Wikidata:Project chat/Archive/2020/01


Добавьте . Алёна Пескова (talk) 14:09, 1 January 2020 (UTC)

  •   Done. You can write sush messages at project chat in Russian. (Вы можете оставлять подобные сообщения на русскоязычном форуме Викиданных, если это будет Вам более удобно. --Ksc~ruwiki (talk) 14:26, 1 January 2020 (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. Ksc~ruwiki (talk) 14:26, 1 January 2020 (UTC)

Quarry query to get edits made by an user

Hi all,

I'd like to take the liberty of bringing up a subject I addressed on the Request a query page. I would like to have the list of items (subclasses of place of worship (Q1370598) in France (Q142)) for which the user Peter17-Bot add the property coordinate location (P625) (e.g.

Tagishsimon proposed to use Quarry, but he didn't know exactly how to do, just like me. However, he found a Quarry query not very dissimilar.

So if anyone knows how to do that, it will be very helpful! I think it's a case that we will see less and less, with the usage of EditGroups.

Thank you, Tubezlob (🙋) 10:22, 2 January 2020 (UTC) (and happy new year! 🎆)

This section was archived on a request by Tubezlob (🙋) 18:32, 2 January 2020 (UTC) I did the work manually, after discussion on the Request a query page

Copy statement types between articles

Hi! I started editing Wikidata items manually yesterday (I'm quite active on English Wikipedia, making mostly small edits for grammar, HTTPS, references, etc.) and have a question: is there any way I can do stuff faster? I've made in a few new items from pages in the category wikipedia:Category:Articles without Wikidata item, and it feels really slow. There are many statements that should have a value for basically ever member of a group; for example, all water polo players should have a country they play for, should probably have a date or year of birth, maybe have a competition they were in, etc. Is there any way to copy a group of statements from one item, then edit the values for another article, maybe a way to copy the source code if that exists? Thanks! DemonDays64 (talk) 01:13, 1 January 2020 (UTC)

@DemonDays64:, check out the WE-Framework gadget --Trade (talk) 01:25, 1 January 2020 (UTC)
@Trade: thanks for the tool, though it's not really what I was thinking—I was hoping to do what one could do with copying Wikitext between articles and changing the values on Wikipedia, rather than using a limited structure template. DemonDays64 (talk) 01:47, 1 January 2020 (UTC)
@DemonDays64: I haven't used WE-Framework, but there is also moveClaim.js, which allows single claims to be copied or moved to other items. -Animalparty (talk) 01:37, 1 January 2020 (UTC)
@Animalparty: oh, that seems useful. Where/how would I install it? It doesn't have instructions. Thanks DemonDays64 (talk) 01:49, 1 January 2020 (UTC)
Make a common.js page, like User:Ghouston/common.js. Ghouston (talk) 02:20, 1 January 2020 (UTC)
Thanks. This points to a pervasive problem in Wikidata: aspects of it are often maintained by and tailored to people deeply immersed in coding, scripting, and programming, and often poses a high entry barrier to users who don't speak robot. @DemonDays64: To use moveClaim.js, create the page User:DemonDays64/common.js, then add the text importScript( 'User:Matěj Suchánek/moveClaim.js' ); to the first line. Other tools and gadgets can be activated by your Preferences, or found at Wikidata:Tools. Structurally, Wikdata is still in its infancy. Happy New Year! -Animalparty (talk) 04:32, 1 January 2020 (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. Ghouston (talk) 03:21, 3 January 2020 (UTC)

single best value constraint

Arsos, Larnaca (Q56448937). At property population (P1082) there is "single best value constraint".

"This property should contain a single “best” value with the same determination method qualifier. Of the current multiple values, one should be marked with “preferred” rank."

But, due to political issue of Cyprus (Cyprus dispute (Q245831)), there is not census after 1973. And is useless to add as best value the 1973 census. Maybe North Cyprus made a census but no data added to Arsos, Larnaca (Q56448937) or to Arsos, Larnaca (Q4796610). Xaris333 (talk) 18:35, 27 December 2019 (UTC)

sounds like a good case for marking the item as “exception to constraint”. - PKM (talk) 19:23, 27 December 2019 (UTC)
But there are almost 200 items with the same issue. Xaris333 (talk) 19:54, 27 December 2019 (UTC)
What do you mean by no best value? Are there multiple estimates and there's no way to decide which is better? Ghouston (talk) 23:27, 28 December 2019 (UTC)
By "best" we can just mean most recent, so on Arsos, Larnaca (Q56448937) the best is 1973. Ghouston (talk) 23:29, 28 December 2019 (UTC)
But that has no sense. We are almost in 2020 and the best value is 1973? And there was a war in 1974. The population of that areas left. Xaris333 (talk) 16:15, 29 December 2019 (UTC)
It may not be a good value, but it is still the best one we have. Circeus (talk) 06:51, 1 January 2020 (UTC)

Can Q26449717 and Q15621234 - both for family name Arai - be merged ?

Hi, Q26449717 and Q15621234 appear to be for the same item (= Japanese family name Arai), with both of them being assigned as Property:P734 in Q609469 for example, and thus creating a Defaultsort-conflict on Commons (c:Category:Akino Arai).
But as my knowledge of Japanese language is somewhat limited, I might be missing an aspect that requires this to be two independent Q-items.
I would be grateful to get suitable input from people with more insight & knowledge whether these two Q-items indeed can be merged, or not.
Carrying out the merge-operation itself will be no problem for me to do.
-Archie02 (talk) 22:46, 29 December 2019 (UTC)

No. Both of them are not same item. One is written in Hiragana and the other is written in Kanji. --Afaz (talk) 02:39, 30 December 2019 (UTC)
If you use "Arai"(Q15621234) or "あらい" (hiragana), this includes "荒井"clan and "新井" clan and "新居" clan and "相良" clan. if you use "新居"(Q26449717), this means only "新居" clan. thanks--Afaz (talk) 03:00, 30 December 2019 (UTC)
Indeed. In the sinosphere's context, this is the same reason we have separate items for McNeil (Q16876719), McNeill (Q4275864) and MacNeil (Q21452079). These are essentially names with distinct spellings. Circeus (talk) 06:59, 1 January 2020 (UTC)

Correct qualifier for P4724

Maitron ID (P4724) is a combined directory of biographical entries across multiple encyclopedias (e.g., women, Africans, anarchists; see website [under "dictionnaires"], each published as separate volumes). I'd like to add entry's associated encyclopedia on WD as a qualifier to the P4724 property for easier lookup, but what would be the correct qualifier to use? described by source (P1343) throws an error when used as a qualifier on the property. czar 01:33, 1 January 2020 (UTC)

This sounds like something that would be better done as a refernece using stated in (P248). As a qualifier, statement is subject of (P805) is perhaps the next best thing. --Tagishsimon (talk) 13:53, 1 January 2020 (UTC)

Fix request

Please go to Europe (Q46) and correct eurpei to europei (masculine plural of Italian) on denomyn section. Thanks! -- 12:17, 1 January 2020 (UTC)

@   Done! DemonDays64 | Talk to me 15:16, 1 January 2020 (UTC)

How do you edit statements on mobile?

Hi! I’m on an iPhone, using Safari (advanced mode on). At the top of each item, there’s a little edit button, but that only allows editing the description, aliases, and name of the item. How can I edit the actual statements on a mobile browser without switching to desktop view? DemonDays64 | Talk to me 05:49, 1 January 2020 (UTC)

I don't think you can at the moment because the desktop interface doesn't translate very well to mobile. More information is available on this issue on Phabricator (the bug tracking software for Wikimedia projects): T95878. Husky (talk) 22:52, 1 January 2020 (UTC)
@Husky: dang, that sucks :-/. Thanks for the reply, though! DemonDays64 | Talk to me 01:39, 2 January 2020 (UTC)

Britannica does not have an article

Related talk: Wikidata:Project_chat/Archive/2016/11#Encyclopædia_Britannica's_'empty-ish'_concepts_in_Mix'n'Match

Currently, we have a problem with Encyclopædia Britannica Online ID (P1417) property which is used in Wikpedia projects. The case is the following. Articles in Wikipedia use P1417 as source to a topic of considered article and provide link with URL to Britannica. Readers click to the link with expection to find an article on particular topic but they found a message like Britannica does not currently have an article on this topic. It leads to user complains: readers consider Wikipedia as a project with less quality, Wikipedia editors start to think that Wikidata is less reliable source.

The problem is known, and as I see (link to talk above), some guys already tried to resolve the problem in the same way.

I tried to start to solve the problem via bot processing to mark no-article links as deprecated. As a result, Wikipedia stopped to capture such links and readers do not see no-article links. However, GZWDer stopped my activity because it may be controversial and s/he does not support it. The question is, how we are going to solve the problem? I see two scenarios here: 1) We mark no-article links somehow (example, as deprecated) and Wikipedia stops to use that. Readers become less unhappy, we keep more links via Wikidata in Wikipedia, editors consider wikidata as more reliable source. 2) We do not change anything in Wikidata, we exclude all of P1417 links from Wikipedia articles, readers loose links to Britannica, editors consider wikidata as less reliable source.

Which path will we choose? Bsivko (talk) 00:19, 30 December 2019 (UTC)

Thanks for bringing this up. One anecdote: When I was trying to determine the precise meaning of some Wikidata item (in the science context) - to add a described by source (P1343) statement - I read some Wikipedia pages and clicked some external ID links to check whether they all are about exactly the same concept. But on said encyclopedia I found no identifying information. So I felt like removing this link (but did not do it). End of anecdote. In the discussion linked above support for keeping such links is justified with 1) the property storing identifiers and 2) an analogy to thesauri. About 1: What's the point of an identifier if it has no associated metadata? Just imagine a Wikidata item without labels, descriptions, statements, or any wiki links. About 2: Thesauri usually list synonyms and therefore help in understanding what a term is about. I'm not sure I would call a random list of words a thesaurus. Therefore I think that a) those values being identifiers is questionable and b) could even be considered spam. So, I tend to think that those values should be dealt with (deprecation or removal). Toni 001 (talk) 08:44, 30 December 2019 (UTC)
In 2016 this was discussed. In my opinion I propose to tag them with a new property, see Wikidata:Property proposal/type of external page.--GZWDer (talk) 09:20, 30 December 2019 (UTC)
Not to argue, but to present some contrasting opinions:
  • If I click on such a link, I do not have an "expectation to find an article on a particular topic" -- I expect to find their page on the same concept. If it's a "directory" page rather than an "article", that's perfectly fine, as long as it's the identical concept.
  • I don't in any way see this as a reason to "consider Wikipedia as a project with less quality" or "think that Wikidata is a less reliable source". We have an article and they don't -- if anything this means we're more comprehensive / higher quality / more reliable! —Scs (talk) 14:23, 1 January 2020 (UTC)
  • We can think anything, but in practice, readers complain. Bsivko (talk) 08:29, 2 January 2020 (UTC)

Shipibo language code

Hi, I was trying to introduce the statement name in native language for Guillermo Arévalo Valera (Q4068966). I couldn't as Shipibo-Conibo (Q2671988) is right now not a Wikimedia language code (P424). Shipibo-Conibo is the 5th most spoken language in Peru. Is this something that can be resolved fast? ISO 639-3 (Q845956) for Shipibo-Conibo is shp. Thanks in advance for the answer. --Cbrescia (talk) 19:04, 30 December 2019 (UTC)

IIRC that's handled on the side of mediawikiwiki:, not WikiMedia. Presumably you only have to create a Phabricator: ticket and wait for the next MediaWiki version offering shp. Or ask for help on metawiki:, they have i18n experts. 8000 speakers might be not enough for the purposes of MediaWiki.
Apparently Shipibo-Conibo (Q2671988) isn't known on enwiki, DAB w:en:shp doesn't mention it, and the IETF language tag (P305) reference here sucks, it should be the IANA registry. – 15:00, 2 January 2020 (UTC)

WikiProject India Newsletter #2

Hi Project chat/Archive/2020,

Happy New Year for those who are living in UTC+05:30 (Q6828) (in advance or belated for the rest!). You are receiving this message as you are one of the participants of WikiProject India on Wikidata and/or a subscriber of the Wikidata India newsletter. You can find our second issue of the quarterly newsletter here, where you can get a quick overview of the Wikidata activities related to India over the last 3 months. If you do not want to receive this kind of notification further, you can remove your username from here.

MediaWiki message delivery (talk) 18:46, 31 December 2019 (UTC)
(on behalf of WikiProject India on Wikidata.)

I think it would make more sense to add this to the weekly summary then to add it as it's own topic for the project chat for future editions. ChristianKl❫ 08:06, 2 January 2020 (UTC)
  Resolved by ChristianKl, thanks. – 16:25, 2 January 2020 (UTC)

Europeana <-> WD connected. Next step create Wiki templates in more languages

I have connected 160 000 WD objects Europeana entity (P7704) with en:Europeana now we need Wiki templates in more languages. We have it up running on sv:Wikipedia as Mall:Europeana / usage and I did a test on en:Wikipedia see link If you think en:Europeana adds value to Wikipedia dont hesitate to add a template in "your wiki"

  • Task T241677 about creating more Wiki templates
  • Task T240290 connecting WD with Europeana

- Salgo60 (talk) 12:43, 2 January 2020 (UTC)

What to put when I know someone was born in Germany, but not which Germany?

I'm trying to edit in a place of birth for Otto Kirchner (Q80683451). According to this source (ctrl F for "Kirchner, Otto), he was born in Germany in 1846. However, I can't figure out how to enter that without error; Germany was not one thing in 1846, so I don't know if he was born in the Wikipedia:German Confederation or whatever, and even if he was, what part (a necessary field) he was born in. Please help! Thanks. DemonDays64 | Talk to me 20:24, 2 January 2020 (UTC) (please ping on reply)

@DemonDays64: According to [1] he was born in Frankfurt (Oder) (Q4024), which thus was in Kingdom of Prussia (Q27306) at time of his birth. Ahoerstemeier (talk) 21:02, 2 January 2020 (UTC)

Restoring incorrectly resolved redirects

Someone merged member of the Swedish Riksdag (Q10655178) into member of the Folketing (Q12311817), which was incorrect and I have restored it now. Unfortunately User:KrBot "resolved" all those redirects before that, so now all Swedish MPs have been turned into Danish MPs. This also affected all other links to Q10655178, including elections and the speaker of the house. As this concerns hundreds of items, could someone restore them in an automated fashion? Väsk (talk) 23:41, 2 January 2020 (UTC)

I think the bot's operator, @User:Ivan A. Krestinin:, does this when required. Ghouston (talk) 05:40, 3 January 2020 (UTC)
It is a VERY bad idea to modify the links poiting to the redirect exact for this reason - especially if it is done just few hours after the wrong merge. Ahoerstemeier (talk) 08:37, 3 January 2020 (UTC)
Redirects need to be resolved at some point as results were otherwise missing in queries. —MisterSynergy (talk) 09:24, 3 January 2020 (UTC)

Property for "shut down", as in a website?

Dot Earth (Q5299386) is about a blog that shut down a few years ago. What property is appropriate for a website's time of shutting down, as an opposite of inception (P571)? I can't find one under "shut down" or something similar. Thanks! DemonDays64 | Talk to me 04:14, 3 January 2020 (UTC) (please ping on reply)

There is dissolved, abolished or demolished date (P576), but I'm not sure if the blog counts as "shut down" or not. It stopped receiving updates, but is still online. If it was a book, we'd say it was still in publication. Ghouston (talk) 05:35, 3 January 2020 (UTC)
How about using significant event (P793) as in [2]. Ahoerstemeier (talk) 14:46, 3 January 2020 (UTC)

Find a list of subproperties of a property

In my quest to find the equivalent of inception (P571), a subproperty of start time (P580), I tried to find the subproperties of end time (P582) to see if one of those would be appropriate as an answer to Wikidata:Project chat#Property for "shut down", as in a website?; however, I cannot figure out how to do this. How can I find all subproperties of a given property? Thanks!  – The preceding unsigned comment was added by DemonDays64 (talk • contribs) at 04:20, 3 January 2020‎ (UTC).

I'm no SPARQL expert, but something like:

SELECT ?item ?itemLabel 
  ?item wdt:P1647 wd:P582.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }

(although that won't do recursion if any of these have subproperties of their own.) - Jmabel (talk) 07:17, 3 January 2020 (UTC)

Based on the return of that, date of the latest one (P7125) may be what you want. - Jmabel (talk) 07:20, 3 January 2020 (UTC)

János Bobula (needs help from hungarian speaker)

items Q832773 and Q832771 seem to have some data that crosses each other. But because most the information is in a foreign language, I don't want to risk doing it on my own. I was hoping someone (who speaks Hungarian) could look and fix the two items (and their commons and wp articles). They may even be father and son. So, you know, adding that kind of data to their items would be great. Quakewoody (talk) 16:54, 3 January 2020 (UTC)

2020-01 split

Should Islamic court (Q3538915) be split into disambiguation page and court sub-class? Visite fortuitement prolongée (talk) 21:28, 3 January 2020 (UTC)

Plural of units


in the last time when I read the properties and the values of items I have seen that if the property has a value like year or a weight then in some languages it is another form of the unit if it is one as if it is more than one. For example at a age in years. As far as I know after I have asked it a few weeks agos are the units items. Are there ways to show the correct form of a unit. This is something what were also interesting for persons. I think in German and also in other languages there are differnt names of the same occupation for man and woman and if you could show there the more correct one that were great. -- Hogü-456 (talk) 18:12, 26 December 2019 (UTC)

It really bothers me when duration (P2047) tells me that a song lasts 220 Second and etc. Hope there's a way to correct that. --Trade (talk) 23:28, 26 December 2019 (UTC)

  • This task is basically about choosing the right lexeme to represent a unit. It's possible to implement this properly but it likely takes some work and there are many areas in Wikidata that take work. ChristianKl❫ 12:30, 28 December 2019 (UTC)

And genetive case (5 минут, neither "5 минуты" nor "5 минута"), and paucal (Q489410) for a small amounts (5 килограммов, but 2 килограмма)! So it's not a single Lexeme to choose, ChristianKl... --Infovarius (talk) 19:26, 3 January 2020 (UTC)

  • As Infovarius above, many languages have more than two cases (singular/plural). E.g. Russian and Polish are similar, but there may be languages having more nuances. Wostr (talk) 19:42, 3 January 2020 (UTC)
You basically have to look up the noun that has "subject item" for the unit in the language that you care about, then chose the form of the lexeme based on the grammatical features at which you are looking in the particular case. ChristianKl❫ 16:16, 4 January 2020 (UTC)

Featured portal in en wiki

In en wiki the featured portal process has been discontinued. en:WP:Village pump (proposals)/Archive 138#RfC_about_marking_the_Featured_portals_process_as_.22historical.22.

How could wikidata remove featured portal Badge from interwiki links?Guilherme Burn (talk) 17:52, 3 January 2020 (UTC)

Guilherme Burn There are 121 items marked as featured portals on enwiki - it should be straightforward to just run a bot to remove them all.
Looking at the discussion there, there was a suggestion for some kind of "not quite featured" badge, and the portals do still seem to be tagged in some way - eg the hollow star at w:Portal:American Civil War. We could replace the badges with a more specific one to reflect this. Andrew Gray (talk) 12:52, 4 January 2020 (UTC)

Assistance on Merging

I recently found that both Q18841134 and Q14934146 are all about the Convention for the Suppression of Unlawful Acts against the Safety of Maritime Navigation, and I wish to merge the former item to the later item. Can anyone assist on this? I am not familiar with here so I appreciate any assistance for this.廣九直通車 (talk) 08:18, 4 January 2020 (UTC)

Sparql query for located in the administrative territorial entity (P131): broad/narrow

Hello all. In this query listing organizations in the US, how to return a column for US states (not towns)? For example, an organization in Albany, New York, would show up in the query results as "New York" (not "Albany"). How to do this? Suggestions welcome. Thanks. -- M2545 (talk) 20:16, 4 January 2020 (UTC)

@M2545: This, I think. Note also Wikidata:Request a query ;) --Tagishsimon (talk) 20:46, 4 January 2020 (UTC)
@Tagishsimon: T h a n k s. Much appreciated! -- M2545 (talk) 20:59, 4 January 2020 (UTC)

Not included if the property in using as reference property

Hello. title (P1476) has single value constraint (Q19474404). At property talk page there is "List of this constraint violations: Database reports/Constraint violations/P1476#Single value, SPARQL, SPARQL (new)". But I think these lists not include results that the property is used for reference. Xaris333 (talk) 20:41, 4 January 2020 (UTC)

User warnings

Are you doing something like {{uw-disruptive1}} on IP talk pages more than one week after the fact? – 10:12, 28 December 2019 (UTC)

There are no strict rules on this, WD:Vandalism page just says that "warnings can be left". And since we don't have tools like Twinkle, the task appears to be tedious and is usually neglected, unless the vandal is really restless. IMHO, there's no much point placing such warnings for incidents older than several days, chances are that the situation will not recur. If it does, there are still evidences of the past and still the appropriate measures can be taken. (It would be great if we had a more developed system like that of enwiki, but that would require some level of automatization). — Mike Novikoff 21:45, 3 January 2020 (UTC)
  The enwiki system isn't perfect, some substituted templates have to be manually fixed if an IP uses them. – 09:25, 5 January 2020 (UTC)

Need an expert on Swedish parishes

Are Anundsjö parish (Q2857809) and Anundsjö parish (Q16496957) labeled properly as to which is the contemporary parish and which is the historic parish that was split/joined. --RAN (talk) 18:28, 5 January 2020 (UTC)

Link to subsection

Hello. The Q10517184 correctly directs to articles in Portuguese, French and Swedish. However, it links to a disambiguation page in English. Is there a way to link a Q-code to a subsection of an article? Instead of linking to en:Show of hands, it should link to en:Voting methods in deliberative assemblies#Show of hands. Thanks. Mateussf (talk) 16:40, 5 January 2020 (UTC)

  Done. Someone else fixed it. Now it links to a redirect, so it works perfectly. Thanks. Mateussf (talk) 19:41, 5 January 2020 (UTC)

Pico da Neblina

Hola. En el "pico da Neblina" en la declaración "situado en la entidad territorial administrativa" correctamente se recoge "Amazonas" (estado de Brasil). Sin embargo, cuando la Ficha de Montaña de la Wiki española extrae el dato aparece "Amazonas (Colombia)". No logró dar con la solución. Gracias y un saludo. --Urdangaray (talk) 10:51, 30 December 2019 (UTC)

Is this still an issue? / O problema ainda existe? Mateussf (talk) 21:00, 5 January 2020 (UTC)

  • Error in replyto template: Username not given. See Template:Replyto for usage. Doesn't look so to me. - Jmabel (talk) 06:20, 6 January 2020 (UTC)

Has part

Hello. If a administrative territorial entity is divided to sup-administrative territorial entities, on the entity we are using located in the administrative territorial entity (P131). Should we also use has part (P527)? For example, Paphos District (Q59133). I have added both properties, not sure if the second is necessary. Xaris333 (talk) 21:16, 5 January 2020 (UTC)

For ATEs, contains administrative territorial entity (P150) is the appropriate property, not P527. --Yair rand (talk) 03:40, 6 January 2020 (UTC)

Copyright on a list of formulas?

Hi. In the context of defining formula (P2534) I've been wondering about the following question: How many of the formulas of a reference work can be entered into Wikidata, and under which conditions? As an example, see the defining formula of the magnetic flux density. The formula itself is simple enough to not fall under copyright (at least that's what I understood from reading about copyright pertaining to formulas in Germany and the USA). The formula also has a reference. That reference (a part of ISO/IEC 80000 (Q568496)) consists of tables of (physical) quantities, units and defining formulas. Using a suitable SPARQL query that takes into account the references one can generate a list of all defining formulas - ordered by item number - in that reference. I'm wondering whether the result of this SPARQL query, which very closely reproduces one column of the reference, has any relevance in copyright, say in the context of "copy of a significant fraction of a work" or the European database copyright. One relevant point to consider is that, while in the context of (physical) quantities, as of today, the source mentioned above is the most frequently used (based on a sample of a few hundred items that I've reviewed), in school I learned the formula with exactly those letters (F for force, ...) and in that form, and there are many other sources that could (and hopefully will be) added in the future to exactly those formulas.

From this discussion I wish to understand whether it is safe with respect to copyright to continue adding formulas by systematically looking them up in a reference work. Thanks. Toni 001 (talk) 12:34, 4 January 2020 (UTC)

Results of a SPARQL query can be sorted by the item ID, so it's possible to recreate an ordered sequence (which is subject to database copyright if I understand correctly) only IF 1. items were created in that order, AND IF 2. the created item IDs were increasing (which is a system question I think but usually they are with manually created items). So IMO yes, if you want to avoid problems do the import in a random manner and don't use external IDs which contain the copyrighted order. But I'm no expert. --SCIdude (talk) 14:48, 5 January 2020 (UTC)
Thanks for the explanation so far.
For the source in question Wikidata already contains most of the quantities; I've been adding a few for completeness and because of a structural need (use in calculated from (P4934) statements). So the Wikidata Q-numbers are indeed in an "arbitrary" order. But then (and that's what I had in mind above), each reference is accompanied by a section, verse, paragraph, or clause (P958) entry containing the reference's item number ("6-21" in the example above). Ordering by this entry (which requires some SPARQL coding) makes it possible to sort just like the reference. This would then - following your explanation - fall under copyright, correct? If that's the case, would removing the item numbers (and using page numbers instead) solve the problem? Another way out might or might not be to say that the reference contains next to each formula a precise textual description, which is missing in the SPARQL result.
Yet another explanation is that matching the reference's item number to Wikidata's item number is a "creative" process in that in some cases it requires deep knowledge of the subject (say in electromagnetism), discussions with fellow editors, ...; this process creates a link set which might receive copyright protection. This linkset is then released into the public domain by the editors who enter it into Wikidata. Toni 001 (talk) 16:21, 5 January 2020 (UTC)
Creating these external IDs that contain the reference's item number in my opinion amounts to an import of that data(base). And the rules for import are clear, see:
I don't know about the procedure if you find copyrighted material on WD. Maybe others can help? --SCIdude (talk) 07:12, 6 January 2020 (UTC)
On a practical note, I see no use having external IDs that just recreate a specific order. However if you mean external IDs solely used as reference in a reference statement I'm not sure if these touch any database copyright (because that's what the external DB has as interface to the outside). Full import would be if the external database's entities are recreated as items here, with all accmpanying information. Please correctme if I'm wrong. But effectively, why have external IDs other than for reference if the items themselves are open formulas? If you have two external IDs on every formula you just created a mapping between those databases, which might even be own work from curation with you yourself (=WD) the owner. --SCIdude (talk) 07:29, 6 January 2020 (UTC)
Thanks for the useful links. About the purpose of the section, verse, paragraph, or clause (P958) values: Those are included "just" to provide a "good" reference, in the same way that other references might include a page and a paragraph number. But then, they can additionally (by not being random strings - a choice made by the source) be used for sorting: On the one hand I found this useful to compare Wikidata and the source side-by-side, but on the other hand made me wonder about copyright and start this discussion.
The documents linked to above (and articles linked therein) mention "repeated and systematic extraction" of "insubstantial parts of the contents of the database": If I understand that correctly then creating new items and filling out all the data (formulas, symbols, units) for a substantial fraction of entries might be problematic - no matter whether done one by one. However, this is not what happens: Thanks to the many volunteers who have been writing Wikipedia articles since long before Wikidata's birth, followed by the (automatic) creation of Wikidata IDs for each existing Wikipedia article, we already have items for most entries in the source. The first process then is to find out which Wikidata item corresponds to which item in the source; the result of this matching can be stored with a described by source (P1343) statement with section, verse, paragraph, or clause (P958) qualifier. That's not simply an import, as matching of, say, labels is not reliable; in some cases I spent more than an hour reading about the subject to match one item. Once that is done, I might add the same source (again with item number for completeness) to the formula or even enter the formula from the source, if it did not exist yet.
In summary, looking at each individual class of steps and how the source is used in each step, I don't see any "substantial copying". Toni 001 (talk) 12:43, 6 January 2020 (UTC)
Correct, you are mapping the Wikipedia (database) to that reference which is curation, a creative process that can't be automated. But I would like to hear a more legal opinion on databases that carry open data. --SCIdude (talk) 13:53, 6 January 2020 (UTC)

How should forms be tagged with grammatical features?

If I have a lexeme, e.g. perro (L34109), that needs four forms: perro (masculine singular), perros (masculine plural), perra (feminine singular), and perras (feminine plural). I don't understand what grammatical features to mark these as having, though; there's ones for masculine/feminine and ones for plural/singular, but there are also single ones for masculine/feminine singular/plural. How should these be used? Should the form perro be tagged as masculine singular and masculine singular? Should it only have the first two? Should it only have the last? Should something be merged? I'm confused. Thanks! DemonDays64 | Talk to me 15:59, 5 January 2020 (UTC) (please ping on reply)

@DemonDays64: Consensus at Wikidata talk:Lexicographical data#Use of combined grammatical features like “second-person plural” (permanent link) was that the combined items should be avoided, and eventually deleted. --Lucas Werkmeister (talk) 11:53, 6 January 2020 (UTC)

Contemporary constraint

We have it set up so that we get a contemporary constraint error when a child is born after a parent is dead, we need to allow a child to be born up to nine months (lets make it 10 months) after a father is dead. We have several situations where a soldier inseminates his wife before going off to war, where he dies. Can we do this? --RAN (talk) 19:11, 5 January 2020 (UTC)

Don't forget sperm donors and in vitro fertilization. Given that humans have been born from sperm frozen for 40 years, the constraint may need to be relaxed even further, or at least an appropriate qualifier added. -Animalparty (talk) 20:55, 5 January 2020 (UTC)
Surely individual exceptions would be sufficient in this case? Circeus (talk) 17:12, 6 January 2020 (UTC)

Wikidata weekly summary #397

Formatting problem with P244 Library of Congress Authority ID

I reported this already on the property talk page, but in case it doesn't get seen there ... It appears that newly created Library of Congress Authority ID numbers are incompatible with the current format regex because they start with 2020 and the regex says that the third digit of the year must be 0-1. So the regex needs to be fixed to be compatible with new and future numbers.  – The preceding unsigned comment was added by Silentdivas (talk • contribs) at 17:10, 7 January 2020‎ (UTC).

