Wikidata:Project chat/Archive/2022/03

From Wikidata
Jump to navigation Jump to search

Q513149 and Q59677066

I became aware of the data object Pavel Novotný (Q59677066) via an the User Sofie Geneea. There are probably two people described there. In addition, this data object is also linked to identifiers to which the data object Pavel Novotný (Q513149) is already linked. I would be happy if someone could take care of this matter. --Gymnicus (talk) 11:45, 2 March 2022 (UTC)

Thanks for reporting, see Talk:Q59677066 Vojtěch Dostál (talk) 09:12, 5 March 2022 (UTC)

How to download a single shape from OpenStreetMap to upload to Commons?

Hi all

I'm working on using Kartographer to create a maps for Wikipedia articles for species using Wikidata. I wrote Wikidata:Map_data with Simon Cobb a few years ago but don't remember it well. I'm getting stuck trying to download shape files for countries and other areas individually, e.g Sicily, does anyone have experience in downloading from OSM who knows how to do it? I've read https://wiki.openstreetmap.org/wiki/Shapefiles but it doesn't explain how to download a single shape file. Any ideas?

Thanks

John Cummings (talk) 13:59, 28 February 2022 (UTC)

Map
Principal areas of Wales
@John Cummings: Downloading? Why not just create a map like the example on the right?
See mw:Help:Extension:Kartographer for more examples. Multichill (talk) 19:22, 28 February 2022 (UTC)
Thanks @Multichill: for the suggestion, this is a really amazing possibility but I don't think it covers everything that's needed. To explain, I'm working with Alicia at WMSE to try to create distribution maps on species articles on en.wiki (which would hopefully get picked up by other languages as well). We've made a pretty good prototype here using Kartographer which uses QIDs and Commons geoshapes. I wasn't aware you could pull shapes from OSM which is amazing and certainly would work some of the time, however species distribution is quite 'messy' and doesn't follow political boundaries a lot of the time, e.g see this species distribution map for a red panda which is a habitat type at a certain elevation. Marine species are certainly not going to follow political boundaries a lot of the time.
I think I can start with using OSM shapes for a lot of plant species but I also really want to make good instructions for importing geoshapes from other sources as well so its easy for people to add them. I wrote Wikidata:Map data with Simon Cobb a couple of years ago which covers using map data from Commons but its very vague on how to find and import them and I'm really struggling to work it out.
Ideally what I'd like is a simple template you can include in the infobox where you can list QIDs in different groups (allowing for things like native and invasive ranges) and then the user can just click save and a map appears.
What do you think is a good way to proceed with this?
Thanks
John Cummings (talk) 10:11, 1 March 2022 (UTC)

Request for feedback on Wikidata-based project

Hi all, I've been working on a project that makes use of Wikidata as a structured tagging system for a social RSS feed aggregator. I think there is a lot more potential for further integrating Wikidata but am in want of some ideas. Inspired in part by https://en.wikipedia.org/wiki/Wikipedia:Wikipedia_is_an_MMORPG, I am especially interested in ways to more richly use Wikidata in order to create a sense of space, as in a territory for exploration, though any and all feedback concerning the project is appreciated. Apologies if this isn't the right place for this, I couldn't find a place for Wikidata-adjacent discussion. Nivekuil (talk) 01:36, 2 March 2022 (UTC)

Processing of federated queries changed?

From the SPARQL endpoint of the 20th Century Press Archives (Q36948990) at http://zbw.eu/beta/sparql/pm20/query, I regularily run federated queries include data from WDQS. Since yesterday noon, these queries result in "Bad request". Strpped down example: 

   select * where {
     service <https://query.wikidata.org/sparql> {
       ?item wdt:P4293 "pe/000012" .
     }
   }

execute

The endpoint is on the list of supported endpoints. Any ideas what may have changed? --Jneubert (talk) 09:41, 19 February 2022 (UTC)

As of today, still, a "Bad request" is returned from queries which were working for years now. Cheers, --Jneubert (talk) 09:52, 23 February 2022 (UTC)
I think this should better be reported at Wikidata:Report a technical problem/WDQS and Search. —MisterSynergy (talk) 18:57, 23 February 2022 (UTC)
Seems to work again - cheers, --Jneubert (talk) 14:27, 2 March 2022 (UTC)

Is Kartographer broken?

Tracked in Phabricator
Task T302853

Hi all

I'm doing some work on Kartographer maps and yesterday they all went blank (a few didn't go blank but the shapes on the maps dissapeared... is anyone aware of an existing bug? Or should I file something on Phabricator?

Here's an example:

Thanks

John Cummings (talk) 08:45, 2 March 2022 (UTC)

The maps on the example page loaded for me eventually, but only after a while, and possibly with missing features. Looks like there’s a phabricator task now: phabricator:T302853 --Lucas Werkmeister (WMDE) (talk) 09:21, 2 March 2022 (UTC)
Thanks very much Lucas Werkmeister (WMDE). John Cummings (talk) 09:53, 2 March 2022 (UTC)

Is there a need for a property “reference supports qualifier”?

Hello! I am currently thinking about proposing a new property “reference supports qualifier” and would love to hear some opinions of fellow contributors of Wikidata before starting a proper proposal.
Especially for politicians I've noticed that when citing news articles these in most cases only mention either a start date or an end date of a position. This can lead to the situation that after a term ends a contributor qualifies an existing statement (which already has a reference) with end time (P582) but forgets to add a new reference supporting the ending date. This causes the already existing reference to give the impression that it supports both start and end date. With the proposed property it would be possible to restrict a reference to a subset of qualifiers and therefore make it detectable by machine whether/which qualifiers are lacking references.

An example for the application I have in mind:

position held
Preferred rank Mayor of New York City edit
start time
end time
2 references
stated in The New York Times
title Lorem Ipsum (English)
publication date
supports qualifier start time
stated in The Washington Post
title Dolor Sit (English)
publication date
supports qualifier end time
add reference
add value

There was already a proposal 2019 but from what I can tell it mainly failed due to missing examples rather than opposition to the general concept. --Nw520 (talk) 19:21, 6 March 2022 (UTC)

@Nw520 This problem also popped up during WikidataCon 2021 and I support your proposal which does sound elegant. My question would be how to show that a reference only supports the main value but none of the qualifiers? --Emu (talk) 19:43, 6 March 2022 (UTC)
@Emu: Actually, that was me who brought that up, but it was only recently that I finally felt the need to address this issue. Regarding your question:
What do you think about 1. supports qualifiernovalue? (Bad: if in the future another contributor adds a new qualifier to statement that is actually supported by the reference and doesn't update the reference this snak becomes incorrect, though I do think that it's rare that an already existing reference supports a qualifier that is added some time in the future) 2. Not using this property at all would be another idea but that would be ambiguous since not having a snak with this property could then imply not checked yet or none. 3. Yet another idea (and the most precise one) might be an additional property not supported qualifier, but IMO that would be overkill. --Nw520 (talk) 20:07, 6 March 2022 (UTC)
After some thinking about this case I'd say that the property should only be used to make positive claims (“supports”) and should not be used to make any negative claims (“does not support (any qualifier)”). This entails that a reference not having this qualifier would, indeed, be ambiguous and might mean that the reference does not support any of the qualifiers or that nobody has checked yet. Do you think, that supports qualifiernovalue might be better? --Nw520 (talk) 10:11, 7 March 2022 (UTC)
Personally, I find the property quite useful and would support a suggestion. --Gymnicus (talk) 19:44, 6 March 2022 (UTC)
Yes, this is really needed. Currently, we are sort of forcing reusers to use all the sources, where some might not even mention the statements they are reusing. We should do better, and this seems to solve many obvious use cases (I have encountered situations similar to your example many times.) Ainali (talk) 21:55, 6 March 2022 (UTC)
Yes, this is definitely needed. Ayack (talk) 09:44, 7 March 2022 (UTC)
In the Wikidata Telegram chat a user remarked that it wouldn't be possible to distinguish between qualifiers that share the same property (but have different values). From my experience I'd say that such statements are very uncommon and that most occurrences are because of botched QuickStatements. However, it indeed is a limitation. Since I can't think of any other model I would claim that the limitation is acceptable. --Nw520 (talk) 10:07, 7 March 2022 (UTC)
This would be super useful for lots of position held (P39) statements which tend to have quite a few qualifiers from different sources. Popperipopp (talk) 14:30, 7 March 2022 (UTC)
I've created a proposal at Wikidata:Property proposal/Supports qualifier 2. --Nw520 (talk) 19:41, 8 March 2022 (UTC)

Data Reuse Days: 35 sessions to discover how Wikidata's data is reused in cool projects

Hello all,

The Data Reuse Days is a series of online events that will focus on the use of Wikidata's data: during the different sessions taking place from March 14th to 24th, you will discover plenty of projects using Wikidata's data, how they gather, polish and display the data, what challenges they encounter on the way, and how they give back to the community. You can also attend presentations and workshops to learn how to create tools using Wikidata, how to use Wikidata on the other Wikimedia projects, and participate in various discussions with editors, data reusers and members of the Wikidata development team.

You can have a look at the schedule on this page or on the calendar view. Most sessions are taking place online on Jitsi, no registration or software needed. All events are free and open for everyone to join. The entire initiative, including the presentations and discussions on various channels, is covered by the Code of Conduct for technical spaces.

Each session is independent, and while the schedule can look a bit intimidating, we encourage you to look for the sessions that could be interesting for you, pick a few of them, and join spontaneously if you have time. To help you navigate the schedule, here are a few suggestions on where to start:

Note about the calendar features: you can add a single session to your personal calendar, or even embed the full schedule. Please note that we're experimenting with using Diff's calendar feature, that is still in a test and improvements phase - some bugs may appear. If you notice anything strange, let me know so I can report them to the dedicated team at WMF.

If you have any questions, or if you would like to propose a last-minute presentation (there is still space left in the lightning talks sessions, or for interactive sessions such as editathons), feel free to contact me.

We are very excited to welcome you all in this new experiment of a remote Wikidata event, and to start plenty of interesting discussions between people editing, maintaining and using Wikidata's data! Cheers, Lea Lacroix (WMDE) (talk) 08:46, 3 March 2022 (UTC)

Help: Merge

Can someone assist me with merging Q111077834 and Q26807837? They relate to the same person. I tried but for some reason the merge hasn't succeeded. Thanks. - Naushervan (talk) 12:53, 3 March 2022 (UTC)

✓ Done I haven't problem. ValterVB (talk) 13:42, 3 March 2022 (UTC)
Thank you, @ValterVB! - Naushervan (talk) 13:58, 3 March 2022 (UTC) 13:57, 3 March 2022 (UTC)

Potential "Airspace Violation" property, data interesting/useful?

If we had an airspace violation property, then we could possibly add data regarding when countries airspaces get violated by their neighboring or other countries. Pros and cons, well the good thing about it we might get extra data regarding how countries behave in a "military way" to their country neighbors and other countries. As in pros/cons, the cons might be that we'll be very dependent on statements(primary sources, right?) by countries, so statement disputed by (P1310) and statement supported by (P3680) might happen a lot by both sides(aggressor country target country), and the target country might be reporting on the violation while few independent sources might be able to confirm or disconfirm whether it happened or not, unless another country can confirm or disconfirm an incident. Imatotalcraterafteryournuclearweapons (talk) 13:21, 3 March 2022 (UTC)

Help us with a new tool for creating presentations of Wikidata

Hello everyone,

Wikidata has done a fantastic job curating valuable data. Developers have been able to create a number of cool presentations of Wikidata, even integrating Wikidata with data from other websites. But creating them takes significant programming effort. We aim to develop tools that make it easier for programmers and non-programmers to create such presentations. We would like to understand your experience with Wikidata to better improve our tool. It would help if you can fill out our survey: https://mit.co1.qualtrics.com/jfe/form/SV_cvZKKlRu2S7C9Fk This survey covers 4 main topics: how you are currently using Wikidata; whether you have or want to create presentations of Wikidata; if so, what these presentations are; and how creating them can be made easier for you.

Thanks! Tarfahalrashed (talk) 20:23, 4 March 2022 (UTC)

Varlamov - Skype username

Q4103885 - why Skype before email and website and far away from VK, TikTok, YouTube, Facebook, Telegram? War4UA (talk) 20:40, 3 March 2022 (UTC)

Because skype does not belong to social media? But I do not think we can affect the order in which properties are shown (is there any public algorithm for the order?)--Ymblanter (talk) 21:40, 3 March 2022 (UTC)
This gadget will sort statements into a specific order. Use it by adding importScript( 'MediaWiki:Gadget-statementSort.js' ); to Special:MyPage/common.js. Bovlb (talk) 22:26, 3 March 2022 (UTC)
Looking at the details, the section isn't called "social media", and if it were, why is "social media followers" not there? In my interface the section is called "Identifiers", and the Skype username is just that, an identifier, like the Wikimedia username which is listed in that section. War4UA (talk) 22:42, 3 March 2022 (UTC)
MediaWiki:Wikibase-SortedProperties determines the order of sorting of properties on all items, but there is a technical requirement that all the "identifier" type properties have to be in the end section. In this case, Skype is not an identifier property and is grouped with other properties linked to addresses/communications (phone, email, etc). However the others (tiktok, telegram, etc) are defined as identifier properties, and for various reasons most of the identifier properties are just in alphabetical order. (It might make sense to group all the social-media identifiers together but that hasn't been done yet.) Why it's like this is a bit unclear, though. The Skype property is quite old, and I think if it were created today it would be an identifier. Andrew Gray (talk) 23:17, 3 March 2022 (UTC)
"if it were created today it would be an identifier" - is there a way to change such a situation when it is determined to be wrong based on current knowledge? War4UA (talk) 11:19, 4 March 2022 (UTC)
It is technically possible to change the type of a property, but it's complicated and needs to be done by a developer (see eg phab:T255241). If you want to do this I would recommend posting on the property talkpage to get agreement (there may be an obvious reason not to do it which I am not thinking of). Andrew Gray (talk) 21:17, 4 March 2022 (UTC)

Description is changed

Exist some tool that show all the description changed in a specifici language? (If possible with indication if bot or normal user, and if changed or new addition.) It's ok also using of API --ValterVB (talk) 21:16, 3 March 2022 (UTC)

Does this tool meet your need? Bovlb (talk) 22:19, 3 March 2022 (UTC)
Thanks @Bovlb:, not completely, but it is a good starting point. I'll must work on it. --ValterVB (talk) 08:18, 4 March 2022 (UTC)

Request

Hello dear Wikidata administrators

‏I have two accounts with the following usernames: ‏1- @arian_aboutalebi is an Extended confirmed users account created in Persian Wikipedia ‏2- @zaghmarz1 which was created 8 years ago (2014) but I forgot my password by February 2022 . ‏In February, dear @ladsgroup admin blocked me due to spam behavior

‏A few days before I was blocked, I remembered my @zaghmarz1 account password and decided to continue with this account. And in a few days I made 112 edits to update and upgrade the information in Wikidata. Please see 112 edits for me and then tell me which version was wrong? ‏Even if, for example, one of the 112 edits is an error, 111 useful edits have been made, and I think blocking this is not the right thing to do, and we should consider a large number of good and useful edits. Only in one of the editions did I send a message to the esteemed administrator of @bovlb and request that he add the mzn Wikipedia page of Mr. Jamal Aboutalebi ( Iranian politician and environmentalist ) to his item a few days after its creation. Is this spam behavior or requesting to add a resource to a page in Wikidata !? Please help me unblock ‏Please read what I said on the page below and then respond if possible. Thank you for your attention

https://m.wikidata.org/wiki/User_talk:Zaghmarz1

Already blocked, see discussion on WD:AN. Why do people think that spamming multiple fora will get them a more favourable hearing? Bovlb (talk) 22:13, 10 March 2022 (UTC)

Updating country during the Ukraine war

Lots of people are going around updating the country of various places as the war develops. Is this good? Is it clear that for example Snake Island (Q155868) is part of Russia (Q159) right now? Is military control of a location the same as being in that country? Is it even clear that Russia intends to annex all this territory? I propose we not update country country (P17) for all these items until things are more clear. Though stopping all the changes in a million different places will be real hard. Thoughts? BrokenSegue (talk) 02:55, 28 February 2022 (UTC)

Has any reliable source claimed Q155868 is part of Russia now? If not, that statement should be removed based our current practices. If there exists a source stating that Russian has claimed Q155868 as their territory, then we have qualifiers that can be used to better describe the situation (eg. statement supported by (P3680) and statement disputed by (P1310)). --Shinnin (talk) 03:52, 28 February 2022 (UTC)
I mean there are sources that it's controlled by Russia. It's rare for sources to be so specific. Part of the problem is that country (P17) is kinda vague about what it means. BrokenSegue (talk) 03:59, 28 February 2022 (UTC) @Benbaruch: who made the edit in question. BrokenSegue (talk)
A country occupying a piece of land is different from the said country claiming to own that piece of land. Is there a reliable source that says Russian has claimed Q155868 as part of Russia? Is there any source, Russian or other, that claims Russia has, or intends to, annex Ukrainian land at all? Last I checked, Russia was claiming something completely different. This is an ongoing situation, exact words are important here. Unless there is a statement from Russia stating that Q155868 is part of Russia, we shouldn't model Q155868 as part of Russia at all. Not even with qualifiers. The occupation can, of course, be modeled with significant event (P793), if there are sources to support it. --Shinnin (talk) 05:44, 28 February 2022 (UTC)
No, it is not good and must be reverted on sight. We can add occupation indeed if there are reliable independent (meaning not Russian, not Ukrainian) sources, but I assume (based on my experience with similar edits on the English Wikipedia) that in most cases these sources do not exist.--Ymblanter (talk) 19:55, 28 February 2022 (UTC)
There was such an edit war when there was the nagorno karabakh war. Just wait until it is clear, with independent sources, as is it the practice on many wikipedias. Bouzinac 💬●✒️●💛 11:21, 5 March 2022 (UTC)

Sidebar - language links - from additional wikidata items

Hello, wikispecies:Hydrobates matsudairae is linked to Q28122588, but all the different language wikipedias are linked to the synonym/combination Oceanodroma matsudairae Q785281. I could manually add [​[en:Oceanodroma matsudairae]​] etc to the wikispecies page so that there would be a link to enwiki in the sidebar, but that would not reflect any additional links added to wikidata. How instead could I make appear in the sidebar, in addition to any language links from Q28122588, all those from Q785281 (and Q28122592)? Ideally through a similar mechanism to [​[en:Oceanodroma matsudairae]​], but instead [​[wikidata:Q785281]​] & [​[wikidata:Q28122592]​], or similar. Thank you, Maculosae tegmine lyncis (talk) 01:39, 6 March 2022 (UTC)

Can the public create a "Tree of Knowledge" in Wikidata?

I'm basically looking for a centralised tree of knowledge, where you can collaborate with the public to construct mind-maps within each topic As with Wiki pages - lots of hyperlinks, anyone can edit the articles. However, wikipedia presents information as long passages of text. Often it is easier to present information as a logical flowchart or mind map, rather than as passages of text. I'm familiar with the existence of specific 'wiki' sites, however that is not what I mean. I'm looking for a site containing "mind maps", rather than passages of text (as you see in a traditional wikipedia article)

I'm looking for a free, open site where the public can collaborate simultaneously to create a universal "Tree of Knowledge"

Any suggestions?

