2022 (UTC)[reply]

statements for places / qualifiers for countries[edit]

There is a disagreement between HarryNº2 and myself concerning statements of the following nature:

place of death
Preferred rank Linz edit
located in the administrative territorial entity Upper Austria
country Austria
0 references
add reference


add value

The question concerns the qualifiers. I consider them to be redundant. The modeling should be done within the object, in this case Linz (Q41329).

My esteemed colleague disagrees and wants to have a “general discussion”, hence this topic. --Emu (talk) 21:06, 5 August 2022 (UTC)[reply]

They're redundant. Using values as qualifiers when the same values are represented in the main statement value item should be done very sparingly and for good reason. There does not seem to be a good reason here. HarryNº2 is wrong in this matter. --Tagishsimon (talk) 21:14, 5 August 2022 (UTC)[reply]
I agree that this is redundant. In my opinion, it's also easier to query for it without using qualifiers. Ainali (talk) 21:34, 5 August 2022 (UTC)[reply]
Some Wikimedia projects use this information in their infoboxes. Smaller Wikipedias in particular do not have an article on each location that they can link to it. The information is even more important if the place of birth or the place of death is a certain street or hospital, for which there is no article in most Wikipedias. Therefore, the information on P131 and P17 is necessary, also due to the changeable history of many countries and municipalities, whose borders have shifted continuously and will shift in the future. Linz, for example, was not always in Austria, but sometimes in the Austrian Empire. The same applies to almost all European places, whose affiliation has constantly changed. Besides, it was not about Linz, but about a small unknown place in Austria called Seewalchen am Attersee. --HarryNº2 (talk) 22:09, 5 August 2022 (UTC)[reply]
Grateful that you've thrown all of the arguments against the wall on the off chance one will stick. Looking at the first 300,000 P20 statements on WD, P17 is a qualifier on ~1% of them, and P131 is used as a qualifier on 0.6% - https://w.wiki/5YNZ . Smaller wikipedias alleged to have a dependency on these qualifiers are going to be sorely disappointed. You talk about "the information on P131 and P17 is necessary" and we say: yes; they would be much better off getting their P131 & P17 data from the value item. It is the case that borders move, and that is a fair argument for qualifiers when that has occurred; on the other hand, in this instance, Lisa-Maria Kellermayr (Q113344024), I'm unaware of Austria's borders having moved. Seewalchen am Attersee is not unknown to WD. It has valid P17 and P131 data in it - Seewalchen am Attersee (Q876227) - and in each case, only a single statement, so there is no room for ambiguity in this case. Good try, but you still lack a good reason and are going against the established pattern b/c it is your whim, not b/c it is neccesary or helpful. --Tagishsimon (talk) 22:51, 5 August 2022 (UTC)[reply]
What are the numbers supposed to tell us? That a bot has created most of the articles in Wikidata, which is not able to insert the data on country (P17) and located in the administrative territorial entity (P131), as this requires a lot of research work. And Austria-Hungary (Q28513) used to divide itself into Cisleithania (Q533534) and Lands of the Crown of Saint Stephen (Q696970). Before that the country was called Austrian Empire (Q131964). There are heaps of errors in Wikidata with this information, especially about Austria (Q40). And this is exactly the problem with people who were born or died in the past, or when it comes to nationality in the past. If you study the history of Austria and its individual states, you will understand my point. And that is just one of many examples, especially when it comes to a European country. --HarryNº2 (talk) 08:45, 7 August 2022 (UTC)[reply]
I changed Seewalchen am Attersee to Linz because the immediate reason we are having this conversation is a very tragic one. I didn’t want to implicitely associate this tragedy with some technical question about modeling. Especially as you yourself said that this is about the general principle. --Emu (talk) 22:56, 5 August 2022 (UTC)[reply]
Where should one draw the line as to when the disclosures should be made and when they should not. The fact that often no information is given is also due to the fact that it is sometimes very difficult to name the affiliation of a place to a certain time, which sometimes requires a lot of research work. Even in the actually good articles on places in the German-language Wikipedia, it is often not stated when the place belonged to which area or country. For example, on Commons, the place of birth and death are displayed in the info boxes with the information from Wikidata with the details P17 and P131. If the place has no pictures, it has no category and is not linked on Commons. --HarryNº2 (talk) 23:10, 5 August 2022 (UTC)[reply]
Your sentence 1 question is hard to answer. But then you go back to throwing stuff against the wall. Neither "no information is given is also due to the fact that it is sometimes very difficult to name the affiliation of a place to a certain time" nor "it is often not stated when the place belonged to which area or country" are causes to add qualifiers to P20, but instead invite the value item's data to be improved. And I'm not seeing any place of birth or death info in the commons infobox for Douglas Adams (Q42) - https://commons.wikimedia.org/wiki/Category:Douglas_Adams - nor any chasing back of P131/P17 trees from e.g. his 'educated at' locations.
Dealing with, shall we say, the Schleswig–Holstein question where the P131 and/or P17 for a place of birth or death has changed since the birth or death, there appear to be two main options: use the date of birth or death to estabish which P131 and P17 value from the place item should be used; or denormalise the data as qualifiers in the place of birth/death statement. It's reasonable to say that for WD, the priority should be in ensuring that the place item has a dated history of P131 and P17 data - benefitting all items using that place item - rather than 'improving' a single statement in a single item. However a case can certainly be made for the qualifier approach, and a suggested test is: that the P17 or P131 value has changed since the date of birth or death and that the change is material. And so that means probably not trying to reflect UK local government reorganisations, for instance, through P20 qualifiers, b/c it's immaterial if it was this urban council or that unitary authority. Given the stats I produced above, I think the mission looks Sisyphean and the probability it will be used is small, because who wants a dependency on data values found in such a tiny fraction of items? --Tagishsimon (talk) 23:45, 5 August 2022 (UTC)[reply]
And even in those cases, the qualifiers would just compensate for a shortfall in programming as (in theory) the necessary information could be extracted from the dated information on the location item. --Emu (talk) 23:50, 5 August 2022 (UTC)[reply]
Just passing by. Is there transitive information exploited by any infobox in a query way? Eg if someone is born in year Y in the town T, can infobox query and show the country where he was born ? I'd be curious about how the infobox can implement transitive informations. Bouzinac 💬●✒️●💛 07:47, 6 August 2022 (UTC)[reply]
@Bouzinac: Look at eg infoboxes on Commons, and the library they call on. Something like c:Category:Belvoir Castle can trace an entire chain of administrative units in its location line (and link them to Commons categories where available), even though P131 on the corresponding item only gives a value for the local parish. This could be extended to other properties with a place value. Jheald (talk) 10:44, 6 August 2022 (UTC)[reply]
Interesting. I tried to find where in the code did the Q816295#P131 be transformed into Belvoir, Melton, Leicestershire, Midlands de l'Est, Angleterre, Royaume-Uni ...? Bouzinac 💬●✒️●💛 11:42, 6 August 2022 (UTC)[reply]
Makes me think of Tagging for the renderer on OpenStreetMap. Don't. Fix your infobox. Multichill (talk) 17:26, 6 August 2022 (UTC)[reply]
People can't fix their infoboxes unless they know how to. Bouzinac would like to look at some existing code but has not found it yet. If you can help with that, please do. (I can't, unfortunately, because I also have no idea how/where it's done) - Nikki (talk) 00:11, 7 August 2022 (UTC)[reply]
It is in here: https://commons.wikimedia.org/wiki/Module:WikidataIB - Utility function number 25. But this function does not help with displaying the correct country for the specific time period when an event occured. Since that module can take into account qualifiers, it could possibly be added. Sotho Tal Ker (talk) 17:40, 7 August 2022 (UTC)[reply]
I indeed see a (very complicated) function location. What I was thinking about (but I'm not skilled enough in Lua...) is how to show a calculated continent (P30).
Let's say there is an infobox about rivers, eg that one https://fr.wikipedia.org/wiki/Module:Infobox/Rivi%C3%A8re.
How should the infobox be calculating what continent is the river in, as there are almost no direct P30 statements on rivers' items as many WD people did remove it from items ?
So that frwiki infobox does not show any continent eg in Veauvre whilst this one does Petite Rivière Saint-Jean (because of direct P30), the both having the same infobox module. Bouzinac 💬●✒️●💛 13:01, 8 August 2022 (UTC)[reply]
How should the infobox be calculating what continent is the river in, as there are almost no direct P30 statements on rivers' items as many WD people did remove it from items ?
The most obvious way would be to follow located in the administrative territorial entity (P131) until you find a value for continent (P30).
That being said, the semantics of all the properties should ideally be formally specified as part of an entailment regime for Wikidata, instead of having to code their semantics across numerous infobox modules. I hope that's on the development team's roadmap. Silver hr (talk) 19:31, 11 August 2022 (UTC)[reply]
@HarryNº2 Whether or not there is an article in a Wikipedia for a place is totally irrelevant to Wikidata as we can model it here independently of any articles and I'm surprised why you even are bringing it up here. Could you please expand on why you think that would have relevance? Ainali (talk) 19:18, 6 August 2022 (UTC)[reply]
The data are often taken from Wikipedia and are thus related. Moreover, I have only mentioned the German Wikipedia as an example in order to refer to the personal data there, see de:Hilfe:Personendaten. --HarryNº2 (talk) 19:34, 6 August 2022 (UTC)[reply]
@HarryNº2 Well, data should have external references and that is where the actual relation is. Wikipedia only acts like a middleman, and should be cut out of the information transaction if possible. So that argument falls flat to the ground. Ainali (talk) 07:52, 7 August 2022 (UTC)[reply]
Be realistic and look at the data sets for the countries, the information is mostly taken from the Wikipedia articles. That Potsdam today is in Brandenburg in Germany is undisputed. In 1933, Potsdam was in Prussia in the German Reich, that is also a fact. --HarryNº2 (talk) 08:04, 7 August 2022 (UTC)[reply]
The argument is that articles should have both location and country to speed access, else the system has to chase the tree though a hierachy of places. But for a qualifier like this, it does seem overkill, as we wont have searches on people dying in Austria. And for places switching countries, omission is better as you could use the date to deduce the country Linz was in at the time. Vicarage (talk) 18:02, 6 August 2022 (UTC)[reply]
It is redundant and unnecessary. Silver hr (talk) 18:50, 6 August 2022 (UTC)[reply]

Undeletion request for Alan MacMasters (Q15994606)[edit]

I recently added three BBC references to this item. It's notable because it's a hoax Piecesofuk (talk) 06:24, 12 August 2022 (UTC)[reply]

Ping Ymblanter as deleting admin. --Emu (talk) 06:44, 12 August 2022 (UTC)[reply]
The only sitelink, the English Wikipedia one, was deleted after discussion as a hoax. I would recommend to try undeleting it there first, we are not in a good position to investigate references. Ymblanter (talk) 06:52, 12 August 2022 (UTC)[reply]
BBC References: https://www.bbc.co.uk/programmes/articles/1D0xzxYf9ykH9gcYp9BnYW4/5-everyday-objects-with-curious-histories, https://www.bbc.co.uk/bitesize/topics/zxwxvcw/articles/zjm4bdm, https://www.bbc.co.uk/newsround/56192812; The Scotsman: https://www.scotsman.com/lifestyle/food-and-drink/scottish-fact-week-electric-toaster-1548627; Daily Mirror: https://www.mirror.co.uk/news/uk-news/made-in-the-uk-the-life-changing-everyday-innovations-1294240
There'e plenty more. They're all repeating the hoax which apparently originated in the Wikipedia article. The item needs to be undeleted and flagged as a hoax to 1) record the fact that this is a hoax and help prevent it from being re-added to Wikidata/Wikipedia and 2) record that all these supposedly reliable sources were taken in by a fake Wikipedia article Piecesofuk (talk) 07:15, 12 August 2022 (UTC)[reply]
I would say the item should be undeleted, and possibly flagged as an instance of human whose existence is disputed (Q21070568). Do we have credible sources confirming it is in fact a hoax? To be honest, just because a gang of Wikipedians declares something false or a hoax, that doesn't mean it's necessarily so (this should be blatantly obvious), and obviously things can be Wikidata notable even if not Wikipedia notable. There's also a conceptual difference between a person being a hoax and an alleged picture of them being a hoax, which appears more likely. In addition to the news items above, one can find claims of an Alan MacMasters having invented the toaster in a 2015 book published by Amberley Publishing (Q47653625) and a 2016 book published by Dorling Kindersley (Q1245484). Now, since all of these sources are younger than the alleged hoax, it's quite possible that subsequent sources have perpetuated a fiction that originated on Wikipedia. That would be unfortunate. But Mr. MacMasters, real or not, remains a clearly identifiable conceptual or material entity that can be described using serious and publicly available references, and so meets Wikidata's lilliputian notability threshold. And I don't think I've ever seen anyone on Wikidata argue that truth is more important than verifiability (although that might be nice). Wikidata fundamentally ONLY cares about data, good or bad. A pile of dog turds with an external identifier is grounds for import. -Animalparty (talk) 07:45, 12 August 2022 (UTC)[reply]
Not sure how credible this site is but this was posted yesterday https://wikipediocracy.com/2022/08/11/wikipedias-credibility-is-toast/. I think we should have an item "longstanding Wikipedia hoax" and anything listed at https://en.wikipedia.org/wiki/Wikipedia:List_of_hoaxes_on_Wikipedia should be an instance of that; I expect most of these have been deleted from Wikidata, might be worth looking at undeleting them? Piecesofuk (talk) 10:02, 12 August 2022 (UTC)[reply]
Also I couldn't find any record of Alan Macmasters in any newspaper archives or genealogy websites, as far as I can see he never existed. Piecesofuk (talk) 10:04, 12 August 2022 (UTC)[reply]
I restored the item and changed the first property.--Ymblanter (talk) 19:14, 12 August 2022 (UTC)[reply]
Thanks for that. However, none of the original references support the claim of "human who may be fictional" (they all presume the human to be real). I added hoax (Q190084) based on this posting, and there may be more reliable press in the near future to plant a further nail in the coffin of this likely hoax. Other claims like humanness or birth date should probably be deprecated and/or qualified appropriately, so that it's clear this was a hoax that several credible sources perpetuated as real. -Animalparty (talk) 20:51, 12 August 2022 (UTC)[reply]
Thanks for undeleting. I agree that the references support the claim that he was an instance of human and not fictional human. Every statement apart from the hoax statement should be deprecated, not sure what an appropriate reason for deprecation should be, is there an existing one or should a new one be created? Piecesofuk (talk) 04:35, 13 August 2022 (UTC)[reply]

decentralized identification products[edit]

i'm working on a review of decentralized identification products https://www.wikidata.org/wiki/Q95350968 and would like to get some help

https://docs.google.com/document/d/1AaqMxO467xzg7_Z7bMen8I75XJgmMPheaf-QcZtIyvE/edit#

we've been working on a schema and have a list of open source products, would be great to have a pro help us with this

Proposed Schema: Transferable or delegatable? Pseudonymous, KYC etc… : type of classification Network availability : eth, btc, sol etc Token Utility: description Dynamic metadata : y/n Supply limit: is there one y/n if yes how much Decentralized metadata: y/n Unique per wallet : y/n Known integrations : list Security Audit Status : date Repository : url Wallet integration : list Proof mechanism : proof of ??? Ansondparker (talk) 13:55, 12 August 2022 (UTC)[reply]

