Open main menu

Wikidata:Project chat/Archive/2015/12

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


Mix'n'match daily report

For those of you interested in Mix'n'match and external catalog reconciliation, I now produce a daily report of things that might require manual attention. Some of that is Mix'n'match specific, but others may be of general interest (e.g. multiple items using the save external ID). --Magnus Manske (talk) 13:50, 25 November 2015 (UTC)

@Magnus Manske: Very useful. Thank you! Would it be possible to generate URLs for the identfiers in the "unrecognised external ID" part, to make them link to the database ? Jheald (talk) 14:57, 25 November 2015 (UTC)
Could disambiguations be included as well? By the way, I don't get any results after clicking on "Disambiguation links", it just says it's loading and should take ~30 sec. It used to work, maybe some change disabled it. Matěj Suchánek (talk) 15:40, 25 November 2015 (UTC)
You make my day :) GerardM (talk) 10:42, 1 December 2015 (UTC)


What is the best way to link to documents in Wikisource? I can add the full link to a document, but wouldn't it be easier to link to a category of subjects, like we do for images? There are a few portals for authors, but not for subjects of source material. Am I doing it properly at Q6273866?  – The preceding unsigned comment was added by Richard Arthur Norton (1958- ) (talk • contribs) at 19:42, 30 November 2015 (UTC).

@Richard Arthur Norton (1958- ): No. The correct approach is to make a new item for the article, then use main subject (P921) to link to the person the biography is of.
There should not be a direct sitelink between a person and a biographical article because (i) an article is not a person; (ii) the same person may have several biographies; but there can only be one sitelink. Jheald (talk) 19:53, 30 November 2015 (UTC)
see eg Tylor, Edward Burnett (Q16081140) / Edward Burnett Tylor (Q141037) for example, using the property described by source (P1343) to link back to the article. Jheald (talk) 20:06, 30 November 2015 (UTC)
Why can't we have "Subject:Foo Smith" at Wikisource to link all the material about that person? They have "Author:Foo Smith" at Wikisource, if the person wrote material, and there you can list material they are the subject of. We have "Category:Foo Smith" at Wikimedia Commons even we just have two items, and sometime only one item in anticipation of more. Or should I just take an image of the original page in the book and add it to Wikimedia Commons to avoid problems with Wikisource? Do other people recognize this as a problem? --Richard Arthur Norton (1958- ) (talk) 20:02, 30 November 2015 (UTC)
The material can be linked to from the Wikidata item on a person by using the described by source (P1343) property, as described above.
It is also (additionally) appropriate to sitelink to the "Author:Foo Smith" page, for authors.
You'll have to take it up with Wikisource as to whether they have pages for subjects of articles too; but it's possible they may just be using Wikidata to index the main subject. There would also again be the problem that there can only be one sitelink, so one could not link both to "Author:Foo Smith" and to "Subject:Foo Smith"
But if Wikisource were to have "Subject:" pages, redirecting to "Author:" pages if the subject was an author, then there would be no objection in principle to a sitelink to such a "Subject:" page, since it would be in direct 1:1 relation with the Wikidata item. Jheald (talk) 20:16, 30 November 2015 (UTC)
This could also be achieved through the use of categories, such as "Category:Foo Smith", on Wikisource. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:38, 1 December 2015 (UTC)
It looks like "Category:Foo Smith" in Wikisource has different purpose depending on which project you look into. I have therefor this far avoided those pages. A category "About Foo Smith" maybe should not be linked with a category "Works by Foo Smith". Some projects may even combine those two subjects in the same category. -- Innocent bystander (talk) 15:29, 1 December 2015 (UTC)

Showcase items

Hey everyone,

Harmonia Amanda and I are thinking about adjustments to the showcase item process. Currently, making a item a showcase item takes a very long time due to a unclear process.

The new process is described here, I already added some notes to the talk page about things I want some opinions about. Please leave your comments there, so we can highlight our best content better. :)


Sjoerd de Bruin (talk) 12:34, 1 December 2015 (UTC)

Template extraction tool ?

Are there any standard tools to eg take a list of pages (eg from PagePile), scan those pages for a particular template, and give some or all of its arguments ?

I feel sure that somebody must have done this, but I couldn't see anything on the "External tools" page. (Or did I miss something?).

I suppose it's easy enough to write a bespoke scraper; but I couldn't help thinking that someone must have written a standard tool. Jheald (talk) 14:19, 1 December 2015 (UTC)

There is
It should be possible to combine it with Wikidata:Database reports/templates and items with 0 claims or similar. --- Jura 14:25, 1 December 2015 (UTC)
Also Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:35, 1 December 2015 (UTC)
For my own purpose I have written a js-tool, linked on my userpage, that uses the SPARQL-Endpoint to produce a list of Wp-articles, where I read some parameters out of the infobox, correct them if needed and then push them into wd. I use that tool at the moment to read P2044 (elevation above sea level) of some lakes. I´m afraid this tool is only usable for those who know how to write SPARQL and javascript, because it is an unfinished tool and maybe it will never be finished. (The reason is, that people write into infoboxes lot of things and it is difficult/not possible to automate the reading. So copy that js-file into your own userspace and improve the tool the way you need (I´m using this tool in preview-modus, therefore I don´t know if it works in userpage/common.js --Molarus 17:21, 1 December 2015 (UTC)

Community Wishlist Survey

MediaWiki message delivery (talk) 14:38, 1 December 2015 (UTC)

Expansion depth limit exceeded

When using this version of {{Constraint:Units}} where list contains Q-templates ({{Q}}), you get the message Expansion depth limit exceeded. If you roll back Ivan's last edit on that template (which resulted in above version of the template), you can use Q-templates again. I asked Ivan what was the cause and he doesn't know. What he figured out is the following:

  • {{#invoke:Constraints|getCaption|units|123456}} has expansion depth 2.
  • {{Q|123456}} has expansion depth 6.

Combining these two: {{#invoke:Constraints|getCaption|units|{{Q|123456}}}} (which is what happens in the background in the above mentioned version of the template together with a Q-template, you suddenly have a expansion depth of 41 (which is above maximum). Anyone an idea what's causing this? Used modules: Module:Constraint and Module:I18n/constraints. Mbch331 (talk) 21:24, 1 December 2015 (UTC)

Wikinews, Wikispecies and MediaWiki now have data access

Hey folks :)

Wikinews, Wikispecies and MediaWiki now have data access. Meta will follow on 15th as we had to delay it to not interfere with the most important fundraising days.

Cheers --Lydia Pintscher (WMDE) (talk) 14:21, 2 December 2015 (UTC)

Conflicting descriptions for language nds

I am trying to merge Q19616710 into Q7316913 but get the titular error - how can this be resolved? Nikkimaria (talk) 17:08, 2 December 2015 (UTC)

Tried and done. --ValterVB (talk) 17:22, 2 December 2015 (UTC)
Thanks. Nikkimaria (talk) 19:26, 2 December 2015 (UTC)

A tricky one

The September 11 attacks (Q10806) are currently described as located in the administrative territorial entity of New York. Obviously, that is not strictly true (e.g. the Pentagon is not in New York). I'm not sure how this is best addressed, but I imagine someone with more Wikidata experience knows how a case like this should be handled. - Jmabel (talk) 17:15, 2 December 2015 (UTC)

Maybe multiple located in the administrative territorial entity (P131) with applies to part (P518)? Matěj Suchánek (talk) 17:34, 2 December 2015 (UTC)
I removed the located in the administrative territorial entity (P131) claim and add location (P276)=New York City (Q60) and location (P276)=Arlington County (Q107126). located in the administrative territorial entity (P131) should not be used because September 11 attacks (Q10806) is not a geographical location. --Pasleim (talk) 19:03, 2 December 2015 (UTC)
Thanks! - Jmabel (talk) 00:20, 3 December 2015 (UTC)

Is there a property for something like performed at or spoken at?

Hi all

I'm looking for a property that would allow you to show that a person had performed or spoken at a particular event or place e.g a person had spoken at a conference e.g TED, or a performer had the Sydney Opera House. Does this exist?


--John Cummings (talk) 10:30, 3 December 2015 (UTC)

participant (P710) to link to people participated in a given event (e.g. ⁠) and participant of (P1344) to link to events people have participated in (e.g. ⁠) which might be what you are looking for, the events having the location. Thryduulf (talk: local | en.wp | en.wikt) 12:05, 3 December 2015 (UTC)
Thanks, Participant seems a little bit vague, it could refer to someone who simply attended a conference, I guess I'm looking for something that specifies they were active e.g a speech or performance but if its widely used then this will be fine. John Cummings (talk) 12:17, 3 December 2015 (UTC)
You could use the P794 (P794) qualifier I suppose to do that, but see also the section immediately below as an answer to that question might help with this also. Thryduulf (talk: local | en.wp | en.wikt) 12:45, 3 December 2015 (UTC)
There's also speaker (P823), connecting an event to a person, which again could be used with a qualifier. Jheald (talk) 13:15, 3 December 2015 (UTC)
Thanks very much, so to use an example e.g the TED conference in 2014, you would use speaker (P823) on the conference item to link the speakers with the qualifier 2014, but how would you link to TED on the speakers' item? Would you use the same property or use another? John Cummings (talk) 13:26, 3 December 2015 (UTC)
Not necessary to have inverse property: this can be deduced by logical query. Better avoid to create the link person->event because for people like politicians or actors this will create hundreds of new statement in their items and will lead to heavy sets of data to load when opening the item. Inverse properties are a nightmare to keep up-to-date and they just create big and redundant data sets.
If you want to find all events where a person spoke just use a query to extract the events which have property speaker (P823) with the value corresponding to the item of the person. Snipre (talk) 15:56, 3 December 2015 (UTC)

How to note the principal participant?

How do you note which of a list of participants is the principal one, e.g. star guest, guest of honour, headline act, principal speaker, etc? Opening of the Liverpool and Manchester Railway (Q7096523) has a large list of participants, but I can't find a way to note that Arthur Wellesley, 1st Duke of Wellington (Q131691) (prime minister at the time) was the guest of honour. It would be possible to mark that statement as preferred rank, but that feels a bit blunt and doesn't give an indication of why. Thryduulf (talk: local | en.wp | en.wikt) 12:43, 3 December 2015 (UTC)

I think you could make a item for "guest of honour" and use it with P794 (P794) as qualifier. Sjoerd de Bruin (talk) 12:47, 3 December 2015 (UTC)

Make WikiProjects more prominent

I think we have hidden the WikiProjects below to many layers of weird links. Can a few people comment on a suggested change of the community portal that would make the WikiProjects more visible. I hope that more people will sign up for the projects when they are more visible: Wikidata_talk:Community_portal#Wikiproject_Links --Tobias1984 (talk) 13:38, 3 December 2015 (UTC)

+1. We need to work together inside WikiProjects. Currently we find solutions on data modeling in discussion (how can I import that data, which property can I use for that, which are the differences between this property and this one,... ) and all that knowledge disappears in archives. Snipre (talk) 15:49, 3 December 2015 (UTC)

Tool for Disambiguation

Just to increase the visibility: request for a new tool --ValterVB (talk) 18:38, 3 December 2015 (UTC)

Commons category format

Any idea why Commons categories don't link correctly? The formatter URL looks correct to me. - Jmabel (talk) 04:48, 3 December 2015 (UTC)

Convenience link: Commons category (P373). - Jmabel (talk) 04:49, 3 December 2015 (UTC)
What do you mean with "don't link correctly? I checked a few items and couldn't find a problem. If it doesn't work on some id's and does work on others, please provide examples. Mbch331 (talk) 06:29, 3 December 2015 (UTC)
This is probably just a usual thing, that sometimes external links get shown by the UI, and sometimes they don't. (Nobody seems to quite know why). Sometimes refreshing the page can then make the link show, but sometimes it doesn't. Sometimes it can take anything up to 5 refreshes to make the link show; sometimes it doesn't show even then. I don't know what causes the problem, but it seems it can happen for any property with an external link, eg it very often happens for identifier properties as well (though Commons category and Commons gallery may be hit more often than others -- perhaps because they may be further down the page (?)). If all else fails, there's always Reasonator. If you haven't activated it already, I strongly recommend the gadget that installs Reasonator as an extra link at the bottom of the sidebar. Reasonator appears to have no problem with external links. Jheald (talk) 09:21, 3 December 2015 (UTC)
That sounds to me like a cache issue, have you tried ?action=purge ? Popcorndude (talk) 13:06, 3 December 2015 (UTC)
It sounds to me like the same problem I reported in phab:T115794. Ever since August, I've been having issues with gadgets and common.js stuff which touch the statement section of the page not working consistently. - Nikki (talk) 20:03, 3 December 2015 (UTC)
I have also seen this problem. Reguyla (talk) 21:02, 3 December 2015 (UTC)

Interlanguage links won't show up, but it says it's already attached?

I've been going around the unconnected pages in Simple English, connecting the ones that should be connected so the interlanguage links will show up. Today I stumbled across Punjabi language. It doesn't appear to be connected, as there are no interlanguage links in the left hand column. However, when I try to connect it I get the error message, "The page you wanted to link with is already attached to an item on the central data repository which links to Punjabi language on this site. Items can only have one page per site attached. Please choose a different page to link with." Going to Q58635 shows a link to simple from that item, and it goes to the page I'm trying to link, so why don't the links show up on simple? ONUnicorn (talk) 21:31, 3 December 2015 (UTC)

Sometime can happen, if you make a null edit the interlink will appear. --ValterVB (talk) 21:43, 3 December 2015 (UTC)
It's on
For a full purge, use followed by the title. --- Jura 21:52, 3 December 2015 (UTC)
Thank you! ONUnicorn (talk) 22:05, 3 December 2015 (UTC)

Property for 'copyright status'?

The 'copyright status of a work' is something that Wikimedians REALLY care about, and it is a crucial piece of metadata that we work with every day, but there is no way to model that in Wikidata yet (as far as I'm aware). I also assume that this kind of model would be a pre-requisite for things like structured-metadata on Commons. Would it be sensible to create a new property specifically to map the 'copyright status' of any item that is classified as an 'instance of' anything that is a 'work' (painting, book, etc.). For starters, I assume it would be relatively simple for a bot to run through all the existing items that have a 'creator' with a 'death date' of 100 years ago or more - and tag them as 'copyright status -> public domain'. Equally, (with some important exceptions), to check for artists who are not dead or artworks that are created in the last 50 years, and tag them as 'in copyright'.

Relevant qualifiers could include an 'end date' (perhaps several end-dates for different countries!); a 'proprietor (Q16869121)' (in cases where the named author is different from the copyright holder (e.g. a commissioned work). The Wikidata model allows for conflicting claims to be mapped within the same statement - this gives us the ability to acknowledge that different people often have different interpretations of the same law. Note: this does not account for claims of 'fair use' or 'copyright claim on the artwork digitisation' because these are about representations of the work, not the work itself. Given how much Wikimedians discuss about copyright I would be surprised if this hasn't come up before in some way or another so I didn't want to create a property proposal directly. Wittylama (talk) 13:40, 1 December 2015 (UTC)

In terms of how things are done on Commons, something like "PD-art-70"" would need to be split. The copyrights status is "PD", and that is qualified by the reason "art-70" (or some such). It can also be qualified by "applies in"="United States", and perhaps by "does not apply in"="France". We also need to distinguish copyright status ("Andy Mabbett is the copyright holder of this image") from licence ("Andy Mabbett has licensed this image as CC by-sa 4.0"). I'm not sure that it's necessary hough to add a copyright status property to an artwork here we have the author and their death date - in such cases, the copyright status can be computed. I suggest starting with properties for 'copyright status' and 'licence', and then see what qualifiers we need, as they are rolled out. Let me know if you need help drafting proposals. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:34, 1 December 2015 (UTC)
To be able to "compute" the 'copyright status', you also have to use such things as "published in". -- Innocent bystander (talk) 15:32, 1 December 2015 (UTC)
And also things like "copyright assigned to", e.g. Peter Pan (Q107190) "copyright assigned to" Great Ormond Street Hospital (Q1349705). Thryduulf (talk: local | en.wp | en.wikt) 17:54, 1 December 2015 (UTC)
Surely that assigns ownership of the copyright, so is already covered? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:11, 1 December 2015 (UTC)
Well, "assigned to" and "ownership" is probably not the same thing. I cannot sell, give away or remove my copyright as long as I am under the current jurisdiction. -- Innocent bystander (talk) 08:01, 2 December 2015 (UTC)
Yes, but that's doable. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:11, 1 December 2015 (UTC)

Apparently we already have all usable licenses? See Creative Commons Attribution-NonCommercial (Q6936496) for example. Jane023 (talk) 14:04, 4 December 2015 (UTC)

@Jane023: We can here have items about any work, also non-free. -- Innocent bystander (talk) 17:35, 4 December 2015 (UTC)
That is not the question, however. The question is whether we have items for all licenses currently in use on the projects. --Jane023 (talk) 18:00, 4 December 2015 (UTC)

I'm not sure where this discussion has concluded.... Is there a consensus to at least start a formal property proposal for "copyright status"? Wittylama (talk) 13:30, 7 December 2015 (UTC) [ping user:Jane023, User:Pigsonthewing]

Protect Q3620669

Is it possible to protect Q3620669? People try to enter [1] over and over again. But tt0078788 is not the id of the redux version. --Jobu0101 (talk) 22:47, 2 December 2015 (UTC)

Instead of protecting the item, there are three easier ways to solve the issue:
  1. Convince yourself that it is really the wrong Id. Looking at it seems to me that IMDb is handling both the 1979 and 2001 version under the same Id.
  2. Remove the wrong Id from the place where it gets imported from, i.e. it:Apocalypse_Now_Redux#Collegamenti_esterni
  3. add the right value to Apocalypse Now Redux (Q3620669) or if no IMDb identifier exists for that movie, add IMDb ID (P345)=novalue --Pasleim (talk) 23:42, 2 December 2015 (UTC)
@Pasleim: Thank you for your steps. What to you think about Q3346699? We created a IMDb ID (P345)=novalue claim because the IMDb ids belong to Q21468403 and Q21468405. But the import bot didn't care. And we shouldn't remove them from enwiki because there are no separate articles for the two parts. --Jobu0101 (talk) 18:08, 4 December 2015 (UTC)
I contacted the IP. Hope it will read my message. I don't know what else we can to to prevent such edits. --Pasleim (talk) 18:55, 4 December 2015 (UTC)

the mess with P357 (P357)


P357 (P357) Title-string has been replaced by title (P1476) a long year ago, but at this moment, 73437 items have a link to P357 (P357) :(

to make it worse, roughly half of those are used as sourcing info, to indicate the title of articles online, for ex, which can be deduced from the number of items with P357 (P357) property, which are currently 35844 ([357]&pagepile=&statementlist=&run=Run&mode_manual=or&mode_cat=or&mode_wdq=and&mode_find=or&chunk_size=10000).

Well, the biggest problem is, it is impossible to change the property from P357 to P1416 automatically, since a language has to be indicated, and, some items may have several source info with P357.

The second big problem is to be able to find the items where the property is used in Source, and find it in the item... I tried on Ireland (Q27), and I just couldn't... the item doesn't even load completely in my browser... :(

So, what do we do, now, to clean up this mess, and try and remove this OBSOLETE property ?

Is there any way to ease this ? or shall it be done by human-hand-eye processing only… and last for as long as the removal of old P107 property ?

Thanks for any tool suggested to try and ease this process. --Hsarrazin (talk) 21:13, 2 December 2015 (UTC)

I've started to write a tool to ease to process of moving such statements. Should be finished in the next few days. --Pasleim (talk) 21:32, 2 December 2015 (UTC)
I have seen a large use of this property together with "imported from: English Wikipedia". The title-property is then used to indicate the name of the used article. In those cases, the language should be English. -- Innocent bystander (talk) 05:19, 3 December 2015 (UTC)
BTW: Do we need a title for sources like that? --Kolja21 (talk) 06:06, 3 December 2015 (UTC)
"Imported from: English Wikipedia" gives almost no information at all, so why not. -- Innocent bystander (talk) 13:23, 3 December 2015 (UTC)
@Innocent bystander, Kolja21, Pasleim, Hsarrazin: We can't do an automatic conversion because we don't have the language value as statement. We could define the language from the language of the work in some case. But we are waiting all the extension of the language list with the addition of an undermined language to fully convert all uses of P357 (P357). This is a phabricator bug (T78006) Snipre (talk) 16:13, 3 December 2015 (UTC)
@Snipre: I offered you a small set of claims that can be changed, but I do not say that other claims are easy to change. -- Innocent bystander (talk) 16:23, 3 December 2015 (UTC)
@Innocent bystander:I prefer to work once on that problem but if you want to do some corrections already now, don't wait on me. Snipre (talk) 19:16, 3 December 2015 (UTC)
@Snipre: - I know that it can't be automated.
However, a semi-automated tool that could find those hidden uses of the P357 (P357), ask the contributor what language it is, then create a new title (P1476), and erase the deprecated property would be very useful, either on a one-shot use (on item found), or through a semi-automatic interface ;) --Hsarrazin (talk) 17:31, 3 December 2015 (UTC)
@Snipre: I already replace this property when I come across it, but I do not search for it on a daily basis. Replace it with a "unknown language" when we in many cases can identify the language looks to me like the wrong way to go. -- Innocent bystander (talk) 19:23, 3 December 2015 (UTC)

Tens of thousands can be changed automatically. These are only in references:

PREFIX pr: <>
PREFIX prov: <>select ?prop ?title (count(?ref) as ?count) where {
  ?item ?prop ?statement . 
  ?statement prov:wasDerivedFrom ?ref . 
  ?ref pr:P357 ?title
group by ?prop ?title having (count(?ref) > 100)
order by desc(?count)

Try it!

--JulesWinnfield-hu (talk) 10:05, 6 December 2015 (UTC)

Property for "This painting was depicted on currency note <country x> <year y>"

Does this exist? If yes, where? If no, how on earth do I propose this one?! --Jane023 (talk) 09:37, 5 December 2015 (UTC)

"present in work"? --- Jura 09:42, 5 December 2015 (UTC)
@Jane023: Do the opposite; on the Wikidata item about the note, use depicts (P180). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:43, 5 December 2015 (UTC)
Thx for the replies! You both appear to share the opnion that we need utems for all denominations of all currencies. I didn't get that far in my thinking yet. Is asethere some datbase we can read in with such things? Perhaps all we want are the more 'artistic' denominations bought and sold on e.g. e-Bay. --Jane023 (talk) 14:21, 5 December 2015 (UTC)
Commons? There is, e.g., c:Category:Euro banknotes with categories for specific denominations and series. You could create corresponding items and link these. --- Jura 14:28, 5 December 2015 (UTC)
OMG sounds like work, but I think you are right. That is probably a good place to start. I will see what I can do with that, thanks! --Jane023 (talk) 15:00, 5 December 2015 (UTC)
@Jane023: To make so many similar-ish items all at once, it may be worth planning it out off-line, then creating a file to drop on QuickStatements. Jheald (talk) 17:11, 5 December 2015 (UTC)
Yes at first I was thinking of that, but unfortunately it's not so easy. I am currently just trying to track down all the articles which already have an item and adding english labels. --Jane023 (talk) 17:18, 5 December 2015 (UTC)
We need two sets of items, one for generic note values like "Bank of England five pound note", etc, and one for "Bank of England Series D five pound note", et al. See en:Bank of England note issues for details. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:07, 5 December 2015 (UTC)
Can you provide us with a reference for this need? --- Jura 22:03, 5 December 2015 (UTC)
It's our choice, no need for a reference, a consistent argumentation should be enough. But I agree, there is classes of coins for example, and eaxh type of coin describes something different and deserves it's own item. author  TomT0m / talk page 15:02, 6 December 2015 (UTC)
If it's our choice, you should be able to provide us with a series of items illustrating it and a link to the discussion where we made this choice. --- Jura 15:25, 6 December 2015 (UTC)
I mean, it's our choice to be made. I guess if we had a ready made decision we would not need this talk. My preference is to crate the item for a lot of reasons it seemed I already wrote a tons of times. So I guess if you do not agree you'll have to do the work of arguing against this :p author  TomT0m / talk page 16:43, 6 December 2015 (UTC)
I think there are various options and they are worth evaluating. I don't think we are in a position to say "we need" yet. --- Jura 17:10, 6 December 2015 (UTC)

List of missing data?