Thank you

Vitreology (talk) 07:53, 6 March 2022 (UTC)

Individual official Website on (Pierre Weil (Q3387348)). Url is now used by a different owner - Original is archived on Wayback Machine

Hello every one - I am at a loss here, I removed the hijacked url for property 'official website', (I realized I should not have, Sorry), I tried to enter the web archive url, but it does not seem to be acceptable. Could someone help populate the entry according to proper way? Archived text for English is here: [1] and multilingual here: [2]. Thanks --DDupard (talk) 11:16, 5 March 2022 (UTC)

Think it's good now, the used qualifier properties are somewhat high in the suggestion list. Sjoerd de Bruin (talk) 11:54, 5 March 2022 (UTC)
Thank you Sjoerd de Bruin - The Various corresponding Wikipedia pages still respond with the wrong (old) url when the automatic template 'official website' is used . ....? ;)--DDupard (talk) 13:46, 5 March 2022 (UTC)
I've purged the FR page - can't see a URL now. I've set the hijacked site statement to normal rank: it was true, it is no longer true, there is an end date. Deprecated rank is for values that were never true. The <no value> statement is preferred rank, and so will be served as the answer to truthy questions, e.g. by templates. --Tagishsimon (talk) 14:22, 5 March 2022 (UTC)
Even purged, It still goes to the wrong website/content. The problem is that the new owner or (I would describe it as the 'new occupant') of the url address is in control of that url. The previous owner/occupant is deceased. The archive is at Wayback Machine, but there is a legal "imbroglio" now, since the new owner seems to have acquired some rights, he could ask Wayback Machine to delete the archived content. From what I understand at Wikipedia article 'Wayback Machine' here :[3]. I guess the solution would be a transfer of the content at another site by rightful owner heirs. --DDupard (talk) 11:46, 7 March 2022 (UTC)
The purge comment related to the language wiki pages, which are now free of the URL. The URL in WD still goes to the wrong place, but that's not an issue for WD; WD merely memorialises that the URL in the past served the official website. --Tagishsimon (talk) 11:52, 7 March 2022 (UTC)

Aliases on names

I’ve seen that a lot of names have aliases in english like on Q4218918. I don’t see use in having aliases with (first name) and (given name) as I think this is information that is (should be) part of annotations/properties and should not be part of aliases. So I would tend to remove them. Some users seam to have the opinion that this is valid data and should not be removed. So I would like to get some input if this aliases are valid or if they can be removed. Thanks —DaSch (talk) 19:49, 6 March 2022 (UTC)

These are important for finding these items when there are a lot of items with the same label! Please don't remove them. Sjoerd de Bruin (talk) 20:06, 6 March 2022 (UTC)
But it’s also part of the description. And there should only be one entry for every first name. So the description is sufficient to distinguish them from other data. That’s at least what if seen when searching for names in wikidata. The aliases are actually quite confusing as they aren’t aliases but labels with additional data. Which from my understanding is complete nonsense for a database. —DaSch (talk) 20:12, 6 March 2022 (UTC)
Additionally your assumption that they can be use to find items with the same label is wrong. The aliases of Q18220903 and Q12173670 are the same. So the alias is useless. Only the description adds information to allow to distinguish these two! —DaSch (talk) 20:20, 6 March 2022 (UTC)
No, DaSch, you are completely wrong. Aliases allow items to be found. Sure, more than one item may be found, but that's fine. Removing aliases makes the item harder to find. Once a list of items has been found, then yes, the description should help to disambiguate. It really is an application of Postel's law - the robustness principle - to provide as many hooks into the item as possible for the benefit of discoverability. Please do not remove aliases from items. --Tagishsimon (talk) 20:34, 6 March 2022 (UTC)
Well. That’s really bad. So we are fixing bad search engine by adding aliases? So I’ll have to add a filter for misused aliases. Maybe we should thing about adding some kind of search/indexing hints instead of polluting aliases? DaSch (talk) 20:54, 6 March 2022 (UTC)
You could easily have 71 items labeled "Peter" and the one you are looking for (the given name) could be at the end.
If you think you need to filter these aliases, maybe you are using aliases incorrectly. Can you provide a sample of your usecase? --- Jura 21:12, 6 March 2022 (UTC)
I’m for using data from wikidata for a keywording engine. I would like to display these keywords together with aliases to the user. With these kinds of indexing hints the aliases are displaying a lot of redundant data to the user, which in many cases makes the actually useless. After looking at some other examples I must say that many items have a lot of redundant aliases. For me an alias is something like Helicopter for german Hubschrauber. But looking on the aliases at Q985 it’s completely redundant. It’s manual stemming which is something search engines can do automatically. DaSch (talk) 21:25, 6 March 2022 (UTC)
If you are looking for pretty looking additional labels, maybe other name properties are more suitable for you, e.g. name (P2561) or short name (P1813). At Wikidata, these aren't indexed though. --- Jura 21:30, 6 March 2022 (UTC)
Also, the basic idea is to that the user has label and description to identify a concept, not label and (all) aliases. --- Jura 21:44, 6 March 2022 (UTC)
Yeah. But in the end aliases are not really aliases but spelling variants and indexing hints. It’s something we do through proper indexing without the need of any manual work. For me it looks like aliases are miss-used because of software limitations. Seams like I overestimates the data quality in wikidata. Actually I’ve seen this type of aliases in former projects and they were always an indicator for some issues in the data structure and handling. DaSch (talk) 21:54, 6 March 2022 (UTC)
There is a perpetual proposal to make suggestions of values for item-datatype properties linked to the relevant type constraints, but somehow that seems too complex to implement. --- Jura 22:16, 6 March 2022 (UTC)
Well, then I’m very sorry for you having to work around this limitations. I‘ll replace add needed aliases as names and remove aliases from my UI. DaSch (talk) 22:23, 6 March 2022 (UTC)
I agree that these aliases should not be there. Aliases are provided as part of our content to Wikidata consumers, they are primarily not for internal use, AFAIK. Why would be want to give our consumers weird auxiliary versions with brackets? Vojtěch Dostál (talk) 21:17, 6 March 2022 (UTC)
What alternative do you propose? --- Jura 21:20, 6 March 2022 (UTC)
Add a new field that contains additional labels for indexing. DaSch (talk) 21:26, 6 March 2022 (UTC)
In some specific cases there are very useful. When you edit some items manually and exists too much items with same label, some kind of disambiguate is needed. Another usecase is name of street - there exists dozens or hunerts of streets with same name and even if there is descruption, the right one is ususally on the seventh page of results.. JAn Dudík (talk) 13:04, 7 March 2022 (UTC)

Wikidata weekly summary #510

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Discussions
    • New requests for permissions/Bot: EnvlhBot 3. Task/s: add dictionaries IDs to French lexemes
    • Closed request for permissions/Bot: IndoBot (Approved). Task/s: I would like to import all Indonesian schools, more than 100000. The data includes school type, location, and coordinates as well as external identifiers
  • Tool of the week
    • Wikxhibit is a tool that allows anyone, even non-programmers, to create cool presentations of Wikidata, and other sources of data on the web, only using HTML and without any additional programming. Are you interested in creating presentations of Wikidata? We would like to understand your experience with Wikidata to better improve our tool. It would help if you can fill out our survey https://mit.co1.qualtrics.com/jfe/form/SV_cvZKKlRu2S7C9Fk
  • Other Noteworthy Stuff
    • Wikidata dumps: Due to technical issues the JSON and RDF dumps for the week of March 1st couldn't be properly generated (phab:T300255#7746418). The situation is expected to get back to normal this week.
    • Item with QID 111,111,111 was created
    • Job openings:
      • The development team at WMDE is looking for a Senior Software Engineer to develop and improve the software behind the Wikidata project. Apply here!
      • The WMF Search Platform team is looking for someone to maintain and develop Wikidata Query Service. Apply here!
      • Natural History Museum, Berlin (Q233098) is looking for someone for project- and data management especially for Wikidata related stuff about the museums collections Job Description (German)
  • Development
    • Data Reuse Days: Preparing for the upcoming Data Reuse Days. Join us! A lot of exciting sessions are coming together.
    • Lexicographical data: Continued work on the new Special:NewLexeme page. We are getting close to the point where it can create a Lexeme.
    • REST API: Continuing to work on the first endpoint to read Item data.

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.

Main page welcome message cut off on mobile

Just a notice if anyone would be interested in fixing it.

Lectrician1 (talk) 14:30, 7 March 2022 (UTC)

A problem in merging concepts in one article and confusion in crossWikis

Hi, an article and its Wikidata item may be a merged item of two or more concepts, and each concept should have a specific item in Wikidata (Unrelevant to the article that contains them). Here we should create an item like "Merged item of Concepts" to the Wikidata item of the merged article. Otherwise we have a serious problem (confusion) in "other languages" section of merged article if in another language we have an article for each (unmerged) concept.

Let me give an example. This article is a merged article for the Sine and Cosine concepts: https://en.wikipedia.org/wiki/Sine_and_cosine . In Persian language there exists a separate article for Sine and for Cosine. Here we cannot create a correct crossWiki for "Sine_and_cosine" article in English, because it should point only to one article at a one time (not point to both of them), i.e., it should point to either سینوس or کسینوس in Persian, and cannot point to both of them at the same time.

So I think we should have an item like "Merged Concept of" in the Wikidata item an article like "Sine_and_cosine" and this item causes crossWikis of سینوس and کسینوس articles in Persian language point to the article "Sine_and_cosine" automatically. Thanks, Hooman Mallahzadeh (talk) 16:46, 7 March 2022 (UTC)

Something like Wikipedia article covering multiple topics (Q21484471), then? --Tagishsimon (talk) 16:58, 7 March 2022 (UTC)
@Tagishsimon: Can it solve the crossWiki problem I mentioned above? I think in other languages section we should choose for Persian language, one of articles سینوس and کسینوس as options. I think, as a scenario, there should exist a list for the articles that are merged, and lets the user to choose one of them for the same language. Hooman Mallahzadeh (talk) 17:08, 7 March 2022 (UTC)
@Hooman Mallahzadeh Yeah, use Wikipedia article covering multiple topics (Q21484471) and indicate the two (or more) subjects with main subject (P921). At the moment, the Wikipedias won't use this data to display more interwiki from related items, but they could do it in principle. I think it would be very valuable, not only for Wikipedia article covering multiple topics (Q21484471) but also other cases where we have two separate items but Wikipedias might profit from interwiki being displayed. However, it's not up to the Wikidata community to approve and develop such a function. Vojtěch Dostál (talk) 18:05, 7 March 2022 (UTC)
@Vojtěch Dostál@Tagishsimon I have added a Quantifier and two Properties to the Wikidata item https://www.wikidata.org/wiki/Q13647261, please inspect that. But a list of merged concepts is not shown on the English Wikipedia article https://en.wikipedia.org/wiki/Sine_and_cosine.
To do that, may I submit a new task at Phabricator? Hooman Mallahzadeh (talk) 18:48, 7 March 2022 (UTC)

Permanent duplicated items

I just mostly-merged Q22521049 into Kamilukuak Lake (Q1723080). I added reciprocal permanent duplicated item (P2959) properties for the two. Following the instructions at Help:Merge#Items to be merged with sitelink conflicts, I also made Q22521049 an instance of Wikimedia duplicated page (Q17362920), with a P642 qualifier pointing at Q1723080.

But now there's a constraint warning at Q22521049: "An entity should not have a statement for permanent duplicated item if it also has a statement for instance of with value Wikimedia duplicated page."

So what's the right way to do this? Is permanent duplicated item (P2959) unnecessary or deprecated, and is Wikimedia duplicated page (Q17362920) preferred? —scs (talk) 13:35, 6 March 2022 (UTC)

@scs: I took a look at other comparable data objects and now you should probably do it with the statement instance of (P31)Wikimedia permanent duplicate item (Q21286738). Then there is no error with the constraint of the property permanent duplicated item (P2959). (Examples: Q22828260, Q31801653 and Q24702261) --Gymnicus (talk) 13:44, 6 March 2022 (UTC)
If the article on arzwiki is duplicate, the wiki should be told that it should be merged. When it's been merged, the items can also merged. So this is not a permanent copy and permanent duplicated item (P2959) is not appropriate. (If the article was not duplicate, it would be neither Wikimedia duplicated page (Q17362920) nor Wikimedia permanent duplicate item (Q21286738).) --Matěj Suchánek (talk) 09:43, 8 March 2022 (UTC)

Invitation to Hubs event: Global Conversation on 2022-03-12 at 13:00 UTC

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

Hello!

The Movement Strategy and Governance team of the Wikimedia Foundation would like to invite you to the next event about "Regional and Thematic Hubs". The Wikimedia Movement is in the process of understanding what Regional and Thematic Hubs should be. Our workshop in November was a good start (read the report), but we're not finished yet.

Over the last weeks we conducted about 16 interviews with groups working on establishing a Hub in their context (see Hubs Dialogue). These interviews informed a report that will serve as a foundation for discussion on March 12. The report is planned to be published on March 9.

The event will take place on March 12, 13:00 to 16:00 UTC on Zoom. Interpretation will be provided in French, Spanish, Arabic, Russian, and Portuguese. Registration is open, and will close on March 10. Anyone interested in the topic is invited to join us. More information on the event on Meta-wiki.

Best regards,

Kaarel Vaidla
Movement Strategy

--YKo (WMF) (talk) 04:04, 8 March 2022 (UTC)

Take a look at Raymond Theodore Wolf (Q107218828) and at Gertrude A. Wolf (Q107297904) where the Legacy link gives a 404 error, but other links with that property in other entries are ok. Can anyone see what I am doing wrong? --RAN (talk) 14:19, 7 March 2022 (UTC)

This link ([4]) works for me, but this one ([5]) and this one ([6]) do not. Isn't the formatter URL (P1630) for Legacy.com person ID (P8367) wrong? --Matěj Suchánek (talk) 09:33, 8 March 2022 (UTC)
I've run into this problem several times before and only recently found a solution. View the source code for the obituary page and search for "GBID". This is the number that should be used with Legacy.com person ID (P8367).
--Quesotiotyo (talk) 18:06, 8 March 2022 (UTC)
Thanks, weird stuff going on! --RAN (talk) 00:58, 9 March 2022 (UTC)

How to represent billionaires on Wikidata?

I was looking for ways to encode information about billionairehood to billionaires on Wikidata that make them easily retriavable.

I have added instance of (P31) statements to billionaires pointing to the item billionaire (Q1062083), scraping the information from English Wikipedia.

I understand this is less than ideal, but both occupation (P106) and has characteristic (P1552) seemed imprecise. Do you have any suggestions?

See https://www.wikidata.org/wiki/Wikidata:Edit_groups/QSv2/77358 . for the Quickstatements batch.

Best, TiagoLubiana (talk) 19:45, 4 March 2022 (UTC)

Try net worth (P2218)? --Tagishsimon (talk) 20:11, 4 March 2022 (UTC)

Net worth is good, but a bit too good; frequently the exact amount is not known or shown in a database. The idea of a billionaire is also a bit complicated; it is not always tied to a billion units of currency. TiagoLubiana (talk) 22:54, 4 March 2022 (UTC)

If your source asserts they are a billionaire, but you have no source for a monetary value, then net worth (P2218) works as <unknown> with a qualifier of minimum value (P2313) = $1,000,000,000 (or whatever the currency is). I think the only reasonable alternative is has characteristic (P1552). --Tagishsimon (talk) 23:07, 4 March 2022 (UTC)
Before we ask how, we should ask: "Should we?". Why the arbitrary value of billionaires? If we should have that, then we should have millionaires and thousandaires as well. Or really any value for what the person is estimated to be worth. Which brings us back to net worth. Ainali (talk) 23:20, 4 March 2022 (UTC)
I am proud to tell I'm a billionaire (in Sao Tome dobra ^^ ). Bouzinac 💬●✒️●💛 11:20, 5 March 2022 (UTC)
  • If they appear in a list like "Forbes Billionaires" you can add "award=Forbes Billionaires" with "point in time=2022". There are several lists online, this would be less subjective. --RAN (talk) 07:05, 5 March 2022 (UTC)
One might have thought that Wikidata:WikiProject Q5 (assuming someone knew to look for it) might have some guidance for how to cover biographies, living or dead. But no. It's bereft of practical information. Wikidata:WikiProject Biographical Identifiers gives some basics on basic human properties, but devotes the page largely to external identifiers. Do we really lack a centralized framework for best practices on how to structure and organize the millions of items on individual living or dead humans? I wouldn't be surprised if we don't: as the blind lead the blind through the wild west. 08:31, 6 March 2022 (UTC)
  • General comment on the thoughts:
    • Why billionaires? Billionaires are rather powerful people, with relevance to understand the world politics and economics. It is a gross synonym of "super super wealthy". Maybe it would be better to use Q17087953; "billionaire", though, is what is used in en.wiki.
    • has characteristic (P1552) seems reasonably good, easy to query and less troublesome. I'll revert the batch and use "has quality". TiagoLubiana (talk) 15:15, 9 March 2022 (UTC)

The wikidata page about Picasso's Violin Glass Pipe Inkwell contains an image from wikimedia. I'm skeptic about the copyright status of that image. The wikimedia page says :

This work is in the public domain in its country of origin and other countries and areas where the copyright term is the author's life plus 70 years or fewer.


But Picasso died in 1973 ...

In my opinion, the copyright statement is wrong and the picture should not be displayed.

Am I missing something ?


Laurent.Claessens (talk) 08:39, 9 March 2022 (UTC)

I nominated it for deletion at Commons. Ghouston (talk) 10:29, 9 March 2022 (UTC)

Why Wikidata have been created ?

Hello,

I'm creating the Wikidata MOOC in french. Some beta testers are asking me the information "why" Wikidata have been created, and I'm not sure to provide the good answer, because this is not clearly explained in the meta pages. Was Wikidata originally created to be able to provide up-to-date data in the infobos of the language versions of Wikipedia, or with a broader objective of centralizing redundant information?

(I'm writing in english in order to share more easily this message among non-french-speaking contributors)

Thanks for your help :-) --Amélie Charles WMFr (talk) 14:17, 9 March 2022 (UTC)

@Amélie Charles WMFr The principal use case in the early days were interwiki links, see en:Wikidata#Initial_rollout. However, there has always been the idea to move beyond that. --Emu (talk) 18:27, 9 March 2022 (UTC)
@Amélie Charles WMFr: Sadly looks like a couple of the most interesting papers are closed (to me) access - [7] & [8], but authors Denny Vrandeči & Lydia Pintscher may be able to help? The first few pages of Wikidata: A Free Collaborative Knowledge Base may help. tl:dr the genesis was recognition that Wikipedia failed to provide direct access to most of its data; data was submerged in 30 million articles in 287 languages with no mechanism by which facts could be picked out. Interwiki language links & population of infoboxes were early & obvious applications of the platform conceived ... difficult to unbundle these as drivers from the (easy in retrospect) sheer obviousness of the utility of a centralised facts store. (And there's a sense in which WD follows on from Commons as central store for images ... serving all wikimedia platforms from a single source didn't begin with WD.) I'll point you at Scholia as a source of literature on WD. hth. --Tagishsimon (talk) 18:54, 9 March 2022 (UTC)
The idea for the project dates back to at least September 2004 when the domain was registered and the name "Wikidata" was first mentioned. There are some related postings by Erik Möller on the wikimedia-l mailing list (e.g. [9]) and related early project drafts at metawiki [10] that do already contain a motivation for the project. As much as I know, the technology for Wikidata was not really there back then, and the idea really gained support no earlier than 2010/2011. —MisterSynergy (talk) 19:39, 9 March 2022 (UTC)

Wikipedia article List of Deputies from Alpes-de-Haute-Provence

Hi,

Could someone please link "List of Deputies from Hautes-Alpes" on the English version of Wikipedia to "Liste des députés des Hautes-Alpes" on the French version of Wikipedia? It won't let me proceed. Thanks!

