Wikidata:Project chat/Archive/2016/03

Latest comment: 8 years ago by Piotrus in topic Please merge

P570

 

Hi, how can I set value as unknown in this property?--ԱշոտՏՆՂ (talk) 18:53, 25 February 2016 (UTC)

See File:SnaktypeUI.png Matěj Suchánek (talk) 19:32, 25 February 2016 (UTC)
Thanks, Matěj Suchánek jan--ԱշոտՏՆՂ (talk) 19:37, 25 February 2016 (UTC)
Why would you ever want to set date of death (P570) to unknown? We have "Precision" so worse case you have to set it to the right millennium. Multichill (talk) 22:55, 26 February 2016 (UTC)
Not always. Person can die somewhere in the end of 20th century, but it can also be 21st century. --Edgars2007 (talk) 23:22, 26 February 2016 (UTC)
There are many historical people for who our only information is that they were alive in a given year, say 1480. We do not know if they were aged 20 or 80 in that year, or whether they died aged 21 or 121. Their date of death (and similarly birth) is truly not known to any precision. Thryduulf (talk: local | en.wp | en.wikt) 22:23, 27 February 2016 (UTC)
That's not totally true, it's just an information display problem. You can enter a date with a precision of the century, and only the century is displayed. The correct way would be to show the two centuries in the UI, but it's not done yet (if somebody wants to code, it's on my TODo list since forever /o\) author  TomT0m / talk page 12:39, 1 March 2016 (UTC)
I don't think there is any consensus about what the precision means. It is especially confusing because the before and after attributes are present and documented, but not in use and always set to 0. At least in the case of days, some Phabricator discussions have indicated that after = 0 and after =1 would have the same meaning. That is, when before = after = 0, this does not indicate an exact value; there is no method to indicate an exact value. Jc3s5h (talk) 13:11, 1 March 2016 (UTC)

Wikidata edits uselessly repeated in Wikipedia watchlist: 173 lines for a single edit

Hello,

Can anyone kindly explain me why a single edit to teacher (Q37226) (teacher) is shown 173 times in my fr Wikipedia watchlist? Basically, every teacher in my Wikipedia watchlist shows the edit, which I think should not be the desired behaviour, as it clutters the Wikipedia watchlist (there are plenty of these, all the time). Am I doing something wrong? (I don't think I did anything specific for this) I've been directed to phab:T90435 ([Epic] Wikidata watchlist improvements (client)) but I'm not sure it covers this point.

What I see in my Wikipedia watchlist (click to expand)
  • (diff | hist) . . D Abraham Adolf Fraenkel (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Adolf Butenandt (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Adolph Wilhelm Hermann Kolbe (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Adolphe Hatzfeld (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Akira Suzuki (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Albert Wigand (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Albrecht Kossel (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Alexandre Schmemann (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Alfred Grosser (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Alfred Jeanroy (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Alfred Läpple (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Alfred Wegener (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D André Vésale (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Andréas Lovérdos (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Andréas Papandréou (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D António Marinho e Pinto (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Aurélien Barrau (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Burrhus Frederic Skinner (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Carl Friedrich Wilhelm Claus (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Catherine Lalumière (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Cecil Roth (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Christiane Nüsslein-Volhard (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Christine Brooke-Rose (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Christine Buci-Glucksmann (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Christopher Pissarides (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Christos Yannaras (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Chrístos Tsoúntas (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Claude Chabauty (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Conrad Moench (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Constantin Carathéodory (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Cornelius Castoriadis (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Cristian Preda (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Daniel Cohn-Bendit (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Dimitri Kitsikis (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Dominique Bucchini (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Donald Nicol (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Eduardo Lourenço (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Eftýchios Bitsákis (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Emil Adolf von Behring (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Emmanuel Kriaras (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Emmanuel de Pastoret (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Erich Auerbach (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Ernst Nolte (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Ernst Otto Beckmann (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Ernst Robert Curtius (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Euclide Tsakalotos (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Eugen Korschelt (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Evángelos Venizélos (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D Ferdinand Braun (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)
  • (diff | hist) . . D François Jeanneau (Q37226) ; 16:18 . . Alain Schneider (discuter | contributions) (?Alias [fr] défini(s) : enseignante, professeure, professeur, pédagogue, éducateur, éducatrice, formateur, formatrive)

Place Clichy (talk) 17:39, 29 February 2016 (UTC)

All those articles have a link to teacher (Q37226) (probably through the infobox). So if you want to see Wikidata edits in your watchlist, you'll see 1 edit on all articles that are on your watchlist and have a link to the edited item. You can turn off showing Wikidata edits in your settings on frwiki. Mbch331 (talk) 17:56, 29 February 2016 (UTC)
@Mbch331: I do not wish to turn off showing all Wikidata edits, I only wish to see Wikidata edits on items on my watchlist, not items linked to them. For instance, if i have fr:Abraham Adolf Fraenkel in my frwiki watchlist, I would understand to see edits on Abraham Fraenkel (Q61043), but I should really not see edits labelled Abraham Adolf Fraenkel (Q37226). Place Clichy (talk) 10:19, 1 March 2016 (UTC)
@Lea Lacroix (WMDE): Can you say anything about this? Mbch331 (talk) 11:23, 1 March 2016 (UTC)
I'll look into it. How often is this happening for people? Is it super annoying because it happens all the time? Or is this a rather exceptional case? --Lydia Pintscher (WMDE) (talk) 12:01, 1 March 2016 (UTC)
The problem is that there isn't really a way to know that it wasn't the label of "Étienne-Louis Boullée" that was changed to "enseignante", but the label of Q37226 which is "teacher". This doesn't exactly described the same sample above, but it is similar to something I found odd at Meta when viewing changes at Wikidata from there. A possible solution could be:
  • to move the QID to the space after the name of the user and include the current label of the item.
  • Maybe there should also be an indication of the link to that item (is it the one of the article or a linked one).
  • For linked items, a consolidated view could be preferable (1 diff not 173).
This at least from my POV. Maybe Place Clichy wants to suggest something else.
--- Jura 12:21, 1 March 2016 (UTC)
@Lydia Pintscher (WMDE): This is happening all the time, I can't remember when it started. My frwiki watchlist has 13 465 items on it, and I'm having a real hard time reading it due to such cluttering from WD edits appearing dozens of times, for items I'm not even following. I understand that changes (especially on labels) on "property" items can affect the Wikipedia article and are therefore legitimate on the WP watchlist, but there seem to be several things worth a cleanup :
  1. As I wrote here, showing changes to a mother class of a property are imho overkill. Whenever someone edits human (Q5) or person (Q215627), it's mayhem.
  2. The language issue has been reported many times before (only show edits to labels/records in the local Wikipedia language)
  3. Consolidating these 173 lines on 1 line (for 1 diff) would be a great improvement
  4. The label issue mentioned by Jura1 would also be an improvement but I would give it a lower priority and I can't tell the best way to display it: in my example above, the watchlist shows Abraham Adolf Fraenkel (Q37226) which is misleading. fr:Abraham Adolf Fraenkel (the article in my frwiki watchlist) is linked to Abraham Fraenkel (Q61043), and this item links to teacher (Q37226), which is where the edit took place.
Again, I can't tell if (especially point 1) is a bug, the expected behaviour, or something already listed in the watchlist integration improvement! Place Clichy (talk) 14:50, 1 March 2016 (UTC)
I understand, that this isn't solution to the problem itself, but you may try using this gadget, that allows filtering. Very nice in such cases. If you're interested, then copy content from this page to your common.js page. --Edgars2007 (talk) 15:02, 1 March 2016 (UTC)
Thanks, folks. That is useful. --Lydia Pintscher (WMDE) (talk) 15:53, 1 March 2016 (UTC)

The Signpost - two items need merge please

Can someone help with the merge of these two pages, please?

Thank you very much,

Cirt (talk) 09:35, 1 March 2016 (UTC)

Help:Merge
--- Jura 09:46, 1 March 2016 (UTC)
  Done--Ymblanter (talk) 12:21, 1 March 2016 (UTC)

Postage stamps

Hoi, how do I indicate that someone has been honoured with a stamp? For instance people on the great Americans series? Thanks, GerardM (talk) 07:37, 28 February 2016 (UTC)

I guess you could use a property like present in work (P1441), however this one is restricted to fictional entities. So either expand the scope of that one to include all kind of entities, or propose a new property.--Micru (talk) 10:06, 28 February 2016 (UTC)
The reverse. Create an item for the stamp, and use "depicts (P180)". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:29, 28 February 2016 (UTC)
depicted by (P1299) -- LaddΩ chat ;) 15:14, 28 February 2016 (UTC)
Good grief. How did that ever get created? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:22, 29 February 2016 (UTC)
Obviously the stamp depicts.. How to name such stamps ? Thanks, GerardM (talk) 19:34, 28 February 2016 (UTC)
Something like "1980 17p Moorhen" (year, value, subject). Possibly also prefixed by country. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:00, 1 March 2016 (UTC)

SPARQL report for item constraints

I'd like to add SPARQL report link to Template:Constraint:Item along with current WDQ link. It would look like this: User:Laboramus/Test/P2549 (compare with Property_talk:P2549). Basically the change would be adding one "SPARQL" link where it says "Report:". Are there any objections to this? Does anybody see any problems with the concept or implementation? --Laboramus (talk) 02:02, 1 March 2016 (UTC)

  • Looks good. As WQS doesn't display labels by default, maybe a column for these could be added too. BTW even the Autolist link could be changed to use SPARQL as source.
Eventually we might want to support exceptions= and items= as well.
--- Jura 09:20, 1 March 2016 (UTC)
Looks good to me, too. BTW, related suggestion. Maybe we can add this kind of reports for (some) constraint templates? The eqiuvalent SPARQL should be this one. Yes, it would fit better at Property documentation template, but it hasn't the needed functionality yet. --Edgars2007 (talk) 12:42, 1 March 2016 (UTC)
Yes, I think eventually we'll get there too. --Laboramus (talk) 07:34, 2 March 2016 (UTC)

Wikidata metalevel mark (Q22985125)

22985125 I want to see if someone has any explanation for this item. @Abc82: you seem to be starting something with this, what's your point ? author  TomT0m / talk page 11:42, 2 March 2016 (UTC) @Abc82: author  TomT0m / talk page 11:56, 2 March 2016 (UTC)

Everything is probably going to "break" for 10 or 15 minutes later this month

This is early notice for everyone, and a request to share the news:

The Ops team is planning a major change to the servers, (very) tenatively scheduled for Tuesday, 22 March 2016. One probable result is that when this happens, all wikis will be in read-only mode for a short time, likely less than 15 minutes for all editors. You will be able to read pages, but not edit them. "All wikis" means all of the WMF wikis, including Wikidata, Meta, Commons, the Wikipedias, and all the sister projects. It may affect some related sites, such as mw:Wikimedia Labs (including the Tool Labs). There will also be no non-emergency updates to MediaWiki software around that time.

Many details are still being sorted out. I am asking you to please share the word with your friends and fellow contributors now. This will be mentioned in m:Tech/News (subscribe now! ;-) and through all the other usual channels for Ops, but 99% of contributors don't follow those pages. If you are active in other projects or speak other languages, then please share the news with your fellow contributors at other projects, so that whenever it happens, most people will know that everything should be back online in 10 or 15 minutes.

Thanks, Whatamidoing (WMF) (talk) 16:37, 2 March 2016 (UTC)

identifier datatype is available and conversion is in progress

Hey folks :)

We have enabled the new datatype for identifiers now. New properties will start showing up at Special:ListProperties/external-id. As announced we will also start converting existing properties with datatype string that should be external identifiers. We will start with the first eleven from User:Addshore/Identifiers/0. We will do more over the coming days. There are still a lot of properties to go through so please help with that so we can get this over with quickly to cause minimal disruption to 3rd parties relying on our data. You will also see the identifiers that are already converted moving to a new section at the bottom of the item pages. This might need a purge still though to clear the cache.