At the moment I´m working on a Lua module for showing WD-data of a cycling competition as a table in Wikipedia. I´m using the Lua Error Handling function pcall (protected call) to encapsulate errors. "Errors" means: properties or values are missing. If I turn things around and don´t show what is right only what is wrong, I would get a list of all "errors", that means a list of all missing data. For generalizing this to all sport competitions and maybe for everything, a way to formulate how the items and properties should be connected (w:Ontology language) and a software to translate this into lua-modules is needed. Such modules could be inserted into the discussion pages to show users and bots which data is missing in an item or in a group of items. "Go bot, find me the latest unemployment figures at www" --Molarus 04:03, 6 December 2015 (UTC)

Catalogue number for NMHH film rating

Hi, I would like to add the catalogue number of NMHH film rating (P2363) as qualifier to claims:

This catalogue number is featured on movie posters and home media covers along with the rating itself, and also would help finding the film's record in the .xlsx given as source on the property page. My issue is that catalog code (P528) is not designed for use as a qualifier, and it has a constraint requiring catalog (P972) as its own qualifier. What would you suggest? Is there a property I could use, or should I propose a new one? – Máté (talk) 08:56, 6 December 2015 (UTC)

How about The Man from U.N.C.L.E. (Q14488725) catalog code (P528) NFT/22948/2015 / catalog (P972) National Media and Infocommunications Authority (Q541497)? Mbch331 (talk) 09:37, 6 December 2015 (UTC)
Well, on the one hand, it seems viable, on the other hand, these numbers rather seem to correspond to decisions about the ratings than the films themselves (one movie may have multiple ratings due to multiple cuts or reconsideration with multiple catalogue number), and the two values wouldn't be directly connected. – Máté (talk) 11:15, 6 December 2015 (UTC)
(MPAA film rating (P1657) has a similar number called "certificate #" [for it's 49390], it is listed on and IMDb. The NMHH number is called "nyilvántartási szám" which roughly translates as registry number. I'm sure there is something similar for most other content rating systems and properties. – Máté (talk) 19:28, 6 December 2015 (UTC))

Help needed to improve anti-vandalism tools

Hello, You may know ORES, We use ORES to build anti-vandalism tools (Learn more). Based on automatic revert detection we were able to build an MVP and we have some high quality classifiers online you can use (WD:ORES). In order to improve the anti-vandalism classifier we need you to go through some edits and determine whether they are damaging to Wikidata and if they are ill-intended edits or they are just newbies/honest mistakes. This would help us distinguish between newbies and vandals and also improves our data to make precise and adequate vandalism detection classifier. Please go to Wikidata:Edit labels install the gadget and do a workset. Thanks Amir (talk) 02:51, 7 December 2015 (UTC)

Spam? P1343: Armenian Encyclopedia (Q2657718)

Sorry, but I do not find it useful to see on many items the claim described by source (P1343): Armenian Soviet Encyclopedia (Q2657718). What's the point? Can someone remove these claims in case there is not a specific id mentioned? Michiel1972 (talk) 20:48, 6 December 2015 (UTC)

As a wikisourcian, I find it very useful to list all the encyclopedias and dictionaries that have an article about "John Smith". Does it have to be done on Wikidata ? With described by source (P1343) ? I don't know. Pyb (talk) 20:58, 6 December 2015 (UTC)
The property is a bit of an oddity. Entries for Nordisk familjebok (Q678259) are similar. --- Jura 21:00, 6 December 2015 (UTC)
if each country or language has one relevant encylopedia, all the basic items can potentially already receive hundreds of such claims. In case there is a pagenumber mentioned or ID I think it can be useful, but just a general claim "this item is mentioned in encylopdia X" does not feel right and useful. Michiel1972 (talk) 22:17, 6 December 2015 (UTC)
Usually, it's just under the keyword <insert label of item in original language of encyclopedia>. --- Jura 22:21, 6 December 2015 (UTC)
Hi. Soviet Armenian Encyclopedia is the biggest Armenian language encyclopedia. It has only 36,767 articles. I generate the list from Armenian Wikipedia (see w:hy:category:նյութեր Հայկական սովետական հանրագիտարանից). I think just a general claim "this item is mentioned in encylopdia X" is much more better then nothing. BTW, is there a limit on the number of claims in described by source (P1343)?--ԱշոտՏՆՂ (talk) 09:25, 7 December 2015 (UTC)


Is there a tool to easily convert SPARQL queries to SQL queries and Vice versa? --XXN, 21:17, 6 December 2015 (UTC)

Not really -- they're running on two rather different databases, which each either (in the case of SPARQL) has its own properties defined, or (in the case of SQL) has its own particular table-structure defined.
I believe there are adaptors available, to make an SQL database look like a SPARQL database accessible with SPARQL-language queries; I am not sure whether or not the reverse is also true. However, as far as I know, nobody has set up such an adaptor for the MediaWiki SQL databases, and (again, so far as I know) there is no way that I know of to access the information in those tables from a SPARQL query.
It would be very nice if it were possible -- eg to return the Q-numbers for the Commons categories for every image, where we have an P18 image property for a person, but not a P373 commons category. But, so far as I know at least at the moment, that kind of investigation still needs a purpose-written bot or script on toolslab. Jheald (talk) 23:12, 6 December 2015 (UTC)
Thank you for your answer, Jheald. --XXN, 23:45, 6 December 2015 (UTC)
The data models do not match well. The sort of SQL one would write to query the data model used by Wikidata or for an RDF store are quite different from that you would use to query a relational database. SQL is for tabular data while Wikidata is more like a big graph of data. Look into learning SPARQL: it is quite powerful without the complexity of SQL (joins etc.). —Tom Morris (talk) 10:59, 7 December 2015 (UTC)

Commons gallery

Jheald has started to add Commons gallery (P935) in all the items which already have this same Commons gallery in "Other sites". My impression was that there is a consensus not to do this? - Brya (talk) 05:37, 2 December 2015 (UTC)

I am not aware of any policy or any consensus against using Commons gallery (P935).
If we look at the historical numbers, one can see that of the 8,469 uses of the property a couple of days ago, 7,676 were on items that also had sitelinks -- ie all except 973.
It is true that on the talk page, in August 2014 User:Ricordisamoa seems to have suggested changing the name to "additional Commons galleries". But that suggestion seems to have attracted no follow-up and was never implemented.
As far as I can see, the apparent reason that this property only had 8,500 uses was that (like so many other properties), nobody had got round to systematically populating the items it applied to.
In particular:
  • I can't see any benefit in having to look up whether an item has a Commons gallery in two different places -- both sitelinks and P935 -- rather than just having all the information available in a single place, on a single property.
  • In the case where an item has two Commons galleries, it seems to me far better to me to record both of them using P935, with appropriate qualifiers to explain the difference between them, rather than one with a sitelink and one with P935 (as User:Brya seems to contend) giving the appearance of a mismatch, a value of P935 that is simply out of date.
  • P935 is also more reliable. Over the last 18 months, Commons users have increasingly been adding WD article-item <-> Commons category sitelinks (see again historical tables linked above), sometimes removing an existing WD article-item <-> Commons gallery sitelink to make this possible. If the Commons gallery is only linked to WD by the sitelink, then we've lost that information. However, if P935 is populated, then it doesn't matter what happens to the sitelink, because the information connecting the item and the gallery is still present in the P935.
  • As a final small point, can I mention that the SPARQL service handles queries involving properties much better than queries involving sitelinks. It extracts property matches directly from a fast-indexed B+ tree. In contrast, sitelink checking involves complex (slow) string manipulation. Larger queries involving properties are therefore far more likely to be successful than similar queries involving sitelinks, without timing out.
For all these reasons, therefore, given that Commons gallery (P935) exists, I think it is entirely appropriate for it to be populated. Jheald (talk) 11:41, 2 December 2015 (UTC)
I support Jheald in this. In the confusing mess of Commons-Wikidata-Wikipedia-interlinking, this is one of the few things that shouldn't be controversial. It will preserve links from Commons-galleries to Wikidata-items/ Wikipedia-articles, that could be lost otherwise, because we have still open questions about this whole interlinking concept (and resulting sitelinks-changes). For reference, the confusion about commons-sitelinks is visualised here:
Sitelinks, Property:P935/Commons gallery, Property:P373/Commons category. (BTW, this is igoring the commons-files linked by wikidata-properties, .) --Atlasowa (talk) 14:38, 2 December 2015 (UTC)
There's one last block of 10,000 links to go.
With 80,000 values of Commons gallery (P935) now in place, even though this discussion is underway, there doesn't seem much point in having 80,000 done and 10,000 not done. So I hope nobody objects if I now drop the final 10,000 onto QuickStatements, so at least then there will be consistency, that we can then review. Jheald (talk) 15:36, 2 December 2015 (UTC)
Final 10,000 now underway... Jheald (talk) 17:11, 2 December 2015 (UTC)
If this is felt to be a good idea, can we then at least get rid of the redundant Commons in the "Other sites"? - Brya (talk) 17:52, 2 December 2015 (UTC)
@Brya: Trying to tell Commons what sitelinks it could no longer have to/from the rest of the world would probably go down extremely badly there -- even more so than the (now failed?) attempt to insist that Commons categories' sitelinks should have to go to/from wiki-categories rather than wiki-articles.
It's possible there might be a move towards more WD-article <-> Commons-category sitelinks, rather than WD-article <-> Commons-gallery sitelinks. But (IMO) any such move would need to come from Commons, rather than be pushed from here; and indeed I think might still be quite highly controversial for many people on Wikidata. Probably best to let sleeping dogs lie. Jheald (talk) 18:37, 2 December 2015 (UTC)

And just when you thought the Commons sitelink mess couldn't get any worse ...

It turns out that User:Tulsi Bhagat has created a vast number of new items for Commons galleries that had no sitelinks to an existing wiki -- without checking whether we already had an item for the same person/thing. It seems that almost the entire final block of 10,000 items that I have been adding P935s for (the ones with the highest Q-numbers), and possibly quite a lot of the earlier ones too, are these new orphan items.

What to do with these, especially if we already have an item, sitelinked to a Commons category ?

-- eg: item Gregory Peck (Q108366), site-linked to c:Category:Gregory Peck; with new item Q21167568 created, site-linked to c:Gregory Peck.

Whatever happens, we only ought to have one item for Gregory Peck (Q108366), and that should have both a Commons category (P373) and a Commons gallery (P935) set.

But the merge script (and presumably tools like Duplicity) won't let us just merge the two items Q108366 and Q21167568, if the two have two different sitelinks to Commons.

So what to do?

  • There's currently no item for "Category:Gregory Peck". It certainly was once part of the grand plan (of some people at least) to see WD category-items created systematically for all such Commons categories, and force Commons categories to sitelink to them rather than real articles. So call this option 1: create a new item "Category:Gregory Peck"; force the Commons cat to sitelink to that; and then merge the new Commons gallery item into the Gregory Peck article item.
  • But in more recent time the trend seems to have been away from that, and instead to let Commons categories sitelink to WD article-items. This is often more popular on Commons, because it gives direct sitelinks to/from Commons cats and wiki articles. And it is often the most efficient solution in terms of Wikidata items, since no extra WD items need to be created, if there are no wiki-categories and no Commons galleries to need extra WD items.
However, in this case there is a Commons gallery as well as a Commons category.
So as option 2: Does the gallery in fact need a Wikidata item? Could it just not be sitelinked to anything on Wikidata? -- ie just delete the Commons gallery sitelink; merge the new Gregory Peck item into the existing Gregory Peck item; and leave the Commons gallery an un-sitelinked orphan.
  • But then perhaps (even in a time of arbitrary access) maybe it is useful for every page on every wiki (including Commons) to be sitelinked to an item on Wikidata -- so Wikidata searches can be used to describe all pages on all wikis; and so that any templates on the wiki-page don't need to rely on any literals stored on the wiki-page, but can be written to (by default) navigate entirely using information stored on Wikidata.
This could give option 3: leave the Commons category sitelinked to the main item Gregory Peck (Q108366), let that hold the main P373 and P935 pointers to the Commons content; but keep the new Q21167568 as a stub item with just instance of (P31) = "Commons gallery" and main subject (P921) = Q108366.

What do people think? Should the new Q21167568 created by User:Tulsi Bhagat be kept, or should it go away? And do we think Commons galleries do "need" a site-linked wikidata item, or is it okay if they are "orphans" without one?

I honestly don't know what is the best way forward. But, by whatever way or other, the present situation does definitely need to be cleared up. Jheald (talk) 19:52, 2 December 2015 (UTC)

We have a sleepy project Wikidata:Wikimedia Commons, a closed Wikidata:Requests for comment/Commons links without recommendations and no clear guideline. Why exactly is it necessary to duplicate sitelinks to a property. Easier SPARQLing is not a reason at all. Sitelinks to a commomscat should be removed and replaced by the gallery link. --Succu (talk) 20:41, 2 December 2015 (UTC)
We had this discussion recently in german at Wikidata:Forum/Archiv/2015/10#Commons Category under "Andere Websites", Succu. There is a list of reasons why "remove and replace commonscat by commons-gallery sitelinks" would suck. --Atlasowa (talk) 10:20, 3 December 2015 (UTC)
There is very little point in having a Wikidata item with just one outgoing link. This, admittedly, makes it possible to get from Wikidata to that project, but not to get from that project to Wikidata or anywhere else. I would merge these two items without giving it any thought.
        Also, it is a lot more helpful for Commons categories to link to content pages than to categories. Hardly any users are interested in categories beyond the purpose they serve in organising local content (here and there, there are users focusing on categories as a purpose unto itself; these make the life of the users-who-are using-categories-to-organise-local-content miserable). - Brya (talk) 04:18, 3 December 2015 (UTC)
I want to underline this last paragraph by Brya: it is a lot more helpful for Commons categories to link to content pages than to categories. Hardly any users are interested in categories beyond the purpose they serve in organizing local content. Also, lots of these wikidata items for categories are only linked to enwiki-categories and are useless for other projects (and enwiki categories are a huge mess). --Atlasowa (talk) 10:20, 3 December 2015 (UTC)
I tend to Jhealds option 2: leave the Commons category sitelinked to the main item Gregory Peck (Q108366), let that hold the main P373(Property:Commons category) and P935(Property:Commons gallery) pointers to the Commons content; and delete the new Q21167568 item for Commons-gallery that serves no real purpose/use. (I hope i got this right) --Atlasowa (talk) 10:21, 3 December 2015 (UTC)
A little perspective form someone who works mostly on Commons: until phabricator:T49930 ("Allow accessing data from a Wikidata item not connected to the current page - arbitrary access") is not resolved there is very little use for wikidata from Commons perspective. Commons is only sleepy because we are not connected to wikidata yet, but once that is fixed there should be much more use of wikidata by Commons. For example: c:Template:Authority control, c:Template:Creator, c:Template:Institution, c:Template:Artwork, c:Template:Book should be rewritten to use wikidata data and many of them already store the Q-codes for related articles. In my opinion interwiki links from commons categories to wikipedia articles will be best handled by some template/gadget or extension that for each category will look-up the q-code of wikidata article and display separately interwikilinks to wikipedia categories (if any) and wikipedia articles. By "look up", I mean using Property:P301 of wikidata category sitelinked to that commons category. That was the original plan as far as I understood, that was just put on hold until phabricator:T49930 is fixed. Alternative approach would be to use some commons template to store q-code of wikidata article for each category, but that system would be harder to keep in synch with information stored on wikidata.
As for interwiki links to Commons galleries: they are of little importance in most cases. They were used in the earlier era of Commons but majority are not maintained and horribly outdated as compared to the categories.
Lately Jheald reported some statistics about commons-wikidata links at Commons:Village_pump. I am surprised at large number of cross namespace links, since it was my understanding that the consensus was not to allow those, so it is a bit more predictable which information will be found where. Current approach might result in unpredictable sitelinks not being used in favor of more specific properties. --Jarekt (talk) 18:19, 8 December 2015 (UTC)

List of orphan galleries

  • Here's a query for items with a Commons gallery, but no sitelinks to any Wikipedias. In all, it returns 7576 hits.
Some of these are legitimate -- eg for books, where we have a gallery of pictures -- and there are statements that can be made about them. These typically have lower Q-numbers.
However, the great majority -- probably at least 7500 items -- are the ones created by Tulsi's bot. These are the ones in alphabetical order from Commons:Reto Wikimedia Semana i Fall 2015/Campus Ciudad de México (Q21081482) onwards.
Some of these we may be able to tie to existing items. Some (eg for books, events, etc) may be worth turning into proper items with a more complete basic set of descriptive statements. Others it's not clear what will be the best approach to deal with. Jheald (talk) 17:47, 3 December 2015 (UTC)
  • To start with, here are 91 of the above that already have an item with a Commons gallery (P935) property pointing to them. Merging may not make sense in all of these cases, but it does appear to for most.
If anyone can think of any cunning ways to help match any of the rest of the 7500 to plausible matches, now is the time for ideas. Jheald (talk) 18:04, 3 December 2015 (UTC)
One set of "orphan" galleries that it may be worth keeping items for are those of the WikiCommons Atlas project. I have now added instance of (P31) = Wikimedia Commons atlas page (Q21623879) to these, to make them easy to find. (eg open Wikimedia Commons atlas page (Q21623879)     in Reasonator, and expand the "instance of" tab, in the incoming links section. Jheald (talk) 20:21, 3 December 2015 (UTC)

Geographic entity associated with a list

e.g. List of landmarks in King County, Washington (Q6624909): what would be the property here to associate with datum King County (Q108861)? Should I simply use located in the administrative territorial entity (P131)? The list does not have a geographic location; the items do. Are there special properties related to lists in this respect? And is there somewhere I should have been able to look this up? - Jmabel (talk) 00:56, 3 December 2015 (UTC)

Use is a list of (P360) = cultural property (Q2065736) with qualifier located in the administrative territorial entity (P131) = King County (Q108861)
This gives a specification for the list, and if you can then look at the list item in Reasonator, it will be able to show a list of all items that currently correspond to that specification.
The property located in the administrative territorial entity (P131) should not be used on its own as a main property except on items that are actually locations themselves -- so not on list items; also not on event items that happened to take place in a particular location. (For the latter use location (P276) instead, then P131 on the actual location).
You could conceivably use applies to jurisdiction (P1001), but I don't think anybody ever does -- it would be pretty redundant, once you've specified the P360.
However, you should set the general country (P17) = United States of America (Q30), so that the item is included in searches for all items that relate to the United States. Jheald (talk) 09:37, 3 December 2015 (UTC)
Thanks. To reiterate the latter part of my question: is there somewhere I should have been able to look this up? - Jmabel (talk) 07:46, 8 December 2015 (UTC)

Is there any intent to link to Wikia projects

This may seem like a stupid question and I apologize if it has already come up but is there any interest or intent of linking pages to other things outside the WMF sphere of influence like Wikia projects? Maybe even something basic like linking in the Other sites box? There are a lot of items that apply to various Wikia projects, especially on commons and ENWP and I think it could be a good thing to consider doing at some point if there is interest. Reguyla (talk) 21:09, 3 December 2015 (UTC)

I don't think linking to Wikia in the sitelinks section is a good idea. Wikia is a non-WMF site. Linking to outside WMF domain should only be done through statements. Mbch331 (talk) 21:18, 3 December 2015 (UTC)
What Mbch said. I don't think there is any interest in having 'sitelinks' to Wikia (like we have to wikipedia and WMF projects) but no objection in principal to linking to Wikia using a suitable property - though no one has, as yet, proposed such a property.
Wikia could, of course, add such links to their articles and can write code to add data from wikidata to their Wikia articles and wouldn't need permission from wikidata to do it.. Joe Filceolaire (talk) 06:48, 5 December 2015 (UTC)
As a Star Trek-nerd I would love to see a way to link to memory-Alpha and memory-Beta. -- Innocent bystander (talk) 13:37, 5 December 2015 (UTC)
Currently, we might already have too many sitelinks. There are a few WMF sites that might work better on separate instances .. --- Jura 16:05, 5 December 2015 (UTC)
Since there doesn't seem to be stand out arguments against the idea maybe I will propose adding a property for it and see what they say. Thanks for taking the time to comment. Reguyla (talk) 14:51, 8 December 2015 (UTC)

Trying to merge...

Q1561020 with siblicide (Q838633). Wanted to make a redirect but can't because Q1561020 isn't empty. But I can't delete anything anymore. Someone help please. -- 23:22, 7 December 2015 (UTC)

They seem to be properly redirected. Popcorndude (talk) 02:09, 8 December 2015 (UTC)

Wikidata weekly summary #187

Notability of photographs

Hi everyone, I would like to get some input. Sic19 has started mass creating items about photographs. I'm not sure about the notability of these works. For me notability is something that slides over the years. You first want to have the UEFA Europa League (Q18756) players before you start (mass) creating Eerste Divisie (Q610823) players. Another expect is how it connects to other items. An item in itself might not be terribly notable, but connects a lot of other items ("structural need" on the notability page). A good way to avoid the notability discussion (and generally make people happy) is to create really good items: complete and sourced. I don't think any object in a GLAM (Q1030034) is notable, at least not now, what are the views on this? Take for example Rijksmuseum (Q190804), it has well over a million objects in their collection. Having an object for every object is a bit too much in my opinion. Structured data on Commons is on hold, but I don't think we should try to replicate that here. It's a shame, take for example the Tropenmuseum images I uploaded a couple of years ago. It would be great to be able to add more information to these photographs.

On the other hand the photographs Sic19 is adding are quite unique. These are not some random snapshots from your smart phone. As far as I can see these are all pre-1900 and part of a historic collection. I can sure be convinced that these photographs are notable enough for Wikidata. What kind of guidelines could apply here? (Yes guidelines, I'm not a big fan of rules.) How do we find the balance here?

Say we assume these items are notable, the quality of the data should definitely be improved:

  • Labels are incorrect. "NLW3365149" at Q21670427 looks like an inventory number, that shouldn't be in the label
  • Descriptions should be expanded. Something like "photograph by John Thomas"
  • inventory number (P217) is missing, this will give a huge number of constraint violations
  • The same Commons category (P373) shouldn't be on every item

And to make it really good:

  • Source the data
  • Add the medium
  • Add the dimensions
  • Make a link to the depicted person
  • Add a described at URL (P973) to the original source.