✓ Done -- Reise Reise (talk) 16:11, 9 March 2022 (UTC)

commonscat linking - mina luna

hello; i created a commons category for the photos (uploaded by others) of Mina Luna Vincent. you DO have a wikidata -thing for her as Mina Luna (with no commonscat linked as far as i can tell). BUT i cannot find any way to add my brand new commons cat to it? i have now spent more time trying to do this, than i spent creating the commonscat in the first place! i have had enough, i bloody give up; this particular function (adding a commonscat) on wd is FUBAR. somebody else can connect it. also; it would be lovely if somebody here made adding commonscats easier for ppl who are non-experts @ wikidata... -__- Lx 121 (talk) 16:39, 9 March 2022 (UTC)

@Lx 121: This seems to work: WD diff and commons diff. --Tagishsimon (talk) 16:59, 9 March 2022 (UTC)

URL encoding snafu

on Q492362#P10431 WD is URL encoding as + in the hyperlink it constructs with P10431's value, where the destination website http://finds.org.uk is holding out for the unescaped

So WD is serving https://finds.org.uk/datalabs/terminology/object/term/AMMUNITION+POUCH where the website wants https://finds.org.uk/datalabs/terminology/object/term/AMMUNITION POUCH in a situation in which the formatter URL provides https://finds.org.uk/datalabs/terminology/object/term/ and the P10431 stored value is AMMUNITION POUCH

Inconvenient. Is there a solution? --Tagishsimon (talk) 16:19, 9 March 2022 (UTC)

ArthurPSmith has a tool that takes care of such. See Property:P4342#P1630 for an example of how it's used. Infrastruktur (talk) 18:04, 9 March 2022 (UTC)
Thanks Infrastruktur. --Tagishsimon (talk) 18:10, 9 March 2022 (UTC)
@Tagishsimon: I've added a preferred formatter URL that should work; however it will take a day or so to take effect, as usual for formatter URL changes. ArthurPSmith (talk) 18:37, 9 March 2022 (UTC)
Much obliged, @ArthurPSmith:. Thank you. --Tagishsimon (talk) 09:43, 10 March 2022 (UTC)

The Call for Feedback: Board of Trustees elections is now closed

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

The Call for Feedback: Board of Trustees elections is now closed. This Call ran from 10 January and closed on 16 February 2022. The Call focused on three key questions and received broad discussion on Meta-wiki, during meetings with affiliates, and in various community conversations. The community and affiliates provided many proposals and discussion points. The reports are on Meta-wiki.

This information will be shared with the Board of Trustees and Elections Committee so they can make informed decisions about the upcoming Board of Trustees election. The Board of Trustees will then follow with an announcement after they have discussed the information.

Thank you to everyone who participated in the Call for Feedback to help improve Board election processes.

Best,

Movement Strategy and Governance
--YKo (WMF) (talk) 03:41, 10 March 2022 (UTC)

Two "Wellmann" entries

When I tried to link the English and German Wikipedia articles on the German surname "Wellmann" through this page, I got an error alerting that there is an existing entry for the English Wikipedia article here. Is there a procedure that has to be followed in order to merge the two Wikidata entries? MarqFJA87 (talk) 09:29, 10 March 2022 (UTC)

@MarqFJA87: The two data objects Wellmann (Q13360766) and Wellmann (Q56246205) cannot be merged because they describe different entities. While the data object for the English-language Wikipedia article describes the family name Wellmann, the data object for the German-language, Low German-language and Chinese-language Wikipedia articles describes a Wikimedia disambiguation page. A surname is not a Wikimedia disambiguation page, which is why the two data objects cannot be merged. --Gymnicus (talk) 09:35, 10 March 2022 (UTC)

Unofficial/official when referring to Discord servers

There seems to be a number of items that are used as values of has characteristic (P1552) in qualifiers of Discord server numeric ID (P9345): verified Discord server (Q105192287), Discord Partner (Q105197552), verified and partnered Discord server (Q105222072) and unofficial Discord server (Q105222350). However, unofficial Discord server (Q105222350) seems to be ambiguous: the relationship with the other possible values implies that it is for servers that are not considered official according to Discord, but it doesn't seem to take into account whether the server is considered official according to the subject or operator of the server.

For example, on the item Among Us (Q96417649), there are five Discord servers linked, but for some of them I can't tell from the information present whether they are considered official according to Innersloth, or just random unaffiliated Among Us servers. There are "references", but they only contain the date, without any other information (which doesn't seem very appropriate).

Would it make sense to create additional items to avoid ambiguity? My assumption is that this property should only be used where the server is actually official according to the subject, but maybe that's not the case across the board. Overcast07 (talk) 21:28, 6 March 2022 (UTC)

I think we should probably only link official ones. If an unofficial server is notable we can make its own item and link it. BrokenSegue (talk) 00:40, 7 March 2022 (UTC)
@BrokenSegue: Do you mean official according to Discord or official according to the subject of each server? Overcast07 (talk) 15:37, 10 March 2022 (UTC)

Incorrect "public domain" for Q30880804

The Bars on the outskirts has an image which is said to be in public domain because

This work is in the public domain in its country of origin and other countries and areas where the copyright term is the author's life plus 70 years or fewer.

See wikimedia

This is wrong since the author died in 1982  – The preceding unsigned comment was added by Laurent.Claessens (talk • contribs) at 14:27, March 10, 2022 (UTC).

@Laurent.Claessens: Well spotted. It turns out the image on Commons - commons:File:Artgate Fondazione Cariplo - Guzzi Beppe, Baracconi in periferia.jpg - has an OTRS ticket on it indicating the copyright holder has CC licenced the image. Commons also had a PD-art template, which I've removed ... this is where Commons' life 70 notion came from for this image. So I think both the item and the Commons image record are good now. --Tagishsimon (talk) 18:04, 10 March 2022 (UTC)

"no source" as a reason of deprecation

Why is there no such reason? The main objective of any Wikimedia (ok, almost any) project is to give sourced data and nevertheless we have no (or is there?) way to indicate that a statement should be challenged because there's no source proving it. I would suggest citation needed (Q3544030) for such a role. --Infovarius (talk) 23:33, 6 March 2022 (UTC)

I believe unsourced information can just be deleted and don't have to be deprecated. BrokenSegue (talk) 00:22, 7 March 2022 (UTC)
With all due respect, a huge Fuck no to this. There's a spectrum. There are metric shit tons of unsourced statements on Wikidata that can be sourced, and often are verified in various external identifiers, but it has been historically (and still is) often prohibitively time consuming and arduous to create and/or and add the neessary references (Help:Sources requires up to 10 additional steps to create a book item just to verify the statement "Fred Smith died in 1869"). Citation-wise, Wikidata is probably fairly close to a flaming bag of feces, even if the vast majority of presently unverified statements are verifiable. A single user with a bot can easily import hundreds of thousands of items from some database of minor planets or molecular compounds or human genealogy. If they forget to tell their bot to add stated in (P248) to every statement, we should not throw all that data into the garbage toilet. Relatively few human beings have the time and energy to create and add the requisite citations to every statement, which is why the 'powers that be' at Wikidata should be constantly trying to make more tools available to automatically generate, add, and edit references. I say, if you can find a source, just fucking add it. If, after a good faith, valiant effort, you can find no credible evidence of the a given statement, deletion or correction might be considered (like if an unreferenced statement said the 4th child of Fred Smith was 'Jennie Smith', but credible sources point to a Jeannie Smith, just fix it). Keep in mind that like almost every Wikimedia project, Wikidata is largely populated and curated by untrained unpaid amateurs volunteering on their spare time, doing the best that they can. Be a wrench not an axe. Tighten and tweak data, don't destroy it. Infovarius Help:Deprecation describes how data should be deprecated rather than removed, including cannot be confirmed by other sources (Q25895909). Deprecation as I understand it functionally hides data from many external sources, even if it's true. Deletion is even more destructive. -Animalparty (talk) 05:12, 7 March 2022 (UTC)
I would note that the very popular drag and drop gadget only adds imported from Wikimedia project for a few select languages (why is that?). It seems extremely dramatic to delete any and all claims without a reference, regardless of whether the claims made are controversial or not. Moebeus (talk) 03:31, 9 March 2022 (UTC)
Other Wikimedia project is not a valid source for any statement, it's just a temporary measure to indicate where it was imported from. Wostr (talk) 09:40, 9 March 2022 (UTC)
I think not been able to confirm this claim (Q21655367) can be used as a reason for deprecation in these cases. - Valentina.Anitnelav (talk) 10:48, 7 March 2022 (UTC)
  • More and more bots are adding sources, but they are computationally intense, and take time to complete their tasks. They use the external identifiers as their sources. I am sure if you look at your watchlist you will see one in action every day. They do a much better job than humans, they add the date that the information was last checked, and provide a link to the source, give them time. --RAN (talk) 15:06, 7 March 2022 (UTC)
  • We recently had an attempted purge of ethnicity and religion data that was unsourced. The deletion started and was stopped, and the data was added back. It takes more time to properly source, than to delete, but having no information serves no one. --RAN (talk) 21:13, 7 March 2022 (UTC)
    Yeah. Stopping and reverting those deletions set a dangerous precedent and was a bad idea. --Emu (talk) 09:04, 9 March 2022 (UTC)
    No information is better than unsourced information. However, the requirement to add sources should be introduced for any new data, old data should be gradually sourced rather than deleted on the only basis of being unsourced. Wostr (talk) 09:40, 9 March 2022 (UTC)
    P91 was also stripped off of items for a lack of sources without any warning, instead of adding easily findable sources or asking Wikimedia LGBT to help out. The data was never added back. So instead volunteers have had to spend time trying to add them back (with sources to prevent it happening again) for the last 2 years. This stripping of P91 has impacted usage farther on down the line and has even been noted offwiki when someone who is out noticed that they had been "shoved back in the closet". Yupik (talk) 08:32, 11 March 2022 (UTC)

BLAKE3 (Q81575705) is Creative Commons CC0 License (Q6938433) but Creative Commons CC0 License (Q6938433) constraints complain that the item needs to be an instance of or a subclass of something

...so I add it as a "work"(loosely described as "physical or virtual object made by humans") and BLAKE3 (Q81575705) definitely fits that definition, unless too broad. I'd say that would be appropriate for SHA-3 (Q1190947) as well, including all checksum (Q218341) algorithms. Do the constraints need fixing for Creative Commons CC0 License (Q6938433)? Maskingself (talk) 11:02, 11 March 2022 (UTC)

@Maskingself: I think this item is confusing the algorithm with a piece of software that implements the algorithm. BLAKE3 is an algorithm. I'm guessing you are saying some implementation of it is CC0. You may want to make another item for the implementation and have it be an instance of software (Q7397) and attach the copyright to that. BrokenSegue (talk) 15:18, 11 March 2022 (UTC)
I just did that myself. see BLAKE3 (Q111191355). BrokenSegue (talk) 15:26, 11 March 2022 (UTC)
@BrokenSegue Thank you for your work. I now have a better understanding of what "implementation of (P4428)" is and how it is used in practice. Maskingself (talk) 16:05, 11 March 2022 (UTC)

Claudio Tadeu Daniel-Ribeiro

I am Cláudio Tadeu Daniel-Ribeiro, a Brazilian physician and researcher and have had my name inserted in Wikipedia when data on the Brazilian National Academy of Medicine (Academia Nacional de Medicina, ANM), to which I am affiliated, was being edited. I would like to introduce more data concerning my carrier and activities and, to start, I would like to correct my name (which is Cláudio, instead of Claudio, as it is informed at the ANM page) and i am not succeeding in doing so. May I ask your help.  – The preceding unsigned comment was added by Cláudio Tadeu Daniel-Ribeiro (talk • contribs) at 13:56, 21 February 2022‎ (UTC).

Two "Christine D. Bacon" entries

Hello,
For Christine D. Bacon, in addition to Christine D. Bacon (Q79988125) we find Christine D. Bacon (Q36521704) ! at the moment there is only one article attached to Q79988125! It would be necessary, perhaps, to operate a merge.... which I could not do... It’s, of course, the same person. RuB (talk) 13:04, 12 March 2022 (UTC)

@RuB: Please make yourself familiar with Help:Merge#Gadget. Duplicate items on WD are very common & the merge process is straightforward. I've merged these two. --Tagishsimon (talk) 13:22, 12 March 2022 (UTC)
@Tagishsimon: Thank's a lot ;-) RuB (talk) 15:06, 12 March 2022 (UTC)

API

Is there an api-version of this question? Testkonto 42 (talk) 14:28, 13 March 2022 (UTC)

Yes, but while you could use the API, it's more common (as well as flexible) to use SPARQL to search for things.
MWAPI: https://www.wikidata.org/w/api.php?action=query&format=json&list=search&srsearch=haswbstatement:P5736=1497
SPARQL: https://query.wikidata.org/#SELECT ?item ?itemLabel WHERE { ?item wdt:P5736 "1497" SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". } } Infrastruktur (talk) 15:34, 13 March 2022 (UTC)
Takk, Infrastruktur! In this specific case, I think the first version fits my need better than the latter! Testkonto 42 (talk) 19:44, 13 March 2022 (UTC)

Universal Code of Conduct Enforcement guidelines ratification voting open from 7 to 21 March 2022

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

Hello everyone,

The ratification voting process for the revised enforcement guidelines of the Universal Code of Conduct (UCoC) is now open! Voting commenced on SecurePoll on 7 March 2022 and will conclude on 21 March 2022. Please read more on the voter information and eligibility details.

The Universal Code of Conduct (UCoC) provides a baseline of acceptable behavior for the entire movement. The revised enforcement guidelines were published 24 January 2022 as a proposed way to apply the policy across the movement. You can read more about the UCoC project.

You can also comment on Meta-wiki talk pages in any language. You may also contact the team by email: ucocproject(_AT_)wikimedia.org

Sincerely,

Movement Strategy and Governance

Wikimedia Foundation --YKo (WMF) (talk) 04:23, 7 March 2022 (UTC)

As a gentle reminder, the voting period is about halfway through, and will be closing about 7 days from now. YKo (WMF) (talk) 06:33, 14 March 2022 (UTC)

List of victims of the Russian invasion of Ukraine (Q110999040)

Here is a SPARQL-powered list of victims of the Russian invasion of Ukraine (Q110999040) : https://w.wiki/4wtu

Louperivois (talk) 04:02, 14 March 2022 (UTC)

Moving P2190 - C-SPAN Person ID forward from string to numeric IDs, what are the next steps?

It seems there is general consensus on the talk page for C-SPAN person ID (P2190) to move from string to numeric IDs and C-SPAN organization ID (P4725) already uses numeric form. I'm sitting on the resolved string to numeric matches (TSV) to make the move quick once the mechanism to switch is clarified. Would deprecating the existing property string value and adding the numeric as the new property numeric value the preferred approach? Is an existing bot or tool to deprecate the old values and add new ones or should quick statements be used? Please chime in on the property talk page to clarify a process for changing the Identifiers. Wolfgang8741 (talk) 16:41, 7 March 2022 (UTC)

Using templates in talk page headers breaks navigation from history and watchlist pages. --Tagishsimon (talk) 16:57, 7 March 2022 (UTC)
Fixed, and thank you for the information though no warning is issued to users. This seems more a technical limitation than an issue users should need to avoid. The template could be converted to static string on save with the template value to avoid the issue for those unaware. Now that knowledge has been shared with me... I'll avoid templates in headers. Wolfgang8741 (talk) 19:01, 7 March 2022 (UTC)

Haven't heard anything further on the property talk page, I'm uploading the matched numeric IDs to P2190. Once the two Quick statement batches are complete (mentioned on the property talk page), what is the suggested way to deprecate all the string IDs for the property? I haven't deprecated at scale before. Wolfgang8741 (talk) 15:01, 14 March 2022 (UTC)

Wiki Loves Folklore 2022 ends tomorrow

International photographic contest Wiki Loves Folklore 2022 ends on 15th March 2022 23:59:59 UTC. This is the last chance of the year to upload images about local folk culture, festival, cuisine, costume, folklore etc on Wikimedia Commons. Watch out our social media handles for regular updates and declaration of Winners.

(Facebook , Twitter , Instagram)

The writing competition Feminism and Folklore will run till 31st of March 2022 23:59:59 UTC. Write about your local folk tradition, women, folk festivals, folk dances, folk music, folk activities, folk games, folk cuisine, folk wear, folklore, and tradition, including ballads, folktales, fairy tales, legends, traditional song and dance, folk plays, games, seasonal events, calendar customs, folk arts, folk religion, mythology etc. on your local Wikipedia. Check if your local Wikipedia is participating

A special competition called Wiki Loves Falles is organised in Spain and the world during 15th March 2022 till 15th April 2022 to document local folk culture and Falles in Valencia, Spain. Learn more about it on Catalan Wikipedia project page.

We look forward for your immense co-operation.

Thanks Wiki Loves Folklore international Team MediaWiki message delivery (talk) 14:40, 14 March 2022 (UTC)

Wikidata weekly summary #511

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Other Noteworthy Stuff
    • WDQS outage on 06 March: users may have unexpectedly had requests blocked. Incident report here.
  • Development
    • Getting ready for Data Reuse Days
    • Mismatch Finder: Discussing the next batches of potential mismatches with MusicBrainz data and some remaining Freebase data
    • Lexicographical data: Continuing work on the basic version of the new Special:NewLexeme page, focusing on putting in the base data about the new Lexeme
    • REST API: Continuing coding on the basic version of the GET Item endpoint

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.

Queries for metadata of the presentations of Wikidata Data Reuse Days 2022

As Wikidata Data Reuse Days 2022 is coming, I decided to create a Wikidata item for each presentation and structure the information that is presented in the description of each presentation in this page in Wikidata.

Now, we can query its data as shown in the queries below. The queries that I liked the most are Graph of all talks, edges are created through "follows", "followed by", "main subject" and "speaker" and Tree view of the subjects grouped by instance of and sorted by most discussed classes (thanks to this query I knew that an instance of W3C Recommendation (Q2661442) was one of the subjects of one of the presentations).

Queries

Additional information

I also planned to add the following properties to the items of the presentations, but I didn't have more time to add them.

Next steps

Once the recordings are uploaded to Commons. I think each presentation can be linked to the video at Commons by using the property video (P10) at Commons.

According to the description of some the presentations (see this one, the recordings are going to be uploaded to Wikimedia Deutschland’s Youtube channel, but if help is needed to upload them to Commons, let me know since I can help uploading them and also linking them to their corresponding Wikidata items of each of the presentations.

