Wikidata

Wikidata:Project chat/Archive/2021/06

< Wikidata:Project chat‎ | Archive‎ | 2021
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

Help with property proposals

Hi

I would like to make my first property proposals but I don't know how to do that. I would like to add a external identifier for MonumentalTrees.com (Q107003167) and for themes in Flemish organization for Immovable Heritage (Q2091956) example. Is there someone who could help me with this so I could learn how to make property proposals.  – The preceding unsigned comment was added by Joeykentin (talk • contribs).

Try to find three statements with each you want to add. If the subject items don't exist yet, create them.
For the first, you then may want to go ahead and create it at Wikidata:Property_proposal/Authority_control. --- Jura 06:53, 28 May 2021 (UTC)
I tried yesterday but I didn't know what to do that is why I am asking for help I don't get what I need to da with the template Jhowie Nitnek 11:24, 28 May 2021 (UTC)
Can you add the samples here? --- Jura 11:41, 28 May 2021 (UTC)
@Jura1:Do you mean with samples, examples? Jhowie Nitnek 14:59, 28 May 2021 (UTC)
A few (three) examples of how each identifier would be used (what values on which existing items). These are handy when you start the form on Wikidata:Property proposal/Authority control. --Matěj Suchánek (talk) 15:40, 29 May 2021 (UTC)
@Jura1: For MonumentalTrees.com I have:
Sample 1: Schone Eik: 1861
Sample 2: Technobeuk: 4048
Sample 3: Chêne Crahay - Eik Crahay: 4489
For Flemish organization for Immovable Heritage Themes ID I have:
Sample 1: Alsemberg: 14064
Sample 2: Antwerp: 13182
Sample 3: Arrondissement of Halle-Vilvoorde: 16210 Jhowie Nitnek 22:01, 29 May 2021 (UTC)
  • The first one is indeed a bit tricky. The question is if there are multiple trees with the number 1861 or if that number is unique in the database. In the later case, the "1861" would be fine as value. The question then is if there is a possible formatter url that would work for all values. If not, this isn't a problem (leave it blank in the proposal), but when adding values to items with the new property one would want to include the url in the reference section.
For the second property, try to fill in the form with the three samples and leave out anything you aren't sure (other participants can later complete it). It should be similar to most other proposal in Wikidata:Property proposal/Authority control. I don't think any on [1] already covers the same.
You can ask questions also in Flemish/Dutch at Wikidata:De_kroeg. --- Jura 12:27, 1 June 2021 (UTC)

About wikidata schema technology

Hello, I want to know which schema technology is using wikidata? JSON or something other?  – The preceding unsigned comment was added by 103.198.29.178 (talk • contribs).

Importing Reign Period Data of Monarchs from Wikipedia to WikiData

There are currently about 9500 humans in Wikidata who are Monarchs (rulers, kings, queens). To be specific, they are instances of humans (Q5) who held a position (P39) that is either a Monarch or an instance of a subclass of a Monarch (Q2116).

Problem to be addressed

Almost none of the 9500 Monarchs in Wikidata have the data for their reign years. This is arguably a very important piece of information from a historical perspective which will enable querying for Monarchs by the period that they ruled in. This information is indeed in Wikipedia in a semi-structured form in the sidebar (Infobox royalty) that is typical of Wikipedia entries of Monarchs, but is not reflected in the corresponding Wikidata entities.

Goal of the project

The goal of this project is to use a bot to migrate the Reign period data that already exists for a Monarch in Wikipedia and update the property qualifiers start time (P580) and end time (P582) of the position held statement (P39). This would bring the reign period information in WikiData in synch with the data that is already in corresponding Wikipedia articles.

Steps to be taken

  1. Retrieve the Introduction section of the Wikipedia articles which contains the Reign information and parse it to get the reign start and end years.
  2. Use the values thus retrieved from Wikipedia to update (using a bot) the P580 and P582 property qualifiers of position held (P39) statement with a value that is an instance of or subclass of Monarch (Q116). The bot would also add the corresponding Wikipedia page as a reference for the Position Held statement using the Wikimedia Import URL property (P4656).

The second step above would be performed using a bot. We would first try this for a few dynasties and verify that the updates are being done correctly before performing the update on the remaining Monarch entities. R balaji69 (talk) 11:35, 1 June 2021 (UTC)

Thanks, @ChristianKl. I was going to submit a request for bot approval once I got feedback on this chat. As for references, I am proposing to have the bot put the corresponding Wikipedia page of the Monarch (from which we are getting the reign years) as the reference to the position held statement (P39) using the Wikimedia Import URL property P4656. Since the automation process knows about the source Wikipedia page URL, the reference would also be added in an automated way. I have put that information now above in the project description. R balaji69 (talk) 11:35, 1 June 2021 (UTC)
Cannot this tool do the collecting dates job ? --Bouzinac 💬●✒️●💛 11:51, 1 June 2021 (UTC)
@Bouzinac, I looked at this tool. Looks interesting, but for this project, it is not clear to me if it can be used. If I understand correctly, the tool asks for a single Property to load on the WikiData side. In this case, the information to be loaded are two values, the reign begin year and the reign end year as two Property Qualifiers of P39 (Position Held) of the Wikidata entity. I am not sure how to make that happen with a single Property Input that this tool is expecting. Is there a read-only version of this tool which just extracts the specified Infobox parameters that can be downloaded?

Properties for Price One Penny (Q106923678)

Hi! I'm new to Wikidata. I would like to import data from my database Price One Penny (Q106923678) to Wikidata. I've requested 5 properties in Wikidata:Property_proposal/Generic and 4 are now ready to be created. However, I do not have the rights to do so.

Should I contact Wikidata:Property creators individually to ask to create them? What's the protocol? Thank you very much! Marianika (talk) 18:32, 1 June 2021 (UTC)

The property creation process is slow and back-logged, I'm afraid. Be patient. Just curious, why are these requested under generic rather than under authority control? Bovlb (talk) 18:57, 1 June 2021 (UTC)
@Bovlb: I modelled my proposals on those for Wikidata:Property_proposal/WeChangEd_ID and Wikidata:Property_proposal/British_Book_Trade_Index, both of which were made under Generic. Only the first said "Originellement proposée sur Wikidata:Property proposal/Authority control" and there was a lot of debate beneath it, so I had assumed this meant it had not been requested at the right place. Marianika (talk) 19:58, 1 June 2021 (UTC)
Thanks. We could probably explain the sections a bit more clearly. Authority Control should (as I understand it) cover any external identifier. Bovlb (talk) 23:37, 1 June 2021 (UTC)
@Jura1: Thanks! Marianika (talk) 19:58, 1 June 2021 (UTC)

Modelling NUTS regions

I noticed Wikidata's coverage of Nomenclature of Territorial Units for Statistics (Q193083) codes were lacking in a few regards. While trying to improve the situation, I ran into some challenging modelling decisions that I'm not sure how to resolve. I left a comment at the discussion page of NUTS code (P605) but I don't think many users monitor that page.

1. Modified codes Some NUTS regions have been recoded. For example, in the 2016 version of NUTS, Upper Normandy (Q16961) was recoded from FR23 to FRD2. As you can see, I modelled this using the qualifiers start time (P580) and end time (P582). I also used stated in (P248) in the references to link the modification to a specific NUTS version. Question: Is there a better way to relate the NUTS region codes to the various versions of the NUTS standard?

2. Date specificity Related to 1, we can do better than state just the year when a given code starts/ends to be valid, and specify the date instead. But I'm not sure which date to use as there are three dates associated with each NUTS version: a) entry into force which "represents the date when the regulation has legal existence" (source), b) date of applicability which means that "the regulation is also applicable; it can be fully invoked by its addressees and is fully enforceable", and c) the date at which it "shall apply, with regard to the transmission of data to the Commission" (source) which is usually at January 1 the year efter it entered into force. Question: Which date should we use as a qualifier for start time (P580) and end time (P582)?

3. Multiple current codes for single regions NUTS consists of three (four if you count countries) levels, going from large regions to small ones. Some regions have codes on multiple levels. For example, Estonia (Q191) is a NUTS 0 region, a NUTS 1 region, and a NUTS 2 region at the same time. I began modelling this with applies to part (P518) as a qualifier, with one of the NUTS levels as values, as can be seen in the Estonia item. Question: Is there a better way to express which level the code refers to? Popperipopp (talk) 19:45, 1 June 2021 (UTC)

[Update] Support process review - May 2021

Hello,

This is a short update on the Wikidata support process review we started last year.

As a reminder, our goal is to improve the way we (the Wikidata development team at Wikimedia Germany) currently support the Wikidata community with collecting and reacting to bug reports and feature requests.

Based on your feedback, we have been working on incorporating your inputs into redesigning the Wikidata:Contact the development team page. We suggest changing its name to make it less focused on the development team and a central place for all discussions about technical issues on Wikidata, where editors are welcome to help answer and support others. We introduced sub documentation pages so that it’s hopefully easier to navigate bug reports, and we will work on a better integration of multiple languages.

You can already have a look at Wikidata:Report a technical problem (it’s a draft page, still in construction)

We invite you to let us know if you feel like something important is missing by leaving a comment on this talk page by June 10th.

If you have any questions please feel free to ask.

Cheers,

-Mohammed Sadat (WMDE) (talk) 08:57, 27 May 2021 (UTC)

  • We already have a support page at Wikidata where contributors help each other. It's regrettable that now that there are three full time staff members employed to communicate about technical questions from the community the dedicated page Wikidata:Contact_the_development_team is being discontinued and a second phabricator is set up.
Did any community members specifically request discontinuing Wikidata:Contact_the_development_team? Please point me to the requests.
Maybe a course of improvement for issues directly reported on Phabricator could be to do a periodic summary by staff on these, e.g. 3 months and 1 years after issues where created there. A user reporting issues there found that it's unclear if anything happened when they went through doing that. The same could obviously happen if technical issues were reported in "phabricator 2"/"bugzilla 3". --- Jura 10:53, 27 May 2021 (UTC)

Data is free to use; are images also free to use ?

Hello.

I want to download painting artworks from wikidata. I am able to write a python script which requests for public domain paintings with images. I'm also able to get the image URL.

Basically, I get a JSON full of items like


{ 

       "painting": {
           "type": "uri",
           "value": "http://www.wikidata.org/entity/Q20016794"
       },
       "picture": {
           "type": "uri",
           "value": "http://commons.wikimedia.org/wiki/Special:FilePath/Pieter%20de%20Hooch%20-%20Skittle%20Players%20in%20a%20Garden%20-%20Waddesdon%20Manor.jpg"
       }

}

First question : is the image itself an entity with properties ? How to programmatically retrieve the image license ?


The introduction of wikidata states that

Wikidata also provides support to many other sites and services beyond just Wikimedia projects! The content of Wikidata is available under a free license, exported using standard formats, and can be interlinked to other open data sets on the linked data web.

Second question Since the images (example) are from wikimedia (not wikidata), is there a garanteed that the images linked from wikidata are free ?  – The preceding unsigned comment was added by Laurent.Claessens (talk • contribs) at 08:26, June 2, 2021‎ (UTC).

The images are not part of Wikidata, so are not covered by the Wikidata licence. They are links to Wikimedia Commons. If Commons:Reusing_content_outside_Wikimedia doesn't answer your questions, you could try asking at Commons:Village_pump. Bovlb (talk) 14:56, 2 June 2021 (UTC)

Mixup in Super Eco 2700-class container ship (Q107002552)

Not sure how to solve this, as I'm not very active here, leaving it to more experienced Wikidata editors: English Wikipedia's article en:X-Press Pearl is about an individual ship of the Super Eco 2700 class which recently caught fire. German Wikipedia on the other hand, where there's a general tendency to prefer articles about classes of freighter ships instead of individual ships, has de:Super Eco 2700 about the class of more than 10 ships, and the X-Press Pearl incident is just mentioned in a "Zwischenfälle" (incidents) paragraph. Super Eco 2700-class container ship (Q107002552) is now a mixup of the two, with Super Eco 2700 (the type) as the label, but X-Press Pearl as an alias (which is clearly wrong, as that's not the name of the class as a whole), and also an IMO number for the individual ship. I think there should be two items, one for the class and one for the X-Press Pearl. Gestumblindi (talk) 20:09, 1 June 2021 (UTC)

Someone merged two items about different things. I undid that. @Gestumblindi: --- Jura 20:21, 1 June 2021 (UTC)
@Jura1: Thank you! I added "vessel class" to X-Press Pearl (Q107002557): container ship built in 2021 though it tells me now that there are "some potential issues" with that statement, not sure how serious these issues are, looks fine to me... Gestumblindi (talk) 20:29, 1 June 2021 (UTC)
Not sure either. I made a few more for ships of the class.
@Gestumblindi: --- Jura 08:01, 2 June 2021 (UTC)
Seems we already had all but X-Press Pearl (Q107002557) --- Jura 12:57, 3 June 2021 (UTC)
I've changed the item about the class to a subclass of ship class (Q559026). I've used this query to look at the types of items that link to "ship class" and ship type (Q2235308). I've also linked to Super Eco Ship (Q11313672), which is probably an unrelated concept ship (class)? --Azertus (talk) 12:27, 3 June 2021 (UTC)

Opinions on legislature / elections

Hi all,

A "legislative term" is, as I see it, restricted to national legislative chambers. However, in certain countries, there are assemblies that are elected, but are not chambers. For instance, Guardian Council or Assembly of Experts for Leadership in Iran.

My main objective is to be able to provide election information for a politician that has been elected several times in a row.

My understanding is that for a chamber, one would have filled "part of" and used the "legislative term" items, such as in Barack Obama.

I have created 1982 Assembly of Experts Election, but I am not very happy with it, as an election in not a term, it is more a point in time.

How would you recommend to do it for, say, a Member of the Guardian Council ?

Thanks

Zejames (talk) 06:54, 3 June 2021 (UTC)

If I understand you, you are talking about parliamentary term (P2937). I think it would make sense to broaden it to terms of any governmental assembly. I don't understand the subtleties you are referring to of Iranian government but we often have corresponding election and term items (e.g. 2004 United States House of Representatives elections (Q2539374) and 109th United States Congress (Q168778)) so you could just make another item to represent the term separate from the election for the term. BrokenSegue (talk) 07:31, 3 June 2021 (UTC)

Vandalism on some articles

Someone removed twice the link to be-tarask article from Q1215892. I guess vandals will do it one more time. Can someone protect it? Moreover, this is not the only article removed. Is there any filter to see changes of be-tarask links only? Dymitr (talk) 05:35, 2 June 2021 (UTC)

Not directly, but we do have a filter for "new user removing sitelink". Try this version of Recent Changes and search for "be_x_oldwiki". Bovlb (talk) 15:57, 2 June 2021 (UTC)
Thanks. Added this page to my bookmarks. --Dymitr (talk) 18:42, 2 June 2021 (UTC)
@Dymitr You can also use https://wdvd.toolforge.org/index.php?lang=be-x-old&limit=50&sitelinks=on (WD:WDVD, limited to be-tarask sitelink removals). Lucas Werkmeister (talk) 17:54, 3 June 2021 (UTC)

Simple question for some people

(Not for me though... :) How can we link İzzet Akay (Q20724828) with Category:Films by İzzet Akay (Q32723579)? Thanks in advance. --E4024 (talk) 14:18, 4 June 2021 (UTC)

One way would be to add a claim of significant person (P3342) on the Category item. From Hill To Shore (talk) 15:09, 4 June 2021 (UTC)
category combines topics (P971)?
category combines topics (P971) seems more appropriate than P3342. See for instance Category:Films directed by Frank Hall Crane (Q8453938), Category:Films directed by Jože Gale (Q26261646), or Category:Films produced by Steven Spielberg (Q8457512) (creation of additional items might be needed to accurately describe the relationships). -Animalparty (talk) 16:20, 4 June 2021 (UTC)
Thanks to both. This person is not "director" or "producer" but "cameraman". Please give a helping hand if you can, directly at the concerned items; I am not the type of person who learns quickly how to fish. --E4024 (talk) 16:33, 4 June 2021 (UTC)

Candidates for deletion?

Q25085714, and its main subject, Q28017217, both seem to be spurious 'made up' items for the purpose of personal amusement and/or trivia collecting. On English Wikipedia, both "List of" and category have been deleted, and it seems to me there is no evidence "presidential school" is anything real, discrete, and notable, even by the minutest of standards Wikidata employs. Is deletion warranted for either of these items? -Animalparty (talk) 16:12, 4 June 2021 (UTC)

nominate it for deletion BrokenSegue (talk) 17:05, 4 June 2021 (UTC)
I do not have any idea about the school, but empty category items can not be notable, as categories do not exist outside the Wikiverse. I deleted the category.--Ymblanter (talk) 19:13, 4 June 2021 (UTC)

Nothing to do with the above

Ping Ymblanter, but I think Sude Nisa Bodur (Q107036039) should be deleted even before empty cats, because the item is only invented to make a link to an unnotable person. Thx. --E4024 (talk) 19:41, 4 June 2021 (UTC)

You've already nominated this item for deletion. Bringing your complaint here is just canvassing. There are hundreds of requests for deletion. We sensibly keep them on their own page, rather than this one. Please respect that convention. --Tagishsimon (talk) 19:56, 4 June 2021 (UTC)
Please don't exaggerate. Why would I canvass? Who is she? What importance does she (or the item) have? I was just trying to attract attention to the same issue you said: "There are hundreds of requests for deletion." The reason of that backlog is the fact that such obvious cases are still there, only because a bot says "there is a link". And the winner issss... The link is added by an LTA to keep "his" own item a few days more... It is not even about this very item but about the LTA whose images are being deleted from Commons everyday and whose item is still here. (BTW I did not even comment in that person's own item for deletion, as it is / as the two are so clear cases. Neither he nor his sister are of no interest to me nor to Wikidata.) Best. --E4024 (talk) 20:07, 4 June 2021 (UTC)
"Why would I canvass?", you say, as you canvass some more: "I was just trying to attract attention". smh --Tagishsimon (talk) 21:22, 4 June 2021 (UTC)
The DR is withdrawn. E4024 (talk) 22:54, 4 June 2021 (UTC)

Conflations of United Kingdom (Q145) and United Kingdom of Great Britain and Ireland (Q174193)

GPinkerton (talk • contribs • logs) has recently been making adjustments to United Kingdom related items that has created conflations. Previously we had United Kingdom (Q145) and United Kingdom of Great Britain and Ireland (Q174193) with a transition in 1927 to represent the formal secession of the Irish Free State (Q31747). Now we have both United Kingdom (Q145) and United Kingdom of Great Britain and Ireland (Q174193) starting in 1801 and conflating their roles. [2] This breaks a number of citizenship and other related date constraints throughout Wikidata, though GPinkerton cites it as an attempt to repair other citizenship constraints in other areas.[3] This also leaves us with a tangle on what to do with Irish Free State (Q31747). If we merge United Kingdom (Q145) and United Kingdom of Great Britain and Ireland (Q174193) to remove the conflation, we will have Irish Free State (Q31747) being the successor to the current United Kingdom (Q145). On a related point, GPinkerton seems to be doing something similar to Parliament of Great Britain (Q2739604).[4] Can I have the community's thoughts on how to untangle this mess? From Hill To Shore (talk) 18:39, 31 May 2021 (UTC)

I think From Hill To Shore has got this backwards; the problem here is not conflation but a false dichotomy. The mess is caused by the parameters that flag as a mistake the adding of "citizenship of" properties to people from the the UK who lived before 1927. This is fixing the mess. Since Q145 and Q174193 are the same entity (the latter being a historical long-form name for the former), they should clearly be merged. Absolutely nothing happened to anyone's citizenship in 1927, and anything that creates this impression should be removed. GPinkerton (talk) 18:49, 31 May 2021 (UTC)
To respond to the point above, having the Irish Free State as a successor to the present UK is entirely correct. At (or rather shortly after) the partition of Ireland (not in 1927), the present-day UK's territory shrank to exclude the short-lived Southern Ireland, hitherto a part of the UK. But only one new state (the Irish one, in its various iterations) started in the 1920s. The British state (i.e., Q145) carried on regardless in a smaller form. GPinkerton (talk) 18:54, 31 May 2021 (UTC)
N.B. that the errant item was created by a bot in 2012 and a year later humans began to populate it as though it were a country rather than an official name that is now historical. GPinkerton (talk) 19:13, 31 May 2021 (UTC)
As I understand it, the UK was considered a "continuing state" through the creation of the Irish Free State, and the latter was a newly created state. There was no question of "successor states". The 1927 event was just a name change. Ghouston (talk) 00:17, 1 June 2021 (UTC)
Primarily, yes. There was some legislation that went along with it but most external sources treat it as a continuing country. The problem comes with how we represent it in Wikidata. A straight forward merge of the items is impossible as many language versions of Wikipedia have separate articles for before and after 1922/1927 (1922 being the creation of the Irish Free State and 1927 being the formal recognition of the change by the UK). I don't have strong feelings on this except that what we have now is a broken mess; either the previous state is restored or we find an alternative method of untangling the situation. From Hill To Shore (talk) 01:07, 1 June 2021 (UTC)
From Hill To Shore what legislation? 17 Geo. 5. c. 4 only changes the long name for what it describes itself as the "Parliament of the United Kingdom" and spells out the meaning of the short form accordingly. " Formal recognition of the change by the UK" happened years before, before the Free State even existed. Complete legislative independence was not achieved until later, after the Statute of Westminster made Ireland, Canada, etc. sovereign states. Whatever the case, 1927 is a mistake for anything other than a change of official long name. GPinkerton (talk) 03:30, 1 June 2021 (UTC)
@GPinkerton: I am not sure why you are trying to start an argument over this as we are in broad agreement on the sequence of events. There was legislation in the early 1920s to set up the Irish Free State and there was legislation in 1927 to remove a vestigal claim to Ireland implied by the old name. 5 years to formally renounce a claim to territory is fairly quick. The UK only renounced its claim to France in 1800, 242 years after the English lost their last French possession of Calais. However, all of that is irrelevant to the discussion about how we represent the information in Wikidata. From Hill To Shore (talk) 13:35, 1 June 2021 (UTC)
From Hill To Shore There was no "vestigial claim to Ireland implied by the name" any more than there is a vestigial claim to Northern Ireland in the name "Republic of Ireland". As you have yourself said just above, the UK existed before 1800 and unquestionably existed before 1921, so as you have now outlined, the way Wikidata handled this before changes were made is ideal: 1707 as the starting date. Later, someone changed this and caused a mess. I have changed it back. Now the errors that accreted while United Kingdom of Great Britain and Ireland (Q174193) has been treated as historical country are obvious, and should be fixed. GPinkerton (talk) 16:21, 1 June 2021 (UTC)
I think you are being argumentative for the sake of being argumentative. There is no dispute on the sequence of events, just that you didn't consider the wider implications for Wikidata by your edit. I don't think that I have had any previous dealings with you but looking at your history on another site makes me wary about engaging with you; I don't need the stress.
I suggest you move on and address the points raised by the editors below. From Hill To Shore (talk) 16:57, 1 June 2021 (UTC)
From Hill To Shore "you didn't consider the wider implications for Wikidata by your edit" why do you keep saying this? What are these implications you think I didn't consider and how can they be reconciled? Are you saying that there are insurmountable problems presented by using the proper date? (problems that are not problems in the case of the USA, France, etc. ...) If so, what are they? GPinkerton (talk) 18:49, 1 June 2021 (UTC)
That was the case, and although Wikidata needs to have items for each name, due to the separate Wikipedia articles, it doesn't need to treat them both as representing sovereign states. United Kingdom of Great Britain and Ireland (Q174193) could be an item for a name, or a period of history, or something Wikipedia-specific. Ghouston (talk) 06:43, 1 June 2021 (UTC)
We can also question whether Irish Free State (Q31747) and Republic of Ireland (Q27) should be considered separate states, or just a change of constitution and name of a continuing state. Ghouston (talk) 06:47, 1 June 2021 (UTC)
  • Please avoid changing the current consensual version of the item until the discussion determined a new consensus. --- Jura 12:31, 1 June 2021 (UTC)
  • Agree with Jura. GPinkerton's date changes seem to me to be unhelpful and and their explanation - false dichotomy, fixing the mess - arrogant and ill considered. Although it is the case that modelling of various UK concepts in wikidata leaves something to be desired, unilaterally deciding that one item, which hitherto had been modelled as the entity apparently established in 1927, is now to be the omnibus entity dating back to 1707 is not the way to go; at least, a) not without discussion and consensus for a change and b) not without considering broader aspects than seem to have occurred to GPinkerton, including sitelinks, statements, and uses of the values. The trouble with being here for 5 minutes and deciding on & implementing a change affecting hundreds of thousands of items, is that you're certain to have underestimated the complexity and consequences. --Tagishsimon (talk) 12:45, 1 June 2021 (UTC)
    +1. Another example in the likes of the topic which drew many drawbacks is the case of Estonia (Q191) vs Interwar Estonia (Q2174038) [the latter being only historical period (Q11514315) and not historical country (Q3024240) .... Bouzinac 💬●✒️●💛 12:57, 1 June 2021 (UTC)
  • Bouzinac Estonia beginning in 1918 is instructive; the equivalent of the UK beginning in the 1920s would be Estonia beginning in the 1990s. Obviously both situations would be absurd and counter-factual and would prevent citizens of either country being labelled as such for the majority of both nations' lives. GPinkerton (talk) 16:36, 1 June 2021 (UTC)
    Estonia commencing in 1918 may appear absurd and to a certain point, I would have agreed. But Estonian do not share this point of view as they explicitly state that their State has commenced in 1918, as per their state continuity of the Baltic states (Q7603672)... I would say states have juridic continuances with changing of constitutions, losing wars, gaining wars, etc.
    About 1927 : you wouldn't compare UK 1928 and UK 1926 because their range does not have the same span : not the same juridic perimeter, the same area, the same population number... In 1926, Ireland was still officially (to English pow) in the UK. Bouzinac 💬●✒️●💛 19:18, 1 June 2021 (UTC)
  • Bouzinac Sorry I may have been unclear: to reiterate: I agree with Estonia beginning in 1918. But I fear your comment about 1927 is not correct. UK 1928 and UK 1926 have exactly the same territory. The Irish Free State started to exist in 1922 and its territory stopped being part of the UK at that time. This event was agreed to by the UK in 1921. The UK in 1928 had precisely the same juridic perimeter, the same area, the same population number (more or less!). "In 1926, Ireland was still officially (to English pow) in the UK." is completely untrue. As I say, nothing happened to the UK (or Ireland) in 1927. Partition of Ireland happened years earlier. GPinkerton (talk) 19:43, 1 June 2021 (UTC)
  • Tagishsimon I don't understand what you mean by "trouble with being here for 5 minutes and deciding on & implementing a change affecting hundreds of thousands of items, is that you're certain to have underestimated the complexity and consequences." Are you saying that I've been here for 5 minutes (why?) or that I don't understand that changing items affects other items? Either way, I don't know why you're making these assumptions. As I have explained, the fictitious 1927 set-up (which does not appear to be justified anywhere) was causing problems, so in order to resolve these problems, (like the problem of the UK not being the state from which Ireland became independent which From Hill To Shore has pointed out) I returned the property to its original state. Whatever "the entity apparently established in 1927" was, it most definitely was not the United Kingdom. GPinkerton (talk) 16:24, 1 June 2021 (UTC)