Wikidata works in terms of properties and not in terms of whatever natural language terms you can think of for your schema. If you want to engage with Wikidata it makes sense to read through our existing properties and try to express what you want to express in those. ChristianKl ❪✉❫ 09:00, 13 August 2022 (UTC)[reply]

How does merging from a connected Wiki work?[edit]

The Viscount Who Loved Me (Q112061793) and The Viscount Who Loved Me (Q111190660) have Tag: Sitelink Change from Connected Wiki. What creates that merge? Is there a tool that's running on other Wiki's that does unclean merges on Wikidata? ChristianKl ❪✉❫ 08:10, 13 August 2022 (UTC)[reply]

Wikimedians with Commons Creator pages[edit]

I always see deletions of entries for Wikimedians with Commons Creator pages. I think we should keep people with Commons Creator pages. One day their images will fall into the public domain and be free from the restrictions of the creative commons license. The only way to track that, will be through their Wikidata entry which has information on birth and residence, to match them to an obituary when they die and calculate their post mortem auctoris (pma). Or is it so far into the future that Wikidata and Wiki Commons will not be around by then? For comparison, look at how many authors we have information on, where we only have a name, and no birth and death information. I think if contributors want to be identified by their real names and not their login ID, we should be tracking that. RAN (talk) 18:13, 6 August 2022 (UTC)[reply]

That’s not the only way, people do find out about the expiration of copyrights even without Wikidata. They have managed to do so for many decades.
In any case, we have discussed similar ideas many, many times. If we create a Wikidata item for every person that manages to create something on Commons, this would effectively mean that we could eliminate WD:N and accept uselessness of our data. We already tolerate a significant number of items about Wikimedians that are in violation of our own guidelines (mostly probably because no admin wants to be the bad guy). Let’s apply the same standard we use for people from other volunteer projects: “serious and publicly available references”. Self-reported information and meta stuff within Wikimedia project can never meet that criterion. --Emu (talk) 23:16, 6 August 2022 (UTC)[reply]
If having a Commons Creator page automatically means that we keep the page of a person, that would make it very easy for spammers to create a Common Creator page to get notability and the SEO guides will recommend doing so to get a Wikidata entry to stick. ChristianKl ❪✉❫ 07:59, 7 August 2022 (UTC)[reply]
Exactly. Lymantria (talk) 08:50, 7 August 2022 (UTC)[reply]
So you can then nominate things on Commons for deletion, then get things deleted here? It really isn't difficult to do. Thanks. Mike Peel (talk) 13:09, 7 August 2022 (UTC)[reply]
@Mike Peel: The amount of work a SEO person has to do to create a Common Creator page that won't get deleted when nominated for deletiong relatively small. ChristianKl ❪✉❫ 08:57, 8 August 2022 (UTC)[reply]
@ChristianKl: Not sure I understood, but a) it's not trivial, particularly discovering how to do so, and b) I would expect that it would be deleted quite quickly if it really is spam? Thanks. Mike Peel (talk) 09:56, 8 August 2022 (UTC)[reply]
Wikimedia Commons have massive backlogs. Assumption "would be deleted quite quickly if it really is spam" is wrong Mateusz Konieczny (talk) 10:48, 8 August 2022 (UTC)[reply]
@Mike_Peel: Discovering how to do so stays hard till one of the SEO people writes a guide of "Here's what you need to do to get on Wikidata" and people offer the service on Fiverr. ChristianKl ❪✉❫ 11:47, 8 August 2022 (UTC)[reply]
  • I think some nuance is missing here. It seems that the comments here are going for two extremes, neither of which will work:
    • If we have a blanket rule to keep all Wikidata items for creators with Commons creator pages, we have a big open door for spam.
    • If we have a blanket rule to delete all Wikidata items that rely on Commons creator pages for their Wikidata notability, we will break many Commons creator pages that pass Commons inclusion criteria.
    The solution will need to be somewhere in between: deletion requests should be handled on a case by case basis, and both Wikidata and Commons should accept such deletion requests. The decision to keep or delete the Wikidata item / Commons creator page should consider incoming links to both the item and the page. The outcome of a deletion discussion on Wikidata or Commons should inform but not prejudice the outcome of the other deletion discussion. (I came here off the back of an off-wiki discussion with Mike Peel.) Deryck Chan (talk) 11:11, 8 August 2022 (UTC)[reply]
    I just looked at Commons own description about who is supposed to get a creator page:
    "Anybody who is an author or creator of works hosted on Commons and meets notability requirements for Wikidata items, Commons categories or for Wikipedia articles. In most cases creator templates should be linked with Wikidata items."
    If there's no external link to a commons category or a Wikipedia article and we think that an item doesn't qualify under Wikidata notability requirements, any commons creators page that exists seems to be in violation of the Commons policy. ChristianKl ❪✉❫ 11:57, 8 August 2022 (UTC)[reply]
    You wrote: "I just looked at Commons own description", but that is marked as a proposal and doesn't have the authority of Wikilaw. Now is the time to discuss our options. Some people are worried that having a Creator_template/Wikidata_entry pair will become a backdoor for spammers and bots. What is we had a minimum number of Commons entries before you got a Creator_template? We often encounter people loading a single image of themselves and creating a wikidata entry, but they quickly get deleted. --RAN (talk) 20:00, 8 August 2022 (UTC)[reply]
    Wouldn’t that result in even more spam on Commons (if you need 50 pictures, well, just upload 50 pictures)?
    In any case, this doesn’t solve the problem that we would favor Wikimedians and effectively hand out Wikidata items for contributions. As I have said before, this is tantamount to nepotism. It hurts our credibility and also our data quality because most information on Wikimedians is 100% self-reported. This kind of data is problematic in many ways. --Emu (talk) 20:15, 8 August 2022 (UTC)[reply]
Every movie I have seen lists all the production people, I have never thought of it as a form of nepotism. --RAN (talk) 05:38, 12 August 2022 (UTC)[reply]
We do, too. At the appropriate location: In the history. --Emu (talk) 05:43, 12 August 2022 (UTC)[reply]
  • If you would limit the items to non-self reported data, that may help? If death dates are not publicly reported someplace, it would be tricky for privacy reasons anyway. If you use the same logic across Wikidata that may have wider implications of course... Effeietsanders (talk) 20:59, 8 August 2022 (UTC)[reply]
    @Effeietsanders You mean items with little more than instance of (P31)human (Q5)? Hm, that kinda might defeat the purpose of Wikidata. I’m not sure we have to keep this kind of information on a general-purpose database anyway. If I understand MisterSynergy correctly, structured information of Wikimedians could be stored in another database (yet to be developed). The alternative would be to allow only well-sourced items about Wikimedians. (Sidenote: It’s probably not that hard for a prolific Wikimedian to give an interview in the local newspaper about their contributions – at least in some parts of the world. If the newspaper is serious, this would create WD:N #2 notability anyway. Not to mention authority records, etc.). --Emu (talk) 22:18, 8 August 2022 (UTC)[reply]
    Note that there is no need to develop anything new, Commons already has its own wikibase instance, where structured information about (non-notable) authors can be stored.--Jklamo (talk) 08:21, 10 August 2022 (UTC)[reply]
    This sounds like the solution. Commons has its own rules about notability and its own Wikibase instance where it can keep structured data about things it finds notable. Win-win. Silver hr (talk) 16:41, 11 August 2022 (UTC)[reply]
    @RAN I don’t know if that’s true but this seems to be a solution for you copyright expiry problems? --Emu (talk) 10:22, 13 August 2022 (UTC)[reply]
    Maybe XX uploads AND accound old YY years. This would eliminate contemporary spammers (Item about creator will be notable after YY years ;-) JAn Dudík (talk) 21:04, 8 August 2022 (UTC)[reply]
    If it's not a policy and Commons currently lacks a policy to decide who deserves a Commons creator page that would be a strong reason currently not to use that information for our decisions about notability at Wikidata. ChristianKl ❪✉❫ 07:22, 9 August 2022 (UTC)[reply]
    @ChristianKl: You've raised a great point. We can't let "fulfils an interwiki structural need" be an unlimited notability criterion. WD:N should make an exception for situations where it is found that the concerned interwiki link goes to a part of a sister project with zero inclusion criterion (or zero enforcement thereof), and deletion from Wikidata should be permitted. Deryck Chan (talk) 22:00, 10 August 2022 (UTC)[reply]
    I do see "fulfills a structural" need as a statement that's about a structural need of Wikidata. Otherwise it wouldn't make sense to have the point about sitelinks separately. ChristianKl ❪✉❫ 08:28, 13 August 2022 (UTC)[reply]

How do you change your default observations from Cc-by-NC to cc0 or cc-by-Sa[edit]

Asking for iNaturalist iNaturalist (Q16958215) users, thks Jane023 (talk) 14:07, 10 August 2022 (UTC)[reply]

It is unclear what you're asking. In the absense of a clarification, I can only point you to the text at the bottom of every page: All structured data from the main, Property, Lexeme, and EntitySchema namespaces is available under the Creative Commons CC0 License; text in the other namespaces is available under the Creative Commons Attribution-ShareAlike License; additional terms may apply. Silver hr (talk) 15:38, 13 August 2022 (UTC)[reply]

SPARQL query times out[edit]

The following simple SPARQL query succeeds: 

SELECT ?item ?itemLabel 
WHERE 
{
  ?item wdt:P31 wd:Q55983715.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}

However, I want to further constrain the results to instances of Q729 or one of its subclasses: 

SELECT ?item ?itemLabel 
WHERE 
{
  ?item wdt:P31 wd:Q55983715;
        wdt:P31/wdt:P279* wd:Q729.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}

This query times out. What am I doing wrong? --2A02:8108:50BF:C694:9542:45CE:E09E:3014 21:28, 13 August 2022 (UTC)[reply]

1. Asking the question on the wrong forum - use Wikidata:Request a query for SPARQL queries 2. The optimizer sometimes makes poor choices and needs to be given a hint or two; see Wikidata:SPARQL query service/query optimization:
SELECT ?item ?itemLabel 
WHERE 
{
  ?item wdt:P31 wd:Q55983715. hint:Prior hint:runFirst true.
  ?item wdt:P31/wdt:P279* wd:Q729. hint:Prior hint:gearing "forward".
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!
--Tagishsimon (talk) 21:33, 13 August 2022 (UTC)[reply]
Thank you! – Unfortunately I find it very cumbersome to find the right forum here at Wikidata (it’s even tricky to find Project chat given that searches prefer items over meta pages). --2A02:8108:50BF:C694:9542:45CE:E09E:3014 21:39, 13 August 2022 (UTC)[reply]
No problem. If you had an account you could watchlist them both, but if you want to continue as an IP user - which is fine - then not so much. Project chat is the third link on the left side menu (I guess, for computer users, maybe not mobile.) Request a Query slightly more obscure, navigation-wise. --Tagishsimon (talk) 21:43, 13 August 2022 (UTC)[reply]

How to properly reference the fact the height of Marcus Rashford is disputed[edit]

On 14 September 2019 at about 12:00 local Q5214544 (Daniel Meirion Walker) on Q5465835 (Football Focus) claimed the height given on the English Wikipedia of 180cm (or 5' 11" / 180cm) did not seem credible. With regards to a statement of his height (P2048), 180cm, which is referenced, I have managed to add a property P1310 that this height is disputed, but my expertised does not go beyond how to model or add a reference to the P1310 property. Before I make a complete mess of this it is time for me to back off and ask for help. thankyou. Djm-leighpark (talk) 09:00, 13 August 2022 (UTC)[reply]

@Djm-leighpark: supports qualifier (P10551) is the property you are looking for, I guess. I would model this like this: Q22951255#P2048 (reference section) - Valentina.Anitnelav (talk) 10:12, 14 August 2022 (UTC)[reply]
Thanks for sorting that. Djm-leighpark (talk) 14:05, 14 August 2022 (UTC)[reply]

Functional program data type[edit]

For data such as power series, there is currently only mathematical notation supported to represent these, and unfortunately these are not strictly machine-interpretable; notations such as superscripts and subscripts are ambiguous, and have different meanings in different contexts. For algorithms, there is nothing other than string types, and including arbitrary code fragments in general-purpose programming languages in Wikidata would be actively dangerous, as people will be tempted to execute them, and code in pseudocode is less than useful, as it is not machine-interpretable or capable of being manipulated by systems such as program provers.

Functional programs in a LISP-like language such as Scheme are unambiguous, and meaningful for machine interpretation. If they are constrained to a safe subset of those languages, they are also safe to interpret directly, offering other potential uses, such as following the general approach of Structure and Interpretation of Computer Programs and Structure and Interpretation of Classical Mechanics. They could also be converted into TeX notation automatically, in much the same was a GNU Lilypond notation can be converted into MIDI or sound files.

Implementation would be relatively simple: all that is needed is an S-expression parser with an post-pass that checks the parsed forms against a allow-list of symbols and syntax productions. Although the general case of proving safety (including, of course, program termination!) is insoluble, we don't need to be able to define all Turing-complete functions, just relatively small subsets (for example, consider only a subset restricted to numbers and keywords such as let, letrec, lambda, *, -, +, /, expt, one-letter variable names, cons, first, last, simple conditionals and logic expressions -- trivially safe). I'd be happy to write the parser and safety-checker, in any reasonable language of the dev team's choice, if the proposal is accepted.

Note also that the programs can also be pretty-printed for output.

Oh, and lazy semantics (such as in https://docs.racket-lang.org/lazy/index.html) can be used to represent infinite series and so on... The Anome (talk) 11:43, 10 August 2022 (UTC)[reply]

WikiFunctions is currently in development. ChristianKl ❪✉❫ 15:00, 10 August 2022 (UTC)[reply]
@ChristianKl: Sounds fantastic. Will it eventually be fully integrated with Wikidata? The Anome (talk) 13:57, 14 August 2022 (UTC)[reply]
There's a desire to have Wikidata integration. How that will look like exactly likely needs a decent amount of discussions to find consensus on Wikidata. From my perspective those discussions will be had once WikiFunctions is further along.
I personally imagine that most items descriptions should be created in the future via WikiFunctions. WikiFunctions will be much better than bots creating inconsistent descriptions. ChristianKl ❪✉❫ 17:43, 15 August 2022 (UTC)[reply]

Question about P968[edit]

Hello there, Wikidata editors.

I am new to Wikidata and don't have much experience, but I have a question.