Multichill (talk) 14:13, 6 December 2015 (UTC)

The photographs are the work of John Thomas 1838-1905, an artist with a ULAN identifier, and they are listed in Europeana. I would suggest that either of these would be enough to meet criteria 2 for notability (i.e. clearly identifiable material entity). Yes, I agree that the data needs improving but the item needs to exist before it can be edited. But, I agree with the suggested improvements.--Sic19 (talk) 14:33, 6 December 2015 (UTC)
It seems to me that the general approach seems to work fairly well, but shouldn't there be a way to do this directly at Commons? --- Jura 15:27, 6 December 2015 (UTC)
Creating wikidata items is advantageous as it offers the possibility of using tools, e.g. crotos, timeline etc., to organise and display the photographs. As far as I know this isn't possible at Commons.--Sic19 (talk) 15:48, 6 December 2015 (UTC)
The posibility to create for example a list-article on a set of items (with or without sitelinks) looks like a large opportunity to me. -- Innocent bystander (talk) 18:26, 6 December 2015 (UTC)
Just to update, I have added an inventory number to each item to resolve the constraint violations, removed the commons category and expanded the descriptions. Will now look at making overall improvements to the data quality.--Sic19 (talk) 14:34, 9 December 2015 (UTC)

Property:P527 or Property:P1012

Hello. What I need to use for add-ons? Iniquity (talk) 09:35, 9 December 2015 (UTC)

How to mark that there is no IMDb ID at the moment

On my user page I put a few people without IMDb ID (P345) claim. Some of them seem to have no IMDb entry. Is there a way to mark that in Wikidata? Maybe with novalue snak with some quantifier as timestamp? So people should see when a user checked for the last time if the item has a IMDb ID (P345) claim. So people don't need to waste their time if the last check is just a few days ago. --Jobu0101 (talk) 23:28, 7 December 2015 (UTC)

As you suspect: novalue snak + timestamp. That's the way we do it with other authority files. --Kolja21 (talk) 00:51, 8 December 2015 (UTC)
@Kolja21: Do you have an example where it is set that way? --Jobu0101 (talk) 08:03, 8 December 2015 (UTC)
@Jobu0101: Example: Samuel Marion Tucker (Q18912727). --Kolja21 (talk) 00:06, 10 December 2015 (UTC)

@Kolja21: I was asking because there is also a different common use of novalue: Camilla and Rebecca Rosso (Q54863). In this case the editor wants to emphasize that the item will never have an IMDb id. How do we distinguish those uses of the novalue snak? --Jobu0101 (talk) 09:26, 9 December 2015 (UTC)

We can't predict the future, so I would treat both cases in the same way. --Kolja21 (talk) 00:11, 10 December 2015 (UTC)
We can't predict it but in some cases it is very likely that it won't happen. See for example Nymphomaniac (Q3346699). Here it won't happen because the two parts it constis of have IMDb ids. --Jobu0101 (talk) 00:26, 10 December 2015 (UTC)
@Kolja21: Is that correct? --Jobu0101 (talk) 00:26, 10 December 2015 (UTC)
@Jobu0101: Yes. It's unlikely that a professor of computer science will have an IMDb id, but it's possible. --Kolja21 (talk) 00:55, 10 December 2015 (UTC)
@Kolja21: He has won an Academy award. So it is not unlikely that he will get an id at IMDb. He hasn't got an id yet but will maybe later. The unlikely item is Nymphomaniac (Q3346699). --Jobu0101 (talk) 01:29, 10 December 2015 (UTC)
Well than both cases are good examples where novalue is useful. --Kolja21 (talk) 01:39, 10 December 2015 (UTC)
The problem with adding a novalue to Camilla and Rebecca Rosso (Q54863) is that it violates the current constraints. Since it's unlikely (but not impossible) that groups of persons have an IMDb ID, they aren't allowed to have an IMDb ID. (And novalue is a value for IMDb ID (P345) so it violates the constraints.) If constraints need to be changed community consensus is needed first. That can be gained here (if there are enough respondents) or on the talkpage of the property.
And regarding the use of novalue when the person currently doesn't have an IMDb ID (and with that I mean the person has no entry), it makes people lazy. There is a value filled in, so they won't check IMDb even if there is a date mentioned when it was last checked. Mbch331 (talk) 08:55, 10 December 2015 (UTC)
-> Here we go! Queryzo (talk) 09:37, 10 December 2015 (UTC)
For VIAF we utilise novalue and qualify with retrieved. The verification process was updated to allow for the value.  — billinghurst sDrewth 14:08, 10 December 2015 (UTC)

Person of the Year

Is there a property? --Jobu0101 (talk) 20:27, 9 December 2015 (UTC)

Use award received (P166) with Time Person of the Year (Q207826) for her (see Willy Brandt (Q2514) --Succu (talk) 20:35, 9 December 2015 (UTC)
Thanks. --Jobu0101 (talk) 08:52, 10 December 2015 (UTC)

Sum of all Paintings update

I put an update at Wikidata talk:WikiProject sum of all paintings#More paintings and painters. Maybe you like it and want to help out too? Multichill (talk) 20:44, 9 December 2015 (UTC)

No encuentro pagina

No encuentro una pagina que esta en Wikipedia y quiera añadir informacion de Wikidata...ayuda

@ ‎Ostmar2015: ?Que es la pagina? —Justin (koavf)TCM 06:02, 10 December 2015 (UTC)
@ ‎Koavf: Marcelo Ostria Trigo
Sigo tratando de editar/añadir información biografica a la página Marcelo Ostria Trigo...
Aqui hay articulo por Wikipedia pero no hay entrada en Wikidata hoy. Es posible crearla. —Justin (koavf)TCM 06:27, 10 December 2015 (UTC)
@Ostmar2015: Disculpa, te entiendo: ?quieres mas informacion de este persona para escribir un articulo de e'l en Wikipedia? —Justin (koavf)TCM 06:28, 10 December 2015 (UTC)
@ ‎Koavf: Es mi primer dia tratando de entender la diferencia entre Wikidata y Wikipedia y como usar los dos. Lo que queria hacer es añadir la información que aparace en la columna derecha de la página de Wikipedia en páginas biográficas...algo que la página Marcelo Ostria Trigo no tiene...
@ ‎Koavf: Justin, acabo de crear una página en Wikidata para Marcelo Ostria Trigo pero no veo los cambios en la página de Wikipedia...que estoy haciendo mal?
Ostmar2015, faltaba por añadir la {{Ficha de persona}} al principio del artículo. Ya debería verse correctamente. -- Agabi10 (talk) 07:19, 10 December 2015 (UTC)
@ Agabi10: Otra pregunta, cuando no aparece el valor personalizado en las declaraciones, no hay manera de agregar el valor? en este caso, el alma mater no aparece..

Event of award received

Instead of (or maybe in addition to) mentioning the point of time of an award received (P166) I want to point out at which event that happened. For example Birdman (Q13255497) received its Academy Award for Best Picture (Q102427) at 87th Academy Awards (Q16773560). This is more information than just the time and includes the time because 87th Academy Awards (Q16773560) could have set a claim which tells the day of the event (which hasn't set so far). Do you think that change is reasonable? Which property would you use to set the event as quantifier? --Jobu0101 (talk) 08:52, 10 December 2015 (UTC)

Adding this as a qualifier is perfectly fine. Thanks, GerardM (talk) 09:15, 10 December 2015 (UTC)
@GerardM: So which property shall I use? --Jobu0101 (talk) 09:32, 10 December 2015 (UTC)
significant event (P793)? Queryzo (talk) 09:47, 10 December 2015 (UTC)
If I check Property talk:P166 I would say statement is subject of (P805). significant event (P793) isn't in the list of allowed qualifiers for award received (P166). Mbch331 (talk) 09:50, 10 December 2015 (UTC)
Thanks. And which property shall I use to enter at 87th Academy Awards (Q16773560) the date the event took place? --Jobu0101 (talk) 10:05, 10 December 2015 (UTC)
point in time (P585). Mbch331 (talk) 11:05, 10 December 2015 (UTC)
@Mbch331: Wikidata model is incomplete and we won't achieve completeness if we discuste everything extensively. Just add it to the documentation, describes your intentions, and if nobody objects, Voilà ! author  TomT0m / talk page 10:50, 10 December 2015 (UTC)
@TomT0m: I was just saying what I found where. I wasn't saying he should go to the talk page and discuss it there. Mbch331 (talk) 11:05, 10 December 2015 (UTC)

Sr. as part of the birth name

Is it ok to add "Hamden Foo, Sr." as a birthname? No one is known as Sr. at birth. You become one when Jr. is born. So is ok to use it as a retronym? --Richard Arthur Norton (1958- ) (talk) 16:24, 10 December 2015 (UTC)

In the US there is conflicting information and confusion about whether suffixes are (1) part of the name or (2) a description (just like brown eyes, weight 65 kg, etc.). Also, in some families individuals keep their suffixes for life, while in other families, suffixes are reassigned after the death of an ancestor. Finally, many people who have a descendant with the same name do not use the suffix "Sr." I would simply not use suffixes as part of the name. I would use a separate property, or if one does not exist, propose such a property.  – The preceding unsigned comment was added by Jc3s5h (talk • contribs) at 21:18, 10 December 2015‎ (UTC).

Harassment consultation

Please help translate to your language

The Community Advocacy team the Wikimedia Foundation has opened a consultation on the topic of harassment on Meta. The consultation period is intended to run for one month from today, November 16, and end on December 17. Please share your thoughts there on harassment-related issues facing our communities and potential solutions. (Note: this consultation is not intended to evaluate specific cases of harassment, but rather to discuss the problem of harassment itself.)

Regards, Community Advocacy, Wikimedia Foundation
 – The preceding unsigned comment was added by MediaWiki message delivery (talk • contribs). 17 November 2015 (UTC)
adding comment to get this archived. --Lydia Pintscher (WMDE) (talk) 09:55, 11 December 2015 (UTC)

Robert de Niro

I don't speak hy. Can you tell my why there are two articles ([2] and [3]) in hywiki? --Jobu0101 (talk) 15:58, 8 December 2015 (UTC)

check Wikidata, e.g. Q21198666. --- Jura 16:09, 8 December 2015 (UTC)
And not even one article about Robert De Niro (Q951321) in hy :-( Multichill (talk) 18:24, 8 December 2015 (UTC)
Hi Jobu0101. There are two standardized literary forms of Modern Aremnian. But we use the same wiki ( :-(--ԱշոտՏՆՂ (talk) 19:45, 8 December 2015 (UTC)
Are there plans to spilt them into two wikis? At least you should link the two items like Jura1 did here. --Jobu0101 (talk) 20:41, 8 December 2015 (UTC)
Yes, there are (we have some problems with iso code). I will work on it--ԱշոտՏՆՂ (talk) 08:01, 9 December 2015 (UTC)
Jura refers to Zaven (Q21198666). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 00:07, 10 December 2015 (UTC)

Jobu0101, Jura1, do we also need this?--ԱշոտՏՆՂ (talk) 11:17, 11 December 2015 (UTC)

Property:P460 is symmetric. So, from time to time, bots add missing halves. It allows to find all content starting from the main item (Q36949). --- Jura 11:26, 11 December 2015 (UTC)

API sandbox examples should have a Wikidata-prefix when used with titles-argument

They don't work properly now with "API|Main Page". [4]

The "Wikidata:API|Wikidata:Main Page" works [5]

Also not mentioned is Property-prefix, that is needed with property ids

Well it seems that it's safest to use the Q123456 id's with the API

--Pasixxxx (talk) 18:20, 10 December 2015 (UTC)

If you think this is problem, you should rather tell this the developers. Matěj Suchánek (talk) 19:32, 10 December 2015 (UTC)

Wikidata language links mysteriously not showign up on German WP article

Hi guys. I found a German Wikipedia article that does not link any of the other language articles in the left sidebar: w:de:Suprafluidität – even though it is included in the Wikidata item: Q106667. Other language versions of the article include all the links, eg.: w:Superfluidity. I did not find the {{noexternallanglinks}} magic word in the de article and am at my wit's end as to what is causing this issue. Can you people help me troubleshoot and fix this? – 07:27, 11 December 2015 (UTC)

Never mind, aude in the IRC chat found the solution: they simply added ?action=purge to the article URL to flush the server cache :). – 07:31, 11 December 2015 (UTC)

Wikidata on The Signpost

Questionning and critisms about Wikidata on the SignPost : en:Wikipedia:Wikipedia_Signpost/2015-12-02/Op-ed Congrats folks :) author  TomT0m / talk page 18:11, 7 December 2015 (UTC)

  • Please suggest improvements to the wikipedia bot import issue on this RFC. ArthurPSmith (talk) 21:36, 7 December 2015 (UTC)
  • Everyone should read this. These are strong criticisms and if we don't modify our policies we will definitively have some big difficulties to spray the use of WD in WP. We should improve the quality of our sources and I think that the change of the licence should be serously considered. Snipre (talk) 09:10, 8 December 2015 (UTC)
  • This piece is very much similar to what people wrote about Wikipedia at the start. There is not much in there except for a singing around of the opinion of others.. Thanks, GerardM (talk) 15:56, 8 December 2015 (UTC)
Yes, but Wikidata came after Wikipedia and this means that we didn't learn anything from the Wikipedia experience. Then when Wikipedia started and even now no other projects are relying on Wikipedia: if Wikipedia is good or not doesn't bother anyone outside of Wikimedia. This is not the objective of Wikidata which from the start said that Wikidata will support Wikipedia. At the end Wikidata can improve at its own speed but in that case we will be isolated and Wikipedia will never use Wikidata. That's what you want for Wikidata ? Snipre (talk) 22:30, 8 December 2015 (UTC)
@Snipre: Obviously, we all want good data on Wikidata; and ideally we want to see Wikidata data used as widely as possible, not least because that way it will get the most scrutiny, the most feedback, and the most people interested to improve and extend it.
But we also have to recognise that Wikidata is a wiki, that anyone can edit at any time, and not all the data here is going to be right.
(Actually, there probably isn't a database in the whole of the humanities that doesn't need some cleanup, but that's a different story).
I can see why the "strong gatekeeper" model can seem attractive -- thinking along the lines of "let any contradiction into a mathematical system, and the result is that no conclusion can be relied on".
But the truth is that in reality that is not how a wiki works -- we will get contributions from various sources, of varying quality, and some of it will contain errors. This is the bazaar, not the cathedral.
Saying it is all the fault of bots is just putting on blinkers. Any human editing is going to introduce mistakes too -- the more confident we are as individuals that our work is 100% error free, the more likely we are to be deluding ourselves. It's all too easy to accidentally pick the wrong homonym from a drop-down list, or accidentally mis-key a date or a number, or be fooled by a language difference, or give a slightly incorrectly chosen value for a property.
A couple of weeks ago, I spent an entire weekend un-merging various sites of heritage significance that had been merged together by people using the merge tool, despite often being located anything up to 300 km apart. Next on my list are various suburbs in England, accidentally identified as incidences of "battles" by GerardM. After that, there will be who knows what next to clear up -- painters where birth dates have been muddled up with activity dates, perhaps.
Where users already have strong data -- eg German wiki and French wiki, in many areas -- they are probably quite right not to trust our data without question, but instead to look at it either as suggestions to be further investigated, or to use in templates that will keep an existing value, but flag if a Wikidata value becomes different from it.
On the other hand, where data may currently be weak (eg many Commons categories?), and where applications may not be quite so sensitive, even flawed data may be worth having as generally a good enough approximation to the truth. Already, Wikidata and SPARQL are making it possible to do some incredibly powerful searching and querying. In some applications, having the results as a set of first suggestions, accepting that they may need further filtering and investigation, may in itself already be hugely valuable.
There are several reasons that I am less than convinced by the thesis that we need to slow down material being added to WD and switch to a model of much "stronger gatekeeping". In particular:
  • I am not convinced by the proposition that bot added content is significantly worse that any of the other content already here.
  • In some particular subject areas there may be 'gold-standard' databases that we can extract material from, that can be freely reusable CC0, but in most subject areas such things are rare. The alternative to mass additions from Wikidata is likely to be ad-hoc user additions which probably will mostly still come from Wikipedia, and even when not will probably tend to be added in an unsystematic and often error-prone way.
  • Item quality is not the only virtue. There is also a virtue in comprehensiveness. People who come to Wikidata not unreasonably assume that we will have a similar level of coverage as Wikipedia. Compared to what is in Wikipedia, they would be entirely justified (IMO) to be really very disappointed indeed in how spotty and thin and incomplete our coverage is here on Wikidata. For some applications, reasonable comprehensiveness with occasional inaccuracy may be far preferable to 100% accuracy but poor coverage. Worst of all, of course, is if people think they have got the results of a reasonably comprehensive search, if in fact those results contain serious gaps.
  • Finally, but perhaps most significantly, getting more data in from Wikipedia is actually a hugely help in assimilating data from external sources. The more comprehensive our data already is, the more identifiers we have, the more properties it contains like eg roughly accurate nationality, occupation, dates of birth & death, works created, etc, in many cases all that is what it makes it possible or not possible to accurately match an entry in an external database (which may often have a significantly different label) -- and also to pinpoint mismatches that may have occurred.
Yes, we should be doing all we can to improve the data we have; and to try to work out what kinds of mistakes most often get made, to warn people adding data (both by hand, and by higher volume methods) of different kinds of "gotcha" patterns that they would be well advised to be wary of.
But my own view is that we should very definitely be trying, as urgently as possible, to capture as much as possible of the huge amount of data in infoboxes, templates, categorisations, etc on Wikipedia that is not yet in Wikidata -- and that (at least in most subject areas) calls to restrict to only data from independent external sources are utterly utterly misguided, and typically bear no relation to either what is desirable, what is available, or what is still needed in order to utilise such sources effectively. Jheald (talk) 23:49, 8 December 2015 (UTC)
Jheald Just give one answer: what is the purpose of Wikidata ? If your goal is just to have a database full of unreferenced statements and which is not used by anybody, yes your argumentation is logical.
But when the main users of data say they don't consider WD as reliable, continuing to import unreferenced data is just waste time and energy. There is a strong issue among Wikipedians about data quality (and I don't speak about errors, but about verifiability). How can you answer that problem ? Yes we improve the data but at the same time you continue to import unreferenced data ? Where is the logic behind that ?
Again what is the purpose of doing that ? For who, for what ? We can't say that we aim to provide data to Wikipedia if we are not able to hear what Wikipedians say. We can't say that we improve our data if we don't enforce your criteria for data import and if we don't delete unreferenced data when referenced data is available. Just to be clear, data sourced by Wikipedia is not sourced according to Wikipedia rules.
Wikidata is becoming unreliable for me not because of its young age but because the original objective of Wikidata, to be a reference database for Wikipedias, is completely abandoned in fact. I can't continue to promote Wikidata in Wikipedia when I see that no improvements are done in data import to select reliable and sourced databases in priority.
If you promote the current data import policy, you have to accept the criticisms of the article saying that we are doing some bullshit and that Wikidata has to be avoided in Wikipedia for the next 5-10 years. Snipre (talk) 01:11, 9 December 2015 (UTC)
@Snipre: One of my personal main goals for Wikidata at the moment is to help me deal with 50,000 maps being georeferenced, to help me put them in the right categories when I upload them to Commons. So from the co-ordinates coming out of the georeferencing, I want to be able to look up what features are near by, what kind of features they are -- eg castles, cathedrals, battlefields, towns, cities, etc; and what their Commons categories are, in order to give me as good a first guess as possible at what categories to put them in, given that the images themselves come with no title or text at all, other than the book they were scanned from, that might contain anything up to 50 different maps. I don't need perfection: I need a good first guess, that can then be manually curated further.
Another area I have done a fair amount of work in is cross-referencing of artists to records in different external catalogues -- eg these from the UK "Your paintings" list. Links like these, eg to the Your Paintings site, or RKD artists, or other external references are hugely valuable when we can match them up to Commons categories and show them on Commons templates. If the occasional wrong link gets in, that's too bad; and obviously we need to do our best to find them and correct them. But the real value is the good links. And this is the classic use of a wiki -- a source of "good enough" information, that you can use to click through to more detailed "gold standard" external sources for more. The more such links we can provide and link to Commons categories the better.
I am also interested in Wikidata as an index to Wikipedia, for searches like SPARQL, to see what coverage we have in particular fields, places or times. This is where the amount of information missing from Wikidata can be intensely frustrating, if I know that I and other searchers are not getting back all the hits that they should. For example, until recently the tree of the UK administrative subdivisions (eg for England) was patchy to such a degree that it was essentially not usable -- most elements simply had located in the administrative territorial entity (P131) = England, making the tree un-traversable; and items at all levels -- including even two top-level NUTS-1 regions of England -- were simply not identifiable for what they were, missing even P31 statements. Now, all the elements down to district level are appropriately identified as what they are; with the appropriate P131 values, so now the tree is usable, and can now be used to start searching for returns from particular parts of the country. All of the information needed to do that came from the excellent pages on en-wiki, and I feel no compunction at all about having used that as a source, which now gives a working set of pages, and items at each level that now have the right P31s to describe what they are. A small success story, but very typical of how important a degree of comprehensiveness monitoring is; and of how much basic information we still need to report, to make Wikidata searches capable of generating even the most simple of useful returns.
I wrote at the top of my previous post, that fundamentally it is important to recognises that Wikidata is a wiki. It is not a closed, cathedral-like authority. Results from wiki should always be treated with caution, and defensively. Established wikis like fr-wiki and de-wiki should treat our information with suspicion, if they have existing information of their own (and even a reference here is not necessarily some magical filter, because referenced information too can all too often be mis-coded, or applied to the wrong object, or be coded in terms of items which don't quite correspond to local-language terminology). To look for Wikidata to be some kind of infallible authority is simply not a realistic aim. But to treat Wikidata like any other wiki -- as a "good enough" summary, for purposes where a "good enough" summary is good enough, and as a prompt for further research and investigation when better than "good enough" is needed -- that is an appropriate role for a Wiki, one that Wikidata is certainly stepping up to; and in which it is becoming more and more useful, as it is becoming more and more comprehensive and less patchy. Jheald (talk) 09:25, 9 December 2015 (UTC)
Jheald WD will never be a reference by itself. I don't know where do you find that idea. WD is just a telephone book for data where people can look for data and after they can go to the primary sources if needed. As Wikipedia. Wikipedia doesn't say the truth, it say what others said. What is the critical point is the links between all these data repositories. Your definition of wiki is really interesting: just a bad tool without any ambition to become something better than a simple bunch of code lines. Your description of a wiki gives as impression that you treat WD as a tool or a game and you don't have a higher vision. Snipre (talk) 18:05, 10 December 2015 (UTC)
Snipre, the announcement of the closure of Freebase included the following passage: [...] we've decided to help transfer the data in Freebase to Wikidata, and in mid-2015 we’ll wind down the Freebase service as a standalone project. Freebase has also supported developer access to the data, so before we retire it, we’ll launch a new API for entity search powered by Google's Knowledge Graph. Loading Freebase into Wikidata as-is wouldn't meet the Wikidata community's guidelines for citation and sourcing of facts -- while a significant portion of the facts in Freebase came from Wikipedia itself, those facts were attributed to Wikipedia and not the actual original non-Wikipedia sources. So we’ll be launching a tool for Wikidata community members to match Freebase assertions to potential citations from either Google Search or our Knowledge Vault[2], so these individual facts can then be properly loaded to Wikidata. In fact, though, the procedure here is often exactly the same as in Freebase, which fed or feeds the Google Knowledge Graph, Bing Snapshot etc.: attribution to Wikipedia, nothing else. For Google and Bing, this was deemed good enough, maybe because they do their own internal crosschecking, or because they simply decided it was good enough to earn money with. So when you ask "For who, for what ?", one possible answer is: search engines. Both Google and Yandex have contributed funding to this project; AI2, the third funder, is the institute of the founder of Microsoft, another major player in the search engine market. I know what has been said about the purpose of Wikidata, but if you look at what is actually being done, it seems to me to be more consistent with that use than with the intent to create a central database that can reliably inform the various Wikipedia projects. Because of its lack of reliable sources, Wikidata content is for the most part unusable according to the policies of Wikipedia. Andreas JN466 16:26, 10 December 2015 (UTC)
Andreas The question is not what the "investors" wanted but what the community wants. I still believe than if we decide to choose quality we will able to get out of the current situation. Google, Bing and the others can use WD, they don't say to the community what they have to do. So even if the creation of WD is a baby of big compagnies which want to do money, the work is done by a community. Google, Bing abd the others will benefit from that work and for nothing but this is the principle of Wikimedia: everyone can use the common work to do profit, nobody prevent you to print wikipedia articles and to sell them as books. Same for data from WD. Even if WD is a economical bet for those compagnies I find that WD is a good concept for Wikipedia too. So I can work with the idea of working for big compagnies even without being paid because I found some personal interests which have not economical worth. All data which was imported was not done by Google, Bing and the others and we can safely remove all unreferenced data without loosing any contract.
I understood the reason of licence: the licence gives the possibility to big compganies to reuse data with no restriction. Again this is not a problem for me. The only problem I found with the license is the prevention to use official data because the minimal requirements is the citation and I am in favour of changing the license to have more potential data sources and to respect the author right to be cited. I don't know if the choice of the license is a community option or a choice of the owner of wikidata. This should be clarified and I hope Lydia can gives some informations about that. Snipre (talk) 17:48, 10 December 2015 (UTC)
Snipre, Denny once said, "an early decision for CC0 can be easily changed by a relevant community, once it exists. So I am more convinced then ever that CC0 is a good license, for starters." Lydia, as you know, is firmly opposed to any licence change.
On the value of crowdsourced work, Google alone makes close to $200 million a day from advertising. The third-quarter (July–September) results for 2015 in that table list advertising revenue of $16.8 billion, i.e. over 15 billion Euro. A significant part of this can be attributed to the value Wikimedia content in the Knowledge Graph adds to Google's search engine results pages: it keeps users on Google pages, because they don't have to click through to Wikipedia or some other site any more if they find the information they were looking for on Google's own page. And this increases the statistical probability that a user may click a paying ad on the Google page. I know we all work on Wikimedia projects without the expectation of payment, but once the work done here leads to an accumulation of such astronomical sums at Google (and others platforms like it) it no longer feels compatible with the principles of social justice.
It always happens at the beginning of technological revolutions that a lot of money ends up in the hands of a few, while those who do the actual work on the shop floor get very little (or nothing, in our case). It was so in the industrial revolution in the 19th century, and it is happening again now, in the digital revolution, in a slightly different way. Long-term, I am convinced this will change, just like the exploitative excesses that were common in the early days of the industrial revolution were curbed. And the sooner it happens, the better. Communism has been a complete failure as a political system, but there was a reason – real abuses – why it arose when it did. And the work of unions 100 years ago was a critical factor in getting us to the more just labour arrangements of today. Andreas JN466 14:06, 11 December 2015 (UTC)
Andreas But you have to agree that Google, Bing and others already gave some money to Wikidata so we got part of the money. Perhaps not proportionally to what they earn but I think we can't criticize the use of the data to do money when we explicitly provide that right. We can't publish data or articles under the CC0 or CC-BY or CC-BY-SA and be offended when seeing people doing money with our work. We can't blame people for the consequences of our choice.
Then stay the question of the CC0 license. For me this license is very bad because the citation is a primary right of the data authors. This prevents to use a lot of data from official sources and we are facing here an ethical problem. Then the citation is a non negotiable point in term of data use: this is the only way to judge the quality of the data and to prevent misuses like comparing data from 50 years ago with current data. I have a scientific point of view and only data with the same quality can be used in any comparison or kind of analysis.
But before to continue to discuss about this topic we have to know who can decide about the license choice. If this is not a choice of the community, this should be clearly stated and let then people decide if they want to continue to work inside WD. Personally when I look at the direction taken by WD I think I will choose to use another system because WD is not respecting its foundation principles, citation of sources to respect author minimal rights and to offer a reliable database which respects the Wikipedia principles. The French Wikipedia will decide in the next months about the use of WD and currently I can't support the use of WD without a lot of limitations. I hope the English Wikipedia will do a similar vote in order to put WD in front of a choice: to continue like now and be an isolated project in Wikimedia or to be ambitious and to increase its data quality and be an important player in the Wikimedia world. Snipre (talk) 15:49, 11 December 2015 (UTC)
Snipre, of course I understand that it's our own fault for making content available under CC-BY-SA. :) Similarly, it was workers' fault in the 19th century to sign contracts stipulating they work 80 hours a week all year round, without days off. They signed them, and were bound by them. But the legal landscape eventually changed to prohibit such arrangements, because they were against human dignity. This was a process that took decades. I don't think it will be any different in the case of unpaid crowdsourcing labour.
On the topic of the CC0 licence, please see today's exchange with Denny on the mailing list: Denny, me, as well as [6] for background.
I completely agree with you regarding the use of unsourced Wikidata content to populate infoboxes in Wikipedia. You can't import infobox content from Wikipedia in Wikidata and then use that same Wikipedia-derived content as a source in Wikipedia, bypassing the need for any external source backing up that data. Using Wikidata content is okay if and only if the statement in Wikidata cites a reliable non-Wikipedia source backing the data up; moreover, that source must come across to Wikipedia and be visible there so that readers can verify the data. Bonne journée. Andreas JN466 14:59, 12 December 2015 (UTC)
If I remember right, dear GerardM, you are one of the supporters advocating a fast addition of claims from wikipedias without referencing to a reliable source. --Succu (talk) 22:44, 8 December 2015 (UTC)
I am glad to have written a rebuttal to this obvious FUD. I am also glad that I can argue why it is FUD without calling names. Sources are not reliable by definition, in the same way not having sources makes it obvious incorrect. The argument for adding more data is very strong particularly when it is accompanied by increasingly sophisticated processes that flag errors. As you know, quality has many aspects, this piece of trash only knows about one, the one that helps us the least. In the mean time I have (again) blogged about quality in my rebuttal and, it is easy and obvious to recognise where we are strong and where we are weak. I do that and, I point to the way forward for all of them. Now I am proud to support the addition of data from sources that are trustworthy I do not need to snipe from the side to make my mark. Thanks, GerardM (talk) 07:00, 10 December 2015 (UTC)

en:Persian art

clearly has the wrong code - the corresponding language articles are all wrong. Thanks Johnbod (talk) 03:07, 11 December 2015 (UTC)