Cheers --Lydia Pintscher (WMDE) (talk) 18:09, 17 February 2016 (UTC)

For the move to happen on items, it seems that the statement with the property needs to be edited. At least purging by users doesn't seem to be sufficient. The RDF version doesn't seem to include the URL either.
--- Jura 18:34, 17 February 2016 (UTC)
Thanks. Will look into it. RDF isn't implemented yet but coming soon. --Lydia Pintscher (WMDE) (talk) 18:36, 17 February 2016 (UTC)
I'm currently mass purging all items affected by the initial round of property conversions. If no further problems occur, VIAF ID (P214) will get done early tomorrow, other properties will follow. Cheers, Hoo man (talk) 20:17, 17 February 2016 (UTC)
Works now at Opuntia ficus-indica (Q144412) (Tropicos ID (P960)) --Succu (talk) 20:20, 17 February 2016 (UTC)
Looks good.
--- Jura 09:16, 18 February 2016 (UTC)

Out of interest, what is the plan if a site has different URLs for humans and for linked data JSONs? Which URL would a SPARQL search return? Jheald (talk) 01:31, 18 February 2016 (UTC)

We will provide both based on the different statements on the property. --Lydia Pintscher (WMDE) (talk) 11:07, 18 February 2016 (UTC)

Is there a way to enable formatter URLs also for other property types? (e.g. monolingual text or strings)
--- Jura 18:01, 18 February 2016 (UTC)

I think so, yes. --Lydia Pintscher (WMDE) (talk) 19:29, 18 February 2016 (UTC)
Do you think this could be roled out at the same time as the RDF part?
--- Jura 07:14, 19 February 2016 (UTC)
I'll check. --Lydia Pintscher (WMDE) (talk) 09:30, 19 February 2016 (UTC)
So technically it is a bit of a pita but if absolutely needed we can make it work it seems. Do you have some cases where it'd be needed? --Lydia Pintscher (WMDE) (talk) 11:55, 2 March 2016 (UTC)

Great! On User:Multichill/Paintings creator no authority control I have a list of properties that could probably be converted right away. Did anyone already try to do an intersection on Category:Properties with string-datatype & Category:Properties with unique value constraints & Category:Properties with single value constraints? That should return a bunch of obvious candidates. Multichill (talk) 13:52, 19 February 2016 (UTC)

I don't think that has been done yet. Good idea. --Lydia Pintscher (WMDE) (talk) 15:59, 19 February 2016 (UTC)
Hi Multichill - I was thinking of doing something like that, however, deciding what the actual criteria should be has been a little frustrating. Could you review the discussion at User:Addshore/Identifiers and comment there on what you think we ought to do? A criterion as you suggest makes some sense, though I'm not sure on the necessity of Category:Properties with single value constraints. We did go through pretty much every property in Category:Properties with string-datatype one way or another. I have some analysis on usage patterns here. ArthurPSmith (talk) 18:51, 19 February 2016 (UTC)

Is there a way of marking possible properties to migrate to this datatype? p:P717 might be one to consider changing. Thanks. Mike Peel (talk) 22:17, 25 February 2016 (UTC)

Railways Archive event ID (P2478) can also be converted (I would have proposed it as an identifier data type had I known that was imminent). Thryduulf (talk: local | en.wp | en.wikt) 01:28, 26 February 2016 (UTC)

Hi, is ISNI to be converted soon ? I was working on some items and the only identifier left unconverted was ISNI. Thanks for the migration. It really looks neat now :) --Hsarrazin (talk) 00:05, 3 March 2016 (UTC)

Thanks :) We're continuing to convert things as they are moved to "good to convert" at User:Addshore/Identifiers/0, User:Addshore/Identifiers/1 and User:Addshore/Identifiers/2. --Lydia Pintscher (WMDE) (talk) 08:09, 3 March 2016 (UTC)
  • Unless we want to leave the linking to the gadget, I don't think we can convert this one for now. Some development is still needed. Not sure when this will happen.--- Jura 08:15, 3 March 2016 (UTC)


I have a question which I also has added to WD:DEV: What will be result of this conversion when there is more than one formatter URL (P1630) for a property. P:P2182 has three, depending on your preferred language. -- Innocent bystander (talk) 08:22, 3 March 2016 (UTC)

In this case it should take the preferred one. We can look at making the link depend on the user's language but that is definitely not trivial and I don't know yet when we can do that. --Lydia Pintscher (WMDE) (talk) 08:26, 3 March 2016 (UTC)
I do not care what happens here in the GUI at Wikidata, that is of minor importance. It is what will be transported to the Clients that bother me. svwiki as a project has a POV for its own language. And I guess fiwiki has the same kind of POV, but for another language. -- Innocent bystander (talk) 08:30, 3 March 2016 (UTC)
There is https://phabricator.wikimedia.org/T107407 about this.
--- Jura 08:44, 3 March 2016 (UTC)

List of data licenses for different government's data

Hi all

Is anyone aware of a list of the different data licenses governments use? I'm thinking perhaps there may be something similar to the Freedom of Panorama page on Commons.

Thanks

John Cummings (talk) 14:23, 1 March 2016 (UTC)

just a note this is something I've run into in my work (not that I've seen a good solution) - we have our own rather ad hoc list of governments with "Crown" copyright rules, the US government employee rules, etc. ArthurPSmith (talk)
@ArthurPSmith: thanks, do you think its worth starting a page with this information on? I can see it being useful for people wanting to import data from different sources. --John Cummings (talk) 15:40, 2 March 2016 (UTC)
In Sweden (Q34) You have to look at each document! -- Innocent bystander (talk) 15:45, 2 March 2016 (UTC)
Would it be helpful to create a page on Wikidata called something like 'Government data licensing' that listed all the licenses used by different governments, with the aim of highlighting data that can be imported into Wikidata? I feel like this would be very useful but I'm a bit stuck with the name. John Cummings (talk) 16:36, 3 March 2016 (UTC)
It would be nice, but I am afraid that the list will be very short?
United States: Free in many cases.
Europe: No, with a few exceptions!
Rest of the World: Probably not!
-- Innocent bystander (talk) 17:22, 3 March 2016 (UTC)

Scripts loaded in common.js not working in Chrome

When using Wikidata in Chrome I'm missing several functions. These are all loaded from scripts in my common.js. When I check my browserconsole I see the following error: Refused to execute script from 'https://www.wikidata.org/w/index.php?title=User:Magnus_Manske/suggestor.js&action=raw&ctype=js' because its MIME type ('text/x-wiki') is not executable, and strict MIME type checking is enabled. (for each non-working script 1 line). I don't know how to make them work under Chrome. I don't have FireFox present on my work PC. Does anyone know how to solve this? Mbch331 (talk) 22:39, 1 March 2016 (UTC)

Solved it myself. The solution is to change the ctype part of the url to ctype=text/javascript, then it does work. Then the content type is set to text/javascript instead of text/x-wiki and the scripts are executed. Mbch331 (talk) 20:37, 2 March 2016 (UTC)

Wiki mark-up in string-type properties

Several users in frwiki, and apparently other wikis as well are pushing for the use of Wiki-markup in Wikidata, like image caption : [[John Doe]] and [[John Smith]] at the {{numeral|2}} edition of the competition.

That makes the statement unreadable anywhere but in a particular Wiki and on the whole I think it is really bad idea. But still, it would be useful if string-type properties supported more formatting options, like italics, superscripts, line breaks, and even, for some properties, internal links. Should we have a more explicit policy about this. Other thoughts ? @Lydia Pintscher (WMDE), Molarus:.--Zolo (talk) 07:09, 2 March 2016 (UTC)

I am very much against introducing formatting options in the string datatype as is because this defeats the whole point of what we are doing here. We can think about introducing other datatypes for a rare number of cases where it does make sense to have formatting but I doubt there will be significant ones. Can you say what the usecases are that they want to use formatting for in strings? Image captions will be handled as we make progress on Commons support. Any other cases? --Lydia Pintscher (WMDE) (talk) 11:44, 2 March 2016 (UTC)
Italics, subscript, superscripts, and line breaks would have relatively wide use for inscription (P1684) and quotation (P1683), and, I think, more occasional use for other monolingual text properties. Actually, it would also sometimes be useful in for labels as well (eg foreign loanwords should usually be italicized in French). --Zolo (talk) 12:36, 2 March 2016 (UTC)
It seems to me it would be reasonable to introduce another string-based datatype (like the mathematical formula type) that is a "formatted string" or "interpreted string" for certain properties. Many programming languages have two different string types, one for raw text (single quoted strings in perl, ruby, etc.) and one for text allowing run-time substitutions (double-quoted strings). ArthurPSmith (talk) 14:37, 2 March 2016 (UTC)
I have inserted this code in a lua module: string.gsub(s, "{{(%w+)}}", function (n) return frame:expandTemplate{ title = n } end) to make templates in image captions work. And now I am adding code for dealing with references, but the formatting will be in the code, not in Wikidata values. --Molarus 16:30, 2 March 2016 (UTC)
@Molarus: the {{Templates}} markup is in Wikidata, which is only usable in (one) Wikipedia, and should be avoided. --Zolo (talk) 17:56, 2 March 2016 (UTC)
It is no problem, because the templates are used in a monolingual text, which is seen only in the wiki that inserted the template into the monolingual text. And if no template is inserted into the image caption, the code does no harm. A problem might arise someday if a value is used in different ways, but the solution is here to have similar code or to filter formatting code out of the value. --Molarus 18:16, 2 March 2016 (UTC)
which is seen only in the wiki that inserted the template: I don't agree with you. For example there are 10 Wikimedia wikis that would display an image description in German (de.wikipedia, de.wikisource, de.wikiquote, de.wikibooks, de.wikinews, de.wikiversity, de.wikivoyage, wikispecies, commons) and I am not counting here usages outside Wikimedia worlds. Tpt (talk) 21:27, 2 March 2016 (UTC)
OK, there are some wikis that could show that info too. If they have the same template as wikipedia, they will see the same image caption. Where is the problem? A problem could arise if I´m using a value in different ways, but I don´t see that as a possibility for captions or text in general. Another possible problem could arise if there are templates with the same name but different functions. By and large, I see the possibility of small problems in the future if formatting code is inserted into wikidata text (!) values. On the other hand, the users get the same features they are used to in wikipedia. I think, we should try to make editors happy, while we pay for that with more work. --Molarus 06:41, 3 March 2016 (UTC)
The problem is that nobody is expected to treat the property as formatted text, nor is there any guarantee that if they do, that they have a template with the same name and if they do have one, that it will work the same way. Wikidata is designed to store data which can be reused by everyone, it's not designed to store data tailored to one specific project - that can be stored locally. Ultimately, I don't think Wikidata is even the right place for image captions, it would make more sense as structured data on Commons (which I think was what Lydia was also implying). - Nikki (talk) 09:34, 3 March 2016 (UTC)
There are even more than that, as smaller wikis may want to fall back to German if the info isn't available in their language yet (there are 15 other languages shown as falling back to German on commons:File:MediaWiki fallback chains.svg...). - Nikki (talk) 09:34, 3 March 2016 (UTC)
It was actual meant for frWP to show something like 1er. I wouldn´t mind to have "sup" instead of the template, but I don´t like to "correct" other peoples text. Another possible use case are right-to-left names in captions, e.g. names of persons. There is a need for formatted text, I´m sure. By the way, pattern matching is central to lua and to this problem. It´s magic! I would never use the parser function #property --Molarus 11:54, 3 March 2016 (UTC)
  Strong support to a formatted text datatype (with visual editor?). About allowing templates, I am not so sure, at least not at the beginning. --Micru (talk) 10:55, 3 March 2016 (UTC)
In taxonomy, titles of books and papers that include names in italics are commonplace, so markup would be useful. - Brya (talk) 11:38, 3 March 2016 (UTC)
A new datatype for such values would be sensible. It could either be qualified by the home wiki (like we currently qualify langauges of works) or the markup could be expressed like [[:fr:John Doe]], with [[John Doe]] forms causing an error message. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:15, 3 March 2016 (UTC)

How to cite Turkish census data?

I'd like to add the population of all 81 Turkish provinces to Wikidata. The data is gathered as part of the Address-Based Population Registration System (Q6042918). There is an interactive query tool on biruni.tuik.gov.tr. How to cite the population count properly? Would stated in (P248)/described by source (P1343) be sufficient or do I need to include retrieved (P813) or title (P1476)? --1-Byte (talk) 10:08, 2 March 2016 (UTC)

@1-Byte: apparently query tool is an interface for the report generated by Address-Based Population Registration System (Q6042918) , so you should source the report with stated in (P248) as Help:Sources suggests (by creating item for the report). -- Hakan·IST 10:44, 2 March 2016 (UTC)
@1-Byte: Use Help:Sources but follow the database section: Address-Based Population Registration System (Q6042918) is a database, not a legal text. Snipre (talk) 13:02, 2 March 2016 (UTC)
@Snipre: It is a live updating database, note that you'd be sourcing the 'report' generated by Query Tool at the time of import. -- Hakan·IST 13:20, 2 March 2016 (UTC)
@HakanIST: Is the report generated by the system an official document ? With some authors, ISBN, DOI or other identifiers ? By report we mean a document which is not generated by an automatic system from a continuously updating database. Snipre (talk) 18:54, 2 March 2016 (UTC)
@Snipre: I found an official press release which includes some tables about provinces. These documents are now included as 2015 Turkish census (Q22987520). The data about districts is not included though. --1-Byte (talk) 21:40, 2 March 2016 (UTC)
No I think you don't understand the difference: a document can't be generated by an automatic system and what you have for 2015 Turkish census (Q22987520) is press release not an official document providing all the data at the end of 2015. As example have a look at the press release of census data for Switzerland there but this can be used to source only what is in hte press release and not other data from the database. Your press relaese just mention some data of the database at a specific date. It can be sued as source for the data you have on the press realease but not for the missing data from the text.
A document is something like that or like that which can be retrieved by everyone in the wolrd just by mentioning its title or one of its identifiers like ISBN. It's something which contains all the data you want to cite.
Why don't you use the data structure for databases in source section ? Snipre (talk) 22:53, 2 March 2016 (UTC)
────────────────────────────────────────────────────────────────────────────────────────────────────@Snipre: I see, this is definitly not optimal. However there is an option to download the tables at the end of the release (not dynamically generated). Would that be OK (at least for the provinces)? Regarding the source section: According to Help:Sources#Databases databases without an identifier should be treated like webpages. I tried to follow the advice in this section. Would you mind helping me by fixing the references in Adana Province (Q40549)? --1-Byte (talk) 08:50, 3 March 2016 (UTC)

Inspire Campaign: Making our content more meaningful

 

The second Inspire Campaign has launched to encourage and support new ideas focusing on content review and curation in Wikimedia projects. Wikimedia volunteers collaboratively manage vast repositories of knowledge in our projects. What ideas do you have to manage that knowledge to make it more meaningful and accessible? We invite all Wikimedians to participate and submit ideas, so please get involved today! The campaign runs until March 28th.