Examples for comparison

Jura1 the problem I have been having is that marking, say Henry Napier Bruce Erskine (Q5726222) as country of citizenship (P27) United Kingdom (Q145) is causing an error message because "the earliest start date of the United Kingdom is 12 April 1927". This is obviously wrong. I'm struggling to see how it could be imagined otherwise. Are we to say Calvin Coolidge is not an American citizen because he was born under a different flag to modern Americans? GPinkerton (talk) 19:50, 1 June 2021 (UTC)
The solution to your problem with Henry Napier Bruce Erskine (Q5726222) is to use the other item. No reference to Coolidge needed. --- Jura 20:08, 1 June 2021 (UTC)
@Jura1: but that misrepresents the subject as a citizen of a state that no longer exists. This is obviously wrong. That is the problem, not a solution to it. Furthermore, it creates the problem that William Erskine (Q8008651), the former's father, would, under the same logic be labelled as United Kingdom of Great Britain and Ireland (Q174193). But this wrong, because the latter was born before the union of 1800, and it would be quite wrong to think that, doubtless unbeknownst to himself, his citizenship somehow changed midway through his life simply because the official name of the United Kingdom changed. GPinkerton (talk) 20:15, 1 June 2021 (UTC)
Nature of citzenship evolves over time. Isn't it that some people with "British" passport aren't even allowed to enter Great Britain? --- Jura 20:25, 1 June 2021 (UTC)
@Jura1: I don't know what you're referring to but my answer prima facie would "no, of course not, where did you get that idea?" and then "what relevance would that have to the subject?"? How does it change over time? Are you suggesting people in the 18th and 19th centuries should not have country of citizenship (P27) or? GPinkerton (talk) 20:35, 1 June 2021 (UTC)

Solution?

I suggest United Kingdom of Great Britain and Ireland (Q174193) should lose (if necessary) its current status as an instance of historical country (Q3024240) and sovereign state (Q3624078) and should gain former name (Q29569274) and/or historical period (Q11514315). What problems/objections could this raise? GPinkerton (talk) 21:01, 1 June 2021 (UTC)

I support this change. There's a tendency to be overly impressed by name changes, I think, hence the attraction for some people of the idea that a new UK was created in 1927. But there's not even much of a name change, since few use the full name, so the United Kingdom prior to 1927 was still the United Kingdom afterwards. In a previous discussion, I found a reference to support the continuing state point of view: https://publications.parliament.uk/pa/cm201213/cmselect/cmfaff/643/643.pdf (a sizeable PDF document), which was a discussion on what would happen if Scotland became independent. On page 130, Lidington says: "If we look at analogous examples, when Ireland established the Irish Free State in 1922 the United Kingdom continued to exist. It was accepted as such. The Free State and subsequently the Irish Republic became new countries. The same applied when India, which as a dominion had been a founder member of the United Nations, separated from Pakistan. India was accepted as a continuing state; Pakistan was the new state and had to apply to join the international organisations. The same took place when Eritrea became independent from Ethiopia, when South Sudan became independent from Sudan, when Malaysia and Singapore separated. If you look at recent European history, it is very striking that at the time of German unification the Federal Republic of Germany continued to exist and was accepted as such and what happened in international law and in terms of membership of organisations was that new Länder from the former German Democratic Republic became part of that continuing Federal Republic of Germany." Ghouston (talk) 22:52, 1 June 2021 (UTC)
I don't think I can add anything further on this. My main concern was that a big change affecting an item linked over 1 million times needed to be discussed by the community. We have now had the opportunity for discussion and Ghouston's initial proposal above and your refinement of it seems to be the only option being discussed. Unless one of the others want to step in with a new proposal (even if it is to maintain the status quo) then we have a new consensus. From Hill To Shore (talk) 19:57, 4 June 2021 (UTC)
What does that implies for the current Q145#P571 value and for the current Q145#P1365 ? I am not comfortable with a current important state being "simply" preceded by a historical period (Q11514315) (I was rather in favor of a historical country (Q3024240) as a previous value. Okay, the 1927 thing might be a bit far-fetched, hair stretched as we like to say in French, but I would like the current UK item being correctly linked to the previous "real state", perhaps not the 1927stuff. Bouzinac 💬●✒️●💛 20:15, 4 June 2021 (UTC)
I'm deeply unconvinced. First, I don't think we've had anything like an articulation of the supposed problem for which a 'solution' is required. The top of the discussion speaks of "The mess is caused by the parameters that flag as a mistake the adding of 'citizenship of' properties to people from the the UK who lived before 1927. This is fixing the mess." which, at best, is to suppose that everyone agrees with an unspoken assertion that there has been just a single sovereign state since 1707 - something which I'd dispute. Nor is assigning country of citizenship to UK people all that hard; three date ranges versus the individual's dob & dod. We've managed to do this - assign an appropriate one or more UKs to P27 for the majority of UK citizens.
There is, for me, a clear requirement to be able to point, in an unambiguous way, on law items, to the country/state in which laws are enacted and to which they apply. UK laws in 1707-1801 cover different areas that 1808-1922 (or 1927), cover different areas than 1922/7 onwards. They were made by parliaments of different names. Countrywide laws in any of those period apply to the country, and that being the case it is useful to have a country item to point to. Antedating the inception date of a country invites the mistake of linking a law created before the existance of that country, with the boundaries of that country, which is erroneous.
iirc, one of the changes I objected to was setting the inception date for what hed been the 1927 onwards UK, to 1707. Obviously this made mincemeat of the inception and dissolution dates of the two preceding UK items. But it also seems as partial as can be. The union of E&W & S produced version 1 of a UK. Why does not the union of this UK with Ireland produce a version 2 UK on exactly the same basis. Why does the loss of most of Ireland sometime later not have the same effect, producing a version 3?
I'm aware of some of the comparisons - with the USA, or France - which have been used to argue against a separation of the UK into a set of sovereign state items having distinct periods. But going back to my point about the lack of articulation of the supposed problem, I don't see that any real analysis of the way in which states/countries are modelled on wikidata has been done. A quick look right now shows the truthy value of inception for Germany as 1949, France as 1958, Russia as 1991. Clearly here the model is, at least to some extent, respecting that fundamental changes to a sovereign state - a new Reich, a new Republic - mark the inception of a new sovereign state. But for the UK, we're being asked to overlook the addition and loss of Ireland, and the changes to the membership of the legislature, as being insignificant; and asked instead to buy the idea that because the term 'United Kingdom' has been applied in each of the three periods, there is only a single continuing sovereign state.
For me, until we have an understanding and articulation of the rule base by which we model the evolution of sovereign states, any change to the UK items status quo is merely arbitrary and capricious, and we should avoid the arbitrary and capricious. And in order to understand and articulate the rule base, work need to be done to analyse our model. Right now all we have is a suggested profound change being driven by one person particular prejudice. --Tagishsimon (talk) 20:37, 4 June 2021 (UTC)
@Tagishsimon: The current set up is arbitrary and is not the original set-up, which used 1707. 1707 is the proper date because 1707 is the date that the United Kingdom first began to exist. Merging it with other "states" (Ireland was not exactly sovereign in 1800...) does not make a new state. The current set up has us being asked to buy that a change in country's official name is somehow the same thing as beginning a new state, or that a change in state's territory is somehow a constitutional difference. The idea that the present UK is different to the early 20th century one is fundamentally misunderstanding the nature of the constitutional situation. What you call "the addition and loss of Ireland" is in fact only, for constitutional purposes, "the addition of Ireland". Nothing has happened to the 1801 union, it was not repealed or whatever; the kingdom of Ireland is all still part of the UK as before. It's simply smaller than it was. The loss of Hong Kong did not create a new state, even though the UK's population decreased by as much, if not more, than in 1922. The loss of the Philippines does not create a new United States. Our inception date for Japan is 11 February 660 BCE; Shogunate, Meji period, MacArthur all notwithstanding. It seems to me as though arbitrary and capricious best describes the current situation. It's strange to see the current situation for the UK (a state which has existed for centuries) as anything but aberrant and untenable. Have you ever seen anyone write "the United Kingdom came into being in the late 1920s"? It's absurd to suggest. What possible grounds are there for maintaining the bizarre caprice that is the status quo? Wikidata cannot invent its own definition for when countries came into being, it has to reflect historical reality. GPinkerton (talk) 02:23, 5 June 2021 (UTC)
To remove the capricious and arbitrary use of 1927 as an inception date, return the United Kingdom back to its original inception date, and demote its historical iterations' long names to official names and/or historical periods properties would seem to be the best solution, and the one that applies to all other nations surveyed thus far. It's clearly nonsensical to insist that people who lived in France before de Gaulle and the Vth republic should have to have any other country of citizenship than simply "France" purely because the state's official name changed. Now, including all the French kingdoms from the Roman empire onwards is clearly far more tenuous than, say, listing the modern French state's first foundation in 1789, or listing the UK's inception date at 1707. The fact that both France and the UK have had the same name (barring official nomenclature) between the 18th century and today should be enough to demonstrate that the UK then was the same country as the UK now. Bulgaria did not exist as a state at all for long swathes of the Middle Ages, yet we accept the traditional foundation in 681. The fact that the 1st and 2nd Bulgarian Empires have since risen and fallen, and then the principality, then the kingdom, then the people's republic, then the modern republic seems as immaterial as the enormous flux in Bulgaria's territorial extent over that time, which changed far more radically and more frequently than the UK's has in its 300 years. Let alone the fact that the 18th-century United States was a fraction of the size of the modern state, yet we don't need to link its citizens to a new Wikidata item every a new state joined the union, or a new territory was annexed, or incorporated, or organized, or whatever. Similarly, the fact the UK's official name (or, more precisely, the official name of its monarch and of its parliament) has changed a number of times makes/made no difference to anyone's citizenship or the applicability of any laws. Everyone in this discussion has used this item's name to refer to the state existing long before the 20th century. Where is it stated that the United Kingdom did not exist before 1927, or that a new one started that year? Where does this idea come from? GPinkerton (talk) 03:43, 5 June 2021 (UTC)
Rules of citizenships varies greatly by time and by law (not necessarily linked to the change of a constitution). France's list of political systems in France (Q29837670) many change of republics do have importance. Let's say there is a general concept 'France' and many regimes that have been part of it. The guy born before 1789 was a subject of the King of France, thus cannot be "French" but "subject of the French king". Plus its french citizenship was non sensical (there were no passports at this time, no ID cards, no real border police). That's why Q7742#P27 might appear nonsensical. As well as Q44279#P27. Perhaps the solution is that country of citizenship (P27) might accept only current and recent countries [citizenships/nationalities // meeting timescales of state and of the individual] and perhaps there would be a need of a P that would collect the "regime under which the people lived in" [or something like ethnic group (P172) ... That point can I really understand. The other point removing historical country (Q3024240) is nonsensical for me too (leaving aside the question of correct inceptions dates). Bouzinac 💬●✒️●💛 05:09, 5 June 2021 (UTC)
@Bouzinac: The guy born before 1789 was a subject of the King of France, thus cannot be "French" but "subject of the French king". Plus its french citizenship was non sensical (there were no passports at this time, no ID cards, no real border police). I'm sorry this sounds a bit silly. Being a French subject is not different to being a British subject (a status that existed until the late 20th century), and that has nothing to do with whether a country is pre- or post-revolution or whether the UK is the same country as the country called the UK for the past 300 years. It can be argued that applying a modern conception of citizenship can be anachronistic in the past, but to argue that people stopped or started belonging to countries in the way that "citizenship" implies because of mere change in government is stretching it again. Roman citizens remained Roman citizens when the Republic became an Empire, and none of them would have noticed any difference. René Descartes was exactly the same amount of French "citizen" as Victor Hugo, and neither of them lived in the Vth Republic. (The One True France, according to the logic that has the One True UK existing from 1927 only and then completely without historical precedent.) GPinkerton (talk) 13:52, 5 June 2021 (UTC)
The w:United Kingdom did not exist by that name until the w:Acts of Union of Great Britain and Ireland 1801-01-01. And Great Britain didn't exist by that name until the w: Acts of Union of England and Scotland 1707-05-01. Prior to 1603 England and Scotland were separate countries. When Queen Elizabeth I died in 1603, King James VI of Scotland became simultaneously King James I of England. From 1603 to 1707 England and Scotland were separate countries but with the same monarch, ignoring the w:Wars of the Three Kingdoms, which ended with Charles I losing his head, leading to the short-lived w:Commonwealth of England, 1649-1660.
I'm no great scholar of UK history, but I believe the following summarizes my understanding and the sources I just checked:
  • From pre-history to 1603, England, Scotland, and Ireland were separate countries; Wales was for centuries many different countries, that were absorbed by England at different times, especially during the w:House of Plantagenet, if I understand correctly.
  • From 1603 to 1707, England, Scotland and Ireland were separate kingoms (or Commonwealths, for a short period), which shared the same ruler (Stuart kings, Lord Protector Cromwell, then William and Mary, then w:Anne, Queen of Great Britain).
  • In 1707 England and Scotland joined to become Great Britain during the reign of w:Anne, Queen of Great Britain.
  • In 1801 Great Britain and Ireland combined to become the United Kingdom, officially the United Kingdom of Great Britain and Ireland.
  • In 1921 Ireland was partitioned, with most of the island becoming the Republic of Ireland and the official name of the UK being changed to the United Kingdom of Great Britain and Northern Ireland.
I don't know what that says about the issues being discussed here, except that the UK was NOT formed in 1707. DavidMCEddy (talk) 05:08, 5 June 2021 (UTC)
I think all you can do is look at each transition and decide whether there is a continuing state, successor state, or newly created state. Changes of name or territory aren't definitive. Ghouston (talk) 12:45, 5 June 2021 (UTC)
This change of the UK in 1927, I can't see is anything more than a change in the way the UK described itself, in its full name. The en:Royal and Parliamentary Titles Act 1927 seems to have taken effect on 13 May 1927, when a royal proclamation took place. There wasn't any change of government on that date. It seems like a very strange date to choose for the foundation of the UK, and could do with some references to support it. Ghouston (talk) 12:52, 5 June 2021 (UTC)
@DavidMCEddy: I'm sorry but your analysis above leaves much to be desired. The United Kingdom did not exist by that name until the Acts of Union of Great Britain and Ireland 1801-01-01. This is wrong. The "United Kingdom of Great Britain" is referred to as such in the Acts of Union of 1707, and in numerous subsequent 18th-century laws. To claim it did not exist until 1801 is simply untrue and ignores the country's own name for itself (the United Kingdom) throughout the 18th century. In 1800, Ireland and the United Kingdom made legislation that made Ireland a part of the United Kingdom, and the UK's official name changed to be the "United Kingdom of Great Britain and Ireland". It changed again in 1927, being renamed to the "United Kingdom of Great Britain and Northern Ireland". At no point between 1800 and today did a new British state get created, and at no point since 1707 has the UK not had the name "UK". Surely if, as you claim "the UK was NOT formed in 1707", you should be able to find reliable sources saying as much. If so significant a country as the UK had sprung into existence between WWI and WWII, surely some some evidence can be found for this. I've seen not a shred so far. If someone told me that USA's inception was not in the 18th century, but, say, in 1960 when the Union with Hawaii was effected, they'd need to produce a pretty good source to make this change from the common sense, historically accepted date of American independence. So should it be with the UK. Where is the evidence? GPinkerton (talk) 13:40, 5 June 2021 (UTC)

Copyright of Classifications

Hello,

I am interested in Classifications. There is a classification International (Nice) Classification of Goods and Services (Q193988) also known as nice classifikation and in this classification there are the most goods and services who exist included. What is the copyright status of this classification and do you think that is classification is helpful for Wikidata. I thought that it is in the public domain because it is a treaty but there I am not sure. From my point of view if for the objects mentioned in the classification items would exist then there is a wide coverage in Wikidata for that topic.--Hogü-456 (talk) 21:07, 4 June 2021 (UTC)

From https://www.wipo.int/classifications/nice/en/faq.html:
Do I need to pay WIPO if I want to use a database version of the Nice Classification in my web service?
No, you can download the NCL files, but you have to acknowledge WIPO's copyright if you plan to use the NCL on your site. For more details, please refer to the conditions of use in the Download and IT Support area.
--SCIdude (talk) 07:50, 5 June 2021 (UTC)

Need property "display-authors"?

Wikiversity:Anti-rivalry, contagious diseases and news displays an error message for Q107063587, saying "Explicit use of et al. in: |author4= (help)" with "(help)" linking to w:Help:CS1 errors, which seems to say that if I want "et al." with a list of authors, I should use "|display-authors=", if I understand correctly.

In fact, Q107063587 has 44 authors. The URL suggested using only the first 3 followed by "et al.". That's what I entered into Wikidata. However, Wikiversity (and apparently also Wikipedia) don't like that. I'm guessing that this error would disappear if a property named "display-authors" were added and assigned a value "3" in Q107063587.

However, I don't know what to do with this, so I thought I'd ask here.

Ideally, I think someone might write code so when "et al." was entered for author, contributor, editor, interviewer, subject, or translator, AND the Wikidata item had at least that many items for that property, Wikidata would automatically add the appropriate "display-authors:", "display-contributors:", "display-editors:", "display-interviewers:", "display-subjects:", or "display-translators:" in what it sent to other projects. Then I wouldn't have to do anything different from what I did, which seemed to me like the sensible, intuitive thing to do in that situation. I refuse to enter all 44 authors into Q107063587 ;-)

Comments? Thanks, DavidMCEddy (talk) 02:00, 5 June 2021 (UTC)

As Wikidata is structured data, we will eventually link Interim Estimates of Vaccine Effectiveness of BNT162b2 and mRNA-1273 COVID-19 Vaccines in Preventing SARS-CoV-2 Infection Among Health Care Personnel, First Responders, and Other Essential and Frontline Workers — Eight U.S. Locations (Q107063587) to items about each author. Those links will allow users of our data to see all of the published works by a specific author. If the author doesn't have an existing item, it can be recorded as text, which will help bots or users to match it to the author item when it is created.
For that reason, Wikidata will be collecting all the author names for a publication. However, if you don't wish to record all 40, enter those that you can manage. A bot or another user will come along eventually and add more authors. From Hill To Shore (talk) 10:02, 5 June 2021 (UTC)
@DavidMCEddy: I have managed to fix your error. Set author 4 as a named individual on Wikidata then set display-authors=3 on the cite template at Wikivetsity. That will leave the correct information here but show "et al" at Wikivetsity. From Hill To Shore (talk) 10:31, 5 June 2021 (UTC)
I see it. Thanks. DavidMCEddy (talk) 13:11, 5 June 2021 (UTC)

Question on an edit

Hi, Newbie here, I added data about a subject "Oduwacoin" two weeks ago but it seems it's yet to be reviewed, I'm not sure how things work here but will I call in an admin to review it or it will get reviewed automatically, or is it reviewed already?Lumared (talk) 09:26, 5 June 2021 (UTC)

@Lumared: there is no "review" process here but you can request review as you just have. I took a pass at cleaning it up for you. One way to learn how to contribute is to look at a similar example and copy how it was done there. So for example look at bitcoin (Q131723) and make Oduwacoin (Q106914879) look like it. BrokenSegue (talk) 13:34, 5 June 2021 (UTC)

Items by numer of labels or descriptions

How I can find items with greatest number of labels or descriptions for each instance of (P31) or subclass of (P279)? Eurohunter (talk) 11:03, 3 June 2021 (UTC)

Hi Eurohunter, I'm by no means an expert (still learning), but I think this does what you want: 
SELECT ?item ?itemLabel ?cnt WHERE {
  {
    SELECT ?item (COUNT(?label) AS ?cnt) WHERE {
      ?item (wdt:P31/(wdt:P279*)) wd:Q810519;
        rdfs:label ?label.
    }
    GROUP BY ?item
  }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
ORDER BY DESC (?cnt)
Try it!
. This should count the labels.
This is the same query, modified to count descriptions. 
SELECT ?item ?itemLabel ?cnt WHERE {
  {
    SELECT ?item (COUNT(?desc) AS ?cnt) WHERE {
      ?item (wdt:P31/(wdt:P279*)) wd:Q810519;
        schema:description ?desc.
    }
    GROUP BY ?item
  }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
ORDER BY DESC (?cnt)
Try it!
.
I've noticed that barge (Q16518) reaches ship type (Q2235308) through two paths, so in this last query, it shows as having double the amount of descriptions it actually has (40 instead of 20). I don't know if that can be fixed using a different property path or if this means there's a better (correct) way to do this. Any experts can chime in? --Azertus (talk) 23:23, 6 June 2021 (UTC)
@Azertus: It works. Thanks. Eurohunter (talk) 12:55, 7 June 2021 (UTC)

Wikidata weekly summary #471

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • You can now add a title to a query visualisation by adding #title:YourTitle to your SPARQL code (phab:T225883, example query with a title)
    • Added a new constraints type to indicate that a certain Property should only be used on Lexemes of a certain language (phab:T200689)
    • Working on not checking the format constraints via SPARQL to take that load off the SPARQL endpoint and make it technically better (phab:T176312)
    • Working on making the API-Sandbox not suggest edits to real Items (phab:T219215)


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.

said to be the same as (P460)

How said to be the same as (Q66209246) is inverse label item (P7087) for said to be the same as (P460)? said to be the same as (Q66209246) isn't even a property. So where is inverse label item of (Q66205187) for said to be the same as (Q66209246)? Eurohunter (talk) 16:31, 6 June 2021 (UTC)