@Johnbod: Fixed mismatch of interwikis by rolling back Persian painting (Q850114) to former English interlink, en:Persian painting. However en:Persian art remains dangling all by itself, because another existing article, en:Arts of Iran, grabs all interwikis associated to that subject (via Iranian art (Q2718840)). Consider merging en:Arts of Iran to en:Persian art. Other suggestions welcome. ---- LaddΩ chat ;) 04:26, 11 December 2015 (UTC)
"Persian art" has replaced en:Arts of Iran, which covers literature, music etc as well, and should take all links for the main art history article. Actually en:Persian painting redirects to "Persian art", and is not a good equivalent for the French and German articles, which lack an English equivalent, so I'd zap that one too. Johnbod (talk) 20:58, 11 December 2015 (UTC)

Done Johnbod (talk) 04:41, 12 December 2015 (UTC)

Also, the French Peinture de maison de café and German equivalent need their own number, they are not at all the same as the others there. Johnbod (talk) 04:34, 12 December 2015 (UTC)
@Johnbod: You should not leave unlinked Wikidata items, nor should you leave Wikipedia items with no corresponding Wikidata entry. Since you assigned en:Persian art to the Wikidata labelled "Iranian art", I merged the statements of the empty Q21704493 into Iranian art (Q2718840). ---- LaddΩ chat ;) 14:19, 12 December 2015 (UTC)

CommonsDelinker removes sources

This was a strange way to replace a file. In this specific case, I cannot see that anything important is lost. But there are maybe other cases? -- Innocent bystander (talk) 18:34, 11 December 2015 (UTC)

In this case it indeed isn't a problem, but in other cases it might. Somehow the bot needs to check if there are qualifiers and/or sources and make sure they get copied to when updating the statement. (I think the bot currently removes the old statement and adds a new statement instead of actually updating the statement). Mbch331 (talk) 20:17, 11 December 2015 (UTC)

Tōkai Nuclear Power Plant (Q738945) and Tokai Power Station (Q11527936)

Q738945 is Tokai No.2 Power Station while Q11527836 is Tokai Power Station, which locate next to Q738945 and is the first one there, according to infobox of Japanese Wikipedia article

However, for most other language's Wikipedia (except Japanese/Chinese Wikipedia) Their entry in Q738945 is "Tokai Nuclear Power Plant" which include both Q11527836 as Tokai-I and the Japanese-Q738945 as Tokai-II.

Should The Q738945 remove Chinese/Japanese link and create a third wikidata entry for them as they are not matching? If this is to be done then English/etc. Wikipedia article on the matter would lost their link to the language of the article subject? Is there any better way to do this?

C933103 (talk) 09:42, 12 December 2015 (UTC)

I think they should be split. They who want to still have interwiki between Tokai-I and Tokai Power Plant can use templates like Template:Interwiki extra (Q21286810). -- Innocent bystander (talk) 10:00, 12 December 2015 (UTC)

Closing in towards 1 million links from article-items to Commons categories

Links from article-items to Commons categories are now less than 8,000 away from hitting the 1 million mark, according to these latest numbers. (See also historical comparisons and details of actual queries run).

Commons categories
Commons galleries
total linked
Wikidata articles
992,272 94,446 1,026,391
Wikidata categories
353,717 249 353,749
total linked n/a n/a n/a
props: 1,376,447 items /
1,197,999 pages

A handful of caveats

  • "Articles" and "Galleries" in the above actually include everything that is not a category -- so some of the sitelinks in the article-items / commons galleries box are actually sitelinks from template items to Commons templates. This accounts for most of the 3700 sitelinks that don't have a Commons gallery (P935).
  • The figures make no attempt to take account of the constraint violations listed on e.g. the constraint vio page for P373, that may imply some double-counting -- in particular, there appear to be quite a few cases where one or more elements of a class have a property Commons category (P373) set for a particular Commons category, as well as there being one on the item for the class itself. All of these have been included above.

A breakdown of P373 by P31 ( suggests these million article-item -> commonscat links include the following leading classes:

  • people: 222,874; taxons: 88,069; churches: 31,691; railway stations: 15,027 ...
  • administrative territorial entities: 191,407 -- breakdown:
  • items with heritage status: 74,316 -- breakdown:

The user-script wdcat.js may also be of interest, which adds a Reasonator link to a Commons category page whenever it is the target of a P373.

Getting to a million is no small achievement, but it's still only a quarter of all the categories on Commons. It would be really good to push this forward some more, so we can make it a lot easier for external programs to find the right Commons categories to put images into; and because these identifications will become particularly valuable for Commons templates, once they become able to systematically draw from all Wikidata items early next year. Jheald (talk) 22:01, 6 December 2015 (UTC)

A version of this comment also posted to Commons village pump. Jheald (talk) 23:16, 6 December 2015 (UTC)
Hi, I am trying to help like this. Is this the proper way to link categories form Commons to Wikidata? Regards, Sturm (talk) 01:36, 7 December 2015 (UTC)

Article2Category sitelink blocking WikipediaCategory2CommonsCategory

@Jheald, Sturm: This blocks linking CommonsCategory to ptWpCategory.

Jheald's table reduced and colored: green=preferred, red=potentially blocking preferred links, yellow=redundant.

Commons categories
Commons galleries
Wikidata articles
Wikidata categories
188 10:14, 13 December 2015 (UTC)

Property talk:P527

Hello. Do you agree with Property talk:P527#Class has part non-class and [7]. Xaris333 (talk) 02:40, 8 December 2015 (UTC)

To clarify the linked edit, my understanding is that the has part (P527) statements contradict the subclass of (P279) statements going in the other direction. --Yair rand (talk) 04:49, 8 December 2015 (UTC)
I guess I must re-link my proposition Wikidata:Property proposal/Generic § subclass then :) It's a proposition that would allow to link a class to their subclass in some cases. I tend to agree with Yair Rand that a subclass Csub of a class C cannot be a part of C. author  TomT0m / talk page 08:17, 8 December 2015 (UTC)

I use a lot this property. I just want to know if I am using it in correct way. In its description says that object that is part of this subject (and it is the inverse of part of (P361)). I want for example, a property in Cypriot Cup (Q245970) where I can list all the seasons of the Cup, like 2013–14 Cypriot Cup (Q15094037). Do we have something else? (Yair rand, I am sorry for opening a conversation here, but sometimes is difficult for me to argument in English. I didn't understand your arguments so I thought it would be better just to ask the community to say their opinion). Xaris333 (talk) 08:55, 8 December 2015 (UTC)

This is not really a correct usage. A specific game is a part of a cup instance like 2013–14 Cypriot Cup (Q15094037). But Cypriot Cup (Q245970) is essentially a type of competition which occurs every year. So we have . author  TomT0m / talk page 08:59, 8 December 2015 (UTC)

I have responded at greater length at on the property talk page. But I do think Yair rand and TomT0m are right, to the extent that if one finds a statement such as

one needs to be able to interpret it without ambiguity as saying that "every baryon contains a quark", rather than "each quark is a kind of baryon" (incorrect).

However I also agree with TomT0m that we do need a property that can sometimes act as an inverse of instance of (P31) and subclass of (P279) -- in particular, to be able to specify a closed list of elements or subdivisions that a larger class contains.

The main discussion about how to move forward with such a property seems to be at Wikidata:Property_proposal/Generic#.28Disjoint.29UnionOf_.28or_any_better_name.29, and a couple of ideas look promising. One would be to introduce a new property "union of", so that a class like Muses (Q66016) could be "union of" "instances", with qualifier quantity (P1114) to give the total number of instances, and a qualifier "has element" for each element, eg Clio (Q103968), Calliope (Q103975) etc. Similarly nucleon (Q102165) could be "union of" "subclasses", with qualifier quantity (P1114) = 2, and qualifier "has element" = proton (Q2294), neutron (Q2348).

Alternatively, more generically, one might introduce a property-valued property "object of", to be able to add descriptive qualifiers, so eg Muses (Q66016) "object of" instance of (P31) could then again have qualifiers for total number of instances and each of those instances; or nucleon (Q102165) "object of" subclass of (P279). This mechanism could be extended to other properties, for example Grade I listed building (Q15700818) as "object of" heritage designation (P1435), to which one might want to attach as qualifier the proposed property "expected completeness".

So I think that we should probably hold off from deleting P527 statements of this kind (even if we agree that it is not an appropriate use of P527), until there is another property that has been created, ready to take up the role. Jheald (talk) 10:18, 8 December 2015 (UTC)

Sorry I don't really get your point. You seems to mix a lot of subjects here ... Any of the usecase should be explained on his own. Maybe you're going a little bit too much generic ? If we end with a "we should have a triple "item/property/value properties qualified with a qualifier one" we will just have replicated the Wikidata data model in Wikidata :) We already have a quantity (P1114)   qualifier which was originally designed to note the number of instances a class have for example. author  TomT0m / talk page 10:59, 8 December 2015 (UTC)
It seems to me the issue is perhaps mostly one of display - the information A subclass of B (or instance of, etc) is already in wikidata, but we want to be able to see when looking at B that it has a subclass A. "What links here" almost does this, but could we maybe consider an addition or option in wikidata display that shows all the properties an item is a target of, and a (limited) list of the source items for each such property? I think this would address most of the above concerns except the issue of whether the resulting list is complete or not (are these all the known subclasses or instances). ArthurPSmith (talk) 14:30, 8 December 2015 (UTC)
No it would not be enough. There is (might be) several way to complete the superclass, for example if you class human, you can have one set of subclasses "man/woman/other" which covers the set of all humans. But there is other way to class humans, by age for example : adults/childrens, which also covers the set of all humans. author  TomT0m / talk page 14:37, 8 December 2015 (UTC)
One thing I would add, which is in addition to the case that TomT0m gives of classes that can be broken down in different ways into different groups of sub-classes, is that even for classes of individuals such as Muses (Q66016), if the class has a distinctive tightly-defined enumerable membership, I think there is a security in listing that membership on the item, where any changes can be directly seen in the item history, rather than just depending on the P31 statements on the different member items. If the membership information is only scattered across all the different items, then its completeness is vulnerable to whatever happens on those items -- including possible deletions, inappropriate edits, bad merges, or just somebody choosing to make another class "preferred" with the effect that our class is suppressed in searches.
Wikidata being a wiki that anybody can edit is a very good thing; but as a result, like DNA, it is sometimes a good thing to have a bit of redundancy and defensive thinking in how we store information. Jheald (talk) 17:33, 8 December 2015 (UTC)
This is a very clear argument in favor of this, thanks Jheald. So then would "object of" as a property be sufficient? It seems TomT0m wants to allow for two or more "object of" claims for each way of breaking things down into different groups of sub-classes, would that work? I guess maybe an extra qualifier indicating somehow the reasoning behind the sub-class groupings? ArthurPSmith (talk) 16:28, 9 December 2015 (UTC)
Even if it would be useful to list some subclasses, are we in agreement that, on the long term at least, P527 is not the right property for this? --Yair rand (talk) 18:43, 13 December 2015 (UTC)

cuando no aparece el valor personalizado

cuando no aparece el valor personalizado en las declaraciones, no hay manera de agregar el valor? en este caso, el alma mater de la persona no aparece ..  – The preceding unsigned comment was added by Ostmar2015 (talk • contribs) at 11 dec 2015 00:17‎ (UTC).

@Ostmar2015: En estas situaciones simplemente debes crear el elemento. Puedes utilizar Wikidata Item Creator o New Item. Allan Aguilar (talk) 16:04, 13 December 2015 (UTC)

Merge Q2551525 and Q131681

I came across Q2551525 (Wasserspeicher in German). I'm not familiar with German but it looks very similar to a reservoir in English. Should it be merged with Q131681(reservoir, Stausee in German)? Or should it be linked somehow? I can't quite grasp the difference and would appreciate if a German speaker could take a look. —Tom Morris (talk) 12:25, 11 December 2015 (UTC)

water reservoir (Q2551525) (artificial and natural reservoirs) seems to be a superclass of reservoir (Q131681) (artificial lakes for storing water). I set the subclass attribute in the latter for now, although I think someone with more knowledge of the subject should have a look. --Srittau (talk) 13:17, 11 December 2015 (UTC)
And while looking at it, I have also come across Q12041539, Nádrž in Czech. Google Translator did not help much. The Wikipedia article seemed to hint at a container for liquids, but it also references reservoirs as an example. Maybe someone speaking Czech can help here? --Srittau (talk) 16:04, 11 December 2015 (UTC)
(cs-N) Yes, you are right: nádrž = whatever you can store watter in. Adjectives change its meaning (vodní ... = water reservoir (Q2551525), palivová ... = fuel tank (Q1411232), přehradní ... = reservoir (Q131681)). Matěj Suchánek (talk) 16:32, 11 December 2015 (UTC)
Okay, from what I understand, to me it seems the most logical structure would be:
Although it is a bit unclear to me at the moment, what to do about non-water liquid storage. --Srittau (talk) 17:33, 11 December 2015 (UTC)
And there is also artificial lake (Q3215290). All of the items have a different link to cswiki, so it is impossible to merge them. It would be great to establish the differences between them though. @Matěj Suchánek: Any hints? Andreasm háblame / just talk to me 01:33, 12 December 2015 (UTC)
Added to the list. Matěj Suchánek (talk) 09:30, 12 December 2015 (UTC)
A rather complex topic, it seems. I also found Ziegelteich (Q198732) and added it. From what I understand, excavation lake (Q358976) is an artificial groundwater lake, created by removing material, e.g. for surface mining. Ziegelteich (Q198732) (a term I had never heard before) is a special case of that, created by removing loam. --Srittau (talk) 14:59, 12 December 2015 (UTC)
I also moved reservoir (Q131681) under artificial lake (Q3215290), since that seems to make sense. --Srittau (talk) 15:04, 12 December 2015 (UTC)
And thinking about it a bit more, is artificial lake (Q3215290) really a sub-class of water reservoir (Q2551525)? There are artificial lakes, created by accident (Salton Sea (Q503301), for example), or just for recreational purposes. I wouldn't consider those to be "water storage", which to me indicates that the water is going to used for some purpose (drinking water, to power engines, or for firefighting) later. --Srittau (talk) 15:08, 12 December 2015 (UTC)
I've added „Tagebaurestloch“ (abandoned mining pit / remnant mine lake)--Kopiersperre (talk) 21:53, 13 December 2015 (UTC)

Incremental JSON dumps

While JSON dumps are the recommended database download format, there seems to be no incremental dump available for this format. It should be straightforward to provide a incremental dump based on the modified date of items. This would lead to a lot of saving on bandwidth and server load to keep a current database snapshot.
Pravincar (talk) 06:47, 13 December 2015 (UTC)

@Hoo man: What do you think? --Lydia Pintscher (WMDE) (talk) 08:40, 14 December 2015 (UTC)

Splitting two-person article

I have created an article in English Wikipedia for w:Adela of Pfalzel by moving and editing w:Adela and Irmina, as w:Irmina of Oeren already existed. There is a German Wipedia de:Adela von Pfalzel and several other items in other wikipedias. I know very little about Wikidata, so could someone please sort this out? Perhaps it's a case of eliminating the link from English wiki to the two-person record which already exists (there's a Spanish article for the pair of them), and then linking Adela's article to the existing Wikidata record for her? Maybe moving the article wasn't the right approach: I was going to "Split" it, but then found the pre-existing article for one half of the split.

I don't watchlist this page, so if anyone wants to contact me please drop a note on my talk page on English Wikipedia. Thanks. PamD (talk) 21:59, 13 December 2015 (UTC)

@PamD: I fixed the links to enwiki in your comment above and updated the links on the Wikidata items: Irmina of Oeren (Q289060), Adela (Q266756), Adela and Irmina (Q19118084). In some of the other WP, they are mother and daughter. --- Jura 22:18, 13 December 2015 (UTC)
...and I fixed the :de: link. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:17, 14 December 2015 (UTC)

TODAY/NOW function for "Retrieved" property

Not sure whether it is readily possible, however, it would be really convenient if there was the ability to type "@today" or "@now" when using retrieved (P813) rather than having to work out and type the date into place. (Yes I know that it is a minor issue, though when doing multiples every keystroke saves time and improves accuracy). Thanks for any consideration.  — billinghurst sDrewth 03:24, 14 December 2015 (UTC)

This is phabricator:T76859. --Lydia Pintscher (WMDE) (talk) 08:39, 14 December 2015 (UTC)

Questions about 'image' property (P18)

A couple of periodical questions at ru-wiki about 'image' Property:P18 due to its data usage in infoboxes.

  1. One infobox — one image (including collages which are physically still a single image file).
  2. Template:Infobox person may be illustrated by something giving any kind of idea of that person historical representation but not by something connected with that person. Say Андрей Рублёв (Andrei Rublev) is OK for image (P18) like this but not OK to illustrate him with his famous work.

Both rules are long standing local consensuses and currently are not negotiable.
Yet I thought that the first might be a consensus here as well, as image (singular form) suggests so the ability to add to image any amount of images is simply an interface glitch.
Similarly "image of relevant illustration of the subject" (from P18 definition) might suggest that the second ru-wiki rule is close to what the original P18 purpose was.
So instead of keep pre-filtering WD output for infoboxes at ru-wiki, as it done right now, something could be decided or clarified at the very "place of origin". --NeoLexx (talk) 12:02, 14 December 2015 (UTC)

I am not sure, but I think you want to just have one image in an item's P18 property, which is what I always understood to be the basic general idea of P18. Recently (like within 6 months or so) it was discussed here with no consensus. --Jane023 (talk) 12:07, 14 December 2015 (UTC)
If there are multiple images, you can set one to preferred and select that. For persons, ideally we would have images for several different years (using qualifier "point in time").
For people, personally, I only add portraits of the person, not their work. Works would generally have their own items which can get images as well.
Image selections for locations are a bit tricky. .. --- Jura 12:22, 14 December 2015 (UTC)

Instance of for texts

I've found this thread, but its conclusions at this moment are widely ignored: can we make a final decision? I've opened a new thread here. --Epìdosis 12:34, 14 December 2015 (UTC)

Wikidata weekly summary #188

Disambiguation page spam

If the article in one single language version is a Wikimedia disambiguation page (Q4167410) then the descriptions in all languages get filled with "disambiguation page" although they have full articles (e.g. Islamic religious police (Q976863)). Additionally there is no easy way to remove this if an article is not anymore an disambiguation page (e.g. Q332918). I've started to fix those examples, but I fail because it's not possible to to show all languages at once.--Kopiersperre (talk) 21:42, 13 December 2015 (UTC)

Use Labellister gadget (can be switched on in preferences) or datatrainer (add mw.loader.load( '//' ); to your common.js and then choose More>Empty on the item. Mbch331 (talk) 21:54, 13 December 2015 (UTC)
@Mbch331: Thanks. By the way: What's the difference between importScript and mw.loader.load?--Kopiersperre (talk) 23:21, 13 December 2015 (UTC)
The importScript method imports and executesthe script immediately, the other after page load. The importScript method is deprecated and doesn't always work anymore. Mbch331 (talk) 05:09, 14 December 2015 (UTC)
@Kopiersperre: little advice in addition: before you change such an item with its descriptions, and before you remove the claim "P31=disambiguation" it is useful to check if there is a reference given for the claim. Here there was huwiki stated ("imported from"). When you then have a look at the linked page hu:Vallási rendőrség (egyértelműsítő lap) you see that it is in fact a disambiguation page. This needs to be moved to an other (new, or existing but different) item.
i fixed it already in this case, also for the mswiki page which was wrongly linked too; see now Q21714378 and Polis agama Islam (Q21714374); and i also removed all disambigation descriptions from the "Absperrorgan" item with Labellister. Holger1959 (talk) 02:11, 14 December 2015 (UTC)
@Holger1959: Thanks.--Kopiersperre (talk) 08:47, 14 December 2015 (UTC)

Tools to add external references to existing statements

If I wanted to add external references including qualifiers as described in Help:Sources#Databases to already existing statements, which tools could I use then? Manual editing via the web front end is not an option here, since I would like to add a 5-figure number of references… —MisterSynergy (talk) 08:44, 14 December 2015 (UTC)

QuickStatements (Q20084080) could help if you know the item number and the statement you want to update. (Just add the statement the way it should be including the source. It won't duplicate the statement, but will add the source.) Mbch331 (talk) 11:34, 14 December 2015 (UTC)
I always thought that QuickStatements cannot add “qualifiers” to statements, at least once I tried this I did not succeed. I only managed to add import references like P143:Qxxxx (with Qxxxx being a Wikipedia item) as a reference, but in this case I’d like to add references composed of 3 to 5 parts (stated in (P248), a database/authority control property, retrieved (P813), and optionally language of work or name (P407) and title (P1476)). How would I have to format the QuickStatements-input in order to get that running? —MisterSynergy (talk) 12:00, 14 December 2015 (UTC)

Here's how to do it. Qxxx = Item you're changing, Qyyy is the value for P143, Sccc is the property id of the database/authority control property. Strings need the quotes (otherwise it won't work), dates need the format I used (/11 means precision = day) and monolingual tekst needs the languagecode as used on Wikidata before the text (with the text in quotes).