All proposals are welcome - research projects, technical solutions, community organizing and outreach initiatives, or something completely new! Funding is available from the Wikimedia Foundation for projects that need financial support. Constructive, positive feedback on ideas is appreciated, and collaboration is encouraged - your skills and experience may help bring someone else’s project to life. Join us at the Inspire Campaign and help your project better represent the world’s knowledge! I JethroBT (WMF) 19:55, 2 March 2016 (UTC)

Instance vs Subclass

For a project I'm working on I created a sort of browser that shows sub classes and instances of any item. What I've noticed is how often people confuse the two. I'm about to work extensively on getting these right wherever I see an issue.

However before I start I would like to make sure I'm right and the community agrees with me.

For example a Vegetable is defined as being the part of a plant that you eat. There are pages and pages of instances of vegetable for example beet (Q165437) or asparagus (Q28367) which are all defined as instances of vegetable, but in my mind they should all be subclasses.

This is better summed up as if it were code. If Vegetable had a method called rot obviously it would not be the class Vegetable that rots but an instance of a vegetable that rots. It's also evident that it's not the class tomato that rots but an physical instance of tomato. Hence tomato being an instance of vegetable does not make sense, it should be a subclass.

Does everyone agree with me? Any ideas? Thoughts? I'm also worried that some people might not understand the difference between instance and subclass and might change it back. By Thewormsterror.

AFAIK, there is a loose interpretation of instance here, we don't use it only for real life unique instances, but it is more like a tree of classes where the last element is an instance. It is improbable that we will have real instances of beetroot, so it is not a big deal. Perhaps it would be more useful if combined with quantity (P1114) but it is quite unlikely that we could source the number of beetroot in the world, other than with weight estimates.
In the end I think it is a question of practicality, do you have some practical application in mind where it would make a difference to consider beetroot a subclass of vegetable instead of an instance?--Micru (talk) 10:46, 3 March 2016 (UTC)
I don't agree with this and I push help:Classification. This loose interpretation is not useful for any purpose and we totally lose the interest of having two properties. If we want to take that approach, w should just create a new property is a and add it to the items that is a superproperty of both instance of and subclass of. Plus the dichotomy instance of (P31) / subclass of (P279) has been used to solve problems about for example differenciate the type of administrative division in some project out here and the divisions themselves (paris / french city / french administration division) which are a mess without using the dichotomy. The most important feature is that subclass is transitive, whereas instance of is not : an instance A of an instance B of some class C is not an instance of C, so that Paris is not a type of french administrative division. author  TomT0m / talk page 12:42, 3 March 2016 (UTC)
I just read your Classification document. I agree with you almost completely TomT0m, I just have some hesitations about things being both instances and subclasses, I don't disagree I just have to think about it more. Being quite knew to wikidata though, I don't know how things work here. As a new user can I just go and change things as I see fit and assume no moderator will flag me or change them back? I know this matters little but I have a lot of qualifications when it comes to making models, so I know what I am doing. I plan on using wikidata a lot in the future for various projects, and the accuracy of the data is important to me. Thewormsterror (talk) 18:07, 3 March 2016 (UTC)
@Thewormsterror: I'd just like to point out the original example above with "Vegetable" seems wrong to me - either the instance of (P31) statements should be changed to subclass of (P279) or vegetable (Q11004) should be replaced in the statements by something that has the form of a metaclass - "type of vegetable" or something like that. All that said, the bio hierarchy is a bit complicated and there are some people with strong opinions there so I would consult a bit (eg. here or talk pages or an RFC) before trying any major changes. ArthurPSmith (talk) 21:02, 3 March 2016 (UTC)
@ArthurPSmith: It's not really bio hierarchy, because a vegetable is an edible plant or a part of an edible plant. It goes across taxons. To me it's obvious that an asparagus should be a subclass of vegetable. I'll wait a week and unless I get any strong objections in this thread I will change it. Thewormsterror (talk) 21:37, 3 March 2016 (UTC)

Admin notice

Please undo the merger of Philip Francis Little (Q7183561) (en:Philip Francis Little, Newfoundland Premier) and Philip Little (Q21794051) (de:Philip Little, painter).--Kopiersperre (talk) 20:00, 3 March 2016 (UTC)

It can be done in the edit history of the items. Please advise if there is a problem with that.
--- Jura 20:10, 3 March 2016 (UTC)

Q23000000

It's created! Just let you know to who are interest about milestones!--DangSunM (talk) 02:14, 4 March 2016 (UTC)

Importing the day of birth into a local wiki

Is it possible to import just the day of birth into a local wiki instead of whole date with Property:P569? --Janezdrilc (talk) 19:17, 28 February 2016 (UTC)

@Janezdrilc: Do you mean determine of the name of the day of the week of the month of the year, or do you mean that you just want the dd value. If the later, pull the whole date value, and parse it to give the day value, see mw:Help:Extension:ParserFunctions.  — billinghurst sDrewth 22:29, 28 February 2016 (UTC)

That's really great. That is what I need: {{#time:d. xg|{{#property:p569}}|sl}}. It works. Thank you. --Janezdrilc (talk) 12:12, 29 February 2016 (UTC)

@billinghurst: There seem to be a problem. Please try to copy this code...

: {{#if: {{#property:p569}} | [[:Kategorija:Rojeni {{#time:j. xg|{{#property:p569}}|sl}}]] }}
: {{#if: {{#property:p569}} | [[:Kategorija:Rojeni leta {{#time:Y|{{#property:p569}}|sl}}]] }}
: {{#if: {{#property:p570}} | [[:Kategorija:Umrli {{#time:j. xg|{{#property:p570}}|sl}}]] }}
: {{#if: {{#property:p570}} | [[:Kategorija:Umrli leta {{#time:Y|{{#property:p570}}|sl}}]] }}

... into next articles: s:sl:Vinko Gaberc, s:sl:Janko Kessler, s:sl:Vatroslav Holz, s:sl:Janez Parapat. There are occuring combinations of working/nonworking functions. --Janezdrilc (talk) 21:16, 1 March 2016 (UTC)

I had tried some as well. Interestingly in some cases it breaks when using "from=" in #property.
--- Jura 21:18, 1 March 2016 (UTC)

@Jura1: I was thinking if there is some MediaWiki page or Module, css or js file or something else to update on Slovene Wikisource? Could be something of that simply missing? --Janezdrilc (talk) 22:43, 1 March 2016 (UTC)

I couldn't reproduce what I just wrote. {{#time:d. xg|{{#property:p569}}|sl}} and {{#time:d. xg|{{#property:p569|from=Q76}}|sl}} work on Q76 at enwiki, but not on slwiki. I think it's probably something to fix for the devs. As they are working on localized dates, maybe they could look into this as well or are already doing it. @Lydia Pintscher (WMDE):.
--- Jura 23:00, 1 March 2016 (UTC)
Noted. Thanks! :) --Lydia Pintscher (WMDE) (talk) 11:45, 2 March 2016 (UTC)
{{#property:P569}} produces: 13 januar 1886, but the time function wants {{#time: xg | 13 january 1886 }} with a "y". I think Wikidata returns the local name, but the time function wants the english name. A #switch should do the translation. --Molarus 02:06, 2 March 2016 (UTC)

@Molarus: I've been testing some dates on sl.wiki. If the date include following months: apr, sep, nov, dec, the translation is done correctly. If the date include remaining months: jan, feb, mar, may, jun, jul, aug, oct, the translation is done incorrectly. This applies in all articles. Some other wikis also have problems with translation, including de, cs, pl, ru, sk, uk. --Janezdrilc (talk) 11:28, 4 March 2016 (UTC)

I think Molarus is right. It seems it needs to be done with LUA. Module:Wikidata has {{#invoke:Wikidata|formatStatementsE|item=Q76|property=p569|displayformat=raw}} (+1961-08-04T00:00:00Z).
With {{#time:d. xg|{{#invoke:Wikidata|formatStatementsE|item=Q76|property=p569|displayformat=raw}}|sl}}, it works here (04. avgusta), but requires the same version of Module:Wikidata to work elsewhere. Ideally the entire thing is done in one LUA function. It would also need to check if the date has day-precision.
--- Jura 11:27, 4 March 2016 (UTC)

Revert Q6581072

Would someone revert the edit of Pigsonthewing on that item? Apparently, they don't want to discuss what they are doing and per Property_talk:P2559 we are not ready yet.
--- Jura 02:57, 4 March 2016 (UTC)

I have undone the edit, as the instructions should remain until the usage notes statement is supported (which I hope is being worked on as a matter of priority) but I cannot see any evidence that you (Jura1) or anyone else has actually tried to discuss this with Pigsonthewing so saying he doesn't want to discuss is not necessarily accurate. Thryduulf (talk: local | en.wp | en.wikt) 13:41, 4 March 2016 (UTC)
Thanks. Someone removed what might be the evidence you are looking for.
--- Jura 13:46, 4 March 2016 (UTC)

Help Commons categories

Where is the rght location in a item like Q880229 for the Commons category: Identifiers or statements? And for links to Wiktionary pages. Thanks--Pierpao (talk) 13:45, 4 March 2016 (UTC)

For Commons, you can use either (click "add"), it gets displayed depending on how the property is defined.
For Wiktionary pages, I don't think any properties are available yet.
--- Jura 13:53, 4 March 2016 (UTC)

Merge problem

I made a mistake and can't undo it. I've merged Q21041415 into Q7232681, thinking it was the same thing. The only page it existed of Q21041415 was this. I've press undo, but the wikidata item now only redirects. To clarify, one is a sports association of roller hockey, and other is of field hockey, I've no idea that there were two different federations.--Threeohsix (talk) 19:47, 4 March 2016 (UTC)

I undid the merge for you. Next time, go to the item that is a redirect (In this case: Portuguese Field Hockey Federation (Q21041415)). It will lead you to the page you merged it into (in this case Portuguese Roller Sports Federation (Q7232681)), then click on the link that points to the redirect page (Q21041415). You'll only see a link. Click on history and compare the last edit with 2 edits before. Than choose undo. Now the item is back into the state it was before you started the merge procedure. Mbch331 (talk) 20:55, 4 March 2016 (UTC)

Where is the Universal Language Selector

Hi everyone, I'd like to change my language preferences but I can't find the "Universal Language Selector" which is supposed to be present "at top of pages" as mentioned here.--Kimdime (talk) 23:04, 4 March 2016 (UTC)

 
In this example, it's the "English" link --YMS (talk) 23:16, 4 March 2016 (UTC)
Hi YMS, so apparently I was misled and asked the wrong question. I'm totally fine with the interface language, what I wanted was to change the label and description default language interface. Thanks to Help:Navigating Wikidata/User Options I managed to do it through the babel infoboxes, but I'd rather not have those boxes in my user page, I'd like to do it through the preferences but somehow I don't manage it that way (didn't find any "Internationalisation" section in my user profile).--Kimdime (talk) 23:37, 4 March 2016 (UTC)

Decoding P1172

Geokod (P1172) has Type:class= administrative territorial entity (Q56061) as a constraint. That is fine. But the code itself gives more information.

  • A code of type: NN00000 is a County.
  • A code of type: XXNN000 is a Hundred, Tingslag or Skeppslag.
  • A code of type: XXXXNN0 is a Municipality. Thereof: XXXXX80 is a City and XXXXX90 is a Market town.
  • A code of type: XXXXXXN is a (subclass of) Parish if N=1,2,3/is a Municipalsamhälle if N=4,5,6,7,8/is a "Market town who does not form a municipality" if N=9 (There are two exceptions for N=3)

Can this information be used here in any way? -- Innocent bystander (talk) 07:42, 3 March 2016 (UTC)

Not with the normal constraints, but if you know how to write SPARQL, you can use Template:Complex constraint. Mbch331 (talk) 11:14, 3 March 2016 (UTC)
Many of these have "instance of:former municipality in Sweden". It is technically correct, but "former municipality" has a number of subclasses, with more detailed information! -- Innocent bystander (talk) 12:59, 3 March 2016 (UTC)
I just discovered that administrative territorial entity (Q56061) maybe is a little to limiting for this property. WD do not regard Swedish parishes as an administrative entity. (They used to be, but are not anymore.) And unincorporated Market towns is more of a "populated place" than a administrative territorial entity (Q56061). I have changed the constraints now, but I do not want to have "subclass of any item in this list". I prefer "is instance of any item in this list". Is that possible? -- Innocent bystander (talk) 10:00, 4 March 2016 (UTC)
@Innocent bystander: From what I understand you just have to create an item "type of geokod populated place", mark
⟨ County ⟩ instance of (P31)   ⟨ type of geokod populated place ⟩
 ;
⟨ Hundred ⟩ instance of (P31)   ⟨ type of geokod populated place ⟩
 ; ...
then with sparql do something like
SELECT ?place WHERE {
    ?admintype wdt:P31 wd:Qtype_of_geokod_populated_place .
    ?place wdt:P31 ?admintype
}
Try it!
that will select any place that is an instance of anything that is an instance of the types of geokod
@TomT0m: SPARQL is far far beyond my knowledge. -- Innocent bystander (talk) 17:50, 5 March 2016 (UTC)

Metrolyrics

I wanted to propose a new property for lyrics at MetroLyrics (Q6824428), but then found, that it's kind of part of full work available at URL (P953), see example at Never Gonna Give You Up (Q57). So, what should I do? Propose the property or not do that and not waste everybody's tyme at proposal page? Theoretically, I'm fine with P953, but property for Metrolyrics would ease life, I think (it could then be used at Wikipedia templates etc.). --Edgars2007 (talk) 20:05, 4 March 2016 (UTC)