  Done --Lucas Werkmeister (talk) 17:59, 7 January 2020 (UTC)

Sidebar cleanup

Can we get rid of the "Print/export" section of the sidebar? I don't think Wikidata items typically have any use for printing, exporting as PDFs, or being placed in books. Also, the "Cite this page" bit doesn't really seem necessary. Attribution for Wikidata items isn't strictly required, and it's unlikely that anyone's going to use that citation style in the first place. --Yair rand (talk) 21:18, 6 January 2020 (UTC)

  • "Cite this page" seems quite useful if someone wants to cite wikidata in an academic paper. - Jmabel (talk) 22:16, 6 January 2020 (UTC)
    • Is it likely that an individual item will be cited? --Yair rand (talk) 18:04, 7 January 2020 (UTC)
      • There are papers that discuss Wikidata's ontology, so significant classes like Q5 have some likelihood of being cited directly. ArthurPSmith (talk) 19:35, 7 January 2020 (UTC)
        • In such situations, the items themselves are the topic of discussion. Is it at all likely that such a paper would make use of Special:CiteThisPage? Is it used sufficiently frequently that it's worth using up a spot in the sidebar for that link? --Yair rand (talk) 16:56, 8 January 2020 (UTC)

(Q81105988) Summerperiod

In Q81105988 I have a period where I define the start and enddate. However I cannot find the correct value for summertime period (railways). Are there any examples of timetableperiods?Smiley.toerist (talk) 22:48, 7 January 2020 (UTC) In Q81105988 I have a period where I define the start and enddate. However I cannot find the correct value for summertime period (railways). Are there any examples of timetableperiods?Smiley.toerist (talk) 22:48, 7 January 2020 (UTC)

Your valid in period (P1264) idea with value duration (Q2199864) makes no sense for me, it doesn't match what P1264 says. How about using duration (P2047), suggested as "See also" in P1264? Unrelated issue, it was certainly summer (Q1313) in Belgium, but for a railway timetable I'd expect something independent of geography. Admittedly I found nothing remotely related to "Q2+Q3". – 12:44, 8 January 2020 (UTC)

You can test a new feature to highlight mismatched references

Hello all,

As part of our work on data quality, and how we can make sure that data and references match, we have been working on a new notification to notify editors when they are editing values without updating the reference.

You can learn more about the feature and give feedback on this page. For now, the feature is deployed on Based on your feedback, we may be able to enable it on in February 2020.

If you have comments or questions, please add them to the dedicated talk page. Thanks, Lea Lacroix (WMDE) (talk) 12:39, 8 January 2020 (UTC)

check my merge?

Hi, I just recently merged a Q for an enwiki disambiguation page with a Q for the thing it was about at Q16659419. Did I do the right thing, or are disambiguation pages supposed to have their own Q #s? Thanks, Rachel Helps (BYU) (talk) 22:59, 8 January 2020 (UTC)

@Rachel Helps (BYU): The latter, I'm afraid. I've reverted your merge. The disambiguation item subject is a disambiguation page pointing (sort of) to an article on the class and to articles on instances of the class, but which is not itself a description of the class. The railway hotel item subject is the class. Thanks for checking it out here, here. General rule; disambiguation items should point to DAB pages and should not seek to be definitions for the class of things that are the subject of the DAB article. --Tagishsimon (talk) 23:16, 8 January 2020 (UTC)
@Tagishsimon: Thank you! Rachel Helps (BYU) (talk) 23:20, 8 January 2020 (UTC)

Last includes not displayed on Wikidata:Property proposal/Authority control due to "template_include_size_is_exceeded"

Hello, the page Wikidata:Property proposal/Authority control is currently included in the category Category:Pages_where_template_include_size_is_exceeded, so at the moment the about last 10 proposals are not displayed as include, but rather only the link to the proposal is shown. In addition, they do not show up in the table of contents at the begin of the page. Is there anything that could or should be done about this? (e.g. increasing the include size limit for this page). --M2k~dewiki (talk) 02:33, 9 January 2020 (UTC)

In my opinion we need many more users who can create properties (sysops and propertycreators). If you look at [3] the number of people who are allowed to create properties did not change much over the last 4 years but Wikidata has grown considerably during the same period. --Pasleim (talk) 17:15, 9 January 2020 (UTC)
We do have a slight backlog in Category:Properties_ready_for_creation but it's not much worse than we've often had in the past. Most of those are proposals pretty new (within the last month or so). What we really need I think is for more people to comment on proposals and help decide whether they should go forward! Wikidata:Property_proposal/Overview is a good place to start if you'd like to help - you certainly don't have to be a property creator to help! ArthurPSmith (talk) 18:26, 9 January 2020 (UTC)

Map of the U3A in Portugal

Hello @Tagishsimon:! Would you help me again? I am developing the list of universities of the third age in Portugal [4], but in addition of the list I would like have a map of the country with the several items inserted in it. Is it possible? How can I do that? Thank you, GualdimG (talk) 20:20, 9 January 2020 (UTC)

@GualdimG: Good question. Listeria will not provide a map. I have not found a template/module that takes a wikidata query and renders returned coordinates on a map. So right now all I can think of is pulling the coords out in a report, and using them in Predefinição:Location map+.
Is there a better solution? I'd like to know, too. --Tagishsimon (talk) 20:32, 9 January 2020 (UTC)
Thank you @Tagishsimon: for the quick answer! I´ll try that, but when you write "pulling the coords out in a report", could be the list of coordinates taken from the List of Wikidata? GualdimG (talk) 21:56, 9 January 2020 (UTC)

P569 greater than P570

The report Wikidata:Database reports/items with P569 greater than P570 will not update manually or automatically, can someone take a peek and see what the problem is? Perhaps the year change caused a problem. --RAN (talk) 21:42, 9 January 2020 (UTC)

Duplicates 2020-01-08

  1. Robert Baum : Q57553760 = Q63861459 21:56, 8 January 2020 (UTC)
  2. Friedrich Haseldorf : Q1439536 = Q64691233 21:56, 8 January 2020 (UTC)
  3. Stephan von Dorpat : Q55936766 = Q64705846 21:56, 8 January 2020 (UTC)
  4. Tilman Bredenbach : Q16403265 = Q55088971 21:56, 8 January 2020 (UTC)
  5. Axel de Vries : Q89411 = Q55236808 21:56, 8 January 2020 (UTC)
  Done 01:07, 9 January 2020 (UTC)
Done by User:Fixer88. --Succu (talk) 19:49, 10 January 2020 (UTC)



Redirected files are still working!

If CommonsDelinker cannot do better than this, then I think we can do without this kind of service! Many image-statements have rank, qualifiers and sources. It seems like all of it is lost because of bad programming! Redirects in file-namespace obviously works, CommonsDelinker doesn't! 62 etc (talk) 16:47, 9 January 2020 (UTC)

@Sextvåetc: Did you try to contact the maintainer of the bot? --Pasleim (talk) 17:18, 9 January 2020 (UTC)
@Pasleim: CommonsDelinker is not an ordinary bot. It is operated by the sysops on Commons! 62 etc (talk) 17:25, 9 January 2020 (UTC)
Well, you got us there, User:Sextvåetc. I guess it's just plain impossible to talk to them. (btw, piping your name to something that is not your name is much less than useful.) -Tagishsimon (talk) 18:28, 9 January 2020 (UTC)
See also Wikidata:Project chat/Archive/2019/12#CommonsDelinker and legend. Ghouston (talk) 03:58, 10 January 2020 (UTC)
It's quite understandable that bugs like get ignored. If people feel the need to be rude and aggressive in reporting bugs, don't expect the developers to invest time in solving the problem. Multichill (talk) 15:50, 11 January 2020 (UTC)

Duplicate item

It is possible to duplicate item for change only the label? For example, Canon BJ-200EX (Q81386668) and Canon BJC-70 (Q81643514) have the same values, only the printer name is different. I could've to duplicate items for create item for Canon BJ-200 (not 200EX) and change only the printer name, instead to create item from scratch? --2001:B07:6442:8903:9C02:13D4:6DFE:9DD1 16:26, 10 January 2020 (UTC)

Help:Split_an_item. --SCIdude (talk) 17:49, 10 January 2020 (UTC)
Thanks, but it is a little bit complicated. I will create new items from scratch... Thanks again! -- 17:54, 10 January 2020 (UTC)
I you have an account, you can make a common.js file like User:Ghouston/common.js and import the script User:Magnus_Manske/duplicate_item.js. This adds a "Duplicate this item" entry in the sidebar. Ghouston (talk) 10:45, 11 January 2020 (UTC)

Is there documentation for how to create Wikidata Tours?

Hi all

I'm planning on writing more Wikidata:Tours however I can't find any instructions for how to create them. Do these exist? If not I could create it if someone could talk me through it. My main questions so far are:

  • What are the steps to create a tour?
  • What are the commands that are allowed in tours? There seem to be some comments in the text that do actions e.g <!-- Action - go to the Earth item page on testwikidata --> but I can't find a list of them

@Thepwnco: and @Bene*: who worked on the existing tours, also @Lea Lacroix (WMDE):, @Daniel Mietchen:, @Multichill:, @Pigsonthewing: and @Spinster: who are subscribed to the Tours Phabricator tasks.


--John Cummings (talk) 09:29, 10 January 2020 (UTC)

All our tours are based on a Javascript file written by Bene*. Thanks to this file, a tour can be created with two pages:
  1. A normal wiki page where all texts are stored, e.g. testwikidata:Wikidata:Tours/References
  2. A javascript file where all steps are defined, e.g. testwikidata:MediaWiki:Guidedtour-tour-wbreferences.js. Here you need to define how the item page should look at start and to which html objects the texts will be attached to.
Everything in <!-- and --> are html comments and do not have any other purpose. If you write the texts for a tour I can help you out with the javascript. --Pasleim (talk) 14:40, 10 January 2020 (UTC)
@John Cummings: Further docs (but possibly outdated or incomplete) are at mw:Extension:GuidedTour and mw:Extension:GuidedTour/Write an on-wiki tour.
Also note that because it requires creating (and probably editing!) a page in the Mediawiki: namespace, you'll need to coordinate with an interface admin (the only people who can edit that namespace). (See also, list of test.wikidata interface admins). HTH. Quiddity (talk) 18:06, 10 January 2020 (UTC)
Thanks very much @Quiddity: and @Pasleim:, a couple of questions:
  • The javascript file and the page in the Mediawiki namespace, are these the same thing?
  • Does someone have to use test Wikidata? This seems like it would complicate things significantly
  • How can you create a button on the Wikidata:Tours page, it seems like its not just a link to a page.
Thanks again, --John Cummings (talk) 13:38, 11 January 2020 (UTC)
@John Cummings: :) 1) Yes. 2) Not necessary, but if you want to be able to edit a tour yourself easily whilst configuring it, that's the easiest place to do so. I.e. You might be able to request temporary IA (interface admin) user-rights on that wiki (note it's one of the most restricted usergroups, because it gives the ability to inject js on everyone's machine). However it's probably best just to work with Pasleim per their offer above. 3) it's just using the template {{StartTour}} to create a link in a blue box; a manually written link would work the same. HTH. Quiddity (talk) 20:02, 11 January 2020 (UTC)

Possibility of automatic updating/adding of dutch municipality data


Is there a way to automatically add new data to dutch municipalities in the future? For example population numbers for 2020 or 2021. The source would most likely be The site has an API and updates automatically ( link with some info).

There is also the problem of new municipalities emerging in the future, as they will have to be matched again or created as an entity. any idea how this could be handled?

Thanks for your time! Antoni1626 (talk) 14:32, 8 January 2020 (UTC)

@Antoni1626:. It seems quite possible to write a bot that would do a task like that. I actually did a couple of things with the CBS data and Dutch municipalities. You could even do simple edits without a bot using a tool like QuickStatements. Fortunately matching the CBS data with Wikidata items is easy because they all have the CBS codes as a statement. Let me know if you need some help. Husky (talk) 13:10, 12 January 2020 (UTC)

Complete list of PMIDs

Is there a way to get a complete list of items with PubMed ID (P698)? If I use a regular query it times out. --Cavernia (talk) 19:36, 10 January 2020 (UTC)

There are over 26 million items that have that property, and I don't think the query service can handle that many items in a single query. Your best option is probably using the database dumps. Vahurzpu (talk) 20:23, 10 January 2020 (UTC)
Maybe you should try your luck at Wikidata:Contact the development team. --Succu (talk) 22:18, 10 January 2020 (UTC)
This is a better tool to fetch such data. It will take about 15-30 hours to complete a list.--GZWDer (talk) 22:20, 10 January 2020 (UTC)
We have linked data fragments for this. Multichill (talk) 15:41, 11 January 2020 (UTC)

Research topic?

On the one hand I understand why an "ordinary" query for this sort of thing times out and can't work.
But on the other hand I don't see anything wrong with someone trying to run such a query: though somewhat extreme, it seems to me precisely the sort of thing Wikidata is for.
(Me, I've got some similarly too-big, "impossible" queries I've been trying to figure out how best to run.)

Is it reasonable to contemplate the existence of a hypothetical, better way to perform queries like this, that's both convenient and efficient? (And that doesn't kill the servers.) It might require a new form of query, it might require some new database indices, it might require cooperation from the user, but it seems to me worth thinking about.

Vahurzpu mentioned database dumps, but those are too big to be reasonably downloaded and worked with these days, I think.

GZWDer mentioned WDumper, but suggested it might take 15-30 hours, and that seems like a long time. (Or perhaps 250-500 results/sec is a perfectly reasonable performance rate to expect; I truly don't know.)

One idea I had -- although, again, I have no idea if this is reasonable or feasible -- would be to break a "huge" query up by using a LIMIT clause, and then run the query multiple time with some kind of a START_AT clause, and with some kind of back-end support so that the database engine is not having to run the whole query each time and subsample the results. —Scs (talk) 12:57, 11 January 2020 (UTC)

I think if you're doing Big Data research you'll have a big-RAM (>=32 GB) machine as a prerequisite and know that there are several ways of working with knowledge bases that are faster locally than what you get from the query server (even the query server uses suboptimal libraries as I understand it). Also if you're doing real research you know that there might come up additional questions, and that makes working on a local copy of WD a must. --SCIdude (talk) 07:31, 12 January 2020 (UTC)

Thanks for the help. I found a workaround now. --Cavernia (talk) 12:50, 12 January 2020 (UTC)

Merge 3 items

It seems these 3 items refer to the same institute (Botanical Research Institute):

Is there a recommended method of merging in cases like this one ? Kpjas (talk) 14:19, 12 January 2020 (UTC)

@Kpjas: if there is no (1) transformations of an organisation (one element would describe the organisation before the transformation, the other after; or something similar), (2) there is no problem of the existence of two articles in one Wikimedia project (e.g. for some reason there would be two articles on the same topic in one Wikimedia project), you just need to merge all the elements into one, which I will do (and clean up the final element). Wostr (talk) 15:18, 12 January 2020 (UTC)
@Wostr: Thanks for info and the merge. Kpjas (talk) 15:23, 12 January 2020 (UTC)

World rankings on wikidata?

Hi all! I do a lot of work on cue sports players and tournaments on a couple wikipedias, but it was suggested that wikidata might be a good place for a poor current solution.

Scenario: We have certain infoboxes that currently wright items for a current world ranking across different sports/lists. For example, the en:Template:infobox snooker player uses en:Template:infobox snooker player/rankings to generate the paramter for ranking. It uses a code for the following:

{{#if:{{Infobox snooker player/rankings|{{{name|<includeonly>{{PAGENAME}}</includeonly>}}}}} | {{nowrap|<!-- If template returns rank --> {{Ifnumber | {{Infobox snooker player/rankings|{{{name|<includeonly>{{PAGENAME}}</includeonly>}}}}} | <!-- then returns rank to infobox --> {{Infobox snooker player/rankings|{{{name|<includeonly>{{PAGENAME}}</includeonly>}}}}} | <!-- If template returns name then recall template with new name --> {{Infobox snooker player/rankings| {{Infobox snooker player/rankings|{{{name|<includeonly>{{PAGENAME}}</includeonly>}}}}} }} }} (as of {{Infobox snooker player/rankings|Date}})}} }}

However, I know we could do this with an instance on Q313507, which then would only need to be updated once for all wikipedias.

Question: Is there a quick way/bot to update each member from similar to en:Template:infobox snooker player/rankings?

Part 2: If so, there are other similar entries on en:Template:infobox pool player which take data from this and this?

I realise this is a lot to ask for, but I was hoping wikidata would be a great fix for this. The issues are that these rankings are quite often updated, so it would need to be something quick to update, not take several hours per update. Thanks for all your help. Lee Vilenski (talk) 18:31, 12 January 2020 (UTC)

I'm not sure whether there'd be a copyright issue with copying all of the rankings from the website. It seems to be run by a British company. Ghouston (talk) 06:51, 13 January 2020 (UTC)

Playbill external identifiers

The regex for theatrical properties Playbill person ID (P6132), Playbill venue ID (P6113), and Playbill production ID (P5833) seem unnecessarily cumbersome, and can be perhaps streamlined. Ideally, just the numerical portion would be used, assuming the numerical suffixes are still unique within venue/person/production urls. For instance, for Lin-Manuel Miranda (Q1646482) it is much easier to copy, paste, and enter "0000006865" instead of "lin-manuel-miranda-vault-0000006865". Can someone fluent in robot take a look at this, and see if it's possible, and even preferable, to concatenate all existing values to just their numerical suffixes via bot? Would there be any downsides? Thanks, -Animalparty (talk) 05:26, 13 January 2020 (UTC)

It needs to have a working link at least; works, but doesn't. Ghouston (talk) 06:34, 13 January 2020 (UTC)

category for files created with program X

There is category for pictures taken with camera (P2033). I'm searching for something similary for files created by software X, but wasn't able to find it in the search or in a what links here search for Wikidata property about Wikimedia categories (Q18667213). Is there something like this? (I already checked the usual suspects like GIMP, blender, Inkscape, etc.) --D-Kuru (talk) 12:22, 12 January 2020 (UTC)

Yeah, it would be the property I'm after. Looks like an upcoming property proposal --D-Kuru (talk) 20:48, 12 January 2020 (UTC)
I created Wikidata:Property proposal/category for files created with program. If you know that the property actually exists or if you have any comments or want to support you are more than welcome to do so! --D-Kuru (talk) 16:18, 13 January 2020 (UTC)

Why 'former X' is subclass of 'X'?

I have found cases where:

Does this makes sense? IMHO, the condition of subclass of (P279) is not fullfilled: it is not true that all instances of 'former X' are instances of 'X'.

Regards, --Albert Villanova del Moral (talk) 13:53, 13 January 2020 (UTC)

Do you have an example? Ayack (talk) 15:06, 13 January 2020 (UTC)
There are plenty of cases, for example:
--Albert Villanova del Moral (talk) 19:08, 13 January 2020 (UTC)
Yes, because the class is "X", not "current X". Circeus (talk) 19:34, 13 January 2020 (UTC)
The main issue I see with these "former XXX" items is that they are used in P31 statements on geographical or administrative entities while the notion of "former" should only be expressed with date qualifiers, never with "former XXX" items. Ayack (talk) 19:55, 13 January 2020 (UTC)
All Xs are current Xs. If I was a police officer yesterday, and then I stopped being one, then I'm no longer a police officer; I'm a former police officer and not a police officer. --Yair rand (talk) 06:04, 14 January 2020 (UTC)
The "former X" items are kind of nonsense, but they're convenient tags. "Former X" is taken as just "X", even in cases where qualified by date qualifiers, while allowing cheaper queries than checking the date data directly. Basically pretend that the word "former" isn't there in any statements using the items, and formulate things so that the assumption works. --Yair rand (talk) 03:11, 14 January 2020 (UTC)
Yes, John von Neumann (Q17455) is still described in Wikidata as human, but has a date of death to express the "no longer current" aspect. I'm surprised that nobody has created a class for deceased humans yet. See also Wikidata:Project chat/Archive/2019/04#Classes for defunct entities. Ghouston (talk) 04:39, 14 January 2020 (UTC)

Wikidata weekly summary #398

@Lea Lacroix (WMDE): Could you tell us where we can find some doc on tainted references (in its current design) and how we can test it please? Thanks. Ayack (talk) 19:48, 13 January 2020 (UTC)
@Ayack: you'll find a short description of the project and how to test it on this page. If there's any kind of information missing, let me know. Lea Lacroix (WMDE) (talk) 10:09, 14 January 2020 (UTC)
Thanks Lea Lacroix (WMDE). So this has nothing to do with "signed statements"? Or is it a first iteration? Ayack (talk) 14:35, 14 January 2020 (UTC)
In short: while doing research about the initial project that was called signed statements, we discovered that the main need for people is to be able to highlight and prevent mismatched references - regardless of the provenance of the source, encrypted or not. That's why we started with this. The original idea of signed statements described in the 2019 roadmap may evolve with time, based on what we learn on the way :) Lea Lacroix (WMDE) (talk) 15:33, 14 January 2020 (UTC)

Neue Strategy Liaison für deutsche Sprache

(This text aims to raise awareness for the final phases of the Wikimedia Strategy 2030 in german language)

Anfang 2017 begann, mit Auslaufen der alten Strategie, die Feststellung des Bedarfs und die Entwicklung einer neuen Strategie für die Wikimedia-Bewegung bis 2030. Jetzt, nachdem viele Einzelpersonen und Arbeitsgruppen Vorschläge erarbeitet haben, die immer wieder durch Feedbackrunden überarbeitet wurden, werde ich euch nächste Woche darum bitten, die finalen Vorschläge für die Strategie Wikimedia 2030 durchzuschauen und zu diskutieren. Meine Rolle in dem Prozess, der bis Juni andauert, wird es sein, eure Rückmeldungen zu sammeln und weiterzugeben, Chancen und Risiken sollen aus euren Antworten sprechen. Außerdem bin ich jederzeit für Hinweise und Fragen verfügbar, von „was ist überhaupt eine Strategie“ bis zu „wie wird mich das im Datensammeln und -veröffentlichen beeinflussen“. Jedenfalls freue ich mich schon auf euch und wünsche mir eine fruchtbare Diskussion. Weitere Infos wird es hier sowie auf Meta und Wikipedia:Wikimedia2030 geben. CJackel (WMF) (talk) 10:52, 14 January 2020 (UTC)

Hi @CJackel (WMF): and welcome to Wikidata! The Wikidata:Project chat is usually an international talk page where discussions take place in English. In order to reach the German-speaking community, I can suggest to add your message to the German-speaking page Wikidata:Forum.
If you need anything regarding community communication on Wikidata, feel free to reach me. Cheers, Lea Lacroix (WMDE) (talk) 11:01, 14 January 2020 (UTC)
It's not easy to keep up. Although I have a few hundred edits on Wikidata with my private account, I was not aware of the Forum. Thank you for your advice, Lea Lacroix (WMDE)! Cheers CJackel (WMF) (talk) 11:07, 14 January 2020 (UTC)
@CJackel (WMF): did you use Wikidata in German or in English? Do you have insight of why it might have been hard for you to be aware of the forum and how it might have come to your awareness easier? ChristianKl❫ 14:13, 14 January 2020 (UTC)
ChristianKl, I am not totally aware of which language I use. I suppose, both. Mostly fixing interwikis, adding descriptions, pictures or citations. This is more or less not language specific. --CJackel (WMF) (talk) 14:30, 14 January 2020 (UTC)
On top of the page the main language you are using at any given point is listed. There's a bunch of stuff the UI does depending on how that language setting is made. ChristianKl❫ 15:00, 14 January 2020 (UTC)
I never changed anything in the preferences, I am only doing so in Wikipedias and other projects in languages that do not use the Roman alphabet. This is probably why I was not aware of the Forum. Regards --CJackel (WMF) (talk) 21:25, 14 January 2020 (UTC)

From 4 Identifier Properties to 1 Identifier Property

Until December 2019 the Media Arts Database (Q54760023) had different identifiers and URIs for different material in their database. There was Media Arts Database manga ID (former scheme) (P6369), Media Arts Database anime ID (former scheme) (P6368), Media Arts Database console games ID (former scheme) (P7117) and Media Art Database author ID (former scheme) (P3231). But they have changed their IDs and the structure of their URIs without redirection. Now there should be only one identifier property for all material. My question is, how should we deal with the old IDs? Is it appropriate to create a new property "Media Art Database ID" and mark the old IDs as deprecated? (In addition, I have got a mapping file for the video game IDs from old to new, but not for the other stuff.) Diggr (talk) 09:41, 15 January 2020 (UTC)

Given that there are no redirects i don't think it makes much sense to keep the old ID's. I would propose a new property, and once that gets accepted deprecate the old properties and start moving the old id's over to the new ones, deleting the old claims in the process. Husky (talk) 11:29, 15 January 2020 (UTC)
Any chance the old ones are worth keeping in terms of finding things in online archives? - Jmabel (talk) 18:26, 15 January 2020 (UTC)
The standard procedure in Wikidata is to keep old identifiers even when a website goes offline. It seems to me like it would be good to have a help page that describes our procedure. ChristianKl❫ 08:45, 16 January 2020 (UTC)
Added a quick line to Wikidata:Identifiers: Special:Diff/1099321573 (feel free to update/move/etc. :-) Jean-Fred (talk) 11:17, 16 January 2020 (UTC)
Thanks! I've created a property proposal for the new identifier Wikidata:Property proposal/Authority control#Media Art Database ID Diggr (talk) 13:24, 16 January 2020 (UTC)


Those of you who are active on enwp, can take a look at en:Lillehammer Municipality and en:Lillehammer. The first one is connected to an item about a "municipality" and the other is connected to an item about a "municipality in Norway". Even the items tells they said to be the same as (P460)! There is only one municipality with this name in Norway, so ... There is also an item about a populated place with this name! 62 etc (talk) 07:26, 16 January 2020 (UTC)

In most cases, its the Cebuano Wikipedia which causes a duplication - apart from the correct thing that there's a separate item about the settlement and the municipality. One way it could be solved would be to link en:Lillehammer Municipality to Lillehammer (Q101341) and en:Lillehammer to Lillehammer (Q3745117), and change that article slightly to make sure that the municipality is handled in another article. However, as most Wikipedias don't separate the two concepts of settlement and administrative unit, there is one one article, and that is linked to the municipality - so can be confusing the long German article is linked to a three sentence stub in English. BTW: Lillehammer might be a rather prominent example, but only in Germany there are still 2000+ real duplicate municipality items [5] I am slowly merging... Ahoerstemeier (talk) 08:27, 16 January 2020 (UTC)
Even in cases where there is only one article, there is often a need to be two items, since the data about the municipality does not match that of the settlement. Not only the population and area differ, also their history. Mark, that can be said when there really is real data about the settlement. In Norway there in most cases is. In most cebwiki/geonames-cases there is not. 62 etc (talk) 14:25, 16 January 2020 (UTC)
Even in the case where there is no exact data on a settlement - like in Germany - there is one perfect reason to split settlement and municipality. In most cases the settlement once established will always be there, maybe growing, maybe even merging into another one. A municipality as a legal entity usually is created much later than the settlement, and it might get merged into a neighboring one at any time, or changing name, or changing its administrative center. Placing all into one item here would require lots of applies to part (P518) qualifiers and become a total mess. Ahoerstemeier (talk) 22:01, 16 January 2020 (UTC)

2020-01 Transnistria

What are the difference between Transnistria (Q3537754), Transnistria (Q907112), Pridnestrovian Moldavian Soviet Socialist Republic (Q7242857), Transnistria (Q81644211) ? Involved contributors: @Quakewoody, Pridnestrovian editor: Visite fortuitement prolongée (talk) 21:53, 12 January 2020 (UTC)

