Wikidata:Project chat/Archive/2019/04

From Wikidata
Jump to navigation Jump to search

This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.


Merging duplicated scientific articles



  1. Is it safe to use merge.js to do the merging ?
  2. How did it happen in the first place that the duplicated items are created at the same time and on top of they are updated exactly at the same time ?

Kpjas (talk) 12:01, 31 March 2019 (UTC)

Yes, they can be merged. After merging, check the result and delete any duplicate authors/author name strings etc. I've seen such duplicates before, presumably it's a glitch during the import. Maybe Daniel_Mietchen (the creator of both) can guess why it happened. The subsequent duplicate updates would be because they were both processed the same way by scripts. Ghouston (talk) 19:47, 31 March 2019 (UTC)
I've merged a number of such duplicates. Sometimes it's because one gets entered via DOI, the other via pubmed ID. Or sometimes they're just exact duplicates presumably due to some sort of query/database synchronization issue. ArthurPSmith (talk) 15:57, 1 April 2019 (UTC)

Wikidata weekly summary #358

P463 (member of) and Q1439921 (list of regicides of King Charles I)

The use of member of (P463) for list of regicides of King Charles I (Q1439921) seems a bit odd (it doesn't seem to be "organization or club", as the property says; see Oliver Cromwell (Q44279) for an example). Are there no better properties to use? And should list of regicides of King Charles I (Q1439921) really be an instance of Wikimedia list article (Q13406463)? --Njardarlogar (talk) 07:53, 27 March 2019 (UTC)

"Regicides" is kind of a weird word usage here. It should probably be instance of (P31) signatory (Q28008347) of (P642) Death warrant of King Charles I (Q19756605). Circeus (talk) 16:08, 27 March 2019 (UTC)
member of (P463) list of regicides of King Charles I (Q1439921) is not useful, because list of regicides of King Charles I (Q1439921) in itself doesn't give any indication of what the set in question means. The value of P463 should be a substantive item (ie an actual organisation of club), not a list-type item, which is a different sort of item.
instance of (P31) signatory (Q28008347) is also unhelpful, because a person should only have instance of (P31) human (Q5). All other information about the person should be conveyed by specific properties.
One possibility would be to use participant (P710) with a new item "signing of the death warrant of Charles I", qualified by subject has role (P2868) = signatory (Q28008347).
significant event (P793) could also be an alternative. (Do we have any hard-and-fast rules for preferring one or the other of P710 and P793 for events involving people?)
Pinging @Andrew Gray:, to see if he has any thoughts. Jheald (talk) 21:36, 27 March 2019 (UTC)
  • @Jheald: I think when I added these it was showing as "member of: regicides of Charles I", ie the item was about the group, and has since got converted back into a list. But it's been a while and I don't really remember.
I suspect I went with member of (P463) because this is a definable group of people and they're a member of it. One problem with participant (P710)/significant event (P793) is that we have a lot of well understood but informal groups of people like this, and they're not always defined by something they "participated" in, or by an "event". See, eg, Berkeley Mafia (Q4354584) or Wolseley ring (Q724756), neither of which would be well-described using participant (P710)/significant event (P793). member of (P463) seems a lot more natural, constraints notwithstanding, and I wonder if the more appropriate method would be to loosen those constraints to include groups like this.
Poking around other similarly informal groupings, Disney's Nine Old Men (Q241920) and traitorous eight (Q1883987) have everyone connected using part of (P361), which is another possibility for groups. But it seems odd to have two different properties for the same basic concept purely depending on whether we identify the target item as an "organisation" or not - there will be many potential groups where it's borderline. Andrew Gray (talk) 00:06, 28 March 2019 (UTC)
  • PS: @Circeus:: it is a bit of an odd word, I grant, but it's also the one pretty much universally used by historians for this particular group of people, so I feel we should definitely retain it as a label if possible. Note that they didn't all sign the death warrant, some were involved in various other ways. Andrew Gray (talk) 00:09, 28 March 2019 (UTC)

Thinking back on this, I think part the issue has more to do with the unusual list nature of the main article more than the semantics of it. There is a group that can be conventionally defined. It just happen to be in a list format on Wikipedia, because they are connected to an event rather than an organization. It would probably be a not unreasonable option to actually switch the item title back to "Regicides of King Charles I", which is certainly less awkward when used with part of (P361). Circeus (talk) 18:19, 2 April 2019 (UTC)

How long does a P1630 (formatter URL) change take to propagate?

This morning I changed the formatter URL (P1630) for British Library system number (P5199) (diff), rewriting the old one with a new one, and the change seems to have propagated almost immediately -- within a few minutes at most.

But this afternoon I changed the formatter URL (P1630) for UK Parliament thesaurus ID (P4527) (diff), changing the rank of the old one to 'normal' down from 'preferred', and adding a new one with preferred rank -- but nothing seems to be changing.

Is there an issue here, or do I just need to be more patient? Jheald (talk) 19:16, 28 March 2019 (UTC)

Additional: Purging a page where the property is being used (ie adding ?action=purge on url) updates the url link to the now preferred option. (hat-tip & thanks to Andrew Gray)
But is this something that is having to be done manually? I am not convinced that, when P1639 is changed in the second way above, the signal is being sent out to pages where the property is being used that they need to be regenerated. Sometimes such a signal can take a while to propagate, which is okay, but here is it being sent out at all? Jheald (talk) 19:33, 28 March 2019 (UTC)
@Jheald: Indeed, at the moment entity pages aren’t updated when the formatter URL changes – that’s the subject of T112081. (Similarly, their RDF in the query service isn’t updated when a formatter URI for RDF resource (P1921) is edited either; I don’t think we have a task for that yet.) --Lucas Werkmeister (WMDE) (talk) 10:23, 29 March 2019 (UTC)
@Jheald, Lucas Werkmeister (WMDE): given that, shall I just knock together a script to purge all the affected pages over the next few days, or is there a more elegant workaround? Andrew Gray (talk) 22:00, 31 March 2019 (UTC)
@Andrew Gray: I’m not sure if I should encourage that (it sounds like a lot of extra server load), but I’m not aware of any better solution. How many pages (roughly) are we talking about? --Lucas Werkmeister (WMDE) (talk) 10:46, 1 April 2019 (UTC)
@Lucas Werkmeister (WMDE): ~5500, I think. I can spread it over a longer period or a few batches, rather than try and do them all in one rush. Andrew Gray (talk) 15:24, 1 April 2019 (UTC)
@Andrew Gray: okay, that sounds acceptable. (I think mw:API:purge doesn’t actually re-render the pages anyways, so that would be spread out to whenever each page happens to be visited the next time.) --Lucas Werkmeister (WMDE) (talk) 16:08, 1 April 2019 (UTC)
@Andrew Gray: If you've got a script, could you do the BL sysnums as well -- it seems I was overly-optimistic above, and their new formatter hasn't propagated either. Thx. Jheald (talk) 09:39, 2 April 2019 (UTC)

Second line and language box

Hello. When you add media legend (P2096) you have to add the language. The problem is that when you are writing the legend the language box appears. And if you need two lines for the legend, then you can't see the second line because of the language box (you can not see what you are typing). The language box must be some mm lower. Xaris333 (talk) 08:19, 30 March 2019 (UTC)

I wonder why P2096 value should originally be languoid-less, can you explain more here? --Liuxinyu970226 (talk) 10:45, 2 April 2019 (UTC)
What is a "languoid"? - Jmabel (talk) 15:37, 2 April 2019 (UTC)

New accounts and QuickStatements

Quick question: are there any restrictions before new accounts can use QuickStatements -- eg account must have existed for a certain number of days, must have made a certain number of edits? Jheald (talk) 12:28, 30 March 2019 (UTC)

Account needs to be autoconfirmed, so 4 days and 50 edits. --Tagishsimon (talk) 17:02, 30 March 2019 (UTC)
 Resolved Jheald (talk) 09:35, 2 April 2019 (UTC)

New to Wikidata, could do with some advice re merging WebP, a web image file format by Google


While I'm not new to Wiki -- I work every so often in EN, DE and Commons -- I'm new to Wikidata, and could do with some advice.

Google has an -- unloved by some, but that is besides the point -- image format for websites called WebP, and distributes programs for converting images to and from this standard. The same program allows for both lossy and lossless conversion, and images converted to the format, both lossy and lossless, have the same file extension .webp. The touted advantage is that WebP compresses files better than the JPEG or PNG formats: file size reductions between 23% and 40-odd% are claimed.

Somehow the WebP format has ended up with three Wikidata entries: WebP Lossless Q683670, WebP Lossy Q45989100 and WebP Extended Q45989477, which seems pretty ludicrous because they are all inside the same file!

What I would propose is to

  • merge all three of them into one single entry
  • rename the entry plain WebP
  • put into the description lossless and lossy file format, and reference extended
  • put both lossy and lossless into the Statement (I looked at JPEG Q2195, it seems possible)

If anyone is interested, the whole file format is described at, in case there is something important I'm overlooking.

Ta very muchly!

--Peter NYC (talk) 23:53, 31 March 2019 (UTC)

幹事長/総書記, 간사장/총비서, 秘書長/總書記, Secretary-general/general secretary

The distinction between general secretary and Secretary-general is a bit unclear and as a result they are currently being used interchangeably. Could someone with Japanese/Korean/Chinese skills have a look? Moebeus (talk) 12:54, 1 April 2019 (UTC)

ja:総書記 looks like specified for communist party (Q233591) (and that article generally mentioned Communist Party of China (Q17427)), while ja:幹事長 is more commonly used.
zh:總書記 is always pointing to, and only to, the political party (Q7278), while zh:秘書長 can also include government departments, NGOs, companies, and (don't laugh) yourself.
For Korean, better to ask at Wikidata:사랑방 (you may translate your questions via [1]). --Liuxinyu970226 (talk) 11:38, 2 April 2019 (UTC)

--Liuxinyu970226 (talk) 10:28, 2 April 2019 (UTC)


Could someone make or tell where/how to make CIN ID property for politicians in BiH? It is this website. Example for URL identificator is profil.php?profil=85, simply a number. --Obsuser (talk) 16:31, 1 April 2019 (UTC)

@Obsuser: You would need to post a "property proposal", at Wikidata:Property proposal/Authority control. Let me know if you need help with that. However, note that the site says (according to Google Translate) "more than 120 officials and politicians... by 2014, that number in the database has grown to 200. Some of the politicians from this base are no longer active or have died because of which their profiles no longer found in the database." Those numbers are low, for justifying a new property, and we need to be sure that the IDs are not reused. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:57, 1 April 2019 (UTC)

Mix'n'match scraper question

parameters used for a follow level in Mix'n'Match

I'm a frequent user of the mix'n'match scraper, but in creating catalogs, I've only used keys and range levels. I understand what the follow levels should be used for, but haven't been able to figure out what should go where. Would anyone be able to explain, or point to an explanation? Thanks! Trivialist (talk) 18:10, 1 April 2019 (UTC)

@Trivialist: I also find this hard. Here's one I did a while back, for which I took a screenshot, so it must have been sort of working or almost working. Hopefully it helps.--99of9 (talk) 23:17, 1 April 2019 (UTC)
@99of9: Thanks! Trivialist (talk) 00:11, 2 April 2019 (UTC)
I also tried many times, succeeded exactly once, and did not take any notes of what I did ^_^" (don’t even remember which catalog it was). Thanks 99of9, that may come in handy :). Jean-Fred (talk) 08:32, 2 April 2019 (UTC)

Facto Post issue 22

It is at w:User:Charles Matthews/Facto Post/Issue 22 – 28 March 2019. Of particular interest for those involved with Wikibase is the link to a blogpost by Michael Dales, who has been contract programmer to ScienceSource (Q55439927) for the past six months, giving an outside view of the MediaWiki API aspects. Charles Matthews (talk) 09:16, 2 April 2019 (UTC)

Popular classifieds website Friday-Ad

I'm looking to add the Friday-Ad to wikidata. It's a very large UK classifieds and commununity website, popular down in the south of England. It's a leading marketplace for motors, pets and all sorts of second hand goods. You can view it here:

Its similar to Gumtree ( and Preloved  – The preceding unsigned comment was added by Miloatfriday (talk • contribs) at 12 March 2019 14:13 (UTC).

This section was archived on a request by: |1=Sjoerd de Bruin (talk) 18:47, 8 April 2019 (UTC)

Request for protect:Q179294

This page is being vandalized by User:Jesamsex and other IPs, socket puppets of User:Unypoly for a long time, by adding a duplicate Korean version article w:ko:환자 (역사) and separate other Asian language versions.--Zhxy 519 (talk) 01:38, 18 March 2019 (UTC)

Q56145599 means Eunuch public official. It is a subclass of eunuch. It is not vandalism.  – The preceding unsigned comment was added by 2001:2d8:e08e:4f67::12ef:40a1 (talk • contribs) at 03:50, 20 March 2019‎ (UTC).
Please ban this IP. It should be no doubt a socket puppet of User:Jesamsex.--Zhxy 519 (talk) 02:42, 24 March 2019 (UTC)
@Zhxy 519: I don't hear arguments from you about the subject. Please can you repeat them? --Infovarius (talk) 14:49, 29 March 2019 (UTC)
Sorry I don’t get the “argument” you mean. And how do you mean to repeat it?—Zhxy 519 (talk) 02:38, 30 March 2019 (UTC)
@Zhxy 519: "Argument" in the sense of stating a case. You have asserted they are sockpuppets, but you haven't provided any evidence, or if you have neither Infovarius nor I can find it. - Jmabel (talk) 04:37, 30 March 2019 (UTC)

@Jmabel, Infovarius:Ah OK.

  1. w:ko:사용자:Jesamsex. This blocked userpage in ko.wikipedia showing no doubt that Jesamsex is a puppet.
  2. Whois showing above IP is from Korea, where Unypoly comes from.
  3. Another Whois also comes from Korea, which is rolling back directly towards this page.

--Zhxy 519 (talk) 00:49, 31 March 2019 (UTC)

@Twotwo2019: Is your this edit meaning that you agree to merge both? --Liuxinyu970226 (talk) 02:52, 31 March 2019 (UTC)
@Zhxy 519: and I meant argument about equality of these subjects. To me, non-speaker of Korean, seems that both senses have right to exist. --Infovarius (talk) 10:48, 1 April 2019 (UTC)
  1. There is no difference between the sock puppet’s work and the original one. He or she doing this for only mischief. I have already started a delete discussion in Ko.wikipedia. You can be more patient.
  2. Even if in Korean these two are different, the sock puppet can’t move some many languages to only fit his minor one.—Zhxy 519 (talk) 00:44, 2 April 2019 (UTC)
@Zhxy 519: Hmm, why not? The very meaning of Wikidata items (and sitelinks in it) is to serve the most correct correspondence as possible. If you are afraid to lose some inter-links you can use redirects for restoring usual (but not very exact) linking. --Infovarius (talk) 09:00, 2 April 2019 (UTC)
Since even in Korean the sock puppets’ edits could be considered improper, I strongly doubt that if they really understand the content in other languages. Which means, they could be hardly capable of making edits to “serve the most correct correspondence as possible”.—Zhxy 519 (talk) 02:29, 3 April 2019 (UTC)

Question about property

We have this property said to be the same as (P460) wich is in the description stated to be "this item is said to be the same as that item, but the statement is disputed" do we have a property for this is exactly equal to I ask because I do some work on military ranks and they have diiferent names in Army, Navy and Airforce but they are on the same rank level and have roughly the same command responsibilities/ duties Andber08 (talk) 11:20, 2 April 2019 (UTC)

how about exact match (P2888)? MSGJ (talk) 14:13, 2 April 2019 (UTC)

The exact match (P2888)? only allows URLs and as far I understand its not ment to be used internaly on WikidataAndber08 (talk) 16:00, 2 April 2019 (UTC)

@Andber08, MSGJ: No, exact match (P2888) is for linking to external URL's that are the same conceptually. Within Wikidata, if two things are really identical they should be merged into one item. However, the presence of sitelinks may prevent a merge, so we have the permanent duplicated item (P2959) property to handle those cases. In this case however, said to be the same as (P460) is exactly the right property to use - from some perspectives these ranks are the same, but they do have distinct labels and other distinctions like military branch. ArthurPSmith (talk) 16:03, 2 April 2019 (UTC)
"Having roughly the same responsibilities" is not the same thing as "being the same thing". --Yair rand (talk) 23:16, 2 April 2019 (UTC)

Usuario discusión:Pepa.charro

Hola, Soy Pepa Charro y he escrito en varias ocasiones para eliminar la fecha de nacimiento del perfil que han creado sobre mi persona. NO ES MI FECHA DE NACIMIENTO

✓ Done, eliminado--Ymblanter (talk) 20:07, 2 April 2019 (UTC)

Change a property datatype

Is it possible to change a property datatype post-creation?

Currently Irish Grid Reference (P4091) is set to the external identifier datatype. However, this should really be string as it's just a coordinate value (the Irish equivalent of OS grid reference (P613)).

I've already removed Wikidata property for authority control for places (Q19829908) from the property --SilentSpike (talk) 15:48, 2 April 2019 (UTC)

the way to go is by replacing the property with entirely new one through Wikidata:Property proposal. Circeus (talk) 20:55, 2 April 2019 (UTC)
It may be possible for the developers to fix a case like this, where the value is not changing, just the datatype. Try the Contact the development team page... ArthurPSmith (talk) 20:59, 2 April 2019 (UTC)

Help with Hebrew needed

It looks like the NNL IDs associated with Orientalism (Q1574967) are for Hebrew translations of this work. Can someone fluent in Hebrew make sure these values are on the correct item(s)? Thanks. (I've separated the English first hardcover and paperback editions from the work item.) - PKM (talk) 22:17, 2 April 2019 (UTC)

Translation property

Hello wikidatans, I was looking for a property to check the langauges a work (ietm) is translated to. Does such a property exist? My search about translation properties did not yiel to sufficient results. If it does not exist I'd like to suggest it to be created. --Sky xe (talk) 00:01, 29 March 2019 (UTC)

@Sky xe: The only property I'm aware of is has edition, but that of course covers any edition or reissue, not just translated ones. The way it's being used is also quite random, with some editors indicating language using a qualifier, while others do not. Personally I think that what you're describing is a perfectly reasonable/very useful query scenario and would support a proposal if you make it. EDIT: to add to the confusion, I should mention that modified version of and derivative work are also sometimes used to "hack" the translation issue. There may be others as well, of course. EDIT2: edition or translation of deserves a mention of course, going in the opposite direction. Moebeus (talk) 00:26, 29 March 2019 (UTC)
For what it's worth, here's a query for the translations currently in Wikidata, looking for edition or translation of (P629) and a different language of work or name (P407) :
SELECT (year(?date) AS ?year) ?old_work ?old_workLabel (GROUP_CONCAT(DISTINCT ?old_lang_label; SEPARATOR = ' / ') AS ?old_langs) 
    (GROUP_CONCAT(DISTINCT ?new_lang_label; SEPARATOR = ' / ') AS ?new_langs)  ?new_work ?new_workLabel (year(?new_date) AS ?new_year)
  ?new_work wdt:P629 ?old_work .
  ?new_work wdt:P407 ?new_lang .
  ?old_work wdt:P407 ?old_lang .
  FILTER (?new_lang != ?old_lang) .
  OPTIONAL {?old_work wdt:P577 ?date} .
  OPTIONAL {?old_work wdt:P571 ?date} .
  OPTIONAL {?new_work wdt:P577 ?new_date} .
  OPTIONAL {?new_work wdt:P571 ?new_date} .
  ?new_lang rdfs:label ?new_lang_label FILTER(lang(?new_lang_label) = 'en') .
  ?old_lang rdfs:label ?old_lang_label FILTER(lang(?old_lang_label) = 'en') .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en,fr,de,it,es,nl,la,pl,ru,uk". }
} GROUP BY ?date ?old_work ?old_workLabel ?new_work ?new_workLabel ?new_date
ORDER BY ?year str(?old_workLabel) str(?old_work) ?new_year str(?new_workLabel)
Try it! Jheald (talk) 09:18, 29 March 2019 (UTC)
@Moebeus, Jheald: Thank you a lot for your answers. The work around to get the translations of a work seems to be not satisfying enough, as it does not easily answer questions as: "give me a list of most translated works" or "sort the languages based on translation counts of certain work (e.g. book)". I think it makes sense to judge the translation movement of important works with a direct subject -> property -> object format. I'll suggest it to be created and hope it will be accepted.

BTW: I would suggest the name "Translated to" (=language) to avoid ambiguity with existing properties. Would be glad and thankful for initial feedback! Sky xe (talk) 15:27, 3 April 2019 (UTC)

@Sky xe: Looking forward to see your proposal! I would suggest you clarify how narrow or general you want the property to be, like will it apply to revoiced work, etc. Moebeus (talk) 15:47, 3 April 2019 (UTC)

Porn link in Q3070476

Property:P856 in Q3070476 points to a porn site. Apparently the old weblink has been re-used by a porn operator. What should we do with it? My suggestion is that we delete it completely as it does not add any useful information to the item. Csigabi (talk) 17:59, 30 March 2019 (UTC)

Have you looked whether has indexed the page? If so, we could link there insteadly. ChristianKl❫ 19:59, 30 March 2019 (UTC)
Looks like that has now been done. Is there any way to kill the link that now leads inappropriately to a porn site? (I understand the need to keep the URL for the record, but is there any way to kill it as a live link?) - Jmabel (talk) 08:18, 31 March 2019 (UTC)
  • The correct thing is to set an end date. Whatever application one builds can take this in account. Wikidata doesn't or shouldn't delete information that doesn't fit a presentist POV. --- Jura 10:20, 31 March 2019 (UTC)
    • End date is there, but so is a live link to a porn site. What would happen if someone similarly took over a URL and had a site with malware? I think we ought to have some mechanism by which we can prevent a live link in situations like this. - Jmabel (talk) 17:36, 31 March 2019 (UTC)
      • @Jmabel, Csigabi, ChristianKl: Better continue the discussion there with the devteam (sorry, should have pinged but totally missed). Lydia gave a first answer and thinks the links on Wikidata are not directly consulted by most people, so it’s not much of a problem) author  TomT0m / talk page 18:05, 31 March 2019 (UTC)
        • To clarify: my point was that most people consume Wikidata's data through other tools and services. So if we solve it locally here in the Wikidata UI then we've only solved a very minor part of the problem. --Lydia Pintscher (WMDE) (talk) 17:35, 2 April 2019 (UTC)

Maybe I'm way off, but wouldn't simply moving the url from Property:P856 to archiveURL using the Internet Archive wrapper solve the problem? The original link is preserved and we don't help the current pornsite and it's rankings? Moebeus (talk) 15:55, 3 April 2019 (UTC)

[Breaking change] Empty containers in JSON outputs will be serialized as empty object "{}"

Hello all,

This change is relevant for everyone using APIs, JSON output and dumps in their tools or gadgets.

Currently, when an object is empty (for example descriptions and aliases), it is rendered as an empty array [] in JSON. (Example) We want to serialized it as an empty object {} instead. This change will ease the deserialization process and bring more consistency in our code as some other places already use JSON objects instead of arrays.

The impact of this change will be in JSON outputs (Special:EntityData) and JSON dumps, as well as the output of wbgetentities, wbgetclaims and editing APIs.

If you’re maintaining tools that use Special:EntityData, you may want to check your code to make sure that it reflects this change, e.g. items with no labels, descriptions, aliases or sitelinks are properly deserialized by your tool.

You can already test your code against our test system on, for example on this item. According to our stable interface policy, the change will be enabled four weeks after this announcement, on April 30th.

In the meantime, if you have any issue or question, feel free to leave a comment in this ticket.