I think a specific property would be beneficial; even if others disagree, having the discussion is not a waste of time. Please go ahead. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:48, 4 March 2016 (UTC)
I also think a specific property would be useful. en:Template:MetroLyrics song has nearly 20,000 uses on enwiki alone. - Nikki (talk) 21:19, 4 March 2016 (UTC)
OK, made the proposal, will see, how things will go. By waste of time I meant something like this :) Yeah, actually it's one of most used ID templates at enwiki and probably the most used template, which don't have property here. --Edgars2007 (talk) 07:48, 5 March 2016 (UTC)

Merge problems

Can someone merge de:Kategorie:Hochschullehrer nach Staat (Q8975789) with en:Category:academics by nationality (Q7004720) ? 92.76.107.255 22:28, 4 March 2016 (UTC)

Seems to me that these shouldn't be merged, with a couple of language conflicts. To me it seems however that several links of Category:Academics by nationality (Q7004720) should be moved to Category:University teachers by country (Q8975789). But my limitations of language knowledge may mislead me here. Lymantria (talk) 17:54, 5 March 2016 (UTC)

Welcome notification

When a user creates a new account on Wikidata, they get a welcome notification. This has also a feature that we don't use: it links to an arbitrary page. Where do we want to point new users to? Help:Contents, Wikidata:Introduction, Wikidata:Tours...? Matěj Suchánek (talk) 09:15, 5 March 2016 (UTC)

I was bold and set Wikidata:Introduction as the target. Matěj Suchánek (talk) 17:23, 6 March 2016 (UTC)

How to use honorary doctorate (Q11415564)

Hi,

how do I specify the university of an awarded honorary doctorate? My case: Eva Persson was awarded an honorary doctorate at Linköping university in May 2011 source (in Swedish). Best regards Ambrosiani (talk) 11:59, 6 March 2016 (UTC)

Use of (P642) as qualifier. Sjoerd de Bruin (talk) 12:14, 6 March 2016 (UTC)
Thank you! Ambrosiani (talk) 12:36, 6 March 2016 (UTC)

Import WP article names and links

Hi there. We're starting m:Wikimedia CEE Spring 2016 soon and currently local chapters are preparing lists of vital articles we would love to see written in other languages. Here's a sample list for Poland, as you can see manually adding new entries is a major pain as one would have to check whether certain article exists in one of the 26 languages, then manually add the links, format the table…

So, I was wondering if this could be done by a template. The template would include all the formatting, all the user would have to do is to fill in the Wikidata Q code for all the links to load (or a – sign to load if there is no article), and then subst the template in place to form a new line in the table. How do I retrieve names and links to articles, and a minus sign if there is no article? Halibutt (talk) 01:41, 3 March 2016 (UTC)

@Halibutt:, is this template+module is suitable for you? --Voll (talk) 16:56, 5 March 2016 (UTC)
@Voll: Yup, that's precisely what I was trying to achieve. Thank you! Halibutt (talk) 21:28, 6 March 2016 (UTC)

Visualization and data contest on Commons Collaborative Economies

From the Commons Collaborative Economies Meeting to be held in Barcelona on 11, 12 and 13 March, we have launched a competition consisting of data visualization from the Open Directory Project of the P2P Value. The five best entries will share a prize of € 1,000.

One of the potentials may be crossing and loading data here, in Wikidata, so they can be reused by all Wikimedia projects. In fact, some reports of this meeting revolve around Wikipedia and its sister projects, and Amical Wikimedia (Catalan Chapter) is collaborating in it.

The link for registration is this, and the deadline is 5 March to 12 March. Come on! --Xavier Dengra (talk) 19:25, 6 March 2016 (UTC)

Register article

I recently met Mark Pesce (Q6769230), in Sydney (Q3130). As a result, he has written an article for The Register (Q1363733), Wikidata makes Wikipedia a database. Let the fun begin. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 04:32, 26 February 2016 (UTC)

Yeah, a good app for data from Wikidata will be the next task. --Molarus 17:23, 26 February 2016 (UTC)
Pity no one turned up to the workshop in Sydney. Funded by donors. Tony1 (talk) 06:42, 4 March 2016 (UTC)
I think you're mistaken. While in Sydney I ran one workshop - at the State Library, who hosted the event and, I understand, provided the venue for free. It was so well attended that extra chairs had to be brought in from adjacent rooms. I also had the meeting which resulted in the above-mentioned article being written. There was a social meetup, in a cafe, and it was indeed a pity that, despite the best efforts of WikimediaAU volunteers to set up and publicise the event, no-one other than me and the organisers turned up. Perhaps it's that you're thinking of. I'm not sure what the total expense for the social event was, but there was no venue hire and no pre-ordered catering, so it would be minimal, if any. Fortunately, every other event on the three-week tour - talks, workshops and social meets where I was available to answer Wikipedians questions about Wikidata and my GLAM work - was also well attended, and well received. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:26, 7 March 2016 (UTC)

What's the correct property for the building of an institution

I'm trying to figure out what is the right way to enter the information relating to the building which houses an institution. Ex .: the Museu Nacional d'Art de Catalunya (Q861252) is inside Palau Nacional (Q2341107) or Louvre Museum (Q19675) is in Louvre Palace (Q1075988). My questions are:

Thanks, --Amadalvarez (talk) 15:07, 26 February 2016 (UTC)

When Amadalvarez commented, one of the constraints needed fixing. This has been done since then.
--- Jura 12
  • 01, 4 March 2016 (UTC)

Thanks to everybody. I will use location (P276) and change when I found any wrong use of any other propertry, as in Museu Nacional d'Art de Catalunya (Q861252) or Louvre Museum (Q19675), for instance,. --Amadalvarez (talk) 09:00, 7 March 2016 (UTC)

Property proposals need attention

There are many proposals for property that have been sitting for months waiting for comments (favourable or otherwise) and/or attention from property creators. Transportation, Unsorted, Generic and Natural Science categories seem to me (subjectively) to be the slowest moving ones. If we are to move the project forwards we need to respond to proposals for properties when people make them - one of the best ways to demotivate people is to ignore them. Thryduulf (talk: local | en.wp | en.wikt) 19:30, 28 February 2016 (UTC)

The problem is that there are several people requesting properties without actually using them. Now we have to do maintenance on a lot of them before they were even used: Wikidata:Project_chat#Dormant_properties.
Don't hesitate to ask an administrator or a property creator to look into specific proposals, if you plan to use them. Maybe not for external identifiers, as we have not sorted these out yet.
--- Jura 17:51, 29 February 2016 (UTC)
Well if proposals were not left lingering for months before being created then more of them would be used quicker (but see my comments elsewhere for why I disagree that dormant properties are actually as big an issue as it is claimed). For example at WD:PP/EVENT there are proposals that I've just marked as ready that have been sat with unanimous support since November, and a property I proposed in January has attracted no comments favourable or otherwise. At Wikidata:Property proposal/Transportation there have been properties marked ready for almost two weeks with no action.
I've tried pinging people to follow up comments they left several months ago without any response. I've tried pinging various property creators (on and off wikidata) previously with no response. Thryduulf (talk: local | en.wp | en.wikt) 13:50, 7 March 2016 (UTC)

Wikipedia articles to Wikidata items

Anybody knows a tool for conversion of Wikipedia articles list

France
Russia

to Wikidata items list:

Q142
Q159

--Voll (talk) 17:52, 5 March 2016 (UTC)

Linked items. --Edgars2007 (talk) 17:54, 5 March 2016 (UTC)
Hmm, I don't understand how it works. I have found two-step path: create PagePile and use it in Autolist or your Linked items (better). --Voll (talk) 18:48, 5 March 2016 (UTC)
Input them with wikilinks: [[Francija]], [[Krievija]] (with marked lv.wikipedia). --Edgars2007 (talk) 18:59, 5 March 2016 (UTC)
Heh, it works. Thank you. --Voll (talk) 19:44, 5 March 2016 (UTC)
Another option is the Wikipedia Tools addon for Google Sheets. If your list is in a spreadsheet, then you can do something like this to get the QIDs:
A B
1 France =WIKIDATAQID("en:"&A1)
2 Russia =WIKIDATAQID("en:"&A2)
--Hardwigg (talk) 02:34, 7 March 2016 (UTC)
Looks interesting (the addon wholly, not just this example), thanks. Will play around. --Edgars2007 (talk) 07:31, 7 March 2016 (UTC)

Probably said it before, many times, but I don't know

A property such as P1449 nickname can take many statements, and each statement has a language. What I want to do is pick out the statements by language. E.g. just the English. How to do it? 194.75.238.182 10:29, 6 March 2016 (UTC)

What's your extraction method ? Sparql, lua, a wikitext infobox ? author  TomT0m / talk page 17:13, 6 March 2016 (UTC)
Just used regular template with
| nickname1 = {{{{{|safesubst:}}}#if:{{{{{|safesubst:}}}#property:P1449}}|{{replace|{{{{{|safesubst:}}}#property:P1449}}|,|{{repeat|<br />|{{#if:{{{lflf|}}}|{{{lflf|}}}|1}}}}}}}}
I manage it works like
| population_as_of = {{{{{|safesubst:}}}#if:{{{{{|safesubst:}}}#invoke:Wikidata|claim|P1082|qualifier=P585}} | {{{{{|safesubst:}}}#time:Y|{{{{{|safesubst:}}}#invoke:Wikidata|claim|P1082|qualifier=P585}}}} | ? }} {{lc:{{#invoke:Wikidata|getRawQualifierValue|P1082|P459|FETCH_WIKIDATA}}}}
| population_point_in_time = {{{{{|safesubst:}}}#if:{{{{{|safesubst:}}}#invoke:Wikidata|claim|P1082|qualifier=P585}} | {{{{{|safesubst:}}}#time:Y|{{{{{|safesubst:}}}#invoke:Wikidata|claim|P1082|qualifier=P585}}}} | ? }}

(but I don't know what the '?' does. 194.75.238.182 15:05, 7 March 2016 (UTC))

SPARQL: items without any statements

How do I request a list of items with SPARQL without any statements, but with a sitelink to a specific project? Sjoerd de Bruin (talk) 15:04, 28 February 2016 (UTC)

SELECT ?item ?wiki WHERE {
  ?wiki schema:about ?item .
  FILTER(SUBSTR(STR(?wiki), 1, 25) = 'https://cs.wikipedia.org/') .
  FILTER NOT EXISTS {
    ?item ?prop ?statement .
    ?statement wikibase:rank ?rank .
  } .
} LIMIT 1
Setting higher limit times out. Using nl.wikipedia.org (propably area of your interest) times out anyway. Matěj Suchánek (talk) 17:36, 28 February 2016 (UTC)
As an alternative, maybe Wikidata:Database_reports/without_claims_by_site can help you.
--- Jura 17:41, 28 February 2016 (UTC)
I've requested Dutch reports on the talk page. Sjoerd de Bruin (talk) 12:58, 1 March 2016 (UTC)
I had seen it and made the change. If Pasleim implements it, it will be there later today.
--- Jura 13:03, 1 March 2016 (UTC)
I know only Quarry way for it. --Voll (talk) 08:42, 29 February 2016 (UTC)
This one seems to work better:
SELECT ?item ?wiki WHERE {
  ?wiki schema:about ?item .
  FILTER(SUBSTR(STR(?wiki), 1, 25) = 'https://cs.wikipedia.org/') .
  FILTER NOT EXISTS {
    ?p a wikibase:Property .
    ?p wikibase:directClaim ?pdirect .
    ?item ?pdirect ?statement .
  } 
} LIMIT 10

--Laboramus (talk) 21:07, 6 March 2016 (UTC)

Times out for larger wiki's. Sjoerd de Bruin (talk) 21:18, 6 March 2016 (UTC)
That probably requires to phab:T129037 and phab:T129046 for comprehensive fix. --Laboramus (talk) 23:21, 7 March 2016 (UTC)

water hardness (Q192905)

May someone please separate Water hardness and Hard water? The latter is the opposite to soft water (Q10856440) and should get its own item.--Kopiersperre (talk) 22:17, 29 February 2016 (UTC)

  Done In hard water (Q22988272) Andreasm háblame / just talk to me 14:41, 2 March 2016 (UTC)
@Kopiersperre, Andreasmperu: Following this separation, the five articles moved to hard water (Q22988272) (en:Hard water, es:Agua dura, ro:Apă dură, simple:Hard water, ta:கடின நீர்) no longer have interwiki links to and from the 36 articles at water hardness (Q192905), whereas their articles cover exactly the same subject (for instance, en:Water hardness redirects to en:Hard water, and the term hardness is ubiquitous in the article, same for es:Dureza del agua and es:Agua dura). This goes against the principle of Wikidata phase 1 and hosting interwiki links on Wikidata, and it does not help navigation for the reader. In such a case, what do you think is the most appropriate action:
  • Merging near-identical Wikidata items so that effective interwiki links are provided to the reader?
  • Allowing redirections in site links? (that would be a big change on the spirit of the project)
  • Placing hard interwiki links in wikipedia articles? (that would be phase 1 backwards, and how could this be maintained)
  • Forgetting the idea of interwiki navigation altogether?
All opinions welcome. Place Clichy (talk) 12:40, 3 March 2016 (UTC)
@Place Clichy: Merging non-identical Wikidata items is not a solution because we want to build an ontology following Occam's razor (Q131012) here.
There is a workaround: Place any bullshit above the redirect, link it with Wikidata and delete the bullshit again.--Kopiersperre (talk) 14:11, 3 March 2016 (UTC)
@Andreasmperu
The current site link in water hardness (Q192905) for Chinese/Japanese/Korean/Vietnamese/ Hindi/Estonian/Persian/Hebrew/ Sinhalese/Punjabi Wikipedia are all about Hard Water not Water Hardness according to my understanding and with the help of machine translator for some language. Can you please help review and move those site links to appropriate wikidata item too?C933103 (talk) 20:48, 7 March 2016 (UTC)