These are really different concepts.
The term "Transnistria" is a purely Romanian designation used in two meanings:
  • 1) the occupation regime created on the territory of the modern Odessa region of Ukraine during the Second World War, where genocide of the civilian population was perpetrated;
  • 2) the interfluve of the Dniester and the Southern Bug, it's is the same territory, but here considered not as an occupation regime, but as a region highlighted by the Romanians, which the Turks called Edisan until the 19th century.
Pridnestrovie is:
  • 1) the unrecognized state between Ukraine and Moldova - the Pridnestrovian Moldavian Republic, declaring its independence;
  • 2) the preceding to it Pridnestrovian Moldavian Soviet Socialist Republic - an administrative entity that claimed to become part of the USSR; after the collapse of the USSR, it was transformed into PMR (not as socialist state);
  • 3) the region covering the territory along the Dniester River (and not only the left bank).
When I looked for the already created items, I did not find anything about Pridnestrovie. Only about "Transnistria", which in principle is of little interest to anyone and exists exclusively in the historical context and the context of agressive Romanian nationalism. Pridnestrovian editor (talk) 23:11, 12 January 2020 (UTC)
It also seemed strange to me that the item "Transnistria" uses the elements related to the Pridnestrovian Moldavian Republic. But, since that item is closed for editing, I considered it most reasonable to engage in the design of the correct record under the correct title than to enter into disputes with moderators about access to edit the record, originally related to Transnistria, and not Pridnestrovie. The difference between these concepts and the confusion in terminology I have explained above. Pridnestrovian editor (talk) 23:33, 12 January 2020 (UTC)
Ok, thanks. If we get rid of any hints of "transnistria" and make a good description of the Pridnestrovian Moldavian Republic instead of what we have now, then I am in favor of deleting the new record. Because it was very strange that I could not find a single WikiData item that would display the name Pridnestrovie. I don’t want instead of the name of Poland to display “Auschwitz” or “Treblinka”. Or "Holocaust firebox" instead of the name of Israel. Moreover, I do not want to see this about Pridnestrovie. Pridnestrovian editor (talk) 05:36, 13 January 2020 (UTC)

It does seem to me that we should have a distinct item for the term as used during World War II, distinct from the present unrecognized state in all but name and western border. That said, the present unrecognized state is usually referred to in English as "Transnistria", and I think that should remain its item name. I have literally never read anything where a native English speaker referred to it as Pridnestrovian Moldavian Republic or anything similar. - Jmabel (talk) 06:07, 13 January 2020 (UTC)

I agree that there should be a separate item for Transnistria (the Romanian occupation district). However, only it should be called that way.
As for the name and the western border, there is nothing in common between Transnistria and Pridnestrovie. The Republic of Pridnestrovie located on both banks of the Dniester river: on the right bank are the city of Bendery and the Kitskansky village council of Slobodzeya district, on the left bank are five more districts. Tiraspol, the capital, today is located on both banks.
The name "Transnistria" appeared in 1992 during the Moldovan-Pridnestrovian war, when the aggressive nationalists who came to power in Moldova needed to somehow name the breakaway state. It was then that they recalled the term of Romanian Nazi criminals, because they could not use the "Russian" name. Then it leaked to the West, because (for obvious reasons) the Western media broadcasted Romanian and Moldovan sources in connection with this war. Literally, this term is translated as “through the Dniester,” “on the other side of the Dniester” - from Romania, of course; it is clear from this that the word has no relation to Pridnestrovie even outside the context of Romanian nationalism.
Over time, this word was almost forgotten: official sources used either "Stînga Nistrului" if it was about the administrative division of Moldova, or "PMR" if it was about the breakaway state itself. In the English language, this region does not have a well-established designation simply because even a fraction of the percentage of native speakers hardly heard about it, and the choice of terminology was dictated solely by context. It received a “new life” with the spread of the Internet and the advent of Wikipedia, which became the main distributor of the term in the Internet environment.
This is the main reason for the confusion and misunderstanding that sometimes leads to conflicts. In the Pridnestrovie, this term, applied as the name of this region, is considered a gross insult and disrespect for the locals and the history of their country (more about this, for example, here: 1234). Therefore, I want to deal with this misunderstanding. I do not think that in English or any other world-class language there is at least one name of the country, which would be an insult to its population. Pridnestrovian editor (talk) 09:27, 13 January 2020 (UTC)
We are here not to rewrite history, but to reflect what's in the sources, even if this does not seem right to someone. For some languages there are lists of official names of countries and unrecognised territories – so it's easier to select a name – but for the rest, we should have a name that is used in most of the sources written in that language. BTW there is also Transnistria Governorate (Q167641) that describe the territory during WW2. Wostr (talk) 12:33, 13 January 2020 (UTC)
That's right, plus we need to remember about the neutral tone of descriptions. Therefore, in this case the only permissible naming must be taken from official sources that established by official legal regulations. Pridnestrovian editor (talk) 14:54, 13 January 2020 (UTC)
By the way, precedents with the correction of names are here even for unrecognized states: 1. And this despite the fact that Karabakh is a real geographic area, the name of which does not bear any second meaning or offensive tinge, in contrast to the ugly Romanian word “transnistria”. Pridnestrovian editor (talk) 16:54, 13 January 2020 (UTC)
Official sources from the entity involved are hardly "neutral." - Jmabel (talk) 17:34, 13 January 2020 (UTC)
So this is not an official source, but a news site. The same articles were published on other sites, but on Russian or Romanian. I brought the Pridnestrovian resource only because there is a qualified translation into English. Pridnestrovian editor (talk) 18:29, 13 January 2020 (UTC)
If the 'Transnistria' is the most common name in English, it should be a label of an item. Period. Wostr (talk) 18:04, 13 January 2020 (UTC)
In English, the most common name of the subject is "Stînga Nistrului" or "Left bank of the Dniester", as follows from official documents of Moldova. This is natural, given that the main area where English is used in relation to this subject is the issue of resolving the frozen conflict and the diplomatic relations of the participants in the negotiation process, including EU and USA. All the use of "transnistria" comes down to Internet dumps, the source of which is Wikipedia, nothing more. To declare on this basis that the subject under discussion, which almost no one had ever heard of outside of ex-USSR, has some kind of an established designation in the language, it is simply meaningless. A designation may appear only with recognition of the independence of this state and the inclusion of its name in official norms and standards. Now we should be guided by the official documents of the state that is described here, and the self-name of the people, and not engage in promoting biased and offensive terminology. Pridnestrovian editor (talk) 18:29, 13 January 2020 (UTC)
It's clearly you POV that should not be transferred to WD. Only a thorough query of reliable sources (academic books, scientific articles etc.) in English can solve this problem. Wostr (talk) 20:54, 13 January 2020 (UTC)
My POV? I ask you to argue. And on which subject do you need reliable sources? On the subject of using the word "transnistria" in English? This is already funny. Pridnestrovian editor (talk) 21:48, 13 January 2020 (UTC)
I can pretty much guarantee that no native English speaker who didn't speak Romanian as well would ever consider "Stînga Nistrului" (or "Stânga Nistrului" if they picked up their Romanian in Romania rather than in Moldova). You would see Transdniestria, Transdniester, and other such variants. For what it's worth: en-wiki, which we usually follow on article names, begins its article Transnistria "Transnistria or Transdniestria, officially the Pridnestrovian Moldavian Republic"; Britannica opts for Transdniestria and begins their article "Transdniestria, also spelled Transnistria, also called Pridnestrovskaia Moldavskaia Republic and Pridnestrovie"; National Geographic opts for Transnistria. - Jmabel (talk) 01:00, 14 January 2020 (UTC)
The Romanian "Stînga Nistrului" will not be considered, but the Romanian "Transnistria" will be? It's funny. Probably, you did not see the use of the terminology of official Moldova because you are not a diplomat or politician participating in the negotiation process. More English in relation to this region is almost never used. The fact of the matter is that in various sources you can find any names that were called PMR: Transdniester, Trans-Dniestr, Transdniestria, Pridnestria, Pridnestrovia, Pridnestrovie, Dnestr Republic, Nistrenia, Nistria, Transnistria, TMR, PMR, Left Bank of Dniester, Stinga Nistrului, Cisdniestria, etc. Meanwhile, the only name of this state - if we are talking about this particular state, as can be understood from the description of item Transnistria (Q907112) - is defined in its legislation and self-name. Pridnestrovian editor (talk) 13:26, 14 January 2020 (UTC)
By the way, here are the legislative acts defining the correct name of the Pridnestrovian Moldavian Republic:
Pridnestrovian editor (talk) 17:36, 14 January 2020 (UTC)

Cisdniestria (belatedly added in comment above is an almost completely unused term (150-odd Google hits), and "cis" is the opposite of "trans" in Latin, Romanian, and English. It would presumably refer , if used, which it basically isn't, to the (from a Western or Central European point of view) "near" side of the Dniester, which is to say the rest of Moldova. Adding it yourself to Transnistria (Q907112) does not make it otherwise. - Jmabel (talk) 17:20, 17 January 2020 (UTC)

True, this is an attempt to literally translate the original name of Pridnestrovie. I tried to collect all designation of the PMR and Pridnestrovian land that are found in English sources, except for the nuances of transliteration, such as Dnestr/Dniestr/Dniester/Dnjestr/... About Western Europe: the toponyms is not created in accordance with the geographical location of native speakers of a language. The "reference point" is usually the country or people to which the region belongs. An example here is Ukrainian Transcarpathia: in relation to Europe, this is also not "trans-". In the case of Pridnestrovie, by the way, this is another aspect that makes all names with "trans-" offensive - this is an expression of the claims of Romania from the time of Nazi dictator Antonescu. Pridnestrovian editor (talk) 20:28, 17 January 2020 (UTC)

Weird sitelink glitch

When clicking on the sitelinks in the article the en sitelink takes you to rather than I tried fixing this in the Wikidata items list of highest paved roads in the world (Q3834274) and extreme points of Europe (Q2344897) by deleting the sitelinks and putting them back, and by purging the Wikidata pages but no luck. Any idea what is wrong? Abductive (talk) 09:05, 15 January 2020 (UTC)

Fixed by ast:Special:Diff/2745627; the astwiki page defined an oldschool-type interwikilink via the ast:Template:VT template. —MisterSynergy (talk) 09:35, 15 January 2020 (UTC)
Thanks. Why is this VT template still in use? Is it common on other Wikipedias? Abductive (talk) 05:42, 16 January 2020 (UTC)
@Abductive: It's a hatnote template, and hatnotes are fairly ubiquitous on most large Wikipedias. The English Wikipedia's version of the template doesn't have this issue, because the Lua module on which it relies automatically prefixes an extra colon at some point. Jc86035 (talk) 02:59, 18 January 2020 (UTC)

Delete (Q82105340)

I accidentaly crested an data-item Q82105340 wich already existed Q47267. Can this data item be removed, to avoid clutter?Smiley.toerist (talk) 12:36, 16 January 2020 (UTC)

@Smiley.toerist: no – the item should be merged into the other one instead. See Help:Merge for how to do it. --Lucas Werkmeister (talk) 13:26, 16 January 2020 (UTC)
  Merged, please read the help page linked above for future reference. --Kostas20142 (talk) 18:12, 17 January 2020 (UTC)

Am I on the right track with Zoom H6 Handy Recorder (Q81525088)?

Hi all! I was just wondering if anyone could take a look at Zoom H6 Handy Recorder (Q81525088) and give me some pointers before I go too much further with this and potentially add the wrong properties to the other few entities. I think I've pretty much got the hang of the basics but I'm still kind of struggling with the whole subclass thing for some reason. It seems to make sense for a second but then I lose it when I try to apply it to sound recording device (Q66087652). I'm getting warnings with storage device (P4788) and writable file format (P1073) when I use them on instances of it.

I'm not too sure about the first because I thought that subclasses transfer, and because SDXC card (Q50993355) -> Secure Digital card (Q466977) -> memory card (Q183731) -> recording medium (Q193395) it should be fine, but I get "Values of storage device statements should be instances of data storage (or of a subclass of it), but SDXC card currently isn't".

The second one, I think that writable file format (P1073) just needs the electronic machine (Q2858615) class added to the value type constraint.

I didn't just want to button mash until something worked though so if anyone has any suggestions, they would be much appreciated! Cheers in advance --Aluxosm (talk) 16:46, 16 January 2020 (UTC)

They look pretty good, I have a few ideas to improve them a bit which I'll edit directly. Ghouston (talk) 05:21, 17 January 2020 (UTC)
The property writable file format (P1073) is only supposed to be used on software items, not devices. However, I've suggested expanding its usage at Property_talk:P1073#Devices. Ghouston (talk) 05:27, 17 January 2020 (UTC)
Ohhh it all makes sense now! You can of course have instances of Zoom H6 Handy Recorder (Q81525088) and they need to inherit sound recording device (Q66087652). Thank you so much Ghouston (talkcontribslogs) and D-Kuru (talkcontribslogs)! --Aluxosm (talk) 07:01, 17 January 2020 (UTC)
You are welcome. Not usre for what, but you are welcome! --D-Kuru (talk) 09:58, 17 January 2020 (UTC)
D-Kuru (talkcontribslogs) - Haha I thought you'd seen this topic and that's why you made the edits to Zoom H4n Handy Recorder (Q8074152) and Zoom H2 Handy Recorder (Q3576049). I didn't realise that you'd made loads of edits to those items before so they were probably on your Watchlist. Either way, after seeing what you did, and then what Ghouston did with Zoom H6 Handy Recorder (Q81525088), it all clicked (my confusion with the subclass thing) and I'm very thankful for it ;) --Aluxosm (talk)
@Aluxosm: You can use {{re|USER}} or {{ping|USER}} to not have the "(talk • contribs • logs)" section included.
I did actually neither see this post nor have it the items on my watchlist. I saw your edits on my images on Commons. Then I thought that I might recheck the subclass of and instance of section since I cofused both of them some time ago. But it seems to be all fine and well now. Just need to add that property for writable file formats
--D-Kuru (talk) 11:34, 17 January 2020 (UTC)
Ah, amazing! Cheers for the heads up on {{re|USER}}, I was just thinking that something like that should exist. --Aluxosm (talk) 11:47, 17 January 2020 (UTC)

Change to regex constraint not being reflected?

A couple of days ago, I changed the format as a regular expression (P1793) constraint on MoEML ID (P6060) (this diff), because the final numeric part of the identifiers can have multiple digits.

But I am still seeing constraint warnings as at Q7308717#P6060, even when such claims were added some time after the constraint was updated. And OpenRefine isn't taking the change on board either, flagging such identifiers as issues before upload.

Can anyone cast any light on what's going on here? Jheald (talk) 09:22, 17 January 2020 (UTC)

This, I think: [7]. Why is it specified in two places? Ghouston (talk) 10:02, 17 January 2020 (UTC)
Been there, tripped over that. --Tagishsimon (talk) 01:49, 18 January 2020 (UTC)

Creating dummy items for easier use

Is there any way to create some kind of dummy item. The purpose would to have a template for a new items that you don't miss anything. I know that you can duplicate an item, but you have to change the values and you have to recheck if there might be a property that would fit for the item that was not included in the copied one. So far I have a list on a userpage where all properties that could fit are listed. If possible I would like to create a virtual item that is filled with all the information you can find and you can publish it when you save it once everything is ready to go --D-Kuru (talk) 19:56, 13 January 2020 (UTC)

With Cradle it is possible to create forms. In the form you can enter all values and properties needed. It is possible to create new ones in that way. I currently dont know a tool, which helps adding properties with a form to a existing item. I think this were something what would help people editing Wikidata. -- Hogü-456 (talk) 20:13, 13 January 2020 (UTC)
I don't think so. If you have several items with identical statements put the statements in a class and make your items instance of it. The items will inherit the statements. Same with object-oriented programming. --SCIdude (talk) 07:51, 14 January 2020 (UTC)
@Hogü-456: Cradle will may help along for now since it looks like that it at least suggests properties.
@SCIdude: The idea is more of itmes with instance of X that you can not really do with a main item. Eg. notable buildings in city X. I don't think that an item with "buildings in city X" is a good idea. For each building you have to enter a bunch of information and they can differ from each other (eg. coordinate location, used by, adress, heritage building ID, architect and so on). You could copy an item, but you would have to change every value.
--D-Kuru (talk) 21:47, 15 January 2020 (UTC)
I have misunderstood your question, sorry. Alternatively (for many items) I would use a spreadsheet for input and create QuickStatement commands from it. --SCIdude (talk) 07:53, 16 January 2020 (UTC)
From what I can tell from the project page I would say that this is pretty much what I'm looking for, but it seems a bit complicated right now. --D-Kuru (talk) 19:07, 18 January 2020 (UTC)

Wiki Loves Folklore

Hello Folks,

Wiki Loves Love is back again in 2020 iteration as Wiki Loves Folklore from 1 February, 2020 - 29 February, 2020. Join us to celebrate the local cultural heritage of your region with the theme of folklore in the international photography contest at Wikimedia Commons. Images, videos and audios representing different forms of folk cultures and new forms of heritage that haven’t otherwise been documented so far are welcome submissions in Wiki Loves Folklore. Learn more about the contest at Meta-Wiki and Commons.

Kind regards,
Wiki Loves Folklore International Team
— Tulsi Bhagat (contribs | talk)
sent using MediaWiki message delivery (talk) 06:15, 18 January 2020 (UTC)

Property for item that is listed on a page

You can only link one article from wikipedia to one item. What if a list article contains more than one item? Is there a property like is listed on? --D-Kuru (talk) 21:48, 15 January 2020 (UTC)

There is part of the series (P179) and has list (P2354), but I think we should not lend too much weight to individual Wikipedia articles/lists: some items that are only redirects to lists on English WP have individual articles on other language Wikipedias, and vice versa. Are you asking how to link individual items to Wikidata that only appear on lists? -Animalparty (talk) 22:33, 15 January 2020 (UTC)
For example Adolphe Thiers (Q5738) is linked with President of the French Republic (Q191954) which in turn is linked with list of Presidents of France (Q29580): there's no need for a direct link between Adolphe Thiers (Q5738) and list of Presidents of France (Q29580). Ghouston (talk) 02:45, 16 January 2020 (UTC)
@Animalparty: I have a wikipedia list page and want to link an existing item to that page. Example: I have de:Liste der denkmalgeschützten Objekte in Pressbaum that lists heritage buildings in the area of the city de:Pressbaum. Then there is Q37825991 that is a small entrance tower to the Second Vienna Mountain Spring Pipeline. Is there a way to link this item to the list? There is Cultural heritage monuments in Pressbaum (Q1858294) which is the Wikidata Item for the list on Wikipedia. Maybe with instance of and the list item as item (does not really sound right here)
@Ghouston: With Q37825991 I don't see a link to the list on de.wikipedia
--D-Kuru (talk) 19:20, 18 January 2020 (UTC)
It looks like it's fine as it is. There's no single item for "Denkmalgeschütztes Objekt in Pressbaum", but Q37825991 has heritage designation (P1435) and located in the administrative territorial entity (P131) statements which in combination declare the same thing. Ghouston (talk) 05:40, 19 January 2020 (UTC)

Categories item that don't linked to any Wikipedia category

Hello. Do you thing is useful to create items for categories about geographical items (so the category item will connect with geographical item with topic's main category (P910)), even though there is not yet any Wikipedia category in any language? Xaris333 (talk) 17:05, 19 January 2020 (UTC)

  • I don't think so. As far as I'm concerned, "Category" items exist only as a structural necessity to describe entities in WMF projects; they don't have any notability. - Jmabel (talk) 17:29, 19 January 2020 (UTC)
  • There are two cases where we routinely create such category items: when there is a category in a different Wikimedia project (excluding Commons); and when there is both a Commons category and gallery so that both can be linked to (or there are both topic and list items that link to the same commons category). Just creating category items without any sitelinks isn't particularly useful though. Thanks. Mike Peel (talk) 18:37, 19 January 2020 (UTC)

Merge two items for Easter date calculation?

Computus (Q245003) and computus (Q56318457) appear to be the same. The English Wikipedia article w:Computus was shifted from the former to the later today. Neither Wikidata item has anything on its discussion page. Jc3s5h (talk) 09:32, 17 January 2020 (UTC)

This items discribe the same thing, so it needs to be merged.--AcademDoctor (talk) 10:05, 17 January 2020 (UTC)
Q245003 is "Computus"; Q56318457 is "calculation of the date of Easter". If they are the same they should be merged to Q245003, as that is the original item, but the French Wikipedia has two articles. Peter James (talk) 13:06, 17 January 2020 (UTC)
The merge fails because there are two different French articles. Q245003 is associated with w:fr:Comput but Q56318457 is associated with w:fr:Calcul de la date de Pâques. I don't read French well enough to know how the linkages should be changed. Jc3s5h (talk) 21:46, 17 January 2020 (UTC)
@Jc3s5h: w:fr:Comput ("the calculation of calendar elements used by the Christian churches") appears to be a sort of overview page that discusses both indiction (Q1339127) and computus (Q56318457), with more attention paid to the latter. It's hard to make a good comparison because the English Wikipedia article is much more detailed (and also because my understanding of French is not particularly good), but both frwiki articles contain information that is also included in w:en:Computus. The ways I could see to model this in Wikidata would be to either keep Q245003 as some sort of parent to the various related topics or to turn it into a disambiguation item. (Some of the statements on Q245003 would need to be changed either way.) Jc86035 (talk) 06:48, 18 January 2020 (UTC)
Hello. w:fr:Comput is a kindalike general article about religious dates calculations/computations (comprised of Easter date, indeed). And w:fr:Calcul de la date de Pâques is a w:fr:Comput technique for Easter. Please note that English word "computer" is derivated from latin computo : to count. Bouzinac (talk) 22:58, 19 January 2020 (UTC)

Archive URL?

After the change of archive URL (P1065) only a handful of web archives are valid entries. Previously this was used for links to archives of any kind. An archive is where documents are physically stored, and possibly copies are made digitally available. This is not the same as a digital copy of something already available on the web. It is neither the same as an archive item, which more often than not refers to content of an archive box, not the actual document.

I believe this property was intended to be a general property for archive URLs, but is now turned into a property for web archive URLs. There are thousands of archives around the world, but only a handful of real web archives. Trying to maintain a single format regexp will quickly become a nightmare.

Perhaps a better description of the difference is that documents in a real archive is born on paper, while documents in a web archive is born digitally. The online information about a document in a real archive may not contain the document at all, it may only contain metadata about the document.The difference is not just functional, it is physical.

I believe we need a property for deep links into a real archive, which P1065 isn't anymore. Jeblad (talk) 17:45, 12 January 2020 (UTC)

But from the property proposal, this property was always intended as an URL to an archived version of a web page. Wostr (talk) 17:49, 12 January 2020 (UTC)
If you want to understand what a property is intended to do, it's useful to look at the property proposal. In this case the description in the proposal is "The URL in the same reference but archived on or some other archive service (," The intention of the property isn't to link to meta data of physical archives. Ideally, we would have a bot that adds an link to every usage of reference url. We currently don't have a bot, but the property was created with that use-case in mind as Enwiki has a bot that actually protects the information in enwiki references that way against link rot.
It's a worthwhile use-case to save data about which libraries actually have physical copies of a document and it might make sense to have a Wikidata property to store that information but it's a different use-case then what archive URL (P1065) was designed to handle. ChristianKl❫ 21:19, 12 January 2020 (UTC)
Use archives at (P485). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:19, 13 January 2020 (UTC)
  • You can expand the regex. --- Jura 08:28, 20 January 2020 (UTC)

Deprecation and preferred rank

I believe reason for deprecation (P2241) and reason for preferred rank (P7452) should always be valid qualifiers - that is, they should never cause a constraint violation on any property regardless of that property's list of allowed qualifiers. Do you agree, and is this possible in the software? - PKM (talk) 20:40, 13 January 2020 (UTC)

Yes, it's quite annoying to add these properties to the list of allowed qualifiers whenever I want to deprecate/prefer a statement using a property that does not have these qualifiers already added as allowed. Wostr (talk) 20:59, 13 January 2020 (UTC)

Lucas Werkmeister (WMDE)
Jarekt - mostly interested in properties related to Commons
John Samuel
Yair rand
Jon Harald Søby
Was a bee
Peter F. Patel-Schneider
ZI Jony
  Notified participants of WikiProject property constraints

  •   Support Have had the exact same thought before --SilentSpike (talk) 21:16, 13 January 2020 (UTC)
  •   Support Hope it's possible. --SCIdude (talk) 07:47, 14 January 2020 (UTC)
  •   Support I wonder if there are more such "administrative/meta" qualifiers. --Jneubert (talk) 10:35, 14 January 2020 (UTC)
  • I agree. ChristianKl❫ 14:11, 14 January 2020 (UTC)
  • Probably mapping relation type (P4390) too? --SCIdude (talk) 16:36, 14 January 2020 (UTC)
    • Yes! - PKM (talk) 21:08, 14 January 2020 (UTC)
  •   Support no objection. Regards, ZI Jony (Talk) 17:35, 14 January 2020 (UTC)

I have created a new Phabricator task T242797 for this request. I've called it "universally allowed qualifiers". Thanks, all! - PKM (talk) 21:32, 14 January 2020 (UTC)

  •   Support No objection. The only valid contraint should be that reason for deprecation (P2241) sould have not normal or a prefered rank. --Fralambert (talk)
  •   Support, Sometimes I feel like we could use more freetext explanation properties, like generic comment (DEPRECATED) (P2315), or more specific syntax clarification (P2916). --Jarekt (talk) 04:24, 15 January 2020 (UTC)
    • I don't see what that has to do with this topic, but free text doesn't have any structured data to it, so I don't think it belongs on Wikidata. --Yair rand (talk) 19:55, 15 January 2020 (UTC)
  •   Support Amen to that. Moebeus (talk) 12:56, 15 January 2020 (UTC)
  •   Support Yes. --Sannita - not just another sysop 21:37, 15 January 2020 (UTC)
  •   Weak support for such a feature.   Oppose for the samples given above, e.g. VIAF ID (P214) shouldn't have both. BTW, it's fairly simple to add an allowed qualifier to several properties. --- Jura 08:27, 20 January 2020 (UTC)

Correctly representing archaic or debunked "racial" term entries

Entries like Q714773 "Nordic race" are held on Wikidata unqualified, and so are both incorrect and may be open to misuse to promote debunked theories of Nordicism or other race hate-related propaganda. This should be qualified within the entry itself as being historical language which cannot be applied to humans in any logical way. Raising here as this has been a current issue for "racial" categories on Commons, ref c:Commons:Village_pump#Correctly_representing,_but_not_promoting,_"racial_theories"_used_in_Nordicism_and_Nazism. -- (talk) 15:42, 17 January 2020 (UTC)

@: This is an issue with all racial categorizations of people, not just this particular item, so I think this would have to be dealt with by modifying the items higher up in the tree like race (Q3254959). For other items, like homeopathy (Q81058), it makes sense to have some sort of "not valid" distinction on the item, but the issue is applicable to all human racial classifications, so it might not make as much sense to add statements on each individual item. I think you could add a statement to indicate whether the subject of the item is in contemporary use, but I don't know if it would be appropriate to go further than that. Jc86035 (talk) 17:26, 17 January 2020 (UTC)
A core principle of Wikidata is that Wikidata is not the judge of what's the correct description. We rather go with what serious sources say about what someone's ethnicity happens to be then trying to make a judgement of our own about what the ethnicity happens to be and note diagrament when different serious sources disagree. It's worth noting that we do have a higher standard for sources for ethnicity then we have for other properties given that ethnicity is special in the way it can be misused. ChristianKl❫ 19:36, 17 January 2020 (UTC)
You are missing the point. These are not "ethnicity" and this is not a debate about validity or whether Wikidata should be a "judge". You may as well be tagging a record for someone born in India a "wog" or start using a people classification of "teddy bears". These are false "racial" types which have been long debunked, are only ever used by white supremacists, neo-Nazis or other racist extremists, and should never be used to classify people at all. -- (talk) 21:55, 17 January 2020 (UTC)
It sounds to me like you lack knowledge of how Wikidata works with race and try to propose policy without understanding the status quo. There are plenty of censi that happened and that do record information about race. I would be surprised if any census exists that classifies people as Teddy bears. On the other hand I would be suprised if no country has historically had a census that classification like "Nordic race". ChristianKl❫ 23:00, 17 January 2020 (UTC)
@ChristianKl: That censuses have recorded information about race does not actually demonstrate any scientific basis in the concept. Censuses are not a reliable source for scientific consensus, and we would look at scientific publications to determine whether to classify human racial categorizations as social constructs or biological concepts.
How to actually classify people by their ethnicity or racial identity is an altogether separate issue to how race itself is classified, and I think it would be best resolved by making the constraints of ethnic group (P172) stricter and cleaning up its usage. Jc86035 (talk) 02:40, 18 January 2020 (UTC)
Whether the concept has a scientific basis is irrelvant for the question or not the concept can be used. The claim that Harry Potter is a horcrux doesn't have a scientific basis but it's still information we store in Wikidata because there are sources that describe Harry Potter in that way.
To the extend that there are studies that papers about how a term like "Nordic race" is used in the scientific literature and you want to store that knowledge in Wikidata, feel free to use it for claims on Nordic race (Q714773). Fæ didn't reference any such sources. Correctly, modeling what a paper leads to a different discussion then the one we are having here. ChristianKl❫ 11:03, 18 January 2020 (UTC)

Make race (Q3254959) subclass of "obsolete concept"? Visite fortuitement prolongée (talk) 22:00, 17 January 2020 (UTC)

Yes. However, all entries like "Nordic race" need to have descriptions which make it clear that they cannot be applied to people. The risk is that we will have categories and media on other projects tagged with these Wikidata types as if they were taxonomies of humans that the projects support. It may be that we will need a cross-project guideline if these do get misused. -- (talk) 22:52, 17 January 2020 (UTC)
@: I don't think it would make sense to pursue a cross-project guideline, since such a thing would be hosted on Meta and would have to go through the dreadful Meta RfC process. I think proposing similar guidelines on both Wikidata and Commons would be an easier way of achieving such goals.
I think it would still make sense to indicate that people have identified as being part of racial groups (as opposed to ethnic groups), since this is obviously still fairly commonplace (e.g. people self-identifying as white). This would most likely either require splitting uses ethnic group (P172) based on the value, or broadening its definition to something like "ethnic or racial group" and periodically manually checking values which might not be appropriate (since it might be difficult to do so with a constraint violation while also retaining the current constraint violations). However, a factor that could complicate this is that contemporary usage does not necessarily distinguish race from ethnicity (e.g. "white" would be more or less synonymous with "European American" in the US). Jc86035 (talk) 02:52, 18 January 2020 (UTC)
The goals of commons and Wikidata are quite different when it comes to descriptions. In commons you want the person who edits a picture make an editoral judgement about what's the best way to classify the picture. In Wikidata we rather want to describe the different ways that external sources describe the person and we are open for external sources disagreeing and we listing multiple classification. ChristianKl❫ 08:59, 20 January 2020 (UTC)
  • While I appreciate your concerns, I do not really think that there is currently an issue. I can't really see any instance consistent misuse of these items. That being said, I think that the higher standards for sources that we have for ethnicity should be enough, as per ChristianKl, and if you are aware of any misuse that can be proven as such by sources, we can fix it. Also I would be against using such descriptions - stating where a term can and cannot be used would be out of item descriptions scope, unlike with property descriptions and property use. Regarding "obsolete concept", I feel that we cannot clasify it as such - even a quick look at the English Wikipedia article will show some conteporary uses of the concept, so it might not be an option. All in all, I think that the best course of action is to use our current policies to ensure that these items won't be misused - if you think that there is a broader gap in the policies, please do share with us your ideas. --Kostas20142 (talk) 00:53, 18 January 2020 (UTC)
  • There is the neutral failed hypothesis (Q66970557) such concepts could be made instance or subclass of, with references to the refutation(s). --SCIdude (talk) 08:13, 18 January 2020 (UTC)

PetScan problem

Oh generous gods of Wikidata, have mercy on my Luddite soul. I'm not really understanding why this query isn't returning any results. I know there is at the very least articles for Leslie Lamport (Q92613) on both projects in this category. So I know I should get at least one result minimum. But nothing. GMGtalk 18:19, 17 January 2020 (UTC)

@GreenMeansGo: The name of category you inputed contains invisible left-to-right mark (Q1022245). It should be removed in your query. ({{ping|Magnus Manske)) PetScan really should automatically remove it as it is found in many places in MediaWiki interface.)--GZWDer (talk) 19:13, 17 January 2020 (UTC) --Tagishsimon (talk) 19:33, 17 January 2020 (UTC)
Hmm. Well I'll be damned. I never would have figured that out in a million years. GMGtalk 20:13, 17 January 2020 (UTC)
Resolved now. --Magnus Manske (talk) 14:59, 20 January 2020 (UTC)

2020-01 Imperium in imperio

Should Imperium in imperio Q3796835 and state within a state Q65168291 be merged? Involved contributors: @Epìdosis, Borvan53:. Visite fortuitement prolongée (talk) 21:14, 17 January 2020 (UTC)

  •   Support They are the same concept, as far as I know. However I would like to read the opinion of someone more familiar. --Kostas20142 (talk) 00:58, 18 January 2020 (UTC)
  •   Support --Epìdosis 07:56, 18 January 2020 (UTC)
  •   Weak support The meaning is the same. But, in most languages, both expressions (latin and vernacular) are used. So why not... Borvan53 (talk) 12:16, 18 January 2020 (UTC)
  •   Oppose one item seems to be about a phrase, the other about a concept (political sciences?) --- Jura 08:23, 20 January 2020 (UTC)

Add/replace references?