Rdrg109 (talk) 20:15, 13 March 2022 (UTC)

@Rdrg109 Given the recent discussions about the notability of Wikimedians, this endeavour could very well be perceived as an open provocation to parts of the community. I for one am very unhappy about a lot of new items that seem to fail WD:N. --Emu (talk) 21:08, 13 March 2022 (UTC)
I second the concerns. What if this task wasn't done for navel-gazing Wikimedia purposes, but by a SEO operative trying to increase information on their company? Every board member, press release, product, meeting and event given a Q-ID ("Joe Schmo", "hiring of Joe Schmo", "promotion of Joe Schmo", "2019 third-quarter sales presentation by Joe Schmo", "2022 Toyota Sale-a-Thon event at the Ogden Car Barn in Utah", etc. ad nauseum)? I think there would be significant pushback if such an operation were discovered, even if well-intentioned. -Animalparty (talk) 21:12, 14 March 2022 (UTC)

Scraping Javascript Databases

Is there a way to scrape a database like the Indian NGO Darpan (ngodarpan.gov.in), where every entry is only opened by javascript:void(0); with no extra URL per entry? I guess not, but perhaps you know a workaround? Newt713 (talk) 21:13, 13 March 2022 (UTC)

That javascript reads data from their API and displays it in the browser. You can investigate the requests with your browser's dev tools—e.g. Firefox or Chrome—and try to make similar requests to the API (be gentle, go slow). As long as there is no cookie stuff in the way, you will probably receive JSON objects so that you do not need to scrape anything from an HTML page. —MisterSynergy (talk) 21:25, 13 March 2022 (UTC)
Thanks. But it looks like there is a server generated one time identifier I can't get around. --Newt713 (talk) 07:09, 14 March 2022 (UTC)
@Newt713: I might be able to help, I've done scraping, but the website seems to be down. Germartin1 (talk) 07:24, 14 March 2022 (UTC)

You first need to retrieve a csrf token in a separate get request, in order to request the payload with a post request. There is also a session cookie required. Here is a very simply demo Python script which works for me:

demo Python script 
import requests


API_PREFIX = 'https://ngodarpan.gov.in/index.php/ajaxcontroller/'


def get_csrf_token(session:requests.Session) -> str:
    response = session.get(
        url=f'{API_PREFIX}get_csrf'
    )

    payload = response.json()

    return payload.get('csrf_token')


def get_profile(session:requests.Session, external_id:str, csrf_token:str) -> dict:
    response = session.post(
        url=f'{API_PREFIX}show_ngo_info',
        data={
            'id' : external_id,
            'csrf_test_name' : csrf_token
        },
        headers={
            'X-Requested-With' : 'XMLHttpRequest'
        }
    )

    payload = response.json()

    return payload


def main() -> None:
    session = requests.Session()

    external_id = '20398'
    csrf_token = get_csrf_token(session)

    data = get_profile(session, external_id, csrf_token)
    print(data)


if __name__=='__main__':
    main()

—MisterSynergy (talk) 08:39, 14 March 2022 (UTC)

Thank you. On a single entry it works, and good to know this is possible. But I'm not at that level of scripting :) But Germartin1 found a set on GitHub I'm working with now. All the best --Newt713 (talk) 17:53, 14 March 2022 (UTC)

Updating a URL formatter

I have just updated the URL formatter for Royal Museums Greenwich artwork ID (P9131) to one which now works.

Could somebody remind me how to get all usages on Wikidata and Commons to update? Jheald (talk) 17:36, 14 March 2022 (UTC)

@Jheald: Seems to be a PyWiki thing here from Mr. G: https://github.com/generalist/wikidata-misc but note 24 hour assertion at Property_talk:P1630#Formatter_URL_cached --Tagishsimon (talk) 18:36, 14 March 2022 (UTC)
Thanks @Tagishsimon:, that looks exactly the thing. Any idea if/when usages on Commons (eg through template c:Template:Artwork) will update as well ? Jheald (talk) 18:49, 14 March 2022 (UTC)
@Jheald: Sadly, no. Nothing obviously useful at https://commons.wikimedia.org/wiki/Commons:Tools --Tagishsimon (talk) 21:16, 14 March 2022 (UTC)
@Jheald, Tagishsimon: I just tested purging a Commons page using the ID (File:Jan Porcellis - Dutch Vessels in a Strong Breeze.jpg) and it didn't seem to pick up the new one, so I would uneducatedly-guess "after a day the URL formatter cache expires, and then they'll start updating when the Commons page cache is updated" - which is to say whenever the Commons page is edited or otherwise purged. I have no idea if the normal job queue would do this but I guess not. Andrew Gray (talk) 21:25, 14 March 2022 (UTC)
(ec) @Tagishsimon, Andrew Gray: As you say, let's not panic, but wait and see the effect of updating wikidata first. AFAIK, the {{Artwork}} template tends to be pretty responsive when it comes to ordinary content changes on Wikidata, so it's quite possible it may just all work. Otherwise User:Jarekt may have some ideas. Jheald (talk) 21:27, 14 March 2022 (UTC)
Jheald This update did require manual change in the code. Thanks for alerting me. --Jarekt (talk) 02:18, 15 March 2022 (UTC)

Deleting of GND redirects – Right or wrong?

In the request for comment Handling of stored IDs after they've been deleted or redirected in the external database a consensus emerges that identifiers that are given a bad rank should also be kept. As the administrator Emu states here, the problem with the data object is that it is too general. As a result, it cannot be applied directly to more precise cases, so that the deletion of GND redirects, such as here, here and here, is considered okay by the administrators, because both users die with themselves Kolja21 and Epìdosis speak out for such deletions. I, on the other hand, am against such deletions and would like to know what the community thinks here. --Gymnicus (talk) 09:07, 14 March 2022 (UTC)

 Info More context: Wikidata:Administrators'_noticeboard#Please_make_a_decision --Emu (talk) 09:44, 14 March 2022 (UTC)
@Emu: To be honest, I see no reason for this link now. I asked the question quite neutrally here, so that this is more of an opportunity for voting. But if you consider this linking is necessary, then so be it. --Gymnicus (talk) 09:57, 14 March 2022 (UTC)
Gymnicus is blocked for "Removing content from pages". He wants to prove that he was treated unfair. We had a long discussion where I explained to him that there are cases in which the deleting of a wrong or outdated identifier (after correcting the given source like Commons, Wikipedia or Wikisource) makes sense. Otherwise we would need to check the wrong ID again and again.
--Kolja21 (talk) 13:25, 14 March 2022 (UTC)
The “long discussion” was about one or two comments where he explained to me that without the deletions the maintenance work would collapse. He does exactly the same thing here in the comment. Only point here is that the problem isn't coming from the GND redirects, but from the bot creating the maintenance pages. Because the KrBot2 cannot recognize ranks, which leads to incorrect messages on the maintenance pages. If this breaks the maintenance work, then you should change the bot's script so that it recognizes the ranks. Then there will be no more false reports. I would like to end my comment with a statement from MisterSynergy (source) about Tn's, which are also a thorn in Kolja21's side: “Eigentlich gehören sie nicht hier her, aber mit missbilligtem Rang stören sie formell nicht – allerdings verstehen das nicht notwendigerweise alle Datennutzer.” (english: “They don't really belong here, but with deprecated rank they don't formally lower - though not necessarily all data users understand that.”) --Gymnicus (talk) 18:22, 14 March 2022 (UTC)
Data users like VIAF and others. And again: It's not only a problem for VIAF and bots. We need to check the IDs and the source given. We need to fix the errors here and in Commons, Wikipedia and Wikisource. It's a multi-stage maintenance procedure you have no idea of since you never participated in maintenance work. All you do is trolling around. --Kolja21 (talk) 18:37, 14 March 2022 (UTC)
MisterSynergy also gave you a tip (source), which referred directly to Wikisource, on the subject of Wikimedia Commons, Wikipedia and Wikisource: “Es ist leider nicht einfach möglich das Wikidata-seitig zu evaluieren, es müsste in Wikisource passieren.” (english: “Unfortunately, it is not easy to evaluate this on the Wikidata side, it would have to be done in Wikisource.”) – The solution on the Wikipedia, Wikisource and perhaps also Wikimedia Commons side are GND matching categories, like the category Category:GND different on Wikidata (Q55746867) which is already included in the German-language Wikipedia. Such a category can recognize the ranks, which can be seen, for example, in the data objects Wikipedia (Q52) and New York City (Q60). Although both data objects have several statements about the identifier, the German-language articles for the two data objects are not listed in the category Kategorie:Wikipedia:GND in Wikipedia weicht von GND in Wikidata ab because the category only compares the GND identifiers with the best ranking. This means, for example, for the German-language article Josef Goldschmidt (Bankier), which is listed in the said category: If in the associated data object Josef Goldschmidt (Q89200255) the statement Josef Goldschmidt (Q89200255)GND ID (P227)1174026251 is placed on the deprecated rank and the qualifier reason for deprecated rank (P2241)redirect (Q45403344) is added. Then the German-language article falls out of the category because then only the GND identifier 1066746923, which has the normal rank and is contained in the German-language Wikipedia article, is checked. Finally, it remains to be said: If you use the entire spectrum of maintenance options, then you do not need to delete GND redirects. --Gymnicus (talk) 08:42, 15 March 2022 (UTC)

How to merge M. L. J. Abercrombie (Q6712823) with Minnie Abercrombie (Q15994255) - both same person

I've tried using the Merge gadget to merge Q6712823 into Q15994255. They are both about the same person, just different aspects of her career. The final version should be Minnie Abercrombie (Q15994255) because there are a number of Wikipedia pages (different languages) with this name. I got an error message when I tried the merge: A conflict detected on enwiki: Q6712823 with enwiki:M. L. J. Abercrombie, Q15994255 with enwiki:Minnie Abercrombie

I tried both ticked and un-ticked answers to the statement: Always merge into the older entity (uncheck to merge into the "Merge with" entity)

How to merge the, please?--MerielGJones (talk) 12:50, 15 March 2022 (UTC)

OK, the merge has now been done by @Epìdosis:, so no problem now. Thanks. --MerielGJones (talk) 12:57, 15 March 2022 (UTC)

Request for property creation review

I am not sure if this is the right place to ask, let me know if I should take this request elsewhere. The properties our institution proposed seems ready for creation, I am wondering if a property creator can kindly spare some time look it over and see if there are any problems and perhaps help with the property creation? Any help is greatly appreciated.

Here are the links to the property: [[11]] [[12]] [[13]]

Hsuaniwu (talk) 02:24, 11 March 2022 (UTC)

I think there is just a backlog of items to create Category:Properties_ready_for_creation. Since it's a lot of work for the property creators, I think it's worth to wait a little longer, and it will happen. --Newt713 (talk) 08:36, 13 March 2022 (UTC)
Totally understandable and thanks for the clarification Hsuaniwu (talk) 01:46, 14 March 2022 (UTC)
What is your institution? It's undisclosed on your user page. SilentSpike (talk) 10:02, 13 March 2022 (UTC)
I have added institution to my page, sorry still not too familiar with how everything works yet Hsuaniwu (talk) 01:52, 14 March 2022 (UTC)
Ah thank you, no apology needed, was just curious since you mentioned "our institution". For reference, with regard to whether disclosure is required: my understanding is that the Wikimedia terms of service only require disclosure of any paid editing (usually done on userpage) where local wiki policy does not say otherwise (which is the case on Wikidata). However, generally it's probably good practice to disclose affiliations whether paid or unpaid anyway. SilentSpike (talk) 23:02, 15 March 2022 (UTC)

Updates of maps in the Wikidata pages

In the Wikidata page of Häagen-Dazs (Q1143333), can you remove the obsolete map File:Häagen-DazsMap.png, because the company left the Russian market? The updated SVG map has already been put there. The obsolete PNG map is superfluous.

In the Wikidata page of Domino's Pizza (Q839466), can you remove the obsolete map File:Domino's pizza world map.PNG and replace it with the updated map File:Domino's world map.svg, because the company left the Russian market?

31.200.20.40 12:00, 15 March 2022 (UTC)

Everything has already been done in relation to data object Häagen-Dazs (Q1143333). It is not necessary to delete the map which you call “obsolete”, because here in Wikidata statements that are no longer up-to-date are also retained. The statement regarding the no longer valid map was also marked with the qualifier end time (P582)2022. --Gymnicus (talk) 12:08, 15 March 2022 (UTC)
the Häagen Dazs map is wrong : it excludes Canada, and is dated 2012 !! see File:Häagen-Dazs_Map.svg Hsarrazin (talk) 12:56, 15 March 2022 (UTC)
I will ask the original uploader, who is active, to make the corrections and update the SVG map. Since the file name does not have a fixed date and is thus not static, it can be updated. 31.200.20.40 15:38, 15 March 2022 (UTC)
Regarding data object Domino's Pizza (Q839466), the updated card should be added with preferred rank. In addition, the no longer up-to-date card should be supplemented with the qualifier end time (P582)2022. --Gymnicus (talk) 12:10, 15 March 2022 (UTC)
Can you do the same in the Wikidata page of Domino's Pizza (Q839466)? 31.200.20.40 12:16, 15 March 2022 (UTC)

United World Wrestling ID

Hi there, I have a question regarding International Wrestling Database ID (P2727). It appears that United World Wrestling (UWW) changed their website and the latest data is now located at a different location. For example, for Irina Rîngaci (Q106603858) the latest data (which includes 2022 data) is here and the old location of the data is here (does not include 2022 data). This URL is also used in the Wikidata property and that now points to what appears to be a database with stale data. I think UWW updated their website because this page now uses the new location and not the old location anymore. Is there a process for updating International Wrestling Database ID (P2727) to use the URL which points to the new location, or is there anything else I should know if I were to make changes to that property? Thanks, Simeon (talk) 22:52, 15 March 2022 (UTC)

@Simeon: I updated the formatter URL (P1630) to point at the URL you listed above. BrokenSegue (talk) 22:58, 15 March 2022 (UTC)
Thanks! Should there perhaps also be corresponding changes to URL match pattern (P8966) and source website for the property (P1896)? In articles (on en.wiki at least) the external link now points to the new location so thanks for helping with that! Simeon (talk) 23:08, 15 March 2022 (UTC)
Yes, those should be updated in a similar manner. BrokenSegue (talk) 00:55, 16 March 2022 (UTC)
Ok, that makes sense. I have updated both accordingly with "preferred rank". Simeon (talk) 12:00, 16 March 2022 (UTC)

Leadership Development Working Group: Apply to join! (14 March to 10 April 2022)

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

Hello everyone,

Thank you to everyone who participated in the feedback period for the Leadership Development Working Group initiative. A summary of the feedback can be found on Meta-wiki. This feedback will be shared with the working group to inform their work. The application period to join the Working Group is now open and will close on April 10, 2022. Please review the information about the working group, share with community members who might be interested, and apply if you are interested.

Thank you,

From the Community Development team
--YKo (WMF) (talk) 05:36, 16 March 2022 (UTC)

János Piry Cirjék (Q1236709)

I have tried to link it to Commons (exactly the same name), I don't know what's wrong. It's bl... difficult to do that. --Io Herodotus (talk) 13:28, 16 March 2022 (UTC)

@Io Herodotus: Because you don't have to add “János Piry Cirjék” but “Category:János Piry Cirjék”. There is no gallery named “János Piry Cirjék” for him, but only a category named “Category:János Piry Cirjék”. --Gymnicus (talk) 13:47, 16 March 2022 (UTC)
If I add "category", the word category appears twice!
As I said, it's difficult Io Herodotus
(talk) 13:58, 16 March 2022 (UTC)

✓ Done thank you. --Io Herodotus (talk) 15:21, 16 March 2022 (UTC)

languages spoken, written or signed (P1412) is *not* an identifier

At the moment we have the following (incorrect) chain:

languages spoken, written or signed (P1412)
instance of (P31) = Wikidata property to indicate a language (Q18616084)
subclass of (P279) = Wikidata property for an identifier (Q19847637)

How best to fix this? P1412 should not be coming up in lists of identfiers. Jheald (talk) 15:05, 16 March 2022 (UTC)

I would remove the P279, there is only one identifier in the list. The rest are general properties. Sjoerd de Bruin (talk) 15:07, 16 March 2022 (UTC)
@Sjoerddebruin: Thanks! Jheald (talk) 17:03, 16 March 2022 (UTC)

Easily build useful websites with data from Wikidata and across the web

Wikidata has a ton of useful data about a wide variety of concepts. To really make that data shine you often want to present it in a way that is purpose-built for a domain or type of data, like a directory of video games or actors. We built a tool to make it easier to build such presentations, even for non-programmers, only using HTML and without any additional programming.

If you are intersted in learning how to build cool presentations of data from Wikidata and other sources of data on the web, please let us know when you are available by filling out this form.

Tarfahalrashed (talk) 19:51, 16 March 2022 (UTC)

Gatehouse Gazetteer covers more than just Wales

Gatehouse Gazetteer (Q59259501) refers to a gazetteer for castles etc in Wales, but Gatehouse Gazetteer place ID (P4141) more correctly indicates that it covers England, Wales, Isle of Man, Channel Islands etc, and includes the country name in the data entry. Should Gatehouse Gazetteer (Q59259501) be changed to indicate the broader scope, as there is not currently an English equivalent?

I have a dataset with the name/url mapping for all 5000 entries of Gatehouse, I could import it if given help, as I'm new to Wikidata. Vicarage (talk) 10:32, 16 March 2022 (UTC)

Renamed and added extra juristictions. Vicarage (talk) 09:15, 17 March 2022 (UTC)

The recent process of changing the identifier for C-SPAN person ID (P2190) from a string to a numeric ID which functionally in creating the URL is backwards compatible, but I've identified multiple users of templates for the property now added to Template:ExternalUse on P2190. The text for this template suggests notifying other projects (and in this case applicable affected templates), but there isn't documentation of how to overcome the language barrier to notify projects across multiple language Wikis and projects. As further integration of Wikidata grows this will be further compounded. Are there 1. existing mechanisms that could be added to notify all projects listed in the template:ExternalUse that would also enable translation or some standard templates to notify of a "discussion on the property" "deprecation of parameter" "merge of templates" or other events that warrants notification, discussion, and faciliates coordination? Wolfgang8741 (talk) 15:17, 17 March 2022 (UTC)

Number of Ukrainian refugees via Wikidata?

Hello, I would like to learn how to use Wikidata for the numbers of Ukrainian refugees by country. I started a conversation here. Would you like to help me? Ziko (talk) 11:35, 12 March 2022 (UTC)

I'll reiterate what I have written elsewhere here too. I would like us to get into the habit of storing the latest data point on Wikidata and all older ones in the Data namespace on Wikimedia Commons. The way we have for instance population (P1082) in countries and municipalities is a bit arbitrary regarding how granular data we store and not very nice when handling the items. Also, items like COVID-19 pandemic in the United States (Q83873577) where we have number of deaths (P1120) with daily updates quickly becomes unwieldy. For population, we have tabular population (P4179) but really, we should have some generic qualifier to point to a data file on commons for a property so that we don't need to create specific "tabular X" properties for all these kinds of properties (a substantial number of properties can have large amounts of time bound data). Ainali (talk) 11:06, 13 March 2022 (UTC)
Thanks for your insight, Ainali! Would you like to start on Commons and show us how it works? Ziko (talk) 15:40, 15 March 2022 (UTC)
@Ziko One earlier conversation regarding the topic of refugees in general. https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2017/02#refugee. There are many conversations, I can post to your talk page if you like if I find any other previous conversations...maybe best if I find some previous conversation that lead to more insight into the matter. Maskingself (talk) 12:01, 18 March 2022 (UTC)
No, I think I got it wrong. That conversation is about adding that a specific individual was a refugee. I'll continue the searching. Maskingself (talk) 12:25, 18 March 2022 (UTC)

Fort Nelson, both a museum and a fort

Fort Nelson (Q5471742) is a military museum in Hampshire, the latest use of Fort Nelson (Q17650048). Is there really merit in having 2 entries, with mapping clashes when I try to add wikipedia and other external pages, or should the former be given instances of fort and museum, and the latter deleted?