Qxxx	P143	Qyyyy	S248	Qzzzz
Qxxx	P143	Qyyyy	Sccc	"tttttt"
Qxxx	P143	Qyyyy	S183	+0000000yyyy-mm-ddT00:00:00Z/11
Qxxx	P143	Qyyyy	S407	Qfffff
Qxxx	P143	Qyyyy	S1476	en:"vvvvvvv"

Best is to prepare it in Excel or another spreadsheet program as there need to be tabs as seperators (not spaces). Mbch331 (talk) 12:17, 14 December 2015 (UTC)

Thank you! I’m somewhat familiar with QuickStatements, I will try this later on this day or tomorrow. I am not exactly sure what I tried last time, I only remember that QuickStatements added many independent references with one statement each instead of one reference with grouped statements. I will report here whether it works. —MisterSynergy (talk) 12:25, 14 December 2015 (UTC)
  • There are a few bots that do that already.
Some new external validation function of special:constraint report might be doing the same. Where can we read more about it? --- Jura 12:35, 14 December 2015 (UTC)
Yes I saw some bots doing that, but I’d prefer to do that by myself if possible. After an initial run where I want to equip all affected existing statements there will be a constant need to add a couple of these to new cases. It wouldn’t be good to depend on bot operators for that reason. Special:Constraint Report is very interesting (thanks!), but I don’t see how it could help here. Regards, MisterSynergy (talk) 12:50, 14 December 2015 (UTC)

Okay, I made a test using QuickStatement and it did not work out properly. QuickStatements input was:

Q1351518	P21	Q6581097	S248	Q21008628
Q1351518	P21	Q6581097	S2091	"12329"
Q1351518	P21	Q6581097	S183	+00000002015-12-14T20:59:05Z/11
Q1351518	P21	Q6581097	S407	Q1860
Q1351518	P21	Q6581097	S1476	en:"Eric Murray"

Result: here (look at the first statement about gender/sex; meanwhile reverted by myself). The affected statement had three separate references with only one statement each, which is not what I wanted to have. Any suggestions for improvements? @Magnus Manske: does QuickStatements actually support the addition of complex references that are composed of multiple statements? If so, how should the input look like? If not, could we have this feature? —MisterSynergy (talk) 23:16, 14 December 2015 (UTC)

One more idea to use QuickStatements that did not work:

Q1351518	P21	Q6581097	S248	Q21008628	S2091	"12329"	S183	+00000002015-12-14T20:59:05Z/11	S407	Q1860	S1476	en:"Eric Murray"

This single-line approach follows QuickStatements instructions as far as I understand. The result was basically the same as with the separate lines as tried last night [8]. Would it help to use some kind of brackets to group the reference Sxxx’s into one reference instead of creating multiple refs? —MisterSynergy (talk) 17:32, 15 December 2015 (UTC)


How to make a wordlist german-paschtu based on items having both languages? I have only the article list. Output should only be like: Baum - ونه (if possible description, but not so important).

Many thanks for help, Conny (talk) 09:39, 14 December 2015 (UTC).

No idea. I know how to get the list of items rather easy (autolist: link[dewiki] and link[pswiki]), but how to get the labels of those item, I don't know. Mbch331 (talk) 11:44, 14 December 2015 (UTC)
here (save and open in decent text editor for proper UTF8). Done via database query. --Magnus Manske (talk) 20:27, 14 December 2015 (UTC)
Is there a way to do this in SPARQL? My attempts either timed-out or didn't work otherwise.
This way we could include P31/P279 in the output and possibly directly exclude people and internal stuff. --- Jura 11:41, 15 December 2015 (UTC)
Vielen Lieben Dank Magnus! Regards, Conny (talk) 17:15, 27 December 2015 (UTC).

Cross-wiki idea: Crowdsource photos

I suggested this at the English Village Pump Idea Lab and it was suggested I bring it here as this would be a project-wide project not limited to the English Wikipedia.

I would like to have a feature that allows editors to request photos to use in articles, and for other editors to register as willing and available to take photos in their area. Editors can take a look at local requests and see if they have any photos or go out and take them. They have this feature on Find a Grave and it's incredible... I requested a photo of a relative's grave on the other side of the country and within a month it had been uploaded by someone nearby. You can see the Village Pump discussion for suggestions, and it was noted that current options like this and this are poorly integrated across Wikipedia and underutilized. (The latter is rather inefficient as it just puts a template on the talk page that says "It is requested that a photograph be included in this article to improve its quality.")

This would be more of a geographic-specific resource rather than BLP ("I need photos of a celebrity!!!"). For example (and this is just an example of how it could be used), I just worked on the article on Auregnais, the extinct dialect of Norman French from the island of Alderney. There's nearly no record of this language and it's now only visible in certain signs on Alderney. No photos of street signs from Alderney are available on Flickr or any other free source. I would love to be able to request a photo that would alert a Wikipedian in the Channel Islands/Normandy of what I'm looking for. And I'd be perfectly willing to take photos of anything people wanted in my area, and before I travelled, I'd take a look to see what photos were requested in that area. I envision this feature being requestable on articles and also send alerts to people who signed up for it, with an additional centralized project on Commons that lets you browse requests. It would be really cool to have a map with pins showing "Photo requested." A little icon in the top corner of the article could indicate current requests, so random people who visited the article in any language, such as Aurigny (French article on Alderney) would also see it, and if interested learn how to upload a photo to Commons and register as a photo contributor. I think there are a lot of Wikipedians throughout the world who would be willing to go out and take a photo someone on the other side of the world requested. What do you think? Wikimandia (talk) 00:39, 9 December 2015 (UTC)

@Wikimandia:, I think it's an extraordinary idea but... I also think it does not fit Wikidata's scope right now and it should be developed through Commons or an external app. I'd suggest taking it to ¿IdeaLab?. Cheers. Strakhov (talk) 20:19, 12 December 2015 (UTC)
@Wikimandia: I agree. Wikidata does good work in certain aspects of geography, but you won't find many in WD who care about pictures. You need to seek help in Commons, and in the various Wikiprojects about photography, geography, and illustration. When there's at least a network of links between Commons and the hundreds (thousands?) of relevant Wikipedia categories, guidelines and Wikiprojects, WD people might be persuaded to help in regularizing the network. Or the picture people who set up the links can learn WD themselves. A new mobile app for Commons could also include features to guide smartphone users to desired sites. I think this prospect is more distant, however, given the slow and fitful development of Wikimedia Foundation's various mobile apps. Jim.henderson (talk) 15:07, 16 December 2015 (UTC)

Merge certain Medal of Honor categories

I found several entities that I recommend be merged. merging . Although some sites use one or the other, each pair of categories are the same and I think it would be better if they were merged.

  • Merge Q7009305 (Category:Army Medal of Honor recipients) and Q13299617 (Category:United States Army Medal of Honor recipients)
  • Merge Q13270109 (Category:United States Marine Corps Medal of Honor recipients) and Q7009311 (Category:Marine Corps Medal of Honor recipients]]
  • Merge Q7009297 (Category:Air Force Medal of Honor recipients) and Q13299612 (Category:United States Air Force Medal of Honor recipients)

I would recommend keeping the United States X Medal of Honor recipients format. If there is a better place to submit this please let me know. Reguyla (talk) 22:23, 14 December 2015 (UTC)

Help:Merge mentions the gadget you can use to merge them. --- Jura 22:54, 14 December 2015 (UTC)
Thanks I'll read up on that. Is this an unambiguous case that I can just do or does something like this need to be discussed first? Reguyla (talk) 19:41, 15 December 2015 (UTC)
Separate items for the same thing are (too) frequent. The ones I checked seem all US only, so I'd go ahead. --- Jura 13:32, 16 December 2015 (UTC)

Q17272400 and Q7366 (song) delete/merge

In Şarkı (Q17272400), the Swedish 'Musikverk' is associated with 'Şarkı (edebiyat)', a form in Turkish classical music/literature. 'Musikverk' seems to mean song, therefore it should be associated with song (Q7366). I could not change Q17272400. Please help. I am not aware of any article in a non-Turkish wikipedia that could be associated with 'Şarkı (edebiyat)'. Bulgu (talk) 21:55, 15 December 2015 (UTC)

Can't be done because Q7366 is already associated with sv:Sång. For now, I've removed sv:Musikverk from Q17272400 . Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:37, 15 December 2015 (UTC)

History and timelines via WikiData?

Hi there - I read through several pages, including FAQs and Projects, but couldn't find an answer, so am hoping you could help me. I'd like to translate w:History of mathematics into a WikiData timeline. What template or (not sure what they're called) should I copy and paste to get started? - Thanks; LeoRomero (talk) 18:53, 16 December 2015 (UTC)

Copy references

Is copy references broken again? It is not appear in pages. Xaris333 (talk) 19:59, 16 December 2015 (UTC)

It has been disabled. Matěj Suchánek (talk) 20:25, 16 December 2015 (UTC)

What country to use

on my talkpage I got a not that country of origin (P495) would be better for sport clubs then country (P17) and actually I think that is correct. But when I do a quick survey on what is filled in at the moment, I see that country (P17) is actually used more often then country of origin (P495). The comparison is not completely fair, as I made 25000 edits in the last 24 days on filling country (P17). What would be the best way to go from here?

When there is a clear preference for one option, I can maybe create a python script to make that happen. Edoderoo (talk) 20:01, 16 December 2015 (UTC)

There is also "country for sports". --- Jura 20:13, 16 December 2015 (UTC)
I see, you mean country for sport (P1532) that is meant for people and maybe national teams. Inter Milan (Q631) is not representing a country, but is clearly an Italian club. country (P17) or country of origin (P495) would be Italy (Q38), and location (P276) would be Milan (Q490). I think that country for sport (P1532) would stay blank on a club, and on Italy national football team (Q676899) it would be (actually is ;-) filled in. Edoderoo (talk) 20:33, 16 December 2015 (UTC)
I think the best would be to define for what exact purpose we want to relate sports clubs to “their” countries. This is not entirely clear to me. Depending on the outcome, location (P276) or a subproperty thereof might also be an option. —MisterSynergy (talk) 20:20, 16 December 2015 (UTC)

make a chatting website please

I think to make a chatting website is best it should work like facebook twitter etc DR MAKUYAHUNDI (talk) 04:11, 15 December 2015 (UTC)

@DR MAKUYAHUNDI: Hello. Unfortunately we are unable to help you with that and this page is only for discussion of Wikidata matters. If you meant live chat, we have Wikidata:IRC, but not much more, because we are not a social network.--Jasper Deng (talk) 04:13, 15 December 2015 (UTC) --Atlasowa (talk) 21:20, 17 December 2015 (UTC)

Categories by country

I want to say that Category:German films (Q7141777) is part of Germany (Q183). Shall I use part of (P361) for that? Queryzo (talk) 13:57, 17 December 2015 (UTC)

Categories generally only use category's main topic (P301) and/or category combines topics (P971). --- Jura 14:45, 17 December 2015 (UTC)
That way? Queryzo (talk) 14:51, 17 December 2015 (UTC)
Yeah. Maybe "country of origin" could be there too. Other sample: Q6833944. --- Jura 14:58, 17 December 2015 (UTC)

What IdeaLab campaigns do you want to see?

Hey folks. I’m seeking your help to decide on topics for new IdeaLab campaigns that could be run starting next year. These campaigns are designed to attract proposals from Wikimedia project contributors that address a broad gap or area of need in Wikimedia projects.

Here’s how to participate:

With thanks,

I JethroBT (WMF), Community Resources, Wikimedia Foundation. 23:02, 17 December 2015 (UTC)

Quantifiers for awards

Hi! Cate Blanchett (Q80966) received (award received (P166)) the Academy Award for Best Actress (Q103618) at the (statement is subject of (P805)) 86th Academy Awards (Q13419956). But which quantifier do I use to state that she acted in Blue Jasmine (Q2907178)? And which quantifier do I use in Blue Jasmine (Q2907178) to state that Cate Blanchett (Q80966) was the actress? --Jobu0101 (talk) 23:02, 11 December 2015 (UTC)

Same question for nominated for (P1411). See American Hustle (Q9013673). There we have nominations for best actor and actress and I want to enter Christian Bale (Q45772) and Amy Adams (Q481832). It would be smart to take the same quantifiers. --Jobu0101 (talk) 23:16, 11 December 2015 (UTC)
In the case of Cate Blanchett use winner (P1346). In the other case there is no property. Mbch331 (talk) 06:18, 12 December 2015 (UTC)
We should have quantifiers for all cases. Are you sure that we don't have them already? If you are we should create them as fast as possible. What I don't like about winner (P1346) is that you can't use it in the nominated case even though it is exactely the same situation. It would be reasonable to use the same quantifiers for award received (P166) and nominated for (P1411). --Jobu0101 (talk) 08:39, 12 December 2015 (UTC)

I'm working on tools which generate en:87th Academy Awards#Awards automatically using the data from Wikidata. So it is of immense importance for me to have the possibility to store that kind of data in Wikidata. See also here for a German list which is already generated automatically using the data from Wikidata. --Jobu0101 (talk) 08:48, 12 December 2015 (UTC)

You can use for work (P1686), I added in Cate Blanchett (Q80966) on property award received (P166) and nominated for (P1411). For the movie I don't think is a good idea to add "best actor/acress" because the award is won by the person not by the opera. --ValterVB (talk) 09:55, 12 December 2015 (UTC)
@ValterVB: Thank you very much. I already thought the same thing about your second point. It is not completely correct to say that the movie won the award for best actress. Do you think we should still mention in Blue Jasmine (Q2907178) that Cate Blanchett (Q80966) won an award with that movie with a different property? Doing so would facilitate award counting for movies but we would have redundant information on Wikidata. By the way: Redundant information seems to be generally accepted in Wikidata. There are many properties like follows (P155) and followed by (P156) or P7 (P7) and P9 (P9) which come in pairs. So I think the correct way would be to have some (may new) property for that case. --Jobu0101 (talk) 11:37, 12 December 2015 (UTC)

ValterVB Josve05a LydiaPintscher Ermanon Cbrown1023 Discoveranjali Mushroom Queryzo Danrok Rogi Escudero Mbch331 Jura Jobu0101 Jklamo Jon Harald Søby putnik ohmyerica AmaryllisGardener FShbib Andreasmperu Li Song Tiot Udi Oron ~ אודי אורון CennoxX Harshrathod50 U+1F350 Bodhisattwa (talk) Shisma Wolverène Tris T7 TT meEsteban16 Antoine2711 Hrk6626 TheFireBender V!v£ l@ Rosière /Murmurer…/

  Notified participants of WikiProject Movies: I'd like to solve this problem and need your help for it. --Jobu0101 (talk) 10:52, 17 December 2015 (UTC)

What solution do you suggest?
Maybe you want to get the views of @GerardM: who is "WikiProject Awards" .. --- Jura 11:00, 17 December 2015 (UTC)
An award is not an award ceremony. An award may be handed out an an award ceremony but that is NOT a given. If you want to connect an award ceremony, make it a qualifier for the Award Received property. Thanks, GerardM (talk) 11:21, 17 December 2015 (UTC)
@GerardM: Thank you for your answer. I don't understand what you mean. It's correct, an award is not an award ceremony. But I didn't use it that way. We have statements like Cate Blanchett (Q80966) received (award received (P166)) the Academy Award for Best Actress (Q103618) at the (statement is subject of (P805)) 86th Academy Awards (Q13419956). Here statement is subject of (P805) is used as quantifier. I think that's no problem. We don't state here that 86th Academy Awards (Q13419956) is an award. Do you agree?
But here are my questions:
  1. How do we mention in Blue Jasmine (Q2907178) that Cate Blanchett (Q80966) won the Academy Award for Best Actress (Q103618) for that movie?
  2. How do we mention in American Hustle (Q9013673) that Amy Adams (Q481832) was nominated (nominated for (P1411)) for the Academy Award for Best Actress (Q103618) for that movie?
--Jobu0101 (talk) 00:53, 18 December 2015 (UTC)
I think "subject of" is not appropriate. It is not what it is. I would not add it on the movie for it is implied. Thanks, GerardM (talk) 05:21, 18 December 2015 (UTC)
@GerardM: Is there already a property you would use as quantifier instead of statement is subject of (P805)? And about my two quesitons you say you wouldn't mention it at all? At this point I don't think so. I think within a movie item we should talk about all awards related to that movie. Other webpages do so too like IMDb and the Wikipedia articles of those movie. See enwiki (they also created a separate article for awards and nominations) and dewiki. --Jobu0101 (talk) 07:26, 18 December 2015 (UTC)
What a Wikipedia does is not reallyu relevant. When you state that an award was won by an actor or whatever, by inference the fact is known for the movie. For me it is redundant. Thanks, GerardM (talk) 10:17, 18 December 2015 (UTC)
@GerardM: I agree (like I did before), it is a redundant information. But as I pointed there already out Wikidata is full of redundant information and it seems that's the way people want Wikidata to be. For the sake of completeness and transparency it's convenient to have all awards related to a movie mentioned in the movie object. If we were strict like you proposed we also would have to delete Academy Award for Best Picture (Q102427) in movie items because this award is won not by the movie itself but by the producers. And what about statement is subject of (P805)? Do you have a substitude or do we need a new property for it? --Jobu0101 (talk) 10:30, 18 December 2015 (UTC)

The last point brings me to a further question: Which quantifier do we use to state in Nebraska (Q3337646) that Albert Berger (Q15782026) and Ron Yerxa (Q15842992) are the persons nominated for the Academy Award for Best Picture (Q102427)? --Jobu0101 (talk) 12:34, 18 December 2015 (UTC)

  • I had a look at Special:WhatLinksHere/Q18615010 to find a qualifier that could be used to state that "Film ABC" received the "DEF best acting performance award" for the performance by (name of actor). criterion used (P1013) seems to be the most likely. --- Jura 13:33, 18 December 2015 (UTC). Obviously, (name of actor) received the "DEF best acting performance award" for work "Film ABC" is the preferred way of stating this. --- Jura 13:38, 18 December 2015 (UTC)
@Jura1: Thank you for your help. I just added to American Hustle (Q9013673) the actress Amy Adams (Q481832) with criterion used (P1013). Do you think that's okay that way? --Jobu0101 (talk) 14:08, 18 December 2015 (UTC)

Wanted: Tool to generate Wikipedia article, from Wikidata

I seem to recall a tool to generate the code for a new Wikipedia article, from a Wikidata item, but my Google-foo is failing me. Where is it, please? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:09, 12 December 2015 (UTC)

Only things I can think of are Reasonator (Q20155952) and Extension:ArticlePlaceholder (Q21676433). Mbch331 (talk) 18:51, 12 December 2015 (UTC)
Thank you, but it was something else. Maybe on toolserver? @Magnus Manske: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:38, 12 December 2015 (UTC)
See ru:HD 207129 for example. It uses "tool" ru:Шаблон:Типовая статья о звезде. Another example: ru:Отто, Пауль. — Ivan A. Krestinin (talk) 20:41, 12 December 2015 (UTC)
Thank you; though interesting, that's not the one I was thinking of either. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:02, 12 December 2015 (UTC)
Maybe this one? Sjoerd de Bruin (talk) 08:21, 13 December 2015 (UTC)
Yes, that's the one. I'd forgotten that it was for biographies only (but still very useful). Thank you. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:19, 14 December 2015 (UTC)

Yes for those with geo-referencing knowledge, it would be great to get a PrepBio thingy going for buildings as well. --Jane023 (talk) 12:02, 14 December 2015 (UTC)

There is also AutoDesc, which can generate MediaWiki syntax. --Magnus Manske (talk) 19:03, 16 December 2015 (UTC)
@Magnus Manske:Thanks, [9] has malformed dates, and not {{Authority control}} Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:21, 18 December 2015 (UTC)

Cheyenne Mountain

There are items Cheyenne Mountain (Q1071445) and Q5094585, both described as instance of (P31): mountain (Q8502). Most of the wp articles in the first item define it as a mountain, but mention also the complex. Shouldn't there be only one item for the complex and one item for the mountain like in en and fr Wikipedas? How to cope with this right? --Venca24 (talk) 10:30, 17 December 2015 (UTC)

Given that most labels, descriptions and statements of both are about the mountain, you might want to create a new item for the bunker and move relevant sitelinks there. Once done, merge the two initial items and fix conflicting labels/descriptions. --- Jura 10:49, 17 December 2015 (UTC)
Now there is Cheyenne Mountain (Q1071445) for the mountain and Cheyenne Mountain Complex (Q21767547) for the military base. --Venca24 (talk) 08:16, 18 December 2015 (UTC)
In the meantime, I discovered another item: Cheyenne Mountain Air Force Station (Q5094587). --- Jura 13:40, 18 December 2015 (UTC)

Is there a page explaining how organisations can import their data into Wikidata?

Hi all

Is there a page explaining the different ways organisations can import their data into Wikidata? I know there are few options e.g Mix n' Match + QuickStatements. If there isn't a page I'd be very happy to work with people to create one.


John Cummings (talk) 16:08, 17 December 2015 (UTC)

WD:Data donation was started a big ago. --Izno (talk) 17:53, 17 December 2015 (UTC)
WD:Tools may be interesting also. --Izno (talk) 17:54, 17 December 2015 (UTC)
Perfect, thanks very much, I do some work on them. John Cummings (talk) 09:40, 18 December 2015 (UTC)
A more general one is here. --- Jura 09:55, 18 December 2015 (UTC)
Yes it'd be great if you could help improve the data donations page. --Lydia Pintscher (WMDE) (talk) 10:17, 18 December 2015 (UTC)

"set in period" or just use "depicts"

Do we really need set in period (P2408) when we could just claim that the creative work depicts (P180) = Edwardian era (Q185852)? Possibly narrative location (P840) could also be merged in to depicts (P180). In which case depicts (P180) would have to double-up as "describes" to cover written as well as visual works. That way we could search for representations of a given thing in one property only, instead of three or more. Danrok (talk) 01:58, 18 December 2015 (UTC)

depicts (P180) is intended only for the persons, scenery or objects that are represented in a visual work (Mona Lisa (Q12418) depicts (P180) Mona Lisa). So a claim depicts (P180) => Edwardian era (Q185852) is wrong as Edwardian era (Q185852) is an abstract concept. --Casper Tinan (talk) 11:14, 18 December 2015 (UTC)
You're just stating what the existing situation is. Consider that there may be another way to handle this, as is the case on IMDb, they use "plot keywords" [10]. Danrok (talk) 15:21, 18 December 2015 (UTC)
To represent has many meanings in both English and French : to depict a person or a scenery, to mimic something, to serve as a sign or symbol of something, to act on behalf of an organisation or another person, etc. In other languages, these concepts can be designated by very different words and people would not consider them as related. Merging the different properties corresponding to these different meanings on the pretext that the same word is used for them in English (or in French by the way) just goes against Wikidata's principles. Casper Tinan (talk) 16:15, 18 December 2015 (UTC)

legal form (P1454) just another "type of" property?

Surely, the legal form (P1454) is not needed. We can already simply state that an item is an instance of (P31) public limited company (Q5225895). Does anyone know why we have this legal form (P1454) property? Danrok (talk) 03:12, 18 December 2015 (UTC)

Because, when you check the "what links here"/ctrl-shift-J, you see that P31 of those is filled with other info already. You are right that there are many ways to fill in the data, and for some reason it has been choosen to use these fields. instance of (P31) is so general, that legal form (P1454) is more dedicated to it's function, usually used on companies and alike structures. Edoderoo (talk) 07:10, 18 December 2015 (UTC)
@Edoderoo: Within the domain of businesses instance of (P31) was previously automatically populated in a simple way, e.g. instance of (P31) = company (Q783794) based on Wikipedia categories. That being better than nothing until such time as the data could be improved by hand, or a using a smarter method. Danrok (talk) 15:03, 18 December 2015 (UTC)
legal form (P1454) is a sub-property of instance of (P31). Specialized sub-properties are useful to clarify the relationship and enable easier querying. --Srittau (talk) 10:06, 18 December 2015 (UTC)
Besides, the legal form of a company may vary over time whereas instance of (P31) is intended for statements that remain (almost) always true. Casper Tinan (talk) 11:19, 18 December 2015 (UTC)
@Casper Tinan: "instance of (P31) is intended for statements that remain (almost) always true." That simply isn't true. Just about everything changes over time, yes? We have point in time (P585), start time (P580) and end time (P582) to solve this. As is shown in the property documentation here [11]. Danrok (talk) 15:09, 18 December 2015 (UTC)

Non transcribed titles

Hi, I want to add title (P1476) to History of the Arkanar Massacre (Q4204793), but the films title is russian, so there is the original title "Трудно быть богом" and the transcribed title "Trudno byt bogom". How do I add it properly? Queryzo (talk) 12:44, 17 December 2015 (UTC)

There is ISO 9:1995 (P2183) for some transliterations of Cyrillic. A more generic property proposal is asleep at Wikidata:Property_proposal/Generic#Transliteration_or_transcription --- Jura 13:09, 17 December 2015 (UTC)
So is this correct and I have to wait for an qualifier called transcription? Queryzo (talk) 13:59, 17 December 2015 (UTC)
Looks good. If you can do ISO, you can use the qualifier P2183. --- Jura 14:03, 17 December 2015 (UTC)
That way? Queryzo (talk) 14:54, 17 December 2015 (UTC)
@Queryzo:: please see my comment at Wikidata:Property_proposal/Generic#Transliteration_or_transcription. --- Jura 13:50, 19 December 2015 (UTC)

Wikimania 2016: call for proposals is open!

