Wikidata:Project chat/Archive/2019/09

From Wikidata
Jump to navigation Jump to search

Broken constraint on subclass of (P279)

I can't figure out what the value-requires-statement constraint (Q21510864) constraint is trying to achieve here, but it's broken due to multiple property (P2306) values. As a pretty widespread property it would be good to get it fixed. --SilentSpike (talk) 10:43, 2 September 2019 (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. --SilentSpike (talk) 11:17, 2 September 2019 (UTC)


They are the same article published in two editions of a journal. I don't know whether they should be merged, or there should be an property to link these two items. --GZWDer (talk) 20:45, 22 August 2019 (UTC)

✓ Done --Liuxinyu970226 (talk) 14:07, 29 August 2019 (UTC)
  • In the logic of @Daniel Mietchen:'s bot, I think they probably should remain separate. The items are per published article, not per work. In merged form, the page numbers and volume indications become meaningless. Still, in one way or the other, these need to be linked. --- Jura 07:52, 1 September 2019 (UTC)

Changing a statement in multiple articles

Sometimes several items have the same statement with the same incorrect item linked. For example on items with heritage designation (P1435):Grade II listed building (Q15700834), any with located in the administrative territorial entity (P131): Ashdon (Q26001112) should be Ashdon (Q1832065) (62 items) and any with located in the administrative territorial entity (P131):Luton (Q11797266) should be Luton (Q21683239) (78 items). QuickStatements can be used to add statements but is there a tool that can be used to change them? Peter James (talk) 15:13, 31 August 2019 (UTC)

@Peter James: The very crude way to do it is PetScan - you can feed it a SPARQL query then say "remove all claims P1:Q1234, add claims P1:Q5678, for all items". However, this will lose any sources, qualifiers, etc. If there aren't any of those, then great!
You can also use QuickStatements to do exactly the same thing (remove one claim, add another). Again, you'll lose sources/qualifiers, but you can also add them back in using QS.
Finally, a very sophisticated approach is to use something like wikidata-cli - this will let you target specific statements & claims, so you can keep qualifiers intact while switching the main values. It's fiddlier to work with but substantially more powerful. Andrew Gray (talk) 17:59, 31 August 2019 (UTC)
In case *all* items should be moved, you can ask at WD:RBOT for someone to create a script to move those items for you. A smart script can move sources and qualifiers too. Edoderoo (talk) 07:46, 1 September 2019 (UTC)

for undo need time estimate

For the option to undo a group of edits, please post a time estimate. In mid-to-late July, I saw that the original edits by someone else had taken 1-2 minutes for one group of 34 edits and my undo was not selective, so it should have been relatively simple and I assumed it would take a couple of minutes. Twenty minutes later, only 8 edits had been done. Over an hour after starting, only 8 edits are done. Eighty-five minutes after starting and after opening a new browser (Arora), it's still only 8. This leaves me with a dilemma: when my session time is about to run out and I have to log off, I don't know whether to stop undoing or let it keep undoing while I'm not logged in, if it will even permit that. (A discussion with the other editor suggested that all my 34 edits were completed, although I don't know when.) If the process would have run more than a minute, some rough time estimate before I started would have been helpful. Nick Levinson (talk) 21:06, 31 August 2019 (UTC)

  • From my vague memory, I think it runs without the browser window being open and can be started a second time. --- Jura 10:32, 1 September 2019 (UTC)

Item update request

On singer-songwriter (Q488205) please add “piosenkarka” (Polish) on female form and “cantautore” (Italian) on male form. Thanks!!! -- 14:28, 1 September 2019 (UTC)

Hi. Done for the Polish. But "Male form" property don't exist, and it is already the Italian label of the item. Cheers, Nomen ad hoc (talk) 14:32, 1 September 2019 (UTC).
male form of label (P3321) Visite fortuitement prolongée (talk) 14:56, 1 September 2019 (UTC)
Ah, didn't know! Sorry. Nomen ad hoc (talk) 15:05, 1 September 2019 (UTC).

item fix request

writer (Q36180) have 2 form for italian female form: “scrittice” and “scrittrice”. Please delete “scrittice” (without R) and leave “scrittrice” (with R). Thanks! -- 16:56, 1 September 2019 (UTC)

Federation - your input is needed

Hi everyone :)

We do want to allow more and more people to run their own Wikibase instance. As part of that we want to allow people to deeply connect these Wikibase instances among each other and especially with Wikidata. We also want to allow them to benefit from work done in other Wikibase instances and especially Wikidata, like the definition and refining of properties. As people talk and think about this it is often referred to as Federation. However there is the problem that we don't talk about the same thing when we say Federation but instead about a multitude of related features. In order to move the development forward we need to sort out what people actually want and need to make sure we're actually developing something useful. I opened a page for collecting your input at Wikidata:Federation input.

Cheers --Lydia Pintscher (WMDE) (talk) 18:27, 1 September 2019 (UTC)

Language barriers - your input is needed

Hi everyone :)

In Wikidata people from many different backgrounds work together. As Wikidata grows we can't rely on everyone speaking English anymore. With the development of Wikidata Bridge (the ability to edit Wikidata directly from the other sister projects) we will see another surge in contributions from people who potentially will have to discuss with each other about the content we have without speaking a shared language well enough for that. It was also one of the major topics identified in the session about obstacles to more Wikidata adoption in Wikipedia at Wikimania 2019.

This is not an easily solvable topic but one we need to think about and your input would be very helpful. I opened a page for it at Wikidata:Language barriers input.

Cheers --Lydia Pintscher (WMDE) (talk) 18:28, 1 September 2019 (UTC)

Qualifying features/qualities

Given the feature lightfastness (Q909369), how do we indicate that a substance has "poor" lightfastness or "excellent" lightfastness as stated in our sources? I can't seem to find items for "good", "poor", "fair", etc,. as measures of quality. Perhaps I am missing something? And what qualifier property would we use? - PKM (talk) 20:15, 1 September 2019 (UTC)

"Subject has role" as statement

subject has role (P2868) says it should be used as a qualifier, but when it is used as a statement (as in carmine (Q320617)) there is no constraint violation. I believe "subject has role" used as a statement would be better as use (P366). What do you think? And should there be a constraint? - PKM (talk) 21:04, 27 August 2019 (UTC)

There should absolutely be a constraint: the property is mean solely to qualify the target of a statement, so it makes semantically no sense at all for it to be a main statement! Therefore I have replaced the baffling "conflict-with instance of (P31) film (Q11424)" that was there with the appropriate scope constraint. Circeus (talk) 23:21, 27 August 2019 (UTC)
@Circeus: thank you. Now to find someone who can make a bot to change all the "Subject has role" statements to use (P366) statements, preserving values and references. - PKM (talk) 02:34, 28 August 2019 (UTC)
Actually, a bot is a bad idea now that I think about it; some may be acting roles rather than uses. - PKM (talk) 02:45, 28 August 2019 (UTC)
There are 3751 uses right now, apparently mostly in the field of biochemistry. It was proposed with the biochemical/biomedical field in mind, so maybe someone from a related project could say more about the use of subject has role (P2868) and if it may be replaced with use (P366).

Andrew Su
Marc Robinson-Rechavi
Pierre Lindenbaum
Michael Kuhn
Dan Bolser
Timo Willemsen
Salvatore Loguercio
Daniel Mietchen
Ben Moore
Alex Bateman
Vojtěch Dostál
Andra Waagmeester
Elvira Mitraka
David Bikard
Dan Lawson
Francesco Sirocco
Konrad U. Förstner (talk)
Chris Mungall (talk)
Kristina Hettne
Karima Rafes
Finn Årup Nielsen
Jasper Koehorst
Till Sauerwein
Amos Bairoch
Was a bee
Muhammad Elhossary
Damian Szklarczyk
Yusra Haider
Hannes Röst
Kritika Dusad
T.Shafee(evo&evo) (talk)
Pictogram voting comment.svg Notified participants of WikiProject Molecular biology Saehrimnir
Jasper Deng
Egon Willighagen
Denise Slenter
Daniel Mietchen
Emily Temple-Wood
Pablo Busatto (Almondega)
Antony Williams (EPA)
Devon Fyson
Samuel Clark
Tris T7
Robert Giessmann
Cord Wiljes
Jonathan Bisson
Charles Tapley Hoyt
Peter Murray-Rust
Pictogram voting comment.svg Notified participants of WikiProject Chemistry Tobias1984
Doc James
Daniel Mietchen
Andrew Su
Projekt ANA
Pavel Dušek
Was a bee
Chris Mungall
Dr. Abhijeet Safai
Sami Mlouhi
Netha Hussain
Abhijeet Safai
Shani Evenstein
ZI Jony
Pictogram voting comment.svg Notified participants of WikiProject Medicine - Valentina.Anitnelav (talk) 07:03, 28 August 2019 (UTC)

There are also quite a few position or occupation statements using subject has role (P2868) that should be moved to position held (P39) or occupation (P106). subject has role (P2868) is also used to indicate the role of a game character in terms of game mechanics (player character, etc.), but this may be expressed as a qualifier to characters (P674). - Valentina.Anitnelav (talk) 07:18, 28 August 2019 (UTC)
The chemical items that have "subject has role" mainly have it because of the ChEBI import---their ontology use it to indicate type of usage of that compound by humans. For example drugs have some medical use, or pesticides have some agricultural use. As there is no other species than humans using chemical compounds, the "use" property fits perfectly too IMHO. --SCIdude (talk) 19:16, 28 August 2019 (UTC)

object has role (P3831) seems likely

Wikidata:Database reports/Constraint violations/P3831 tells me that there are 4 items that use P3831 as main statement. --Liuxinyu970226 (talk) 04:35, 2 September 2019 (UTC)

@, Alotrootropico, Moebeus, Shsucoll:^^ --Liuxinyu970226 (talk) 04:36, 2 September 2019 (UTC)
@Liuxinyu970226: I fixed those four violations, three were mistakes, one vandalism. Moebeus (talk) 06:40, 2 September 2019 (UTC)

"File usage on other wikis" for Wikidata items?

Is there a way to find "File usage on other wikis" for Wikidata items?

If there is an article devoted to a specific Wikidata item, that will appear. Also, "What links here" will identify other Wikidata items that reference a particular wikidata item.

Items in commons.wikimedia include "File usage on other wikis". I'd like to see that in Wikidata.

Example: Earlier today I edited the Wikipedia article on w:Kenneth Snelson to include a link to Triple Crown sculpture (Q66839784). I don't know how to find that from Wikidata.

More generally, it would be nice to be able to identify all the Wikimedia Foundation project links to a particular Wikidata item. If there aren't many, a user might want to check them all to see, e.g., if a related reference might be useful in some of those other articles.

Thanks, DavidMCEddy (talk) 04:05, 31 August 2019 (UTC)

--- Jura 04:21, 31 August 2019 (UTC)

Thanks very much.
May I ask what are the "Statement: P123, ... Title, Label: en" in *
Anyway, I'm glad to know how to get that information. I think it would be easier to find if it were the same as Wikimedia Commons, but at least it can be found.
DavidMCEddy (talk) 22:06, 31 August 2019 (UTC)
These are specific pieces of data which are used on a page. --Matěj Suchánek (talk) 06:40, 2 September 2019 (UTC)

"Please give a reason" spam from IPv6s

e.g. [1] should we set up an AbuseFilter to hold up so? --Liuxinyu970226 (talk) 04:56, 2 September 2019 (UTC)

Wikidata weekly summary #380

Wikidata read-only period on September 10 at roughly 05:00 AM UTC

Due to on-site maintenance for some of our database servers, there will be a read-only period for Wikidata on September 10. You can read more in phab:T226778 and phab:T230788. This means that the wiki can be read but not edited. The read-only window is 05:00AM UTC to 05:30 AM UTC, but if everything goes well the actual period of no editing will be shorter. When editing on Wikidata, it will be apparent that the changes are not working. If you've got any questions, the Phabricator tickets are probably the best place to ask. /Johan (WMF) (talk) 07:58, 21 August 2019 (UTC) Fake timestamp to not have it archived before the maintenance: 05:30, 3 September 2019 (UTC)

It is September first and I notice regularly moment of read-only statuses. It follows that edits are lost and consequently data integrity is compromised. Thanks, GerardM (talk) 12:28, 1 September 2019 (UTC)

Qualifiers for P571

Is there any discussion or project to talk [deeply] about ontology for inception (P571) ?. Its talk page is not enough.

In 90% of cases P571 is just a single value property, in some cases together with sourcing circumstances (P1480) or collection or exhibition size (P1436) or earliest date (P1319) or refine date (P4241), all of them related with precision of value itself. These are not problematic and are well-defined in Help:Dates.

However, in cases like organizations, countries, municipalities, etc. that may have several "creation dates" about for changes, merging or splitting, it is not clear how to record nor which qualifiers are the most representatives to use. The present collection is large. Beside the above mentioned (the most used), if we look for the real use, there are several qualifiers that may be use alternatively (P3831, P2868, P4649, P793,..). Others are really strange: P580, P582,...(?). The P571 is a "point in time", not an event with start-end date. In some cases related with territorial items (specially state (Q7275)) the different values of P571 are showing the changes of the basic form of government (P122) instead of changes in territory or jurisdiction. Two examples: Afghanistan (Q889) or Spain (Q29). In opposition, Germany (Q183) has no P571; all its recent changes track are in replaces (P1365). Anyway, if somebody has any information about a previous discussion, a project page, are interested in this topic or may help me in any way, I'll appreciate. Thanks, Amadalvarez (talk) 15:28, 2 September 2019 (UTC)

Thanks, @Jura1:. I'd seen and I think is one of the solutions for one of the multiples situations. It could be a good starting point to make a more global analysis of this topic. I asked about P571, but it's obvious that the scope must include replaces (P1365), the boundaries with significant event (P793) and, as the talk mention, how it appears in the items of the new and the previous country. But I don't want focus in state/country only. For instance, when a territorial unit (municipality, region, etc.) change its name (not the boundary nor dependency), do we must create a new item ? And the item of its governor, is it common for both periods/names ? Label should show the difference, so, it seems that we will need two. Is it? or it seems excessive ?. I'd would like to talk/debat/discuss about the best solution and, be able to have a guideline clear and complete. The specific solutions for specific problems are usefull for a short time and we have a mix of "good individual criteris". Do you think we can find a space (and interested people, of course) to explore this common debat about this ontology ?. Thank you, Amadalvarez (talk) 21:17, 2 September 2019 (UTC)
  • I'm sure you can dig up discussions, but in any case, you probably won't get around accommodating globally all approaches, e.g.: 1 item, 2 items (present, past), 1 item (all)+1 (past). Try a consistent approach depending on the country/field you are editing and the WP articles present. --- Jura 22:01, 2 September 2019 (UTC)
It's probably a wise proposal. In fact, there are only 26 countries with qualifications in the P571; 3 other countries without P571, one of them with its history at P31 (NZ), another at P1365 (GER) and the third without any foundation information (FRA). The rest has only one unqualified data. If I ever build a reasonable proposal, I'll let you know. Thanks,Amadalvarez (talk) 06:07, 3 September 2019 (UTC)

Wikipedia article covering multiple topics

It seems that its properties for this type (P1963) value is currently under contesting, as of now, its P1963 values say that main subject (P921) and different from (P1889), as well as P921's Wikidata usage instructions (P2559) says "List subjects which the article covers. Do not use P527 "has part" as parts of an article are the title, sections and paragraphs."

But now one user pointed me that such P1963 values caused problems at some articles, and has part (P527) don't need to be "main" subject in common sense, for the time being I don't know which is the actual consensus on this matter. --Liuxinyu970226 (talk) 22:37, 30 August 2019 (UTC)

Note that I watched its linked-in pages, something e.g. grapefruit and pomelo (Q3406575) are also using has parts of the class (P2670). --Liuxinyu970226 (talk) 23:53, 31 August 2019 (UTC)

@Dhx1: Maybe you can explain more here? --2409:8902:9321:BE18:ABE0:E214:E02D:B7D2 23:46, 31 August 2019 (UTC)
Many of these items are conflation (Q14946528) of multiple distinct entities/topics into one due to Wikipedia often combining similar topics together to aid readability of the encyclopedia. has part (P527) and has parts of the class (P2670) would only be appropriate if referring to items which themselves are a conflation (Q14946528) or Wikipedia article covering multiple topics (Q21484471), or otherwise these properties should be referring to title (Q783521), paragraph (Q194431), etc. The property different from (P1889) is probably the best to use for items which are conflation (Q14946528). If the item is Wikipedia article covering multiple topics (Q21484471) then main subject (P921) is probably also appropriate to use. Dhx1 (talk) 22:25, 2 September 2019 (UTC)
@Dhx1: But then, @Eurohunter: tells me on their talk page that "There are articles about series of cars or other products which has more and less common items.", and that user prefers not to disallow P527 usage for Wikipedia article covering multiple topics (Q21484471) items. --Liuxinyu970226 (talk) 03:02, 3 September 2019 (UTC)
@Liuxinyu970226: @Eurohunter: Have a look at Soundtracks of Grand Theft Auto: San Andreas (Q55675394) and items referenced by different from (P1889). I reorganised the Wikipedia article links to Wikidata items depending on which Wikipedias use any of the variety of conflations of topics. Actual items we should have in Wikidata in the future are Soundtracks of Grand Theft Auto: San Andreas (Q55675394) (all the music contained in the game), Grand Theft Auto: San Andreas Official Soundtrack (Q64547511) and Grand Theft Auto: San Andreas Official Soundtrack Box Set (Q5595150) which are two publications/products released with some or all of the works contained within Soundtracks of Grand Theft Auto: San Andreas (Q55675394). Thoughts? Dhx1 (talk) 11:54, 3 September 2019 (UTC)

Reopening date

Hi. Amsterdam Sloterdijk Zuid (Q2347349) for example was open from 1890 to 1905 and again from 1956 to 1985, but currently ony the first two are given as date of official opening (P1619) and date of official closure (P3999). How should the latter two dates be entered into Wikidata? --HyperGaruda (talk) 20:21, 2 September 2019 (UTC)

Although the nl-lemma mentiones several stations on that location, it were actually multiple stations on multiple rail ways in multple periods. I would suggest to create a single wikidata item for every station in history, and maybe also a "container" to link to the nl-lemma that can be used to link to the historical stations. Edoderoo (talk) 09:42, 3 September 2019 (UTC)

How to present time as a range? (problem of Chinese timekeeping 時辰)

Li Fook Lam (Q7226094)'s exact times of birth and death are recorded on his gravestone, but in Chinese form ja:十二時辰, so they would be ranges of 2 hours. The gravestone reads "生於同治壬申年七月十八日巳時 **民國壬辰年正月十六日卯時", which should be 21 Aug 1872 (I'm not sure. cant find a reliable converter.) and 11 Feb 1952 (this is reliable).--Roy17 (talk) 12:38, 3 September 2019 (UTC)

or (dates/date of birth)

The date of birth of Yosef Aharon Rabinovits (Q11728295) is 1852 or 1853. How to update it properly. Geagea (talk) 18:41, 1 September 2019 (UTC)

As long as there is source for each, both should be added separately. --Matěj Suchánek (talk) 06:33, 2 September 2019 (UTC)
The source says (more then one) 1852 or 1853. Geagea (talk) 11:38, 2 September 2019 (UTC)
@Geagea: Perhaps add both dates with the same source, but apply qualifier sourcing circumstances (P1480) minimum (Q10585806) to 1852 and sourcing circumstances (P1480) maximum (Q10578722) to 1853? Dhx1 (talk) 12:50, 2 September 2019 (UTC)
And what in case whem it's 1852 or 1854 (not between 1852 to 1854). Geagea (talk) 12:56, 2 September 2019 (UTC)
Seems that we need new Q for "or". Geagea (talk) 14:10, 2 September 2019 (UTC)
Sjoerddebruin, you added 1850s to the date. Well it's not wrong but the Wikidata statement should say "1852 or 1853". In some way I see the date issue her some equivalent to othe date template in Commons which have "or". Geagea (talk) 20:13, 3 September 2019 (UTC)

We need a uniform way to handle record labels vs record companies vs long articles about the history of a label

In May 2019 Warner Bros. Records re branded to Warner Records, Warner Bros. Records Inc. became Warner Records LLC, and now it's a complete mess. Long story short: Most Wikipedia articles about larger labels are about more than just the label and usually will cover company history, the founders, the sub-labels, actual releases, etc. Wikipedians (and others!) have a legitimate wish to keep these articles together, but it makes it a nightmare to keep track of the correct, chronological data for music releases, legal records, etc. My suggestion: I think articles like these need to be moved to a meta-grouping like "history of Warner Bros. Records" or something similar. I'm not hung up on the name, VERY open to suggestions, but this is an issue that needs to be fixed or we'll end up reverting each other for ever. What do you guys think? Moebeus (talk) 21:48, 2 September 2019 (UTC)

I agree, but cannot recommend a good solution. There is a similar issue with firearms manufacturers such as Steyr Arms (Q591359) that have changed many names in past or merged with other compaines. --M11rtinb (talk) 10:19, 3 September 2019 (UTC)
We have the same problem with publishers and imprints. - PKM (talk) 20:30, 3 September 2019 (UTC)

external ID URL not working

The molbio property MEROPS enzyme ID (P3717) was never used and I would like to. But you see in the property example the value does not show as link, although the property has a formatter URL. What could be the reason? --SCIdude (talk) 13:41, 8 September 2019 (UTC)

The statement about formatter URL (P1630) has been deprecated. --Matěj Suchánek (talk) 16:01, 8 September 2019 (UTC)
Matěj Suchánek Jura added a normal rank URL (thanks) but still no link (I reloaded the page). What am I missing? --SCIdude (talk) 07:24, 9 September 2019 (UTC)
Purge? --- Jura 07:34, 9 September 2019 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. SCIdude (talk) 08:48, 9 September 2019 (UTC)


Request protection of Q63245258 as it continues to be vandalized with links that do not meet notability requirements. This has been an ongoing issue. 01:35, 9 September 2019 (UTC)

Please see edits by which are only on this property. 02:23, 9 September 2019 (UTC)
✓ Done by Mahir256. --Lucas Werkmeister (talk) 13:02, 9 September 2019 (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. Lucas Werkmeister (talk) 13:02, 9 September 2019 (UTC)

Commons Village pump: One more Wikidata case

There is a discussion on Wikimedia commons about how vandalism on Wikidata is harming that project. The example given was HMendez (talkcontribslogs)'s edits on Wikidata entry Q3071895 were not reverted till six years afterward. Here is a link to that discussion. MorganKevinJ(talk) 00:34, 29 August 2019 (UTC)

In fact it is label vandalism, while linked enwiki article has average 1 visitor per day (source), ptwiki article 17 per day (source) and commons category less than 1 visitor per day. From my point of view just making a mountain out of a molehill.--Jklamo (talk) 09:37, 29 August 2019 (UTC)
  • Vandalism has been happening on every wiki, since wikis were invented. I just reverted one at the English Wikipedia which has gone unnoticed since 2015-06-21.[2] Ghouston (talk) 12:39, 31 August 2019 (UTC)
  • It's simple. Vandalism is everywhere. Vandalism fighters are also everywhere. Depending on how many you got of each, vandalism can be kept under control.. or not. A large wiki like enwiki refuses to use data from Wikidata because it could be vandalized. If the vandalism fighters from enwiki would dedicate some of their time to Wikidata, this problem wouldn't exist. Not only that, it would help all the wikis using Wikidata. Chicken and egg. Many big wikis don't want Wikidata because of vandalism. Wikidata has more vandalized information because the big wikis don't want it. Perhaps a watchlist that shows Wikidata changes that affect a particular wiki would help. Alexis Jazz please ping me if you reply 15:05, 31 August 2019 (UTC)
Not very original I suppose: [Epic] Wikidata watchlist improvements (client). Alexis Jazz please ping me if you reply 15:16, 31 August 2019 (UTC)
... and I suppose all is perfect in Wikimedia Commons. I regularly compare IRL some Wiki-users to politicians. We have similar people on FR Wiki. Jérémy-Günther-Heinz Jähnick (talk) 09:29, 4 September 2019 (UTC)

Compound given names

Louis-Julius (Q50270823) was changed from "Louis Julius" to "Louis-Julius", should we have a separate entry for the French hyphenated version of each duo, or should it be just one entry? --RAN (talk) 05:07, 3 September 2019 (UTC)

"Louis Julius" doesn't sound like a valid combination for a given name item, they should be used separately with qualifiers. Sjoerd de Bruin (talk) 12:04, 3 September 2019 (UTC)
We know your opinion, please wait for more people to respond before you begin changing entries please! The whole point of asking a question is to get consensus first. --RAN (talk) 15:38, 3 September 2019 (UTC)
My thinking is that they should be two separate entries - "Louis Julius" as a double name and "Louis-Julius" as a hyphenated double name. Where I come from it's common to ask if people spell their double names with or without a hyphen and it can also give an indication as to how it's pronounced. Same goes for last names. EDIT: The term "compound given name" is a bit confusing to some - to me "Louisjulius" would be a compound given name, while "Louis-Julius" is a double name. For Spanish speakers the same difference is not there. Moebeus (talk) 12:15, 4 September 2019 (UTC)

Roadmap update: improve the connection between values and references

Hello all,

As you may have seen it in the development roadmap or discussed with us at an event, we are working on a project called "signed statements". During the research that we are currently running to understand better the various stakeholders' needs for this feature, it occurred to us that a first step should be tackled first: informing editors when there is a mismatch between a value and the reference attached to it. This is why we decided to start working on tainted references before moving forward with signed statements.

So, what is it going to look like? The first step is to inform the editor, when they change a value without changing the reference, that something may be wrong and that they should double-check if the value and the reference are still matching. The editor will have the option to change the reference, or to state that the value is still correct (for example, they could have been just fixing a typo). During this first phase, the notification will happen only for the editor who made the change. The second step is to make it appear to all users (a bit like the constraints) so other people can monitor the mismatched statements and help fixing them.

Our goal is to improve the data quality and the connection between values and references by making it clearer when a problem occurs, without adding too much extra work on the editors' shoulders.

The development will start next week, and we expect a first version to be tested in the upcoming weeks. As usual, you can follow the progress of the development tasks in the Weekly Summary. You can also have a look at the Phabricator board. If you have any question, feel free to reach me. Cheers, Lea Lacroix (WMDE) (talk) 12:32, 3 September 2019 (UTC)

I find it unfortunate that, the intention for creating this feature wasn't shared before it was coded up.
How does the feature deal with statements that have multiple qualifiers that are backed by different references? (Maybe the reference for when a marriage ended is not the same reference as the one for the marriage location) ChristianKl❫ 12:38, 3 September 2019 (UTC)
@ChristianKl: If you read Lea's message you would have noticed that "development will start next week", so it hasn't been "coded up" at all yet. Also this is something I've seen discussed here on wiki before so it shouldn't be much of a surprise; in fact there has been a phab ticket on at least part of this since 2014: - T76232 "nudge when editing a statement to check reference". ArthurPSmith (talk) 17:47, 3 September 2019 (UTC)
As a first step as long as any part of any reference on the statement was changed we consider the edit fine and don't show a warning. We might tweak that later based on your feedback but I'm not sure we can make it much smarter than that. But I'm open to ideas. --Lydia Pintscher (WMDE) (talk) 18:48, 3 September 2019 (UTC)
I already expressed my opinion and I know that few persons will agree with it, but the fact is that putting more constraint when modifying a statement will reduce vadalism and that will be critical if WD want to become a central hub for data. We can't ask people to share their data if we can't ensure the integrity of this data.
If we don't want to accept that the change of the value of a statement is followed by the deletion of the related reference, then we should at least limit this possibility to registered users. IPs should not be able to perform a change of a value without changing the reference. Clicking twice a button is not enough to limit vandalism. Snipre (talk) 22:58, 3 September 2019 (UTC)

What methods of adding data will this apply to? The GUI? Some or all of the various APIs? The ability to edit Wikidata through a Wikipedia that displays Wikidata article in an article?

A different point is the claim in a closed Phabricator thread that one could indicate that an existing reference supports a change by changing retrieved (P813). That only applies to online sources, not paper sources. Jc3s5h (talk) 22:46, 3 September 2019 (UTC)

It will be used on the GUI, but not on APIs at that point. Maybe later if people request it. As for the mw:Wikidata Bridge, it will include a similar feature to make sure that people see the existing references and update them if needed.
I believe that the use of "retrieved" was only mentioned as an example of what one could change in a reference. Lea Lacroix (WMDE) (talk) 09:39, 4 September 2019 (UTC)
A rather common use case is for an editor to use a paper source, such as Dictionary of National Biography (Q1210343), is used to add a birth or death date, but the original editor is completely oblivious to the existence of the Julian calendar (Q11184). So a later editor must change the date to correctly interpret the source. No "retrieved", because it's paper. No change in the information about the source. Jc3s5h (talk) 14:40, 4 September 2019 (UTC)
  • Thanks for doing this. --- Jura 00:15, 4 September 2019 (UTC)


see also (P1659) - I do not understand why must have a single value. There may be more properties that are helpful. For istance Commons creator shows 4 errors. Please let's discuss there: Property_talk:P1659#Single_value--Pierpao (talk) 06:18, 4 September 2019 (UTC)

Label for taxon

In the example Alkanna tinctoria (Q856964), should the preferred EN label be the taxon name or the common name? (In this case, the EN wiki article is titled with the taxon name.) I looked at the project page but I didn't find best practices for labels. Thanks! - PKM (talk) 20:21, 1 September 2019 (UTC)

Wikidata:WikiProject Taxonomy/Tutorial says the label should “usually” be the taxon name, so I’ll follow that. - PKM (talk) 21:25, 4 September 2019 (UTC)

Relation between an interchange and the roads it joins together

From what I've briefly seen it seems like there's a bit of a mess with Japanese road items.

Part of the problem is the incorrect use of of (P642) on interchange items. Is there a better property than connects with (P2789) to fix these? I feel like that property would rather describe the relation between the roads themselves (e.g. ROAD connects with ROAD via INTERCHANGE). --SilentSpike (talk) 12:02, 20 August 2019 (UTC)

I wanted to ask about that on for quite some time as well. For example on Fai Chai Intersection (Q16434047) I also used connects with (P2789), and adding the direction by a qualifier. Additionally for those roads ending at the intersection, added the intersection as terminus (P559). As other properties there's located on street (P669), which also not fit that well. There's also the question how to link streets when there is no named intersection - or where I have that problem, to connect the canals in Bangkok. Ahoerstemeier (talk) 12:42, 20 August 2019 (UTC)
  • (interchange) "part of" (some road)
and (interchange) "part of" (some other road) ? --- Jura 21:11, 20 August 2019 (UTC)
    • This was my initial thought, but I'm not sure if the interchange is really part of the road unless it's a major meets minor road situation. --SilentSpike (talk) 15:09, 22 August 2019 (UTC)
      • Maybe the road project participants have some suggestion. --- Jura 15:16, 22 August 2019 (UTC)
@B.O.B. jp: --Rschen7754 00:48, 27 August 2019 (UTC)
Better to Alexis900 (talkcontribslogs) Asqueladd (talkcontribslogs) BeneBot* (talkcontribslogs) Detcin (talkcontribslogs) Dough4872 (talkcontribslogs) Gz260 (talkcontribslogs) Happy5214 (talkcontribslogs) Imzadi1979 (talkcontribslogs) Jakec (talkcontribslogs) Labant (talkcontribslogs) Liuxinyu970226 (talkcontribslogs) Ljthefro (talkcontribslogs) mxn (talkcontribslogs) naveenpf (talkcontribslogs) Puclik1 (talkcontribslogs) Rschen7754 (talkcontribslogs) Scott5114 (talkcontribslogs) SounderBruce (talkcontribslogs) TCN7JM (talkcontribslogs) TimChen (talkcontribslogs) Bodhisattwa (talkcontribslogs) Daniel Mietchen (talkcontribslogs) Tris T7 TT mePictogram voting comment.svg Notified participants of WikiProject Roads plus (per WD:A) @Jianhui67, Nikki, Okkn, Penn Station, Sotiale:. --Liuxinyu970226 (talk) 04:21, 2 September 2019 (UTC)
My view would to be to use connects with (P2789) on highway items, with the intersecting highway as the value and the interchange as a qualifier (like SilentSpike suggests). I would not say an interchange is "part of" a road, and part of (P361) is already used for parts of multi-state highways (like Interstate 35 in Texas (Q2445457) and Interstate 35 (Q94150), though that particular example isn't tagged). In the US, the value of terminus (P559) is normally the highway(s) which the subject ends at, with the appropriate direction qualifier. terminus location (P609) might be useful, though this is usually used for the city/town or state/country (for routes ending at such a border) at the terminus. -happy5214 14:14, 5 September 2019 (UTC)


Q351651#P1472 shows a constraint (Commons Creator page). Why? Is namespace qualifier required?--Pierpao (talk) 11:26, 3 September 2019 (UTC)

Just ignore it, server issues. Sjoerd de Bruin (talk) 12:07, 3 September 2019 (UTC)
I wouldn't ignore, we should have it fixed... --Matěj Suchánek (talk) 17:17, 5 September 2019 (UTC)

Bad titles/labels created by SourceMD

It seems there's an unknown number of wrong titles and labels in new items of scientific articles created by SourceMD.