I'm not entirely sure where to ask this question, but here goes:

Lots of Wikidata entries have very incomplete or even invalid references that are repeated a number of times, e.g. in the genre (P136) entry at Santa Claus: The Movie with just a retrieval date, possibly imported by a bot. I'd like to replace all occurences of this reference with a complete one (w/ title, language, stated in aso.), but doing it manually in hundreds of pages is impossible. I'm using the copy/insert references add-on quite a bit, but even that is unncessarily cumbersome when this should be done with a script. Also, it doesn't work across entries on different pages.

Is there such a thing as a script that can do this? Asav (talk) 22:16, 17 January 2020 (UTC)

That sounds like a job for WDQS and Quickstatements. The first can find items with references of interest; the second can be used to change the refs - albeit at the cost of deleting and remaking the statement. It's not trivial, since you'll need a SPARQL query - but Wikidata:Request a query - and you'll need to be able to turn the report output into a quickstatements input - see its help. hth --Tagishsimon (talk) 01:46, 18 January 2020 (UTC)
@Asav: This is a very common reference format for imported statements, and there are hundreds of thousands of these from IMDb and MusicBrainz. I don't think it's worth fixing these specifically, because neither database would be considered a reliable source. It would probably be better to remove these entirely once better references have been imported. Right now, they're not doing anyone any harm, and they're not really any worse than the millions of imported from Wikimedia project (P143) uses. Jc86035 (talk) 03:28, 18 January 2020 (UTC)
@Tagishsimon: Thank you for the heads up! I've looked at both, but I don't think I'd be able to puit together such a script. If somene could make a sample that works on one page, I'd be able to amend it to worsk on other pages as well. Do you know who'd be able to wtite such a sample script?
@Jc86035:You're absolutely wrong on this count. Literally tens of thousands of Wikipedia infoboxes in multiple language editions rely heavily on Wikidata references, so not fixing this simply is not an option. I'm sorry, but your your suggestion is not helpful at all. Asav (talk) 07:51, 18 January 2020 (UTC)
@Asav: Sorry, I didn't realize that you meant that the references were showing up in Wikipedia articles. I think not displaying these references would be possible in a Wikidata infobox, and it would probably be a better long-term solution than fixing all of these references which will (hopefully) eventually no longer be necessary. (I'm not sure if nowiki allows user-generated content, since the relevant policies don't seem to mention it, but the English Wikipedia doesn't.) I think for this particular issue it would be sufficient to add a filter in w:no:Mal:Infoboks film/wikidata and its counterparts by using w:no:Module:String. (Edit: The template uses a Wikidata module, so this is incorrect. 08:41, 18 January 2020 (UTC)) Alternatively, if the main issue is that it is unacceptable to show "bare URLs" in references, then the infobox could filter out all such references; and if IMDb is considered acceptable on nowiki (for whatever reason) then I would probably go with what Tagishsimon has suggested.
I would also note that there is an active project to integrate Citoid in Wikidata, so it will soon(-ish) be somewhat easier to generate valid references. Jc86035 (talk) 08:17, 18 January 2020 (UTC)
Again, not fixing a problem means not solving it. This needs a valid solution, not to be hidden away. If you have any constructive suggestions as how to fix the reference search/replace problem, please contribute them. Thank you! Asav (talk) 08:24, 18 January 2020 (UTC)
@Asav: The "valid" long-term solution is to remove the references and replace them with better ones. Obviously, that is not currently happening on any sort of meaningful scale, and probably won't happen for some time. Hiding invalid Wikidata references has always been a perfectly acceptable option; the infobox you're currently dealing with already uses that option, since w:no:Module:Wd shows a reference "only if at least "stated in", "reference URL" or "title" is present". Wikidata infoboxes will probably always have to do filtering on statements and references to avoid this sort of issue, e.g. because Wikidata doesn't have any policies or guidance on what sources to use, and so some references used in Wikidata will most likely always be considered unacceptable on some Wikipedias; or e.g. because of the use of valid deprecated statements on items. Jc86035 (talk) 08:41, 18 January 2020 (UTC)
I would also note that w:en:Module:WikidataIB omits references altogether purely because of the issue of formatting references in an acceptable way. Jc86035 (talk) 08:51, 18 January 2020 (UTC)
I'll just disagree with you, so I'm not following up this particular thread anymore, except to note that this is not relevant to just one type of references (e.g. IMDB) or one type of infobox, but a general problem that needs to be fixed.
Is there at least an API that I could access with an external program that would help med search/replace references? If so, where could I find its specifications, and what would I have to do to get access to it? It would only be used manually to access one page at a time, so there'd be no batch processing with all the risks involved. Asav (talk) 09:40, 18 January 2020 (UTC)
What you want to do is practically only possible with bot code, I'd say. In the past I have done similar modifications to references with pywikibot and it works fairly well, but from my experience you need to adapt the code quite often. Thus, there is no click-and-run ready-to-use script which you just needed to start, because it is simply not that easy of a task.
In the end you'd like to have a reference that matches one of the models defined in Help:Sources, probably "Web page" or "Database" types. Some of the missing reference qualifiers can indeed be complemented automatically (stated in (P248) based on the URL/ext ID ref qualifier, title (P1476) by crawling the source, etc), while others would require to verify the actual information (such as adding missing retrieved (P813) ref qualifiers). For external databases one can technically even try to crawl the source pages and scrape the information to be verified from it, but that is a real hassle and only makes sense in my opinion if there are really many cases from this external database to process. —MisterSynergy (talk) 10:00, 18 January 2020 (UTC)
Thank you; that seems like a viable solution, provided that pywikibot can handle Wikidata entries. What I was thinking is somewhere along the lines:
1. Enter a valid reference to an entry (including title, retrieval date, stated in aso.)
2. use this reference as a source (the existing copy/insefrt reference script already has som reusable code for that), then
3. search for a particular target string (e.g. a barebone reference, such as url only) and
4. replace all occurences of the target with the source.
But I'm not entirely sure you'd need a bot for that; maybe a script would be sufficient as you'd only handle one page at a time? Asav (talk) 10:14, 18 January 2020 (UTC)
The Wikidata:Bots policy is unfortunately not very helpful when you want do decide whether you need a bot account or not. From my experience, admins have different views on the matter, so in order to be on the safe side you probably want to use a bot account for such a task; that means that you need to request approval via Wikidata:Requests for permissions/Bot once you have a functioning script that can demonstrate its potential over there. Could be helpful anyways.
In pywikibot, you don't really need to re-use code from the copy/insert reference script. Using PAWS, you have already a working pywikibot installation available where only need to create the actual bot script. Something of the order of 100 lines of code would probably already be sufficient. —MisterSynergy (talk) 19:27, 18 January 2020 (UTC)
  • If IMDb ID (P345) was used in the sample above, the entity could easily include a statement that can provide further details for any reference using it. In general, I don't think reference URLs should be replaced. Obviously, one could complete them. --- Jura 09:25, 20 January 2020 (UTC)
That's not really the core of the problem, unless the data gathered from IMDB were to replace (i.e. also remove) the current barebones references, which mostly consist only of the URL and retrieval date. But I was using film entries mostly as an example because I've been working on a movie infobox based solely on Wikidata.
However, the problem is rooted much more deeply than this. There are probably hundreds of thousands incomplete references in Wikidata that need to be amended with title (P1476), stated in (P248), reference URL (P854), publication date (P577), page(s) (P304) (for printed articles) and ideally language of work or name (P407). (See discussion thread below). Asav (talk) 03:12, 21 January 2020 (UTC)


Hello. I want to discuss the structure of items of dam (Q12323).

1) Do all items must have instance of (P31)->dam (Q12323)?

2) How to show the type of the dam according to its structure? Please read w:el:Dam#Types of dams. It's not so clear, but it says the types of the dams. We have:

a) arch dam (Q890545) b) gravity dam (Q3497167) c) arch-gravity dam (Q357679) d) barrage (Q350495) e) embankment dam (Q1244922) f) rock-fill dam (Q11350418) (and concrete-face rock-fill dam (Q20180822)) g) earth-fill dam (Q11285593)

It's confusing because some items are about the shape and some about the materials. In wikidata we have that rock-fill dam (Q11350418) and earth-fill dam (Q11285593) are subclass of (P279) of embankment dam (Q1244922). But Wikipedia article don't. There is also dome dam (Q56631985).

Xaris333 (talk) 10:17, 18 January 2020 (UTC)

  1. yes
  2. I think that rock-fill dam (Q11350418) and earth-fill dam (Q11285593) will be always poor stubs. Only one stub, in Japanese, exists among these two topics ! Such articles should be merged inside embankment dam (Q1244922). Best solution, for me : put every thing as a subclass of (P279) of dam (Q12323). More eover, shape and material are always closely linked.
  3. Please also notice beaver dam (Q4878217) and all topic related with it... Borvan53 (talk) 12:32, 18 January 2020 (UTC)

But how we are going to add the information about the shape and material? Xaris333 (talk) 14:18, 18 January 2020 (UTC)

Also a dam
Don't forget that things like Oosterscheldekering (Q1893089) are dams too (and Dam Square (Q839050) used to be one). They whole waterworks ontology here seems to be a bit messy. A dam is a divider with water on both sides. So Longitudinal control dam (Q1816121) is a dam to ("strekdam" in Dutch) and breakwater (Q215635) often too. The current description of dam (Q12323) doesn't seem to cover this. Multichill (talk) 17:44, 18 January 2020 (UTC)

Another dead end in Wikidata. The mess will continue. Xaris333 (talk) 11:05, 19 January 2020 (UTC)

This is type of dam in ja-wiki. Classification by Japan Dam Foundation.

Category Subcategory Abbreviation
concrete dam (Q11302949) (concrete dam) gravity dam (Q3497167) (gravity dam) G
hollow gravity dam (Q11366848) (hollow gravity dam) HG
arch dam (Q890545) (arch dam) A
arch-gravity dam (Q357679) (arch-gravity dam) GA
multiple-arch dam (Q11341352) (multiple-arch dam) MA
buttress dam (Q1355334) (buttress dam) B
embankment dam (Q1244922) (fill dam) earth-fill dam (Q11285593) (earth-fill dam) E
rock-fill dam (Q11350418) (rock-fill dam) R
combine dam (Q11303142) (combine dam) GF
trapezoid-shaped CSG dam (Q11411941) (trapezoid-shaped CSG dam) CSG

--Afaz (talk) 13:32, 19 January 2020 (UTC)

How do I add things?

Maybe this is a stupid question, but I used the help button, went through the item tour, and searched through Help:Contents and found absolutely no information on my problem, so I have no idea where else to look for help.

I am simply trying to add *エルニーニョ・南方振動* as the Japanese-language Wikipedia entry for Q7939. When I click edit, it allows me to specify the language but not the page. It won't let me publish without a page. How do I add the page??? -- Mocha2007 (talk) 19:59, 19 January 2020 (UTC)

I eventually figured out the issue. Apparently it doesn't like it if I try to type in the whole code (ja) instead of just the first letter (j). Unfortunately there's yet another issue. I have no idea what to do, if someone else knows how to take care of it, they can. -- Mocha2007 (talk) 20:13, 19 January 2020 (UTC)
The problem is that the Wikipedia article is already site-linked to El Niño Southern Oscillation (Q14524818). If the item El Niño (Q7939) is a better match, it can be moved by deleting it from the first item before adding it to the second. Ghouston (talk) 21:55, 19 January 2020 (UTC)
These are not identical. El Nino is part of the oscillation with the same name which also includes other phenomena. Looking at the jawiki article the sitelink is in the correct place and nothing should be changed. --SCIdude (talk) 07:23, 20 January 2020 (UTC)
@Mocha2007: Wikidata collects concepts not articles. Do you understand that El Niño Southern Oscillation (Q14524818) and El Niño (Q7939) are different things, and that the article should point to El Niño Southern Oscillation (Q14524818) rather than El Niño (Q7939)? Have a look at the statements in these two items, especially has part (P527) and part of (P361). --SCIdude (talk) 07:45, 20 January 2020 (UTC)

Should the Foundation call itself Wikipedia

There is a RfC at Meta: Should the Foundation call itself Wikipedia From the reasoning of it:
The m:Communications/Wikimedia brands/2030 movement brand project plan states as a foregone conclusion that the endpoint of its process will include calling the Foundation and affiliates by the name Wikipedia, admitting, among other obvious issues, that "Wikipedia France," will likely be confused with the French Wikipedia. The community consultation it cites in support apparently did not clearly include this information when it was put to its respondents, and therefore it is not representative of authentic community sentiment on the question. To the extent that the question may have been implied, community agreement with the proposals did not achieve majority support.
Feel free to take part in that RfC. Grüße vom Sänger ♫ (talk) 15:16, 20 January 2020 (UTC)

Wikidata weekly summary #399


Is the alternate image, Property:P6802 meant to display in the Wikidata infobox if the primary image is missing? It currently does not. I thought the idea was that if we do not have an image of the person or thing something closely related was meant to display, like a cover of a book they wrote or a work of art they created. Currently we display images of gravestones and coats of arms if no primary image is available, and we have the option of displaying all of them by ticking a box in the infobox even when the primary image is present. For instance at Robert Ensko (Q7344166) since we have no image of him, the image of his book would display. --RAN (talk) 18:58, 20 January 2020 (UTC)

This is a question probably better asked at Commons:Template talk:Wikidata Infobox. In my opnion, infoboxes should not display tangential images, which can easily attract cruft and trivia. Some people would rather have infoboxes display every conceivable iota of information imaginable. Looks matter. Less is more. And in the case of Mr. Ensko, I'd argue it's better to have no image than an image of a book. Inboxes should display only the most exemplary images, not any old image available. -Animalparty (talk) 21:38, 20 January 2020 (UTC)

Wikimedia 2030 community discussions: Influence our shared future!

From today on to 21 February, the last facilitated round of movement discussions on the Wikimedia 2030 recommendations will be held. I created a dedicated page: Wikidata:Strategy 2018-20. These conversations offer the opportunity to review the movement strategy draft recommendations, and discuss how Wikidata community would be affected as well as how well they align with the strategic direction.

Wikimedians across the world have been shaping the 2030 strategy since 2017. The first phase was aimed at establishing a shared strategic direction: that by 2030, Wikimedia will become the essential infrastructure of the ecosystem of free knowledge, and anyone who shares our vision will be able to join us. This vision of Wikimedia’s future is shared by all of us, irrespective of background (such as home wiki, culture, etc.) or contributing model (some of us don’t edit, and yet, do take part in fulfilling the Wikimedia mission).

Wikimedia 2018–2020 is all about recommendations that answer the question: what systemic changes in our worldwide movement are needed to advance this vision? These new draft recommendations are intentionally broad and focused on long-term impact. However, they will inevitably be familiar to many of you. Your previous feedback was taken into account, and the recommendations are based upon both 2017 and 2019 discussions reports.

Since July 2018, a group of more than 90 volunteers from across the Wikimedia movement have worked to produce various drafts of recommendations to support this change. This effort is now reaching its final stage, and we invite you to review and discuss the new strategy recommendations. This final set synthesizes all previous drafts, and will offer a clear and condensed product for your review.

After the discussions, the recommendations will be finalized, and presented at the Wikimedia Summit in Berlin. Soon after that, the implementation phase will begin.

You can learn more about the process of forming these recommendations and the next steps in the Signpost’s text by Risker, and in a dedicated FAQ on Meta-Wiki.

SGrabarczuk (WMF) (talk) 21:59, 20 January 2020 (UTC)

Commons category question

When we have things like University of Washington Tacoma (Q7896583) and Category:University of Washington Tacoma (Q49217861), which one should the Commons category be connected to as a sitelink? - Jmabel (talk) 03:42, 26 January 2020 (UTC)

@Jmabel: The Commons category should be connected as a sitelink to the category item. If no category pages existed for the university on other wikis, then the Commons category link could be connected to the main item. Mahir256 (talk) 05:21, 26 January 2020 (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. Jmabel (talk) 06:08, 26 January 2020 (UTC)

Reducing the number of property proposal topics?

Cross-posting from Wikidata_talk:Property_proposal since I'm not sure that page is watched by many people: We currently have 13 different topics for property proposals - do we really need all of these? I quite like the approach that the status updates use, where the properties are just divided into 'general datatypes' and 'external identifiers' - would that simple split also work here? Thanks. Mike Peel (talk) 18:41, 19 January 2020 (UTC)

As far as i've noticed Authority control exist for external identifiers while Generic exists for everything else --Trade (talk) 19:26, 19 January 2020 (UTC)
For example Wikidata:Property proposal/gained territory from is in /Place but not /Generic. There are links to /all and /Overview below the topics, but nothing that just combines the non-identifier proposals, although the overview page can be sorted by category. Peter James (talk) 19:51, 19 January 2020 (UTC)
  • There are certain properties where domain knowledge is required to model them well. For people who are specifically interested in linguistics and our lexeme project it's helpful to have a list of properties proposals that deal with lexemes. Other scientific topics are similar in that we don't want properties to be created without expert input. The page for WikiCommons also seems to be valuable because it's a cluster that's interesting to certain people.
Another concern is that it isn't fun to go through a list of 120 properties to comment. I think it's more likely that people like commenting on stale proposals if they come in a chunk with similar topics. ChristianKl❫ 09:42, 20 January 2020 (UTC)
I would support merging "Sport", "Transportation" and "Sister Projects" into Generic as they aren't used much currently. ChristianKl❫ 09:08, 20 January 2020 (UTC)
  • Authority control might need splitting in two as it gets too long.--- Jura 09:14, 20 January 2020 (UTC)
    @Jura1:I think we need a better definition of what exactly constitutes an 'authority control identifier' vs a 'normal external identifier'.--Trade (talk) 09:09, 21 January 2020 (UTC)

Wiki by curid?

Most wikis are linked by page title. Recently the two above were created using MediaWiki revision id. While there are cases where the revision id is needed, in general, wikis aren't linked that way.

The question doesn't seem to have been considered during the property creation discussion.

@ديفيد عادل وهبة خليل 2, Jeb, M2k~dewiki, Looniverse, Karl Gruber, DannyS712: @Pebaryan, Krabina, Mfchris84, Gerwoman:

Can you comment on this aspect?

Unless there are some good arguments, I'd replace the two properties with standard ones. --- Jura 12:28, 20 January 2020 (UTC)

Using stable page ID instead of changing page title
Hello @Jura1:,

also Property:P6228 uses the ID of the page for example.

The page ID (curid) is given on pages like and does not change over time, while the revision id, e.g. changes with every revision. (

One advantage of using the page ID instead of the page title is that the page ID never changes, while the page title might became invalid if the page is moved on the other wiki and the old title is deleted. So, if a property uses the page title instead of the page ID (given by curid), every time the page title changes, also the property in wikidata would have to be updated (which probably never happens and the links will became invalid, since the people editing and moving pages on other wikis might not be aware that they would also have to update the wikidata property). --M2k~dewiki (talk) 12:53, 20 January 2020 (UTC)

  • @M2k~dewiki: Ok, yes "curid" is the "page id" and not the "revision id". I had thought it was the later. (e.g. Wikidata:Project Chat has pageid=394 [9], but my comment above revisionid=1101869279 [10]
Obviously, same seems to apply to P6228.
I guess the question then is if we should change page_name-properties to (new) page_id-properties for most other wikis. --- Jura 13:01, 20 January 2020 (UTC)
@Jura1: Hej Jura, M2k~dewiki mentioned all the good arguments, why i have chosen the page-id to link these wikis with wikidata and mind that it is not the revision id. for other cases like articles are splitting up, this also will have impact on the corresponding wikidata items, but therefore it will make no difference if we choose page-title or page-id. imho page-id seems to be a little bit more stable. furhtermore i think it's totally ok, to have both here on wikidata, that some properties linking external wikis are using the page-titel (which is more human-friendly to read and also to edit) and others are linking by the page-id. Mfchris84 (talk) 13:07, 20 January 2020 (UTC)
Page IDs may be probably more stable, but on most wikis less visible. In any Wikimedia project, if a page is deleted and restored, then a new ID is created, and there are deletion logs for titles but not for IDs so it is not always possible to find them, unless a revision ID or title is known or an index exists. Which is better as an identifier probably depends on the software used and whether there are many page moves or temporary deletions. Peter James (talk) 14:29, 20 January 2020 (UTC)
@Peter James: ad if a page is deleted and restored, then a new ID is created. In my opinion, usually if a page is "deleted", it is only marked as deleted, just not visible for regular uses anymore (but for admins), if it is restored then it is marked as visible for everyone again. The page ID (also the version history and all revisions) should stay the same. --M2k~dewiki (talk) 15:01, 20 January 2020 (UTC)
@Peter James: as i understand you correctly, you're also pointing out, that both - page titles and page ID has its own advantages and disadvantages, am i right? what do you prefer? if a page is deleted and maybe created under a new page id - who guarantees that the page title will be same? i personally don't disagree with properties linking external wikis by the page title, but i think page id does it also in a good way. each property has its own properties and attributes and at least each own discussion process before creation. --Mfchris84 (talk) 15:38, 20 January 2020 (UTC)
Usual I work with article ID, not with lemma or with page ID. The cause, is that the lemma can be changed and the link is broken. The article ID is the most stable connection to a wiki. --K@rl (talk) 16:01, 20 January 2020 (UTC)
It used to be that new IDs were created when pages were restored; looks like it was changed (at least in Wikidata) in 2016 (the page ID of Q22002272 changed; Q22162805 didn't). They can still disappear if there are history merges. Probably the main advantage of IDs is that they are unlikely to be reused for unrelated pages. Peter James (talk) 16:16, 20 January 2020 (UTC)
@Karl Gruber: what is the difference between an article ID and a page ID? Could you please give an example for both? Thanks a lot! --M2k~dewiki (talk) 16:25, 20 January 2020 (UTC)
I know it's not the correct definition, but I meaned the version of an article --K@rl (talk) 18:01, 20 January 2020 (UTC)
@Karl Gruber: do you mean the revision ID (oldid), e.g. vs. the page ID (curid), e.g. as discussed above? --M2k~dewiki (talk) 18:37, 20 January 2020 (UTC)
@M2k~dewiki: Correct this ID is not in the sense of a wiki connection. sorry but in German. Auch die Verbindung zum Regiowiki ist der wert aus der "Seiteninfornation". lg K@rl (talk) 21:32, 20 January 2020 (UTC)
I found WikiSkripta ID (P3471) that also uses curid. I indicated that there with Property:P3471#P1552. --- Jura 06:42, 21 January 2020 (UTC)

Question: Does this apply to all MediaWiki external identifiers? Or only to a selected few? --Trade (talk) 09:00, 21 January 2020 (UTC)

  • We seem to have only 4 using MediaWiki Page ID (Q83336089). Most others are MediaWiki page title (Q83336425)-based. Some of the later properties have external-id as datatype others string-datatype. When creating the external-id datatype, some opposed converting string-properies to external-id due to the stability issue. Most properties created for wikis since however do use external-id datatype. Wikimedia wikis generally use sitelinks (which are Q83336425 based). Not sure if this answers your question. --- Jura 09:08, 21 January 2020 (UTC)

stated in (P248) in vs. reference URL (P854)

When citing a news article as a source should i include a link to the URL or should i create an item for the news article and link to it using stated in and published in (P1433)? --Trade (talk) 01:09, 18 January 2020 (UTC)

You could create an item for the article, but unless you think the reference is likely to be used for other means beyond solely as a reference, it might be more efficient to just use something like stated in (P248) --> The New York Times (Q9684), with reference URL (P854), along with optional qualifiers like publication date (P577) and page(s) (P304). Of course, some people may argue that every article ever written should have its own item, but I don't have the time nor interest for that. -Animalparty (talk) 01:27, 18 January 2020 (UTC)
NB: I've said it before, and I'll say it again: creating references is still one of the most important but most neglected aspects on Wikidata. It's often prohibitively tedious to make good citations. Aside from the few tools like Wikidata:SourceMD that quickly and conveniently create journal article citations, we're basically medieval monks, laboriously transcribing information for the greater good. Probably most statements on Wikidata are unsourced directly, and I'm sure a large amount of items have no statements referenced, even if references exist. Thus, nobody should pretend Wikidata is reliable. -Animalparty (talk) 01:37, 18 January 2020 (UTC)
^I agree. Please don't add barebones references to items. Please use title (P1476), stated in (P248), reference URL (P854), publication date (P577), page(s) (P304) (for printed articles) and ideally language of work or name (P407). Asav (talk) 12:09, 18 January 2020 (UTC)
@Animalparty:, Wikipedia already have an easy-to-use reference generator. I wish they'll make it work on Wikidata items as well. --Trade (talk) 19:16, 18 January 2020 (UTC)
Oh, I know about those. Wikidata is still quite a clumsy toddler, big on data but weak on data management and user ease, despite being increasingly assumed to be the all-knowing mainframe of all-that-is-Wiki. -Animalparty (talk) 19:53, 18 January 2020 (UTC)
@Animalparty:WD had a weird thing where the references refuses to appear and i hate to keep reloaded the page over and over again until it works. There's only so much time i can justify spending on wrestling the references --Trade (talk) 23:21, 22 January 2020 (UTC)on each item.

Bad contemporary constraint violations when dates have different granularity

Re: Maison Hamot (Q83192538), Chimpanzee (Q29212866)

Is there anything that can be done about bad constraint violations like "contemporary constraint The entities Chimpanzee and Maison Hamot should be contemporary to be linked through manufacturer, but the latest end value of Maison Hamot is 1999 and the earliest start value of Chimpanzee is 20. century."