Thanks for your attention, Lea Lacroix (WMDE) (talk) 10:21, 2 April 2019 (UTC)

@Lea Lacroix (WMDE): I don't think that this can be a breaking change, maybe just say "import change" is enough? --Liuxinyu970226 (talk) 10:24, 2 April 2019 (UTC)
It's a change that could easily break something, so the wording seems appropriate. Ghouston (talk) 10:49, 2 April 2019 (UTC)
According to the definition of the stable interface policy, a breaking change is "a change to an API or data format that violates guarantees given or widely assumed before". In that case, the impact is mostly for people reusing Wikidata's data on their tool, still it is a change on something that was assumed otherwise. Lea Lacroix (WMDE) (talk) 11:01, 2 April 2019 (UTC)
that's very counter-intuitive. Data type is not something that depends on object state (empty or not empty). I've never seen rule like "if array is empty, serialize it as empty object, rather than an empty array" before. Not sure how exactly it shall be handled in strong type based language tools and libraries (such as Jackson for Java). Do developers have any examples or advises? Would be much better just to omit value completely (do not output empty object/array at all). -- Vlsergey (talk) 11:32, 2 April 2019 (UTC)
@Vlsergey: Nonempty labels, descriptions, aliases, or claims are already serialized as objects (e. g. { "en": { "language": "en", "value": "Douglas Adams" } } or { "P31": [ … ], "P18": [ … ] }). The fact that those objects turned into arrays when empty was a bug, which we’re now fixing. --Lucas Werkmeister (WMDE) (talk) 12:43, 2 April 2019 (UTC)
Ohh, seems i misunderstood the change. If all objects will be objects, and arrays stay arrays (whatever content is), that's good. -- VlSergey (трёп) 13:17, 2 April 2019 (UTC)
@Lea Lacroix (WMDE): according to the policy you're supposed to notify the Pywikibot list. I don't see an email yet. Did you forget to do this or is it stuck somewhere in moderation? For the pywikibot part I created Phab:T219891. Multichill (talk) 16:06, 2 April 2019 (UTC)
Hello @Multichill: I sent the email yesterday in the same time as the others, so I guess it's still in moderation. Lea Lacroix (WMDE) (talk) 08:04, 3 April 2019 (UTC)
This is a great move! If we are to change dump format, then I humbly suggest the addition of revision ids to the entities dumped. I think this would make it easier for data consumers to evaluate freshness and keep data in sync after importing a dump. See the corresponding patch. − Pintoch (talk) 12:43, 3 April 2019 (UTC)

Properties for external MediaWiki wikis

Hi! There are a number of properties for MediaWiki wikis outside of the Wikimedia ecosystem. I just created Wikidata property linking to external MediaWiki wiki (Q62619638) for such properties to have a way to track them (I added it to many properties, but I probably missed some). Now, a discussion has come up about one that I proposed – Lokalhistoriewiki article (P6520) – because I used the article title for the URL formatter instead of the Page IDs. Now, the question doesn't just concern this ID, but rather all of these IDs. Most of them in fact use tha article titles, but some use the page IDs and some use specialized IDs (e.g. in external Wikibase instances or other MediaWiki extensions that give stable identifiers). So the question is what we should use? As far as I can see, there are two alternatives:

  1. Use the Page IDs. A page ID for a page on a MediaWiki website can be found by appending ?action=info to the URL, and can be linked to via <wiki URL>/Special:Redirect/page/<ID>. The advantage with this is that the page ID is stable – if a page gets moved, the page ID stays the same. The disadvantages are that they are non-intuitive to find, and the links they produce are not very telling.
  2. Use the page titles. The page names are of course very easily available to everyone, and the links produced are easy to read and usually say something about the subject of the article. The disadvantage is that these are not stable – pages can be moved, and the redirects that are created can be replaced with other pages, basically creating a form of link rot (Q1193907) here at Wikidata (and elsewhere where these properties may be used). There are ways to mend this – someone skilled enough could create a bot on Tool Labs that keeps track of page moves on these external wikis and automatically changes the values of the statements here on Wikidata when pages are moved. This will essentially be mimicking what happens when a page in any Wikimedia project is moved.

These properties are basically just sitelinks for external wikis, and hopefully one day Wikibase would be able to handle this natively. However, I'm not aware of any development on this at the moment, so at the moment these are the options I believe we have. So, what do you think is the best solution here? Page IDs or article titles? Jon Harald Søby (talk) 13:50, 1 April 2019 (UTC)

Pictogram voting comment.svg Comment I had raised similar questions in Wikidata:Property proposal/RegiowikiAT ID (Also, I had a SPARQL query there that returns similar results to Wikidata property linking to external MediaWiki wiki (Q62619638)). Jean-Fred (talk) 14:50, 1 April 2019 (UTC)
I like the idea of using IDs instead. Perhaps a bot can fetch the IDs in case it's too hard for regular users to find it. NMaia (talk) 22:42, 3 April 2019 (UTC)

Honorary doctorates

I recently noticed honorary doctor of the McGill University (Q62061080) being added to items on my watchlist. While I'm sure that was done in good faith, it strikes me as the wrong way to model the data. What do others think? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:58, 3 April 2019 (UTC)

There are a few different ways that honorary doctorates are handled, and it would be a good idea to pick a single system. Items named like this one may come from Wikipedia, e.g., Category:Honorary doctors of the McGill University (Q8943449) or Category:Honorary doctors of the University of Oxford (Q7982096). Ghouston (talk) 19:48, 3 April 2019 (UTC)
Also discussed briefly at Property_talk:P166. Ghouston (talk) 19:54, 3 April 2019 (UTC)

Data model intent on ranking and Wikidata documentation on it contradiction

Quoting mw:Wikibase/DataModel#Ranks_of_Statements

  • Preferred statements refer to the most important and most up-to-date information that should be used per default in most contexts (example: only most recent population figures for Berlin would be shown in the Wikipedia infobox for Berlin). Note that there may be multiple preferred statements. This may imply a multi-valued property (e.g. a person's children), or a disagreement (diverging population figures given by different sources).

Quoting Help:Ranking : « The item for Barack Obama has two values listed as children; both values should be given the normal rank because neither value is more "correct" than the other »

It’s an inconsistency. How do we solve this ? author  TomT0m / talk page 11:48, 28 March 2019 (UTC)

I think the example of multiple children being "preferred" is relevant only if there are other values to which these need to be preferred. Equally valid multiple values, like single values, are normally "normal". - Jmabel (talk) 15:08, 28 March 2019 (UTC)
Does not make a lot of difference in practice indeed if there is no other information, but your interpretation is contradicted by the Normal rank description in the data model document Normal statements contain relevant information that is believed to be correct but that may be too extensive for showing it by default (example: historic population figures for Berlin for many years).. The children are really neither. However one application could be that in crowded infoboxes we could exclude by default normal rank informations. I know in frwiki for example Wikidata infoboxes offers a way to disactivate one field of data present on Wikidata if a contributor feels like it’s not really a relevant information. author  TomT0m / talk page 17:16, 28 March 2019 (UTC)
Again, "normal" is… normal. In most multi-value cases, where all values are equally appropriate and all shown, they are all marked as "normal". "Deprecated" indicates there is something wrong with the particular value. So if we need to distinguish "here are a bunch of valid values, e.g. some of them historical, but there are too many to show routinely," then we mark the ones that are worth showing as "preferred" and leave the others that are true but not equally useful as "normal". - Jmabel (talk) 19:31, 28 March 2019 (UTC)
@Jmabel: We don’t really understand each other. I don’t really see an answer to my question in your statement, so it remains, but whatever, I guess this is not a big deal. I don’t really take back what I said about a potential use case however. Not really thinking about the fact that there is actually multiple value in a particular case could be valuable. I mean, if we consider there is a « main » child (say child1 (whatever that means, probably not the best example) for a person (say person1) with multiple value for the « child » property (say child2 and child3), and we decide child1 should be preferred and child2 and child3 should be « normal ». If we have another person2 with only one child child4. By your rule, child4 should be « normal », but what if it’s in exactly the same position relative to person2 that child1 is to person1. We could use « preferred » rank to indicate that, and a Wikipedia could choose to use only the values indicated as « preferred » in its person infobox because it aims to be parcimonious, and we know the criteria that implies the child is on « preferred » rank, and it do not really want the « normal » one in its infobox — again, just a (bad and un-thought of) criteria, but say we choose the « main » children are those raised by the person and not those he did not even know he had, if he’s a man. The advantage is that it’s a consistent rule, the rank does not depend on the fact that there is actually something to rank. And the definition of « normal » is in that spirit, it does not refer to the fact that there are other values, it depends on the fact that it may not be chosen in an infobox. On the example, we may not want to rely on the number of children to choose not to show the children unknown to the genitor, and not show unknown children in the infobox of the genitor even if it’s his only child. author  TomT0m / talk page 20:04, 28 March 2019 (UTC)
@Jmabel, TomT0m: I am pretty sure that the paragraph written about preferred rank after a long discussion of a specific case related to children listing. This is a bad idea to use example in the definition, an example should come after the definition.
We can have a list of three children with two defined as preferred rank and one considered with normal rank if there is a parmeter allowing to distinguish the different children. One can be the authority of the source defining the filiation: if several recognized sources mention that a person X has 2 children and one obscure source mentions 3 children then to take account of the obscure source without promoting it, it can be worth to use rank to distinguish the quality of the sources. It can be used to distinguish other situations like illegitimate child especially when no legal way is available to confirm the filiation. Snipre (talk) 19:45, 29 March 2019 (UTC)

@Jmabel, Snipre: Actually I think there is a link from the question on the usage of ranks and the section #A lot of partners above. Could rank be used to distinguish between the importance of partnership ? My problem with this is the conflict with the fact that one use of the ranks is the history of the information. If a partner is important but historic, putting it in the « normal » rank make the information of relative importance lost compared to an unimportant historic partnership. This seems to be a conflict of usage of ranks, but at the same time seems to be the usecases the ranks were supposed to address. (Also curious to hear if Lydia Pintscher (WMDE) have an opinion of this, on account of the dev team.) author  TomT0m / talk page 17:26, 4 April 2019 (UTC)

Collegiate athletics conferences

There are many inconsistencies between different collegiate athletic conferences and I would like to discuss and form best practices for them all to adhere to.

  • It seems most if not all use Has Part to connect to the member institutions - I think this works. This allows for and should require listing of Start Time and End Time which gives context and allows for previous members to be listed
  • Some conferences use Has Part on their member universities (Big 10, ACC, while others use it on the university's sports teams (SEC). I think it is more accurate to use Has Part with the full university, as membership in these conferences does have minor relations outside of just athletics.
  • Big 10 is currently using the qualifier Mascot on its Has Part property. This is not valid. Though I do think it adds nice context and as I'm new to Wikidata, I look to others for the best solution (removal of Mascot, or inclusion as some kind of exception)
  • I have not found any University items that have a property listing their athletic conference. I would think that Member Of would be a valid property to include an athletic conference and would be helpful data for those items.
  • I can imagine cases in which knowing "past membership" of conferences would be helpful. Listing Start and End on the conference items is a good start, but it could also be helpful if there existed some kind of "Past Member" or "Left Membership Of" type of Property that could be added on the appropriate university items. Does such a thing exist?

Thanks! Kushboy (talk) 18:59, 3 April 2019 (UTC)

There is probably a need to standardize whether universities or athletics programs are connected via has part (P527) and member of (P463) (and indeed both properties should be in use, not just one). As for your last point... there's not much to discuss since you already point yourself the obvious (and well-agreed) solution in your first point. Circeus (talk) 16:38, 4 April 2019 (UTC)

SA Rugby player ID (P3678) and Springboks Rugby Player ID (P6662)

@TheMightyPeanut, Gabbg82, Eihel, ZI Jony: Hello.Both are on the same subject.Please correct.Thanks --David (talk) 07:41, 4 April 2019 (UTC)

@TheMightyPeanut, Gabbg82, Eihel, ديفيد عادل وهبة خليل 2: Springboks Rugby Player ID (P6662) SA has been changed as Springboks. Regards, ZI Jony (Talk) 12:09, 4 April 2019 (UTC)

Agregar Información Fiscal Española

Buenas tardes,

Me interesaría volcar buena parte de la información fiscal de nuestra web. Es un contenido serio y bien trabajado. De las múltiples secciones, la que me propongo compartir es la que hace referencia al IRPF. Actualmente está en nuestro servidor ( dicho servidor es de una empresa con ánimo de Lucro, es decir meramente comercial ).

Sin embargo, la información de los blogs son de acceso gratuito. Bien, mi pregunta, es si puedo volcar ese contenido, y de que manera; considerando que la fuente es una web de carácter comercial.

Un Saludo y gracias

Josep navarro

PD: Dejo el enlace de la sección del blog.

Rough translation (per Google): Good afternoon, I would be interested in dumping a large part of the tax information on our website. It is a serious and well-worked content. Of the multiple sections, the one that I propose to share is the one that refers to the IRPF. Currently it is on our server (this server is from a company with the spirit of Profit, that is, purely commercial). However, the information on the blogs are free access. Well, my question, is if I can dump that content, and in what way; Considering that the source is a commercial website. A greeting and thanks Josep navarro PS: I leave the link of the blog section Bovlb (talk) 18:05, 4 April 2019 (UTC)

Likely duplicates - Spanish diplomat and writer Eulogio Florencio Sanz and Spanish diplomat, writer and politician Eulogio Florentino Sanz

Two individual articles on Spanish wikipedia block a merge, but it looks to me like Eulogio Florentino Sanz and Eulogio Florencio Sanz are duplicates. Could someone with a flair for Spanish poetry and/or diplomacy take a look? Moebeus (talk) 02:35, 5 April 2019 (UTC)

  • Certainly the same person. es:Eulogio Florencio Sanz is an uncited stub (not that es:Eulogio Florentino Sanz is well-cited, but, oh well...), and in the unlikely event that it has any significant content that is not in the other article that should be merged. In any case, it should be redirected and the items here should be merged. - Jmabel (talk) 03:11, 5 April 2019 (UTC)

Data about historical places in Wikidata

Can you help put Wikidata:WikiProject Historical Place up to date with current practises and latest properties? We are helping the GLAM sector in Finland define the data structures for national authority files for historical places, and hope to present the best practises from Wikidata. – Susanna Ånäs (Susannaanas) (talk) 15:36, 5 April 2019 (UTC)


It looks like someone decided to make an entry of their self. I'm assuming that this isn't allowed and the item needs to be deleted, but what is the actual policy on this? Sario528 (talk) 16:05, 5 April 2019 (UTC)

Wikidata:Autobiography applies here. It does not prohibit that, but unless the person is notable, the item should be deleted (and it is not well seen to create articles of oneself). I've deleted it. Esteban16 (talk) 17:34, 5 April 2019 (UTC)
Thanks for taking care of it, also for the policy links. Sario528 (talk) 22:51, 5 April 2019 (UTC)

Wrong link

In wikipedia in Spanish we have Superman (Max Fleischer). It is the same subject than Superman (1940s cartoons). But I can not link them. I think that the element Q39081974 must be deleted. I do not know how to do it.--Chamarasca (talk) 17:05, 6 April 2019 (UTC)

I've merged the two items (Wikidata:Merge) Andrew Gray (talk) 17:20, 6 April 2019 (UTC)

Most "exact" property

What will be the most exact property for an elected member to the Norwegian Constituent Assembly (Q2305249). I have member of (P463), position held (P39), participant of (P1344) or may be even subject has role (P2868)?. The point here will be that the person was elected and not just qualified. Pmt (talk) 14:52, 6 April 2019 (UTC)

I would normally recommend using position held (P39) with a value for "member of the Norwegian Constituent Assembly" (you may need to create a new item for this). I'm not sure what you mean by elected vs qualified, though - is qualified = appointed? If so, a qualifier on the P39 entry might be the best way to express this. Andrew Gray (talk) 17:25, 6 April 2019 (UTC)
@Andrew Gray: Thank you! (by qualified I ment appointed yes) Pmt (talk) 18:10, 6 April 2019 (UTC)
I think this is best, then. A elected in (P2715) qualifier of "appointed" might work for those, pending anything better (I've not done much with mixtures of appointed/elected people before) Andrew Gray (talk) 20:08, 6 April 2019 (UTC)
For cases where someone was appointed rather than elected, I think elected in (P2715):novalue coupled with appointed by (P748) would be better. For Norwegian Constituent Assembly (Q2305249), we can use elected in (P2715):Norwegian Constituent Assembly election, February–August 1814 (Q1772809). --Oravrattas (talk) 06:23, 7 April 2019 (UTC)


Could somebody review Talk:Q48685#Judaism? The way it is now it's just bogus information. Reality was far more complex than that. Tgeorgescu (talk) 21:33, 6 April 2019 (UTC)

reservoir filled by other reservoirs

Hello. If a reservoir (Q131681) is filled by a river, we can add that reservoir item->lake inflows (P200)->river item. But I don't know what to do in the case that the reservoir is not filled by a river. I need someone to explain me how to show that a reservoir (Achna Reservoir (Q22991968)) is filled through watercourse (Q355304) that connect the reservoir with two other reservoirs (Kouris Reservoir (Q6434705) and Kalavasos Reservoir (Q62607262)). The reservoir is part of the Southern Conveyor Project, which carries water from the southwest side of Cyprus to the southeast part of the island, over a distance of 120 km. Xaris333 (talk) 16:47, 6 April 2019 (UTC)

Seems to me like the inflow just happens to be the outflow of another reservoir. Not entirely sure how that's problematic. Circeus (talk) 03:19, 7 April 2019 (UTC)
Not exactly. The water department choose when to the first reservoir needs water and send to it from the second and third reservoir. Xaris333 (talk) 12:33, 7 April 2019 (UTC)

Templates and Categories - A question.

What is the inverse of (P1696) the property known as Wikimedia template that populates this category (P4329)?

I was looking at Q6476774 which contained only a single statement. The statement had an issue, and it was "topic's main category (P910) is Category:Wikipedians by language (Q4655215)." The issue with that statement is that it did not satisfy the distinct values constraint (Q21502410) of P910. I went to Category:Wikipedians by language and saw it had the statement "P4329 is Template:Babelcat (Q60674924)." Now, Template:Babelcat had only two statements, the second being: "P910 is Category:Wikipedians by language."

I want to fix Project:Babel, but this issue is going to bother me before all else. My preference would be to propose such an inverse Wikidata property (Q18616576) for P4329 and an inverse constraint (Q21510855) to it. I believe I have seen this be a thing before, but I can't remember where. Would I be using corresponding template (P2667) for this? –MJLTalk 00:26, 7 April 2019 (UTC)

  • Aren't these user categories populated by a MediaWiki feature, not a Wikidata property value? --- Jura 16:11, 7 April 2019 (UTC) seems to be about something else. --- Jura 16:15, 7 April 2019 (UTC)

Move members of Writer in Prevezanika Chronika to Contributed to creative work, Prevezanika Chronika

Hello to all expert users of Wikidata, How easily (i.e. automatically) can subjects which are currently in the set of writer in Prevezanika Chronika (Q42405008), currently a subclass of writer (Q36180), be moved to the contributed to creative work (P3919) with qualifier Prevezanika Chronika (Q16329615) ???
Actia Nicopolis (talk) 18:49, 8 April 2019 (UTC)

One possible way: The result of this query:
SELECT ?item { ?item wdt:P106 wd:Q42405008 }
Try it!
can be downloaded and exported to QuickStatements, which can add new statements with qualifiers and remove obsolete ones. Matěj Suchánek (talk) 15:54, 11 April 2019 (UTC)

Dear Matěj Suchánek, thank you very much for your help and suggestion.
Well, it took me some hour of trying different things with no positive result.
I see that you are a programmer and I realised that it is not me being a fool and not understand.
The English say, it is all Greek to me.
We, Greeks, say, it is Chinese to me...
I now, come to my problem and request to you.
I tried the query. I got 180 results. I downloaded a csv file. But after that the Quickstatements is beyond my capasity !!!!!!
Could you, please, delete from these 180 persons the (statement Q42405008) from (occupation P106)
and add to all these 180 persons (statement Q16329615 ) into (contributed to creative work P3919)

I would very much appreciate it.
Which city of the Czech Republic do you live?

Best wishes, Nikos Actia Nicopolis (talk) 18:04, 13 April 2019 (UTC)

✓ Done (with a small mistake made in haste). The goal was something like Q### P3931 Q16329615 and -Q### P106 Q42405008. My home town is Prague (Q1085). Matěj Suchánek (talk) 18:39, 13 April 2019 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 18:39, 13 April 2019 (UTC)

Interested in Q26126631

Which country is called "United Kingdam"? --Liuxinyu970226 (talk) 23:37, 12 April 2019 (UTC)

There is also ja:Template:Country_alias_United_Kingdom. I think the misspelling could be deleted in ja:wp and then here, but this has to be discussed in ja:wp. 07:32, 13 April 2019 (UTC)
Raised at ja:Wikipedia:Help for Non-Japanese Speakers#Template:Country alias United Kingdam. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:23, 13 April 2019 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:29, 13 April 2019 (UTC)

Reasonator is down

Hoi, it is important that our tools are functional. Reasonator is one of the most useful tools to appreciate the data you are working on. Thanks, GerardM (talk) 06:12, 14 April 2019 (UTC)

Looks like the same situation that’s also affecting QuickCategories and Ordia, which I’ve reported as T220912. --Lucas Werkmeister (talk) 15:31, 14 April 2019 (UTC)
@GerardM: it’s back now. --Lucas Werkmeister (talk) 16:57, 14 April 2019 (UTC)
I noticed, was about the inform the chat about it.. Thank you! GerardM (talk) 17:08, 14 April 2019 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 18:00, 15 April 2019 (UTC)

Duplicats? Italian composer, guitarist Charles Michael Alexis Sola and Italian composer, guitarist Carlo Michele Alessio Sola

I think Charles Michael Alexis Sola and Carlo Michele Alessio Sola are probably duplicates, but two separate articles on Catalan wikipedia are blocking a merge. If someone who parla català could have a look, that would be cool. Moebeus (talk) 13:00, 16 April 2019 (UTC)

@Moebeus: The same contributor, Vulcano, created both of the Catalan Wikipedia articles; the former in 2016 and the latter in 2010. The language used is almost the same, the dates of birth and death are the same, the links are the same. I don't speak Catalan but I think it would be appropriate to redirect the newer article (Charles) to the older article (Carlo). Jc86035 (talk) 13:45, 16 April 2019 (UTC)
@Moebeus: I've redirected the newer article to the older one and merged the items. However, the item now has two different death dates; the English Wikipedia article uses January and the Catalan Wikipedia article uses February. I don't know which one is correct. Jc86035 (talk) 10:26, 17 April 2019 (UTC)
@Jc86035: Thank you for all the help, super-thorough as always. I wouldn't know about the death date, but I think the way you left it with the comment on the Discussion page is plenty enough for others to pick up from if they know something. Thanks! Moebeus (talk) 11:24, 17 April 2019 (UTC)
This section was archived on a request by: Moebeus (talk) 11:24, 17 April 2019 (UTC)

Barry Barish

For Barry Barish there are two items that need to be merged. The problem is at Commons.. No clue what to do. Thanks, GerardM (talk) 12:02, 19 April 2019 (UTC)

"The problem is at Commons" Wrong venue. Try c:Commons:Village pump. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:10, 19 April 2019 (UTC)
To be fair, there's still work to do here. I've emptied commons:Category:Barry C. Barish into commons:Category:Barry Barish and nominated the former for deletion. We might as well get rid of Q60439408 no matter how that turns out. --RexxS (talk) 15:16, 19 April 2019 (UTC)
I merged the two regular items. ArthurPSmith (talk) 16:03, 19 April 2019 (UTC)
"there's still work to do here" Indeed. But not until the matter is (was) resolved at Commons. Thanks for doing that. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:18, 19 April 2019 (UTC)
Thank you all. GerardM (talk) 06:39, 20 April 2019 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 18:36, 20 April 2019 (UTC)

Some discussions about disputed territories

Emergency discussion about @C933103:'s Golan Heights (Q83210) edits needed

Look here, do we know what he did here? This user simply removed country (P17) Syria (Q858) which, although I undid him, therefore and thereafter means that:

  1. C933103 is a Judaic, that supports Israel's illegal seize of this area, and oppose any claiming from Syria government;
  2. C933103 withdrawn NPOV and support Trumpism (Q31838499), meant that this user also agreed Trump's recognization of so-called "this is an Israel territory"
  3. C933103 is starting an Intercontinental Edit War, just based on bad Arabic-Israel relations
  4. C933103 is an Anti-Muslim-ism.