Item: A Novel Monoallelic Nonsense Mutation in the NFKB2 Gene Does Not Cause a Clinical Manifestation (Q64100401)


A Novel Monoallelic Nonsense Mutation in the Gene Does Not Cause a Clinical Manifestation

and the original paper [3]

A Novel Monoallelic Nonsense Mutation in the NFKB2 Gene Does Not Cause a Clinical Manifestation

the problem is that all (?) italicized texts in articles' titles are left out.

Kpjas (talk) 09:42, 4 September 2019 (UTC)

@Daniel Mietchen: since you created the linked item, can you look into this? ChristianKl❫ 10:13, 4 September 2019 (UTC)
I don't find problem in output of Crossref/EuropePMC API.--GZWDer (talk) 12:08, 4 September 2019 (UTC)

Adding text qualifier to Q17252?

Hi! I noticed there is a Tibetan flag used at Q17252. I added the date qualifiers (specifying when it was used by a particular government) but it would be nice to add a text qualifier too (stating that currently there is no official flag used to represent Tibet in any of the countries that control territory in the region, that the Central Tibetan Administration uses it as their flag, and that the PRC banned it in 1959)

This template is used in the Commons, but the question on whether/how to amend Wikidata material should go here. WhisperToMe (talk) 17:12, 4 September 2019 (UTC)

I think Tibet (Q2444884) is the item for the state that used it. Ghouston (talk) 05:11, 5 September 2019 (UTC)
qualified with "followed by"="no value". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:01, 5 September 2019 (UTC)

Change multiple labels at once?

Hello! Is there a way to change multiple labels at once? I know the script nameGuzzler.js, but it seems to allow adding new labels, not changing existing ones. Korg (talk) 10:23, 5 September 2019 (UTC)

  • QuickStatements. --- Jura 10:25, 5 September 2019 (UTC)
Thank you! Korg (talk) 11:09, 5 September 2019 (UTC)

Cameroon administrative regions

A professor here at Oxford has spent time in archives in Cameroon (Q1009) making a table of the country's administrative regions, as best as can be done with the surviving records. The results are in a paper and a data table which are both on open access. I don't foresee having the time to import this to Wikidata but I'm raising awareness here in case anyone else can work on it, given our currently sparse coverage of Cameroon. MartinPoulter (talk) 11:24, 5 September 2019 (UTC)

Hi, what's the use ? I believe department of Cameroon (Q2306813) are already on databse : Bouzinac (talk) 12:20, 5 September 2019 (UTC)

Knut Hamsun

Take a look at birth place for Q40826#P19. I'm sure this irks a lot of users! :D (Bad-bad-me, now you can't stop thinking about it!) Jeblad (talk) 15:02, 4 September 2019 (UTC)

Now someone has added Q35494082 as yet another location. It is a 230 km long valley, which contains (or actually goes through) the municipalities Q113438 and Q488395. Q35494082 should be merged with Q1292655. Jeblad (talk) 15:59, 4 September 2019 (UTC)
It's me. The nobel prize website reference claims Hamsun was born in Gudbrandsdalen :) (which is a very common toponym in Norway) :) Bouzinac (talk) 20:42, 5 September 2019 (UTC)

Nutrition Data

I noticed there is a ton of useful nutrition data on the sidebar of food products like Big Mac, but properties for those things do not seem to exist in Wikidata. Should properties be created for this kind of data? I think having that kind of data in Wikidata would be really useful, but I wasn't sure if this had already been disucssed before. U+1F360 (talk) 15:32, 30 August 2019 (UTC)

Pictogram voting comment.svg Comment @U+1F360: collects such data. --Marsupium (talk) 07:50, 6 September 2019 (UTC)
Pictogram voting comment.svg Comment I would really like to see nutritional info included, not sure how to structure it though, have been looking at beers like Windhoek Lager (Q2538549) Iwan.Aucamp (talk)

"references" bei "date of birth" and "date of death" - German speaker here?

Bei Personendaten erscheint in Wikidata neben dem Geburts- oder Todesdatum mit Referenz ein Wimpel. Klickt man auf den Wimpel erscheint: Suggestions "single best value constraint" (This property should contain a single “best” value. Of the current multiple values, one should be marked with “preferred” rank.) UND "citation needed constraint" (Statements for date of birth should have at least one reference.) - Die Folge: Das Datum erscheint doppelt in der Infobox und in den Commons wird weder die Kategorie Births noch Deaths erzeugt. (Beispiel: Category:Ingeborg Rönnblad) - Ist beim Klick auf den Wimpel neben dem Datum NUR "citation needed constraint", erscheint das Datum ordentlich und die "Kategorien 1876 births" und "1941 deaths" werden erzeugt. (Beispiel: Category:Lisa Ranft) - Wo kann ich "single best value constraint" finden und entfernen um ordentliche Daten zu erhalten? Adelfrank (talk) 18:51, 5 September 2019 (UTC)

@Adelfrank: Die "single best value constraint" sagt aus, dass bei mehreren Werten ein Wert den bevorzugten Rang erhalten sollte (siehe Help:Ranking/de). Bei Lisa Ranft (Q20652760) wurde das Datum 11. August 1876 mit dem bevorzugten Rang markiert (ersichtlich an dem Symbol links neben dem Wert - das nach oben zeigende Dreieck ist hervorgehoben), bei Ingeborg Rönnblad (Q4976323) wurde bisher noch kein Wert mit dem bevorzugten Rang markiert. Wie Ränge gesetzt werden, wird unter Help:Ranking/de#Ränge_anwenden beschrieben. Es gibt in Wikidata auch ein deutschsprachiges Forum: Wikidata:Forum. - Valentina.Anitnelav (talk) 09:17, 6 September 2019 (UTC)
Danke. - Ich verstehe es nicht und weiss nicht was ich machen kann/soll. Wikidata ist ein Buch mit 7 Siegeln für mich. Wenn beide Werte stehen, ist das Ergebnis fehlerhaft (s. oben). Um nur "citation needed constraint" zu erhalten müsste ich irgendwas machen, nur was ist mir nicht klar. Wie Ränge gesetzt werden, ist toll erklärt: Um einer Aussage einen Rang zu verleihen, gehe zu einer Aussage und klicke auf [bearbeiten], - ich weiss weder ob ich auf edit klicken soll oder auf den Wimpel und [bearbeiten] gibt es schon mal garnicht, da ALLES english ist. Unter edit gibt es noch add qualifier, dazu nur eine einzige Vorgabe: sourcing circumstances. Klicke ich das an, muss irgendwas eingegeben werden, was wissen die Götter. Wenn man bei Wikidata kein perfektes english beherrscht, wird man echt kalt gestellt. Mit meinem Schulenglisch sehe ich hier weit und breit kein Land in Sicht. Trotzdem Danke für den Tipp u.a. dass es ein dt. Forum gibt, viell. versuche ich da mal mein Glück, Licht ins Dunkel zu bekommen. Schönes WE wünscht Adelfrank (talk) 09:55, 6 September 2019 (UTC)
@Adelfrank: Die Oberflächensprache kann auf Deutsch gesetzt werden: ganz oben im Menü, links neben dem Nutzernamen, auf "English" klicken und dann "Deutsch" als Sprache auswählen. - Valentina.Anitnelav (talk) 10:07, 6 September 2019 (UTC)
@Adelfrank: Das deutsche Forum ist übrigens Wikidata:Forum. Ahoerstemeier (talk) 10:10, 6 September 2019 (UTC)

Danke Valentina, wieder was Neues für mich. Auch Dir Ahoerstemeier, hat Valentina mir auch schon oben erklärt. Ich werde das Ganze mal dorthin übertragen, dennn mein Problem ist für mich weiterhin ungeklärt. Adelfrank (talk) 10:16, 6 September 2019 (UTC)

Jetzt habe ich endlich gefunden, wo ich draufklicken muss.
✓ Done Adelfrank (talk) 11:10, 6 September 2019 (UTC)

WorldCat Identities for people lacking other identifiers

There seems to be a small but significant number of authors/subjects who are indexed on WorldCat Identities, but apparently not indexed by any other contributing database (e.g. VIAF, Library of Congress, ISNI, etc.). A current example is Ernestine Hara Kettler (Q66712318), at Kettler, Ernestine Hara 1896-. I think WorldCat is one of the most useful databases for biographies, in that it shows not only works by the subject but works about a subject (crucial in finding references and ascertaining identity) and provides a very helpful directory of libraries and other institutions to actually find the works (for old codgers like me who actually use offline print sources). WorldCat Identity pages are normally accessed through Wikimedia Authority control templates via the VIAF or LCCN ID. Is there a way to link such "non-indexed" people in Wikidata? And if not, why not make one? -Animalparty (talk) 20:37, 6 September 2019‎ (UTC)

I have been adding them as "described by url", but I am not opposed to any other creative solution, including a field synthesized from other field information to create the Worldcat url for all entries. --RAN (talk) 22:42, 6 September 2019 (UTC)

Undo at diff pages

Have anyone noticed that undo link is currently missing from diff pages? (Or is it only me?) Rollback and restore are still there, but not the undo. You can still use it from the page history, but then it's impossible to undo a series of edits that are not the latest (unless you undo each edit individually). — Mike Novikoff 20:56, 10 September 2019 (UTC)

I may add that it doesn't seem to be the case in this ("Wikidata:") namespace, but surely does in the main and "Template:" namespaces (at least). — Mike Novikoff 21:40, 10 September 2019 (UTC)
Yeah, I noticed the missing undo link, too, and was wondering why it was changed. Reverting vandalism has become a bit harder now. --Kam Solusar (talk) 22:50, 10 September 2019 (UTC)
In my case, I don't see it on the property namespace. It has been reported to the development team. Esteban16 (talk) 23:57, 10 September 2019 (UTC)
@Mike Novikoff, Kam Solusar, Esteban16: that is T232214 (affecting all entity namespaces – Item, Property, Lexeme) and should be fixed later today. (In the meantime, you can still find undo links in the history.) --Lucas Werkmeister (WMDE) (talk) 09:07, 11 September 2019 (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) 17:25, 12 September 2019 (UTC)

Use of scholarly article (Q13442814)

scholarly article (Q13442814) has the description "article in an academic publication, usually peer reviewed", and is a subclass of scientific publication (Q591041), article (Q191067) and scholarly work (Q55915575) but many instances are letters or news articles. There is also academic journal article (Q18918145), which is a subclass of Q13442814 but doesn't mention peer review, and is used similarly. How should these be used? It looks like what is needed is an item "article in an academic publication" that is only a subclass of article (Q191067), of which the others are a subclass and which can be used until the type of article is determined. Peter James (talk) 20:15, 5 September 2019 (UTC)

Any examples of scholarly article (Q13442814) being used on news articles. --Trade (talk) 23:31, 5 September 2019 (UTC)
It tends to be used for anything taken from an academic publication, even if it's in a news, letters, or book reviews section, e.g., Why most published research findings are false: author's reply to Goodman and Greenland (Q21563418). Ghouston (talk) 23:53, 5 September 2019 (UTC)
  • Yeah, it's somewhat suboptimal. As it's used for any type of article in such publications, article(Q191067) would probably work better. --- Jura 08:48, 6 September 2019 (UTC)
Addedd it at User:Research_Bot/issues#use_of_scholarly_article_(Q13442814)_in_P31. --- Jura 15:45, 8 September 2019 (UTC)

How to create a new property

Hey everyone,

how can i make a new property? You know, i work on Muslims notables biographies... a Property is needed for "madhhab". --Ruwaym (talk) 08:40, 7 September 2019 (UTC)
@Ruwaym: Try WD:PP. --Liuxinyu970226 (talk) 09:29, 7 September 2019 (UTC)
@Liuxinyu970226: Property proposal/madhhab, I hope not seem like trolling! --Ruwaym (talk) 10:58, 7 September 2019 (UTC)
@Ruwaym: Doesn't movement (P135) cover this? The property proposal is currently very incomplete and will fail in this form. Multichill (talk) 09:43, 8 September 2019 (UTC)
@Multichill: I guess not. Salafism, Sufism (branches of them like Malamatiyya) , Islamic modernism etc are movements, But w:madhhabs are few schools backs to early Islam. If there is no active user who can help me about that property, you tell me what is needed, here or on my talk page. --Ruwaym (talk) 10:45, 8 September 2019 (UTC)

Item update request

Please add “Amen” as last words on John Paul II (Q989). Thanks! -- 19:41, 14 September 2019 (UTC)

Any source? Nomen ad hoc (talk) 19:47, 14 September 2019 (UTC). -- 19:56, 14 September 2019 (UTC)
✓ Done. Nomen ad hoc (talk) 20:00, 14 September 2019 (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. Nomen ad hoc (talk) 20:00, 14 September 2019 (UTC)

New API parameter to add tags on edits summaries

Hello all,

As you know, it is possible to add tags in the edit summaries (eg "mobile edit"), especially for edits made by tools (eg "openrefine"). These tags are used to more easily find specific edits and monitor the entering data, for example on the Recent Changes where one can filter all edits made via the VisualEditor, or all edits that remove a sitelink.

So far, tool builders had to use hacky ways to make their tools automatically add the tag to the edit summaries. In order to improve this process, we started adding a tags parameter to a bunch of APIs, to allow the developers to integrate the tags directly in their API actions. Here's the list of impacted APIs:

  • wbeditentity
  • wbsetaliases
  • wbsetdescription
  • wbsetlabel
  • wbsetsitelink
  • wbcreateclaim
  • wbremoveclaims
  • wbremovequalifiers
  • wbremovereferences
  • wbsetclaim
  • wbsetclaimvalue
  • wbsetqualifier
  • wbsetreference

This change has already been deployed, so you can start using the new tags parameter right away.

You can find more information about tags and how to create some new ones on mw:Help:Tags (this page would probably need some love, feel free to improve it). If you encounter any issue, feel free to let a comment in the related ticket.

@Magnus Manske, Pintoch, Pasleim: I guess you're some of the people that can be interested in this change :)

Cheers, Lea Lacroix (WMDE) (talk) 10:33, 28 August 2019 (UTC)

@Lea Lacroix (WMDE): yes I saw that, it is great! Many thanks. We should add support for the feature in Wikidata-Toolkit (Thank you Lucas Werkmeister (WMDE) too!) − Pintoch (talk) 10:45, 28 August 2019 (UTC)
Changed my entire OAuth tag sytem to make use of that. Now I can't use WDFIST bedause E1:The tag "wdfist" is not allowed to be manually applied. Changing everything back now. WTF? --Magnus Manske (talk) 08:34, 2 September 2019 (UTC)
OAuth tagging has been working fine and is secure, why to change it? --Matěj Suchánek (talk) 08:49, 2 September 2019 (UTC)
@Magnus Manske: tags can be applied via the API only if they have been created for that specific purpose. OAuth tags cannot be applied manually, they are automatically added if the action was performed using OAuth authentication. So for OAuth applications there is generally nothing to change (tags continue to be added automatically). This change is useful for other applications which cannot use OAuth (such as OpenRefine, which cannot embed OAuth secrets in its source code). Or if you use the same generic OAuth provider for multiple applications (such as Widar), we could create tags for each of these applications and you could apply these tags via the API (so the edits would be tagged by the generic OAuth tag and the specific application tag). If you want such tags to be created let me know. − Pintoch (talk) 12:58, 9 September 2019 (UTC)

New termbox making it hard to copy and paste QIDs on mobile?

I just wanted to check first if it's me or not: I think that since recently it became rather hard to select the QID when on a mobile device, in order to copy and paste the QID. My guess it has to do with the new termbox (which is pretty awesome, by the way). Basically, I can't tap or doubletap on the QID. It works with labels and other content, but not the QID. I am running Chrome on a Pixel 3a. Is this known or should a bug be filed to fix this? --Denny (talk) 18:33, 4 September 2019 (UTC)

Not known. A ticket would be <3. --Lydia Pintscher (WMDE) (talk) 18:40, 5 September 2019 (UTC)

Thanks for checking, filed T232356.. --Denny (talk) 15:49, 9 September 2019 (UTC)

Can't enter labels in Montenegrin

It looks like our site can't support labels in Montenegrin (ISO 639-3 cnr), when entering, it only warns me that The language code "cnr" is unknown. Please use a language code known to the system, such as "en". -- 08:34, 6 September 2019 (UTC)

  • It works only for monolingual strings, e.g. at Q7700307#P1705. --- Jura 17:38, 6 September 2019 (UTC)
    • Label supports are only happened if and when your language is supported by MediaWiki software itself, which as I've checked, non of the core-0-mostused group messages are translated, so far not yet.
Notifying most active editors of Incubator test Wikipedia: @ExplodingPoPUps, Bellatrix10, Zanović, Srdjan m, Ookuninusi:@Je7roi, Romaine, Incnis Mrsi, Markus cg1, Orbwiki107:@StevenJ81, Rovoobob, Ivarhcp4ever, Ooswesthoesbes, KOMITA:@Nesmir Kudilovic, Lujki, WescoM, Ego and his own, Acamicamacaraca2:@Katxis, Lontschar, Porfirius Maximus, MarkoM5, Crnogorac98:@Lojbanist, Nikola Krivokapić, Pop Milo Jovovic, MontenegrinSportEditor, AT44:@Adrian Hernandez, Akhemen, Krassotkin, Gogash86:@Vlad5250, НиколаБ, DannyS712, KHMELNYTSKYIA, DaveZ123:@Green Giant, PlyrStar93, Silverije, Eniisi Lisika, Чръный человек:, anyone I pinged should have time to translate those messages, if you want to see and edit labels in your language. --Liuxinyu970226 (talk) 01:24, 7 September 2019 (UTC)
@Amire80, I thought you were going to activate this, at least for this type of purpose. StevenJ81 (talk) 14:06, 9 September 2019 (UTC)
I did activate it for translatewiki, with crn-latn and crn-cyrl.
Wikidata has labels in British English, and in quite a lot of cases they are definitely useful, such as autumn (Q1314). So there can be labels in Montenegrin, too. --Amir E. Aharoni (talk) 16:01, 9 September 2019 (UTC)

Wikidata weekly summary #381

Norwegian Wikipedia

Norwegian Wikipedia (Q191769)

Wikipedia in Bokmål and Riksmål (Q32176383)

w:en:Norwegian Wikipedia

Now I have a headache. Alexis Jazz please ping me if you reply 22:31, 22 August 2019 (UTC)

@Alexis Jazz: This looks too hard to resolve unless if one day this Meta-Wiki RFC can be proceed, more informations may be provided by langcom members: @Amire80, N-true, Antony D. Green, Maor X, GerardM:@Jon Harald Søby, Klbroome, MF-Warburg, Evertype, Millosh:@SPQRobin, santhosh.thottingal, Satdeep Gill, StevenJ81, Doc James:. --Liuxinyu970226 (talk) 22:45, 22 August 2019 (UTC)
Also the nominator of that RFC @Jeblad:. --Liuxinyu970226 (talk) 22:48, 22 August 2019 (UTC)
@Liuxinyu970226: Brilliant. I don't like it. Wikidata:Requests for deletions#Q32176383. Alexis Jazz please ping me if you reply 23:06, 22 August 2019 (UTC)
  • Doesn't Norway have two versions of the language, Bokmål and Nynorsk? --RAN (talk) 03:22, 23 August 2019 (UTC)
It's not very complicated. There used to by Norwegian Wikipedia (Q191769). Then in 2004 Norwegian (Nynorsk) Wikipedia (Q2349453) split off, and a short while later Norwegian Wikipedia (Q191769) became Wikipedia in Bokmål and Riksmål (Q32176383) after a vote. It kept (and still keeps) the URL, for partly historical, partly political and partly technical reasons, but it is not the same thing. (The problem in relation to Wikidata is that a lot of the connected Wikipedia articles cover Wikipedia in Bokmål and Riksmål (Q32176383) and Norwegian (Nynorsk) Wikipedia (Q2349453) in the same article, but that's just another variation of the Bonnie & Clyde problem.) Jon Harald Søby (talk) 06:11, 23 August 2019 (UTC)
@Jon Harald Søby: And why exactly wouldn't it be the same thing? Alexis Jazz please ping me if you reply 06:27, 23 August 2019 (UTC)
@Alexis Jazz: Are Serbia and Montenegro (Q37024) and Serbia (Q403) the same thing? Jon Harald Søby (talk) 07:18, 23 August 2019 (UTC)
@Jon Harald Søby: You didn't answer my question. Alexis Jazz please ping me if you reply 08:45, 23 August 2019 (UTC)
@Alexis Jazz: Norwegian Wikipedia before the split was the Wikipedia of both Nynorsk and Bokmal. It's now just Bokmal (/Riksimal) Wikipedia. Thus, there are two entities and the analogy which Jon has given to you is a valid one. Serbia is the successor state of FR Yugoslavia / Serbia and Montenegro, but Serbia is not FR Yugoslavia / Serbia and Montenegro. --Millosh (talk)
@Millosh: By that analogy, the Dutch Wikipedia ceased to exist when fywiki was launched. Except that's not true either. Plus, according to Wikipedia, nowiki was already de facto written in Bokmål/Riksmål. And even so, a change in scope does not mean the project was killed off. Alexis Jazz please ping me if you reply 11:02, 25 August 2019 (UTC)
@Alexis Jazz: In both cases, it's about different things, no matter what the political decision was. If the idea is to structure WikiData entities logically, then they should be treated differently. If not, they shouldn't be. Also, I really don't want to spend any minute more on this issue, which implies that I don't care what will be done. --Millosh (talk) 12:19, 25 August 2019 (UTC)
@Millosh: I don't object to adding start and end dates to the language of the work. In fact, I just did. Still not a new project. Alexis Jazz please ping me if you reply 13:22, 25 August 2019 (UTC)
In my opinion this is one continuous project, and it should have one page at nowiki and one item here. Still, unless the articles at nowiki are merged the items can't be merged. Also note that the RFC has nothing to do with the current articles at nowiki, and should not be used as argument for any specific actions on neither articles nor items. Jeblad (talk) 10:33, 23 August 2019 (UTC)
@Jeblad: What do you mean, "one continuous project"? StevenJ81 (talk) 12:47, 23 August 2019 (UTC)
The articles w:no:Wikipedia på norsk (Q191769) and w:no:Wikipedia på bokmål og riksmål (Q191769) both refer to the same project, and I believe the later page should be a redirect to the first as this is still the correct name given the “no” prefix. That prefix is wrong, but changing that is another discussion that can not (and should not) be decided here. Jeblad (talk) 13:09, 23 August 2019 (UTC)
Bonnie & Clyde solution? (and I don't mean a crime spree). - Jmabel (talk)
Didn't know bokmål was different from riksmål (besides nynorsk) Bouzinac (talk) 06:33, 24 August 2019 (UTC)
"riksmål" isn't a language, it's a sociolect for those that think everything was better when Norway was under Danish rule. ;-) Moebeus (talk) 13:46, 24 August 2019 (UTC)
I had been wondering why langcom considered it inappropriate to identify it with the accurate language code. --- Jura 13:49, 24 August 2019 (UTC)
@Jura1: By checking IANA's language subtag list:
Type: language
Subtag: nb
Description: Norwegian Bokmål
Added: 2005-10-16
Suppress-Script: Latn
Macrolanguage: no
Type: grandfathered
Tag: no-bok
Description: Norwegian Bokmal
Added: 1995-08-23
Deprecated: 2000-02-18
Preferred-Value: nb
Type: language
Subtag: nn
Description: Norwegian Nynorsk
Added: 2005-10-16
Suppress-Script: Latn
Macrolanguage: no
Type: grandfathered
Tag: no-nyn
Description: Norwegian Nynorsk
Added: 1995-08-23
Deprecated: 2000-02-18
Preferred-Value: nn
Type: variant
Subtag: hognorsk
Description: Norwegian in Høgnorsk (High Norwegian) orthography
Added: 2010-01-02
Prefix: nn
Comments: Norwegian following Ivar Aasen's orthographical principles,
  including modern usage.

Very interesting, why those who say their orthography as riksmål don't request such a tag? --Liuxinyu970226 (talk) 15:56, 30 August 2019 (UTC)

A strange thing is that Norwegian officially has (at least) two ways to write the language, but only one way to speak it. But the diversity in how people speak the language is actually much larger than the ways they write it. I cannot easily see if a text is written in nynorsk or bokmål. But there are many forms of Norwegian I do not understand. Not only in the southwest, where they often write in nynorsk, but also in the north where they write in bokmål. IP (Hej!) 07:17, 26 August 2019 (UTC)
Support merging, there are just two Norwegian Wikipedias, No reason to have 3rd. -- 06:06, 29 August 2019 (UTC)
  • I wonder why people would think that changing a valid P31 value to Wikipedia:Stub (Q4663261) could be acceptable. --- Jura 13:35, 30 August 2019 (UTC)
@Jura1: I was about to say, I just reworked both items. And why, well, because the value was actually invalid? It's not a Wikipedia language edition, that's Norwegian Wikipedia (Q191769). This is merely a short description of a part of its history. I was going to look up the best value for that sort of thing, but can't think of anything similar right now. Alexis Jazz please ping me if you reply 13:39, 30 August 2019 (UTC)
I don't think your changes are helpful. Wikipedia:Stub (Q4663261) is probably helpful for you on Wikipedia, but it's not really valid around here. --- Jura 13:44, 30 August 2019 (UTC)
@Jura1: Not worth arguing over, I've removed the claim. Alexis Jazz please ping me if you reply 13:47, 30 August 2019 (UTC)
Can you revert your stuff. Deleting valid claims is considered vandalism around here. Wikipedia might refactor articles to the current view of the world, but Wikidata doesn't work that way. --- Jura 13:51, 30 August 2019 (UTC)
@Jura1: No, those claims were invalid. And if you falsely accuse my good faith edits of being vandalism once more, I'll be reporting you. Alexis Jazz please ping me if you reply 14:01, 30 August 2019 (UTC)
  • By reading two nowiki articles, I would say that I also vote  Support merging both, there are really nothing about histories that we must need to have separated two, and nearly Ctrl+C and Ctrl+V "synced" articles to describe, just use one article to describe domain usage history is already Okay. --Liuxinyu970226 (talk) 15:46, 30 August 2019 (UTC)
    @Jmabel: To me, saying this as Bonnie and Clyde case will only make troubles for the future created Wikipedias, they will have to re-judge if they want a link to be fall under one item, and so makes them more complex troubles when maintaining interwiki links. --Liuxinyu970226 (talk) 15:49, 30 August 2019 (UTC)
    • Why? Everyone but no-wiki uses the "Bonnie and Clyde" item, no-wiki uses the "Bonnie" item and the "Clyde" item. Presumably it would be pretty obvious to any future Wikipedia that they go the "normal" way, not the way used by one wiki. - Jmabel (talk) 16:39, 30 August 2019 (UTC)
      • There are essentially two ways of doing this. None of them is today's approach with a mostly blank item or one that is an instance of "Wikipedia stub". This might work at Wikipedia, but not here. Also it's odd that has no mention at all of Nynorsk (Q25164). Why is this omitted? --- Jura 17:33, 30 August 2019 (UTC)
    @Jmabel, Jura1: That said again and again, I've read both nowiki articles, and I now believe that their contents are Ctrl+C and Ctrl+V copied, not enough for merging suggestion? --Liuxinyu970226 (talk) 22:41, 30 August 2019 (UTC)
  • There was a project for all language variants at nowiki. At one point a new project nnwiki emerged for the Nynorsk language variant. For a short time Nynorsk was allowed at both projects. Somewhat later users at nowiki redefined the project to be “Bokmål and Riksmål”, yet kept on using the “no” macro code. There are now one article about the previous language situation, and an article about the current language situation. As a project nowiki is one continuous project, but when it comes to use of specific language variants Nynorsk is discontinued. There are two articles, and given that Wikidata is supposed to support the actual articles in Wikipedia further discussion about whether there should be two items are moot. Further discussions about merging the articles should be left as an issue for the community at nowiki. What can (and should) be discussed here is how the articles should be described. Such articles isn't the real thing, it is a description of the real thing. The article Norwegian Wikipedia (Q191769) is about the real thing (ie the project), given the use of the “no” language code, and Norwegian Wikipedia (Q191769) is an article about the language history of the real thing.
    Note that a discussion about use of “no” as a language prefix for nowiki does not belong on Wikidata, it should be resolved at Meta. I would prefer use of “nb” as a language prefix, but that discussion does not belong here.
    There are really four language variants in Norway. Two are official, “Nynorsk” and “Bokmål”. “Riksmål” is semi-official, and somewhat backed as “Moderat Bokmål”. The form Moderat Bokmål is between Bokmål and Riksmål. Both Bokmål and Riksmål is allowed at nowiki. “Høgnorsk” is unofficial, but allowed at nnwiki. Høgnorsk is an older form of Nynorsk. These are all written form of Norwegian, as spoken Norwegian mostly follow local dialects or sociolects. Jeblad (talk) 14:37, 1 September 2019 (UTC)
    • If what nowiki article they are connected to is correct, then Q32176383 is about, while Q191769 is about both and Their relationship to each other is kind of the same as between Q31/Q55 and Q476033, respectively, common history and still very related, but currently distinct. However, since almost all other wikis link to Q191769, it does seem likely that the nowiki articles should swap their Qs (either that, or everyone else). It does also seem strange that both present themselves as a Wikipedia project. There has been only one It just reduced its (sub-)language scope at one point. (I certainly didn't feel like moving from one project to the other back then.) Ters (talk) 06:07, 2 September 2019 (UTC)
      • @Ters: Yes, the question is when something becomes something new. When an organization changes its name but nothing else, does it become a new company? If it changes ownership, does it become something new? If it fires half of its employees? Did Wikimedia Commons cease to exist when it stopped accepted GFDL-licensed photos and should the current Wikimedia Commons be treated like a new thing? I'm not quite sure myself where to draw the line. If nowiki has purged or translated half of its articles after the nnwiki split (I have no idea, but I suspect not), that would help to make a case for the current nowiki being something new. Right now, I haven't even been able to pinpoint an actual date when the vote to restrict the wiki to Bokmål/Riksmål became policy. In fact, did it ever? Alexis Jazz please ping me if you reply 08:08, 2 September 2019 (UTC)
        • Seems to have been decided by vote on 3 april 2005 [4]. There was discussion about a re-vote, but 20 april 2005 it was decided not to go forward with it [5], which I think validates the original vote. - Soulkeeper (talk) 08:23, 2 September 2019 (UTC)
          • Also, the Norwegian wikipedias don't run on official policy so much as tradition, best current practice, and bickering. The closest thing to an official decision that we usually get, is when everyone is completely fed up with discussing the topic, and just want to go on with their lives. Whatever seems to be the closest thing to consensus at the time, is what usually ends up being the BCP going forward, unless or until someone starts making a stink about it again. - Soulkeeper (talk) 08:34, 2 September 2019 (UTC)
            • According to enwiki, nowiki was already de facto Bokmål/Riksmål. If the outcome of the vote was never ratified, it's essentially still only de facto Bokmål/Riksmål. Alexis Jazz please ping me if you reply 15:30, 2 September 2019 (UTC)
              • ¯\_(ツ)_/¯ That's how we roll. - Soulkeeper (talk) 16:28, 2 September 2019 (UTC)
      • @ters: Article (and item) Norwegian Wikipedia (Q191769) is only about nowiki, nnwiki did not exist. It had although Norwegian Nynorsk. he nnwiki project was never a part of nowiki, nnwiki was and is an independent project of nowiki. Jeblad (talk) 21:24, 9 September 2019 (UTC)
        • That is not immediately obvious. The title is ambiguous, and about half the history section is about after nnwiki split out, including what must be considered a joint award. The next section is just as valid post-split, jointly or separately, as before the split. The section about the domain names for Norwegian Wikipedia mentions as well. The paragraph about is shorter, but that is because it came into being too late to have a domain. The only section that excludes nnwiki is the section for statistics, but that section blurs the line between the nowiki that was, which contained nynorsk, and the nowiki that is. The first sentence clearly is about the nowiki that was, but is the first sentence in an article always more correct than the rest? As it is, its meaning is outnumbered. Furthermore, other wikis disagree on what the item is about. If anything, more items are needed. Ters (talk) 05:44, 10 September 2019 (UTC)
  • Is w:Wikipedia:No original research also policy on Wikidata? Alexis Jazz please ping me if you reply 08:10, 2 September 2019 (UTC)
    I do not know, but it does not matter! As long as at least one project has two different articles Wikidata should have two items. A merge of the article is not our thing to decide here, it is up to the local project. IP (Hej!) 08:46, 2 September 2019 (UTC)
Actually it does. Even if Wikidata is forced to have items for articles, the statements on those items should be verifiable. Alexis Jazz please ping me if you reply 15:30, 2 September 2019 (UTC)
@Alexis Jazz: Why? We allow linking to Wikiversity which is known as an original research site, so why this wiki can't allow? --Liuxinyu970226 (talk) 21:47, 2 September 2019 (UTC)
@Liuxinyu970226: I'm not really familiar with Wikiversity, but two things: regardless of having an item or not, is Wikidata forced to import statements that can't be verified? And second, wikiversity:Wikiversity:Original research suggests that original research still should be verifiable or at least be falsifiable. Whether Norwegian Wikipedia ceased to exist in 2005 appears to be a matter of nothing but opinion over definitions. You can't prove it happened and you can't prove it didn't happen. I don't think Wikiversity would allow that. Being unable to prove either way, I believe the burden of proof is on those who say it did happen. For comparison: if someone claims God exists, the burden of proof is on them. Not on everyone else to prove God doesn't exist. "God doesn't exist" is falsifiable (easy: show him), while "God exists" is not falsifiable. (you can't check every atom in every universie in every dimension) - Alexis Jazz please ping me if you reply 09:59, 5 September 2019 (UTC)
The situation is quite simple; there are two articles and it can't be resolved here. Jeblad (talk) 21:27, 9 September 2019 (UTC)