Shouldn't there be a property formatter URL mailto:$1? Why do I have to write mailto: before the email address? I am new here, correct me if I am wrong, QuickQuokka (talk) 23:30, 10 August 2022 (UTC)[reply]

Also, shouldn't it use the official RFC 5322 email regex, (?:[a-z0-9!#$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*|"(?:[\x01-\x08\x0b\x0c\x0e-\x1f\x21\x23-\x5b\x5d-\x7f]|\\[\x01-\x09\x0b\x0c\x0e-\x7f])*")@(?:(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?|\[(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?|[a-z0-9-]*[a-z0-9]:(?:[\x01-\x08\x0b\x0c\x0e-\x1f\x21-\x5a\x53-\x7f]|\\[\x01-\x09\x0b\x0c\x0e-\x7f])+)\]) or official W3C RegEx, /^[a-zA-Z0-9.!#$%&’*+/=?^_`{|}~-]+@[a-zA-Z0-9-]+(?:\.[a-zA-Z0-9-]+)*$/ QuickQuokka (talk) 23:48, 10 August 2022 (UTC)[reply]
As to the first question/suggestion, it has been discussed but failed to raise any interest; subsequently an objection was raised that all consumers of WD P968 data - such as language wiki templates - would need to be changed should WD change its data format. See property talk & an RfC. --Tagishsimon (talk) 00:04, 11 August 2022 (UTC)[reply]
@Tagishsimon: Again, I am new to Wikidata and this may be a stupid question: Is it possible to use 1 formatter url if it matches the current regex and another if it uses my proposed regex? QuickQuokka (talk) 08:50, 11 August 2022 (UTC)[reply]
No, not possible to do that afaik. --Tagishsimon (talk) 08:56, 11 August 2022 (UTC)[reply]
I don't see that RFC 5322 specifies a regular expression directly, however, the W3C page does, and obviously we should defer to the standard instead of making up our own email regex. Silver hr (talk) 15:59, 13 August 2022 (UTC)[reply]
@Silver hr This regex defined by W3C is trivially wrong as it doesn't match for instance é@example.org. If I remember correctly, email grammar cannot be expressed as regex (meaning there is no regex matching exactly every email).
A good solution is to use a regex matching more than only valid emails, for example as defined in RFC 4287: The Atom Syndication Format (Q47471082):
# Whatever an email address is, it contains at least one @
atomEmailAddress = xsd:string { pattern = ".+@.+" }
— Metamorforme42 (talk) 13:28, 15 August 2022 (UTC)[reply]
In my opinion, we shouldn't keep an inferior solution just because some work is needed to adapt to a better solution, so I would support this change if there is a good reason for it. At first glance, the mailto: part seems redundant. Silver hr (talk) 15:53, 13 August 2022 (UTC)[reply]
@Silver hr what work do you think should be done before making the change? How do you plan to get every third party who might use the property to change? Who do you think is supposed to do that work? ChristianKl ❪✉❫ 11:58, 15 August 2022 (UTC)[reply]
Can I also point up the conjunction of "an inferior solution" (it's not explained why it is inferior) with "if there is a good reason for it". Just so much tiring hot air. --Tagishsimon (talk) 12:05, 15 August 2022 (UTC)[reply]
The first part of my sentence was general. The second applied to the specific issue, and it was meant to invite User:QuickQuokka (and anyone else) to elaborate on what benefit he or she thought the change would bring.
Also, if reading tires you, might I suggest taking a break? Silver hr (talk) 21:26, 15 August 2022 (UTC)[reply]
@Silver hr: I just think it's clunky and unintuitive. Also, btw, MediaWiki has a magic word where it you can add different text options for different genders. I know you use he/him pronouns because I typed {{gender:Silver hr|he/him|she/her|they/them or other}}. QuickQuokka (talk) 21:34, 15 August 2022 (UTC)[reply]
what work do you think should be done before making the change?
I don't know, I haven't looked into it.
How do you plan to get every third party who might use the property to change?
If you're talking about WMF projects, I suppose by notifying them. Again, I haven't looked into it.
Who do you think is supposed to do that work?
I'm not sure what work you're referring to exactly. If you mean the work of adapting to the change, those who use the property, obviously. Silver hr (talk) 21:20, 15 August 2022 (UTC)[reply]
  • On an email regex, there's a nice discussion here - "If you thought something as conceptually simple as validating an email address would have a simple one-size-fits-all regex solution, you’re quite wrong. ... Allowing invalid addresses to slip through may be preferable to annoying people by blocking valid addresses. For this reason, you may want to select the “simple” regular expression." I think keeping the regex simple here is preferable to trying to be overly specific. ArthurPSmith (talk) 16:10, 15 August 2022 (UTC)[reply]

Removing an Identifier for unscientific reasons[edit]

Hey everyone. I am just importing an Arabic Wikipedia discussion, w:ar:ويكيبيديا:الميدان/سياسات#ضبط استنادي, to here, where it's related. An Arabic Wikipedia user requested that. Briefly : "Do we can remove National Library of Israel J9U ID from Identifiers of the articles related to Arab people?" Looks like they resent when see that ID in Authority control of Arab people, or find it mocking. I don't know why exactly, but looks like it related to 74 years, 2 months, 3 weeks and 5 days of mutual conflict, or w:International recognition of Israel? Anyways, It seems unscientific reason. I will translate any answer, or ping that user to see comments. With Regards. Ruwaym (talk) 00:48, 14 August 2022 (UTC)[reply]

We cannot. Whatever is going on between Israel and Arabic countries, it is not a conflict which should influence WDs data. Arabic wikipedia is quite capable of taking decisions about what wikidata they import into their wikipedia, without seeking to influence the data held at source. It would, obvs, be appalling were Arabic wikipedia to start censoring Israeli IDs, but whatever, it is a decision for them. WD absolutely will not interest itself in the user's request, which is completely antithetical to pretty much everything the wikimedia movement stands for. --Tagishsimon (talk) 01:45, 14 August 2022 (UTC)[reply]
Arab Wikipedia is free to decide in their authority control template which properties they want to import and could also write logic to make that decision based on other Wikidata properties to make the template behave differently for Arab people.
Wikidata itself is pluralistic and won't remove data on those grounds. ChristianKl ❪✉❫ 08:26, 14 August 2022 (UTC)[reply]
Hello everyone,
I am an admin from Arabic Wikipedia.
In the Arabic version, we do not add or remove information based on personal opinions. This is a pillar, and it is out of the discussion. Any Israelian-sourced information is welcomed as long as its source is reliable.
The person who proposed this got the following answer:
Hi Ramy,

Just a comment that what you are proposing is contrary and unacceptable in Arabic Wikipedia, because it collides with one of the five pillars of the encyclopedia: Wikipedia adopts a neutral point of view.

Please note that we, in the Arabic Wikipedia, do not consider editors's opinions or their orientations regarding conflicts and disputes, whatever their form: political, religious, ethnic, cultural, economic or sportive. We simply present the well-documented information with its sources (except for fringe theories).

So, if you have a personal opinion of a conflict, whatever that opinion is, whatever that conflict is, Wikipedia is not the place for it. We simply don't discuss our personal thoughts or allow them to influence our contributions here.
I am just wondering why Ruwaym did not translate this. I am also referring to the block he got on English Wikipedia for "Disruptive editing: competence issues as well as repeated ethno-nationalist aggression and personal attack". He is also got blocked in Arabic Wikipedias (twice) for dishonesty when delivering others' opinions, exactly as he did above.--Michel Bakni (talk) 10:42, 14 August 2022 (UTC)[reply]
@ChristianKl @Tagishsimon Do i said i translated anything from anyone? Do i said i repsenet the Arabic Wikipedia? Do i said at all I'm an "Arab" or "Jewish"? Do i did any wrong? Wait, there is a value judgment based on block log. Do you know Michel Bakni accused me for "Lack of honesty and distorting the image of the Arab community in a foreign forum". Do you know it's not the first time that Micheal Bakni falsely accused me? Ruwaym (talk) 14:36, 14 August 2022 (UTC)[reply]
Michel Bakni is right on this. Your use of "they" made it seem like the discussion has already reached consensus on Arabic Wikipedia. Moreover, the discussion has nothing to do with Wikidata. Every language project is free to display whatever identifiers they want from Wikidata. Please stop this meaningless discussion here as it has nothing to do with Wikidata. Vojtěch Dostál (talk) 15:07, 14 August 2022 (UTC)[reply]
Yeah, Please stop unscientific methods of talking. Ruwaym (talk) 15:46, 14 August 2022 (UTC)[reply]
This "meaningless " discusion is done some hours ago. I got infinite block in Ar wk because of "Lack of honesty and distorting the image of the Arab community in a foreign forum", and labled as a "vandalism-only account". Time to celebrate it! Ruwaym (talk) 18:49, 14 August 2022 (UTC)[reply]
@Michel Bakni You requested blocking me for ininite and you said w:ar:ويكيبيديا:إخطار الإداريين/منع/الحالية#منع: Ruwaym "...In an irrelevant foreign forum, Wikidata, which resulted in a clear insult to the image of the Arab community as a whole, and the appearance of the Arabic version of the encyclopedia in a fanatical and unacceptable appearance (and this can be seen from the responses to his question)." Do you think I'ts not a clear insult or a fanatical and unacceptable behavior if this "foreign forum" knows i got blocked infinite just because of this "meaningless " discusion? and your prejudice, value judgment and cross-wiki hounding users? Why a user blocked in Ar wk because of a Wikidata talk?!
@-revi, Nikki, Mahir256: Feel free to block me here too or ask a global lock/block for my account. I am R W of telegram groups. Ruwaym (talk) 10:15, 15 August 2022 (UTC)[reply]
Ruwaym, you are forwarding troll postings. Why it would be useful in any way? If Michel Bakni correctly described situation then "Lack of honesty and distorting the image of the Arab community in a foreign forum" was a justified reason for block, and I would also block you for trolling Mateusz Konieczny (talk) 14:26, 15 August 2022 (UTC)[reply]
@Mateusz Konieczny I'm sorry, ok I stop. Discusion is done for me. Ruwaym (talk) 17:22, 15 August 2022 (UTC)[reply]

Merge[edit]

Please, merge Q28874524 and Q3181354. There are 2 name of 1 species of bird. Thanks. Zagrebin Ilya (Talk page) 07:48, 14 August 2022 (UTC)[reply]

Merging is not appropriate for such pairs of homotypic names. See Wikidata:WikiProject Taxonomy/Tutorial § Homotypic names for details. William Avery (talk) 12:04, 14 August 2022 (UTC)[reply]
William Avery, OK. Zagrebin Ilya (Talk page) 08:14, 15 August 2022 (UTC)[reply]

Wikipedia References/Cites, Categories, and Infoboxes[edit]

Are any of the following Wikipedia items used in WIkidata ?

  • references and cites
  • categories
  • info boxes
  • projects
  • tables and list articles

Wakelamp (talk) 13:43, 15 August 2022 (UTC)[reply]

I'm not sure what exactly you mean with Wikipedia items. Are you speaking about Wikimedia internal item (Q17442446) that relate to Wikipedia? ChristianKl ❪✉❫ 15:16, 15 August 2022 (UTC)[reply]
  • We imported data from Wikipedia to create the first entries, you will see a reference "imported from Wikipedia" and the date when a fact has been added. A bot still uses Categories to add info to Wikidata, I just saw a bot crawling through the alumni categories to add in educated_at= to Wikidata. And it works the other way, most non-English Wikipedias use an auto-generated info box, drawing the info from Wikidata. --RAN (talk) 00:02, 16 August 2022 (UTC)[reply]

São Paulo Museum of Art (Q82941)[edit]

I am trying to change the official site claim from http://masp.art.br to https://masp.org.br (you can even see the first link redirects to the second one) but I am really struggling to change it. I think the page is locked from editing. Can someone do it for me then? I asked it on its talk page months ago but apparently no one saw it. AnAkemie (talk) 02:47, 20 August 2022 (UTC)[reply]

The existing claim should be left in place, with a new claim of peferred rank added (done). It is, I guess, for this reason that editing an official website value is not permitted. --Tagishsimon (talk) 02:51, 20 August 2022 (UTC)[reply]
Actually, it's semi-protected as a high-use page. In any case, thanks to you both for getting the fix made. JesseW (talk) 02:53, 20 August 2022 (UTC)[reply]
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Emu (talk) 22:42, 21 August 2022 (UTC)[reply]

occupation vs role[edit]

to be a widower as an occupation? https://www.wikidata.org/w/index.php?title=Q109374369&type=revision&diff=1709394696&oldid=1528039972 78.55.59.122 10:46, 21 August 2022 (UTC)[reply]

I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Emu (talk) 22:41, 21 August 2022 (UTC)[reply]

instance of human AND instance of role[edit]

Rudy Baldwin https://www.wikidata.org/wiki/Q108239980 ? What is this item about? 78.55.59.122 10:50, 21 August 2022 (UTC)[reply]

I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Emu (talk) 22:41, 21 August 2022 (UTC)[reply]

A large number of proposed properties[edit]

Hello. There are 60 properties waiting to be created on Wikidata, which is a lot, especially with the activity of some of the project's contributors growing. The lag in this situation is unacceptable, since individual pages no longer display properties and now only text links to them. Can this situation be resolved? MasterRus21thCentury (talk) 20:26, 3 August 2022 (UTC)[reply]

There are also quite a few proposals which have been around for months & are clearly not going to be passed; they should be put out of their misery. And other proposals for which properties have been created, but which still linger as proposals. I couldn't see any guidance on when & how to close down stalled proposals having sufficient opposes to make it unlikely they'll ever be passed, nor on how archiving proposals works. --Tagishsimon (talk) 09:15, 4 August 2022 (UTC)[reply]
Looks like we ought to make the rules a bit more clear/complete, wouldn't you say? Silver hr (talk) 14:34, 4 August 2022 (UTC)[reply]
If there's no consensus and the proposal seems stale (certainly if it's been 6 months since last comments, or probably generally when it's been over a year since first proposed), then edit the proposal: put "not done" in the status field and ping all those who commented on the proposal to let them know what you've done. That automatically archives them. "withdrawn" status is also appropriate if the original proposer no longer supports the proposal, that also should automatically archive them. ArthurPSmith (talk) 19:04, 4 August 2022 (UTC)[reply]
Yes, in such a situation, a new order of creation of properties is needed in order to eliminate the queue. For example, with active voting, the creation time is reduced, and with a long one, it is increased. This also happens when discussing candidates for good and featured Wikipedia articles, whose discussion period is at least a month. MasterRus21thCentury (talk) 20:01, 4 August 2022 (UTC)[reply]
Marking proposals as done or not done is the task of people with the property creator right. User with that right are supposed to understand our current defacto norms. There's no need for formal rules. There's just a need for property creators to do the work. ChristianKl ❪✉❫ 20:57, 4 August 2022 (UTC)[reply]
That's right, but they themselves are inactive. Now they need a significant increase in activity in order to eliminate the growing debt. I could have helped, but I was stripped of my property creator flag in May and told to reapply in November. MasterRus21thCentury (talk) 06:32, 10 August 2022 (UTC)[reply]
Almost 1/2 of the properties in the queue right now have been marked as "ready" since last week. So the number of waiting properties looks the same as one week ago, but actually dozens of properties have been created in the meantime. It may not be as bad as it looks. Vojtěch Dostál (talk) 07:18, 10 August 2022 (UTC)[reply]
Another thing is that some properties are on the shelf for creation after reaching a consensus since April-May-June and have not yet been created by any of their authorized participants. Moreover, the same problem applies to the situation when a large number of properties on the topic of a particular country. Currently, 12 out of 55 "waiters" are sources for Russian sites. Since last year, I have been placing special emphasis on the development of properties related to Russia - now there are 116 more than in August 2021, that is, 137 against 253. Therefore, at present, it is necessary to limit the number of waiting lists at the level of 20-30 properties, so that there are no visible lines. MasterRus21thCentury (talk) 08:27, 10 August 2022 (UTC)[reply]
@MasterRus21thCentury Right now, there seems to be only 6 ready properties which were created in May or earlier Vojtěch Dostál (talk) 09:05, 10 August 2022 (UTC)[reply]
While a bunch of those are marked as ready I consider it questionable to label a property without an English description as 'ready' and wouldn't have created it when I was a property creator. I believe that properties should have an English description that explains what a source.
Putting things that are not Wikiprojects in Wikidata into the Wikiproject tab also makes things harder for property creators. ChristianKl ❪✉❫ 09:18, 10 August 2022 (UTC)[reply]
@ChristianKl That's understandable, but I'm in favor of tightening up the property creation policy, which includes shortening the time between being marked Done and creating it. Moreover, in Wikidata, the participants with these rights - more or about a hundred - are the creators of properties and administrators, but they do not put this work on stream, even if there is a significant consensus for its creation in a particular property. MasterRus21thCentury (talk) 15:34, 15 August 2022 (UTC)[reply]
Yes, I agree that the property creation policy should be tighter. Ideally, that includes rules that clearly say that properties that have no English description are not ready, so can easily marked undone instead of clogging up the list without any actions. ChristianKl ❪✉❫ 15:49, 15 August 2022 (UTC)[reply]
@ChristianKl At the very least, a general vote by Wikidata members is needed on a possible tightening of the policy regarding the timing of the creation of properties, so that later it does not drag out for weeks and months. The responsibility of property creators and administrators should also be much higher than it was before. MasterRus21thCentury (talk) 20:27, 15 August 2022 (UTC)[reply]
@MasterRus21thCentury Remember, property creators and administrators are volunteers without a contract. They don’t have any legal or moral obligation to use their rights. You can’t really make them if they don’t feel like it. You can only incentivize them. --Emu (talk) 21:38, 15 August 2022 (UTC)[reply]
I understand all this very well and even wrote to some of them on their talk pages. However, I don't see any clear progress yet. MasterRus21thCentury (talk) 22:05, 15 August 2022 (UTC)[reply]
If you would understand this very well, then it would sense to take actions that make it easier for property creators to create the properties you want to have created. In this case you don't seem to care enough about the properties being created to add the missing English descriptions. Ideally, good descriptions that tell the reader what the organisations of the identifiers are about. ChristianKl ❪✉❫ 06:43, 16 August 2022 (UTC)[reply]