Is this another aspect of different understandings as to what "20. century" actually means, or is it something else? - PKM (talk) 22:48, 19 January 2020 (UTC)

  • I think the contemporary constraint understands "20. century" as 2000 (see the date at [11]). This is after 1999 on Q83192538#P576. User:Abián is currently doing a survey on the way users understand constraint violations. Maybe you want to participate? --- Jura 08:20, 20 January 2020 (UTC)
    • Actually I think "20. century" is understood as whatever the underlying date was input as -- might be 2000, might be 1901, might be 1950 -- depending on when the statement was created, and whether the inputting user overrode the default. It's not particularly satisfactory. (The other place this comes up as a particular annoyance is in the query output, where the default currently does not deliver the complexity of the date in a form the GUI can interpret). Jheald (talk) 11:11, 20 January 2020 (UTC)
      • @Jura1, Jheald: thanks for that insight. Since the date is given as "3e quart 20e siècle" in the source, I changed the date to 1950 with precision = century, keeping refine date = third quarter, and the constraint has disappeared. I'll remember that trick. - PKM (talk) 19:59, 20 January 2020 (UTC)
      • And from what I've seen coming across in Tabernacle, the input "20. century" is indeed recorded as "2000[6]" in the underlying data (unless the editor changes it manually). - PKM (talk) 20:08, 20 January 2020 (UTC)
        • You might want to drop a note to Abian. Otherwise he might omit it from his study. --- Jura 15:30, 21 January 2020 (UTC)

Gobelins Manufactory

Looking at the identifiers attached to Gobelins Tapestry Manufactory (Q732312), it appears that this item is used for both the tapestry-weaving business (it has artist IDs) and the building in which it was established. Neither use is throwing constraints because Gobelins is an <instance of> a subclass of manufactory (Q380342), which is both a “facility” and a “group of humans”.

Are we comfortable with “manufactory” as structured? Should we split “Gobelins” into separate items for the business/creator and the building? If so, which should be the current item? I’m reluctant to mess with the Gobelins item without input from someone in France. The main Commons category is about the building, and there's no Creator template yet. - PKM (talk) 15:38, 21 January 2020 (UTC)

  • Note I am the one who added “group of humans” to “manufactory” based on Getty AAT who class “manufactory” as an organization. So it’s “facility” we may want to remove. - PKM (talk) 15:50, 21 January 2020 (UTC)
  • It would likely be good to have two items named "manufactory". ChristianKl❫ 16:23, 21 January 2020 (UTC)
    • +1 on this point. - PKM (talk) 20:13, 22 January 2020 (UTC)
  • A factory is a facility which is typically housed in one or more buildings or part of a building and operated by an organization. Ghouston (talk) 05:04, 22 January 2020 (UTC)
  • Thanks, all. I will split "manufactory" and then split "Gobelins". - PKM (talk) 20:13, 22 January 2020 (UTC)
    •   Done Gobelins Tapestry Manufactory (Q732312) (tapestry makers/artist, now with Commons Creator template which was a red link before) and Gobelins Manufactory (Q83437773) (factory/historic building). I think I sorted the external identifiers properly. Categories are separated in Commons as well. The Institution template is attached to the building, but we really ought to have a separate item for the museum in Wikidata. Work for another day. Please help disambiguate these two Gobelins items in your language! PKM (talk) 22:50, 22 January 2020 (UTC)

Listeria Bot

Is something wrong with listeria bot? It made many "updates" to some wikidata list in el.wikipedia but are wrong. It deleted many informations that are still in the items. Xaris333 (talk) 19:04, 21 January 2020 (UTC)

This is often caused by lag on the replication server(s). Also will often have a log on showing the most recent changes, which can be a pain in the <dot> when using such a query to improve the data set. Edoderoo (talk) 07:38, 22 January 2020 (UTC)

Vandalism, not sure what to do

See contributions Special:Contributions/ which are all cases of vandalism. I'm new to Wikidata. What is the standard procedure for someone without rollback privileges in cases like this? Where does one report vandalism? Popperipopp (talk) 11:22, 22 January 2020 (UTC)

@Popperipopp: The usual place is the Administrators' noticeboard (shortcut WD:AN). I’ve left a note there on your behalf. --Lucas Werkmeister (talk) 14:01, 22 January 2020 (UTC)
@Lucas Werkmeister: Thanks! Popperipopp (talk) 15:03, 22 January 2020 (UTC)

Discussion on the "right way" to encode exceptions

This is the third place I've sought advice. This follows on from en:Template talk:Infobox power station#Overriding turbine height for a particular wind farm and Wikidata talk:WikiProject Energy#Wind turbine height for a particular wind farm. Each place has a slightly different focus, and this will be different again, but the same underlying problem.

The question for this forum is How should exceptions be represented in Wikidata?

The background is that the en:template:Infobox power station can create an infobox completely populated from Wikidata. For a wind farm, properties of the wind turbine such as power, tower height and rotor diameter are drawn from the wind turbine model specified on the wind farm. The specific problem is that the en:Warradarge Wind Farm (Warradarge Wind Farm (Q78439596)) uses V136-3.45 MW (Q71812982) wind turbines that can be built on a range of different height towers. I am trying to find the "right" way to express (for the enwiki infobox and for Wikidata which could end up in other infoboxes or uses) the actual height of the towers - a property of the wind farm, a qualifier on the turbine model link in the wind farm, or (the only method that currently "works") a local override in the enwiki article itself. Hub height will be 84m on V136 3.6W turbines[12][13] Thank you for any advice or assistance, and thanks to Rehman and others who have set up the infobox to work from wikidata. --ScottDavis (talk) 13:51, 22 January 2020 (UTC)

Using a height property on the wind farm seems fine to me if all wind turbines in the wind farm have the same height. A height qualifier to the specified model also seems fine, and can used if more than one model with different heights are used. Why do you say that these methods don't work? I think that current use of clearance (P2793) is wrong. It is used for the available height under a bridge. --Dipsacus fullonum (talk) 16:45, 22 January 2020 (UTC)

New product manager for Wikibase

Hello to everyone!

I’m Sam, the new product manager for Wikibase. I recently joined the Wikimedia Deutschland team to tend to the emerging Wikibase ecosystem so it can continue to thrive and grow. A lover of all things Internet, I bring previous experience in open source software with Mozilla and spent several years in the consulting industry.

My mission is to drive Wikibase enhancements that will make it possible for more people, projects, and institutions to set up knowledge bases and access the power of linked open data. Lydia Pintscher continues to serve as the product manager for Wikidata and we will be working closely together going forward.

To learn more about our roadmap for 2020, or just to “meet” me virtually, please join the upcoming Wikidata/Wikibase office hour on Telegram this Wednesday the 22nd. You can also check out the vision and strategy papers for Wikidata/Wikibase for more information.

You are welcome to email me at to share your feedback on Wikibase or start the conversation about whether it is the right tool for your institution or open knowledge project.

--Samantha Alipio (WMDE) (talk) 13:58, 22 January 2020 (UTC)

Movement Learning and Leadership Development Project


The Wikimedia Foundation’s Community Development team is seeking to learn more about the way volunteers learn and develop into the many different roles that exist in the movement. Our goal is to build a movement informed framework that provides shared clarity and outlines accessible pathways on how to grow and develop skills within the movement. To this end, we are looking to speak with you, our community to learn about your journey as a Wikimedia volunteer. Whether you joined yesterday or have been here from the very start, we want to hear about the many ways volunteers join and contribute to our movement.

To learn more about the project, please visit the Meta page. If you are interested in participating in the project, please complete this simple Google form. Although we may not be able to speak to everyone who expresses interest, we encourage you to complete this short form if you are interested in participating!

-- LMiranda (WMF) (talk) 19:00, 22 January 2020 (UTC)

How to reference?

There was some debate on Wikidata:Bistro about how to add the reference for the spelling "Écublens" (spelling at Q69716#P1705).

This seems incorrect, as the spelling is only in some pdf hosted on the domain.

What do you think? --- Jura 08:47, 20 January 2020 (UTC)

@Jura1: not pinging the person who added this claim, not a good idea.
Both the spelling Ecublens and Écublens can be found on the website How would you structured that information? List explicitly the 9200 pages with Ecublens and the 247 pages Écublens? That seems a bit overkill. Any other ideas?
Cheers, VIGNERON (talk) 11:46, 23 January 2020 (UTC)
To discuss a different hypothesis, it might be preferable to start a separate discussion and ping the users who initially participated. --- Jura 16:50, 23 January 2020 (UTC)


Is present in work (P1441) only meant for fictional people? The prose description says so but the constraints allow "human". I have been using it when a name appears in a list or a news article, but not a full description where I use: "described by source". Am I using it properly? --RAN (talk) 17:42, 22 January 2020 (UTC)

The property was proposed and created for fictional persons in fictional works, so it isn't correct to use it for real persons. Besides I see no point in linking from a person to lists or news articles that mentions them. That could give extremely many links for some people. --Dipsacus fullonum (talk) 19:54, 22 January 2020 (UTC)
What about a real person in a fictional work? E.g. Lee Harvey Oswald (Q48745) in Libra (Q1093807)? - Jmabel (talk) 23:36, 22 January 2020 (UTC)
Best practice would be to create a new item for the fictionalized Lee Harvey Oswald (Q48745) and link it to the real Lee Harvey Oswald (Q48745) via fictional analog of (P1074). But as long as the person appears as a character in a (fictional) story (so not as an actor in a film or as a person described/mentioned in a biography/obituary/newspaper article etc.) it is okay to use present in work (P1441) directly on the person's item (quasi as a shortcut). - Valentina.Anitnelav (talk) 09:09, 23 January 2020 (UTC)
  • Maybe a Wikipedia notable person is mentioned in many works that have entries in Wikidata, but we have children and grandchildren of famous people mentioned in obituaries and biographies, but they are not the subject of the biography ... they just get a mention, hence the confusion with the property. --RAN (talk) 01:05, 23 January 2020 (UTC)


wikidata project chat don't link to Wikipedia English village pump and link is in correct please fixet Amirh123 (talk) 08:04, 23 January 2020 (UTC)

Somebody missed a colon and re-sitelinked it with a [[en:, hopefully fixed. Ghouston (talk) 08:54, 23 January 2020 (UTC)

8th birthday of Wikidata in October 2020

Hello all,

Wikidata birthday celebration in Marseille in 2018

As you may know, every year around October 29th we celebrate the launch of Wikidata together. In 2017 and 2019 we organized a Wikidata conference in Berlin, and in 2020 just like in 2018, we encourage local Wikidata communities to organize events in their area.

Although October 2020 seems to be very far in the future, now is a good time to start thinking about it, especially if you need funding.

➡️ What kind of event can you organize?

The main goals of the Wikidata Birthday 2020 are:

  • share knowledge among your local community: encourage people to share what they know about Wikidata and its related tools
  • onboard new people: introduce new people to Wikidata
  • celebrate the achievements of the Wikidata community

Examples of events:

  • Meetup in your town with your local community
  • Kick-off a Wikidata section in your user group
  • An introduction to Wikidata for beginners, a training about Wikidata tools...
  • Hackathon, edit-a-thon, transl-a-thon...
  • Give an introduction to Wikidata to a local partner (university, library...)
  • Wikibase workshop or install-party
  • Online events are also welcome: editing contest, challenge, livestream...

➡️ What do you need to start an event?

It doesn't have to be a big thing like a conference. I encourage you to not start on your own: try to find other people in your area who could be interested: other Wikidatians, but also Wikipedians, your local open source/hacker/open data community...Together, you can start thinking about the following things: what do you want to organize, for what target audience, to achieve what goal?

➡️ How to connect with other event organizers?

The other events organizers will be your main source of support and inspiration. If you want to get in touch or ask questions, don't hesitate to:

➡️ How to get funding or Wikidata swag?

Wikidata birthday cake

You will find a summary of the information on this page. Here are the main points to consider:

  • In order to get any kind of funding or swag, you will need to present a detailed project of your event (what are your goals, how many people do you expect, main logistics details, etc.)
  • If you want to apply for an event grant at WMF (requesting more than $2000), the deadline is February 2nd!
  • If you want to apply for a rapid grant at WMF (between $500 and $2000), the deadline to apply is July 15th
  • Wikimedia Deutschland cannot fund events taking place outside of Germany. However we can provide a reasonable amount of swag. Because sending packages is complex in various countries, we will try to dispatch most of the swag during the Wikimedia Summit in April.

If you have questions, feel free to contact me or to write on this talk page.

In 2018, almost 40 events have been organized for the Wikidata birthday around the world - I'm super excited to see what can happen this year!

Cheers, Lea Lacroix (WMDE) (talk) 09:42, 23 January 2020 (UTC)

Bess - given name and family name

There is a single site for both in the English Wikipedia, but two separate Commons categories. This currently prevents merging the items:
Q16275032 is a name, links wikipedia and commons:given name - Q37526137 is a surname, links commons:family name.
What to do about it, merge both or proper items for family name and given name? Are there similar cases with other names? Frlgin (talk) 20:35, 23 January 2020 (UTC)

How many unique articles across all Wikipedias?

English Wikipedia hit 6 million articles today, and it got me curious - how many unique articles are there? I.e. if one would translate all articles across all Wikipedias to English, how many articles would that be? I would think that's a theoretical max size for English Wikipedia.

From I see that there are 51,919,749 grand total articles across all languages of Wikipedia. But how many of them are unique? I would think the easiest way to estimate that would be counting sitelinks to Wikipedia main space on Wikidata items. Something similar to the table "Sitelinks per item" in User:Pasleim/Sitelink statistics but specifically for Wikipedia main space. Ping User:Pasleim. Renata3 (talk) 03:10, 24 January 2020 (UTC)

  • Maybe starting out from the number of items (75 million),
removing items without sitelinks (48 million) [14]
removing categories (4 million) [15], disambiguations (1 million) etc.
removing items that don't link to a Wikipedia
etc ..
eventually gets you there. Not sure how large the impact would be if one also allows for English Wikipedia covering in a single article or list what is covered in multiple articles elsewhere.
  • Another approach could be to check the overlap between enwiki and some of the other Wikipedias. --- Jura 11:30, 24 January 2020 (UTC)
I expect there are 19,630,868 unique articles across all Wikipedias [16]. --Pasleim (talk) 13:56, 24 January 2020 (UTC)
Very cool, Pasleim! I have no experience in coding, so excuse me if it's a silly question. But does the query exclude all titles with ":" in them? So articles like Avatar: The Last Airbender are excluded from the count? Not a big deal, just try to understand the caveats. Also, it seems the count includes dab pages. Any way to exclude them? Or count them manually and then subtract them out from the 19M figure? (it's funny -- I was guesstimating 20M). Renata3 (talk) 16:34, 24 January 2020 (UTC)
Renata3 There are about 1.3m disambiguation pages, and while it's not easy to figure out how many have at least one WP link, my guess is that the vast majority do. So that would suggest a total count around 18.3-18.5m.
You're right that this will miss out any articles with a "real" colon in the title, but on the other hand it will also count any sitelinks which are redirects as valid articles, so maybe it balances out. :-) Andrew Gray (talk) 19:17, 24 January 2020 (UTC)
But I think, there is no such thing as a theoretical max size as initially stated, since the world is constantly changing, new companies and organisations are founded, new people (politicians, sports people, actors, ...) come into relevant positions, new things are discovered, new films, albums, books are produced, etc. Also if you change the definition what is notable for wikipedia (e.g. rise or lower the criteria in some sections or language versions) also the number of articles might change (grow or shrink). --M2k~dewiki (talk) 14:18, 24 January 2020 (UTC)

Google Data Search

Google have a new feature, a data search:

which is possibly relevant for us. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:23, 24 January 2020 (UTC)

Open call for Project Grants

Greetings! The Project Grants program is accepting proposals until Feburary 20 to fund both experimental and proven projects such as research, offline outreach (including editathon series, workshops, etc), online organizing (including contests), or providing other support for community building for Wikimedia projects.

We offer the following resources to help you plan your project and complete a grant proposal:

With thanks, I JethroBT (WMF) (talk) 18:38, 24 January 2020 (UTC)

what is allowed to have an item?

On Help:Items#Notable_items you find this text: "Wikidata items are said to be notable. This means at least one of three things:

  • a corresponding page on a Wikimedia site such as Wikipedia, Wikivoyage, Wikisource, Wikiquote, Wikiversity, or Wikimedia Commons,
  • a clearly identifiable conceptual or material entity that can be described using serious and publicly available references or
  • the item fulfills a structural need. For more information look at our notability policy."

I'm thinking about the second one. What is this eg. "clearly identifiable [...] material entity" that can be described using "serious and publicly available references". I read there is Wikidata:Notability, but that does not help in these edge cases. A small list of things I'm thinking about:

  • streets - QID: Yes, they are unique and I guess notable enough
  • lines of public transport - QID: Yes, some of them have whole Wikipedia articles about them and I think you can call them notable
  • art in public places - QID: Yes, as long as they are build to stay where they are

the cases on the far edge I'm interested in:

  • trees - In some cities each tree has it's own number and there is public information about the species, where it is planted, etc. But why should a single tree (or a bunch of them) be notable even they have their very own ID (except they may break down and kill someone)
  • charging stations (with public access) - All charging stations have an ID and information that is available for the public where you can check things like available sockets and power
  • all buildings in general - There is a lot of information about buildings in general, but do they really deserve their own item? Well known buildings like the Eiffel Tower, the Golden Gate Bridge or the Fenghuoshan Tunnel may is. I have seen that buildings like firestations and churces get their own item. But what about that not so special house next to the church that was there before the church which burned down several decades ago. What about every other house in town? Every house used by some family will not be notable, but are eg. skyscrapers even they serve the same purpose?
  • factories - if the area is used by a well known brand name, maybe. If the area is used by a small one-person-company, probably not.
  • restaurants/stores - for the well known brand if it's a big store, maybe. I would agree if it's a shopping mall. For the small asian restaurant around the corner that opend a week ago, no

These are just my personal thought. I try to get the unsharp "notable" more in focus. Eg. self-promotion items of some artist, probably not. A photographer from Commons - these items already exist even you could argue about the notability here. But since they are allowed to have images of themselves on Commons, why not an Item on wikidata.
I would love to hear your thoughts on this topic! --D-Kuru (talk) 19:50, 13 January 2020 (UTC)

@D-Kuru: Our notability requirement, point 2, says "described using serious and publicly available references" - note the plural. If something is described only in a single database, then I think it does not qualify. I believe we also have a consensus that a facebook page or other similar self-curated source also doesn't confer notability. Some of these things belong elsewhere - perhaps other wikibase instances. Streets and possibly public transport lines I would mostly keep out of Wikidata. Public art I think is fine to add, at least for clearly identifiable unique artworks. Trees - some individual trees may be added (we already have some like Neuhaus Oak (Q114810)) but for the most part, I don't think so. Charging stations I'd class with public transport lines - maybe, but probably better elsewhere. Buildings/factories/restaurants etc. could belong here, we certainly have a lot of them already. But not every house! ArthurPSmith (talk) 20:27, 13 January 2020 (UTC)
  • There are quite a few articles about streets in major cities on en-wiki, and even more Commons categories; necessarily these get Wikidata items. Special:Search/Streets in Manhattan will bring up several hundred examples (plus a comparable number of false positives). - Jmabel (talk) 01:10, 14 January 2020 (UTC)
  • I think there are plenty of cases where data about streets is useful. named after (P138) allows us to store information that OpenStreetMap doesn't store well.
I feel like it would be great to have items for all restaurants in a given city if we would have a strategy for keeping the items up to date. Whether or not that's the case is however different in different cases. As a result I think it's good to have bot discussion when bigger chunks (>10,000) of new content are planned to get added. ChristianKl❫ 11:06, 14 January 2020 (UTC)

Thank you for your comments! In some parts people disagree (well, who would have thought, right?). Eg. If streets are not important (other than storing data OSM can't handle well), but keep every restaurant (for whatever purpose) that would be strange. I guess that you have to discuss about every item individually which makes adding content harder since you have to create a first item and then discuss about it. If you plan for something larger you want to work on it and not wait for months or maybe even see the discussion die out and nobody cares in the end (something I saw a lot on Commons).
@ChristianKl: But where would you start a discussion for such items?
Are there any more opinios on that? if you completely disagree it's even better so that I get a view from the other side!
--D-Kuru (talk) 21:39, 15 January 2020 (UTC) is the page where a lot of relevant discussion like the one about whether we want to import all German companies happen. ChristianKl❫ 20:22, 17 January 2020 (UTC)
@ChristianKl: Thanks for the note, I will take a look at it! --D-Kuru (talk) 19:01, 18 January 2020 (UTC)

It is what it is. Trying to make "all or nothing" rules about various types of item is, in my opinion, unhelpful. For example, the British Government publishes a database of all bus stops in the UK. Each of the bus stops listed is therefore "a clearly identifiable conceptual or material entity that can be described using serious and publicly available references" and some day I intend to get that information into Wikidata. Does that mean all bus stops, everywhere in the world are suddenly "notable"? No, because not all bus stops are "a clearly identifiable conceptual or material entity that can be described using serious and publicly available references". Can I say "Well if that bus stop is notable then this building next to it must also be notable because a building is more substantial than a bus stop"? No. The bus stop is a clearly identifiable conceptual or material entity that can be described using serious and publicly available references, but the building might not be. If the building is notable according to Wikidata's definition, it would be because it's a clearly identifiable conceptual or material entity that can be described using serious and publicly available references, not because it's bigger than an entity we have an item for or just because it belongs to a class of items we've decided should all be included. Waggers (talk) 13:34, 17 January 2020 (UTC)

@Waggers: "All or nothing" more or less has to be in use for at least some categories, because of the structural need criterion. If a bus stop is notable, then the next stop on the line is also notable because of that criterion (to allow for valid adjacent station (P197) statements), and then the next stop, and then the next stop. This issue also applies to (certain groups of) people, streets, creative works, and pretty much any other group of entities where not all the items are Wikipedia-notable but most or all of the items are related to each other in some way. Because only one of the criteria needs to be met in order for an item to be notable, the policy does currently allow for items which are only notable per the structural criterion. Jc86035 (talk) 17:34, 17 January 2020 (UTC)
@Waggers: Sure there can't be an all or nothing approach in the first place since you have to start somewhere. However, would a bus stop in the UK be more notable (or notable at all) just because it has a number and is listed on an official page? What about bus stops that are not listed by the government? Eg. a bust stop in the middle of Africa, that is called "Toto stopped here". The bus station does not have a number, but a sign and a shelter and the bus stopped at this position since 1982 and it is an important bus stop in the area. Is this bus stop now less notable because it is not listed on some governmental page? Or what about bus stops that maybe are in the UK, but are located in Scotland where no english man dares to set foot into the area?
To condense it: It has to be all or nothing since one is not more or less important than the other based on the simple fact if there is a reference for it or not.
To have an example ready: EV charging station at Kasernenstraße 13a, Eisenstadt (Q83009020)
I added pretty much everything there is to add for a charging station. You can find it on some sort of official map (bad example, there are much better ones made by electric companies) and it has a serial number. So it is clearly identifiable and in the whole world there will note be a second charging station with the same serial number since said number includes the country code. And yet, it is just a charging station. One of many and many more to come. Is this now a notable item or do I have to ask for the item to be deleted?
--D-Kuru (talk) 19:01, 18 January 2020 (UTC)

I'm trying to get a clear idea of this myself, before I start adding additional items. The example I'm looking at:

  • I found an OpenStreetMap POI for a piece of public art in my city, which had a Wikidata tag.
  • The Wikidata entry is Two huskies (Q56679029)
  • I found the name of the sculptor, via a news article used as a reference on the item
  • From this I found his website with a list of additional public artworks
  • Now, is the artist notable? Maybe not for a Wikipedia article. But his has four other pieces of public art, and having a Wikidata entry for the scupltor could link them together, which I believe aligns with the overall purpose of Wikidata

Just trying to get additional opinions before I start creating... -- Chuq (talk) 13:59, 25 January 2020 (UTC)

Bollywood Hungama person ID (P3910)

Hi, all. The property Bollywood Hungama person numeric ID (P3910) should be updated. The website Bollywood Hungama (Q4939834) changed its URL format. Both the formatter URL and the format as a regular expression properties should be changed. (Also, note that the website switched from http to https.)

So, the new formatter URL should be added and the old one should be marked as depreciated. I am starting this discussion to see if any one objects before actually making the changes. Best wishes. --Meno25 (talk) 11:01, 21 January 2020 (UTC)

I can't imagine why anyone would object to this. Just go ahead and update the URL.--Trade (talk) 12:20, 21 January 2020 (UTC)
  • Sure, please use Wikidata:Property proposal for the new scheme/property. Wikidata can't handle two properties in one. --- Jura 13:15, 21 January 2020 (UTC)
Thank you for your help. Request submitted at Wikidata:Property proposal/Bollywood Hungama ID 2. --Meno25 (talk) 22:48, 25 January 2020 (UTC)

What is badly wrong with UvA Album Academicum-identificatiecode (P6815)?

Checking Special:WhatLinksHere/Property:P6815 I learn:

  • The German-Romanian Herta Müller, born in 1953, was a special professor of literature and philosophy in the last year of the Second World War. The Doopsgezind Seminarium was also one of her employers, starting in 1827. Assuming that she will remain active for a while, this Nobel Prize winner will soon have a career that spans two centuries.
  • The drummer Christoph Schneider cannot compete with that, but this multi-talent was certainly a prodigy with a professorship at the age of two.
  • Karl Wolff pales by comparison, but in the context of post-war sentiments it is remarkable that this SS General, convicted of crimes against humanity, could teach at the Dutch Universiteit van Amsterdam (UvA) in 1955.
  • Finally, Pál Turán. This Hungarian may have taught at the UvA Faculty of Mathematics and Natural Sciences in 1973, but none of the Wikipedias says anything about this.

I didn't have to search as these were the top four of the list with P6815. Are these people just confused with namesakes or is something else going on?

UvA has recently added two leading zeros to its ID code which may have kickstarted these nonsens. Anyway, now it looks as if biographies have become intermingled. Hope someone can untie the knots! — bertux 15:41, 23 January 2020 (UTC)

I don't think it has anything to do with the leading double zeros. They just seem to be cases of mistaken identities. The reference of employer/University of Amsterdam refers to Kurt Müller while it has been have been put at the item of Herta Müller. I don't mind adjusting these cases but it won't be tonight. (@Hannolans: for completeness sake), Ecritures (talk) 16:41, 23 January 2020 (UTC)
I don't understand with what you mean with Pál Turán. This Hungarian may have taught at the UvA Faculty of Mathematics and Natural Sciences in 1973, but none of the Wikipedias says anything about this.. The info he was a university teacher at the UvA is mentioned in the UvA Album Academicum. Why would you want wikipedias to mention it before it is 'true'? Ecritures (talk) 16:41, 23 January 2020 (UTC)

I had a quick look with tabernacle to show the professions of the items connected to this property, and most of them were university teacher (Q1622272), so I think that the first four items checked were all in error are a coincidence, or due to the way the results were sorted. The problem is probably not as huge as mentioned, to my idea. Edoderoo (talk) 16:47, 23 January 2020 (UTC)

They are in statements, not identifiers - added in and - I moved the statements mentioned here to the correct items (creating new items where I couldn't find one) with the exception of Pál Turán where it is the same person, but as those are just the first links to the property, there are probably more. I looked at some items, and many were new items; of the existing items it's possibly only those with lower Q numbers where this has happened - Cornelis de Vos (Q474934) is another. Peter James (talk) 16:51, 23 January 2020 (UTC)
I found out that they were used as a source, indeed. Sometimes correct, but sometimes not sure why they were added. Every single Dutch person knows that Laura Dekker has never been a teacher at the UvA, and the source was about someone with the same lastname, but that is not a good reason to call a search a good match. This is just a wrong upload with OpenRefine that has not been noticed before, and needs to be fixed by just removing the edits of Hanno Lans that added them. Edoderoo (talk) 17:01, 23 January 2020 (UTC)
Ah, well, I think in OpenRefine some later refined matched qids were still present. But it is simple to check as every person should have an UvA id. Please don't remove the batch. --Hannolans (talk) 17:10, 23 January 2020 (UTC)
This is the listeria list with employees of the UvA. Some of them don't have a UvA-code and are suspicious. I will have a look at them. --Hannolans (talk) 17:12, 23 January 2020 (UTC)
It's the ones without occupation (P106)/subclass of (P279):university teacher (Q1622272) that are more likely to be wrong - I found 391 items (now 390 as I moved the statements from Marcell Jansen (Q152940), a footballer). Peter James (talk) 18:02, 23 January 2020 (UTC)
Some weren't added in this edit group. Sharadindu Bandyopadhyay (Q117737) has "employer: University of Amsterdam" without any references, added by GerardM[17], and I couldn't find anywhere else that this is mentioned. Peter James (talk) 18:36, 23 January 2020 (UTC)
I verified all the items without profession is university teacher or researcher and moved around 20 people. I think it is more or less fine now more or less. Would like to have the mix n match correct as than we can do some iterations and checks.--Hannolans (talk) 00:17, 24 January 2020 (UTC)
Thanks for the corrective support! Edoderoo (talk) 20:12, 25 January 2020 (UTC)

D links on my Wikiversity Watchlist


Occasionally, when someone here makes a change to, for example, Q25670, I receive a message in my watchlist like " D Draft:Metadefinitions (Q25670); 12:50 . . MichelBakni (discuss | contribs) (Language link changed ...". Not only did I remove the link from WikData to Metadefinitions here on Wikiversity after it was moved to a Draft: namespace, but the Q ID for metadefinitions was deleted. So does this message mean that WikiData tracks every item entered where a link originally existed or is this some kind of database residual? --Marshallsumter (talk) 02:57, 25 January 2020 (UTC)

As you can see on digital object identifier (Q25670), it is stated as connected to v:Digital object identifier. The enwikiversity page is deleted, but the Wikidata item isn't updated about that. Maybe you see the changes to the item in your watchlist because the deleted page is still in the watchlist? --Dipsacus fullonum (talk) 19:20, 25 January 2020 (UTC)
I have now confirmed this, and reported it on Phab:T243697. It doesn't matter if the deleted page is in the watchlist. --Dipsacus fullonum (talk) 20:06, 25 January 2020 (UTC)
It seems that v:Draft:Metadefinitions actually uses Wikidata to fetch a sitelink from Q25670. It happens in the Cite book template. The template uses the sitelink to link from the text "doi" in citations with a doi parameter, so it is correct to show updates to Q25670 in the watchlist. --Dipsacus fullonum (talk) 01:24, 26 January 2020 (UTC)

Ranks update

Does Sparkql allow the standard SQL Update command or else is there a tool that can massively change the rank of a statement ? --Yanik B 13:30, 25 January 2020 (UTC)

If everything else fails: WD:RBOT could be worth a try then. Edoderoo (talk) 20:13, 25 January 2020 (UTC)
As far as I know SPARQL 1.1 Update is not supported. --Succu (talk) 20:48, 25 January 2020 (UTC)

Author name string

After I Create a Q entry for an author, should I delete the duplicated name in the field called "author name string" at the entry for the book they wrote? We created "author name string" for when we do not have a Q entry for that person ... so what to do once we create one? Keep as duplicate or delete? See for instance Scots in Sweden (Q83590701) --RAN (talk) 05:27, 25 January 2020 (UTC)

see Wikidata:Property_proposal/Archive/39#P2093. --- Jura 07:12, 25 January 2020 (UTC)
tl:dr. Delete. --Tagishsimon (talk) 05:42, 26 January 2020 (UTC)

Links between Lexemes and Wiktionary?