Do we agree his these behaviors??? -- 09:23, 2 April 2019 (UTC)

What? Golan Heights (Q83210) still maintain territory claimed by (P1336) by Syria (Q858) after my edits, and my edits was only a partial revert to another editors' edit a few hours ago that claimed Syria (Q858) have object has role (P3831) of de jure (Q132555) on Golan Heights (Q83210) which I don't really think that's the case (at least according to my understanding which have a chance of not aligned to fact), and thus I have raised the question to the editor that committed the previous edit on the talk page while making a revert on the edit. C933103 (talk) 09:36, 2 April 2019 (UTC)
Wrong, after Your edits, the P17 Syria value was unfairly lost temporary. --2409:8902:9300:2967:8520:E396:FDC6:D58D 09:43, 2 April 2019 (UTC)
It looks like that we have troubles about ethnic conflicts here, so helps from administrators that can speak Arabic and Hebrew languages needed: @علاء, باسم, Deror_avi, יונה_בנדלאק:. --Liuxinyu970226 (talk) 10:13, 2 April 2019 (UTC)
This verion looked correct. maybe lebanon need to be removed from the territory claimed by (P1336). they have a claim about very small part of the golan (Shebaa farms (Q1133272) only). so it probably better to put the claim in Q1133272 and not on the Q83210. - yona b (talk) 10:50, 2 April 2019 (UTC)
i agree to what @יונה בנדלאק: saysباسم (talk) 11:10, 2 April 2019 (UTC)
lebanon need to be removed from the territory claimed by (P1336). they have a claim to (Shebaa farms (Q1133272) which is not a part of the Golan. Deror avi (talk) 19:51, 2 April 2019 (UTC)
IIRC that is again a bit like the Kashmir situation below, where one side claim a place is part of the larger region while the other side doesn't recognize such position. C933103 (talk) 03:23, 3 April 2019 (UTC)
@C933103: I'm just wondering why this area is like Kashmir, is PR China also claiming Golan? --Liuxinyu970226 (talk) 03:48, 3 April 2019 (UTC)
Because IIRC Lebanon (Q822) think Shebaa farms (Q1133272) is not part of Golan Heights (Q83210) but Israel (Q801) think it is? Not sure about UN/Syria's position. C933103 (talk) 04:51, 3 April 2019 (UTC)
No, the Israel government also don't think this farm is a part of Golan, both are different disputed areas. -- 08:34, 3 April 2019 (UTC)
@ Hmm, you're probably meaning "isn't", isn't there a typo in your sentence? --Liuxinyu970226 (talk) 15:59, 3 April 2019 (UTC)
@Liuxinyu970226: Doesn't appear so to me. "don't think this farm is a part of Golan" == "think this farm is not a part of Golan" == "think this farm isn't a part of Golan". All equivalent. Are you saying the matter is otherwise? - Jmabel (talk) 22:18, 3 April 2019 (UTC)
While figuring out how to deal with the subject of the Golan Heights is important, I think it's quite clear that the OP/IP should be immediately permanently banned, unless the content of the post was somehow a result of some serious mistranslation. --Yair rand (talk) 05:41, 3 April 2019 (UTC)
Sure, I rolled back to the normal status of Golan Heights (Q83210), so that Syria is shown normally. -- 08:34, 3 April 2019 (UTC)
The user who started the discussion seems to be using dynamic IP that I don't think banning can make any difference. C933103 (talk) 13:21, 3 April 2019 (UTC)

de jure (Q132555)?

Someone keeping adding de jure (Q132555) back to the page as Syria's role on the territory but they continually refusing to provide any source for it. Actually, what is the applicability of this role when it come to a conflicted territory? According to which law was Syria (Q858) having de jure (Q132555) role on Golan Heights (Q83210)? C933103 (talk) 03:37, 3 April 2019 (UTC)

Yes, please read their constitution: s:ar:دستور سوريا. -- 08:38, 3 April 2019 (UTC)
Going by this logic, Israel would also have de jure sovereignty on Gaza strip or West Bank, China would also have de jure sovereignty on Taiwan and vice versa Taiwan would also have de jure sovereignty on China. I am not going to say whether it is a right thing to do or not to add the "de jure" role to all these countries involved in conflicts, but I want other wikidata users participating in the discussion consider what is the right time for the "de jure" role to be applied onto countries and with these examples listed out in mind. C933103 (talk) 13:18, 3 April 2019 (UTC)

Pictogram voting comment.svg Comment Can we please continue discussions at Wikidata:Property proposal/Recognition, recognized by, not recognized by, jurisdiction status? It looks like by proposal clauses, de jure and de facto can be used for disputed territories, but with specified properties instead of sourcing circumstances (P1480), subject has role (P2868), object has role (P3831) (3 most confusing properties of the entire Wikidata world). --Liuxinyu970226 (talk) 04:47, 11 April 2019 (UTC)

I can agree, but how about C933103? Can that user agree? -- 03:16, 17 April 2019 (UTC)

Kashmir (Q43100)

@Liuxinyu970226: In your recent edits to the Kashmir (Q43100) page, you added China as a country that own Kashmir and also an English description saying that "former princely state, now a conflict territory between India and Pakistan (and in rarely seen cases, the People's Republic of China". However there are a few problems with the edit.

  1. The official position of China is that they don't recognize the part they control as part of the region of Kashmir, given the view point should China still be added as a P17 to the Kashmir region?
  2. I don't think China claim any part of the area that were actually part of the former princely state that the new description seem to imply?

C933103 (talk) 09:47, 2 April 2019 (UTC)

"now administered by three countries: India, Pakistan, and China." is wrong in your opinion? --Liuxinyu970226 (talk) 10:10, 2 April 2019 (UTC)
Read the entire sentence.... C933103 (talk) 10:30, 2 April 2019 (UTC)
@C933103: "The official position of China is that they don't recognize the part they control as part of the region of Kashmir" so that Kashmir is not a PR China territory? So there's no problem about China at all? --Liuxinyu970226 (talk) 10:37, 2 April 2019 (UTC)
The best area to discuss your second point, TBH, is Talk:Q230830, you should talk this problem here by creating it. --Liuxinyu970226 (talk) 10:37, 2 April 2019 (UTC)
That is why you would never hear Chinese media say something along the line of "Chinese controlled part of Kashmir". But then the problem is that India see those Chinese-controlled area as part of Kashmir. The situation is actually a bit similar to the Southern Kuril Islands where Japan doesn't recognize those islands as part of Kuril Islands instead they see them as islands associated to the Hokkaido.C933103 (talk) 11:27, 2 April 2019 (UTC)
@Liuxinyu970226, C933103: This seems like the lack of informations about Aksai Chin (Q230830), hence I re-introduced India on P17 of that item, and added both PRC and India as P1336, that said, for such disputed territories that controlled areas on both sides are not fully, we need P17 values same as P1336. -- 01:02, 3 April 2019 (UTC)
Thx. --Liuxinyu970226 (talk) 04:10, 3 April 2019 (UTC)


Anything Google+ is now in essence dead, and Google+ ID (P2847) could be deleted wholesale. I'd like to remove the Google+ ID from the self-proclaimed G+ murderess Emma Blackery (Q15994935).[2] What's the plan? – 00:12, 3 April 2019 (UTC)

Google+ pages are full of useful information, and the internet archive has done extensive archiving of the public pages. I recommend against removing the IDs. --Yair rand (talk) 01:40, 3 April 2019 (UTC)
I'm also strongly against the removal of Google+ ID (P2847): we don't delete items about people when they die... -- Brasig (talk) 06:26, 3 April 2019 (UTC)
Enterprise pages are still functioning C933103 (talk) 04:44, 3 April 2019 (UTC)
Why would we delete the property just because the resource is down? Is it not worth archiving? —Justin (koavf)TCM 06:52, 3 April 2019 (UTC)
@Koavf, Yair rand, Brasig, C933103: Google+ is now dead (or most of it is). However, a lot of the content has been archived over the years (and more recently), so I've changed the formatter URL to point to the Wayback Machine (Q648266). Jc86035 (talk) 15:03, 4 April 2019 (UTC)
Barely better than nothing,[3] many photos and videos didn't make it, unlisted collections are of course lost, and ordinary collections not listed on the overview page are unreachable via "view all". If you keep this Gcemetry it shouldn't be an "ID" anymore, maybe "Former Google+ page" would be clearer. – 23:34, 10 April 2019 (UTC)

URL shortener for the Wikimedia projects will be available on April 11th

Hello all,

Having a service providing short links exclusively for the Wikimedia projects is a community request that came up regularly on Phabricator or in community discussions.

After a common work of developers from the Wikimedia Foundation and Wikimedia Germany, we are now able to provide such a feature, it will be enabled on April 11th on Meta.

What is the URL Shortener doing?

The Wikimedia URL Shortener is a feature that allows you to create short URLs for any page on projects hosted by the Wikimedia Foundation, in order to reuse them elsewhere, for example on social networks or on wikis.

The feature can be accessed from Meta wiki on the special page m:Special:URLShortener (will be enabled on April 11th). On this page, you will be able to enter any web address from a service hosted by the Wikimedia Foundation, to generate a short URL, and to copy it and reuse it anywhere.

The format of the URL is followed by a string of letters and numbers. You can already test an example: redirects to

What are the limitations and security measures?

In order to assure the security of the links, and to avoid shortlinks pointing to external or dangerous websites, the URL shortener is restricted to services hosted by the Wikimedia Foundation. This includes for example: all Wikimedia projects, Meta, Mediawiki, the Wikidata Query Service, Phabricator. (see the full list here)

In order to avoid abuse of the tool, there is a rate limit: logged-in users can create up to 50 links every 2 minutes, and the IPs are limited to 10 creations per 2 minutes.

Where will this feature be available?

In order to enforce the rate limit described above, the page Special:URLShortener will only be enabled on Meta. You can of course create links or redirects to this page from your home wiki.

The next step we’re working on is to integrate the feature directly in the interface of the Wikidata Query Service, where is currently used to generate short links for the results of the queries. For now, you will have to copy and paste the link of your query in the Meta page.

Documentation and requests

Thanks a lot to all the developers and volunteers who helped moving forward with this feature, and making it available today for everyone in the Wikimedia projects! Lea Lacroix (WMDE) (talk) 11:44, 3 April 2019 (UTC)

That's great news! Thank you, Lea, for bring it to our attention. I have taken the liberty of copying much of your post to Wikidata:URLShortener, so that we can use it as the basis of Wikidata-specific documentation. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:03, 3 April 2019 (UTC)
Also, please see Wikidata:Property proposal/WMF short URL. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:46, 3 April 2019 (UTC)
@Lea Lacroix (WMDE): Any chance we can get /Q1, /Q2, /Q5, and maybe some others to point to their respective Wikidata items? --Yair rand (talk) 02:12, 4 April 2019 (UTC)
We can't reserve URLs in advance, they will be created incrementally :) Lea Lacroix (WMDE) (talk) 08:22, 4 April 2019 (UTC)
@Lea Lacroix (WMDE): Will the first users be restricted from creating links with one-character strings? It seems to me that it would be appropriate to reserve those to allow them to be used for future Wikimedia projects, instead of having them link to things like random Phabricator comments; e.g. If a lot of them already exist, deleting them now (since stewards can do that) would presumably be less disruptive than deleting them later. Jc86035 (talk) 15:00, 4 April 2019 (UTC)
Hello Jc86035, thanks for your comment. The URLs with one character have been created on purpose before the deployment of the feature. Some of them may contain easter eggs ;) After the deployment, URLs will continue to be created following an incrementing pattern. It won't be possible to choose a particular URL. Lea Lacroix (WMDE) (talk) 15:35, 4 April 2019 (UTC)
@Lea Lacroix (WMDE): Just for clarification: The concern won't matter as much if the first user-generated URL will have a string longer than one character. However, my main concern was that if at some point the WMF wanted to repurpose e.g. (which points to w:en:Alan Turing) for linking to WikiApiary or a new Wikimedia project, then any links to the short URL would end up being broken, so (in my opinion, at least) it would be better to remove the Easter eggs sooner instead of breaking any accumulated inbound links after several years. Jc86035 (talk) 12:37, 8 April 2019 (UTC)
I've filed a Phabricator ticket. Jc86035 (talk) 17:21, 8 April 2019 (UTC)
^.^b But not yet enabled on Meta, a nice test case could be this recently (2019) discussed here. – 08:13, 11 April 2019 (UTC) WFM, nice, this SHOULD be implicitly class="plainlinks" and SHOULD NOT require a captcha. – 21:42, 12 April 2019 (UTC)

Hello all,

The feature is now live at m:Special:UrlShortener :)

I already filled a ticket about the limit of characters being too small for some queries, if you spot some bugs or have ideas for improvement, feel free to create tickets as well.

Enjoy shortening! Lea Lacroix (WMDE) (talk) 14:45, 11 April 2019 (UTC)

Léa, can you please update File:Screenshot URL shortener.png so that it shows an URL which the feature actually accepts? If I understand correctly, * is not supported, thus the image is a bit misleading. Thanks, MisterSynergy (talk) 17:00, 11 April 2019 (UTC)
Done, thanks for noticing! Lea Lacroix (WMDE) (talk) 06:53, 12 April 2019 (UTC) .. redirects to obscurification (L45083) --- Jura 18:52, 14 April 2019 (UTC)

Variant spellings

I've created carabiner (L44813) (English noun). How would I indicate the alternate spelling "karabiner" (for the generic form of English)? Jc86035 (talk) 16:20, 4 April 2019 (UTC)

Look like a new Wikidata:Property proposal is needed (there's quite some need in the area of lexicographcial data). Circeus (talk) 16:42, 4 April 2019 (UTC)
@Circeus: Would a property even be appropriate in this situation? realize/realise (L5412) seems to be a good example for this, but -ize/-ise verbs are conveniently only used with -ise in British English, whereas "karabiner" doesn't have that sort of clear usage difference. (The software prevents the addition of more than one spelling variant with the same language code.) Jc86035 (talk) 17:05, 4 April 2019 (UTC)
@Jc86035: But as Mahir points out, the issue is this is not a regional spelling. Things like foreign transliteration also cause this issue. In French, kosher has over a half-dozen recorded alternate spellings (the switch in transliteration styles for Chinese proper names also comes to mind). Circeus (talk) 17:46, 4 April 2019 (UTC)
As a note, it is this problem of adding multiple equally acceptable variant spellings with no specific regional or other notable distinction between them which hinders me (and possibly others) from adding Bengali lexemes more in earnest. Mahir256 (talk) 17:20, 4 April 2019 (UTC)
@Mahir256, Circeus, ArthurPSmith: Is this a bug, or just a limitation? I'm not totally familiar with the Lexeme structure, but it seems to me that either adding additional Forms would be appropriate in these cases, or the software should allow multiple representations with the same language code for each form (and the same for lemmas). Jc86035 (talk) 10:51, 5 April 2019 (UTC)
Would a possible approach be to add alternate spellings as additional forms with a P31 or other property to indicate they are rare/alternate? I've done something like this for plurals where there are several alternative pluralizations of the same singular English word. ArthurPSmith (talk) 20:13, 4 April 2019 (UTC)
But that requires choosing one spelling to be primary over the others, because it's the one appearing in the lemma. —Rua (mew) 21:39, 6 April 2019 (UTC)
I agree with ArthurPSmith, adding variant as forms is the best way to go, I added them (more statements can be added, for instance citations specific to these forms). @Rua: there is two possibilities: one lemma is clearly the main one ("carabiner" here), or none is and then you add both as main lemma (like in realize/realise (L5412)). Cheers, VIGNERON (talk) 21:46, 7 April 2019 (UTC)
@VIGNERON, Mahir256: Perhaps a property like "alternative form to" (linking from one form to another form) would be appropriate? If multiple forms are equally popular then the property could be used in all directions. Jc86035 (talk) 12:42, 8 April 2019 (UTC)
@Jc86035: isn't that already the case? If several forms have the same grammatical features, then they de facto are equivalent. True, it's a bit implicit but I'm not sure if we need to store a data to say it explicitly (and working on Breton lexemes with a sh*tload of equivalent forms - at least 2 and up to 20 for per lexeme -, it would be a lot of work to add this statements and I must say I don't really see the need right now). Cheers, VIGNERON (talk) 13:56, 8 April 2019 (UTC)

Finding items missing label and/or description in a particular language?

Is there a way to get a list of all items that do not have a label in a particular language? And the same for description? It would be useful to know for me (Dutch speaker) but also others who can contribute by writing labels for their language. —Rua (mew) 21:37, 6 April 2019 (UTC)

@Rua: Here you go! (You can change to missing descriptions instead of missing labels via a drop-down and adjust both the "for languages" and "with labels in" boxes to something else if desired.) Mahir256 (talk) 22:04, 6 April 2019 (UTC)
@Mahir256: Is it supposed to return false positives (entries that do have a Dutch label already) sometimes? —Rua (mew) 19:45, 7 April 2019 (UTC)
@Rua: Not sure, since I don't regularly use the tool myself, but I would guess that any delays in updating the Query Service would contribute to the presence of false positives on occasion. Mahir256 (talk) 19:54, 7 April 2019 (UTC)

How are 'suggested' items ranked?

When I go to add a new property, a list of 'suggested' items for that property appears: e.g. typing "male" into sex or gender (P21) brings up options such as male organism (Q44148), male (Q6581097), Mâle (Q1347276), etc., apparently always in that same order. What is the method these are ranked, and can it be tweaked to offer smarter suggestions first? Like, if instance = human, can Q6581097 (or Q6581072) be prioritized over the non-human equivalent? Similarly, when trying to add a statement of "child (P40)" to a human, the top suggestion after typing "child" is father (P22) ("Father (child of)"). If possible, re-weighting the suggestions would make editing just a little more convenient. -Animalparty (talk) 23:10, 6 April 2019 (UTC)

  • Values for P21 depend on the one-of-constraint on the property. Suggestions of additional statements were debated at Wikidata:Suggester_ranking_input. --- Jura 13:03, 7 April 2019 (UTC)
    • Additional info: For the one-of-constraint to work you currently still need to enable the beta feature for it in your preferences. We are working on taking it out of beta features and making it available for everyone. --Lydia Pintscher (WMDE) (talk) 17:55, 7 April 2019 (UTC)

Please revive Q62002130

He is a famous prodigy and developer of Aheui (Q55731562)  – The preceding unsigned comment was added by [[User:{{{1}}}|{{{1}}}]] ([[User talk:{{{1}}}|talk]] • [[Special:Contributions/{{{1}}}|contribs]]).

  • @Bencemac: as deleting admin. --- Jura 17:14, 7 April 2019 (UTC)

Because he is the devoloper of Aheui(Aheui (Q55731562)), it meets third criteria of the notability policy: fulfills some structural need, for example: it is needed to make statements made in other items more useful.  – The preceding unsigned comment was added by 2001:2d8:21b:b5ed::1057:f0b0 (talk • contribs) at 8 April 2019 (UTC).

w:ko:아희 is spam tagged as (Google translation): "The source of this document is heavily dependent on the primary data. Please edit this document and mark the second and third materials as footnotes. (October 29, 2017)". No references here for Aheui (Q55731562). – 08:36, 11 April 2019 (UTC)

Vandalism dashboard

Sorry to say this listing doesn't do what it says on the tin. Very few of the edits listed are damaging in any way and even fewer could be called "vandalism". I just made two minor edits myself, and both were listed as "vandalism": one was correcting a spelling, the other was reverting a damaging edit. It looks like the ORES system is delivering far too many false positives. Bhunacat10 (talk) 11:35, 7 April 2019 (UTC)

Yes. We are working on improving that but we need people to help us with judging a set of edits we can train the system on. If you'd like to help please go to and do a set. The sooner we have the campaign completed the sooner we can run a new training and get better automatic judgements here. --Lydia Pintscher (WMDE) (talk) 17:58, 7 April 2019 (UTC)
@Bhunacat10: just to clarify – what do you think it says on the tin? Because I don’t think this is intended as a dashboard of vandalism – it shows you all unpatrolled changes affecting a certain language, some of which might be vandalism. If it gave you a different impression then we might need to clarify this somewhere. --Lucas Werkmeister (WMDE) (talk) 09:38, 8 April 2019 (UTC)
Lucas Werkmeister I'd expect something headed "Vandalism dashboard" to be displaying likely instances of vandalism. What drew my attention was this: as you are probably aware the English Wikipedia is embarking on a project to add locally defined short descriptions to all its >5M articles, in the end intending no longer to draw these from the descriptions in Wikidata. One of the principal justifications is a perception that vandalism to Wikidata descriptions is rife and often undetected. Now in the lengthy discussions surrounding this initiative, the "vandalism dashboard" is frequently linked, in a spirit of "just look here to see some of the terrible vandalism on Wikidata!"
So I look and I see very few such instances. I see a display including ORES scores and colour-coded according to those scores. But even among those edits given an ORES score close to 1, most of them look fair-to-middling. A few are damaging through what looks like ignorance; very few indeed appear to show malicious intent. In fact ORES seems to be scoring simply on the basis of whether the contributor is an IP, a new account (new to Wikidata), or an account with many edits; and not really assessing the merits of the edit itself.
On enwiki we're used to ClueBot NG, an AI algorithm built up over many years that in general does a good job of spotting vandalism and often auto-reverting it. So when I see this dashboard I'm led to think "how can they hope to detect vandalism if it means searching for needles in this haystack?" Anyway I'll set to and work on the training set as Lydia asks, and hope that the improved system when implemented will help Wikidata get on to vandalism quickly and satisfy sister projects that that is the case. Bhunacat10 (talk) 16:58, 8 April 2019 (UTC)
Typical. First they invent Wikidata to offload data onto, then they decide they want to keep data to themselves after all... —Rua (mew) 17:09, 8 April 2019 (UTC)
"the English Wikipedia is embarking on a project to add locally defined short descriptions to all its >5M articles, in the end intending no longer to draw these from the descriptions in Wikidata." Mind-boggling. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:14, 9 April 2019 (UTC)
Then again, de.Wikipedia stores its authority control ("Normdaten") values locally, too: [4]. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:16, 9 April 2019 (UTC)

Undu some merging of Q1136860 into Q1082861 (by Rua)

Rua and I disagree about the (incl. the historical) content of Q1136860 and Q1082861 and that is why I write this chat/discussion. (Low frequency has it's own item Q17156810)

Quote fra Rua: "Langbølger (Q1082861) is about the longwave broadcasting band, and has always been about that before I even made any edits, because it always had "radio broadcast band" in the description. So Q1136860 is a duplicate."

My judgement is that:

  • Q1082861 should be about the longwaves as defined by ITU, as before
  • Q1136860 should be about the longwave (broadcasting) band as defined by ITU, as before.

The elder history of Q1082861 shows that it is about the longwaves (as defined by ITU; 10 km - 1 km; correspond to 30-300 kHz). Contents: GND=4166718-9 (Source: Imported from english Wikipedia) refers to Kilometerwelle, Radiofrequenzbereich. commons:Category:Longwaves. Kilometerwelle found here; de:Frequenzband#Übersicht.

The elder history of Q1136860 shows that it is/was about the longwave (broadcasting) band (as defined by ITU; refers to the range center carrier wave to center carrier wave: 148.5-283.5 KHz, interval with channel AM-modulation 153-279 kHz; 9kHz (+-4.5kHz) channel bandwidth). (ought to have commons:Category:Longwave radio broadcasting).



--Glenn (talk) 19:37, 7 April 2019 (UTC)

@Glenn: But then, you have to try explaining what's Japanese "長波" (by separating this link as an own item and add instance of (P31) Wikipedia article covering multiple topics (Q21484471)?!), as it's a mixed article of all items you mentioned. --Liuxinyu970226 (talk) 05:07, 8 April 2019 (UTC)
Quote: "...LF(Low Frequency)またはLW(Longwave, Long Wave))とは、30 - 300kHzの周波数の電波をいう[1][2]。波長は1 - 10km..."
It mentions first: 30 - 300kHz and 1 - 10km. "长波" means longwave according to so it should be in Q1082861. --Glenn (talk) 15:19, 8 April 2019 (UTC)
We should go by the primary topic of the article, not try to match the English translation. The article defines itself as being about the range 30-300 kHz, so it's clear which item it belongs to. —Rua (mew) 15:52, 8 April 2019 (UTC)
@Okkn: Is this true? That Japanese article matches low frequency more than langwave? -- 01:49, 9 April 2019 (UTC)
Sorry, I don't understand the difference between "longwave" and "low frequency" well, but the Japanese article matches the definition of "LF" (30–300 kHz) in w:Radio spectrum. "長" means "long", and "波" means "wave". --Okkn (talk) 05:25, 11 April 2019 (UTC)

New RfC: Countries, subdivisions, and disputed territories

I've started a new RfC at Wikidata:Requests for comment/Countries, subdivisions, and disputed territories. Comments would be most welcome. Thanks. --Yair rand (talk) 07:25, 8 April 2019 (UTC)

Read-only mode for up to 30 minutes on 11 April

10:56, 8 April 2019 (UTC)

New status of constraint: "suggestion"

Hello all,

As you may know, we currently have two levels of property constraints: mandatory constraints and "normal" (non-mandatory) ones. In order to allow more flexibility and subtlety in constraints definition, we are going to introduce a new constraint level, suggestion. This way, editors can distinguish the really crucial constraint violations from the ones that only suggest additional edits that would be nice to make.

When setting up a constraint rule, you could define it as a suggestion by adding a qualifier to the constraint definition statement with the property constraint status (P2316) and the value suggestion constraint (Q62026391).