Dear Wikidata users,
the call for proposals for Wikimania 2016 is open! All the members of the Wikimedia projects, researchers and observers are invited to propose a critical issue to be included in the programme of the conference, which will be held in Italy, in Esino Lario, from June 21 to 28.
Through this call we only accept what we call critical issues, i.e. proposals aiming at presenting problems, possible solutions and critical analysis about Wikimedia projects and activities in 18 minutes. These proposals do not need to target newbies, and they can assume attendees to already have a background knowledge on a topic (community, tech, outreach, policies...).
To submit a presentation, please refer to the Submissions page on the Wikimania 2016 website. Deadline for submitting proposals is 7th January 2016 and the selection of these proposals will be through a blind peer-reviewed process. Looking forward to your proposals. --Yiyi .... (talk!) 09:07, 19 December 2015 (UTC)

How can I count how many statements are on Wikidata?

Dear all, I have a question: I usually do a lot of presentations around Italy about Wikidata, but there is one statistic I never get my grasp on, which is how many statements are on Wikidata? Is there a tool that can give me a reasonable approximation that may answer to my question? Thanks. Sannita - not just another sysop 14:09, 18 December 2015 (UTC)

“We reached 70M statements in September.” (Wikidata:Third Birthday#Numbers) No idea how they measured… —MisterSynergy (talk) 16:09, 18 December 2015 (UTC)
@Sannita, MisterSynergy: See -- LaddΩ chat ;) 21:26, 18 December 2015 (UTC)
@Laddo: That's exactly what I was looking for, thank you very much! :) --Sannita - not just another sysop 10:29, 20 December 2015 (UTC)

Get involved in Wikipedia 15!

This is a message from the Wikimedia Foundation. Translations are available.

As many of you know, January 15 is Wikipedia’s 15th Birthday!

People around the world are getting involved in the celebration and have started adding their events on Meta Page. While we are celebrating Wikipedia's birthday, we hope that all projects and affiliates will be able to utilize this celebration to raise awareness of our community's efforts.

Haven’t started planning? Don’t worry, there’s lots of ways to get involved. Here are some ideas:

Everything is linked on the Wikipedia 15 Meta page. You’ll find a set of ten data visualization works that you can show at your events, and a list of all the Wikipedia 15 logos that community members have already designed.

If you have any questions, please contact Zachary McCune or Joe Sutherland.

Thanks and Happy nearly Wikipedia 15!
-The Wikimedia Foundation Communications team

Posted by the MediaWiki message delivery, 20:58, 18 December 2015 (UTC) • Please help translate to your languageHelp

Any ideas for something nice we as Wikidata could do for our big sister? :D --Lydia Pintscher (WMDE) (talk) 11:46, 20 December 2015 (UTC)

Relation of a government agency to geographic entity

I've recently created Q21774060 (City of Seattle Landmarks Preservation Board). Right now I have located in the administrative territorial entity => Seattle, but that seems unsatisfactory. It would seem that there ought to be a specific way (maybe a qualifier?) to relate a specific government agency to the government of which it is an agency. Perhaps we need to connect them via a distinct entity for the Government of Seattle distinct from the entity for the city, even if there is no wikipedia article corresponding to that intermediate entity.

I'm very open to suggestions. Still trying to get my head around this. Also, is there anywhere that we present normal schemas for data about members of particular classes (in this case government agency)? How do people go about learning this? I see little or nothing in the community portal to get me past a very basic level. - Jmabel (talk) 23:37, 18 December 2015 (UTC)

A board is part of an organisation or it is an organisation. This board is located in Seattle and why would that be an issue? When it is part of something else fine. When there is another logical distribution of the land that makes sense upto the level of a state or a country than fine. GerardM (talk) 05:58, 19 December 2015 (UTC)
OK, so there is nothing inaccurate about it then for geographic location, but it still fails to express the fact that it is part of the city government. A federal courthouse in Seattle would also be a government organization in Seattle, but would not be part of the city government. And, again, is there anywhere that we present normal schemas for data about members of particular classes (in this case government agency)? How do people go about learning this? - Jmabel (talk) 06:29, 19 December 2015 (UTC)
applies to jurisdiction (P1001)? That's what I use. --Srittau (talk) 06:53, 19 December 2015 (UTC)
Thanks, and good to know. That still doesn't answer my more general question about schemas and where this can be learned, though. - Jmabel (talk) 08:09, 20 December 2015 (UTC)
In the Netherlands polders are not necessarily part of a municipality, they are part of a "water board?". This water board is a law on its own so that is what I use for polders. When a courthouse is in a place, it is still in that place even when it is owned by a courtsystem. Thanks, GerardM (talk) 09:14, 20 December 2015 (UTC)

Credit-linked note, etc.

Currently, Credit-linked note (Q1139416) have 8 different language's wikipedia entry. In English wikipedia, the entry wrote: 'credit-linked notes have been marketed as "minibonds" [...]', and in Chinese&Yue Wikipedia, the entry wrote (rough translation): 'credit-linked notes have been marketed as "minibonds" or "Octavenotes" in Hong Kong and "Structured Note" in Taiwan.' and the Chinese Wikipedia entry have an infobox saying "credit-linked notes" is how the thing being called in Mainland China whereas "minibond" is how it is called in Hong Kong and "structured note" is how the thing being called in Taiwan. As such, it is presented in a way which "credit-linked notes", "minibond", and "structured note" seem to be locale variation in term of how to describe the same thing, which cause an automatic conversion of article title and content, making the article for "structured note" become "minibond" for readers from Hong Kong, and "structured note" for readers from Taiwan. However, note that when a reader type "structured note"('s chinese name) into Chinese wikipedia's search field and click enter, they will be redirected to another article that's talk about structured note (Structured note (Q3566070)), but it seem like both are just describing something nearly the same, and why both article exist is just because their creator does not aware of existence of others. Note that Q3566070 have three language versions wikipedia entry and English wikipedia have the entry too. (As a Note, there are also separate chinese wikipedia entry for the "Morgan Stanley Octave Notes Incident" which interwiki-linked to Yue Wikipedia's "Octave Notes" page, and on the other hand Chinese Wikipedia have a "Octave Notes" disambiguation page that lead to either the event page or the credit-linked note page but that is another topic.)

However, There is an separate article on English&Yue wikipedia known as "minibond" (Minibond (Q16254016), despite the Yue Wikipedia entry were pending to merge with the "credit-linked notes" page. The English version of minibond page claim "It is a misnomer to describe the Minibond as credit-linked note because of the Synthetic Collateralised Debt Obligations hidden in the three-layered structure (as all series of Minibonds issued in 2004 and thereafter are)." and it is probably the reason why it separate from the "credit linked note" page. It used old-fashion bidirectional in-page interwiki link to link to the Chinese Wikipedia page "雷曼兄弟迷你債券事件", i.e. "Lehman Brothers Minibond Incident". However, the "Lehman Brothers Minibond Incident" is bankruptcy of Lehman Brothers (Q3269580) which correspond to English Wikipedia Entry: Bankruptcy of Lehman Brothers.

In short, when you are in Hong Kong, when you browse the Chinese Wikipedia Entry "Minibond", and then click English Wikipedia interwiki link, you will get directed to an English Wikipedia entry "Credit-linked Notes", in which you can read it is same thing as minibond, but when you click into minibond article, it'll tell you minibond is not credit-linked notes. Furthermore, In the minibond article, when you click Chinese Wikipedia interwiki, you will get the Chinese Wikipedia entry "Lehman Brothers Minibond Incident" and if you click back the English Wikipedia entry from this Chinese page you will go back to the minibond English Wikipedia entry. However, If you click from Chinese Wikipedia via let say Japanese Wikipedia to English Wikipedia, then you would go from "Lehman Brothers Minibond Incident" to "Lehman Shock" in Japanese Wikipedia to "Bankruptcy of Lehman Brothers" in English Wikipedia.

How to resolve the series of conflict in the process?

edit: To help understanding,

  Chinese Wikipedia Yue Wikipedia English Wikipedia Japanese Wikipedia Greek Wikipeda
Mainland China Hong Kong Taiwan
Q1139416 Credit-linked note Minibond Structured note Credit-linked note Credit-linked note N/A N/A
Q3566070 Structured note N/A Structured note N/A Structured note
Q16254016 N/A Minibond Minibond N/A N/A
Q3269580 Lehman Brothers Minibond Incident N/A Bankruptcy of Lehman Brothers Lehman Shock N/A

In which Chinese Q3269580 and English Q16254016 established bidirectional old-fashion interwiki link and you can see there is a number of mismatch in this table..

C933103 (talk) 09:30, 19 December 2015 (UTC)

Here are links to the items you mention: Q1139416, Q3566070, Q16254016, Q3269580
If the items at Wikidata weren't virtually empty (other than the bankruptcy one), one might have wanted to add an additional column for the concept of each item at Wikidata.
Once this is defined (and statements added to the items), one would want to link the corresponding articles at Wikipedia. It can be that this matches the current sitelinks, requires creating new items and/or needs moving around sitelinks.
We generally tend to link Wikipedia articles based on their title. So, if it's called "minibond incident", but is about the bankruptcy in general, it would still end up on an item for the incident.
Independently of the sitelinks provided by Wikidata, wikis could use "old style" interwikis to link articles that could be related (probably not a great idea, but it's up to them). --- Jura 14:30, 20 December 2015 (UTC)

Fusion problems

Can someone fusion en:Category:Former mines in Germany (Q8467385) with de:Kategorie:Stillgelegtes Bergwerk in Deutschland ?

Can someone fusiion en:Category:Former mines in the United States (Q8467396) with de:Kategorie:Stillgelegtes Bergwerk in den Vereinigten Staaten ? 16:05, 20 December 2015 (UTC)

  Done. And it wasn't a "problem". Jared Preston (talk) 17:23, 20 December 2015 (UTC)

Meta now has access to the data on Wikidata

Hey folks :)

We've enabled data access for Meta last night. Welcome to Wikidata, Meta! Questions and coordination is happening at Wikidata:Meta-Wiki.

Cheers --Lydia Pintscher (WMDE) (talk) 19:14, 16 December 2015 (UTC)

Maybe a small step, but doesn't that mean it's all done now? Congrats! --- Jura 22:53, 16 December 2015 (UTC)
@Jura1: That's a lot of it, but there is still no functionality for incubator: or s:mul: yet. —Justin (koavf)TCM 09:03, 17 December 2015 (UTC)
Is anything planned for them? Should these two get data access only? --- Jura 09:36, 17 December 2015 (UTC)
@Jura1: There are plans for both. Right now, Wikidata:Incubator is a redlink and there is discussion and a ticket about (I've skipped beta.v as it's a moribund project which I think should be closed and merged into Meta.) How desirable or feasible Wikidata integration is for these two is an open question but there is definitely a conversation going on about them. —Justin (koavf)TCM 09:48, 17 December 2015 (UTC)
Incubator seems to be a temporary stage. Adding sitelinks to it at Wikidata might create more problems than it solves, but it could help if each page there had a way to define which item it relates to. (Maybe through a magic word).
s:mul seems to be for languages that are unlikely to get a separate wiki anytime soon. I suppose it shouldn't be more complicated than the other Wikisources we already have. Add this next week? ;) --- Jura 09:59, 17 December 2015 (UTC)
There's problem with s:mul that it can have several pages for one work (because of multiple language nature). So semantically we want to have several sitelinks from s:mul in one item. --Infovarius (talk) 14:57, 21 December 2015 (UTC)

Wikidata Universe Tree

The Wikidata Universe Tree uses part of (P361)/has part (P527) relations (as well as some subproperties and similar properties), both explicit and implied by classes, to organize everything into this nice little tree structure.

Some odd stuff shows up, some as a result if misuses of has part (P527), some due to ambiguity of item meaning, some due to mistakes in the class structure, some from simple errors.

Questions/comments/complaints/ideas/bug reports are most welcome. --Yair rand (talk) 01:49, 21 December 2015 (UTC)

You don´t fill variables in Sparql-queries, don´t you? How do you use them? Are you asking the queries basicParts, partsFromSuperClass, ... at the start of the code and use the return values for any item that somebody might enter? By the way, I have used a different way to connect to the endpoint "//" while you are using "//". Should I change my url? --Molarus 03:25, 21 December 2015 (UTC)
@Molarus: I'm not sure what you mean by "fill variables". The script does replace the $1, etc with the item ID (or other relevant data) for whatever item is being expanded. Regarding which URL, the user manual mentions both options, but doesn't say which is preferred, so I assume both are equally okay. --Yair rand (talk) 03:39, 21 December 2015 (UTC)
That was what I thought first, but I couldn´t find how you do this. Now I have looked again and I think you do this in function getQuery with replace. By the way, I have thought some time ago about an alternative to wdt:P31/wdt:P279* and part of and found P403 (mouth of the watercourse). I think it should be possible to build lots of nice datatrees with that property, do some math (e.g. distance to the ocean) and ask for a way from one place to another place for a ship. --Molarus 10:25, 21 December 2015 (UTC)

Password Strength RFC


We have started an RFC on meta to increase password requirements for users that have accounts which can edit MediaWiki:Common.js, have access to checkuser or have access to Oversight.

These types of accounts have sensitive access to our sites, and can cause real harm if they fall into malicious hands. Currently the only requirement is the password is at least 1 letter long. We would like to make the minimum be 8 letters (bytes) long and also ban certain really common passwords.

By increasing requirements on passwords for accounts with high levels of access, we hope to make Wikimedia wikis more secure for everyone. Please read the full text of the proposal here, and make your voice heard at the RFC.

Thank you

(On behalf of the WMF security team) BWolff (WMF) (talk) 08:30, 21 December 2015 (UTC)

Delivered using the distribution list

I was for try to the wikidata

Please help for me Thanks – The preceding unsigned comment was added by IkhsanNH (talk • contribs).

With what do you need help? You're question is too general at the moment. Mbch331 (talk) 08:56, 21 December 2015 (UTC)

Weight property

Surely it would make more sense to have at least two properties for mass (P2067), those being "net weight" and "gross weight" since this is the normal way of doing things. Having just one mass (P2067) is both limiting and may lead to muddled up data because it is vague by definition. Danrok (talk) 15:05, 17 December 2015 (UTC)

What, in your mind, is the difference between net and gross for the kind of objects we are describing? --Izno (talk) 17:48, 17 December 2015 (UTC)
Well, I was just looking around and found empty weight (Q6753082) which shows a way to handle this using P794 (P794) empty weight (Q6753082), which is effectively net weight. Danrok (talk) 19:24, 17 December 2015 (UTC)
User:Danrok's understanding of net weight seems to be the opposite of my understanding. Gross weight would be the total mass of a container and its contents, such as the mass of a box of cereal, including the box, while net weight is just the contents (in the case of cereal, only the stuff you actually eat). Jc3s5h (talk) 19:34, 17 December 2015 (UTC)
@Jc3s5h: It can work both ways. The exact meaning/method depending on context, industry standards, etc. There's more to this than cornflakes. For trucks, and cargo containers the term is tare weight (Q1638155) (effectively net weight) also known as unladen weight, then we have others like kerb weight, and max. payload weight. Plus another long list if we were to get in to ships, aircraft and spacecraft. Danrok (talk) 20:28, 17 December 2015 (UTC)
@Danrok: American Heritage Dictionary 3rd ed. defines sense 2 of "tare" as "1. [pronunciation omitted] Abbr. t. The weight of a container or wrapper that is deducted from the gross weight to obtain net weight. 2. A deduction from gross weight made to allow for the weight of a container." [Definition related to chemistry omitted]. If we are to have properties for net and gross weight, they should be defined in accordance with reliable sources such as dictionaries. If you can produce reliable sources that have a contradictory definition, then we should not have such properties because the contradictions would make them meaningless. Jc3s5h (talk) 16:42, 18 December 2015 (UTC)
@Jc3s5h: You are singing to the choir. I am well aware of how tare weight and net weight work in terms of vehicles. I have qualifications\licences to drive trucks, dumpers, excavators, fork lifts, cherry pickers, vehicle mounted cranes, etc. In any case it is written here: net weight. Net weight = "the weight of a vehicle without that of its fuel, cargo, personnel etc". Danrok (talk) 17:57, 18 December 2015 (UTC)
Wictionary is not a reliable source. Jc3s5h (talk) 20:26, 21 December 2015 (UTC)
The word nett is an old English word for clean. That's why it can be used to mean a vehicle which is unladen, i.e. clean and empty. Danrok (talk) 18:12, 18 December 2015 (UTC)

Fusion problems

Can someone fusion de:Kategorie:Christentum in Katar (Q20192598) with en:Category:Christianity in Qatar (Q7481314) ? 18:00, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in Oman (Q6924960) with de:Kategorie:Christentum in Oman ? 18:03, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in Kosovo (Q7481192) with de:Kategorie:Christentum im Kosovo(Q19340345) ? 18:08, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in North Korea (Q7481206) with de:Kategorie:Christentum in Nordkorea (Q18560769) ? -- 18:12, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in South Korea (Q7481207) with de:Kategorie:Christentum in Südkorea ? Mir42326rwe235 (talk) 18:15, 20 December 2015 (UTC)

  Done. The first four by me, the fifth by User:Mbch331. Jared Preston (talk) 18:22, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in Kuwait (Q7481190) with de:Kategorie:Christentum in Kuwait ? Mir42326rwe235 (talk) 18:59, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in Jordan (Q6924819) with de:Kategorie:Christentum in Jordanien ? Mir42326rwe235 (talk) 19:01, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in Bosnia and Herzegovina (Q7481392) with de:Kategorie:Christentum in Bosnien und Herzegowina ? Mir42326rwe235 (talk) 19:04, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in the Republic of Macedonia (Q7481359) with de:Kategorie:Christentum in Mazedonien ? Mir42326rwe235 (talk) 19:07, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in Bahrain (Q7481408) with de:Kategorie:Christentum in Bahrain ? Mir42326rwe235 (talk) 19:12, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in Libya (Q7481304) with de:Kategorie:Christentum in Libyen ? Mir42326rwe235 (talk) 19:12, 20 December 2015 (UTC)

  Done the rest.--Ymblanter (talk) 19:17, 20 December 2015 (UTC)

Can someone fusion en:Category:Christianity in Portugal (Q7481227) with de:Kategorie:Christentum in Portugal ? Mir42326rwe235 (talk) 19:22, 20 December 2015 (UTC)

  Done--Ymblanter (talk) 19:27, 20 December 2015 (UTC)

@Mirw42326rwe235: What exactly is the problem for you, that makes you unable to do it yourself? The standard approach here should be to go to the German category and click "Links hinzufügen" and follow the directions. If that doesn't work, for some reason, you can go to the item in Wikidata (which you must be able to find since you've pasted the Q numbers here) and edit the list of Wikipedia links at the bottom manually. Is this not working for you for some reason? Jon Harald Søby (talk) 20:51, 20 December 2015 (UTC)

It's the way people handle this kind of stuff. Instead of learning them to merge items, they just do the merge and say "done". That doesn't help the user. But this user, that already posted like 50 times here with different IP's, also doesn't seem to give any answers to questions about this. Sjoerd de Bruin (talk) 08:41, 21 December 2015 (UTC)
  • BTW Help for merge is at Help:Merge. It also includes countless lists of merge candidates. Items for categories are probably of lowest priority. Help for sitelinks at Help:Sitelinks. The graphic at Items by status gives an idea of the number of items already merged (or deleted). --- Jura 15:20, 21 December 2015 (UTC)
But, Jura1, all of these items only had one link in German. So then it shouldn't be necessary to do anything but to add an interwiki link the normal way (like you would do with an unconnected page), and the software will do the merging automatically. It's as simple as that, which is why I don't understand why Mirw42326rwe235 is having any problems. Jon Harald Søby (talk) 18:27, 21 December 2015 (UTC)
I'm not aware that it does that or should do that. Do you have a sample for such a merge? This (from above list) seems to be a standard item merge. --- Jura 18:46, 21 December 2015 (UTC)
"Mirw42326rwe235" is just another account in a very very long list of sock puppets blocked on the German Wikipedia. That's why this user can't merge items. I would propose not the encourage this user anymore and just archive these topics on sight. Multichill (talk) 22:15, 21 December 2015 (UTC)

lists of episodes

list of Twin Peaks episodes (Q2257179) ist currently defined as Wikimedia list article (Q13406463), like many others. Wouldn't it be better to have a quantifier like "episodes list"? Queryzo (talk) 15:06, 21 December 2015 (UTC)

No. It is the not a good idea. A list of episodes is just a list. Thanks, GerardM (talk) 15:11, 21 December 2015 (UTC)
Yes. A specialized sub-class sounds pretty useful to me. "Episode list" is also a useful concept to have. --Srittau (talk) 16:24, 21 December 2015 (UTC)
I added of (P642): Twin Peaks (Q2085) as qualifier for is a list of (P360) --ValterVB (talk) 17:11, 21 December 2015 (UTC)
With "series" as qualifier, Reasonator can build a list of episodes: [12]. --- Jura 17:32, 21 December 2015 (UTC)
I have changed of (P642) to part of the series (P179) as Jura suggests. This matches the usage on the episode items themselves, eg Episode 1 (Q5383524), Episode 29 (Q1979506), and therefore will allow Reasonator to build a list. Jheald (talk) 21:13, 21 December 2015 (UTC)

fusion problems

Can someone fusion en:1915 with de:1915 ? 15:09, 21 December 2015 (UTC)

They are already linked together since the third day of Wikidata [13] --Pasleim (talk) 16:09, 21 December 2015 (UTC)
It was a cache issue (I didn't see the interwiki too). I have purged the page. --Voll (talk) 19:05, 21 December 2015 (UTC)

Sourcing circumstances

Is there a "sourcing circumstances" that correspond to "near" for a location to handle: "Mrs. Raiche was a native of Wisconsin, having been born near the town of Beloit. "--Richard Arthur Norton (1958- ) (talk) 16:06, 21 December 2015 (UTC)

You could look for other sources which might give more information. Another possibility would be to claim Rock County (Q500939) as the place of birth, assuming she was born within that county for certain. Danrok (talk) 05:31, 22 December 2015 (UTC)

Wikidata weekly summary #189

Wikidata weekly summary #186

Deleting items when entities have been dissolved? - considered vandalism

@Floscher: The Regierunsbezirke in Saxony have not been dissolved? [14] 18:59, 19 December 2015 (UTC)

Yes, they have. Just like this item will dissolve shortly... [15] --Floscher (talk) 19:02, 19 December 2015 (UTC)
To elaborate a bit on that: These items are completely superfluous, as the same meaning can be conveyed by setting instance of (P31) Regierungsbezirk (Q22721) and located in the administrative territorial entity (P131) Saxony (Q1202). --Floscher (talk) 19:04, 19 December 2015 (UTC)

They are not.

are distinct sets of entities. Like there are

  • counties of the states of the United States (Q47168)
    • counties of New York
    • counties of Ohio
    • etc
  • counties of the United Kingdom (Q1379030)


  • provinces of Italy
  • provinces of Spain

The sets are defined by different laws. You have no clue what you talk about. Go and revert your vandalizm! 19:12, 19 December 2015 (UTC)


There seems to be some lack of knowledge and some ignorance. Why are these items from Germany, below the state level, treated with immediate deletions, without any discussion with the creator(s), while at the same time it is common practice to have items for each of the subclasses of U.S. counties?