Site links to Commons pages should point to pages within the same namespace and others should be moved to properties like:

I realize that there is still no consensus on how to deal with Commons sitelinks to Category pages, so untill that is resolved I would just copy those links to Commons category (P373) instead of moving them, but no pages on Commons which are in Creator or Institution namespaces are suppose to have interlanguage links, so those should be just moved. Can anybody with bot skills could help with this task? --Jarekt (talk) 20:27, 4 March 2016 (UTC)

@Jarekt: I agree with you. Maybe we can define the following rules. I can help with the bot work later on:
What do you think? If we have consensus on this I can fill a bot request and run the task. -- T.seppelt (talk) 19:19, 7 March 2016 (UTC)
@T.seppelt: I am not that familiar with use of instance of (P31) and how consistent it is. I assume instance of (P31) = Wikimedia category (Q4167836) test is a reliable way to identify Q-items associated with wikipedia categories. I do not know how one detect Q-items associated with wikipedia articles (I assume that in most cases instance of (P31)Wikimedia category (Q4167836), but there might be a better way). Last December User:Jheald had excellent report on how many wikidata pages link to Commons pages through sitelinks and properties. He called those "Wikidata articles" and "Wikidata categories", so let me stick with this nomenclature. Also at this point I would like to concentrate on ensuring that every commons sitelink can also be found in namespace-specific properties: Commons category (P373), Commons gallery (P935), Commons Creator page (P1472), Commons Institution page (P1612) of "Wikidata articles", and verifying that those properties point to the right namespaces on commons. Also no sitelinks should be pointing to Commons Creator and Institution namespaces.
  • Case #1: mostly Agree; however:
  • the sitelink should be removed from the item: I would not remove them at this stage, since we have over 250k of them. I would remove them after we know what will happen with Sidebar links on Commons.
  • the category should be added as Commons category (P373) claim in case the entity doesn't already has one Agree
  • the sitelink should be moved to the entity which is linked by using topic's main category (P910) in case the entity has one Agree, but I would not be creating new "Wikidata category" enteties that do not link to wikipedia articles.
In general I see this task as cleaning outlines as oppose to transforming the database. --Jarekt (talk) 21:15, 7 March 2016 (UTC)

Wikidata weekly summary #199

SPARQL problems list

I've started a page that tracks problematic SPARQL queries: mw:Wikidata query service/Problematic queries. Note that these are not every query that anyone has problems with or does not know how to implement (everybody is still welcome to ask about those here or in SPARQL queries page :) but those we know and can reasonably prove that they should work, but they do not (i.e. either time out or produce exceptions or wrong results). Which means it's either a bug or at least a missing feature in the engine.

Not every query that times out is a problem - there are many queries that are legitimately too hard for current beta and some that would be too hard for any engine (full scans of billion-entry database usually are not very fast :). But a query that we can reason that it should be easy but still is not working is something that needs work and should be tracked.

The page is not replacement for Phabricator and other bug tracking, so if you think you found a bug, still submit it to Phabricator, but I think it would be useful to collect all queries that we have trouble with in one place so we could discuss them and prioritize the work that needs to be done.

If you have such queries please add those to the discussion page and I'll triage them and submit to appropriate places. --Smalyshev (WMF) (talk) 23:11, 7 March 2016 (UTC)

Repeated article in sk-wp

In sv-wp there is a repeated article w:sk:DOSBox and w:sk:Dosbox. Can anyone merge them? --Micru (talk) 15:59, 2 March 2016 (UTC)

  Done Articles merged. JAn Dudík (talk) 07:12, 8 March 2016 (UTC)

Tool to batch-process properties

Is there any tool like Autolist or QuickStatements that can be used to scan and/or change properties on properties? There are no less than 118 properties where comment (DEPRECATED) (P2315) needs to be replaced with Wikidata usage instructions (P2559)... -- LaddΩ chat ;) 00:03, 8 March 2016 (UTC)

One possibility:
But there most probably is a better way of doing this. --Edgars2007 (talk) 07:45, 8 March 2016 (UTC)
Neither Autolist nor QuickStatements act on property items such as Property:P1144, only on main namespace items. That's the problem... -- LaddΩ chat ;) 15:09, 8 March 2016 (UTC)

Best practices for biographies

If I want to create a perfect biographical item am I supposed to add all properties relevant to biographies even if the info is unknown (e.g. signature (P109) of almost all biblical characters), obviously not important (e.g. hair color (P1884) of Robert A. Heinlein (Q123078)) or certainly being without value (Instagram username (P2003) and astronaut mission (P450) for anyone dead before 2010 or 1957 respectively), or do we aspire to add only possibly helpful properties? DGtal (talk) 07:15, 8 March 2016 (UTC)

Instagram username (P2003) on all items would be annoying. Sjoerd de Bruin (talk) 07:38, 8 March 2016 (UTC)
Yes, that wouldn't be a very good idea. We generally try to put only those properties, which value is known. Of course, there are some cases, when we do it differently. Novalue/some value/unknown value (from your post I assume, you know what those are) are for other cases (from preventing bots to put that property again and again, for example). --Edgars2007 (talk) 07:50, 8 March 2016 (UTC)
I think they are mostly useful for items where a value for the property would be likely or expected to exist and adding the statement provides extra information (for example, astronaut mission (P450): novalue on an astronaut would distinguish them from an astronaut whose missions are missing, date of death (P570): unknown would distinguish them from someone who is still alive) or where a source for the unknown or novalue can be provided (e.g. an interview where someone says they refuse to make an Instagram account there would be a source for a Instagram username (P2003): novalue statement, a book saying that a historical person's hair colour is unknown would be a source for a hair color (P1884): unknown statement). - Nikki (talk) 09:09, 8 March 2016 (UTC)

Islamic calligraphy (Q656831) and Islamic calligraphy (Q23011124)

Islamic calligraphy (Q656831) is the wikidata item for Arabic calligraphy for English Wikipedia whereas Islamic calligraphy (Q23011124) is the wikidata item for Islamic calligraphy. However the article for Islamic calligraphy in some other (10+) language Wikipedia is now listed under Islamic calligraphy (Q656831). Is there any convenient way to move all those Islamic calligraphy pages to Q23011124 by one click or a few click that also remove the original wikidata item's i18n name without needing to repeat the deletion-addition process? C933103 (talk) 19:55, 7 March 2016 (UTC)

Q656831 seems to be a mix. It might be cleaner to create a new item for "ArabicIslamic calligraphy" and sort links from Q656831 to the new item and Islamic calligraphy (Q23011124). Afterwards Q656831 can be deleted. --- Jura 21:17, 7 March 2016 (UTC)
Fixed label in my comment above.
--- Jura 23:51, 7 March 2016 (UTC)
@Jura, either way, so that can only be sorted manually?C933103 (talk) 14:48, 9 March 2016 (UTC)

@Jura so they can only be sorted manually?C933103 (talk) 14:46, 9 March 2016 (UTC)

At Special:Preferences#mw-prefsection-gadgets, there is a gadget "Move" sitelinks from one item to another. Once it's activated, an icon next to the sitelink allows to move them. The same for statements doesn't exist yet.
--- Jura 16:37, 9 March 2016 (UTC)

Belarusian help needed

Please can someone who reads and writes Belarusian translate the property proposal at Wikidata:Property proposal/Place#SOATO so that the wider community can decide on its merits? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:47, 9 March 2016 (UTC)

Google Translate says the description is "Code System designation of objects administrative and territorial division". There's a mention of a SOATO on en:OKATO, I don't know if it's the same system or if the different countries have different systems with the same name. A Russian speaker might be able to help find out - we have more of those than Belarusian speakers. - Nikki (talk) 19:51, 9 March 2016 (UTC)
It is useful and important identifier for Belarusian administrative and territorial division. OKATO != SOATO. I have translated and supported the request. --Voll (talk) 21:29, 9 March 2016 (UTC)

Language Specific Statements

I don't know if this has come up before as I'm quite new to wikidata, and I don't know how to go about asking for this feature, or if it is even doable.

For what I've seen wikidata tries to do 2 things. The first is describe the concept. The second is associate that concept with specific language information such as a word/description/aliases.

The first is covered quite well and I've seen that we can ask for new properties for this concept side of things. However I haven't figured out how to create or ask for language specific statements.

The reason I bring this up is that I was thinking of how to add Chinese measure words in wikidata. For those of you who don't know what they are I'll explain quickly. For almost every "noun" in Chinese there is one or more possible measure words that go with it. The measure words are rarely predictable, and change between singular and plural forms. In english there are also measure words like swarm for bees or school for fish. Because of the way the measure word system works in Chinese getting the right measure word is crucial for proper understanding (this is even more true orally since a lot of characters sound the same).

I hope people can realize the problem here. I eventually see semantic engines tying into wikidata (I've started building the building blocks of one recently), and without measure word information this will almost never work for Chinese. I do know that wikidata is quite english centered (just because of the majority of people who contribute) and that's why I think this problem might not have come up before. Though I think it could benefit all languages to have language specific statements as we could use them to describe language oddities. I'm sure there are 100s of use cases that I haven't thought of yet.Thewormsterror (talk) 06:26, 9 March 2016 (UTC)

I think the datatype "monolingual text" will be the solution. You should request a property for measure words (not language specific) and with the datatype monolingual text, you need to specify the language when adding a value. Mbch331 (talk) 06:51, 9 March 2016 (UTC)
Wikidata is still not including "nouns" in the way Wiktionary does. Almost every item here today only cover the term Fish. The items does not (yet) tell anything about how "fish" is written in a grammatical way.
But Yes, some properties look language-centric in some aspects here. I cannot imagine any way to use demonym (P1549) in my language for example. It does not only have to do with grammar, otherwise I could have added gender/numerus/article as qualifiers for my language, but that is not always enough. Another language-specific property is named after (P138). In mercury (Q925) it tells it is named after a Roman god. The name I know for this element has nothing to do with a Roman god. I have no source, but I have always thought it got its name after "liquid silver". Another property is P558 (P558). In kilometre (Q828224) it has two values. These two values are separated by qualifiers for latin script and cyrillic script. If all languages with latin or cyrillic script uses the same unit symbol is as far as I know not verified. And of course, there are more scripts than these two. In some cases, there will be a lot of qualifiers like those.
-- Innocent bystander (talk) 07:09, 9 March 2016 (UTC)
That can't work since there's more than one most of the time, for singular, plural, a dozen, a piece and sometimes 3 or 4 might work for the singular form.  – The preceding unsigned comment was added by Thewormsterror (talk • contribs).
So you can't use zh labels now? Currently the label is just one field, but eventually it could be extended.
--- Jura 09:46, 9 March 2016 (UTC)
@Thewormsterror: Wikidata is concept centric, not language and words centric. There is a project dedicated to words and their different form, structured wiktionary. See Wikidata:Wiktionary. You may want to participate :) author  TomT0m / talk page 12:34, 9 March 2016 (UTC)
    • Isn't the core of some of the proposals that structured Wiktionary is to be at Wikidata?
      --- Jura 16:33, 9 March 2016 (UTC)
      • I would prefer structured Wiktionary to be part of Wiktionary at some point. Same with structured metadata of Commons files, for example. Of course, both should use the Wikidata data model and integrate with Wikidata. Sometime in the far future ... --Srittau (talk) 17:04, 9 March 2016 (UTC)
        • @Jura1 yes, let's say "core Wikidata" then. What I meant is that wikidatas data model will have to be extended and that the current one is not done for storing those kind of datas. @Srittau: it does not matter where the data is stored, it'll be available for rendering (and probably eventually editing) on wiktionary as well. But what's for sure is that multiplying the number of wikis may be bad for community because of fractioning people, rules and so on. author  TomT0m / talk page 17:33, 9 March 2016 (UTC)
          • @TomT0m The problem I see with having it on wikitionary has to do with data retrieval. Let's say an outside app wants to perform a specific SPARKLE query such as : `get me all animals with their associated measure words for chinese`. If the data was split between wikitionary and wikidata there would have to be 2 requests and I'm not even sure what the request to wikitionary would look like if there were 100s of animals, especially if the words didn't match up perfectly (scientific names in wikidata, non scientific in wikitionary). Basically it would be a total mess at being reliable. Again without the measure word a word is kinda useless in Chinese. For example 马 is a horse. The measure word is 匹, which is a measure word only for some equids. If an app ties in wikidata for semantic and conceptual analysis without that measure word the 马 really has little meaning. Previous suggestions seem to be good but have disadvantages. The first that @Mbch331 pointed out using monolingual text qualifier would basically put word relationship data in the concept area and might be confusing to a lot of people who don't speak chinese, pretty much 1000s of pages would now have a measure word statement. Another disadvantage is the Chinese label might change between aliases. Because of China's many regions and rich language heritage very often there are 2-5 words for every concept, often with their own different measure words. Sounds like a pain? Well it is. Another suggestion was to include them in the label. This doesn't quite work either because you also can't really put '匹马' you would have to write '一匹马' because the amount is vital to the measure word. 2 horses has a different measure word and I think many horses would have even a different measure word. The best thing to do think is to include all measure words for horse (somehow) and then put properties on the concept of each measure word describing the measure word. While I'm not specifically hunting for a solution right now, I did think it was worthwhile to at least open this discussion. As for how this is done i'm not sure how the database works on wikidata but if it's like many semantic databases it could be done with named graphs on n-quads I believe. We could then represent this to the user as a qualifier on the label.Thewormsterror (talk) 22:43, 9 March 2016 (UTC)

Ordered sets

At Wikidata:Property proposal/Place#next crossing Joshbaumgartner wrote "What we really need is a good way to do ordered sets in Wikidata. That way each crossing could be a part of the set of crossings, and the order could be defined as part of that set. Sets can be ordered numerically, temporally, spatially, administratively, etc. and an item could be part of any number of sets..." (permalink to discussion).