The suggestion constraint would get a specific icon to make it clearly different from the mandatory and non-mandatory constraints. The current icons look like this:

Screenshot icons for Wikidata constraints.png

The feature will be enabled on May 6th. Starting from this date, you will be able to define constraints as suggestions and they will be displayed as such for all logged-in users.

Please note that this is the first roll-out of the feature, some issues may appear. You are also very welcome to give feedback, if some things are not working as you wish they would, we can still fix them. In that case, feel free to ping me or write a comment in the related ticket.

Thanks to the people who made this request, this feature was suggested during one of the “pink pony session” at Wikimania and we hope that you will enjoy it!

Cheers, Lea Lacroix (WMDE) (talk) 12:17, 8 April 2019 (UTC)

Great news! I'm sure going to add a lot of these new suggestions. Are you also going to make use of this in the property suggestion interface? Multichill (talk) 18:53, 8 April 2019 (UTC)
Hello Multichill, can you tell me more about what you have in mind? Do you mean the properties that are suggested when creating a new statement? Lea Lacroix (WMDE) (talk) 08:58, 9 April 2019 (UTC)
@Lea Lacroix (WMDE): Exactly. If an item for example has J. Paul Getty Museum object ID (P2582) and that property contains a suggested constraint for main subject (P921), it should offer main subject (P921) as a suggested statement on The Shade of Samuel Invoked by Saul (Q20179282).
The negative form would also be nice to prevent certain properties for example in this case to prevent the suggestion of described at URL (P973). Multichill (talk) 11:11, 11 April 2019 (UTC)
Thanks for the details. Update: no, it's not something we had planned so far, but it looks interesting. Feel free to fill a Phabricator tickets with ideas and examples, and to ping me in there :) Lea Lacroix (WMDE) (talk) 13:35, 11 April 2019 (UTC)
Phab:T220875. Multichill (talk) 09:25, 13 April 2019 (UTC)
Wondering if this will be applied to obsolete Wikidata property (Q18644427) or not. -- 05:32, 9 April 2019 (UTC)
Hello, what do you mean? I'm not sure I understand what you would like to have. Lea Lacroix (WMDE) (talk) 08:58, 9 April 2019 (UTC)
Maybe slightly off topic, but why do IPs like me don't get those apparently helpful suggestions? – 08:47, 11 April 2019 (UTC)
Hello, and thanks for raising this question. Providing the constraints for non logged-in users is actually part of our next steps! We wanted to enable the "suggestion" level first, in order to make the whole constraint system more stable and less scary for newcomers. But soon after it's enabled, we'll come back to the community and ask them if they agree with displaying constraint notifications for IPs. Lea Lacroix (WMDE) (talk) 13:35, 11 April 2019 (UTC)

Fair use and non-free content

Wikidata currently does not have any policies relating to copyright. No RfCs have been held on the matter, except Wikidata:Requests for comment/Exclusion of pages in the file namespace, which resulted in a consensus to disallow local file uploads.

Should Wikidata have policy or guidance on how to deal with fair use content (or, strictly, copyrighted content covered by the fair use doctrine of US copyright law)?

While Wikidata is licensed CC0, there are already many statements which would have to be removed under a strict prohibition of non-free content. For example, Aurangabad district (Q43086) uses quote (P1683) in references to quote from Census of India 2011 (Bihar): Aurangabad District Village and Town Directory (Q55971592) (PDF), which contains prose which was not released into the public domain. (In India, government works are copyrighted for sixty years plus the remainder of the final calendar year.) quote (P1683) is used on over 48,000 items, although not all uses contain copyrighted or copyrightable text.

Several other properties which could potentially be significantly affected include inscription (P1684), first line (P1922), last line (P3132), P5482 (P5482), usage example (P5831), award rationale (P6208) and musical quotation or excerpt (P6670).

Since in practice Wikidata already contains fair use content, should Wikidata give a latitude similar to e.g. the English Wikpedia's non-free content policy, or should a different approach be taken? Currently, whether or not this is intentional, the lack of policy on the matter means that virtually anything that could conceivably be allowed under the fair use doctrine is supposedly usable in Wikidata. (This is not an RfC, although it may be appropriate to create one.) Jc86035 (talk) 12:21, 8 April 2019 (UTC)

Is this really fair use content? I think to remember discussions saying that hen raw data are extracted from content, the license no longer apply. And for the list of text properties you mention, the quotation right apply - which is a copyright exception different of fair use and stronger than fair use. But IANAL, we might need a lawyer point of view here. Cheers, VIGNERON (talk) 14:05, 8 April 2019 (UTC)
@VIGNERON: As another non-lawyer: The "it's now data" reasoning might not be applicable here, since prose and other creative works constitute more than basic facts and data points; w:en:Threshold of originality#United States might be useful. I think fair use is relevant for quotation; w:en:Right to quote#Europe indicates that the US has effectively implemented the quotation exception as part of the fair use doctrine.
I thought the whole point of having all these discussions about whether anything here is actually legal was that there wouldn't be any lawyers at our disposal to begin with. Jc86035 (talk) 15:33, 8 April 2019 (UTC)
@Jc86035: thanks for mentioning the case Feist Publications v. Rural Telephone Services, now I remember where I read a lawyer perspective about data and citation : m:Wikilegal/Lexicographical Data (written by the WMF legal team which is not exactly "at our disposal" but if we have a specific use case, we can always ask them). As always for legal matter, it's not crystal clear but you should find references to back up your sentence « Wikidata already contains fair use content » (and even the more simple « Wikidata contains content », Wikidata is know to be containing data and not content). Cheers, VIGNERON (talk) 16:28, 8 April 2019 (UTC)
@VIGNERON: If only one example is needed, Harry Potter and the Philosopher's Stone (Q43361) (still copyrighted, of course) uses first line (P1922), so Wikidata does already contain fair use content. I think the distinction that Wikidata is "a database" is not enough to exempt all of its data/content from copyright, since it could be said that everything on the Internet is stored in databases to begin with.
The original point of this, and of my original post on the English Wikipedia, was that I was trying to figure out if it would be appropriate to use fair use content through musical quotation or excerpt (P6670). If fair use is problematic for that property, then presumably it would be problematic for other properties which quote from copyrighted works, which is why I didn't focus on the music notation properties. Jc86035 (talk) 16:44, 8 April 2019 (UTC)
« so Wikidata does already contain fair use content » but why and how? I fail to see the logic in your syllogism, I don't see how this conclusion can be inferred from theses premisses (and by the way, the item Harry Potter and the Philosopher's Stone (Q43361) is a total mess, most statement belong to the edition of the book, not the work itself, I'll try to clean this). For me « A quotation is either non eligible to copyright or fall under the quotation right, this item contains a quotation, so this item does not contain fair use and possibly not even content (depending on your view on the first premiss) ». Cheers, VIGNERON (talk) 17:20, 8 April 2019 (UTC)
@VIGNERON: My reasoning:
  • Wikidata servers are in the US, so for all intents and purposes only US copyright law applies to them.
  • In spite of Wikidata's function as a database, a text box in the CC0 parts of Wikidata is not inherently different to text on a normal MediaWiki page. Thus, text on Wikidata items is not inherently uncopyrightable due to Wikidata's medium.
  • The fair use doctrine is part of US copyright law.
  • That the Berne Convention mandates the right to quotation is not relevant because US copyright law enshrines the right to quotation through the fair use doctrine.
  • The fair use doctrine does not remove the copyright from original works; rather, it is a legal exception allowing limited use of copyrighted works without infringing on the original copyright holder's rights. As I understand it, the quotations do not become "owned" by the copyright holders of works which use those quotations. (Otherwise, if all of the lines from a copyrighted work were quoted separately in different CC0 works for valid reasons, the entirety of the original work would effectively fall out of copyright.)
  • As such, the text of Philosopher's Stone is not automatically placed into the public domain upon its quotation in a database. Rather, it can be used within another work (Wikidata, in this case) on the grounds that it is not infringing copyright, even though the original copyright still applies.
  • Therefore, the quotation of the first line of Philosopher's Stone in its Wikidata item is fair use, even if Wikidata as a whole is licensed CC0.
  • Further, American reusers of Wikidata content are not affected by any fair use content, since they may also use said fair use content under the fair use doctrine. Reuse within the EU would always be legally shaky (even if there were no fair use content) due to the recognition of database rights within the EU and the massive amount of imports of non-CC0 identifiers and other non-CC0 data into Wikidata.
Note that my understanding of this issue is mainly based on my experience at Commons. Jc86035 (talk) 17:47, 8 April 2019 (UTC)
It's also worth noting that there is a possibility that use may be de minimis, distinct from fair use. Sufficiently small or incidental copyright takings may be de minimis without requiring any further fair use analysis, or any specific exemption in copyright exemption based systems. Jheald (talk) 18:42, 8 April 2019 (UTC)
It may be hard to argue that anything in Wikidata is incidental, except when it appears in images. An alternative to accepting fair use would be to say that Wikidata is supposed to be purely CC0 or public domain, and copyrighted data can be deleted if desired, just like fixing any other mistake. Ghouston (talk) 01:45, 9 April 2019 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── @VIGNERON, Jheald, Ghouston, Nikkimaria: I've begun an RfC. Jc86035 (talk) 09:05, 10 April 2019 (UTC)

[Breaking change] "wbcheckconstraints" API new result status

Hello all,

This is an announcement for a breaking change to the output format of the WikibaseQualityConstraints constraint checking API, to go live on (most likely around 12:00 UTC). It affects all clients that use the wbcheckconstraints API action.

We are adding a new status for constraints, in addition to regular constraints and mandatory constraints: suggestion constraints indicate possible improvements to a statement, but are not inherently problematic like other constraint violations. This implies a new status for constraint results as well: in addition to 'violation' for violations of mandatory constraints and 'warning' for violations of regular constraints, as well as several statuses that are not violations, there is now 'suggestion' for violations of suggestion constraints. The default value of the status API parameter is changed to include this status as well (from violation|warning|bad-parameters to violation|warning|suggestion|bad-parameters), and it can appear as the "status" of a result in the response.

API consumers that are not interested in suggestion constraints can specify a non-default value for the status API parameter, e. g. the old default violation|warning|bad-parameters, to avoid getting responses including this status. Others should decide how to handle it, and update their code accordingly if necessary.

According to our stable interface policy, the change will be enabled 4 weeks after this announcement, on May 6th, and a test system will be set up latest on April 17th on

If you have any question or issue, let us know in the related ticket. Cheers, Lea Lacroix (WMDE) (talk) 12:28, 8 April 2019 (UTC)

Hello all,
As promised, you can now test the new "wbcheckconstraints" API result status on Here's an example of a constraint suggestion (you need to be logged-in to see the notification). Lea Lacroix (WMDE) (talk) 12:39, 17 April 2019 (UTC)

Wikidata weekly summary #359

A miracle, I do like a new ID, Libregamewiki ID (P6666) also has a nice number.:-) 09:15, 11 April 2019 (UTC)


I am working on a list of paintings using wikidata data. See nl:Gebruiker:Vincent Steenberg/Lijst van werken van Roelant Savery. Some dates, however, are rendered incorrectly. "17th century" for example is rendered as "1700s". This is incorrect and should be "1600s" or better "17th century". How do I change this? Rehards, Vincent Steenberg (talk) 20:09, 8 April 2019 (UTC)

@Vincent Steenberg: This is a combination of two problems, one of the Wikidata user interface, one of ListeriaBot that creates the tables on nl:Gebruiker:Vincent Steenberg/Lijst van werken van Roelant Savery:
  • Citing from Help:Dates#Known issues: “See Phab:T95553. The user interface accepts and presents dates with precision 7 (100 years) strangely, enticing editors to enter incorrect dates.” That's why the user input created for example for the inception (P571) (mainsnak) on Landscape with Trees and Roots (Q22243304) the following internal value (see here):
    	wikibase:timeValue "1700-01-01T00:00:00Z"^^xsd:dateTime ;
    	wikibase:timePrecision "7"^^xsd:integer ;
    with "7" meaning "century". On Wikidata this is rendered as "17. century".
  • ListeriaBot on the other hand seems just to take the year and put an "s" at the end which gives "1700s".
Workaround (which also leads to perhaps more meaningful internal mainsnak values): Give ((in the mainsnak)) a year that lies in the middle of the range known, here that would (between 1603 and 1613) be "1608" (if all we knew was it's e.g. 17th century it would be "1650"). In the interface it's possible to input this by first typing the year and then changing the precision by clicking on it. From Listeria you'll then get "1608s".
If you need better output then Listeria's XXXXs that can be done with qualifiers and querying them with SPARQL and if you need help with that feel free to ask. Cheers, --Marsupium (talk) 21:18, 8 April 2019 (UTC)
@Marsupium:, Thanks very much for your explanation. I'm beginning to understand I think. I'll try and tweak some of these dates. Hopefully that'll work. Regards, Vincent Steenberg (talk) 21:52, 8 April 2019 (UTC)
And if that fails I have two hands and a keyboard. Vincent Steenberg (talk) 21:54, 8 April 2019 (UTC)
@Vincent Steenberg: I've updated my robots some time ago to handle imprecise inception. I'm currently re-running a lot of the collections to import the missing dates. The logic works like Marsupium describes, so 1608 with precision century qualified with earliest date 1603 and latest date 1613. The collections I've worked on (and will be working on) are listed on my user page. While I'm at it, I'm also adding title statements. So the collections at Wikidata:WikiProject sum of all paintings/Property statistics that are very green for title, have already been processed and the red ones are still to do. That will probably fill some holes in your lists. Multichill (talk) 20:55, 9 April 2019 (UTC)
Multichill's comment at 20:55, 9 April 2019 (UTC) seems to conflate centuries and decades. Jc3s5h (talk) 01:24, 11 April 2019 (UTC)
@Jc3s5h: That's not the case, 1603 is in the 1600s (Q42995) and 1613 is in the 1610s (Q46050). Both are in the 17th century (Q7016). Multichill (talk) 08:56, 11 April 2019 (UTC)

Reactions with molecules or atoms

How do I represent reactions from one molecule/atom to another. Like the famous oxygen reaction with hydrogen to water. At the moment there is no qualifier to do this, or is there? And which one is it?

I searched a bit on the Help page but I did not find anything Juliansteinb (talk) 20:16, 8 April 2019 (UTC)

There are a few items that are instances of chemical reaction (Q36534), and some subclasses of it, and maybe instances of the subclasses. Ghouston (talk) 00:27, 9 April 2019 (UTC)
As far as I can tell, there is currently no property for this. You want to go to Wikidata:Property proposal. Circeus (talk) 00:29, 9 April 2019 (UTC)
pinging Chemistry folks here: ArthurPSmith (talk) 17:30, 9 April 2019 (UTC)

Jasper Deng
Egon Willighagen
Denise Slenter
Daniel Mietchen
Andy Mabbett
Emily Temple-Wood
Pablo Busatto (Almondega)
Antony Williams (EPA)
Devon Fyson
Samuel Clark
Tris T7
Pictogram voting comment.svg Notified participants of WikiProject Chemistry

@Juliansteinb, ArthurPSmith: We are already started the discussion about that topic but we never came up with a solution, see Wikidata_talk:WikiProject_Chemistry#Modelling_a_chemical_reaction?. So if someone can propose an good idea, why not. Snipre (talk) 18:32, 9 April 2019 (UTC)
There are many ways to represent a reaction: 2 H2 + O2 → 2 H2O; H2 + 12O2 → H2O. --Leiem (talk) 01:44, 10 April 2019 (UTC)

Displaying label and description in user's preferred language, on item talk pages

I've created a Phabricator ticket, requesting that item labels and descriptions be displayed on item talk pages, in the user's preferred language.

In the meantime, does anyone know of a user-script or other tool that will do this? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:31, 9 April 2019 (UTC)

T220487 merged to T53044 - which was created on Jul 9 2013! Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:32, 10 April 2019 (UTC)

International Directory of Company Histories

How to create item for a journal? The "journal" published histories of company and constantly renewed the articles in the new volumes. They have individual ISBN and may be the whole series had ISSN. Thus, how to create wikidata item? One volume (book) an item? 10:52, 9 April 2019 (UTC)

I would recommend one item for the series/journal with instance of (P31):book series (Q277759), and an item for each seperate book in the series, using part of (P361). Andrew Gray (talk) 11:50, 9 April 2019 (UTC)


Is tone/news an appropriate Guardian topic ID (P3106) for news (Q38926)? Usually, the property is used where the subject of the Wikidata item is the same as the subject of the news articles tagged with the topic. However, in this case the news articles are not specifically about news; the topic is a catch-all for all news articles. (There are some other topics – e.g. tone/features, /obituaries, /letters, /quizzes, /blog, /extract, /help, /competitions, /polls – which would also currently throw a constraint violation because their IDs begin with "tone".) Jc86035 (talk) 16:07, 9 April 2019 (UTC)

Reminder: Wikidatacon application deadline is April 29

Since this has already slipped off Project Chat I thought I should remind people - if you're interested in attending Wikidatacon in October, there's an application deadline coming up soon - application is via a Google form here. Lots more details on the Wikidatacon 2019 talk page. It would also be nice if people could post a reminder about the application deadline in the other language versions of this page... ArthurPSmith (talk) 20:01, 9 April 2019 (UTC)

multiple editions of a book?

I just created a Wikidata entry for the 1957 edition of "The effects of nuclear weapons" (Q63072754).

There are also a "revised edition" published in 1962 and a third edition from 1977.

Is there some special way to note all three editions with the same Wikidata Q number? Or some way to connect multiple entries, so it's easy to find the second edition from the first, etc.? Or I should just create three separate Wikidata items for the three separate editions and not try to connect them?

Thanks, DavidMCEddy (talk) 04:31, 10 April 2019 (UTC)

  • Editions of a book should normally each be different items, all relating to the item for the underlying work. See Wikidata:WikiProject Books. - Jmabel (talk) 05:01, 10 April 2019 (UTC)
Thanks. DavidMCEddy (talk) 11:46, 10 April 2019 (UTC)
I'm having problems with the following:
* The effects of nuclear weapons (Q63072754): 1957 book
* The effects of nuclear weapons (Q63079864): written work produced in 3 editions: 1957, 1962, 1977
What needs to happen to fix these two, so each points appropriately to the other, and neither has "potential issues"?
Conveniently, all three editions are US government documents, which means they are in the public domain. Moreover, scanned images of all three are available.
However, a Wikipedia article that mentions them (Samuel Glasstone) does not yet provide a link to any of these three. I can fix that, but I think I should get this right first.
Thanks, DavidMCEddy (talk) 13:14, 10 April 2019 (UTC)
@DavidMCEddy: I modified both items to match the classification scheme of WD. Wikidata has no book concept but a work concept and an edition concept. For particular exemplar usually we create an additional item. Snipre (talk) 16:10, 10 April 2019 (UTC)

RfC on non-free content

Pursuant to this discussion a few sections up, I have started an RfC concerning Wikidata's approach to non-free content, since currently the lack of policy on the matter creates ambiguity on whether non-free content can be used under the fair use doctrine of US copyright law. Jc86035 (talk) 10:24, 10 April 2019 (UTC)

Pantone colors

Is there any to note Pantone Matching System (Q749816) colors? I'm thinking of color (P462) or official color (P6364), where a specific color is specified. For example, Princeton University specifies that "Princeton Orange" is Pantone (PMS) 158. Trivialist (talk) 02:58, 11 April 2019 (UTC)

The instructions for official color (P6364) say that the value must be an item which is an instance of (P31) color (Q1075). So far it's simple, silver (Q1090) fails, silver (Q317802) passes. LightGoldenrodYellow fails, if the unnumbered rgb.txt X11 colours aren't complete don't even think about Pantone 158, unless it's Pantone 448 C (Q24885519).
Just kidding, create an item for Pantone 158 with a notability of "you said so", but better explain and discuss what you plan, if it's about hundreds or thousands of new items (my rgb.txt has only 500 lines / colours.) – 09:39, 11 April 2019 (UTC)
Nah, nothing that complex, just wondering if there was a suitable qualifier to use for such things. Trivialist (talk) 00:55, 12 April 2019 (UTC)
Somehow that interests me, and I love subversive editing. Is there some site for the Pantone colour numbers, i.e., individual pages per value? Maybe we could twist that into one of those (too many) IDs favoured here… – 21:22, 12 April 2019 (UTC)
Pantone's site has individual pages for its color IDs—here's Pantone 158 C. Trivialist (talk) 01:39, 14 April 2019 (UTC)
Nice, maybe suggest Pantone color by number or similar as new property. – 00:13, 18 April 2019 (UTC)

A Wiki alphabet

A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

Some amusing trolling:

  • The language of which LGBT-unfriendly country was chosen for the link to a famous LGBT computer scientist ?
  • Which city not in the People's Republic was chosen for the only link to a page in Chinese ?
  • How impressively language-neutral is the link to a famous language-divided city ?
  • Where did at least one other link originally go to, before it ended up linking to 'technical error' ?

Oh, and props to the sharp shooting of international wikidata user of mystery, who managed to bag himself QX ... beats nerve agent VX, I suppose, pointing to a dead page on a dead wiki.

:-) Jheald (talk) 21:31, 11 April 2019 (UTC)

  • The language you're referring to which is my native language is actually pretty gay-friendly wiki. Look at Homosexuality Same-sex marriage and Homosexual behavior in animals. Aren't we supposed to not judge languages by their countries? (isn't it called racism and prejudice?)
  • The cities are chosen on merit that they hosted Wikimedia events specially, Wikimania or Wikimedia hackathon.
  • Some letters are intentionally was left out, like "I", "l", "0", "1" because they are confusing to readers.

BestAmir (talk) 09:13, 12 April 2019 (UTC)

Ahem. Thanks Amir. I feel suitably schooled and chastened. Thanks for putting me right. Jheald (talk) 16:11, 12 April 2019 (UTC)
What Wikimedia community, instead of complaining about the Easter eggs, already moved forward with the feature and created their template to use it more easily? English Wikipedia :-) Lea Lacroix (WMDE) (talk) 09:18, 12 April 2019 (UTC)

@Jheald: See phab:T220432 (declined), which I filed. I think the resolution has mostly been satisfactory, considering the apparent software constraints of modifying the extension to allow for it to be changed. The reason is a dead link is because I used a shell script to reserve links for the 600 Incubator, Beta Wikiversity and Multilingual Wikisource languages (excluding projects without language codes), in addition to creating redirects for basically all active projects and other subdomains. Jc86035 (talk) 09:25, 14 April 2019 (UTC)

Case-insensitive properties

Some identifiers found in Wikidata, such as Property:P638, are case-insensitive and often subject to duplication in both upper-case and lower-case (or even mixed) forms. One such example is found in Q13561329. What is the recommended course of action to solve such a problem:

  • Modified the syntax regex to accept one letter case only, or
  • Ask for Wikidata to check for such duplicates somehow?

--Artoria2e5 (talk) 22:25, 11 April 2019 (UTC)

I think it's best to normalize to one form. Ghouston (talk) 01:03, 12 April 2019 (UTC)
By a bot ? A constraint ? It could puzzle a naive user anyway, if naive users are entering ids.
@Lucas Werkmeister (WMDE): Do you think it would be doable to add a constraint with a « non normalized » regex together with a « normalized » id, to put a useful constraint violation message to the user in the UI ("your ID is ok but not normalized, the correct format is [regexp…]. Usually this means replacing upper case letters with lower case or the converse.", for example. author  TomT0m / talk page 09:35, 12 April 2019 (UTC)
There's a bot that changes the format of ISBNs, as a precedent (User:DeltaBot). Ghouston (talk) 10:02, 12 April 2019 (UTC)
@TomT0m: I’m sure it’s possible with enough effort, but I’m not sure it’s worth it. (It should probably be discussed with Wikidata:WikiProject property constraints first, anyways.) What you can do right now, though, is define a syntax clarification (P2916) on a format constraint (Q21502404) which is shown to the user to explain the regexp, e. g. “all lowercase” for a regexp that prevents uppercase characters. (You can also have multiple format constraints on a property, instead of combining all requirements into a single complicated regexp – that could make violations easier to understand, too.) --Lucas Werkmeister (WMDE) (talk) 10:54, 12 April 2019 (UTC)
In the case of P638, the target site normalises to upper case; we should follow suit. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:16, 12 April 2019 (UTC)
In this case, post of those protein ID's get created by @ProteinBoxBot:. It would make sense to raise the topic on that talk page or speak with people who maintain that bot. ChristianKl❫ 16:16, 12 April 2019 (UTC)

Multiple Start and End Times

I was hoping to add the start and end times to when a specific person was the mayor of a town. He has been the mayor multiple times (three times total). Is there a standard for how we would handle qualifying the Position Held: Mayor property with multiple starts and multiple ends? Kushboy (talk) 23:33, 11 April 2019 (UTC)

@Kushboy: Three different position held (P39) statements, each one with its own start time and end time. Jheald (talk) 23:43, 11 April 2019 (UTC)

Phd (and other kind of) thesis advisors modelling

A relevant question about the advisors of a thesis : How do we link the thesis, PhD thesis and other kind, to the thesis director and so on ? Atm the custom is to link the doctor and its thesis masters with doctoral advisor (P184) View with SQID and doctoral student (P185) View with SQID is supposed to be the inverse property. The point is that a thesis advisor is linked to a specific thesis, if the doctor has several PhD we must find a way to link the director to the species itself. We can qualify the doctoral advisor (P184) View with SQID statement with the thesis, and we can link the director into the thesis item and broaden the property domain.