Inverse doesn't mean opposite in this context. It's more clear if you think of a property like "parent". Say you link two items with the property parent: Simba - has a parent - Mustafa. It should now be possible to say something about Mustafa, i.e. in the other direction, without necessarily adding a second statement. The inverse of parent would be child: Mustafa - has a child - Simba.
So since "same as" is a symmetrical property: A - same as - B, the statement when viewed from the other direction, B - same as - A, still should be called "same as". Hope this helped?
By doing it this way we don't have to duplicate every statement from the other direction, but it can still be displayed to the user as if such a property/statement existed. Anyone else reading, if I made any mistakes in the above, please correct me! --Azertus (talk) 23:38, 6 June 2021 (UTC)
@Azertus: Okey but I think its still misslleading in current form and it need better solution. Do you have any idea? Eurohunter (talk) 15:55, 7 June 2021 (UTC)
The English labels seem to make sense. Maybe the pl translation you are viewing is somewhat off?
The use of said to be the same as (Q66209246) is somewhat limited as P460 is generally symmetric.
Not sure if inverse label item of (Q66205187) actually works, but that's another problem. --- Jura 16:41, 7 June 2021 (UTC)

Search results

Why I can find "live result" of phrase like "Category:Songs written by Basshunter" in the left sidebar search but when I press enter there is no results in index.php?search? What I'm missing? What is this search for? Eurohunter (talk) 11:18, 3 June 2021 (UTC)

Maybe the "Category:" prefix takes precedence and only searches through Wikidata internal categories. --Matěj Suchánek (talk) 12:18, 3 June 2021 (UTC)
@Matěj Suchánek: Was it reported to fix already? Eurohunter (talk) 15:58, 7 June 2021 (UTC)
Dunno, it's better to ask / report it anyway. It should be considered, though, that somebody might actually expect the kind of results you are getting. --Matěj Suchánek (talk) 09:39, 8 June 2021 (UTC)
@Matěj Suchánek: Thanks. Reported. Eurohunter (talk) 09:59, 8 June 2021 (UTC)

No results from MWAPI + RDF query?

https://w.wiki/3S7D returns no results, but https://w.wiki/3S7F, just the MWAPI query without the triple (?item wdt:P31 wd:Q1549591) does. I expected wd:Q1461 to show up in both, but no results ever appear whenever I include a triple. What gives? Nivekuil (talk) 11:35, 4 June 2021 (UTC)

Issue is statement rank: Manila (Q1461) has a Preferred Rank P31 - highly urbanized city (Q29946056) - and so looking for wdt:P31 wd:Q1549591 fails ... Q1549591 in Q1461 is not truthy. p:P31/ps:P31 works: https://w.wiki/3S7Z . Wikidata:Request a query is a better place for this sort of question.
@Seav: might like to come here to explain why exactly Manila now has a preferred rank for one of several equally valid P31 values ... the diff seems indistinguishable from vandalism. --Tagishsimon (talk) 12:07, 4 June 2021 (UTC)
Thanks for the help :) Nivekuil (talk) 00:10, 5 June 2021 (UTC)
The convention for local government units in the Philippines is to use the legal/official class for the entity (ex.: province of the Philippines (Q24746), highly urbanized city (Q29946056), component city (Q106078286), municipality of the Philippines (Q24764), barangay (Q61878)) as either the single normal-rank value for the instance of (P31) statement, or as the preferred value if there are other valid values. This is what is expected by the Template:PH wikidata (Q19921792) template in the English Wikipedia. I guess the template can be updated to ignore unexpected P31 values but at the cost of added complexity and computation. —seav (talk) 06:41, 7 June 2021 (UTC)
@Seav: It is really bad practice - the very worst practice: basically vandalism - to have a single template on some language wiki somewhere peversely drive differential ranking of equally valid P31 values on a wikidata item. That template is not the only consumer of wikidata, and other consumers - in this case Nivekuil - have a valid expectation that equally valid values will not be downranked and hence invisible to a wdt: query. Please add "complexity and computation", or, in other words, please fix your badly designed and broken template, so that it does not require wikidata to be degraded for other users, and fix P31 statements that have been demoted to make the broken template work. smh. --Tagishsimon (talk) 11:27, 8 June 2021 (UTC)
Another interesting topic : should wikiarticles be more important/ than wikidata or the reverse ? Should Wikidata outweigh wikiarticles ? Bouzinac 💬●✒️●💛 11:34, 8 June 2021 (UTC)
Should wikidata items be frigged to make badly designed language wiki templates work? Just let me think for a minute. --Tagishsimon (talk) 11:49, 8 June 2021 (UTC)
In that case, no. But what happens if langX wiki says "this is true" and langY says "that is true". Which statement would be set as preferred ? Bouzinac 💬●✒️●💛 12:00, 8 June 2021 (UTC)
Absent other corroborating sources, neither would be preferred. Both would be referenced to their source. The normal caveat emptor would apply. We see this 19 times a day with DoBs and DoDs. Where there is compelling evidence that one is wrong, then we would demote the wrong value. Where there are two values, one more precise than the other, we might well promote the more precise. What we do not do, where a thing is an equally valid instance of X, Y and Z, is set X as preferred merely because someone somewhere has a special interest in the class X and no interest in classes Y & Z. --Tagishsimon (talk) 12:42, 8 June 2021 (UTC)

Problem with wikidata info

Not sure if this request should be here or in a different place, so sorry in advance if I'm in the wrong place. There is something strange with the wikidata of the article "Ramala" in the Spanish Wikipedia, as it shows the country as "Ottoman Empire" instead of "State of Palestine". I've checked the same article in many other wikis (English, French, German, Italian, etc.) and none of them have such problem. Indeed, to my inexperienced eyes, nothing seems to be wrong in the wikidata of "Ramallah" itself. This is driving some editors mad, and I'm sure you'll find it quite easy to solve, so I'm here to ask for help. Thanks a lot in advance.--11koyo11 (talk) 10:38, 3 June 2021 (UTC)

@11koyo11: Fixed Eurohunter (talk) 10:57, 3 June 2021 (UTC)
@Eurohunter: no you didn't. These edits are incorrect. The deprecated rank is only for things that turned out to be incorrect, not for statements that used to be valid, but are not valid anymore. Please have a look at Help:Ranking#Deprecated_rank. Multichill (talk) 16:00, 3 June 2021 (UTC)
So many people get this wrong that I wonder if we should rename deprecated in the UI to something more evocative. Like "wrong". BrokenSegue (talk) 16:32, 3 June 2021 (UTC)
@Multichill: I thought it is to manage display in Wikipedia infoboxes etc. because when something is incorect then should be simply removed. Eurohunter (talk) 06:55, 4 June 2021 (UTC)
You might want to re-read Help:Ranking then. It's one of the key differences to Wikipedia. --- Jura 07:09, 4 June 2021 (UTC)
@Eurohunter: It's not incorrect Ramallah was located in the Ottoman Empire. ChristianKl ❪✉❫ 18:07, 5 June 2021 (UTC)
@ChristianKl: Yes. It was there but infobox lists actuall contry than all historic changes. Eurohunter (talk) 18:47, 5 June 2021 (UTC)
So what? That doesn't change the fact that the statement is correct. We have preferred rank to mark the current statement so that a infobox that wants the best rank gets the current one. ChristianKl ❪✉❫ 20:35, 8 June 2021 (UTC)

Construction period for buildings

Fairly new to Wikidata, so apologies if this is obvious. I've been expanding items relating to listed buildings in Cheshire, UK, including adding the date they were built (property "inception"). Sometimes, though, a building takes several years to build (and the start and end of the contruction period are known). How would I reflect that in the item? An example would be Q26656494, which I've listed as inception=1900, when its National Heritage List for England entry states that construction was 1898–1900. Dave.Dunford (talk) 17:07, 5 June 2021 (UTC)

And a followup, if I may. How would I record an approximate date (e.g. NHLE often says "c.1900")? Dave.Dunford (talk) 17:23, 5 June 2021 (UTC)
@Dave.Dunford: there's a few options. you can make a "significant event" so e.g. significant event (P793) start of manufacturing (Q27136782) or groundbreaking ceremony (Q1068633) or end of manufacturing (Q59913255) and then attach a point in time (P585) qualifier to that event statement. you can see this was done on United States Capitol (Q54109). For uncertain dates wikidata supports dates with different level of precision so you can say "1900" the century or decade or year. You can also indicate that the date is only "circa" by using a qualifier on the date e.g. sourcing circumstances (P1480) circa (Q5727902). BrokenSegue (talk) 17:31, 5 June 2021 (UTC)
Sagrada Família (Q48435) might give some inspiration. Multichill (talk) 21:17, 5 June 2021 (UTC)
Thanks – vey helpful, and useful to have some exemplars. Dave.Dunford (talk) 08:24, 6 June 2021 (UTC)
@Dave.Dunford: For more on approximate- (and other vague-) dates, see Wikidata:Extended Date-Time Format Specification, its talk page, and the linked ticket Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:32, 8 June 2021 (UTC)

IMA Number, broad sense

Hello, I need help on IMA Number, broad sense (P484). Now we have this format constrain: IMA(19[6-9]\d|20\d{2})(-\d{3}[a-z]?|-[A-Z]|-[a-z]{3}| s\.p\.). I would like it to be true for these instances: IMA1984-057, IMA2014-057, IMA1984 s.p., IMA2014 s.p., IMA1984-L, IMA2014-L, IMA84-L and IMA14-L. 'L' can be any letter. Thx n Regards --Chris.urs-o (talk) 16:58, 8 June 2021 (UTC)

Wanted: User script to convert "author name string" to "author"

Suppose I have in my clipboard the QID of an author; and I am viewing the item about one of their works, where their name is recorded using only author name string (P2093).

I would like a user script that will give me a button next to the P2093 value. When clicked, that button would open a dialogue where I would paste the author's QID.

The script would then generate an author (P50) statement, using the series ordinal (P1545) & affiliation string (P6424) qualifiers (including any multiple values), reference(s) and the string value (for subject named as (P1810)), if any, from the statement.

It should reject the change if a corresponding P50 already exists.

It might also remove the P2093 statement.

An example of such changes is in this diff.

Does such a script exists or, if not, could someone kindly make one? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:13, 8 June 2021 (UTC)

As I assume you know, Author Disambiguator does this on a larger scale as an external service. Doing it with a gadget would certainly be useful though. ArthurPSmith (talk) 19:05, 8 June 2021 (UTC)

Citation needed constraint for sex or gender

See discussion. --Epìdosis 19:21, 8 June 2021 (UTC)

Opens only at certain dates/events

How to tell that tramway station 京橋停留場(Q28685195) opens only on certain events/days ? Bouzinac 💬●✒️●💛 19:30, 8 June 2021 (UTC)

This sort of thing? https://www.wikidata.org/wiki/Q6373#P3025 --Tagishsimon (talk) 19:34, 8 June 2021 (UTC)

Q466784 (American Soccer League)

According to the descriptions, Q466784 (American Soccer League) is a redirect page. This is indeed the case in enwiki and cawiki, but not in the others (de, fr, nl, it, pt).

It seems to me that this item should be split. If someone can give me a tool to do that in a few minutes, I'll do it, otherwise I hope someone else picks this up.

Sincerely — bertux 12:19, 4 June 2021 (UTC)

It is not a redirect, but a disambiguation page in en:. The item is for a league, whose correct en: link is en:American Soccer League (1921–1933), but the pages on de:, fr: etc. are conflations discussing all the ASLs. The problem is that the pages on other wikis are so detailed as to become conflations (Wikipedia article covering multiple topics (Q21484471)): fr: has a page for the 1921 league, but not the other ones, for example. Circeus (talk) 16:18, 4 June 2021 (UTC)
Yes, disambiguation, my bad. Surely there is a mess, but what are possible solutions? What about removing the items that are no proper disambiguation pages? — bertux 16:32, 9 June 2021 (UTC)

labels of wikipedia lists and their lemmas

What is the difference between the label for a Wikipedia list object and the name of that Wikipedia list in a given language (de for the remainder of the question)? See (old version before correction of label) [5], which had a label of Liste der denkmalgeschützten Objekte in Innsbruck-Innenstadt/H–K and a sitelink to de:Liste der denkmalgeschützten Objekte in Innsbruck-Innsbruck/H–K (mind the difference -Innenstadt vs. -Innsbruck). What would be the sense of letting both names differ? I want to access the name of the list. If I assume both names are equal

  • {{#invoke:Wikidata|claim|P2817|id=Q1856206|language=de}}

will yield the same value as the much more expensive (2 instead of 1 Wikidata-access)

  • {{#invoke:Wikidata |sitelinkOf|Q{{#invoke:Wikidata|claim|P2817|id=Q1856206|parameter=numeric-id }}}}

see Query for mismatches in Austria and for lang=de (~160).

there are mismatches due to:

Are there any constraints regarding this topic? Should there be more constraints? --Herzi Pinki (talk) 13:11, 9 June 2021 (UTC)

I tried the above mentioned accesses to wikidata and it seems that they are both done with a single expensive object call. --Herzi Pinki (talk) 13:35, 9 June 2021 (UTC)
  • The problem at Q1858636 is that a sitelink to a category is on an item for a list. The English label seems to come from the category.
In general, there shouldn't be any renaming due to lists on Wikipedia being re-scope. The sitelinks should just go to a new item.
If you want feedback on language specific differences (de), please use Wikidata:Forum.
The usual label in English for lists is "list of <something special>"--- Jura 17:58, 9 June 2021 (UTC)

Check if redirect

Is there a template to check if an object is a redirect? Queryzo (talk) 08:36, 7 June 2021 (UTC)

You could add this tool to your own common.js https://www.wikidata.org/wiki/User:You/common.js via this code importScript('User:Matěj Suchánek/checkSitelinks.js'); : it will add a button and tell you whether the wikiarticle contains a redirect.--Bouzinac 💬●✒️●💛 11:16, 7 June 2021 (UTC)
Nice tool too! But I am searching for a template, which gives true for {{Check if redirect|Q105966850}}. Queryzo (talk) 07:52, 11 June 2021 (UTC)
Some page information can only be accessed from Lua, which a template can use as a backend. For querying pages on the same project, there is the Lua Title library (mw.title.new('Q105966850').isRedirect). But this isn't possible for pages on another wiki. Then, there is a hack like mw.wikibase.getEntity(item).id ~= item, but it needs to load all item data. --Matěj Suchánek (talk) 09:46, 11 June 2021 (UTC)

Great, Matěj Suchánek can you insert the new code into Module:Wikidata, so that I can use in Wikitext? 

function p.isRedirect(id)
	return mw.title.new(id).isRedirect
end

Queryzo (talk) 18:53, 11 June 2021 (UTC)

  Done --Matěj Suchánek (talk) 11:55, 16 June 2021 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Matěj Suchánek (talk) 11:55, 16 June 2021 (UTC)

atomic mess

Looking at a few elements such as hydrogen (Q556) and lithium (Q568) I notice that the property mass (P2067) is being misused for atomic weight, a.k.a. relative atomic mass. What should be used is a property analog of the item relative atomic mass (Q41377). But I am unable to find an existing appropriate property. Jc3s5h (talk) 12:09, 9 June 2021 (UTC)

You ignored the preceding talk. -DePiep (talk) 19:04, 9 June 2021 (UTC)
@Jc3s5h: These are merely missing the unit, which should be dalton (Q483261). See aluminium (Q663) for where this is done properly. They should also have determination method (P459) qualifiers to specify the conditions under which these (average) masses are determined. Presumably the issues are mainly because the chemical element properties were added very early in Wikidata's history and aren't up to the usual standards in some cases. ArthurPSmith (talk) 13:37, 9 June 2021 (UTC)
Why is "relative atomic mass" a 'misuse'? You mean to say that is not a mass? @ArthurPSmith:: no the unit is not "missing", the relative atomic mass is defined to be dimensionless (i.e., no unit). This is how the source defines it. (If old trem "atomic weight"is confusing, then forget that one. 'Relative atomic mass' is fine) -DePiep (talk) 19:00, 9 June 2021 (UTC)
Is this based on the question of whether a chemical element corresponds to an individual atom or some collection of them? I mean, the dalton mass unit is defined as "1/12 of the mass of an unbound neutral atom of carbon-12 ..." (see en:Dalton (unit)) so that seems the perfect unit for this purpose otherwise. ArthurPSmith (talk) 19:12, 9 June 2021 (UTC)
As indicated in the citation given the lithium item, https://www.degruyter.com/document/doi/10.1515/pac-2015-0305/html the atomic weight is based on the relative abundance of isotopes in natural terrestrial materials. The atomic weights are expressed in daltons. By stating the "mass" rather than the "relative atomic mass" or "atomic weight" we are being fuzzy about whether we are referring to the average atom on the surface of Earth, the most common isotope, or whatever. Jc3s5h (talk) 00:56, 10 June 2021 (UTC)
I don't think so. I think it is agreed that an element item represents the typical mixture of isotopes. For specific atoms create a separate item. --SCIdude (talk) 06:38, 10 June 2021 (UTC)
@Jc3s5h: determination method (P459) relative atomic mass (Q41377) might be suitable then as a qualifier. We generally don't add special properties for every version of how one might measure something. There is a Chemistry WikiProject that might be a better place to discuss this in more detail. ArthurPSmith (talk) 12:08, 10 June 2021 (UTC)

Abbreviation of a name - can it have its own item?

I created item Q106825653 for the name "Md.", which is an abbreviation used by many authors for the name Muhammad. Another editor merged this item into Muhammad (Q19693229). Is that correct? The author Md. Rizwanul Islam (Q106288113) does not use Muhammad in his works, he uses the abbreviated form Md. as his first given name. His item now states that his given name is Muhammad, but that seems incorrect to me. UWashPrincipalCataloger (talk) 16:32, 10 June 2021 (UTC)

We have items for nicknames like Bob (Q18105736) (which are often used directly as given names) so an abbreviated name that is used the same way seems like it should have an item also, yes. ArthurPSmith (talk) 17:28, 10 June 2021 (UTC)
@UWashPrincipalCataloger: There are even data objects for abbreviations of first names. Please refer Chr. (Q21146553) --Gymnicus (talk) 18:31, 10 June 2021 (UTC)
Is it possible then for an administrator to undo the deletion of the item I created, or do I have to create it again? UWashPrincipalCataloger (talk) 18:41, 10 June 2021 (UTC)
It doesn't need an administrator, but I have reverted @Quesotiotyo's change to Md. (Q106825653). Bovlb (talk) 18:56, 10 June 2021 (UTC)
"Md." belongs as a statement on Muhammad (Q19693229), and perhaps an alias as well. There is no need for it to have its own item. Chemical element symbols do not have their own items. Neither do U.S. state postal abbreviations. There is no separate item for "Jas." or "Tho.", even though those were common abbreviations for "James" and "Thomas" for many centuries, and undoubtedly there are individuals whose first name was only ever recorded that way. If Md. Rizwanul Islam's actual first name is not known, "unknown value" should be used.
--Quesotiotyo (talk) 19:12, 10 June 2021 (UTC)
@Quesotiotyo: I do not think so. Nowhere in the available data is it stated that his name is Muhammad (Q19693229). Instead, only the abbreviation Md. (Q106825653) is used. So we don't know what the abbreviation stands for. The abbreviation Chr. (Q21146553) for example does not necessarily stand for Christian (Q18001597), it can also stand for Christin (Q21132614), Christina (Q1083457) or Q17689481. So if you don't know exactly what the abbreviation stands for, it makes perfect sense to use the abbreviation data object. --Gymnicus (talk) 19:29, 10 June 2021 (UTC)
I agree with --Quesotiotyo. Others use Mhd. No need to make separate items for these. In Turkey we have something called "göbek adı" (belly name) which is a grandparent's name that they give to the newborn. As -generally- it is an old-fashioned name, the youngster does not use it, hiding it behind an initial like M. (It can be Mehmet, Mahmut, Mustafa, Mümin, etc etc etc). --E4024 (talk) 19:33, 10 June 2021 (UTC)
@Quesotiotyo: Chemical symbols and US state postal abbreviations are set by an authoritative oversight body; there is no such body for names generally nor for abbreviations of names (though some countries may have such things). Names are a very personal thing, and there is no reason why they need match anybody else's conception of length, pronunciation, character set or any other criterion you may choose. In the US we have Prince (Q7542) who at one point had a legal name that is not representable in Unicode, and X Æ A-Ⅻ Musk (Q93418989) for instance. ArthurPSmith (talk) 12:58, 11 June 2021 (UTC)

filtering the History of an element on IP contributors

Is there any way to easily filter the history of an element on IPs only ?

On many elements, vandalisms are made by IPs, and on some IPs only vandalize... like Empire State Building (Q9188)

being able to quickly toggle between all edits, and IP edits in history would help analyse whether a specific item is a target for vandals... and thus, know which items need semiprotection...

is there a way or a tool to do so ? or should I propose it on Phabricator ? --Hsarrazin (talk) 13:16, 11 June 2021 (UTC)

I think this could be a good idea for a tool. Fralambert (talk) 15:49, 11 June 2021 (UTC)
I am not exactly sure what exactly you are looking for. If this is just about some IP-username highlighting on page histories in the web UI, you can get this done with some custom CSS in Special:MyPage/common.css. —MisterSynergy (talk) 16:13, 11 June 2021 (UTC)

Provocation

In the case: provocation defence (Q3566104), I have merged all of provocation (Q257931) except "All entered languages" of the Q257931 Because i can't add and edit at Q3566104 more languages, how can i add, Thanks --Bmt3s (talk) 07:31, 8 June 2021 (UTC)

Are you trying to merge them? If so, see Help:Merge. We normally merge to the older item not the newer item. — Martin (MSGJ · talk) 08:42, 8 June 2021 (UTC)
It seems that provocation defence (Q3566104) was originally a legal term (with English label "provocation defence") and distinct from the general term. Perhaps a merge is incorrect here. — Martin (MSGJ · talk) 08:44, 8 June 2021 (UTC)
So the merge is done and most statements on the item are about provocation. The subclass one is about provocation defense, as is the enwiki link. It seems the original poster did not understand what was said above. What's more, there is also provocation (Q6132400) so we now have a duplicate and a missing item. --SCIdude (talk) 06:56, 9 June 2021 (UTC)
I have undone the merge. There were a couple of languages which were incorrectly identified, which I have fixed to the best of the ability. Apart from that I think we are done here. The three items are all for different purposes and should not be merged. — Martin (MSGJ · talk) 21:04, 13 June 2021 (UTC)

Entity untangling help

I just stumbled upon a bit of a mess with Q21517518. It seems as though an import of a VIAF ID (P214) went a bit wrong. James E. King (botanist born in 1940) had information about James A. King (sociologist? born in 1940) imported on top of him. After that, automated edits (and manual ones like mine) added ID's and statements like date of death (P570). I have seen some odd use of initials and interesting career moves before but I'm pretty sure that this isn't one of them and it just needs splitting in two. I thought I'd ask for another pair of eyes on it before that though. I also thought it would be useful to highlight an example what can go wrong when bots fight over each other! Aluxosm (talk) 06:56, 12 June 2021 (UTC)

@Jura1: Glad I asked; I would have just split one of them off and kept the "original". I did have a quick look but couldn't find those docs (think I searched for "split entity" or something). Thanks for the pointer! Aluxosm (talk) 12:40, 13 June 2021 (UTC)
Here are the two entities it was split into: James E. King (Q107212758) and James A. King (Q107212764). Aluxosm (talk) 12:46, 13 June 2021 (UTC)

Taxon merge?

