Wikidata:Project chat/Archive/2018/09

From Wikidata
Jump to navigation Jump to search


Honorable Citizens

Do we have property for en:Honorary citizenship? I mean a property to add all the Honorable Citizens of a municipality to the municipality items. And about the opposite, to add to the item of a person that he/she is a Honorable Citizens of a municipality can we use award received (P166) and how? Xaris333 (talk) 03:08, 1 September 2018 (UTC)

It's done the second way with many instances, like . Ghouston (talk) 03:12, 1 September 2018 (UTC)
So this way is enough? We don't need the first way? Xaris333 (talk) 03:31, 1 September 2018 (UTC)
It seems like enough to me. It's similar to other awards. Ghouston (talk) 11:06, 1 September 2018 (UTC)
Yes, this seems best. Andrew Gray (talk) 14:25, 1 September 2018 (UTC)
By the way, "honorable" and "honorary" are distinct words, and one cannot be substituted for the other. - Jmabel (talk) 16:20, 1 September 2018 (UTC)

Thanks all! Xaris333 (talk) 18:56, 1 September 2018 (UTC)

This section was archived on a request by: Matěj Suchánek (talk) 09:57, 3 September 2018 (UTC)


I simply don't get what the heck is he doing here? He is not even properly explaining what is the purpose for this? And why is undoing the translations in other languages? Why not just edit the description? This is redundant. He is also accusing me of repurposing the matter. I don't know why? I just asked him to elaborate. Harsh Rathod Poke me! 17:41, 2 September 2018 (UTC)

@Harshrathod50: We try not to conflate family name items with given name items if at all possible; you don't see us merging Smith (Q21079662) with Smith (Q1158446), for example. If you must combine the two—which under the current system for modeling names is discouraged—then Jura's last edit summary ('please don't repurpose items. If it's used for both, P31 should be name (Q82799)') is a perfect succinct explanation of what you should do. @Harmonia Amanda: can explain more on this subject. Mahir256 (talk) 18:03, 2 September 2018 (UTC)
Yes, this is the explanation I was asking for. Now I get everything. Thanks. Harsh Rathod Poke me! 19:14, 2 September 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:57, 3 September 2018 (UTC)

Question about mass adding properties into wikidata

So some of my suggested properties for wikidata have recently be approved, how can I add those properties into related wikidata entries relatively quickly? C933103 (talk) 23:18, 29 August 2018 (UTC)

@C933103: it depends what your data source is. If you want to extract them from Wikipedia articles, try Harvest Templates. If you want to import the data from a tabular file, try OpenRefine. For more background, see Wikidata:Data_Import_GuidePintoch (talk) 10:35, 1 September 2018 (UTC)
@Pintoch: I am talking about properties of external identifiers. Also, it seems like that Harvest Templates tool cannot be used to harvest wiki not run by WMF?C933103 (talk) 14:51, 1 September 2018 (UTC)

Please delete Q56501558

Unfortunately I created this duplicate to Q4214852.--Carl Ha (talk) 00:11, 7 September 2018 (UTC)

Mahir256 Merged Q56501558 into Q4214852. --- Jura 04:27, 7 September 2018 (UTC)

This section was archived on a request by: --- Jura 04:27, 7 September 2018 (UTC)

Merge instructions - where?

So I tried to add an article to an item and got a warning message saying it was already in another - but if there were really both the same merge them. But the warning did not link to any description of how to merge them. By trial and error, I got two sets of two into one set of four (Q1863530) - with a mostly empty item still hanging around: (Q27664520). How do I get rid of that? Which one should I have chosen as the merge target? Why doesn't the warning message give any guidance? Rmhermen (talk) 16:17, 28 August 2018 (UTC)

Rmhermen, the page Help:Merge explains the different ways to merge items. And, the target item should be the one who is not obsolete. Esteban16 (talk) 18:30, 28 August 2018 (UTC)
Why doesn't the error message link to this explanation, though? Rmhermen (talk) 14:40, 30 August 2018 (UTC)
Doesn't it: MediaWiki:Wikibase-error-sitelink-already-used? Matěj Suchánek (talk) 13:18, 3 September 2018 (UTC)

How to mark or report items which seem to be the same?

A colleague who is currently working on a mapping between STW Thesaurus for Economics (Q26903352) and Wikidata sometimes finds more than one counterpart for a STW descriptor. In obvious cases, he merges the items and goes on. In other cases, items cannot be merged, because they are linked to different Wikipedia pages, or a deeper understandig of the concepts would require more language knowledge - or the items look just too large and complex. One such case is the cluster of fishing (Q14373), fishery (Q180538) and commercial fishing (Q11202642). Please understand that we do not want to discuss this particular example here, but introduce it merely as an illustration of the problem. More generally, it is out of scope of the mapping work he is charged with, to solve such conflicts or invest time into a discussion about a proper solution.

What we could and happily would do is to mark such clusters we come across, in order to hint other users interested a the particular item to other items which seem to be the same. I've experimentally marked the three items mentioned above with said to be the same as (P460) inter each other. I'm not sure, if this is the right way (at that point, the dispute is only in our heads, and we don't want to make it explicit with arguments and counter-arguments). If somebody can point out a better practice of marking the involved items, which does not require much more work, or where to report such items on a maintenance page, we are happy to follow suit. We can also deal with the issues completely on our side of the mapping, but feel that would mean wasting insights we gained along the way. Jneubert (talk) 14:09, 30 August 2018 (UTC)

@Jneubert: "Said to be the same as" is not the correct sollution; a fishery (a venue) is not the same as "fishing" (an activity). In your example, "commercial fishing" is a subclass (or subcategory, if that helps) of "fishing" (which can be commercial or not). It would seem that you and your colleague need to be clearer about what precisely the thesaurus is, in such cases, describing. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:26, 30 August 2018 (UTC)
Does the STW contain a mapping between German and English? – Looking at your example, it is impossible to do a mapping de:STW = en:STW = en:WD = de:WD. Assuming that the STW mapping en:STW = de:STW is correct and that you will be able to do the mapping en:STW = en:WD with reasonable reliability, I would suggest that you focus on mapping en:STW = en:WD and output a maintenance list of all the cases where de:STW != de:WD. – Does that make sense?
I don't like the idea of using said to be the same as (P460) in this case. The problem here lies not so much in the fact that someone would claim that the concepts are the same, but more in fuzzy language links between different Wikipedias, in bad translations on Wikidata, etc.
--Beat Estermann (talk) 14:31, 30 August 2018 (UTC)
Thanks, Andy, Beat. The example was chosen because it is complex (the class and English description for fishing (Q14373) says it is an activity, but the German description says it is an economic sector) - this cannot be solved "en passant" in a straightforward way. The STW is concept-based (according to SKOS and ISO-25964), which means that the English and German labels should denominate exactly the same "unit of thought" (which, in the example, comprises the economic sector and the activities within that sector). It might be broken in some places, but it's the basic principle, and will do our best to fix errors (e.g., misleading labels or scope notes). For Wikidata, in my understanding the basic idea (one concept/item, multiple labels and descriptions in different languages) is the same, but due to historic reasons (Beat mentioned some) there are more places to fix things. In the mapping, we can take into account that Wikidata items may be ambiguous (e.g., by using a close match (Q39893184) mapping relation type (P4390), but we cannot fix more complex issues in Wikidata itself.
We can avoid said to be the same as (P460) - but is there another, more proper way to mark or report such clusters with seemingly overlapping items? Jneubert (talk) 15:28, 30 August 2018 (UTC)
Let me guess: Your dictionary entry refers to fishing industry (Q635139), as in "agriculture, fisheries, and forestry" / "Landwirtschaft, Fischerei und Forstwirtschaft". – You may have missed a WD entry. If this is the case, you should re-think your matching approach (in this case, if you match against "Fischerei" you'll get a false positive because the corresponding Wikipedia articles in German and English are not attached to the same WD item). --Beat Estermann (talk) 14:44, 30 August 2018 (UTC)
Thanks, obviously we indeed missed that one - which just adds to the complexity of the cluster. Jneubert (talk) 15:37, 30 August 2018 (UTC)
I think people are getting hung up here on what turns out to be a bad example, because in this case the answer would be, "Yes, there is a reason for all of these items to be distinct." The original question was how to mark things when there are doubts. - Jmabel (talk) 15:25, 30 August 2018 (UTC)
I don't think it is a bad example, because the various terms and definitions across languages are indeed a big mess in this case (if you were to match against the German version you would need to link to a different Q-number). Maybe you could give us a couple of further examples we can analyse in order to see whether the approach I suggested above would work for all of them. If you don't mind linking against messy data, you could just pretend that English is Wikidata's main language and ignore the other labels, outputting a list of problematic cases for the rest of the community to work on at a later point in time. If you're interested in linking against clean data, I would suggest that you invest some time in cleaning the data up before adding the STW identifiers. Having a list of problematic cases ready may help you mobilize some further helpers from the community. Linking a shitload of external identifiers (you are not the only ones) against messy data does not make much sense after all. By the way, what other data are you adding to the WD items apart from the STW identifiers? And do you have any specific use cases in mind for which your data ingest could be useful? (this could also be a good starting point for any reflections about the importance of having clean data in Wikidata in the first place.) --Beat Estermann (talk) 20:23, 30 August 2018 (UTC)
My colleague came across several such clusters of mostly rather high-level items with multiple language links. I hesitate to introduce them into this discussion, because we're not looking for a solution of the particular issues.
On the gereral level - if I didn't get you wrong - you suggested above to link in respect to language-specific labels/terms, and keep a maintenace list for exceptions. For the latter, we are currently implementing a mechanism. But that still is intended to work on a concept level, across all languages, with the main purpose of being able to provide users with more links to Wikipedia on STW pages (example). We at that point have no resources to get involved into Wikidata clean-up before linking, but would just pragmatically from our side offer links to items and WP pages which seem useful.
Sometimes cleaning up Wikidata would not work, because the mess results from reality. I brought up a - for an economics library quite fundamental - case some time ago: business administration (Q2043282), business economics (Q24208053), and business studies (Q5001951) are closely related and to some extent interchangeable concepts. In that discussion, my concern was about a in my eyes misplaced sitelink to (insofar quite similar to the above example). I was advised to take it to the talk page at the German Wikipedia (what I did), but without any response. What I also did - and what probably is of more lasting value - is interlinking the three according Englisch Wikipedia pages, which up to that point had been completely unrelated, with "See also" links, so readers and maintainers can at least easily discover the related pages.
Perhaps, a "see also" link could serve that purpose on Wikidata, too. There is see also (P1659), but this is restricted to properties. @Srittau: challenged that restriction on Property_talk:P1659, but without much response. So we're still stuck here. Jneubert (talk) 07:31, 31 August 2018 (UTC)
It seems just unlikely that someone on a Wikipedia talk will shed much light on this. At some point, one needs to determine what a Wikidata item is about and then, if some labels or sitelinks don't match that, move them to the appropriate items. Trying to fill up the Wikidata items with all subjects that might also be covered by one of the 256 Wikipedias is just not helpful.
--- Jura 18:28, 31 August 2018 (UTC)

@Pigsonthewing, Beat_Estermann, Jmabel, Jura1: Thank you for your contributions. I've removed the preliminary said to be the same as (P460) entries. Additionally, I did a bit of cleanup (moved de:Fischerei sitelink and adapted the German label of fishing (Q14373)). Since there currently seems to be no common way to mark such clusters of items, we will keep track of those which come up during the STW mapping process on our side. If we stick with the experimental publicly accessible exception list, that at least would allow others to trace dubious cases. Jneubert (talk) 16:28, 3 September 2018 (UTC)

Template:Property proposal content

Where is the boilerplate guidance text for {{Property proposal}} stored? I want to make (or propose) a change. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:35, 1 September 2018 (UTC)

Wikidata:Property proposal/Proposal preload? Template:Property proposal/doc? Matěj Suchánek (talk) 09:56, 3 September 2018 (UTC)

Two different Wikidata items referencing the same Wikipedia article

Prior to 2018-09-01 the Wikipedia article on "w:Historical Statistics of the United States" was a thinly veiled advertisement for the commercial version of this historical project that was introduced in 1949 by the US Census Bureau, then spun off to a commercial version after the 1975 edition. I added that 1975 edition to Wikidata as Q56424104 and then modified that Wikipedia article to also cite Q56424104.

However, I'm not allowed to link from Q56424104 to that Wikipedia article, because:

Could not save due to an error.
The save has failed.
The link enwiki:Historical Statistics of the United States is already used by item Q5773964. You may remove it from Q5773964 if it does not belong there or merge the items if they are about the exact same topic.

What do you suggest be done to resolve this conflict, if anything? I think it's appropriate for both Q56424104 and Q5773964 to reference this same Wikipedia article.

Other questions

  1. Is Wikidata currently inputting standard CC0 data series like the many published by the US government, to eventually make Wikidata a "one-stop shopping bazaar" for standard international data series? If the answer is "partially", what are the obstacles to faster progress?
  2. What might be an appropriate role for Wikidata for inviting the academic community to start publishing things like this within the Wikimedia system?

The US Census Bureau publishes many series monthly, quarterly or annually, as do many other agencies internationally. The US Census Bureau published the first edition of Historical Statistics of the United States in 1949 and updated it in 1954, 1960, 1965 and 1975. The 1975 edition was 2 volumes. The next update appears to have been the 2006 edition, Q5773964, which appeared in five volumes -- and is locked behind a paywall. Rather than pay Cambridge U. Pr. to access it online, I went to a local library that has it in its "reference" department. Thanks, DavidMCEddy (talk) 20:23, 1 September 2018 (UTC)

@DavidMCEddy: I'm not sure on the specifics of this case, but in general if this could be considered a single "work", there should be one Wikidata item for the work as a whole, and then separate items for each edition that is notable. The "work" item should be the one that links to the enwiki article, and the edition items would then link to the work item via edition or translation of (P629). On the other hand, if this should be considered as two (or more) separate "works" or "projects" or some other type of entity, there is at least a succession sequence, so the item for the older project should be linked to the item for the newer project via replaced by (P1366) or followed by (P156); only the item for the current state of the project would be directly linked to the enwiki item. The single-value constraint for wikidata items to wikipedia articles in a given language is strictly enforced (though there is a workaround via redirects). On your other question, if I understand it correctly, there are structural limitations to how many statements can be attached to an item in Wikidata - more than 1000 and the user interface becomes pretty unworkable - but we also allow linking to tabular data in the Wikimedia Commons project, so large statistical datasets would probably be best accommodated that way - as long as they meet the other requirements (like CC0 compatability) for import. ArthurPSmith (talk) 12:30, 4 September 2018 (UTC)

I would like to create a report on new articles: users, wordcount, categories, bytes for a specific wikiproject

I know that I can find out information about wikipedia articles (wikidata items) and their properties on Wikidata. I would like to find information on user contributions per month for a specific Wikiproject. This would help with wikiproject planning and management. I am not a developer. Can you point me in the right direction if Wikidata is not the place to get that info? thx MauraWen (talk) 15:49, 2 September 2018 (UTC)

@MauraWen: Wikidata does not carry metadata about user actions in other wikimedia projects; the data in Wikidata is about the concepts that are described in articles, not about the articles themselves. What you are looking for may be available however from the Wikimedia Statistics pages; please check there. ArthurPSmith (talk) 12:55, 4 September 2018 (UTC)
thx @ArthurPSmith: for answering my question. I eventually figured out that this is a metadata issue, but glad to hear Wikimedia statistics is the best place to get that information. MauraWen (talk) 13:10, 4 September 2018 (UTC)

What is the URL of the RDF representation a property?