What is the policy for re-use of old buildings under the original name? Vicarage (talk) 13:44, 16 March 2022 (UTC)

Having two distinct items for building and institution is better, especially when you start to add properties like inception date, founded by, etc. Ayack (talk) 15:30, 16 March 2022 (UTC)
Should the editor reject adding the same Wikipedia page to both articles then? Why does it insist on a one-to-one mapping when the Wikipedia page says "after a long career as a military building, its now a museum"
Forts are often built and rebuilt on the same site, so would most usefully have the inception of each phase in its life, whether it be fortification, prison, then museum. Indeed I see I can add inception as a qualifier for the museum and fort properties of Fort Nelson (Q5471742)Vicarage (talk) 16:37, 16 March 2022 (UTC)
Wikidata is more granular than Wikipedia. Combining building/structure and organization (museum) into one item is not a good idea. The changing use of the building over the course of time can be recorded using has use (P366) and relevant qualifiers.Jklamo (talk) 09:07, 18 March 2022 (UTC)
Thanks. I'll keep updating the 2 items to reflect this and explore the issues that develop. Vicarage (talk) 11:53, 18 March 2022 (UTC)

Oneway only for a train station

Hello, I wonder how to model the fact that Gulleråsen (Q11973210) collects passengers only in one way, not the other way. Any idea? Bouzinac 💬●✒️●💛 08:21, 18 March 2022 (UTC)

Maybe about adjacent station (P197). If it only shows one adjacent station in one direction, then that shows you can only board in one direction. --Gymnicus (talk) 08:32, 18 March 2022 (UTC)
That's problematical, defining P197, wrongly, as being valid only if passengers can board at this station to alight at the adjacent station. In the situation in which passengers can only alight at this station, the adjacent station still exists as an adjacent station. I did, briefly, think about qualifiers for P197 statements, but again that seems wrong b/c the issue is with this station, not the adjacent station. --Tagishsimon (talk) 08:41, 18 March 2022 (UTC)
@Tagishsimon: That you can only get in or out is a different case. In the example of Bouzinac the point is that the traffic line only goes in one direction and not in both, at least that's how I understand it and that's how it is represented in the data object Gulleråsen (Q11973210) with the Qualifikator has characteristic (P1552)one-way traffic (Q786886). This can be represented using the property adjacent station (P197). --Gymnicus (talk) 09:47, 18 March 2022 (UTC)
Hmm. I wonder. Half tempted to agree, half not. We'd be using P197 on Station1 to point to Station2 b/c passengers can board at 1 and alight at 2. We'd omit a P197 from Station1 on Station2 b/c passengers cannot board at 2 and alight at 1. Station1 remains adjacent to Station2. The Station2 record now has no indicator on it of the adjacent station from which passengers can arrive. It would seem to be a 'where can I go to when boarding'-centric approach which disregards the 'from where might I have come when alighting'-centric approach. One station is adjacent from the perspective of boarding. The other station is adjacent from the perspective of alighting. --Tagishsimon (talk) 09:57, 18 March 2022 (UTC)

I asked the question because, according to that map and to OSM trains go both ways but only stop/pick passengers in the direction of Frogneseteren. Hence, this actual template would give the current result :

North / westbound Gulleråsen South / eastbound
Vettakollen
towards Frognerseteren 		Oslo Metro Line 1 Gråkammen
edit wikidata

and the general Oslo map would render as such Map of Oslo Metro (query)

Do you guys know of other subways which serve passengers only in one way ? --Bouzinac 💬●✒️●💛 10:12, 18 March 2022 (UTC)

@Bouzinac: So directly I can't think of any other example in relation to subways. On the other hand, I know from my own experience that something like this definitely exists in tram and bus transport. In addition, I also came up with a kind of line management that provokes such cases. These are ring lines (circle route (Q145179)). It is usually the case that the clockwise line has a different name than the anti-clockwise line. Examples of this are the S41 and S42 in the Berlin S-Bahn network or line 63 or 61 of the Budapest tram. The Glasgow Subway is also a ring line. On the outer ring track, orange is used to drive clockwise, while on the inner ring track, gray is used to drive counter-clockwise. --Gymnicus (talk) 11:10, 18 March 2022 (UTC)
Circular lines are OK and as far as I know, their model looks OK, see for instance Map of Glasgow Subway (query). Besides, there are many specific/peculiar subways, as is the case for New York and Séoul, a nightmare to model. For that station X, train may go to Y or Z depending on the time/the day. A solution was to model specific services, such as in
North / westbound Porte des Postes South / eastbound
CHU - Centre Oscar-Lambret
towards CHU - Eurasanté 		Lille Metro line 1 Wazemmes
towards Quatre Cantons - Grand Stade
Montebello
towards Saint-Philibert 		Lille Metro line 2 Porte d'Arras
towards CH Dron
Montebello
towards Lomme - Lambersart 		Q110395973 Porte d'Arras
towards Roubaix - Grand-Place
edit wikidata
but still New York and Séoul's subway model are unsatisfiying, actually. Bouzinac 💬●✒️●💛 11:42, 18 March 2022 (UTC)
@Bouzinac: Well, I knew from the start that the Glasgow Subway map would look good. There could have been an error in the template {{Adjacent stations}}. However, this is not the case because the outer and inner lanes were not implemented as different lines, despite their different color code, as can be seen, for example, at the stop Buchanan Street subway station (Q3269428). That's why we should rather look at the example of the S-Bahn Berlin (Berlin S-Bahn (Q99654)), because there you can't represent lines S41 and S42 as one line. That's why the template {{Adjacent stations}} gets certain difficulties as you can see at the example of the train station Berlin Westhafen station (Q465972):
North / westbound Berlin Westhafen station South / eastbound
Berlin Beusselstraße station
(terminus) 		S41 Berlin-Wedding station
Berlin Beusselstraße station S42 Berlin-Wedding station
towards Berlin Beusselstraße station
edit wikidata
I don't know if that's the right representation. Because that's how it looks, at least to me, as if line S41 starts and ends in the station Buchanan Street subway station (Q3269428) just like line S42. But that would also be the same for all other stations on these lines. What I'm saying is that the template has a few weaknesses. --Gymnicus (talk) 12:59, 18 March 2022 (UTC)
Yes, that template is really fit and should be fit for only subway stations, not railways stations as they are really more complicated to model : eg other pb with services departing station X to Y or Z depending on the train mission that are typically rare in subways which mostly go always A to B to C without jumpings, as in most RER Paris lines and probably other suburbans trains. I've tried to correct data on

Bahnhof Berlin Westhafen(Q465972). --Bouzinac 💬●✒️●💛 14:21, 18 March 2022 (UTC)

What should be fixed in Lexeme:L4127?

The Lexical category of Lexeme:L4127 is Q1084. I expected it to be displayed as 'noun', like in Lexeme:L6191. However, in Lexeme:L4127, it is displayed as 'Verb', which is confusing. What should be fixed? Intolerable situation (talk) 12:55, 18 March 2022 (UTC)

@Intolerable situation: For me, the lexeme hit (L4127) is also a noun and is also displayed as such. --Gymnicus (talk) 13:01, 18 March 2022 (UTC)
@Gymnicus: Then why is its lexical category displayed as 'Verb' (yet links to Q1084) to me? Intolerable situation (talk) 13:19, 18 March 2022 (UTC)
@Gymnicus: Oh, it is now displayed as 'noun'. Sorry. Intolerable situation (talk) 13:20, 18 March 2022 (UTC)
@Intolerable situation: All good that happens. You might have had both hit (L4126) and hit (L4127) open and then looked at the wrong one and wondered. --Gymnicus (talk) 13:24, 18 March 2022 (UTC)

Modelling Toys and Toy Lines

On my discussion page Trade pointed out that there are items conflating fictional characters and toy lines/brands/toys (esp. dolls) (link to Discussion: Dolls vs. fictional characters). We decided to split items into one item for the doll and one item for the fictional character. In the case of Barbie (Q167447) we now have Barbie (Q167447) for the brand and the newly created item Barbara Millicent Roberts (Q111242136) for the Barbie character but now there appear some other issues. I will take Barbie as an example, but I think these questions also concern other doll lines (like He-Man (Q550404)/Masters of the Universe (Q519710)).

  1. I think there still exists a conflation of concepts at Barbie (Q167447), namely between doll brand and Barbie doll type (model?). Do we need a third item for Barbie, having one item for Barbie as a brand, one item for Barbie as a doll model (besides Francie (Q5479798), Ken Willson (Q737939), Skipper (Q1995127)) and one item for Barbie as a fictional character? Oh, and we also have Barbie (Q98148278) for the Media Franchise.
  2. How to link Barbara Millicent Roberts (Q111242136) and Barbie (Q167447) (as a doll model)?
  3. How to link individual Barbie doll models (e.g. Francie (Q5479798), Ken Willson (Q737939), Skipper (Q1995127)) to the brand? Is brand (P1716) appropriate here? e.g. Francie (Q5479798)brand (P1716)Barbie (Q167447) (brand)
  4. How to link toy brand and toy line (e.g. Barbie (Q167447) and The Marvelous World of Shani (Q107675961)? Is brand (P1716) appropriate here? e.g. The Marvelous World of Shani (Q107675961)brand (P1716)Barbie (Q167447) (brand) (In this case I would remove the subclass of (P279) statement). Some sidenote: according to the English Wikipedia article en:The Marvelous World of Shani Shani dolls are only retroactively considered Barbie dolls. In this case - should The Marvelous World of Shani (Q107675961) be considered of the Barbie brand?
  5. How to link individual Barbie doll models to the toy line (e.g. Shani to The Marvelous World of Shani (Q107675961)). Should we use brand (P1716), too?

I'm not familiar with modelling of products, so thanks for every input. - Valentina.Anitnelav (talk) 10:58, 16 March 2022 (UTC)

Thanks for raising this!
For 1/ a third item seems reasonable
but that begs the question, do we need to also split Ken Willson (Q737939) in two?
For 2/ how about Barbara Millicent Roberts (Q111242136)based on (P144)Barbie (Q167447) ?
For 5/, not sure about brand (P1716), how about part of the series (P179)?
Jean-Fred (talk) 14:07, 16 March 2022 (UTC)
"but that begs the question, do we need to also split Ken Willson (Q737939) in two?" - Yes, in the current approach we also need to split Ken Willson (Q737939) in two (one item for the puppet model and one item for the character)
I like the idea to use part of the series (P179) to link to lines like The Marvelous World of Shani (Q107675961). I still have to think about using based on (P144) for 2) but it might work. - Valentina.Anitnelav (talk) 09:29, 17 March 2022 (UTC)

Barbara Millicent Roberts (Q109225717) is a thing now, courtesy to @Arlo Barnes, Valentina.Anitnelav, Jean-Frédéric:. Barbie the Doll have official biography published but Mattel while her appereance in media vary from work to work so it makes sense to consider the character group to be a seperate entity. --Trade (talk) 22:31, 16 March 2022 (UTC)

I made the 'Millicent' entity in order to improve Tefillin Barbie (Q16205089) (although then I became unsure how best to do so due to issues similar to the ones raised here). Since it's a kitbash, it depicts a variant of the Barbie character, but can't properly be said to be part of the Barbie line or brand but is said by the artist to be based on "Halloween Hip Barbie 2006", which is (do we think individual releases like that should have dedicated items?). I'm not too familiar with this particular fashion doll line despite it being the most famous one, so I'll leave questions of proper ontology to those who do; but right now I'll note that the two Rogers entities need more work to distinguish them, as they come off somewhat as duplicates. Arlo Barnes (talk) 23:43, 16 March 2022 (UTC)

Do you think we should merge the two items? @Arlo Barnes:--Trade (talk) 01:13, 17 March 2022 (UTC)
That seems best to me; if a distinction becomes important later, a new item can always be created. Arlo Barnes (talk) 01:40, 17 March 2022 (UTC)

@Infovarius: --Trade (talk) 16:29, 17 March 2022 (UTC)

What do you think about this? subclass of (P279)Barbie Doll (Q97593385) may be a bit redundant here as it already follows from instance of (P31)doll or action figure model (Q111282474) and brand (P1716)Barbie (Q167447) (I think).
Hm, maybe 'has quality'='gender expression (Q15404978) 'of' 'female'? Arlo Barnes (talk) 06:00, 19 March 2022 (UTC)
I like 'has quality'='gender expression (Q15404978) but I'm not sure about using 'of' as a qualifier here. Maybe somebody can think of a better one? - Valentina.Anitnelav (talk) 13:30, 19 March 2022 (UTC)

Another aspect of dolls is the moulding, with different individual doll models often sharing a mould. However, I'm uncertain where this info could be sourced from, and if it's actually useful to record in Wikidata. Arlo Barnes (talk) 06:00, 19 March 2022 (UTC)

field of work (P101)

There is a significant structural issue with field of work (P101): it is intended to be an inverse/reciprocal property and yet it does not function as one. Its property constraint classes are fundamentally flawed, because the current list suggests an ultimately open set of people, places, and things. It is extremely useful to indicate the academic discipline(s) of a scholarly work. Grouping by disciplines allows bibliographic data in wikidata to be more easily compared to library purchasing activity, book jobbers/vendors, and classification systems such as LC.

However, the property field of work (P101) is confusingly flipped with Works in this Field, a (proposed?) inverse property which exists as a data item but not a property. Works in the Field should appear on Agent (e.g., instance of = human) items, and Field of Work should appear on the Patient (e.g., instance of = journal article) items.

There is also the genre (P136) property that needs to be related to academic discipline, so that the scholarly output of faculties of creative writing, fine arts and performing arts are also accessible through academic discipline.

Academic Discipline is a distinct and broader property than subject. It tends to correspond with academic departments, programs, and the foci of academic societies. Disciplines are likely to be found as descendants of knowledge (Q9081), science (Q336), or scholarly method (Q17079481).  – The preceding unsigned comment was added by Joeeasterly (talk • contribs) at 18:14, March 18, 2022 (UTC).

yeah works shouldn't have field of work (P101). A work doesn't do work so it doesn't have a field of work. Shold use main subject or something. BrokenSegue (talk) 17:49, 18 March 2022 (UTC)

Join the Community Resilience and Sustainability Conversation Hour with Maggie Dennis

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

The Community Resilience and Sustainability team at the Wikimedia Foundation is hosting a conversation hour led by its Vice President Maggie Dennis.

Topics within scope for this call include Movement Strategy, Board Governance, Trust and Safety, the Universal Code of Conduct, Community Development, and Human Rights. Come with your questions and feedback, and let's talk! You can also send us your questions in advance.

The meeting will be on 24 March 2022 at 15:00 UTC (check your local time).

You can read details on Meta-wiki.--YKo (WMF) (talk) 09:48, 19 March 2022 (UTC)

Edits from Microsoft's Open Data Team

Hello, I work with Microsoft's Open Data Team. Our goal is to maintain and improve data accuracy for Bing Maps. We import Wikidata and occasionally when tracking down issues with map data we find the problem is sourced from Wikidata. Currently we don't have a mechanism for addressing these issues at the source.

Our Team would like to establish a pipeline where maps issues we find that are sourced from Wikidata result in edits/updates to those relevant Wikidata features. To establish this pipeline the Open Data Team would like to express interest in joining the Wikidata community as editors. This would entail training up 5-10 people on Wikidata policy regarding editing and sourcing edits. The resources available in the Help section will be instrumental in this training. Our interests are in transparency and in data accuracy.

I wanted to reach out to the community before we start editing to raise awareness around our interest in performing organized edits to Wikidata. I want to create a discussion here to field any concerns or questions.  – The preceding unsigned comment was added by Open Data Team (talk • contribs) at 21:58‎, 14 March 2022 (UTC).

Sounds v.promising. Welcome to wikidata. There's a great deal wrong with WD geodata, so lots to do. tbh, for me, the best pipeline is you fixing issues that need fixing, with good references and appropriate use of statement rank (e.g. to deprecate erroneous but referenced data, or to prefer a more accurate statement). Let us know what help you need. --Tagishsimon (talk) 22:32, 14 March 2022 (UTC)
Definitely sounds great. Personally, I've been looking at trying to bridge the microsoft and wikidata knowledge graphs through the Bing entity ID (P9885) property. We currently maintain mappings from wikidata to the google knowledge graph (see Google Knowledge Graph ID (P2671)). Unfortunately the Bing APIs are too slow/expensive for us to maintain the mapping to Bing (iirc would cost me thousands of dollars). It'd be cool if you could provide some support there. But in general sounds like an exciting effort and I hope we can be supportive of it. Might be wise to read this page on paid editing and the terms of use on disclosure to make sure you don't run afoul of the rules. BrokenSegue (talk) 00:32, 15 March 2022 (UTC)
There's also Help:Usernames and accounts ... WD is quite lassaiz faire, but your username might fall foul of en.wikipedia's rules, which in a nutshell says "The username should represent one person; do not use your organisation's name". --Tagishsimon (talk) 04:11, 15 March 2022 (UTC)
@Tagishsimon: Wikidata's policies are quite explicitely written to allow organizations to have accounts. While the username wouldn't be allowed in Wikipedia it's fine for Wikidata.
That said, it would still be preferred that if you train 5-10 people to edit Wikidata that everyone of those creates their own account and note that the person is working on the Microsoft's Open Data Team on their userpage. ChristianKl ❪✉❫ 20:03, 19 March 2022 (UTC)

Mathematical Notation in Aliasses or Labels

How is it possible in Wikidata to enter special characters to items. I tried to add new Items through using arXIV to QuickStatements and in these there are sometimes special characters. I usually try to add not yet such ones, when I do not trust the Label as proposed. Now I have not looked good enough before. An example is the item Spatially-resolved, substrate-induced rectification in C₆₀ bilayers on copper (Q111286372). There the 60 in C60 should look like C60. When I copy it it did not work. Is there a Wikitext editor extension for Wikidata items or is only plain text allowed. I have seen it in some items in alliases with a lower stepped number at other combinations.--Hogü-456 (talk) 19:44, 18 March 2022 (UTC)

I don't think any formatting is supported in labels and descriptions. This means you are restricted to what is allowed in plain Unicode, and there is a couple of explicit alphanumeric super- and subscript letters available. For instance you could type subscript 60 as U 2086 and U 2080 which will render as ₆₀. That said, I'd encourage not getting too creative in the label fields, there are some properties that allow Latex, such as defining formula (P2534) that is a better place for things that need a certain formatting, there's likely one for chemical formulas as well. Infrastruktur (talk) 00:06, 19 March 2022 (UTC)
Yes, approximate in plaintext as best you can in the label field (or if infeasible, put in something like [formula excised] in the title where the unacceptable bit is), duplicate that as a statement with title (P1476), and as a qualifier to that use title in LaTeX (P6835). I mocked it up for you (Hogü-456) in the example you gave, and you can also look at the property description for more examples (edit them without making changes / saving to see the LaTeX source). Arlo Barnes (talk) 06:10, 19 March 2022 (UTC)
Wikidata allows no formatting but does allow all of unicode. C₆₀ would be unicode way to express it. ChristianKl ❪✉❫ 21:00, 19 March 2022 (UTC)

Cause of death on a death certificate usually lists multiple morbitities

See: Sophia Weber (Q63973123) and File:Sophia Weber (1815-1891) death certificate.pdf where a death certificate lists two major causes and a contributing cause. Modern certificates have the same layout. I can switch dementia to "medical condition", and other ideas to resolve the error messages. --RAN (talk) 21:30, 19 March 2022 (UTC)

  • The suggestion to remove the single value constraint was made multiple times on the talk page without anyone arguing for keeping it. I just removed it. ChristianKl ❪✉❫ 23:09, 19 March 2022 (UTC)

What is the opposite concept of alternative fuel vehicle (Q8449254)?