Middle Spotted Woodpecker (Q275591) and Middle Spotted Woodpecker (Q27074888) seem to be duplicates and should be merged, but since I am not a taxon expert and both have quite some history I somehow hesitate to merge them myself. Ahoerstemeier (talk) 09:28, 13 June 2021 (UTC)

I'm not an expert either. But one has parent taxon Dendrocoptes (Q24033503) and the other Dendrocopos (Q723059) which suggests they may not be the same. — Martin (MSGJ · talk) 12:04, 14 June 2021 (UTC)
Thanks, didn't spot that small spelling difference, so they are different taxons, but synonyms. Interestingly, while Dendrocoptes medius catches all the wikipedia links, some wikipedia articles have the wrong taxon name. Ahoerstemeier (talk) 18:53, 14 June 2021 (UTC)

Wikidata weekly summary #472

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Other Noteworthy Stuff
    • Wikidata Community/Diversity 2021 Survey has been published. The results are meant to serve as a baseline to see how the community might (not) change in the future.
    • WikidataCon 2021: A sustainable future for Wikidata. Information about "the conference theme, its three-day program structure and a special project on diversity taking place before the conference itself".
    • The call for candidates for 2021 Wikimedia Foundation Board elections has begun on June 9. The last date to apply is June 29, 2021.
  • Development
    • Working on no longer using the Query Service to evaluate Constraints Checks regular expressions (this should make our checks faster and allow for further improvements) (phab:T176312)
    • Changing the rate limits for assigning Item IDs further, which should result in even fewer Q-IDs being skipped in the future (phab:T284538)
    • Finalizing the concepts for a tool to help work on mismatches between Wikidata's data and other databases/websites/...

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.

Wikidata Community/Diversity Survey 2021 has been published

Hello,

In March, many of you participated in our first community survey (see announcement). We analyzed and summarized the results -- you can have a look at the slides on Wikimedia Commons.

Your answers helped us to get a better understanding of where editors come from and which gender and age they have, as well as what activities you do on Wikidata.

The results are meant to serve as a baseline to see how the community might (not) change in the future. We also hope that the data might inspire community activities around increasing diversity and give some more insights in who your fellow editors are and what they do.

If you have any questions, please do not hesitate to ask. Cheers, -Mohammed Sadat (WMDE) (talk) 15:10, 14 June 2021 (UTC)

IP Masking Engagement – New Update

Hello Wikidata community,

The Anti-Harassment Team has updates on IP Masking here. 

In this update, there are details on the proposal for sharing IP addresses with those who need access and there’s also an update on tool development.

Please pass by and check the new details out and give your feedback.

STei (WMF) (talk) 15:19, 14 June 2021 (UTC)

Kazakh in Turkey (Q64427352)

Kazakh in Turkey (Q64427352). There are Kazakhs in Turkey who came quite recently from their home country. I wonder when they could develop a Turkish version of Kazakh language... Is there anybody here who can remove my ignorance on this issue? --E4024 (talk) 15:26, 14 June 2021 (UTC)

References for values openly supplied by the individual themselves

I am involved in a consultative process to learn how Indigenous artists feel about having information about them and their artistic practice stated in Wikidata. As part of this process we are guiding participants as they add statements on their own person item and we collect feedback from them over the course of the process. The response is so far positive and the process is very instructive, but we face certain challenges with references. Sometimes, we can find external references in support for their statements. Other times, these external references don't exist (either because Indigenous artists don't benefit from the same media coverage as mainstream artists, because the information is part of the oral knowledge within an Indigenous community, or because the information isn't documented anywhere, but the artist wishes to make this information available in Wikidata). Among other things, we lack often external references for gender, year of birth and ethnic group, but the artist are stating these values themselves. Is there a property that could be used to source a statement to the individual themselves? Fjjulien (talk) 20:29, 8 June 2021 (UTC)

How should be ensured that it's really the person making claims? If the person provides information in Wikidata, it goes public. So this info could be put on a web or Facebook page and that page referenced. Is that too much to ask for? --SCIdude (talk) 08:10, 9 June 2021 (UTC)
@Fjjulien: While Wikipedia looks down on sources that are controlled by the subject of an article, those sources are okay for Wikidata in circumstances like this. While an Indigenous artist might not have mainstream media coverage, in most cases they will have their own webpresence, whether that's Facebook/Twitter or Soundcloud. If the artist wansts gender/year of birth/ethnic group to be publically known it would make sense to have that information on their own webpresence.
If you discover during your consultative process that there are artist that currently don't have any kind of webpresence, can you tell us more about them? ChristianKl ❪✉❫ 14:00, 9 June 2021 (UTC)
@SCIdude: "How should be ensured that it's really the person making claims?" – In this particular case, we are guiding the artists over videoconference, and they are indeed providing these values themselves. But I acknowledge there are many other circumstances where it would not be possible to assert if the person has indeed made the claim themselves. Fjjulien (talk) 15:26, 9 June 2021 (UTC)
@SCIdude: @ChristianKl: You are right. Most of the time, these artists have their own website from which certain statements can be referenced. This is entirely possible with properties such as "occupation", often possible for "ethnic group", but more difficult for "sex or gender", "date of birth" or "country of citizenship". Artists do not usually not share this kind of information on their website. However, when we asked in a safe context and explained why this information can be useful for disambiguating entities, we are finding that some are happy to disclose this information. I wish there was a way to source these statements as self-declared by the subject (even though such a reference would not be verifiable). I guess, I could create an item for this consultative process and then use stated in (P248). But that's a rather intricate solution that would only meet the need in these few particular items. If there's no right solution, then so be it. We'll leave the claims without references for now.
Please note – We have been reaching out to performing arts associations, inviting them to collect year of birth and self-identification from their members + seeking their consent for disclosing this information in Wikdata. This has worked really well with the catalog for the project of the Conseil québécois du théâtre (Q100443909). We are currently in discussion with the Indigenous Performing Arts Alliance. We hope that we can build enough trust with the association and their members to perform a similar batch upload. The consultative process described above is part of building trust (and gaining confidence that Wikidata is deemed by Indigenous performing arts community as an appropriate knowledge base for publishing this kind of information about them). Fjjulien (talk) 15:52, 9 June 2021 (UTC)
Can you tell us more about who suggested this project? While it may be acceptable in some languages of Wikipedia's that users write articles about themselves, it's not ok in most. Wikidata isn't designed either as place of primary publication. --- Jura 18:24, 9 June 2021 (UTC)
@Jura: This project was suggested by members of the advisory committee for the project grant Modelling and Populating Performing Arts Data in Wikidata. Members of this advisory committee felt it was not ok for Wikidata to feature mostly white men and they devised a project by which they could identify obstacles that led to the under-representation of Indigenous artists. They thought that, by involving Indigenous artists in the project, they might find a few champions who would then encourage other Indigenous artists to get involved with Wikidata. Wikidata is indeed not designed as a place for primary publication. But if that's an obstacle to greater participation by Indigenous Peoples, then perhaps it should be addressed. Fjjulien (talk) 01:55, 11 June 2021 (UTC)
@Fjjulien: Maybe. But this can only work if an account is linked to an identity. Just imagine I create such an artist account and make false claims. Would that be good for Wikidata? Note I'm not against such "identified" accounts: they would also allow data curation by specialists---their claims could have the same authority as referenced claims (the remaining problem being the inherent difficulty of extracting the author of a claim). --SCIdude (talk) 09:09, 11 June 2021 (UTC)
@Jura: Indeed, the "advisory committee" is the same as the project's "core working group" (the terminology evolved over time – sorry about the confusion). Members of this group devised this project although not for themselves: for the hundreds of Indigenous artists who were not involved in the project. We put out a call for participation, and then we guided these artists as they created their own items. We used the "user acceptance test" methodology to collect their verbal and non-verbal feedback. We also took time to specifically inquire about their feelings vis-a-vis properties denoting aspects of identity/belonging to a group. So far, we created two items: Mike Alexander (Q107139441), Monday Blues (Q107155126). We were able to rely on these artists' websites to source the ethnic group (P172) statements, but not for citizenship, gender and year of birth: Indigenous artists are using their website to describe their practice, not their feeling about being a citizen of a nation state that has purposely and repeatedly attempted to assimilate and destroy their culture. Hence my question for finding other ways for referencing these statements as coming from the persons themselves. If no mechanism exists to meet this need, then so be it. This will be another finding from this project that we can share back with the Indigenous arts community (and possibly discuss at the WikidataCon). Thanks for your response. Fjjulien (talk) 21:31, 11 June 2021 (UTC)
  • @Fjjulien: There's an RfC about signed statements that's about creating functionality in Wikidata that proposes a way a person or organization can claim to create a statement. It might be worth to read it and comment how you think it interacts with your use-case. ChristianKl ❪✉❫ 20:44, 12 June 2021 (UTC)
@ChristianKl: Thanks for pointing this RfC out. I will certainly read, and I will comment if I have anything relevant to contribute. Fjjulien (talk) 03:35, 15 June 2021 (UTC)

Notable Nigerian Sport Ambassadors

I have applied for a new project on the above subject matter. you suggestions will be appreciated. the link is https://meta.wikimedia.org/wiki/Grants:Project/Rapid/kid_keen_47/Notable_Nigerian_Sport_Ambassadors  – The preceding unsigned comment was added by Kid Keen 47 (talk • contribs) at 18:59, June 14, 2021‎ (UTC).

Also at Wikidata:Project_chat/Archive/2021/05#Grant_request:_Nigerian_Sports_Ambassador_on_wikidata. --- Jura 08:16, 15 June 2021 (UTC)

Separating Hansard (Q524352) into separate concept items

Hansard is the name of a long running publication in the UK that provided transcripts of the proceedings of the Parliament of the United Kingdom (Q11010). The name was then used for similar publications in other territories of the British Empire, or the successor countries. Item Hansard (Q524352) represents the modern UK database of hansard and is tied to Hansard (2006–March 2016) ID (P2170). However, the site links and some of the statements link the item to the similarly named publications in other countries. We also have Hansard 1803–2005 (Q19204319) and Hansard (1803–2005) ID (P2015) for a database hosting historic copies of the UK Hansard. I think Hansard (Q524352) needs splitting into an item covering Hansard in general (with potential links to items for each country), Hansard the UK publication from 1812 and Hansard the modern UK database (to link to Hansard (2006–March 2016) ID (P2170)). Does this sound like a reasonable idea? Should I keep Q524352 as representing one of these concepts or move all statements to fresh items and delete the original? From Hill To Shore (talk) 00:42, 15 June 2021 (UTC)

It sounds as though splitting might not leave enough connection between the present-day British Hansard and the legacy version of the same thing; Hansard in general would not be enough to specify which country's parliament is referred to. Additionally, while there will be a great deal of cross-over in membership between the pre- and post-database aera at Westminster, that will not be the case between Canberra (say) and Westminster. I think there ought to be way of grouping the original Hansard with its legacy editions, data, people. while keeping separate the Australian, Canadian, etc. GPinkerton (talk) 07:52, 15 June 2021 (UTC)

Usecases for qualifiers defining exceptions to constraints

Please help expand https://phabricator.wikimedia.org/T236295 --- Jura 06:08, 15 June 2021 (UTC)

Incorrect Merge

It looks like the item Linearity Q1753188 [6] was erroneously merged into laser beam welding (Q937468). Will somebody please help revert? — The Erinaceous One 🦔 06:47, 15 June 2021 (UTC)

@Burakzdd, The-erinaceous-one: I have reverted the merge but that has left tr:Doğrusallık without a Wikidata item. Is anyone able to read the language and able to confirm where it should be linked? Some of the content suggests linearity (Q1753188) may be a good match but the article also seems to cover other topics, so it may need to sit on a separate item. From Hill To Shore (talk) 07:17, 15 June 2021 (UTC)
There is a Google Translate plugin for Chrome that offers translation of any page you are on. This one translates to: "Linearity is the property of a mathematical relationship ( function ) that can be represented graphically as a straight line" and the other meanings are mostly just applications of the principle (that's why math is fundamental btw), so I added it to Q1753188. --SCIdude (talk) 16:12, 15 June 2021 (UTC)
So sorry. That's my fault. Linearity must be related to Doğrusallık. But I couldn't. --Burakzdd (talk)

Nature News: scholarly article (Q13442814) or article (Q191067)?

I've just disambiguated a writer for the News division of Nature (Q180445) by the name of Rex Dalton (Q107241085). As far as I can tell, all of his works (query) are news articles, which (if I'm understanding this correctly) are not peer reviewed. Apart from Q28302723 (which I've already changed) they're currently all instances of scholarly article (Q13442814). Would it make sense to do a batch edit and change them all to article (Q191067)? In this case that would be easy enough to do, but unfortunately only a few (like this one) have the string "NEWS" as part of the DOI so I don't see an easy way to fully automate this for Nature news articles from different authors. Aluxosm (talk) 09:35, 15 June 2021 (UTC)

I think it would be fine to add such a P31 statement. I also think "scholarly article" should be changed to "article in scholarly journal". --SCIdude (talk) 16:27, 15 June 2021 (UTC)
@SCIdude: Thanks for your input. To be clear though, I would be replacing the P31 statement, not adding to it. Also, I'm not sure about the renaming, a similar entity already exists — academic journal article (Q18918145). All very confusing! Aluxosm (talk) 11:50, 16 June 2021 (UTC)
Yes, article (Q191067)? is more appropriate for these and, by extension, many more. Looking at the results to the linked query, I also noticed that there seem to be quite a few duplicates? For the one example I checked out, California approves $1.5 billion campus (Q58926806) seems to be correct while California approves $1.5 billion campus (Q46529753) contains a DOI that leads to an entirely different article? (Edit: it’s fewer than I first thought. They just happened to cluster at the top of the list)Karl Oblique (talk) 21:37, 15 June 2021 (UTC)
@Karl Oblique: Good spot! I'll have to add that to the to-do list. To further complicate matters, I've just come across news article (Q5707594) 😬. Need to have a think before I go much further. Aluxosm (talk) 11:50, 16 June 2021 (UTC)

Are VIAF entries supposed to be unique?

We have had Edmund Glover (Q21455684) and Edmund Glover (Q5339502) with the same VIAF and WorldCat_Identities_ID. I think both belong exclusively to the actor, the DNB entry does not mention that the actor was also a painter, so they seem to be different people with coincidental birth and death dates. If we had a reminder that a VIAF is used in another entry, like we do for Findagrave entries, it would help find potential duplicate entries, and help identify inadvertent conflations. --RAN (talk) 13:32, 16 June 2021 (UTC)

I assigned them exclusively to the actor. --RAN (talk) 15:30, 16 June 2021 (UTC)

Q822835 Bernarda Morin

I have some questions about this person:

  • her birth name is Vénérance. She was born in Quebec, Canada. When I try to add her birth name to wikidata; "Canadian French" is not an option for (mandatory) language. How to proceed?
  • her english wikipedia page is titled "Mother Bernard Morin"; but "Bernard" is a masculine name; surely this should be "Bernarda"?

Thanks. Fbax (talk) 14:09, 16 June 2021 (UTC)

@Fbax: :
  • in the dropdown menu, you can select "fr-ca", I did a test here: Q4115189#P1477.
  • yes, to be checked with references but it's likely a spelling mistake
Cheers, VIGNERON (talk) 16:48, 16 June 2021 (UTC)

Repeated informations

In Lina Medina (Q202307), number of children (P1971) have a problem: as you can see in the image, the number 2 is repeated. It should be "1 and 2", not "1, 2 and 2". Why there is this problem?

https://i.postimg.cc/MZPd8j68/Lina-Medina.jpg

--93.35.184.189 15:28, 11 June 2021 (UTC)

Sorry, fixed. --Infovarius (talk) 21:32, 17 June 2021 (UTC)

Sixth chord

Could I please have some assistance with Q28089342 and sixth chord (Q515989)? I'm thinking that we ought to split this into three items: the classical music sense; the modern popular music sense; the conflation of the two as a disambiguation page. I'm having a little trouble assigning the various sitelinks and redirects to these three buckets. Thanks, Bovlb (talk) 00:05, 16 June 2021 (UTC)

This is done, sort of. A dab page item is not possible, since there are actual articles about the conflated concepts which are technically not dab pages and full of useful content. I have moved everything into a new item, since the old one was not properly defined anyways. —MisterSynergy (talk) 16:11, 17 June 2021 (UTC)

Item edit request

Is there semeone that can kindly edit Tatar (Q25285)? "Tt" on Wikipedia entries is wrong: please delete "Татар теле" and write "Tatar tele". Many thanks in advance!!! --93.35.184.189 14:41, 17 June 2021 (UTC)

Help with merging

The New York Kouros (Q28810545) and New York Kouros (Q29383870) are the same statue, but I'm scared to merge these two items myself because I don't want to mess anything up. I know it may sound lame, but if someone could help hold my hand through this process, that would be great. Thanks, Tyrone Madera (talk) 16:21, 17 June 2021 (UTC)

@Tyrone Madera: I've done the basic merge. There was a reference on one statement for Q28810545 that pointed to Q29383870, which I had to delete before the merge could be done. You can complete cleaning up the merged item - there are some duplicate or not-quite-duplicate statements there now. ArthurPSmith (talk) 17:15, 17 June 2021 (UTC)
@ArthurPSmith: Thank you! Tyrone Madera (talk) 17:20, 17 June 2021 (UTC)

Subway modelling on adjacent station (P197)

Hello, I have been unfriendly reverted on this https://www.wikidata.org/w/index.php?title=Q1340440&oldid=1442360329 edit. Adding qualifyers to adjacent station (P197) is useful : the qualifyer might be redundant but very useful to distinguish whether there are multiples lines served on a station (eg on other examples : https://www.wikidata.org/wiki/Talk:Q462201). I have been reverted because I did on one item, not on all 200+ Moscow subways. I don't have time to edit all of them very quickly before being reverted again by an individual. Bouzinac 💬●✒️●💛 20:48, 15 June 2021 (UTC)

@Michgrig: Per P197#P1855, there is consensus that connecting line (P81) is a valid and necessary qualifier for adjacent station (P197) and so your revert of Bouzinac was not helpful. --Tagishsimon (talk) 21:19, 15 June 2021 (UTC)
@Tagishsimon, these are just examples, not the clear instruction. As I see it, if a station belongs to one line, adding the line as a qualifier to adjacent station (P197) is simply redundant. Any automation must be flexible enough to get a single value from the connecting line (P81) property. If a station belongs to multiple lines, then qualifiers are needed. You can find examples of this in Moscow Metro (e.g., Park Pobedy (Q1810822)) and on railways in Moscow Oblast (e.g., Dmitrov (Q4162925)). Michgrig (talk) 21:29, 15 June 2021 (UTC)
@Tagishsimon P.S. Please note that in the examples that you gave the link to, all of the stations belong to more than one line. Therefore, using the line qualifier fits into my logic. Michgrig (talk) 21:33, 15 June 2021 (UTC)
The trouble with your logic, @Michgrig:, is that it takes no account whatsoever of expectation by others of the data model. The connecting line is unambiguously an attribute of any reasonable model for adjacent stations, and it is completely unreasonable to suggest that the adjacent station <--> connecting line model should differ according to whether there is a single connecting line or multiple connecting lines - that one model should hold the data as a qualifier of P197, the other as a main property of the station, and that those querying for station data, for instance to draw maps, must make already complex queries more complicated because of your prejudice against storing valid data. I can point you at evidence that the connecting line qualifier is an expectation of all of the examples of this art that I've seen & been involved in in the last few years. I can point you at the property definition which specifies that the connecting line qualifier is an expected or sanctioned qualifier for the property. What, your prejudice aside, can you point me at to support your view? --Tagishsimon (talk) 22:00, 15 June 2021 (UTC)
I do not have any prejudice. In my support, I can point you at the Occam's razor principle: entities should not be multiplied without necessity. Please give me links to the discussions that support your point of view and to examples of ready maps that use it. Michgrig (talk) 22:10, 15 June 2021 (UTC)
Does Occam have anything to say about why two distinct models are preferable to one? Because that - a single model - seems to be highly desirable. I'm not aware of any discussion for or against the idea that WD should have two distinct models for adjacent station; no-one before you has suggested that would be a sensible way to go. Ready maps that use it: [7], for instance. --Tagishsimon (talk) 22:43, 15 June 2021 (UTC)
@Tagishsimon OK, at least I've seen a live example of using this. I agree to follow this "single model" and will try to add qualifiers to all stations of Moscow Metro and railway.
However I still have one question to @Bouzinac: why in the middle of tons of edits in stations of China, you decided to edit one Moscow Metro station? Michgrig (talk) 21:07, 16 June 2021 (UTC)
Some rapid transit (Q5503) and tram system (Q15640053) eg : Strasbourg have been modeled this way and it's perfect. But it's true it is not yet the case for every subways/tramways systems. And modeling takes some time (it took me 3 days to build the Shanghai map, in finding, cleaning, modeling). Bouzinac 💬●✒️●💛 05:28, 16 June 2021 (UTC)
@Bouzinac: Moscow Metro was modelled without such qualifiers and it's perfect too: https://w.wiki/UE8 (even smaller query!) --Infovarius (talk) 21:54, 17 June 2021 (UTC)
The way of writing the query is really nice, but there might be some weirdnesses : look closely to

Перово(Q2550994) and Новогиреево(Q662318) in the map ? Data might perhaps need some data cleansing (but where? didn't spot what was wrong, perhaps the duplications of coordinates in the items or a better filter of coordinates?), but you definitely proved there is no need to add qualifyers. отличная работа ! --Bouzinac 💬●✒️●💛 22:22, 17 June 2021 (UTC)

Botched police raids

I am trying to aggregate "botched police raids" https://www.wikidata.org/wiki/Special:WhatLinksHere/Q107238003 I was searching for the phrase on Google and matching them to our entries, but it seems easier to just use my memory, and then find a find a reference using the wording "botched raid" in the news. Can anyone think of others to tag? --RAN (talk) 03:05, 15 June 2021 (UTC)