For example the property P279 (subclass of):

According to usage in other entities, it should be

But there no RDF content obtained at this URL by content negotiation, only HTML is gotten:

wget --save-headers --header='Accept: text/turtle'


wget --save-headers --header='Accept: application/rdf+xml'
wget --save-headers --header='Accept: application/ld+json'

My use case

I have a web tool that is completely generic about LOD data source , for example here is what I get from a Wikidata URI :

 – The preceding unsigned comment was added by Jmvanel (talk • contribs) at 07:55, 3 September 2018‎ (UTC).

Thanks to Mister Synergy for the Special:EntityData URL's , that are documented at the right place in the doc. . This answers to the original question in the title.

However, there is still an issue with respect to Semantic Web good practices. Namely, that an URI should be dereferentiable and subjected to RDF/XML, Turtle, and HTML content negotiation . This indeed is the case for URL's with prefix . But properties, as they are used in triples, are not. For instance, this entity uses a property with URL . But this latter URL only provides HTML, as demonstrated above. Instead, this well behaved URL should be used in Wikidata RDF data: . It is probably the same with every property; prefix should be used . But it would be probably more coherent if would be used. It is also a well behaved URL prefix.

MisterSynergy (talk) 09:04, 3 September 2018 (UTC)
The canonical URI for the property with ID P279 is just like the canonical URI for the item with ID Q42:, abbreviated wd:P279. wdt:P279 (…/prop/direct/P279) is just one of several predicates related to that property (there’s also /prop/, /prop/statement/, /prop/qualifier/value, etc.) – all of these predicates have different meanings, and we can’t just replace them all with the entity URI because that one happens to resolve better for certain use-cases. However, we can change the other URIs to be more well-behaved – I’ve filed phabricator:T203400 for this. --Lucas Werkmeister (WMDE) (talk) 14:33, 3 September 2018 (UTC)

Thanks for filing an issue, Lucas; since it's not clear to me , shall we continue discussion here or in phabricator:T203400 ?  – The preceding unsigned comment was added by Jmvanel (talk • contribs) at 17:21, 3 September 2018‎ (UTC).

@Jmvanel: I guess it’s easier to do it here, if you have any more comments. --Lucas Werkmeister (WMDE) (talk) 10:34, 4 September 2018 (UTC)

P248 instead of P143

Hi, I noticed that a large number of items use stated in (P248) to cite Wikipedias, instead of using imported from Wikimedia project (P143) just to notice that the information comes from there (unnecessary IMO if there exists a valid source). Wikipedias are not considered valid sources: Help:Sources#Different_types_of_sources. It was noticed in Greek Wikipedia that in relevant articles these wikipedias are cited in the references section through automated infoboxes. I believe that a bot that will change all these stated in (P248)  English Wikipedia (Q328) to imported from Wikimedia project (P143)  English Wikipedia (Q328) would be of great value. -Geraki (talk) 13:01, 3 September 2018 (UTC)

There are a little more than 14.000 of such references (query1, query2). I can fix them later this day. —MisterSynergy (talk) 13:17, 3 September 2018 (UTC)
All done now. The final ~200 cases where cluttered so much that I had to repair them manually after the bot did most of the work :-/ —MisterSynergy (talk) 22:54, 3 September 2018 (UTC)

It would be nice to make this a constraint violation for stated in (P248) P.a.a (talk) 15:59, 3 September 2018 (UTC)

  • As I understand it, "stated in" is correct for manual addition of an entry, "imported from" when scraped by a bot. Is that not correct? - Jmabel (talk) 16:20, 3 September 2018 (UTC)
    • No, this is not correct (although the English description of imported from Wikimedia project (P143) might indicate it). Wikipedias typically discard imported from Wikimedia project (P143) references as “not considered valid sources” (as indicated above), and consider stated in (P248) references to be always serious. It does not matter whether the referenced value was added by hand or bot. You can thus import from Wikipedia by hand, and indicate it with imported from Wikimedia project (P143) in the reference. —MisterSynergy (talk) 16:46, 3 September 2018 (UTC)
      • And the Spanish says the same as the English, and both are rather specific and prescriptive about it. If it is wrong, them someone should certainly change it. - Jmabel (talk) 19:43, 3 September 2018 (UTC)
        • English description is changed; someone proficient in Spanish could do the same for the other label. —MisterSynergy (talk) 22:54, 3 September 2018 (UTC)

Good work MisterSynergy! Thanks. -Geraki (talk) 16:36, 4 September 2018 (UTC)

Wikidata weekly summary #328

Starting Wikidata Query Library

Yes we have Query Examples but it is a very good idea to start a Portal or Libaray that store all kind of queries sorted by subject, Category etc.. and Featured query, Start your first query, New queries, Request for a query etc.. Can we do a query contest? --Ranjithsiji (talk) 10:29, 4 September 2018 (UTC)

There's a parallel discussion on Facebook, where a Query: namespace has been proposed, with one query per page, using Wikipedia- (or Commons-) style categories to aid findability. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:59, 4 September 2018 (UTC)
Not sure if this has been raised elsewhere, but a "Query" namespace has been part of Wikidata for a long time - but it's unused because the original approach to querying was abandoned in favor of the SPARQL service we have now. See this page for some details and links; you can also see a trace of the Query namespace when you do an advanced search, it's one of the options (but there's nothing there). ArthurPSmith (talk) 17:50, 4 September 2018 (UTC)

Do you speak German/French? Help WLM 2018 Switzerland!

Hello guys!

Do you see these giant "de" and "fr" bubbles? If you understand German/French, you can be very useful in helping Wiki Loves Monuments in Switzerland in translating as monuments as possible in English. We want to see the "en" bubble growing!

Even providing an English label to just 2-3 monuments will be very appreciated... but pay attention: it's a drug! :^) Thanks you everybody and have fun! --Valerio Bozzolan (talk) 15:16, 4 September 2018 (UTC)

Implementation of a new Wikidata-to-OSM service for P402

Please, do you support the implementation of a new Wikidata-to-OSM service for OSM relation ID (P402)?

Explaining. The new service will solve the cited problem of "only relations": will be possible to link OSM Map Features represented by nodes or ways.

The new server must be hosted at the Openstreetmap-side, but needs the support of the Wikidata community also. The service will redirect Wikidata item "Q" identidiers by$1 (or or other) to the respective geometry. Examples:

  • Relation. The Wikidata item that represents the concept of the country Brazil (Q155), on OSM is a relation, will redirect to
  • Way. The Wikidata item that represents the concept of the Fraternity Bridge (Q2679759), on OSM is a way, will redirect to
  • Node. Wikidata item of the Number zero survey marker of the city of São Paulo, (Q10325364), on OSM is a node, will redirect to

See more details at the Proposal-QID. --Krauss (talk) 13:39, 29 August 2018 (UTC)

  • It looks like a nice idea. Have you thought about hosting it on Toolforge? NMaia (talk) 15:47, 29 August 2018 (UTC)
@Samwilson: You can indeed use toolforge:hub to generate URLs to OSM relations from Wikidata IDs (example: ) but it doesn't work for ways or nodes as we don't have properties for those (right?). Note that if we had those properties, the link could simply include the different possible properties (example:,PXXXX,PYYYY )-- Maxlath (talk) 11:11, 5 September 2018 (UTC)
  • I don't really understand what this is intended for. What does it have to do with OSM relation ID (P402)? What will you do when multiple OSM objects have the same Wikidata ID? Personally I would really like to see OSM support searching by Wikidata IDs, so that we can enter Qxxx or wikidata=Qxxx into the search box, that would give us short copyable URLs as a side effect. - Nikki (talk) 12:47, 30 August 2018 (UTC)

The OSM assumption that Wikidata Q-numbers represent "unique and eternal" identifiers for their objects is blatantly wrong. That's why they need to implement unique IDs for their elements, so we can manage the links to their elements from within Wikidata. This issue concerns for example all the items of museums, archives, libraries, cultural venues and the like which confound the geographical object and the organization. This typically is the case for Wikidata items that were created based on Wikipedia entries and their categories (i.e. tens of thousands of items). On Wikidata, we will end up disentangling these items and their statements, which will "break" the OSM-WD link that they have set on the OSM side with 50% probability. That's why we should be able to assign the OSM-ID explicitly on the WD-side and have an automatized service in place that handles the discrepancies between OSM and WD whenever objects and identifiers get moved around on the WD or the OSM side. --Beat Estermann (talk) 13:53, 30 August 2018 (UTC)
@Beat Estermann: "unique and eternal" is indeed strange but I'm not sure it's « blatantly wrong ». At least a Qid is unique, that's for sure and most of the item it's perennial (but indeed not eternal), especially if you keep in mind that OSM have current objects.
For the GLAMs who are a geographical object and organization, that's why I'm often creating two items, one for each (and I do the same on OSM too). Or splitting when there is only one item and I need to put separate informations (typically for inception (P571) who is the date of construction in the context of a building and the date of creation for an organisation).
For assigning from the WD side it would be ideal, but how to do? As a reminder, when you edit an OSM object, the identifier can change too (which is quite painful... in this sense, it's very understandable than an OSMer thinks that Qid are "eternal").
Cdlt, VIGNERON (talk) 19:23, 30 August 2018 (UTC)
Using Wikidata IDs on OSM objects is the currently best way to guarantee a link between OSM objects and Wikidata concepts. The 'burden' of maintenance is on editing the OSM objects, like it is done for transit relations. OSM relation ID (P402) is not the best way to have a permanent correlation because also relations can break and change ids in OSM. The "permanent id" discussion can be solved by a service, yes, but to actually solve it the service would need to be as much automatic as possible, otherwise it's simply a case of moving the problem around. Perhaps the OSM API can be extended to deal with this. --Sabas88 (talk) 09:14, 4 September 2018 (UTC)

Is it possible to add ebird data to Wikidata?

Hello everyone, eBird is a website by Cornell Lab of Ornithology. eBird has a lot of data about birds. If we can able to add them to Wikidata it would be a great achievement for the birders. Because through query service we can find exactly how many number of birds found (from eBird checklist) in this area in so and so the duration of time. I had shared this with Ijon. He suggested me to write here and ping Pigsonthewing. Let me know if it is possible. I can write to contact Cornell Lab also. --Gopala Krishna A (talk) 03:56, 4 September 2018 (UTC)

Note we have eBird taxon ID (P3444) and eBird hotspot ID (P5200) which can link wikidata items to the eBird website (and about 13,000 taxa have already been linked). Pulling in all their data is beyond the scope of Wikidata; it may also not be legally allowable - I couldn't find a rights statement on the eBird site, but Wikidata data is CC-0 so the licensing would need to be compatible. I'm also not sure the data structure of this site is right for this kind of information - you would presumably have to create a statement on each bird item for every observation (with location, date, count)? This might be something that could be done via tabular data in Commons however. ArthurPSmith (talk) 13:10, 4 September 2018 (UTC)
Further to Arthur's very good reply; there's not much point in us trying to replicate what eBird do. In time, it's to be hoped that services like eBird will see the value in making their data available as linked-, open-, data, and providing a SPARQL end-point, so that people can make federated (combined) queries across their database and ours. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:41, 4 September 2018 (UTC)
Thanks everyone for your valuable time and reply. I would also recommend eBird to start a SPRQL based query search on their website. --Gopala Krishna A (talk) 08:03, 5 September 2018 (UTC)

contains administrative territorial entity: value type constraint

Hello. Limassol Municipality (Q28870916) -> contains administrative territorial entity (P150). It shows value type constraint (Q21510865). Before some time, there was not such problem. I guess something changed with the value of a property of Quarter of Limassol Municipality (Q29463880). Like third-level administrative country subdivision (Q13221722) or third-level administrative country subdivision (Q13221722) or about an item which those two are subclass. Because I have created many items for third-level administrative country subdivision for Cyprus, I want to know what to do, so the structure will be correct. Xaris333 (talk) 09:47, 5 September 2018 (UTC)

  • If what you are writing is correct, it should be subclass, not an instance. Fixed that > works now. --- Jura 11:08, 5 September 2018 (UTC)

@Jura1: thanks. But how can I connect Quarter of Limassol Municipality (Q29463880) with Limassol Municipality (Q28870916)? I did it with located in the administrative territorial entity (P131) but I get item requires statement constraint (Q21503247). Xaris333 (talk) 11:31, 5 September 2018 (UTC)

  • Maybe the "designation"-P31 should be used there .. Also, I don't think P131 should be used for that. I changed it to "part of". --- Jura 11:35, 5 September 2018 (UTC)

Ok for part of, but I don't think P31 -> designation for an administrative territorial entity (Q15617994) is correct. Xaris333 (talk) 12:34, 5 September 2018 (UTC)

Please Merge

10 minutes in and I'm still at a loss trying to learn how to easily merge two items. This is my first time using or even looking at WikiData and I'm not even sure if I'll return and I hope you'll forgive my laziness. Someone please merge Karlis Johansons (Q56226186) with Kārlis Johansons (Q1387109). I created the English stub using Google Translate from the Latvian stub. Thanks in advance. ~ JasonCarswell (talk)

Already done by Mahir256. --Liuxinyu970226 (talk) 10:44, 11 September 2018 (UTC)
This section was archived on a request by: Liuxinyu970226 (talk) 10:44, 11 September 2018 (UTC)

Sourced invalid statements

I have seen many cases like this:

< Temple of Athena Nike (Q384813) View with Reasonator View with SQID > located in the administrative territorial entity (P131) View with SQID < Acropolis of Athens (Q131013) View with Reasonator View with SQID >

Obviously Acropolis of Athens (Q131013) is not an administrative unit. But there is a reference for this statement, although not an authoritative one: . In fact even the source does not support the claim, it lists the Acropolis as "sonstige Fläche" (to my understanding it means "other"?) Another example would be the statement that an uninhabited islet is a "settlement" just because it is included in census records with 0 population (in fact that would be proof that it is not a settlement). What do you think about the removal of sourced statements that contradict themselves, other sources or common sense? -Geraki (talk) 17:05, 4 September 2018 (UTC)

I guess the statement with P131 should be

< Temple of Athena Nike (Q384813) View with Reasonator View with SQID > located in the administrative territorial entity (P131) View with SQID < Acropolis (neighborhood) (Q2336187) View with Reasonator View with SQID >

. Xaris333 (talk) 09:53, 5 September 2018 (UTC)

Nope, that is still a neighborhood. Administrative units in Greece start at the municipality level. -Geraki (talk) 20:41, 5 September 2018 (UTC)
Xaris333 (talk) 22:02, 5 September 2018 (UTC)

About the 0 population: if an administrative territorial entity has 0 popolulation, is not a problem. The administrative territorial entity is still exist as a legal entity, even thought no one is living there at the moment. Xaris333 (talk) 09:52, 5 September 2018 (UTC)

I am sure I did not use the wording "administrative territorial entity" but "uninhabited islet" and "settlement" (which by definition means an inhabited/populated location). A location with no population presently and in the past, is by definition not a human settlement (Q486972)... -Geraki (talk) 20:41, 5 September 2018 (UTC)
You spoke about settlement that it is included in census records with 0 population. I didn't know that countries use no administrative territorial entities at their census. Usually a municipality is slpited to quarters, parishes, etc... Xaris333 (talk) 22:05, 5 September 2018 (UTC)
The U.S. uses many non-administrative places in their census. See en:Census-designated place. - Jmabel (talk) 04:28, 6 September 2018 (UTC)
I spoke about uninhabited islet that it is included in census records with 0 population, and the conception it is a human settlement (Q486972) (which is not an "administrative territorial entity" per se). There may be some administrative territorial entities called "settlements" in some countries but that would be different items. Do not forget that there are settlements that are not included individually in census records but they are still settlements, and the concept of "administrative entities" is somehow modern and with completely different scope from the common definition of human settlement (Q486972). -Geraki (talk) 08:24, 6 September 2018 (UTC)
  • Geraki, please don't forget to advise the contributor about this. @Arch2all:--- Jura 13:04, 6 September 2018 (UTC)

Odd identifier, how to do it

There is a proposal for an odd numerical identifier. It is odd because it is used, but it can't be used in an URL in the usual way.