Personally I think this is good idea. As a starting point for discussion (as like Josh I've not managed to find any previous anywhere) I conceive that one way to achieve this would be items (possibly in their own namespace "set"?) that have a title and description in each language and statement of which items are included in the set. At it's simplest it could just be this, and all interaction with it is via (external) tools that allow you to sort the items by label in a selected language or by the value of any property that all (most? at least n?) members of the set have. This could apply to all the items or only to subsets (e.g. the subset of crossing of River Thames (Q19686) that have located in the administrative territorial entity (P131) Greater London (Q23306) or a subdivision).

Alternatively, the set page could define the properties that are relevant to that set. For bridges that could be e.g. number of spans (P1314), start time (P580), length (P2043), etc. These could then be used to sort the set by any of these. My guess is that this would be more work to program but easier to interact with.

A third suggestion is that each set could be more limited, e.g. "List of bridges over the River Thames by length" or "List of tunnels under the River Thames, downstream to upstream" where the order is defined as part of the set. This will generate lots more sets of course, but it might be easier to work with?

All of these should automatically generate a series ordinal for the set, based on however it is sorted. They should also be dynamically updated when a new item is added to the set.

For any of these, it would be useful if the existing item pages had an automatically generated statement/list of which sets the item is a member of.

In addition to obvious benefits internally to Wikidata, there would be obvious benefit to other projects if a list only needed to be maintained in one place rather than needing to update e.g. the lists of longest bridges on 200+ wikis. Thryduulf (talk: local | en.wp | en.wikt) 01:05, 10 March 2016 (UTC)

What's come in mind immediately when talking of automated management is SPARQL. It occurs that SPARQL can handle lists (see for example http://www.snee.com/bobdc.blog/2014/04/rdf-lists-and-sparql.html ) but I don't see at that point any plans to handle this in Wikibase core (@Lea Lacroix (WMDE):? - @Markus Krötzsch:? any opinion ? author  TomT0m / talk page 10:06, 10 March 2016 (UTC)
we can do something like this with SPARQL and the existing structure of wikidata if there is a suitable relation (instance of, part of) between the set members and the set as a whole. Obviously people are managing using Listeria etc - for example I created this automated list of nuclides in Wikidata. There is also the lingering Wikidata:Property_proposal/Generic#.28Disjoint.29UnionOf_.28or_any_better_name.29 proposal that is related to this. I've also been thinking it might be useful to have "map" or "list" datatypes also in wikidata to support time-series and the like; this seems like a related proposal. But maybe another entity type besides Item and Property would be the right solution for some of these issues too? ArthurPSmith (talk) 15:09, 10 March 2016 (UTC)
You seem to be looking for aggregates of data. This is not supposed to go into items but instead it should be created through queries. --Lydia Pintscher (WMDE) (talk) 16:34, 10 March 2016 (UTC)

Add company to wikidata

Hello,

I am new here to wikidata, can anyone advise me on how to add a company's information onto wikidata? Your help is much appreciated, thank you!  – The preceding unsigned comment was added by ‎SuzehXD (talk • contribs). 14:39, 1 March 2016‎ (UTC)

Property for British Film Institute URL

The BFI are interested in matching items in their database to Wikidata items, which I'm helping to organise. I've proposed a property for the BFI identifier, but their is no usable pattern to get from the ID numbers to the url of the page on the BFI. I'm thinking of proposing another property "British Film institute URL" to allow us to link to the pages, but wondered whether is better to just add the url as a reference to a BFI identifier statement instead of creating a new property? NavinoEvans (talk) 12:19, 10 March 2016 (UTC)

Does the URL include the identifier in one way or the other? How many items is this likely to cover? 1000+ or 10000+ ?
--- Jura 13:38, 10 March 2016 (UTC)
@NavinoEvans: It may be worth expending some effort in explaining to the BFI the advantage of them creating such URLs, even if only as redirects, with a view to eventually providing linked data via them. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:44, 10 March 2016 (UTC)
@Jura1: No, the identifier is not included in the url anywhere. The size is over 100,000 items, but will include many obscure titles which won't be in Wikidata.
@Pigsonthewing: Well worth a try :) I'll bring it up in the next discussion and see how it goes. NavinoEvans (talk) 15:31, 10 March 2016 (UTC)
Given the size, I tend to agree with the suggestion. BTW, Wikidata is open to obscure titles ;) There is already a list of lost films.
--- Jura 15:45, 10 March 2016 (UTC)
Thanks Jura, will report back once we have some news about the url structure. Good to see the list of lost films, thanks for the link :) I'm all for adding obscure things to Wikidata too - you never know what someone will find useful! NavinoEvans (talk) 21:50, 10 March 2016 (UTC)

Hall of fames

I noticed, that we don't have anything specific for hall of fames. Checked Rock and Roll Hall of Fame (Q179191) and National Baseball Hall of Fame and Museum (Q809892). Both are used at award received (P166) (logical, of course). Baseball guys suggested member of (P463) (which also could be OK from my point of view). So I wanted to hear your thoughts before adding such statements to items. We should continue adding to award received (P166)? Or create a new property for inclusion at hall of fames? And what about Hollywood Walk of Fame (Q71719)? Could that also qualify for P166 (if we use it for hall of fames)? --Edgars2007 (talk) 16:47, 9 March 2016 (UTC)

It should be award, not membership - the latter implying something that anyone might join. Perhaps we need items for "Membership of the Foo Hall of Fame"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:00, 9 March 2016 (UTC)
OK, then award. I don't understand, why we would need "Membership of the Foo Hall of Fame"? --Edgars2007 (talk) 17:50, 9 March 2016 (UTC)
Because an inductee is awarded "Membership of..." (which should have a suitable instance of (P31) value), not the HoF itself. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:33, 9 March 2016 (UTC)
@Pigsonthewing: Agree, logical. Any idea for property, what to use at National Baseball Hall of Fame and Museum (Q809892) item to add "Membership of ..."? --Edgars2007 (talk) 06:52, 10 March 2016 (UTC)
The other way round - use conferred by (P1027) on the "Membership of ..." item. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:53, 10 March 2016 (UTC)
I wanted to include something on item about HoF, so people can find that membership item, but this also will be fine, thanks. --Edgars2007 (talk) 10:22, 11 March 2016 (UTC)

List of interesting item numbers

Hi, is there a page that maintains a list of interesting item numbers (eg. Q42, Q2001, Q12345) that have special meaning in Wikidata? I started a page to note the ones I've seen, but don't want to duplicate efforts if one already exists. See: meta:Wikidata_Easter_eggs. Thanks. -- Fuzheado (talk) 01:09, 11 March 2016 (UTC)

Wikidata:Humour --Pasleim (talk) 08:20, 11 March 2016 (UTC)
 Like --Edgars2007 (talk) 10:21, 11 March 2016 (UTC)

Please merge

Sorry, merging instructions are too complicated; I just have 10 years of Wikipedia experience and Wikidata is beyond me. I thought about using Special, but there are no instructions which number to put first and which second, so I don't want to mess something up. Anyway:

seem to cover the same topics, at least in Polish and English.--Piotrus (talk) 06:21, 11 March 2016 (UTC)

@Piotrus: Are you sure? First category is for 'civic and political organisations', second is for 'social organisations'.--Kopiersperre (talk) 08:28, 11 March 2016 (UTC)
@Kopiersperre: Sorry,I don't visit WD often. Yes, I am reasonably sure. I'll start a CfD discussion on en wiki about the best English name, but yes, those concepts are the same. PS. Done, see en:Wikipedia:Categories_for_discussion/Log/2016_April_4#Category:Civic_and_political_organizations. --Piotrus (talk) 05:09, 4 April 2016 (UTC)

Dormant properties

The following properties are sparsely used (≤ 10 times):

To the people, who have requested these properties: Use them!--Kopiersperre (talk) 11:22, 12 February 2016 (UTC)

I will try to populate some of them, importing some values from Wikipedias. --Edgars2007 (talk) 12:42, 12 February 2016 (UTC)
I have gone through the first column but without big success (HarvestTemplates):
Matěj Suchánek (talk) 15:43, 12 February 2016 (UTC)
It seems this is limited to string properties. The more problematic item-datatype-properties were skipped.
--- Jura 16:37, 12 February 2016 (UTC)
Striked out also those, which are most probably used more than 10 times, according to pagelinks. --Edgars2007 (talk) 16:58, 12 February 2016 (UTC)
Left a message on Wikidata talk:Flemish art collections, Wikidata and Linked Open Data for MSK Gent work PID (P2511). Mbch331 (talk) 19:31, 12 February 2016 (UTC)

Why are sparsely used properties probematic? Thryduulf (talk: local | en.wp | en.wikt) 21:39, 12 February 2016 (UTC)

Shouldn't the proposer of a property interested in using it? --Succu (talk) 21:47, 12 February 2016 (UTC)
The proposer of a property have identified a need based on one or two entries they are working on, that they extrapolate will be useful for other items. They are not required to commit to using the property on a large number of items within any given time frame. If they do, that's great, but that doesn't answer the question - why sparsely used properties are problematic? Thryduulf (talk: local | en.wp | en.wikt) 20:15, 13 February 2016 (UTC)
... Or it was approved only last month after 50 weeks of waiting and the proposer is still learning how to use wikidata. Dispenser (talk) 22:29, 13 February 2016 (UTC)
@Dispenser: Many of the above properties have low P numbers and were created some time ago. I think the proposers just lost their interest.--Kopiersperre (talk) 19:48, 14 February 2016 (UTC)
I was speaking of myself. 29 of the 206 listed properties were only approved in 2016. Dispenser (talk) 18:14, 15 February 2016 (UTC)
@Dispenser: probably a little late to add this comment, but I would like to urge all property proposers to keep track of their proposal regularly and if nobody's commented please reach out to individuals you think would care and have them say something under your proposal. To be created normally we would look for around 3 supporting comments from other people. It should also ideally come with a well-filled out property documentation template - see Template:Property_documentation for details on that. Good examples are important so we understand how it will be used - the example should include the item the property would appear on, the value you would set, and some way to verify (a reference) that the value is correct. When there has been sufficient discussion and at least a week has passed, you can change the "status" field to "ready" which will make it easier for a property creator to spot that you think it is done. Please answer questions from property creators, if there are any, as soon as you can to speed the process along. Thanks! ArthurPSmith (talk) 21:13, 18 February 2016 (UTC)
I asked the relevant WikiProject to comment on HowLongToBeat and they said it was an unreliable source for Wikipedia. Dispenser (talk) 16:09, 26 February 2016 (UTC)
Well, that's definitely a good way to proceed, and then let the discussion sort out what we should do. Thanks for persevering! If you have other property proposals I would definitely recommend you advertise similarly. ArthurPSmith (talk) 15:00, 29 February 2016 (UTC)

If you can quickly get to the raw data behind any of these properties, please consider importing it into mix'n'match. I did Italian Chamber of Deputies dati ID (P1341) and P1222 (P1222) (though the latter doesn't seem to work with the URL pattern for some IDs, so I deactivated it again for now). If you can get the data but can't import for some reason, tell me and I'll try, but I don't scale too well ;-) --Magnus Manske (talk) 14:48, 16 February 2016 (UTC)

I have started now with Geokod (P1172). I cannot see how you could be able to import that data into Wikidata in any automatic way! -- Innocent bystander (talk) 17:12, 2 March 2016 (UTC)
Imported SABR person ID (P2482) to the mix'n'match. Is there an easier way to confirm multiple auto matches ? .-- Hakan·IST 12:12, 5 March 2016 (UTC)

Isn't ISO 639-6 code (P221) being withdrawn from ISO already?C933103 (talk) 19:49, 7 March 2016 (UTC)

@Jura1: FYI, I don't think this will work (about pinging), at least, it shoudn't. That's why let's ping them here: Snipre and Tobias1984 - take a look at this edit. --Edgars2007 (talk) 08:16, 9 March 2016 (UTC)

@Edgars2007: Thanks for the ping. Yes, some properties are dormant but I don't see a huge problem with these properties, especially because they don't do any harm. Some of these properties are even pretty essential, but are just waiting for someone to import more data. For example, not many people edit Geology pages, so gathering the identifiers for the sedimentary units in the Netherlands will take some years. --Tobias1984 (talk) 19:22, 9 March 2016 (UTC)
@Tobias1984: It would be helpful if you could sort the ones you had created and want to keep, but haven't used yet at User:Addshore/Identifiers/0/1/2.
--- Jura 11:28, 10 March 2016 (UTC)
@Jura1: They are in the convert column. DSM and Pubmed Health probably don't have an ideal 1:1 relationship with Wikidata. I missed some of the discussion about the identifier datatype, so I am not sure how to proceed. --Tobias1984 (talk) 19:06, 11 March 2016 (UTC)

Numismatic objects (coins) on Wikidata

Hi! Is there an interest in adding individual coins to Wikidata? They are very well-documented, have clear data and identification numbers etc., so it wouldn't be totally chaotic. One could probably import data from an online database such as http://en.numista.com or http://colnect.com/en/coins/.

I tried to make a template here, 1 Krona - Gustaf V (Q23002046), but lots of properties I don't know which ones to use or if new ones should be created.