That looks like a very subjective item. I can see a lot of future arguments about whether a raid that doesn't meet all of its objectives should be called "botched." There isn't any practical definition of when a raid goes from being not-botched to being botched, so views could differ significantly. We will likely end up with cases where one journalist says a raid was botched and another who says it wasn't botched. Is it possible to come up with a more precise concept that covers what you are wanting to map? From Hill To Shore (talk) 07:25, 15 June 2021 (UTC)
Hmm, botched police raid (Q107238003) = failed police raid, when a police raid (apart from the fact that this term is already imprecise, seems to be US-centric and may have little reflection in the techniques of police work in other countries) can be called "failed"? Wostr (talk) 11:51, 15 June 2021 (UTC)
  • When the phrase appears in a news article it is 100% objective. GHits are 230,000 for "botched raid" with quote marks. Sorting through these hits to see if they correspond to Wikidata entries is hard, it is easier to use your memory then search for the phrase "botched raid" in conjunction with that event. See for instance "botched raid"+"ruby ridge" to get Q1604442. The early ones I added have the phrase in the corresponding Wikipedia article. It is no different than the aggregation we do for "police brutality" which can be argued is subjective but when a reliable source uses the phrase it is objective. --RAN (talk) 12:53, 15 June 2021 (UTC)
  • That's a very poor understanding of news media you have there, RAN. As a simple for instance, where a news article reports source A saying "x was a botched police raid" and source B saying (in effect) "x was not a botched police raid", the news story is not 100% objective proof of a botched police raid. There's very clearly a wide and complex spectrum from successful/good police raid through to botched/unsuccessful/bad police raid. You've coined an item which has a ridiculous combination of label and description (does botched=failed? does failed=botched?). Your item has no references. Your "but it seems easier to just use my memory, and then find a find a reference using the wording 'botched raid' in the news" is nothing more than the Garbage In element of GIGO. There is no definition of scope, no criteria whatsoever for evaluating whether instances are in scope of out of scope. Really, the whole thing is risible. --Tagishsimon (talk) 13:18, 15 June 2021 (UTC)
And I see there's more misunderstanding on your part. You have coined raid (Q107238062) with has part(s) (P527) values of "botched raid", "military raid" and "police raid". THIS IS NOT HOW WIKIDATA WORKS. "botched raid", "military raid" and "police raid" may indeed be subclasses of a raid, but WD does not list subclasses using has part(s) (P527), but rather by the subclass having a P279 pointing to the class. I fear to explore how much further you've gone with this line of work; how much more damage you've inflicted on WD. I see worrying signs of complete nonsense on police raid (Q97368680), where that item is now both a subclass of military raid (Q476807) and different from military raid (Q476807). The impression left is that you are well meaning & enthusiastic, but lack basic competence, and so unwittingly leave a trail of mess in your wake. --Tagishsimon (talk) 13:50, 15 June 2021 (UTC)
  • STOP YELLING! And you can be less pompous too by saying I "lack basic competence", keep your personal animosity toward me to yourself. Give me an example of one I listed where a reliable source used "botched raid" where it is contradicted by another reliable source saying it "not a botched raid". We also already have Category:Police misconduct (Q8768020) and Category:Police brutality (Q8768009) where I am sure you can find someone arguing that there was no misconduct or no brutality. No one else demands that 100% of the sources have to agree to have an entry. If you hate the phrase "botched raid" remove it from each Wikipedia article, or better yet, make your argument in the Wikipedia article on Breonna Taylor on why it was not a botched raid, remember to add in your reliable sources. Data integrity is important, but just a reminder, we have over 100,000 data points from reliable sources that had to be deprecated and marked as "typographical error" or "implausible" or "contradiction". We also have over 100,000 data points from reliable source for date of birth or date of death where two different years or two different dates in the same year are given. We average five entries a week where someone died before they were born. We have a list of humans living over 120 years waiting to have corrections made, or marked somehow as having a fictional lifespan. Despite your concern, I don't see you participating in any of the projects that make corrections for these errors. --RAN (talk) 12:52, 16 June 2021 (UTC)
  • When the phrase appears in a news article it is 100% objective – no, it's just source X stated that incident Y was "botched", nothing more. News articles are far from being objective. Also I don't think that such imprecise and poorly-sourced data should be added to WD, it's far from the concept of structured data, it looks like a tag, not a proper class. Police actions or operations could be classified more precisely with classes like police operation ended with the death of a civilian(s) – such classes can be properly structured, do not carry subjective opinions with them. However, (1) not every police operation ended with the death of a civilian could be classified as a police misconduct or police brutality; (2) with botched police raids class, I really don't know what kind of information you want to add to WD? What happened during the operations classied as botched police raids that allows them to be called botched or failed? Without more precise definition, the only thing that botched police raid (Q107238003) means now is police operations which have been called "botched" in the media, i.e. it means really nothing. Wostr (talk) 14:18, 16 June 2021 (UTC)
I think you are trying to distinguish the difference between absolute truth and the information found in reliable sources. We can only record information found in reliable sources. Again, they are called "botched" because that word appears in the reliable source and the corresponding Wikipedia article. You are welcome to add in reliable synonyms for the concept if you do not like use of the word "botched". Some concept entries in Wikidata have a half-dozen synonyms. You are welcome to edit each Wikipedia article which contains the phrase and remove it, if you think the source is unreliable. It is the concept that is important, even if reliable sources use synonyms. You can write The New York Times and complain of their use in the following articles: https://www.nytimes.com/search?query=botched+raid and the Washington Post: https://www.washingtonpost.com/newssearch/?query=botched%20raid&btn-search=&sort=Relevance&datefilter=All%20Since%202005 which contains both military raids and police raids. If you do not want the concept in the WikiUniverse, you have to stop the use by reliable sources first. You wrote: "it's just source X stated that incident Y was …" but isn't that how every data point in Wikidata is added? That is why some entries have three contradicting years of birth. Again this is no different than how we add entries to Category:Police misconduct (Q8768020) and Category:Police brutality (Q8768009). -RAN (talk) 16:01, 16 June 2021 (UTC)
Yes, reliable sources! News articles are not such sources in most situations. Also, we are not writing an encyclopedia here; in Wikipedia citing a source is allowed and you can write that X, Y and Z called something "botched". In Wikidata we are trying to establish a structured database with data that is objective (neutral), accurate and unambiguous, some information may not be needed here or have to be written differently than in Wikipedia. So botched is not an adjective that has any meaning in this situation in relation to the situation. It can be "botched" in so many different ways, that – as I said – the only thing that this class means in Wikidata is that someone in source X said that this operation was botched, nothing more. I don't think that we need such information in Wikidata. What we need, in my opinion, is a set of objective, structurally linked information why someone might think an operation was botched. And I don't have a problem with botched police raid term at all if it appears in the media, even in Wikipedia in a proper context, however, I have pointed out to you what ambiguities I see with using this term in Wikidata, I tried to show you a better, in my opinion, approach to classification of police operations which did not go as planned. Sadly, you seem to have some kind of "mission" to add this information exactly like this, no matter what the arguments against. Also, you failed to answer my questions and I don't see any willingness to find a common ground here, so I don't think there's a room for a constructive discussion here. Wostr (talk) 16:43, 16 June 2021 (UTC) PS I don't know what do you mean by linking Category:Police misconduct (Q8768020) and Category:Police brutality (Q8768009) – these are just Wikipedia categories and I see no connection to this problem; every project has its own rules, guidelines and – the most importantly – way of presenting knowledge. Wostr (talk) 16:53, 16 June 2021 (UTC)
@Richard Arthur Norton (1958- ): The key difference with misconduct is that there is a factual standard to apply. There will be a standard of conduct that has either been breached or not breached. We can have accusations of misconduct, investigations of misconduct, charges of misconduct, trials of misconduct and verdicts about a misconduct trial. All are factual statements with objective criteria. Similarly, brutality is excessive use of force beyond what is normally allowed. We will normally have a set of rules on a level of force allowed in a society, which are either breached or not breached. There can be accusations of brutality, investigations of brutality, charges related to brutality, trials regarding brutality and verdicts about the trial. Again, these are factual statements linked to objective criteria.
For botched police raid, it is unclear what standard is being applied. There is no objective standard for judging success or failure of the raid. Each raid will have a unique set of goals to be achieved and witnesses will apply their own subjective assessment of the outcome.
Here is a hypothetical scenario. A police raid is conducted on a drugs cartel and $1m worth of drugs are seized, 10 gang members are arrested but 3 escape and another person related to a gang member (but not an active participant in the gang) is killed.
Source A says it was a botched raid as some of the gang members escaped and an innocent person was killed.
Source B says it was a successful raid as a large amount of drugs and many of the gang members were captured. The killed person was involved in the crime (even if it was to the extent of not revealing the activities to the police) so was not an innocent party.
Source C says the results were inconclusive as many but not all of the goals were met.
We create an item about the incident and mark it simultaneously as an instance of botched police raid and an instance of successful police raid. A later editor creates an item for inconclusive police raid and adds that to the item.
What useful data can we draw from that item? From Hill To Shore (talk) 18:18, 16 June 2021 (UTC)
  • Can you give an example of the raids listed here at Wikidata that are inherently ambiguous? It seems people do not like the term "botched" because it sounds colloquial, even though it is the exact word used to describe the events in reliable sources. --RAN (talk) 18:48, 16 June 2021 (UTC)
    • A rename alone won't fix the problem. "Failed police raid" is equally ambiguous and is a subjective judgement by the source. The best we can map from that is "this source thinks it failed." Your earlier comparison to age data doesn't really fit here. A birth date is a fixed value; while sources disagree, we can assess the quality of the sourcing and come to a preferred value (that preferred value may even be a date range covering the sourced dates). The same situation doesn't apply here where there is no preference; an item could be equally failed, succeeded and inconclusive all at the same time. I'll take a look later to see if any of the current raid examples are flawed but I'd like to declare now that it is a straw man argument. It is not the current minimal implementation that is my primary concern but the potential expansion of an item with unclear parameters. You will probably have applied some level of consistency to the current data as you have a clear picture of what you want to map. The problem comes when future editors try to interpret the item and apply their own subjective judgement. If you can rationalise the concept you want to map into objective statements then there won't be an issue. From Hill To Shore (talk) 20:03, 16 June 2021 (UTC)
      • Looking at Harding Street raid (Q65083136) there are several objective statements that could be made. 1) Police raid involving an illegal warrant, 2) Police raid involving a fatality, 3) Police raid based on a false report of a crime, 4) Police raid involving injury of police personnel, 5) Police raid using "No knock" tactics, 6) Police raid involving falsified evidence. While quite a few sources do use the word "botched," that doesn't change the fact that a decision to classify items using this or a similar term relies to a high degree on the subjectivity of sources. Sooner or later there will be an item where the sources are divided on the success or failure of a raid and we will end up in a quagmire. It is better to change tack now while the concept is linked to a few items and agree a less subjective concept to apply. From Hill To Shore (talk) 21:29, 16 June 2021 (UTC)
English Wikipedia dropped religion and at one point sexuality from infoboxes because of something similar to "sources are divided", but we kept both at Wikidata with references. If potential-for-abuse is a reason to not add something, we should drop date_of_birth and date_of_death. I have to correct a half-dozen a week from people changing those dates randomly as vandalism. I only catch the ones that cause person to die before they are born. Subtle vandalistic changes go undetected, which is why English Wikipedia does not use Wikidata information for infoboxes. Every field has the potential to be abused. --RAN (talk) 17:45, 18 June 2021 (UTC)
I think you are missing the points that we are raising here. It is not a question of abuse. It is a question of whether this effort will produce meaningful data. I already explained about age information above, please read what I said there. With religion, a person or location can be part of one religion for a period of time and then convert to another religion at another point in time. We can set date ranges or other qualifiers to explain the relevance of the situation. Here you will have many items that are both "failed raid" and "successful raid" at the same time. It would be like allowing claims of both "good man" and "bad man" on a single item for a controversial politician. From Hill To Shore (talk) 18:01, 18 June 2021 (UTC)

Merging duplicates

There is an instance of multiple items for the same subject, The items Q106977647, Q107261414 and Q107176477 are all the same and should be deleted or merged to the more elaborate one which is Q107177453. I would have done this but I really don't know how it's been done here. Lumared (talk) 10:52, 18 June 2021 (UTC)

@Lumared, I've merged. Please check and remove duplicated statements. Michgrig (talk) 12:00, 18 June 2021 (UTC)

Ignored request

Why my request was ignored? --93.35.184.189 13:48, 18 June 2021 (UTC)

"Tartar tele" is used by four different language labels on Tatar (Q25285). Which language are you wanting to change? From Hill To Shore (talk) 14:01, 18 June 2021 (UTC)
In that page, when I click on "Tatarça / Татарча" (is the number 35) I enter on this page, but the page dont have language. It should be update the wikidata item to best reflect what https://meta.wikimedia.org/wiki/List_of_Wikipedias page have or update the page to best reflect the wikidata item have. --93.35.184.189 14:36, 18 June 2021 (UTC)
Are you talking about linking the Tatar article about the Tatar language to articles in other languages? If yes, then we cannot do this because you have another article, https://tt.wikipedia.org/wiki/%D0%A2%D0%B0%D1%82%D0%B0%D1%80_%D1%82%D0%B5%D0%BB%D0%B5 that is linked with other languages. It's up to the Tatar community to decide which article to keep and which to delete. Michgrig (talk) 15:03, 18 June 2021 (UTC)
You dont understand. There is a page on WIKIMEDIA server (https://meta.wikimedia.org/wiki/List_of_Wikipedias) that link the wrong page. --93.35.184.189 15:41, 18 June 2021 (UTC)
We can change the linked page on the Wikidata item, if it is wrong. What is the right page? From Hill To Shore (talk) 16:30, 18 June 2021 (UTC)
It looks like our item links to tt:Татар теле but the Meta page links to tt:Tatar tele. Which page is correct? If the page linked here is correct then I am not sure if I can help. It sounds like a problem at Meta. From Hill To Shore (talk) 16:41, 18 June 2021 (UTC)
@From Hill To Shore: https://tt.m.wikipedia.org/wiki/Tatar_tele --2001:B07:6442:8903:A9D7:ABF5:D770:AF9 16:46, 18 June 2021 (UTC)
I have run the articles through a translation tool and they appear to be duplicating the same subject. One is written in Latin script and the other is written in Cyrillic script. I am not comfortable making this change for an unknown IP editor. The Tatar community needs to decide on a solution; we can't implement the preference of one unknown editor. From Hill To Shore (talk) 17:11, 18 June 2021 (UTC)

Correct way to link items of floklore to places

Hi all. I am struggling to find a reliable standard for a specific information. If an item of a tradition or a character is linked to a certain area, how do you state it?

For example:

Do we need a specific new property for that? Would you rather use a more generic one with some appropriate qualifier? Please notice that these items are usually neglected, consisting of minimal information and one simple link to a local wiki... so this improvement would IMHO really help to scale up their quality.--Alexmar983 (talk) 20:17, 18 June 2021 (UTC)

Probably indigenous to (P2341), I guess. --Epìdosis 20:19, 18 June 2021 (UTC)
location of creation (P1071), significant place (P7153) and set in environment (P8411) may also be useful. Folklore may be indigenous to one area but be associated with a different location. For example, a Japanese folk tale about China would be indigenous to Japan but set in China. Perhaps we use either location of creation (P1071) or indigenous to (P2341) to signify where the folklore comes from (indigenous to (P2341) may be useful where the location of creation is not known) and then set in environment (P8411) to identify where the story is set. From Hill To Shore (talk) 21:00, 18 June 2021 (UTC)

Scale of objects

I need to know the scale of objects. Please include the page "scale of objects".  – The preceding unsigned comment was added by 104.13.58.13 (talk • contribs) at 23:01, 18 June 2021 (UTC).

Wikidata is a structured data project that anyone can edit. You are welcome to create an item if you think we are missing something. There is always more work to do than there are participants to do it. New editors are welcome. From Hill To Shore (talk) 23:05, 18 June 2021 (UTC)

Watchlist request

Q12140 and Q179661 were recently edited by newcomers who didn't quite grasp the concepts. The first is medication (e.g., aspirin) and the second is medical treatments (including non-drug treatments, such as surgery). I've reverted the errors, but it might be helpful if a few people kept an eye on those articles for a few days. WhatamIdoing (talk) 15:59, 19 June 2021 (UTC)

Please provide input here or on Meta and during an upcoming Global Conversation on 26-27 June 2021 about the Movement Charter drafting committee

Hello, I'm one of the Movement Strategy and Governance facilitators working on community engagement for the Movement Charter initiative.

We're inviting input widely from users of many projects about the upcoming formation of the Movement Charter drafting committee. You can provide feedback here, at the central discussion on Meta, at other ongoing local conversations, and during a Global Conversation upcoming on 26 and 27 June 2021.

Further details and context about these questions is on Meta along with a recently-updated overview of the Movement Charter initiative. Feel free to ask questions, and add additional sub-sections as needed for other areas of interest about this topic.

If contributors are interested in participating in a call about these topics ahead of the Global Conversation on 26 and 27 June, please let me know. Xeno (WMF) (talk) 17:23, 19 June 2021 (UTC)

The three questions are:

  1. What composition should the committee have in terms of movement roles, gender, regions, affiliations and other diversity factors?
  2. What is the best process to select the committee members to form a competent and diverse team?
  3. How much dedication is it reasonable to expect from committee members, in terms of hours per week and months of work?

Family vs. surname

I have added a CoA image to Salvagnoli but my edit was reverted because "it is a surname not a family". However in this case there's only a family called Salvagnoli (based in Empoli, Tuscany) as can be seen here. As a rule, CoAs should be put only in family pages and not in surname pages? Also similar cases should be edited? Thanks in advance.--Carnby (talk) 18:27, 19 June 2021 (UTC)

I would say yes, coat of arms only belong to family items and not to surname items. Family and surname are different concepts. Coat of arms belong to families, not to their surnames. Even if you showed that 100% of people carrying this surname are part of Salvagnoli (Q81722003), I'd say that the coat of arms has no direct affiliation to the surname. In fact, if I wanted to rename myself to Salvagnoli, nobody would stop me, yet I would have no relationship to that coat of arms. Vojtěch Dostál (talk) 20:04, 19 June 2021 (UTC)
Thank you--Carnby (talk) 06:46, 20 June 2021 (UTC)

Wikidata page addition

Please, I want a wiki data page to be created for the author, Michael Bassey Johnson. Maestro333 (talk) 03:44, 20 June 2021 (UTC)

Wikispecies pages not yet linked to wikidata items

Hello, three parts/requests/desiderata; how can I create/see a list of:

1 pages existing on wikispecies not yet linked to a wikidata item (both where the wikidata item exists and where it does not)
2 pages existing on wikispecies not yet linked to a wikidata item (where the wikidata item already exists (with the same name))
3 pages on eg ja.wikipedia (or en.wikipedia), in category X (/and its subcategories), that are not yet linked to wikispecies items
I have tried to use PetScan to achieve the above, but without success, thank you, Maculosae tegmine lyncis (talk) 08:16, 20 June 2021 (UTC)
@Maculosae tegmine lyncis: 1) Duplicity Tool might be a way to go: https://wikidata-todo.toolforge.org/duplicity.php?wiki=specieswiki&mode=list . 2) This is complicated. What label language do you want to pair the WikiSpecies sitelinks to? It's probably doable with a Wikidata Query. 3) This is doable with Petscan, see https://petscan.wmflabs.org/?psid=19385875 for Category:Poaceae genera. Vojtěch Dostál (talk) 09:08, 20 June 2021 (UTC)
@Vojtěch Dostál: Thank you. 1) Seems to work brilliantly (have made the first link); 2) I guess Latin usually (but (1) includes this and is more than good enough, at least for now); 3) This seems to have the same issue I encountered - the above "see...petscan...19385875" for Category:Poaceae genera includes on the list Phragmites - but the en.wikipedia page Phragmites is already linked to wikispecies page Phragmites (Q1976487), Maculosae tegmine lyncis (talk) 09:52, 20 June 2021 (UTC)
@Maculosae tegmine lyncis: Yeah, sorry, the correct code for Wikispecies in Petscan seems to be "specieswiki": https://petscan.wmflabs.org/?psid=19386258 Vojtěch Dostál (talk) 11:11, 20 June 2021 (UTC)
@Vojtěch Dostál: Ah yes, I see - as in the wikidata edit summaries where it says "Added link to [specieswiki]", now I get it. Thank you very much indeed, once again, Maculosae tegmine lyncis (talk) 11:20, 20 June 2021 (UTC)
@Vojtěch Dostál, Maculosae tegmine lyncis: I've also looked at that report recently. As has been noted/discussed over on species:Wikispecies:Village Pump#Wikidata to do list, it hasn't been updated since 2 months. @Magnus Manske was pinged there about that... --Azertus (talk) 11:27, 20 June 2021 (UTC)

Rank for software

Hello! I noticed that quite a lot of items about software (for example SeaMonkey) have one instance of (P31)-claim marked as preferred while all other claims (this quite often includes claims with free software (Q341) as value or other "broad" values) are marked as normal. Since the preferred statement isn't truer or more up to date than the other ones I think that the ranks here have been used incorrectly. Should this be corrected, i.e. all statements set to rank normal? (Please ping me) --Nw520 (talk) 10:49, 17 June 2021 (UTC)

I agree this is confusing at best. I think whether it should be corrected depends on what people are using these statements for which isn't clear to me. Does it practically matter at the moment? BrokenSegue (talk) 19:06, 20 June 2021 (UTC)
@BrokenSegue: Thanks for your comment. I don't think that this is a major or pressing issue, however, I do think that this modelling can lead to erroneous queries (i.e. SPARQL-querying with wdt:P31 vs p:P31) and might motivate other users to adopting this (wrong) use of ranks. --Nw520 (talk) 19:09, 21 June 2021 (UTC)
Absent a credible reason for preferred rank (P7452) it's hard to support what seems to be the arbitrary promotion of a single statement. As you note, it's unhelpful for wdt: queries. My general experience is that such promotions are done to support a template, somewhere, on some other wiki; and with no consideration for the effect on other uses of the data. For my money, unsupported promotions should be demoted on sight. --Tagishsimon (talk) 19:30, 21 June 2021 (UTC)

Wikidata item for Wikidata API

Is this the Wikidata API?

If not, then what is? Is there an official website for the Wikidata API? Is it correct that there is one MediaWiki API for accessing anything in the Wikimedia platform, including Wikidata? Thanks. Blue Rasberry (talk) 22:07, 20 June 2021 (UTC)

@Bluerasberry: See also Wikibase API (Q106877126). It is based on the Mediawiki Action API but goes into more detail on what's specific for Wikibase. The API is also self-documenting so the Wikidata API endpoint tells you what it does with some examples. ArthurPSmith (talk) 14:05, 21 June 2021 (UTC)

Wikidata weekly summary #473

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Other Noteworthy Stuff
    • Interlanguage links now work for multilingual Wikisource. phab:T275958
    • User:Nikki/LexemeEntitySuggester.js now supports searching for senses for the translation property.
    • The call for candidates for 2021 Wikimedia Foundation Board elections has begun on June 9. The last date to apply is June 29, 2021.
    • Since the beginning of the month, phabricator tickets about language codes (and names of languages) were reviewed and triaged to better reflect their content and current status. Some work was done to better identify the steps for such changes, highlight and address bottlenecks and system issues. Patches for a few codes were contributed and are being released. Phabricator "Language codes" workboard provides an overview, phab:T284856 attempts to identify maintenance steps and phab:T284276 determine turn-around times. phab:T284808 should finally close a gap in termbox language handeling. A way to better address some or all aspects of changes of language codes applicable to Wikipedia editions still needs to be found. Don't hesitate to request the addition or update of language names (e.g. the name of Dutch in Danish) or missing language codes, notably for monolingual strings (see Help:Monolingual text languages).

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.