The proposed identifier is Wikidata:Property proposal/Norwegian war refugees register ID. The ids are numbers assigned to war refugees in Sweden during WWII. The best source for them are at Arkivverket Digitalarkivet: Den Kgl. Norske Legasjons Flyktningskontor*, but that is an archive of scanned papers. That is, plain old paper trail. Some pointers; National Archival Services of Norway (Q6516420), and Digital Archives (Q11154625).

I wonder if this is a kind of identifier that lacks an inherent source, but can be given a source. To make an example; Norwegian war sailor register person ID (P4972) has an inherent source as it can create links by using formatter URL (P1630), that is$1. The Norwegian war refugees register ID would not be able to use a link like that, but it can use an "URL-source" that points to a source property. That URL-source can then be appended a hash (#) and the identifier, like described for "cool URIs".[1]

For an entry about "Fjellbu, Arne" we would then give a source URL as, which would then be used to rewrite the for the identifier like [ 32915](or simply 32915). We link to the source, but only add the id as a fragment identifier. In the general case we must probably adjust the link somewhat because there might be specific anchors we would like to hit.

I believe this could solve a lot of similar problems, without to much clever hacks, but the problem remains – shall we do it? Perhaps there is already a better way to do it? Jeblad (talk) 19:05, 5 September 2018 (UTC)

Read-only mode for up to an hour on 12 September and 10 October

13:33, 6 September 2018 (UTC)

Please split and merge

I guess that no label (Q54859371) should be split and some part merge/move with no label (Q54006375) or no label (Q48246688). And maybe deleted, see w:en:Wikipedia:Articles for deletion/MJ Tamondong and w:en:Wikipedia:Sockpuppet investigations/Matrix 2. Visite fortuitement prolongée (talk) 20:37, 10 September 2018 (UTC)

Probably best to delete all three, I think. Andrew Gray (talk) 21:08, 10 September 2018 (UTC)
✓ Done Mahir256 (talk) 21:30, 10 September 2018 (UTC)
This section was archived on a request by: Visite fortuitement prolongée (talk) 20:40, 12 September 2018 (UTC)

Why does this not work?

The collection list Wikidata:WikiProject sum of all paintings/Collection/Shchusev State Museum of Architecture doesn't work. There are items in the Museum, see Q56596912 or Q56596977.--Carl Ha (talk) 08:19, 12 September 2018 (UTC)

  • @Carl Ha: ✓ fixed --- Jura 10:07, 12 September 2018 (UTC)

This section was archived on a request by: --- Jura 10:07, 12 September 2018 (UTC)

Revert merge

Can someone revert my item merge? Ilham151096 (talk) 08:20, 12 September 2018 (UTC)

This section was archived on a request by: --- Jura 10:01, 12 September 2018 (UTC)

Link to commons wiki and conflict between gallery and category

Hi could you please remind me what is the standard procedure when I need to merge items such as Q2557858 and Q21162042? Shlìuld I remove the link to the category, keep the gallery in the final item.

Please consider that the majority of it-N users are quite practical nowadays, they basically avoid the creation of galleries on commons, they consider them difficult to update and time consuming, if they invest on something are the categories so most of the work of improvement of local items ends up connecting to commons categories.--Alexmar983 (talk) 19:24, 24 August 2018 (UTC)

If there's a Commons gallery, sitelink to it from the main item.
If there's also a Commons category, create a separate item for it with a label starting "Category:", and relate it to the first item by category's main topic (P301)/topic's main category (P910) in each direction.
Also add Commons gallery (P935) statements on both items linking to the Commons gallery, and Commons category (P373) statements pointing to the Commons category. On most Wikipedias, the software will follow these, in preference the sitelinks, so adding the P373 will be enough to make the sidebar link point to the Commons category.
On the Commons category, use c:Template:Wikidata infobox to create an infobox. This should automatically follow the sitelink and then the P301 to draw data from the main item.
A bit of a chore, yes; but it's the system we've come to, and it seems to work reasonably enough.
If there is no Commons gallery, and there is no category item already in existence (eg linked to a Wikipedia category, or to a category on a sister project), then it is acceptable to follow the simpler strategy of linking the Commons category straight to the main item. But this should be done only when the two conditions above are met; otherwise the full rigmarole as previous described is necessary. Jheald (talk) 23:11, 24 August 2018 (UTC)
Creating a new item for the category, with only a Commons sitelink, isn't permitted by Wikidata:Notability. It would be better to keep the category sitelink than replacing it with the gallery, in my opinion. But we've been through all this: I suggested a new policy of giving Commons sitelinks priority over galleries. But people said why not just make a category item for the sitelink, and if the notability policy forbids it, change the policy. So I proposed changing the notability policy, but it got no consensus. Ghouston (talk) 01:22, 25 August 2018 (UTC)
That restriction in WD:N is history. We now have thousands of items with just a wikilink to a Commons category, supporting an infobox there, and nobody is going to delete them. WD:N hasn't been changed yet, but the clause has a footnote noting that its continued application is questionable and pointing to open discussion on the talkpage. In practice the restrictive clause is dead, the message just hasn't reached the brain yet. Jheald (talk) 08:28, 25 August 2018 (UTC)
ok but this is getting confusing. I mean, I need to give newbies very clear tasks. And from this point of view, as long as the information is clear it can always be fixed later by bot. In any case the level of sophistication required by current users is increasing. They expect wikidata infobox and structured data more and more. So just to clear, current consensus is that i put the gallery not the category in the "other projects" section and I don't merge?--Alexmar983 (talk) 19:47, 25 August 2018 (UTC)--Alexmar983 (talk) 19:47, 25 August 2018 (UTC)
The two items you mentioned should be merged, since neither is a category item. Personally, I'd just leave the Commons category sitelink in place and add a Commons gallery (P935) statement for the gallery. Ghouston (talk) 00:11, 26 August 2018 (UTC)
The gallery hadn't been properly updated since 2011 (and only had low-resolution images when more + higher resolution ones were in the category), so I've redirected it to the category and merged the two items. thanks. Mike Peel (talk) 17:38, 26 August 2018 (UTC)
Wikidata infobox has a "qid" attribute which allows to draw in the information from whatever data item you assign it. I have used it on several categories leaving the link open for a gallery if and when they exist. It fixed a problem at source where the template assumes the link is a gallery and also pulls in the category if it is assigned. Since then, a bot has come through and re-assigned the link to the category and it is (as far as I know) simply not needed and ugly at source. It is an old and unnecessary war which I personally will defer to the gallery makers because the good galleries take a lot more work. The galleries with only one little fuzzy image on them are, I think, a response to being told to make the gallery, where the nice galleries are made out of choice and inspiration. Please leave this war with the galleries being the "winner". Also, the Category hack was beautiful, it worked just fine but I am not happy with the additional data entry when software can handle it just fine as it is. Having to choose gallery or category is a terrible thing. I did it a few times and I did not like it.--RaboKarbakian (talk) 20:49, 1 September 2018 (UTC)
There is w:Raising the Flag on Iwo Jima and c:Category:Raising the Flag on Iwo Jima on Q117693. Now I created c:Raising the Flag on Iwo Jima but interwiki doesn't work. If I change the sitelink I'll break the Wikidata infobox I think. @Mike Peel: now what? Alexis Jazz (talk) 16:00, 7 September 2018 (UTC)
@Alexis Jazz: It looks like you created a page that should be in the Commons: namespace, not a gallery. Most of it is about the copyright status, and the part that is a gallery does not look useful to me. So I'd suggest moving it to c:Commons:Raising the Flag on Iwo Jima, and maybe creating a new item here with instance of (P31)=Wikimedia project page (Q14204246). Thanks. Mike Peel (talk) 16:24, 7 September 2018 (UTC)
@Mike Peel: I wasn't sure if it should be in the Commons or gallery namespace. It doesn't really answer the question of what to do with gallery pages. Wikidata infobox needs the sitelink, so a gallery can never be sitelinked? Alexis Jazz (talk) 16:35, 7 September 2018 (UTC)
Personally, I'd first ask whether the gallery is actually useful, or if it should just be deleted/redirected to the category. If it is useful, then create a new item with instance of (P31)=Wikimedia category (Q4167836), set category's main topic (P301) and topic's main category (P910) to link the two items, move the category sitelink to the new category item, and add the gallery sitelink to the item. Thanks. Mike Peel (talk) 16:38, 7 September 2018 (UTC)

Wikidata item popup preview does not work?

I put my mouse on an link to an wikidata item, but in the preview window, what it show is only . What's wrong with it?C933103 (talk) 04:32, 1 September 2018 (UTC)

Which gadget does this? Matěj Suchánek (talk) 10:01, 3 September 2018 (UTC)
MediaWiki:Gadget-popups.js ? C933103 (talk) 11:43, 7 September 2018 (UTC)


Do we have a property to indicate the biome (Q101998) or ecosystem associated with a place (park, nature preserve, or any similar)? I'd like to indicate that the biome found at Konza Prairie (Q6430461) (<instance of> research station) is tallgrass prairie (Q7680132). If not, I'll propose a new property for "biome, ecosystem, or habitat". - PKM (talk) 20:30, 6 September 2018 (UTC)

Javascript hooks are now stable

Hello all,

In order to allow better gadget integration, JavaScript hooks documented in the hooks-js.txt file delivered together with Wikibase source code are considered stable. We just added this information in the stable interface policy.

If you have any question about this, feel free to ping me. Lea Lacroix (WMDE) (talk) 16:22, 7 September 2018 (UTC)

Lua error

I created this property proposal page, but it gives me this Lua error: Lua error in Module:Property_proposal at line 219: bad argument #1 to 'gsub' (string expected, got nil). Why? --Malore (talk) 18:30, 10 September 2018 (UTC)

Fixed.--GZWDer (talk) 19:38, 10 September 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 08:47, 13 September 2018 (UTC)

Renaming of wikidata element needed

Hi everyone,

I didn't manage to do it myself neither to find where to file a request for renaming. The issue is explained here, the wikidata element Q17269 should be reverted to its original name (Tibet Autonomous Region) to avoid confusion with element Q17252 (Tibet). Thanks for any support. --Dominique Roux (talk) 19:11, 12 September 2018 (UTC)

Sorry, the title seems to have changed automatically after editing the English label, it just needed some time to reflect the change. Now item Q17269 appears as Tibet Autonomous Region, problem solved.--Dominique Roux (talk) 19:17, 12 September 2018 (UTC)
This section was archived on a request by: Mbch331 (talk) 09:03, 13 September 2018 (UTC)

changed name

What can he do? Regards, Conny (talk) 19:57, 7 September 2018 (UTC).

Q7542#P1828 and Property:P1828#P2302 (add all three and a constraint exception) —MisterSynergy (talk) 20:28, 7 September 2018 (UTC)

Q355304 class hierarchy

watercourse (Q355304)

According to @Мастер теней:, a watercourse should be directly a subclass of no label (Q4114033), omitting the previously used body of water (Q15324) intermediate step. I don't see where is the actual difference between these two concepts anyway, I only noticed it because this change breaks many constraints which rely on Q15324. Can someone who is more fluent in the ontology concept have a look into this? Ahoerstemeier (talk) 23:25, 7 September 2018 (UTC)

How to model a museum label as a reference?