Open Call to Join the Northern and Western Europe (NWE) Regional Funding Committee[edit]

The Wikimedia Foundation’s Community Resources team and the Northern and Western Europe (NWE) Committee invite you to apply to become a committee member in the Northern and Western Europe (NWE) region. The NWE Committee supports the Wikimedia Foundation Funds programs' participatory grantmaking practice. The deadline for applying is September 4, 2022.

The NWE Committee reviews grant proposals from affiliates and other community members within the Northern and Western Europe region who are seeking grant funding from the Wikimedia Foundation. Committee responsibilities include: providing thought partnership to help support applicants and strengthen their strategy and proposals; evaluating and recommending promising proposals for funding; and helping recruit and spread the word about grants. Orientation and training is provided for new members.

Serving committee members will have the opportunity to advise on funding decisions and proactively share recommendations and mentorship to support grantees in the Northern and Western Europe region in their development, growth, and sustainability strategies; that contribute to a world in which every single human being can freely share in the sum of all knowledge. You will also strengthen your capacity in participatory decision-making processes while learning about our community's incredible work and the different contexts that influence their work.

Deadline for submitting committee candidacy:

  • September 4, 2022

Please reach out to nwe_fund @ wikimedia  · org for questions about the NWE Committee.

Warm regards,

On behalf of the NWE Regional Funding Committee. Mike Peel (talk) 09:01, 16 August 2022 (UTC)[reply]

"Number of parts" property?[edit]

I'm looking to add a count of the number of components of mathematical links; for example, L7a3 (Q113139246) has 2 components, and Borromean rings (Q894166) has three. But I can't find the right property to use. number of objects (Q41792217) aka "quantity of components" is the concept I'm after, but it's an item, not a property

I guess I could use has part(s) (P527) knot (Q1188853) qualified by quantity (P1114), but is this the right way to go?

Can anyone help me? The Anome (talk) 08:50, 16 August 2022 (UTC)[reply]

has part(s) (P527)knot (Q1188853)quantity (P1114)N seems exactly right to me. Silver hr (talk) 15:40, 16 August 2022 (UTC)[reply]
I would agree has part(s) (P527) + quantity (P1114) looks like the way to go, and knot (Q1188853) does look like the right item for what those parts are called. (Though I'd expect you'd know more about that -- create a new item, if it's not exactly the right item for what the parts are). Jheald (talk) 15:43, 16 August 2022 (UTC)[reply]

Merge request for Toyota[edit]

Q53268 should be merged with Q113537184, Q113537225, Q113541067 and Q113541083 at Wikidata. Q53268 is for Toyota in all Wikipedia languages available for this article while Q113537184 is for Toyota in Swahili Wikipedia language, Q113537225 is for Toyota in Quechua language, Q113541067 for Boarisch, Q113541083 for Brezhoneg and are separated. Please merge them since Q53268 Wikidata is protected.SSHTALBI (talk) 12:36, 17 August 2022 (UTC)[reply]

It seems very misguided indeed to have created a series of new items one each for each sitelink, and to come here asking for them to be merged, rather than, for instance, coming here and asking for sitelinks to be added to a protected page. Please do not do that again. Even if creating a new item to host the sitelinks had been necessary - it was not - then a single item would have sufficed. --Tagishsimon (talk) 12:41, 17 August 2022 (UTC)[reply]

@Tagishsimon Since these pages have been created, you should add them to the Toyota Wikidata page. Please add them.SSHTALBI (talk) 12:49, 17 August 2022 (UTC)[reply]

"you should add them". Go away. --Tagishsimon (talk) 12:51, 17 August 2022 (UTC)[reply]

Toolhub Taxonomy: Call For Feedback[edit]

Hello!
The Wikimedia Toolhub team wants to hear from the Wikidata community. The team is working to improve tool discovery/search on Toolhub. This is a catalog of 1500+ tools used every day in a wide variety of workflows across many Wiki projects.

Toolhub data model describes what pieces of information Toolhub records contain about tools.

We've put together a feedback page[0] that describes this data model and the taxonomy. Please consider taking 5-10 minutes to leave feedback.
We'll very much like to hear from the Wikidata community. The period for giving feedback ends on the 21st of August.

Thank you.

[0] https://meta.wikimedia.org/wiki/Toolhub/Data_model/Feedback SSapaty (WMF) (talk) 14:31, 17 August 2022 (UTC)[reply]

items for individual birth / death certificates / registries[edit]

Notability of individual birth / death certificates[edit]

After several months of open discussion, the RfD for five items at Wikidata:Requests_for_deletions#birth_certificates still aren’t resolved:

Users have requested a general discussion that doesn’t seem to have happened yet, so let’s have it! Personally, I would limit the notability of items like these to exceptions like Barack Obama birth certificate (Q14527788).

ping User:Nomen ad hoc, RAN, VIGNERON, Hsarrazin --Emu (talk) 15:58, 10 August 2022 (UTC)[reply]

To me it seems they are notable if they are used on another item via the structural need provision of our policies. Wikibase works in a way where it's useful to have information spread over multiple items instead of having a lot of information in a single item. ChristianKl ❪✉❫ 16:15, 10 August 2022 (UTC)[reply]
I don't think structural need applies here because the item isn't needed to be able add the reference. I don't think we should create an item for any/every document used as a reference. I would limit it to cases where lots of items reference the same document and there is a clear benefit to having a separate item that outweighs the downsides of having a separate item. - Nikki (talk) 10:22, 14 August 2022 (UTC)[reply]
@Nikki You could add all the information in the item directly as a reference instead of a separate item. That would however be worse for Wikidata because it increases the costs whenever the item gets edited. ChristianKl ❪✉❫ 15:49, 17 August 2022 (UTC)[reply]
  •  Keep I don't think we need a Wikidata entry for every document stored at Commons, but I can see the use for some, so that we can use them as references for statements like birth and death dates. We aren't being overwhelmed with them, we have 5, and we should be experimenting with different ways of referencing statements. I have seen examples of people adding the image of the birth and death certificate as a reference, and I have seen more like the above, where they create a Wikidata entry, both are good experiments. --RAN (talk) 05:31, 12 August 2022 (UTC)[reply]

Notability of individual registries[edit]

Connected with those cases but also different are other cases such as Berlin death register (Q104508368) or death record of the Lutheran City Church of Vienna (Q99896173) (see this query – they aren’t about individual certificates but rather about (death registry of a specific church or within a city). Most were created by myself before this discussion. The modeling is bad at the moment, but before fixing it to something like in birth registry of the Jewish Community of Vienna (Q99662601), I’d like to have some input. Should we keep those? Should we get rid of them? --Emu (talk) 15:58, 10 August 2022 (UTC)[reply]

Ping Sapfan, Jura, Jc3s5h from the last discussion. To be perfectly honest, I don’t fully understand if you agreed on an all encompassing back then … --Emu (talk) 16:09, 10 August 2022 (UTC)[reply]

Popular data objects[edit]

On the page Wikidata:Main Page/Popular are seven data objects, which are designated as popular or as a highlight. Normally the DeltaBot refreshes the page every day. However, there has been no renewal since August 5, 2022. Unfortunately, also Pasleim, who runs the DeltaBot, has not been active in Wikidata or any other Wikimedia project since May 5, 2022. That's why you can't ask him why the bot doesn't update the page anymore. What can we do now? --Gymnicus (talk) 14:57, 16 August 2022 (UTC)[reply]

We can run it as well. Pasleim has published most of his scripts, including popularItems.py. —MisterSynergy (talk) 17:59, 16 August 2022 (UTC)[reply]
@MisterSynergy: Ah very good. Does the bot do it every day now? --Gymnicus (talk) 18:05, 16 August 2022 (UTC)[reply]
This was a one-time run for now. Honestly, I am not eager to take on another bot job, so if someone else wants to, I'll happily return this job again. I just wanted to note that the code is available and still running as is (but it looks very dated and would deserve some improvements). —MisterSynergy (talk) 18:12, 16 August 2022 (UTC)[reply]
@MisterSynergy: It looks like the DeltaBot has started working again after the edit of your bot. So no further processing is necessary from you or your bot. --Gymnicus (talk) 09:29, 17 August 2022 (UTC)[reply]
@Gymnicus, Pasleim: okay for now. I would otherwise be able to take over this job as well, pretty much immediately. —MisterSynergy (talk) 20:21, 17 August 2022 (UTC)[reply]

How to get the langwiki-link from QID?[edit]

Given a QID, how can I retrieve its langwiki link? For example, phosphorus (Q674) on dewiki is de:Phosphor. ])] ] DePiep (talk) 08:37, 15 August 2022 (UTC)[reply]

Something like one of these; depends if you are interested in de wiki only, or all language wikis.
SELECT ?item ?itemLabel ?sitelink ?article WHERE 
{
  VALUES ?item {wd:Q674}
  ?article schema:about ?item ;
  schema:isPartOf <https://de.wikipedia.org/> ; 
  schema:name ?sitelink .
  SERVICE wikibase:label { bd:serviceParam wikibase:language 'en' }
}
Try it! 
SELECT ?item ?itemLabel ?lang ?sitelink ?article WHERE 
{
  VALUES ?item {wd:Q674}
  ?article schema:about ?item ;
  schema:inLanguage ?lang ; 
  schema:name ?sitelink ;
  schema:isPartOf/wikibase:wikiGroup "wikipedia" .
  SERVICE wikibase:label { bd:serviceParam wikibase:language 'en' }
}
Try it!
--Tagishsimon (talk) 10:09, 15 August 2022 (UTC)[reply]
You can also simply parse https://www.wikidata.org/wiki/Special:EntityData/Q674.json and extract it from there. In the future, the REST API might perhaps provide direct access—but it's not ready yet, see Wikidata:REST API feedback round. —MisterSynergy (talk) 10:31, 15 August 2022 (UTC)[reply]
I think we do have more direct ways to get the link but I don't know how or where we might have documented them. ChristianKl ❪✉❫ 15:23, 15 August 2022 (UTC)[reply]
Thanks all. So no inline option then (module:WikidataIB?, after all it's just similar to QID-PID value). -DePiep (talk) 19:56, 15 August 2022 (UTC)[reply]
I think it's rather that nobody knows the inline way. I vaguely remember there having been one sometime. ChristianKl ❪✉❫ 22:26, 15 August 2022 (UTC)[reply]
Well, it seems that the question was poorly written in the first place. Based on the second comment by User:DePiep and their recent contributions, I suspect that they might perhaps be talking about MediaWiki template scripting using Lua modules, and Module:WikidataIB in particular. If I had known this earlier, I would have given a different answer than above. Clarification would help here. —MisterSynergy (talk) 22:58, 15 August 2022 (UTC)[reply]
I don't see what is "poorly written" in my OP, let alone misleading. I actually am expecting that Wikidata has a way to retrieve a single interwiki link [QID, lang-wiki] value from Wikidata. After al, that value is with the item. The example, peritem phosphorus (Q674): [qid=Q674, langwiki=dewiki] to return https://de.wikipedia.org/wiki/Phosphor or de:Phosphor. DePiep (talk) 14:56, 16 August 2022 (UTC)[reply]
Well, it does matter whether you want to retrieve this information at a Wikimedia-external location (e.g. at home, or on some webserver), or within an environment such as Toolforge where you have easy access to plenty of data sources, or from within an environment such as Lua modules on Wikibase clients. It would thus be useful if you could tell us your specific needs—the advice given will usually be much more helpful then.
Anyways—based on your recent contributions—let's assume you are talking about Lua module calls from template pages. There is indeed a function mw.wikibase.getSitelink( itemId, globalSiteId ) in the Scribunto interface of Wikibase clients (i.e. it is available on all WMF wikis that are connected to Wikidata, including Wikidata itself). Unfortunately, this function cannot be called directly from template code, so some Lua module needs to be wrapped around it in order to make it available for usage in templates—and that does not seem to be the case for some reason. However, it's a rather simple community task to make this available. —MisterSynergy (talk) 17:18, 16 August 2022 (UTC)[reply]
Thanks MS, sorry for the confusion. Your deduction of my need is correct: in this case, my need is for item research/documentation in WD WikiProject Chemistry. For now, I understand your pointing to Lua and its templates is the route, I see what is missing in this (mainly template wrapping & documentation). So, not easily solvable from here, I'll continue my quest otherwise. DePiep (talk) 06:47, 18 August 2022 (UTC)[reply]
  • ┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Nearby: for a regular property, there is parser function #statements: eg mass (P2067):
{{#statements:P2067|from=Q674}} → 30.973761998±0.000000005 dalton. (as expected). -DePiep (talk) 09:08, 19 August 2022 (UTC)[reply]
@DePiep Here is how you can do it through the API: https://www.wikidata.org/w/api.php?action=wbgetentities&format=json&ids=Q674&props=sitelinks&sitefilter=dewiki Ainali (talk) 13:09, 19 August 2022 (UTC)[reply]
Thanks! will research (API/JSON new to me). DePiep (talk) 13:11, 19 August 2022 (UTC)[reply]

Notability gadget[edit]

I have thrown together a user script User:Bovlb/notability.js that attempts (in a conservative, mechanical way) to assess whether an item (apparently) meets our notability criteria. The interesting part of the script is the SPARQL query, which you can play with here. The query is a work in progress, so feedback is welcome. Known defects:

  • Because it relies on WDQS, the tool is subject to replication lag.
  • For structural need, no account is taken of links not represented in WDQS (but which appear in, say, Page Information).
  • I believe we should be ignoring references that just say "imported from Wikipedia", but I'm not sure what the criteria should be for a good reference.
  • Some sources are clearly better than others.
  • The sitelink test does not approach the complexity of the actual rule.