Wikidata Discord invitation link expired

The Discord link on Wikidata:Discord is expired, please update it. --ΟΥΤΙΣ (talk) 01:01, 22 June 2021 (UTC)

Duplication

both exist. Why? Q1249802 links only to the Spanish (and Latin) Wikipedia entries and is clearly incomplete. Surely Q1249802 should be merged into Q9648? GPinkerton (talk) 19:11, 20 June 2021 (UTC)

one refers to the physical island and one the state on the island. Spanish wikipedia has an article for both so they cannot be merged. We have the same for other identically named places (e.g. ireland and ireland). BrokenSegue (talk) 19:25, 20 June 2021 (UTC)
@BrokenSegue: That seems odd. What about Malvinas Islands (Q23305213)? [How many more?] GPinkerton (talk) 01:45, 21 June 2021 (UTC)
(Stalker) These Islas Malvinas belong to Chile, not a source of dispute between Argentina and UK. (Well, they had two entries in WD, which I merged a couple of minutes ago. :) --E4024 (talk) 01:58, 21 June 2021 (UTC)
@GPinkerton: A simplified way of looking at it is that the Falkland Islands as a political entity came into existence in 1833. That is what we have mapped in one item. However, the physical land mass that the people live on has most likely been there for millions of years. If we need make a statement in another item that references the islands 2,000 years ago, we can link the statement to the physical item. If we need to make a statement about the government of the islands, we can link to the modern political entity.
Similarly if we use Britain as an example, we can create separate items for the Celtic tribes, Romano Britain, the Anglo-Saxon kingdoms, the Danelaw, the Kingdom of England and so on, each with their own start and end dates. However, there should be a separate item to describe the land mass that all of those political entities sat on. Tribes and kingdoms come and go with a high frequency. The land mass remains mostly unchanged for millions of years. From Hill To Shore (talk) 06:17, 21 June 2021 (UTC)
@From Hill To Shore: Should there be similar duplicate items for everywhere? I found Sri Lanka (Q854) and Sri Lanka (Q4526612), and Isle of Man (Q9676) and Isle of Man (Q27508141) but should there be similar copies for Japan, New Zealand, Denmark, China? How specific does it go? Does everywhere has its own physical geography doppel-item? It seems a little like having separate items for a notable person and for that person's physical body. GPinkerton (talk) 06:41, 21 June 2021 (UTC)
The duplication only appears where the physical land mass and the country share a one to one relationship. That is most common with islands. However, all countries should have a related item for the physical land mass. In the case of Denmark it is probably an item about the main European land mass, with several other items to represent Denmark's various islands. To better understand the concept, it may be useful to turn the question around so you can see the situation from another angle.
Hypothetical: I want to create an item to explain about an animal population that lived two million years ago. Fossil records show that the animals died out one million years ago. The animals lived in the physical location where China sits today. Would it be correct for me to make a statement that the animals lived in the People's Republic of China? From Hill To Shore (talk) 07:34, 21 June 2021 (UTC)
@From Hill To Shore: I don't know. It would make sense to say that fossils are found in the country they are found in. What is the alternative for the Chinese case? For the case in hand, did Falkland Islands wolf (Q205935) live (and become extinct) on Falkland Islands (Q9648) or on Falkland Islands (Q1249802)? The existence of the creature was not known until after the population of the islands, so it would make sense that it lived in Falkland Islands (Q9648). GPinkerton (talk) 15:25, 21 June 2021 (UTC)
If your question concerns islands, an island is a different concept than a state/country. Republic of Ireland (Q27) is not Ireland (Q22890), United Kingdom (Q145) is not Great Britain (Q23666). New Zealand (Q664) is neither North Island (Q118863) nor South Island (Q120755).
Otherwise a simple rule is: If you can describe each thing using different properties, these are two concepts which might (depends on whether the concept is notable) have two different Wikidata items. (And this is not a duplication.) --Matěj Suchánek (talk) 07:45, 21 June 2021 (UTC)
@Matěj Suchánek: the short description of New Zealand describes it as "situated on two main and around 600 smaller islands in the southwestern Pacific Ocean". My question is, if there are two separate items for the Falklands, are there two separate items for the "two main and around 600 smaller islands in the southwestern Pacific Ocean"? GPinkerton (talk) 15:32, 21 June 2021 (UTC)
Replace "island" with "archipelago" in my initial statement. Falkland Islands (Q1249802) had probably existed before the first humans, Falkland Islands (Q9648) since 1833.
There is one more simple rule: Wikidata relies on authorities (we call the data "claims", not "facts"). What authorities claim, Wikidata is supposed to mirror. (No problem when multiple authorities disagree.)
I am not answering your question, I am not an authority. --Matěj Suchánek (talk) 15:50, 21 June 2021 (UTC)
Wikidata is about single concepts and based merely on Wikiarticles. Generally speaking, one/++ wikiarticles = a wikidata item. There is different wikiarticles about Falklands. You could tell that Falkland islands are an island, geographically speaking ; distinct from the government body, politically speaking. The same applies to many cities : some Wikidata items are about the city, municipality ; some about the city, as a city. --Bouzinac 💬●✒️●💛 07:08, 21 June 2021 (UTC)
@Bouzinac: but this is a group of islands that are not connected by anything except their human-imposed name, which is why I'm interested in whether New Zealand and Japan have equivalent duplicate items. If not, why not? GPinkerton (talk) 15:25, 21 June 2021 (UTC)
@GPinkerton: I am very concerned about some of your edits here. You have now removed the Falkland Islands from the South American continent.[8] Setting aside the fact that it could be interpreted as an attempt to make a political statement, the propper way to deal with statements that are incorrect is to use Help:Deprecation. If you find yourself deleting a statement on Wikidata, you will have taken the wrong action in 99% of circumstances. From Hill To Shore (talk) 08:04, 21 June 2021 (UTC)
@From Hill To Shore: What is concerning about that edit? Islands cannot be part of continents, by definition. In this particular case the claim is clear jingoísmo. Nowhere on New Zealand (Q664) do we list those islands as being part of Australia, and neither should we do the same for other islands. GPinkerton (talk) 15:25, 21 June 2021 (UTC)
@GPinkerton: As I said, the primary concern is that you are blanking statements. This is not Wikipedia where editors debate what information is or is not included. If a piece of information on Wikidata is wrong, the default option is to either use the deprecated rank to mark an incorrect answer or use preferred rank to set the best answer out of a group of plausible answers. Blanking in the majority of situations is just damaging Wikidata. This is a fundamental concept here, so I suggest you take this to heart. From Hill To Shore (talk) 16:10, 21 June 2021 (UTC)
@GPinkerton: What is concerning is that you are making stuff up as you go along, based on whatever today's whim happens to be. w:en:South America and sources therefrom make it clear that the Falkland Islands are consided a part of South America. But you're that guy on the interenet who knows better than anyone else; and in your arrogance, you're damaging wikidata. --Tagishsimon (talk) 15:32, 21 June 2021 (UTC)
@Tagishsimon: you claim "w:en:South America and sources therefrom make it clear that the Falkland Islands are consided a part of South America" but can you ("in your arrogance") please identify a "source therefrom" that makes this geographically extraordinary claim? GPinkerton (talk) 15:38, 21 June 2021 (UTC)
https://www.cia.gov/the-world-factbook/countries/falkland-islands-islas-malvinas/
https://www.nationalgeographic.org/encyclopedia/south-america-human-geography/print/
Try dealing with the central point: there's consensus on the EN wiki article that the islands are part of the continent. What is your source authority supporting your removal of the statement? --Tagishsimon (talk) 15:41, 21 June 2021 (UTC)
That would be the basic meaning of the word. I note that neither of your sources deals with non-human geography. Such sources include Easter Island (Q14452) as "South America", a clear oxymoron, since continents and islands are separate things, each defined by not being the other: "The main land, as distinguished from islands, islets, or peninsulas; mainland." (O.E.D.) Easter Island in particular is notable for being as far away from a continent as it's possible to be on land. GPinkerton (talk) 15:56, 21 June 2021 (UTC)
I haven't looked for sources behind this claim, but it doesn't surprise me at all. Some definitions of 'continent' are 'any land mass sitting on the continental plate', in which case, e.g., Great Britain and Ireland are part of the continent of Europe. In that model, it's not possible for any land mass to not part of a continent/on a continental plate. WhatamIdoing (talk) 15:44, 21 June 2021 (UTC)
@WhatamIdoing: if that's the case, shouldn't there be different items for these differing definitions? Do we have an item for Australia that includes both New Zealand and Tasmania but which is separate from the Commonwealth of Australia? GPinkerton (talk) 15:49, 21 June 2021 (UTC)
In this subsection we're discussing your unsourced removal of a statement, GPinkerton. Whataboutism does not cut it. You asked for sources. I've given you sources. Please now respect the process and accept your removal was in error. --Tagishsimon (talk) 15:51, 21 June 2021 (UTC)
@Tagishsimon: your insulting behaviour is not likely to engender respect, and please do not confuse yourself with process. GPinkerton (talk) 15:58, 21 June 2021 (UTC)
You were wrong to think the Falkland Islands are not part of South America. Even had you been right, you were wrong in the way you went about removing the statement. Now you are wrong in ducking the discussion, preferring to complain about butthurt. Wikidata does not work on pigheadedness. It works on sources. Where is your source? --Tagishsimon (talk) 16:09, 21 June 2021 (UTC)
@Tagishsimon: more insults lead me to believe you're very personally involved in this issue, so if you want to push this little political claim that's fine, I'll not stand in your way. Have two separate items (Falkland Islands (Q9648) and Falkland Islands (Q1249802)) and call them both South America. We can call them both overseas territories too with equal justification. All sources agree on that. GPinkerton (talk) 16:17, 21 June 2021 (UTC)

The archipelago is definitely in South America. Multiple sources agree on this matter. You may add a qualifier claiming it's disputed if you can find a source that disputes it but I very much doubt you can. BrokenSegue (talk) 18:32, 21 June 2021 (UTC)

@BrokenSegue: an archipelago could only be a part of a continent if using definition that diverges from the common-sense definition of a continent used by, inter alia, the: Concise Oxford Dictionary of World Place Names (6 ed.): "South America: The southern half of the American land mass." Argumentum ex silentio is a fallacy but there is no room in land mass for archipelagos hundreds of miles away from the nearest land mass. Cuba (Q241) is noted as being in North America (Q49) but Iceland (Q3740828) is apparently in Europe (Q46). This seems against all reason, so obviously this is a systemic political (rather than geographic) usage of "continent". Is there a different "South America" item of which the Falklands would not be a part, and a "North America" of which Cuba (Vancouver Island, Puerto Rico, Greenland, etc) would not be a part? It seems odd to have an apolitical item for an entity but then categorize that item as being part of a political unit (like the political unit that would include Easter Island and the Falklands together with actual continental South America continent). GPinkerton (talk) 19:00, 21 June 2021 (UTC)
Cuba is a part of Central America. The Florida Keys are in North America. Sometimes we simply lack the apolitical item. Do we have an item for the island cuba resides on? Or the peninsula that Florida resides on? Maybe not (I didn't check). If you think it useful you can feel free to make an item for those concepts. Otherwise, I have run out of patience here. Hopefully someone else will explain the rest to you. Please be more willing to accept that you could be mistaken here. BrokenSegue (talk) 19:07, 21 June 2021 (UTC)
Should the property for Cuba be changed then? Central America (Q27611) is presently described as "southernmost region of North America, lying between Mexico and South America and comprising Panama, Costa Rica, Nicaragua, Honduras, El Salvador, Guatemala, and Belize" - does that need to be changed too? At the moment, it's described as an instance of subcontinent (Q855697). To me, it doesn't seem like this is consistent. GPinkerton (talk) 21:00, 21 June 2021 (UTC)
No, the existing statements shouldn't be changed. If you have a source that contradicts a statement already on the item, you can add a new statement on the same property with the contradictory claim (make sure to record the reference where your alternative claim is stated). Wikidata is meant to report the claims of different sources, even where those claims disagree. In some cases where a property has a single value constraint, we grant one of the claims a preferred ranking. If more details or claims emerge later then the preferred ranking may switch to another statement. From Hill To Shore (talk) 21:07, 21 June 2021 (UTC)
@From Hill To Shore: where are the sources that support the statement that Cuba is in North American and not in Central America? Or indeed, that any particular island should be considered part of either and not both or neither? GPinkerton (talk) 21:38, 21 June 2021 (UTC)
That doesn't matter. Don't change or remove the existing statements. If the existing statements do not have a reference listed, that will likely mean your contradictory statement is immediately of higher quality and should be marked with the preferred rank. Another editor may arrive later and add a reference to the original statement; that editor or another editor may then change which statement has the preferred rank. The key point is that you shouldn't change or remove statements, even if the statements are "wrong." Just add the correct statement alongside the incorrect statement. If you choose not to insert references for the statements that you add, you run the risk of another editor coming along and deprecating your unreferenced statement or marking an alternative statement as preferred due to better sourcing. From Hill To Shore (talk) 21:49, 21 June 2021 (UTC)
@GPinkerton: Ok, looks like Cuba is generally considered to be in North America. You can find sources for this claim. I'm not a continent expert and the very concept of "continents" are a weird blend of geology and human culture (which is why people will disagree on some intermediate cases and wikidata can never be fully consistent). But many cases are non-ambiguous and wikidata's job is just to list statements that sources support. BrokenSegue (talk) 05:21, 22 June 2021 (UTC)

Template request

Requested template: {{Speedy delete}}

Rationale: Many RfDs in progress (as of 06:51, 22 June 2021 (UTC): 386), and I didn't see any template that allows for speedy deletion.

Description of request: One could mark anything that is obvious vandalism or empty with {{Speedy delete}}, when one isn't an administrator. Alfa-ketosav (talk) 06:51, 22 June 2021 (UTC)

wikidata has no "speedy deletion" policy so this is putting the cart before the horse. BrokenSegue (talk) 06:56, 22 June 2021 (UTC)
Oops. Wrong template name as well, as I found 2 pages which already have the template transcluded. Source: Special:WhatLinksHere/Template:Speedy delete. But it's still not that good that one has to open an RfD for everything, even if it is no more than vandalism. Alfa-ketosav (talk) 07:38, 22 June 2021 (UTC)
See Wikidata_talk:Requests_for_deletions/Archive_2#Splitting_this_process_into_speedy_deletions_and_deletion_discussions?. Another proposed solution is Proposed deletion.--GZWDer (talk) 10:18, 22 June 2021 (UTC)

Q107309201 and botanical expert

Can someone help me fill out Prunus pumila var. besseyi (Q107309201) I have never created an entry for a cultivar of a plant. We need the entry because English Wikipedia is redirecting Hansen's bush cherry to Prunus tomentosa instead of the correct Prunus besseyi. There appears to be some confusion, for instance some images online show Prunus tomentosa 'Nanking' (red berries) instead of Prunus besseyi 'Hansen's' (purple berries). I think the cultivar is significant enough to have its own entry here at Wikidata. For instance see: https://northernridgenursery.com/products/hansens-bush-pie-cherry-2-pack (shows Nanking) and https://www.starkbros.com/products/fruit-trees/cherry-trees/hansens-bush-cherry (shows Nanking). While https://www.gurneys.com/product/hansens_bush_cherry (Shows Hansens) --RAN (talk) 17:12, 21 June 2021 (UTC)

According to this website Hansen's Bush Cherry is a common name for Prunus pumila var. besseyi (Q3408648) (basionym: Prunus besseyi (Q15624054)) and not a cultivar. --Succu (talk) 17:57, 21 June 2021 (UTC)
@Succu: I'll merge them, can you make the correction at English Wikipedia so the redirect goes to the proper one. It is a shame that some catalogs show the wrong image. I'll take some photos to help correct the bad information in Google Image. --RAN (talk) 18:10, 21 June 2021 (UTC)
Done. --Succu (talk) 18:40, 21 June 2021 (UTC)
You got Hansen's Bush Cherry, but, w:Hansen's bush cherry still goes to the wrong one! --RAN (talk) 23:57, 21 June 2021 (UTC)
An enWP problem, RAN. --Succu (talk) 20:03, 22 June 2021 (UTC)

External ID property on five items about humans out of millions marked as no value

Out of millions of items about humans that have no GND identifier in Deutsche Biographie (Q1202222) five got the mark "no value" between 2020-05-14 20:44 and 2020-05-15 07:26 :

  1. Q350488: novalue , 2020-05-14 22:22 https://www.wikidata.org/w/index.php?title=Q350488&diff=1181573060&oldid=1181572834
  2. Q594005: novalue , 2020-05-14 20:44 https://www.wikidata.org/w/index.php?title=Q594005&diff=1181528194&oldid=1181528103
  3. Q1169400: novalue , 2020-05-14 22:31 https://www.wikidata.org/w/index.php?title=Q1169400&diff=1181578224&oldid=1181578089
  4. Q1308206: novalue , 2020-05-15 07:20 https://www.wikidata.org/w/index.php?title=Q1308206&diff=1181874820&oldid=1181874618
  5. Q1635497: novalue , 2020-05-15 07:26 https://www.wikidata.org/w/index.php?title=Q1635497&diff=1181877896&oldid=1181877775

All marks added by one user. It seems s/he abandoned the idea to mark more items of humans in such a way. The situation persists since more than one year now and is related to the Property:P7902 which has many other issues in Wikidata.

What do other think about this? Probably one can add such values to any external id property for which no value exists. Then hundreds of statements - one per property - could be made on the millions of itemts about humans. 2 million items have a VIAF ID, but no other property has that much. Just an approximation; 5 million items about humans times 100 "no value" per item would mean 500 million new statements like the five presented above. And that is only for humans, the "no value" concept could be expanded to other classes of items.

What is the value of "no value" in external id properties, and how will the the truthness be checked? What is the reference to add for such a statement? HumanAFuser (talk) 17:58, 22 June 2021 (UTC)

You've started the same discussion here: Property talk:P7902#P7902 clean-up - format violations - novalue - longterm strategy. --Kolja21 (talk) 18:05, 22 June 2021 (UTC)
In this cases reference usually contains only retrieved (P813) stating the date on which the user has found that there was no ID for the item. Anyway, I tend to think that these statements are superfluous for the reasons you state and I would support their removal. I would make one exception: if two people are connected through different from (P1889) because they are easy to confuse, and one has a certain ID while the other hasn't, using such statements might be somewhat helpful. --Epìdosis 18:03, 22 June 2021 (UTC)
Looking at these items, I think there is a misunderstanding of what the intention was. These appear to be items with a GND ID (P227) but no Deutsche Biographie (GND) ID (P7902). I can't check how many items use GND ID (P227) on my mobile, but that will be the absolute limit on this application, not the number of human (Q5).
Whether tracking which GND ID (P227) lack Deutsche Biographie (GND) ID (P7902) in this way is a good use of resources is a separate question. From Hill To Shore (talk) 18:43, 22 June 2021 (UTC)
No intention is stated next to the statement. Please provide evidence re "misunderstanding". RE "I can't check how many items use GND ID (P227) on my mobile, but that will be the absolute limit on this application" - Why? There is also Sächsische Biografie (GND) ID (Property:P1710) and WorldCat which follows VIAF and LC. Re statistics:
Statistics 2021-06-17 - Most used VIAF-related IDs on instances of humans
Property name Property id items distinct IDs ID claims WDQS
VIAF 214 2126799 2156651 2158834 https://w.wiki/SVe
ISNI 213 1171088 1180388 1181527 https://w.wiki/SVc
GND 227 1071079 1074509 1074514 https://w.wiki/Rkt
LC 244 1026681 1028352 1028432 https://w.wiki/SVN
DtBio 7902 626536 627645 627645 https://w.wiki/Rkq
~600 000 for DtBio vs ~1 000 000 existing GND, i.e. 400 000 "no value"? There is not any hint DtBio wants to cover all WD-GNDs. BTW: GND itself has 5 milllion+ primary human IDs (i.e. incl. duplicates not merged yet).
If a human has two GND IDs, but in DtBio only contains one, would one then also state "no value" for the ID that is not in DtBio alongside the GND value which is in DtBio? Or does the "no value" statement for ID X depend on existence of ID Y in DtBio? HumanAFuser (talk) 19:29, 22 June 2021 (UTC)
There may be an issue of misinterpreting comments in another language. Your reaction suggests you have not quite understood what I was saying.
In terms of what I said about "misunderstanding," your original comment characterises this as being applied to all instances of human (Q5), when it appears to be limited to instances of human (Q5) that have GND ID (P227) but lack Deutsche Biographie (GND) ID (P7902). Rather than a misunderstanding, it may be a case that you were aware but didn't include that detail in your original explanation. Either way, there is no harm done.
For "why" I can't access certain data on my mobile, it is simply that I can't do it. I normally access Wikidata through a PC and the tools I use to do so aren't present on this mobile. There may be other ways to access the data but I don't know how to do so; I was going to check later on my PC when I get to it. I simply said that I haven't checked the scale of usage of GND ID (P227) to try to avoid misunderstanding.
Finally, I stated that whether this method of working is a good idea or not is a separate question. This means I have not made a judgement on whether the original implementation was good or bad. There is no need to go on the attack. From Hill To Shore (talk) 20:04, 22 June 2021 (UTC)
What other language, I see only English here. Re "There is no need to go on the attack." - like implying some other user was involved in "misunderstanding" and now "misinterpreting"? "Wikidata:No personal attacks : Do not make personal attacks anywhere in Wikidata. Comment on content, not on the contributor. Personal attacks do not help make a point; they only hurt the Wikidata community and deter users from helping to create a good site. Repeated or egregious personal attacks may lead to blocks."
Re: "your original comment characterises this as being applied to all instances of human (Q5)" - the section headline tells otherwise, using the numeral five in "External ID property on five items about humans out of millions marked as no value". HumanAFuser (talk) 20:13, 22 June 2021 (UTC)
Suggesting that your unusual reaction to my previous comment may be due a language issue on a board frequented by users with different native languages was an attempt to assume good faith. It was the very opposite of a personal attack. If you are saying that your native language is English then I am very confused by your reaction to my first comment. There is no need to treat this like a confrontation.
I am getting deja vu of a very similar conversation I had last year with another user that had a focus on GND items.
I am going to withdraw now as this confrontation is pointless. Hopefully some other editors will step in and address your points. From Hill To Shore (talk) 20:36, 22 June 2021 (UTC)
@From Hill To Shore: Yeah, it’s abundantly clear that this is yet another sock of a user than was globally banned back in 2015. They now seem to have entered the “accusation phase” of this sock. Discussing anything is pretty pointless at this point. --Emu (talk) 21:17, 22 June 2021 (UTC)
"Wikidata:No personal attacks : Do not make personal attacks anywhere in Wikidata. Comment on content, not on the contributor. Personal attacks do not help make a point; they only hurt the Wikidata community and deter users from helping to create a good site. Repeated or egregious personal attacks may lead to blocks." HumanAFuser (talk) 04:46, 23 June 2021 (UTC)
@From Hill To Shore: "Wikidata:No personal attacks : [...] Comment on content, not on the contributor. " . This would even include commenting on yourself, not? "I am very confused by your reaction" - How does it further the objectives of the WMF that on WD Project Chat a user tells s/he is confused about a reaction? Wouldn't it be more helpful to ask for help and precisely describe what additional information would be helpful? HumanAFuser (talk) 05:20, 23 June 2021 (UTC)

Human item having deprecated statement DtBio withdrawn identifier value GND incorrect identifier value

https://www.wikidata.org/w/index.php?title=Q94762256&type=revision&diff=1446631366&oldid=1286583860 - same value for P227 and P7902, each deprecated, but different reasons stated, no sou

Deutsche Biographie (GND) ID: 

Deprecated rank
Property / Deutsche Biographie (GND) ID: 107872013 / qualifier
		
reason for deprecation: withdrawn identifier value
Property / Deutsche Biographie (GND) ID: 107872013 / reference

retrieved: 22 June 2021
Timestamp	+2021-06-22T00:00:00Z
Timezone	+00:00
Calendar	Gregorian
Precision	1 day
Before	0
After	0

GND ID 

Property / GND ID: 107872013 / rank	Property / GND ID: 107872013 / rank
Normal rank
Deprecated rank
Property / GND ID: 107872013 / qualifier
		
reason for deprecation: incorrect identifier value

What do others think about this? Why is the incorrect/withdrawn value for DtBio marked "withdrawn identifier value" and the same incorrect/withdrawn value for GND marked "incorrect identifier value"? Was it correct for DtBio and only withdrawn? Was it not withdrawn from GND? HumanAFuser (talk) 19:59, 22 June 2021 (UTC)

Note: Another similar edit less than 24 hours earlier, here removing deprecated statements, while on the item above they have been created [9] HumanAFuser (talk) 20:09, 22 June 2021 (UTC)
Why don't you ask me directly? These edits were a first step: Checking the IDs and mark GND as incorrect identifier value (Q54975531) and Deutsche Biographie as withdrawn identifier value (Q21441764). Second step: Looking for citable sources. Result: Eric L. Dreifuss (Q94762256) (1943-2007), lawyer. Obituary in the Neue Zürcher Zeitung (Q658945). If there is no citable source the item should be reported @WD:RfD. --Kolja21 (talk) 00:41, 23 June 2021 (UTC)
PS: There are different reasons why an ID is broken. It's the same job as working with tracking categories like Category:Articles with dead external links (Q7045997). It would be easier and faster to delete every weblink that is offline but you have to check each of them. --Kolja21 (talk) 01:00, 23 June 2021 (UTC)
1) "Why don't you ask me directly?": Because you recently several times resorted to reply in German, e.g. at Property talk:P227: 13 June 13:09 20 June 12:51 22 June 18:29 and Property talk:P7902: 22 June 16:49 ("ping" not linked) and stopped pinging in your replies, whilst before you replied in English and pinged more frequently.
2) "Second step: Looking for citable sources": Not a single of the "citable sources" even mentions the Tn, not to speak of declaring that the Tn (=not individualized name) referred to a specific Homo sapiens. HumanAFuser (talk) 05:09, 23 June 2021 (UTC)

Criminal Convictions

I am supervising a project which has transcribed a historic register of convicts, uploaded relevant details to Wikidata and where there are photo, added those to Commons.

Example https://www.wikidata.org/wiki/Q107243202

What we are struggling with is the notion of conviction and sentence.

In Scots Law (and English as far as I can tell) previous convictions are noted as a block of four elements which are always together:

  • Date (eg 28 March 1872)
  • Place (eg The High Court, Edinburgh)
  • Offence (eg Theft)
  • Sentence (eg 3 months imprisonment)

There appears to be no way to model these in Wikidata at present.

We can use P1399 (Convicted Of) to link to a crime (such as Theft) but no way to record have an associated date, place, or sentence. So, we can't query and retrieve data on, for example, sentences compared male to female, or where most prisoners were sentenced, or construct timelines using sentence date.

We have around 280 records to add which would use the format above.

Before I make a proposal for a property - or amendment to the existing P1399 - can anyone suggest how presently the four-aspect comviction noted above could be modelled?

And, if not, what the best route to fixing this is - ie new property or amendment of P1399?