How should I model a museum label (Q6940727) as a reference? Using stated in (P248): museum label (Q6940727) raises a violation of this value type constraint because museum label (Q6940727) isn't an instance but a class and I don't plan to create an item for each museum label. Change the constraint? Any ideas? Thanks a lot in advance! --Marsupium (talk) 08:18, 14 September 2018 (UTC)

Maybe use the item for the museum itself for the value of stated in (P248), and maybe add a quote with the label text. Using museum label (Q6940727) as the value would be like using book (Q571), not very specific. Ghouston (talk) 08:40, 14 September 2018 (UTC)
I suppose a museum may not be an acceptable value either. Ghouston (talk) 08:40, 14 September 2018 (UTC)
@Ghouston: Hm, okay, both does not seem to fit too well. Perhaps one could indicate the location somehow? Or we need a property like "source is instance of"? --Marsupium (talk) 15:27, 14 September 2018 (UTC)
Haha, #Tombstone as information only three sections beneath this gives the solution: type of reference (P3865). Continuing there … --Marsupium (talk) 15:53, 14 September 2018 (UTC)
This section was archived on a request by: Marsupium (talk) 15:53, 14 September 2018 (UTC)

Wiktionary links for Hangul syllables/Unicode characters

Just a reminder that links to Wiktionary pages in the main namespace are disallowed on Wikidata. Please remove those which you have already added, whether to items about Hangul syllables or to other items, and do not add any more until that restriction is lifted, if that restriction is ever lifted someday. Mahir256 (talk) 19:30, 31 August 2018 (UTC)

@Mahir256: Thanks for the reminder. These statements have been removed. @Okkn: I think an exception should be made for Unicode characters to be linkable to Wiktionary pages. This could be discussed by the administrators. Wiktionary contains descriptive information and users may visit it for more information. KevinUp (talk) 19:48, 31 August 2018 (UTC)
@Mahir256: I understand we usually don't add sitelinks to Wiktionary pages in the main namespace, because they are not concepts, but is it really not allowed to add Wiktionary sitelinks? I think WD:N just says Wiktionary sitelinks are not valid to fulfill the notability criteria, but does not say we must not add Wiktionary pages to already created items, right? When it comes to Unicode characters, including Hangul character (Q55809450) and Chinese character (Q53764738), they are clearly notable by themselves, and they are concepts rather than representation. Or if we add sitelinks to Wiktionary pages, the function of mw:Extension:Cognate does not work...? --Okkn (talk) 01:30, 1 September 2018 (UTC)
That makes sense. If links to Wiktionary pages are disallowed on Wikidata, then the input box that allows users to add Wiktionary links should be removed from Wikidata. KevinUp (talk) 05:09, 1 September 2018 (UTC)
The restriction only applies to Wiktionary pages in the main namespace, as I stated initially; links to categories/templates/Wiktionary: pages/etc. (except those in the citation, user, and all talk namespaces) are allowed. Mahir256 (talk) 21:15, 1 September 2018 (UTC)
This is a matter for the community at large to decide; not just administrators. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:43, 2 September 2018 (UTC)
Is it possible for us to take a vote regarding this matter? KevinUp (talk) 16:43, 2 September 2018 (UTC)

The discussion above was from my talk page. I wish to gain wider consensus regarding this matter. As mentioned by User:Okkn, Unicode characters such as Hangul character (Q55809450) and Chinese character (Q53764738) are clearly notable by themselves, and are concepts rather than representation. Does this affect the function of mw:Extension:Cognate? KevinUp (talk) 08:18, 2 September 2018 (UTC)

Also, I have begun to link Unicode characters on Wiktionary to Wikidata. Wikidata is useful for obtaining structural data regarding Unicode characters while Wiktionary is useful for obtaining the meaning and usage of Unicode characters. KevinUp (talk) 08:18, 2 September 2018 (UTC)

WD:N state that the reason why wiktionary pages are excluded because wiktionary interlanguage links are done by Extension:Cognate with the reason on the extension page saying that the use of this extension is because wiktionary links are based on homograph, not concept. However, this assessment does not apply to each individual Unicode characters, since each individual Unicode characters are a concept in themselves, with many different data for each individual unicode code points. Especially for Pan-CJK characters. You can click into any wiktionary page about each individual CJK characters page and see how many values there that can be stored centrally, and there's even a non-WMF wiki known as glyphwiki that store even more info about individual CJK character codepoints. Because of such a number of data that can be centrally managed at wikidata, and because each Unicode code point are representing the same, single concept of a single Unicode code point instead of being merely homonym, I suggest changing the policy to allow the storage of those data within wikidata. C933103 (talk) 14:31, 2 September 2018 (UTC)

Well said indeed. Recently, Wikidata:WikiProject CJKV character has been created to provide structural data for the 88,889 CJKV characters available in Unicode 11.0. Prior to this, only a small number of Unicode items exist. For comparison, please compare (Q3595028) with wikt:雨. KevinUp (talk) 16:43, 2 September 2018 (UTC)
I believe the restriction was a request from the Wikidata developers, not something that community consensus can address. @Lea Lacroix (WMDE), Lydia Pintscher (WMDE): can you comment? ArthurPSmith (talk) 12:38, 4 September 2018 (UTC)
The reason we asked for this is that we have seen a huge amount of confusion because people didn't understand the difference between concepts and words - in essence the difference between Items and Lexemes here. Often it is not understood how the linking done in Wiktionary's main namespace is different from the one in Wikipedia's main namespace for example. Now there might be rare cases where it still makes sense to add Wiktionary main namespace pages to items as sitelinks. (And the one you mention here looks like it might be one of them.) For those it is totally fine to make an exception from my side. --Lydia Pintscher (WMDE) (talk) 19:28, 5 September 2018 (UTC)
  • Wouldn't they better fit directly in Wikidata's dictionary namespace? --- Jura 04:26, 10 September 2018 (UTC)

Another type of wikidata concept that I believe should have wiktionary links: chengyu (Q28060424)+Yojijukugo (Q7010524)+Sajasungoh (Q17017012) (actually why is there at least three different wikidata item for same thing that exists in different language/geographical area), with examples like no label (Q5356314). C933103 (talk) 14:22, 9 September 2018 (UTC)

Item merger


I think I spoted two idems which should be merged, but don't know how to do it. I'm basing this on the fact that the Wikipedia pages are about the same topic. I assume that it got fragmated at some point.

If someone could smush these two togeter it would be helpful. Thanks.

--2A00:EE2:204:7500:D461:77CF:8760:7131 19:40, 3 September 2018 (UTC)

Silver tax 1816

What property should be used to indicate that a person contributed to the Silver tax 1816 (Q25429073). I can find participant of (P1344), subject has role (P2868), sponsor (P859) and may be contributor(s) to the creative work (P767). Making this diffcult is that it was a forced tax and I can not see that the Properties mentioned will cover a forcement.

The Silver Tax 1816 was established by the newly created independent Norwegian state (1814). It is a common name for the compulsory deposit, which was issued in 1816, to acquire a foundation fund of 2 million specidaler for the newly created Norges Bank (Q970769). Pmt (talk) 06:05, 8 September 2018 (UTC)

Open Location Codes aka Plus Codes

Is wikidata has a P for Open Location Codes. Any license or compatibility issues with that ? Can we incorporate that ? --Ranjithsiji (talk) 06:51, 8 September 2018 (UTC)

We had a discussion about Plus codes earlier this year but there was no consensus to create a property for it, see Wikidata:Property proposal/Plus codes --Pasleim (talk) 12:13, 8 September 2018 (UTC)
Perhaps a gadget that supplemented the coordinate entry box with a conversion tool from Plus codes back to WGS84 coordinates would be a nice little task for the Code-In? (Similarly for what3words codes or other sorts of coordinate references?) Mahir256 (talk) 22:09, 9 September 2018 (UTC)
I support @Mahir256: it is a necessary one. How we can implement that. --Ranjithsiji (talk) 05:38, 10 September 2018 (UTC)

Puerto Rican nationality

Today, I have removed the country of citizenship property on many Puerto Rican people items, because they wrongly value "Puerto Rican" as the nationality, but it should be "American". AFAIK, Puerto Ricans are citizens of the U.S, althought to get real citizenship (such as voting) they must live in the mainland. When valued "Puerto Rican", the property displays a message error. There must be many more items with this mistake, and I think a bot could easily fix them. Esteban16 (talk) 00:20, 30 August 2018 (UTC)

  • This query gives 1009 items which are citizen of Puerto Rico. Perhaps some of them predate the US takeover in 1899. Before that they would have been Spanish? Ghouston (talk) 01:35, 30 August 2018 (UTC)
  • It would probably be better to change all the citizenship claims to residence (P551) of Puerto Rico, instead of just deleting the link with Puerto Rico. It may not be correct in every case, but it's a start. Ghouston (talk) 01:40, 30 August 2018 (UTC)
Ghouston: That is a good idea, too. And yes, they had Spanish citizenship for a while. The Jones–Shafroth Act granted Puerto Ricans American citizenship on 1917. Esteban16 (talk) 02:07, 30 August 2018 (UTC)

@Esteban16: However, according to Wikipedia article w:en:Puerto_Rican_citizenship, "The United States government also continues to recognize a Puerto Rican nationality". @Ghouston: I don't think it is a must for everyone with Puerto Rican nationality to keep living in Puerto Rico? C933103 (talk) 02:25, 30 August 2018 (UTC)

  • It's definitely not "a must". I would think Puerto Rican citizenship should be recorded in Wikidata, although possibly (from 1917 onward) in conjunction with U.S. citizenship. Jmabel (talk) 03:29, 30 August 2018 (UTC)
  • The article also says Spain continues to grant citizenship to Puerto Rico citizens. Interesting. Ghouston (talk) 06:32, 30 August 2018 (UTC)
Then it seems removing the property from some items was wrong, and I think as same as User:Jmabel, the nationality should be recorderd. Esteban16 (talk) 11:04, 30 August 2018 (UTC)
w:en:Puerto_Rican_citizenship is a seriously flawed article, and should be ignored. I don't know what the situation is for Puerto Rico citizenship (if it currently exists), but for US states, the 14th Amendment to the United States Constitution states in part "All persons born or naturalized in the United States, and subject to the jurisdiction thereof, are citizens of the United States and of the State wherein they reside." So if a US citizen resident permanently moved from Puerto Rico to a state, she would become a citizen of that state. If a US citizen resident of a state permanently moved to Puerto Rico, he would cease to be a citizen of the state. Citizenship in a state is not a nationality. Jc3s5h (talk) 13:07, 30 August 2018 (UTC)
@Jc3s5h: You are certainly right about the states here, but I believe you are wrong about Puerto Rico (which is not a state). I believe someone born in Puerto Rico retains Puerto Rican citizenship even if they leave (while also being a citizen of any U.S. state they may reside in), and in particular this is relevant to the Spanish recognition of Puerto Rican citizenship. Do you have any citation to the contrary? - Jmabel (talk) 15:16, 30 August 2018 (UTC)
Jmabel, I do not have sources concerning Puerto Rican citizenship. But of course the burden of producing reliable sources falls upon those who want to state a claim in Wikidata; it is not the responsibility of other editors to disprove claims which lack such sources. Jc3s5h (talk) 15:30, 30 August 2018 (UTC) (in Spanish, cited in the English-language Wikipedia article) looks pretty solid to me. There is a specific certificate of Puerto Rican citizenship, and rights to favorable status for gaining Spanish citizenship based on place of birth being Puerto Rico (not needing to renounce prior citizenship) or any other Ibero-American territory. The portion there about "naturales de Puerto Rico estamos incluidos" suggests strongly (though it doesn't state outright) that this would be about place of birth rather than place of residence, so a different criterion than a U.S. citizen becoming a de jure citizen of whatever state they reside in.
I don't have time right now to check out every source of the en-wiki article; do you have specific claims there that you doubt? " a seriously flawed article, and should be ignored" is a pretty vague critique, and does not leave me a lot to respond to, short of researching or re-researching an entire article myself. - Jmabel (talk) 20:49, 30 August 2018 (UTC)
Unfortunately I don't speak Spanish. When I read some of the English sources from the article, the sources seemed biased toward a considerable degree of separation between Puerto Rico and the rest of the United States, and didn't seem to clearly support the claims that were just before the footnotes. After I wrote my comment above, I located the current directions (written in Spanish) from the Puerto Rican government to apply for a Puerto Rican certifiate of citizenship. I ran it through Google Translate, and the gist of it sees to be that a US citizen who was born in Puerto Rico, or a US citizen who was born elsewhere but has lived in Puerto Rico for at least one year immediately prior to applying, is eligible for the certificate. Jc3s5h (talk) 00:02, 1 September 2018 (UTC)
And apparently any benefits that Puerto Ricans may obtain in Spain are limited to "natural born" citizens of Puerto Rico, not just any US citizen who spends a year there. Ghouston (talk) 04:41, 1 September 2018 (UTC)
@Ghouston: That is correct. I hadn't realized that you could get the certificate just by being resident, so this may work out differently than I thought: if it provides no information beyond place of birth, it's less relevant. Not stated in the article I had cited, and I hadn't tried following up a ton of sources. Probably should be clarified in the en-wiki article. - Jmabel (talk) 16:14, 1 September 2018 (UTC)
@Jc3s5h: Thank you for now bringing up something substantive, rather than just a claim that the en-wiki article is crap. - Jmabel (talk) 16:14, 1 September 2018 (UTC)
There are also plenty of references in that Wikipedia article supporting what the article describe.C933103 (talk) 22:10, 30 August 2018 (UTC)

Please don't try to standardize for the sake of standardizing. As Puerto Rico doesn't really fit the traditional categorization of sovereign states and nationalities, it is unhelpful to make a blanket rule on whether Puerto Ricans should be recorded as having Puerto Rican citizenship or U.S. citizenship. If in conflict, consult biographical sources about the person. Deryck Chan (talk) 22:26, 30 August 2018 (UTC)

Deryck Chan, people raised in Puerto Rico are commonly referred as "Puerto Rican", regardless if they live there or not (i.e. Ricky Martin), and without providing a nationality in the case they reside in Puerto Rico. Esteban16 (talk) 00:00, 31 August 2018 (UTC)
That's also the case for a lot of places without official citizenships, like Scotland and Tasmania. Ghouston (talk) 04:32, 31 August 2018 (UTC)
Except it's more like sub-sovereign entities like Jersey and Hong Kong, where citizenship of these entities is legally recognised as a different kind of citizenship from citizens of the main country of that sovereign state that holds it. It's a different class of citizenship and it would be against the principle of original research if we refuse to recognise these citizenships categories. Deryck Chan (talk) 09:20, 31 August 2018 (UTC)
This has been discussed enough but no solution has been provided. According to the property, the statements could be a state, a nation or a fictional country. It seems Puerto Rico does not belong to any of them, and I think it should not be considered a nationality, as they are American citizens. If we are not agree about this, then consensus should be made. The property currently has issues in all those items, and it should be fixed ASAP. Esteban16 (talk) 22:33, 31 August 2018 (UTC)
From the article, it seems like one can at the same time be an PR citizen & US citizen & PR national & US national.C933103 (talk) 11:34, 1 September 2018 (UTC)
May I presume that "fictional county" above means to say "fictional country"? - Jmabel (talk) 16:18, 1 September 2018 (UTC)
Yes, Jmabel, it was a mistake. Esteban16 (talk) 16:26, 1 September 2018 (UTC)

Most of us agree with the fact "Puerto Rican" is not a nationality. Then, should I proceed with requesting the bot task to remove the property from the items? Or, should it belong to the property? Esteban16 (talk) 18:36, 2 September 2018 (UTC)