Bovlb (talk) 01:14, 18 August 2022 (UTC)[reply]

  • Interesting.
  • I think it's a valid approach since Wikidata notability is indeed predominantly a technical condition that can be mechanically assessed.
  • Functionally the script works for me as desired, it seems.
  • However, the assessment is a still way too simple in my opinion.
—MisterSynergy (talk) 13:15, 19 August 2022 (UTC)[reply]
Thanks. Which aspects should be made less simple?
For references, I was thinking about:
?x ?p ?statement .
?statement prov:wasDerivedFrom ?ref .
# P854 reference URL
# P248 stated in
# P3680 statement supported by
?ref pr:P854|pr:P248|pr:P3680 ?source .
For sitelinks, I quail at the idea of implementing the entire rule in SPARQL. Which part do you think is most important to include? Bovlb (talk) 15:45, 19 August 2022 (UTC)[reply]
Well, there are plenty of possible improvements—let's start with some rather simple suggestions.
—MisterSynergy (talk) 15:54, 19 August 2022 (UTC)[reply]
OK. I'll do the first when I have a chance. For the second, I agree those are weak identifiers, but in most cases they do answer the "Which John Smith?" question. Maybe the tool should provide three levels (red, yellow, green). Bovlb (talk) 16:22, 19 August 2022 (UTC)[reply]
—MisterSynergy (talk) 16:00, 19 August 2022 (UTC)[reply]
I agree that we have not formed a solid process around them, but I like the idea of having them. In particular, I think it's better to a have a community-maintained attribute like this rather than baking a list into a tool. Bovlb (talk) 16:24, 19 August 2022 (UTC)[reply]
  • I don't think that the presence of a official website (P856) claim should result in a green/notable result. It's not exactly the subject-independent proof of existence that we otherwise expect ;-)
—MisterSynergy (talk) 16:00, 19 August 2022 (UTC)[reply]
I agree it's weak, but again I think it's enough in most cases to answer the "Which John Smith?" question. Bovlb (talk) 16:25, 19 August 2022 (UTC)[reply]
Sure it might be helpful for identification, but it does not confer notability by itself. —MisterSynergy (talk) 17:50, 19 August 2022 (UTC)[reply]
@MisterSynergy Updated. Bovlb (talk) 19:22, 19 August 2022 (UTC)[reply]
While this tool may be useful to identify items that may not meet notability, I hope it doesn't become a driving tool saying "these items must be deleted". I do a lot of genealogical work on minor historical figures (children & spouses of politicians, locally prominent businesspeople etc.), many of which have no external identifiers or devoted websites, and whose biographical info is only found in digitized newspapers or books. Wikidata still hasn't figured out how to quickly and repeatedly generate complete bibliographic citations from a URL (English Wikipedia is lightyears ahead on this). I don't always have the time to create an item for a historic book or newspaper to source "Jane Smith, daughter of Judge Joe Smith, died on July 3, 1850" even if it's clearly visible to anyone who takes a second to Google it. I hope people working on assessing notability are working equally hard at improving the process of documenting notability. I've said this may times on Wikidata. Make it easier to plug in references, and you'll get better, more reliable data. -Animalparty (talk) 22:48, 19 August 2022 (UTC)[reply]
In response to MisterSynergy's suggestions, I have made the tool show red, amber (yellow) and green to reflect the range of views in our community. The intention is that if it shows green, pretty much everyone would agree that it's notable; if it's amber, then a significant portion of our community would feel it's notable, but many would disagree.
Do you have some examples of items that (you believe) the tool would look unfavourably on? Bovlb (talk) 00:23, 20 August 2022 (UTC)[reply]
I agree that we don't want a tool like this to be a driving force to delete items that the community doesn't want deleted, but I am hopeful that by laying out the notability criteria in query form, we can collectively come up with a better understanding of where the community stands. Obviously a simple query like this cannot handle nuance like whether a URL is a good reference or just a self-published puff piece, so it is perforce conservative. Bovlb (talk) 04:43, 20 August 2022 (UTC)[reply]
Not sure if your tool looks unfavorably on this, but one structural need in Wikidata is for people who act as familial links. So I would suggest that if person A is green, then any father/mother/spouse/child of A should be at least amber. Dogfennydd (talk) 07:04, 20 August 2022 (UTC)[reply]
As currently implemented, any incoming link (including familial) is enough to satisfy N3 and turn it green, but no account is taken of outgoing familial links. What change would you propose? Bovlb (talk) 14:15, 20 August 2022 (UTC)[reply]
Thanks for clarifying: no change proposed, since (I think) these properties are required to be reflected on both linked items. Dogfennydd (talk) 20:18, 20 August 2022 (UTC)[reply]
To make this discussion more accessible to those who prefer not to wade through my SPARQL, I have prepared a short English summary of the currently-implemented rules. User:Bovlb/notability.js/doc As before, this is intended to start discussion and gather feedback. Bovlb (talk) 14:49, 20 August 2022 (UTC)[reply]
Do you have an idea whether your tool basically does the same thing as https://www.wikidata.org/wiki/User:Pasleim/notability? ChristianKl ❪✉❫ 21:35, 20 August 2022 (UTC)[reply]
It is different:
  • The new user script by User:Bovlb operates within the browser on item pages and gives the user an idea which of the criteria in WD:N are (probably) met. For such an assessment, the user needs to install the script to their Wikidata environment, and they need to open the item page in question to see the report. It fundamentally operates on individual item pages.
  • User:Pasleim's notability report is a daily bot-updated list of items that meet all of these criteria at the time of reporting:
    • newly created yesterday (relative to the day when the bot edits the report page),
    • creation edit is unpatrolled (i.e. the item was created by an IP user or a registered newcomer account),
    • main namespace,
    • no sitelinks,
    • fewer than 3 claims,
    • and no backlinks.
The source code of Pasleim's bot script is accessible at /data/project/pltools/reports/notabilityCheck.py on Toolforge. —MisterSynergy (talk) 22:26, 20 August 2022 (UTC)[reply]
Thanks. That’s a helpful explanation.
There’s no reason my SPARQL could not be used in a similar way, to generate reports on recent creations. The user script could also be extended to annotate lists of items (recent changes, user contributions). Bovlb (talk) 15:39, 21 August 2022 (UTC)[reply]

Wikidata weekly summary #533[edit]

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Discussions
    • Open request for adminship:
      • Estopedist1 (RfP scheduled to end after 21 August 2022 12:36 UTC)
  • Events
    • Upcoming:
      • Next Linked Data for Libraries LD4 Wikidata Affinity Group Wikidata Working Hour August 15, 2022: Fourth session of our summer/fall project working with diverse children's book metadata. We'll be covering manually creating publisher items: showing how to create items, add statements to items, add references, and use gadgets. You are, as always, welcome to bring your own data to work on. All sessions in our project series will be recorded for those who cannot attend. Links will be added to the event page when available. Event page
    • Ongoing:
  • Tool of the week
    • KnolCase is a case-based knowledge management tool for gathering information about subjects of interest and organizing these into case files.
  • Development
    • Took part in Wikimania and the Wikimania hackathon
    • Continuing work on making it possible to add sitelinks to redirects under some special conditions (phab:T278962)
    • Lexicographical data: continuing to address the feedback from testing
    • REST API:

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.

@Mohammed Sadat (WMDE) https://www.wikidata.org/wiki/Wikidata:Property_proposal/provides_data and https://www.wikidata.org/wiki/Wikidata:Property_proposal/type_of_external_page are on the list "New property proposals to review" but both proposals have been closed more than a year ago. What happened? ChristianKl ❪✉❫ 06:44, 16 August 2022 (UTC)[reply]
Hi ChristianKl. DeltaBot didn't update the list in the past week, and I overlooked it. This should be fixed in the next issue. -Mohammed Sadat (WMDE) (talk) 20:05, 16 August 2022 (UTC)[reply]
This week's list does not contain the proposals that didn't appear on this list, so a whole week of proposals has never been announced. That needs to be fixed somehow. UWashPrincipalCataloger (talk) 16:46, 22 August 2022 (UTC)[reply]

Subclass cycle for "botanical name"[edit]

∞ effectively published name (Q108592268)

I don't know enough about taxonomy to fix this, but I thought it worth pointing out so someone else who does can try. JesseW (talk) 20:37, 21 August 2022 (UTC)[reply]

https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Taxonomy might be the better place. ChristianKl ❪✉❫ 09:15, 22 August 2022 (UTC)[reply]
Copied there, thanks. JesseW (talk) 14:57, 22 August 2022 (UTC)[reply]

Wikidata weekly summary #534[edit]

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
    • List of created Wikidata items - a tool which combines Xtools pages created API with Wikidata API to get the list of items created by a user with their label.
  • Development
    • Lexicographical data: continuing to address feedback from testing. We are almost done now.
    • REST API: continued investigating the use of PATCH and JSON Patch and worked on a proof-of-concept
    • Birthday planning and prepping is in full swing

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.

Overall item count correction[edit]

A quick note in case anyone is wondering why the main page and Special:Statistics are now reporting ~99,000,000 items instead of the ~99,950,000 a couple of days ago - there was a bug in the script used to count this which meant that the stats had got out of sync with the database over the last few months (phab:T315693). This has now been fixed and the count shown is correct - no items have been removed!

(Note that the headline "data items" count on the main page is made up of all non-redirect items, lexemes, and entity schemas - there are about 98.3m items proper, 675k lexemes, and 350 entity schemas, not counting redirects.) Andrew Gray (talk) 18:22, 22 August 2022 (UTC)[reply]

Thx Bangbang.S 01:57, 23 August 2022 (UTC)[reply]

Possible duplicates[edit]

Help needed from French speakers – can you take a look at the frwiki article for Raymundus Jordanus (Q93266293) to see if it is the same person as Raymundus Jordanus (Q4202811)? I think they are, but I don't speak French and would like to confirm before merging the items. Thanks! –FlyingAce✈hello 06:42, 26 August 2022 (UTC)[reply]

Yes, same, merged. --Tagishsimon (talk) 07:56, 26 August 2022 (UTC)[reply]
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. --Matěj Suchánek (talk) 16:10, 28 August 2022 (UTC)[reply]

Separate entries for sociology and wiktionary[edit]

Hi all, I am new to wiktionary and want to better link it up with areas of interest. However, I have noticed that sociology on wikidata has a fairly well occupied data from around the wiki project, but the reflecting wiktionary has its own entry. Is that normal? I can see that the sociology page has the ability to add a wiktionary entry, so I am not sure why the dictionary-focused aspect of wiki has its own entry. Can these be merged/ is there a work around to add it to the main page of sociology? Jamzze (talk) 13:09, 21 August 2022 (UTC)[reply]

The Wikitionary page you linked is for the English term in sociology. The previous Wikidata item is however about the general concept of sociology independent of the English language. Wikionary likely has similar pages for other languages. ChristianKl ❪✉❫ 18:31, 21 August 2022 (UTC)[reply]
I don't understand why that is a problem? On the Sociology page I linked to, it has Wiki projects linked to it across various languages. Such as Wikiquote in English, Spanish, German, etc. Should I not then be allowed to link the English version of the sociology Wiktionary to it, as well as others from other languages if they want to? Jamzze (talk) 19:14, 22 August 2022 (UTC)[reply]
The Wiktionary link I put in leads to a general category of English Sociological dictionary terms.
The Wikidata link leads to the overall general Wikidata entry for Sociology.
I want to be able to link the English version of Wiktionary's Sociology category on Wikidata to the general Sociology Wikidata entry so they can be found more easily. Jamzze (talk) 19:21, 22 August 2022 (UTC)[reply]
The English version of Wiktionary's Sociology category is properly linked to Category:Sociology (Q6642550) which has a category's main topic (P301) of sociology (Q21201). This is the conventional modelling of categories on WD. --Tagishsimon (talk) 19:32, 22 August 2022 (UTC)[reply]
@Jamzze There are items like Category:fr:Sociology (Q32858256) and Category:de:Sociology (Q32858249) and each of those has a link to the English version of Wikitionary as well. The current setup allows sidelinks from the French Wikidictionary version of the English language to the English Wikidictionary version of the English language. ChristianKl ❪✉❫ 13:02, 23 August 2022 (UTC)[reply]

Use of P6954[edit]

Is there a reason why both registration required and login are allowed values? Trade (talk) 01:24, 23 August 2022 (UTC)[reply]

You should probably ask @Eihel: who added login (Q16572632) to online access status (P6954) back in Feb of this year. JesseW (talk) 03:11, 23 August 2022 (UTC)[reply]
Hello Trade Hello and thank you, JesseW for letting me know. Subtlety is difficult to recognize, I grant you. The first value, registration required (Q107459441), means that access is possible by registering an account with the platform, then connecting with this same account. The second value, login (Q16572632), means that you cannot personally register with the service and access is only provided through another organization or certain condition, so a login is required in addition to having to log in.
As these values give a different restriction to a smaller population, I added this extra value. That said, registration required (Q107459441) encompasses both concepts. Regards. ―Eihel (talk) 20:38, 23 August 2022 (UTC)[reply]
That distinction makes sense, but it's not (yet) expressed in the descriptions (or statements) attached to those items (login (Q16572632) and registration required (Q107459441)). @Eihel -- could you clarify this in the descriptions? JesseW (talk) 20:51, 23 August 2022 (UTC)[reply]

Bulk upload of words and it's meanings[edit]

Hello, I have an igbo language newly created words. I want to know how I can add these words and it's meaning in wikidata in bulk. Let's say 500 words and it's meanings. How can I add or upload these bulk words in wikidata. How can I automate the process.

Your contributions will be highly appreciated

Adding bulk lexemes and senses into wikidata[edit]

Hi, I have these language words and meaning which I want to add to wikidata... Please how can I automate the process of adding bulk words(lexemes) and it's senses into wikidata.

TemTechie (talk) 14:14, 24 August 2022 (UTC)[reply]

Thanks

@TemTechie: You asked this question almost a month ago and @Multichill, ArthurPSmith: gave you good answers then. Mahir256 (talk) 14:21, 24 August 2022 (UTC)[reply]

The 2022 Board of Trustees election Community Voting period is now open[edit]

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

Hi everyone,

The Community Voting period for the 2022 Board of Trustees election is now open. Here are some helpful links to get you the information you need to vote:

If you are ready to vote, you may go to SecurePoll voting page to vote now. You may vote from August 23 at 00:00 UTC to September 6 at 23:59 UTC. To see about your voter eligibility, please visit the voter eligibility page.