Odd Commons category sitelink

Abandoned Barge Act of 1992 (Q28405208) is sitelinked to commons:Category:Barges in the United States, which is not altogether unrelated but seems wrong. It looks like the connection was made by EdgarsBot (talkcontribslogs) and turned into a sitelink by Pi bot (talkcontribslogs). Normally I'd just drop the bad sitelink, but I suspect that there is some other item to which the Commons category should actually link. - Jmabel (talk) 00:13, 6 September 2019 (UTC)

There's Category:Barges of the United States (Q8289889), but that's linked correctly to another Commons category. I can't see any item for "Category:Barges in the United States", so it should probably just be unlinked. Ghouston (talk) 00:27, 6 September 2019 (UTC)
It traces back to the Commons category link in en:Abandoned Barge Act of 1992, added by @Jphill19:. I've removed it [6] [7], and created Category:Barges in the United States (Q67198955). Thanks. Mike Peel (talk) 17:36, 8 September 2019 (UTC)
In this particular case, I completely agree with that removal, but there are definitely times when it makes sense for the See also section of a Wikipedia article to link a Commons category that would not be an appropriate sitelink for the article. If the mere presence of such a link is taken to imply that there necessarily should be such a sitelink, we have a process problem. All the more so if that is done by a bot. - 04:22, 9 September 2019 (UTC)
I have yet to find a good example of that situation. In most cases, the Commons link isn't actually useful or relevant (such as this case), or it indicates that there's an issue either on Wikipedia or Commons that needs to be resolved (e.g., by creating a new commons category). Thanks. Mike Peel (talk) 05:46, 9 September 2019 (UTC)
I don't think Wikidata is in a position to tell another project that they need to create a category or article. - Jmabel (talk) 15:11, 9 September 2019 (UTC)
Wikidata content is quite often used to say "there's something missing here, perhaps fix that?" - for example, en:Wikipedia:WikiProject Women in Red/Redlist index. Thanks. Mike Peel (talk) 06:26, 10 September 2019 (UTC)

New WikiProject - Textile arts

We’ve kicked off a new project - Wikidata:WikiProject Textile arts. Please join us if you’d like to help improve coverage of sewing, weaving, knitting, lacemaking, needlework, and other related crafts. - PKM (talk) 03:18, 10 September 2019 (UTC)

Connection of fictional and real locations

@Loominade, Shisma, Nvrandow: (see Topic:Uxejvxhjztld09m7 and User talk:Loominade#Q59818657 for the beginning of that discussion). Apparently some people do not see any problems mixing fictional and real locations in one "hierarchical" statement like:


I have no opinion about the last one, but the first two definitely will make (at least) my life more complicated. parent astronomical body (P397) is used in infoboxes in wikipedia for showing list of exoplanets, and it was never implied that one should filter out items with instance of (P31) fictional planet (Q2775969) (and even that is not enough, because Andor (Q66724357) parent astronomical body (P397) Procyon (Q13034)). My understanding is that hierarchy should be optimized for simplifying most frequent usages, rather than exotic ones. I'd suggest to use parent astronomical body (P397) to connect either functional object to fictional object or real object to real object. Connection between fictional object and real prototype can be expressed via fictional analog of (P1074) Ghuron (talk) 14:09, 6 September 2019 (UTC)

Valentina.Anitnelav Thierry Caro Shisma (talk) Arlo Barnes (talk) Tsaorin (talk) 16:37, 12 November 2019 (UTC) Nomen ad hoc SilentSpike (talk) 10:40, 9 May 2021 (UTC) Pictogram voting comment.svg Notified participants of WikiProject Narration & TomT0m Valentina.Anitnelav Shisma (talk) ElanHR Arlo Barnes (talk) Maria zaos (talk) EEMIV (talk) Pictogram voting comment.svg Notified participants of WikiProject Fictional universes--Loominade (talk) 15:02, 6 September 2019 (UTC)
related discussion--Loominade (talk) 15:06, 6 September 2019 (UTC)
Strictly speaing, I guess a proper qualifier might be useful. I think, that qualifier would be nature of statement (P5102), though I'm not clear what the target would be.
BUT it's worth remembering that every single statement about a fictional item that is not in relation to the fictional nature itself (such as present in work (P1441) or first appearance (P4584)) is in the exact same situation, so I don't see why fictional locations ought to be somehow singled out any more than Harry Potter (Q3244512) being a British resident... Circeus (talk) 17:27, 6 September 2019 (UTC)
@Circeus: The difference between Axanar (Q57032278) parent astronomical body (P397) Epsilon Eridani (Q194394) and Harry Potter (Q3244512) country of citizenship (P27) United Kingdom (Q145) is that normally in wikidata we do not produce a list of people who has both lion and unicorn on their passport (it will be both incomplete and meaninglessly huge). Building a list of exoplanet for a specific star is certainly more practical query, in theory every wikipedia article about extrasolar system would benefit from one. Also in sci-fi, portrait of star that hosts fictional planet sometimes differs from what we know about that star now. For instance in Andromeda (Q513228) star Epsilon Tucanae (Q2980984) is described as "located in 90pc from solar system and radiates 78L⊙" Ghuron (talk) 06:28, 7 September 2019 (UTC)
@Ghuron: for your hypothetical query for exoplanets you should make sure that your child object actually is instance of (P31) extrasolar planet (Q44559) for reasons other than it might be fictional. it could also be a exocomet (Q2855300), a circumstellar disk (Q3235978), or even a spacecraft (Q40218) (at some point in the future 😬). all these are yet allowed by the constrains of parent astronomical body and you don't want them in your list.--Shisma (talk) 07:31, 7 September 2019 (UTC)
@Shisma: Not all exoplanets has instance of (P31) extrasolar planet (Q44559) (see MOA-2007-BLG-192Lb (Q860312) as example), so I have to use wdt:P31/wdt:P279* wd:Q44559 And as long as I'm in sparql - that's fine, but when I'm doing infobox on lua, I'm much more performance constrained (expensive function calls, limitations on number of loaded wikibase objects, etc). When spacecraft (Q40218) will have extrasolar parent astronomical body (P397), wikidata will unlikely be relevant to anyone except few AI, specialized in history. And I still don't understand what is wrong with creating fictional star that will have fictional analog of (P1074) like it was proposed in previous discussions. --Ghuron (talk) 07:46, 7 September 2019 (UTC)
So just filtering out all items with instance of (P31)/subclass of (P279)* fictional entity (Q14897293) would be too expensive? The problem with creating new items for fictional analogues for other statements than P31 and P279 is that there would be the need to duplicate another huge portion of wikidata (besides classes also individuals). There is potentially the need to duplicate any place/administrative territory. - Valentina.Anitnelav (talk) 08:36, 7 September 2019 (UTC)
@Valentina.Anitnelav: the asterisk after P279 is the problem, I don't know how many items should I fetch in order to prove (or disprove) statement. I was thinking about that stuff for some time, and it looks like I need to define issue more clearly. In fact it is quite unlikely that statements like "fictional item->relation->real item" should cause a lot of problems for infoboxes. For instance, when I'm building a list of exoplanets I'd rather use child astronomical body (P398) not parent astronomical body (P397). So the real problem is how to distinguish Epsilon Eridani (Q194394) child astronomical body (P398) Axanar (Q57032278) from Epsilon Eridani (Q194394) child astronomical body (P398) Epsilon Eridani b (Q867895) And here the idea of Circeus might work. I'm not sure about nature of statement (P5102), but if we can put from narrative universe (P1080) Star Trek multiverse (Q18043309) as a qualifier for the first statement - it would probably solve majority of issues. Ghuron (talk) 13:04, 8 September 2019 (UTC)

So the real problem is how to distinguish Epsilon Eridani (Q194394) child astronomical body (P398) Axanar (Q57032278) from Epsilon Eridani (Q194394) child astronomical body (P398) Epsilon Eridani b (Q867895)

correct me if i'm wrong but that can be solved by setting the value to depreciated so it won't be returned in a query. isn't that what the rank is meant for? --Shisma (talk) 15:26, 8 September 2019 (UTC)
@Shisma: rank would definitely work for me, but I always thought that deprecated rank is used to indicate incorrect (or no longer correct) statements (e.g. there are 3 measurements of parallax (P2214) for Epsilon Eridani (Q194394), but the one that comes from Gaia DR2 can be considered as most accurate one). In that sense the statement you've added is totally correct. It just has nothing to do with observable universe :) --Ghuron (talk) 15:52, 8 September 2019 (UTC)

I always thought that deprecated rank is used to indicate incorrect (or no longer correct) statements

it is incorrect in that it is fictional. at least it is less correct then the other values. we could also set the other two values to preferred which seems more appropriate while it is also more work, when we are expecting a large number of child objects as the star we know most about (Sol (Q525)) has. 😬 --Shisma (talk) 16:56, 8 September 2019 (UTC)
Thank you for your clarification, Ghuron! I never thought about inverse statements in this context. I agree with Shisma that Epsilon Eridani (Q194394) child astronomical body (P398) Axanar (Q57032278) should be deprecated. (Statements about fictional universes should be limited to items about fictional objects). - Valentina.Anitnelav (talk) 06:37, 9 September 2019 (UTC)
We can also specify reason for deprecation (P2241) qualifier, but I couldn't find any applicable value at list of Wikidata reasons for deprecation (Q52105174). Maybe we should create one? Ghuron (talk) 08:02, 9 September 2019 (UTC)
@Ghuron: well there is fictional speculation (Q63077779) which seems to have precisely that meaning though the wording could be improved. Epsilon Eridani (Q194394) child astronomical body (P398) Axanar (Q57032278) / reason for deprecation (P2241) fictional speculation (Q63077779)--Loominade (talk) 13:22, 9 September 2019 (UTC)

Ok, so it looks like we are close to reach consensus here. Let me summarise it and see if anyone has any objections:

  1. There are no problem when fictional entity (Q14897293) has statement that uses concrete object (Q4406616) (e.g. Axanar (Q57032278) parent astronomical body (P397) Epsilon Eridani (Q194394) as well as all examples from my initial post)
  2. It is possible to have reverse statement where concrete object (Q4406616) utilizes fictional entity (Q14897293) (e.g. Epsilon Eridani (Q194394) child astronomical body (P398) Axanar (Q57032278)) but it should have deprecated rank and reason for deprecation (P2241) fictional speculation (Q63077779) qualifier

Right? Ghuron (talk) 14:12, 9 September 2019 (UTC)

[corrected your summery] agreed--Loominade (talk) 14:38, 9 September 2019 (UTC)
I disagree. I don't think it's appropriate for real entities to have statements pointing to fictional entities like that. A fictional universe can adapt parts of the real world, but that doesn't affect the real objects. --Yair rand (talk) 19:50, 9 September 2019 (UTC)
@Yair rand: I do not see "ideal" approach here, but this one is at least a) consistent b) cause minimum practical problems for both astronomers and "fictional universes". If you see any non-discussed issues - please share them with us --Ghuron (talk) 08:15, 10 September 2019 (UTC)
One problem is also that "fictional universes" does not always follow normal logic. One problem I remember from an article on svwp, is that a fictional armed force only recruited a handfull men each year. The armed force still had thousands of men in service. Another problem I remember from StarTrek is that Bajor has X moons but one less in an alternative timeline. Picard is married and father of a son and still Captain of Enterprise in one timeline, but Admiral and unmarried in another timeline. 622 017 074 (Hej!) 12:53, 10 September 2019 (UTC)
this problem has been addressed and solved here. I don't see how this is a relevant problem in this particular discussion--Loominade (talk) 13:23, 10 September 2019 (UTC)

Conversion of units

To be able to convert from one unit to another, in a safe and predictable way, we need some kind of normalized description of an unit. This is almost the same as Wolfram Language unit code (P7007), but it refers the actual items, and not application specific entries. In that case we also loose precision as we end up with square foot (Q857027), using an undefined foot, where we really should use for example a Swedish foot of 296.9 cm. An example from nowiki is a square Norwegian mil. The entry for Scandinavian mile (Q42764) can't be used, it is an item for the history about the different Scandinavian mil. A fun example for nerds; take a look at the units for w:Magnetic complex reluctance.

So how should this be done? The bulk of the units is composed of a reference to a base or derived unit (which could be implemented as an item) and that item has a power component (which could be implemented as an qualifier). A list of such referenced units are conceptually multiplied together. It is possible to make derived units by merging and simplifying such lists, and when ordered they can be used for further lookup of normalized names. Only the base unit needs a conversion to standard unit (P2442), unless we want to materialize into the item. But if we chose to materialize, then we should not refer to derived units. Because the scheme is rather compute intensive without materialization, the results should be cached in the clients, otherwise it would easily eat the whole load budget for a page at Wikipedia. In this respect, note that the computed units will not change over the pages where it is used.

I would propose

  1. we create a property “unit derivation” that holds references to items for units, and a “unit power” that goes as qualifier. All items should have this qualifier, even those that would otherwise have a power of one (1).
  2. we create properties “unit prefix” that points to an item used as a prefix for this unit, and an “unit offset”, as it might be necessary to recompute the prefix and offset for a computed unit. [A missing point here; there should be a statement that holds “unit precision”. If a unit is defined with a given precision, then a value should not be given with a higher precision.]
  3. we create properties “unit compression” that points to an item defining the algorithm, with a “unit scale” and a “unit log”.

The first makes conversions of length, weight, and most physical properties possible. The second makes conversion of temperature and time possible. The third makes conversions to neper, bel, and decibel possible.

Most likely I have made some important omissions. Jeblad (talk) 20:56, 8 September 2019 (UTC)

  • Did you check conversion to SI unit (P2370) and the way it's included in the configuration for query server/normalized triples? --- Jura 22:23, 8 September 2019 (UTC)
That property holds an implicit reference to a unit and a conversion factor for the prefix. (Neglecting multiplication of two slightly different units, which gives a square length for example.) It will be pretty hard to unwind that to get to a normalized unit representation for non-SI units. The core problem is the assumption of a single system where all occurrences can be referred. Even if something is measured, and you have some quantity of units, it does not imply that you can recalculate that into a quantity of some other unit. For example; we know there is 240 “mansheight” on a “pilskudd”, this is old norse length measures, but we don't know how this relates to meter in the SI-system. Current guesses is 200–500 m. A little warning; some Wikipedias has articles about old measures for old length, area, volume, and weight that is wildly speculative in both value and precision. Those values are imported to Wikidata, and should be reviewed. Jeblad (talk) 00:44, 9 September 2019 (UTC)
Use of precision can have some really weird consequences, like the definitions used by the law Lov om metrisk Maal og Vægt (1875) [8] in Norway. This defines a “tomme” (“inch”) as 26 mm,(page 19) and then goes on to define a “kvadrattomme” (“square inc”) as 7 cm² (actually 6.76 cm²). The original definitions are somewhat better. ;) Jeblad (talk) 01:30, 9 September 2019 (UTC)
  • The current approach works mainly for units that convert to SI. It relies on a separate items being used for units that have the same name but a different conversion.
Temperatures are currently not supported (phab:T174353, a query to convert between °F and °C is at Property_talk:P6591#Query with °F/°C conversion).
There is a request at phab:T228645 to use the same for % and ‰. A similar approach could be used for some historic units, but I wonder if we have that many measures that could use conversion. --- Jura 07:59, 10 September 2019 (UTC)

Adding data of a business person


I want to add relevant data about Jason Washington. I have got all the reference/citation links about each information. Please let me know how to start.  – The preceding unsigned comment was added by GoldenLannister (talk • contribs) at 20:22, 9 September 2019 (UTC).

@GoldenLannister: Please see: Wikidata:Tours. But check also Wikidata:Notability to make sure at least one of the criteria for inclusion are met. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:37, 10 September 2019 (UTC)

SourceMD bot re-launch

I am about to re-launch SourceMD bot today.

The updated version has bug fixes, new features, and new tests, but there may be issues cropping up. If you see an issue (starting today), please add it to

Thanks! --Magnus Manske (talk) 13:51, 10 September 2019 (UTC)

Mayor redux

I can't find the previous discussion decided a few weeks ago. There are two ways we can represent, lets say, a mayor of a locality. We can create a specific "position=Mayor of X" like we have for state governors. This is useful if there are multiple mayors to be added. We also use a generic "position=Mayor" then qualify it with ... what was the final discussion on what field to use for adding in the locality? Was it "located in the administrative territorial" or was it something else like "of="? We use this if we only have one to add. --RAN (talk) 21:58, 6 September 2019 (UTC)

The previous discussion was at Wikidata:Project_chat/Archive/2019/08#University_presidents. I personally am in favor of the simpler mayor (Q30185) of (P642) City Name, placed in position held (P39) rather than occupation (P106) (which should be a more inclusive, enduring profession such as "politician") as mayoral stints are often short spells within broader careers, not occupations in themselves. The alternative is to create a distinct item for every single city/town/municipality on Earth that has ever had a mayor (and ditto for vice mayor, city council member, ombudsman, etc.), which seems like needless busywork that doesn't improve clarity. -Animalparty (talk) 22:44, 6 September 2019 (UTC)
I think that the solution as Animalparty prefers is good. I asked a question like this in the German WikidataForum. But using qualifiers in queries is not possible without a knowledge in SPARQL. So there should be a solution to make it easier to query qualifiers. If somebody with a knowledge of the QueryService knows a solution it were great if the one could tell how it could work or develope a tool for it. -- Hogü-456 (talk) 19:02, 8 September 2019 (UTC)
Agree.Vojtěch Dostál (talk) 19:53, 9 September 2019 (UTC)
Don't agree. If (and only if...) there exists an item about a given mayor position, I see no reason not to use it... Nomen ad hoc (talk) 18:33, 11 September 2019 (UTC).

"To create a distinct item for every single city/town/municipality on Earth that has ever had a mayor": that was just done for French towns. Nomen ad hoc (talk) 20:19, 9 September 2019 (UTC).

Defining birth place original name

When we are talknign about place of birth (P19), we can define the country in which this "place of birth" was at that time (if it's in different than the current country) by adding qualifier country (P17). How about if this place had different name at that time? How can we define that, so we can use that value in Wikipedia Infoboxes? --StanProg (talk) 09:03, 9 September 2019 (UTC)

In the place's item You should use official name (P1448) with qualifiers start time (P580) and end time (P582). See, for example, Saint Petersburg (Q656). And using country (P17) as a qualifier in personality's item is not a good solution. Better use this property in the place's item with the same qualifiers. --Ksc~ruwiki (talk) 11:57, 9 September 2019 (UTC)
Be aware, that official name (P1448) is only for official names; I highly doubt that Saint Petersburg (Q656) ever had had an official name in Hungarian, Spanish or Armenian (while the Wikidata item claims it did…). IMHO the unofficial translations of official names should be added as qualifiers of the respective statements, not as separate statements.
Also, there is already in the proposal discussion an unanswered question of spelling: E.g. Saint Petersburg (Q656) was spelled Санктъ-Петербургъ at the begining of 20th century, not Санкт-Петербург as it is today.
And finally, not only the name changes with time, also the area does. Hence people born in the area of today's Saint Petersburg (Q656) in early 1920s shouldn't claime to have been born in Петроград, if their place of birth was not part of the city at that time.--Shlomo (talk) 03:17, 10 September 2019 (UTC)
"I highly doubt that Q656 ever had had an official name in Hungarian, Spanish or Armenian..." I've resolved that by moving the relevant values to literal translation (P2441). Each save took a very long time! Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:50, 10 September 2019 (UTC)
unless "sankt" is Romanian or Hungarian for "saint", the ones in those languages are in no way literal translations. They're transliterations. Circeus (talk) 15:38, 11 September 2019 (UTC)
If the official name is in Russian, the literal translation should be made from the Russian word “санкт”, not from the English “saint”. But “санкт” is not a Russian word, neither is “Петер” nor “бург”; actually, the Russian (official) name is already a Russian transcription of a (probably) German word/phrase, so that a Hungarian or Romanian transcription of this German phrase is a perfect literal translation of the Russian name.--Shlomo (talk) 16:35, 11 September 2019 (UTC)
See also stated as (P1932). Circeus (talk) 14:23, 9 September 2019 (UTC)
Thank you both. I also looked at stated as (P1932), but I think this is not suitable for our case, since it's not multilingual. --StanProg (talk) 14:55, 9 September 2019 (UTC)

Occidental Mindoro New Website

Good Day!

May we wish to update the domain name address of Occidental Mindoro from to

Thank you & Have a Blessed Day!

Frank John Iniego admin

Occidental Mindoro (Q13867) official website (P856) is now updated. --Bluemask (talk) 08:39, 11 September 2019 (UTC)

Clarification regarding purpose of afflicts (P689)

The description of afflicts (P689) is currently "type of organism which a condition or disease afflicts" but there is also an example where the value of this property is an organ (i.e. liver cirrhosis (Q147778) afflicts (P689) liver (Q9368)) - can we change the description to include "part of organism" (e.g. "type of or part of organism which a condition or disease afflicts")? I tried to do this but it was reverted as someone saw it as vandalism / disrupting behavior.

More examples where afflicts is not an organism but a part of an organism:

Iwan.Aucamp (talk) 11:46, 11 September 2019 (UTC)

I think the liver example was leftover by accident when the property was being discussed. You can clearly see in the discussion that the property at first included no provision for organs, then it did, then a different property was proposed by another user. Clearly that property didn't make it through for some reason (but I can't find a trace of that proposal). Actually changing the scope of a property, I've found, is like talking to a wall on here (don't even get me started about changing a property name). I think it's more sensible to accept defeat, delete the liver cirrhosis example and head over to WD:PP to make a proposal. Worse come to worse it'll be voted down with prejudice because "afflicts (P689) already exists", and then no one can argue against your changes. Circeus (talk) 15:52, 11 September 2019 (UTC)
@Circeus: Just to note again these and many more use it similarly to liver cirrhosis (Q147778): sodium cyanide exposure (Q21175308), Trinitrotoluene exposure (Q21175396), rheumatoid arthritis (Q187255). Thank you though for input and advice and I will proceed as you suggest. Iwan.Aucamp (talk) 17:52, 11 September 2019 (UTC)

Importing names & aliases from Wikipedia


I'm particularly interested in Norwegian Wikidata entities. There are a lot of aliases missing from a lot of entries, and I had an idea about supplementing them that might also be valid for other languages. What about going through all Norwegian Wikipedia articles, finding all links, grabbing the text from those links, and adding them as aliases to the article -> entity being pointed to? For Norwegian, some heuristics could be added. If the entity being pointed to has a name which is capitalized, it should be considered a proper noun, and thus the text within the link should also be capitalized. This rule will of course not work for e.g. German.

Here is a valid example:

In the Norwegian article about Albert Einstein, there is a link "Keiser Wilhelm-instituttet for fysikk", which points to The corresponding Wikidata entity, "Kaiser-Wilhelm-Gesellschaft", does not have the Norwegian alias "Keiser Wilhelm-instituttet for fysikk".

I' not sure it will be 100% fool proof, but I think it will be pretty good.

I'm able to extract this data myself (e.g. JSON), but unsure how to feed it into Wikidata.

Kind regards, Emil

If you are sure your aliases are good, and you have the QId of the item, then quickstatements provides a method for adding aliases. --Tagishsimon (talk) 10:06, 11 September 2019 (UTC)
  • Something like this is being done regularly from one or several Asian language Wikipedias. Interestingly it also yields labels in Latin script languages (e.g. for items about Europeans with articles in these languages, but not elsewhere). As far as I can tell, the Latin script part works quite well and I'm not aware of complaints about the others. Maybe you could adapt the code for nowiki. I try to recall the name of the bot .. --- Jura 10:09, 11 September 2019 (UTC)
  • Be aware - when a Wikipedia article discusses a class of objects (say "Chinese porcelain") there will often be many redirects that represent the types of Chinese porcelain that are discussed in the article. In Wikidata, each of these types should be (and may already be) a subclass of "Chinese porcelain". In these cases, where the redirects have been added as aliases to the main article, those aliases all have to be removed. I do this sort of work a lot. So please be careful that the redirects don't represent subtypes of the main topic. Thanks. - PKM (talk) 19:23, 11 September 2019 (UTC)

Properties, right use and moving values

Hey, I wonder how Wikidata works. Does everybody know, what properties and qualifiers use and when? Sometimes people propose new properties and there is a discussion, that such a case may be handled by already existing property. So I wonder how accurate Wikidata are?

This takes me to another question. If a more specific property is approved, those it means that data/values will be moved from general property to this specific one? --Juandev (talk) 18:13, 11 September 2019 (UTC)

Wikidata:WikiProjects typically document the properties (and qualifiers etc) used in various different domains. ArthurPSmith (talk) 19:16, 11 September 2019 (UTC)
There are many different types of properties and identifiers. Some identifiers are used to link items to movie databases, some are used to link items to online newspaper topics while others are used to link humans and organizations to their respective social network profiles. --Trade (talk) 20:23, 11 September 2019 (UTC)

possible new tool: VIAF identifier importer

Hi folks -- I'm not sure quite the right place to ask this, so if there's a better place to do so, please let me know.

I've been developing a tool to make it easier to import VIAF-linked identifiers into Wikidata. It's far enough along now that I'd like to get community input on the utility of the tool and how we might leverage it to be useful for the Wikidata community as a whole. Right now, it's hosted on my own web space. I'd prefer not to link to it publicly, so I've made a quick showing how it works.

It looks up a Q-item and a VIAF ID, then looks at all the other identifiers linked from VIAF. It formats them as necessary, validates them against the known format as a regular expression (P1793) associated with the identifier, and spits out the appropriate QuickStatements-formatted data for all the identifiers it validated.

Not explained in the video are what happens to data that don't work out: if the tool doesn't know what to do with an identifier or it fails the regex check, it's noted as an error at the bottom and not put into QS list.

Does this look like something that would contribute to the Wikidata world and not already accomplished elsewhere? If so, what would be an appropriate development path to bringing it into production?

Some areas for development I have in mind include:

  • adding rules for handling some identifiers it currently doesn't know what to do with
  • there could be a bot that does the same thing for records that already have VIAF ID (Q19832964) statements, and by-passes QuickStatements and just adds the extra ids in automatically

What do you think?

Thanks - Kenirwin (talk) 00:37, 11 September 2019 (UTC)

Keep going. It's nice. Thierry Caro (talk) 10:38, 11 September 2019 (UTC)
Can you add references to the imported claims? I think "stated in (P248): Virtual International Authority File (Q54919), VIAF ID (P214): (source VIAF ID), retrieved (P813): (date of data retrieval)" should be sufficient, and it is fairly simple to add such references to the QuickStatements input. From my experience VIAF has way too many mixed profiles about different persons, thus I worry about data quality and it would be easier to identify where potential mistakes come from if there were such references. —MisterSynergy (talk) 10:46, 11 September 2019 (UTC)
@Kenirwin: First of all, thank you very much for your work in this area. I add loads of authority control IDs and any new tool is very welcomed. I have one major question. Are you aware of this tool, which I am using at the moment? Can you please explain how your tool will be better than this one? From what I can see in your YT film, your tool at the moment seems to be a bit more complicated to use than Magnuss's (because with his tool you just get a pop up on the WD element page, you confirm VIAF identity and that's it), but maybe I'm wrong. And I absolutely support what @MisterSynergy: wrote about the references. The lack of references is the main problem with the current tool and it should be adressed in any new one. Please tag me if I can be of more help. Powerek38 (talk) 11:29, 11 September 2019 (UTC)
Thanks @Powerek38: and @MisterSynergy: -- I wasn't aware of the other tool, so I'll take a look at that. And we can definitely build in some references. If there are things I can learn from Magnus's tool, I'll see what I can improve in my own. This is exactly why I wanted to come to the community with the proof-of-concept level tool before trying to develop it much further. Thanks for the feedback! _Kenirwin (talk) 18:57, 11 September 2019 (UTC)
@MisterSynergy: -- can you tell me how to use/install Magnus' script? It doesn't give any instructions for use, and I don't know what to do with that code. Thanks -Kenirwin (talk) 00:32, 12 September 2019 (UTC)
You need to add mw.loader.load( '//', 'text/javascript' ); // [[User:Magnus Manske/authority control.js]] to Special:MyPage/common.js, and then reload an item page. You'll see a new "Authority control" entry in the "Tools" menu in the left navigation bar. --MisterSynergy (talk) 06:55, 12 September 2019 (UTC)
@Kenirwin: this toll can be really really useful for me, please tell me when it's ready for use. Note that VIAF also contains sex or gender (P21) and ORCID iD (P496) under "about" in the bottom of the VIAF page, both of which could be useful to harvest. Good luck, DGtal (talk) 09:51, 12 September 2019 (UTC)

Should iPhone 5s, model A1533 (Q66816688) be linked to smartphone (Q22645) via subclass of (P279) or instance of (P31)

To me it seems a iPhone 5s, model A1533 (Q66816688) should be instance of (P31) of smartphone (Q22645) and subclass of (P279) of iPhone (Q2766). If every smartphone (e.g. Xiaomi Mi 2 (Q2355401)) is also a subclass of (P279) of smartphone (Q22645) then it is unclear where actual sub categories of smartphones (e.g. modular smartphone (Q18611337) and phablet (Q521097)) should go.