@Ghouston, C933103, Jmabel, Jc3s5h, Deryck Chan: Pinging users who participated in the thread. Esteban16 (talk) 01:06, 3 September 2018 (UTC)
I don't see why it being not a nationality is relevant. If there is a relevant legal status as citizenship of Puerto Rico, shouldn't it be included? We can have multiple citizenships given for a single person, and the constraints can be modified if they don't reflect the concept correctly. --Yair rand (talk) 04:15, 3 September 2018 (UTC)
I can't see how you come up with the conclusion that most of us agreed on that.... C933103 (talk) 06:06, 3 September 2018 (UTC)
I disagree with the summary that 'Most of us agree with the fact "Puerto Rican" is not a nationality.' In fact I am of the view that we should allow Puerto Rico to be listed as a country of citizenship, particularly for people who self-identify as such by reliable sources or by participation in international event as Puerto Rican (rather than USA, or Spanish in the past) representatives. This doesn't exclude them from also being listed as having USA (or Spain, or any other country that the person holds citizenship of) as country of citizenship. If we need to "fix" the constraint then I would broaden P:P27's value type constraint. Deryck Chan (talk) 08:58, 3 September 2018 (UTC)
off-topic a bit, how to treat sub-nation such as Puerto Rico seems a more off-shot discussion. For example, Bermuda had their own en:British passport (Bermuda), as well as the existence of en:British National (Overseas). So, given reliable source to verify, how to properly treat those "sub-class" of nationality of an empire or "federation of state"? British National (Overseas) were either de facto stateless (some ethnic non-Chinese minority thus they can exchange the status from BNO to BC) or ethnic Chinese. The Chinese, after 1997 had de jure automatically qualified (or became) as China PR citizen. But it seem wise to list them under value "British Hong Kong" as depreciated value. Also, for all other former colony, at that time there is no separation of nationality in the law of the empire, thus how to treat them, filling en:Gold Coast (British colony)? or British subject for those obtained British nationality before independent ? Matthew hk (talk) 06:42, 3 September 2018 (UTC)
@Deryck Chan, C933103: I thought most of us were agree because of the provided arguments. Anyways, I still consider the value should be fixed, regardless if it is accepted as a nationality or if it is changed/removed. Esteban16 (talk) 00:56, 9 September 2018 (UTC)
@Esteban16: Fixed it for you [2]. I've added value type constraint (Q21510865):class (P2308):dependent territory (Q161243), which captures Puerto Rico, British Hong Kong, and the other citizenships of sub-sovereign countries we've discussed here. Deryck Chan (talk) 09:07, 10 September 2018 (UTC)

Google Dataset Search

Google has introduced a beta Google Dataset Search available at which may be of interest here. It was reported in Nature yesterday. There is no Q item for it yet. Abductive (talk) 18:20, 6 September 2018 (UTC)

no label (Q56524784) Blue Rasberry (talk) 15:24, 10 September 2018 (UTC)

Dataset search tools

Hi all

Google have recently released Dataset Search, which looks really interesting. What other data search tools are available to find data that might be suitable for Wikidata?


John Cummings (talk) 17:53, 7 September 2018 (UTC)

Wikidata weekly summary #329

Excessively long names

When the Pawn... (Q2003805) has the alternate name When the Pawn Hits the Conflicts He Thinks Like a King What He Knows Throws the Blows When He Goes to the Fight and He'll Win the Whole Thing 'fore He Enters the Ring There's No Body to Batter When Your Mind Is Your Might so When You Go Solo, You Hold Your Own Hand and Remember That Depth Is the Greatest of Heights and If You Know Where You Stand, Then You Know Where to Land and If You Fall It Won't Matter, Cuz You'll Know That You're Right. How can I add this? —Justin (koavf)TCM 04:27, 8 September 2018 (UTC)

This is not possible due to length limit by the software. There is a phabricator ticket for it phab:T154660. --Pasleim (talk) 12:18, 8 September 2018 (UTC)
I@Koavf: You could cheat if you wish for any part of the string to show in the search results, and just complete the title in the next line. The artist might object, however. - Bossanoven (talk) 21:54, 10 September 2018 (UTC)

Recursive application of continent (P30)

I asked a user to remove "continent: Asia" from items about Indian villages on the basis that there would be no ambiguity about which continent they're on, since India is entirely in Asia and the villages would be noted, via located in the administrative territorial entity (P131) and country (P17), as being part of India and the continent could be inferred from there. Yet I have seen the recursive application of located in time zone (P421) to places at the village level for many countries (thanks @Nikosguard:!)—which I contend is rather inefficient—and am concerned that performing such a recursive application for other properties is just extra work whose benefit, when compared to the vast majority of other ways by which to improve items for geographical entities—such as adding some demographic info, adding specific coordinate locations, or adding authority control of any sort—is quite low.

The questions I wish to ask are this: Is it necessarily appropriate to recursively add continent (P30) to subnational entities of countries that are entirely in a single continent? Where should we draw the hard line between the explicitly enumerable and the implicitly inferrable? Mahir256 (talk) 21:58, 9 September 2018 (UTC)

† or to entities located in federal subjects of Russia/provinces of Turkey/etc., where the federal subjects/provinces are entirely in a single continent, or other situations like this at a lower administrative level in the appropriate countries

Three questions.
  • Question1. Is Wikidata following waterfall model? Is property of parent data not needed for child data. And same problem with the located in time zone (P421)timezone thing. For example, India follows single timezone. If it is in waterfall no need to add the timezone to all geographical entities in countries where single timezone is present.
  • Question2. thinking from a single Q item. Property of the geographical entity includes the continent, planet, solar system, galaxy, galaxy cluster, universe. So adding these properties is very much valid (We don't know what is outside universe. If something there then that must be included in Wikidata).
  • Question3. We dont have a list of confirmed Properties for a Q(At least I cannot find one. If there please pardon me and redirect me to that. I am new to the scenario).
So I am open for more instructions. Please help me. --Ranjithsiji (talk) 05:34, 10 September 2018 (UTC)
  • It's understandable to have an intuition that if an item X is a subclass of Y, then all of the claims on Y would also apply to X. However, this isn't officially the case: if a claim isn't on X, it doesn't exist. Tools won't return a claim from Y for a query about X. See also Wikidata:Project_chat/Archive/2018/04#Inherited_values. On the other hand, putting a continent on every location doesn't really seem necessary or desirable. Ghouston (talk) 07:01, 10 September 2018 (UTC)
If the inheritance not work then the data point is not existing in X. So If the continent is not on X then it is not on X. No inheritance. So can we add? Is the data incomplete without it? Basically it is a property of every geographical location in wikidata. So what is the decision? --Ranjithsiji (talk) 09:37, 10 September 2018 (UTC)
The existing consensus from previous discussions, as I understand it, is to limit it to countries or country subdivisions, except for Antarctica and places where it's not obvious and I think that makes sense. While we don't have any automatic inheritance of statements, we've always considered items to inherit information from other items, e.g. we generally only link to the most specific values for instance of (P31) and located in the administrative territorial entity (P131). Sometimes we intentionally duplicate things because it's very useful (e.g. country (P17)) but I don't see why we should do it for continents (or timezones). - Nikki (talk) 11:30, 10 September 2018 (UTC)

This isn't limited to the continent - for countries which have just one timezone, villages might have that statement as well even though it should be possible to be retrieved. Or the postal code, or even the country can also be retrieved from the items higher in the administrative subdivision tree. But there is no such automatic inference, which would require be a totally new software feature of "calculated statements". Especially as for countries like Turkey or Russia, the continent is not trivially inferred from the country. Ahoerstemeier (talk) 10:35, 10 September 2018 (UTC)

We need automatic inheritance for statements. Because that is the better way of keeping data more accurate and problem free. It is a high priority required feature. -- Ranjithsiji (talk) 14:39, 10 September 2018 (UTC)

If you're interested in this problem, you may want to look at Wikidata:Property_proposal/Generic#value is inherited from. - PKM (talk) 22:09, 10 September 2018 (UTC)

Ιnception of a municipality

Deryneia (Q777760) -> inception (P571) -> 1994

Hello. An item is about a human settlement. That settelement is also a municipality. By using P571 as a property, it means that the municipality was create at 1994 or the human settlement was created at 1994? The first is correct, the second not.

Of cource I have the option

< Deryneia (Q777760) View with Reasonator View with SQID > instance of (P31) View with SQID < municipality of Cyprus (Q16739079) View with Reasonator View with SQID >
inception (P571) View with SQID < 1994 >

but I just wanted to know about the first question. Xaris333 (talk) 08:27, 10 September 2018 (UTC)

Settlements and municipalities should be seperate items for this reason. Sjoerd de Bruin (talk) 09:53, 10 September 2018 (UTC)
From a data point of view, that's the way it should be, and that's the reason why geonames did duplicate them all - and thus via ceb-Wikipedia we already have lots of these in here. However, almost no Wikipedia ever does this duplication, so to be pedantic we'd even need a third item here which indicates that the linked Wikipedia articles cover both aspects. But even if we keep both aspects together in one item for simplicity, the inception can be clarified by using the qualifier applies to part (P518) -> municipality of Cyprus (Q16739079). Ahoerstemeier (talk) 10:26, 10 September 2018 (UTC)

It's a problem that we have only one article to most of Wikipedias... I think is better to decide what to do in Wikidata, and maybe wikipedias follow us later, if we decide to have two articles... Xaris333 (talk) 10:29, 10 September 2018 (UTC)

  • No, it's not a problem that Wikipedias are aimed at how humans think, rather than at some kind of pure epistemology. By the nature of Wikidata, we need to always be "splitters", distinguishing closely related concepts because they may have different attributes. But, in general, that is not how people want an encyclopedia to work. If you look up Seattle in an encyclopedia, you don't want separate articles to cover its pre-incorporation period, its brief period as an incorporated town, the brief period after that township was dissolved but before it incorporated as a city, and its period as an incorporated city. If you are managing it in a project like Wikidata, then these distinctions matter. - Jmabel (talk) 18:06, 10 September 2018 (UTC)

I mean it is a problem because one of the item will not have sitelink for a wikipedia, but the other will have. Morevore, it will be problem beacuse, for example, some wikipedia articles fetch data from wikidata. Is easiest to have all the information to a wikidata item than in two. I don't know if we can made template of wikipedia to take some informations for one templae and some from the other. Xaris333 (talk) 18:13, 10 September 2018 (UTC)

  • Ah. That makes more sense. Just so long as we don't think we should try to impose a "splitter" approach on the Wikipedias. - Jmabel (talk) 19:28, 10 September 2018 (UTC)

songs on albums - published in or part of?

It seems like published in is the wiser choice. - Bossanoven (talk) 21:46, 10 September 2018 (UTC)

@Bossanoven: We have tracklist (P658) and "on the tracklist of" as an alias for published in (P1433) (@Moebeus: regarding that alias); do either of those properties help? Mahir256 (talk) 21:57, 10 September 2018 (UTC)

Ideally we would have a property for specifically linking music tracks to tracklists, but in the absence of that P1433 is a lifesaver, and matches with Property:P577 publication date, helpful for when a track is present on several releases. "part of" is too broad in my opinon. Moebeus (talk) 22:04, 10 September 2018 (UTC)

Seattle Landmarks

How should we indicate buildings (and possibly districts) that have Seattle Landmark status? I see the one case

< Seattle Center Monorail (Q2618100) View with Reasonator View with SQID > appears in the heritage monument list (P2817) View with SQID < list of Seattle landmarks (Q1837959) View with Reasonator View with SQID >

but that seems wrong to me: we should be recording that it is present in an actual government list of landmarks, not a Wikipedia list article, especially because there is a possibility that the latter might eventually exist in multiple languages and be better maintained in one than in another. - Jmabel (talk) 19:31, 10 September 2018 (UTC)

heritage designation (P1435)? --Pasleim (talk) 20:15, 10 September 2018 (UTC)
@Pasleim: Clearly heritage designation (P1435), but what would the value be? Is it simply that we need to create a new item for the concept of a designated Seattle Landmark? - Jmabel (talk) 22:19, 10 September 2018 (UTC)
I've created Seattle landmark (Q56572098) --Pasleim (talk) 09:37, 11 September 2018 (UTC)

How to represent label changing over time?

Entities such as places or organisations (and sometimes even people) often change names over time. The time span during which a given name has been used could be represented by qualifiers attached to the label statement, but there is currently no way to attach qualifiers to the label statements. What is the alternative? I've not found any property such as "former name" to which time qualifiers could be attached. -- Bvatant (talk) 23:47, 10 September 2018 (UTC)

Game Publishers with No Information

For the past few months I have been adding instance of (P31) video game publisher (Q1137109) and industry (P452) video game industry (Q941594) to all companies/labels listed as a publisher for a video game in prominent game databases like MobyGames and Gamebase64. Yesterday I realized that this is probably not right to do for companies that aren't primarily focused on video games, like the retail companies Sears (Q6499202) (published some arcade games) and Zellers (Q845330) (published some bootleg console games). I am in the process of going through my history to remove these inappropriate claims, but now I don't know what to do when I can't find any information about the publisher for a game besides that it apparently published some or several video games. Often times the only evidence I find of its existence is in video game databases, which cite either images of old disks and cartridges games displaying an apparent publisher name, or a brief mention in the title screen of the game itself. Should I put/leave instance of (P31) video game publisher (Q1137109) and industry (P452) video game industry (Q941594) because they are related to the only verifiable information I have? Or should I leave those blank because I can't verify the company focused primarily on video games? And in that case, what should I put in the description of the item? Rampagingcarrot (talk) 19:16, 7 September 2018 (UTC)

@Rampagingcarrot: Well, industry (P452) would be correct in any case, wouldn't it? A company can be in more than one industry, and producing a video game makes the statement correct, I would think. (Not sure about this.) Regarding descriptions, I would recommend leaving them blank until an actual general source about the publisher is available. --Yair rand (talk) 18:37, 12 September 2018 (UTC)


We're currently not distinguishing between prayer (Q40953) as an act and as a work (e.g. as in Lord's Prayer (Q23393)). I believe we should have a prayer with

subject > subclass of (P279) View with SQID < work (Q386724) View with Reasonator View with SQID >

and start moving some statements over. Then we can start correctly modeling translated prayers, etc. I'm asking first since it's bound to affect a large number of items. What do you think? --Azertus (talk) 10:30, 11 September 2018 (UTC)

Why doesn't Wikidata allow two qualifiers for a statement?

Hi all

Can someone explain why Wikidata doesn't allow two qualifiers for a statement? Is this a restriction used in the model of how Wikidata should be structured, or is it a techncial limitation or something else?

As an example, Shrek (Q483815) has a list of voice actors, with the qualifers describing which character they voiced, however this isn't accurate, these actors only voiced the characters in English. It seems like the model doesn't allow you to record which character the voice actor played and the language version they played it in. Maybe this isn't the most important example I could have come up with but it feels like a clear example. Maybe I'm missing something?