On Wikipedia, in the "Tools" part of the sidebar, there is a link to the Wikidata item for the article you are reading. Is there any way that I can link Wiktionary articles to Wikidata Lexemes (or vice versa) like this? DemonDays64 | Talk to me 19:49, 25 January 2020 (UTC)

@DemonDays64: This would not be appropriate, as 1) a single Wiktionary page can correspond to multiple lexemes and 2) interwiki links between Wiktionary pages, which are already handled via mw:Extension:Cognate, would be broken if any links to Wikidata items (or lexemes, were that possible) were made. The closest we may be able to get with respect to such a linkage is arbitrary access for lexemes, which IIRC has yet to be enabled. Mahir256 (talk) 05:33, 26 January 2020 (UTC)

Content deletion by ‎Sapphorain

@‎Sapphorain: Can you explain what you are trying to do and why? --- Jura 13:36, 23 January 2020 (UTC)

A few pointers:

  • labels, descriptions and aliases usually duplicate content from one or several statements.
  • labels and aliases are used to make content searchable
  • The string "🇺🇸" is not a flag, but a regional indicator symbol. Depending on the tool it functions as "US".
  • If one pastes it into search box, it finds the corresponding item.

--- Jura 13:47, 23 January 2020 (UTC)

I am not deleting anything. I removed duplicate instances of unicode symbols for countries, which were listed as "aliases". The unicode symbol is already entered in the element at Unicode character (P487). (I recall that an alias is an alternative name, and that a name is a word or phrase by which a person, thing, or class of things is known, called, or spoken to or of. Thus a unicode symbol representing the flag of a country does not qualify to be an alias, as it is not a name.) Sapphorain (talk) 13:49, 23 January 2020 (UTC)
  • As mentioned above, it's not necessary to delete aliases or labels if there is a statement with the same information. --- Jura 13:52, 23 January 2020 (UTC)
  • Wikidata aliases include symbols and abbreviations such as US to make item searchable by these. The lexical definition for the word "alias" might not help you much for this.
    Technically, I think the regional indicator symbol represents the country and not the flag, is that correct? It's a like the string "US", but in one character.
Is there some added value in these deletions? --- Jura 14:04, 23 January 2020 (UTC)
  • As mentioned above, the symbol has been removed from aliases not only because it appears elsewhere on the page, but because it does not satisfy the definition of an alias. The fact that it represents the country doesn’t change that fact: it is a picture, and not a word or phrase. Sapphorain (talk) 14:12, 23 January 2020 (UTC)
    • It doesn't satisfy your definition. Besides, it's a regional indicator symbol, not a "picture". Anyways, what would be the added value of such deletions? --- Jura 14:15, 23 January 2020 (UTC)
  • It is not my definition, it is the definition you will find in every good dictionary. The added value of such a deletion is thus simply to correct a mistake. Sapphorain (talk) 14:30, 23 January 2020 (UTC)
    • As said, you are using the lexical definition. Anyways, what would be the disadvantages for you? --- Jura 14:32, 23 January 2020 (UTC)
  • A definition is necessarily lexical. A "lexical definition" is what I call a pleonasm. The disadvantage would (should?) be for everybody, not just for me: one should never like to let a mistake stay. Sapphorain (talk) 14:41, 23 January 2020 (UTC)
  • Ok, so there is no practical disadvantage for you. Let's see what others think. --- Jura 14:44, 23 January 2020 (UTC)
  • A mistake is a practical disadvantage for everybody. Sapphorain (talk) 14:52, 23 January 2020 (UTC)
  Keep the aliases, IMHO. Sapphorain, the information at WD:Glossary#Alias and Help:Aliases is more useful than a dictionary definition here, and both of those pages emphasize the usefulness of aliases for the practical purpose of searching. --Lucas Werkmeister (talk) 15:28, 23 January 2020 (UTC)
  • As Lucas points Help:Aliases explicitly that calls for unicode characters. While it's not ratified as policy it seems like it's the current status quo of how we do things and much more important then the definition of any dictionary. ChristianKl❫ 22:35, 23 January 2020 (UTC)
  • I find this type of justifications very problematic. Precision should be a priority in such a project. What’s next? If it appears convenient we’ll mix up paintings with photographic pictures? Coats of arms with family names?Sapphorain (talk) 09:17, 24 January 2020 (UTC)
It is your conception of the limited scope of aliases that is problematic, Sapphorain. The guidance is clear and the purpoe is easily understandable, but you choose to ignore them both. --Tagishsimon (talk) 09:51, 24 January 2020 (UTC)
I'll grant you that. Our points of view are so exactly opposite it's not worth continuing a discussion. Sapphorain (talk) 10:36, 24 January 2020 (UTC)
  • Maybe Help:Aliases could use an update. Many things usually added in that part of an item aren't mentioned. --- Jura 11:16, 24 January 2020 (UTC)
    @Jura1: I would agree with that. I think this is more of a general problem – in practice, a lot of Wikidata's conventions (particularly those related to labels, descriptions and aliases) are unwritten, which (as this discussion probably demonstrates) can be less helpful than properly codifying everything, even if doing so might be seen as overly bureaucratic. Jc86035 (talk) 13:42, 26 January 2020 (UTC)
    • Many guidance pages are written to help people start adding things .. not necessarily to delete everything that might not be supported. Wikipedia is generally much more in need of content deletions/updates as it only supports a single current view. --- Jura 17:18, 26 January 2020 (UTC)

Language is shown as unknown for English (Q1860)

This is weird. In search, when I look up a lexeme, e.g. learn, in search it says "Unknown language". On the page for the lexeme (e.g. learn (L4176)), it says the Q id rather than "English". I noticed this maybe a week ago.

What is causing this bug? DemonDays64 | Talk to me 06:27, 26 January 2020 (UTC) (please ping on reply)

@DemonDays64: Interesting. This doesn't seem to be happening for other languages (e.g. Special:Search/lexeme:chat), and it also isn't happening when the interface is localized into different languages, which is pretty unusual. I don't know why it's happening, though, so I've filed a bug. Jc86035 (talk) 08:47, 26 January 2020 (UTC)

UK local newspapers and distribution areas

I am interested in expanding the dataset of UK local newspapers (?item wdt:P31 wd:Q1868552. ?item wdt:P17 wd:Q145.) In particular, I want to query distribution areas by postal outcode to answer the question "which local newspaper(s) can be bought in Postal Area NG13?" etc.

Is there a precedence to how a distribution area might be stored?

The (Dundee) Evening Standard states 'place of publication' P291 of Dundee Q123709 but this is a little vague, and perhaps is not directly equivalent to 'distribution area'. Jnicho02 (talk) 14:15, 25 January 2020 (UTC)

You have at least three issues: 1) finding decent sources specifying distribution areas 2) finding an appropriate property to record the distribution area; perhaps operating area (P2541), perhaps KML file (P3096) with a qualifier:
⟨ item ⟩ KML file (P3096)   ⟨ foo.kml ⟩
subject has role (P2868)   ⟨ QId for an item meaning distribution area ⟩
and then 3) putting together a mapping between postcode areas - on which our coverage is pitiful (cf. NG postcode area (Q6953990)) - and whatever scheme for recording distribution areas. You're right that place of publication (P291) is not appropriate. I've not found any more appropriate properties than the two I mentioned. Worthy but very uphill task. --Tagishsimon (talk) 02:13, 26 January 2020 (UTC)
Thank you. 'Operating area' sounds right for a newspaper....and yes, it'll take a fair bit of legwork to compile, but I may have a willing supplier. Jnicho02 (talk) 13:54, 27 January 2020 (UTC)

Allowed characters in labels/descriptions

Where can I find a list/regex of all supported characters in labels/descriptions/aliases? I recently managed accidentally to include an object replacement character in a label. Is this something that should be locked down a little further? Danielt998 (talk) 12:23, 27 January 2020 (UTC)

Wikidata:Glossary doesn't have definitions for Q numbers or L numbers

Hi all

I'm writing some documentation and noticed in Wikidata:Glossary doesn't have definitions for Q numbers and L numbers which feels like a noticeable gap. There is an entry for Items but this doesn't really talk about the Q and L number as a way of uniquely identifying things. Does anyone have any suggestions of definitions?


John Cummings (talk) 14:50, 27 January 2020 (UTC)

  • How about "QID" see "item" ? --- Jura 14:57, 27 January 2020 (UTC)
They do now. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:50, 27 January 2020 (UTC)
Thanks very much Andy, I've added a bit more detail for the Q number, perhaps others can help with that. --John Cummings (talk) 16:54, 27 January 2020 (UTC)

Wikidata weekly summary #400

How to indicate an identification is questionable ?

Suppose I have an item for somebody in one context -- eg Richard Strange (Q16847651), a man from Hereford who was chosen as MP between 1414 and 1420.

Suppose I have another source, with an identifier -- eg STRA14 -- which fundamentally relates to the indexing of a 1598 book, that said that "Richard Strange, son to Roger Strange" was buried in the church of Blackfriars in London. ([18], p.276).

The external site says that the two are the same person; but doesn't give evidence; nor does a quick Google search turn up any evidence; and it seems to me that prima facie the identification is unlikely. (And the external site has made other errors -- its identifications aren't necessarily gold-standard reliability).

How to indicate this on the item? I'm wary of using sourcing circumstances (P1480) since it's not the source that says the identification is questionable. I can't use statement disputed by (P1310) since I don't actually have a source that the identification is not correct. (Also it's possible the external site has supporting evidence I'm not aware of). But how else might one indicate that this is an identification that doesn't feel right? Jheald (talk) 08:51, 22 January 2020 (UTC)

@Jheald: I think list of Wikidata reasons for deprecation (Q52105174) has several promising candidates as reason for deprecation (P2241) qualifier values: not been able to confirm this claim (Q21655367), source known to be unreliable (Q22979588), cannot be confirmed by other sources (Q25895909), possibly invalid entry requiring further references (Q35779580), or implausibility (Q16886573). (Though none of those let you specify exactly why you think the statement is implausible / needs confirmation etc.) --Lucas Werkmeister (talk) 11:03, 22 January 2020 (UTC)
@Lucas Werkmeister: Thanks, Lucas. Those are exactly some of the things I might want to say about the statement. But I'm not sure I want to deprecate it, since their website is explicitly making the connection to our Wikipedia article. I suppose the question to the wider community is: do we have any way to tag a statement with the equivalent of "citation needed" or "further evidence required", other than deprecation? Should we? Jheald (talk) 11:11, 22 January 2020 (UTC)
I do believe we should have a rank between the normal rank and the deprecated rank for statements where we are uncertain. That includes poorly sourced statements and also statement created by machine learning algorithms without human verification. ChristianKl❫ 15:03, 22 January 2020 (UTC)
Please note that there is a whole infrastructure for giving the determination method (P459) of a statement, used by those databases that rely a great deal on electronic annotation (molbio). Note also there is no reason to believe that statements made by humans are any more reliable than electronic annotation. --SCIdude (talk) 16:57, 22 January 2020 (UTC)
@SCIdude: Even so, there may be statements the algorithm asserts with 99.999% probability, and others with only 90% probability. It would be nice to be able to record the 90% ones as suggestions, rather than just throw them away. This may particularly be an opportunity in eg machine identification of items in images. Jheald (talk) 17:35, 22 January 2020 (UTC)
@Jheald: Have we mentioned nature of statement (P5102) as a possible solution to the original question? Seems packed full of handy values. --Tagishsimon (talk) 00:36, 23 January 2020 (UTC)
That may be the best way to go. As per an earlier thread here, I get confused as to the exact demarcation between nature of statement (P5102) and sourcing circumstances (P1480). But maybe P5102 with "possibly unreliable source" or "possibly unreliable identification" could be a way to go. Jheald (talk) 11:11, 23 January 2020 (UTC)
Just to add, I'm also finding this with some of the genealogical links, added with eg the Peerage import, and (if we're lucky) also matched or matchable to sites like Wikitree etc. Sometimes the generalogical relations proposed may be the best available, but at the same time be ultimately derived from quite weak sources (eg unreferenced accounts maybe written two centuries later), and may eg sometimes appear to squeeze in more generations than there is actually time for. In such cases, do we have preferred ways to indicate "best information available, but may not be quite right"? cc @Salgo60:. -- Jheald (talk) 11:11, 23 January 2020 (UTC)
That property has some use, but if the data user asks for a truthy value, they wouldn't see the qualitication when it's applied to an item with normal or higher rank. I think it would be good to have a new rank that's not truthy but which is also distinct from deprecated as using the depricated rank is also misleading here. ChristianKl❫ 14:23, 28 January 2020 (UTC)
@Jheald: I tried sourcing circumstances (P1480) possibly (Q30230067) when I connected some external identifiers were they havnt added information so I could be 100% its the same person, any opinions if that is good bad? - Salgo60 (talk) 11:25, 23 January 2020 (UTC)
@Salgo60: I think P1480 should only be used if the source itself describes the relationship as "possible". But I'm not 100% certain about that. (Also, what if some sources caveat the relationship, but others present it at face value). So I am not sure. Jheald (talk) 11:35, 23 January 2020 (UTC)
Personally, I would feel uncertain enough about this link that I would just mark it deprecated & create a qualifier reason such as "connection considered unlikely"? But that's not a good solution for the general case! Andrew Gray (talk) 20:30, 24 January 2020 (UTC)

Splitting and merging fictional characters

At the moment we have an odd situation where some fictional characters who appear in different adaptations of the same work have two or three items. For example, the protagonist of Little Women is at Jo March (Q27902552), but we also have items at Jo March (Q63882404) and Jo (Q63535345), each of those for the character in a specific musical version.

My feeling is that in general, we should merge these into one item per fictional character, unless the adaptation is dramatically different - when it comes down to it, everyone who's played Jo March is basically playing the same character.

For real people in works of fiction, it makes sense to use the real person's item for things like character role (P453), and only split out a fictional-character item where they're really substantially different - eg as well as the real Julius Caesar (Q1048), which is also the character who normally appears in fiction, we have items for Julius Caesar (Q3188509) (in the Asterix stories), Gaius Julius Caesar (Q5517445) (in the Rome TV series), and Caesar (Q5016966) (in the Xena TV series).

Does this sound reasonable? If so, I'll start merging some of the extra character items... Andrew Gray (talk) 12:57, 26 January 2020 (UTC)

The point about having separate items for fictional counterparts of real people, distinct from the items for the actual real people, and only letting fictional characters interact with other fictional characters, is important. Otherwise we find statements like this getting batch-added to the items for real people. (Much as I have enjoyed The Last Kingdom (Q18085820) / The Saxon Stories (Q7762369), I don't think we really ought to be asserting that Æthelflæd (Q235250) daughter of Alfred the Great and ruler of the Mercians had her head turned by Uhtred of Bebbanburg (Q7878021)).
As for merging the entries for the same fictional or fictionalised character when they appear in different works, that seems appropriately parsimonious and prudent (unless we have to deal with multiple articles on any Wikipedia for the different character versions). But it's probably then necessary to qualify statements according to which work or series of works they apply to -- eg statements about Mina Harker (Q2570236) should probably distinguish whether they apply to Dracula (Q41542) or Dracula (Q62033168) or The League of Extraordinary Gentlemen (Q95029). Jheald (talk) 21:36, 26 January 2020 (UTC)
pinging relevant projects:

Valentina.Anitnelav Thierry Caro Shisma (talk) Arlo Barnes (talk) Tsaorin (talk) 16:37, 12 November 2019 (UTC) Nomen ad hoc   Notified participants of WikiProject Narration TomT0m Valentina.Anitnelav Shisma (talk) ElanHR Arlo Barnes (talk) Maria zaos (talk) EEMIV (talk)   Notified participants of WikiProject Fictional universes Beat Estermann
Vladimir Alexiev
Birk Weiberg
Daniel Mietchen
Klaus Illmayer
Vero Marino
Anju A Singh
Simon Villeneuve
Gregory Saumier-Finch
Gabriel De Santis-Caron
Raffaela Siniscalchi
Hirok Raja
  Notified participants of WikiProject Performing arts

@Beat Estermann, Strakhov, EncycloPetey, Infovarius: (as participants in different discussions related to this question (User_talk:Beat_Estermann#Characters and User_talk:Infovarius#mythical_characters) they could be also interested in this one)
From the perspective of modelling narrations: If characters are substantially different they should not be merged (e.g. the Mina Murray from The League of Extraordinary Gentlemen (Q95029) is a substantially different character from the Mina Murray in the original novel and would deserve an own item if someone would like to model the narration behind The League of Extraordinary Gentlemen (Q95029)). Also adaptations of mythical/folklore characters in modern fiction (comics, books) would deserve an own item if somebody would like to model the narration behind it (mostly they have different characteristics and different relationship to other characters and the items for the mythical character should not be cluttered with statements about the comic character). I'm not sure about items that are basically the same apart from a few differences (There are some thoughts about this issue at Wikidata:WikiProject_Narration#Characters_appearing_in_multiple_works_(adaptations,_spin-offs)_with_different_characteristics). One could solve this with a qualifier (e.g. present in work (P1441)), but that could become quite messy. In any case: characters that are direct adaptations of an original character should be linked to the original character using based on (P144) and said to be the same as (P460).
Characters are also important for modelling works in the performing arts and their productions. Two character adaptations may be substantially different roles: different voice types, different arias (characters are used to indicate which roles involve which arias or songs (see e.g. Der Vogelfänger bin ich ja (Q3023536))), different importance in the play... Of course also this could be modelled with qualifiers and retrieved with more complex queries. (I pinged the project Performing arts for these questions)
As some of the items are already around here some time: Merging items for different character adaptations/roles could be an issue for the stability of Wikidata-Ids (maybe there is already some project focussing on building an opera database using characters to indicate different roles and linking to Wikidata). - Valentina.Anitnelav (talk) 10:56, 27 January 2020 (UTC)
We have instances where this has gone to silliness, such as having Medea (Q174278) (daughter of King Aeëtes of Colchis in Greek mythology) separated from Medea (Q60607452) (theatrical character in Euripides' tragedy "Medea"). While a separation of known historical persons from fictionalized representations makes sense, a separation between a figure from mythology and the stories of those myths makes no sense. --EncycloPetey (talk) 17:22, 27 January 2020 (UTC)
Valentina and EncycloPetey, Konggaru Starry K. Erne Mogilevich Santer AldNonUcallinme? Thibaut120094 Shikeishu C933103 Sight Contamination -Zest Vulphere Sakretsu Jean-Frédéric Tris T7 TT me
Wallacegromit1 Jeanjung212 Bagas Chrisara ミラP CrystallineLeMonde
Nicereddy Shisma (talk) MatrosMonk Bwk24
  Notified participants of WikiProject Anime and Manga should also be mentioned, simply because Japanese fictional characters will also be affected by this section. --Liuxinyu970226 (talk) 01:05, 29 January 2020 (UTC)

"metamathematics theroem" (Q19673396)


I was looking for subjects concerning metamathematics and ran across this entry. Initially, I assumed "theroem" was a misspelling of "theorem", but these days in metamathematics I thought it might be real, not necessarily in the real number sense. So, I searched Google scholar and found "About 5,240 results (0.05 sec)" using the apparent misspelling "theroem" in article titles about various "theorems"; e.g., "Dilworth's Theroem and its Applications" about Dilworth's Theorem and "Generalization of Smirnov's theroem" about Smirnov's theorem. What to do? --Marshallsumter (talk) 18:30, 26 January 2020 (UTC)

  • Maybe I'm missing something, but what is the reason to believe this is anything other than a misspelling/typo? - Jmabel (talk) 02:32, 27 January 2020 (UTC)
  • Likewise, I can't see any indication that it's anything other than a misspelling. Ghouston (talk) 02:39, 27 January 2020 (UTC)
If you see a mathematician spell "if" as "iff", then it is most likely not a misspelling. "If the sun shines then I am happy" does not mean the same thing as "Iff the sun shines then I am happy". 62 etc (talk) 11:56, 27 January 2020 (UTC)
Right, but does that have any bearing on this? - Jmabel (talk) 17:04, 27 January 2020 (UTC)
On spelling in math, there is! 62 etc (talk) 13:01, 28 January 2020 (UTC)

Creating instructions for how to write Wikidata Tours

Hi all

I've started some instructions to try and make it more clear how to write Wikidata:Tours, I think I understand the basic steps but would really appreciate some help in writing the details. User:John Cummings/How to create Wikidata tours


John Cummings (talk) 11:55, 27 January 2020 (UTC)

Also where is the Javascript file that defines the structure of the tour kept? It would be really helpful to link to examples. I cannot find any obvious link in the wiki page or any of the exitsing snippets of documentation. @Pasleim:, @Quiddity:, @Multichill:, @Lea Lacroix (WMDE): any ideas? --John Cummings (talk) 12:45, 28 January 2020 (UTC)
An example for a javascript file you find under MediaWiki:Guidedtour-tour-wbstatements.js. It is the file used for the statement tour. --Pasleim (talk) 15:54, 28 January 2020 (UTC)
Thanks very much @Pasleim:, --John Cummings (talk) 22:42, 28 January 2020 (UTC)

Hijri dates

Is it possible to enter values of Hijri calendar to properties like inception (P571)? On many Islamic monuments the Hijri year is noted which cannot be converted exactly to (a single) Gregorian date, at least if day and month are missing. --RolandUnger (talk) 12:51, 27 January 2020 (UTC)

@RolandUnger: Unfortunately, that’s not supported yet. For now, maybe you could use stated as (P1932) as a qualifier to clarify some of the cases? --Lucas Werkmeister (WMDE) (talk) 13:12, 27 January 2020 (UTC)
The lack of Hijri dates is causing problems with people appearing to be over 120 years old or being born after they died. One value, usually birth, is in the Hijri calendar but calculated as if Julian or Gregorian, and the other value in the Julian calendar. I have been converting the dates to the Julian calendar to clear out all the errors. --RAN (talk) 13:57, 28 January 2020 (UTC)

When people request permissions for to be different kinds of admin what are the rules around votes

Hi all

I know someone who is requesting to be an interface admin, its not clear if only other interface admins can vote or if it is anyone. Is there a general rule for this or is it decided for each kind of admin? I can't find anything written down about this but maybe I'm looking in the wrong place?


--John Cummings (talk) 15:23, 28 January 2020 (UTC)

The rules differ for each kind of permission. Normally the rules are defined on the voting page. E.g. for interface admin, on Wikidata:Requests for permissions/Interface administrator it's written that registered users with a minimum of 100 local non-automated edits can vote. --Pasleim (talk) 15:58, 28 January 2020 (UTC)
@Pasleim: thanks very much for the explanation, sorry to miss the instruction. p.s I left a note on your talk page about the tours, let me know what you need. Thanks again, --John Cummings (talk) 17:09, 28 January 2020 (UTC)

Qualifiers for scheduled service destination (P521) when used for airports

Hi! Following this week's very inspiring example query from WD weekly summary, I've started to add P521 to some airports, beginning with my native Poland. As far as I can see, the only widely used qualifier so far has been operator (P137) to name the airline flying the route in question. I think some more qualifiers are needed and I want to ask for your advice on how to structure them (p -> q). What I need is to find the way to indicate that the connection presented using P521.

  • cargo flights
  • seasonal flights
  • charter flights.

Example: The route from Warsaw Chopin Airport (Q465071) to Verona Villafranca Airport (Q1433197) is operated by Wizz Air (Q335147) only in the summer season. How do we state this in Warsaw Airport's item? Thank you in advance for any input. Powerek38 (talk) 18:17, 28 January 2020 (UTC)

scheduled service destination (P521) offers applies to part (P518) as a pq - that might take an item value for e.g. "cargo flights" &c. Nothing is offered for time, but consider time period (P2348) with suitable items (e.g. the property gives an example of 2016-2017 one-year-period (Q30201458) but you can coin your own - "Last Wednesday of May to third Sunday of September", &c). If the P521 allowed qualifiers constraint needs a new pq adding (hint: it does) then add it with an explanatory note on the talk page. --Tagishsimon (talk) 23:18, 28 January 2020 (UTC)

Indigenous peoples of the United States

For Ethnicity="Indigenous peoples of the United States", I thought we had a field that contained "tribal affiliation", but I do not see it. I checked a few important historical Native Americans but most are not even marked Ethnicity="Indigenous peoples of the United States", which I am adding now. How do we handle tribal affiliations, should that be ethnicity? --RAN (talk) 23:53, 26 January 2020 (UTC)

indigenous peoples of the United States (Q49297) isn't a ethnic group (P172), but many different ethic groups that all happen to live in or partly in the United States. Likewise for ethnic groups in Europe (Q1464982), ethnic groups in the Philippines (Q2002622), Ethnic groups in Latin America (Q5404241), ethnic groups in Russia (Q2625447) etc. I don't think any of these should be used with P172. --Dipsacus fullonum (talk) 01:39, 27 January 2020 (UTC)
This is an important best practice to define for Wikidata:WikiProject Indigenous peoples of North America, but unfortunately no work has been done in this area (on-Wiki, anyway) so far. - PKM (talk) 01:42, 27 January 2020 (UTC)
I realize that en-wiki does not make rules for Wikidata, but you still might find it useful to read en:Wikipedia:WikiProject Ethnic groups: we gave this a lot of thought some 15+ years ago. - Jmabel (talk) 02:34, 27 January 2020 (UTC)
indigenous peoples of the United States (Q49297) is an ethnic group, just a very broad one. Obviously the smallest unit is preferred. There just has been no consistency so far in how we have been handling entries. --RAN (talk) 05:39, 27 January 2020 (UTC)
I read the article en:Native Americans in the United States and I couldn't find any mention about whether it is a single ethnic group (defined in en:Ethnic group as a category of people who identify with each other) or not. Anyway I agree that the smallest unit is preferred. --Dipsacus fullonum (talk) 15:59, 27 January 2020 (UTC)
PS. If a reliable source says that someone "identifies themself as native American", then that can be stated with P172, but not in my opinion if the source says that someone "is a native American", as that can mean that they belong to an unspecified native ethnic group in USA. --Dipsacus fullonum (talk) 16:18, 27 January 2020 (UTC)
Yes, people sometimes say "native American" to mean born in the modern USA instead of "indigenous peoples of North America", so we have to be careful. The way I am handling it is to put in the smallest unit we have for Ethnicity=, which is their tribal association, then I make sure that the tribe eventually is a subclass of "Indigenous peoples of the United States", some currently are and some are not. Something like: "Indigenous peoples of the United States" > Nation of tribes > tribe > Moses Yellow Horse. --RAN (talk) 19:08, 27 January 2020 (UTC)
In any case, I would support a new property “tribal affiliation” which defers the question of whether a tribe is an ethnic group or a government (I think we need three structures, people - government/organization - reserve/reservation) but even if we agree to that it will be a long time before we have three items for every tribe. - PKM (talk) 00:00, 29 January 2020 (UTC)
Much as with nation states, tribe as identity and tribe as government/citizenship/enrollment are two separate (albeit correlated) issues. To further complicate the matter (and, again, there are plenty of analogues elsewhere) some such as Duwamish (Q491721) have a strong identity but are not generally recognized on an institutional/governmental level; conversely, there is Mandan, Hidatsa, and Arikara Nation (Q3285430) as a governmental entity, but I imagine no one identifies as a "Mandan, Hidatsa, and Arikara," they identify as one of the three. - Jmabel (talk) 01:16, 29 January 2020 (UTC)
@Richard Arthur Norton (1958- ): I know the item currently says that, but I think it's incorrect. Indigenous "person" (or "people" in plural) of the United States might be an ethnic group, but indigenous peoples (or "people" in the singular) of the United States would be either a class of ethnic groups or a group of ethnic groups (both of which are different from a larger ethnic group that includes other ethnic groups). The handling of these things is kind of a mess. It's also unclear in general whether ethnic groups are supposed to be structured as classes or groups... --Yair rand (talk) 05:54, 29 January 2020 (UTC)

Need help for a merge

Hi, I tried merging Q8599880 ("Category:Liverpool and Manchester Railway locomotives") into the older item Q6576955 (list of Liverpool and Manchester Railway locomotives), but I failed spectacularily. The Merge two Items page told me I need to solve conflicting issues in labels and descriptions first, which turned out to be maybe a hundred individual conflicts (added a year ago in bulk by a bot), which I then wasn't allowed to remove because "as an anti-abuse measure, you are limited from performing this action too many times in a short space of time". So I tried a few times and then noticed that each attempt removed only a single (!) conflict, and after some time didn't remove anything at all anymore (see history of page Q8599880). So I'm giving up now and hope somebody with more knowledge or more rights will finish this merge. I hope I did not cause any inconvenience. --Alias Sobriquet (talk) 19:28, 27 January 2020 (UTC)

They should not be merged. One points to a category. The other to a list. A list does not equal a category - e.g. this: does not equal this: . I've reverted your edits. (No harm done). --Tagishsimon (talk) 19:33, 27 January 2020 (UTC)
But the Commons category contains the pertinent pictures for the Wikipedia article. The Wikipedia article item (Q6576955) has a slot for commons items (under "Other sites") where usually the relevant commons category is added. Commons is full of categories linked in such a way to Wikipedia pages via Wikidata. Are all those combinatons faulty, or am I misunderstanding some facet of this? --Alias Sobriquet (talk) 19:46, 27 January 2020 (UTC)
You'd have to point to some instances where a commons cat points back to a wikidata item for a list. That would seem to me to be a wrong thing. By way of a comparator, list of compositions by Johann Sebastian Bach (Q375345), Category:Compositions by Johann Sebastian Bach (Q7236116) and ... sitelinks should point to the item's subject - a list, a category. The list item is linked to the commons category via Commons category (P373). At the very least, look at the instance of (P31) values for your two items; clearly different things. --Tagishsimon (talk) 20:01, 27 January 2020 (UTC)
The separate item is necessary here as there's also a Wikipedia category, but sometimes a Commons category can be added to the same item as an article - is it different if the article is a list? There are also properties list related to category (P1753) and category related to list (P1754) for links between categories and lists. Peter James (talk) 17:55, 29 January 2020 (UTC)

Language add request

Please open Yumiko Ashikawa (Q4023235) and add Venetian (Q32724) on the language section. Thanks! -- 20:11, 28 January 2020 (UTC)

What is your source? ChristianKl❫ 12:36, 29 January 2020 (UTC)

Wikidata Glossary entry for Property is really technical

Hi all

I'm working on the Wikidata glossary and the description of a property is extremely technical and not really understandable for someone without a computer science background, currently it reads

Property (also attribute) is the descriptor for a data value, or some other relation or composite or possibly missing value, but not the data value or values themselves. Each statement at an item page links to a property, and assigns the property one or several values, or some other relation or composite or possibly missing value. The property is stored on a page in the Property namespace, and includes a declaration of the datatype for the property values. Compared to linked data, the property represents a triplet's predicate.