Also if iPhone 5s, model A1533 (Q66816688) is related to smartphone (Q22645) via subclass of (P279) should these not also be sub-classes instead of instances:

There is some further discussion on the matter here: Talk Page for iPhone (Q2766)

In Help:Basic membership properties there is also a similar case

Maybe there is some resources on how this should be used that I am not aware of that provides guidance. From my point of view I would like to have different ways of finding specific smartphones (iPhone 7, Galaxy A5 2017, etc...) from the smartphone object than finding things like modular smartphone and phablet form the smartphone object. Maybe the right solution is to not use subclass of (P279) for modular smartphone (Q18611337) and phablet (Q521097) and use something different there

Iwan.Aucamp (talk) 09:51, 11 September 2019 (UTC) Iwan.Aucamp (talk) 18:00, 11 September 2019 (UTC)

An individual phone would be an instance of smartphone (Q22645). iPhone 5s, model A1533 (Q66816688) represents a phone model, including all of the phones that are instances of that model, so it's a subclass. It should be an instance of smartphone model (Q19723451). Since iPhone 5S (Q14850354) is being treated as a series of models, it can be linked using part of the series (P179). Ghouston (talk) 09:18, 11 September 2019 (UTC)
So you are saying if there was an entry for say Steve Job's iPhone then that would be related to iPhone (Q2766) via instance of (P31)? How do you propose to relatemodular smartphone (Q18611337) and phablet (Q521097) to smartphone (Q22645)? Would it not be better if they were related in different way than actual models of smartphones? Iwan.Aucamp (talk) 09:51, 11 September 2019 (UTC)
Apologies, I missed your point of part of the series (P179) - will try make proposal for linking below. Iwan.Aucamp (talk) 09:55, 11 September 2019 (UTC)
Do you have any input on other similar cases? Iwan.Aucamp (talk) 10:17, 11 September 2019 (UTC)
Individual phones don't usually have items in Wikidata, although occasionally people have imported museum catalogues which include such things (I'm not really convinced that museum catalogues are useful here.) But in principle, you could have an item for an individual phone which would be an instance of iPhone 5s, model A1533 (Q66816688). There are more likely to be items for individual ships or aircraft, which are also often built as particular models. Vehicle models like Tesla Model S (Q1463050) are the same, it can be an instance of automobile model (Q3231690) and a subclass of everything else. Programming languages are a different kind of thing. I'm not sure whether Ruby (Q161053) should really be considered a single language, since programming languages typically develop over time and form a family of related languages, even if they are referred to colloquially as a "language" singular. Ghouston (talk) 01:15, 12 September 2019 (UTC)
What I find a bit difficult is that iPhone 5s, model A1533 (Q66816688) is part of the series iPhone 5S (Q14850354), but that in turn is part of the series iPhone (Q2766). I.e., nested series. Ghouston (talk) 09:24, 11 September 2019 (UTC)
Identity is a complex subject but what Ghouston said describes our general approach to the topic even if some old statements derivate from best practices. ChristianKl❫ 17:14, 12 September 2019 (UTC)

Duplicate entries

I believe Mangon, Johann (Q27586756) and Johannes Mangon (Q1697986) refer to the same person. Can this be verified and if so do the entries need merging? (Also, where is the right place to report something like this?) Carcharoth (talk) 12:54, 11 September 2019 (UTC)

One describes a biographic article, the other the human. No, they shoudn't be merged. Sjoerd de Bruin (talk) 13:02, 11 September 2019 (UTC)
OK, can they be linked in some way so people realise they are about the same person? Carcharoth (talk) 13:22, 11 September 2019 (UTC)
They are. Q27586756#P921 and Q1697986#P1343 --Tagishsimon (talk) 13:25, 11 September 2019 (UTC)
Thank you. As a human, I find "also known as" more obvious than 'main subject' or 'described by source' (see comments below), so made that change. Carcharoth (talk) 13:31, 11 September 2019 (UTC)
@Carcharoth:If I write a book about say Barack Obama, and I use "Obama, Barack" as the title of my book, then I cannot say the book is also known as the person, nor can I say the person is also known as the book, the one is a person, the other is a book which subject is person. This is the same here. Mangon, Johann (Q27586756) is a literary work about a person called Johannes Mangon (Q1697986) - or stated differently as a human would, Johannes Mangon (Q1697986) is the subject of the literary work Mangon, Johann (Q27586756). Hope that helps clear things up. Iwan.Aucamp (talk) 13:49, 11 September 2019 (UTC)
Thank you (though I did know that already). What I am trying to say that the way Wikidata presents its information can be confusing to humans. Unless you are familiar with Wikidata, it can be difficult to work out what is going on. You and I understand it, but is there a way to make the link more obvious and not buried in the data? Carcharoth (talk) 14:00, 11 September 2019 (UTC)
Would this change be acceptable? Carcharoth (talk) 14:02, 11 September 2019 (UTC)
tbh, I think you're missing the point, Carcharoth. Wikidata depends, ultimately, on a shared understanding of the ontologies within it. The convention for linking a biographical article item to the item for the subject of the biography is as has been done in this instance. Your adding dubious aliases is not how it is done, and will not IMO lead to any greater understanding. On the contrary, it risks confusion by making previously obviously distinct items appear at first glance to be similar. Amending the description as you have works & might be improved by tacking his DoB and DoD at the end of the string. --Tagishsimon (talk) 14:07, 11 September 2019 (UTC)
Ah, I see they are linked already. But it is not immediately obvious. I guess machines parsing the data would see the link. Is there a way to make the link more obvious to humans? On 'Johannes Mangon' it was fairly easy to find "Johann Mangon" under 'Also known as', but when looking at 'Mangon, Johan' I had to scroll down to find 'main subject ' = "Johannes Mangon". I made these edits and this edit to help people make the link. Carcharoth (talk) 13:30, 11 September 2019 (UTC)
I really don,t see how it could have been any more clear even before you went ahead and muddied things up. There is a very clear description of the item ("entry in the Allgemeine Deutsche Biographie") and I really don't understand how anyone seeing that would mix up this instance of (P31) biographical article (Q19389637) with a real person... You're just looking for a reason to get confused at this point. Circeus (talk) 15:57, 11 September 2019 (UTC)
  • We had a lot of problems with these items being seen as identical and people attempting to merge .. The earlier label me it clear that they are different. Maybe @Simon Villeneuve: wants to comment. --- Jura 18:17, 11 September 2019 (UTC)
    As I have said before on the French village pump about this, I think that putting the acronym of the publication between parenthesis in the label is a false good idea. It is not only a Wikipedian reflex, but also a degradation of the knowledge base (breaking of the "no parenthesis in the label" rule) to try to solve a human fail.
    I don't have a perfect solution for this. Maybe that one day, Wikidata will offer a colour code of (the background, the text, only the label, only the background of the title of the page, etc.) for every item with the principals "instance of/subclass of" classifications (for example, green for plant (Q756), blue for body of water (Q15324), white for written work (Q47461344), red for human (Q5), etc.). Then, humans will immediately see that the two items (white and red in our case) are different and don't have to be merged. Simon Villeneuve (talk) 20:14, 11 September 2019 (UTC)
    The label here would just reflect the page title from WS.
    Anyways, I like the part about colors. Maybe we could start with one for p31=Q5 --- Jura 11:46, 13 September 2019 (UTC)
    +1. Nomen ad hoc (talk) 13:12, 13 September 2019 (UTC).

Monolingual text labels

Q14874254#P1448 if you are using non-English UI, you might still see Chinese (Singapore). Where can people translate it?--Roy17 (talk) 13:28, 12 September 2019 (UTC)

It's not clear to me what you want. If you want to add an offical name in another language you would add a new statement. ChristianKl❫ 17:13, 12 September 2019 (UTC)
That name is defined in the cldr extension (you can find that with MediaWiki code search, but as far as I can tell it doesn’t come from upstream CLDR data, so I think it’s manually added to the LocalNames/LocalNamesEn.php, LocalNames/LocalNamesDe.php etc. files. I’m not sure if the preferred process is submitting Gerrit changes for those files, though – translatewiki:CLDR#Localised language names also has some other instructions. --Lucas Werkmeister (WMDE) (talk) 10:28, 13 September 2019 (UTC)
I think French has had a few translations added. Unfortunately discussions on Wikidata:Bistro aren't searchable .. --- Jura 11:39, 13 September 2019 (UTC)

similar problem

Look at any date of birth statements in Chinese (zh), you'll see some nonsense like "31 12 1999". Where is the configuration? It should be YYYY年MM月DD日 .

A general question would be, where is all the configuration located? Where to go to fix these bugs, or at least report them? Some ppl must know it because for example that Chinese (Singapore) label is translated to German but not to French. dob format is also correct in German.--Roy17 (talk) 19:12, 12 September 2019 (UTC)

You can report them in Phabricator. Plenty of them have already been. --Matěj Suchánek (talk) 11:40, 13 September 2019 (UTC)

First images from the UNESCO archives now on Commons

Dear all

C:Category:Media files produced by UNESCO: 2019-09

I’m very pleased to say that the first batch of 100 very high resolution photos (plus descriptions) from the UNESCO archives has been uploaded to Commons by our friends at Wikimedia Sweden. There are some amazing photos in this first batch including Italian National Archives material being washed and dried in a train station after a flood, and the installation of a Henry Moore sculpture. We would really appreciate it if you could take 5 minutes to add some to Wikidata items, show what is possible with structured data on Commons etc so that UNESCO will be encouraged to make more images available.

C:Category:Media files produced by UNESCO: 2019-09

Thanks very much

John Cummings (talk) 18:57, 12 September 2019 (UTC)

Is there a page that helps Chrome users viewing TIFF pages? --SCIdude (talk) 07:38, 13 September 2019 (UTC)

Activity policies

Hello. I have recently created a series of items to be used as values for activity policy in this place (P5023). You can see a list here. Would you be nice enough to try to find better labels in the English language and then in the other ones you may know? What we have needs to be rationalized and then expanded to cover various wordings, I believe. Thierry Caro (talk) 03:50, 13 September 2019 (UTC)

Commons might have signs and categories one could add. --- Jura 10:56, 13 September 2019 (UTC)

Two items, one sitelink


Just noticed Mabel Addis (Q67436667) & Mabel Addis (Q67436668) − two items with both a sitelink to en:Mabel Addis :-)

Leaving it like in case someone needs/wants to investigate :)

Jean-Fred (talk) 15:21, 13 September 2019 (UTC)

Thanks at diff pages

Thanks to everyone involved for correcting the previous issue with undo, and now we seem to have the similar problem with the "thank" function which is now still available from the history pages, but no longer from the diff ones. — Mike Novikoff 20:01, 14 September 2019 (UTC)

I can see them, do you have a sample? Note that they are not displayed when the diff spans multiple revisions. --Matěj Suchánek (talk) 08:33, 15 September 2019 (UTC)
Now I can see them too. Seems like it was either a temporary glitch or just my mistake, sorry. — Mike Novikoff 07:05, 17 September 2019 (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) 15:40, 19 September 2019 (UTC)

Item change request

In Wojtyla (Q8029293) please change Wojtyla to Wojtyła. There is a multiple languages to change. Thanks! -- 06:04, 15 September 2019 (UTC)

You mean Wojtyła (Q56541347)? --SCIdude (talk) 11:34, 15 September 2019 (UTC)
Wojtyla (Q8029293) is intended for Mary T. Wojtyla from Chicago. For Karol J. Wojtyła from Wadowice, use Wojtyła (Q56541347). Welcome to the world of poorly structured data :'( --Shlomo (talk) 11:39, 15 September 2019 (UTC)
Yes, so change family name Wojtyla to Wojtyła (Q56541347) in John Paul II (Q989). Thanks!!! -- 18:19, 15 September 2019 (UTC)

✓ Done Please consider creating an account and login in to edit. --- Jura 22:36, 15 September 2019 (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) 15:40, 19 September 2019 (UTC)

Vandalism from an IP. Help

An IP is currently vandalising items on Alan Turing (Q7251). See the modifications done. Can anybody please take action. John Samuel (talk) 11:51, 17 September 2019 (UTC)

✓ Done. Next time you should ask on Wikidata:Administrators' noticeboard. Ayack (talk) 11:59, 17 September 2019 (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) 15:40, 19 September 2019 (UTC)

Batch deletion

Could someone please delete all the items here? I created them for GeneDB a while ago but the reference genome for that species changed, all the gene and protein IDs are changed as well, and there is no good mapping (for redirects) for them. I will create new items some time this week. Thanks! --Magnus Manske (talk) 14:09, 17 September 2019 (UTC)

✓ Done all deleted.-- Hakan·IST 16:26, 17 September 2019 (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) 15:40, 19 September 2019 (UTC)

Lexicography page move request

Please can someone with translator rights move Wikidata:Lexicographical data/FAQ to "Wikidata:Lexicographical data/Wiktionary FAQ", so that we can use the existing page name (and split out relevant content) for a more general Lexicography FAQ? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:41, 17 September 2019 (UTC)

✓ Done --Esteban16 (talk) 00:15, 19 September 2019 (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) 15:40, 19 September 2019 (UTC)

Stage name or pseudonym?

Italian singer Lorenzo Cherubini is best know as “Jovanotti”. In this case, I should use stage name (Q1055303) or pseudonym (P742)? -- 20:08, 13 September 2019 (UTC)

A stage name is a kind of pseudonym. And there's only a broad property about pseudos. So I would suggest pseudonym (P742), with object has role (P3831)=stage name (Q1055303) as a qualifier. Nomen ad hoc (talk) 20:23, 13 September 2019 (UTC).
@Nomen ad hoc: thank you for that elegant suggestion! What do you think of a similar construction pseudonym (P742), with object has role (P3831)=trade name (Q1417728) for fashion designers who work under a variant of their names (e.g. Adele Filene (Q354423). formerly "Juliusburger" running her fashion business as "Adele Jay")? I've been looking for a better way to model this situation. - PKM (talk) 20:43, 13 September 2019 (UTC)
That seems accurate! Nomen ad hoc (talk) 20:58, 13 September 2019 (UTC).
A trade name normally refers to a business, not a person. Occasionally a person runs a business as a sole proprietorship, in which case they amount to the same thing. But usually a business is a corporation or partnership, so the entity referred to by the trade name is different than any individual human. Jc3s5h (talk) 14:57, 14 September 2019 (UTC)

educated at (P69)

Are there any way to use educated at (P69) to indicate that someone who have been homeschooled? --Trade (talk) 22:28, 13 September 2019 (UTC)

educated at (P69)+no value + has cause (P828) homeschooling (Q83602)? Circeus (talk) 04:52, 14 September 2019 (UTC)
Back in 2016 I suggested using a generic "homeschooled" item here (Property_talk:P69#Adding_some_non-institutional_exceptions) and people seemed okay with it, but I never quite got around to following it up. Andrew Gray (talk) 17:11, 14 September 2019 (UTC)

Proposed change of categories notability policy

Hi. Perhaps you would be interested in commenting Wikidata talk:Notability#Change to 1. 4. regarding Commons. Cheers, Nomen ad hoc (talk) 16:52, 14 September 2019 (UTC).

Editors who are interested in that may also be interested in Wikidata:Properties_for_deletion#Property:P373. Thanks. Mike Peel (talk) 18:05, 14 September 2019 (UTC)
You're welcome :). Nomen ad hoc (talk) 19:19, 14 September 2019 (UTC).

Stood for election

There's no such property? That a person stood for election an election, with qualifiers constituency XYZ, votes won XX, result elected/lost/disqualified...? I checked some high profile politicians who lost elections but couldnt find such statements.--Roy17 (talk) 10:28, 14 September 2019 (UTC)

I think you are looking for candidacy in election (P3602), right? —MisterSynergy (talk) 10:40, 14 September 2019 (UTC)
@MisterSynergy: which qualifier should be used for election result/outcome?--Roy17 (talk) 10:48, 14 September 2019 (UTC)
And also what's the qualifier for party represented? Sometimes people have more than one parties and they represent one of them in elections. Sometimes they belong to some parties but stand as an independent... So I think there should be a qualifier for political association in an election?--Roy17 (talk) 11:03, 14 September 2019 (UTC)
I would say represents (P1268) for the party of electoral list (and "no value" the candidacy is an independent). Nomen ad hoc (talk) 11:27, 14 September 2019 (UTC).
parliamentary group (P4100) is generally used for their party affiliation once elected, so it would make sense to use the same property for elections to indicate the group they're standing for. Andrew Gray (talk) 17:07, 14 September 2019 (UTC)
We had a lengthy discussion about this the other day and came to the conclusion that it wouldn't make sense, at least not for non-mono-/bi-party countries. Not sure why it wasn't fix afterwards. --- Jura 17:24, 14 September 2019 (UTC)
Not my field of work, thus I don't really have an idea. In such cases, I usually try to evaluate the current situation by having a look at actual qualifier use for the property in question. For instance this query counts numbers of qualifiers on all 47571 candidacy in election (P3602) statements, which might be a point to start from. —MisterSynergy (talk) 18:53, 14 September 2019 (UTC)
Thx a lot! Your query link is very useful. It should come as part of the standard package on property talk. special:diff/1013281209 I added the info. Thx to all.--Roy17 (talk) 10:22, 15 September 2019 (UTC)

Remove or deprecate obsolete external identifiers

Please express your opinion in this discussion. --Epìdosis 13:25, 15 September 2019 (UTC)

Conflation of trans status with gender identities

gender identity (Q48264) includes items such as transgender female (Q1052281). But in almost all cases, the gender identity of cisgender women is given as female (Q6581072) rather than cisgender female (Q15145779). This is inconsistent.

(Examples: cisgender woman: Ursula K. Le Guin (Q181659); transgender woman: Chelsea Manning (Q298423).)

The root of the problem is that sex or gender (P21) conflates cis/trans status with gender. "Transgender woman" isn't a gender any more than "Australian woman" is a gender.

Fixing this sounds like a big change. I suppose a bot could do it. What do you think? Marnanel (talk) 15:40, 11 September 2019 (UTC)

  • I think that most cisgender women identify simply as a "woman" not as a "cisgender woman". - Jmabel (talk) 16:09, 11 September 2019 (UTC)
  • The problem is that while the cis/trans dichotomy is defined, at least in theory, as opposing subsets of "a thing", there is as far as I know, no clear definition of what that "thing" is. What is the category to which these two things belong? It's not really "trans status" any more than it is "cis status". Both these are names for the subsets and not the name for the set. GMGtalk 17:39, 11 September 2019 (UTC)
  • Pinging @Kaldari:, who was working on a new data model to address this issue. I'll make the same comment here as I did to them: Ideally I would like cisgender to be included as a parallel to transgender, rather than singling out being transgender as a "status". But I realize it is not possible to do this accurately as most cis people do not explicitly state that they are such, some trans people are stealth, etc. Also of course many people are still ignorant of the word cisgender and some will complain of it being a neologism, etc. Funcrunch (talk) 18:03, 11 September 2019 (UTC)
    • The data model I'm proposing is at User:Kaldari/Sex or gender. It would address Marnanel's concerns I believe. As to whether or not to mark people as "cisgender", I agree with Funcrunch that ideally we would indicate both, but practically, I don't think it would be doable. Kaldari (talk) 18:20, 11 September 2019 (UTC)
      • Probably a stupid question, but, so cis/trans male/female is a nice neat system that makes sense in a vacuum. a taxonomy...doesn't that make "intersex male" or "genderqueer female" in the trans category by default? If intersex is a sex category with no corresponding gender, and genderqueer is a gender with no corresponding sex category. Doesn't that make the whole system diverge significantly from the way that "cis" and "trans" are normally defined? I'm just not sure I understand how these classifications come together neatly, assuming they do. GMGtalk 20:11, 11 September 2019 (UTC)
  • I am at the start of my own experiment to collect conversations on this topic at Wikidata:WikiProject LGBT/gender after the model of conversation collection at en:Wikipedia:Manual of Style/Gender identity. I do not think this issue has an obvious easy answer besides listing the positions of the discourse in a central place and sending people to a central discussion location. I think that Wikidata:WikiProject LGBT should be the place to convene this discussion. Blue Rasberry (talk) 19:52, 11 September 2019 (UTC)
  • I would be very hesitant in a database to describe someone as "cisgender" in terms of identity rather than just presentation if they have not asserted such. How can we presume that a person who has stated nothing on the subject embraces their assigned/biological gender, any more than we could presume they do not? - Jmabel (talk) 20:22, 11 September 2019 (UTC)
  • For a previous discussion on this exact topic, see Wikidata:Project_chat/Archive/2019/03#Let's_talk_about_gender. -Animalparty (talk) 18:46, 12 September 2019 (UTC)
    • Thanks; I didn't find that when I searched before I posted this. I have now linked that discussion from Talk:Q48264, so it's easier to find. Marnanel (talk) 13:20, 13 September 2019 (UTC)
  • The models described seems to miss an important point. How important is the gender_question for the person in question? Has Chelsea Manning said anything that tells us how important this question is to her? My gender is mainly important in my relation to my wife and my relatives. In relation to the rest of the world, I would say that it does not make much any sense at all. I do not wear womans clothes or mens clothes. I simply wear clothes. When I visited the hospital last time, I was in a room for two. My bed neighbour was first an older man and he was replaced by a young woman. Gender is not that important for all of us, not even my doctor. I first of all identify myself as a human. Ethnically I belong to a suppressed minority you can find in several nations in Nothern and Wester Europe, but I do not identify myself as such. I have Swedish citizenship, but I do not identify myself as a Swede. Such things are not important to me. I have other things I consider very important, but none of these things are among them. 622 017 074 (Hej!) 19:04, 12 September 2019 (UTC)
  • P21 deliberately doesn't only include gender information. It is not conflating things. This issue has been discussed numerous times. --Yair rand (talk) 19:33, 15 September 2019 (UTC)

Should nervous system (Q9404) have part of (P361) human body (Q23852)

I removed part of (P361) human body (Q23852) from nervous system (Q9404) but it was reverted and claimed to be vandalism. I don't quite get how this relation is appropriate. Surely not all organisms with a nervous system can be listed there? Iwan.Aucamp (talk) 19:00, 11 September 2019 (UTC)

It seems body (Q170494) would be more appropriate for which human body (Q23852) is a subclass. Even Invertebrata (Q43806) have nervous system (Q9404). ElanHR (talk) 19:57, 11 September 2019 (UTC)
I'm going to remove it again seeing as the person who made accusations of vandalism seem uninterested in discourse. Iwan.Aucamp (talk) 21:09, 11 September 2019 (UTC)
@Andreasmperu: While I remember that you don't like to engage in much discussion, I do see this way of reverting a user and not engaging with them problematic. To me it doesn't seem like it leads to an integration of new users that want to engage with our project. ChristianKl❫ 12:34, 12 September 2019 (UTC)
User:Andreasmperu Have again reverted my change with a comment "Once again, do not delete statements with valid references". Not sure how to resolve this as long as User:Andreasmperu is not willing to have a discussion. So I guess I will just have to discuss with others and ask that User:Andreasmperu has the courtesy to honour the consensus of those who are interested in discourse.
@Was_a_bee:, seeing as you originally added the statement, please also weigh in here and maybe @ChristianKl: or @ElanHR: also has some advice and input.
So to be clear, the claim in question is nervous system (Q9404) part of (P361) human body (Q23852). This has reference Terminologia Anatomica (Q286567).
"Terminologia Anatomica (TA) is the international standard on human anatomical terminology." (Terminologia Anatomica (Q286567)). What it describes when it describes things like little toe, index finger and heart it is specifically describing these parts of human (Q5) anatomy and human body (Q23852). It does not describe the parts of dog anatomy, nor ant anatomy, nor sponge anatomy, nor plant anatomy, nor cat anatomy. It is not making the claim that every heart is part of a human body.
So the nervous system entry in TA does not say that every nervous system is part of a human body, what it says is that a human body has a nervous system, it does not try and claim that cats don't have nervous systems, nor does it claim that a cat's nervous system is part of a human body.
The claim nervous system (Q9404) part of (P361) human body (Q23852) however does imply to me that everything that has part (P527) nervous system (Q9404) also is part of (P361) human body (Q23852). Now as far as I know cats do have nervous systems and those nervous systems of cats are in not part of human bodies (regrettably I have no citation for this ...).
Maybe I'm wrong, and I'm willing to hear arguments that I'm wrong, so if anybody is interested in defending the statement nervous system (Q9404) part of (P361) human body (Q23852) please do so.
I think better statements may be:
Iwan.Aucamp (talk) 14:10, 14 September 2019 (UTC)
@Okkn: Also pinging you as the only remaining member of Anatomy project that was not pinged Iwan.Aucamp (talk) 14:19, 14 September 2019 (UTC)

Tobias1984 (talk) Andy Mabbett (Pigsonthewing); Talk to Andy; * *Andy's edits TypingAway (talk) Daniel Mietchen (talk) Tinm (talk) Tubezlob Vincnet41 Netha Hussain Fractaler Tris T7 TT me Photocyte Nomen ad hoc GoEThe (talk)
Pictogram voting comment.svg Notified participants of WikiProject Biology

Pinging biology project, I really would like to hear any argument for keeping this claim. Iwan.Aucamp (talk) 16:33, 14 September 2019 (UTC)
Hello @Iwan.Aucamp: I understand well what you say. The distinction between “human specific” and “whole species”is always the point that can provoke this form of controversy in Wikimedia. If I'm going to say the conclusion first, I agree with you philosophically. But I think there are difficulties to separate all connections between “human specific” and “whole species”anatomical entries from practical point of view. From practical/pragmatical points, not by philosophical points, anatomy in Wikimedia have been generally firstly "human" anatomy, and secondary "animal" anatomy. Possible practical refinement I can imagine is changing term which is used for connecting items. For example, from "part of" to "is also a part of" or something like that. (Sorry, I'm not good at English D:)
From the philosophical point of view, I know the most best ideal is that there are species specific anatomical Wikidata entries (for example, human brain, salmon brain, mosquito brain, and so on) for all species. And connecting them within their own groups. In Wikidata, one (perhaps only) successful example which is simlilar to this kind of ideal is gene/protein data which have been editing by User:ProteinBoxBot. Here is brief explanation of their data structure concept (w:en:Template:Infobox_gene#Data_structure). But even this most successful realm, whole species are not treated. Data of only some species (called model organism by biologists) are registered.
Important thing here is numbers. So I list up some important numbers relating to this issue.
*A: Human anatomical terms count is, in roughly worded, 7,000 (this is in Terminologia Anatomica 98 ID (P1323) and TA2 ID (P7173)). If detailedly counted, it's 79,000 (in Foundational Model of Anatomy ID (P1402))
*B: Number of known biological species is roughly 1,000,000.
From these, we can calculate needed number of anatomical entries for ideal Wikidata, by A×B. Though this is very rough Fermi estimate, it is 7 billion pages or 79 billion pages. Even if creating 10 pages per second by a bot automatically, this needs 22 years or 250 years.
Actually, currently (I suppose also in near future) there is not such external database which contains all species anatomical terminology.
FYI, currently the most prominent database which challenges cross-species anatomy is Uberon (UBERON ID (P1554)). Although I don't understand well about their design yet. --Was a bee (talk) 18:02, 14 September 2019 (UTC)
@Was a bee: thank you for your input. I agree that having intermediate items to link nervous system (Q9404) to each organism with a nervous system (Q9404) as we do for with human nervous system (Q11392181) is not practical, but I think a better option would be to have specific intermediate items for prominent things like human body, and for the case of say southern three-banded armadillo (Q75070) rather just put has part (P527) nervous system (Q9404) than nervous system (Q9404) part of (P361) southern three-banded armadillo (Q75070). Even if there was no human nervous system (Q11392181) I would still say human body (Q23852) has part (P527) nervous system (Q9404) is better than nervous system (Q9404) part of (P361) human body (Q23852). Also note I did not create human nervous system (Q11392181) - but given it is an appropriate item to exists I still think my proposal is acceptable and optimal. Even the wikipedia page for nervous system (Q9404) is not specific to humans. Iwan.Aucamp (talk) 18:25, 14 September 2019 (UTC)
To be more explicit maybe, I would be completely happy to not have human nervous system (Q11392181) part of (P361) human body (Q23852) and just have human body (Q23852) has part (P527) human nervous system (Q11392181) Iwan.Aucamp (talk) 18:26, 14 September 2019 (UTC)
And if needs be I would be fine if human nervous system (Q11392181) was deleted and we just have human body (Q23852) has part (P527) nervous system (Q9404). But I still think that nervous system (Q9404) part of (P361) human body (Q23852) is not appropriate. Another property maybe, but still I think inverting the relation and just having human body (Q23852) has part (P527) nervous system (Q9404) is more optimal. Note however that human nervous system (Q11392181) and unless someone goes and deletes it we should incorporate it. Iwan.Aucamp (talk) 18:30, 14 September 2019 (UTC)
I'm sure there are other use-cases for human nervous system (Q11392181); such as being the main subject of scientific papers and books. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:06, 14 September 2019 (UTC)
I also think it is useful, and I'm not suggesting it should not exist, I'm just trying to point out that arguments against the existence of human nervous system (Q11392181) is not valid arguments for the existence of nervous system (Q9404) part of (P361) human body (Q23852) in my view and that given it does not exists nervous system (Q9404) part of (P361) human body (Q23852) should still not exist. Iwan.Aucamp (talk) 19:43, 14 September 2019 (UTC)

If user:Andreasmperu persists, this matter should be raised on the admin noticeboard. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 00:02, 15 September 2019 (UTC)

Iwan.Aucamp Proposal 000

So to make it clearer and to resolve this issue quicker, here is my exact proposal, please vote below it.


I propose the following changes:


 Support I think that the relation nervous system (Q9404) part of (P361) human body (Q23852) is not appropriate as to me it implies everything that has part (P527) nervous system (Q9404) is also then part of the human body - or something, it is actually very painful to think of it for me. Not ever nervous system (Q9404) is part of a human body (Q23852) - to my knowledge this is just factually false. For example cat's have nervous systems, they don't have human bodies. Iwan.Aucamp (talk) 20:15, 14 September 2019 (UTC)

 Support - Jmabel (talk) 21:33, 14 September 2019 (UTC)

 Support While I agree with @Was a bee: that a corollary to this approach would be the A x B issue, I think notability takes care of this. Other specific versions (canine, feline nervous systems, etc.) can be created when there is a structural need such as topics of books (as @Pigsonthewing: described). ElanHR (talk) 23:59, 14 September 2019 (UTC)


No opposition, changes have been made. Iwan.Aucamp (talk) 19:34, 15 September 2019 (UTC)

Note that this problem also exists in many similar anatomic structure items which don't distinguish between the general version and the human-specific version. There's a lot of cleanup to do. --Yair rand (talk) 19:40, 15 September 2019 (UTC)

Inclusion of free text description statements?