Best,

Movement Strategy and Governance

This message was sent on behalf of the Board Selection Task Force and the Elections Committee

MNadzikiewicz (WMF) (talk) 13:05, 24 August 2022 (UTC)[reply]

10 037 954 direct instances of human - thousands of instances of subclasses of human[edit]

  1. https://w.wiki/5avc - 10 037 954 statements declaring direct instance of human
  2. https://w.wiki/5avk - 34 265 statements declaring instance of subclass of human
  3. https://w.wiki/5bLp - 295 instance statements on items which have a) a statement for direct instance of human and b) for instance of a subclass of human // query added 2022-08-20 89.14.88.55 14:23, 20 August 2022 (UTC)[reply]

77.13.6.6 14:16, 18 August 2022 (UTC)[reply]

Actually, that second query finds all classes on items known to be an instance of human, not just the subclasses of human, but it is identifying an actual problem. The problems seem to fall mainly into a couple of classes:
Bovlb (talk) 16:01, 18 August 2022 (UTC)[reply]
re https://w.wiki/5avk - "?item wdt:P31/wdt:P279+ wd:Q5 ." - it finds statements that declare that an item is an instance of a subclass of human. Items ("humans") can be listed more than once, e.g. one-year-old male (Q108887238) is listed four times. - 1702 left. 89.14.88.55 14:08, 20 August 2022 (UTC)[reply]
4594 results of the second query were about instances of video game developer (Q210167) (type of company) which was a sublass of software developer (Q183888) (human profession). I removed the Q183888 statement so they will no longer show up in the query results.
Here is a variation of the second query, which gives a better overview of what types of items are used and how common they are:
SELECT ?type ?typeLabel (COUNT(?item) AS ?count)
WHERE {
    ?item wdt:P31/wdt:P279+ wd:Q5 .
    ?item wdt:P31 ?type.
    SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }.
}
GROUP BY ?type ?typeLabel
ORDER BY DESC(?count)
Try it!
--Shinnin (talk) 16:45, 18 August 2022 (UTC)[reply]
All too easy for people to make mistakes when the names are ambiguous like *developer. A tool that spots the biggest number of affected items is jolly useful, so the low-hanging fruit can be collected. BTW is there a tool that shows all the entries with any errors (of the kind that flag (!) in the editor) for a given property? Vicarage (talk) 16:55, 18 August 2022 (UTC)[reply]
Here's the diff on worker (Q327055) which deals with the 'ambassador to a country' type issue (it removes Q5 as a subclass value, leaving 'person'). Who knows? 2341 instances left, if so. --Tagishsimon (talk) 17:17, 18 August 2022 (UTC)[reply]
@Vicarage: We have constraint violation reports. For example Wikidata:Database_reports/Constraint_violations/P106 shows all constraint violations for occupation (P106). You can find the link to a constraint report page from the property talk page. --Shinnin (talk) 18:06, 18 August 2022 (UTC)[reply]
I was meaning that if you were interested in Danish authors, which ones had problems with their range of properties, rather than focussing on a single property. Vicarage (talk) 19:35, 18 August 2022 (UTC)[reply]

Re third query - https://w.wiki/5bLp - can the P31 that does not link to Q5 be moved by a tool to subject has role (Property:P2868)? prisoner, anonymous, anonymous master, refugee, missing person, child, missionary kid etc. each is a role. Editors can then decide to move it to some more specific property, e.g. occupation. The query by User:Shinnin, "16:45, 18 August 2022 (UTC)" shows 136 results for human, these should be fixed if the other statement is move to subject has role. 89.14.88.55 14:38, 20 August 2022 (UTC)[reply]

Child is not something that's supposed to be listed separately as it's easily deduceable from the age. human fetus (Q26513) is more complex because it's debateable whether or not we want to consider a fetus that was never born to be human (Q5). I think it's valid to use instance of (P31) in those cases.
When it comes to prisoners, place of detention (P2632) with somevalue would be a way to express their status.
For missing persons date of disappearance (P746) with somevalue would be appropriate. ChristianKl ❪✉❫ 09:55, 21 August 2022 (UTC)[reply]
This does not answer the question. 89.12.28.223 13:10, 25 August 2022 (UTC)[reply]

Again Re third query - https://w.wiki/5bLp - can the P31 that does not link to Q5 be moved by a tool to subject has role (Property:P2868)? 77.13.39.226 12:33, 22 August 2022 (UTC)[reply]

Thoughts about a new «protection status of a natural space» property[edit]

Hi:

I'm reflecting about the need of a new property for protected natural areas, analogue to intangible cultural heritage status (P3259) and heritage designation (P1435) properties used in cultural heritage.

These are just first thoughts and I would like feedback about usefulness and enhancements. I chose to ask here instead of opening an RFC because the idea needs to be more developed first. Feel free to comment in the discussion page. Thanks. —Ismael Olea (talk) 22:00, 23 August 2022 (UTC)[reply]

@ Tagishsimon : Answering to your removed answer:
Forgot to add to the document: yes the combination of located in protected area and located in/on physical feature is key part of the natural protection schema. And yes, the Caenlochan (Q5016835) example is paradigmatic too. My thinking is instancing Caenlochan SSSI as an protected area instead the proper designation type value. This value would be used with the «natural protection status» property. This would solve any ambiguity like the UNESCO cases and would align with the Nationally designated areas inventory model. I know CDDA database is not worldwide but the data model is flexible enough to add any country designation type without the potential nightmare of subclassing. And would help to make SPARQL queries simpler. —Ismael Olea (talk) 10:36, 24 August 2022 (UTC)[reply]
New properties are not created via RfC's but via the property proposal process. ChristianKl ❪✉❫ 07:45, 25 August 2022 (UTC)[reply]
I know. This is just a first analysis to conclude if a new property and its ontology model is valid or not, before writing a formal proposal. —Ismael Olea (talk) 09:48, 25 August 2022 (UTC)[reply]

Elections[edit]

I tried to find a suitable item to use with P31 in Q10430116. I finally used public election (Q40231), but realised that probably is intended for public elections for a public office. In this case, they are electing a bishop in Suomi/Finland. I do not know who are eligable to vote, the article doesn't tell. But I guess it is a limited group of people, since the winner got 308 votes. Donald Duck and Micky Mouse normally get more votes in public elections here, where I live. Testkonto 42 (talk) 07:03, 25 August 2022 (UTC)[reply]

The description of Q40231 said something about it being for public election so I renamed it appropriately. ChristianKl ❪✉❫ 08:06, 25 August 2022 (UTC)[reply]

edit protected - LGBT (Q17884) - instance of group of humans as subclass of human[edit]

LGBT (Q17884) - item is edit protected. 89.12.28.223 13:24, 25 August 2022 (UTC)[reply]

Wikidata as repository of study results?[edit]

Hello! Wikidata has an extensive coverage of scientific publications, and I am wondering if it may be an interesting idea to also develop a framework for adding the study results in a structured manner? Say I perform a meta-analysis, charting the results (e.g., regression coefficients or risk estimates or concentrations) from ten studies. To make my data available for others, I add the results in a structured manner to Wikidata, including meta-data such as study group charactertics and methods and time point (if repeated design). If someone later wants to repeat my meta-analysis, or just want to compare results related to a specific subject, they can now query Wikidata and get the data directly. Would this be an interesting idea to explore in Wikidata, or would it be better to make a separate Wikibase? Ajarmund (talk) 18:55, 3 August 2022 (UTC)[reply]

I don't believe Wikidata is an appropriate venue for publishing original research. I encourage you to submit the results of your meta-analysis for publication in an appropriate scientific journal. When it gets published, an item about the article can be added to Wikidata, and any statements describing the results of the research can be added to the appropriate items as well, with reference to the article item. Possibly even the details of the study as well, although I'm not sure whether that would be notable. Silver hr (talk) 19:25, 3 August 2022 (UTC)[reply]
Thank you for your quick reply; just to clarify: the idea would be to add the results of existing studies already published (and already being Wikidata items). For example, to add the actually measured cytokine levels that are reported in Male fetal sex is associated with low maternal plasma anti-inflammatory cytokine profile in the first trimester of healthy pregnancies (Q99590222) as statements with qualifiers. Ajarmund (talk) 19:40, 3 August 2022 (UTC)[reply]
I just happened to be online :). I suppose in the case of research studies that produced articles describing them, the studies themselves would be notable under criterion 2 of Wikidata:Notability. And it seems we already have such items, e.g. 18F-PM-PBB3 PET Study in Tauopathy Including Alzheimer's Disease, Other Dementias and Normal Controls (Q61902219). So I see no reason that items that are instance of (P31)clinical trial (Q30612) or a related class couldn't have statements describing them in detail.
There is the problem of reliability, though. Wikidata is a wiki that anyone can edit, which implies that any statement it contains at any point in time can be incorrect. So if your goal is to have a repository of reliable study data that scientists can use, I don't think that goal can achieved through Wikidata. Silver hr (talk) 20:31, 3 August 2022 (UTC)[reply]
Generally, listing study results in the items of the relevant scientific papers is a good idea if they can be listed in a structured way where multiple people are likely noting them in the same structure.
It's worth thinking about whether study results should belong to the item of the clinical trial or the item of the academic paper. ChristianKl ❪✉❫ 07:26, 4 August 2022 (UTC)[reply]
Seems to me a very poor idea indeed, for at least a number of reasons: WD is not designed for tabular data, and I'd presume 'study results' would tend to be exactly that. 'Study results' shorn of their context & qualifiers, seems plain wrong - the presentation of a partial picture. The diversity of forms of 'study results' would also seem to be an enormous ontological challenge; really, in my experience, no two papers have the same sort of content and I cannot imagine how data would be collected in WD in a way which made any real sense. --Tagishsimon (talk) 09:12, 4 August 2022 (UTC)[reply]
Thank you for pointing out important issues. I agree with you that study results need context, such as in the form of qualifiers. I would say that wikidata offers a unique platform to do exactly this as concepts can be referenced quite precisely across entries. I am not exactly sure what you mean by tabular as I associate it with data presentation rather than the data itself (e.g., long vs wide table formats); columns in a table are generally translatable to qualifiers? Another thought is that study results are already prevalent in wikidata - they are just associated with specific items and having the study source as reference. For example, the reference range of serum potassium level (Q658883) is kind of a study result which may have benefitted from more context/metadata (method, info about study population; human, adults, number of participants, sex ratio, ...). The question, then, is how this information is best organized; as part of the item of the source or as separate and more specific items? How can context best be preserved? How can the results be queried in a useful way? How can they be organized so that contributions can be made efficiently? -- Ajarmund (talk) 00:31, 8 August 2022 (UTC)[reply]
@Ajarmund How about you create an example item of how you expect this to be modeled? ChristianKl ❪✉❫ 09:03, 10 August 2022 (UTC)[reply]
Good idea! Here is an example https://test.wikidata.org/wiki/Q225956 -- Ajarmund (talk) 10:17, 10 August 2022 (UTC)[reply]
I can't say I'm in favor of conflating research studies and articles that describe them in the same item. We should generally err in the direction of more granularity. Also, I'm not enthused about having 3 maternal age properties; having the study populations as instances would enable the use of mean, Q1, and Q3 as qualifiers for maternal age (BTW, why maternal age and not just age?). Finally, the properties study population, reference group and outcome should have the item type. Other than that, I suppose I don't see a reason why this couldn't be in Wikidata, especially if someone finds it useful. Silver hr (talk) 16:50, 10 August 2022 (UTC)[reply]
Thank you for valuable feedback, @Silver hr! I agree with you and guess there are at least three strategies then that may produce "clearly identifiable conceptual or material" entities, with various degrees of granularity;
Example: three outcomes (say crp, BMI, cholesterol) measured in two study groups at two time points
  • Option 1: 16 items
    • One item for each study group (i.e., two items)
    • One item for each condition (i.e., time points -> two items)
    • One item for each measurement (i.e., 3x2x2=12 items)
  • Option 2: 5 items
    • One item for each study group (i.e., two items)
    • One item for each condition (i.e., time points -> two items)
    • One item collecting the study outcomes (i.e., one item)
  • Option 3: 7 items
    • One item for each study group (i.e., two items)
    • One item for each condition (i.e., time points -> two items)
    • One item for each "type" of study outcomes (i.e., crp, BMI, cholesterol -> three items)
Option 1 results in the most granularity, whereas option 3 perhaps is closest to how we conceptually think about results (i.e., the measurement of something where the focus is on something). In general, it will produce items with duplicated names, but as far as I know, that's not a problem in itself
-- Ajarmund (talk) 12:34, 12 August 2022 (UTC)[reply]
@Ajarmund
When adding new data to Wikidata, we try to minimize the number of new properties that we create. In your test Wiki item you didn't try to build on existing properties. Instead of creating a new material age property we would likely reuse mean age (P4442). That does need an agreement to increase the scope, but I don't see a reason why we would create here a new property.
Wikidata is also inherently about linked data. If you have a string "Pregnant women carrying a male fetus" that string is only readable in English. If you instead create https://test.wikidata.org/wiki/Q226003 a German user will see the German translation. In Wikidata the Q226003 would also be further described via structural data. Different studies that investigate the same thing would link to the same Q226003 item.
The question about whether study results should be modeled in an item about the clinical trial or in the item about the paper about the trial is also open. ChristianKl ❪✉❫ 09:33, 20 August 2022 (UTC)[reply]
@ChristianKl Thanks for the feedback - you're right, the example was not very thoroughly made. I will happily contribute with data but find it difficult to decide what structure is best. But since the idea didn't seem to cause much enthusiasm in the community, it might be wrong time or place
Another option is be to use items such as serum C reactive protein level (Q10438898), add values through statements with observed in (P6531) (or something better) and have the paper as reference.
-- Ajarmund (talk) 21:55, 20 August 2022 (UTC)[reply]
I wouldn't say enthusiasm in the community is necessary to do this, as long as there's at least one person willing to do it. If you are that person, go for it. Silver hr (talk) 15:29, 25 August 2022 (UTC)[reply]

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ As long as fields can be edited by academic rivals, ideologically-motivated bad actors, or random bozos who just want to cause trouble, keeping study results on WD is far more dangerous than just having entries on the articles themselves. DS (talk) 13:29, 12 August 2022 (UTC)[reply]

An attempt to summarize the discussion;

  • Published study results are, in general, notable
    • They can (by definition) "be described using serious and publicly available references" (notable)
    • Having study results available within the strucure that Wikidata offers, has many potential use cases
  • There are ontological challenges related to how study results would be structured (scattered over multiple items) to ensure that
    • the relevant and necessary context is conserved,
    • there is consistency across items, and
    • they are accessible by querying
  • Wikidata can be edited by everyone, including malign actors, so that false data may be introduced (either by accident or intent)

--Ajarmund (talk) 10:23, 17 August 2022 (UTC)[reply]

Murder case[edit]

Can someone help with Hall–Mills murder (Q4157206), I can't figure out how to model the people to get rid of the error flags. The ones I can find are all set up for the trial with defendant and prosecution, and not victim and perpetrator. RAN (talk) 15:57, 22 August 2022 (UTC)[reply]

I would add the statement Hall–Mills murder (Q4157206)perpetrator (P8031)...object has role (P3831)suspect (Q224952) with disapproved rank for the three suspects. --Gymnicus (talk) 19:08, 22 August 2022 (UTC)[reply]
I found similar problems modelling Almería Case (Q8342043). Would be useful to know any best practice available. —Ismael Olea (talk) 22:03, 23 August 2022 (UTC)[reply]
Not sure about your particular question, but I find it wrong that two different events (a murder and a legal case) are conflated into a single item. The item looks like it primarily describes the legal case related to the murder, so a proper classification IMO would be instance of (P31)legal case (Q2334719)relative to (P2210)murder (Q132821), unless there's a better property to describe what a legal case is about. Silver hr (talk) 15:44, 25 August 2022 (UTC)[reply]
  • There must be a famous case that can used as a model, that has no error flags. I am not against splitting all murders into a murder case and a trial. --RAN (talk) 20:34, 25 August 2022 (UTC)[reply]

artificial intelligence[edit]

artificial intelligence (Q11660) is currently being used for both artificial intelligence, the branch of computer science, and artificial intelligence, the category of computer software. Is there some way that these can be split into two items without causing confusion? Nosferattus (talk) 19:36, 22 August 2022 (UTC)[reply]

The answer is yes for both this specific question and in general. If you find an item that conflates two or more concepts, the proper thing to do is try to find whether one of the conflated concepts already has a separate item, and move the relevant statements there. If you can't find such an item, create it. Silver hr (talk) 15:49, 25 August 2022 (UTC)[reply]

instance of role as subclass of human[edit]

Victim was subclass of human - https://www.wikidata.org/w/index.php?title=Q10436169&diff=1664310702&oldid=1664310455 . But victim is a role, an instance of a human can assume that role, but still is a human. What can be technically done to stop instances of roles from being used as subclass of human? 89.12.119.74 08:38, 23 August 2022 (UTC)[reply]

I might be wrong, but I don't see a way to do this with the existing constraints. What you're asking for is a general disjointness constraint, and I agree it would be useful to have one. In the meantime, you could put a Template:Complex_constraint on the talk page for human or role and specify a query that finds items that are both instance of role and subclass of human. This constraint wouldn't stop people from doing this though, it would merely cause a bot to produce a list of violating items. Silver hr (talk) 18:50, 25 August 2022 (UTC)[reply]

A question about a potential merge or sitelink shuffle-around[edit]

Hello! New Wikidata user here, although I've been on enwiki for a while. I came across the Wikipedia article en:Kabouter Plop, and wondered how, being a Belgian TV show, it doesn't have a Dutch or French version – but it does (nl:Kabouter Plop), just not linked from the English version. I discovered that these are linked to different Wikidata items, and thus don't share the same pool of interlanguage links: English, Esperanto and Italian articles are linked from Kabouter Plop (Q2832088) (the main character of the TV show), while the Dutch and French articles are linked from Kabouter Plop (Q71828087) (the show itself). I read through Help:Merge and Help:Sitelinks, and out of an abundance of caution decided to not attempt merging these items without asking first.