The help page on Wikidata for properties has a much more understandable :

A property describes the data value of a statement and can be thought of as a category of data, for example "color" for the data value "blue". Properties, when paired with values, form a statement in Wikidata. Properties are also used in qualifiers. Properties have their own pages on Wikidata and are connected to items, resulting in a linked data structure.

How would people feel about replacing the more technical glossary description with the simpler help page description and then finding a home for the current glossary description on the help page?

There are are other glossary terms which are similarly technical.


--John Cummings (talk) 12:36, 29 January 2020 (UTC)

I think the more understandable statement is an improvement. You might add into the statement that properties can also be used for references. ChristianKl❫ 15:54, 29 January 2020 (UTC)

"academic journal article" versus "scholarly article"

Hi, all,

What's the difference between scholarly article and academic journal article? The latter says that it is different from "scholarly article", but in English has "scholarly article" as an alias. I haven't found usage instructions.

To make it worse, the discussion appears to be more about "scientific article" than "academic journal article". Pinging Mtrognitz, DarTar, Antrocent.

Sorry if I'm being dense; I'm inexperienced here. I just want to create a couple of references for items, but the learning curve is steep.

Which to use? Do we care about differentiating between articles in scholarly journals, chapters in scholarly monographs, and some other kind of "scholarly article"?

Pelagic (talk) 10:21, 28 January 2020 (UTC)

  • It might be worth looking at three questions separately:
  1. Are there 2 (or more) concepts that can be described separately?
  2. If so, are the labels/descriptions adequate and consistent across languages?
  3. If so, which item to use in instance of (P31) articles in periodicals? Even if there are several possible ones, it might not be efficient to actually use them with P31.
The answer to #3 is that currently >99% use Q13442814 even for articles or journal sections that aren't necessarily scholarly or scientific. --- Jura 12:17, 28 January 2020 (UTC)
1) Apparently yes, since there are Wikipedia articles about "scientific article" and not all academic journals are scientific 2) They don't seem very consistent, and the subclassing is questionable. 3) That was just the choice of the operators of the bots that created the items. Perhaps they'd have been better off using Q18918145, since it's valid regardless of whether the journal is "scientific" or not. You also find articles in scientific journals that aren't actually scientific articles, such as news and obituaries. Ghouston (talk) 23:40, 28 January 2020 (UTC)
I've reset the English label on scholarly article (Q13442814) back to "scientific article", and made it a subclass of academic journal article (Q18918145), instead of the reverse. Ghouston (talk) 00:16, 31 January 2020 (UTC)

Is there a version of {{P|}} and {{Q|}} that shows both the name and the description?

Hi all

I'm writing a lot of the missing Wikidata:Tours and one thing that would be really helpful is to have a way of showing both the name of a Property and its description using a template like {{P|}} and {{Q|}}. Including the description in the instructions makes it easier for people to know which of the properties and items to chose from from similarly named one without having to worry about Q numbers (I'm writing a tour about the Q numbers also). It would also make it much easier for the translation of the documentation because it would mean the translator doesn't have to translate a plain text description and would also stop instructions becoming incorrect if a description changed.

If this functionality does not yet exist how difficult would it be for someone to add this functionality to the the existing templates and who could I ask?

Thanks very much

--John Cummings (talk) 11:28, 28 January 2020 (UTC)

@John Cummings: Template:Q+? --Tagishsimon (talk) 12:10, 28 January 2020 (UTC)
You might also be interested in Template:Autodescription. --Tagishsimon (talk) 12:11, 28 January 2020 (UTC)
Thanks very much @Tagishsimon:, very helpful, if only Q+ and P+ allowed show_p=no to hide the Q number like Q and P do, that would perfectly replicate what the interface shows. @JakobVoss: could you pretty please with sugar on top add this functionality? John Cummings (talk) 12:24, 28 January 2020 (UTC)
@John Cummings: I've made {{LD}} for this, adding documentation shortly. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:29, 28 January 2020 (UTC)
@Pigsonthewing: bless you, two questions, will the label automatically change to the users language if its used on a translated page? Also is there a way to make it link to the property page instead of being plain text (both are useful) --John Cummings (talk) 14:33, 28 January 2020 (UTC)
@John Cummings: It does now; and it soon will. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:42, 28 January 2020 (UTC)
@Pigsonthewing: 👍 👍 👍 --John Cummings (talk) 14:56, 28 January 2020 (UTC)
Done. See also {{Entity description templates}} for all related templates (that I could find!). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:00, 28 January 2020 (UTC)
Regarding the translations, they automatically change to the user's language regardless of the language of the page. It actually looks a bit odd, e.g., to be reading a page in English like this one, and all of the labels are in some other language. Ghouston (talk) 00:01, 29 January 2020 (UTC)
Thanks very much @Pigsonthewing: and @Ghouston:. --John Cummings (talk) 12:08, 30 January 2020 (UTC)
As the template's documentation notes, the language can be specified. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:01, 30 January 2020 (UTC)

Tower Block Archive - Aberdeen as first trial run

Hi, the Edinburgh College of Art has the Tower Block Archive, a "fully searchable database of around 4,000 images of every multi-storey social housing development built in the UK in the postwar decades. The photographs were largely taken in the 1980s by Miles Glendinning." Kat Breen, a database assistant, has been seconded to the project to support the images being hosted to Wikimedia Commons once we have stable DOIs for the images being hosted in the University of Edinburgh's Datashare repository. We are looking at 63 images of housing developments in Aberdeen as our first upload but would like these housing developments to also link to Wikidata items for these housing developments. Two questions have arisen in terms of mapping these housing developments to Wikidata.

  1. We would like to geolocate the developments on a map but some no longer exist or may be split in different locations. As an approximation of a geolocation, could we use the first part of the postcode (e.g. AB4) and use the lat & long of the centre point of that postcode where we don't have the more specific full postcode (and use its centre point)? Boundary information would be a huge and complicated task esp for demolished blocks.
  2. We have the original commissioning authority commissioned by (P88), original location authority authority (P797)and present day location authority authority (P797) detailed in our metadata so we are looking at how best to model this on Wikidata. Could we use authority (P797) for original location authority AND present day location authority as two different values with date qualifiers (where known)?

All being well if the Aberdeen import goes well we can look to roll out the pics and link to data items for whole of Scotland's housing developments (England could be done if we could keep Kat on). Any thoughts then let me know, Stinglehammer (talk) 16:45, 28 January 2020 (UTC)

For your questions:
  1. Yes I don't see why not, you'd just ideally want to reflect the inaccuracy somehow - most likely with a determination method (P459) qualifier (may need to create an appropriate item for the value). Would suggest also adding a corresponding postal code (P281) statement.
  2. Yes, in fact, I would say you definitely should capture such historical data. See start time (P580) and end time (P582).
Hope this helps, just my personal response, someone more knowledgeable may have better suggestions. --SilentSpike (talk) 17:35, 28 January 2020 (UTC)
I should also say, re: coordinates, that I'll be amongst those volunteering to find accurate ones. Depending on how you count these things, there are only ~650 geolocatable items in Scotland that do not have coordinates 1. The addition of even 4k new items some or all of which have coordinate issues would just be something we work through until it's done. Didn't want you going away with the idea that I'm being stampy footed about you having to put in all the effort. --Tagishsimon (talk) 19:14, 28 January 2020 (UTC)
Many thanks all! Will be working with Kat tomorrow afternoon so might be back to iron out any more kinks as they arise then but this is all super helpful. Thanks again, Stinglehammer (talk) 15:03, 29 January 2020 (UTC)
It looks like some have Canmore entries, which will have coordinates, and some have full postcodes which are still current (presumably a tower block gets all, or almost all, of a postcode to itself?), so that should cut down the number to locate. Even with postcodes that have been retired, the NRS census indexes may be able to track which postcodes they were merged into. A "determination method" qualifier on the coords could be used to flag ones that might need confirmation, as SilentSpike said. A "location" statement for what neighbourhood or part of town the blocks were in can also be useful, for aggregation, and eg for putting the image categories on Commons into appropriate one level up categories for location. Jheald (talk) 10:45, 30 January 2020 (UTC)
Sounds like a good project, and a great media donation. If you're going to be creating Wikidata items about extant tower blocks, it would be worth getting the QIDs added to the corresponding objects in OpenStreetMap. User:Edward can probably apply his tools with that.
It might also be worth considering the creation of an external ID property, though I note that some of the entries in the archive are for sets of tower blocks (examples: [20], [21]). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:17, 30 January 2020 (UTC)

step by step

hello there, I'm a new member here, could you tell me step by step guide for a beginner like me?

--PutriAmalia1991 (talk) 12:01, 30 January 2020 (UTC)

@PutriAmalia1991: Check out Wikidata:Tours. Vahurzpu (talk) 14:31, 30 January 2020 (UTC)

Bots work requests

Do we have a page to request work to be done by a bot? --Juandev (talk) 14:45, 30 January 2020 (UTC)

@Juandev: Wikidata:Bot requests (not to be confused with Wikidata:Requests for permissions/Bot, where bot operators request account permissions). --Lucas Werkmeister (talk) 15:03, 30 January 2020 (UTC)

Order of properties

How come properties are sorted in a specific order in while external identifiers are sorted after the date they were added? --Trade (talk) 21:15, 30 January 2020 (UTC)

Because nobody has assigned a sort order to them. You may be interested in Wikidata:Requests_for_comment/Sort_identifiers. Ghouston (talk) 23:35, 30 January 2020 (UTC)


I stumbled across Q83752609 when doing some queries on countries today. I have never heard of such a country, no Wikipedia articles exists and enwp article with that name has been deleted once for being a ”blatant hoax”.

Is it notable enough for Wikidata and, if so, how should it be represented?

Related items: Q83817395 and Q83753823 / Autom (talk) 21:07, 29 January 2020 (UTC)

Notifying the author @DominusVilicus:. /Autom (talk) 21:17, 29 January 2020 (UTC)
The only sources are the supposed state's own websites (one of which is defunct) and a book which is not found by Google, Goodreads, Trove or Worldcat/OCLC. Also deleted as a hoax on Wikinews: n:A new country formed in northern Australia. It's really poor to waste volunteers' time like this. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:55, 30 January 2020 (UTC)
Some notes: (1) Here is the book in question | (2) The wiki news was a draft post with no content, not sure why it was posted | (3) The other website is no longer defunct DominusVilicus (talk) 01:22, 1 February 2020 (UTC)
This project seems notable enough to me, but perhaps it should have another type of class instance. SoaringMoon (talk) 01:24, 1 February 2020 (UTC)

These and more are now at Wikidata:Requests for deletions#Promatia-related items. Images on Commons are also nominated for Deletion. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:51, 30 January 2020 (UTC)

Wikidata:Glossary entry for a 'policy'

I think it is probably important to have definitions in the Wikidata:Glossary not only for technical parts of Wikidata's structure but also more social parts, especially if people are coming from outside Wikimedia to Wikidata instead of e.g coming from Wikipedia.

What would be a definition for a Wikidata policy? I can find lists of policies but not a definition for what a policy is. Any suggestions on where to find something already written?

Also what other definitions are missing from the glossary? Especially around social aspects of Wikidata.


John Cummings (talk) 11:46, 30 January 2020 (UTC)

There should be something about QuickStatements aka QS. --Trade (talk) 16:16, 30 January 2020 (UTC)
@Trade:, QS added :) --John Cummings (talk) 17:51, 30 January 2020 (UTC)
Maybe something like : A policy is a rule which is developed by the Wikidata community to describe best practices, clarify principles, resolve conflicts, and otherwise further our goal of creating a free and open knowledge base? @Pigsonthewing: what do you think?
--John Cummings (talk) 21:00, 31 January 2020 (UTC)
No "rules", but (drawing on en.Wikipedia) how about "Policies are pages that serve to document the good practices accepted in the Wikidata community. They are standards all users should normally follow, but there is no need to read any policy page to start contributing, and they should be applied using reason and common sense."? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:11, 31 January 2020 (UTC)

Three items need merging

I have discovered that Pulau Breueh (Q2659873), Pulau Breueh (Q24821857) and Pulau Breueh (Q48989768) all refer to the same island. Since I have already failed at merging two items, I know I will fail at merging three. Could somebody take care of this? Abductive (talk) 02:33, 31 January 2020 (UTC)

I merged them. They are listed on Wikidata:Do not merge/12 for some reason, but I can't see any indication that there's more than one entity, an island. Ghouston (talk) 03:55, 31 January 2020 (UTC)


Good afternoon. Is it possible to cancel the entire contribution of this participant in the Tatar language. He adds words that he himself invents. And these words are not in the Tatar language.--Ilnur efende (talk) 14:50, 18 January 2020 (UTC)

@Ilnur efende, Ymblanter: I don't get it. Why isn't there any message on User talk:Maitsavend asking the user what they're doing? I see three admin noticeboard topics and two blocks, but how is this user supposed to know that they're doing something wrong? If you don't tell people they're doing something wrong, how are they supposed to know? Please assume good faith and start with a clear explanation at User talk:Maitsavend to the user what they're doing wrong and how they should correct this. In case that fails, you can go to Wikidata:Administrators' noticeboard. Multichill (talk) 17:28, 18 January 2020 (UTC)
My block wat a year ago, I do not remember anything, but I see that I previously posted at their talk page, without any result.--Ymblanter (talk) 17:51, 18 January 2020 (UTC)
@Ymblanter: Am I missing something here? I see some interaction about blank items, this edit about a block for vandalism and another block for a month. I don't see any mention of what the user actually did wrong. I just see the broad term "vandalism". How was the user supposed to know this is about incorrect labels/descriptions? Multichill (talk) 17:57, 18 January 2020 (UTC)
They actually got a warning [22], and the editing history of one of the pages [23] (there were issues in three items simultaneously) is telling.--Ymblanter (talk) 18:09, 18 January 2020 (UTC)
Dear user. This user is a participant in the Tatar Wikipedia, as User:Marat-avgust, he does the same thing there, with the help of his bot, many participants have already written to him, but he simply ignores us, and continues to replace terms with those invented by him, or from other languages that are not clear to Tatar-speakers.--Ilnur efende (talk) 17:37, 18 January 2020 (UTC)
@Ilnur efende: sure, but Wikidata is not the Tatar Wikipedia. You have to start first to try to sort it out with a user directly. If that doesn't work, someone else can step in. For both user accounts (User:Maitsavend and User:Marat-avgust) I don't see any interaction about this on their talk pages. So please leave the user a message, that can be done in any language you believe the user understands. Multichill (talk) 17:57, 18 January 2020 (UTC)
  • @Ymblanter, Multichill: there seems to be a difference of editorial views among two ttwiki admins about spellings. The name in the sitelink can, but neededn't be the label at Wikidata. It's unclear why a note about that should be a warning (that's what [24] seems to say).
The edit by @Derslek: reflects their view on ttwiki about the spelling. The repeated edit by Derslek at Wikidata is technically incorrect as it adds the spelling already used as alias as label, but doesn't re-add the second spelling as alias (ttwiki uses the second spelling as a redirect).
Maybe I'm missing something, but calling edits by @Maitsavend: "vandalism" seems inappropriate. While it can happen by otherwise unproductive users, I'd expect a Wikidata admins to know better and present excuses to the contributor(s). --- Jura 08:13, 20 January 2020 (UTC)
I do not see anything resembling different views. For me, this is a deliberately introduced typo. Tatar is well documented, I do not see how there could be any differences in spelling of this pretty large Russian city. If not, they should have engaged with the conversation when offered a chance.--Ymblanter (talk) 08:19, 20 January 2020 (UTC)
Agree about that last point, but I don't think Derslek's approach was correct. I don't read the language, but I don't see why Wikidata should second guess ttwiki admins. If they can be referenced, both spellings should be added to Wikidata. If it's a common typo, probably too. --- Jura 08:40, 20 January 2020 (UTC)
    • This is not an ordinary typo, but this user intentionally adds his own words, he changes the literary words to his own, and in Wikipedia he does it with the help of a bot, being one of the administrators. Does not respond to comments, does not respond. They haven't been able to contact him for months, or get a response to his actions.--Ilnur efende (talk) 19:49, 23 January 2020 (UTC)
      • Jura. Maitsavend is just inventing new words and names, that never been in Tatar. It's typical WP:OR. --Derslek (talk) 20:00, 27 January 2020 (UTC)

Jura, Ymblanter Users latest actions: [25], [26], and so on. Is it possible to ban him at all? I suppose that it will be impossible soon to correct all the contributions of this hard-working, but unsocial user. --Derslek (talk) 18:46, 27 January 2020 (UTC)

Now for one month, the next block will be for an indefinite duration.--Ymblanter (talk) 20:49, 16 March 2019 (UTC)

Jura, Ymblanter, is it possible to cancel all user actions? Because it is not possible to fix all this manually. And on other Wikimedia projects, words invented by them are displayed in the Tatar language?--Ilnur efende (talk) 20:36, 27 January 2020 (UTC)
I blocked the user indef. They have zero edits to their talk page, and the edits are clearly problematic. Whoever wahts to unblock them please ask them what they mean by this edit for instance - why did they replace the label which coincided with the name of the Tatar Wikipedia article with a combination which looks like a deliberate distortion of the name.--Ymblanter (talk) 20:55, 27 January 2020 (UTC)

Ymblanter, User:Maitsavend bypasses blocking through other accounts (User:IanraBot, User:Marat-avgust).--Ilnur efende (talk) 18:34, 1 February 2020 (UTC)

Movie extras in P161 values

An unregistered user has added a fairly large number of cast member (P161) uses for an actor who was uncredited or an extra in most of the relevant films (see IMDb). Is this an appropriate use of the property? Jc86035 (talk) 08:28, 26 January 2020 (UTC)

  • large number like 12?
Hopefully, they will add more. Uncredited appearances should have some sort of a reference. --- Jura 10:46, 26 January 2020 (UTC)
Would probably be worth adding some kind of character role (P453):extra (Q658371) to these to clarify they're an uncredited extra, but if it's sourced it seems okay if a little strange. Andrew Gray (talk) 11:54, 26 January 2020 (UTC)
There are a few cases where a notable actor (or someone later notable for something completely different) started out as an extra. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:41, 27 January 2020 (UTC)
Are they reliably cited? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:41, 27 January 2020 (UTC)
@Pigsonthewing: No, they don't seem to be, and I don't think there will be RSes for most of these. This search, for example, returns no reliable sources, and a User:Slimkhezri actually edited the Fandom page that's linked as the first search result. I suspect the linked IP edits may have been self-promotion. Jc86035 (talk) 09:14, 1 February 2020 (UTC)

Adding examples to property proposals

Hi all

I've started two property proposals for external identifiers and I'm struggling to know how to create examples because there's not really any instructions, could someone help me and also explain what is needed and where to put the instructions?


--John Cummings (talk) 18:12, 27 January 2020 (UTC)

@John Cummings: Give us, at the minimum, 1) a URL to the catalog you want to create a property for (that goes in the 'source' field), and 2) values for 'example 1', 'example 2', and 'example 3', mapping items to entries in the catalog, in the form {{Q|itemID}} → [http://URL_to_entry_in_catalog entry_ID_in_catalog], where "itemID", "URL_to_entry_in_catalog", and "entry_ID_in_catalog" are filled in as appropriate. Mahir256 (talk) 18:36, 27 January 2020 (UTC)
Thanks @Mahir256:, I don't understand what you mean. Ideally the instructions should include a little proforma you fill out just replacing the gaps with the information needed. I'd be happy to write it if I understood what I was supposed to do.... --John Cummings (talk) 12:26, 28 January 2020 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────@John Cummings: Here's what an example should look like:

|example 1 = {{Q|6939}} → [ 1338]

  • An existing Wikidata item
  • the little right arrow →
  • a link to the item or page in your proposed data source that would be matched to that Wikidata item using the new property

This displays as:

I'd be happy to help with your property proposals. Looking now! (It took me forever to figure it out.) - PKM (talk) 20:32, 28 January 2020 (UTC)

Thanks very much @PKM:, I'm also trying to think of a generic version to put in a template, for authority control it could be something like this? {{Q|EXISTING_Q_NUMBER}} → [URL TITLE] --John Cummings (talk) 22:04, 28 January 2020 (UTC)
@John Cummings: I asked some questions on your talk page (and will have more after those). Let’s chat there. - PKM (talk) 00:03, 29 January 2020 (UTC)

@PKM: Do we even have a guide on how to create property proposals? --Trade (talk) 17:10, 1 February 2020 (UTC)

@Trade: I don’t think so. It took me over a year to get comfortable with it, and I still mess up “domain” on the first pass about half the time. - PKM (talk) 20:56, 1 February 2020 (UTC)

Wrong identifier values deprecated and kept

Hi all, I noticed some people deprecate but keep wrong identifier values instead of removing them. Example is in Martin Beck (Q6774971) where several identifier values are deprecated with reason for deprecation (P2241) : applies to other person (Q35773207). Is there a good reason for this? In effect, all items could have all existing identifier values (but one) set as deprecated with this reasoning. I would be in favor of removing such values on the spot except for maybe a few non-obvious cases. What does the community think? PS: Last but not least, it leads to problems in Constraint violation tables such as this (deprecated values are still understood by the bot to break the unique value violation). Vojtěch Dostál (talk) 08:00, 30 January 2020 (UTC)

  • .. and then people re-add them, and then you remove them again .. Sounds like Wikipedia. What's the point?
It's a recent bug, if the bot doesn't understand the deprecated values. --- Jura 08:03, 30 January 2020 (UTC)
  • Yes. To be clear, Vojtěch Dostál, the points of retaining deprecated IDs are 1) to signify that some sources, at some times, have used that ID in respect of the subject (which is to say, to convey useful information to our users) 2) to indicate the erroneous nature of the ID and 3) to prevent, or at least deter them from being re-added. It is somewhat the whole point of having a Deprecated rank, that we get to apply it in this useful fashion. Deleting deprecated information really misses the point in a fairly fundamental fashion :( --Tagishsimon (talk) 09:22, 30 January 2020 (UTC)
    • "to signify that some sources, at some times, have used that ID in respect of the subject" is that always the case, or is the error sometimes ours? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:22, 31 January 2020 (UTC)
      • Dunno Andy. Sounds unlikely :) --Tagishsimon (talk) 16:13, 31 January 2020 (UTC)
  • The knowledge that a certain identifer value is wrong is valuable and often worth storing. The constraint vioaltion table shouldn't behave that way. It shouldn't list cases where there's exactly one statement of the highest rank. ChristianKl❫ 13:10, 30 January 2020 (UTC)

Thank you all of you. Tagging bot operator @Ivan_A._Krestinin: who is the bot's maintainer :) Ivan, what do you think? Vojtěch Dostál (talk) 15:00, 30 January 2020 (UTC)

  • It looks like this one will keep appearing. Some errors get corrected at VIAF and we can delete them once fixed, but other conflations of similar people just keep getting mixed up again and again, so worth keeping them. It can also be good to add in the other person to Wikidata that is causing the conflation and use different_from=(add in second entry) to both entries. It will help the next person understand how the conflation occurred and prevent them from being merged. --RAN (talk) 05:56, 1 February 2020 (UTC)

part of (P361) vs. member of (P463)

I have a dilemma. When should we use part of (P361) and/or member of (P463)? Both property have a lot in common, but are also quite different. Example Terry Jones (Q166159) can be part of (P361)Monty Python (Q16402) and/or member of (P463)Monty Python (Q16402). part of (P361): "object of which the subject is a part", Terry is a part of Monty Python (Q16402). member of (P463): "organization, musical group, or club to which the subject belongs", Terry is a member of that comedy group. By the looks of it, there is more people struggling with the same problem. - Premeditated (talk) 18:43, 30 January 2020 (UTC)

Generally, "member of" would be used for human groups, being more specific than "part of". But some people prefer "part of" because it has in inverse property "has part": you need the inverse if you want to list the members in a Wikipedia template. Ghouston (talk) 23:56, 30 January 2020 (UTC)
member of (P463) is a subproperty of affiliation (P1416) which in turn is a subproperty of part of (P361). If you use member of (P463) and someone asks the query service for all part of (P361), the result should still show up. If someone is a member then the idea is to use the more specific member of (P463). ChristianKl❫ 19:41, 1 February 2020 (UTC)
I have some doubts that it works that way by default. I tried it on Commonwealth of Nations (Q7785), it didn't find the members:
SELECT ?x ?xLabel WHERE {
 ?x wdt:P361 wd:Q7785
 SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en" } }

Try it!

Ghouston (talk) 23:23, 1 February 2020 (UTC)

cuisine (P2012)

Hello. Would you mind if I widen this to chef (Q3499072)? We could have something like Kelly Rangama (Q83871997)Réunionese cuisine (Q3006981). Thierry Caro (talk) 15:52, 28 January 2020 (UTC)

Q185441 and Q28806256

I have noticed, that these two appear to be about the same topic. Does anyone with more experience see a reason not to merge the two? If not, I would kindly ask to do it, as I do not feel confident enough to do so myself. AntiCedros (talk) 23:37, 31 January 2020 (UTC)

No. One is an item for a member state. The other is an item for a list of member states. A list of member states != a member state. One is a list. The other is a country. Welcome to wikidata :) --Tagishsimon (talk) 23:43, 31 January 2020 (UTC)
They shouldn't be merged but there are sitelinks that could be moved - at least one (ca:Estats membres de la Unió Europea) in Q185441 looks like (and is categorised as) a list. Peter James (talk) 23:52, 31 January 2020 (UTC)
Seems a bit weird though, as no language but ia (where both appear to be stubs) appears to have both articles. AntiCedros (talk) 00:10, 1 February 2020 (UTC)
Every article has a Wikidata item, but there are much more WD items than articles. --SCIdude (talk) 08:29, 1 February 2020 (UTC)
Every Wikipedia article should have a WD item, not all of them do. Not being an expert on the subject but it's not 100% automatic and can appear a little random. Quite a few articles are currently without QItems, across the different wiki-languages, I believe a list exists somewhere. Some of them have special characters in the url and that might be a reason, but that's certainly not the case for all of them. Moebeus (talk) 03:54, 2 February 2020 (UTC)
Very far from 100% automatic. Linking articles to items (obviously) happens after the creation of the article, with varying amounts of lag. However has 6M articles; Wikidata has sitelinks for all but 39,461 at the most recent count, earlier today - see ... so that's 0.66% of articles with no item. There's no special reason (e.g. special characters in the article title) that precludes a sitelink from an item; it's just bots or people have not got around to finding and dealing with the need for the sitelink. Every so often one of our users, for better or worse, links all unlinked articles to new items - see - and so far as I'm aware, does this for many language versions. --Tagishsimon (talk) 06:12, 2 February 2020 (UTC)
I don't know how other Wikipedias handle this, but on dawiki new articles will not be linked to new items by bots before they are at least 3 weeks old. In the meantime they are listed by a bot on a list of new articles without a Wikidata item, giving time to human users to either link them to existing items or create new items with appropiate statements. --Dipsacus fullonum (talk) 02:25, 3 February 2020 (UTC)

Reasonably quick way to resolve "non-compliance property" issue?