I am only able to find a wikidata entry for alternative fuel vehicle, but unable to find wikidata entry for a concept for internal combustion engine vehicles or fossil fuel vehicles or vehicles with conventional energy. Given they are much more dominant, wikidata entry on such vehicle should already exists but I am unable to locate such entry, only able to find entity for vehicles of individual types of fossil fuels. Anyone able to locate what is the entry for conventional fossil fuel powered internal combustion engine vehicles? C933103 (talk) 09:30, 19 March 2022 (UTC)

The question : If we have a parallel hierarchy of engine, like

⟨ internal combustion engine ⟩ disjoint union of (P2738) View with SQID ⟨ list of values as qualifiers (Q23766486)  View with Reasonator View with SQID ⟩
list item (P11260) View with SQID ⟨ conventional fuel combustion engine ⟩
list item (P11260) View with SQID ⟨ alternative fuel combustion engine ⟩

maybe we don’t need to repeat it ? isn’t it a bit redundant ? author  TomT0m / talk page 17:43, 19 March 2022 (UTC)

    • I think such a statement should only added if you go through the subclasses of vehicle has no instance_of statements vehicle (Q42889) and find that the relationship really holds for all the items we have. Currently, I do think that there are rockets that don't have an internal combustion engine but that do use a patrolium based fuel and are thus no alternative fuel vehicles. There are also vehicles like bicyles that don't run on any fuel. ChristianKl ❪✉❫ 19:54, 19 March 2022 (UTC)
      @ChristianKl I did not mention the more generic class « vehicle » but only « internal combustion engine vehicle ». author  TomT0m / talk page 14:37, 20 March 2022 (UTC)


Why was my data deleted?

Why was my data deleted?  – The preceding unsigned comment was added by 911poyabt (talk • contribs).

Presuming any items you created (not "my data") have been deleted, that would likely be because they did not meet WD:N or else lacked statements and/or references which would have allowed WD:N to be evaluated - an issue we see with the item you created, Q111305489. --Tagishsimon (talk) 12:55, 20 March 2022 (UTC)

Manfredi

The item Manfredi (Q52746351) refers to 3 brothers.

The brothers Eustachio (1674-1739), Gabriele (1681-1761) and Eraclito (1682-1759) Manfredi were professors at the University of Bologna, of astronomy, mathematics and medicine-cum-geometry. Eustachio was the author of the famous Istituzioni astronomiche, published as the second volume of his works. The reference is "http://www.minorplanetcenter.net/db_search/show_object?object_id=13225"

Question: Should we have 3 items ?

If there is only one item, it is impossible to complete some datas, like birth, death etc.

--Io Herodotus (talk) 02:48, 20 March 2022 (UTC)

  • Q52746351 is an institute, you need to create an entry for each brother. You can then add each brother under "named for" at the institute. --RAN (talk) 17:40, 20 March 2022 (UTC)

No labels in Wikidata generic tree

The tool Wikidata generic tree doesn't show labels since a while. I was already writing Magnus but he didn't answer. I do not find a portal to report that issue. Can anyone help or does anyone have alternative tools? Bigbossfarin (talk) 17:46, 20 March 2022 (UTC)

Proposed config change: remove changetags right from users

Tracked in Phabricator
Task T303682
Resolved

Hi everyone! I’d like to propose a change to the wiki configuration related to change tags.

Background: MediaWiki has two rights related to change tags: applychangetags allows users to add tags (like OpenRefine [3.5] or Wikidata user interface) to their own edits and actions as they perform them, while changetags lets users add and remove tags after the fact, on their own edits and actions as well as others’. (A third right, managechangetags, controls who is allowed to create and delete change tags, and is limited to administrators.) By default, MediaWiki assigns both rights to all users; any registered user can add or remove arbitrary user-defined tags on anyone’s edits and actions. (Such tag changes are logged at Special:Log/tag.) Exempt from this are software-defined tags, which cannot be added or removed manually; this includes all the OAuth tags. So, for example, because OpenRefine [3.5] is a user-defined tag while quickstatements [2.0] is a software-defined tag, anyone can add the OpenRefine tag to a QuickStatements edit, but it’s not possible to remove the QuickStatements tag. You can probably imagine that this can result in confusion, and I personally find it weird that MediaWiki lets any registered user do this by default.

Proposal: Remove the changetags right (but not the applychangetags right!) from all users, and instead add it to other, more limited groups – maybe confirmed/autoconfirmed users, maybe even more limited than that. Some Wikimedia wikis have already made similar changes, namely Wikimedia Commons (T134196) and Meta (T283625), as well as English (T97013), French (T98629), Croatian (T270996), Russian (T136187), and Turkish Wikipedia (T264508).

Open questions:

  • Special:Log/tag is currently mostly populated by WE-Framework gadget – apparently this gadget (sometimes?) tags its edits after making them. This might break due to the config change, but looking at the source code, I think this might not actually be needed anymore – the extra action: 'tag' probably dates back to when the action: 'wbeditentity' API that actually made the edit had no tags parameter yet (it was added in T229917), and when the tags parameter was added to WEF’s action: 'wbeditentity' call, the separate action: 'tag' call wasn’t removed. @Vlsergey: can you check / confirm this, and maybe remove the call that is hopefully no longer needed?
    • Update: I’ve submitted a pull request to remove the unneeded (I think) API call.
      • Update: This was merged and deployed.
  • Which group(s) should the right be limited to? Autoconfirmed/confirmed users, i.e. most people? Admins and bots, like on the other wikis where it was removed from users? Something in between, like rollbackers?
    • Update: So far, it looks like we’re going for admins only (without bots).

If the community agrees on this change, and the open questions are resolved, the next step would be to open a Phabricator task – see meta:Requesting wiki configuration changes.

Thoughts? :) Lucas Werkmeister (talk) 13:25, 28 February 2022 (UTC)

 Support I think rollbackers (and approved bots?) might be the right group here for this. Maybe this should be an RFC though? ArthurPSmith (talk) 17:58, 28 February 2022 (UTC)
I don't think an RFC is necessary unless people object to it. It can be easily changed again if necessary. - Nikki (talk) 19:11, 28 February 2022 (UTC)
 Support Agree with above. Sjoerd de Bruin (talk) 19:07, 28 February 2022 (UTC)
 Support It sounds reasonable to me. - Nikki (talk) 19:15, 28 February 2022 (UTC)
  •  Oppose the added value of this request seems somewhat limited and the impact on improving data quality on Wikidata close to nil.
    Given that the developers are already much challenged with a backlog of handling requests from the community from 2019/2020/2021, it's better not to add more to the queue.
    There are other requests with clearly added-value and where the community has already invested much time and efforts in the analysis of the needs and the planning of the improvements. --- Jura 08:08, 1 March 2022 (UTC)
    This is a trivial change to the config, nothing Wikidata developers would have to deal with. --Matěj Suchánek (talk) 08:45, 1 March 2022 (UTC)
They can propose changes to be deployed, including during the regular backport windows (explained at the link I provided above), which take place every Monday to Thursday, as Matěj mentions below. --Lucas Werkmeister (talk) 18:56, 1 March 2022 (UTC)
    • Sorry, but this is silly. Are you suggesting the configuration should not be changed even if communities request that, because (it is claimed that) it takes (some) developers' time? So we will first wait for the major stuff to be done and then we can deal with this trivial request? Do you know that there is reserved time specifically for this kind of requests every Mon-Thu? --Matěj Suchánek (talk) 09:36, 1 March 2022 (UTC)
      It's just not a priority.
      At least we seem to agree that it requires Wikimedia Foundation developers to handle them.
      Also, as a change to Wikidata org, it would probably also require Wikidata product management to look into it, review and approve it.
      We have other configuration changes that are lingering as neither have time to deal with them. So we don't really want them to work on requests that only have limited or no added value and almost nil impact on improving data quality on Wikidata. --- Jura 09:45, 1 March 2022 (UTC)
it requires Wikimedia Foundation developers to handle them. The need for such changes is well known, and there is an established and well-functioning procedure for handling them. Note that changing user rights is specifically mentioned as one of the types of requests at meta:Requesting wiki configuration changes.
it would probably also require Wikidata product management to look into it, review and approve it. I’ve already had a brief discussion with Lydia about this change, and she’s okay with it if the community decides it.
We have other configuration changes that are lingering That’s not a valid reason for opposing this request, in my opinion. But if you tell me which other simple configuration changes you have in mind, I can take a look at them.
--Lucas Werkmeister (talk) 18:56, 1 March 2022 (UTC)
  •  Support I regularly have to come back and clean up tags added by new users [14]. --Matěj Suchánek (talk) 08:45, 1 March 2022 (UTC)
  •  Support, but please do not bundle this with any existing custom purpose group such as rollback as this would make things more difficult to govern. So, the three options in my opinion are: (1) bundle with autoconfirmed (i.e. make it available to all trusted users), or (2) bundle with admin (make it available by request), or (3) make a new custom group for this where users can request access based on some policy to be written. —MisterSynergy (talk) 10:46, 1 March 2022 (UTC)
I would lean towards (2) for now (so far it doesn’t sound like there is a need for (1)). (3) is a good option for the future if we need it, but I don’t think we need to start with it. --Lucas Werkmeister (talk) 18:56, 1 March 2022 (UTC)
The WE-Framework change was merged and deployed. I found another gadget in Special:Log/tag that adds tags separately, Infobox export gadget, but it turns out that only old versions of the gadget do this. I’ve left a notice about this at Help talk:Infobox export gadget § Stop tagging edits after they were made; in the meantime, since the old gadget versions are only used comparatively rarely (if you look at the recent changes, you’ll see way more InfoboxExport-tagged edits with no corresponding log entry, coming from wikis that use the latest version), I think we can still go ahead with this change. --Lucas Werkmeister (talk) 13:18, 13 March 2022 (UTC)

Wikidata weekly summary #512

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Discussions
    • New requests for permissions/Bot:
      • PodcastBot. Task/s: Upload new podcast episodes, extract: title, part of the series, has quality (explicit episode), full work available at (mp3), production code, apple podcast episode id, spotify episode ID. Regex extraction: talk show guest, recording date (from description)
      • AradglBot. Task/s: Create between 100,000 and 200,000 new lexemes in Aragonese language Q8765
    • Closed request for permissions/Bot: EnvlhBot 3 (approved). Task/s: add dictionaries IDs to French lexemes
  • Tool of the week
    • Linked People project let's you explore the family trees of all known people at Wikipedia/Wikidata.
    • Gene of the Day (gene-wordle) uses Wikidata for gene names and crafting an answer list by number of sitelinks.
  • Development
    • Lexicographical data: We're continuing with the work on the new Special:NewLexeme page. We worked on saving a valid new Lexeme with the new page. We are now focusing on the suggesters for language and lexical category so editors can select the right Item for them.
    • Data Reuse Days: We ran sessions on how to use Wikidata's data programmatically and the best practices around it. Slides and videos are available already (see above).
    • REST API: Continuing coding on the basic version of the GET Item endpoint. We have the very initial version of the get item endpoint ready and are now adding more parameters to it.

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.

Simple editor for Linux/python/command line ecosystem

I want to make a series of simple edits from the Linux command line, but all the suggested tools lead me down rabbitholes of npm, docker or mass copies of data between systems. Is there a tool with the simple syntax 

wikidata Q5470717 P3134 "2225973"

Which would add a statement without caveats, as I might do in the web page. I use python, but would much rather have a command line tool, than have to code up a function call. Obviously I'd have to log in and get some token stored. Vicarage (talk) 09:28, 20 March 2022 (UTC)

@Vicarage: Wikibase-cli is rather good. --Tagishsimon (talk) 12:58, 20 March 2022 (UTC)
I had high hopes of that until I saw it needed to be installed with npm or docker, which are whole ecosystems I know little about. I'd find java OK at a pinch, but python is my ecosystem of choice. Vicarage (talk) 15:07, 20 March 2022 (UTC)
@Vicarage: I also know pretty much nothing about the Node.js ecosystem, but npm itself should be easily installable via your package-manager-of-choice (https://linuxconfig.org/install-npm-on-linux), and thereafter installing/updating wikibase-cli is a single command-line action, and there's no need to know or care what the underlying code happens to have been written in. It's an amazing piece of software, and worth spending a couple of minutes checking whether it installs as trivially for you as it did for me.--Oravrattas (talk) 17:18, 21 March 2022 (UTC)
With Mint apt can't find it, and I can't face finding a repository. Quickstatments is working well enough for me. Vicarage (talk) 19:59, 21 March 2022 (UTC)
@Vicarage I have a proposal to make regarding investigating npm (Q7067518) on English Wikiversity (Q22808092), I will soon post my proposal on your talk page. To my knowledge npm has been considered[by whom?] "weak" in that packages in the system can be compromised or that not enough effort has been made to make the system "invulnerable"(no system can be made perfect when human error is a factor in security) to intrusion by attackers. If you want proof and think this is "spreading rumors" then I will attempt to back up my claim but these are mostly opinions by leaders or personnel at privacy software projects. They might not be even "credible" but considering that technological data on Wikidata is extremely lacking, I'd say hearing more opinions and viewpoints as long as they have something meaningful to say, is helpful. I have no complaints regarding Wikibase-cli, on the contrary, I'm prepared to "go deep", creating a new educational resource on English Wikiversity about it, but I won't do it by myself as I am interested in many topics already and I got my hands mostly full, but in my proposal it will be explained how this can be achieved, in small steps. Maskingself (talk) 22:30, 21 March 2022 (UTC)
You're almost describing Quickstatements (QS), which takes tabular data as input, so you can use either LibreOffice Calc or any commandline text editor you like to edit such files. Wikibase-CLI can do things QS can not do, but QS is nice and easy for uncomplicated and simple edits, and matches well with other unix tools such as awk. It is possible to use QS from the commandline, search the help page for API. Remember that you are responsible for any any edits made in bulk, so make a few test-edits and check them before making a whole lot. Infrastruktur (talk) 15:11, 20 March 2022 (UTC)

determination method or standard (P459) = Tor Browser (Q15397253)

I'm kind of a perfectionist(to the limits of my knowledge of what "perfect" is) and I try to make my references as accurate as possible but I'm attempting to not overdo it to prevent making my own life so painful that I'd rather not add/edit things anymore.

Is this going too deep? determination method or standard (P459) = Tor Browser (Q15397253) in the references field? I also got an error regarding Tor Browser (Q15397253)(constraint?). Should I use Tor (Q202044) instead or neither of the two? What I'm trying to say is: I fetched the price using a virtual private network service (Q56240391) and I also used Tor Browser (Q15397253) to verify that the price is the same using both methods and to say that in data in Wikidata.

I was fetching the price in Euro (Q4916) for the video game (Q7889): Grand Theft Auto V (Q17452) and I'd like to know if there's a consensus of how to add prices so it looks consistent across all or most of the uses specifically in regards to the video games category.

If I'd like to focus on "maintaining" the price (P2284) property for a few video game titles, does anyone have an opinion regarding how much or how little I should edit? If we take only the price of ie. Grand Theft Auto V (Q17452) with the currency Euro (Q4916) would it be excessive to update it every day(new price (P2284) added each day)? Would it be better once a week or once a month? Without pondering about it too much I'd say once a day allows one to track price changes 'pretty well' but will burden the database more. Does anyone have an opinion regarding this? I've asked many things now, hope this is a helpful post. Maskingself (talk) 22:10, 21 March 2022 (UTC)

@Maskingself: a few things. I don't think it makes sense to update the price for video games. video games are sold at thousands of retailers and the price is constantly varying at each of them. maybe if a game has an MSRP in various currencies at various points in time then we should list that when it changes. as for the reference content, I think this is way too much detail. just the URL and retrieval date are sufficient. BrokenSegue (talk) 23:21, 21 March 2022 (UTC)

Universal Code of Conduct Enforcement guidelines ratification voting is now closed

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

Greetings,

The ratification voting process for the revised enforcement guidelines of the Universal Code of Conduct (UCoC) came to a close on 21 March 2022. Over 2300 Wikimedians voted across different regions of our movement. Thank you to everyone who participated in this process! The scrutinizing group is now reviewing the vote for accuracy, so please allow up to two weeks for them to finish their work.

The final results from the voting process will be announced here, along with the relevant statistics and a summary of comments as soon as they are available. Please check out the voter information page to learn about the next steps. You can comment on the project talk page on Meta-wiki in any language. You may also contact the UCoC project team by email: ucocproject(_AT_)wikimedia.org

Best regards,

Movement Strategy and Governance
--YKo (WMF) (talk) 04:01, 22 March 2022 (UTC)

Articulated tram specification

I just created Q111328751. This is an articulated tram with 7 parts. There is also Q45856843 with 5 parts. Is there a way to specify the number of articulated section in tram types?Smiley.toerist (talk) 09:44, 22 March 2022 (UTC)

I am not aware of specific property, but you can use has part(s) of the class (P2670) tram/vehicle secion (seems that relevant item does not exist) quantity (P1114) qualifier. Jklamo (talk) 10:03, 22 March 2022 (UTC)

Please help: Question about the full list/look-up table for the properties in Wikidata base

Dear Sir/Madam,

I hope this message finds you well.

My name is Di Ma, an AI researcher in the UK. We have recently used some data in Wikidata base to conduct research in graph representation learning. For the node properties, we noticed that we have to manually use the website address (e.g., [[17]]) to get the description of the property. So we need to manually change the relation ID each time and access to the website to get the name of the property one-by-one. This would lead to a heavy workload when the number of properties is big (e.g., more than 500 properties).

Could you please kindly let us if the Wikidata has provided a website/full list directly showing all the names of all properties? If not, could you please give us some suggestions about how to get the property name more quickly? We are not sure if there is any tool or scripts that can be used to obtain this more efficiently.

Many thanks for your kind help!

I look forward to hearing from you.

Best regards,

Di Ma

@Di (Dee) Ma: a WDQS SPARQL report would seem to be the soluton, if I understand correctly - report, results. --Tagishsimon (talk) 19:54, 22 March 2022 (UTC)

"Expedia" hotels in Indonesia

A user has recently added a lot of hotels like Alam Sari Homestays (Q111332323). I doubt their notability, what do you think?--Môi âm hộ (talk) 20:36, 22 March 2022 (UTC)

WD:N suggests they very clearly are notable for wikidata: " a clearly identifiable conceptual or material entity. The entity must be notable, in the sense that it can be described using serious and publicly available references". --Tagishsimon (talk) 21:07, 22 March 2022 (UTC)
It says "homestay" but this is an actual hotel, not just someone renting out a room in their home. Hotels will be the subject of public info and in the case of the one you noted, quite a bit of info is already there. It looks good. Bluerasberry (talk) 21:31, 22 March 2022 (UTC)
They are notable via our rules and likely useful within wikivoyage. ChristianKl ❪✉❫ 22:33, 22 March 2022 (UTC)

Automatic Update from Connected Wiki

When moving a Wikipedia article from article namespace to user name space while suppressing the redirect, the Wikidata link will be automatically deleted (example: [18][19]). However when following the same process but moving the article to project namespace, the Wikidata link updates and needs to be deleted manually (example: [20][21][22]).