--John Cummings (talk) 09:50, 12 September 2018 (UTC)

  • Yeah, you'd have to create a separate item for other language versions. --- Jura 09:58, 12 September 2018 (UTC)
  • Jura is correct - you can do this but only by creating a new statement to hang the new qualifiers off. The key thing here is that while a statement can have many different qualifiers, they all apply to the whole statement. To do otherwise we'd need to have qualifiers applying to other qualifiers, which doesn't work with the existing data model. Andrew Gray (talk) 11:54, 12 September 2018 (UTC)
Ah it seems link someone worked it out, you have two qualifiers for the the same statement, I assume that you could add a third if they were the same character in other languages (I think Antonio Banderas may have been Puss in Boots in the Spanish version also).
  • Voice actor
    • Mike Myers
      • character role: Shrek
      • language of work or name: English
John Cummings (talk) 18:37, 12 September 2018 (UTC)

Generating a list of references for a set of literature works

Hello, after reading Wikidata:How to use data on Wikimedia projects I'm still clueless about how to get a query trying to fetch all literature works about epistomology that Wikidata includes. The query itself surely should be improved, but my main helplessness pertains to how to make a wiki bullet list or even a table. I'm open with a solution that requires writing Lua modules, just show me the relevant documentation if what I'm aiming at is currently possible and documented. :) Cheers, Psychoslave (talk) 15:30, 12 September 2018 (UTC)

Wikidata:WikiProject Women/Wiki monitor/frwiki has an example of generated table using template:Wikidata list

Creating best practice example items for different subjects

Hi all

Nav Evans, the good people of Wikimedia Deutschland UX team and myself have been working on ways to try to make modelling items on Wikidata easier to understand, easier to learn and for items to be modelled more consistently.

As I'm sure many of you have experienced each of these issues causes quite a few problems, e.g trying to query a subject and the data is modelled in two/five/ten different ways....

Our first approach is to make it easier to find an example of well a modelled item for a subject e.g Douglas Adams as a well modelled item for an author. We're doing this through two properties, one to label the example item on the item for the topic, and one to label the example item as a good example for a topic e.g

  • Author has model item = Douglas Adams
  • Douglas Adams = model item for Author

We've started two property proposals to make this happen (not exactly inverse properties), I hope you will support them


--John Cummings (talk) 18:44, 12 September 2018 (UTC)

Splitting country of citizenship (P27)

I just realized that the English description for P27 is "country of citizenship: the object is a country that recognizes the subject as its citizen.". And then, in some other languages, the word "nationality" also appeared in description. Such description would cause numerous problems:

  • For instance when someone is a citizen of a non-widely recognized country, and then when that non-widely recognized country appear as a value for the property, someone who are against the existence of that country might dispute it and replace the value with countries that claims to control the area
  • Or if the subject is a citizen of a dependent territory or other area that are not a country in itself, the property cannot be used to reflect such situation
  • And see w:WP:Citizenship and nationality for one of the attempted explanation for difference between "nationality" and "citizenship"

As such, in order to properly describe those subjects, please separate the property for "nationality" and the property of "citizenship" into multiple properties. C933103 (talk) 09:49, 30 August 2018 (UTC)

  • "Nationality" seems suspiciously like a way of associating people with countries based on somebody's arbitrary judgement that they belong with that country. If it's just country of birth, we already have a separate property for it. If it's not country of birth, how is it defined exactly? Ghouston (talk) 10:55, 30 August 2018 (UTC)
  • I don't think the "country of birth" part in that proposal make a lot of sense since there are countries that do not determine nationality according to country of birth. There seems to be multiple definition for what the term "nationality" could mean. In some cases it is a legally defined term that comes with rights and duties, while in other cases it might also be the case, and in some cases it could even be arbitrarily decided by individuals. For instance, For British person, they are UK citizens but they are English/Scottish/Welsh/N.Irish/etc. nationals according to my understanding although they are not legally defined if I didn't miss anything. It is obviously helpful to include such information in wikidata but then "how" would be the problem. Another example is that, for a citizen in Hong Kong, as Hong Kong is not a country so it cannot be a value in P27, the P27 value for those subject can only be China/UK/India/nil/etc., with an additional property saying the subject is a "permanent resident" of Hong Kong which might not sufficiently explain the status of it being equivalent to citizenship of Hong Kong in most cases.C933103 (talk) 11:37, 30 August 2018 (UTC)
Ad 1 and 2: Since when are contradicting statements a problem on Wikidata? – If a fact is disputed and there are valid sources for either view, just enter the contradictory statements. I don't think we'll be able to resolve the type of cases you are mentioning by refining the definitions.
Ad 3, I don't see the problem with country of citizenship (P27); the labels and descriptions look ok. There are a few (pseudo-)nation states that confound the concepts of citizenship and nationality/ethnic group, and this is reflected in their language (thus the primary label of "country of citizenship" in French is "pays de nationalité").
There also is ethnic group (P172), which in my understanding largely covers the concept of "nationality" stripped of its meaning of "belonging to a country".
Cheers, Beat Estermann (talk) 13:20, 30 August 2018 (UTC) If anybody wonders what I mean by pseudo-nation state, here you go: Willensnation (Q694366)
  • How about when the value is clearly not a country like, let say, French Polynesian? The French translation have been adjusted according to what you said, however in English it would still say "Country of citizenship: French Polynesia" when French Polynesia isn't a country.
  • P172 is ethnicity. It is still a very political term and sometimes there are even special legal definition that restrict it in some rather strange ways. Like if someone from Taiwan try to claim his ethnicity is Taiwanese, then he could be considered as a radical separatist by China, and if the situation occurs within China then the person could face severe administrative punishment. C933103 (talk) 13:57, 30 August 2018 (UTC)
Ad 1: country of citizenship = France; nationality / ethnicity = Polynesian.
Ad 2: why do you think the term "nationality" is less prone to that type of problem?
--Beat Estermann (talk) 14:13, 30 August 2018 (UTC)
  • So, a French Polynesian if supposed to fill France in P27 as country of citizenship, then how about when someone immigrated to French Polynesia? I am pretty sure ethnicity won't change due to immigration, but what sort of property would describe the subject's status as part of French Polynesia? Like if the subject is an Indian immigrating to French Polynesia, his ethnicity will still be Indian regardless of his status, and wouldn't be an ethnically Polynesia according to my understanding.
  • Using Taiwan as example, in 1992 cross strait talk a consensus have been reached that would enable both sides to express themselves as being part of China and by extension to that the P27 of Taiwanese subject can be described as "Republic of China" in Chinese label to show its different from People's Republic of China. As long as individuals continue to claims the scope of "Republic of China" is not an entity separated from mainland China then it would be slightly more politically acceptable. (Although still not strictly acceptable)
    • Note:The situation described above was only to illustrate how ethnicity can also be politically sensitivity, the proposed nationality-citizen split will most probably not be used on subject of Taiwan.
      • Edit: Actually I should have used myself as an example. Given that myself is a
        1. People's of Republic China national, which mean People's Republic of China citizen
        2. "British National Overseas", which is NOT British Citizen and would not be treated as UK National by numerous authorities including EU
        3. Permanent Resident of Hong Kong, which also represent most of the citizen right in the territory
        4. Ethnically Han Chinese, belong to the sub-ethnic group of Canton
      • In the current wikidata structure, if I somehow become notable enough to get a wikidata entry about myself, that entry will only feature P27: PRCHina, Ethnicity: Han Chinese, Permanent Resident: Hong Kong, which is not very descriptive about the actual situation of such a subject like myself. People who are familiar with the situation of the territory can easily infer what these value mean however that would require such background knowledge to fully understand these wikidata values.
C933103 (talk) 22:09, 30 August 2018 (UTC)
The only thing you would seem to be lacking is a way to represent "British National Overseas". Maybe just use British National (Overseas) (Q5613380) in country of citizenship (P27)? No, I guess that wouldn't work because of the way the value of country of citizenship (P27) is formulated to be a country. Ghouston (talk) 00:04, 1 September 2018 (UTC)
@Ghouston: And how do you think the Taiwanese problems here? Really they are Republic of China nationalities, NOT Taiwan nationalities! -- 03:13, 14 September 2018 (UTC)
I guess having a separate item for the citizenship/nationality would help clarify what it means. Ghouston (talk) 04:18, 14 September 2018 (UTC)
See prior discussions Wikidata:Property proposal/Nationality, Wikidata:Project_chat/Archive/2017/08#How_to_look_up_persons_nationality.3F. "Nationality" is not a workable property. --Yair rand (talk) 19:56, 30 August 2018 (UTC)
What about only allow the use of such property when they are legally defined?C933103 (talk) 22:09, 30 August 2018 (UTC)
If the property is created, people will use it any way they like. It's quite possible that people not familiar with Wikidata will pick it instead of citizenship. Ghouston (talk) 04:28, 31 August 2018 (UTC)
Naming it something like legal nationality could help? C933103 (talk) 10:17, 31 August 2018 (UTC)

Pointing out that the "country of birth" is irrelevant to citizenship in many, many cases. E.g. millions of Palestinians have been born refugees in Jordan and Lebanon for decades but are stateless persons. They still have an ethnic group to which they belong (the Palestinian people) but no citizenship rights and certainly none in their countries of birth. —Justin (koavf)TCM 00:08, 1 September 2018 (UTC)

I've seen somewhere recently the suggestion of using citizenships instead of countries in P27. So instead of P27 = United States of America, you'd have P27 = United States citizen. That would allow creating items for arbitrary citizenships and nationalities that aren't states, e.g., citizen of Puerto Rico or British National (Overseas) (Q5613380). That would be better than having separate properties for "country of citizenship" and "nationality". It would also allow describing citizenships better in their own right. E.g., Australian citizenship has an inception date of 26 January 1949. However, this still leaves the problem that people were commonly described as "Australian", even in the 19th century, and many are marked up that way in Wikidata. Ghouston (talk) 00:38, 1 September 2018 (UTC)

Additionally, American citizens are also citizens of states (for those who live in states--i.e. not D. C., Guam, CNMI, Puerto Rico, or USVI). —Justin (koavf)TCM 01:10, 1 September 2018 (UTC)
It seems like a good idea. C933103 (talk) 04:34, 1 September 2018 (UTC)

I have always used both country of citizenship (P27) and ethnic group (P172) for people. That serves for all purposes. Let's take as examples revolutionaries for national independence:

< Rigas Feraios (Q319684) View with Reasonator View with SQID > country of citizenship (P27) View with SQID < Ottoman Empire (Q12560) View with Reasonator View with SQID >


< Rigas Feraios (Q319684) View with Reasonator View with SQID > ethnic group (P172) View with SQID < Greeks (Q539051) View with Reasonator View with SQID >

(he died before the creation of a nation-state called Greece). In many cases the most important property would be ethnic group (P172) (when someone looks for "German philosophers" usually looks for a list containing Immanuel Kant (Q9312) rather than ommiting most of those that would be included in such a list. -Geraki (talk) 18:28, 4 September 2018 (UTC)

We already have country of citizenship (P27) (legal) vs ethnic group (P172) (ethno-cultural). Creating an extra "nationality" property will make it even more confusing, as "nationality" (as the English language uses it) is a combination of "citizenship" and "ethnicity" with a little bit of self-identification. Deryck Chan (talk) 09:05, 10 September 2018 (UTC)

I'm not sure which which properties I can add to Fundação Alexandre de Gusmão (Q56604803)

Hello! I'm creating Fundação Alexandre de Gusmão (Q56604803)) and I'm not sure which instance of (P31) I could add to this item. To explain, this is a foundation under Brazil's Ministry of External Relations (Brazil) (Q1193807). I can't think of similar institutions in Brazil to copy and paste.

Also, there are two institutions "below" (I'm not sure the correct term for it), one is a library, and the other is a research institute. I'm going to create both eventually, but I'm wondering what property I could use for them.

Thanks for all the help! Tetizeraz (talk) 13:32, 13 September 2018 (UTC)

  • It depends what it does. I added some statements. It seems to act as a publisher, but "publishing house" seems suboptimal. You could add library and research institute. --- Jura 14:41, 13 September 2018 (UTC)
Hi Jura1 (talkcontribslogs), thanks for the help! Tetizeraz (talk) 00:05, 14 September 2018 (UTC)
For the "institutions below" you could maybe use empresa matriz (P749) <=> subsidiária(s) (P355)? The English label for "empresa matriz" is "parent organization" with various aliases that allow for more than just conventional "empresas". Moebeus (talk) 00:30, 14 September 2018 (UTC)


Hi, I was wondering about this for a long time and I didn't find the answer, so hopefully someone will help me. I'm adding Slovak descriptions, and sometimes I use Descriptioner tool to add repetitive descriptions and I'm using SPARQL Query. I need to know 2 things: is there a possibility to specify, that I want only one item to one property? For example when I want to add description Belgian tennis player, it will mostly return Belgian players, when I will specify country of citizenship: Belgium. But there will be also e. g. French-Belgian players, which have Belgium in this property, but they also have another country there. So if I can specify just e.g. Belgium. Second thing: Is there a possibility to return only items with some specified description? For example I will choose description Belgian tennis player and I will add Slovak description, so I don't need to worry about inaccurate descriptions, because English ones are nearly always accurate and correct. Thanks for the answers. --Luky001 (talk) 21:23, 13 September 2018 (UTC)

Maybe you could avoid the issue and just add the translation of "tennis player" as the description. It would only be a problem if there are multiple tennis players with the same name. Ghouston (talk) 00:59, 14 September 2018 (UTC)
It's standard practice here to include the country. - Nikki (talk) 08:41, 14 September 2018 (UTC)
If year of birth and perhaps place of birth are known, they may work better as disambiguators than citizenship(s). Ghouston (talk) 01:15, 14 September 2018 (UTC)
You can use ?item schema:description "Belgian tennis player"@en . to select items with the English description "Belgian tennis player". If you want to include multiple descriptions (e.g. when there's more than one way to write something or if you want to use multiple languages), you can do something like values ?desc { "Belgian tennis player"@en "belgian tennis player"@en "belgische Tennisspielerin"@de } ?item schema:description ?desc . (which would find items with any of those three descriptions) - Nikki (talk) 08:41, 14 September 2018 (UTC)
Thank so much Nikki, pity I didn't ask sooner. It solves (at least I think) all problems, which I experienced during adding descriptions for many items at once. --Luky001 (talk) 13:30, 14 September 2018 (UTC) as reference for human

Seemingly is in a lot of cases reference for the statement 'human'. But mentions only "gender", but does not mention, whether a living being is human or an animal.

Though - normally - the statement 'human' is correct, is not the correct reference for the assertion at issue.

Of course, normally we can deduce from other information, that a certain item is (1.) a living being and (2.) human. But such deduction is something different than a reference. -

However GND ID (property:P227) consist of the information 'Typ[:] Person (piz)' - e.g. -: German 'Person' = English 'person' ≈ English '(a certain) human (i.e.: an individual)' ≈ German '(ein bestimmter) Menschen (ein Indiviuum)'.

So, if references for the statement 'human' are need, could be substitute by GND ID.