Yesterday, over on the talk page for P5009 ("complies with (P5009)"), I asked if there was a property to use when an item does not comply with another item. Additionally:

The Bechdel test (Q4165246) was used as an example when P5009 was first proposed, but it appears as though we still need Q45172088 to provide a counterpart to Q45171911. More generally, it would be nice to be able to use Wikidata to express the compliance table here: w:Comparison of electoral systems#Comparisons, and to more-easily collaborate on all of the citations associated with that table. I was able to successfully mark the relationship between Copeland's method (Q5168347) and monotonicity criterion (Q6902035) (because Copeland complies with Monotonicity), but (for example) it doesn't yet seem possible to mark the relationship between instant-runoff voting (Q1491219) and monotonicity criterion (Q6902035) (which doesn't comply).

What's the fastest way to resolve something like this? Should I propose a new property sooner rather than later, should I just wait for an answer over on the talk page for P5009, or should I start getting in the habit of raising issues like this over on this project chat page? -- RobLa (talk) 17:59, 30 January 2020 (UTC)

It's not unusual to be ignored on property talk pages: most likely only a few people are watching them. Asking here is good. I see that has quality (P1552) is being used, e.g., on Hackers (Q13908) with the special item that you mentioned Q45172088. The advantage of a dedicated property like complies with (P5009) is that you don't need to create such special items, you can use the property with Bechdel test (Q4165246) directly. Ghouston (talk) 23:52, 30 January 2020 (UTC)
Thanks for the response, Ghouston! Yes, I agree that an advantage of a property would be the ability to combine it with an item to make a statement about an item. Judging from your response, the property that I'm seeking doesn't seem to exist, but I hope you agree that it should. I guess I should learn about writing a property proposal for the "fails compliance with" property. -- RobLa (talk) 01:32, 31 January 2020 (UTC)
Proposal made: Wikidata:Property proposal/fails compliance with. -- RobLa (talk) 03:08, 31 January 2020 (UTC)
  • Wouldn't this work better with something like test taken (P5021) and test score (P5022). I know it's not a test, but the structure seems to work better: a main value describing the aspect being checked and qualifier of the result (y/n/don't know/partially/sometimes etc). --- Jura 06:03, 31 January 2020 (UTC)
  • Such as test taken (P5021) monotonicity criterion (Q6902035) / test score (P5022) "0 percent"? It doesn't seem very appealing to me. Ghouston (talk) 06:31, 31 January 2020 (UTC)
  • I agree with Ghouston; I think P5021 and P5022 are a poor fit for criteria compliance. A "test" as used by P5021/5022 is a one-time evaluation, administered to a human. The implication is there is a date the test administered/taken and score received. The result of a "test" is not a mathematically-provable thing. Granted, the "Bechdel test" isn't designed for mathematical rigor, either, but election methods and the criteria they are frequently the subject of mathematical scrutiny. -- RobLa (talk) 06:42, 31 January 2020 (UTC)
  • With "like", I didn't mean you should use that property. It would rather look like "aspect (P9021) monotonicity criterion (Q6902035) / has quality (P1552) pass (Q123456) ". This way one could also state that it's unknown or not one of only two options. Even the table mentioned above has more different values than "compliant" and "not compliant". --- Jura 07:36, 31 January 2020 (UTC)
  • Hmmm, ok, that's a more interesting idea. I don't know how to design such a thing, and it seems as though we have many true pass/fail criteria that don't benefit from the complexity. What do others think? -- RobLa (talk) 16:43, 31 January 2020 (UTC)
  • I think that this can best be achieved by adding qualifiers to a new "doesn't comply" property, e.g., has cause (P828) or something to indicate the degree of non-compliance. The advantage is that you still have the binary complies/doesn't comply statements that can be interpreted without needing to interpret arbitrary qualifiers. Ghouston (talk) 22:33, 31 January 2020 (UTC)
  • I don't think the qualifier should be arbitrary. The advantage of the approach is that if qualifiers aren't read, no context is lost and we don't have to read two different properties to get the entire info. --- Jura 15:02, 3 February 2020 (UTC)

Not a quick solution, but a principled one would be to change the data model so as to allow statements which negate any existing property. [27]. Ls1g (talk) 04:17, 1 February 2020 (UTC)
Would that require a patch to Wikibase? That may be a little too principled for me.  :-D -- RobLa (talk) 04:51, 1 February 2020 (UTC)
Yes it would, and understandably that's a little much. :-D But just wanted to note it in case the issue reoccurs in other contexts. Ls1g (talk) 02:07, 2 February 2020 (UTC)

Quotations from copyrighted works

Wikidata does not have an exemption doctrine policy, so only copyleft and public domain text may be used in Wikidata under the global licensing policy.

usage example (P5831) and quotation (P1683) are currently used with fairly lengthy values (the longest use of the former, at Szwajcaria (L39612), is 534 characters in length). This poses a problem for these properties, since the vast majority of creative works in existence are currently still copyrighted in the US; it is virtually certain that some uses of these properties violate local policy or global WMF policy. Nevertheless, there is obvious utility to providing usage examples and quotations from after 1924.

Currently, the properties do not have any relevant constraint violations or usage notes which restrict the values which can be used. I think it would be appropriate to properly address the issue.

Without any Wikidata policy changes, resolutions would mostly be limited to those that significantly limit usage of the properties.

  • A simple approach would be to limit quotations to those that are original or that are quoted from PD works. This would be workable and is the approach that quotation or excerpt (P7081) takes, but it would clearly be detrimental to the usage of the properties.
  • Quotations from copyleft works would also be allowed by the global licensing policy and could be allowed here, although this does go against the spirit of Wikidata's CC0 licensing and could theoretically cause copyright issues in some countries.
  • It could be possible to allow quotations which are not complicated or expressive enough to be copyrightable (I believe this would be the justification for allowing titles of copyrighted works to be used). However, there is no explicitly defined limit in US copyright law, so this would probably be difficult to manage without having to carefully prune uses.
  • For usage examples, it would also be possible to introduce alternate properties (e.g. "used in work"; "used in URL"; "usage examples in Wiktionary") with which to state the work that the usage example would be taken from, in lieu of actually providing an example. I don't know if this would be an improvement on the current situation.
  • Technically, doing absolutely nothing would not be a significant violation of the terms of service, and in all likelihood no legal action against the WMF would actually result from this particular issue. However, actively doing this would be a relatively significant departure from the typical practice of taking copyright law extremely seriously.

Any policy change would presumably involve the introduction of an exemption doctrine policy.

  • It would be possible to use such a policy solely to provide a (very low) length limit on quotations from copyrighted material, since US law does not explicitly provide one. While it would be arguable that this would not qualify as an exemption doctrine policy, it is possible that virtually any policy based on a hard length limit would not be sufficient to prevent quotations that would be copyrightable.
  • It would be possible to generally allow fair use under such a policy, but this would most likely be undesirable due to the resulting limitations on the general reuse of Wikidata data.
  • Only allowing fair use in certain contexts, such as only allowing it for particular properties, could be a more tenable option, since reusers would be able to omit such data to avoid licensing issues. This would obviously make Wikidata's licensing situation more complicated, and some reusers would face an additional burden.

I'm leaning towards avoiding any significant policy changes at the moment, but without that it would seem difficult for the properties to actually be useful in the short to medium term. Jc86035 (talk) 13:31, 26 January 2020 (UTC)

I'm not sure what is you meaning of so much text, but your first, simple approach, is what we can use right now. All the rest is out of scope. Of course there might be uses that are a violation of copyright, they should be deleted immediately. Exactly as is done on a daily base on Commons with images that are out of license-scope. Edoderoo (talk) 14:01, 26 January 2020 (UTC)
@Edoderoo: I thought it was worth being thorough here, especially because every available option seems to have significant drawbacks. Furthermore, any narrowing of the acceptable content for these properties – which combined have about 60,000 uses – would result in a lot of work to be done, and it would probably be difficult to automate this process. Jc86035 (talk) 11:47, 27 January 2020 (UTC)

Look at the Berne Convention

Extracts from the Berne Convention (text:; text and commentary:

  • Article 9.2 "It shall be a matter for legislation in the countries of the Union to permit the reproduction of such works in certain special cases, provided that such reproduction does not conflict with a normal exploitation of the work and does not unreasonably prejudice the legitimate interests of the author."
  • Article 10.1 "It shall be permissible to make quotations from a work which has already been lawfully made available to the public, provided that their making is compatible with fair practice, and their extent does not exceed that justified by the purpose, including quotations from newspaper articles and periodicals in the form of press summaries."

The commentary on these articles suggests to me that taking quotes from copyrighted works for purposes of illustration is lawful, provided that the interests of the copyright holder is not compromised. Therefore, particularly if the work concerned is available to the public free of charge (for example The International Vocabulary of Metrology) taking quotes from appropriate copyrighted sources for purposes of Wikidata definitions is permitted under the Berne Convention. Martinvl (talk) 22:42, 26 January 2020 (UTC)

  • The problem isn't with doing that as such. The problem is with doing that and then asserting that all of our content is CC-0. - Jmabel (talk) 02:31, 27 January 2020 (UTC)
    @Jmabel, Martinvl: Yeah, I'm not sure if it would be possible to use this alone to justify lengthy quotations, especially since the articles don't actually state that the quoted/reproduced material would be rendered in the public domain. These also would seem to overlap with the permissions that the US fair use doctrine grants. Jc86035 (talk) 13:47, 27 January 2020 (UTC)
    These provisions in Berne just allow countries to include a "fair use" provision in their copyright laws. They aren't likely to have any direct legal effect in any particular country. The discussion about creating a fair use policy for Wikidata is still open at Wikidata:Requests_for_comment/Non-free_content, but there hasn't been a lot of enthusiasm. Ghouston (talk) 05:25, 28 January 2020 (UTC)
  • The UK and some other countries distinguish de minimis takings from fair dealing. In the UK the case-law test is first whether there has been a "substantial taking" of copyright content, and then only if there has been a substantial taking, to ask whether that taking can be justified by one of the fair dealing exceptions. The assessment of substantiality is qualitative rather than quantitative. We might therefore perhaps accept de minimis takings, even without a formal WMF EDP. But if a taking is noteworthy enough to be the object of a statement here specifically for it, then arguably it is not de minimis or insubstantial; and a collection of such statements even less so. Jheald (talk) 12:19, 28 January 2020 (UTC)


  • Wikiquote is under a "Creative Commons Attribution-ShareAlike License" why are we not harmonized with them? They are compliant with Wikimedia Foundation regulations for Fair Use of copyrighted material. --RAN (talk) 18:54, 28 January 2020 (UTC)
Anyone can argue anything, read any comments thread in any and every online posting. In the end it is the Wikimedia Foundation that determines the legal status of each project and the license it is released under. We should also consider displaying Fair Use images here at Wikidata, as done in English Wikipedia. --RAN (talk) 14:05, 29 January 2020 (UTC)
@Richard Arthur Norton (1958- ): Are you suggesting we host non-free images within WikiData (since they are certainly not allowed on Commons)? - Jmabel (talk) 16:24, 29 January 2020 (UTC)
Not host, somehow display the image without hosting it, since we already store a url for an image. Disable download. See Property:P6500. --RAN (talk) 22:20, 30 January 2020 (UTC)
Embedding images from 3rd party sites isn't a great idea. They can change or delete the image any time they like, or detect the embedding and display an error image [28]. Ghouston (talk) 06:21, 31 January 2020 (UTC)
  • Wikiquote has different scenario's of reuse then Wikidata and for us it's worthwhile to find a model that works for our reusers. ChristianKl❫ 09:56, 4 February 2020 (UTC)


This site in the strategy paper meta:Strategy/Wikimedia movement/2018-20/Recommendations/Promote Sustainability and Resilience includes some ideas about the diversification of the revenue channels and there is something about the API where they suggest to think about a premium service for large API-users. That for your information if you havent read it through until now. Is there a page in Wikidata with a overview about the biggest donations and Grants for Wikidata. As far as I have seen there is a employee at Wikimedia Germany for this topic. Maybe it is also possible to create a page with the costs for Wikidata. So that it is more clear how many the API-use of the large API-users costs. -- Hogü-456 (talk) 17:28, 29 January 2020 (UTC)

  • I think that's okay as a funding source. We at the moment have a state of affairs where certain queries can't be run on Wikidata because they time out. It would be possible to have a system where some large API-users can pay to run those queries. Seperately we could have a permission system that allows us to run some of those queries for the sake of Wikidata for free. ChristianKl❫ 16:57, 4 February 2020 (UTC)

Visits of Q18693398

I was looking at these yearly stats. I am not sure why Body Fat Index of Love (Q18693398) has such a huge amount of visits.

The page has no recent edit and, as I was discussing here, there is no correlation with the visits at the article on the local wiki. Also, there are no specific temporal spikes. There are more visists around April-May but also later. To me, there is no clue there.

It might be a useless curiosity but sometimes these things help to find some bugs, issues, misplaced connectivities or bizzare off-wiki uses... Do you have any idea?--Alexmar983 (talk) 00:36, 31 January 2020 (UTC)

Another interesting aspect is the low percentage of mobile traffic. Usually, off-wiki traffic for articles on Wikipedia can be related to some apps on smartphones (so it was discovered in other similar cases, if recall correctly)... so I am surprised that the figure in that column is so much below avarage in this case.--Alexmar983 (talk) 00:43, 31 January 2020 (UTC)
the visits have spikes around Tuesday. That's also unusual, less expected than fro example week ends.--Alexmar983 (talk) 01:24, 31 January 2020 (UTC)
I have also checkd the incoming links if it had some unusual presence on navboxes or main pages, but nothing.--Alexmar983 (talk) 01:26, 31 January 2020 (UTC)
  • A few have been deleted: Q_64830548 (178,212 pageviews, rank #3), Q_63749024 (76,789 pageviews, rank #14). To find the item, remove "_". --- Jura 07:17, 3 February 2020 (UTC)
Jura but this is not related to the entry of this Finnish movie... right? I am left with my curiosity :)--Alexmar983 (talk) 20:48, 5 February 2020 (UTC)


Is occupation=photographer only reserved for people who are professional photographers or does it include people who own a camera that that took photos as a hobby and they have images at Wikimedia Commons. --RAN (talk) 15:56, 29 January 2020 (UTC)

That's a discussion to be had at the talk page of the property. It seems to me that the view that's expressed there is that the property is about the subject's dayjob. ChristianKl❫ 16:04, 29 January 2020 (UTC)
That view was expressed by one user, in 2015. I think it's wrong; equally I don't know where the dividing line should be placed. We use occupation for e.g. suffragettes, for few of whom it was their 'day job'. (And of course, that usage is also questionable). More widely we use it for a range of activist roles, where many described have other day jobs. For me, RAN's example doesn't merit occupation=photographer any more than the idea that we're all cooks since we once made a piece of toast. Wikipedia usefully comments in its Occupation DAB: "Occupation commonly refers to: Occupation or job, one's role in society, often a regular activity performed for payment" - my emphasis. That may be a useful yardstick; certainly wider than day job. --Tagishsimon (talk) 16:16, 29 January 2020 (UTC)
Not that I at all want a Wikidata item about me, but let me take myself as a middling example. I'm a Commons photographer. I mainly make my living developing software, but I have had photos I've taken used prominently as billboard advertisements, covers of books, in major newspapers such as Ha'aretz and the Seattle Times, and in one book from National Geographic; I've won an award from an architectural organization for my documentation of Mid-Century Modern architecture in Seattle; and I have at least one photo on permanent exhibition in a museum. I suspect that this is probably typical of the more prolific Commons photographers, as is the fact that I mainly make my money elsewhere. We describe H. Ambrose Kiehl (Q38605082) as an "American civil engineer and photographer" and list ; I don't see any sense in which he was more of a photographer than I am, or than a good several dozen Commons photographers are. It seems that if you were going to have an item about me, my work as a photographer is of much broader interest than my work as a software developer. - Jmabel (talk) 17:19, 29 January 2020 (UTC)
It's not a matter or right or wrong but about how we decide to set the scope of the property. That discussion is better had on the page of the property as it's a place where in future people will look if they want to know. ChristianKl❫ 13:06, 30 January 2020 (UTC)
It comes down to common sense, but sense isn't all that common. All scientists and academics write to communicate, but it would be redundant, frivolous clutter to add writer (Q36180) to every entomologist, art historian, and mathematician. Many people take and publish photographs incidental to other roles (e.g. scientific articles). My rule of thumb is that an occupation is worth adding only if it's a role commonly and consistently associated with the subject. In the case of H. Ambrose Kiehl (Q38605082), given that the University of Washington has a collection dedicated to his photographs, I think it is appropriate to add photographer (Q33231). For people known solely or predominantly from posting photos to Flick or Commons, or who take photographs rarely or incidentally, amateur photographer (Q21694268) might be more appropriate. It is also not mandatory that occupation (P106) be completed at all. -Animalparty (talk) 21:09, 29 January 2020 (UTC)
Yes, they do have such a collection as a representative example of amateur photography, in no small part because (unlike most amateurs) he took good enough notes so that we know what he photographed. As it says on the page you linked, "The photographs are typical of those found in many family albums of that period and illustrate everyday family life at the turn of the century." - Jmabel (talk) 00:37, 30 January 2020 (UTC)
  • Perhaps occupation can be used once somebody has achieved some recognition in the field? Being paid is one form of recognition. Being able to publish an article in a peer-reviewed journal or receiving a prize would be another. I don't see a difficulty with using it on a graduate student, for example. We've even got it set to criminal (Q2159907) for people like John Wilkes Booth (Q180914). Ghouston (talk) 00:51, 30 January 2020 (UTC)
I think Occupation=amateur photographer (Q21694268) is a good compromise for now. We can always remove them globally in the future, if we make that decision. I want to be able to distinguish the professional photographers from the amateur ones as I am adding in identifiers from the various photographer databases. I will reserve Occupation=photographer for those that have images in museums and those that operated photo studios. We may develop something like Hobby= in the future. --RAN (talk) 05:19, 30 January 2020 (UTC)
"those that have images in museums and those that operated photo studios" would leave out the bulk of photojournalists working since the end of the film era, since studios as such are no longer common. (Some have work in museums, but many not, I'd guess.) I suspect there is at least one staff photographer for a genuinely major newspaper or magazine who wouldn't meet those criteria. - Jmabel (talk) 05:36, 30 January 2020 (UTC)
  • Yes! Sorry for ignoring photojournalists! The database I am working with seems to contain photographers from the 1800s to 1920s. --RAN (talk) 06:01, 1 February 2020 (UTC)
Perhaps a separate property would be better than creating an item "amateur X" for every occupation X. association football player (Q937857) -> amateur footballer, criminal (Q2159907) -> amateur criminal? But I think that either way it puts too much weight on being paid verses other forms of recognition. Ghouston (talk) 05:48, 30 January 2020 (UTC)
  • IMO the broad understanding of occupation (P106), ie as representing occupation (Q12737077) "label applied to a person based on an activity they participate in" is helpful, for queries to be usefully broad and inclusive finding aids. (ie otherwise queries based on P106 would miss out too many people, that probably are wanted). I therefore think User:Jeblad's 2015 proposal to change the scope to "the subject's day-job", ie specifically work (Q6958747), would have been unhelpful, and I am glad it was not taken up. Jheald (talk) 10:10, 30 January 2020 (UTC)
  • First; Wikidata or any system that model semantic relations will not work properly if we start to use predicates with dual meaning. From Oxfords Advanced Learners Dictionary (1989):
    3 [C] (a) (fml) job; employment: ‘what's your occupation?’ ‘I'm a dancer’Please state your name, age, and occupation. (b) activity that occupies a person's (esp spare) time; pasttime: She has many occupations including gardening and wine-making.His favorite occupation is reading.
    What some try to do is to use both (a) – your job, and (b) – your hobby in the same statement. That creates a mess. Owning a camera and shooting pictures are usually a hobby, only when you start to put a real effort into it, and you are known as a photographer, then you should be named a photographer.
    The template entries using occupation (P106) is now changed from “yrke” (“work”) to “beskjeftigelse” to be somewhat more descriptive for how the property is actually used. The discussions about “yrke” (“work”) has been one of the more contentious debates at nowiki, and resurfacing this discussion might very well end with removal of its use. Jeblad (talk) 12:15, 30 January 2020 (UTC)

Should having works in Wikisource lead to "occupation:writer" claim?

Currently, every person who has an Author page at wikisource is having the claim "occupation: writer" added to their wikidata page. That doesn't seem right to me. Having had just one or two magazine articles or letters published is not the same thing as being a writer as an occupation. The confusion seems to have chiefly arisen because Wikisource automatically adds Author pages to categories such as "Women writers." Should there somehow be standards for what is an Occupation that means WS categories don't automatically translate into occupations? Levana Taylor (talk) 22:45, 4 February 2020 (UTC)

We are a bit confused about what qualifies as an "occupation". See the discussion above Wikidata:Project_chat#Occupation=photographer. I suppose a lot of people may write things throughout their lives, like SMS messages to their friends, and we wouldn't consider them writers any more than we'd consider them chefs for cooking their own dinners, even though they may spend quite a bit of time on these things and develop some expertise. How many magazine articles, books, scientific papers, etc., do you need to write before you are a "writer"? In this case, Wikisource has apparently already classified them as writers. Ghouston (talk) 22:52, 4 February 2020 (UTC)
Yikes, I didn't see that an almost identical discussion had been already started just days ago re: photographer. Okay, so let me clarify: I think general discussion of what is an occupation should be in that thread, and this discussion should question whether it's desirable to automatically translate Wikisource "writer" categories into WD "occupation" statements. Levana Taylor (talk) 23:19, 4 February 2020 (UTC)
Have you any estimate or hunch about the the false positive ratio in such an exercise? By your lights what % of wikisource authors are not occ=writers? --Tagishsimon (talk) 00:10, 5 February 2020 (UTC)
I can tell you that for the magazine I'm working on, there are author pages for nearly all the contributors, and I've tried to put a semi-complete list of all their publications on their pages (though I'm not done with the project). That allows me to estimate that about 10% (not more) have only a couple of publications, in most cases only their contributions to this magazine. I don't know what the percentage of occasional authors is for WS overall but it is certain to be far, far less. Levana Taylor (talk) 00:50, 5 February 2020 (UTC)
Without knowing how "occupation" should be used in Wikidata, it's impossible to say whether copying the entries from Wikisource is a good idea or not, or to say that something is a false positive. Ghouston (talk) 00:57, 5 February 2020 (UTC)
True! I have moved this section as a subsection of the other discussion. Levana Taylor (talk) 01:05, 5 February 2020 (UTC)

Thought in favor of separating author from occupation as writer: in the work of some scholars attempting to uncover the underdocumented lives of women in history, a woman writer is a woman who wrote anything, however little, and no matter whether it was published. For such scholars, women who wrote as a profession, or who had a reputation as writers (for example, those whose works circulated in French salons) are of interest in a different way. WD can't capture every nuance, but seems like this is one that could be codified by having only the latter qualify as an occupation. (the former is a member of the category of woman writers, though). Levana Taylor (talk) 01:12, 5 February 2020 (UTC)

  • Note to closer, please cut and past a copy of this to talk page of Photographer or add a link to it. --RAN (talk) 03:11, 8 February 2020 (UTC)
  • @Levana Taylor: do you mean English wikisource? Do they have speech creators/performers? Because in ru-wikisource we have Category:Speeches (Q7564984) (and in en-ws too as I see) and probably some people only authored speeches, I wouldn't call them "writers". Also we have Category:People by behavior (Q6819158) in which Category:Writers is one of ~100 other occupations... --Infovarius (talk) 11:47, 10 February 2020 (UTC)

Academic journal with different series

Can anyone please help me with this.

For instance this article appeared in The Zoologist 1900, in the 4th series, volume 4, etc. I can add volume 4, but I don't find how I can add that "4th series".

Thanks in advance, --Dick Bos (talk) 12:01, 29 January 2020 (UTC)

Interesting, it's described at en:The Zoologist. It would be possible to make a separate item for each series, e.g., "The Zoologist series 1", but I'm not sure if that's best. Ghouston (talk) 12:08, 29 January 2020 (UTC)
There are journals named in that way, like Deep-Sea Research. Part 1: Oceanographic Research Papers (Q15765364) or Journal of Polymer Science Part A (Q2411132), although it's not exactly the same. My inclination would be to create a separate item for each series, and declare it to be a scientific journal (Q5633421) in its own right. The "parent" item The Zoologist (Q7776911) would also remain a scientific journal (Q5633421), and they could all be linked with has part (P527) and part of (P361). Perhaps not perfectly logical, but is there anything better? Ghouston (talk) 22:13, 29 January 2020 (UTC)
Yes. I noticed this about 18 months ago, when doing some reconciliation/matching work for some of the journals available through the Bioheritage Diversity Library. I did try to ask at WikiProject Periodicals at the time (see archived threads: Journal splits and Journals that change their name, which include some more examples), but got no real feedback. It's an area where we could do with a much firmer style guide. As far as I can see, we seem to be very inconsistent, both with journals being split into multiple parts, and for journals starting "new series", with the volume number being re-set to 1. It also doesn't help that some different external IDs can be inconsistent, sometimes splitting the journal sometimes not (particularly for journals that undergo name changes, but may or may not restart their numbering). I don't have an answer for this also when some sources/wikis wants to say that a journal was "original founded in 1836", and some wikis may have an overall article for its entire history since that time, but it may have gone through a lot of new series and/or part reconfigurations since then. Jheald (talk) 10:26, 30 January 2020 (UTC)
There's also the angle of what is most helpful for people wanting to cite the journal (eg from wiki-based templates), and/or analyse data about journals. Is it helpful to be able to cite <Overall title> with additional parameter <4th series>, or do we require them to specify <Overall title, 4th series>. Would everyone tend to get it right, either way? Pinging @Andrew Gray: for any thoughts. Jheald (talk) 10:30, 30 January 2020 (UTC)

John Vandenberg (talk) 09:30, 2 December 2013 (UTC) Aubrey (talk) 12:15, 11 December 2013 (UTC) Daniel Mietchen (talk) 12:47, 11 December 2013 (UTC) DarTar (talk) 01:37, 15 January 2014 (UTC) Maximilianklein (talk) 00:23, 28 March 2014 (UTC) Mvolz (talk) 08:10, 20 July 2014 (UTC) Andy Mabbett (Pigsonthewing); Talk to Andy 22:17, 27 July 2014 (UTC) Mattsenate (talk) 17:26, 14 August 2014 (UTC) author  TomT0m / talk page JakobVoss (talk) 14:25, 16 June 2016 (UTC) Mahdimoqri (talk) 08:04, 5 April 2018 (UTC) Jsamwrites Dig.log Sic19 (talk) 22:46, 12 July 2017 (UTC) Andreasmperu Nomen ad hoc Pete F (talk) 99of9 Mfchris84 (talk) 09:02, 26 November 2018 (UTC) Runner1928 (talk) 17:22, 1 December 2018 (UTC) Wittylama (talk) 09:55, 22 December 2018 (UTC) Jneubert (talk) 07:30, 22 February 2019 (UTC) --Juandev (talk) 20:28, 27 April 2019 (UTC) VIGNERON (talk) Uomovariabile (talk to me) 08:46, 24 June 2019 (UTC) SilentSpike (talk) Ecritures (talk) Tfrancart (talk) Dick Bos (talk) 10:47, 30 January 2020 (UTC) --Rdmpage (talk) 09:56, 15 May 2020 (UTC) Clifford Anderson (talk) 15:19, 5 August 2020 (UTC) Parobis1 (talk) 00:02, 28 August 2020 (UTC) Susanna Giaccai (talk) 15:24, 5 December 2020 (UTC) Zblace (talk) 19:40, 12 December 2020 (UTC) Alessandra.Moi (talk) 17:23, 20 February 2021 (UTC) ArthurPSmith (talk) 18:06, 22 February 2021 (UTC) Alessandra Boccone (talk) 10:37, 24 February 2021 (UTC) Erfurth (talk) 20:44, 28 February 2021 (UTC) Mgrenci (talk) 21:59, 10 March 2021 (UTC)  Notified participants of WikiProject Periodicals

Citations are a good point. I suppose the way it's cited would depend on which citation style is in use, and it would be nice to be able to generate arbitrary styles from a Wikidata item. In that case, perhaps the series should be a separate property, not mixed up with the journal name. I guess it would be worth looking at how citation software/databases handle it. Ghouston (talk) 00:07, 31 January 2020 (UTC)
  • Here's another example, Annals and Magazine of Natural History: Series 1 - 13 [29]. Ghouston (talk) 04:49, 6 February 2020 (UTC)
  • en:Template:Cite journal has a field "series", which is described as "series or version: When the source is part of a series, such as a book series or a journal where the issue numbering has restarted." I think a new property should be created for it, maybe called "journal series id" to distinguish it from part of the series (P179). Ghouston (talk) 01:06, 14 February 2020 (UTC)