Related : Which WikiProject for this kind of questions ? I found WikiProject Education and indexed it on Help:Modelling Any other idea ? @Ehitaja: Runner1928 (talk) 22:45, 10 November 2015 (UTC) ArthurPSmith (talk) 17:40, 11 November 2015 (UTC) —M@sssly 10:56, 18 January 2016 (UTC) DarTar (talk) 04:19, 18 February 2016 (UTC) Abreu Guilherme (talk) 23:59, 19 March 2016 (UTC) Netha --Daniel Mietchen (talk) 19:18, 6 June 2018 (UTC) Tris T7 TT me Tris T7 (talk) Vahurzpu (talk) 04:07, 29 April 2019 (UTC) Jinoytommanjaly (talk)

Pictogram voting comment.svg Notified participants of WikiProject Education author  TomT0m / talk page 16:38, 12 April 2019 (UTC)

Depends on the timeframe. E.g. works published in Amoenitates Academicae (Q473804) are often ascribed to Carl Linnaeus (Q1043) and not to his scholars. --Succu (talk) 21:24, 12 April 2019 (UTC)

Wikimedia Foundation Medium-Term Plan feedback request

Please help translate to your language

The Wikimedia Foundation has published a Medium-Term Plan proposal covering the next 3–5 years. We want your feedback! Please leave all comments and questions, in any language, on the talk page, by April 20. Thank you! Quiddity (WMF) (talk) 17:35, 12 April 2019 (UTC)

A third tab?

Pages like Talk:Q14211, and all of our property pages, contain very useful information - but that obscures and overwhelms discussion. Should we request that the devs add a third tab, perhaps called "Documentation"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:11, 13 April 2019 (UTC)

A long time ago I implemented a footer for properties, see for example bottom of OMIM ID (P492) based Property talk:P492/footer. Never really took off. Maybe having an extra link for documentation is better. Shouldn't be too hard to add it with a bit of JavaScript. Multichill (talk) 20:15, 13 April 2019 (UTC)
Symbol support vote.svg Support I haven't seen that before, looks super useful! Is it implemented using Listeria? Would love to see some documentation or a tutorial on how to implement something like that for other types of items, like cover versions of songs. Moebeus (talk) 20:44, 13 April 2019 (UTC)
Started as a User:ListeriaBot job. ---Succu (talk) 20:55, 13 April 2019 (UTC)
Oh my: Talk:Q20895456. --Succu (talk) 21:04, 13 April 2019 (UTC)
I was curious what the values were as part of Wikidata:Wiki-wetenschappers, might import a bunch of data tomorrow. :) Sjoerd de Bruin (talk) 21:07, 13 April 2019 (UTC)
This patch needs a subclass of it's own. Regards --Succu (talk) 21:15, 13 April 2019 (UTC)


Hello. What value genre (P136) must have for:

Xaris333 (talk) 12:03, 14 April 2019 (UTC)

None? genre (P136) is meant for individual works and performers, not professions. The equivalent here, and it's already being used on this item, is field of this occupation (P425). Circeus (talk) 12:52, 14 April 2019 (UTC)
So genre (P136) is not about actors? (I know that is for individual works and performers, not professions. I am just trying to find how (and if) this property could work for actors). Xaris333 (talk) 13:45, 14 April 2019 (UTC)
No. Even if a particular actor worked overwhelmingly in a particular genre, it would be rare to have one who worked entirely in a particular genre. I could imagine having some way to associate an actor with genres or bodies of work with which they are particularly associated (e.g. Humphrey Bogart with film noir, Laurence Olivier with the works of Shakespeare), but in both cases these are loose associations rather than inherent characteristics. It would still be better to get this by some sort of query on the genres of the works with which they are associated. - Jmabel (talk) 14:53, 14 April 2019 (UTC)

Should Q29234537 and Q5322 be merged?

capacitors (Q29234537) and capacitor (Q5322) look the same but they have a different English Wikiversity page. Not sure if they should be merged or not TB5ivVaO1y55FkAogw1X (talk) 20:20, 14 April 2019 (UTC)

Wikiversity has at least three, there's also electronics capacitors (Q60172786). Ghouston (talk) 00:27, 15 April 2019 (UTC)

New dashboard: external IDs and their usage on Wikidata

Hello all,

As you might know, there are plenty of tools analyzing Wikidata’s content and its usage across the Wikimedia projects.

The new dashboard we would like to present you today, the Wikidata Identifier Landscape, is focusing on the external identifiers and their usage on Wikidata. Its different views allow to answer questions like: how much do our external identifiers overlap with each other? With how many statements are they currently described? Which identifiers represent certain topic areas? On what type of items are they usually used? How can we map the galaxy of our external identifiers?

WD EId Network.png

Map of the external identifiers galaxy


On the dashboard, you can browse through different tabs:

  • the Similarity Map presents a global overview of the overlap in the usage of Wikidata identifiers
  • the Overlap Network visualizes all Wikidata external identifiers in a network of nearest neighbors
  • the Tables section allows you to check directly the number of items using a certain external identifier, and the number of items using two external identifiers of your choice
  • the Identifier Classes tab maps the relationships between the external identifiers belonging to the same class (eg chemistry, cultural heritage, feminism…)
  • the Particular Identifier tab provides for each identifier a map of its neighbors and some examples of items where this identifier is used

On every tab, the descriptions gives you more information about the calculation method and the result. You can also check the documentation page.

Screenshot from ExID dashboard - Mérimée ID.png

Map of the external identifiers neighbors of Mérimée ID (P380)


If you have any question, if you find a bug of if you have a request for future development, feel free to ping me or to comment on this Phabricator task.

And of course, feel free to share the dashboard with people or projects that could be interested in playing with identifiers :) Cheers, Lea Lacroix (WMDE) (talk) 07:10, 15 April 2019 (UTC)

Merge Q13217298 and Q12508

Please see long simmering debate to Q13217298. Q13217298 and Q12508 are obviously doubles. WHo can repair that?Kipala (talk) 08:28, 15 April 2019 (UTC)

Link? It appears that they are not the same; one (Giza Pyramids (Q13217298)) is a subset of the other (Giza pyramid complex (Q12508)). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:37, 15 April 2019 (UTC)

Merge Seedot

Hello, it seems that Q49061495 should get merged into Q32006. Could somebody do it, please? --GeXeS (talk) 13:19, 15 April 2019 (UTC)

✓ Done, you can read Help:Merge next time to see how you can do it yourself. Sjoerd de Bruin (talk) 13:32, 15 April 2019 (UTC)
Thank you! I've looked it up in Help, but I must admit it seemed too complicated. --GeXeS (talk) 13:40, 15 April 2019 (UTC)
@GeXeS:, you might try this: in your user preferences, select Gadgets and check the Merge tool. It adds a "merge with..." option to the More dropdown at the top of the page. Really easy. - PKM (talk) 19:27, 15 April 2019 (UTC)
@PKM: Ah, there it was! Thanks a lot, I'll try using it the next time :o) --GeXeS (talk) 06:09, 16 April 2019 (UTC)

Wikidata weekly summary #360

missing artworks

Please find discpage. Regards, Conny (talk) 17:52, 15 April 2019 (UTC).

URL shortener on Wikidata Query Service


Starting today, the URL shortener on Wikidata Query Service has been switched to Wikimedia URL shortener. This means you can now create and post links to queries like this: . Please note that since WDQS uses the anonymous API, the limit is currently 10 short URLs per 2 minutes, or one per 12 seconds. Which means, if too many people do that at once, it would fail. In that case, please wait for a bit and try again.

Smalyshev (WMF) (talk) 18:40, 15 April 2019 (UTC)

to editors and interested on tropical cyclone ontology

Hi. As I suspect that the Wikidata:WikiProject Tropical cyclones is inactive, I invite from this chat, to interested people on cyclone/hurricane topic, about a change in the use of properties explained on its discussion. Thanks, for yours suggestions in those page. Amadalvarez (talk) 22:10, 15 April 2019 (UTC)

Unable to add Commons category to Wikidata item

I'm working on the Amazon Mechanical Turk page on Wikipedia, and would like to use a Wikidata Infobox. Unfortunately, when I try to add the category in the "Other Sites" > "categories" field, I get this error message:

A page "Category:Amazon Mechanical Turk" could not be found on "commonswiki".
The external client site "commonswiki" did not provide page information for page "Category:Amazon Mechanical Turk".

I have tried adding the Wikidata Infobox template, and it comes up poorly populated - the label, description and aka are included, but none of the actual data (Q733115).

I've gone through all the tutorials and templates and am still very confused. Any help would be much appreciated. - Seazzy (talk) 01:26, 16 April 2019 (UTC)

It's true that Commons doesn't have any "Category:Amazon Mechanical Turk". I'm not entirely sure what you are trying to do. Ghouston (talk) 02:26, 16 April 2019 (UTC)
You'd need to create the category first on Commons before linking it to Wikidata. Ghouston (talk) 02:29, 16 April 2019 (UTC)
@Seazzy: I think I know what's happening. First, you or someone else needs to create c:Category:Amazon Mechanical Turk (assuming there is enough media to actually populate such a category. When you create the category, include c:Template:Wikidata infobox. Then come here and edit Q733115 to add that category in the Other Sites field. Once you go back to Commons, you will find it populated. Note that Q733115 doesn't have much data on it now, so not much data will display in the infobox; if there are only four statements made at Wikidata, not a lot can be imported elsewhere. Let me know if you have more questions. —Justin (koavf)TCM 02:30, 16 April 2019 (UTC)
en:Template:Wikidata Infobox also exists, and may be causing some confusion. It's possible to add that directly to the Wikipedia article, but it's not well-tested at the moment, and things like the Alexa rank of websites aren't included in the infobox yet, so they won't show up. Thanks. Mike Peel (talk) 06:49, 16 April 2019 (UTC)
Thanks for your help with this, everyone. I should mention that I regularly facilitate editing workshops at various edit-a-thons, and am hoping to integrate Wikidata Infoboxes into trainings for new editors. The process is very complex for beginners at this point, so I'm hoping that contributing to Wikidata items and linking the data through templates can become more streamlined. - Seazzy (talk) 14:23, 16 April 2019 (UTC)

Place for articles in small languages

Allow me to poke on a probably age-old topic which is new to me: We are creating activities around Saami languages in Wikimedia Finland. We have now the capability to add labels in Inari and Skolt Saami, which are both official minority languages in Finland. (Thank you for everyone who participated!) When we start planning the use and addition on labels, we become curious where to store the articles. We acknowledge that the languages will probably never be able to support their own Wikipedias. What is the status of incorporating Incubator in Wikidata? Could a multilingual Saami Wikipedia be an option to host all Saami languages? What would the technical solution look like?

Another thing enabled by the addition of the language tags is the capability to navigate the Wikidocumentaries project in those languages (translations to sms or fallback for the interface don't exist yet). It would be awesome to be able to display articles that exist in the Incubator and include them whenever there has been a contribution in that language, otherwise rely on language fallback. – Susanna Ånäs (Susannaanas) (talk) 07:47, 16 April 2019 (UTC)

When languages are added, there are two options. When a mono lingual code is added, it is to enable the inclusion of a string of text is specific fields open for this. You always have to add the code for that language.
The other, more productive option is to enable a language in MediaWiki. This does not mean that a language "must have" a project but that multilingual projects like Wikidata are open for content in that language. So when Saami languages are enabled as such, it follows that in you may include localisations and that you can add labels that will show in the Wikidata user interface but also in tools like Reasonator. I am in favour of having these languages accepted in the MediaWiki user interface. It enables the language for sharing on the topics Wikidata has information on. Thanks, GerardM (talk) 09:59, 16 April 2019 (UTC)

I get easily confused with the peculiarities of languages administration, but as far as I understand, this is what we did already. Now that we have the codes, we are figuring out how we can connect texts made in these languages to data recorded in these languages, as they do not have their own projects. – Susanna Ånäs (Susannaanas) (talk) 14:16, 16 April 2019 (UTC)

Help needed: Mix'n'match scraper

I'm looking for someone much better at technical stuff that I am to create a Mix'n'match scraper in order to harvest and add to Mix'n'match some new data from This would really help me in adding statements for the newly created play ID (P6679). If someone feels like doing it, please ping me. Thank you in advance. Powerek38 (talk) 10:34, 16 April 2019 (UTC)

URL-to-statement converter

Is there a Wikidata tool that can quickly add a well-formed identifier statement to an item based solely on URLs? This would be similar to the MusicBrainz interface, which can automatically recognize valid URLs when they are input into the external links boxes. Something like this could potentially remove a large proportion of the delay involved in manually adding lots of identifier statements to individual items. Jc86035 (talk) 15:34, 16 April 2019 (UTC)

Amen! I've been looking for this too, it would be an enormous timesaver when adding any identifiers, but especially ones like spotify that add query parameters to the end adding an extra-extra step. Moebeus (talk) 15:49, 16 April 2019 (UTC)
@Moebeus: I've noticed that w:ru:Википедия:WE-Framework (which I previously installed) does have a popup for searching all of the external identifiers and adding several all at once. I haven't tried it, though. Jc86035 (talk) 16:02, 16 April 2019 (UTC)
  • Just add it with described at URL (P973) and wait for Krbot --- Jura 16:51, 16 April 2019 (UTC)
    @Jura1: I don't think I knew about that. I've added some autofixes since the list isn't very complete at the moment. Jc86035 (talk) 09:27, 17 April 2019 (UTC)

Short URL parameter in Template:SPARQL

I have added an optional |shortURL= parameter to {{SPARQL}}; it takes a value of just the URL slug, thus: |shortURL=37V, and renders, for example:

#Author name strings 
SELECT ?item ?itemLabel
  ?item wdt:P2093 "Bill Thompson" .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }

Try it! (

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:15, 17 April 2019 (UTC)

Statistics about scholarly articles

I am intersted in statistics about the already imported and yet to import articles. Is this somewhere documented? If not, how many scholarly articles are there in total, and how many have already been imported? 18:16, 12 April 2019 (UTC)

There are over 100 million scholarly articles indexed by Crossref; Wikidata has about 20 million (Wikidata:Statistics has not quite 19 million but is several months out of date on that figure I believe). ArthurPSmith (talk) 19:11, 12 April 2019 (UTC)
There are some pertinent statistics on the front page of Scholia: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:09, 12 April 2019 (UTC)
A subset: Wikidata:Database reports/Constraint violations/P356. --Succu (talk) 21:14, 12 April 2019 (UTC)
Note that the articles currently imported form a mostly arbitrary collection. These imports are done by a few editors on their own, without coordinating about any criteria for inclusion. It seems to me that the goal is just to import as many as they can given (or rather despite) the recurrent complaints about this mass-flooding (the constraint violations pointed out by Succu being one example). − Pintoch (talk) 15:06, 14 April 2019 (UTC)
Some of the constraint violations claimed there look bogus - if you click on some of the "Format" violation DOI's, they work just fine. The regex needs to be fixed I guess. Also, why are there 2 "Format" sections in the constraint report? In any case, the overall constraint violation rate is at most around 0.2% (in the 30,000 range out of almost 17 million). Also on the order of 10,000 of these should have been cleaned up in the last few days with fixes for the <...> problem. ArthurPSmith (talk) 16:59, 15 April 2019 (UTC)
Note I corrected the first format constraint on DOI (P356) so it allows the '[' and ']' characters, which are perfectly valid. That should remove a lot of the constraint violations. I'm also running a batch to clean up some remaining '%' URL-encoding issues. There do seem to be some other persistent problems with DOI's in Wikidata, I'm going to continue to work on fixing these. But at least the number of violations should be much reduced the next time that constraint check runs. ArthurPSmith (talk) 18:21, 16 April 2019 (UTC)

The constraint page at Wikidata:Database reports/Constraint violations/P356 is currently 1,800,890 bytes - even my high-spec Dell XPS struggles to open it. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:58, 18 April 2019 (UTC)

It's currently dated April 6 - 12 days old. ArthurPSmith (talk) 17:11, 18 April 2019 (UTC)

Rivers with several named sections

There is some ambiguity with river entities. E.g. Nile river Q3392 is claimed to be 6,690 km, but Nile River proper actually starts at the junction of While and Blue Niles.

So the length actually applies to the main watercourse of Nile basin (Nile<-White Nile<-Bahr al Jabal<-Albert Nile<-Victoria Nile<-Akagera<-Ruvubu<-Ruvironza) rather than to Nile River itself. The same issue applies to quite many world famous rivers like Amazon Q3783 (Amazon<-Rio Solimões<-Ucayali<-Tambo<-Ene<-Apurimac) or Yangtze Q5413 (Yangtze<-Chang Jiang<-Jinsha<-Tongtian<-Tuotuo).

Would it make sense to create entities for each named section with their proper lengths and put the total lengths into the basin entity only?  – The preceding unsigned comment was added by Yaugenka (talk • contribs) at 18:10, 15 April 2019‎ (UTC).

Ideally, it should be clear what a particular item represents. If it's a particular stretch of the river, it should have an appropriate description and length. The English Wikipedia article en:Nile in its lead says "The Nile, which is about 6,650 km (4,130 mi)[n 1] long" and then "The river Nile has two major tributaries, the White Nile and Blue Nile." So how long is the entire river, the Nile plus the White Nile (which is the longer branch?) Or does the 6650km length already include both? Ghouston (talk) 00:12, 16 April 2019 (UTC)
I think that perhaps two items would be useful, one for the stretch of river from the confluence of the White and Blue Niles to the sea, known as the Nile, and another for the "Nile river system" which would include all tributaries. These two concepts seem to be conflated in common use, so it's tricky. I assume that the "length" of a river system can be taken as the longest path from a source to the mouth, rather than the recursive addition of the length of all tributaries. I suppose the enwiki article is about the entire river system, based on the map and length that it gives. Ghouston (talk) 22:48, 18 April 2019 (UTC)
Don't you think that "river system" is somewhat equal to "river basin"? There are "river basin" enities already (e.g. Nile basin Q2887548). btw the 6,650km length is not from the Lake Victoria to the sea but from the most remote spring before the lake (see the river chain in the first post of mine).

Separating person and incident

In case I've done something wrong here, was I correct to separate Junko Furuta (Q1204879) and murder of Junko Furuta (Q61872806)? I got reverted by an unregistered editor on both items, but I reverted back because they left the newer item empty. Jc86035 (talk) 09:54, 18 April 2019 (UTC)

@Jc86035: No. I believe this was a proper split indeed. I made a few edits and removed the significant event (P793) statement as it is now connected using the more specific statement is subject of (P805) instead. Circeus (talk) 22:01, 18 April 2019 (UTC)

how to use talk page?

how to use talk page? any resource


{{页面}} Jiangchaowiki (talk) 08:07, 19 April 2019 (UTC) 页面 – Exactly the same way you use this page. - Jmabel (talk) 15:33, 19 April 2019 (UTC)

Gadget: Links count

I've activated the "Links count" gadget (MediaWiki:Gadget-linkscount.js), but can't see any difference (I've purged and done a browser restart). I've also installed it on Wikispecies, with the same result. Is it working for anyone else? Where should the count appear? @Ebrahim: who added it to this wiki. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:08, 25 April 2019 (UTC)

Hey there, click on "What links here" on your left side toolbox of the page you like then click on "(count)" link, it should work. Thanks −ebrahimtalk 15:04, 25 April 2019 (UTC)
@Ebrahim: Ah, now I see it - the link was not where I expected to find it. Thank you. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:01, 25 April 2019 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:37, 25 April 2019 (UTC)

Representing criminal inquiries in Wikidata

Many Wikipedia articles about events like disappearance of Madeleine McCann (Q639251) and murder of Junko Furuta (Q61872806) contain much more detail about the ensuing legal affairs than the actual incidents. If Wikidata were to represent those, would we need to create separate items for each related investigation and trial? Jc86035 (talk) 11:31, 18 April 2019 (UTC)

Yes, you would need to have an item for the trials. I did created copyright trials. NB I'm still struggling with legal information modelling --Hannolans (talk) 08:47, 20 April 2019 (UTC)

P31 stats

I'm looking for statistics about instance of (P31). That means what are the common items used to identify other items. There are some old statistics at Property talk:P31#Value statistics but this is from 2014.

I tried using SPARQL but that always fails with timeouts. P31 is pretty important so I assume there are statistics somewhere, but where do I find it? --Slomox (talk) 14:54, 19 April 2019 (UTC)

Wikidata:Statistics has at least some top-level analysis, though the main graph there is almost a year old. ArthurPSmith (talk) 16:05, 19 April 2019 (UTC) used to have data on "instance of counts", but somehow it seems borked. @Lydia Pintscher (WMDE): --- Jura 08:47, 20 April 2019 (UTC)


Hello. Which are the icons we are using for feature and good article in wikidata near Wikipedia sitelinks? I mean commons image. Xaris333 (talk) 09:25, 20 April 2019 (UTC)

@Xaris333: According to Firefox's element inspector, the images are SVGs directly embedded in the page CSS, and the fallback is this image. The HTML looks like <span class="wb-badge wb-badge-Q17437798 wb-badge-goodarticle" title="good article" data-wb-badge="Q1"></span>. Jc86035 (talk) 18:23, 20 April 2019 (UTC)

Czech check: Are Státní hymny and Národní hymna not covering the same thing?

Seems to me like Kategorie:Státní hymny and Kategorie:Národní hymny are really trying to cover the same thing, but I don't speak the language so I might be wrong. Doing some heavy spring cleaning on religious hymns Vs. anthems Vs. psalms etc. and this came up. Moebeus (talk) 10:03, 20 April 2019 (UTC)

See interwiki links on Wikipedia:Local Embassy (Q6090776); one should be for the Czech Wikipedia. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:40, 20 April 2019 (UTC)
  • Project chat in Czech might be a more suitable venue. --- Jura 16:15, 20 April 2019 (UTC)
Ah, thanks to both of you, I wasn't aware of either of those, I'll give it a go. Moebeus (talk) 16:53, 20 April 2019 (UTC)
"stát" = country, "národ" = nation. The former category collects anthems which a country has, whereas those of the latter are less official (something like Q19659229 maybe). Matěj Suchánek (talk) 18:35, 20 April 2019 (UTC)


Seems that many things are timing-out today. What's happening? --- Jura 10:47, 20 April 2019 (UTC)

@Jura1: phab:T221458 and phab:T221380 are probably related. Jc86035 (talk) 18:21, 20 April 2019 (UTC)

Another conflicted merge: Kwajalein (Atoll)

Kwajalein (Q309172) is clearly primary.
Kwajalein Atoll (Q33688221) is a dup.
But there are distict sitelinks at cebwiki, so a conventional merge fails.

I gather that in these cases, we simply can't do a merge here until the sitelink problem is fixed (typically with a merge and/or redirect on the offending site, I guess). And when for whatever reason that's impossible, we have to settle for permanent duplicated item (P2959) here. Is this true? Scs (talk) 12:37, 20 April 2019 (UTC)

According to Wikipeida, there's a Kwajalein Atoll / municipality comprising 97 islands and islets, and one of the islands is Kwajalein Island, so there are two entities (or 3 if you want a separate item for the atoll and municipality). Ghouston (talk) 12:52, 20 April 2019 (UTC)
Ah. Good point. That distinction isn't really reflected in the way Q309172/Q33688221 are currently tagged, but it looks like it is the distinction that ceb:Kwajalein (pulo)/ceb:Kwajalein Atoll (munisipyo) captures. Scs (talk) 13:39, 20 April 2019 (UTC)

ResearcherID vs Publons

Clarivate have transitioned ResearcherID to Publons. Eg goes to

I think we should keep both, but encourage the population of Publons.