At Wikidata:Property proposal/scope and content, there is some discussion about the possible inclusion of a property for some type of free text description. --- Jura 02:56, 12 September 2019 (UTC)

We have already discussed at length in that property proposal that this is not a "free-text" field. In fact, it isthe opposite—archival metadata, produced by professional catalogers (not Wikimedians) we would like to import into the Wikidata items about those archival collections. Jura seems to be trying to make make some kind of point by repeatedly making these claims, and now he is forum-shopping it over here. Dominic (talk) 17:52, 12 September 2019 (UTC)
Yes, it's not free text just as much as Wikidata item descriptions aren't free text either. Read: it's free text just as much as Wikidata items descriptions are free text.
It seems like a good idea to link to the property proposal from here (as done sometimes by ?Arthur when it's a more general question). The invitation to also comment here isn't mine either .. --- Jura 11:43, 13 September 2019 (UTC)
Echoing @Dominic:s comment about this being a standard field for archivists and archival materials, which is a great reason to support it. I could understand non-archivists confusing this with other description properties, but this happens with a lot of similar properties. What is the argument against non-item taking properties? As mentioned in the property proposal discussion, it seems in the spirit of quotation (P1683), inscription (P1684), and epigraph (literature) (P7150). Could we include a property constraint that requires an ID from an archive? Wskent (talk) 16:54, 16 September 2019 (UTC)
Lots of textual descriptions are standard in many fields. It's just that they aren't included in Wikidata. quotation (P1683) you mention is probably a good comparison, as it's generally used to support a statement (in the reference section) and actual quotes generally stored in Wikiquote and not at Wikidata. Similarly, the proposed free texts can be stored at Wikisource or Wikimedia Commons. These three are actually the type of WMF projects for that type of content. Have tried these projects yet? --- Jura 17:18, 16 September 2019 (UTC)

Qualifier for computer performance (P7256), for maximum and peak FLOPS


As the property has been recently created, I want to establish with a qualifier, the kind of performance named, as Maximum, and Peak. For example, Pleiades (Q2442043) as a max performance of 5.95~ petaflops, and a Peak performance of 7.1~ peraflops...

So, ¿which qualifier (if already existed) should be used? --Amitie 10g (talk) 14:24, 13 September 2019 (UTC)

I think I found it, object has role (P3831). --Amitie 10g (talk) 08:07, 14 September 2019 (UTC)
@Amitie 10g: An alternate qualifier for this case may be determination method (P459). ArthurPSmith (talk) 13:55, 16 September 2019 (UTC)


We all use diffs; so please add labels, aliases and descriptions in your language(s), to diff (Q66371350), which currently only has English. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:58, 5 September 2019 (UTC)

Thank you for informing about this. Esteban16 (talk) 17:28, 5 September 2019 (UTC)
Forgive me for reactivating this thread after it had been archived, but I meant to ask:
Should diff (Q66371350) really be described as specifically referring to as being a "difference page on a Mediawiki wiki"?
"diff" in the sense of "a concise, mechanically-derived delta between two texts" has a much wider usage. Consider:
  • w:en:Diff (which is primarily about the program) also notes that "The output is called a "diff", or a patch, since the output can be applied with the Unix program patch."
  • The Jargon File entry on diff gives sense 1 as "A change listing, especially giving differences between (and additions to) source code or documents".
  • Wiktionary "diff" sense 3: "The output of a diff program. A diff file."
Since Q66371350 is a Wikidata-only entity without any sitelinks, it's hard to be sure, but it seems to me it should probably refer to the more generic sense. (Unless, I suppose, the more-generic sense is adequately captured by patch (Q208218). If so, then Q66371350 is either a dup of Q208218, or perhaps an instance or subclass of Q208218. But Q208218, as written, seems to be pretty narrowly targeted at patches on computer programs. So perhaps both Q66371350 [if it remains specific to Mediawiki] and Q208218 are instances or subclasses of a hypothetical more-generic superentity.) —Scs (talk) 12:47, 15 September 2019 (UTC)
The item is not about the more general sense; it is about the MediaWiki sense. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:19, 15 September 2019 (UTC)
I can't say the wikimedia usage ("link to a special page that shows the difference between two versions of a document") is much different from any other diff in other system i.e. github or even Google docs. Circeus (talk) 01:55, 16 September 2019 (UTC)
@Pigsonthewing: A "diff" like this one? 622 017 074 (Hej!) 08:02, 16 September 2019 (UTC)
Yes. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:42, 16 September 2019 (UTC)
And yet it clearly is. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:42, 16 September 2019 (UTC)

Wikidata weekly summary #382

Theresa May's Resignation Honours

How should we, for example, best link to Theresa May (Q264766) from 2019 Prime Minister's Resignation Honours (Q67206887)? Author? Organiser? Commemorates? Named after? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:35, 10 September 2019 (UTC)

How helpful would it be to suggest an intermediate item for the event: <<2019 Prime Minister's Resignation>> to which the honours item is linked by part of (P361) / has part (P527) ... leaves you the same problem for the event <-> May link, but maybe main subject (P921). --Tagishsimon (talk) 13:55, 10 September 2019 (UTC)
@Tagishsimon: Beltedly, thank you. I'm not sure that's really doing much more than pushing the problem down the road, so to speak. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:24, 16 September 2019 (UTC)

Improving Terminologia Anatomica (Q286567)

So I have done some changes to attempt to improve Terminologia Anatomica (Q286567).

But I think I need some guidance (asked similar questions here)

I Think what makes most sense is to have the following items

The rationale for using Terminologia Anatomica (Q286567) to reference to the instance of (P31) version, edition, or translation (Q3331189) publication date (P577) 1998 instead of using it for "Terminologia Anatomica" instance of (P31) book (Q571) is so that it is more appropriate for citation as it has been used so far.

Please can I get some input here.

Iwan.Aucamp (talk) 13:26, 14 September 2019 (UTC)

Morrigan68 (talk) 17:09, 7 March 2021 (UTC) Aubrey
Viswaprabha (talk)
Maximilianklein (talk)
Jane023 (talk) 08:21, 30 May 2013 (UTC)
Alexander Doria (talk)
Ruud 23:15, 24 June 2013 (UTC)
Jayanta Nath
Yann (talk)
John Vandenberg (talk) 09:14, 30 November 2013 (UTC)
Danmichaelo (talk) 19:30, 16 February 2014 (UTC)
Ravi (talk)
Mvolz (talk) 08:21, 20 July 2014 (UTC)
Hsarrazin (talk) 07:56, 9 August 2014 (UTC)
PKM (talk) 19:58, 10 October 2014 (UTC)
Revi 16:54, 29 November 2014 (UTC)
Giftzwerg 88 (talk) 23:36, 1 January 2015 (UTC)
Almondega (talk) 00:17, 5 August 2015 (UTC)
Jura to help sort out issues with other projects
Skim (talk) 13:52, 24 June 2016 (UTC)
Marchitelli (talk) 12:29, 5 August 2016 (UTC)
Alexmar983 (talk) 23:53, 28 August 2016 (UTC)
Finn Årup Nielsen (fnielsen) (talk) 10:44, 29 August 2016 (UTC)
Chiara (talk) 14:15, 29 August 2016 (UTC)
Thibaut120094 (talk) 20:31, 14 September 2016 (UTC)
Ivanhercaz | Discusión Plume pen w.png 15:30, 31 October 2016 (UTC)
YULdigitalpreservation (talk) 17:35, 10 November 2016 (UTC)
PatHadley (talk) 21:51, 15 December 2016 (UTC)
Erica (ohmyerica) (talk) 19:26, 1 January 2017 (UTC)
Mauricio V. Genta (talk) 05:38, 12 March 2017 (UTC)
Sam Wilson 09:24, 24 May 2017 (UTC)
Sic19 (talk) 22:25, 12 July 2017 (UTC)
MartinPoulter (talk) 09:21, 20 July 2017 (UTC)
ThelmadatterThelmadatter (talk) 01:11, 13 September 2017 (UTC)
Zeroth (talk) 15:01, 16 September 2017 (UTC)
Beat Estermann (talk) 20:07, 12 November 2017 (UTC)
Shilonite - specialize in cataloging Jewish & Hebrew books
Elena moz
Oa01 (talk) 10:52, 3 February 2018 (UTC)
Maria zaos (talk) 11:39, 25 March 2018 (UTC)
Wikidelo (talk) 13:07, 15 April 2018 (UTC)
Mfchris84 (talk) 10:08, 27 April 2018 (UTC)
Mlemusrojas (talk) 3:36, 30 April 2018 (UTC)
salgo60 Salgo60 (talk) 12:42, 8 May 2018 (UTC)
Dick Bos (talk) 14:35, 16 May 2018 (UTC)
Marco Chemello (BEIC) (talk) 07:26, 30 May 2018 (UTC)
 徵國單  (討論 🀄) (方孔錢 💴) 14:35, 20 July 2018 (UTC)
Alicia Fagerving (WMSE)
Louize5 (talk) 20:05, 11 September 2018 (UTC)
Viztor (talk) 05:48, 6 November 2018 (UTC)
RaymondYee (talk) 21:12, 29 November 2018 (UTC)
Merrilee (talk) 22:14, 29 November 2018 (UTC)
Kcoyle (talk) 22:17, 29 November 2018 (UTC)
JohnMarkOckerbloom (talk) 22:58, 29 November 2018 (UTC)
Tris T7 TT me
Helmoony (talk) 19:49, 8 December 2018 (UTC)
Shooke (talk) 19:17, 12 January 2019 (UTC)
DarwIn (talk) 14:58, 14 January 2019 (UTC)
I am Davidzdh. 16:08, 18 February 2019 (UTC)
Juandev (talk) 10:03, 27 February 2019 (UTC)
Buccalon (talk) 15:51, 27 March 2019 (UTC)
MJLTalk 16:48, 8 April 2019 (UTC)
Rosiestep (talk) 20:26, 24 April 2019 (UTC)
Dcflyer (talk) 12:23, 7 May 2019 (UTC)
Susanna Giaccai (talk) 05:56, 29 July 2019 (UTC)
Asaf Bartov (talk) 19:03, 31 July 2019 (UTC)
Msuicat (talk) 17:58, 6 August 2019 (UTC)
SilentSpike (talk) 15:27, 12 August 2019 (UTC)
TheFireBender (talk) 12:40, 20 August 2019 (UTC)
Jumtist (talk) 21:45, 22 October 2019 (UTC)
DrLibraryCat (talk) 18:25, 25 November 2019 (UTC)
ShawnMichael100 (talk) 20:04, 25 November 2019 (UTC)
Lmbarrier (talk) 19:47, 2 December 2019 (UTC)
Satpal Dandiwal (talk) 17:32, 16 December 2019 (UTC)
Rosiestep (talk) 17:08, 14 February 2020 (UTC)
Clifford Anderson (talk) 01:37, 1 April 2020 (UTC)
Discostu (talk) 09:02, 9 April 2020 (UTC)
Subodh (talk)
Iwan.Aucamp (talk) 14:02, 27 April 2020 (UTC)
Алексей Скрипник (talk) 15:31, 4 May 2020 (UTC)
MLeonStewart (talk) 18:04, 11 May 2020 (UTC)
ArielBritoJiménez (talk) 16:17, 31 May 2020 (UTC)
DanielleJWiki (talk) 16:16, 8 June 2020 (UTC)
Ninovolador (talk)
Alex (talk) 06:05, 3 August 2020 (UTC)
Alex_Q (talk) 11:11, 18 September 2020 (UTC)
See the bright light (talk)
Alessandra Boccone (talk) 11:18, 6 November 2020 (UTC)
Uomovariabile (talk) 09:54, 13 November 2020 (UTC)
Pru.mitchell (talk) 08:11, 17 November 2020 (UTC)
Carlobia (talk) 13:34, 26 November 2020 (UTC)
Mathieu Kappler (talk) 11:31, 12 December 2020 (UTC)
Pierre Tribhou (talk) 19:19, 28 December 2020 (UTC)
Alessandra.Moi (talk) 16:54, 20 February 2021 (UTC)
Kind data (talk) 18:09, 23 February 2021 (UTC)
Morrigan68 (talk) 17:11, 7 March 2021 (UTC)
ChristianBRoy (talk) 15:06, 27 April 2021 (UTC)
Antoine2711 (talk) 15:08, 27 April 2021 (UTC)
Pictogram voting comment.svg Notified participants of WikiProject Books

For your attention (hope this is acceptable use if ping project) Iwan.Aucamp (talk) 14:14, 14 September 2019 (UTC)
I believe your suggestion of instance of (P31) book (Q571) should actually be instance of (P31) written work (Q47461344) unless the item represents a specific physical book somewhere --SilentSpike (talk) 14:29, 14 September 2019 (UTC)
Noted, and thanks for the input, I do not have an objection to this. Iwan.Aucamp (talk)
That's excellent. You may be interested in Wikidata:WikiProject Books where the de facto ontology for books & editions is set out. --Tagishsimon (talk) 17:10, 14 September 2019 (UTC)
@Tagishsimon: @SilentSpike: Thank you for the link, I had a look at the ontology at Wikidata:WikiProject Books. Would you have a good example of a similar case to mine? I looked at Mister President (Q3050361) but I feel it is not that similar to my case. Iwan.Aucamp (talk) 18:36, 14 September 2019 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I don't have a good example unfortunately, but the rest of your suggestion seems appropriate to me. My understanding is that basically the written work item is the text as a concept and each edition is a manifestation of that (so because there are different published dates it's correct to call those different edition items). The website item is probable what I'm most unsure about, because it could also be considered an edition of the written work. --SilentSpike (talk) 20:38, 14 September 2019 (UTC)

pinging Anatomy project for their input, specifically on if it is right to assume Terminologia Anatomica (Q286567) refers to "Terminologia Anatomica, 1998 Edition" instance of (P31) version, edition, or translation (Q3331189) publication date (P577) 1998 edition or translation of (P629) NQ000 ISBN-13 (P212) 9783131143617 Iwan.Aucamp (talk) 20:50, 14 September 2019 (UTC)

ChristianKl (talk) 14:41, 8 July 2016 (UTC) Iwan.Aucamp (talk) 15:13, 22 March 2020 (UTC) Was a bee (talk) 14:48, 23 September 2017 (UTC) Okkn (talk) 02:20, 25 October 2017 (UTC) JS (talk) Pictogram voting comment.svg Notified participants of WikiProject Anatomy

I don't have strong opinion about this or deep knowledge about exact differences of each books (basically listed items would be almost the same). There is "Terminologia Anatomica second edition" TA2 ID (P7173). This is very different thing from Terminologia Anatomica 98 ID (P1323). It is like "The Matrix Reloaded" is different from "The Matrix", or "Terminator 2" is different from "The Terminator". What you listed is something like "The Matrix" and "The Matrix DVD version" and "The Matrix Blue-ray version" and "The Matrix Netflix version". So I don't think those edition items are actually useful/need or notabile. at least from anatomy project scope. From books or scientific article point of view, possibly those can have meanings. I don't have opinion or knowledge for that aspects. --Was a bee (talk) 03:08, 15 September 2019 (UTC)


The following items now exists. For authoritative citation Terminologia Anatomica (Q286567) should be used (as per this description of the on line version). This is fine though as it is what was cited as the rest of the items are newly created.

Any critique or improvements are welcome. Iwan.Aucamp (talk) 21:13, 15 September 2019 (UTC)

@Jura1: thanks, I will update accordingly a bit later. Iwan.Aucamp (talk) 05:53, 16 September 2019 (UTC)
@Iwan.Aucamp: Have s look at that list of items describing the different editions of CRC Handbook of Chemistry and Physics. Snipre (talk) 07:38, 16 September 2019 (UTC)
Thank you for editing. By the way, it seems indented list above doesn't reflect the actual relations between each items. Here is basic history of international standard of anatomical nomenclature (w:Nomina_Anatomica#Editions). Honestly speaking, how deep will this go..? There are also many different language versions of this nomenclature. Including all of that? If so, for what...? Sorry, I know there are many visions and scopes. Simply I'm wondering. --Was a bee (talk) 08:06, 16 September 2019 (UTC)


@Jura1: @Snipre: I looked at some other books:

Following this, I think the most common practice is to put edition in parentheses. So I updated the labels as follow:

@Was a bee: I don't plan to make more, I just wanted to put the online edition and print edition and physical copies in the right context, you can see some of the relations above but there are more and if you think there should be more I would be glad to add them - not really sure what others there should be. Iwan.Aucamp (talk) 19:56, 16 September 2019 (UTC)

@Iwan.Aucamp: The current model to describe book in WD is the following: an item about a book can be a written work (Q47461344) or a version, edition, or translation (Q3331189) or a individual copy of a book (Q53731850). Use has edition or translation (P747) to link work item with its editions and edition or translation of (P629) to link an edition item with its work item. Use exemplar of (P1574) to link an exemplar item to its edition item. In principle, reprints of an edition shouldn't have a dedicated item unless a difference exits between the different reprints. Snipre (talk) 20:57, 16 September 2019 (UTC)
So Terminologia Anatomica (1998 edition) (Q67580232) and Terminologia Anatomica (1998 edition) (Q67510334) should not exists and the merge with Terminologia Anatomica (1998 edition) (Q67578843) is recommended unless there is a difference in the text version.
Terminologia Anatomica 1998 on-line version (Q67603718) should be treated as a particuler exemplar of the edition.
described by source (P1343) should not be used to link the different editions/work items. Snipre (talk) 21:08, 16 September 2019 (UTC)
@Snipre: The online edition has differences to the print version so is not exactly an exemplar of Terminologia Anatomica (Q286567) and the about page explains it is not authoritative and that Terminologia Anatomica (Q286567) is all that is authoritative. And the standard Terminologia Anatomica (1998 edition) (Q67578843) is not quite the same as the book that describes the standard in my view, and the distinction here is important specifically because of Terminologia Anatomica 1998 on-line version (Q67603718). I think merging Terminologia Anatomica (1998 edition) (Q67580232) and Terminologia Anatomica (1998 edition) (Q67510334) into Terminologia Anatomica (Q286567) makes sense maybe, I created those because I wanted a place to put ISBNs, which if they were merged I don't know where to put as there cannot be multiple ISBNs on one item without generating a warning, Iwan.Aucamp (talk) 22:09, 16 September 2019 (UTC)
@Iwan.Aucamp: If your reprints have a different ISBN, this usually means that some major changes were done, at least in the format and perhaps in the content. A simple reprint of the same edition doesn't require a new ISBN and as you have to purchase the ISBN numbers I don't think that the publishers paied it for nothing. It is more a new issue of an edition. We don't have so many cases until know in WD so we should start to think how we can handle that case. Snipre (talk) 02:21, 17 September 2019 (UTC)

How do I add audio pronounciation to Lexemes?

I would want to add one or more pronounciation audio for Lexeme L:L8885 based on the collection of pronunciation audio available at hello but was confused where to add it to the Lexeme item. Does anybody have any Lexemes with existing audio pronuncation audio already added to the item?(then I can just do by their example) Ελλίντερεστ (talk) 16:00, 16 September 2019 (UTC)

@Ελλίντερεστ: There's baguette (L7347) for example. Just add a pronunciation audio (P443) statement to the form being pronounced, pointing to the audio file. ArthurPSmith (talk) 19:38, 16 September 2019 (UTC)

Documenting a calculated birthdate

Given an obituary that says a person died age 103 on June 11, 1994, what is the correct way to reference a birthdate of ca. 1891? "heuristic" = "age at death" seems logical but is not allowed by the constraints. - PKM (talk) 21:19, 15 September 2019 (UTC)

Did they die in the United States? You can get a more accurate date from the Social Security Death index. You can also register for a free Familysearch account and add the person to Familysearch and it will look for records pertaining to that person automatically, no matter what country they are from. You can then link that entry you created to their Wikidata profile with Familysearch_ID. --RAN (talk) 19:13, 16 September 2019 (UTC)
I used "stated age at event" as a qualifer on "date of death" but how to I qualify the calculated "date of birth"? - PKM (talk) 20:47, 16 September 2019 (UTC)
How about Property_talk:P569#Qualifier_determination_method_(P459)_use_with_date_of_birth_(P569) ? --- Jura 20:50, 16 September 2019 (UTC)
@Jura1: determination method (P459) is what I needed, thanks! - PKM (talk) 19:31, 17 September 2019 (UTC)

Disambugation page item

How to list items from Wikipedia disambugation page in Wikidata disambugation item? I guess they should be devided into groups like "cities", "people", "music", "albums" etc. Eurohunter (talk) 21:56, 15 September 2019 (UTC)

On a disambiguation page, anything could be listed, and often that is happening. That makes it quite impossible to specify anything else then disambiguation. Edoderoo (talk) 05:50, 16 September 2019 (UTC)
Given that the items on an indvidual page are language dependent we don't list them on the Wikidata item that language independent. ChristianKl❫ 14:41, 16 September 2019 (UTC)
@Edoderoo: @ChristianKl: For example. How to include en:Masterpiece (disambiguation) items in Wikidata and make connections beetwen them. It would contain all Wikidata content so probably more things called "Masterpiece" than in ENwiki. My other idea I got is to make disambugation pages for certain categories like songs and include all songs named "Masterpiece", then item for albums etc. Eurohunter (talk) 20:01, 16 September 2019 (UTC)
Are you asking "how do I make a wikidata equivalent of a disambiguation item for all items with the title [title]"? Because we just plain do NOT do that on wikidata. Circeus (talk) 22:30, 16 September 2019 (UTC)
The answer is still "you don't". If someone wants all items with a given name, they can put that name into the searchbar, there's no need for a curated list. ChristianKl❫ 07:33, 17 September 2019 (UTC)

Please help with query

for example:

persons with their birth name, family name, occupation, gender, country of citizenship, name in native language so it will be ?person wdt:p1477 ?familyName wdt:p734 etc

but do I still have to add the Q - value?  – The preceding unsigned comment was added by BloukopKoggelmander (talk • contribs) at 18:10, 16 September 2019‎ (UTC).

I guess you don't look for a specific value. Then it suffices to give a variable like ?x or ?dummy. --SCIdude (talk) 04:44, 17 September 2019 (UTC)
@BloukopKoggelmander: For query help, you might do better at WD:RAQ. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:17, 17 September 2019 (UTC)

[Resolved][pywikibot] Error when adding a numeric value with unit

Update: I've resolved the issue by providing '' + statement. --Amitie 10g (talk) 23:21, 22 September 2019 (UTC)


I'm creating a semi-automated tool in Python, using pywikibot, to import data from TOP500 database.

It ran fine, until the script attemps to commit a numeric value with a qualifier, namely Pleiades (Q2442043) power consumed (P2791) 4407.00 kW. The script parses the numeric value with the unit, then, the unit is converted to the given statement (kW kilowatt (Q3320608)). Following is the error message:

WARNING: API error modification-failed: Illegal value:
Traceback (most recent call last):
  File "/mnt/c/Users/Davod/Documents/git/wikidata_top500/core/pywikibot/", line 112, in handle
    func(self, *args, **kwargs)
  File "/mnt/c/Users/Davod/Documents/git/wikidata_top500/core/pywikibot/", line 4325, in addClaim
    self.repo.addClaim(self, claim, bot=bot, **kwargs)
  File "/mnt/c/Users/Davod/Documents/git/wikidata_top500/core/pywikibot/", line 1326, in callee
    return fn(self, *args, **kwargs)
  File "/mnt/c/Users/Davod/Documents/git/wikidata_top500/core/pywikibot/", line 7946, in addClaim
    data = req.submit()
  File "/mnt/c/Users/Davod/Documents/git/wikidata_top500/core/pywikibot/data/", line 2080, in submit
    raise APIError(**result['error']) modification-failed: Illegal value: [messages:[{'name': 'wikibase-validator-not-allowed', 'parameters': [''], 'html': {'*': 'Valor ilegal: <a class="external free" href=""></a>'}}]; help:See for API usage. Subscribe to the mediawiki-api-announce mailing list at &lt;; for notice of API deprecations and breaking changes.]

Script Source code. Relevant code is at the addClaim() function, line 109.

Any hints? Thanks. --Amitie 10g (talk) 04:49, 21 September 2019 (UTC)

I am not exactly sure what str2statement(statement) in your script does. However, please make sure that the unit needs to be given as a dereferenceable Wikidata concept URI in string form. In your example it should thus roughly look like this: claim.setTarget(pywikibot.WbQuantity(amount=4407.00, unit='', site=site)) (with "http" rather than "https", and "entity" rather than "wiki" in the dereferenceable concept URI). Mind that I also suggest using named parameters "amount" and "unit" here ("lowerBound" and "upperBound" are also available for Quantity datatype claims). In case of quantities without unit, please either omit the "unit" parameter, or set it to None. —MisterSynergy (talk) 07:42, 21 September 2019 (UTC)
@MisterSynergy: I substantially updated the code (and updated the Git repo), and fixed some issues. However, I haven't tested them on live Wikidata yet (I'm too sleepy) (I'm working on Pleiades (Q2442043)); I'll test them soon.
Regarding str2statement() and str2ptop(), those functions parse an arbitrary string and retrieve the equivalent statement (QXXX) and property (PXXX) (list of statements/properties are found at "" file). --Amitie 10g (talk) 05:01, 22 September 2019 (UTC)
  • I see (re.
  • If I remember correctly, you also need to convert the value you set for "amount" to a numerical format in Python. Probably like amount = float(value[0]) in line 179.
  • To test scripts, please use either a fresh item on (by initializing site = pywikibot.Site(code='test', fam='wikidata'), or edit Wikidata sandbox (Q4115189) in Wikidata directly. Do not test-edit other items, please.
MisterSynergy (talk) 10:22, 22 September 2019 (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) 15:17, 23 September 2019 (UTC)


I seem to be perpetually fighting bots on Paul Marchal (Q58453385). I'm getting rather bored of doing so. Is there a tool to quickly delink all the items to which he has wrongly been added as the author? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:27, 22 September 2019 (UTC)

The VIAF entry looks like it conflated a modern researcher and an historic one. --RAN (talk) 20:58, 22 September 2019 (UTC)
@Pigsonthewing: You can use the author disambiguator tool to remove or move all or selected P50 links a person has - in this case go to this page to make the changes. ArthurPSmith (talk) 16:40, 23 September 2019 (UTC)
Thank you. That's just the tool I envisaged - in fact, I knew the tool, but only in the reverse direction. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:25, 23 September 2019 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:25, 23 September 2019 (UTC)

Siblings born days apart

I just posted some queries about twin-related data, but while those have some potential for cleanup (feel free to go through them and add sibling (P3373) / kinship to subject (P1039) twins (Q14756018) qualifiers), something else came up that I think is worthy of project chat attention: there are pairs of siblings who were born several days apart.

I suppose some of these could be adopted siblings, or half-siblings. But still, those results don’t seem entirely plausible to me… what do you think? --TweetsFactsAndQueries (talk) 16:20, 14 September 2019 (UTC)

Interesting idea!
¿some date precision or calendar model snafu? --- Jura 16:26, 14 September 2019 (UTC)
@TweetsFactsAndQueries: I've fixed a few; but there are at least two false positives: Manuel Beingolea (Q5574906) was born on 7 January 1881; but for his brother Antonio Beingolea Balarezo (Q14568096) we show only "1881", which is plausible. The case of Salem bin Laden (Q1139461) and Bakr bin Laden (Q428820) is similar; can you exclude year-only dates? Also, could you perhaps combine your queries, and show all dates which are, say, six months or less apart? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:48, 14 September 2019 (UTC)
@Pigsonthewing: Here's a histogram of day-precision sibling separation [9] (excluding same-day twins because they go off the scale). --99of9 (talk) 07:39, 16 September 2019 (UTC)
That's interesting, but not whet I meant - almost all of the results are valid. Only those of under, say, 300 days relate to the issue discussed here. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:50, 16 September 2019 (UTC)
@Pigsonthewing: Ok, if you're checking individual cases, here's a query. --99of9 (talk) 05:44, 17 September 2019 (UTC)
Thank you, but "Query timeout limit reached" :-( Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:55, 17 September 2019 (UTC)
@Pigsonthewing: Try again every now and then, it has worked 2/2 for me, but often comes close to 60s. --99of9 (talk) 05:28, 18 September 2019 (UTC)
BTW, your conjecture is correct: Louis César, Count of Vexin (Q3839486) is the illegitimate half-brother of Louis François, Duke of Anjou (Q3839656). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:53, 14 September 2019 (UTC)

Lexemes for names

Are we creating lexemes for 1) given and 2) family names? If so, what are good examples? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:30, 17 September 2019 (UTC)

It's definitely a better idea than creating items for given and family names. See Paulina (L41783).--Shlomo (talk) 06:28, 18 September 2019 (UTC)
  • So far I've been avoiding proper-noun lexemes. One issue is that lexemes are language-specific, so for a name used in many languages you'd have many lexemes (one per language). And of course the number of them is pretty much unbounded, while there is at least a reasonable limit for common nouns and other parts of speech. However, names can have etymologies and can have differing forms (in English that's pretty much limited to singular vs plural, unless you want to include the possessive forms too) so I guess there's some reason to add lexicographic data particularly where those are unique or unusual. ArthurPSmith (talk) 18:11, 18 September 2019 (UTC)

ar:مستخدم:صندوق مستخدم

Hello.This is a user page that has 2٬204 subpages (index) and mostly linked to elements and this are against policy.Please remove their links.Thanks David (talk) 08:23, 18 September 2019 (UTC)

  • What are they? I can't find one that is linked from Wikidata. Some seem to be userboxes like Template:User wikipedia/Bureaucrat (Q6303902). For historic reasons some of these are in username-namespace, not template namespace. --- Jura 10:08, 18 September 2019 (UTC)
    It's an account created so that userboxes can be in user namespace without belonging to one user. en:User:UBX exists for the same purpose, and some of its pages are attached to items such as User:UBX/The Prisoner 2 (Q22706846), which is also linked to the French equivalent in template namespace. Peter James (talk) 10:55, 18 September 2019 (UTC)

When to add brands

User:SCIdude, about this change, is that the recommended way to record brands? That item is a generic class (multiple different types of vaccines against one infection, rather like sedan (Q190578) for automobiles). I don't have a good understanding of how brands work on Wikidata. WhatamIdoing (talk) 14:48, 18 September 2019 (UTC)

Better probably active ingredient in (P3780), and instance of trade name (Q1417728), what do you say? However Coca-Cola (Q2813) inst-of brand is then wrong, should be trade name? --SCIdude (talk) 15:11, 18 September 2019 (UTC)
I see the other issue may be wrong modeling. Instead of brand or TM the reverse of instance-of would be correct, but we don't have that. So just remove from the general item? --SCIdude (talk) 15:22, 18 September 2019 (UTC)
Maybe? I'm still figuring out how to connect items.
The real-world relationship is that there are multiple anti-shingles vaccines, and in practice, all of them are proprietary, name-brand vaccines, which aren't exactly interchangeable. Even if you set out to make as exact a copy as possible, the virus strain you ended up growing would probably be slightly different.
You are going too far. Variants of proteins are usually ignored in WD (and elsewhere), e.g. Reelin (Q13561329) is an instance of a protein, not of a group of variants. --SCIdude (talk) 05:57, 19 September 2019 (UTC)
There's "all the kinds of shingles vaccines" as a group, and then there are the individual vaccines, such as Zostavax. On Wikidata, Q8074572 is "all the kinds of shingles vaccines". Q29006757 is "Zostavax" specifically. Theoretically, there could be an intermediate layer – "any kind of shingles vaccine" > "any live attenuated shingles vaccine" > "Zostavax". But in practice, AFAICT, Zostavax is the only live attenuated shingles vaccine that exists, so we're skipping that step. WhatamIdoing (talk) 00:42, 19 September 2019 (UTC)
Yes, and so Zostavax is an instance of a zoster vaccine. That intermediate class wouldn't even be necessary with multiple attenuated shingles vaccines. --SCIdude (talk) 05:57, 19 September 2019 (UTC)

Discussion and image properties

Hey, I am working on periodicals and I wonder whether there is a property to indicate that the web page or book has images?

Then some online texts has comments and some not. Is there a way how to indicate it here on Wikidata? --Juandev (talk) 11:55, 19 September 2019 (UTC)

WikidataCon Award 2019: help us nominating the best projects and tools

Hello all,

The Wikidata and Wikibase ecosystem is a huge galaxy of exciting content, tools, projects, powered by the communities as well as organisations working with the software and the data. For seven years, people are gathering, starting projects, developing tools, improving the editors' workflows, filling various gaps, working all together to give more people more access to more knowledge.