administrative territorial entities in Germany named "Regierungsbezirk" (government region/government district)
ID Talk Name preceded by established dissolved Notes People involved in creation/deletion
government region of North Rhine-Westphalia (Q829277) Talk Links (5) - named after city
government region of Prussia (Q1850612) Talk Links (28) - 1815-04-30 „Verordnung über die verbesserte Einrichtung der Provinzialbehörden“, named after city
Q21779166 Talk Links (3) - named after city created by, deleted by User:Mbch331, items unlinked by User:Floscher and User:Cycn
Q21779171 Talk Links (3) Kreishauptmannschaft of Saxony 2008 named after city created by, items unlinked by Floscher and Cycn
Q21779180 Talk Links (4) - named after city created by, deleted by User:Mbch331, items unlinked by Floscher and Cycn
Q21779185 Talk Links (3) 2003 named after city created by, deleted by User:Mbch331, items unlinked by Floscher and Cycn
Q21782522 Talk Links (6) - Bezirk and Regierungsbezirk cover the same territory, but are different entities created by, 01:26, 20 December 2015 (UTC)

I totally agree that these should NOT be deleted. I would call it ill considered, only after continuation I would call it vandalism. Thanks, GerardM (talk) 06:52, 20 December 2015 (UTC)
I've undeleted the items for now, so all items can be properly discussed here. If the consensus here is that there are notable, I won't re-delete them, if the consensus here is that they aren't valid, they will be deleted again. I hope Floscher and Cycn will give their opinions here. I acted because Floscher seemed to have a valid point they were redundant. Mbch331 (talk) 08:04, 20 December 2015 (UTC)

Similar items exist for districts in Germany: rural district of Baden-Württemberg (Q20738676) created by User:Dxq189, district of North Rhine-Westphalia (Q20738811) and district of Thuringia (Q20738945) created by User:Dxq298, district of Saxony-Anhalt (Q17302772) created by User:Tamawashi. That are all socket puppets by the global banned user Tobias Conradi which is also editing anonymously under IP's --Pasleim (talk) 10:40, 20 December 2015 (UTC)

As far i know, especially Gbeckmann is systematically working on German district items. maybe he can make a statement here what is best? Holger1959 (talk) 11:38, 20 December 2015 (UTC)

I do not care all to much about what others say. When a type of district existed, it has its place in Wikidata. When it ends at a given date, that is exactly what you indicate. Please assume good faith and examine things on their merit. Thanks, GerardM (talk) 12:21, 20 December 2015 (UTC)
The type of district that is discussed, already is in Wikidata: Regierungsbezirk (Q22721). The question is if it should be subclassed for every state of Germany. And regarding the dissolution of them, I'd argue that the date should be placed on the instances of Regierungsbezirk (Q22721). --Floscher (talk) 13:17, 20 December 2015 (UTC)
(edit conflict) I don't see an advantage of having separate "Landkreis" items for each state. This is something that can be easily queried using located in the administrative territorial entity (P131). Conceptually they are exactly the same. See also the diagram to the right, which makes no difference (except the naming) for each state. The only difference is that some states have "Regierungsbezirke", while others haven't. --Srittau (talk) 12:23, 20 December 2015 (UTC)
Addendum: I am only talking about current forms of government here. It is possible that historical subdivisions might be conceptually different, even if they use the same name. --Srittau (talk) 12:25, 20 December 2015 (UTC)
I can not really figure out why they were deleted, except the opinion they have been edited by sockpuppets. I assume there is a better reason, not mentioned yet? Edoderoo (talk) 12:30, 20 December 2015 (UTC)

Hello, first of all I want to apologize for my overreaction to delete without discussion. At the moment when I saw these changes in my watchlist, I considered them to be quite clear cases, so I just went ahead and requested deletion. In hindsight, I should have discussed it first. My main argument against such items is lack of notability. Let's look at the notability criteria one after another:

  • N1: The first two items in above table have Wikipedia articles, so they are notable, which is ok with me.
  • N2: The items are no clearly identifiable entities, because „Regierungsbezirke“ are the same everywhere in Germany and the entity „Regierungsbezirk in Bundesland XY“ is just an arbitrary subdivision.
  • N3: They also do not fulfil structural need. As I said multiple times yesterday, the same meaning can be achieved by adding instance of (P31) Regierungsbezirk (Q22721) and located in the administrative territorial entity (P131) Hesse (Q1199). And in my opinion the dates of inception and dissolution should not be applied to classes of things. A class of things automatically stops to exist when the last instance of that class stops to exist.

@ I would be interested to hear, if you have anything to say about why these items are notable other than that I don't have a clue what I am talking about or that similar items exist in other countries or for other subdivisions (e.g. Stadtbezirke). Pointing at similar items can be helpful to judge, wheter it's sensible to have such an item. But when creating these items you should also have arguments that these items in particular are notable. --Floscher (talk) 13:08, 20 December 2015 (UTC)

As Holger1959 mentioned, I work quite a lot on German district items. My evaluation is:

administrative territorial entities in Germany under discussion
ID criteria N1 criteria N2 criteria N3 suggestion comment
government region of Prussia (Q1850612) yes to hold administrative territorial entity of former country Prussia
government region of North Rhine-Westphalia (Q829277) yes to hold current administrative territorial entity
no no no to delete Regierungsbezirk (Q22721) is definied. instance of (P31) Regierungsbezirk (Q22721) and located in the administrative territorial entity (P131) Hesse (Q1199) are sufficient. No need for subclasses for every state of Germany. I agree with Floscher.
rural district of Baden-Württemberg (Q20738676)
district of North Rhine-Westphalia (Q20738811)
district of Thuringia (Q20738945)
no no no to delete No need for subclasses for every state of Germany too. I agree with Srittau.

--Gbeckmann (talk) 23:22, 22 December 2015 (UTC)

How do I handle Wikipedia articles with many topics in the same article?

I'm currently trying to work with items of type instance of (P31) building (Q41176) and to do that I look through categories on Wikipedia that contains articles on buildings.

I often come across articles like da:Cloëtta which is an article talking about a family, the business they owned in Denmark, which is now a trademark of a foreign company and talks about two different factory buildings they had in Copenhagen. So, a lot of topics blended together.

I would really like to create an item that points to one of the factory buildings for which I have some data, but should I do that in the existing item containing the link to the article or should I create a new item for the building? If I do that, how do I then get proper linking back to the article?

And, broading the scope of the question, should an article like the example be split into 7 items, three for the persons, two for the buildings, one for the organization and one for the brand?

Thanks in advance for any pointers or viewpoints that can help me to get on :-) --VicVal (talk) 13:56, 22 December 2015 (UTC)

I've wondered about this sort of problem too. Is it ok to add a wikipedia page for, say, the building that is just a redirect to the existing page, and link the wikidata item for the building to the redirect page? Not every item in wikidata has to be linked to a wikipedia article though... ArthurPSmith (talk) 14:55, 22 December 2015 (UTC)
The basic question is - Can these things be described by the same statements? From that it is immediately obvious that 7 different items will be needed as the statements to describe one will not be true about the others. The WP article can probably be linked to the item for the organization but maybe one of the people is more appropriate. Hope that helps. Joe Filceolaire (talk) 15:07, 22 December 2015 (UTC)
Linking from redirects is not possible (though there is a bodge where you link to a page then turn it into a redirect) and in this case I think it is not necessary. Joe Filceolaire (talk) 15:07, 22 December 2015 (UTC)

Navigational card

I have no pretty information card (label+description+image) in navigational popup for some days. Is it my personal issue or a script is broken? --Infovarius (talk) 18:17, 21 December 2015 (UTC)

They are working here for most items, but they don't seem to work on items with Wikimedia-related stuff like lists and categories. Sjoerd de Bruin (talk) 18:33, 21 December 2015 (UTC)

Working again. May be some caching issue. --Infovarius (talk) 15:07, 23 December 2015 (UTC)

Disambiguator tool request

Just in case someone want work on this tool. --ValterVB (talk) 09:26, 22 December 2015 (UTC)

Reasonator does a perfect job in disambiguation. The best part is that it works in any language and uses descriptions based on existing statements. This makes it REALLY relevant to add statements in order to disambiguate. I use it all the time. GerardM (talk) 10:16, 22 December 2015 (UTC)
But Reasonator use Wikidata, the tool is useful to check if all the sitelink are real disambiguation page. --ValterVB (talk) 10:27, 22 December 2015 (UTC)
Feel free to create a Phabricator task for it. Even better would be a database report with non-disambiguations connected to disambiguations. Matěj Suchánek (talk) 08:52, 23 December 2015 (UTC)
Good idea. --- Jura 09:36, 23 December 2015 (UTC)
I do not understand what such a disambiguator is supposed to do. Reasonator does a perfect job. What more do you need? Thanks, GerardM (talk) 10:14, 23 December 2015 (UTC)
We don't want to "disabiguate" (whatever it means), we would like to see if all sitelinks in an item are consistent. Matěj Suchánek (talk) 10:22, 23 December 2015 (UTC)
Yes, for example with this tool I can check immediately if all the sitelink in English (Q182) are disambiguation, (en passant: bclwiki, glkwiki, maiwiki aren't disambiguation page). --ValterVB (talk) 10:38, 23 December 2015 (UTC)

Time series data type?

Has there ever been thought given to having a "time series" data type - that is, an extension of the Quantity data type to allow for a series of values through time, probably best represented as a list of (Timestamp, Quantity) pairs? Looking at the property proposals in Wikidata:Property_proposal/Economics it seems to me the current proposals would all require dozens or thousands of statements with point-in-time qualifiers, when really what is most useful would be a single statement that captures the way the value has changed over time. ArthurPSmith (talk) 15:13, 23 December 2015 (UTC)

different case

Why should disambiguation pages be unlinked to other ones ? We have the issue in French with optic nerve disease (Q2879095), so we cannot have interwikis. It's not very friendly for Wikipedian users.--La femme de menage (talk) 10:00, 17 December 2015 (UTC)
Because, there is a superstition that they can't, @La femme de menage:. I don't like it too. --Infovarius (talk) 14:46, 21 December 2015 (UTC)
True disambiguation pages are there to list the different meanings of a given word or lexeme. Adding a link to a page in another language that is not a disambiguation page implies choosing arbitrarily between these different meanings and should therefore be avoided. However, in the case of optic nerve disease (Q2879095), the french article is not a real disambiguation page but a short article about a class of disease. I can't see why this article can't have interwikis. Casper Tinan (talk) 20:31, 21 December 2015 (UTC)
I guess @Andreasmperu: and @Holger1959: didn't see the question. --La femme de menage (talk) 02:31, 22 December 2015 (UTC)

@La femme de menage: Neuropathie optique looks like any other typical disambiguation page, so it should not be linked to real articles for one of the possible meanings and/or with specific contents. i don't understand French, but i think that fr:Wikipédia:Homonymie and fr:Aide:Homonymie will explain well what disambiguation pages are used for. Here you might want to check out Disambiguation pages/guidelines/fr. Holger1959 (talk) 09:05, 22 December 2015 (UTC)
PS: you can still use old style interwiki links on Wikipedias, like [[en:page title]] on frwiki, which is also used when you want to make links to paragraphs like [[en:page title#paragraph]] (when there is no specific page in an other language, but a part of a different article deals with a topic). Holger1959 (talk) 09:11, 22 December 2015 (UTC)

Sorry, @Holger1959:, I understand what a disambiguation page is supposed to be for English speaking people (ie on Wikidata), but it just happens that on French Wikipedia, they are also sometimes used as a "short page", with links to the included subjects. Please take into consideration Casper Tinan advice just above. "Old fashion links" won't help, as they are erased as soon as input. (Because "interwiki links are now managed on Wikidata" ). I still do not understand why this (pages linked) should be an issue for Wikidata admins - or users. @Andreasmperu:, any chance to have a reason ? --La femme de menage (talk) 21:40, 22 December 2015 (UTC)
La femme de menage The French page is a list of disease, so probably you can delete disambiguation template and solve the problem. --ValterVB (talk) 08:17, 23 December 2015 (UTC)
@La femme de menage: just for ping :) --ValterVB (talk) 08:19, 23 December 2015 (UTC)
If an item is a short page, please remove the disambiguation template. Otherwise, it would be considered a disambiguation page, and as such could not be linked together with non-disambiguations (ontology is totally different). Andreasm háblame / just talk to me 23:54, 23 December 2015 (UTC)

UNESCO and Wikidata

Hi all

I've set up a group of pages on Meta to help share content and data from UNESCO and improve information about UNESCO inscriptions (e.g World Heritage Sites) as part of my project as Wikimedian in Residence there.

I've created a data section as part of it, any feedback or participation would be greatly appreciated.

Also UNESCO have just released 750 images of World Heritage Sites to Commmons, I see from Autolist there are 280 sites without images, thanks very much for Navino Evans and Magnus Manske for adding the sites to Mix n' Match.

Also if anyone has interest in importing data from UIS using a bot please let me know, there is a lot of very useful data available and UNESCO are very interested in adding it to Wikidata.

Many thanks

John Cummings (talk) 10:12, 23 December 2015 (UTC)

Excellent news. What parts do you plan to upload to Wikidata? Potentially, could anything on UIS be made available? --- Jura 11:43, 23 December 2015 (UTC)
Hi Jura, to the best of my understanding potentially everything could be made available, it may be good to start with a specific area, the education data is very important and I think could be imported using a new property some qualifiers. I assume the first step is to map out the data and see if any new properties need to be created. John Cummings (talk) 12:48, 23 December 2015 (UTC)
@John Cummings: Please be aware of the licence problem: Wikidata has a CC0 licence for the data and UNESCO a licence similar to a CC-BY-NC licence (see here). Before doing any large data import check if UNESCO agrees for that data import under CC0 licence.
@Lydia Pintscher (WMDE): We still don't have any way to allow data releases under CC0 licence by organizations like OTRS. Anything planned for that purpose ? Snipre (talk) 12:57, 23 December 2015 (UTC)
Mpfh. I thought this was all handled in OTRS by now... Thanks for pinging. I will visit the WMF office in SF at the beginning of January. I'll add it to my list of things to poke people about. --Lydia Pintscher (WMDE) (talk) 13:57, 23 December 2015 (UTC)
Thanks @Snipre: and @Lydia Pintscher (WMDE):, can someone direct me to some clear guidance on when individual facts accumulate enough to become a copyrightable database? And how much of a database could be imported without breaking the copyright? Its my understanding UNESCO are happy to release data under an open license but a lot of the information inside the UNESCO database will already exist, all the items will exist so only some values would need to be imported. John Cummings (talk) 14:08, 23 December 2015 (UTC)
Please have a look at this statement. I am not a lawyer but my feelings are the next ones: identifier of a database is not a creative work so extraction of the whole list of elements with the corresponding identifier is not a big problem. Then a significant portion is difficult to define but more than 50% of a database is a significant portion especially when no other databases provide similar data.
So the first step is to concentrate on identifier mainly because UNESCO can have some interest to use that relations between its database and wikidata mainly to match its elements with other databases. Then for the rest I advice you to have a clear agreement with UNESCO until we can provide you an official way to release your data under the CC0 license. Or at least to have a clear idea about what data are critical for UNESCO and what not.
If different versions of the databases exists avoid to use the last one: most of the time people don't take care about archives. This allows you to test and to work with a minimal risk. Snipre (talk) 21:25, 23 December 2015 (UTC)

Copy references

I know that copy references has been disable. But is there a way to use it? Or is there something else? Xaris333 (talk) 23:08, 23 December 2015 (UTC)

Done. Xaris333 (talk) 01:00, 24 December 2015 (UTC)

Highest and lowest elevation

Hi every one! I joined the project before a day, and it's wonderful. In my edits, I saw that in the items about cities, there is just "elevation above sea level" - the average elevation, and there isn't "highest elevation" and "lowest elevation". It's a software problem. May someone can fix it.--Gederade (talk) 04:33, 24 December 2015 (UTC)

The closest properties I can find are highest point (P610) and lowest point (P1589) to indicate where the highest and lowest points are. To propose properties for highest and lowest elevation, go to Wikidata:Property_proposal/Place. Popcorndude (talk) 14:16, 24 December 2015 (UTC)

is there a property for associated property ?

For example, on pseudonym (Q61002) it would allow to link to pseudonym (P742) ? --Hsarrazin (talk) 09:57, 24 December 2015 (UTC)

subject item of this property (P1629)Wikidata property (P1687). Matěj Suchánek (talk) 10:13, 24 December 2015 (UTC)
Thanks, could not find it :) --Hsarrazin (talk) 11:17, 24 December 2015 (UTC)

Set time using the API

I keep getting a modification-failed error when I try to add a claim using a time value. Here the error json:

{"servedby":"mw1221","error":{"code":"modification-failed","info":"Malformed input: +2015-12-26T00:54:37Z","messages":[{"name":"wikibase-validator-malformed-value","parameters":["+2015-12-26T00:54:37Z"],"html":{"*":"Malformed input: +2015-12-26T00:54:37Z"}}],"*":"See for API usage"}}

What am I doing wrong? --Jobu0101 (talk) 00:03, 26 December 2015 (UTC)

After reading the API help page I think the + shouldn't be there. Mbch331 (talk) 05:24, 26 December 2015 (UTC)
@Mbch331: Without the + I get an invalid-snak error:
'{"servedby":"mw1136","error":{"code":"invalid-snak","info":"Invalid snak ($timestamp must resemble ISO 8601, given 2015-12-26T09:34:09Z)","messages":[{"name":"wikibase-api-invalid-snak","parameters":[],"html":{"*":"Invalid snak"}}],"*":"See for API usage"}}'
It works now when I don't use precision=14 (seconds) but precision=11 (day) with my original code. The + is important. Strange but enough for the moment since I don't think that I'll need more precision. --Jobu0101 (talk) 09:08, 26 December 2015 (UTC)

Link back to Wikipedia

If I link to an image in Wiki Commons it shows up on the image page that there is a link at Wikidata, see File:Frederick Ambrose Clark in 1916.jpg. Is there any way that Wikidata can show me that I have formed a link to a Q-value in Wikipedia? For instance at w:Draft:Coroner of New York City I have linked people to their Wikidata since they will never have their own articles. --Richard Arthur Norton (1958- ) (talk) 00:45, 27 December 2015 (UTC)

If you want to have all coroners in Wikidata you can. If you want a list shown in a Wikipedia, you can. I would not say that whatever topic will never have an article. What I can say is that linking to redlinks or using lists that include redlinks but are items in Wikidata is something we have for some time. It is just for a Wikipedia to understand the quality implications implied. Thanks, GerardM (talk) 11:58, 27 December 2015 (UTC)
With arbitrary access I believe it's a matter of using/creating right template with data off WD Hakan·IST 12:12, 27 December 2015 (UTC)

Qualifier for buildings that are proposed / planned / under construction

Hello. I'm wondering whether there's a qualifier I could add to statements about buildings that are proposed, planned or under construction, e.g. One Nine Elms or The Pinnacle. These both have 'instance of' statements asserting that they are buildings, even though they're not quite at that stage yet (although the former is under construction, so perhaps that's arguable). In addition it'd be nice to be able to add statements about proposed height, number of floors, and so on, without confusing people doing queries across the database for things like "tallest building in London". --Frankieroberto (talk) 17:09, 27 December 2015 (UTC)

API question

Why does Query: claim[166:106301] and noclaim[31:5] and noclaim[31:(tree[11424][][279])] give no results? It should give 4 results since we get

And we have 156-75-77=4. --Jobu0101 (talk) 17:11, 27 December 2015 (UTC)

Query: claim[166:106301] and claim[31:(tree[11424][][279])] returns for me 79 results. --Pasleim (talk) 17:21, 27 December 2015 (UTC)
Same here now. I think it was some cache problem. --Jobu0101 (talk) 18:21, 27 December 2015 (UTC)

Tool to turn string claims to monolingual claims

It's more than a year ago that the decision was made to deprecate the properties P357 (P357), P387 (P387), P438 (P438) and P513 (P513) with string data type and create corresponding new properties with monolingual data type. However, the move from string to monolingual is tiresome and so many claims with deprecated properties still exist.

To make it less tedious, I created a tool with which it's rather easy to turn string claims to monolingual claims. You find it on Hope you will use it so that we can soon delete the deprecated properties. --Pasleim (talk) 19:20, 23 December 2015 (UTC)

As I said, more than 10,000 can be changed automatically. See my previous comment. --JulesWinnfield-hu (talk) 19:46, 23 December 2015 (UTC)

Thank you so much Pasleim. That's exactly the kind of semi-automatic tool I was dreaming of.
Christmas gift before time... --Hsarrazin (talk) 21:29, 23 December 2015 (UTC)
@Pasleim: - could you please add Latin as possibility of title for periodicals, please ? A lot of scientific periodicals have latin title, like Acta Organologica (Q343172) ;) --Hsarrazin (talk) 21:35, 23 December 2015 (UTC)
+1 Snipre (talk) 23:07, 23 December 2015 (UTC)
@Pasleim: ...and Esperanto and Dutch. Also, in many cases, you will find a qualifier language of work or name (P407) (see this example) providing the sought language; this qualifier should preferably not be ported to the new monolingual statement. Thanks for the tool :) -- LaddΩ chat ;) 23:12, 23 December 2015 (UTC)
but the language of the title can be very different from the language of the review ;)
also, I find a lot of periodicals where the title has been added as transcription in English of the original title, like Belaruskaya Litaratura (Q15766655) that is transcripted from Belarus... what should be done, then ? (for now, I skip, of course) --Hsarrazin (talk) 23:21, 23 December 2015 (UTC)
Could this tool be changed so that it adds an edit summary more specific than "Updated item"? --Yair rand (talk) 23:10, 23 December 2015 (UTC)
also, a lot of items share a same "reference" (like Befolkning i tätorter 1960-2010 - not the only one, but I've seen this one quite a lot). Could the language value be transferred to all items instead of taking the risk to have diff. languages on diff. items ? :)
even better would be to have an item with this reference, I think, so that it would not have to be saved as title in references ;) --Hsarrazin (talk) 00:43, 24 December 2015 (UTC)
Befolkning i tätorter 1960-2010 (sv) can be found in almost every item that uses Property:P775. I know, I am the one who added them! Feel free to change! -- Innocent bystander (talk) 17:32, 25 December 2015 (UTC)
Hi @Innocent bystander:,
I would gladly create an item for Befolkning i tätorter 1960-2010 (it's now Population in urban areas 1960-2010 (Q21855710)), but I would not know how to add properties to it, above those given in reference of those items, as I don't understand Sweedish at all, and it is not my field of best competence ;)
I guess this is an official reference document (like INSEE nomenclature for French cities ;)
and… I do not know how to use other tool than Autolist, to make serial changes on items. and I wouldn't like to change manually more than 2200 items, as it seems from Autolist :)
I did it manually on Finspång (Q1000129), but it's quite tedious ;) --Hsarrazin (talk) 13:28, 27 December 2015 (UTC)
Well, I do not know how "official [as a] reference document" it really is. I have used it as a reference for property:775, since pre-1980-documents do not tell that. And this document is the only source in those cases. It is a poor source for the population-numbers. Such references as Localities 2010 (Q14907217) is much better for that. One problem is that the file probably will be updated and many numbers changed when the 2015-numbers arrive. -- Innocent bystander (talk) 13:53, 27 December 2015 (UTC)
well, official is not the question here... easyness of ref. rather…
probably, the name of the file will be changed if 2015 nbrs are included. Maybe an item for each edition would solve the problem of updates ? what do you think @Innocent bystander: ?
the use of individual manually completed reference for each town seems really very heavy for modifications :( --Hsarrazin (talk) 15:33, 27 December 2015 (UTC)
@Hsarrazin: Do as you wish! Not only the 2015-numbers will be added to the file. The revised 2010-numbers will also be added to the file. That is why it is not a good source for those numbers.
One question regarding Q21855710 is that the xls-file is used with official website (P856). Should not the ref-url-property be used instead? The official page of Statistics Sweden is -- Innocent bystander (talk) 19:22, 27 December 2015 (UTC)
oups, sorry, my mistake, I intended to use reference URL (P854) - don't even know how this happened :/ --Hsarrazin (talk) 19:35, 27 December 2015 (UTC)
I a repetetive failure on Hebrew titles, like Arieh Dulchin (Q4790361) Malformed input: +1969-12-15T05:00:00Z --Hsarrazin (talk) 11:19, 24 December 2015 (UTC)
@Pasleim: I'm encountering a lot of Dutch sources--can that be added as an option? Thanks for your work. —Justin (koavf)TCM 16:00, 25 December 2015 (UTC)
Dutch, Esperanto and Latin are added as additional language options. --Pasleim (talk) 09:28, 26 December 2015 (UTC)

Is there a statistic tool that would allow to see the importance of the task, and to follow the progress ? I'm really incompetent with SPARQL :( --Hsarrazin (talk) 16:26, 27 December 2015 (UTC)

Gender variants for labels?