--Vladimir Alexiev (talk) 12:59, 20 April 2019 (UTC)

Actually you added Publons author ID (P3829) replaces (P1365) ResearcherID (Q7315186) - properties have a designated data type, which is "item" for both replaced by (P1366) and replaces (P1365), so they can't have another property as a value. Thanks for initiating this though, I was just noticing the other day that ResearcherID links redirected to publons. ArthurPSmith (talk) 16:13, 20 April 2019 (UTC)

Wikipedias' 'local embassy' pages

Did you know that, in many cases, you can ask a question about (or request a change on) a Wikipedia in language you don't speak, using your own language, on their 'local embassy' page?

From several recent questions here, it seems this feature is little-known.

See Wikipedia:Local Embassy (Q6090776) for links to the 'local embassy' pages of those Wikipedias which have them. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:44, 20 April 2019 (UTC)

Spam blacklist

Does Wikidata have its own spam blacklist, or does it use the global one? I've been looking for a local one but have been unable to find it. Trivialist (talk) 18:02, 20 April 2019 (UTC)

@Trivialist: It's at MediaWiki:Spam-blacklist. Jc86035 (talk) 18:12, 20 April 2019 (UTC)
@Jc86035: Wow, that's it? Surprised it has so few entries. Thanks for the pointer. Trivialist (talk) 18:14, 20 April 2019 (UTC)

difference between flash fiction (Q5457615) and Flash fiction (Q18235012)

The Japanese Wikipedia has a different article for each item. Is there someone who knows japanese?--Malore (talk) 22:36, 12 April 2019 (UTC)

@Yukida-R, Libertas, Sakamoto3~jawiki: If they know. --Liuxinyu970226 (talk) 23:30, 12 April 2019 (UTC)
@Malore: You can ask at ja:Wikipedia:Help for Non-Japanese Speakers. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:30, 13 April 2019 (UTC)
@Malore: I can't find "フラッシュフィクション" (Flash-fiction) in some Japanese dictionaries, but "ショートショート" (Short-short). "フラッシュフィクション" is only Japanization of "Flash fiction", I think, that is used in America.--SilverSpeech (talk) 00:36, 22 April 2019 (UTC)

Classes for defunct entities

Border Reivers (Q893414) has recently been made an instance of defunct rugby union club (Q63171186) which is a subclass of rugby union team (Q14645593). Apart from the issue that clubs aren't teams, this isn't desirable, is it? We don't have a class like "dead human" for persons. @Blackcat:. Additionally, there are items like Prussian Union of churches (Q163202) and AltaVista (Q433505) which are instances of former entity (Q15893266). Ghouston (talk) 00:17, 17 April 2019 (UTC)

Hello @Ghouston:, I thought it was useful even because there's a Commons subcat for defunct sports teams. Anyway If it's not OK there's no problem for me. -- Blackcat (talk) 07:38, 17 April 2019 (UTC)
Yeah, I don't think it's a good idea, because former entities are handled in Wikidata by adding a dissolved, abolished or demolished (P576) statement. Adding additional instance statements, and class versions for former entities, is counterproductive. It adds no extra information and makes it harder to make queries, since if there are two ways of doing something you have to check both and merge the results. To do it properly that way, you'd basically need a class "former X" for every entity "X", so the dissolved, abolished or demolished (P576) method is a lot easier. Ghouston (talk) 10:57, 17 April 2019 (UTC)
There are such classes that have instances, such as historical country (Q3024240), I think the same would apply. Ghouston (talk) 11:09, 17 April 2019 (UTC)
+1. Defining something as historic or defunct is subjective. The problem is there is no real rule to handle the addition of dissolved, abolished or demolished (P576). Snipre (talk) 18:42, 17 April 2019 (UTC)
I suppose something that doesn't exist any more can be described as "historic" or "defunct", so it's no more subjective than adding a P576. The ambiguity comes in when you are trying to work out whether an item is actually defunct, or whether it continues in a different form. Maybe we have a country with a name change and/or a new form of government: is it the same county, or a different one? The answer is that you can do it either way, or both ways simultaneously with different items. As for whether P576 (and related properties, such as date of death (P570) are sufficient on their own, or need to be augmented with "instance of defunct entity" or "instance of dead human", can only be decided by consensus discussions like this one. If there's no consensus, anyone is free to do it either way. Ghouston (talk) 22:58, 18 April 2019 (UTC)
@Ghouston: What's your take on former municipality and it's subs like former municipality of the Netherlands, former municipality of Switzerland, etc, etc. ? I definitely see your point, but there are thousands of these we need to clean up if this is to be streamlined. Moebeus (talk) 23:46, 18 April 2019 (UTC)
These are more instances of the same thing. Any changes can be done with scripts, so the actual number isn't an issue I think. former municipality of the Netherlands (Q7265977) actually has a description in Dutch that reads "een niet meer bestaande gemeente (gebruik voor claims bij voorkeur item gemeente met einddatum als kwalificatie)", i.e., "a no longer existing municipality (in claims preferably use municipality with and end date qualifier)", so the issue has already been recognized there. Ghouston (talk) 00:19, 19 April 2019 (UTC)
Cool, how do we advance this? I'm up for helping any way that I can, but I *do* see some challenges with script-automation when it comes to locating start/end times (and do we want end cause?). There's a whole lot of these, like former populated place, former protected area, former church, former cultural heritage site, former village, former administrative territorial entity, the list goes on. Would it be a good idea to set up a small project/task force/Trello board to work through all of these in an "orderly fashion" perhaps? Moebeus (talk) 20:50, 19 April 2019 (UTC)
If the end date was missing, I'd just insert an end date with an unknown value. Finding exact dates can be hard work. I'm all for setting up a mini project for it if it's clear that there are no serious objections. Although, objections sometimes don't show up until you start changing things and people see it on their watchlists. Ghouston (talk) 00:05, 20 April 2019 (UTC)
I don't see much added value in deleting such statements. One could easily add two statements on the same item. --- Jura 07:34, 20 April 2019 (UTC)
@Jura1:, I see it as a bit of database cleaning, removing redundant statements, and stopping the proliferation of a pointless parallel hierarchy of former entities. Ghouston (talk) 10:45, 20 April 2019 (UTC)
I think such classes does not make sense if they do not follow a 3 class model : one class for the non defunct entities, one class for the defunct one, and one superclass for both. The interesting thing with this is that it would be very easy in queries to exclude defunct entities of some kind just by querying the « non-defunct » class instances. Without this third item, the defunct classes are not really useful.
For example if « timeless city » is the class for cities, either passed one or still existing one
author  TomT0m / talk page 09:50, 20 April 2019 (UTC)
So we would have 3 entity classes for every entity class that can become defunct, but how would this be more useful than using only dissolved, abolished or demolished (P576) statements to indicate defunct entities? It seems like extra complexity and redundancy for no benefit. Ghouston (talk) 10:40, 20 April 2019 (UTC)
Sounds like a theoretical approach, adding more another possibly unneeded layer. Still, it could work for WikiProject France which has some problems with queries on its regional structure as they use a single instance for some of the defunct layers. --- Jura 10:45, 20 April 2019 (UTC)
Any examples of situations that can't be handled with dissolved, abolished or demolished (P576) statements alone would be helpful. Ghouston (talk) 10:50, 20 April 2019 (UTC)
I don't think the problem discussed at Wikidata_talk:WikiProject_France/Communes#Communes_multi-départementales is resolved. It's partially linked to P31 values, but not primarily. As participants of that project are fond of debating its structure and I don't regularly edit these items, I didn't clean it up. --- Jura 10:56, 20 April 2019 (UTC)
Situations where one level of administrative areas doesn't fit perfectly into a higher level, because some of them cross boundaries? I'm unclear how the old entity classes are helpful. Ghouston (talk) 12:08, 20 April 2019 (UTC)
I think at a given point in time, there should be just one department per commune. Some users might avoid the problem by using one of the identifiers, but it should be possible to do without. --- Jura 10:55, 21 April 2019 (UTC)
@Ghouston: The idea is the same as the one of using the "preferred" rank for up to date information and the "normal" one for older one. In typical query, you don't want to care about the history; Unless you really want it. Ranks allows to do that. In typical query, you don{t want, say for administrative units, to care about the history, unless you really want it. Using the property dissolved, abolished or demolished (P576) or end time (P582) View with SQID qualifiers to check for former entities break that assumption and make mandatory to filter non relevant entities explicitly using a "filter not exists", a minus or something like that. With a different class for current entities, you don't, you just check its instances and don't have to worry about the former one. author  TomT0m / talk page 16:36, 20 April 2019 (UTC)
Thinking about it, if we really want to avoid some parralel class hierarchies, a solution may be to mark a disappeared entity with two statements
Normal ranksubject > instance of (P31) View with SQID < city >
. This would allow the same logic in query in which we want the history using the qualifiers than for example the history of number of inhabitants of a city. Not incompatible with subclasses of "former entity". author  TomT0m / talk page 17:32, 20 April 2019 (UTC)
Yes, you could optimise the data structure for certain types of queries, making it easier by default to write queries that return only the currently existing entities, or making it easier to write queries that return all entities. It's not that uncommon to want queries that also return defunct entities, e.g., it's pretty common for lists of humans to include the deceased. I'm not sure in general that it's a good idea to use a less optimal data structure (e.g., including redundancy, that then ideally would be managed with bots to keep things in sync) to work around limitations in a particular tool. I'd look at whether it can be made very easy to add filtering based on the existence of the dissolved, abolished or demolished (P576) statement, although this is complicated somewhat by using different statements for some classes of entities, like date of death (P570) for humans. Ghouston (talk) 02:55, 21 April 2019 (UTC)
There are other ways that data can be requested from an item, like from a Commons template, and it should give something useful as the instance, not just "former entity". Ghouston (talk) 09:07, 21 April 2019 (UTC)

Request for user script: order snaks

Would be great to have a user script that enables editing the order the qualifiers of one statement and the properties in one reference. I know that the order isn't supposed to carry meaning, but in many cases this would increase clearness on entity pages, its nicer to have start time (P580) qualifiers before a end time (P582) qualifiers. In the serialization this information is hold by the "qualifiers-order" for qualifiers and the "snaks-order" key for references. Thanks for any help with this! (pings appreciated) --Marsupium (talk) 11:45, 17 April 2019 (UTC)

We really need to be able to qualify qualifiers, so it's clear which date a "circa" statement applies to. - PKM (talk) 20:47, 17 April 2019 (UTC)
I think qualifiers apply to the statement they qualify, and there's technically no way to attach a "circa" to another qualifier. If it was going to be possible, it would probably have to be a property of the date itself, like precision or time zone. Ghouston (talk) 02:42, 18 April 2019 (UTC)
  • A default sort order for qualifiers might do? Not sure if I'd like to look at diffs of people trying to order statements. --- Jura 12:59, 21 April 2019 (UTC)

Properties of a reference

Hello. When I add a reference to statement, I usually use 7-8 properties. Is there an easy way, a gadget or something else, that automatically will appear that properties and I have just to complete them? Is hard always to writing the each properties and sometimes I may forgot some etc... Xaris333 (talk) 19:46, 20 April 2019 (UTC)

the closest I know of is the "clone reference" widget, which I haven't found all that useful so far. Circeus (talk) 03:27, 21 April 2019 (UTC)
You can try User:Matěj Suchánek/referenceSuggestions.js. You can clone it and complete this list of suggested properties. Matěj Suchánek (talk) 07:41, 21 April 2019 (UTC)
I already use "DuplicateReferences" but is not useful when there is no reference (or a complete reference) to the item. Xaris333 (talk) 11:37, 21 April 2019 (UTC)

Constraint and geographical object

Hello. Please see Q959341#P159. There is value type constraint (Q21510865). This constraint appears to all Cypriots teams. See type of administrative division of Cyprus Republic (Q60223819). We have either municipalities of Cyprus Republic (Q16739079) either Communities of Cyprus Republic (Q29414133). Aren't these two geographical object (Q618123)? And if they are, where I have to add subclass of (P279)? Xaris333 (talk) 12:08, 21 April 2019 (UTC)

Seeking consensus on a datatype change

Hey all, I'm currently seeking to have the datatype changed for Irish Grid Reference (P4091). All of the details are listed here: Property_talk:P4091#String_Datatype.

A phabricator task is currently pending community consensus and I've pinged some users who may have been interested in the property at some point in time. However, due to lack of response I'd like to invite anyone interested to add their thoughts (please do so on the linked talk page so everything's in one place) --SilentSpike (talk) 15:55, 21 April 2019 (UTC)

Selma Schottlaender (Q63183793)

What is the best upstream strategy to get rid of the error message at Selma Schottlaender (Q63183793) that the Lodz Ghetto is not a proper place of internment? --RAN (talk) 04:30, 17 April 2019 (UTC)

Looks like the problem was caused by the removal of the statement geographic location (Q2221906) subclass of (P279) location (Q17334923), which made every geographic place no longer count as a location. --Yair rand (talk) 04:41, 17 April 2019 (UTC)
Problem existing again. Sjoerd de Bruin (talk) 12:38, 18 April 2019 (UTC)
Pinging @François Malo-Renault, Infovarius:, who performed the changes causing the issues. --Yair rand (talk) 17:29, 22 April 2019 (UTC)

Merging on Wikidata when duplicate Wikipedia articles exist

I would like to merge psychotropic drug (Q20092038) and psychoactive drug (Q3706669), since reputable sources regard them as synonyms (eg. WHO). However, they cannot be easily merged since there are duplicate Wikipedia articles in several languages. I read both Spanish Wikipedia articles, and believe they support the understanding that these are pure synonyms. I think the Wikipedia articles should be merged, but don't know how to go about proposing that in languages I don't speak. Is there a process on Wikidata for merging topics when their are duplicate Wikipedia articles in some languages? Daask (talk) 21:21, 18 April 2019 (UTC)

What about psychiatric medication (Q1572854)? --Succu (talk) 21:29, 18 April 2019 (UTC)
@Succu: psychiatric medication (Q1572854) is defined by intended purpose, whereas medications intended for other purposes, such as cough suppressants, affect mental processes and therefore are psychotropic drug (Q20092038) and psychoactive drug (Q3706669). Daask (talk) 18:38, 19 April 2019 (UTC)
See interwiki links on Wikipedia:Local Embassy (Q6090776). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:10, 18 April 2019 (UTC)
Maybe @MarioFinale, Montgomery: should come here to explain that why es:Psicotrópico and es:Psicoactivo are different. -- 08:47, 19 April 2019 (UTC)
Since at least one Wikipedia has two separate articles, I'd advise keeping two distinct entries here. For the moment, I've added partially coincident with (P1382) to each entity referring to the other (using the WHO as a reference), as that at least indicates a relationship between them, although its original intention seems to be for roads. It's a pity that we can't use synonym (P5973) for this, but it has datatype Sense. Similarly exact match (P2888) would represent the equivalence, but has datatype of URL. Maybe permanent duplicated item (P2959) is the best property to use? --RexxS (talk) 15:39, 19 April 2019 (UTC)
I'd probably go with permanent duplicated item (P2959) in this particular case, but I think we could really use a property "concept overlaps with" (I do use partially coincident with (P1382), but as you say, it seems to want to be about roads. - PKM (talk) 20:52, 19 April 2019 (UTC)
Hi, at least on eswiki those two are different things, psychotropic is used to describe substances that have any impact on the central nervous system (is often used to refer to medication) while psychoactive describes any substance that has some impact on the individual psyche (often used to refer to recreative drugs).
The merging of those two pages has been discussed several times and the community decided to keep both pages as separated terms (for now). MarioFinale (talk) 19:10, 20 April 2019 (UTC)
@MarioFinale: In that case, I guess I'll merge psychotropic drug (Q20092038) into central nervous system agent (Q50429917), make the name "psychotropic" as an alias for psychoactive drug (Q3706669), while attaching at least the Spanish wikipedia article to central nervous system agent (Q50429917). Daask (talk) 13:36, 22 April 2019 (UTC)

Some birth and death locations are no longer a "geographic location"

Some birth and death locations are no longer a "geographic location" and giving an error, see here for example at Arthur Hornbui Bell (Q4799125), they started Friday and I assumed it was temporary. Any ideas why happening and what changed? Something must have changed upstream. --RAN (talk) 15:30, 22 April 2019 (UTC)

#Selma Schottlaender (Q63183793), #Constraint and geographical object. Matěj Suchánek (talk) 16:03, 22 April 2019 (UTC)
  • The query to check this can time out. Also, it only goes through a limited number of levels. --- Jura 17:27, 22 April 2019 (UTC)

Structured data on Commons & Wikidata notability

A heads-up: in a couple of hours, Commons will gain the facility to add "depicts" statements, with QIDs as the value.

We may see colleagues from Commons newly arriving here, to update or create items about things depicted in photographs or artworks. Please welcome them collegially.

We may also see people creating new items for this purpose, some of which may test our current conception of notability; especially in the realm of "structural need". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:55, 18 April 2019 (UTC)

Delayed "for a few hours" as an unrelated bug is preventing deployment. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:04, 18 April 2019 (UTC)
Now postponed to Tuesday 23 April. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:07, 18 April 2019 (UTC)

This is now live. Here's The announcement on Commons and a sample diff. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:07, 23 April 2019 (UTC)

un-deprecate comment (DEPRECATED) (P2315)

What is the process to un-depriciate comment (DEPRECATED) (P2315) property? While working with Copyright statements I occasionally come to some limits of wikidata data model where some statement warrants clarification (in free text), but there is no way to save it. We do have syntax clarification (P2916) "to provide a textual description of the regex syntax" and constraint clarification (P6607) "to provide details on the function or purpose of a property constraint", but nothing useful for clarifying copyright status (P6216) statements. Another approach would be to create new "copyright clarification", but I generally do not like very narrow scope properties. --Jarekt (talk) 13:37, 19 April 2019 (UTC)

I think having a property discussion that works with the same rules as proposing a new property would be the best way. ChristianKl❫ 16:24, 19 April 2019 (UTC)
Thanks --Jarekt (talk) 18:01, 19 April 2019 (UTC)
think in cases where the copyright situation can't be properly modelled using our properties/qualifiers, just adding comments isn't going to work well. Users of our data most likely will never see/read those comments and then try to modify the data by hand on their end. They simply query the data and reuse it without attention to such rare exceptions. The best way to deal with those cases would be to not include incomplete/misleading/incorrect copyright information and wait until the copyright status in question can be properly modelled. --Kam Solusar (talk) 23:39, 19 April 2019 (UTC)
Pinging users who participated that PFD discussion: @Matěj Suchánek, Nikki, Laddo, Izno, Srittau:@Jura1, VIGNERON, ԱշոտՏՆՂ, Snipre, Lymantria:@Pasleim, Thryduulf, Suruena, Arctic.gnome, CC0:@Pppery:. -- 09:16, 22 April 2019 (UTC)
Why not a new property i.e. "copyright clarification"? We don't need to undeprecate another badly named property for that case. --Izno (talk) 12:19, 22 April 2019 (UTC)
We could do "copyright clarification", but how many of those narrow "clarification" properties do we need? I prefer a single generic "clarification" or "comment" than many separate specialized ones. But I might be a minority there. How do others feel? --Jarekt (talk) 13:24, 22 April 2019 (UTC)
@Izno: And why don't you consider just renaming after undeprecate? I ask you because the potential "copyright clarification" will technically very like how P2315 works. --Liuxinyu970226 (talk) 14:56, 22 April 2019 (UTC)
Scope of the property. We are significantly narrowing it and as such we should consider using a new property. It will make it clear both to new users and old (reusers). --Izno (talk) 16:24, 22 April 2019 (UTC)
@Izno: Scope can also be changed, just see original language of film or TV show (P364) for example, its usage was also too confusing before the creation of language of work or name (P407), but should we therefore delete P364? No, P364 is now limited to films only, as (cited from @Jura1:) there are reasons that films do need separated property instead of P407. --Liuxinyu970226 (talk) 13:41, 23 April 2019 (UTC)
We deprecated it because of many factors, not least of which that it was too broad. We should continue down the path where we need to clarify a property. Perhaps a generic "clarification" might be in order, but even that feels too broad. Regardless, please consider starting a PFC discussion to get consensus. --Izno (talk) 16:24, 22 April 2019 (UTC)

Increasing the size of the image in current highlights on the main page

It seems to me that the image current highlights on the main page is often to small and there would be enough space to make it bigger. I asked on Quora about how to go about that in mediawiki and the answer I got was that it requires editing the css. Can someone who knows more about MediaWiki look into the issue on how to make the picture bigger when there's free space for it? ChristianKl❫ 10:08, 20 April 2019 (UTC)

Wikidata:Main_Page/Popular has a talk page. No CSS involved, ordinary wiki text: Admittedly 100px is rather small, but the two column layout with floating images yields in essence four columns, and in my browser window (less than 1600px in full screen mode) wider images could cause havoc. OTOH I never visit the main page of c:, d:, m:, mw:, w:de:, or w:en:. – 04:55, 23 April 2019 (UTC)

User pages

We have got quite a backlog at Wikidata:Database reports/User pages. For admins, help delete obsolete items. For others, help decide which items should be deleted and disconnect invalid links. Matěj Suchánek (talk) 08:16, 22 April 2019 (UTC)

Do we even need copious manual review here? It seems to me the question is: Are there significantly many cases where sitelinks to user pages are ever necessary/desirable? If not, couldn't we just have a bot delete all sitelinks to user pages, and then delete all entities with no remaining sitelinks or claims? (Apologies if I'm overlooking something; I'm relatively new here.) Scs (talk) 11:36, 22 April 2019 (UTC)
There are plenty of items, mostly userboxes for user pages, where some wikis have them in the template namespace and some in someone's userspace. Unless we follow zero tolerance policy, I wouldn't oppose keeping them (so that we don't make people outside Wikidata unhappy). Matěj Suchánek (talk) 16:05, 22 April 2019 (UTC)
Maybe the report can be modified to include only items that have only user pages as sitelinks. Ghouston (talk) 23:38, 22 April 2019 (UTC)
Arwiki seems to have changed where userboxes should be located, from template namespace to subpages of ar:User:صندوق مستخدم. There are currently 1657 pages on the above mentioned list with only one connected sitelink, and that sitelink links to the userspace of ar:User:صندوق مستخدم in arwiki. What to do with them? If there is consensus to delete those items, I would hereby offer to do that. —MisterSynergy (talk) 06:01, 23 April 2019 (UTC)

Why duplicate label names are allowed in some languages, and are disallowed in others?

Are those controlled by some scripts? --Liuxinyu970226 (talk) 10:53, 22 April 2019 (UTC)

Which languages allow this and which don't? Matěj Suchánek (talk) 15:56, 22 April 2019 (UTC)
@Matěj Suchánek: e.g. zh and their zh-* allow, while ca disallows. --Liuxinyu970226 (talk) 13:48, 23 April 2019 (UTC)


do can creat items about sports match in wikidata example Bayern munich vs Werder Bremen in week 30 bundes liga 2018 19 Amirh123 (talk) 11:05, 22 April 2019 (UTC)

Yes, you can definitely. There are already many of them, e.g. VfB Stuttgart versus Borussia Mönchengladbach 2004–05 Fußball-Bundesliga (Q30579939). 12:14, 23 April 2019 (UTC)


Somaliland. The former British Somaliland Protectorate achieved full independence from the United Kingdom in June 1960. ... Somaliland restored its independence in 1991 after the collapse of Somalia as a result of the civil war. Since then it has established and sustained peace and stability  – The preceding unsigned comment was added by Maansoore (talk • contribs) at 11:43, 23 April 2019‎ (UTC).