I'm not sure how to resolve this, since they are technically of different items. In practical terms we want the links in the same place – unlike the case of Seinfeld (Q23733) and Jerry Seinfeld (Q1797473), there isn't enough written about the character to warrant Wikipedia articles separate from the show, but the three articles linked from Q2832088 do start with a sentence along the lines of "Kabouter Plop is the main character of a TV series of the same name", with the rest of the article after the first sentence being about the TV show, so there is an argument (in my opinion an unhelpful one) to be made that the current linking is correct and shouldn't be changed. (The articles linked from Q71828087 are immediately about the show.)

So, do I/we

  • Merge Q2832088 (the character) into Q71828087 (the show) or vice versa, or
  • Keep Q2832088 (the character), but move its Wikipedia links into Q71828087 (the show) or vice versa, or
  • Do nothing in Wikidata, but override it, by adding local interlanguage links into the articles over there? (Obviously the not-preferred option.)

Thanks in advance, Oatco (talk) 21:27, 23 August 2022 (UTC)[reply]

Just found out about Wikidata:Do not merge, and sure enough, these two are in there (Wikidata:Do_not_merge/05#9001_to_10000, number 9178). Oatco (talk) 21:37, 23 August 2022 (UTC)[reply]
I think there's a method that involves creating a redirection on the Wiki, like nl:Kabouter Plop (personage) -> nl:Kabouter Plop, and linking the redirection to the other Wikidata item. Ghouston (talk) 23:49, 23 August 2022 (UTC)[reply]
I would argue that the English article should be rewritten to be about the show itself (it already is about the show; as you mention, only the first line talks about the main character, and the rest of the article talks about the show and the other characters). But that's a matter for enwiki rather than Wikidata :) –FlyingAce✈hello 00:52, 24 August 2022 (UTC)[reply]
Indeed, the items on Wikidata are about single concepts, and a TV show and a character in a TV show are different concepts, and so should have different items. This often causes problems such as the one you describe, which is unfortunate, but only solvable on the wikipedia side. User:Ghouston's suggestion might work. Silver hr (talk) 19:02, 25 August 2022 (UTC)[reply]

regarding the block on my IP[edit]

i wanht to know why do you bloked my IP when i´m trying to open a new account to a real new singer from PuertoRico who moved to Spain and also that he have a lot of participation on my country and out in America. he deserve to be in wikipedia whe are part of a fan page at Spain and we want to open this space and put all his effort and achivments here. Please help us to do it right but do not block my IP. let me kn ow clearly what i need to improve to maked better. thank you for any or all your help. best regards Karlo Rivera FAN PAGE ES (talk) 19:17, 29 August 2022 (UTC)[reply]

@Karlo Rivera FAN PAGE ES: This appears to be related to your block on the Spanish Wikipedia, which is a separate project. If you wish to be unblocked there, you should put {{Desbloquear|1=INSERT REASON HERE}} on your user talk page there. You may wish to read es:Wikipedia:Artículos sin relevancia aparente first. I'm afraid we can't help you here. Bovlb (talk) 17:30, 30 August 2022 (UTC)[reply]
Unblock requested and denied on ESWP. Bovlb (talk) 15:25, 31 August 2022 (UTC)[reply]
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Bovlb (talk) 15:25, 31 August 2022 (UTC)[reply]

trains as instance of passenger[edit]

e.g. Mathura - Bhiwani Passenger (Q39047829) 89.12.28.223 13:33, 25 August 2022 (UTC)[reply]

including a "source" added via a tool reCh, 2019-01-15 https://www.wikidata.org/w/index.php?title=Q39047829&type=revision&diff=834863782&oldid=554227565 89.12.28.223 13:36, 25 August 2022 (UTC)[reply]

I replaced to instance of (P31) = train (Q870). It is OK for you? Michgrig (talk) 16:31, 25 August 2022 (UTC)[reply]
At least for this item no more "instance of passenger". But could this in general be forbidden by the software? There are 21 others https://w.wiki/5cm$ . And for Q39047829 enwiki text contains "Both trains are hauled by a Shakur Basti Loco Shed based WDM-2 diesel locomotive from Bhiwani to Mathura and vice versa." - so it isn't one train. 89.12.28.223 17:21, 25 August 2022 (UTC)[reply]
fixing the other 21 is a good idea. not sure this is common enough to justify a software block. the first few I checked were all added by @Nvrandow: so maybe we can ask them to fix this. BrokenSegue (talk) 18:13, 25 August 2022 (UTC)[reply]
I've fixed them. Nvrandow (talk) 07:12, 26 August 2022 (UTC)[reply]
Thank you. 77.183.4.252 11:29, 26 August 2022 (UTC)[reply]

57x instance of human and instance of anonymous[edit]

https://w.wiki/5cy3 57x instance of human and instance of anonymous Q4233718, but anonymous is a role ascribed by some. Can someone with a tool please move the claims to subject has role (property:P2868)? 77.183.4.252 11:47, 26 August 2022 (UTC)[reply]

Global ban discussion for Velimir Ivanovic/Kolega2357[edit]

Hello all,

In line with the global ban policy, local project communities where the user was active should be informed of the discussion in a prominent public place.

This ban request has been live for about a month now, but proper notification was not made. The ban request will continue for at least a few more weeks, and input is welcome on the RfC page linked in this section heading.

Thank you for your time, Vermont (talk) 02:37, 27 August 2022 (UTC)[reply]

Artificial landform is stated to be a subclass of natural geographic object[edit]

human-made landform (Q19816755) is a subclass of landform (Q271669); following the subclass links leads to geomorphological unit (Q12766313) and natural geographic object (Q35145263). Q271669 says "natural" in the English description, but the English Wikipedia article says it can be natural or artificial. There seems to be no "natural landform" item, so Q271669 should probably be split. Peter James (talk) 16:13, 27 August 2022 (UTC)[reply]

I'd say that geomorphological unit (Q12766313) is the thing that need not be natural. Vicarage (talk) 17:52, 27 August 2022 (UTC)[reply]

connecting wikidata to my rest api to enable automatic update[edit]

hello, please i want to find out if there is a means i can automatic the process of adding and updating lexemes into wikidata directly from my rest api, thats fetching from my database and posting in form of adding to wikidata lexemes.in Such a way that lexemes i posted on wikidata will update itself automatically when my database is been updated. connecting wikidata to my rest api TemTechie (talk) 21:54, 27 August 2022 (UTC)[reply]

Main image for Winchester (city in England)[edit]

Hi, may I suggest that the image for Winchester (https://www.wikidata.org/wiki/Q172157) is changed to one showing a view of the city, not just the cathedral, such as the one at https://commons.wikimedia.org/wiki/File:View_of_Winchester,_Hampshire,_England.jpg. I cannot edit this myself. ITookSomePhotos (talk) 17:42, 26 August 2022 (UTC)[reply]

Done, thanks for the suggestion! JesseW (talk) 18:23, 26 August 2022 (UTC)[reply]
@ITookSomePhotos, JesseW: I guess that as a general rule of thumb, we should follow enwiki in such image questions, hence File:Winchester Montage.jpg may be best Estopedist1 (talk) 06:02, 27 August 2022 (UTC)[reply]
No. There's no very good reason for following EN wiki in this respect, especially if it winds up with WD using a montage which would have been rejected by even a third-rate picture postcard company. --Tagishsimon (talk) 12:28, 27 August 2022 (UTC)[reply]
@Estopedist1 There's a separate property for montage images - montage image (P2716). I think the idea is to keep image (P18) for very straightforward images of something, and then if people specifically want a montage, they can pill that out as needed. Andrew Gray (talk) 20:50, 28 August 2022 (UTC)[reply]

merging Q874059 (Roman Catholic Diocese of Pelplin) and Q3738192 (Bishopric of Culm)[edit]

According to the papal document from 1992: "Dioecesis Pelplinensis (vetere nomine Culmensis appellata)", it means a change of name, not creation of a new entity. In the case of the creation of new entities, the papal bull clearly states this using the term: 'noviter erecta' or 'noviter condita'. Bogumił Szady (talk) 09:12, 28 August 2022 (UTC)[reply]

merging Roman Catholic Archdiocese of Halyč (Q9159034) and Roman Catholic Archdiocese of Lviv (Q1364829)[edit]

According to documents and state of research in 1412 the capital of the diocese was transferred from Halych to Lvov and not a new diocese was established Bogumił Szady (talk) 15:31, 28 August 2022 (UTC)[reply]

There might be a problem because at least the Polish Wikipedia has two separate articles for w:pl:Archidiecezja lwowska i w:pl:Archidiecezja halicka. Kpjas (talk) 19:23, 28 August 2022 (UTC)[reply]

9 seconds sleep between statement uploads[edit]

Hi all, Not sure if this is the right place to ask: We received permission for a bot that uploads the data from the Israeli Film Archive. However during the upload process we get a 9 seconds sleep between the processing of each statement. Is this normal? Running the bot the whole weekend resulted in just 200 new items, which makes the whole thing unfeasibly slow. Is there a way to speed things up? Are we doing something wrong? Keren - WMIL (talk) 09:34, 25 August 2022 (UTC)[reply]

Can't speak for the latency here, but if it were possible to upload the entire item in a single edit, the edit rate would be ~14 times faster. Don't have any info on the nature of your bot's tech; a quick sketch might be helpful. --Tagishsimon (talk) 09:53, 25 August 2022 (UTC)[reply]
As much as I am aware, your bot account does not have stricter ratelimits than any other bot account — thus it should be able to make 90 edits/minute at maximum which is the limit for most accounts.
Thus, my first guess would be that you have not configured your bot framework properly. There are some frameworks such as pywikibot for instance that have a default edit throttle at 10 sec/edit — if you want to go quicker, you need to change this config. It would thus be helpful if you could either verify your setup, or provide some more info about it here. —MisterSynergy (talk) 09:55, 25 August 2022 (UTC)[reply]
If you are using Pywikibot, the relevant parameter is -put_throttle, alias -pt. See https://www.mediawiki.org/wiki/Manual:Pywikibot/Global_Options. William Avery (talk) 10:03, 25 August 2022 (UTC)[reply]
maybe it is better to use OpenRefine ? Geagea (talk) 13:53, 29 August 2022 (UTC)[reply]

Cannot create interwiki link between :en:Adoptium and :de:Adoptium[edit]

Hello, I tried to create interwiki links, both from the English and the German side via the "add links" link under "Languages", and got an obscure error message pointing to Wikidata: "Could not link pages: failed to merge corresponding Items on Wikidata ...". From that message it is not clear what I can do to resolve the issue. I also left a comment on the English article's talk page. Adding interwiki links used to be straightforward for about 20 years ... Any hints to resolve this would be appreciated :-). --Zeno Gantner (talk) 08:14, 29 August 2022 (UTC)[reply]

@Zeno Gantner I moved the sitelink but the work isn’t done. AdoptOpenJDK (Q62633542) is a complete mess and should probably be deleted as conflated. I’m not an expert but maybe use Adoptium (Q106879832) for the organization and Eclipse Adoptium (Q99482986) for one of the software products? --Emu (talk) 08:47, 29 August 2022 (UTC)[reply]
Maybe somebody from WikiProject Informatics has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. can help? --Emu (talk) 08:49, 29 August 2022 (UTC)[reply]

Thank you very much for the quick response and for resolving this! For me to learn, what could I have done myself to resolve the issue? I assume normal Wikipedia editors without much Wikidata exposure might run into similar issues easily. Or was this a rare and exceptional situation? --Zeno Gantner (talk) 08:54, 29 August 2022 (UTC)[reply]

@Zeno Gantner Schwer zu sagen, ich hatte diese Fehlermeldung noch nicht. Man kann Sitelinks am besten verschieben, indem man das Gadget Move verwendet, das man unter Special:Preferences#mw-prefsection-gadgets aktivieren kann. Die Zuordnung der Sitelinks zu den richtigen Items ist eine ständige Herausforderung, aber technisch sollte das eigentlich kein Problem sein. Wenn es noch einmal vorkommt, würde ich es noch einmal ansprechen.
P.S.: Ich kann mich noch erinnern, wie das mit den Interwikilinks vor 19 Jahren war und es war ganz entsetzlich! ;-) --Emu (talk) 17:42, 29 August 2022 (UTC)[reply]

Wikidata weekly summary #535[edit]

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
  • Development
    • REST API: working on the ability to change existing statements on an Item (phab:T306934)
    • Lexicographical data: Finishing up the remaining pieces of the new Special:NewLexeme page based on the testing feedback (subtickets of phab:T307866)

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.


Check out the View it! prototype & join us for a demo + discussion on Aug. 31st 16:00GMT[edit]

A project has been funded by the Wikimedia Foundation as part of the Structured Data Across Wikimedia Work to create a tool called View it! The tool aims to increase the discoverability of images on Commons, give readers and editors access to more images, and encourage contributors to utilize Commons & structured data. View it! also works on Wikidata (there is an example that will be uploaded on the Meta page soon) and will show users any images related to a given item.

Please visit the Meta page if you are interested in trying out the prototype. We are having a demo and feedback session on August 31st at 16:00 UTC, please join us if you wish!