Is there a reason why the automatic update works differently with different name spaces? It would be nice if I didn't have to manually delete the Wikidata link every time I move an article to de:Wikipedia:Artikelwerkstatt (which is a draft space for articles written by LTA-sockpuppets). Johannnes89 (talk) 14:50, 22 March 2022 (UTC)

@Johannnes89: The difference is due to Wikidata notability rules; articles in User namespace do not confer notability on a Wikidata item, and sitelinks are not created. From Wikidata:Notability - "To be valid, a link must not be a talk page, page in MediaWiki namespace, special page, file, translations page, page in User or Draft namespace, page used by LiquidThreads (i. e. page in Thread and Summary namespace), page used by Structured Discussions (i.e., page in Topic namespace), subpage of Portal namespace, or any page that is intended for TemplateStyles (i. e. suffix that contain ".css" and/or ".js")." ArthurPSmith (talk) 15:26, 22 March 2022 (UTC)
Thanks for explaining, that makes sense! So I guess there is no other option but to remove the link manually. Johannnes89 (talk) 10:57, 23 March 2022 (UTC)
Unfortunately I'm not sure that will prevent somebody recreating the Wikidata item later. Can you not use the Draft namespace for this? ArthurPSmith (talk) 17:35, 23 March 2022 (UTC)
Draft namespace is not available at dewiki. --Ameisenigel (talk) 19:28, 23 March 2022 (UTC)

social media followers (P8687) was a terrible idea

I don't know the full history of how it came about, but from my current vantage, merging follower counts on tons of different platforms together into the single property social media followers (P8687) has been needlessly introducing hurdles to what would otherwise be a productive use of this project. Here's the story:

At English Wikipedia, editors expressed willingness to start updating subscriber counts of YouTubers via Wikidata. Yay! This is exactly what we want: taking an important, non-static, easily automatable piece of information, and centralizing it here, where it's now updated thanks to BrokenSegue's bot. The only step left is the seemingly trivial one of fetching the Wikidata information via w:Template:Wikidata and plugging it in. But the use of P8687 throws a giant wrench into that. It's no longer possible to simply fetch its value, since that could fetch e.g. someone's Twitter follower count instead of their YouTube follower count. So we have to check that YouTube channel ID (P2397) is defined as a qualifier before returning any info. The Wikidata template doesn't currently have any way to check just that a qualifier is defined, so despite Thayts' help, it hasn't been possible to get it to work yet.

Yes, it'd be nice if the Wikidata template were a little more flexible, but really, the root of this problem is the merging of follower counts here in a way that makes it extremely difficult to use the data. If we aren't able to overcome the hurdles this decision erected, we're going to lose the opportunity to use Wikidata prominently and usefully in thousands of Wikipedia infoboxes. Others hoping to source follower count information from Wikidata are surely having a similarly frustrating experience. Given all this, I hope we'll reconsider the choice to implement P8687. {{u|Sdkb}} talk 00:58, 23 March 2022 (UTC)

if I understand the situation correctly the only way to have been to make it work would've been to make a property specific to youtube subscribers. that seems needlessly specific to me. would things even work if we did the alternative (that is sometimes suggested) of making a separate item for youtube channels themselves? that said I have zero experience with accessing wikidata information from enwiki. BrokenSegue (talk) 02:06, 23 March 2022 (UTC)
I question the wisdom of having properties whose values can and do change from minute to minute, every single day. There should be periodic, regular, well-defined points in time to reference (compare to annual/decadal censuses for city/country populations), otherwise a poorly operated bot could conceivably add a new value every time someone adds or unfollows a celebrity on Twitter or YouTube, resulting in thousands of needlessly precise snapshots cluttering Wikidata items and slowing down load times (as Wikidatans are loath to remove outdated or redundant data). -Animalparty (talk) 02:30, 23 March 2022 (UTC)
@Animalparty, we considered this question at Wikidata:Requests for comment/Frequency of YouTube follower count data and came up with what I think is a reasonably balanced approach. {{u|Sdkb}} talk 04:24, 23 March 2022 (UTC)
That looks sensible. I hope it becomes familiar to more than the few who took part in the discussion. Like much of Wikidata, consensus and guidelines seem to exist in a million different dimly lit catacombs that can only be located by following an ancient map procured from a mountain-top hermit (after slaying a dragon and solving a riddle). Aside from User talk pages there are only 6 incoming links to that discussion, and none of them are social media followers (P8687), number of subscribers (P3744) or YouTube (Q866). As I am not a bot wrangler, I hope wiser folks than myself can find the resource should they need it. -Animalparty (talk) 04:46, 23 March 2022 (UTC)
Wikidata properties have no sitelinks so there's no way to link ie. a learning resource on English Wikiversity to the Wikidata property itself(if this is good design, I might not have needed to type this sentence).
https://en.wikiversity.org/wiki/Wikidata could be used as a ground for developing a good learning resource about Wikidata, investigating methods of finding consensus and finding current consensus, systemise. Though maybe it shouldn't be defined only by "learning resource", I'm kinda new to Wikiversity and I haven't grasped all the things it's about yet but weighing many different thoughts is one of them. Studying Wikidata from a Wikiversity perspective and also letting it be helpful to Wikidata editors might be a reasonable goal. ie. on the Wikidata learning resource we could add a section such as this:
  • == Properties where formal consensus has been reached == with a subsection for this property
    • === P8687 consensus === where maybe a link to the Wikidata request for comment page is enough.
continuing with other subsections for other properties. Then if you see a property in the future which you are unsure if there's been any consensus about then you can go to Wikiversity to learn about it. I'm kinda more of an interactive learner than a "static page" learner so I'd benefit from learning through interacting and if that is interesting to anybody perhaps we could add a "workshop" section to the learning resource. If not, then all above stands as suggestions. Maskingself (talk) 14:15, 23 March 2022 (UTC)
On making consensuses better known, if Wikidata had a version of w:Template:Please see, I'd put that on the relevant property/item talk pages. That'd need to be imported, though. {{u|Sdkb}} talk 16:20, 23 March 2022 (UTC)

Copy an item

I seldom make an item; just edit old ones. So, after making Commons category Commons:Category:Morris Park station (HR&PCRR) for my new photos I edited item Q110078278 to point to it. But, that's wrong as the Q item is for a future station and the commonscat is for an abandoned station a kilometer away, so we need a new item for the old station. When making new Commons categories or new ENWP stub articles, I can simply copy-paste an existing, similar one and make the changes that are necessary for the new article or category; that's how I made the commonscat for my photos. Is there a way to do that for a WD item, or must I create separately each statement in the new item? In the present case, there will be only small differences between the statements for the future station and the past one. Jim.henderson (talk) 12:37, 23 March 2022 (UTC)

See User:Magnus Manske/duplicate item.js.--Jklamo (talk) 12:53, 23 March 2022 (UTC)
I use moveClaim (Q110793966) for this — Martin (MSGJ · talk) 20:05, 23 March 2022 (UTC)

Template:Wikiversity Template:Wikiversity (Q7219662) doesn't have a Wikidata equivalent

Can I help in porting the English Wikipedia one of Template:Wikiversity (Q7219662) to Wikidata? What is needed to work on templates on Wikidata? Maskingself (talk) 12:58, 23 March 2022 (UTC)

✓ Done Template:Wikiversity — Martin (MSGJ · talk) 20:16, 23 March 2022 (UTC)

Correct address for a Wikidata competition

We, that is Wikimedia Austria, are planning a Wikidata competition in May. It will be part of the International Museum Day (WMCH with WMAT and WMDE) again, as it was in 2020 and in 2021. This year we would like to put the pages with all the information directly on Wikidata instead of Meta. So the question is: What will be the correct placement here. Would it be Wikidata:Events/Museum Day 2022 Wikidata Competition or is something else than Wikidata:Events more fitting? --Manfred Werner (WMAT) (talk) 16:44, 23 March 2022 (UTC)

Wikidata:Events looks right to me! Can we help to advertise it, e.g. on Wikidata:Main Page or in banner perhaps? — Martin (MSGJ · talk) 20:21, 23 March 2022 (UTC)
Thank you, MSGJ! The competition will be in May. Currently we are preparing all the pages of information which will take some more time (we are working with co-organizers basically from all over the world this time). Once we are ready to go public any help in making it seen and advertise it among the community will be very welcome! --Manfred Werner (WMAT) (talk) 21:26, 23 March 2022 (UTC)

Item to be deleted

Stanislaus Vasilevskis (Q111270184) has been created by mistake. The item already exists under (Q17593896). Sorry I don't know where to find the information to delete it. --Io Herodotus (talk) 04:56, 17 March 2022 (UTC)

@Io Herodotus: The easiest would be to merge both items. There is a HowTo an Help:Merge and a nice Tool in your Settings -> Gadgets. If there are any problems, do not hesitate to ask again. But since duplicates are quite common it's good, to be able to do it by yourself. --Newt713 (talk) 11:41, 17 March 2022 (UTC)
✓ Done Pas de problème. Déjà fait. --Kolja21 (talk) 13:48, 17 March 2022 (UTC)
I don't understand why we can't ask a "speedy deletion" like on wikipedia. --Io Herodotus (talk) 03:06, 18 March 2022 (UTC)
Our normal deletion process basically is the equivalent of enwiki's speedy. it's pretty light weight already. BrokenSegue (talk) 03:11, 18 March 2022 (UTC)
@Io Herodotus: Given our rules on Wikidata we don't delete any items in case of doublicate items, we instead merge doublicate items. This allows 3rd-parties outside of Wikidata who have links to an item to still find the correct item after the merge. Providing stable ID's is important in the world of structured data. While in it's unlikely that there's harm done in this particular case by deleting the item, we still have our general policy of how to deal with doublicate items. ChristianKl ❪✉❫ 11:21, 24 March 2022 (UTC)

Regular expression filter for a mediawiki wiki

I added "Unicon (UK)" as a property Fancyclopedia 3 ID (P9307) to Unicon (Q111328625) but it complained that I wasn't meeting the format regex \S{1,128}. I guess it was the space that upset it, but the mediawiki it points to handles spaces and and converts them to underscores.

Should a mediawiki reference always be coded with underscores, or should I change Fancyclopedia 3 ID (P9307) to accept spaces? Vicarage (talk) 13:41, 22 March 2022 (UTC)

  • Generally, the formatting is supposed to be stable. You should only change it if you have a good reason and believe that this won't damage data reusers. ChristianKl ❪✉❫ 12:51, 24 March 2022 (UTC)

electric household appliance (Q20076681)

electric household appliance (Q20076681) currently gets used as superclass for washing machine (Q124441). It's unclear to me how it differs from major appliance (Q12269769) and electric household appliance (Q20076681). Can someone who knows either Russian or Ukranian (where the item currently has labels) add an English label? ChristianKl ❪✉❫ 20:06, 19 March 2022 (UTC)

I tried electric appliance, though I am not 100% sure this is a good translation. Ymblanter (talk) 20:13, 19 March 2022 (UTC)
@Ymblanter: What's the difference with electrical appliance (Q2425052)?ChristianKl ❪✉❫ 23:06, 19 March 2022 (UTC)
electric household appliance (Q20076681) is household appliance (Hausgeraete), electrical appliance (Q2425052) can be anything electric.--Ymblanter (talk) 19:45, 20 March 2022 (UTC)
@Ymblanter: might "electric household appliance" then be a better label? ChristianKl ❪✉❫ 11:24, 24 March 2022 (UTC)
Yes, it is probably better Ymblanter (talk) 19:40, 24 March 2022 (UTC)

How to get the Q-ID from dewiki title?

In quickstatements V1 there was a Tool to convert dewiki pages in a list into their Q-IDs. For example: I need to convert dewiki:Bad (Album) into Q275422 in a list of many items. Bigbossfarin (talk) 14:38, 23 March 2022 (UTC)

@Bigbossfarin: You can use the PetScan tool to do this. Go to the "Other Soures"/"Andere Quellen" tab and paste the list of dewiki page titles (without interwiki prefix) where it says "Manual list"/"Manuelle Liste". Just below that enter "dewiki" in the "Wiki" option box. Then switch to the "Wikidata" tab and select the "Add items, where available"/"Füge Objekte an, falls verfügbar" option near the top. Finally, click the "Do it!"/"Los!" button at the bottom of the page. The QIDs will be listed in the rightmost column of the results.
I hope that this method works for you. Let me know if you have any problems.
--Quesotiotyo (talk) 22:45, 23 March 2022 (UTC)
Thank you very much :) Greetings Bigbossfarin (talk) 23:14, 23 March 2022 (UTC)
@Bigbossfarin: There's also a WDQS approach:
SELECT ?item ?itemLabel ?sitelink WHERE 
{
  VALUES ?sitelink {
    "Bad (Album)"@de
    "Appetite for Destruction"@de
    "In the Land of Grey and Pink"@de
  } 
  ?article schema:about ?item ;
    schema:isPartOf <https://de.wikipedia.org/> ; 
    schema:name ?sitelink .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],de,en". }
}
Try it!
--Tagishsimon (talk) 08:53, 24 March 2022 (UTC)
Out of curiosity (and nervousness;) - why would you want to convert the German article about "Bad" (the album) into "Bad" (the composition) ? If you want to go from the album to the song, wouldn't the logical way be to go via the recording: Bad - Wikidata found on the tracklist of the album? Moebeus (talk) 01:01, 25 March 2022 (UTC)

Do we include "Best paper awards" in the author's, the paper's, or both?

For instance the Frisch Medal (Q1320666) (best paper published in Econometrica (Q375835)). Do we list the 2018 award in The Economics of Density: Evidence From the Berlin Wall (Q57547095) (i.e. the paper) or in Gabriel M. Ahlfeldt (Q41802078), Stephen Redding (Q30066231), Daniel M. Sturm (Q41805459), and Nikolaus Wolf (Q41805060) (i.e. the authors)? Or both? --Bender235 (talk) 18:59, 24 March 2022 (UTC)

To partially answer my own question: I presume we use the for work (P1686) qualifier when included these "best paper" awards on author's entries. But what do we include on paper's entries, if anything? --Bender235 (talk) 02:41, 25 March 2022 (UTC)

Creat a new Item for Amirreza Alizadeh

Peace be upon you. I want to make a new item for Amirreza Alizadeh, but I can not. I had created a case with this subject before, but someone else ruined it over and over again. And now I can not recreate this. Please help me solve this problem. Thank you

Q111248779 has been deleted because it did not mean the notability policy (WD:N). There is no point in recreating a non-notable item. --Emu (talk) 21:59, 30 March 2022 (UTC)

Unfortunately, it is impossible to me to add interwiki links to the Shan Wikivoyage. Do you know what I have done wrong? --RolandUnger (talk) 05:41, 25 March 2022 (UTC)

See phab:T302799. - Nikki (talk) 08:20, 25 March 2022 (UTC)

Yomi (Q1142558)

This item includes two different concepts, one is the equivalent to hell/underworld in Chinese folklore and Taoism, another is the equivalent to hell/underworld in Japanese Shintoism. However, since both concepts use the same Chinese characters "黃泉", they got linked together into same item. I am not sure how to split this up, as despite most Wikipedia article on this wikidata item exclusively cover Japanese Shintoism, the Japanese Wikipedia and English Wikipedia article on this also have a short text covering the use of this term in Chinese folklore, as well as another section of the term being used in translation of Sheol (Q31780)/Hades in Christianity (Q3452850) in the Japanese language. On the other hand, Chinese Wikipedia and Swedish Wikipedia article currently linked under the wikidata entry is mostly about the concept of hell/underworld in Chinese folklore, with short description and link to Japanese Shintoism concept in respective Wikipedia, but the link on Chinese Wikipedia is a recursive redirect and the link on Swedish Wikipedia is a red link. Min Nan Wikipedia cover exclusively the Chinese use. English Wikipedia also have a disambiguation page Huang Quan, but it mistakenly directed readers to the page of Diyu (Q1899406) as "Chinese hell", despite it is Chinese Buddhist concept that is different from the Chinese concept of "Huang Quan". So, how should the Wikidata entry be split up? C933103 (talk) 12:27, 25 March 2022 (UTC)

Connection

Can someone interwiki for en:Forbes list of Russian billionaires 2017 (Q30887737) with German article de:Liste der reichsten Russen ? --188.96.230.188 18:55, 24 March 2022 (UTC)

It is not possible. The Wikidata item that you mention lists billionaires as of year 2017. However the German article contains the general list as of an unknown date. Michgrig (talk) 18:43, 25 March 2022 (UTC)

Fusions ?

Polydamas (Q1205276) and Poulydamas (Q91121087)

Polydamas statue base at Olympia (Q107043450) and Poulydamas statue at Olympia (Q107043451), maybe not, one is for the base only.

--Io Herodotus (talk) 14:44, 27 March 2022 (UTC)

@Io Herodotus: You name the same data object twice here. So there is confusion as to what you want. --Gymnicus (talk) 15:02, 27 March 2022 (UTC)
Sorry I got confused. --Io Herodotus (talk) 15:17, 27 March 2022 (UTC)

[SIGNIFICANT CHANGE] wbsearchentities returns new display section

Hello,

This is an announcement for a significant, but not breaking, change to the response format of the wbsearchentities API. So far, each search result has included "label" or "description" members- which included the label/description of the matched entity- so that it could be shown to users. However, these were plain strings with no indication of the language of the label or description: if language fallback happened (e.g. if the search was in Arabic, but the item only had an English description), then there would be no indication of this in the response. Users of the API couldn’t know what the language of the returned label/description was without a separate API call to retrieve the entity’s data. This is for example necessary for screen readers to read out the label/description in the correct language. (We will follow up with other changes to make use of it in the Wikibase UI, so that screen readers there are getting a better indication of which language to read out.)

We have resolved this with the addition of a new "display" member to the search result structure, which contains a "label" and "description" that are not plain strings, but rather objects with "language" and "value" members (like labels and descriptions in the normal JSON serialization). The old "label" and "description" fields are still there, but deprecated: we recommend that you don’t use them in new code. Please note that, just like the "label" and "description" were always optional (i.e. could be missing for entities that had no label or description available in the request language or a fallback language), so the "label" and "description" in the new response are also optional (under the same conditions).

This change was already deployed to Beta Wikidata, Test Wikidata, and Wikidata. If you have any questions or feedback, please feel free to let us know in this ticket.

Cheers, -Mohammed Sadat (WMDE) (talk) 08:08, 22 March 2022 (UTC)

We only realized after sending this announcement that it also has an impact on users of the wikibase.entityselector.search JS hook (which is a stable interface). That hook provides additional search results to the entity selector, and is used, for example, by WikibaseQualityConstraints to suggest properties or items based on constraint definitions. Currently, code using this hook will typically return objects that look like this:
{
  "id": "Q1",
  "label": "some label",
  "description": "some description",
  "rating": 1,
  "url": "https://www.wikidata.org/wiki/Q1"
}
The returned objects should now look like this instead:
{
  "id": "Q1",
  "display": {
    "label": { "language": "en", "value": "some label" },
    "description": { "language": "en", "value": "some description" }
  },
  "rating": 1,
  "url": "https://www.wikidata.org/wiki/Q1"
}
Based on this search, we believe that user scripts by the following users are affected: Nikki, Matěj Suchánek, Jon Harald Søby, Xaris333, Marsupium, Data Gamer, and Philocypros. (No gadgets appear to be affected.)
Please let us know if you have any questions. -Mohammed Sadat (WMDE) (talk) 15:42, 28 March 2022 (UTC)

How do I add a schema to an object?

Hi! I just created https://www.wikidata.org/wiki/Q111364686 for https://www.pando.com/ and want to add the schema organization , but I can't figure out how! Can you please help me? Thank you! LeslieCarr (talk) 16:03, 25 March 2022 (UTC)

Wikidata weekly summary #513

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
    • Scribe is a keyboard extension based on lexicographical data that can help users remember grammar rules (see blogpost above).
    • WorldlEH is a wordle clone in Basque.
  • Development
    • [Significant Change]: wbsearchentities changed to explicitly return display terms and matched term
    • Lexicographical data: Working on the lookup for language and lexical category and displaying potential errors during Lexeme creation
    • Improved the API response of the wbsearchentities endpoint by adding the language to the labels and descriptions in the API response (phab:T104344)
    • Data Reuse Days: Second and final week - organized, attended and held a few sessions incl. bug triage hour and pink pony session
    • REST API: Continuing work on getting the the data of an Item, we almost have filtering of the data returned by the API and basic error handling is in place. Next up: not returning the data if the client already has the most recent data, and authentication

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.