@Maansoore: Is there an issue with a related Wikidata item? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:45, 23 April 2019 (UTC)
@Maansoore: Just to be clear here, if you're going to edit Q34754, that's fine but ensure that you have sources for claims you make. —Justin (koavf)TCM 17:51, 23 April 2019 (UTC)

imported from Wikimedia project

Today a constrain added to imported from Wikimedia project (P143). [7] I have noticed many items that use GeoNames ID (P1566) that have imported from Wikimedia project (P143) --> GeoNames (Q830106). I know that the constrain is correct beause the property is about Wikimedia projects. How can we solve the problem? Maybe by changing imported from Wikimedia project (P143) --> GeoNames (Q830106) with another property? Xaris333 (talk) 13:11, 7 April 2019 (UTC)

  • The long term plan was to move it to "stated in". If there is a bot operator who wants to do it .. please go ahead. --- Jura 13:14, 7 April 2019 (UTC)
    I didn't realize it's still unresolved... I reverted myself but of course, better it got fixed. Matěj Suchánek (talk) 14:47, 7 April 2019 (UTC)

@Pasleim: Can you do this? Xaris333 (talk) 13:22, 7 April 2019 (UTC)

  • Last year I did some work in this field and I do have bot code, but the fact that User:Pintoch "temporarily" removed the constraint from the property somehow killed my motivation to invest more time here. Without an active constraint, I'm afraid that we will never reach a point where this reference property is in good shape. In August 2018, we had around 6 million imported from Wikimedia project (P143) references with non-Wikimedia project values. —MisterSynergy (talk) 16:04, 7 April 2019 (UTC)
    • I am very sorry about that! I did this because Nikki reported to be on the verge of quitting the project because they were hampered in their work by the delayed opening of references which had constraint violations. I thought that was sad. I have personally no strong feelings about this and will stop standing in the way of progress on this issue. − Pintoch (talk) 17:27, 7 April 2019 (UTC)
    • Let's try to fix them first and re-add the constraint afterwards. Otherwise users loose time trying to figure out what might be wrong and even attempt to fix it manually. --- Jura 17:36, 7 April 2019 (UTC)
      • Do we have anyone who is willing to do this in the current situation? Since I stopped mid last year, nobody has spend any efforts into it. —MisterSynergy (talk) 08:55, 10 April 2019 (UTC)
        • @MisterSynergy: Is there anything that is holding you back from running your bot code? Is it available somewhere? − Pintoch (talk) 12:17, 10 April 2019 (UTC)
          1. Lack of time; I have meanwhile re-counted actual numbers: we currently have 5.1 million references left that need to be fixed, with ~2100 different items as invalid values. If my bot was editing at 60/min and 24hrs a day, this would take 2 months to complete.
          2. The code needs some minor optimizations (not really a problem, but needs to be done—optimally before I share it).
          3. My approved bot task for this job is valid explicitly for constraint violations; technically, we do not have one here any longer. I have been in trouble regarding bot policy violations once, thus I am very careful regarding such questions meanwhile.
          4. Furthermore, for quite some values it would be better to really fix them, rather than just move P143 to P248 in reference qualifiers. For VIAF, I have complemented a lot of useful information (and checked the actual statement against VIAF), but doing this for all values would scale up the entire problem by orders of magnitudes. —MisterSynergy (talk) 18:30, 10 April 2019 (UTC)
          • Ok, well I don't mind using my bot and a server for that, resource-wise. What sort of fixing are you referring to in your last point? Do you have an example? − Pintoch (talk) 18:39, 10 April 2019 (UTC)
            • Special:Diff/787597532 for example. I did check the value against the VIAF references before doing so. —MisterSynergy (talk) 19:11, 10 April 2019 (UTC)
              • @MisterSynergy: I see, yeah that seems a lot harder indeed. I can cook up something if you are reluctant to run your own code, but I will definitely not get into that sort of fixing (I would just translate the properties). − Pintoch (talk) 19:58, 10 April 2019 (UTC)

@MisterSynergy: here is an update on this: I have started to run the task, but I have stopped the bot today, given that the WDQS lag is increasing. Because the issue affects a lot of big items, editing at 30 edits/min might be too much. I will try running the edits at around 10 edits/min once the lag is back to normal. At this rate it should take about a year to migrate everything. − Pintoch (talk) 12:36, 24 April 2019 (UTC)

I think you can edit with max allowed rate on a substantial subset of items with no or only few sitelinks. You can play with (the results of) the queries which I added on your bot task request page, in order to find such subsets. At least for the values which show up more than 10.000 times it would probably make a lot of sense to check the situation first, in order to reduce the required time for the migration. Example: there are ~390.000 items with ~1.12M references imported from Wikimedia project (P143): English Heritage (Q936287), but only 18.5k items of them have one or more sitelinks. I am pretty sure that you won't bring Wikidata down if you edit on such a set with 60/min and finish it in a week or so. The same applies to the reference value China Biographical Database (Q13407958), and possibly some others. —MisterSynergy (talk) 13:46, 24 April 2019 (UTC)
@MisterSynergy: you are certainly right, but I have neither the time nor the interest in doing this fine tuning, especially given the trouble each WDQS high lag episode causes for the community and the admins. If anyone wants to have a go they are more than welcome, but as far as I am concerned I will keep my bot slow. − Pintoch (talk) 14:01, 24 April 2019 (UTC)

How to model a herbarium

We have, for example:

and multiple cases of each pattern. Can something be a herbarium and an organisation? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:06, 17 April 2019 (UTC)

No, 2 items are required. Snipre (talk) 18:38, 17 April 2019 (UTC)
organization (Q43229) has to be present somewhere in the ontology because herbarium must be able to be a subject of properties like employer (P108) and affiliation (P1416). Really, there's no difference between herbarium and any other collection in this regard. As a general rule, Wikidata does not (currently, anyway) separate libraries and museums from the collection they hosts. What makes herbaria peculiar is mostly that they are as likely to be independent institutions of their own as to represent department/collections within a broader institution. Circeus (talk) 21:55, 18 April 2019 (UTC)
I suppose some herbariums are run by separate organisations and some are operated by an organization like a University or government body. If you worked at a herbarium operated by a university, you may be an employee of the university, and the herbarium would be a place of work. I also know of people who work at a museum, but the museum is operated by a local government body and they are employees of that body, not of the museum. Ghouston (talk) 00:41, 20 April 2019 (UTC)
After this merge another question raises: How to model merges / splits of those collection? --Succu (talk) 21:36, 20 April 2019 (UTC)
I'm not sure what the question is? I mean, this item literally seems 100% fine as it is? Former, now merged entities can have their own items such as Dudley Herbarium (Q18352749), which is now for all practical purposes is merged into California Academy of Sciences (Q965731), but unlike the herbarium merged into AD, is still cited independently. (Note to self: we probably need an item for the CAS herbarium... It's overlooked because there's a single CAS page at Wikispecies) If they don't exist (or even if they do), then it's just a matter of assigning proper ranks to the Index Herbariorum code (P5858). See also Herbarium, Botany Unit, Finnish Museum of Natural History (Q56033053) for a more detailed approach. Circeus (talk) 02:28, 21 April 2019 (UTC)
I added some references to State Herbarium of South Australia (Q7603235). BTW Herbarium, Botany Unit, Finnish Museum of Natural History (Q56033053) is missing any references. In my opinion HFR, HEL, HSI and HPP need their own item and should modeled accordingly. I think this makes mapping of type species (Q252730) and follow ups much easier. --Succu (talk) 20:47, 24 April 2019 (UTC)
Somehow related is the employment of Hansjörg Eichler (Q4530123) modeled here . --Succu (talk) 21:39, 24 April 2019 (UTC)

Wikidata weekly summary #361

Linking request

Could someone please link w:hyw:Ֆինլանտա into Finland (Q33)? 2001:999:3:F513:6B30:5E8D:87BC:2FCE 13:56, 24 April 2019 (UTC)

✓ Done Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:23, 24 April 2019 (UTC)

Q7434833 and Q8639236

Should the Wiktionary sitelinks in Q7434833#sitelinks-wiktionary (Category:Proto-languages (Q7434833)) be moved to Category:Reconstructed languages (Q8639236)? Visite fortuitement prolongée (talk) 19:38, 24 April 2019 (UTC) @Infovarius, Pamputt: Visite fortuitement prolongée (talk) 19:44, 24 April 2019 (UTC)

Need help for query

Hello, I need help to build a query: I want to extract all chemicals having an InChIKey with the CAS value if this value is available and the name of the English Wikipedia article if an article exists in WP:en.

Here is the start of the query:

  ?compound wdt:P31 wd:Q11173 ;
            wdt:P235 ?inchikey 

Try it!

I could add a line wdt:P231 ?cas ; to extract the CAS value, but I don't want to get only items having both InChIKey and CAS, I want all chemical with InChIKey (mandatory) with CAS if available (option). Same for the title of the English Wikipedia article. Thank you for your help. Snipre (talk) 09:55, 29 April 2019 (UTC)

You can do this using OPTIONAL { ?compound wdt:P231 ?cas } . OPTIONAL { ?article schema:about ?compound; schema:name ?title; schema:isPartOf <> }.
You can post questions on queries to WD:Request a query. Matěj Suchánek (talk) 11:59, 29 April 2019 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 19:21, 30 April 2019 (UTC)

merge request

Could someone please merge Q62985816 into Chris Kingsbury (Q59588837)? - 17:21, 29 April 2019 (UTC)

✓ Done. --Epìdosis 17:39, 29 April 2019 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 19:21, 30 April 2019 (UTC)

Suggestions based on constraints will be enabled for all users on April 24th

Hello all,

In March, we told you that we want to make the beta feature providing suggestions based on constraints to all users. Some of you gave us feedback and pinged us about unresolved issues (thank you for that!). These issues are now solved:

  • the order of suggestions in the case of a property having a one-of constraint now matches exactly the order of the constraint (phab:T220587)
  • the results appearing in the drop-down menus are now links: you can open them in another tab (phab:T207363)

We're now ready to enable the suggestions based on constraints for all users (including non-logged-in users). It will take place next week, on April 24th. The beta feature will still be available during a few days (in case something goes wrong and the deployment train is rollbacked, for example), but if everything works well and no major issue is reported, we will disable the beta feature one week after the deployment.

You can see more information about the feature on the documentation page. Please note that the feature is not preventing users to add any item they want as a value, you can still type in and choose the item you want, but it is making more useful suggestions, based on criteria defined by the community with the constraints, and hopefully reducing the number of mistakes and improving the user experience when adding new statements.

Thanks again for your feedback! Lea Lacroix (WMDE) (talk) 13:23, 18 April 2019 (UTC)

Edit: the feature will be enabled on Thursday, April 25th. Lea Lacroix (WMDE) (talk) 12:49, 23 April 2019 (UTC)
Edit: the feature is now enabled for everyone!
A simple way to try it: go to the sandbox, add a new statement with sex or gender (P21) or another property having a one-of constraint, click on the value field: the suggestions that are made are now based on the values of the property constraint.

Feel free to test it and give us feedback. Lea Lacroix (WMDE) (talk) 11:56, 25 April 2019 (UTC)

Note on merge fuction

The access to the merge fuction must be reviewed, if numerous items are involved. It can effect the databank structure. Grandfathered mineral (G) got merged with an auxiliary status (Q30146803). Now I have to use circa 35 hours editing time to review the databank on minerals. The Wikipedia projects have a problem with number of editors, editing time, maintenance, trolling and mobbing/power play. Regards --Chris.urs-o (talk) 03:36, 25 April 2019 (UTC)
One easy way to prevent wrong merges - at least those which happen in good faith - is to interlink the items which easily confused to be the same with different from (P1889). In fact, the merge is prevent on any items which are linked by any property, so if the two items share a real relation other than being confused with each other that should be modeled here, and as a side-effect will it also helps to avoid wrong merges. Ahoerstemeier (talk) 08:45, 25 April 2019 (UTC)

Wiktionary & Wikidata: What is the current status?

Hello! I am interested in knowing the percentage of Wiktionary items (from any or all languages) linked to Wikidata. I found this as I was looking online, but I still don't understand much about the current situation.--Reem Al-Kashif (talk) 19:03, 25 April 2019 (UTC)

Hello, Have you read Wikidata:Wiktionary/Sitelinks? Visite fortuitement prolongée (talk) 19:07, 25 April 2019 (UTC)

Please cancel the Armenian "permanent duplicate" items

Because the Western Armenian Wikipedia has created now. -- 01:37, 8 April 2019 (UTC)

Symbol support vote.svg Support These are not actual permanent duplicate, they're just different language contents. Pinging @Nikki, Jura1: to help. --Liuxinyu970226 (talk) 05:04, 8 April 2019 (UTC)
The duplicated hyw-articles need to be moved to the new hyw-Wikipedia project, sitelinked in the main item, and subsequently deleted in hywiki (?) and removed from the duplicated item. Once this is the case, remaining empty items will automatically come to administators' attention for deletion. —MisterSynergy (talk) 11:59, 8 April 2019 (UTC)
Hoi, which status now? It seems that links to hyw.wikipedia can be added now. -- 09:18, 22 April 2019 (UTC)
Symbol support vote.svg Support There's entirely and No doubt that No reason to continue calling them "permanent duplicates". -- 01:53, 27 April 2019 (UTC)

Let me pick a language in which I want to add labels

I do not want a user page (so no babel setup), but I often want to edit labels in languages other than English, German, French and Bavarian. The babel setup is useless anyway since I may edit in for example Japanese which I don't speak but can help with the kanjis. I have no idea why Bavarian is one of the four default shown to me. If this were eurocentric, why not Spanish? If by alphabetical order, why not Arabic? I don't want to use any gadget (LabelLister) either. I am using web version on Windows 10 and I have not customised anything.--Roy17 (talk) 23:32, 23 April 2019 (UTC)

I guess I am shown Bavarian because I tried switching UI to it long ago. Wikidata shows me the current UI language and remembers the three I picked immediately before that.--Roy17 (talk) 23:44, 23 April 2019 (UTC)
"I do not want to use the tools provided to allow me to do X; please provide me with another tool to do X" is not a good way to suggest that either paid devs or volunteers spend their time. Set up a user page; you do not need to put anything else on it. For a language you do not speak, use (for example) |ja=0. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:47, 24 April 2019 (UTC)
@Pigsonthewing: I'll rephrase my words. Multilingual labels, descriptions and aliases should be as editable as wiki sitelinks, that users should be allowed to select a language at least by its code. IP users dont have a user page and cannot use gadgets either. I often want to edit in Chinese family languages (zh zh-hans zh-hant zh-cn zh-hk zh-tw zh-sg gan wuu yue), Japanese and European languages. Running a long list of babels on my user page is obnoxious. LabelLister is not very convenient: it takes much more clicks than the JS interface. What I suggest is nothing new. Users can pick any language they wanna help with on Wikimedia Commons.--Roy17 (talk) 18:34, 24 April 2019 (UTC)
I thought there was a project the developers were working on a year or so back to change that label/description box to allow language selection - it's definitely something I've wanted too, when I occasionally have a name in a language that's not on my list that I'd like to add, for instance. Anybody know what happened on the dev side there? No enthusiasm for it? ArthurPSmith (talk) 21:31, 24 April 2019 (UTC)
@Pigsonthewing: The babel solution is really awkward, though - I sympathise entirely. It's perfectly reasonable as a way to show default or preferred languages, but if you want to do a once-off edit of a label in another language, you have to edit your userpage twice (to add the babel and remove), and it feels weird to say "to edit something, you need to make a public declaration that you do or don't speak it", even for purely trivial housekeeping.
Of the other on-wiki methods, LabelLister only works for languages which are already present. Changing your user interface language will bring up the new label/description section, but means you have to edit with the interface changed, which can be quite daunting. I needed to do this yesterday with a label in Czech that had been put into the English section by accident, and I ended up just doing a one-line QuickStatements batch rather than fiddle with the onwiki language options. We really could do with an easier way to get at these fields. Andrew Gray (talk) 22:02, 24 April 2019 (UTC)
@Andrew Gray: "LabelLister only works for languages which are already present" Not so. I just used it to add a test label here, in a language not already present. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:17, 24 April 2019 (UTC)
@Pigsonthewing: Weird! If I try that, with LL enabled, I'm only presented with entries for the existing languages and no way to tell it to include a new one. I wonder what's different? Andrew Gray (talk) 22:21, 24 April 2019 (UTC)
@Andrew Gray: Which version are you using ("released" or "beta")? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:55, 24 April 2019 (UTC)
@Pigsonthewing: "released", presumably (in preferences > gadgets); I don't see a version in the beta tab and nothing else there is enabled. Andrew Gray (talk) 19:56, 25 April 2019 (UTC)
@Andrew Gray: If I open LabelLister, I see a line at the top, "NEW ! : Change several language in one fell swoop (go to beta version)." ("NEW !" is bright red). If I then select "edit", I get a pop-up inviting me to select "Language code:". This is in Firefox, under Win 10. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:28, 25 April 2019 (UTC)
Still mysterious, but unselecting it, reselecting it, and purging everything I could think of seems to have worked! Andrew Gray (talk) 22:02, 25 April 2019 (UTC)

That's one more question I've found so far, even I've added those language labels in the user page in Meta (the Babels), the recent changes won't allow me to edit the labels in the Chinese language families (zh-CN, zh-HK, zh-MO, zh-MY, zh-SG, zh-TW) but a single Chinese (zh) entry? Seem's cannot find the related discussions around here or Phabricator. Before this changes I can add those labels for the languages above. The changes makes more trouble when dealing with the titles for the movie or the characters for those titles, either in Anime, Magna or Movies. Shinjiman (talk) 06:01, 26 April 2019 (UTC)

Just have some update for that one, seems looks like it may be the issue found from the Babal extension? the zh-HK, zh-MO and zh-TW were rendered with the zh-Hant-* and zh-CN, zh-MY and zh-SG were rendered with zh-Hans-* respectively and suspected these kind of rendering causing the new labels with those languages could not be added, even they are added as the Babel template in the user page. Shinjiman (talk) 02:57, 27 April 2019 (UTC)

Constraint on father that dies before the birth of their child

See Johan Christopher Ruuth I (Q50346736). Shouldn't we add 9 months to the death date of a father before we invoke an error message? I have seen several examples where men going off to battle impregnate their wives before they leave, assuming they may die. --RAN (talk) 06:07, 24 April 2019 (UTC)

Nine months is the average pregnancy, but some last for almost ten. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:49, 24 April 2019 (UTC)
@Pigsonthewing: Agreed! Can you update the constraint? --RAN (talk) 03:02, 25 April 2019 (UTC)
I'm not sure that's possible; I think the best thing you can do is add the item as an exception, on Property talk:P40. Note that Louis X of France (Q8384) is already listed as an exception. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:12, 25 April 2019 (UTC)
@User:Lucas_Werkmeister_(WMDE): It would be good to be able to change the constraint in a way that doesn't require listing individual items as expections. Currently, that doesn't seem to be possible. What do you think? ChristianKl❫ 11:50, 26 April 2019 (UTC)
@ChristianKl: I imagine we could introduce some parameter like “tolerance” or “permitted delta” (quantity data type) that could be specified on this constraint. But that should first be discussed with Wikidata:WikiProject property constraints, I think.
(Side notes: with frozen sperm, the delta can actually be a lot longer than nine months; and, just for reference, there are currently some two hundred fathers who died up to nine months before their child was born.) --Lucas Werkmeister (WMDE) (talk) 13:57, 26 April 2019 (UTC)

Medal "For Courage" in USSR and Russian Federation - two items or one?

Hi! I have got a little problem with Russian Medal "For Courage". I have figured out, that there are two items for this award:

  • Q1970346
    • linked with article on English Wikipedia (and several other language versions), that is treating this award before and after 1993 as one
    • linked with article on Russian Wikipedia, that is considering only USSR era
  • Q16675224
    • linked with article on Russian Wikipedia, that is considering only usage in Russian Federation

What should I do:

  • Save two items and recommend to split article on English Wikipedia; or
  • Merge those items and recommend to merge articles on Russian Wikipedia?

--Krzychu025 (talk) 07:36, 26 April 2019 (UTC)


How do we think about that? Looks like a community-wide consensus is still waiting for this. --Liuxinyu970226 (talk) 00:07, 27 April 2019 (UTC)

[Breaking change] Important for Wikidata tools maintainers: wb_terms table to be dropped at the end of May

new scheme

Hello all,

This is an important announcement for all the tool builders and maintainers who access Wikidata’s data by querying directly Labs database replicas.

In May-June 2019, the Wikidata development team will drop the wb_terms table from the database in favor of a new optimized schema. Over years, this table has become too big, causing various issues.

This change requires the tools using wb_terms to be updated. Developers and maintainers will need to adapt their code to the new schema before the migration starts and switch to the new code when the migration starts.

The migration will start on May 29th. On May 15th, a test system will be available for you to test your code.

The table being used by plenty of external tools, we are setting up a process to make sure that the change can be done together with the developers and maintainers, without causing issues and broken tools. Most of the documentation and updates will take place on Phabricator:

We are aware that this change will ask you to make some important changes in your code, and we are willing to help you as much as our resources allow us to. We hope that you will understand that this change is made to avoid bigger issues in the near future.

Note that this change is not impacting Wikibase instances outside of Wikidata. A dedicated migration plan and announcement will follow.

We strongly encourage you to not wait until last minute to make the changes in your code. If you have any question or issue, we will be happy to help. In order to keep the discussions in one place, please ask questions or raise issues directly in the Phabricator task and board.

Thanks for your understanding, Lea Lacroix (WMDE) (talk) 14:26, 24 April 2019 (UTC)

Xwiki categorization help

I feel like this is a stupid question, but I'm not very tech savvy, so bear with me. Is there any fairly easy way to tell what pages in a category on one project are connected through WD to corresponding pages on another project?

For example, if I looked at en:Category:Nobel laureates, can I somehow pull a list of how many of those English Wikipedia articles have a corresponding page on the English Wikiquote? GMGtalk 20:21, 25 April 2019 (UTC)

GreenMeansGo Sort of, but it's tricky, because Wikidata itself doesn't know about category memberships on the wikis. You can, however...
  • First, use petscan to generate a list of articles in that category for en.wikipedia (first tab, "categories", depth=0 means only use this category, no subcats), and also tell it to give you the results as wikidata pages (fourth tab, "other sources", and check "use wiki: wikidata" at the bottom of the options) example, using en:Category:Nobel laureates in Chemistry with no subcategories.
(You can tweak this to include multiple categories, exclude items in other categories, extend the search to include subcategories, etc. Note that if you include a large depth it can get tied up in knots, because the en.wikipedia category tree is really messy. But let's keep it simple for now)
  • If this gives reasonable results (it should, but best to check in stages), tweak the petscan search to filter for pages that also have a sitelink to en.wikiquote - this is on the "wikidata" tab, under "Has any/all of these site links". example
  • The resulting list is Wikidata items for each page that a) exists in en.wikipedia on that category; b) also exists in en.wikiquote (without checking on how it's categorised there). (If you want to get English Wikipedia pages for those items, go to the "other sources" tab and check "Use wiki: from categories" instead, then rerun.)
Hope that works! It's a bit daunting at first - petscan is quite fiddly - but hopefully a worked example makes it clearer. Andrew Gray (talk) 22:01, 25 April 2019 (UTC)
Wikidata does not know about category memberships, but I understand that the query service does, as it can call through the MediaWiki API? There are some examples in Lucas_Werkmeister’s QuickCategories documentation − but that’s way above my SPARQL-fu, so not sure whether that applies here ^_^'. Jean-Fred (talk) 07:50, 26 April 2019 (UTC)
  • Hmm... thanks all for the responses. I'm going to have to set aside some time to play around with this a bit. GMGtalk 15:07, 26 April 2019 (UTC)

@GreenMeansGo: a little bit easier query: enwikiquote links (the diff from Andrew's query is that I fed "Manual list" in "Other sources" tab and used "Use wiki" option). --Edgars2007 (talk) 08:47, 27 April 2019 (UTC)

Ah. Thanks. If only I could get cat-a-lot to play nice with pet scan. GMGtalk 12:40, 27 April 2019 (UTC)
Quote me on this, I will donate $50 to the charity of your choice for anyone who can make cat-a-lot and petscan work together. GMGtalk 12:40, 27 April 2019 (UTC)
GreenMeansGo I wouldn't do it with Cat-a-lot per se, but here's a solution - @Lucas Werkmeister:, who is an absolute genius, recently wrote m:User:Lucas Werkmeister/QuickCategories, which may do exactly what you want (and @Jean-Frédéric: mentioned above). Specifically, looking at m:User:Lucas Werkmeister/QuickCategories#Copying category members from another wiki, you can run this query):
SELECT ?titleWQ ("+Category:Nobel laureates in Chemistry" AS ?command) WHERE {
  hint:Query hint:optimizer "None".
  SERVICE wikibase:mwapi {
     bd:serviceParam wikibase:api "Generator";
                     wikibase:endpoint "";
                     mwapi:generator "categorymembers";
                     mwapi:gcmtitle "Category:Nobel laureates in Chemistry";
                     mwapi:gcmnamespace 0;
                     mwapi:gcmprop "title";
                     mwapi:gcmlimit "max".
     ?titleEn_ wikibase:apiOutput mwapi:title.
  BIND(STRLANG(?titleEn_, "en") AS ?titleEn)
  ?articleEn schema:name ?titleEn;
             schema:isPartOf <>;
             schema:about ?item.
  ?articleWQ schema:about ?item;
             schema:isPartOf <>;
             schema:name ?titleWQ.
Try it!
...then copying-and-pasting as suggested will let you populate a QuickCategories run to take all the Wikiquote pages which have a corresponding enwiki page in w:Category:Nobel laureates in Chemistry, and add them into q:Category:Nobel laureates in Chemistry. Andrew Gray (talk) 12:58, 27 April 2019 (UTC)
Yo @Koavf:, you are fairly good at computering. All I see is hieroglyphics. Does this make any sense to you? GMGtalk 14:33, 27 April 2019 (UTC)
@GreenMeansGo: Hm, sorry--I'm not really wrapping my head around this. (I also had a hard time sleeping last nite. :/) I've never used most of the tools that manipulate or analyze Wikidata. —Justin (koavf)TCM 16:37, 27 April 2019 (UTC)

Help! How can I edit this darned thing???