Please feel free to reach out if you have any questions! Thanks! JamieF (talk) 16:59, 29 August 2022 (UTC)[reply]

Linking wikidata API to my API as form of adding lexemes[edit]

Greetings!!... I have a Node Express built API with mondoDB Mongoose as it's database where my lexemes and senses are stored ... I wanna ask is there any way I can automate the process of adding or uploading these lexemes/senses that is stored in my API database directly to wikidata in form adding the lexemes/senses??? Maybe through any wikidata API or something like that??? Any idea ?? Do we have any kind of wikidata API which I can use to achieve this... any ideas on how to go about it ... Or is there a script is sort of idea existing already??? Your response will be highly appreciated TemTechie (talk) 17:10, 29 August 2022 (UTC)[reply]

@TemTechie Yes, Wikidata has an API, see mw:Wikibase/API. Special:ApiSandbox is probably a good place to start – all the wb* actions are related to Wikibase (the software behind Wikidata), and the wbl* ones are specific to lexicographical data. Lucas Werkmeister (talk) 20:32, 29 August 2022 (UTC)[reply]
Thank You, I appreciate TemTechie (talk) 21:31, 29 August 2022 (UTC)[reply]

Layout of the “Welcome message” on the gray background[edit]

Hello, I'm copying-pasting my topic and the reply by the wub from Wikidata talk:Main Page as it was suggested by Estopedist1.

Please take a look at the first box with Welcome message ("Welcome to Wikidata ...") on the gray background: the versions in different languages (except English) don't fit the screens of mobile devices. Could someone fix it?

Screencaps (with the version in Russian, but it applies to the other versions, too).

1) https://share.icloud.com/photos/0b7ebBJr4FH-d7zD0sdZkUfZQ 2) https://share.icloud.com/photos/0c9CP7rlf4R1uzT_p22oBBEzQ 3) https://share.icloud.com/photos/0a1_WUdCh16cKksX7vfDHZnKA 4) https://share.icloud.com/photos/0f1s4p66G3y6_nRvl4n-ET1Yw 5) https://share.icloud.com/photos/092jgd2mQ-Q1ieCAVu-t4LWFg

— Pacha Tchernof (talk) 09:23, 1 August 2022 (UTC)[reply]

It is getting cut off in the English page too for me, see screenshot: Wikidata Main Page on mobile, with intro getting cut off. the wub "?!" 23:13, 17 August 2022 (UTC)[reply]

Pacha Tchernof (talk) 19:05, 29 August 2022 (UTC)[reply]

It's not too complex for someone with skills. We should get rid of relative+absolute positioning of and inside the top box, and use something more modern such as flexbox or grid. Also, the image can probably go to the background. Template:Main Page/Header new/styles.css has the current style definition. —MisterSynergy (talk) 20:20, 29 August 2022 (UTC)[reply]
Hold on, the "background image" cannot be used as a background image due to its CC-by-sa 4.0 license. It needs to link to File:Wikidata nodes in white.svg in order to be re-used properly (which currently does not work properly as well).
Maybe we should better design a completely new main page; something that does not rely on ancient positioning techniques. —MisterSynergy (talk) 20:31, 29 August 2022 (UTC)[reply]

Proposal marked as ready since April[edit]

Is it ever ok for a property creator to create a proposal that is marked as ready that they themselves proposed? I ask because a property that I proposed has been marked as ready since April of this year, and no one has created it. I would gladly create it if that is permissible. UWashPrincipalCataloger (talk) 00:31, 28 August 2022 (UTC)[reply]

Which one is it? Jean-Fred (talk) 09:41, 28 August 2022 (UTC)[reply]
@Jean-Frédéric Likely it's Wikidata:Property proposal/onscreen participant (currently the oldest ready proposal) Vojtěch Dostál (talk) 16:54, 28 August 2022 (UTC)[reply]
If so, then I think it should be de-marked as ready, as there is one Oppose comment that the proposer has not addressed, and it thus fails Criterion 3 “All opposing points of discussion should be addressed before creation occurs.”. (Granted, given that this Oppose is from me, I am obviously biased ^_^) Jean-Fred (talk) 17:01, 28 August 2022 (UTC)[reply]
I do think that Jean-Fred is right here. While the proposal criteria don't require that everyone has to be convinced, they do call for all open points of discussion being addressed. Now I do think you have addressed it. ChristianKl ❪✉❫ 15:34, 29 August 2022 (UTC)[reply]
Thanks, Kristian. UWashPrincipalCataloger (talk) 22:04, 29 August 2022 (UTC)[reply]

Conflation at VIAF[edit]

https://viaf.org/viaf/58476802/ and its corresponding ISNI record appears to conflate two people with the same name, both active in Cincinnati (Q43196): Joseph Henry Gest (Q18912379) (artist and museum director, 1859-1935) and Joseph Gest (Q113628883) (city surveyor, 1776-1863, biographed here). The records referred to at VIAF, ISNI, and LCCN concern a map from 1838, which is clearly not the work of the museum director. Since VIAF and ISNI conflate the two, and the LCCN record links to the Wikidata item Q18912379, should Q18912379 be turned into a conflation item and a new item be made for the museum director? Thanks. -Animalparty (talk) 23:12, 26 August 2022 (UTC)[reply]

@Animalparty Please don’t. Items for conflated VIAF entries have been tried without success. The best practice is to create as many items as necessary (one for each person) and do your best to assign the correct identifiers. In some cases, a deprecated rank can be used. In most cases the libraries are also happy for bug reports (although the LC doesn’t respond and ISNI’s process seems to be broken at this time) (CC Kolja21) --Emu (talk) 08:54, 27 August 2022 (UTC)[reply]
 Info I filed an error report for the ISNI (via homepage) and GND (also via homepage) --Emu (talk) 09:04, 27 August 2022 (UTC)[reply]
I have had success with Library of Congress in the past on a conflated individual. I created two WD items, populated them as much as possible with referenced statements and then reported the error to LC using an online form and included links to the two WD items. Library of Congress then created a separate entry for the second individual a few weeks later. From Hill To Shore (talk) 11:27, 27 August 2022 (UTC)[reply]
@From Hill To Shore That’s really impressive. I’ve tried everything (homepage, email, …) and it never went anywhere. But I will try it again sometime, maybe I’m lucky this time! --Emu (talk) 18:57, 27 August 2022 (UTC)[reply]
So in this case we should ignore the recommendation of splitting at Help:Conflation of two people? That page should probably be clarified/updated, and perhaps include best practices for different types of frequently encountered conflations, and a variety of case examples (e.g. internal wikidata conflation vs conflation at external sources). -Animalparty (talk) 17:33, 27 August 2022 (UTC)[reply]
@Animalparty This page is essentially an essay by Jura and doesn’t reflect current practice. This is especially true in cases where there has never been any doubt about the subject of the item – it’s just a wrong Library of Congress authority ID (P244). --Emu (talk) 18:57, 27 August 2022 (UTC)[reply]
Well at least it's written down. Is there anywhere on Wikidata that does codify best practices? This isn't a complaint at you, but it seems that Wikidata largely still operates on hunches and whimsy - the lawless wild west. There are very few rules and guidelines, and if there are any they are often hard to find, buried deep in the bowels of project chat archives or talk pages, or on seldom-frequented informational pages ("essays") or neglected project pages. There has to be a better system than repeatedly asking "anyone know a guy who knows whom to ask to know what to do in this situation?" -Animalparty (talk) 20:20, 27 August 2022 (UTC)[reply]
@Animalparty: For working with authority files, see Help:P227. BTW: GND now distinguishes the museum director from the (older) city surveyor. [31] Kolja21 (talk) 00:33, 28 August 2022 (UTC)[reply]
Yeah, this can be very frustrating. To be honest, I really don’t understand how anything is supposed to work at all given the organizational culture of Wikidata. But it curiously somehow does work and the results are staggering especially consider if you consider the pretty small userbase (let alone the almost laughable number of admins).
But then again, my former main project’s notability criteria fill 85 KB and the index to previous discussions of changes to the notability criteria amounts to a staggering 348 KB. That’s the other extreme. --Emu (talk) 20:39, 27 August 2022 (UTC)[reply]
To report incorrect clustering in VIAF, you can email bibchange@oclc.org The LC name authority file is actually maintained by hundreds of libraries around the world, and any library that contributes name authorities can revise existing ones or create new ones as needed. Therefore there are many other options besides contacting Library of Congress. You can contact any participating NACO (Name Authority Cooperative Program) library. The Program for Cooperative Cataloging's Wikidata Pilot participants are probably a good place to start: https://www.wikidata.org/wiki/Wikidata:WikiProject_PCC_Wikidata_Pilot/Participants. If LC doesn't respond to a report/request, feel free to contact another NACO participant. A list of all NACO libraries is at https://www.loc.gov/aba/pcc/naco/pccliaisons.html. UWashPrincipalCataloger (talk) 21:54, 29 August 2022 (UTC)[reply]
@UWashPrincipalCataloger Cases where the information in the authority file is conflated or wrong are fairly rare. Most problems arise from wrong books that are associated with a Library of Congress authority ID (P244) in the LC catalogue – I presume this can’t be changed by other participants? --Emu (talk) 22:19, 29 August 2022 (UTC)[reply]
@Emu I'm not exactly sure what you mean by "wrong books that are associated with a Library of Congress authority ID (P244) in the LC catalogue". Books don't get assigned a Library of Congress authority ID (P244). Are you referring to LC records describing books that have the wrong person or corporate body given in an access point? For example a book listed under Harrison, George, 1940- when the correct access point is Harrison, George, 1943-2001 ? UWashPrincipalCataloger (talk) 22:48, 29 August 2022 (UTC)[reply]
@UWashPrincipalCataloger I’m not really familiar with the “access point” terminology but I think we mean the same problem: Take n90677014. According to the “source” section, this LCCN refers to Georg Wimmer (Q106521531), the author of the 1989 dissertation Parteien in der Kommunalpolitik. However, this LCCN is also listed as the author of Kinderarbeit--ein Tabu Mythen, Fakten, Perspektiven by Georg Wimmer (Q1506451). That’s the reason why VIAF 51935298 can’t (CC Kolja21, Wurgl). --Emu (talk) 23:04, 29 August 2022 (UTC)[reply]
@Emu OK I see the problem. The same access point (Wimmer, Georg) is given on both https://lccn.loc.gov/2015400825 and https://lccn.loc.gov/90132196. The cataloger at LC who created the cataloging for the 2015 did not realize that these are two different Georg Wimmers. I have now created a new authority record for Wimmer, Georg, 1961- . It will take a few days for it to show up at http://id.loc.gov/authorities/names/no2022106416 and a few weeks to get into VIAF. I have reported the need for LC to change the access point on https://lccn.loc.gov/2015400825 to Wimmer, Georg, 1961- Hopefully they will make the correction within a few days. UWashPrincipalCataloger (talk) 23:49, 29 August 2022 (UTC)[reply]
@UWashPrincipalCataloger Thank you, that sounds very promising! Could you imagine sorting other similar problems out, too? There is some kind of backlog on de.wp (although there is no real list at the moment) and similar problems arise from time to time. But I don’t know how much time this would take. --Emu (talk) 17:44, 30 August 2022 (UTC)[reply]
@Emu I can certainly occasionally help with other problems with Library of Congress/NACO name authorities (or lack of one). That is an integral part of my job. UWashPrincipalCataloger (talk) 23:26, 30 August 2022 (UTC)[reply]

sales vs selling[edit]

I'd be really grateful if somebody could take a look at sales (Q194189) (currently "sale") and selling (Q3380760) (currently "selling").

It would be really useful to have distinct items for "sale" (ie a completed transaction) and "selling" (ie the process of persuading a customer to buy something) -- the two perhaps related by has effect (P1542) and has cause (P828).

Unfortunately, despite the current en-label "selling", selling (Q3380760) seems very confused at the moment, with most of the statements (and as far as I can tell some of the labels and descriptions in many languages) actually seeming to define the item as a completed transaction; while sales (Q194189) seems to be very blurred because many wikis (eg en-wiki but also many others) seem to only have one article for the whole area (sales / selling / salespeople / sales revenue), and sales (Q194189) seems to have picked up a blur of statements and identifiers blurring together a mix of all of these things.

We also have separate items for sales revenue (Q3934352) and for various occupations that would be sub-classes under the general ISCO/ESCO classifications retail and wholesale trade managers (Q108305359) (ISCO 1420), sales, marketing and public relations professionals (Q108289399) (ISCO 243) and sales worker (Q108285732) (ISCO 52) so they should also be kept well clear.

I'd be really grateful if somebody could look at this, and make sure the right items have the right interwiki links (?? or should the interwikis be grouped together ??), I'm just snowed under with other things myself at the moment. Jheald (talk) 10:02, 30 August 2022 (UTC)[reply]

Shortcuts for adding properties with qualifiers[edit]

I want to manually add launch dates to ships, but its tedious doing "add, type sig, select significant item from dropdown, type launched, select ship launching, add property, type poi, select pointintime, type 12 May 1873"

Can I setup a shortcut so I can bring up the entry box with a few keystrokes or clicking a button? Vicarage (talk) 09:18, 24 August 2022 (UTC)[reply]

idk. FWIW, My approach here is to have a spreadsheet template which generates the Quickstatements needed to populate a statement such as this, so that I need only enter the bare data, cut & paste the QS into quickstatements, and run. V.quick, IMO. --Tagishsimon (talk) 09:34, 24 August 2022 (UTC)[reply]
I can see how that would work, but not ideal when flipping between tabs in a browser, and a problem for dates, where the GUI is very flexible in interpreting dates, and QuickStatments very fussy about format Vicarage (talk) 09:43, 24 August 2022 (UTC)[reply]
Would it work to copy from other ships? If you enable the MoveClaim gadget in your preferences, you can copy an existing statement to another item. In that case you'd only need to change the launch date in the item to which you copied the statement to. Simeon (talk) 11:19, 24 August 2022 (UTC)[reply]
I had a look, but as its not drag and drop, you'd have to type in the fiddly Q number. I'm trying a SPARQL query that produces a quickstatements set of placemarker times which I could then edit to paste in the correct values, but the danger is missing some and leaving the incorrect placemarker. Vicarage (talk) 12:54, 24 August 2022 (UTC)[reply]
You can also copy/paste the Q number. Silver hr (talk) 19:10, 25 August 2022 (UTC)[reply]
In addition to the normal copy-pasting a clipboard manager like Ditto also helps to deal further. ChristianKl ❪✉❫ 21:15, 31 August 2022 (UTC)[reply]

Do we have properties for countries a domain's servers are located in?[edit]

A thing I care about lately is where a domain name has its servers, in which countries. Some servers like www.wikidata.org got its servers in the Netherlands(I'm from Norway) so in this example we don't get any Norwegian wikidata servers, serving Norwegians inside of this country.

Do we add such types of data or is it out of scope?

The idea is to add it as a qualifier for official website (P856) in a similar way that we add language of work or name (P407) as a qualifier for it Wewannarock (talk) 14:38, 31 August 2022 (UTC)[reply]

As far as I know there's currently don't have a property. https://www.wikidata.org/wiki/Wikidata:Property_proposal/Generic#service_hosted_by is a proposed property in that direction.
As far as Wikidata's own hosting goes I would expect it to be part of https://meta.wikimedia.org/wiki/Wikimedia_servers . According to that the main servers are in the US and the server in the Netherlands is only doing caching. ChristianKl ❪✉❫ 15:23, 31 August 2022 (UTC)[reply]
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2022/08&oldid=1723300049"