The WikidataCon Award 2019 will reward the greatest projects in a variety of categories. The awarded projects are selected by the WikidataCon Award committee, based on nominations by the community. Everyone is invited to participate and nominate one or several projects. The nomination process is open until October 7th.

What we mean by project: a project can be a community gathering or other initiative that led to great results (WikiProject, event, editathon…), a tool (gadget, script, external tool…) or any other action that led to improving Wikidata’s data, the workflow of its editors or the outreach.

To read more about the award and participate, please check out this page. Thanks! Lea Lacroix (WMDE) (talk) 12:46, 19 September 2019 (UTC)

Property-valued wikibase:quantityUnit statements

The following query lists properties that appear in the unit position of wikibase:quantityUnit statements:

select distinct * where {
  [] wikibase:quantityUnit ?unit .
  ?unit a wikibase:Property .

Try it!

To my surprise the result is not empty. To narrow down the issue one can run the following query to find item-property combinations that are affected:

select ?item ?prop ?amount where {
    wikibase:propertyType wikibase:Quantity ;
    wikibase:claim ?p ;
    wikibase:statementValue ?psv .
  ?item ?p [
    ?psv [
      wikibase:quantityUnit [
        a wikibase:Property ;
      ] ;
      wikibase:quantityAmount ?amount ;
    ] ;
  ] .
} limit 100

Try it!

However, looking at some item pages I can't find where those properties have been entered as unit and how to fix it. Toni 001 (talk) 12:02, 5 September 2019 (UTC)

Yup, looks like fallout from phabricator:T230588 – the reported unit for some of those is P199, which is supposed to be 1 (Q199) (default unit for RDF). --Lucas Werkmeister (WMDE) (talk) 12:58, 5 September 2019 (UTC)
@Lucas Werkmeister (WMDE): It seems happen here too.--- Jura 22:57, 8 September 2019 (UTC)
They also appear as objects of wikibase:geoGlobe Toni 001 (talk) 11:25, 9 September 2019 (UTC)
@Lucas Werkmeister (WMDE): the red stuff at are probably some P used as reference values. --- Jura 22:10, 14 September 2019 (UTC)
@GLederrey (WMF): you seem to be the only user that could do it listed on the page [10] mentioned by @Lucas Werkmeister (WMDE): in phab:T230588. If the import now works correctly, maybe a simple deletion is sufficient (correct triples are available), compare e.g. [11] with [12]. Is delete on Blazegraph DELETE ? --- Jura 11:23, 20 September 2019 (UTC)

Consider math symbols lexemes

Arthur Rubin
Nomen ad hoc
The Anome
Daniel Mietchen
John Samuel
Jeremy Dover
Pictogram voting comment.svg Notified participants of WikiProject Mathematics

Math is a language, and math symbols (science symbols in general) behaves a lot like lexemes:

  • the same symbol can mean a lot of different things, each one would be considered a different sense;
  • some properties used to describe mathematical symbols would be more appropriate as properties for lexemes (for example, Unicode character (P487) and TeX string (P1993)).

Furthermore, it could help to add semantics to mathematical formulas. I don't know if it viable but I think to something like the formulas in excel: selecting the right symbol senses from a dropdown menu in prefix notation. For example, to write the equation "a=b+c" one types "=" and it is presented with all the senses of the "=" symbol; he select the most appropriate one, then Wikidata presents the syntax of the "=" operator, i.e. two operands; the user inserts "a" and it selects the sense "variable", then it inserts "+" and selects the sense "addition"; Wikidata shows the syntax of addition: two addends; the user writes "b" and selects variable and it does the same with "c". I hope I was clear.--Lore.mazza51 (talk) 04:23, 14 September 2019 (UTC)

To annotate formulae with their semantics in Wikipedia articles, consider using the AnnoMathTeX system: The semantically annotated formulae will be stored the system's database and subsequently seeded into Wikidata. See also: [13].--PhilMINT (talk) 10:10, 14 September 2019 (UTC)
@PhilMINT: Thank you for the suggestion, but annomathtex seems to work only with Wikipedia articles, not Wikidata.--Lore.mazza51 (talk) 06:29, 19 September 2019 (UTC)
@Lore.mazza51: Yes, it can be used to employ Wikipedia articles to get semantically annotated formulae for Wikidata. In the future, it will be possible to assign Wikidata QIDs to formula entities.--PhilMINT (talk) 11:42, 20 September 2019 (UTC)

Wrong data

Ukrainian Wikipedia (Q199698) have both official name (P1448) and title (P1476). Why? Please remove title (P1476), thanks!!! --2001:B07:6442:8903:C4AD:B849:2AF8:ED72 15:47, 20 September 2019 (UTC)

Best practices for branch of the military in human entries

What are the best practices for branch of the military in human entries. See for example Henry Crandall Newcomer (Q67907138) where I use the three names of the US Air Force as it changed from its inception to now. Or should we just use the modern name. We seem to be split on the issue looking at a few entries. We have all discussed this before with "citizenship" as the names, and governments, of countries changed, for example Germany. No clear decision was made for citizenship. --RAN (talk) 19:24, 19 September 2019 (UTC)

When you have an item like United States Air Force (Q11223), with an inception of 1947, it wouldn't make any sense to use it for earlier dates. So linking all three versions seems correct. I.e., they are treated as different organizations, not just a single organization with a changing name. Ghouston (talk) 01:16, 20 September 2019 (UTC)
It would be much easier if the error message told me what the correct entry should be, instead of telling me I have the wrong one. Can that be done? It knows the dates ... we just need to have "preceded by" and "replaced by" properly in place. --RAN (talk) 03:36, 20 September 2019 (UTC)

statement remove request

Please remove native label (P1705) from Ukrainian Wikipedia (Q199698), since already available as official name (P1448). Thanks!!! --2001:B07:6442:8903:2DC5:51C:4EDD:7465 13:55, 18 September 2019 (UTC)

statement remove request

Please remove native label (P1705) from Ukrainian Wikipedia (Q199698), since already available as official name (P1448). Thanks!!! -- 19:09, 19 September 2019 (UTC)

I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. SCIdude (talk) 13:19, 26 September 2019 (UTC)

Retrieving nth value using #invoke:WikidataIB

Is it possibly to retrieve an nth value from wikidata within a property. E.g. for a person to retrieve the first listed occupation separately to the second, the third etc? Example:

{{#invoke:WikidataIB| getQualifierValue| qid=Q15136093| P106| pval=Q3809586| qual=P108| name=occupation| fetchwikidata=ALL}}

returns me the whole list of employers when I use it on wikipedia, or just the first one if I use it here. T.Shafee(evo&evo) (talk) 02:03, 21 September 2019 (UTC)

It should be possible. Ask the module maintainers to implement it. --Matěj Suchánek (talk) 09:34, 22 September 2019 (UTC)
@RexxS: can answer better than I, but as I understand it, the order of the items in Wikidata is essentially random, and what's currently the third item might not be the third item in the future. The normal solution, assuming you want to see them all in the end, is to use the prefix/postfix parameters to adjust the returned code. Thanks. Mike Peel (talk) 10:08, 22 September 2019 (UTC)
Already answered on en:Module talk:WikidataIB #Retrieving nth value using #invoke:WikidataIB for English Wikipedia.
I think the WikidataIB module here is out-of-date so probably would be best to sync from Enwp or Commons; you might find the functions in en:Module:String2 worth having here as well. Cheers --RexxS (talk) 11:31, 22 September 2019 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. SCIdude (talk) 13:11, 26 September 2019 (UTC)

Are these chemical compounds identical ?

There are two chemical compounds with identical chemical formula: alpha,alpha-trehalose (Q421773) and trehalose (Q27109935). Are they identical that is candidates for merging ? Kpjas (talk) 12:14, 21 September 2019 (UTC)

They have different external identifiers. I think one is for a particular isomer, the "α,α form". Ghouston (talk) 12:28, 21 September 2019 (UTC)
Use ChEBI as guidance, it has the most detailed ontology. In particular this ChEBI item shows there are three trehaloses, and there is the abstract concept trehalose that contains all of them. Now alpha,alpha-trehalose (Q421773) is the alpha,alpha form and trehalose (Q27109935) is the generic form. The 2 missing items are alpha,beta-trehalose (Q27119787) and beta,beta-trehalose (Q27119788). I'll add the resp. instance/subclass relationships and descriptions. --SCIdude (talk) 13:57, 22 September 2019 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. SCIdude (talk) 13:11, 26 September 2019 (UTC)

deletion request

Please open North Macedonia national football team (Q188568) and delete native label (P1705) since already exist as official name (P1448). Thanks!!! --2001:B07:6442:8903:B821:B516:C633:97DE 16:24, 23 September 2019 (UTC)

I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. SCIdude (talk) 13:17, 26 September 2019 (UTC)

English labels for items with no name in English

What is the community preference for English labels for items that have no name in English?

For example, consider Sonya v tsarstve diva (Q64415956), an 1879 Russian translation of Alice in Wonderland. Should the English label be:

  • Соня в царстве дива (the book's title)?
  • Sonya v carstve diva (the transliteration of the book's title)?
  • Alice in Wonderland (the translation of the book's title, but which is not actually the book's title)?
  • no label?
  • something else?

Beleg Tâl (talk) 20:30, 18 September 2019 (UTC)

I think if there's an official title in a single language the label should be that. Literal translation in other languages can be added using literal translation (P2441). --SilentSpike (talk) 21:50, 18 September 2019 (UTC)
Beleg Tâl, I like the idea of the transliteration. WhatamIdoing (talk) 00:53, 19 September 2019 (UTC)
English Wikipedia uses romanization (not transliteration!!! see w:en:WP:RUS) in such cases. I suggest doing the same on Wikidata.--Ymblanter (talk) 10:56, 19 September 2019 (UTC)
IMO, it should be romanization/transliteration (let's not get overly technical), but "1879 Russian translation of ..." should be the description. Circeus (talk) 17:09, 19 September 2019 (UTC)
I've updated the item's label and description with the romanization as per the English wikipedia guidelines. Unsure how to also add the romanization using a property, but that seems like something that should also be done. --SilentSpike (talk) 22:03, 21 September 2019 (UTC)

Definition of Datatypes

Where can I find the definition of data types. And who can edit them. I want to add a Start time in a precision of minutes. And I wasnt able to choose it. The Special:ListDatatypes says that the precision minute is using the number 13 in the scheme like this +2013-01-01T00:00:00Z. Can someone please tell me if it is possible to add a date in that precision and where I can find the definition of data types. -- Hogü-456 (talk) 18:38, 21 September 2019 (UTC)

To my knowledge, mw:Wikibase/DataModel#Dates and times is the most useful description. Precisions higher than day (11) are not usable until now, and enabling that would not be straightforward at all. See also phab:T57755. —MisterSynergy (talk) 19:09, 21 September 2019 (UTC)

Want to be in contact

I am a first batch student when the college/Polytechnic was started in 1961. Our first Principal was D.K. Datar.

Any informatio you have?

Thanks  – The preceding unsigned comment was added by D.J.S. Singh Saggu (talk • contribs) at 14:20, 22 September 2019‎ (UTC).

  1. Not sure how this relates to this site, please explain.
  2. When you explain, please clarify: what college? - Jmabel (talk) 17:09, 22 September 2019 (UTC)

Best reference for statements generated by AI

Hi everyone,

Based on an ongoing discussion with MisterSynergy and Jura1, w'd like to gather more input here.

Any thoughts on what would be the best solution are super appreciated.

See also here for some digging on the confidence score.

Cheers --Hjfocs (talk) 14:52, 17 September 2019 (UTC)

  • Many data reusers would look at truthy values and do't see qualifers. When it comes to AI generated statements with low confidence, they shouldn't make it on Wikidata. Confidence scores might be valuable to distinguish 0.99 from 0.999 but they shouldn't lead to more statements being created.
On alternative that would allow for more statements being created is to have a forth rank that marks uncertainity that doesn't evaluate to being truthy but that's also distinct to being deprecated.
@Lydia Pintscher (WMDE): can you say something about how much work it would be to have another rank so that the ranks are "preferred/normal/uncertain/deprecated"? ChristianKl❫ 11:02, 18 September 2019 (UTC)
This is interesting project. Even human edit, or bot-assisted human edit, errors unavoidably happen. I suppose human (my own) error rate is roughly 1% or so. (I myself notice my own errors generally through constraint violation report.) So actual performance would be important. If it is 10% error rate, I think it should not be added. If error rate is 1% or less than that, but is even doubtful in some degree, then I hope that the data is added with "possible other candidate(s)". Possible other candidate(s) makes the data easier to fix. (P.S. I know nothing about how your AI actually works. So if my comment is pointless, or meaningless, please ignore me. Thanks) --Was a bee (talk) 13:00, 18 September 2019 (UTC)
What about saying determination method (P459) (e.g. --> automatic assertion (Q28955231) or a subclass of it). After all a specific algorithm is used and should be specified somehow. There are lots of "AI" statements on proteins, generated by bioinformatics centers (UniProt/GOA). They are distinguished from human statements by the value of the determination method (P459) qualifier, and even with human claims there are lots of grades, e.g. NAS (Q23190854) would be equivalent to handwaving (Q2861077). --SCIdude (talk) 15:48, 18 September 2019 (UTC)
Example class tree
automatic assertion (Q28955231)
+-- algorithm type ABC
    +--- specific algorithm XYZ
         +--- XYZ with 95% cutoff be used with determination method (P459). --SCIdude (talk) 16:03, 18 September 2019 (UTC)
  • @Hjfocs: A semaphoring that was suggested at Wikimania, as a temp fix while waiting for Christian's new fourth rank for unconfirmed statements, was that AI-inferred statements could be added with "deprecated" rank, and reason for deprecation (P2241) = "human confirmation required". A Wikidata game could then easily find these, present them to humans, and then accordingly promote them or delete them or change the reason for deprecation. But it would put them out in the open, so they would be shared, and other users could work with them.
I would support a new property for "projected probability according to model", that could either be a qualifier, or perhaps (better) would live in the references, so different models could have different recorded assessed probabilities for the same match. But it's worth noting that such projected probabilities are sometimes not at all well calibrated -- for example, with face-detection, I have seen algorithms sometimes consistently claiming 20-30 identities match with over 90% probability, when at most only one was ever correct. So numbers coming out of the box should be treated with some caution, at least until validated against a real-world test set. Jheald (talk) 13:53, 19 September 2019 (UTC)
  • Thanks everyone for your feedback so far!
    • @ChristianKl: I totally agree that low-confident statements should be completely discarded, and medium-confident ones would require an extra curation step via third-party tools, like Mix'n'match for identifiers (BTW, it's where Soweego bot puts them, see [14]).
    • The 4th rank idea sounds to me very nice and simple.
    • @Was a bee: I'm not sure I got your point about the doubtful data. With possible other candidate, are you referring to some exisiting property which could express that doubt?
    • @SCIdude: I think this looks like a resonable solution that can be already implemented, but I personally prefer the simplicity of the 4th rank suggested by ChristianKl.
    • @Jheald: your option seems like a trade-off between SCIdude's and ChristianKl's, I like it.
    • I agree on your concern regarding the confidence score calibration/real-world validity: it's a hard challenge to deal with. It would be very interesting to understand how/if it changes when an AI is re-trained over fresh data. This is something that needs to be carefully assessed in the next Soweego bot run, for instance.
--Hjfocs (talk) 15:51, 19 September 2019 (UTC)
@Hjfocs: My English is not good, so let me explain. I took an example from your talk page. For example, this edit[15] which is about Discogs artist ID (P1953) on ambiguous artist names.
In this case, what I wanted to say is the edit is like the following. Although ID is wrong, but correct ID appears in qualifier section.
Discogs artist ID
Normal rank 285540
Possible other candidate 391546
0 references
add reference

add value
In other words, putting second (and third) matching ID as qualifier (or data in reference section). By this, corrections by human become easier. As my personal experience which I added thousands of IDs in anatomy region (Terminologia Anatomica 98 ID (P1323)). Through that process, it was not rare that I felt difficulty on judging correct matching between WD-item and external-ID because there are many similar names and obscure concepts in both sides. So if there are second candidate information, I think it would be very useful. By the way, as far as I searched, I found candidate (P726), and possible creator (P1779). But there are no suitable property for this. So I think new property must be proposed and created, if do this. I hope my text is understandable... Thanks. --Was a bee (talk) 11:17, 20 September 2019 (UTC)
A comment on the reference statement given above as an example: I think the statement is not accurate, indeed in respect to stated in (P248) plainly wrong. Twitter does not state anything about the identity of Aaron Black (Q4661852) and The statement "this is the same person" stems from the algorithm involved and should be referenced accordingly. --Jneubert (talk) 13:11, 21 September 2019 (UTC)
  • @Was a bee: thanks for the detailed explanation, now I completely understand your point, and I think it's a brilliant idea. Not sure this would be consistent with the schema, though. @Lucas_Werkmeister_(WMDE): do you think this is feasible, or would it break the constraints check extension?
  • @Jneubert: that's exactly the reason why we opened this thread, based on Topic:V6cc1thgo09otfw5. See the top of this thread.

Hjfocs (talk) 15:10, 23 September 2019 (UTC)

@Hjfocs: I’m not sure I fully understand what you mean, but I don’t think the {{Statement+}} that Was a bee posted would break anything in the constraints system. Constraints like distinct-values constraint (Q21502410) might not fully work as you’d expect them (you probably wouldn’t put any such constraints on the “possible other candidate” property), but I don’t see a big problem with this. --Lucas Werkmeister (WMDE) (talk) 16:06, 23 September 2019 (UTC)

Gender version of the family name (P734)

What is the proper Wikidata practice/policy on the gender versions of family names ? I've been adding some Polish names and there are sometimes (not always) 2 forms like Kowalski (Q3199417) and Kowalska (Q37469735). Are Kowalski and Kowalska separate items by design or are they mutually gender inflection of surname (P5278) ?

The same issue applies to some other (Slavic) languages.

A complicating factor is that in the English speaking world usually the male form is used/preferred or the female form of the name is equated to the male form. For example a woman who is Maria Kowalska by birth, moves to the US and is known as Maria Kowalski.

A large proportion of items of Polish women in Wikidata doesn't have the family name (P734). It'd be good to add missing family names to thousands of items so it is important to me.

Not knowing the policy I have been avoiding creating new items for the female form but on the other hand it is to me utterly weird to assign the Kowalski (Q3199417) name to a woman unless there are other reasons like the one above. Kpjas (talk) 06:30, 19 September 2019 (UTC)

The existence of gender inflection of surname (P5278) suggests that the idea is to create items for both variants, then use that property to link them. If somebody uses both variants, in different places, then you'd add them both, perhaps with start and end dates if relevant. Ghouston (talk) 06:58, 19 September 2019 (UTC)
I also would create a separate family name item for each version with a secondary P31 = Q18972245 / Q18972207 and finally use gender inflection of surname (P5278) to link them together. Moebeus (talk) 15:57, 19 September 2019 (UTC)
In Lithuanian language family name also differ for married and unmarried women. --EugeneZelenko (talk) 14:05, 20 September 2019 (UTC)
This discussion is interesting for the names project. Klaas `Z4␟` V:  07:46, 23 September 2019 (UTC)

Wikidata weekly summary #383

Types of stairs

Hi, I am very busy with WLM so I have no time to dig into that but I feel that the stairs (Q12511) is too generic. Shouldn't we have a distinction between the structure inside a building (or outside but still as part of the building) and the flight of steps in a public road that exists on ts own and is not related to any precise building?

Maybe I have not looked properly into that because I have no left-over time but if two items should be created, this is the sort of things to set up correctly since the beginning avoiding any doubt and with the best possible definitions in all the languages. In italian the second one for example is more properly defined as scalinata (or gradinata). Some items such as Staircase of Santa María del Monte (Q3057824) and Spanish Steps (Q848072) will be described much more precisely. I have however the feeling that other languages do not make this distinction in such as strong way. Of course, in Italian we can use the generic word as well, scala, but this sounds much more precise.--Alexmar983 (talk) 15:58, 23 September 2019 (UTC)

I agree we need a subclass hierarchy for “stairs”. The Getty Art & Architecture vocabulary has a structure we might work with. I can’t take this on right now either, but if anyone is looking for a project to work on this would be a good one. - PKM (talk) 18:22, 23 September 2019 (UTC)
... what is the question or issue here aside from "items need to be created"? I mean, unless you're going to disrupt something like instance of (P31) human (Q5), I'm not sure how something like this, which would touch a fairly specialised corner of wikidata, requires a topic on the Project chat. Circeus (talk) 00:26, 24 September 2019 (UTC)

Adding source to items of metro stations?

We have plenty of items about metro stations which have a lot of statements either unsourced or sourced from Wikipedia. However, for this class of items we have a universal reliable source,, which is maintained up-to-date. It should not be difficult to extract from there and to reference properties P31, P17, and P131 by bot for every item. I have done that for some stations, see Q68272373 as an example, but it presumably needs to be done by bot and not manually. Is this at all feasible? Is there anybody knowledgeable who could help?--Ymblanter (talk) 19:11, 23 September 2019 (UTC)

Merge articles

Could someone complete the merge of Q12300651 into Q11959835? -- 12:11, 29 September 2019 (UTC)

Please have a look at Help:Merge. Cheers, Nomen ad hoc (talk) 12:33, 29 September 2019 (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. Nomen ad hoc (talk) 12:33, 29 September 2019 (UTC)

Personal news: a new role

Over the last few years, more and more research teams all around the world have started to use Wikidata. Wikidata is becoming a fundamental resource. That is also true for research at Google. One advantage of using Wikidata as a research resource is that it is available to everyone. Results can be reproduced and validated externally. Yay!

I had used my 20% time to support such teams. The requests became more frequent, and now I am moving to a new role in Google Research, akin to a Wikimedian in Residence: my role is to promote understanding of the Wikimedia projects within Google, work with Googlers to share more resources with the Wikimedia communities, and to facilitate the improvement of Wikimedia content by the Wikimedia communities, all with a strong focus on Wikidata.

One deeply satisfying thing for me is that the goals of my new role and the goals of the communities are so well aligned: it is really about improving the coverage and quality of the content, and about pushing the projects closer towards letting everyone share in the sum of all knowledge.

Expect to see more from me again - there are already a number of fun ideas in the pipeline, and I am looking forward to see them get out of the gates! I am looking forward to hearing your ideas and suggestions, and to continue contributing to the Wikimedia goals. --Denny (talk) 16:58, 19 September 2019 (UTC)

P.S.: Which also means, incidentally, that my 20% time is opening for new shenanigans. --Denny (talk) 16:58, 19 September 2019 (UTC)

  • When it comes to the proposal for an abstract Wikipedia, I believe that it might be good to start with a system for "abstract descriptions" for Wikidata entries. The complexity of Wikidata descriptions is a lot less then the complexity of full Wikipedia articles and it would be more easy to generate them automatically and build a group of people that use the tools for that purpose. ChristianKl❫ 16:11, 20 September 2019 (UTC)
    • I like that idea! There are bots and tools that do something like this now (generate automated descriptions from Wikidata fields) but making that work across many languages would be a really interesting capability! ArthurPSmith (talk) 17:20, 20 September 2019 (UTC)
    • Thanks! I also was thinking about it, and I think you are right. Sometimes I think that these short descriptions are harder than longer prose, but I think that is only true for a few corner cases - and in most, it should be pretty easy. So, yes, I agree. --Denny (talk) 05:23, 24 September 2019 (UTC)
  • We do currently have bots, but it seems to me that the template descriptions on Reasonator often look better then the bot created descriptions we currently have. With the bots it's hard for multiple people to collabortate on how the descriptions should look like. Descriptions are often out-of-date.
I would guess that 99% of the existing wikidata items can be created with templates and for those cases where people want to have a more natural language description for items that are more complicated they could fallback to the status quo.
If we reach a state where bots don't need to do edits to create descriptions that would reduce the edit load on Wikidata. We might have a separate server that caches descriptions that get auto-created based on the descriptions. ChristianKl❫ 11:24, 24 September 2019 (UTC)

Should we be using document images in the reference field

I just noticed this clever entry John Wayne Gacy (Q298252) where the image of the death certificate is used as a reference for a fact, should we be doing more like this? I think it would be best as a reference rather than as a qualifier as was done in the Gacy entry. Generally I just say the fact came for a birth or death certificate when we have the document at Commons, but maybe we should be displaying the document too like was done for Gacy. What do you think? --RAN (talk) 03:32, 20 September 2019 (UTC)

Probably the best way to do this is to use the file in the associated language Wikisource, and then link to the Wikisource page as use reference URL (P854). This enables the user to locate the associated language as well as use any available translation tools to read the text depicted. Jane023 (talk) 06:29, 21 September 2019 (UTC)
Are you suggesting that we create entries for each document we want to display at Commons in Wikisource? I should make a text based version of the death certificate, or the census, or a birth certificate? --RAN (talk) 17:56, 22 September 2019 (UTC)
The example in the OP has been deleted as a copyvio. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:37, 22 September 2019 (UTC)
I am not sure why you deleted it, when it clearly was not a copyright violation, especially when it was under discussion. It takes the same amount of time to correct a license as it takes to delete it. The item was restored, birth, marriage, and death certificates are not eligible for copyright since they consist entirely of information that is common property and contain no original authorship. --RAN (talk) 16:11, 24 September 2019 (UTC)
I asked for it to be restored, Commons already declared death certificates to be ineligible for copyright. --RAN (talk) 19:00, 22 September 2019 (UTC)
I created an example. I used the image of the death certificate as a reference for cause_of_death at Eddie August Schneider (Q5335826) to show what it would look like, I have not seen it other the Gacy example, which has been removed. --RAN (talk) 20:34, 22 September 2019 (UTC)

Upload dataset


I would like to upload a dataset which contains triples of the form <item, property, value>, for example: <Q42, P69, Q691283> (which refers to "Douglas Adams, educated at, St John's College"). I have 8 million triples extracted from Wikipedia and I would like to contribute with this information. I read about "Primary sources tool" which allows to upload a dataset and Wikidata users review the information and accept or reject each statement if it is erroneous, but the gadget of this tool seems to be broken, and I would like to know if it is possible to upload this information in another way. Could you help me please?  – The preceding unsigned comment was added by JLuzc (talk • contribs).

If you believe all triples are correct, you may use QuickStatements. Please note import of such size need a bot approval, and you should publish the list somewhere for community to review. Alternatively see Wikidata Game.--GZWDer (talk) 05:01, 22 September 2019 (UTC)
@GZWDer: What is the threshold for amount of changes when a bot account is needed? On the Wikidata:Bot requests page it says you don't need bot if you use QuickStatements Iwan.Aucamp (talk) 11:24, 23 September 2019 (UTC)
@JLuzc: To import data is one thing but to add references to the data and relations between items is even more important. Besides IMO proper references (in this case from original Wikipedia) are preferred to imported from Wikimedia project (P143). Kpjas (talk) 07:27, 22 September 2019 (UTC)
We don't have formal rules about the size when bot approvals are needed. I personally think that if you add >10,000 items you should seek some form of consensus before adding the data.
  • I think it's a huge misunderstanding to think of Wikidata as a triple-store. If you enter data it shouldn't be triples but you should have references for your claims which makes the information essentially 5D. ChristianKl❫ 08:02, 24 September 2019 (UTC)

MalaCards bot

I want to make a bot that imports data from MalaCards would be nice as they have somewhat structured symptoms. The bot page suggests bots be discussed here first before a request is made for them so this is area for that discussion. Please weight in and indicate if you think it should not be done. Iwan.Aucamp (talk) 22:57, 22 September 2019 (UTC)

  • The question that always comes *before* the bot question: can you manually enter the information? If yes, you can automate that with any kind of bot/script or tool. If not, you need to work on creating properties, sources, etc. Edoderoo (talk) 12:03, 23 September 2019 (UTC)
  • Anyone can create and run a pseudo-bot (aka. semi-automated tool) to import such data, using your account, and running from your computer, under your control. A bot account is required only for permanently or periodically running, totally automated tools, and, for security reasons, don't use your main account on external servers like Toolforge.
I created a tool for importing data from TOP500 using pywikibot and Python3 and WSL, take a look. --Amitie 10g (talk) 14:28, 23 September 2019 (UTC)∆
MalaCards would be a source where it would be good to have a regular import if we have the data.
Currently, a lot of related medical data is imported by . Raising the issue of how the data is best imported on that talk page might be benefitial. Maybe, you can even cooperate with them to have the script that updates MalaCards run on their infrastructure. ChristianKl❫ 07:18, 24 September 2019 (UTC)
@ChristianKl: Okay, I will work on ETL software for it and maybe engage if it works. I am concerned about licensing though, more so than with the oil well dataset, would the requirement for ETL invalidate the concerns of licensing or would there be potential issues with MalaCards data license? Iwan.Aucamp (talk)

Settlements and municipalities in Brazil

There are three items for Porto Belo in Brazil: Porto Belo (Q22036564) for the settlement, Porto Belo (Q22063826) for the municipality, and Porto Belo (Q1799285) which says it's the municipality but at least some of the identifiers are for the settlement. I haven't checked the sitelinks but in English it's the municipality. Should all three be merged, one of them merged with Q1799285 (which one?), or would it be better to keep all three? Is there a guideline somewhere? Peter James (talk) 15:46, 24 September 2019 (UTC)

Also there are three possible values for Who's on First ID (P6766): 102060643 (county), 404565747 (local admin) and 101958157 (locality). Q22063826 has county, the other two have locality. We don't have an item for "county of Brazil"; is the county the same as a municipality? What is "Local admin", and should it be added to one of the items? Peter James (talk) 15:55, 24 September 2019 (UTC)

“Requests for deletions” proces

Where are the instructions for Wikidata:Requests for deletions?

Specifically, if a merge is possible and appropriate, can I just merge the item and report that I have done that? - PKM (talk) 18:47, 23 September 2019 (UTC)

Yes, please--Ymblanter (talk) 19:03, 23 September 2019 (UTC)
+1. Nomen ad hoc (talk) 19:45, 23 September 2019 (UTC).
Currently, it gets noted automatically and archived shortly afterwards. This is useful as new users tend to ignore Help:Merge.
However, I find that highly disruptive when the user is likely to know how to merge. --- Jura 20:13, 23 September 2019 (UTC)
My understanding is that PKM is talking about closing other's nominations, not their own ones.--Ymblanter (talk) 20:29, 23 September 2019 (UTC)
Of course. Imagine you list an item for deletion and PKM just merges it .. likely with something you didn't think it was appropriate merging with .. (in most cases) the bot currently archives such requests. --- Jura 20:45, 23 September 2019 (UTC)
In my experience, there are many items put up for deletion as duplicates. These just need to be merged. Complicated cases are best left to administrators.--Ymblanter (talk) 21:10, 23 September 2019 (UTC)
Yes, I am talking about others' items, such as [16] which needs an ENwiki sitelink moved and a merge. - PKM (talk) 20:53, 23 September 2019 (UTC)
Whenever approaching an item on RfD, it's worth to think about whether it's a controversial entry or isn't. If it isn't a controversial item, merging makes sense. For admins deleting it makes sense for noncontroversial items that can be merged. If it however is a controversial item, it's worth finding consensus for what should happen with the item instead of taking direct action. ChristianKl❫ 07:53, 24 September 2019 (UTC)
Wikidata:Proposed mergers was created two years ago, but has never been used and there are no links to it. There isn't much discussion at most deletion requests, and I doubt that there would be on a proposed mergers page. Peter James (talk) 21:01, 24 September 2019 (UTC)

Reuse of blank items

I propose to NOT delete the blank items and reuse them for new ones. A bot could identify and categorize the blank items in a specific category and when a user want create new item should check in the category for the still available and when item is sufficiently filled, the user can now delete the category from item. Is it a good propose? --2001:B07:6442:8903:D924:FC3F:3FBC:475F 16:17, 24 September 2019 (UTC)

We do exactly not this to not hurt our data users. Sjoerd de Bruin (talk) 17:12, 24 September 2019 (UTC)
Very bad idea (and we don't have categories on items). Imagine if an item is use outside Wikidata (for instance on Wikipedia), then blanked (by mistake for instance) and then re-purposed and then de-re-purposed as the mistake is spotted. It would be a nightmare to manage. On the other hand, I see no advantage of doing that. Cheers, VIGNERON (talk) 17:20, 24 September 2019 (UTC)
Agree entirely - this would cause immense amounts of confusion and be very difficult to fix when it gets missed up. Creating new items doesn't cost us anything. Andrew Gray (talk) 20:27, 24 September 2019 (UTC)

P1559 name in native language criterion?