Hi. I come from editing an English Wikipedia article. That article contains a number of false and some bad faith/misleading information that comes from Wikidata and that I want to fix. Clicking on the edit symbol next to that information, I ended up here on Wikidata. But despite there being and "Edit" button it does not let me correct any of the false data. According to some help page I should be able to edit existing statements, but apparently I am not.

What can I do to fix that? I've now read through several help pages and articles, and I'm not an inch closer to what I want to do. This whole place seems to function completely counter-intuitive... Wefa (talk) 23:20, 25 April 2019 (UTC)

@Wefa: What is it that you would like to change? There is no protection on the item, so in principle, despite only one edit (your message) on Wikidata, you should be able to edit the item. Mahir256 (talk) 23:22, 25 April 2019 (UTC)
yes I can "edit" the item - but that does not let me change the data fields (stateements? I'm not really sure about trhe terminology here). Look at the page I linked - I need to change the version, release date and software url and remove the additional fraudulent url pointing to rivaling project (fork). But none of the statement boxes gives me an editing opotion. I only have an edit button in the top right corner of the page, and that only lets me edit name, description and aliases. Wefa (talk) 23:33, 25 April 2019 (UTC)
Update: fascinating, Wikidata needs Javascript but does not complain when not having it. Allowing in NoScript (a Firefox script blocker addon) added the statement edit options... Wefa (talk) 23:37, 25 April 2019 (UTC)

Ok, so I can now edit. But... it wont let me save in many cases. The software version isnt 2.27, its 2.29 now. Could not save that the first time - error message suggesting to add another entry instead. 2nd attempt I could save that. And I still can not remove the fraudulent git url - when I tried it removed the whole statement "source code repository"(whose semantic is somewhat unclear to me anyway), which was not my intention. I had to undo it via the version history. How can I just remove the stale (and fraudulent) git link? Wefa (talk) 00:05, 26 April 2019 (UTC)

@Wefa: If you're referring to the "" link, you can click "edit" on the "source code repository" statement, click "remove" (the one not right next to that URL, but the link above that one), and then click "publish". Looking at it, though, I don't think this is intended to be used to access the git repo itself, but as a source for the "git://" URL for the repo. Mahir256 (talk) 00:33, 26 April 2019 (UTC)
@Wefa: If you're referring to the "git://" link, you can simply click "edit" next to that URL and then click "remove", which will remove it and its associated reference, leaving the "" statement unaffected. Mahir256 (talk) 00:37, 26 April 2019 (UTC)
Ok, I just did that. Seems to have worked, thanks (didn't work last time, or I misinterpreted the error message then).
But in the meantime someone else (a user "Visite fortuitement prolongée") re-added the old, outdated software version as an additional claim. Why would he do that? Why just this old version, and not the other dozen or so older versions of the software that were relsaed at some point in the past? This place is strange. Wefa (talk) 01:48, 27 April 2019 (UTC)
"Why just this old version, and not the other dozen or so older versions of the software that were relsaed at some point in the past?" => Because I do not know those versions. Visite fortuitement prolongée (talk) 20:13, 27 April 2019 (UTC)

Bot not currently running, that used to create P373s for Commons categories?

I think there used to be a bot that created a Commons category (P373) statement if an item was sitelinked to a category on Commons. (@Mike Peel: was this one of yours?)

Currently it doesn't seem to be running -- or a least, there seem to be a lot of such items sitelinked but with no P373.

Would it be possible to get it going again? It's a pain to have to check for both types of link in queries. Thanks, Jheald (talk) 12:36, 26 April 2019 (UTC)

See Wikidata:Administrators' noticeboard#Bug on Q729048--Ymblanter (talk) 20:50, 26 April 2019 (UTC)
That was DeltaBot, not one of mine. Thanks. Mike Peel (talk) 20:55, 26 April 2019 (UTC)
Thanks, Mike. @Pasleim: Looks like the underlying query might be timing out, if it is still the one you linked to on Topic:Uaevjjsargkbkeb0. Any chance of having another look at this? Jheald (talk) 22:22, 26 April 2019 (UTC)
BTW, if someone can provide me with a suitable query, then I can write a new script for Pi bot that will do this quite easily. Although I'm hoping that P373 will be deprecated at some point this year so that we can more permanently avoid this issue/duplication. Thanks. Mike Peel (talk) 06:45, 27 April 2019 (UTC)

Rebuttal; the method behind the madness

Ask yourself, what is it that Wikipedians say validates their articles.. It is their citations. Ask yourself, when scientists ask attention for other scientists like Jess Wade does on a daily basis, what is it that makes them scientists.. It is their work and it is in their scholarly papers. How will we know what citations are bogus.. We will know because retractions are finding their way in Wikidata and they are used to identify bogus facts in Wikipedia.

At this time we are only at a fifth of the authors compared to ORCiD and we are lagging behind in their papers. So when I add papers it may escape casual observers what the method behind the madness is.

  • scientists that are in the public eye, people with public records at ORCiD are singled out for processing
  • they and their co-authors are processed with SourceMD in order to complete their papers and add co-authors that are not known yet
  • thanks to the Scholia tool it is known what gender these people are. When it is not known I add genders
    • this helps us understand the gender ratio of scientists in a particular science, in Wikidata, in a set of co-authors
  • science is easily ignored by pundits that want to only express their opinion, this is particularly true in politics and business
    • it is why I am adding the papers of Elizabeth Warren - this may change the exchange of opinions by adding validity to points of view
  • other subjects that I personally care for: maternal health, health care in Africa, climate change, environmental science

As attention is given to publicly known science and scientists, it follows from set theory that when you expand on items in the set, that relations become apparent when you look for them. Also more becomes known from what is published on a given topic. It is not complete but completeness becomes increasingly irrelevant for use in a Wikipedia, for finding sources on a subject.

People have objected that thanks to issues with the Wikidata database, the use of SourceMD, duplicates are created. I have cleaned up the duplicates for authors for a few iterations of the report. This is something I will continue to do.

What I ask of you is the understanding why the use of SourceMD is important, I will keep up with duplicate authors.

Thanks, GerardM (talk) 08:54, 27 April 2019 (UTC)
I agree this is important to do. One suggestion: if you are running for coauthors of an author, you are likely to pick up the same papers for each author. If SourceMD recognizes the papers have already been added, that's not a problem. But if you run two coauthors too close together, you may get the same paper added twice. So maybe in the case of coauthors, running them should be spaced a few hours apart? ArthurPSmith (talk) 16:22, 27 April 2019 (UTC)
The work I do on Elizabeth Warren shows the same co-authors popping up all the time. It follows that it does not make a real difference. Again, one person with his co-authors at the same time. Thanks, GerardM (talk) 17:13, 27 April 2019 (UTC)

How to identify sources to free copies of a scholarly work

Hoi, I regularly find links that provide access to a PDF. How do I add this to an item for a scholarly article ? Thanks, GerardM (talk) 09:26, 27 April 2019 (UTC)

arXiv ID (P818) could be used if it is on ArXiv. But I assume that your question targets cases where there is no version on ArXiv? In this case, reference URL (P854) might be appropriate. 09:33, 27 April 2019 (UTC)
I do not know if they have something of ArXiv, it would be cool when there is (or a backup on Internet Archive).. Thanks, GerardM (talk) 09:35, 27 April 2019 (UTC)
There is full work available at (P953). − Pintoch (talk) 12:13, 27 April 2019 (UTC)

Categories move on the projects does not result in the move on Wikidata?

An example from yesterday: w:en:Category:Places of worship in El Salvador was moved yesterday to w:en:Category:Religious buildings and structures in El Salvador with a creation of a category redirect. However, our Q9505708 still lists the old name and apparently was not affected by the move. Anybody knows what is going on?--Ymblanter (talk) 12:29, 23 April 2019 (UTC)

There is only a partial match between the category redirect pages created yesterday by Cydebot in enwiki,[8] and its wikidata contributions at the same time.[9] E.g.
  1. en:w:Category:Places of worship in Hesse YesY also in Wikidata
  2. en:w:Category:Places of worship in Hawaii County YesY also in Wikidata
  3. en:w:Category:Places of worship in Hawaii YesY not changed in Wikidata
  4. en:w:Category:Places of worship in Havana YesY not changed in Wikidata
Fayenatic london (talk) 12:55, 23 April 2019 (UTC)
Well, this is disappointing so far. I was hoping for
  1. a diagnosis leading to a fix to prevent Wikidata being skipped any more in future when pages are moved in Wikipedias, and
  2.  automated processes to identify the redirects in Wikipedias that have been left linked to Wikidata, and then update Wikidata to the target of the redirect.
Over a thousand "places of worship" categories in enwiki were moved to "religious buildings and structures". Having reviewed them manually for other purposes (checking backlinks), not really looking at Wikidata links, I estimate that about a quarter of the redirects are still linked to Wikidata.
Is there a bot that could at least report which categories listed at en:Wikipedia_talk:Categories_for_discussion/Log/2019_March_21#OPTION_B are linked to Wikidata?
Of course we should also seek evidence as to whether other pages have been moved in Wikipedias without updating Wikidata, unless there is a reason to think it only happened with Cydebot's contribs on 22 April. Fayenatic london (talk) 20:37, 25 April 2019 (UTC)
Could this be a bot approval issue? Cydebot is not an approved bot on Wikidata, so it would have the same edit-limit restrictions as any regular user, so if these changes were made too fast, some fraction of them would be disallowed by the API, I believe. ArthurPSmith (talk) 23:29, 25 April 2019 (UTC)
Thank you, ArthurPSmith, that would make sense. I'll recommend Cyde to seek prompt approval for Cydebot on Wikidata, and recommend the same for other bots that might move or delete categories. At least that should reduce future omissions of this kind.
I don't know the process for bot approval. Could the general process be tweaked to prompt the bot owner to also seek approval on Wikidata where relevant? Fayenatic london (talk) 11:19, 26 April 2019 (UTC)
I don't think this problem is related to edit rate restrictions (IIRC we don't have any right now, and User:Cydebot's edit rate wasn't particularly high at that time of the failed sitelink moves anyway), and it is probably also not related to a (missing) botflag. No idea what's on here, but we should keep an eye on this problem. --MisterSynergy (talk) 14:35, 26 April 2019 (UTC)
Anybody running multiple simultaneous Quickstatements batches quickly hits the edit rate limit, it's definitely there. ArthurPSmith (talk) 15:49, 26 April 2019 (UTC)
It would be a credible cause of skipping these particular category page moves. This is because (i) I had listed a large number of categories for moving, and (ii) there were relatively few direct members of the categories for Cydebot to update in between moving one category page and the next, so that Cydebot was making category page moves at an unusually high rate – and those were the contribs that (should have) affected Wikidata.
This is a pretty unusual combination, which also gives some reassurance that there should not have been many other such cases in the past. I have requested a report at Wikidata:Request a query#Wikidata items linked to soft redirect on enwiki. Fayenatic london (talk) 20:25, 27 April 2019 (UTC)
And now Larske has written this PetScan query which lists Wikidata items linked to enwiki category pages using category redirects. There were 69 "Places of worship" categories, not too many to fix manually. Fayenatic london (talk) 23:06, 28 April 2019 (UTC)

I've filed a request for bot flag at Wikidata:Requests_for_permissions/Bot. --Cyde (talk) 14:25, 26 April 2019 (UTC)

named as (P1810) vs. stated as (P1932) with external identifiers

Names of entries in external databases often do not exactly match the label of an Wikidata item. This is especially true for non-person items, where also the scope of the entry may be slightly different from the scope of the item. So, Computable general equilibrium (Q3589458) is linked via STW Thesaurus for Economics ID (P3911) to 29879-5/"CGE model" in the Thesaurus for Economics, or Dardanelles (Q6514) is linked via PM20 folder ID (P4293) to the press archives folders sh/141034,146008/"Osmanisches Reich (-1923) : Dardanellenfrage" and sh/141111,146008/"Türkei (1923 -) : Dardanellenfrage".

Especially in the latter case, it is very useful to see the external entry title together with the external id link. The same is true for companies, where name changes are relatively frequent. Therefore, I consider automatically adding the according values as qualifiers for the above mentioned properties.

Two qualifiers, named as (P1810) and stated as (P1932), can be used for that. Their descriptions and the according creation discussions indicate that they where originally introduced with other use cases in mind (credits in film industry vs. author names/abreviations in printed sources). Both are now in relatively wide use with identifiers, because their definitions do not make really clear which one to use for the given purpose.

I have a slight preference for named as (P1810), because its description (in some languages) already mentions "name by which a subject is recorded in a database", which seems to come close to external ids. What do you think?

If a consensus for a recommendation to "named as" can be reached, perhaps a usage instruction for stated as (P1932) could be added: "With external identifiers, better use P1810 (named as)." That would not outlaw the probably wide current uses of "stated as", but would be helpful for new users and uses.

Current usage statistics

Property uses in qualifiers uses in references uses in qualifiers for external ids
named as 239477 846607 108020
stated as 4537368 2540 ???

(according to Sqid and this query, as of 2019-04-28) Unfortunately, due to timeouts, it was not possible for me to see how many of the "stated as" qualifiers are used with external id properties.

-- Jneubert (talk) 07:49, 28 April 2019 (UTC)

@Jneubert: As I understand it, the distinction between these two qualifiers is that named as (P1810) should be used to indicate how the subject of the qualified statement is given in the source; stated as (P1932) indicates how the object of the qualified statement was given in the source. In some cases both may appear, eg in the item for a person giving the job they had on a particular film, the credits for a film may name the the person in a particular way, eg "J.S. Smith" rather than "Joan Smith" (--> named as (P1810)); and also state their job in a particular way, eg "lead makeup artist" rather than "makeup artist" (--> stated as (P1932)).
Applied to identifiers, what we want to record in the qualifier is how the name of the subject of the statement is given, so named as (P1810) should be used. Jheald (talk) 09:01, 28 April 2019 (UTC)
@Jheald: This has been my understating as well. As an example I use named as as a qualifier on Discogs artist ID to indicate what name a recording artist is registered under in Discogs, as it often differs from what we have in WD. I use stated as as a qualifier on performer to indicate the actual artist name printed on the label of a music release, in contrast to what the label in WD might say. Thank you @Jneubert : for bringing this up, it's an area that's not very well documented. Moebeus (talk) 10:46, 28 April 2019 (UTC)
I also distinguish subject = "named as" and object = "stated as". - PKM (talk) 20:00, 28 April 2019 (UTC)

What's the policy on Wikipedia articles on single words? - Move to lexeme or wiktionary? (Re:Nynorsk wikipedia "Fasong")

I just stumbled on fasong, a nn:nynorsk wikipedia article (stub) about the Norwegian loan-word "Fasong". While I think it's cool to find it in Wikidata, I can't help but think it would be better suited as a Lexeme, it feels weird/messy trying to model it as instance of = word, or whatever. Do we have a policy/way to handle items like this? Moebeus (talk) 21:43, 22 April 2019 (UTC)

  • If there is an article at WP, we have an item here.
Maybe it's "Create the corresponding lexeme" ? You could link it with subject lexeme (P6254). --- Jura 04:56, 23 April 2019 (UTC)
That should work, I wasn't aware of that property, thanks! Moebeus (talk) 10:57, 23 April 2019 (UTC)
@Moebeus, Jura1: I'm fairly confident that's not the correct property to use here. There's a difference between an item's topic being a word and the topic being about a word (like, suppose the linked article explained things about a famous article/book/entry that was about a word), and that property appears to be for the latter. It's important not to mix the two up. --Yair rand (talk) 17:12, 28 April 2019 (UTC)
@Yair rand: Any suggestions as to how you would go about modelling it? What not to do is useful, what to do is even better! Moebeus (talk) 17:24, 28 April 2019 (UTC)
@Moebeus: My first thought is that the best way to handle this would be to create a new property which could point from the lexeme to the item (or the other way around). However, the handling of some lexeme-items is quite inconsistent. The items for "American" (word) and the phrase "talk to the hand" have linked articles dealing with related phrases in other languages, rather than all being about one phrase in one language, whereas the items Bule (Q4996114) and Sprezzatura (Q545839) are about the lexeme in one language. I think these items in general need to be cleaned up and standardized, and a distinction needs to be made between some different types, and then perhaps a new property can be created to link relevant items to lexemes. --Yair rand (talk) 17:56, 28 April 2019 (UTC)

The specific article in question should be probably converted to an entry on Wiktionary. Might be relevant to mention en:WP:WORDISSUBJECT here, which contains a few more examples. --Njardarlogar (talk) 11:11, 29 April 2019 (UTC)

SPARQL Queries from the Wolfram Language

After executing a query at a button appears to generate, for different languages, code that executes that query.

Version 12 of the Wolfram Language (Q15241057) contains a dedicated function to do that. Example:


endpoint = "";

(* cats *)
query = "
select ?item ?itemLabel where {
  ?item wdt:P31 wd:Q146.
  service wikibase:label {bd:serviceParam wikibase:language \"en\".}
} limit 2

result = SPARQLExecute[endpoint, query];

The result looks like this:

In[5]:= result

Out[5]= {
  <|"item" -> IRI[""], "itemLabel" -> RDFString["Nala", "en"]|>,
  <|"item" -> IRI[""], "itemLabel" -> RDFString["Mr. White", "en"]|>

So, if the community considers it useful to have a tab to generate Wolfram Language code for queries, here is one possible way to do it.

Disclosure about possible conflict of interest: I work on semantic web technologies at Wolfram Research (Q1367937), but this is my personal contribution.

Toni 001 (talk) 10:53, 24 April 2019 (UTC)

I think it's worth submitting a patch for this. phabricator:T207749 has some related patches that could be helpful. --Lydia Pintscher (WMDE) (talk) 11:48, 29 April 2019 (UTC)

Length limit on URL-shortener URLs ?

I have been trying to make a short URL for this query, but I can only get "URL shortening failed".

Does anyone know, is there a length limit on the URLs that can be sent to be shortened? @Ladsgroup: ? Jheald (talk) 14:39, 27 April 2019 (UTC)

For length 1994 I got For lengths greater than 2048 I got a bogus [Object object] or similar. For length 2047 I got an empty string. The error handling is apparently not yet implemented. There must be some limit, browsers and servers have limits, e.g., [10]. Huge limits like the 2 MB in Chrome make no sense for this service, my first guess was 4096, and from there I needed only three attempts to "find" (guess) 2000. Nice question, thanks, please put the result in phab:, it needs a task number.:tongue: 21:14, 27 April 2019 (UTC)
Ticket now filed for this in Phabricator. Jheald (talk) 09:28, 28 April 2019 (UTC)
Turns out there was already a ticket in Phabricator; I've updated the link above. Jheald (talk) 13:53, 28 April 2019 (UTC)
^.^b Habitator terrae has good CCC "what could possibly go wrong" ideas, there's some attack vector for this feature. – 12:35, 29 April 2019 (UTC)
FWIW, worth noting that there was always a limit with the old (non-WMF) shortener - but it was a bit longer, at 3k not 2k. Andrew Gray (talk) 14:06, 28 April 2019 (UTC)

Property NLP ID

Something is wrong with links of NLP ID (former scheme) (P1695). From any item You can get only onto the page of Biuro Usług Promocyjnych (Sopot). On property examples links are different: (for Jan Stanisław Skorupski (Q12904)) and (for Piers Anthony (Q559409). Links are different but result is always Biuro Usług Promocyjnych (Sopot). Is it possible to fix? --Ksc~ruwiki (talk) 17:34, 27 April 2019 (UTC)

@User:Ksc~ruwiki: This ID is generally being maintained by the Polish National Library (I mean, as an authority control identifier, not on Wikidata), which is one of GLAM partners for Wikimedia Polska. If you think something is wrong on the Library's side, you can try writing to Marta or Celina - they are both WMPL staff members working with GLAM partners. Powerek38 (talk) 09:54, 29 April 2019 (UTC)

Awarder does not seem to work

Hoi, Awarder is a tool with a minimal amount of edits. When there are 80 it is already really exceptional. Lately the tool does not seem to work, however it turns out that it is likely throttled because Wikidata cannot cope with its edits. As I had waited for fifteen minutes for the first edit to complete, I moved elsewhere. The next morning I found that the seven edits for the J.T. Knight Prize had completed.

In another place I read that because of a lack of hardware particular changes were not possible. Availability of hardware is often essential for performance. My question is: Wikidata grows faster than expected. Have we factored in this growth beyond expectation, are we in a position to throw additional hardware at the problem. Thanks, GerardM (talk) 05:58, 29 April 2019 (UTC)

Fixing coordinates in Mainland China

I've noticed that for a lot of items/articles for pages in Mainland China, the coordinates are offset due to Restrictions on geographic data in China (Q20716290). Is there a tool that could be used to semi-automatically fix these coordinates (e.g. OpenStreetMap iframe with a "which of these coordinates looks more accurate" picker next to it)? Jc86035 (talk) 09:29, 29 April 2019 (UTC)

Wikidata weekly summary #362

Iron oxide

Can someone familiar with our chemistry ontology determine if we should merge Q26841283 and iron oxide (Q721849), and change the name or aliases of iron(II) oxide (Q196680) to distinguish it? Thanks. - PKM (talk)

Yes, I will try to clarify this. --99of9 (talk) 00:49, 30 April 2019 (UTC)
@99of9: Looks much better now, thanks. - PKM (talk) 19:42, 30 April 2019 (UTC)

Best qualifier to divide P5008 (Wikimedia project focus list) ?

Can anybody suggest a good qualifier to allow an on focus list of Wikimedia project (P5008) group to be divided into batches?

In the past whenever I've wanted to identify a new batch of items, I've simply created a new item to be a new value for the on focus list of Wikimedia project (P5008) property; but this isn't very structured, is creating unnecessary items, and doesn't scale well.

Instead, I'd prefer to add a qualifier, eg batch = n, that I could then pick up with a query to pull out a particular batch of focus items within an ongoing project.

Can anyone suggest if there is already a qualifier that would be a good fit for this role? One option would be series ordinal (P1545) to identify the batch. (Usefully it is string-valued, so values like "1A" would be possible). But perhaps this might be misleading, people might instead expect series ordinal (P1545) to indicate a sequence number within a batch, rather than a batch number itself.

Is there a better option that anyone can suggest? Jheald (talk) 18:00, 30 April 2019 (UTC)

Might edition number (P393) work? Perhaps not, as its restricted to be used for main statements, rather than as a qualifier. Jheald (talk) 21:18, 30 April 2019 (UTC)
issue (P433) could be viable. Jheald (talk) 21:28, 30 April 2019 (UTC)

Olga vs Оља

I am a little confused by Olga in the Latin alphabet vs. Оља in the Cyrillic alphabet, please see Olga (Q38038024) vs. Olga (Q30060526). Should they be merged, or are they two entries for a reason? --RAN (talk) 22:22, 23 April 2019 (UTC)

@Richard Arthur Norton (1958- ): Names in different scripts have different items. See especially bullet point 2 at Wikidata:WikiProject Names#Basic principles. —Galaktos (talk) 23:28, 23 April 2019 (UTC)
@Galaktos: But then, ar:أولغا (اسم) must be separated from the first item, right? --Liuxinyu970226 (talk) 23:35, 23 April 2019 (UTC)
@Liuxinyu970226: possibly – I can’t read the linked article (and I don’t think it makes much sense to send it through Google Translate for this question), but if there are people who are named something like “Olga” but in Arabic script, then I would think that should be a separate item. (Though I’m not even a member of WikiProject Names, so you should probably ask them to be sure.) —Galaktos (talk) 00:46, 24 April 2019 (UTC)
I think this whole name stuff would be better suited for lexemes. Currently, a russian person named Оља often has the given name statement "Olga", which is strictly speaking not correct. 15:54, 26 April 2019 (UTC)
I am sorry but JFYI there is no such Russian name (Оља), there is "Оля". --Infovarius (talk) 15:18, 3 May 2019 (UTC)
@Infovarius: Given context, presumably О́льга. - Jmabel (talk) 15:37, 3 May 2019 (UTC)