Watty62 (talk) 08:33, 23 June 2021 (UTC)

I have added the point in time (P585) qualifier to your example. Likewise you can add a location (P276) and penalty (P1596) qualifier. --SCIdude (talk) 09:24, 23 June 2021 (UTC)
To add to what User:SCIdude has written, here's a query for all the different sorts of qualifiers that are currently being used on convicted of (P1399) statements: https://w.wiki/3Xdy plus also a similar query for the types of qualifiers being used on more generic significant event (P793) statements: https://w.wiki/3Xe2 to give an idea of what's available. (This query is often enough useful that it is made available for any property, via the 'List of Qualifiers' link at the top of the 'Documentation' box on its talk page, as eg at Property_talk:P1399#Documentation).
Also, here's also a tweaked version of the P1399 query that includes a sample value for each qualifier: https://w.wiki/3Xe7
Compare also the list of agreed allowed qualifiers, at [[10]]
From the results:
  • court (P4884) looks to be the best choice for the court that handed down the sentence -- probably better to standardise on that rather than location (P276), which maybe might be reserved for use as an additional bit of information for where a particular court can have multiple seats (eg a travelling assizes).
  • point in time (P585) can be used for the point in time. We should probably underline somewhere that when used as a qualifier to convicted of (P1399), this is for when the conviction occurred, not when the underlying offence was committed (Q: how should that be recorded ?)
  • penalty (P1596) = incarceration (Q853735) with additional qualifier duration (P2047) = <length of time> can be used to record a sentence of a period of imprisonment.
  • The offence should be given as the main value of the statement.
No doubt there are clean-ups and clarifications that could maybe be applied to the data model, and it would be good perhaps to write up a description of recommended best practice, eg on the talk page of P1399. But the above, plus a look at some of the other qualifiers in common or occasional use, should give at least a starting point. Jheald (talk) 09:49, 23 June 2021 (UTC) Jheald (talk) 09:49, 23 June 2021 (UTC)
I suggest setting up a wiki page with a dozen or so real-world examples (regardless of whether they are yet modelled on Wikidata), to see what metadata is available for each. We should include examples from different jurisdictions, as well as less-common patterns, such as sentence increased or decreased, or conviction overturned, on appeal; pardon; parole; release on licence, etc. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:03, 23 June 2021 (UTC)
Thanks SCIdude, Jheald, Andy Mabbett Those are all useful suggestions - and I will follow up, linking any new Talk / Wiki page from here.
Watty62 (talk) 10:21, 23 June 2021 (UTC)
Have we considered the impact on items about living people? There are situations (often with minor offences) in the UK where a conviction is expunged from the public record after a period of time. This is so that reformed former criminals with minor offences (shop lifting, for example) can reintegrate into society without their criminal record barring them from jobs or other activities. We may need to think on how we handle such situations. One option may be to publish everything and rely on the servers sitting outside of UK jurisdiction as sufficient protection. Another option may be to only record convictions of a certain severity (murder, for example). From Hill To Shore (talk) 13:41, 23 June 2021 (UTC)
That's a valid point, From Hill To Shore. In our project we agreed with the archives a 100-year cut off. So, while the register runs to 1939, we're only publishing to 1921 at the moment. Watty62 (talk) 14:56, 23 June 2021 (UTC)

HathiTrust Identifiers

Hello Wikidata Community,

I am the Metadata Analyst at HathiTrust Digital Library and I would like to start adding HathiTrust data to relevant Wikidata entries. At minimum I want to add the HathiTrust ID(s) for monographs, multipart monographs and serials to their already existing entries. It looks like an entity already exists for a HathiTrust ID but unless I’m mistaken I don’t believe it has been widely applied.

The metadata we publicly expose at this time can be seen here with greater description: https://www.hathitrust.org/hathifiles_description

But in short the potentially valuable metadata would be ht_bib_key (our catalog title-level identifier), htid (persistent identifiers for each individual scan in our corpus), ISBN, ISSN, LCCN, OCLC Control Numbers, Title, Author, Publication Place, Publication Date, Whether or not a document is considered a federal document, and the language. Possibly there are others on that page, but those are entities that strike me as most valuable or relevant to this project.

I’m new to the Wikidata community and I don’t assume to know anything, so I’m looking for a lot of guidance here to make sure our contribution to this project is a valuable one.

GDethmers (talk) 12:48, 23 June 2021 (UTC)

Yeah so HathiTrust ID (P1844) already exists which is helpful. So you can start by just manually adding it to any relevant items. A bulk import would be good but it's often hard to automatically match items wikidata items to entries in some external database. People sometimes use Mix'n'match for this. If you want to add support for new identifiers (e.g. ht_bib_key) you would have to request a new property at Wikidata:Property_proposal. Once data is linked the next step would be to copy over other data fields that you mentioned. Maybe Wikidata:Dataset Imports would be helpful. BrokenSegue (talk) 14:30, 23 June 2021 (UTC)
Thanks! I'll follow your suggestions!GDethmers (talk) 15:32, 23 June 2021 (UTC)

Merger requests

Is there an established procedure for requesting mergers of Wikidata items to be done by someone else in the community? I feel that I'd need to have some more experience with Wikidata before I would do that myself in non-trivial situations.
Right now I would like to suggest merging items Q21502003 ("Pohlmann" as a familiy name) and Q175051 ("Pohlmann" as a disambiguation page). --ΟΥΤΙΣ (talk) 01:06, 22 June 2021 (UTC)

You can request mergers here if you are not sure if a proposee merge is correct. In this case a merge is wrong. Family name items have to be separate from disambiguation page items. From Hill To Shore (talk) 06:56, 22 June 2021 (UTC)
Thank you for looking at my request and telling me the appropriate rule/guideline for the situation, From Hill To Shore. Could you please also tell me where I cound find this specific rule, myself? I still don't feel fully oriented at Wikidata. Thanks --ΟΥΤΙΣ (talk) 18:53, 22 June 2021 (UTC)
I am not sure if it is written down anywhere as a rule but it comes from the idea that each Wikidata item represents a different concept. The concept of a family name is different from the concept of a page to disambiguate different topics. Merging the concepts will generate confusion when we want to link to just one of those concepts.
Maybe an analogy will help to explain this better:
Hypothetical: we merge an item for the family name "Jones" with an item for the disambiguation page "Jones." We use the merged item on another item about the singer Tom Jones. A user extracts a report on Tom Jones and sees he is called "Tom" "Jones (disambiguation page)."
We need to keep the items separate so that human (Q5) items don't inherit the concept of being a disambiguation page.
If you have two items describing the same concept, a merge is a good idea. If the two items describe different concepts with the same name, a merge should be avoided. From Hill To Shore (talk) 19:37, 22 June 2021 (UTC)

Thanks for your explanation, Fhts. But in the case of "Pohlmann" (see above), there must be something wrong with the linking of the disambiguation pages: The English disamb page stands alone and links to none of the disamb pages in other languages.
Would you look at it and tell me how to fix it? --ΟΥΤΙΣ (talk) 04:27, 23 June 2021 (UTC)

I agree that the enwiki page is a list of people with that name rather than an article about the name itself. So I have transferred the site link to the other item. — Martin (MSGJ · talk) 04:44, 24 June 2021 (UTC)

Problem with Q215857 (bookcase): both Commons categories Bookcases and Bookshelves refer to it

In WikimediaCommons Category:Bookcases refers to Q215857 (bookcase), as is Category:Bookshelves. That is not right: Q215857 should only refer to Category:Bookcases. But though I have looked into this problem several times, I cannot make it correct. Category:Bookshelves has an item in Wikidata of its own: Q55277362, but it looks it is not well linked/connected to Commons. Could you please help me? JopkeB (talk) 05:46, 23 June 2021 (UTC)

@JopkeB: if I'm understanding you correctly it sounds like the problem is that we don't have an item to represent the concept of a bookshelf but we do have an item to represent a bookcase (i.e. bookcase (Q215857)). the solution is to make a new item for bookshelf then move the references to bookcase in Category:Bookshelves (Q55277362) to bookshelf. BrokenSegue (talk) 06:16, 23 June 2021 (UTC)
@BrokenSegue: Thank you for you answer. I did make Q107326062 (bookshelf), but the problem has not been solved: still Category:Bookshelves is automatically linked to Q215857 (bookcase) and not to this new item. What is not correct yet? JopkeB (talk) 11:40, 23 June 2021 (UTC)
@JopkeB: That's not a detail I'm super familiar with since it involves commons but looking at the category now it seems properly linked to your new item. I'm guessing there's just some delay between the change and commons picking it up. BrokenSegue (talk) 14:11, 23 June 2021 (UTC)

@BrokenSegue: Yes: the problem has been solved! See Category:Bookshelves. Many thanks for making this happen! Discussion can be closed. JopkeB (talk) 09:18, 24 June 2021 (UTC)

Q-titles instead of

Hi all. This is my first question here. We are in the middle of a project aiming to visualize history with large amounts of disparate open sources. In the project, we cooperate with Wikimedia Sweden, and has come quite far in the incorporation of wikidata-datasets along with other sets for our timeline and historical maps.

I have stumled upon a small issue, for which I seem to require some hints. This query:

SELECT ?itemLabel ?at ?itemDescription ?coord ?picture ?article WHERE { 

   ?item wdt:P31/wdt:P279* wd:Q13418847 .   
 ?item wdt:P625 ?coord.
    ?item wdt:P18 ?picture.
     ?article schema:about ?item .
?article schema:isPartOf <https://en.wikipedia.org/> .
 ?item wdt:P585 ?at;
   #wdt:P580 ?start;
   #wdt:P582 ?end;
   SERVICE wikibase:label {
       bd:serviceParam wikibase:language "sv".
   }

} ORDER BY ASC(?at) ?itemLabel

returns a good number of items. But many of the headlines are only q-numbers. For instace the title "Q1822262" does not imply it regards the siege of Rennes, which it's corresponding image does. How can I avoid getting Q-numbers as titles for the items?

Best regards Per Filipsson trackuback.com

Just write something like bd:serviceParam wikibase:language "sv,fr,de,en,zh". : it will search language label in the specified order sv>fr>de>en>zh etc. A lightly better deduplicated query : https://w.wiki/3Y28 --Bouzinac 💬●✒️●💛 11:52, 24 June 2021 (UTC)
Thank you very much. I understand the logic of securing it in multiple languages now.
All the best
/Per

How to add given name to some of 39474 items

https://w.wiki/3Y8V - is there a tool, that can take the first word of the labels and assign a given name item or assist in assigning one in batch for all items containg a certain name as first word? HumanAFuser (talk) 18:26, 24 June 2021 (UTC)

@Jura1:.--GZWDer (talk) 18:34, 24 June 2021 (UTC)

Server switch

Read this message in another language • Please help translate to your language

The Wikimedia Foundation tests the switch between its first and secondary data centers. This will make sure that Wikipedia and the other Wikimedia wikis can stay online even after a disaster. To make sure everything is working, the Wikimedia Technology department needs to do a planned test. This test will show if they can reliably switch from one data centre to the other. It requires many teams to prepare for the test and to be available to fix any unexpected problems.

Unfortunately, because of some limitations in MediaWiki, all editing must stop while the switch is made. We apologize for this disruption, and we are working to minimize it in the future.

You will be able to read, but not edit, all wikis for a short period of time.

  • You will not be able to edit for up to an hour on Tuesday, 29 June 2021. The test will start at 14:00 UTC (07:00 PDT, 10:00 EDT, 15:00 WEST/BST, 16:00 CEST, 19:30 IST, 23:00 JST, and in New Zealand at 02:00 NZST on Wednesday 30 June).
  • If you try to edit or save during these times, you will see an error message. We hope that no edits will be lost during these minutes, but we can't guarantee it. If you see the error message, then please wait until everything is back to normal. Then you should be able to save your edit. But, we recommend that you make a copy of your changes first, just in case.

Other effects:

  • Background jobs will be slower and some may be dropped. Red links might not be updated as quickly as normal. If you create an article that is already linked somewhere else, the link will stay red longer than usual. Some long-running scripts will have to be stopped.
  • There will be code freezes for the week of June 28. Non-essential code deployments will not happen.
This project may be postponed if necessary. You can read the schedule at wikitech.wikimedia.org. Any changes will be announced in the schedule. There will be more notifications about this. A banner will be displayed on all wikis 30 minutes before this operation happens. Please share this information with your community.

SGrabarczuk (WMF) 01:19, 27 June 2021 (UTC)

@SGrabarczuk (WMF): Ther is a wrong links in the header, the link leads to the switch from last year. Could you please fix this grave error? Grüße vom Sänger ♫ (talk) 07:29, 27 June 2021 (UTC)
Those of you who are interested in this thread, please see my talk page on Meta-Wiki where Sänger wrote to me as well, and I replied him there. Thanks! SGrabarczuk (WMF) (talk) 08:44, 28 June 2021 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Matěj Suchánek (talk) 14:01, 30 June 2021 (UTC)

difficulty with regular expression

I changed the property WTA tennis tournament ID (P3469) a little because the formatter was no longer correct. However, this also changed the regular expression. I tried to adapt it via the property WTA player ID (P597) but so far there is still an error by the object Wuhan Open (Q15650921) whether the regular expression should actually be correct. Could someone look over it again? --Gymnicus (talk) 11:39, 25 June 2021 (UTC)

There were two format contraints on the item. I've deleted one. Now Wuhan Open (Q15650921) passes the regex; but Cincinnati Open (Q461834) - a working link - fails b/c no \/[a-z]+. And Indian Wells Open (Q974505) - another working link - also fails b/c hyphen. Possibly need to establish the set of working patterns and then revisit the regex. --Tagishsimon (talk) 17:19, 25 June 2021 (UTC)
Regarding Indian Wells Open (Q974505) it should be said that there are several other data objects that do not satisfy the new regular expression. I still have to improve the link for these data objects. --Gymnicus (talk) 17:23, 25 June 2021 (UTC)

Qualifier anti-pattern

In the examples section of X username (P2002), all examples include the qualifiers number of subscribers (P3744) and point in time (P585). I interpret this as a way of getting around the fact that Wikidata does not have support for nested qualifiers, i.e. qualifiers qualifying qualifiers. The intention seems to be to qualify number of subscribers (P3744) with point in time (P585). Technically, though, both qualifiers apply to the main claim. Surely this must be an anti-pattern? If so, should we modify the examples? Popperipopp (talk) 13:06, 25 June 2021 (UTC)

yes it is an anti-pattern and is intended to be replaced with a social media followers (P8687) main claim which one of my bots populates. I also have an open bot request for permission to do this for youtube channels. The examples should be modified. BrokenSegue (talk) 16:05, 25 June 2021 (UTC)

YouTube video ID

A while back a bot migrated any YouTube_video urls that were at described_at_url to YouTube_video_ID. The only problem is that YouTube_video_ID is expected to be only in instance_of=video, or something similar. So we have to decide if we want to change that so video interviews where instance_of=humans can have have a YouTube_video_ID. Or should we switch back to the url format? See for instance: Robert E. Lang (Q85681784) --RAN (talk) 21:07, 23 June 2021 (UTC)

so lots of uses of youtube video id are for things like trailers or music videos. I don't think described_at_url is right there. also some films/tv shows have them because they are officially hosted on youtube. BrokenSegue (talk) 02:48, 24 June 2021 (UTC)
The intended solution is create a new item dedicated to the YouTube video.--GZWDer (talk) 19:42, 24 June 2021 (UTC)
That may be optimal but it's not super practical and we may have to accept the reality of the suboptimal solution. BrokenSegue (talk) 21:24, 24 June 2021 (UTC)
  • I think we should expand the allowance of videos beyond instance_of=video, maybe we don't need any restrictions. I can see having a video of a particular plant, or a location, or an engineering project. Rarely will someone create an entry just for the video itself. --RAN (talk) 02:39, 26 June 2021 (UTC)

curid properties

Since MediaWiki page ID (P9675) is created, should we migrate 17 existing properties to new properties using page names? This is for consistency.--GZWDer (talk) 09:19, 25 June 2021 (UTC)

What apart from consistency would be the benefit? Consistency is small benefit if you consider the additional work that has to be done in keeping page titles up to date. --Emu (talk) 09:42, 25 June 2021 (UTC)
Maybe add curid to those using titles? --- Jura 09:04, 27 June 2021 (UTC)

Re: Annoying merge limit

I used Special:MergeItems; this has worked before. It doesn't work when both items have descriptions on the same language. Earlier, the descriptions could be removed manually, then Special:MergeItems could be easily used. I have now been reminded of the merge gadget. I will try it the next time. Utfor (talk) 13:03, 26 June 2021 (UTC)

distance education

Hello. How can we add the information that the person took his bachelor, master, phd etc by distance education from a university? For example,

. Data Gamer play 12:23, 27 June 2021 (UTC)

Perhaps use has characteristic (P1552) as a qualifier with the value distance education (Q159595)? --Tagishsimon (talk) 13:30, 27 June 2021 (UTC)

Why do we require that the victim of a crime have "victim" as object_has_role

Look at murder of George Floyd (Q95579249) and we have the property:victim=George_Floyd and then we require below it object_has_role=victim. It appears redundant, and if we want redundancy, shouldn't it be subject_has_role=victim (when we are writing about a person). --RAN (talk) 14:12, 22 June 2021 (UTC)

I can see situations where it would be useful. For example, if we added it to a mass shooting incident, we may have a number of dead victims, a number of wounded victims and a number of unwounded victims of a hostage situation. Assuming that all of those victims have wikidata entries, we would want to distinguish the ones who died from the ones who survived but were wounded and the ones who survived but were unwounded. However, I am not sure whether object has role (P3831) should be a mandatory qualifier where there is only one victim of the incident. From Hill To Shore (talk) 07:25, 26 June 2021 (UTC)

"village/town/city in Lebanon"

Q23828039: "village/town/city in Lebanon"... Is this good? Correct? Necessary? Have we got similars? Why? Is it not good enough to have "village in L", "town in L" and "city in L"? Are they so similar to each other? I have seen Beirut and environs, I could observe clearly that the Lebanese capital was a city, and a beautiful one, against all odds, and Baalbek was a town, and some other places (whose names I do not remember) were probably villages... --E4024 (talk) 23:24, 26 June 2021 (UTC)

There is "village in..." for some countries, surprisingly. If people like their village not to be found by normal SPARQL queries, then this is the way to go. I suspect it rather laziness to do it right. --SCIdude (talk) 06:31, 28 June 2021 (UTC)

language of work or name (P407) and Chinese

Is it only me who think Standard Taiwanese Mandarin (Q262828), simplified Chinese characters (Q185614) and written Chinese language (Q3110592) shouldn't be used with this property? Chinese (Q7850) and Simplified Chinese (Q13414913) seems like more fitting replacements --Trade (talk) 09:57, 28 June 2021 (UTC)

Wikidata weekly summary #474

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • Mismatch Finder: We started building the foundation of the place that will store the mismatches between Wikidata and other databases/catalogs/...
    • Continued work on migrating the checks of regular expressions for constraints from SPARQL to a better solution to take that load off of the SPARQL endpoint (phab:T176312)
    • Made it possible to show language links from multilingual Wikisource to the other language versions of Wikisource (phab:T275958)
    • Added a new constraint type to indicate that a certain Property should only be used on Lexemes with a specific language (phab:T200689)
    • Fixed a weird issue with suggesters popping back up when they shouldn't (phab:T284219)
    • Invalid data will not be handled better in RDF outputs (phab:T285131)
    • Working on lowering the rate limit at which a misbehaving bot can waste new Item IDs to further reduce the percentage of skipped Item IDs (phab:T284538)
    • Working on not suggesting real Item IDs in the API sandbox to avoid accidental edits by people who think it is not making real edits (phab:T219215)
    • Working on reducing the time it takes between entering a value in a new statement and being able to save the statement (phab:T281669)
    • Improved documentation of the usage tracking aspects in API help pages. Usage tracking is an internal mechanism to track which article on Wikipedia etc uses which data from an Item on Wikidata. (phab:T283040)
    • Language codes for monolingual strings "gsw-fr", "ykg", "wya", "osa-latn" were made available: Alsatian, Tundra Yukaghir, Wendat, Osage (phab:T262922, phab:T252198, phab:T283364, phab:T265297)
    • The English name for language code "crh" and the Swedish name for "fa" were corrected (phab:T240350, phab:T281702)
    • Lexeme language codes ha-arab, sux-latn, sux-xsux, gsg, tlh-piqd, tlh-latn, bfi, pwn, enm were added. That is Hausa in Arabic script, Sumerian in cuneiform and Latin-script, German Sign Language, Klingon in pIqaD and Latin script, British Sign Language, Paiwan and Middle English. (phab:T282512, phab:T279557)
    • It was determined that language codes can be activated for Wikidata while being blocked for use on Incubator (phab:T273705), this to avoid projects such as a "British English Wikipedia"
    • Language code "es-419" for Latin American Spanish has been available for labels and descriptions for quite some time (phab:T230786)
    • A possibly confusing Russian mis-translation of the name of the language code for "multiple languages" (mul) is being reviewed (phab:T245927)
    • "en-simple:" can be used instead of "simple:" to link to Simple Wikipedia. Query Services outputs "en-simple", not "simple" for sitelinks to Simple Wikipedia (phab:T283149).
    • There was some discussion about the creation of a language code "en-in" for monolingual strings, but "en-in" as interface language seems to be preferred (phab:T212313)

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.


multiple warehouses

Try as I might I can't see any significant difference between Q181623 and Q1362225. Should they be merged? Rodhullandemu (talk) 12:22, 21 June 2021 (UTC)

I haven't checked all languages but de and nl wikis are both pointing to different articles, so a merge is impossible. Perhaps someone from those wikis can explain the differences we need to map here on each item. From Hill To Shore (talk) 13:15, 21 June 2021 (UTC)
In Dutch (and probably German too), warehouse (Q1362225) seems to refer to a historical type of warehouse (Q181623). In my mind they are almost subclasses, i.e. any "pakhuis" is a "warehouse", but not every "warehouse" is a "pakhuis". --Azertus (talk) 13:51, 29 June 2021 (UTC)
  • Q181623 has "lugar o espacio físico" in Spanish, "building for storing goods" in English, "bâtiment (ou action)," in French: these are at least three different concepts. --- Jura 12:15, 27 June 2021 (UTC)

In the deWP de:Lagerhaus, the german name of the item 181623, is a redirect to de:Lagerwirtschaft, which is an article about storage in general, not the building. The article de:Speicher, the german name of the item 1362225, is connected with the item Q9158768, as well called Speicher, about the general concept of storage places, while the article about a building has a different article name, a for me up to now unknown word, de:Bodenspeicher. Both names have as well disambiguation pages: de:Speicher (Begriffsklärung) and de:Lagerhaus, where different meanings of those words are listed. So I think it's complicated, and we should first decide wich item has wich real world meaning, then sort out the connected articles and disambiguation pages. Grüße vom Sänger ♫ (talk) 12:31, 27 June 2021 (UTC)

Tour de France 2021 participants

Since the 2021 Tour de France (Q61778309) has begun I wanted to start adding the participant in (P1344) property to the riders using the table that is part of https://en.wikipedia.org/wiki/List_of_teams_and_cyclists_in_the_2021_Tour_de_France

Unfortunately, a lot of the riders are now semi-protected which keeps me from editing them. Also, adding this manually isn't a lot of fun. Is there a way someone could automatically add the participant property using the Wikipedia table (from table entry to Wikipedia entry to Wikidata entry)? Maybe there are scripts for this type of tagging which somebody can run. --176.199.18.119 15:43, 26 June 2021 (UTC)

Try this tool https://www.wikidata.org/wiki/Wikidata:Tools/OpenRefine Bouzinac 💬●✒️●💛 14:21, 27 June 2021 (UTC)
If you want to engage with Wikidata it's very useful to register an account. ChristianKl ❪✉❫ 14:13, 29 June 2021 (UTC)

Question on copying references between items

Is there a tool that lets you copy references from a statement on one item to a different statement on another item? I am already using the DuplicateReferences gadget, which allows you to copy references between statements on the same item. I am also using the Move gadget, which lets you copy whole statements to other items or copy references where there is an identical statement already on the other item. Using these tools together, I can copy an incorrect statement and reference to another item, copy the reference to the statement where it is needed and then delete the incorrect statement. I can continue to do it this way if that is the most effective method but I am hoping there is another tool that skips the part where I import and delete the incorrect statement. From Hill To Shore (talk) 20:37, 27 June 2021 (UTC)

  • You could try to query the data of the source item in a format suitable for input with QuickStatements for the target item(s).
Initially I was going to suggest using wikibase-cli (Q87194660), but that might be more complex. --- Jura 14:14, 29 June 2021 (UTC)

Automatic deletion of Wikidata items when all Wikipedia links are removed

I always thought that Wikidata was a database "to represent all the things in human knowledge". Just because an item fails Wikipedia notability doesn't mean it automatically fails Wikidata notability. Wouldn't it be more constructive to add references to these item rather than delete them? Piecesofuk (talk) 09:58, 28 June 2021 (UTC)

Another example is chinese subways stations being deleted by MisterSynergy : they are being built. Under zh wikipedia rules, their articles are deleted but they still are notable, being build and being notable sooner or later. Bouzinac 💬●✒️●💛 13:11, 28 June 2021 (UTC)

  • In both cases MisterSynergy likely ran some query and decided to delete the relevant items. While it might be true that the person in the first example is a major, that information is not on the Wikidata item. ChristianKl ❪✉❫ 13:23, 28 June 2021 (UTC)
  • @Piecesofuk: There is scarce information on Stan Lawlor (mayor in the 1980s) that could be referenced in Wikidata, other than to state he was a former mayor who has tried to return to the job in recent years[13]. It appears that one would have to go back into local Canadian newspapers or some other historic documents to find better references from the 1980s. It sounds like there was no references on the Wikidata item at the time of deletion and the item was 'orphaned', so it would have been fair that it was deleted. Now that some searching on the Internet has been conducted and references found, the item could be recreated, as could items for other mayors of the same council where there is trustworthy references available. --Dhx1 (talk) 01:28, 29 June 2021 (UTC)
    • @Dhx1: My point is about one of the reasons for Wikidata: "Wikidata is a free, collaborative, multilingual, secondary database, collecting structured data to provide support for Wikipedia, Wikimedia Commons, the other wikis of the Wikimedia movement, and to anyone in the world." I interpret this as if there a list of X on a particular language's Wikipedia page, then every item of X should be on Wikidata whether or not the item of X has a Wikipedia page, that is so that the list of X can be reproduced (generated from Wikidata) in any other Wikipedia language. Also for the same reason any entity that is contained within the body of a Wikipedia page is notable enough for Wikidata Piecesofuk (talk) 07:02, 29 June 2021 (UTC)
The general idea is correct, but you can easily run into difficulties when you try to implement this in practice:
  1. Something may be listed on a Wikipedia page in some unstructured way that we cannot see from here. There are currently ~900 (AFAIR) Wikimedia wikis, we cannot search for unstructured notable mentions of entities everywhere. The pages/Wikipedia lists need to actively use data from Wikidata (may be invisible to Wikipedia readers) in order to leave a trace on Special:EntityUsage in the projects.
  2. Often, not much is know about an entity, apart from the fact that it is member of some list. Particularly, it might be difficult to support this membership with a reference.
  3. A somewhat related question that is often being discussed in some way here is to which degreee relatives of notable persons should be deemed notable as well. Their names might be listed in infoboxes and Wikipedia articles sometimes, but does this warrant a separate item? Particularly if nothing but a name is known? We need to consider that this project (as well as Wikipedia itself) is a privacy invasion to peoples' lives; this is to some extent inevitable for public figures, but not for their relatives.
—MisterSynergy (talk) 08:12, 29 June 2021 (UTC)

Some background:

  • I find these cases via the report at User:Pasleim/Items for deletion/Page deleted (and its archives).
  • Items can be removed from this report page manually if deemed notable, or via deletion (a bot purges deleted items once a day).
  • There used to be a backlog of ~100k items a while ago. We are meanwhile down to <20k.
  • Some idea about what is listed in this report (and archives) can be gained from User:MisterSynergy/sysop/pagedeleted stats (updated once per week).
  • Experience shows that a majority of listed cases is abandoned, unused, and completely unreferenced content that usually remains unnoticed und untouched for years. There are also quite some items with potentially problematic content.
  • For roughly four years meanwhile, my approach is to check whether these items technically meet the notability requirements (sitelinks, any sort of external references, identifiers, backlinks, use in Wikimedia projects, etc), in order to delete them in case they don't.
  • During the process, I do not search for external references actively, as this is pretty difficult in many cases anyways. Many of the listed cases contain rather generic information only which does not allow anyone without expert knowledge to tell what the item is actually about.