A new update on the Wikidata Query Service scaling process

Hi all! We have an important update to share with the community regarding the Wikidata Query Service scaling process: we closed our study phase and published a working paper, in which we short-listed and analyzed four potential candidates for replacing Blazegraph.

You can find more information about the full evaluation study process and results here. We would love to hear your thoughts on this on the discussion page. -- Sannita (WMF) (talk) 08:28, 29 March 2022 (UTC)

Wrap-up of the Data Reuse Days 2022

Hello all,

Thanks again to everyone who attended the Data Reuse Days and contributed to its success! We had a total of 45 sessions, with more than 60 different speakers, presenting a broad diversity of projects reusing Wikidata's data. Also a warm thank you to all speakers who presented, and people who helped take notes during the sessions!

If you missed the event, here are several ways to catch up: you can watch the recorded sessions, access the slides on Commons and the collaborative notes for all sessions. You can also access the individual links for a specific session directly from the schedule.

If you attended the event and contributed to something that could be useful for others (create or improve a tool, start a discussion, build a cool query...), please add it to the outcomes page.

As usual, I'm open to feedback, suggestions and ideas that would allow us to improve the upcoming events and try out new things. Feel free to reach out to me on my talk page or at lea.lacroix@wikimedia.de. Cheers, Lea Lacroix (WMDE) (talk) 08:31, 29 March 2022 (UTC)

subclass of (P279) statements on items that aren't classes

Take for example these statements I found:

From what I understand, these items should only be used as the value of occupation (P106) and genre (P136) claims, respectively – never as the value of an instance of (P31) claim. So then these things shouldn't be treated as classes, right? It doesn't make sense to talk about "a rock music". So these subclass statements seem erroneous to me, but they are everywhere, on all sorts of things. Here's more examples:

Are people just using P279 for lack of better property or something? Or am I just misunderstanding it? ―Jochem van Hees (talk) 01:30, 29 March 2022 (UTC)

I thnk you're misunderstanding. The occupation television presenter is a subclass of the occupation presenter; hard rock is a subclass of rock music; and so on. The occupation television presenter is a class. Really am unsure where you think the problem lies. --Tagishsimon (talk) 01:36, 29 March 2022 (UTC)
If something is a class of things, then that's because there can be instances. But nothing has (or should have) an instance of (P31)rock music (Q11399) statement; you should use genre (P136) for that. You don't say "this is a rock music". A genre is therefore not a class, but rather a characteristic of something. And since having a subclass of (P279) statement makes it a class, it shouldn't have that statement. ―Jochem van Hees (talk) 01:58, 29 March 2022 (UTC)
The class of people who are television presenters are the class of people with an occupation television presenter. Why exactly do you think that a class can only be used as a P31 / is only valid if it is used as a P31? --Tagishsimon (talk) 02:03, 29 March 2022 (UTC)
I mean with occupations it's a bit unclear because in English we do say that a person is "a television presenter", however that's not how we record it in Wikidata. Currently Wikidata says that television presenter is a class but also that you shouldn't use it as a class. I guess that can be possible, but I think the other examples highlight more clearly what I mean: British English is a class of what? British English is a language, not a type of language. ―Jochem van Hees (talk) 02:21, 29 March 2022 (UTC)
Is it not a class in the sense that there are subclasses such as British English, Canadian English, &c? A type which has subtypes? --Tagishsimon (talk) 02:25, 29 March 2022 (UTC)

Since I am not entirely sure I follow: The problem is with the label's wording, right? The word "subclass" implies there is a class, etc. Would the problem persist if the label was "form of", "subset of" or some other wording? (The "also known as" offers a range of expressions of varying helpfulness.) British English is a type/ form/ subset/ part/ variety of English. In general, I think fewer and therefore broader properties are a good thing. More and more specialized properties need more documentation, and they lead to more mistakes since it is harder to choose the correct property for a statement. So I am not in favor of creating more properties, just because we have problems with their labels. --Jonas kork (talk) 07:18, 29 March 2022 (UTC)

Hm yeah I didn't really consider P279 having that broader meaning than just classes, but in that case this also leads to other mistakes/inconsistencies. Help:Basic membership properties introduces the property as indicating that it's a class, which apparently is not always true. There's also the statement subclass of (P279)equivalent property (P1628)http://www.w3.org/2000/01/rdf-schema#subClassOf – if you look at that schema under "subClassOf", it says both the domain and range are a class. Furthermore, British English is an example of a human language (Q20162172), so that relationship should be P31, but if you follow the subclass chain you'll see it's also a subclass of human language (Q20162172). And I noticed that genre (Q483394)subclass of (P279)class (Q16889133), so genres are actually classes, even though they're not supposed to be. ―Jochem van Hees (talk) 10:07, 29 March 2022 (UTC)
Currently, in Wikidata we do call things that don't have instances but that are abstract in nature still class. Do you have an argument for why you think we shouldn't do that? ChristianKl ❪✉❫ 11:15, 29 March 2022 (UTC)
That doesn't make sense though; a lot of things are subclasses of object (Q488383), including building (Q41176), creative work (Q17537576), square (Q164), etc. which all should be used as classes. ―Jochem van Hees (talk) 11:43, 29 March 2022 (UTC)
You are right that it might make sense to adapt something in that. In general we use subclass of (P279) for the relationship between rock music (Q11399) and hard rock (Q83270) because it's the same relationship as the one between building (Q41176) and commercial building (Q655686). ChristianKl ❪✉❫ 12:09, 29 March 2022 (UTC)

Abraham Lincoln (1809-1865) president vs. Abraham Lincoln president (1809-1865)

There has been a push recently to add birth and death years (1900-2000) in descriptions to aid in disambiguation, but it has led to confusing label-description combinations like "John Smith husband of Jane Doe (1862-1939)" and "John Smith Mayor of Wikiland (1862-1939)" where the dates appear to be for the spouse in the first case, and the term of office in the second case. My suggestion is to move the dates to the start of the description field so they appear as "John Smith (1862-1939) husband of Jane Doe" and "John Smith (1862-1939) Mayor of Wikiland". That way they will match the order used in almost every encyclopedia, including Wikipedia. For example: "Abraham Lincoln (1809-1865), sixteenth president of the United States of America". --RAN (talk) 17:26, 30 March 2022 (UTC)

The role qualifier seems more useful than the dates, so dates should only be used as a last resort. The role should be in brackets or have a comma, and then Abraham Lincoln (president) (1809-1865) is much the same as "Abraham Lincoln (1809-1865), president", but there is some confusion over the role dates in the former. After all John Smith Mayor sounds like a valid name.
But if you really do have 2 Richard Burton actors, then, yes, put the dates between the name and role. But I suspect you'd need a bot to keep enforcing it

https://en.wikipedia.org/wiki/Richard_Burton_(disambiguation) has 

Richard Burton (footballer) (1889–1939), English footballer
Dick Burton (baseball) (1907–?), American baseball player
Dick Burton (golfer) (1907–1974), English golfer
Richard Burton (cricketer, born 1955), English cricketer
Richard Burton (cricketer, born 1976), English cricketer

Wikidata could have 

Richard Burton (1889–1939), English footballer
Dick Burton (1907–?), American baseball player
Dick Burton (1907–1974), English golfer
Richard Burton (born 1955), English cricketer
Richard Burton (born 1976), English cricketer

Vicarage (talk) 18:09, 30 March 2022 (UTC)

The comma is fine with me, we just need some standard. The comma actually is much better. --RAN (talk) 19:50, 30 March 2022 (UTC)
We could avoid this problem by simply not using years and similar clutter in descriptions. Problem solved. --Emu (talk)
Except, of course, we would lose the disambiguating utility of dob & dod in the description. So more ill-considered idea verging on trolling, than problem solved, IMO. --Tagishsimon (talk) 18:57, 30 March 2022 (UTC)
@Tagishsimon We have discussed this time and time again: The “disambiguating utility” isn’t needed for MnM, it isn’t needed for OpenRefine and it’s of very limited use in other cases. It creates all sorts of problems. The “push recently to add birth and death years” has resulted in temporary blocks. So please don’t call me a troll just because I happen to disagree with this pet project of some users. --Emu (talk) 22:06, 30 March 2022 (UTC)

AMX-30E

Hello, I am trying to get the English AMX-30E wikipedia page to redirect to the Spanish AMX-30E, but it does to Spanish AMX-30. I can't fix it. I have to add that AMX-30E in Spanish wikipedia was a redirection to AMX-30 until a few days ago I separated it to create its own article.81.0.34.181 20:46, 30 March 2022 (UTC)

There was an interwiki link on the English article which have now removed. — Martin (MSGJ · talk) 23:56, 30 March 2022 (UTC)

COVID-19 pandemic in Austria (Q86847911) is full (no, literally)

I tried to add the sitelink jv:Pageblug Covid-19 ing Ostenrik to COVID-19 pandemic in Austria (Q86847911), but I got the error message "The entity is too big. The maximum allowed entity size is 2.93 MB." The long list of case numbers is so big that the item cannot be changed any more - and apparently has been since November last year. What to do about this? - Andre Engels (talk) 21:06, 27 March 2022 (UTC)

Maybe stop trying to break wikidata items by using them as repositories for vast tabular data arrays? Granted it - tabular data - is not something that wikimedia seems to have got its head around storing, but even so, hoping that you can continue to stuff data into an overfull item was never going to work out. For practical purposes, you might split the item's content across a number of new items each dealing with a time limited span of Covid19 in the country. That'll still be a terribly bad way of storing the data.
The issue has been raised repeatedly. I've seen precisely zero interest being expressed in it by WMF, who are probably too busy counting their gold & 'communicating'. --Tagishsimon (talk) 21:17, 27 March 2022 (UTC)
I agree that that's not the kind of thing Wikidata should be for, but I don't feel that I am the one to choose what to do with it. However, for now I have resolved my issue by removing some duplicate references (though I had to log out to do so, logged in the extra Javascript stopped me from getting edit links). - Andre Engels (talk) 05:46, 28 March 2022 (UTC)
More to the point, why doesn't CovidDatahubBot use tabular case data (P8204)? Someone had the foresight to add this back in May 2020. So why keep stuffing these items like a christmas-turkey? Infrastruktur (talk) 06:30, 28 March 2022 (UTC)
I agree with @Tagishsimon and just recently I argued that we should stop storing time series data in this fashion and instead only store the latest value. All other values should be in a tabular data file on Commons (with a generic tabular data property). We could just move the data now, but it would be better if the tabular data also could be used in queries so that it was easier to reuse it on the other wikis. Ainali (talk) 20:34, 28 March 2022 (UTC)
I don’t think we need to make everything accessible via WDQS—particularly when we discuss data that is stored outside of Wikidata. One can easily make separate requests to WDQS (learning where the tabular data is located) and subsequently to the data URL at Commons.
WDQS seems already in a problematic state due to all the features that have been added in the past. —MisterSynergy (talk) 21:06, 28 March 2022 (UTC)
I agree. That is a nice to have, not a requirement to start migrating the data. I think we should just start moving the data over to Commons as soon as someone can build a decent pipeline for it. Ainali (talk) 21:11, 28 March 2022 (UTC)
I also agree that time series data is a bad fit for Wikidata (unless/until there are fundamental changes to the underlying data model). I don't think many people turn to Wikidata in the first place to look for such statistical data sets. Storing the latest value is much better, if "latest value" refers to the sum of all previous cases. Popperipopp (talk) 13:54, 29 March 2022 (UTC)
@Ainali: @Tagishsimon: @MisterSynergy:, @Popperipopp: What would be the downsides of not storing tabular data as statements? I can think of
(1) SPARQL queries for analyzing an arbitrary number of tabular datasets would be more complex. I've noticed that T181319 only considers querying a single dataset, but what if I wanted to query an arbitrary number of datasets in the same query? (e.g. analyze the number of cases (P1603) of countries in Latin America (Q12585) or analyze the social media followers (P8687) of any Member of the Congress of Guatemala (Q18277108), etc.)
(2) More friction for users to add data to Wikidata. Adding a single value will imply modifying the CSV file, so now users need to learn how to edit Wikidata and how to edit CSV files which adds more friction for them to start contributing. If this is done, not only we need to teach people on adding statements, qualifiers and references to Wikidata, but also to correctly format tabular data files.
(3) Consistency is required between tabular data files that store the the same information of different items. With consistency, I mean that the fields need to have the same name or they need to have the same order. For example, If I wanted to query the population of three countries, I would need those tabular data files be structured in the same way in order to be able to retrieve that data consistently with SPARQL. Not only those files need to be structured in the same way, but they also need to use the same units (having population counted in hundreds is different than having it counted in thousands).
Do you consider these as valid downsides? (please let me know so that I can proceed sharing an idea I've been thinking of)
Rdrg109 (talk) 14:35, 29 March 2022 (UTC)
Thanks for your overview Rdrg109. I think you have some good points, and I have a few ideas of ways to mitigate some of these.
(1) Yes, this could be more complex, but depending on how we make the datasets, it doesn't have to be. For example, your second task could easily be data in just one tabular file.
(2) It doesn't have to be more complex. Sure, if you would like to "complete the workflow" yes, it would involve editing both Wikidata and Commons. But one could also imagine a bot that updates time series on Commons whenever the data is changed on Wikidata. That means that a user only ever need to edit Wikidata in one place and not even needing to add a new statement. I would also be happy if a user just kept Wikidata updated and left it to other users to keep Commons in sync if no bots are being made. We shouldn't demand that Commons must be edited just to add the latest data point here.
(3) Yes, this could be troublesome, but that is already kind of the case today. In many places, we have different ways of modeling things that makes it hard to query. That is where wikiprojects are useful, to align this. And those could also help with the data on Commons. There is one upside too, on Commons we could have variants of data files, and in a way this open up possibilities of "alternative modeling" that could live in parallel with each other without hurting other user's workflow or reuse of the data.
-- Ainali (talk) 17:03, 29 March 2022 (UTC)
For consistency, we could introduce a way of specifying schemas in Wikidata for tabular data property values (e.g. new item for "tabular case data schema", which has some property with values which are themselves properties corresponding to headings in the CSV, then we link P8204 to that item with a schema property). SilentSpike (talk) 19:10, 29 March 2022 (UTC)
This is IMO not about advantages and downsides. I think we much rather discuss a common misconception about WDQS.
Many users/editors access WDQS via the web UI and consider this to be the one and only interface to access data; the UI is even expected to present posh results in various forms (tables, graphs, maps, charts, etc.); at the same time, the backend is supposed to do all the heavy lifting, no matter how complex this turns out to be. If such a UI-centered WDQS was the ultimate goal, it would be a very limited service since we would have to balance plenty of interests (data model as discussed here, but also which services to offer, resource usage limits, backend to use, customizations etc.). And there will always be usecases that will still not be covered…
Instead, we need to consider it to be an endpoint which predominantly serves machines that access it programmatically. For machines it is not at all complicated to collect data from several sources, filter and aggregate some new dataset from the different sources/inputs, and present it in the desired form. This is often a highly specific process, but ultimately allows for maximum flexibility since we could work with as few self-imposed limitations as possible. We really need to get away from the idea that WDQS is the ultimate service which is supposed to do everything with some magic; it is neither feasible in many situation, nor even simply possible with SPARQL only.
That said, offline data processing is not an easy task for everyone as it usually requires some sort of proficiency with a suitable programming language. —MisterSynergy (talk) 20:11, 29 March 2022 (UTC)
With regards to Rdrg109's points one and two; That the items are now reaching capacity proves that storing time series in Wikidata is a bad idea, at least how it's done at the moment. But reading MisterSynergy's reply gave me an idea. What if you simply segregated this data to a second Wikibase instance? You could then easily allow for bigger limits on individual items and you could access the data through federation, a standard SPARQL feature.
This will nicely solve most of the problems, except editing big items via web-browser might become problematic, but this isn't a hard problem to fix. The only problem I foresee is moderation, as I expect there will be fewer eyeballs watching the second instance, this can be solved by having people ask for edit permissions on the instance, or require manual review of the first 50 submissions or something similar. Infrastruktur (talk) 22:46, 29 March 2022 (UTC)
Someone wrote a prototype for a service to access CSV data from SPARQL back in 2017, surely federation is a better solution? It doesn't have to be complicated, just disable things like talk pages on the second wikibase instance and things like that to keep the maintenance burden to an absolute minimum. I have a feeling people might not like the idea of splitting up wikidata, but at the rate it is growing that may be inevitable, heck it may even improve performance to split off various special-purpose data. Infrastruktur (talk) 08:48, 31 March 2022 (UTC)
More SPARQL query federation and distribution of data across different databases is definitely something to expect in the future. However, I don't think we should use this to force tabular data into a triple store/graph database. It is simply the wrong solution to the problem. —MisterSynergy (talk) 09:49, 31 March 2022 (UTC)

Chatham Harbour

The page Chatham Harbour page is duplicated. This page here was created by a bot and I am convinced that it should be merged with this page here. I know that there is a tool to do that but I don't have much experience here. Thanks! Nistok1811 (talk) 14:22, 31 March 2022 (UTC)

✓ Done — Martin (MSGJ · talk) 14:25, 31 March 2022 (UTC)

Machine-readable list of all Wikidata redirects?

Is there a list of redirects in Wikidata, so that bots can fetch all redirects in bulk? Say, a CSV file with columns like Q166338,Q17992798 for redirect source and target. The file should get periodically updated. Ideally continuously, with a way to fetch only changes since a given timestamp/cursor, so that clients can update their local copy of the redirects. If no such tool exists today, I would volunteer to write one and put it on Toolforge, but first I’d like to double-check that nobody else has done it already. --Sascha (talk) 10:03, 31 March 2022 (UTC)

The list is long, there's about 3,6 million items. You can retrieve the list with SPARQL in several steps like so, just increase the offset value on each iteration:
SELECT ?a ?b ?modified
WHERE {
  ?a owl:sameAs ?b;
    schema:dateModified ?modified.
}
OFFSET 0
LIMIT 500000
Try it!
Trying to filter this list by a date-range will time out.
Also you can use Linked Data Fragments to get a list of all the redirects: https://query.wikidata.org/bigdata/ldf?predicate=owl:sameAs
This is more efficient but doesn't list the modification times obviously, but you may not need that if you keep track of when the lists were created. Infrastruktur (talk) 14:29, 31 March 2022 (UTC)
@Sascha, Infrastruktur: I wouldn't use SPARQL query above. The order of SPARQL results are undefined unless you use an ORDER BY clause in the query, so you have no guarantee that multiple executions with increasing OFFSET values will give different and complete results. The SPARQL engine (BlazeGraph) comes gives with slice factory service designed to do this in a safe way (documented here), so I recommend instead to use that service: 
SELECT ?a ?b ?modified
WHERE
{
  SERVICE bd:slice
  {
    ?a owl:sameAs ?b .
    bd:serviceParam bd:slice.offset 0 .
    bd:serviceParam bd:slice.limit 500000 .
  }
  ?a schema:dateModified ?modified .
}
Try it!
--Dipsacus fullonum (talk) 21:50, 4 April 2022 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2022/03&oldid=1843467142"