For persons from multilingual backgrounds, is this property used for their names in only their mother tongue, or in both mother tongue and official languages of their countries? For example, if a Japanese was born in US (hence born a US citizen), grew up in Japan, but became famous in US and is known by his English name around the world, do we put only the Japanese name as P1559, or both Japanese and English? For Hong Kong people who speak Cantonese, should their P1559 be only their Cantonese (Chinese) name, or both Canto and English?--Roy17 (talk) 18:21, 24 September 2019 (UTC)

It is "name in native language". In the example you provided, I think it should be their name in English letters (latin script) and not in Japanese writing system. There are cases in which those names (latin-based) are related to another language, but are written in the way/s a certain place does. Esteban16 (talk) 21:39, 24 September 2019 (UTC)

Warning: WDQS is way behind

The query service has been at least 4 hours behind Wikidata all day - and from this grafana page it looks like the problem started about 15:00 (GMT?) yesterday (23 Sep 2019). Does anybody have any idea what's going on? It's been many months since we had a problem as bad as this! ArthurPSmith (talk) 19:34, 24 September 2019 (UTC)

It's mostly a combination of multiple bots/users editing large items. Soon query service lag will influence the maxlag parameter. Sjoerd de Bruin (talk) 21:17, 24 September 2019 (UTC)

Cleaning up Wikidata Entries for Heritage Institutions - Your Help is Needed!

Dear all,

The FindingGLAMs and the Sum of All GLAMs projects (see: WikiProject Heritage Institutions) are currently cooperating to ingest (a large number of) official databases of heritage institutions.

However, in order to avoid ingesting a lot of new data on top of messy data, we would like to tackle existing data modelling issues first and to clean up the existing data. This Google Doc contains a tentative list of data modelling issues related to heritage institutions. Please help complementing it!

The most salient issue is the fact that many Wikidata items describing heritage institutions have an “instance of” statement both for “organization” (or subclass thereof) or “building” (or subclass thereof). - These cases should be tackled as soon as possible, and this is where we need your help! - The coordination and follow-up of this clean-up tasks can take place on a sub-page of this project page.

Who would be ready to help with these tasks?

--Beat Estermann (talk) 19:53, 14 September 2019 (UTC)

Akuckartz Beat Estermann Vladimir Alexiev Ilya Sadads Strakhov Zeromonk Spinster Wittylama Daniel Mietchen Susannaanas Sic19 Jason.nlw Carlojoseph14 YULdigitalpreservation MB-one Ouvrard MartinPoulter Missvain VIGNERON Ainali Birk Weiberg Pmt Mauricio V. Genta Smallison ProtoplasmaKid 2le2im-bdc Rodrigo Tetsuo Argenton Ivanhercaz VisbyStar Patafisik Beireke1 Vahur Puik Ettorerizza Sp!ros Alexmar983 Epìdosis Buccalon Mrtngrsbch Eothan Giaccai NAH User:Fralambert Ipoellet Valeriummaximum Hannes Röst Ahc84 AmarilisMGC Trivialist Fuzheado Pictogram voting comment.svg Notified participants of WikiProject Cultural heritage

Is there is a good reason why issues have been collected in a Google Doc and not onwiki?--Ymblanter (talk) 20:00, 14 September 2019 (UTC)
I usually draft this kind of lists first in a Google Doc; it's easier and quicker, especially if you want to be able to restructure things in the process, and we can discuss directly on the same page. The idea is to eventually document the issues and clean-up tasks on the wiki. See for example this page where I've started to tackle data modelling issues in the context of performing arts data. But feel free to start documenting maintenance tasks on the wiki. --Beat Estermann (talk) 20:13, 14 September 2019 (UTC)
  • Coordination should preferably be done onwiki. --- Jura 20:21, 14 September 2019 (UTC)
Perhaps a Etherpad document will accepted by everyone? --2le2im-bdc (talk) 09:59, 18 September 2019 (UTC)
(conflict) Oh Gosh... last time in Italy they inserted 10000s of files form public archive, it took us two years to clean up. And we have dozend of active users, we still find problems nowadays mostly duplicates... I apreciate your effort of informing or this will become critical aoon. As part of the problem we are in fact, in Tuscany, starting also to separate buildings from institutions, especially museum collections. Once you need order, you just create more order, it's a continous process. But it's a preliminary attempt influenced mostly by WLM and it will take years with our human workforce. I am actually working on a general insertion and clean up of some data from public archive before November, so we could create some query that can be adapted changing the administrative entity.--Alexmar983 (talk) 20:02, 14 September 2019 (UTC)
Have you already considered using property constraints for this purpose? - I would be interested in hearing your thoughts on this. --Beat Estermann (talk) 20:15, 14 September 2019 (UTC)

In order to gauge the size of the problem, I've started to tweak the Property Dashboard template; an overview table of problematic museum items per country can be found here. The template seems to function like Listeria; thus, it is impossible to have several lists on the same page. --Beat Estermann (talk) 20:22, 14 September 2019 (UTC)
@Beat Estermann: Yes, having several templates on one page is just very annoying to handle (the bot has to understand between which portions where to insert the table) − it actually should give a proper error message when that happens. The workaround is to create a sub-page for each dashboard, and transclude it in the page where you wanted to have both. Jean-Fred (talk) 07:06, 16 September 2019 (UTC)
Ok, thanks for the hint regarding the workaround. --Beat Estermann (talk) 18:31, 18 September 2019 (UTC)
Thanks for initiating this work. The problem is broader than GLAM items being both an organization and a building though - art museum (Q207694), library (Q7075), archives (Q166118) and museum (Q33506) are all indirect subclasses of geographic entity (Q27096213). Perhaps the first task is to agree whether these items are supposed to represent the organization, collection or place? Simon Cobb (User:Sic19 ; talk page) 11:12, 15 September 2019 (UTC)
We have library (Q7075) and library building (Q856584), though the first is often used for buildings. I think the distinction is respected generally when a museum/library/etc. has moved from one location to another, or when it is located in a building which is otherwise notable under a different name. But I agree that institutions/organizations are different from buildings, and I am not sure that either one is really a geographic entity. - PKM (talk) 21:25, 15 September 2019 (UTC)
There are some institutions that are essentially location based. The question is if it really creates a problem as it's generally done. Sums of all paintings seems to use the same for both and works well. --- Jura 22:09, 15 September 2019 (UTC)
@Beat Estermann: I really don't see the problem. When I think of any museum I have visited, I think first of the building - many of them are quite striking architectural structures in themselves. Yes there is an organization and collection that are conceptually distinct, but (for the most part) they all reside in the same geographical location! Where there are exceptions to that, if a museum collection relocates, or if a building has its own history of interest that predates (or postdates) the museum then sure, lets create separate items. But barring that, what is the problem with a single item for these concepts at least to start with? ArthurPSmith (talk) 13:49, 16 September 2019 (UTC)
@ArthurPSmith: Yes, exactly. You first think of the building, I first think of the organization, and someone else first thinks of the collection. And when we start adding statements to the item, let's say inception (P571), we all suppose that the others share our assumptions. And that's the problem, for most museums have different dates for the construction of the building, the initialization of the collection, and the creation of the organization. I think some people have tried to counter this problem by adding qualifiers to ambiguous statements, specifying whether the statement applies to the building, the organization, or the collection. But I don't think this is the way forward, because it unnecessarily adds more complexity to the data model and makes the data difficult to query. --Beat Estermann (talk) 16:03, 17 September 2019 (UTC)
Well, the institution is localized . In general, we aren't really interested in where they store their non-exhibited items and they might not even want to share that .. If you want to add details about the architectural history, this might not be the question at hand. --- Jura 07:52, 18 September 2019 (UTC)
I think there may be different issues as libraries/GLAM come in all shapes and sizes. Larger institutions may occupy more than one building and these can be modified over time by different architects and as such should probably have data entries to cover these attributes which dont necessarily fall within the functions or collection activities of the institutions. On the other hand smaller institutions can be moved through a variety of locations by their governing bodies (in Australia's case local councils) so I guess the question is whether wikidata should record these changes of locations? Eothan (talk) 23:46, 16 September 2019 (UTC)
Just for the record, I am not suggesting that we need to create three distinct items for every heritage institution. I am just saying that whenever we create an item, we should decide whether we are describing the organization, the collection, or the building. In many cases, it may be fine to just describe the organization. --Beat Estermann (talk) 18:56, 18 September 2019 (UTC)
This discussion points to a larger issue, Wikidata doesn't currently have a process for the community to agree on schemas for different subjects or an easy way to find the agreed schemas. We can have a discussion here and agree a schema but then it gets archived and no one knows it existed and we are back to the start again, similarly for recording them on Wikiprojects, you have to know they exist and several Wikiprojects may cover the same topic e.g people. My suggestion would be to have a central place for discussions and capture the schema in the Wikidata item for the topic using a new property called something like 'Schema'. --John Cummings (talk) 11:02, 17 September 2019 (UTC)
Hi @John Cummings: We currently use the thematic WikiProjects to keep track of the outcomes of such discussions. Wasn't there another proposal to improve the findability of the documentation on the WikiProjects pages by adding pointers to them on the WD entries for the classes they cover? I currently cannot recall exactly where I heard/read about that proposal. This would be an organic way forward, without centralizing our approach to developing and documenting ontologies from the outset. What do you think? --Beat Estermann (talk) 16:10, 17 September 2019 (UTC)
Agreed with @John Cummings: to the difficult to discuss and validate schema with the community. For GLAM's data models in Wikidata, it would be perhaps required to organize a physical discussion panel in a next WikiConference or to organize one focus on it? @VIGNERON, Beat Estermann: --2le2im-bdc (talk) 09:53, 18 September 2019 (UTC)
I think we can use online calls or real-life meetings (for example the upcoming Wikidata Conference in Berlin) subsidiarily; but let's try to have a structured discussion on the Wiki. After this initial discussion here and the collection of issues in the Google Doc, we are planning to systematically describe the issues, the proposed solutions and the resulting maintenance tasks on the following page: Wikidata:WikiProject_Heritage_institutions/Data_structure/Data_modelling_issues. Its talk page is where we can eventually move this discussion. --Beat Estermann (talk) 18:56, 18 September 2019 (UTC)
This is the property I referred to above and that can be used to link property and class items to their respective WikiProject page(s), where data modelling issues are discussed: maintained by WikiProject (P6104) (query generating a list of such WikiProjects) --Beat Estermann (talk) 06:42, 25 September 2019 (UTC)

Using Wikidata to describe datasets

Today in the AKTS workshop at the eScience conference in San Diego, I will present the idea of using Wikidata to describe datasets in the rare case that does not work. This would allow to include these datasets in dataset catalogs (such as Google Dataset Search) in a collaborative manner and that is beneficial to everyone. This relies on the dataset communities and the Wikidata communities to agree with that approach.

I will open the discussion with the dataset communities at the workshop, and I wanted to open the discussion with the Wikidata community on Wikidata (I guess, the WikiProject Datasets is the right place?). I am happy to answer questions and discuss the proposal. --Denny (talk) 14:15, 24 September 2019 (UTC)

  • We had this some time ago, but I don't think it really got active. Good idea to relaunch this. --- Jura 14:18, 24 September 2019 (UTC)
Thanks for the pointer. I copied your answer to the WikiProject Datasets talk page to keep it in one place. --Denny (talk) 17:45, 25 September 2019 (UTC)
I don't like people doing that (unless context is provided). --- Jura 18:02, 25 September 2019 (UTC)

Import mayors of Colombia?


I noticed that the plateform has extensive data about mayors of the country (all the current mayors at least, and for some areas former mayors too) but these data are quite limited in depth, mostly only the name (given and last name, sometimes together but often separate) and the timespan. Qv. dataset on

Would it be a good idea or not to import these data? It would mean to create skeleton items with at least one serious and publicly available reference (point 2 of WD:N) and needed to indicate the mayor of the municipality item (point 3 of WD:N but with very few statements.

What do you think?

Cdlt, VIGNERON (talk) 17:17, 24 September 2019 (UTC)

I think it would be good to import them if you get the matching of the municipality right. ChristianKl❫ 08:29, 25 September 2019 (UTC)
@ChristianKl: the matching should be fairly easy (there is only 1102 municipality in Colombia, OpenRefine should have no problem with it). Cdlt, VIGNERON (talk) 09:55, 25 September 2019 (UTC)
  •  Support This would fulfill a structural need. Nomen ad hoc (talk) 12:02, 25 September 2019 (UTC).

How do we delete empty items ?

I've moved the content of Q60940956 to merge it with Q8733659. The first is now useless and empty. What should I do with it now ? --TwoWings (talk) 12:14, 25 September 2019 (UTC)

Merge it the proper way so it creates a redirect. Sjoerd de Bruin (talk) 12:20, 25 September 2019 (UTC)

Make items with schema:description "hazardous chemical exposure" that is subclass of chemical hazard (Q21167512) subclasses of hazardous chemical exposure (Q68182219)

Currently there are some items with description "hazardous chemical exposure" that have symptoms (P780). These should be subclass of (P279) or instance of (P31) of physiological condition (Q7189713).

Example items:

Query to find items:

SELECT ?condition ?conditionLabel ?conditionDescrption WHERE {
  ?condition wdt:P279+ wd:Q21167512.
  ?condition schema:description ?conditionDescrption.
  FILTER(LANG(?conditionDescrption) = "en").
  FILTER(STRSTARTS(?conditionDescrption, "hazardous chemical exposure"))
  OPTIONAL { ?condition rdfs:label ?conditionLabel FILTER(LANG(?conditionLabel) = "en") }

Try it!

I want to fix all of them, not sure what best option is. I think best is to make them all subclasses of hazardous chemical exposure (Q68182219) which is a subclass of physiological condition (Q7189713). I can do it with wikidata cli but I don't know if that would be acceptable or if I should first make a bot request before doing it with wikidata cli.

To do it with wikidata cli I would run something like this

curl --silent -H 'Accept:application/json' --data-urlencode query='SELECT ?condition ?conditionLabel ?conditionDescrption WHERE {
  ?condition wdt:P279+ wd:Q21167512.
  ?condition schema:description ?conditionDescrption.
  FILTER(LANG(?conditionDescrption) = "en").
  FILTER(STRSTARTS(?conditionDescrption, "hazardous chemical exposure"))
  OPTIONAL { ?condition rdfs:label ?conditionLabel FILTER(LANG(?conditionLabel) = "en") }
}' | jq -r '.results.bindings[].condition.value' | gawk -F, '{ print gensub("^.+/", "", "g", $1) }' | xargs -I{} -t wd add-claim {} P279 Q68182219

Iwan.Aucamp (talk) 14:24, 22 September 2019 (UTC)

I already tried some edits with wd cli, here are what they look like:

Iwan.Aucamp (talk)

I ran the updates, if someone has a problem with it ping me and we can talk. Iwan.Aucamp (talk) 19:27, 25 September 2019 (UTC)

Would oil wells be notable enough to include here

I would like to import this data:

I think it satisfies Notability Criteria #2 ("It refers to an instance of a clearly identifiable conceptual or material entity. The entity must be notable, in the sense that it can be described using serious and publicly available references.") - so would it be acceptable to import? Iwan.Aucamp (talk) 11:15, 23 September 2019 (UTC)

How many items are we talking about? How's the license of the source data? ChristianKl❫ 14:54, 23 September 2019 (UTC)
@ChristianKl: The import would be about 2000 Items. As Andber08 said the license is Norwegian Licence for Open Government Data (NLOD) 2.0. This is a fairly permissive license, but it does not allow sub-licensing or transferal of license and all data entered here should be CC0 so I'm not sure it is liberal enough. In another post here I ask about importing data from MalaCards for example - but I would guess the same concern exists there? There is some guidelines here Wikidata:Licensing, but maybe there are some specific examples which may act as guidance? Maybe I can email them an ask if I can import it? Iwan.Aucamp (talk) 19:20, 25 September 2019 (UTC)

The source is a Norwegian government directorate so its under this license Andber08 (talk) 15:22, 23 September 2019 (UTC)


For Ford National Reliability Air Tour (Q5467878) and other recurring events, should I be using start_time and end_time, or inception and dissolved_or_demolished? Looking at other similar ones, it is a mixture. For the main event I have been using inception and dissolved_or_demolished, and for the yearly entries I have been using start_time and end_time. Is this correct? --RAN (talk) 23:35, 25 September 2019 (UTC)

RAN: 2016 Summer Olympics (Q8613), 2018 FIFA World Cup (Q170645), 2019 Tour de France (Q30091668), they all have start time and end time , while Tour de France (Q33881) and FIFA World Cup (Q19317) have inception. inception (P571) states that is is "the date or point in time when the subject came into existence as defined", that is to say, when something was officially created. Therefore, you are right. Esteban16 (talk) 01:48, 26 September 2019 (UTC)

What to do with statement without any source?

Hello, some contributors think that we should removed statement for which we cannot find any source. Help:Sources states "majority of statements on Wikidata should be verifiable insofar as they are supported by referenceable sources of information" but do not say anything what wa should do with statement without any source.

So, in the case we cannot find any source, is it better to leave the statement without source or delete it? Pamputt (talk) 12:46, 22 September 2019 (UTC)

  • I think the one of "some contributors" deletes statement that don't have a reference even if they think their are probably correct (meaning: the main tertiary or quartery reference they are using doesn't mention it). (They also delete references in Russian and German, but that's another problem.)
So generally these claims are verifiable by a source that could be added as reference.
Anyways, most claims in Wikidata don't have a reference, but were imported from a Wikipedia that generally has a policy about references.
If the statement is plain wrong and has no reference, I think it can generally be deleted. To avoid re-deleting it, deprecation is also an option. If it has a reference that supports the (incorrect) claim, it should be deprecated. If there are some doubts about it, it can be deprecated as well, until better references are added.
Gradually, the situation is improving as we add references to statements. --- Jura 14:00, 22 September 2019 (UTC)
tl:dr; keep the statement, find & add a source. --Tagishsimon (talk) 14:02, 22 September 2019 (UTC)
  • On a related note to this, some statements don't need a source because the item itself is the source. However, I wonder whether we should really be adopting a method of explicitly marking this to be the case? Otherwise there's no way to know the statement was sourced this way. --SilentSpike (talk) 14:26, 22 September 2019 (UTC)
    In the case we cannot find any source, the claim, or assertion, or hypothesis, etc, is by definition not verifiable, and should not be included in wikidata; it is, also by definition, not an information.
    In some cases, we can keep a statement without a source, if is is completely obvious (« such and such is a human being » ), or if there is absolutely no doubt a source will be found to confirm OR infirm it (« such and such won the Nobel Prize » ). But for a statement such as «  such and such was (a hundred and fifty years ago) a member of the Hungarian Academy of Science » we must have a reliable source. Because if the statement is false, we may never be able to find a source establishing it is false.
    Also, we should be aware that by accepting as a source an importation from Wikipedia in some language we have a big risk of creating circular references: Wikipedia in some languages (like French and Spanish) have infobox templates that directly import declarations from Wikidata (like « such and such is a member of the Hungarian Academy of Science »); then, using this, a French or Spanish contributor will use this importation to put the item in a local category («  Category: member of the Hungarian Academy of Sciences » ). And finally, a wikidata contributor (sometimes a bot) will notice the category and put such and such as a member of the Hungarian Academy of Sciences, with as the reference « Imported from the Spanish Wikipedia » . Sapphorain (talk) 15:13, 22 September 2019 (UTC)
    • Do I understand this to mean that all images taken by Commons contributors should be deleted from Wikidata? That seems absurd. - Jmabel (talk) 17:06, 22 September 2019 (UTC)
    • It's been a day and no one seems to have responded to my remark. How about
      ⟨ FOO ⟩ instance of (P31) View with SQID ⟨ human (Q5)  View with Reasonator View with SQID ⟩
      ? Are you saying we need to find an overt, reliable third-party source for this for every person? - Jmabel (talk) 15:31, 23 September 2019 (UTC)

If the discussion is too astract, let us take an example. On Alphonse Pyramus de Candolle (Q123807), a lot of statements have been deleted for member of (P463) and award received (P166). I did not spent a lot of time to find sources for deleted statement but for two of them, I have found sources. For the rest, I did not take time to search. If the unsourced information is kept, then sources can be added later. If the information is not present anymore, then someone will need to redo the job to find the information by himself/herself. If an unsourced information is not true, then it will be deleted one by one (not systematically). Pamputt (talk) 16:15, 22 September 2019 (UTC)

I brought it up at the time, and one user was warned to discontinue the practice. It was the same user that was deleting all the Findagrave references from Wikidata for the same reason. There was a vote to remove Findagrave as a reference at the English Wikipedia and at least one user began deleting Findagrave, when used as a reference here at Wikidata while they were deleting them at the English Wikipedia. I am sure with a little searching you can find the discussion. Here it is for Findagrave Wikidata:Requests_for_comment/Findagrave_removed_as_a_source_for_information, you may want to check out the current contributions of the editor in question to make sure the practice has discontinued. --RAN (talk) 18:56, 22 September 2019 (UTC)
If you look at the contributions for that editor they are currently removing easily sourced statements from Wikidata, that are currently not sourced, in their past 20 edits. In the time it took to delete it or modify it, I was able to find a source. I don't see any value from removing easily sourced/modified statements on religion or ethnicity, even if English Wikipedia is deleting them. --RAN (talk) 20:52, 22 September 2019 (UTC)
Going back to the subject of this section (which is, in case we forgot: « what to do with a statement without any source » ), I think we should stop referring to an unsourced assertion as an « information », and recall what it is, which is, no more no less: «a statement without any source » If a claim is unsourced, we don’t know whether it is true or false, and it is not an information. The inappropriate use of the word « Information » leads to misunderstandings. For instance the statement «  If the information is not present anymore, then someone will need to redo the job to find the information by himself/herself » presupposes that what is claimed is known to be true, which is not the case. If one writes instead the more correct statement «  If the unsourced claim is not present anymore, then someone will need to redo the job to find (and source) this unsourced claim by himself/herself (or somehow establish that it is false) », then it is not obvious at all that there is any advantage to keep this unverified claim as a declaration, waiting for an hypothetical confirmation or infirmation. Sapphorain (talk) 20:27, 22 September 2019 (UTC)
If we start at universe (Q1) almost every statement is unreferenced, I imagine about >80% of Wikidata would be deleted and >90% if we eliminate Wikipedia as a circular reference. Can someone count how many Q entries have no references at all? At one time I did an English Wikipedia search to show how many English Wikipedia articles were devoid of any references at all, to counter the argument to delete unreferenced English Wikipedia articles. --RAN (talk) 20:42, 22 September 2019 (UTC)
FYI, I did the job on frwiki in November 2016 and about 50% of the articles have no source. Of course they do not have been deleted. Pamputt (talk) 23:11, 22 September 2019 (UTC)
@VIGNERON: could you help to write a query that compute the number of items that have no reference at all and how many have only reference with imported from Wikimedia project (P143)? Thanks in advance (if you have no time for that, do not worry ;)) Pamputt (talk) 23:14, 22 September 2019 (UTC)
@Sapphorain: That a statement without source is no information is simply not true. You at least have the information that the predicate can be applied, and the statement object is a possible value. --SCIdude (talk) 05:47, 23 September 2019 (UTC)
An information is a « knowledge acquired in any manner; facts; data; learning; lore » (Webster’s). A statement without any source is nothing of the sort. The only information I get from a statement without any source, is that somebody stated it. Sapphorain (talk) 08:35, 23 September 2019 (UTC)
One should not add any info on Wikimedia without a source. But you can not always turn that around to "if there is no source given it must be deleted. My proposition: if you delete information, you should have a good reason to do so. There is no source is no good reason for me. The author is known to be unreliable might be a good reason to me. I have serious doubts about the claim might also be a good reason to me. Edoderoo (talk) 12:07, 23 September 2019 (UTC)
Wikidata is work-in-progress (and always will be). In almost 7 years we've got <1% of information that must be here, so we should encourage mass import of data, not discourage it. Of cause, importing well referenced data and verification of unsourced statements is important. But even unsourced (or poorly sourced) data bring some value both for our project and for our clients (especially for items without links on major wikis). If some wikipedias (or google/siri/alexa/etc) do not want to use unsourced data, they always can check presence of stated in (P248)/etc. Ghuron (talk) 12:19, 23 September 2019 (UTC)
I second everything that User @Ghuron: said above. I think that referenced data are the standard we all should strive for. It should be absolutely requied for bots to add only referenced data and we need initiatives to add missing references across WD or in Wikidata projects. Kpjas (talk) 15:25, 23 September 2019 (UTC)
We have not been able to confirm this claim (Q21655367) as a reason for deprecation. If a claim has no source, and you can't verify it on the Internet, is there any reason to keep it? It would make Wikidata into the source of the claim, from the point of view of the rest of the Internet. I suppose eventually somebody else may copy it, and they can then be cited as the source, per Ghouston (talk) 23:47, 23 September 2019 (UTC)
Internet isn't the only source for references .. anyways, the question here seems to be much simpler: it appears that content that can easily be referenced in Russian, German, Swedish, Hungarian or Dutch is simply deleted as it can't be found in the tertiary source the contributor is using (or is referenced in these languages and the contributor deletes the references). --- Jura 06:20, 24 September 2019 (UTC)
@Kpjas: I'm not sure how do you combine "I second everything that User Ghuron said above" with "It should be absolutely required for bots to add only referenced data" :) Yes, we all should strive for getting referenced data. No, we should not forbid importing unreferenced data as long as error ratio is acceptably low. Please tell me what is the feasible way to automatically import date of birth for Q67936223 with reference? And we should have this info here, because it will help us to identify inconsistencies (e.g. check dates with his father) and duplicated items. --Ghuron (talk) 09:31, 24 September 2019 (UTC)
My two cents: 1: While of course references are great, I think most emphasis should be placed on verifiability, not simply verified. Thus, a statement without a reference may or may not be accurate or true, but it is almost always verifiable in theory, even if not currently cited, should one familiar with the subject consult reliable published sources (or spend 20 seconds Googling).
2: We'd have a lot more sourced statements if making and adding references was easier. It's time-consuming enough when most entries need to be added by hand, 1 data morsel at a time, (instance of, gender, occupation, DOB, nationality, spouse, etc. etc. etc.). I predominantly create/expand items for biographies: often times I'll add the basic biographic statements, plus some External Identifiers to link to other databases, and "Described by source" (if the source is already in Wikidata) or "Described at URL". Adding a reference to each statement would be great, but I only have so many hours in a day to waste making the world smarter. When adding references to individual statements, the simplest is identifiers like American National Biography ID (P4823) (In which case I don't bother with "Stated in", because I'm not that pedantic) but for many statements that can't be covered by existing items, I generally don't care to spend extra time creating items for every book (and book edition!), journal article, newspaper page, etc. to give a complete, "proper" reference. Thus, a quick and dirty alternative is reference URL (P854), which minimally provides citation, but bare URLs invite the problem of future linkrot, and obscure important bibliographic data. SourceMD is handy for some situations, but I wish there were more tools to quickly create well-structured references by simply entering an ISBN, DOI, ASIN, or URL (similar to the Wikipedia citation tool for Google Books) and to quickly add the same reference to multiple statements at once. Even better would be the ability to create an item that pulls all existing structured data from a source, such as gender, birth/death date & place, by simply pasting say this URL. Bots that automatically add citations when unsourced data matches external identifiers are also very helpful in this regard (e.g. here). . People who know nothing about writing script should be able to contribute efficiently to Wikidata. The more convenient and user-friendly Wikidata becomes, the more likely we are to have good data. And yes there are a lot of helpful tools already, but they are not always easy to find.
And P.S. just as Wikipedia itself is not a reliable source (data can be changed or deleted at any time), Wikidata isn't either , and no one should use its data completely blindly or uncritically. -Animalparty (talk) 05:16, 24 September 2019 (UTC)
  • I don't think there's any basis for removing unsourced claims outside of claims related to Living Persons just because they are unsourced. Any data-reuser can choose to ignore statements without sources if they think they are not important to that data reuser. At the RfC for the Living Persons there were also opposition to automatically remove unsourced items for Living Persons.
As far as help sources go, it's still in draft state and not accepted policy. It summarizes a best practice that's quite effortful to add full items with books and their authors whenever you want to take information from a book. There's no consensus that that effort is always required. ChristianKl❫ 07:43, 24 September 2019 (UTC)
Can someone look at these edits of a user removing unsourced ethnicity and religion entries. The English Wikipedia ruled that they should be removed at English Wikipedia, but should we be applying that ruling here at Wikidata? For instance she is removing that Amelia Erhard is ethnically German. A quick search shows that John Earhart (1732-1777), one of her immigrant ancestors, was born in Germany. We can work on rules for defining ethnicity, but we should decide now whether we should be removing them, once deleted they are difficult to restore. The editor also removed Amelia's religion as "Episcopal Church", yet I found easily: "Of note, the famous female aviatrix, Amelia Earhart, was born in Atchison and baptized at Trinity Episcopal in 1897. She attended Trinity as a child during her frequent visits to her grandparents’ home in Atchison. Amelia’s grandparents, Judge Alfred and Amelia Otis, were early Atchison residents and founding members of Trinity Church." The issue was brought up with her before, when she was applying the English Wikipedia rule of removing instances of using Findagrave as a reference here at Wikidata. English Wikipedia rules should not automatically apply at Wikidata. --RAN (talk) 13:07, 25 September 2019 (UTC)
The description of ethnic group (P172) reads "consensus is that a VERY high standard of proof is needed for this field to be used. In general this means 1) the subject claims it him/herself, or 2) it is widely agreed on by scholars, or 3) is fictional and portrayed as such" and this property has a constraint citation needed constraint (Q54554025). So I think this removal is not just due to applying English Wikipedia rules in Wikidata but due to rules in Wikidata. See also the related discussion at Wikidata:Project_chat/Archive/2018/01#Cleanup_of_unsourced_“ethnic_group_(P172)”_claims. If you can find sources supporting these statements, just restore the statements with a reference. Valentina.Anitnelav (talk) 13:59, 25 September 2019 (UTC)
Date of birth and date of death have the same citation needed constraint (Q54554025). Wikidata:Project_chat/Archive/2018/01#Cleanup_of_unsourced_“ethnic_group_(P172)”_claims is a discussion between 5 users with no  Support  Oppose !vote. Have a real  Support  Oppose !vote with a quorum and decide whether to remove all unsourced statements with citation needed constraint (Q54554025) or demand that minimally a search should be done before deletion. Ad hoc removal is not the efficient way to do it, it should be handled by a bot if the decision is to be made, so we have internal consistency. The time it took me to find the information was about the same as the time for deletion. Once deleted no one will bother with it again. As for "consensus is that a VERY high standard of proof is needed for this field to be used." you need to link to the ruling where the consensus was established here at Wikidata containing the  Support  Oppose !votes. Anyone can add in the wording of their choice on the discussion page. We need a "high standard of proof" that consensus was actually established. Judging from the diverse opinions above, I don't believe it was established. --RAN (talk) 19:06, 25 September 2019 (UTC)
citation needed constraint (Q54554025) at date of birth and date of death is qualified with constraint status (P2316) suggestion constraint (Q62026391). Ad hoc removal may not be an efficient way, but in my opinion legitimate if in doubt. I agree that doing a quick search would be better than just removing the statement, but the burden of proof generally lies with the editor adding a statement, not the one removing it (this is stated at Wikidata:Verifiability#Responsibility_for_providing_sources - this page is still in development/discussion, but I could not find anyone opposing this on its talk page). - Valentina.Anitnelav (talk) 07:19, 26 September 2019 (UTC)
I don't think we should support ANY decision based on description of the property and/or constraints. People do not spent a lot of time discussing perfect wording of description, normally it just represent opinion of the person who propose it. Constraints are purely utilitarian function, for instance I've put proper motion (P2215) property constraint (P2302) citation needed constraint (Q54554025) not because star proper motion is particularly sensitive data, but because vast majority of P2215-statements do have references and its useful to monitor the remainings. --Ghuron (talk) 11:21, 26 September 2019 (UTC)
Of course the description and constraints of a property are an important basis for deciding how to properly use it and if a use is correct or incorrect. They describe its scope and appropriate use. Sometimes they may be sloppy and need improvement. Constraints specify the correct usage of a property in general (there may be always some exceptions to the rule), so they are a bit more than just "utilitarian functions".
This does not mean that description and constraints can't be changed. If there is consensus that references for ethnic group (P172) are not that important, description and constraints should be probably adjusted. - Valentina.Anitnelav (talk) 14:34, 26 September 2019 (UTC)
Hoi, from my perspective .. the notion that you even consider deleting data from Wikidata based on the notion that Wikidata has no sources is utter destructive. When for instance the Australian Academy of Sciences has a list of award winners associated with that award, that source is great and well placed. When you think that a source has to be on the award winner, you forget how many duplicates exist, how many false friends you can find in Wikipedias.
In the example given a lot of correct information was deleted. For me that is basic vandalism. Thanks, GerardM (talk) 05:37, 26 September 2019 (UTC)

EditGroups: batch archival

Hi all,

EditGroups is a tool which lets you undo groups of edits made by supported tools (see Wikidata:Edit groups for more info). Today I have introduced a change in the behavior of the tool, which should hopefully be unnoticeable for most use cases, but I wanted to announce it here to make sure people do not get bad surprises.