--Villa loga-WB (talk) 09:48, 14 September 2018 (UTC)

  • Maybe the bot operator can explain how it's inferred. Please don't replace valid statements with some resource you consider "more valid". --- Jura 14:08, 14 September 2018 (UTC)

Autmatically Adding Images (P18) to Resources

Hi @all, I'm currently using DBpedia and Wikidata as query endpoints and one thing that would be very helpful for me is having a describing image for each resource in Wikidata. When going through data obtained from DBpedia i noticed, that many concepts in dbpedia have a "dbo:thumbnail" property, that is not reflected by a P:18 (Image) property on wikidat.

For example, The dbpedia entity: has a dbo:thumbnail and a owl:sameAs "".

The entity Q782543 has no image.

My approach now would be to automatically go through the DBpedia concepts I use, and for each concept with a dbo:thumbnail check the corresponding Wikidata item (with owl:sameAs) and set the P31 value to the dbo:thumbnail value, if the Wikidata item has no image so far. I think, from a technical standpoint I am able to do this, but my question is: Is this a valid approach? Or am I missing something that could have negative consequences? --Mackeprm (talk) 11:57, 14 September 2018 (UTC)

  • With "go through" you mean you'd screen the image, decide if it's relevant and, if yes, add a statement to Wikidata? There is a tool "wdfist" for just that.
    We don't want random images from, e.g. to be dumped into Wikidata.
    Consider that a P18-image on an item about a person is expected to depicted that person, not one of its notable works or grave. --- Jura 14:06, 14 September 2018 (UTC)

Thanks, that answers my question! And thank you for the reference to wdfist! --Mackeprm (talk)

Tombstone as information

Since we have to have a reference for religion, I added headstone (Q203443) as an instance of information since a Star of David or a Latin cross will differ between Judaism and Christianity. Yet when I use tombstone as a reference for religion here, I still get an error flag: Max S. Freudenberg I (Q20726682). Can someone see how to fix it, or come up with a more elegant way to be able to use a tombstone as a reference without getting an error flag? --RAN (talk) 14:38, 14 September 2018 (UTC)

Someone asked on IRC the same about interviews yesterday, we have type of reference (P3865) for this. Sjoerd de Bruin (talk) 15:01, 14 September 2018 (UTC)
Thanks a lot, Sjoerddebruin, I've also asked only some hours before in #How to model a museum label as a reference? above. :-)
A following question: It would probably now be good to indicate the location and the date of survey/consultation of the headstone (Q203443) or museum label (Q6940727).
  1. My use of location (P276) has been moved to a qualifier by DeltaBot.
  2. retrieved (P813) does not seem to be made to indicate a date of a physical survey. Should we broaden the scope? Or use point in time (P585)?
Thanks in advance, --Marsupium (talk) 16:10, 14 September 2018 (UTC)
Maybe Help:Sources#Headstones_at_Commons or Q706363#P569 can help. --- Jura 16:39, 14 September 2018 (UTC)

proto-language (Q206577) reconstructed/unattested?

There's currently this item, which is for a common ancestral language. Some of the labels say that it's reconstructed, but at the same time Proto-Norse (Q1671294) is a proto-language but is not entirely reconstructed; there are inscriptions in Proto-Norse. w:Proto-language also says usually unattested, but not always. So what is the case here? Is the label wrong and proto-languages can have written attestations, or is the claim that Proto-Norse is a proto-language wrong? In other words, is proto-language (Q206577) the same as "unattested language" or "reconstructed language" (neither of which have items) and thus redundant to them, or do these items need creating? Rua (talk) 14:54, 14 September 2018 (UTC)

Invitation to Wikidata QuickSheets online training

Hello Wikidata project chat. Over the past half year I have been working on a tool that semi-automates the improvement and citation of Wikidata items. You can read more on this Rapid Grant proposal. I wanted to let you know I will be holding an online training for this tool. This training will take place Tuesday September 18th, from 4:00-5:30PM, Eastern Standard Time. More info on the Dashboard page.--Theredproject (talk) 01:46, 15 September 2018 (UTC)

This section was archived on a request by: Matěj Suchánek (talk) 09:34, 20 September 2018 (UTC)

Wikidata property for enwiki

Does Wikidata have a property number or similar for Wikipedia pages, in the same way that Encyclopædia Britannica Online ID (P1417) gives me the relevant Encyclopædia Britannica page or subreddit (P3984) shows me the subreddit? Wikidata's user interface presents Wikipedia pages separately at the bottom, but I'm trying to get the title of a Wikipedia page in a SPARQL query where a property id is required. This query throws up nothing, and it seems odd that Wikipedia is the only major encyclopedia not represented. I'm aware of Sitelinks and Special:GoToLinkedPage/enwiki/Q… but I don't see how to use either in this context. Thanks, Certes (talk) 16:13, 16 September 2018 (UTC)

@Certes: Here's a simple query that returns the names of 300 random enwiki articles:
SELECT DISTINCT ?item ?page_titleEN WHERE {
  ?article schema:about ?item ; schema:isPartOf <> ;  schema:name ?page_titleEN .
Try it!
Hope that helps. Mahir256 (talk) 16:26, 16 September 2018 (UTC)
Thank you! That's exactly what I was looking for. Certes (talk) 16:38, 16 September 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:34, 20 September 2018 (UTC)

What am I doing wrong?

Hey there,

on Pentax X-5 (Q7165092), I added Commons:Category:Taken with Pentax X-5 as category for pictures taken with camera (P2033) – worked like a charm. So I wanted to do the same with Pentax XG-1 (Q18387564) and Commons:Category:Taken with Pentax XG-1, but I only get "no match was found". I thought it might have to do with Commons:Taken with Pentax XG-1 being marked as a hidden category, so I turned it into a normal category for testing purposes, but Wikidata still does not recognize the category's existence. Am I doing something wrong? Cheers, --El Grafo (talk) 09:08, 18 September 2018 (UTC)

Same problem for Commons:Category:Taken with Pentax X70 and Pentax X70 (Q56651080) … --El Grafo (talk) 09:18, 18 September 2018 (UTC)
✓ solved Nevermind, I just didn't notice that on that first edit I actually entered an item on Wikidata for the category on Commons rather than just the name of the Category as it is done with Commons category (P373). Problem solved never actually existed. --El Grafo (talk) 09:39, 18 September 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:34, 20 September 2018 (UTC)

CollabPad (but it is a special page)

In case you're wondering why your watchlist and other special cases have such a funny header: phab:T204873. Multichill (talk) 19:55, 19 September 2018 (UTC)

This section was archived on a request by: Matěj Suchánek (talk) 09:34, 20 September 2018 (UTC)

wiki markup to display links with labels of items

Hello, The wiki markup [[Q2]] creates a link to Q2. But is there a markup to create a link with the label of the item ? For instance [[Q2||label]] for Earth. Thanks,  – The preceding unsigned comment was added by Hopala! (talk • contribs).

There is the Q template which displays the label in the user's preferred language, e.g. {{Q|2}} produces Earth (Q2) - Nikki (talk) 09:25, 20 September 2018 (UTC)
Thank you very much ! - Hopala! (talk) 09:32, 20 September 2018 (UTC)
This section was archived on a request by: Bencemac (talk) 12:33, 20 September 2018 (UTC)

Q741775 - "Enemy of the people" and "class enemy"

After a discussion concerning the term enemy of the people (as used by Donald Trump) I have seen that the Wikidata item Q741775 confounds two different concepts: the enemy of the people (stemming from Roman hostis publicus, used time and again, most notably in Nazi terminology as de:Volksfeind and in the Soviet Union mainly in the 1930s, Russian: ru:Враг народа. There is also the more specific term class enemy (German: de:Klassenfeind, Russian ru:Классовый враг which redirects to ru:Классовая борьба = class conflict/de:Klassenkampf) which was more widely used in the 1920s after the Revolution. There are some articles that deal with the class enemy such as da:Klassefjenden, hr:Klasni neprijatelj, nl:Klassenvijand, no:Klassefienden, and some that deal with the enemy of the people such als en:Enemy of the people, de:Volksfeind, ca:Enemic del poble, pl:Wróg ludu etc. I would like to separate the two (which means that the English language Wikipedia would have a redirect for en:Class enemy only). Any thoughts on this? --Andropov (talk) 08:47, 9 September 2018 (UTC)

  • As a native English speaker, I would definitely agree that these are two distinct concepts, and should be separated, even if they have sometimes been used in a similar manner as political epithets. - Jmabel (talk) 04:17, 10 September 2018 (UTC)
@Jmabel: Thank you for your input! Could you, speaking of muddled concepts, tell me what the difference between en:enemy of the people and en:public enemy is? Kind regards, --Andropov (talk) 10:14, 15 September 2018 (UTC)
In the U.S., at least, "public enemy" has a connotation of actual criminality. The FBI used to put out a list of the "Ten Most Wanted" criminals, and the person at the head of the list was referred to "Public Enemy Number One." "Enemy of the people" is less specific; the phrase echoes the Ibsen play of that name, where it is used more or less ironically, in that the person who is accused of being an "enemy of the people" is the doctor who correctly diagnoses that the waters of the town's famous baths are contaminated, and his publicizing this hurts the local economy. That irony may or may not be there in any particular usage. - Jmabel (talk) 16:25, 15 September 2018 (UTC)

How to propose renaming a property?

Is there a way to propose renaming a property, or propose replacing one with another? If so, how? Rua (talk) 17:53, 15 September 2018 (UTC)

If the name of the property is incorrect, you can just change the label. If the you want to change the purpose of the label, start with a discussion on the talk page of the property. It might result in a PFD procedure and new property request. Mbch331 (talk) 18:25, 15 September 2018 (UTC)

Q25056785 and Q56537023 refer to the same real life object

  • Q25056785 ("F.A. Brockhaus AG" – without blank between "F." and "A.") consists only of three statements and two interwikis links. This item belongs to w:en:F.A. Brockhaus AG. There is said:
"Friedrich Arnold Brockhaus founded his publishing house in 1805 in Amsterdam, [...]."


  • Q56537023 ["F. A. Brockhaus (Verlag; ab 1817: Leipzig)" with blank between "F." and "A."] consists of few more information and belongs to GND 65191-6. There is said:
"1805 in Amsterdam unter Rohloff & Compagnie gegründet; 1811 Umsiedlung nach Altenburg in Thüringen; 1817 nach Leipzig".

Therefore I guess: This two item should be fused. - However, Q464633 (Bibliographisches Institut & F. A. Brockhaus) belongs to a different real object, because it came in existence not before 1984 - due to an fusion of two distinct companies. The latter item belongs to w:de:Bibliographisches Institut & F. A. Brockhaus. There we can read:

"Als Beginn des Verlages F. A. Brockhaus wird die Gründung des Verlages 'Rohloff & Co.' am 15. Oktober 1805 in Amsterdam durch den Kaufmann Friedrich Arnold Brockhaus angesehen, der zuvor mit einem Handel für Wollstoffe gescheitert war. Da Brockhaus kein Mitglied der örtlichen Buchhändlergilde war, erfolgte die Gründung über einen Strohmann, den Buchdrucker J. G. Rohloff."

--Villa loga-WB (talk) 18:08, 15 September 2018 (UTC)

P values to use for gene structure

Gene structure eukaryote 2 annotated.svg

Hello all. I'd like to try to encode the hierarchical organisation of gene components as summarised in the diagram (right). What are the most logical P properties to use? Is there something more specific than "is part of" / "is instance of", e.g. for the relationship between promotor and regulatory sequence? Evolution and evolvability (talk) 02:53, 15 September 2018 (UTC)

@Evolution and evolvability: The relationship between a regulatory sequence and a promoter in it can be described using part of (P361) and has part (P527). But I think there is currently no way to link between a gene (which has genomic start (P644) and genomic end (P645)) and its regulatory sequences. In addition, genomic start (P644) and genomic end (P645) of genes are currently not only including open reading frames, but also including untranslated regions. So we may have to create some new properties. --Okkn (talk) 06:08, 16 September 2018 (UTC)


Can anyone clarify what the difference between cirque (Q184368) and cirque (Q388227) is supposed to be? One is supposedly a subclass of the other, but they have the same English label, they're both described as valleys caused by glaciers, they have the same Commons category and essentially the same GeoNames code (one is singular the other is plural), so I have no idea which one to use. - Nikki (talk) 09:54, 15 September 2018 (UTC)

Sorry, I do not remember. Chaoborus (talk) 12:14, 15 September 2018 (UTC)
Both have dewiki sitelinks, and the dewiki sitelink for cirque (Q388227) seems to describe it as a general geological shape, possibly but not necessarily caused by glaciers, with cirque (Q184368) being a special form of it (i. e. the reverse of the subclass of (P279) situation on Wikidata). --Galaktos (talk) 23:14, 15 September 2018 (UTC)

Gee fizz

Can someone protect Rick and Morty (Q15659308). It's currently my entire watchlist. GMGtalk 00:20, 16 September 2018 (UTC)

@GreenMeansGo: it has been protected, but requests that need sysop actions are placed on the administrators' noticeboard. Esteban16 (talk) 02:35, 16 September 2018 (UTC)

Make Wikidata P and Q templates universal in Wikimedia projects

I am writing to propose making P and Q templates in Wikimedia projects outside Wikidata, in every language, to redirect to Wikidata. Any comment here in support or criticism of this idea will help future planning.

Is anyone aware of this being discussed before? Can anyone say where they saw this, even if it was another language or project?

Consider {{P}} and {{Q}}. In Wikidata these templates are fundamental to quick discussion of Wikidata items and properties because they convert Wikidata identifiers to human readable text for conversations. Outside of Wikidata in other Wikimedia projects there are often P and Q templates which link to Wikidata. For example, see Commons:Template:P and Commons:Template:Q. Having these templates in place means that irrespective of language or project, editors can use P and Q templates to make connections to Wikidata.

Check out Template:Property (Q19694638) and Template:Wikidata entity link (Q17280715), which are Wikidata's own records of which Wikimedia projects have equivalent templates by any name. Each of these are currently in about 40 Wikimedia projects. In about 20 cases the name is P or Q. In other cases perhaps there is another name, but P and Q are redirects. For some of the cases P and Q go to non-Wikidata uses. I think there are about 800 Wikimedia projects total. If it would be useful to make P and Q templates universal then we should make some plan to set that up or reserve it.

In English Wikipedia en:Template:P currently goes to en:Template:Smiley to make an emoji. In German it goes to en:Template:P for "priority" or the number 1. In French fr:template:P goes to the portal system. It causes big social problems to change templates which get established in Wikimedia community projects. Wherever possible we should propagate templates out now so that in the future we do not have to negotiate for this in every language and every project. Currently I am talking about this on English Wikipedia at en:Template talk:P where there is resistance to the idea.