This must have come up before, but I'm not seeing it in the FAQ.

Will Wikidata ever support gender-matched variants for labels? (perhaps more generically, parameterized labels?) I.e. for languages where e.g. teacher (Q37226) has two different forms, depending on whether the architect's gender is male or female? If so, when is that planned? If not, is there a proposed workaround?

There are already the "also known as" extra labels, but I am looking for a way for Wikidata to be able to provide the appropriate label not just per language, but per language and one or two parameters (I'm thinking of gender, but some languages have inflections based on other distinctions too, like animate or inanimate), e.g. being able to say something like {{Q|37226|gender=f}} and get, for German, 'Lehrerin' rather than 'Lehrer'.

The lack of this means that it is impossible to construct automated lists on certain topics (e.g. anything including the occupation (P106) property) without manual fix-ups. Thanks! Ijon (talk) 17:27, 24 December 2015 (UTC)

Ijon, there is a workaround, even though I am not sure if this is the best solution for using these names: Property:P1549, see for example Q17#P1549. Please let me know if you see any application on how to use these names in wikipedia infoboxes. --FocalPoint (talk) 18:19, 24 December 2015 (UTC)

However, the ultimate solution would be to incorporate the wikctionary data in Wikidata so that we could end up with {{Q|37226|gender=f|number=plural}} and get "teachers", rather than teacher (while in french you would get "enseignantes" rather than enseignant (male) or enseignante (female). --FocalPoint (talk) 10:28, 25 December 2015 (UTC)

@FocalPoint, Ijon: I am not sure it always is as easy as only looking into the Gender of the person. If you use Teacher (f) or Teacher (m) in Swedish (sv) mainly depends on what education (s)he has. A modern teacher is addressed with a gender-free word. The gender-free form of Teacher is identical with the m-form in this specific case. -- Innocent bystander (talk) 11:35, 25 December 2015 (UTC)

Innocent bystander, I fully agree that it is not a simple thing, as each language works differently. From my knowledge of Greek language, all instances of female teacher must have a different form (δασκάλα) than that of a male teacher (δάσκαλος), while there is no gender-free word for a teacher. However, what we are discussing here is whether the existing infrastructure of Wikidata allows for such differences (male/female etc). How you can use it is another question. --FocalPoint (talk) 11:49, 25 December 2015 (UTC)

teacher (Q37226) is an item that refers to either a male or a female teacher. Any Label which refers only to a male teacher or only to a female teacher is wrong. "professor ou professora" would be an acceptable label. Recruitment consultants round the world have developed gender neutral versions of many job titles over the last few years, such as "professor(a)" and I think these would also be acceptable (your mileage may vary). Would "δασκάλα/ος" work for Greek teachers?
Plural versions of job titles are generally not used as labels as they are harder to reuse in infoboxes. Male only and female only job titles are acceptable as aliases but not as labels. There was some discussion as to the creation of a property for the "female version of this job title". I opposed it as it seemed to be an excuse to use male only labels for these items instead of finding inclusive labels. That is how I understand the situation is today. Joe Filceolaire (talk) 03:05, 27 December 2015 (UTC)
I resent the notion that we should include "acceptable labels". There is a way around this and it is called gender.. It is something that Wikidata does not support yet. When we need gender specific labels, use a script to do that as long as Wikidata is impotent in such things. Thanks, GerardM (talk) 12:00, 27 December 2015 (UTC)
@Ijon: The french WP is using a template which defines the correct form by analysing the job and the gender of the item. So the best is to have a post processing of your data which corrects rwa extracted data from WD. Snipre (talk) 21:57, 27 December 2015 (UTC)
For frwiki, it's at fr:Module:Dictionnaire Wikidata/Métiers. Implementing it locally has the advantage that you don't get advice on what form to use from people who don't read or write the language. --- Jura 08:28, 28 December 2015 (UTC)

See also Wikidata:Property proposal/Unsorted#Female form of label (string).--Jklamo (talk) 00:04, 28 December 2015 (UTC)

Joe Filceolaire, every language is different. "δασκάλα/ος" would certainly work (and be quite adequate) for Greek teachers in a text of instructions from the Ministry of Education. However Jane the teacher can only be δασκάλα and Joe the teacher can only be δασκάλος in an infobox. Any other case would get you bad grades from your δασκάλα/ος :) as unacceptable use in Greek language. Jura made a nice proposal, but it is a temporary solution and has to be fed with data for each and every word, rather than reading it from Wikidata. I do not favor any solution, but it would be nice to have an easy to implement solution. --FocalPoint (talk) 22:30, 5 January 2016 (UTC)

Connect to person only appearing in a list

Samuel Davidson Nutt only appears in a Wikipedia list Coroner of New York City how would I connect the two? --Richard Arthur Norton (1958- ) (talk) 19:59, 26 December 2015 (UTC)

Connect where? --- Jura 09:42, 28 December 2015 (UTC)
en:Template:Redwd should do the trick. author  TomT0m / talk page 12:55, 28 December 2015 (UTC)

Wikidata weekly summary #190

Wikidata weekly summary #186

One to two percent of people have gout

We're trying to use the data in Q133087 to populate an infobox at w:en:Gout. One of the problems is the "prevalence". The actual prevalence is "somewhere between one and two percent of people have gout". This field doesn't seem to permit a range of numbers, i.e., "1–2". It also reports this number out as "1 percentage", which is bad grammar. (It should either say "1%" or "1 percent".)

Can someone tell me how to fix these problems in the record, or is this something that needs dev support? WhatamIdoing (talk) 18:09, 24 December 2015 (UTC)

Values for the property prevalence (P1193) are supposed to be unitless numbers between 0 and 1 (1 being 100%), so you would presumably want 0.015±0.005. Currently, the template w:en:Template:Infobox medical condition(new) is using the raw value instead of formatting it into a percentage, which causes this to appear as "0.015±0.005", instead of either "1.5%" or "1.5±0.5%" or "1% - 2%". --Yair rand (talk) 18:21, 24 December 2015 (UTC)
This formatting limitation makes it pretty useless (uninterpretable) for the broad readership of Wikipedia. That's unfortunate, and seems to be good reason to be reluctant to use for now in mainspace. Soupvector (talk) 18:40, 24 December 2015 (UTC)
Converting it to a human-readable percentage should be easy in the template. But how should a range be handled? Prevalence is not a single-number answer. 0.015±0.005 is technically wrong. WhatamIdoing (talk) 20:05, 24 December 2015 (UTC)
A Lua module could access value.lowerBound and value.upperBound to give a result of "1% - 2%". Would this be correct? --Yair rand (talk) 20:54, 24 December 2015 (UTC)
Yes exactly we need the ability for ranges of values. And the range of values typically apply to a specific population. Thus we have 1 to 2% (developed world)Doc James (talk · contribs · email) (if I write on your page reply on mine) 23:00, 24 December 2015 (UTC)
The original proposal for prevalence (P1193) recommended using qualifiers to specify particular groups, suggesting country (P17) for countries and a new property for demographic groups. In practice, the uses have been split between country (P17), location (P276), and applies to jurisdiction (P1001). I recommend standardizing to one of these or a new property (maybe "applies to region"?), and then adding it as a qualifier constraint to P1193 so that the data is consistent.
We also don't yet have an item for "developed world". Is this a term used consistently with the same boundaries? --Yair rand (talk) 23:42, 24 December 2015 (UTC)
The boundaries of "developed world" will probably change over time, and it might not be a sensible term for anything pre-1900. I believe that it is generally taken to mean "nearly all of Europe, US, Canada, and Japan", but different sources are likely to use different lists of countries, especially in different fields (e.g., medicine vs manufacturing). WhatamIdoing (talk) 21:25, 28 December 2015 (UTC)
We also need different demographics. The cancer prevalence in older adults is much higher than the cancer prevalence in children. Some diseases exist almost exclusively within certain ethnic groups. WhatamIdoing (talk) 21:26, 28 December 2015 (UTC)

Commons category to store one item

Should we be creating a Commons category at Wikimedia Commons to store just one item, so it can be linked to Wikidata? --Richard Arthur Norton (1958- ) (talk) 23:18, 26 December 2015 (UTC)

AFAIR I have seen some of these cases lately, but I don’t know whether there’s a policy to do so. Would it be possible to link the file (if image) with image (P18) instead? —MisterSynergy (talk) 10:17, 27 December 2015 (UTC)
If there is an image I do that, but I have been adding in the image of the obituary. I tried adding the text at Wikisource, but there are so few people there, that the few that edit there have made arcane rules. For instance they will not let you annotate the text or to link back to Wikipedia from a name in the text. And unless you cut and paste in a full document, it is immediately deleted, since they do not accept partial transcriptions. I had multiple documents that I was actively transcribing deleted while I was transcribing them. --Richard Arthur Norton (1958- ) (talk) 00:32, 28 December 2015 (UTC)
Creating categories for just one item is a bit of a contentious issue on Commons. Personally, I'd say go for it, especially since categories tend to fill up anyways and they fill a structural need. (Commons cats are not comparable to Wikipedia cats.) --Srittau (talk) 08:53, 28 December 2015 (UTC)
I agree. If the issue comes up, we can always vote at Commons to keep them. I have noticed that categories tend to stay longer, where people tend to delete a page at Commons if it only contains one or two items. So, I never bother making a page. --Richard Arthur Norton (1958- ) (talk) 19:57, 28 December 2015 (UTC)

Would we ever link to the entries at They have genealogical information for every person that is dead. They also have free access to birth, marriage, and death certificate and the census records for these people linked to the profiles. The problem is that there is no landing page unless you are logged in as a free registered user. If you do not have a free account or are not logged in, it takes you to the login page. If you are logged in, you get an amazing genealogical history of that person and free access to all the documents. I have been filling in hundreds of missing birth dates and missing middle names using this resource. --Richard Arthur Norton (1958- ) (talk) 00:45, 28 December 2015 (UTC)

@Richard Arthur Norton (1958- ): There's no login required for access to some records, for example a birth record for this familiar person. If the identifiers seem stable, I would recommend you propose them as a property in Wikidata:Property_proposal/Person perhaps under the Identifiers list (though I can imagine there might be more than one, for each genealogical record related to the person - birth, marriage, death, census, etc). Maybe there are other identifiers associated with individuals as a whole that should be added? Anyway I think a property proposal would do what you want here - adding the identifiers with a proper formatter automatically provides a link to the site. ArthurPSmith (talk) 19:05, 28 December 2015 (UTC)
Anna Augusta Kershaw appears to ask for login-- Hakan·IST 20:20, 28 December 2015 (UTC)
Feel free to propose familysearch person ID as new property. We already have comparable properties (like Property:P1819), inaccessibility without login is not obstacle. --Jklamo (talk) 02:35, 29 December 2015 (UTC)

Tool wanted

Does anybody have a tool I can borrow the code from, to easily make edits like this? -- Innocent bystander (talk) 13:56, 27 December 2015 (UTC)

Quick statements can do it --ValterVB (talk) 15:02, 27 December 2015 (UTC)
in order to avoid precisely the problem with "Befolkning i tätorter 1960-2010" Population in urban areas 1960-2010 (Q21855710), I would certainly recommend to create an item for the source, and use it with stated in (P248), which would ease the edit and/or modification ;) --Hsarrazin (talk) 15:29, 27 December 2015 (UTC)
@ValterVB: There is something I misunderstand here. When I add:
Q21560019	P1448	sv:"Bromma"	S248	Q21856133	S813	+2015-10-24T00:00:00Z/11
The sources are missing! -- Innocent bystander (talk) 20:00, 27 December 2015 (UTC)
@Magnus Manske: Help!!!! :) I don't know why don't work (maybe monolingual conflict with source?). --ValterVB (talk) 20:17, 27 December 2015 (UTC)
The missing S813 statements has been questioned on Magnus' talkpage for some time, but unfortunately no solution is available yet. I've been searching for the sourcecode to try and understand the problem but so far without succes :-( --VicVal (talk) 00:13, 29 December 2015 (UTC)
@VicVal, ValterVB: The problem looks like it is deeper than just only QuickStatments. Query: claim[31:18333556] and noclaim[1448] Gives me 2523 results. It should be less than that, since claim[31:18333556] alone gives me 2523 results and several of them has P1448! -- Innocent bystander (talk) 11:59, 29 December 2015 (UTC)
Yes, something of WDQ don't work with monolingual. If you try with claim[1448] WDQ return only 4 item that haven't value (no value or unknow value). --ValterVB (talk) 12:12, 29 December 2015 (UTC)

I had an equivalent question two weeks ago: Wikidata:Project chat/Archive/2015/12#Tools to add external references to existing statements. I guess there is no such tool at the moment, but it would be great to have one! —MisterSynergy (talk) 20:48, 28 December 2015 (UTC)

P2429 "expected completeness"

expected completeness (P2429) is now available for documenting other properties; please see its talk page for examples. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:36, 29 December 2015 (UTC)

RDF Dumps are using a outdated Wikidata-Toolkit.

I checked the RDF exports in And these file are using a outdated version of Wikidata-Toolkit. These dumps are incompatible with the RDF standards.

See also:,

--Lcy2000 (talk) 11:47, 28 December 2015 (UTC)

Hi Lcy2000, maybe you should move this information to Wikidata:Contact the development team. --Succu (talk) 20:28, 28 December 2015 (UTC)
Thanks for your helpful suggestion. --Lcy2000 (talk) 12:18, 30 December 2015 (UTC)

Can't save an edit

I'm trying to add a head of government for Racine, Wisconsin (Q40340). The mayor's name is John Dickert. For whatever reason, the first time I tried to add him, his name was left off at John Dick. Now when I try to fix it, the save button remains greyed out no matter what I do. I've tried multiple computers, multiple browsers, as well as logging out of my account and editing it anonymously. I have had the same problem in all attempts. Reschultzed (talk) 19:05, 29 December 2015 (UTC)

The problem is that John Dickert don't exist in Wikidata. --ValterVB (talk) 19:14, 29 December 2015 (UTC)
@Reschultzed: To expand on Valter's reply, the propoerty in question must have a value which is an item; first, create an item for John Dickert. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:14, 29 December 2015 (UTC)

New section for sources?

Currently people look at the reference section below statements to find the sources of a statement and seem disappointed when they find the source of statement, but not necessarily the reference they are looking for.

Should there be a separate "source" section? or Shouldn't we simply call the "reference" section "source" section? --- Jura 22:50, 29 December 2015 (UTC)

Originally, it was called "source" but was then renamed to "reference" per Wikidata:Project chat/Archive/2014/04#Proposal: change "source" to "reference" --Pasleim (talk) 23:16, 29 December 2015 (UTC)
Looks like that was meant to solve one potential problem, but created another one. --- Jura 09:43, 30 December 2015 (UTC)

Deleted Population from comune of Italy (Q747074)

Just for info, I have deleted population (P1082) from comune of Italy (Q747074). See at this page for detail. --ValterVB (talk) 13:08, 30 December 2015 (UTC)

+1. No choice if we want to have the minimal respect for author rights. There is a simple way to be free again: this is to set WD under the CC By or CC BY-SA licences like WP. Snipre (talk) 15:26, 30 December 2015 (UTC)
This example shows why using any license except CC0 or a similar public domain license for database works is a bad idea. --Srittau (talk) 15:33, 30 December 2015 (UTC)
@Srittau: Can you explain why ? We have two types of legal requirmeents, the copyrights and the databases right. We should deal with both and this prevents to use large parts of databases under CC0 unless a formal approval of the data source. Copyright in case of data is arguable and in case of isolated use of data pieces we can forget it but use of large amount of data from an unique source is still unresolved. We can spent a lot of time to find a solution to that question. so to save time we can look at the minimal requirements to use data like the ones above and we can see that often the minimal requirement is a CC BY licence.
What prevent us to set WD under a CC BY licence, that's for me a better question. Snipre (talk) 16:01, 30 December 2015 (UTC)
Because, as can be seen in this case, a CC-BY license hinders reuse and limits Wikidata's usefulness, especially in a "linked data"/"semantic web". --Srittau (talk) 16:26, 30 December 2015 (UTC)
It is more or less impossible to claim copyright for anything created here. This would therefor only apply to data that is imported from other places, like Italian statistics. And such data is anyhow worthless if we do not show where it comes from. -- Innocent bystander (talk) 18:04, 30 December 2015 (UTC)

Inflated figures for unreferenced statements

The image above and related stats have been much discussed of late; but I think the numbers are misleading. We have many statements which can be considered their own references, and thus do not need a separate citation. For example, Authority control identifiers are resolvable as URLs. Likewise official website (P856). It's probably unnecessary to give citations for given and family names or gender for many biographical items; and for statements about properties, too. Would it be possible to adjust the figures to take account of these (or, indeed, has anyone already done so)? What other properties are affected? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:51, 29 December 2015 (UTC)

Hey! Please also see If we could come up with a list of properties that we may deem as 'not needing a reference' I could probably gather data for that too! ·addshore· talk to me! 20:23, 29 December 2015 (UTC)
@Addshore: Thank you. Where should that list be compiled (or would it be better to have a meta-property indicating that a property is self-referencing?) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:41, 30 December 2015 (UTC)
The first name, family name and gender are data that ultimately should have references like any other data, so I don't see any reason to "adjust" the diagram in consideration of them.--Shlomo (talk) 20:27, 29 December 2015 (UTC)
You want a reference that Paul McCartney's given name is Paul and his surname is McCartney?!? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:41, 30 December 2015 (UTC)
Actually, I don't care much about the properties family name (P734) or given name (P735), since the whole conception of name properties as item datatypes makes them virtually useless. But if there should be some useful property for names sometime, yes, I would like to see references like for any other data. Or do you think that a statement that Jorge Mario Bergoglio's first name is Francesco (Q2268455) doesn't need a reference? BTW if we make a general rule that given name (P735) doesn't need a reference, how can we prove, that a statement, that "Paul McCartney's given name is Helmut (Q18191106)" is false?--Shlomo (talk) 09:29, 31 December 2015 (UTC)
Personally i would like to see "unreferenced" group splitted in 3 parts: 1) names, 2) authority properties (i.e. references not needed) 3) all other. -- Vlsergey (talk) 23:02, 29 December 2015 (UTC)
To be honest, this graph is fairly useless. There are many statements that don't need referencing, like structural statement. instance of (P31) and subclass of (P279) are often (but not always) uncontroversial, for example. In the vast majority of cases, coordinate location (P625), located in the administrative territorial entity (P131), and country (P17) can be resolved by looking at a map. (Again, there are exceptions, for example in disputed territories or when there are "official" coordinates, like for airports.) I feel that the "it's unreferenced, it's useless" crowd uses this as a mantra without questioning why and especially when quotations are needed. --Srittau (talk) 11:17, 30 December 2015 (UTC)
I don't think it's "useless". but it could be improved, which is what I'm suggesting. Your later point is well-made. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:41, 30 December 2015 (UTC)
I agree we need to somehow identify uncontroversial "structural properties" that do not require references (using some item "Wikidata property not requiring a reference"?)
Regarding authority properties, we already have 79 properties in Category:Authority control properties and 83 instances of "Wikidata property for authority control", and more generally all 423 instances of "Wikidata property representing a unique identifier" (46 of which are also Wikidata property for authority control (Q18614948)). There is no category for such identifiers (though they should all be in both Category:Properties with single value constraints and Category:Properties with unique value constraints), but at one point we'll need to get a complete list in order to convert them to the upcoming new datatype for identifiers (see Add a new datatype for identifiers). -- LaddΩ chat ;) 17:07, 30 December 2015 (UTC)
Is there a way to request a source for a property on WikiData? That would be more helpful to me than claiming that 95% of WikiData is now unsourced, as I know myself that most of the sources are Wikipedia itself (which might be unsourced too). Edoderoo (talk) 21:56, 30 December 2015 (UTC)
AFAIK the only way is to ask for it on the discussion page as for now. But some procedure for marking the disputed statements where a source/reference was requested and deleting such statements if the reference is not added within defined time range could bring Wikidata on another level. It won't work though as long as the disapproved and deleted statements are promptly mass-returned by bots or WiDaR :( --Shlomo (talk) 09:41, 31 December 2015 (UTC)

How to clean up Q707544 ?

Currently (this version), Q707544 is a combination of a list of episodes and an item about a television series (with 17 sitelinks).

  • Most sitelinks are clearly about a "list of episodes" item.
  • Most statements and some of the labels and aliases are for an item about the series.

A clean solution could be to create two new items and move things that apply there. --- Jura 17:46, 30 December 2015 (UTC)

Without really looking into it: a TV series, an episode of a TV series, and an episode list are three (related) concepts to me. Splitting seems to make sense. I think the main reason that this has not happened yet is that at least on the English and the German Wikipedia, the article about "Naruto Shippūden" seems to be integrated into the main "Naruto" article, where Naruto Shippūden is - according to Wikipedia - a sequel to the original Naruto. (In this sense the "Naruto" articles describe not the concept of the series "Naruto", but the concept of the class/set "Naruto TV series", but this would go too far here, I think.) TLDR: yes. --Srittau (talk) 18:12, 30 December 2015 (UTC)
I agree that that we need at least two items.
Generally, as we try not delete items, we would move just parts of the content to a new item.
In this case, the two concepts seem to be so much mixed-up that two new items seem preferable. --- Jura 19:00, 30 December 2015 (UTC)
  Done: Q21898102 (series) Q21882811 (list). We just need to decide what to do with the initial item. --- [[User talk:Jura1|Jura]list] 08:22, 31 December 2015 (UTC)
An examination of the history of Q707544 shows that it was created as "List of Naruto: Shippuden episodes". The new item for the lists should therefore be redirected to it. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:13, 31 December 2015 (UTC)
@Pigsonthewing: Eh, not really. The mess that Jura1 was talking about existed in that item from the start, with sitelinks to both articles about the series and lists of episodes. That said, I wouldn't mind if that item is used for the list of episodes. Jon Harald Søby (talk) 16:29, 31 December 2015 (UTC)
The problem with redirecting it to one or the other is that the labels/descriptions/aliases are likely to be messed up again (in one or the other language). Further, as the statements had been inconsistent with the sitelink over a considerable item, people might not get what they had linked before.
--- Jura 18:07, 31 December 2015 (UTC)


There are properties about awards (award received (P166), for work (P1686), prize money (P2121)). Is there available any page that consists of list of properties usable to describe award items? There are many events when award is received (sport competitions, honors, etc.). It might be helpful if there is description what to assign to generic properties like instance of (P31) or subclass of (P279). I would like to see some wikiproject similar to those available in Category:Properties list in a WikiProject. Paweł Ziemian (talk) 23:07, 30 December 2015 (UTC)

You mean yo say for instance an amount, a statue, a medal ? They would all not be "instance of". It would be good to include that in one way or other GerardM (talk) 09:22, 31 December 2015 (UTC)
@Paweł Ziemian: Once you started adding statements to an item, the property suggester will propose new properties (when you click "add" at the bottom of the list of statements) that other similar items already bear. See Wikidata:Project_chat/Archive/2015/05#statement_suggester_.2F_dropdown_menu_when_adding_a_statement.
You may also create a new template like Template:Authority control properties and put it on the talk page of these properties. -- LaddΩ chat ;) 16:59, 31 December 2015 (UTC)

Tool to move statements fom one item to another item

Hi, I am working on violations of contraints and often I have to move some statements from an item to a more appropriate item. This is quite challenging when the statements are well constructed with qualifiers and references. So my question is to know if a tool exists to perform the move of a whole statement with all data (qualifiers, references) in one click after providing the target item of the move. In one word this is a simialr tool as the Move gadget but for statements and not for sitelinks. Thanks Snipre (talk) 11:39, 31 December 2015 (UTC)

+1 <3 Matěj Suchánek (talk) 11:47, 31 December 2015 (UTC)
I agree; I had to split two badly merged items yesterday, and encountered just this issue. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:04, 31 December 2015 (UTC)
  • This is needed. The solution for sitelinks works well. If there was the same for statements. There is , but it doesn't seem to ready yet.
    --- Jura 11:48, 31 December 2015 (UTC)
  • AFAIK there is no tool, but if multiple statements need to be moved, you can move them to a new item using procedure at Help:Split an item, then merge the new item in the target item. -- LaddΩ chat ;) 16:40, 31 December 2015 (UTC)

History of Wikipedia

I have a question Does wikipedia has a lot of works all the time?  – The preceding unsigned comment was added by (talk • contribs). 19:39, 31 December 2015‎ (UTC)