In short: batches older than one year containing more than 10 edits can no longer be undone.

Long version: because EditGroups caches edit metadata for all edits made by the tools it tracks, the size of the database could potentially grow beyond reasonable and make the tool slow in the future (it contains more than 100 million edits at the moment). I have therefore decided to delete the edits in old batches, only retaining the batch metadata and a few sample edits to get a sense of what the batch does. These batches are marked as archived and can no longer be undone (see this batch for instance). You can consider this the equivalent of MediaWiki's limitation on the edits available in the recent changes feed (which only covers one month of edits). This should not impair usability too much, as undoing old edits is generally hard anyway (since newer edits might interfere), and it is unlikely that a problematic batch would go unnoticed for a long time. In the future it could be possible to un-archive batches by recovering the deleted edits from the contributions feed of the user who performed it, should we want to undo an archived batch. This move should ensure that EditGroups can be run reliably and cheaply in the coming years, adding support for more tools, and not creating dead links in edit summaries (since archived batches are still available at the same URL).

If you have any feedback about that move (or any other aspect of the tool), I would be happy to hear it. Contributions to the tool (in any form) are always welcome. − Pintoch (talk) 19:45, 26 September 2019 (UTC)

The "depicts" we add at Wikimedia Commons

The "depicts" we add at Wikimedia Commons that adds the Wikidata Q number to images, does that became integrated into the exif data when you download the image? --RAN (talk) 12:48, 27 September 2019 (UTC)

  • No. Any EXIF comes from the file itself. - Jmabel (talk) 21:06, 27 September 2019 (UTC)
Thanks!  – The preceding unsigned comment was added by Richard Arthur Norton (1958- ) (talk • contribs) at 23:06, 27 September 2019‎ (UTC).

participating team (P1923)

1994 FIFA World Cup (Q101751) have participating team (P1923). It is permitted to use in videogame items too, like Nintendo World Cup (Q1145485)? --2001:B07:6442:8903:D59C:BDEC:6508:700B 09:12, 27 September 2019 (UTC)

No. As participating team (P1923) should probably be for events only. Thierry Caro (talk) 13:21, 27 September 2019 (UTC)
Plus it does not really make sense to have real world teams stored as participants to a game where they only appear as playable teams. Thierry Caro (talk) 13:24, 27 September 2019 (UTC)
characters (P674) feels like it might be the right sort of property, but it's a bit of an odd way to describe it. Andrew Gray (talk) 13:30, 28 September 2019 (UTC)


Feel free to add yourself to this list. User:Danrok Chase me ladies, I'm the Cavalry (talk) Rama Christian Ferrer Nortix08 Andrawaag Cavernia De728631 Vladimir Alexiev , 7 January 2021 simon.letort Pictogram voting comment.svg Notified participants of WikiProject Ships

  • Is it possible to add a set of property classes to ships of ship launch date and others key start dates; launch is the key disambiguation datum for ships of the same name in en:wiki. Wikipedia is increasingly the site to go to for identifying the vessel prior to uploading images at Commons or elsewhere.

Currently it seems to be done on wikidata via significant event (P793) - see, for instance, HMS Hood (Q220239). The Hood is a good example of some or all of the key dates we need.

We have a problem on this topic (ships), there is a disconnect between Wikipedia and Commons. Wikipedia's prime identifier is the launch date of a ship, however on Commons it is the build date or commissioning date. This stems from the fact that the influential voices that were around when writing the rule book at Commons were either barge people or simple filers.

Barges and watercraft are built, ships are launched. The numbers of ships being added at the Commons register is increasing at a furious rate year upon year.

What I'm talking about here is a set of datums that apply to any watercraft, though it is most keenly required with ships, paddle boats, submarines, oil-rigs etc. It's just too big an issue for having to make a case for adding a significant event (P793) as a discrete property on a case by case basis as it applies to all ships regardless of whether they are notable or not. So it needs to go in with the original template at the time of creation of the Wikidata item even if not filled in at the time.

Surely a bot could do the job of adding this data from the header box-header templates at the top of every Wikipedia article. Broichmore (talk) 16:45, 7 September 2019 (UTC)

  • I didn't mention it, but it would also be valuable to add the IMO number of the ship. This number stays the same throughout the life of the ship even if it's name changes. Broichmore (talk) 16:57, 9 September 2019 (UTC)
  • The following is not really a comment on what Wikidata should do, but that I believe you've misrepresented what Commons intends to do. Although Commons uses the word "built," in practice the date used is almost invariably the launch date. While the launch date is certainly distinct from the period in which the ship was "built," typically it is almost immediately after the end of building. And, yes, ships are "built": we certainly commonly speak in English of where a ship was "built," as well as referring to "shipbuilders." In any case, in practice this works on Commons almost exactly like it does for buildings, for which the intended date to use is the date of completion. Commons categories don't use the keel-laying date for a boat or the groundbreaking date for a building. - Jmabel (talk) 04:33, 9 September 2019 (UTC)
Ship cats are named as follows in Commons "<shipname> (ship, <year built>)", sample: "Category:Queen Elizabeth (ship, 2010)". If the year of completion is not available or can't be found, use the year of the maiden voyage or the year of launching. Of course the preferred notion at Commons ist to ue the completion date of commissioning as the year. I do agree in practice that the great majority of ships launch date is also the completion date. Don't forget the time from launch to completion can be considerable, for warships in years of peace, and once launched work starts on outfitting and rigging etc. Broichmore (talk) 17:54, 9 September 2019 (UTC)
  • @Broichmore: Have you considered using inception (P571)? That's the date generally used for "launch" date for organizations and "launch date" is one of its aliases. ArthurPSmith (talk) 13:28, 9 September 2019 (UTC)
The problem is the shear number of ships involved. Royal Navy 1000+, US Navy 2500+, Ships by name 55,000+. Thanks to meddling by particular Admin's who don't know what they're doing those numbers are not correct. I'm guessing you could add another 20 percent at least to these cats to get a more accurate figure, possibly more. That why we need a standard all encompassing template, and a bot to do the job. Upgrading on a manual basis will not hack it. Broichmore (talk) 17:54, 9 September 2019 (UTC)

Some clarification. "Launch" can take place several years before the ship is completed. Historically hulls were erected on the building ways; the hull was launched by sliding it into the water, and the hull was towed to a fitting-out basis where the superstructure was constructed and the vessel completed. Now ships can be constructed from keel to mastheads in graving docks, and the date when water is admitted and the ship floated out is often referred to as the launch date. Kablammo (talk) 20:13, 9 September 2019 (UTC)

Please realise that the year of launching is hard to find for old ships. And different Wikipedias use different notations for the name giving of the items, so Commons had to make a choice. The Dutch, French, Germans and many others use the year of completion, not the year of launching, as it is much easier to find than the year of completion. For naval ships the year of first commissioning. It was an old discussion, the time I created Commons:Category:Ships by name, now Commons:Category:Ships by name (flat list), in 2008. --Stunteltje (talk) 19:38, 13 September 2019 (UTC)
Yes. what you say is very true. The only thing I can add, to that, is that "notable" ships from build nearly always have a launch date. It's a thing that comes with notability. Other ships become notable, because of an event in their career; sinking or war, or hijack, etc etc. The fact remains that there is a disconnect between Commons and Wikipedia.
A way to resolve that is to add to the wikidata template milestones such as keel laying. launch, build, trial, commissioning, etc.
Also in Commons most items are labelled at some stage "ship", even if they are a yacht or a brig. This is done to aid in the searching out of images from Commons. So we need a way of differentiating types in the wikidata template.
Of value too would be the number of funnels, and their configuration, bow shape, pennant number etc. This would enable a way of identifying unidentified ships.
A lot of this stuff can be found in the Wikipedia header box-header templates; which can be scraped by a bot. We may have to upgrade those templates to cope with some questions.
Resolving these questions is what I thought that wikidata was set up to do. Am I right?
The existing Wikidata template, if there is one, needs to be upgraded, because the job is too big for manually modifying by exception the existing wikidata template, which is the solution given here...
My question, now, is, am I posting this request in the right place. Where should I be asking it. Broichmore (talk) 10:44, 14 September 2019 (UTC)
If we want the way ships are handled here to change, where should I be asking?Broichmore (talk) 11:20, 19 September 2019 (UTC)
On ArthurPSmith's advice (seperate to here) when I find the time, I'm going to move this topic to make a proposal at Wikidata:Property proposal. Broichmore (talk) 19:20, 23 September 2019 (UTC)
Actually looking at the Property proposal page I'm not convinced that's the right place. Broichmore (talk) 10:40, 29 September 2019 (UTC)

Is this a compound surname (Q60558422), double surname (Q29042997) or two surnames used together ?

In Poland we have double names joined by a hyphen that are almost entirely adopted by married women. If a woman is Maria Kowalska by her maiden name and she marries a man with a surname Malinowski she can become Maria Kowalska-Malinowska (or just Maria Malinowska) after the marrige. When she divorces Mr. Malinowski she can become Maria Kowalska again. If she remarries she is free to choose any surname variant she is inclined to.

What is the Kowalska-Malinowska surname ?

Is this a compound surname (Q60558422), double surname (Q29042997) or two surnames used together ?

What is the woman's family name (P734) ?

I guess it is not a good idea to create an item for Kowalska-Malinowska. So what is the right thing to do ?

Kpjas (talk) 13:50, 26 September 2019 (UTC)

The problem is common in Israel too, especially in recent years (30 years ago most women just took their husband's surname). Also note that we need to work out how this effects Wikicommons surname caterories. I'm pinging @Jura1: who is an expert on name items. DGtal (talk) 14:13, 26 September 2019 (UTC)
In this particular case, what you technically have is the equivalent of a Swiss alliance name (Q849167). As far as I can tell, unless the compound name is itself unusually notable (I don't know of a case though), it's just listed as multiple family names.
A good treatment is at Hillary Clinton (Q6294), who is often referred to as Hillary Rodham Clinton, and birth name (P1477), married name (P2562), and family name (P734) IMO accurately reflect the situation. Compare with the treatment for Michelle Obama (Q13133), who is hardly ever referred to by her maiden name. Circeus (talk) 19:04, 26 September 2019 (UTC)
  • Interesting questions. I think there are several aspects to consider:
    • Which name is actually being used by the person, in their works or in works about them? Which is or which are the formal names? Which names can actually be referenced? Which names are used by other members of the family?
    • Which languages or country specific formalizations to consider for the above?
    • What to add to string based properties? what to add to item based properties?
    • Which labels or descriptions or sitelinks of Q60558422 or Q29042997 you refer to (I think some differ)?
Personally, I'd start out with the name actually used by the person in name in native language (P1559). There, one could qualify it, e.g. as [17]. After the string-based property, I'd add each component into P735 or P734. If there is a more specific property, use that.
We had some items created as values for P734 for people from France in the format (e.g.) "Bestel Épouse Courault". I think we should avoid this, unless the person generally makes use of that. As the person could use either, I'd add both names separately into P734. If more family members use that name, maybe "Bestel-Courault" should be created and added into P734. --- Jura 14:19, 28 September 2019 (UTC)

@jura, kpjas: Actually, after a quick investigation, I'm about 99% surte Q60558422 and Q29042997 should not be separate items in the first place. The latter exists only because of a TDKIV entry, "složené příjmení", which lists "dvojité příjmení" (the Czech label in the other item) as a synonym. Circeus (talk) 14:33, 28 September 2019 (UTC)

  • Hard to say. I think we should probably have a series of items describing various types of these. @Moebeus: is fairly active in the field. --- Jura 09:01, 29 September 2019 (UTC)

"Duplicate references" gadget hangs

When I copy a reference from one item to another, the "saving" prompt appears but it hangs and never actually saves.

If I manually add another reference to the item while the "saving" is hung, both are saved.

Is anyone else seeing this behavior? I first noticed it yesterday and it's still happening. - PKM (talk) 20:49, 26 September 2019 (UTC)

yes, I confirm... since yesterday it does not work anymore for me either - pasted refs are never saved :( --Hsarrazin (talk) 08:24, 27 September 2019 (UTC)
Me too… --Harmonia Amanda (talk) 09:07, 27 September 2019 (UTC)
  • I've opened a bug report in Phabricator for this problem. - PKM (talk) 19:29, 27 September 2019 (UTC)

Anyone else have problems with DuplicateReferences?

It doesn't work anymore. Everytime i'm try to duplicate it gets stuck while saving --Trade (talk) 11:40, 29 September 2019 (UTC)

See #"Duplicate references" gadget hangs on this page. —MisterSynergy (talk) 11:42, 29 September 2019 (UTC)

Wikidata:Property proposal/Poster vs. Property_talk:P3383

People at the first think film poster should be expanded. People at the latter are (for reasons I really don't understand) against having film poster be expanded into something actually useful. As it is, theatre posters - despite being super-common - have no category to be filed under beyond the basic image property, and the situtation is regularly getting worse, since more posters are regularly getting added. It feels like "film poster" is bizarrely restrictive, and also encourages misclassification of theatrical posters. At the least, film poster ought to be a sub-property. Adam Cuerden (talk) 05:22, 29 September 2019 (UTC)

  • What prevents you from using image (P18) for theatre advertisements and others? --- Jura 08:29, 29 September 2019 (UTC)
    • image (P18) seems awfully broad, and also doesn't seem appropriate for things that might be mostly typography. - Jmabel (talk) 16:21, 29 September 2019 (UTC)

Nobel Prize

Why is Nobel Peace Prize (Q35637) shown as "part of" Nobel Prize (Q7191) (described in English as a "set of annual international awards"), and not instance or subclass of it? Was this discussed? @TomT0m: who made the change, last month. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:05, 28 September 2019 (UTC)

My usual lecture on prizes is that, ontologically, stuff we name « Nobel prize » are essentially types of prizes. Each year people receives one instance of Nobel Prize. In that sense the Peace Nobel Prize is a subtype of Nobel Prize.
I’d really be curious about the justification of « Nobel Prize » being a « group of price ». The group of prizes would be to me the « Nobel Prizes », in plural. This is consistent with the presentation on the website : "Nobel Prizes & Laureates". But … I may be wrong. author  TomT0m / talk page 12:21, 28 September 2019 (UTC)
I agree with you. But the problem is a) none of the Wikipedia does that and b) if you tried to change it, people would also get uppity at you for not fixing all of the languages (no really; that's happened to me, and on a far less salient item). Circeus (talk) 14:36, 28 September 2019 (UTC)
none of the Wikipedia does that As far as I can tell, no Wikipedia I read formally states what a « prize » here, and Wikipedia article fails to be internally consistent (or, better, are ambiguous): You can read « the nobel price is …» then a few lines later « the nobel prizes … » (Wait what, is there one Nobel Prize or several ? Choose your side !). author  TomT0m / talk page 17:06, 28 September 2019 (UTC)
I must admit I left the stuff in a bad state, because actually there was some discussions in the process going on, and it seemed there was no actual consensus. If I recall well I had a discussion with someone and left a few messages on some user talk pages, the problem being I did not had answers so … this might have stayed as is. Happy we’re moving on. author  TomT0m / talk page 17:06, 28 September 2019 (UTC)
(later addition) @Circeus: There may be a misunderstanding here. Just in case, I did not imply that the fact that ontologically « Nobel price » may be seen as a class of price, Wikipedia or Wikidata should have an item/articel for any of its instance. We can leave perfectly happy with a statement representing the instances. The question that was answered here is wether we should use part of or subclass of. To answer this, we have to think about of the nature of all the concepts. This work should inform us on the correctness of what we are doing. author  TomT0m / talk page 09:35, 29 September 2019 (UTC)
Wikipedia doesn't really make decisions about this. It is not concerned with epistemology, set theory, modeling, etc., except in the articles about those topics. - Jmabel (talk) 20:51, 28 September 2019 (UTC)
Still I agree the Nobel Peace Prize (Q35637) should be a subclass of (P279) Nobel Prize (Q7191). Is there a reason why not? --SCIdude (talk) 16:04, 28 September 2019 (UTC)
I'm unclear if it should be subclass of (P279) or instance of (P31). Most prizes, like C.L. de Carvalho-Heineken Prize for Cognitive Science (Q278240), to pick an example, are instance of (P31) of award (Q618779) or a subclass. Do we consider Nobel Peace Prize (Q35637) to be *a* prize, or a subclass of prizes where each year is a separate prize? Ghouston (talk) 02:12, 29 September 2019 (UTC)
As a prize there is only one Nobel prize for ... Each year there is a new recipient and a ceremony. That is true for any award. Thanks, GerardM (talk) 05:22, 30 September 2019 (UTC)
This seems reasonable. There are also awards like Commander of the Order of the British Empire (Q12201477) which can have many recipients in a single year, but the award itself can be thought of as a single thing. Ghouston (talk) 06:18, 30 September 2019 (UTC)

Mountain elevations taken from (i.e. Crozzon di Lares - Q3698743 )

Hi everybody, I noticed there is a general problem with lots of mountain articles on wikidata: the elevations taken from the articles on are totally wrong.

You can see for instance here that reports in its infobox 2 elavations for the mountain (3,180 and 3,354 m). The correct one is the latter (as stated on, here), but I suppose the robot that uploaded the elevations from took in the wikidata item (here) also the first - and wrong - elevation value (3,180 m).

Do you think is possible to do something to solve this? I erased some wrong elevations from the articles of many mountains of W Alps that I know, but I guess the problem is quite whidespread.

I.e. I wonder if is possible to make a bot that erases all the mountain elevations taken from on wikidata, or that changes the wrong value with the correct one (the second - and higher - of the infobox). -- 06:45, 30 September 2019 (UTC)

Inclusion guideline of Modules

I was surprised by the creation of Q68069630 - a new item with a deprecated Module on the Cantonese Wikipedia as its only sitelink.

It seems that the current wording of Wikidata:Notability does not forbid the creation of such an item, but I wonder if we should change that. Should we edit point 3 of the inclusion guideline so that it aligns with point 2, in other words:

If a link is a module, the item must contain at least two such sitelinks, and any of them must not be a /doc page. Items for non-subpages can be created with 1 sitelink, but shouldn't be created in great numbers.

--Deryck Chan (talk) 17:52, 23 September 2019 (UTC)

Do you have any reason for why you think we should change the rule? The number of items that are created likely aren't substantial and making a rule that disallows the items means that however wants to create an item doesn't get to do what they want. ChristianKl❫ 07:50, 24 September 2019 (UTC)

Just like templates, a lot of modules are bespoke to one particular wiki. Some of them are test modules, deprecated, or fulfilled a certain function in the past and is no longer needed (like the case that brought me to this discussion). These modules have no cross-wiki value and I don't think we should mass-create items for backstage content that have no cross-wiki value. Deryck Chan (talk) 15:59, 25 September 2019 (UTC)

Q68069630 is no mass-item creation and the rule does say "shouldn't be created in great numbers" with is already a rule against mass creations. ChristianKl❫ 14:21, 26 September 2019 (UTC)
@ChristianKl: If you look at Special:Contributions/Myerbee, you can see that Q68069630 is part of a batch creation to indiscriminately add items for yue.wp pages on Wikidata. A rule against mass creations would forbid the inclusion of this module in User:Myerbee's batch creation. If Myerbee had taken the time to examine the module, they would've seen that the module was deprecated and stopped short of creating the item. Deryck Chan (talk) 16:29, 30 September 2019 (UTC)

movies parts of a film festival edition (Q27787439)

First raised at Wikidata talk:WikiProject Movies, but got no answer there :)


I wanted to model the movies that are part of a movie festival (in my case, this year’s slash Filmfestival (Q27981131) − I also have to create items for all these editions ^_^), but I have found neither guidelines nor example items where this is done. Would that be in scope? And how would we do it?

Most obvious would be has part (P527) (but if I recall correctly it has a mandatory inverse property?) I could not find any good subproperty of it (tracklist (P658) and list of works (P1455) have different domains).

What I could find in discussion on the case where a movie is first presented at a festival (such as Our Body (Q56558368) location of first performance (P4647) 2018 Toronto International Film Festival (Q55392201)) but my use-case is different.

The other domain that comes to mind would be artworks and their exhibition history (P608) ; but it feels to me that the relationship belongs more on the festival item rather than on the movie-items.

Any thoughts? :) Jean-Fred (talk) 10:39, 30 September 2019 (UTC)

We do have presented in (P5072) to indicate the inverse of this relationship. – Máté (talk) 11:28, 30 September 2019 (UTC)
  • I'm not sure if there is a answer to this. We haven't really added these and/or discussed which solution should be applied. Feel free to suggest something. --- Jura 14:08, 30 September 2019 (UTC)
    Well, I’m really not too sure whether we should create a new property, inverse of presented in (P5072) (Thanks @Máté:!), or just go with has part (P527)... Jean-Fred (talk)

Allowed units constraint

allowed units constraint (Q21514353) supports an explicit list of units. In some situations it might be appropriate to allow a whole class of units. Is this already possible? Note that there is type of unit for this property (P2876), but this is not a constraint and therefore does not produce any "lightning icons" upon violation (to my knowledge). Toni 001 (talk) 11:57, 30 September 2019 (UTC)

@Toni 001: WikibaseQualityConstraints, the extension powering the live constraint checks and icons, applies unit conversion, so Wikidata sandbox (Q4115189) Sandbox-Quantity (P1106) 1 centimetre is valid even though Sandbox-Quantity (P1106) property constraint (P2302) allowed units constraint (Q21514353) only allows item of property constraint (P2305) metre (Q11573). KrBot, the bot behind Wikidata:Database reports/Constraint violations, doesn’t apply unit conversion as far as I’m aware (but I’m not sure). --Lucas Werkmeister (WMDE) (talk) 12:12, 30 September 2019 (UTC)
@Lucas Werkmeister (WMDE): Understood. The reason I'm asking is this: area (P2046), under (property constraint (P2302) -> allowed units constraint (Q21514353) -> item of property constraint (P2305)) list 16 units, a list that will have to grow as more "areas" are being entered in different units. Instead of maintaining this list I propose that we should be able to specify that all units in the class (or subclass of) unit of area (Q1371562) are allowed. Toni 001 (talk) 14:13, 30 September 2019 (UTC)
I'm not sure if that is really desirable. People might end up using the wrong unit that just happens to be called "foot" as well. Currently, it's a non-mandatory constraint: so exceptions are possible. You could change to a mere suggestion. --- Jura 14:15, 30 September 2019 (UTC)
If some non-unit "foot" were in the class "length unit" then it should be removed and therefore allow the constraint violation to be show. Right now we are maintaining the statements which units are valid "length units", "area units", ... in two places: "instance of" and unit constraints. My proposal would eliminate the latter. Toni 001 (talk) 14:21, 30 September 2019 (UTC)
No, there can be various items for "foot", all units of length, corresponding to different length in various countries or times in history. However, most if not all users would probably use just one of them. P31 has a different function. --- Jura 14:31, 30 September 2019 (UTC)
Now I understand, I had miss that in a first reading of your reply. But both constraints could co-exists: Say the "class-based" constraint could be mandatory, while we maintain the hand-picked list of "most likely" units and make that constraint a suggestion (as you suggested) (say we include the square foot (Q857027) but not Greek square foot (Q25909387) or Roman square foot (Q25909396)). Toni 001 (talk) 14:59, 30 September 2019 (UTC)
Maybe lighthouse range (P2929) is a better example: range is usually in NM. There is a Spanish website that uses km, but I'm not really sure if they haven't mixed it up. --- Jura 15:11, 30 September 2019 (UTC)

add a url to item

How can I add (italian url) to a RMS Titanic (Q25173)? --2001:B07:6442:8903:18A2:EC04:831C:EBD 14:20, 30 September 2019 (UTC)

  • I'm not sure in what capacity you wish to add it. Is it a reference for a particular statement? If not, what is the relation? - Jmabel (talk) 16:20, 30 September 2019 (UTC)

Wikidata weekly summary #384

The consultation on partial and temporary Foundation bans just started

-- Kbrown (WMF) 17:13, 30 September 2019 (UTC)


Can anyone say why parent's brother (Q41805469) is needed, and is marked as "different from" uncle (Q76557)? How are they different? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:33, 28 September 2019 (UTC)

I'm not sure what the problem is. One is about strictly blood relatives, the other include relatives by marriage. Circeus (talk) 14:39, 28 September 2019 (UTC)
(ec) Yes, it seems the distinction here is that parent's brother (Q41805469) is specifically an uncle who is the brother of one of the parents, while uncle (Q76557) also encompasses the brother-in-law of one of the parents (ie someone married to one of the parents' siblings - an uncle by marriage). English blurs the distinction but some languages might distinguish between the two - it looks like the Danish/Swedish labels are different, for example. Andrew Gray (talk) 14:42, 28 September 2019 (UTC)
As much as I'd love to blame the Swedes, the Swedish and Danish "farbror" / "morbror" have their own items: paternal uncle / maternal uncle. The only distinction seems to be that Q76557 includes "parent's brother-in-law" while Q41805469 does not? Moebeus (talk) 01:28, 29 September 2019 (UTC)
The sv label on parent's brother (Q41805469) is "farbror", which is the same as paternal uncle (Q12158205), and doesn't match the English label. Ghouston (talk) 01:53, 29 September 2019 (UTC)
parent's brother (Q41805469) was created by Jura1 in 2017, maybe as some kind of SPARQL helper item. Some of the labels are probably just wrong. I assume it's intended to include siblings of adoptive parents, not just genetic. Ghouston (talk) 02:20, 29 September 2019 (UTC)
It's tricky to get the labels and descriptions in all languages right. Wikidata:WikiProject Parenthood/lists/kinship types is meant to help with that. Maybe a Swedish language contributor would be so kind to double-check. The sparql definitions allow to query the persons who match the definition for an individual. --- Jura 08:28, 29 September 2019 (UTC)
  • I wonder if the casual English language reading Wikidata user (en-2 "user with an intermediate knowledge of English" or en-3 "user with an advanced knowledge of English") can be expected to comprehend the difference between the two items from the current English language label and description. @Circeus, Moebeus, Ghouston, Andrew Gray:: what do you think? --- Jura 08:28, 29 September 2019 (UTC)
    • I wonder why this project seems to be only for English speaking contributors? Aint Wikimedia a worldwide organisation? Edoderoo (talk) 09:12, 29 September 2019 (UTC)
      • Who says it is? Questions on English language project chat happen to be in (or about) .. English. --- Jura 09:39, 29 September 2019 (UTC)
    • My guess is that it's probably unclear to English readers of any level what the difference is, re. the start of this discussion. Maybe employing a property like Wikidata usage instructions could help with that? Moebeus (talk) 10:31, 29 September 2019 (UTC)
      • I think label and description should convey the difference. Maybe Q41805469 needs "excluding other 'uncles' ". --- Jura 10:34, 29 September 2019 (UTC)
      • How about have one item labeled uncle (including in-laws and one item labeled uncle (excluding in-laws)? ChristianKl❫ 20:10, 29 September 2019 (UTC)
        • A solution that doesn't need re-writing Help:Label might be preferable. --- Jura 05:33, 30 September 2019 (UTC)
          • "uncle" as the label in both cases, and descriptions that make the difference clear? Then fixing or blanking the other languages? Ghouston (talk) 06:23, 30 September 2019 (UTC)
You should see the list of relationship options between contacts in iOS 13.. It's insane the level of detail you can get into: "aunt (father's younger brother's wife)" there are some 180 possible labels for a relationship (Edit a contact, choose "related name", then click the label, then choose "All labels"). I think people care a lot about the accuracy of such statements. So they have "uncle" which in their model can be with or without in-laws, but also someone who isn't even your uncle, but just someone you 'call' uncle. And then they have more accurate terms for 'if you (or your language) cares about more detail' where needed, that term gets a 'technical disambiguation' in a label. It's very much speaking language based, mostly because it primarily is information used for (and learned by) Siri. TheDJ (talk) 11:05, 30 September 2019 (UTC)
"Aunty and uncle are terms of address for older people, to whom the speaker may not be related.", from en:Australian Aboriginal kinship. Reuse of words for different concepts. Ghouston (talk) 22:06, 30 September 2019 (UTC)

Wikidata and Wiktionary

Why can't we connect the Wikidata with the words of Wiktionary? I tried it, but the system refused. I don't really know why. Can anybody explain it to me, please?--Porbóllett (talk) 19:08, 27 September 2019 (UTC)

I understand what you say, but on the left side I see: Creating new lexeme link. For me the element (of Wikidata) is crystal clear. But I tell you the truth, I don't know what lexeme really is.--Porbóllett (talk) 17:28, 30 September 2019 (UTC)

I tell you why the connection is fairly helpful for the Wiktionary projects. On an average Wiktionary page you can see the word written on different languages. In many cases this list doesn't contain all language versions what we have now on Wikipedia. But many titles on the Wikipedia projects are exactly the same words which are missing from the dictionary. So when the whole system will be connected by Wikidata, these lists would be completed by the Wikidata connections to other Wiki-projects and filled by the system itself. As we can see this method on Wikipedia pages, where infoboxes connected to the Wikidata and the changes on the WD page changing the whole connected Wikipedia infoboxes. I know this will be working one day. Though it seems a bit difficult for now, but I think in the future of this project-system this would be ordinary for all of us.--Porbóllett (talk) 17:44, 30 September 2019 (UTC)

See Wikidata:Wiktionary for more. —Justin (koavf)TCM 18:00, 30 September 2019 (UTC) ChristianKl❫ 09:07, 1 October 2019 (UTC)
Additionally, @Porbóllett: we do have Cognate extension to handle interwiki links of Wiktionary mainspace pages. The only known exception is Wikimedia main page (Q5296) which you should add Wiktionary main page links to this item. --Liuxinyu970226 (talk) 23:22, 1 October 2019 (UTC)