Sorry - I have a major deficiency in this, in that I do not know the best way to migrate templates across languages and projects. If anyone can point to documentation on how to do this or outline the workflow of how this should go down then speak up. Also, if anyone has an objection to making P and Q templates universal then please say why. Blue Rasberry (talk) 14:04, 5 September 2018 (UTC)

The simplest solution is to do this as parser functions. I already wrote those parser functions when I was at WMDE, but it was turned down as it would interfere with community work on templates. In my opinion parser functions are the right way to do this. Jeblad (talk) 18:42, 5 September 2018 (UTC)
@Jeblad: Thanks, I would not have known to talk about that. What do you mean you wrote those parser functions? You mean for migrating arbitrary templates, or for this case specifically? Blue Rasberry (talk) 01:29, 6 September 2018 (UTC)
@Bluerasberry: As I recall they were parser functions for label and description, but in effect nearly the same. Jeblad (talk) 02:25, 6 September 2018 (UTC)
Most projects have little or no need for such a template, and occupying two of the single-letter template names is really a bad idea, probably disrupting dozens of local workflows and making everyone there memorize new names. I would advise against trying to make these universal. --Yair rand (talk) 00:20, 6 September 2018 (UTC)
@Yair rand: How would you feel about doing this for languages which do not have a P or Q and instead use non-Latin characters? What about for languages which have these letters, but where P and Q templates do not exist? Blue Rasberry (talk) 01:29, 6 September 2018 (UTC)

This is not something that Wikidata can decide - it is exclusively a matter for the community on each individual project. I guess that it is unlikely to be a significant issue to create them on wikis that do not currently have templates by this name (unless it conflicts with any local policies, e.g. I have a very hazy recollection that one of the Indic language wikis doesn't (or didn't) like templates other than in that language's script). On wikis where templates by these names do already exist you need to discuss any proposals to change with the community on that project. Thryduulf (talk) 16:30, 6 September 2018 (UTC)

Any project that want to use Wikidata, or really that enable the client part of the Wikibase extension, will at some point need tools to point to items, properties, and a number of other entities. Now those projects has to implement those tools themselves. In fact, when any project must implement a module to make a template work it is a clear indication that some basic tool are missing. To not implement a tool just because it could be implemented as a Lua module is not very wise, and it blocks smaller projects from progressing.
There are several other modules that are reimplemented again and again, and those should also be replaced by generalized tools and libraries. Now many modules are copied between projects, and more often than not they should not be used at all because they fail even quite basic quality measures (not to say tests, they have none).
Please do request the devs to implement basic tools for accessing entities, it is the best way to help smaller communities to access and use Wikidata. Jeblad (talk) 18:17, 6 September 2018 (UTC)
Whether the tools are or are not needed or wanted is irrelevant to this discussion. This is about only the name for templates that refer to Wikidata properties and items (i.e. the equivalent of {{P}} and {{Q}}). This is something that can only be decided by the individual wikis. It is not possible for a discussion here to determine this for any project other than Wikidata. Thryduulf (talk) 20:50, 6 September 2018 (UTC)
Quite frankly, this does not make sense. Jeblad (talk) 11:52, 10 September 2018 (UTC)
@Jeblad: What doesn't make sense to you? Why do you think it doesn't make sense? Thryduulf (talk) 01:30, 17 September 2018 (UTC)


Is there a property to use to mark a battle in a war, and in relation to fronts and other battles? The part of (P361) piles everything together, and does not structure the information very well. Or can conflict (P607) be used for battles in war? –Susanna Ånäs (Susannaanas) (talk) 11:22, 16 September 2018 (UTC)

What about significant event (P793) per significant or notable events associated with the subject. Pmt (talk) 11:51, 16 September 2018 (UTC)
Thanks, I think it's much better. I realise that this may be the smallest of problems, it seems that most everything about wars is hard to model to the level of Wikipedia infoboxes: Who were the fighting parties, who lead them, how many casualties and prisoners were there on each side in each battle etc. – Susanna Ånäs (Susannaanas) (talk) 11:59, 16 September 2018 (UTC)

Geographical coordinates from image metadata

If we add geographical coordinates to a listing based on the metadata of an image pertaining to that listing, would we add a source for this statement, and, if so, what kind? I originally asked in Help talk: Sources. ARR8 (talk) 14:39, 16 September 2018 (UTC)

  • Not an answer to your question, but from my experience at Commons, a lot of images have inaccurate geocoord metadata in their EXIFs. Before trusting that, always look at a map and see if it is at least highly plausible. This should never be done by a bot and simply trusted. - Jmabel (talk) 17:17, 16 September 2018 (UTC)
Thanks for the reminder. In this specific instance, I have done so. The specific use case here is landmark listings on Wikivoyage which are missing coordinates but have pictures in Commons, so this will be a manual and infrequent process. ARR8 (talk) 02:29, 17 September 2018 (UTC)

deactiveing article preview

How can I hide article preview pop-up permanently?Yamaha5 (talk) 02:44, 17 September 2018 (UTC)

@Yamaha5: disable the “Preview” gadget? or do you still want to have the gadget, but without the preview opening by default? --Lucas Werkmeister (talk) 10:43, 17 September 2018 (UTC)
@Lucas Werkmeister: This preview artcile window should be optional or should be disabled by gadget or prefrencess.
This Window
Yamaha5 (talk) 13:38, 17 September 2018 (UTC)
As Lucas wrote, you need to disable the "Preview" gadget. You can do this at Special:Preferences#mw-prefsection-gadgets --Pasleim (talk) 14:41, 17 September 2018 (UTC)

Wikidata weekly summary #330

TV anime described as film

Why are there many TV anime entry on wikidata being mistakenly described as other things like film or TV drama or whatever other things? C933103 (talk) 13:31, 13 September 2018 (UTC)

Hi C933103. In general, this is because Wikipedia articles cover all media of a given franchise together, and this was used to seed Wikidata initially. For example, Black Lagoon covers 5 media (manga, both anime seasons, a light novel and an OAV) or City Hunter 12 (!) media.
I tend to think the correct way to do that is to have the "main item" about the media franchise, and separate item for each different medium. Some examples: Dragon Ball (Q2020) Naruto (Q642), Ghost in the Shell (Q92580), Cardcaptor Sakura (Q49182)
That’s a lot of work though :)
Jean-Fred (talk) 14:48, 13 September 2018 (UTC)
That would be hard to do interwiki link if wikidata entries are broken down this small...C933103 (talk) 08:03, 14 September 2018 (UTC)
Why so? Most of the interwiki links would be on the franchise item. If Wikipedia articles happen to exist dedicated to a movie or an OAV, then they can be linked to the Wikidata item. This is exactly what happens for Ghost in the Shell (Q92580):
That way, we have a 'proper' data model and reasonable interlinking, no? Jean-Fred (talk) 10:10, 14 September 2018 (UTC)
It would be good to fix these at some point. I tend to avoid them. In general, statements about any of the elements end up combined on some item. Maybe the situation has gotten better recently. --- Jura 10:20, 14 September 2018 (UTC)
If Wikipedia language version A create only an article for a novel version of a series, then Wikipedia language version B create only an article for comic version of a series, and then Wikipedia language version C create both articles then user would not be able to jump between language version A and language version C article on the same series directly. Unlike the current situation that all of them would be linked to same wikidata entry for the series and then only link to individual works when such wikipedia article exists.C933103 (talk) 03:06, 15 September 2018 (UTC)
@C933103: If there are article groups which have this problem, then for each item new placeholder pages should be created on each wiki, added as sitelinks to the items, and then turned into redirects. The alternative is that the concepts are conflated into one item, which is convenient from the WP perspective but incorrect from the WD perspective. Jc86035 (talk) 14:21, 15 September 2018 (UTC)
@C933103: That’s a valid concern, but I think in practice the pattern of linking together the main Franchise item will probably work in many cases, in particular the most complex ones. No Wikipedia links to Naruto (Q26971382) or Naruto (Q26971391), because it is likely that no Wikipedia will ever want to create a separate article for manga and anime.
Where I agree with you and am a bit worried about are cases where « what is the "main" medium » is debatable. For example, the manga Miss Hokusai (Q29940472) vs. its movie adaptation Sarusuberi (Q11580762). French Wikipedia is the only one having an article for both ; and for some language versions is is unclear which one they 'mainly' or 'primarily' talk about. A third item about a franchise would feel like overkill to me.
Jean-Fred (talk) 10:16, 18 September 2018 (UTC)

wikibase and authentication methods

Hello, I'm new with wikidata and I don't know if it's the right place to ask about technical question about wikibase. I wonder if it's possible to have several authentication methods (ex: LDAP, oAuth, SAML, OpenID ...) when installing and configuring a standalone wikibase instance?  – The preceding unsigned comment was added by 2a01:cb1d:89d9:d000:886:6d0f:7907:f52c (talk • contribs).

You probably want to talk to the developers - Wikibase Installation has some info, also the project page outside of wikimedia. I can't find anything directly about authentication, but I expect it's something that could be configured as you like. ArthurPSmith (talk) 18:27, 18 September 2018 (UTC)

Distinction of title of authority (Q7810129) and position (Q4164871)

Despite those items are said to be distinct, I don't see when to use which. There are also some Item that claim to be Instance of both (list). Can someone help me out? -- Dr.üsenfieber (talk) 13:07, 18 September 2018 (UTC)

@Dr.üsenfieber: title of authority (Q7810129) refers to how a person would be addressed - for example "Professor X", "Dr. Y", "President Z" etc.. position (Q4164871) is the label for the job they are in, whether or not they would be addressed that way, and it is usually much more specific, for example "President of Singapore", etc. ArthurPSmith (talk) 18:35, 18 September 2018 (UTC)

How to enter this property?

SO I'm working with herbaria (because of the newly approved Index Herbariorum code (P5858)) and at Oakes Ames Orchid Herbarium (Q55828891) I am trying to enter the collection or exhibition size (P1436).

Entering a number? It's a no go. The property pretends the "units" subproperty/qualifier is optional, but clearly it is not. In fact, it won't save on older entries that have no unit if you attempt editing the property (e.g. Maksim Bahdanovič Literary Museum (Q12337378)). Alright, so I try to enter biological specimen (Q2075980) (I mean, what else am I supposed to use?!)... but it still won't save.

What am I doing wrong? Can I adjust the property so this nonsense will stop without manually adding exceptions for every collection in the world? Circeus (talk) 13:02, 18 September 2018 (UTC)

@Circeus: This should work with or without a unit, see the examples on collection or exhibition size (P1436). What exactly is the error message you are seeing? "won't save" suggests maybe there's something else wrong - if it's a constraint violation it will still let you save but then show a warning flag. ArthurPSmith (talk) 18:30, 18 September 2018 (UTC)
GAH. Turns out it was far more stupid than I thought: I pasted the number as found on IH with the space and all, and what goes on is that numbers (unlike dates) don't have any wiggle room whatsoever when you enter them. Circeus (talk) 19:40, 18 September 2018 (UTC)

What do you think is the main benefit of Mix'n'match?

Hi, I'm preparing for a presentation on Mix'n'match at Wikiconference North America. I have a little experience with Mix'n'match, but I know that I don't know very much compared to what everyone else knows. If you are going to Wikiconference North America and would like to collaborate on the presentation, please e-mail me or leave a talk page message. If you're not going to the conference but have knowledge about it, you can tell me here and I'll incorporate it into my presentation.

On Mix'n'match, I've imported a database there and I've done some matching. As I see it, right now links to external identifiers like those found in library databases are nice to have, but I don't think other sites use them (I'm aware that SNAC pulls from Wikidata, but I don't think it picks up non-authority file links). I think of it like investing in Wikidata without knowing what the payoff will be. Are there sites that currently use links from Wikidata like those matched through Mix'n'match?

I know Magnus created Mix'n'match as a way to create redlink lists--is the Creation candidates part of the site currently used for making new Wikidata items? I haven't heard of many people using it, but what do you think? From Wikidata, the Listeriabot can make redlink lists, like those for the Women in Red project--are there other projects using Wikidata to make redlink lists?

Are there other things about Mix'n'match that you think other Wikipedians and library people should know?

Thanks, Rachel Helps (BYU) (talk) 16:00, 18 September 2018 (UTC)

  • Wikidata:BEIC used mix'n'match to identify missing authors and then import/create items for them. --Nemo 17:28, 18 September 2018 (UTC)
  • Oh, thanks, I'll definitely add that to my presentation. Rachel Helps (BYU) (talk) 18:21, 18 September 2018 (UTC)
The biggest benefit is to find external id's for people, but more recently I have been using it for paintings. It's great to match up Qids for datasets that connect to some external database already (like any of the catalogs in there already, you can easily upload your own dataset, but sometimes you just want the Qids). Jane023 (talk) 18:48, 18 September 2018 (UTC)
Hmm, when would you just want the Qids? Rachel Helps (BYU) (talk) 19:45, 18 September 2018 (UTC)
I always explain it like this to Wikimedians: In Wikipedia you get the basic information about a topic and links to further literature. In Wikidata it's the same thing. You can basic data about a topic and then humans and machines get links to other databases/catalogs/etc to find out more. One of the things Wikidata is good at is being this identifier hub that brings all these different identifiers together. When you have one identifier you can easily find all the others be it as a human in the website or as a machine via the API. --Lydia Pintscher (WMDE) (talk) 18:52, 18 September 2018 (UTC)
okay, so mix'n'match links are useful because the page they link to could provide more information about a subject, which can then be used to improve Wikidata/Wikipedia? Rachel Helps (BYU) (talk) 19:13, 18 September 2018 (UTC)
Yes that is one really important aspect. The other part is what Jane was referring to I believe: When you are building a third party app or are matching catalogs then Wikidata can be very helpful by being the glue between many catalogs. If you have a list of IDs from the German National Library and want say the equivalent IDs in the French National Library. The German National Library might not have the link but you can go to Wikidata to find it because they are all collected in the same item here. (Made up example! They probably link to each other but there are enough combinations where this connection is useful but doesn't exist.) --Lydia Pintscher (WMDE) (talk) 19:21, 18 September 2018 (UTC)
Oh, I can see how that would be useful. Do you know of third party apps or websites that are currently using Wikidata for this kind of thing? Rachel Helps (BYU) (talk) 19:45, 18 September 2018 (UTC)
  • More specifically about "Mix'n'Match": it's mainly a tool to avoid creating duplicates. Let's say you have some identifier about 100 martians that are notable and should be added to Wikidata, you can create the items directly as you know nobody else created them yet. However, if you want add some identifier about 100 British PMs, you should avoid creating new items as they mostly likely already have items. The identifier should be added to the existing item. Many datasets are somewhere between these two.
    There are few other tools to achieve the same. --- Jura 19:24, 18 September 2018 (UTC)
oh, definitely! It's such a basic convenience of Mix'n'match that I didn't even think of it. Rachel Helps (BYU) (talk) 19:45, 18 September 2018 (UTC)