The properties that would be appropriate are these:

available
to do

If this is not desired, feel free to delete my template.

Regards, Jssfrk (talk) 16:35, 4 March 2016 (UTC)

Here is a link to my template coin on Numista: http://en.numista.com/catalogue/pieces6154.html
Jssfrk (talk) 16:37, 4 March 2016 (UTC)
Note that you probably describe a 'type' of coin (and not a coin owned by mr X) so you should use subclass of (P279) coin (Q41207). Look e.g. at one pound coin (Q1439830) Michiel1972 (talk) 17:18, 4 March 2016 (UTC)
And instead of P17 you better use P495 (In changed it in your example). Michiel1972 (talk) 17:23, 4 March 2016 (UTC)
For the start / end date I would probably not use Property:P571 and Property:P576 but start time (P580) and end time (P582) instead, as normal claim or as qualifier with the P31/P279 claim 'coin'. Michiel1972 (talk) 17:27, 4 March 2016 (UTC)
I definitely don't think it belongs on the P31/P279 "coin" statement, it doesn't stop being a coin just because it's no longer minted or no longer legal tender. I also wouldn't add the statements as normal statements because it seems quite vague, it could be referring to the dates the coins were minted or to the dates they were legal tender (which can differ quite significantly, e.g. in the UK, the shilling coin was minted until 1970 but remained legal tender until 1990). For minting dates, I would probably use something like significant event (P793): minting (we don't seem to have an item for that yet) with the dates as qualifiers. I'm not sure about legal tender...
Regarding country of origin (P495) instead of country (P17), I think that depends what the country is supposed to represent, e.g. Ecuador uses the US dollar with coins minted in Canada and Mexico. - Nikki (talk) 19:13, 4 March 2016 (UTC)
In Euroland, each dountry has its own coins, even some non-EU-countries are allowed to mint their own Euro-coins. But all coins are legal to use in all of Euroland (and some other nations too). -- Innocent bystander (talk) 19:19, 4 March 2016 (UTC)
This looks like a sound proposal. I'll be happy to assist you in drafting proposals for the new properties, if you need assistance. I suggest we also need to use properties for the designer(s) and mint. Also an identifer-property for Numista, and other catalogues. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:31, 4 March 2016 (UTC)
I have created a proposal for thickness, the other ones need to be created. Regarding "value", is it meant face value (Q931754)?--Micru (talk) 11:32, 5 March 2016 (UTC)

Thanks for all the valuable input, I have no experience in creating new types of objects from scratch. With value, I indeed mean face value (Q931754). Properties for designer and mint also seem reasonable, as well as different catalogue ID's. However, I have no experience in bot-editing, which I assume is needed in order to make this fly. Is there anyone who can be of assistance? /Jssfrk (talk) 08:34, 8 March 2016 (UTC)

For the lettering on the coin, I think inscription (P1684) could be used - certainly it fits the English description "inscriptions, markings and signatures on an object". Thryduulf (talk: local | en.wp | en.wikt) 11:41, 8 March 2016 (UTC)
Thanks, that's perfect. I'll add it above. I tried to add designer as well, but can't find a property for that! /Jssfrk (talk) 15:21, 8 March 2016 (UTC)
Designer is designed by (P287). For image, I think two image (P18) statements with a qualifier applies to part (P518) obverse (Q257418)/reverse (Q1542661) will suffice rather than new properties, but I'm not 100% certain. Thryduulf (talk: local | en.wp | en.wikt) 13:14, 9 March 2016 (UTC)
For mint, maybe manufacturer (P176) would work? Thryduulf (talk: local | en.wp | en.wikt) 13:16, 9 March 2016 (UTC)
@Jssfrk: Er, yes, I wrote above "I'll be happy to assist you in drafting proposals for the new properties, if you need assistance." Please drop a note on my talk page, when you're ready. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:31, 8 March 2016 (UTC)
@pigsonthewing: Sorry if I was unclear, I am most grateful for your help. When I asked for more assistance, I meant with bot-editing or batch-editing. I'm thinking of making an Excel sheet for Swedish coins 1875-1942 or so, which I then hope that someone knows how to translate into Wikidata pages as a test run. But I don't know how this type of editing works. /Jssfrk (talk) 08:49, 11 March 2016 (UTC)
@Jssfrk: No problem. The "Quick Statements" tool will do that (once the properties are all created!), and I can use that for you if you need me to; or help you to learn to use it. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:34, 11 March 2016 (UTC)

@Jssfrk: Thickness is done: thickness (P2610), for the other properties you have to propose them at wd:pp --Micru (talk) 19:15, 11 March 2016 (UTC)

Project?

We need to create a page to hold the above list of properties, and guidelines for their use. Shall we start "Wikidata:WikiProject Numismatics"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:52, 11 March 2016 (UTC)

Deaths of Wikidata editors

I was saddened, while travelling recently, to learn of the sad passing of Joe Filceolaire, a thoughtful and positive collaborator.

We should consider a process for memorialising Wikidata editors who die, and protecting their user-pages. The equivalent on the English Wikipedia, which I helped to draft, is en:WP:RIP. Would anyone object to importing and localising those guidelines? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:44, 6 March 2016 (UTC)

I highly recommend accepting your idea. DGtal (talk) 13:26, 6 March 2016 (UTC)
I too fully support this proposal. Thryduulf (talk: local | en.wp | en.wikt) 14:53, 6 March 2016 (UTC)
I also support the idea. I sent an email to the family expressing my condolences, and if we have a little memorial here, even better.--Micru (talk) 16:42, 6 March 2016 (UTC)
Sad to hear this. I approve the idea. author  TomT0m / talk page 17:12, 6 March 2016 (UTC)

Thank you, all. I have made a start at WD:RIP. There are still some templates that need fixing up. We will also need input from admins, to protect the user-pages of deceased colleagues. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:47, 6 March 2016 (UTC)


  • I guess the question is now how to localize those guidelines. I think we all agree that the user page is probably best protected. Usual process for deceased users accounts is generally to de-activate accounts. Oddly, this isn't done on a global level. If the family asks for a specific message to be placed on the user page, I think we should consider doing that.
    --- Jura 19:52, 6 March 2016 (UTC)

Differences between statements and constraints

Maybe somebody could generate some report, where could be listed differences (one of them missing should also count as difference) between format as a regular expression (P1793) and {{Constraint:Format}}? I think there could be some mismatches there. --Edgars2007 (talk) 07:37, 7 March 2016 (UTC)

Using some parser functions and category magic could also help. Matěj Suchánek (talk) 12:10, 12 March 2016 (UTC)

Cannot remove an entry

I am trying to remove cs:Kur from Q9304816 (the Czech article is not about Perdicinae subfamily), but something is wrong. When I click the "trash" button, the entry vanishes, but the "save" button above is not clickable. When I refresh the page, the entry is still there. What should I do? --Vachovec1 (talk) 22:09, 9 March 2016 (UTC)

I tried doing the same thing for hi:चकोतरा from Q12425769 (I was planning to expand the article and associate it with en:Pomelo), and ran into the exact same issue. Seems to be a bug? --Hunnjazal (talk) 00:43, 10 March 2016 (UTC)

Same situation today entry can be removed but save button dosnt work --Milicevic01 (talk) 11:38, 10 March 2016 (UTC)

Glad I am not the only one having this problem! I am trying to remove sv:Albury Municipality from Q21922687 so I can merge it to Q1719401 but the "save" link doesn't light up. Was working fine yesterday. Orderinchaos (talk) 12:23, 10 March 2016 (UTC)
@Vachovec1, Hunnjazal, Orderinchaos: Using API, I removed the sitelinks you had tried to remove. Until it's fixed which should be soon, you can use Special:ApiSandbox or the gadget Move. Matěj Suchánek (talk) 13:34, 10 March 2016 (UTC)
We just deployed the fix for this. Sorry that the bug happened :/ Aude (talk) 21:45, 10 March 2016 (UTC)

Thank you for the quick fix. Much appreciated. --Hunnjazal (talk) 07:33, 11 March 2016 (UTC)

Thank you also :) Orderinchaos (talk) 07:33, 12 March 2016 (UTC)

Importing dates from Histropedia

I've been experimenting with exporting data from Histropedia so that it can be used to add missing dates or improve date precision in Wikidata. I've started with a small dataset, just taking dates from the battles and operations of World War II timeline.

I've compared the values for start time (P580), end time (P582) and point in time (P585) to the dates in Histropedia, and generated this list of suggested edits (around 240 in total) - An edit is only suggested if there is either A) no date in Wikidata yet, or B) a less precise date in Wikidata.

Any objections to me making these edits in one go using QuickStatements? There will be plenty more of this type of export in the future, so it would also be good to figure out a general way of feeding back into Wikidata that the community is happy with. NavinoEvans (talk) 12:33, 11 March 2016 (UTC)

I strongly object. The Histropedia "About" page suggests it is based on a combination of Wikipedia and WikiData, neither of which is a reliable source. A section of Wikipedia's Verifiability policy explains why this kind of circular sourcing is a bad idea. Jc3s5h (talk) 12:52, 11 March 2016 (UTC)
To clarify, all of the items on the list have been manually edited by humans within Histropedia because the data obtained from Wikidata was either missing or in-precise. And just to be clear, I'm not proposing adding Histropedia as a reference to these statements, it's just about getting some better data than we have in Wikidata already. Countless statements are being added to Wikidata using Wikipedia categories or bots reading infoboxes already, and this is just an extension of that as all of the edits suggested reflect the contents of English Wikipedia infoboxes (just that they have been entered by humans instead of read by a bot). NavinoEvans (talk) 14:42, 11 March 2016 (UTC)
Nothing is better than data that lacks a reliable source. The bot copies from English Wikipedia infoboxes was a horrible idea, and is one of the reasons Wikipedia won't automatically put WikiData information into articles. Jc3s5h (talk) 15:14, 11 March 2016 (UTC)
But if it didn't happen, how much less useful Wikidata would be today. I strongly disagree with being so picky at this stage in Wikidata's development - later on down the line I agree we should be more picky, then we can output a list of suggestions from Histropedia (or another application) to highlight items the community should look into. But for now, I don't think this slow approach makes sense. I've checked 50 of the suggested edits and found each one to be correct. In the mean time the items I'm proposing to fix have no dates, or very in-precise dates, or completely incorrect dates (often using the wrong property as well, like having a single end time (P582) instead of a point in time (P585) which is very common). On a timeline, mistakes in date editing are instantly visible (e.g. a battle appears before the war even started) so it makes sense that it would be less error prone - the same would go for location data that was manually added to a mapping application.
What would you suggest we do with this data (and other similar exports) ? NavinoEvans (talk) 15:57, 11 March 2016 (UTC)
Jc3s5h: Be aware your concern is only limited to a certain number of Wikipedias. A larger number don't care or do care but are willing to accept bot imports. I personally find it obnoxious that edits like these are lambasted. We can't source the data in mass form because of licensing restrictions; Wikipedias don't want to pick up the slack on top of that? That's bullshit and puts Wikidata between a rock and a hard place for providing any functionality. If Wikipedia doesn't want to use it, that's their problem IMO. --Izno (talk) 17:25, 11 March 2016 (UTC)
Izno But the Wikipedia which provide negative feddback are among the biggest ones so they represent a larger amount of users. Then the objective if WD is not to become a trash bin of the web collecting data from any website (you can use google for that) but to offer reliable data which can be used by WP and other external users.
And just forget the problem of data reuse to think about WD data management: how can you handle contradicory data without any sources ? Even permissive WP will have a problem with WD in that case if there is no possibility to filter data using sources. Snipre (talk) 19:09, 11 March 2016 (UTC)

True, they are the largest. Frankly, though, they're not our main users anyway since they have the number of editors required to be the largest. Not so the little wikis, who are the ones who are going to use Wikidata for its automation value on-wiki.

The data is, for-the-most-part correct (where it is not contradictory). Who cares about reliability at that point? This argument doesn't make sense to me.