Happy to answer questions. —MisterSynergy (talk) 14:44, 28 June 2021 (UTC)

  • Thanks for the link to the Items for deletion report, I'm having a look through it too see what I can remove (one so far). I think more editors should be involved in looking through these lists, and doing thorough reference searches. I think a lot of notable items are getting lost: Yahoo! Shopping! Q8046872 Piecesofuk (talk) 15:51, 28 June 2021 (UTC)
  • Thank you MisterSynergy, your work is really useful and helps Wikidata staying as clean as possible, please continue to do so. But what I find it sometimes annoying is that a deletion is cumbersome. For instance, for a deleted item, it's hard to find out what was it and dispute the deletion action. Bouzinac 💬●✒️●💛 17:49, 28 June 2021 (UTC)
    • The deletion interface pre-fills the deletion reason with one of the labels of the item by default, which then appears in the deletion log and on the deleted item page. This should help a lot already, although some admins prefer to remove this information; sometimes, however, it better should be removed because it can be inappropriate in some way.
      You can always ask the deleting admin (or any admin) in case you are not sure about the deleted item content just from the available information.
      Also, please be aware that undeletion requests do not need to fulfill any particular form or requirements, so in case you think there is something which shouldn't have been deleted, you usually get the item back without much discussion. —MisterSynergy (talk) 18:40, 28 June 2021 (UTC)

Problems when mergin

I tried to merge two obiously duplicated Q, Q64141157, and Q64682082, without success. Any thoughts?--TaronjaSatsuma (talk) 07:15, 29 June 2021 (UTC)

How to model a job vacancy advertisement (Q356251)?

I am involved in a project — Towards an open, zoomable atlas for invasion science and beyond (Q107368720) — that is closely linked to WikiProject Invasion Biology (Q56241615) and currently has an open job vacancy advertisement (Q356251), so I tried to create an item for that: Q107368777. It looks like it was the first instance of (P31) of a job vacancy advertisement (Q356251), so I did not have much to go by in terms of the data model, and suggestions for improvements would thus be welcome. --Daniel Mietchen (talk) 13:50, 29 June 2021 (UTC)

OK, thanks for the feedback. --Daniel Mietchen (talk) 22:15, 29 June 2021 (UTC)

How to understand when an internet page was created?

Does anybody have the tools to tell me if this internet page is newer than our item Jahrein (Q99001082)? Thanks in advance. --E4024 (talk) 17:23, 29 June 2021 (UTC)

The website has been around since at least 2016. However, Internet Archive's Way Back Machine has no copies of that specific page. That could mean it is an old page that has been missed or a newer page that hasn't been copied yet. I'm not sure what other methods are available for checking the age of the page, other than asking the site owner. From Hill To Shore (talk) 17:45, 29 June 2021 (UTC)

Presenting data in a more tabular format

When larger datasets of information are placed in wikidata, it usually results in an item page that scrolls forever, see COVID-19 pandemic in Colombia (Q87483673), which has six large properties with hundreds of values, one for each day. Completely valuable I think, but viewing the data on wikidata isn't great.

At the top of item and property pages, there's usually the big table of languages, showing the name of it in various languages, it's a much neater table.

Why don't we have that as an option to display large amounts of data? In the above example, a table labelled number of deaths (P1120) would have a column labeled point in time (P585), with the values in each row, and a column labelled "Quantity", which is the data type of number of deaths (P1120). And if you mouse-over/click a quantity cell, you could see the references, or other qualifiers that haven't been given their own column. Thoughts on how large datasets should be presented? Supertrinko (talk) 22:55, 23 June 2021 (UTC)

See commons:Category:Tabular data. Wostr (talk) 00:20, 24 June 2021 (UTC)
Commons:Tabular data is a completely failed experiment: established 2016, only ~150 categories, probably fewer than a couple of thousand datasets, no meaningful support for using the data. It's not an answer to the question posed.
fwiw, historically it's been the case that WD items have been considered unsuited to storing large tabular datasets, because large item sizes (Q87483673 is 4.3 meg, ~2000 statements) caused report server lag at RDF serialisation time. Not sure if a recent-ish change in serialisation overcomes this issue. --Tagishsimon (talk) 01:49, 24 June 2021 (UTC)
Incidentally, I was adding references and needed to add the article with DOI 10.1080/15548627.2020.1797280 but refrained because it has 2,929 authors! Will there be quasi-permanent limits to Wikidata? --SCIdude (talk) 09:18, 30 June 2021 (UTC)
Tabular data is essentially the feature that does this. On the other hand, the tooling around it isn't that great and we haven't used it as much. Maybe we need some gadget to display tables inline in Wikidata? ChristianKl ❪✉❫ 15:30, 30 June 2021 (UTC)
yeah, i think there's a few unrelated questions here that are being conflated. how can we improve how data is stored to make it more efficient? and how can we render the data in a more usable manner? they can be worked on independently but doing the latter without a plan for the former seems unwise. a simple solution to the original problem of " six large properties with hundreds of values" though would be just to collapse sections with more than $N values to keep the UI usable BrokenSegue (talk) 17:58, 30 June 2021 (UTC)
+1. Collapsible properties should be set by default...
A kind workaround to presenting tabular data was greatly made by https://ru.wikipedia.org/wiki/%D0%92%D0%B8%D0%BA%D0%B8%D0%BF%D0%B5%D0%B4%D0%B8%D1%8F:WE-Framework Bouzinac 💬●✒️●💛 19:33, 30 June 2021 (UTC)

Help with Iranian entities (Herat)

Mehri Heravi (Q6400736) and Bibi Heravi (Q27825630) have place of birth (P19) set to Herat (Q45313) (Afghanistan) but I think the right item would be Herat (Q2438142) (Iran). Could someone help to verify this? MarioGom (talk) 16:08, 24 June 2021 (UTC)

User:4nn1l2 is the expert. --E4024 (talk) 02:06, 29 June 2021 (UTC)
@MarioGom: Nope. They are correct. The famous Herat (now in Afghanistan) was part of Iran (actually Persia) until 180 years ago. See w:Siege of Herat (1837–1838). Another source of confusion my be the name of Iran which is a rather new term in English and other European languages. The former name of the country was Persia. Both of these poets are actually Persian, not Iranian. Persian is both an ethnicity and a language. Calling these ancient figures Iranian is anachronistic and somewhat ridiculous, as there was no nation-state named Iran back then. Geographically, they were from w:Khorasan; ethnically and lingually, they were Persian. 4nn1l2 (talk) 02:39, 29 June 2021 (UTC)
4nn1l2: Thank you for the detailed explanation! MarioGom (talk) 08:07, 30 June 2021 (UTC)

How to prevent people from making mess (wrong merges)?

When I already fix some items accoding to Wikipedia content or redirects: split, connect or replace with redirects they are messed or merged again after some weeks or months. How to prevent it? They don't care it is two different things/mereaning they only want to link their article with wrong content. Some examples:

  • Sample ≠ sampling
  • Partition ≠ Partitioning
  • Aston Martin in Formula One ≠ Aston Martin F1 Team

I can list many more. Eurohunter (talk) 11:04, 29 June 2021 (UTC)

Adding different from (P1889) with a qualifier criterion used (P1013) having an appropriate value, to both items, might help. --Tagishsimon (talk) 12:14, 29 June 2021 (UTC)
@Tagishsimon: I don't know which criterion used (P1013) add for different from (P1889). I only use different from (P1889) and people just ignore it because they don't know sample ≠ sampling and if they don't know it I have no idea how they arrived to Wikidata page and edit it yet. Eurohunter (talk) 13:44, 29 June 2021 (UTC)
This is just joke. They really can't even read. Eurohunter (talk) 13:45, 29 June 2021 (UTC)
Pretending that you understand Indonesian and are able to tell which Indonesian descriptions are appropriate feels more problematic. ChristianKl ❪✉❫ 14:10, 29 June 2021 (UTC)
I don't know Indonesian of course and I didn't pretend to know it (never said it) but after editing Wikidata and Wikipedia I had a lot of contact with different latin languages and its characteristic so it is obvious case for me (not for everone else? "Konstruktor Formula Satu asal Jerman" - what else it can mean expect "Formula One constructor from Germany"? This is really simple - it just speaks by himself. I can compare texts and contexts in written articles and use other tools - I mean I don't really like when someone say something like "Pretending that you understand Indonesian" - I'm not blind - I have enough experience, knowledge to know and it's also the reason why I never try to edit Arabic or Japanese. I know there are more problematic sentences or words in every language which I'm not able to understand and I'm not trying to do more than I really can - in some cases of Wikidata items I can't even edit in my native language:) Eurohunter (talk) 14:41, 29 June 2021 (UTC)
You accused an Indian user of not being able to read. That user might very well not speak English. Furthermore the user used the app and is not editing directly via Wikidata, so I'm not sure from where you take your conviction that the user saw your edit summary. ChristianKl ❪✉❫ 15:19, 30 June 2021 (UTC)

Fair use images on Wikidata

Heads up to all, I have opened a feature request ticket on Phabricator for some method of representing fair use images (e.g. non–public domain logos) on Wikidata. Feel free to share thoughts at the ticket. {{u|Sdkb}} talk 21:15, 29 June 2021 (UTC)

That seems extremely premature. Do we have a consensus to allow fair use images on this project? From Hill To Shore (talk) 21:27, 29 June 2021 (UTC)
One of the key problems with implementing fair use on this project is that many Wikimedia Projects use automatic imports of Wikidata images (actually Commons images associated with the Wikidata item). Allowing fair use on Wikidata also means allowing fair use on all of those other projects, where in many cases there is no right of fair use (we will effectively be creating copyright breaches in dozens of countries with the push of a button). To allow fair use here, we would have to disable or heavily modify the image distribution system to prevent fair use images from being distributed outside of Wikidata.
A second problem is that Wikidata use the CC0 licence as we are a project of raw data. Most/all information on Wikidata is not subject to copyright. If we start importing images (as we can't host fair use at Commons) then our entire licensing structure for contributions will need to be changed.
If you want to proceed with this, as a minimum it will require an RfC with contributions from multiple Wikimedia Projects. From Hill To Shore (talk) 23:50, 29 June 2021 (UTC)
  • Is the intention to host fair use images on Wikidata, or just link to the ones hosted on other projects? If it's the former, I'm not sure that's a good idea due to the issues raised by User:From Hill To Shore. But I don't have an issue with it if it's the latter. Actually, we already have non-free artwork image URL (P6500). --Stevenliuyi (talk) 00:19, 30 June 2021 (UTC)
  • What discussion? If fair use was accepted at ENWP it can be accepted at any Wikimedia project because there is no reason to not do. @From Hill To Shore: "we will effectively be creating copyright breaches in dozens of countries with the push of a button" - it's totally not true. There is no difference in access to ENWP, DEWP, FRWP or any other WP from any location. Users from all over the world can upload fair use images to ENWP even in all locations where it's "illegal". Eurohunter (talk) 00:27, 30 June 2021 (UTC)
Each wikimedia property makes its own rules on content, Eurohunter. Whilst fair use could be accepted on WD, it would require community consensus. No such consensus exists. (Nor is it likely, IMO). --Tagishsimon (talk) 00:32, 30 June 2021 (UTC)
@Eurohunter: Commons provides a useful summary of the position at c:Commons:Fair use. There will be other pages around Wikimedia with similar explanations but that Commons page has already been translated into several languages, so is a rather convenient starting point. To implement fair use on Wikidata (beyond providing a link to external non-free content) we will need input from those Wikimedia projects who have refused to accept fair use. From Hill To Shore (talk) 00:38, 30 June 2021 (UTC)
@From Hill To Shore: Better to encourage them to apply fair use right? There is no reason for them to not accept fair use - they just refused to accept fair use due to missunderstanding as users all over the world can upload fair use images to ENWP. Communisties decided in according to their local law which doesn't makes sense. Eurohunter (talk) 00:45, 30 June 2021 (UTC)
I'm sorry but you don't appear to understand the complexities of the situation. "Fair use" is an aspect of US copyright law and protects limited usage of copyrighted materials. Many (most?) legal jurisdictions in the world do not have similar legal protections. As English Wikipedia's servers are hosted in the USA, they are subject to US copyright law and copyright protections. However, if English Wikipedia's servers were moved to the UK then UK copyright law would apply and the "fair use" content would have to be deleted as the UK has no "fair use" law. Similarly, Wikimedia servers hosted in other countries have to obey the laws of their host country. Arguing on Wikidata (or any Wikimedia site) will do little good as Wikimedia does not set the law. Whether an individual outside of the USA has any legal protection in their home country when they upload a non-free image to English Wikipedia is something they have to determine themselves. In many cases I expect they are at a theoretical risk of prosecution (though whether any actual prosecutions materialise is a separate question). From Hill To Shore (talk) 01:02, 30 June 2021 (UTC)
@From Hill To Shore: What ąre you talking about? FRWP, DEWP, ESWP and all other versions has also servers located in the United States, on the same servers. Both ENWP and DEWP and other versions are located on the servers in the United States, Netherlands and Singapore. Eurohunter (talk) 11:02, 30 June 2021 (UTC)
I am not sure why you are trying to argue this with me. A few people here have given you links to the policies and explanations. If you think the Wikimedia projects that don't accept non-free content are wrong, you are welcome to try to convince them otherwise. From Hill To Shore (talk) 11:11, 30 June 2021 (UTC)
@From Hill To Shore: You didn't answer. What is "However, if English Wikipedia's servers were moved to the UK then UK copyright law would apply and the "fair use" content would have to be deleted as the UK has no "fair use" law." according to? You just told that ENWP has some seprate and the only Wikipedia servers in US are ENWP. It clearly isn't true. Eurohunter (talk) 12:18, 30 June 2021 (UTC)
I said that the servers being in the US makes them subject to US law. I then gave a hypothetical example of what would happen if the same servers were moved to the UK. If you don't believe me, then fine. If you want to try to convince the other Wikimedia Projects that they are wrong then that is also fine. If you want to continue this discussion, I would suggest that you include other editors. I won't be responding to you anymore on this topic. From Hill To Shore (talk) 12:26, 30 June 2021 (UTC)
@From Hill To Shore: Okey but what is the point of hypothetical example if all Wikimedia projects including all languages are in same location in the US, Netherlands and Singapore? I just don't see point to mention that if fair use is not allowed somewhere so some Wikipedia versions are not allowed to use fair use. Eurohunter (talk) 14:29, 30 June 2021 (UTC)
@Eurohunter: Whether or not something is fair-use depends on how it's used. Using an image as part of a Wikipedia article can be fair use while at the same time hosting the same image alone on WikiCommons isn't fair-use. ChristianKl ❪✉❫ 14:59, 30 June 2021 (UTC)

Meanwhile: https://meta.wikimedia.org/wiki/NonFreeWiki --Tagishsimon (talk) 01:06, 30 June 2021 (UTC)

  • Last year there was a RFC concerning non-free content. There was no consensus to implement a non-free policy. Comments made in that RFC might be relevant to this discussion. --Shinnin (talk) 01:43, 30 June 2021 (UTC)
  • Welp, this got a fair bit more attention than I was anticipating, and some of the concerns here seem to be reacting to a proposal much bigger than what I'm actually seeking (partially my fault for not being more precise with the header language, and partly all of your faults for not reading the phab description as asked). To answer Stevenliuyi's question, it's the latter—my thought would just be for there to be a way to link to fair use images on another wiki, not host them here. As at the example I gave at the phab ticket, just like we link to Commons files, there should be a way to link to w:File:American_Airlines_logo_2013.svg for logo image (P154) at American Airlines (Q32396). We can debate about how precisely to implement, but to say that we don't want to document that piece of information would be to abandon our purpose as a knowledge base. {{u|Sdkb}} talk 03:53, 30 June 2021 (UTC)
If you just want that, then a new property along the lines of Commons compatible image available at URL (P4765) or non-free artwork image URL (P6500) and which does not attempt to display the image would suffice. But if that's what you wanted, there would be no need for a phab ticket. I think you're backpeddling, but not very successfully. --Tagishsimon (talk) 03:59, 30 June 2021 (UTC)
@Sdkb: No, I read the phab ticket before responding. As you said there and have repeated here, you want to implement a similar mechanism to how Commons images are displayed (and not just a link as you half suggest now). Displaying fair use images here brings up the copyright issues. If we give a simple link, we are saying, "this other site offers an image that may be in breach of copyright law but that is their problem and your problem, we just have a link," while displaying a non-free image here says, "we have published an image that may be in breach of copyright law." In the latter case, as you are trying to implement, you need consensus to change fair use policy on Wikidata.
Similarly if the same mechanism for displaying Commons images is used but with a pointer to the English Wikipedia server, the risk of replicating the fair use image on other Wikimedia projects also materialises, which may breach their local policies on non-free images. A larger discussion would be needed before implementation to work out what parts of the Commons image display mechanism are replicated and which parts are locked down.
Your proposal to display the image is the key problem. If you just want a link, you can follow Tagishsimon's advice. From Hill To Shore (talk) 07:25, 30 June 2021 (UTC)
@Tagishsimon, From Hill To Shore: Yikes. I just had to check that this isn't a redlink. At the phab description, I explicitly talked about linking the example, not displaying it. If you want to say I was sloppy in how I presented this, fine, fair enough. But to accuse me of backpedaling after I clarified my intent is a brazen breach of AGF and not helpful for moving the discussion forward. I'm not inclined to create a new property proposal given the reception here, but I hope someone else does so. {{u|Sdkb}} talk 07:43, 30 June 2021 (UTC)
Alternatively we can repurpose non-free artwork image URL (P6500) for any non-free image of the subject (not necessarily artwork) — Martin (MSGJ · talk) 08:04, 30 June 2021 (UTC)
@MSGJ: That sounds like purposefully violating Wikidata consensus about what properties should exist. ChristianKl ❪✉❫ 15:09, 30 June 2021 (UTC)
@Sdkb:To the extend that you are creating a Phabrictor ticket that asks to violate the property creation consensus, that's still bad. There's a place to discuss Wikidata policy and that isn't Phabrictor. ChristianKl ❪✉❫ 15:10, 30 June 2021 (UTC)
@Sdkb: You stated this discussion was "partly all of your faults for not reading the phab description as asked." That could be read as an accusation of bad faith (though probably not your intention). I explained that I had read your description and what your request actually says by requesting the same functionality as Commons images (what it says and what you intended may not be the same). I have not accused you of bad faith editing or of backpedalling, so including me in the ping where you make your own reference to AGF is puzzling. You repeated again in your reply above your intention that we implement this "just like we link to Commons files." My reply was saying that implementing this in the same way as Commons is the core of the problem. Any link here must be made in a different way to Commons or else we need to debate fair use. Rather than continue to talk about good faith or bad faith editing, I'd suggest drawing a line under this and moving on. From Hill To Shore (talk) 09:12, 30 June 2021 (UTC)

disambiguation items following page moves

I recently retitled the human name disambiguation page en:w:Kenny Casey to en:w:Kenneth Casey (disambiguation). There appear to be no entities named "Kenny Casey" on Wikidata in any language, save for the disambiguation page. Should Q17129502 be renamed, and thus every label changed from Kenny to Kenneth (and is there a tool to do this all at once besides tedious copy-paste)? Or should a new disambiguation item be created? -Animalparty (talk) 21:31, 30 June 2021 (UTC)

@Animalparty: Create a new disambiguation page item for "Kenneth Casey", and move the sitelink there. Wikimedia human name disambiguation page items without sitelinks will appear at Wikidata:Database_reports/to_delete/empty_human_name_disambiguation_items, and an administrator will delete them at some point. --Shinnin (talk) 21:43, 30 June 2021 (UTC)
Thanks. Done. Kenneth Casey (Q107386461). -Animalparty (talk) 22:21, 30 June 2021 (UTC)

Deprecated "one of" constraint

I see several such claims at some properties. E.g. https://www.wikidata.org/w/index.php?title=Property:P105&action=history. Do they really help for suggestions? Do they really create violations? --Infovarius (talk) 13:26, 25 June 2021 (UTC)

After my removal no ranks are suggested, but the massive constraint violations are gone. --Succu (talk) 19:10, 25 June 2021 (UTC)
  • Deprecated constraints are (correctly) ignored by the constraint checking gadget, but KrBot still lists them.
See Help:Property_constraints_portal/Entity_suggestions_from_constraint_definitions and Help:Suggesters_and_selectors about the suggestions from contraints (even deprecated) and suggesters in general. --- Jura 09:01, 27 June 2021 (UTC)
Constraints should not misused as suggestions. This is a bad practice. A new property should be created for this purpose. --Succu (talk) 18:58, 28 June 2021 (UTC)
You could file a phab request if you think it should be changed. --- Jura 19:23, 4 July 2021 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2021/06&oldid=1458489270"