I'm not going to argue against that branch, since it's not relevant here. And the obvious answer is "well gee we actually do have to source this". But when a (W|w)iki* needs the information of a particular sort, guess what will happen? They will add the sources. I personally would rather see the big ones start using the data so we can know where our inaccuracies are. Maybe the inaccuracies will be systemic across a topic. Maybe they will be random occurrences. We can't do the job of making sure our data is correct until someone comes along and tells us it's wrong. It is hard work to "source" a statement and we need help doing it because most of us aren't subject matter experts. They're afraid to be the SMEs and I honest-to-god cannot figure out why. Our interface is not scary-enough for the persistent "oh, this data is bad" to be a legitimate argument, quite frankly. And that's without any of the effort that's been made to make better UIs and integration with the client databases (e.g. allowing editing a statement on the (W|w)iki* of interest). --Izno (talk) 19:24, 11 March 2016 (UTC)

There is an ongoing fight about Wikidata's usage on frwiki, with people with very strong opinions that uses every possible ways to slow down stuffs : disinformation, personal users targeting with repeated blocking requests to administrators, trying to push unrealistic requirements for Wikidata usage in RfCs ad nauseam ... This is pretty bad on that side, but it seems that things are slowly settled down and that there is actually only a very limited number of people left in this part, and that other parts of community are globally beginning to actually use Wikidata whatever happens on the politics side which totally lost all forms of legitimacy and credibility as far as I'm concerned. author  TomT0m / talk page 11:50, 12 March 2016 (UTC)
NavinoEvans - this RFC is apropos. In particular, the consensus such as there was seemed to be strongly in favor of adding a source statement of "reference URL" to the exact page from which the information is coming, in this case I would recommend for each date you put in wikidata you point to the precise Histropedia link that is the source unless you have a more reliable source to include. ArthurPSmith (talk) 19:43, 11 March 2016 (UTC)
A source "imported from" "Histropedia" would allow those (wikipedia, external users) who don´t want to use that information to detect this statement. By the way, I´m pro importing this data. --Molarus 01:30, 12 March 2016 (UTC)
Thank you all very much for the feedback. Given the outcome of the RFC and other comments here I feel happy to proceed with this, including as suggested an 'imported from' 'Histropedia' statement and the url of the page it came from. As you say, this will allow anyone to filter out this data if required and gives bots/humans a flag to replace with a proper source later. NavinoEvans (talk) 13:31, 12 March 2016 (UTC)

Files, qualifiers, and licenses

Is it recommended to add qualifiers with copyright license (P275) to statements with data type "commonsMedia" like done here? I understood that statements about Commons files had to be made on Commons, have I failed to keep up with the policies? --Ricordisamoa 23:40, 11 March 2016 (UTC)

That seems like a huge duplication of effort for little gain. My gut reaction is "no, don't do it". But that's not policy, just my opinion. —Tom Morris (talk) 10:54, 12 March 2016 (UTC)
What Tom said. Don't do it Jeblad, just make sure that the information is correct on Commons. Multichill (talk) 11:44, 12 March 2016 (UTC)
As long as we can't get those data from Commons it must be added here. At nowiki we have a clear understanding that we must add the license as part of the credit, and with no credit no reuse of statements fra Wikidata about photographic work. If license and/or creator is removed, then we have to drop reuse of image links from the item, and thus we can't add images to the infoboxes.
Note that proper credit is part of the Norwegian law on copyright, Lov om opphavsrett til åndsverk m.v. (åndsverkloven) §3 which says "named according to good practice" ("navngitt slik som god skikk tilsier"), which is interpreted by us that we should credit the work as the creators says it should be. The law goes on to say "Sin rett efter første og annet ledd kan opphavsmannen ikke fraskrive seg, med mindre den bruk av verket som det gjelder, er avgrenset efter art og omfang." which basically means we can't overrule remove the creator unless he explicitly says soallow us to do so, wherby we need the license. That is either we must have the creators name, or if not, we need the license to say why we don't credit the creator.
There are ongoing and previous threads at w:no:Wikipedia:Tinget#Manglende kreditering av bilder i infoboksen, w:no:Wikipedia:Torget#Byline og dataelementer, and w:no:Wikipedia:Tinget/Arkiv/2007-17#Kreditering i byline. There are a lot of other threads about this issue, but I don't think it is necessary to post all of them. The same issue exists for nnwiki, and the law is also similar in Denmark Bekendtgørelse af lov om ophavsret §3 and Sweden Lag (1960:729) om upphovsrätt till litterära och konstnärliga verk §3 so it also apply for svwiki and dawiki but I don't know what is done there to handle this issue. Jeblad (talk) 13:05, 12 March 2016 (UTC)

Representative image

If you go to Q345641, the WD item for the deceased English philosopher Bernard Williams, the article preview for English Wikipedia shows an image of Shirley Williams, his living ex-wife. This also appears in the "similar articles" function on English Wikipedia which is powered by Wikidata. This isn't helpful for readers as it isn't an accurate picture, and it is possibly a BLP violation (or at least in poor taste) to say that a picture of Shirley Williams is representative of her ex-husband. How does one fix this? —Tom Morris (talk) 10:49, 12 March 2016 (UTC)

Well, the picture that was used on Wikidata was a building. We can't influence images on other projects, though. Sjoerd de Bruin (talk) 10:54, 12 March 2016 (UTC)
Ideally, by sourcing an image of him. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:28, 12 March 2016 (UTC)

I realise this isn't Wikidata related. I've reported it over on Mediawiki.org. Sorry for posting it here. —Tom Morris (talk) 14:56, 12 March 2016 (UTC)

Tool for adding multiple statements for a single reference?

Does any one know of a tool that will allow me to do this for batch edits? e.g. for adding stated in (P248) and retrieved (P813) to a single reference. Unfortunately it appears that QuickStatements does not support this (unless I'm missing something?) - it will allow you to add multiple separate references, but not multiple statements for the same reference. NavinoEvans (talk) 15:45, 12 March 2016 (UTC)

If you can provide data sets you can take your turn in the list of bot request. Snipre (talk) 19:04, 12 March 2016 (UTC)
Many thanks Snipre, will do. NavinoEvans (talk) 23:55, 12 March 2016 (UTC)

Wikidata items for Wikimedians

Hi. I've started a discussion at Wikidata talk:Notability#Are Wikimedians considered notable? about the practice of creating Wikidata items for Wikimedians. --MZMcBride (talk) 20:24, 12 March 2016 (UTC)

See the above RFC for some proposals to reform the property creation process. MediaWiki message delivery (talk) 06:29, 20 March 2016 (UTC) (on behalf of Rschen7754)

Mess with Julian dates

(I guess it has already been discussed but didn't note it)

If I understood it right, dates will now be stored in the same calendar as they are displayed (ie, the one indicated in UI), and the UI now works fine. However, massive clean-up is needed for Julian dates. I have mostly noticed that for European Middle Ages.

Most of those that state "gregorian" appear to be actually the Julian date (example). If a bot changed the calendar to Julian for dates in the 500-> 1582 period or so, that would probably be mostly correct, but not 100%.

In other cases, the date is stated to be in Julian but is actually in Gregorian. here, the calendar should indeed be Julian, but the date is actually 22 August. 31 August is the date in Gregorian.

I don't know what we should do, but we should do it quickly, as the data are currently plain wrong, and it will get increasingly difficult to find a pattern in the wrongness. --Zolo (talk) 08:37, 11 March 2016 (UTC)

It isn't true that the UI is fine. Various bugs on Phabricator have agreed that years before AD 1 should be stored with the convention that the year 0 exists and is equal to 1 BC, that year -1 = 2 BC, and so on. But if you put 2 BC into the UI it is stored as -2, rather than -1 as it should be.
Also, it is impossible to enter a date in the UI that exists in the Julian calendar but not the Gregorian, such as 29 February 1800. Jc3s5h (talk) 12:45, 11 March 2016 (UTC)
Ok, so we may have another batch of things to fix later on. Anyhow, most AD dates are now fixable, and should be fixed. Perhaps a good starting point would be a sparql query for Midlle Age dates with precision >= month and using only Gregorian calendar are they need to be checked. --Zolo (talk) 15:13, 11 March 2016 (UTC)
There is something similar to what you request: Wikidata:Database reports/Complex constraint violations/P569 Jc3s5h (talk) 13:23, 13 March 2016 (UTC)

SPARQL question

Hi, I'm working on an area with very few statements, and the SPARQL manual doesn't seem to have a relevant example. How would I form a query to get a list of all items with a particular bit of text in the label, e.g. ", Western Australia", and the corresponding Q-codes? Thanks in advance. Orderinchaos (talk) 07:38, 12 March 2016 (UTC)

You can try something like this: tinyurl.com/zbvy5vc but it seems to be very slow and times out. Matěj Suchánek (talk) 12:08, 12 March 2016 (UTC)
For finding all items with a particular string in the label, I usually use Autolist (e.g. [1]) or Quarry (e.g. [2]). They're still slow, but they let queries run for much longer, so they don't time out as much. - Nikki (talk) 12:48, 12 March 2016 (UTC)
Just created {{LabelContains}}. Usage is like this (to be used on Wikidata):
SELECT ?item where { {{tl|LabelContains|?item|, Western Australia|en}} }
Which gives
SELECT ?item where { 
     ?item rdfs:label ?itemLabelEn filter (lang(?itemLabelEn) = "en"&& contains(?itemLabelEn,_",_Western_Australia")) }
Try it!
Also timesout for this query, even with "limit 1" added. But works pretty well when limiting to "Australia" with a limit of one result. author  TomT0m / talk page 13:48, 12 March 2016 (UTC)
Thanks very much! :) I'll give that a try later today. Orderinchaos (talk) 08:53, 13 March 2016 (UTC)

Awfull mess on VIAF - 3 diff. persons mixed up

Hello,

When I wanted to check VIAF for Marcus Joseph Müller (Q16127018), I found this, where Friedrich Burgmüller (Q62075), Marcus Joseph Müller (Q16127018) and Johann Heinrich Jakob Müller (Q1666860) are mixed up together.

I don't know who to contact to untangle those. Do we have a contact in VIAF ? Thanks for any help. --Hsarrazin (talk) 21:36, 12 March 2016 (UTC)

Your example is peanuts (Q2065858) when compared to VIAF 28144648313309328319. There are several cases where last month's reclustering went havoc. I approached VIAF via their contact mail ("Send us a comment" at the bottom of the pages) but didn't receive any answer yet. Which from my experience is quite normal for them and does not necessarily imply that the report went unnoticed. -- Gymel (talk) 00:38, 13 March 2016 (UTC)
You may also report such issues at en:Wikipedia:VIAF/errors. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:30, 13 March 2016 (UTC)

Sports in the UK

The meaning of country (P17) is somewhat fuzzy. Is it ok if I change the value in Arsenal F.C. (Q9617) from UK to England ? In most context, it seems to make more sense (Scottish and English leagues are completely separated etc.) --Zolo (talk) 13:22, 13 March 2016 (UTC)

I personally don't think so. This appear to be the same issue as here. We have country for sport (P1532), which may be used for sports clubs, as well (note, that description isn't up to date). --Edgars2007 (talk) 14:49, 13 March 2016 (UTC)
Yes, I saw the other thread too late.
I think that given the fuzziness of "country" the solution may differ from on domain to the next, but it's true that is might make things rater confuse.
Property talk:P1532 states that is can be used for national teams (English football team), but says nothing about clubs (Chelsea FC). I don't know wether or not that should be extended. --Zolo (talk) 15:31, 14 March 2016 (UTC)

Elo ratings

Is there some ongoing bot-task that updates elo ratings of players (Chess for ex.)? Thanks. --Wesalius (talk) 17:37, 13 March 2016 (UTC)

Not an on-going but there was a single-run request at Wikidata:Bot requests/Archive/2015/06#Elo ratings for chess players. Matěj Suchánek (talk) 17:59, 13 March 2016 (UTC)
Thank you. --Wesalius (talk) 18:45, 13 March 2016 (UTC)
Yeah, we should get this task done regularly, otherwise it's pretty useless to have one old rating. Of couse, better than nothing, bet still. On enwiki, the template stopped updating few months ago, also on dewiki it's now updated by user. --Edgars2007 (talk) 18:37, 14 March 2016 (UTC)

Properties data types

ISO 3166-2 code (P300), GND ID (P227), KOATUU ID (P1077) — should its datatype be external identifier, as i.e. GeoNames ID (P1566)?--Avatar6 (talk) 06:07, 14 March 2016 (UTC)

Please comment at User:Addshore/Identifiers and its subpages. Thanks. --Izno (talk) 12:37, 14 March 2016 (UTC)
looks like GND ID (P227) at least has been converted... ArthurPSmith (talk) 15:39, 14 March 2016 (UTC)

Page at... properties, new nomenclature???

I am trying to understand why we have only four properties that start "Page at ..." and they are all ".by" related.

and all created at the same time. Not certain whether it is an ugly precedent, and we can expect thousands of properties like this as websites all over the world come along to fill the same naming pattern.  — billinghurst sDrewth 10:25, 14 March 2016 (UTC)

They were probably all proposed by the same person. I think there may be value in something like this but agreed, I'm not sure this way is the best way to do this, or whether this is the best time to do this. --Izno (talk) 12:39, 14 March 2016 (UTC)

Wikimania: Last chance to register discussions, posters, trainings

The deadline (20 March) is approaching rather quickly. Please visit this page to register your submissions: https://wikimania2016.wikimedia.org/wiki/Submissions . And hopefully some of you can translate this message and post it in other project chat languages. --Tobias1984 (talk) 20:48, 9 March 2016 (UTC)

I am planning to do a very basic pywikibot introduction: https://wikimania2016.wikimedia.org/wiki/Training_sessions/Proposals/Learn_to_use_Pywikibot_for_Wikidata_and_other_projects --Tobias1984 (talk) 23:03, 11 March 2016 (UTC)
Less than a week now. Final bump. --Tobias1984 (talk) 20:59, 14 March 2016 (UTC)

Subclass of Great-grandfather?

I have discovered that relative (P1038)/kinship to subject (P1039) causes us some problems in Swedish. To be able to use such items as great-grandfather (Q2500621)/Grandfather/Aunt, we have to invent new words to our vocabulary. If you know if Máximo Jeria (Q6035647) is "mothers fathers father" or "fathers mothers father" or "mothers mothers father" or "fathers fathers father" to Michelle Bachelet (Q320), then please use such items. "Great-grandfather" does not tell us enough information to describe a relation in proper Swedish! -- Innocent bystander (talk) 13:42, 10 March 2016 (UTC)

We have items like maternal grandfather (Q20776692) and paternal grandfather (Q19682162) - you could always create (and use) more. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:07, 10 March 2016 (UTC)
 Like -- Innocent bystander (talk) 17:31, 10 March 2016 (UTC)

How to fix incorrectly spelled item

I'm sure it's been asked many times, but I'm new here. How should one proceed when a Wikidata item is spelled wrong? I just came across Alexandru Chiculiţă (Q1644850), which should be spelled Alexandru Chiculiță (t-comma, not t-cedilla). Adding the correct spelling as an alias just seems to compound the error. DavidBrooks (talk) 00:52, 14 March 2016 (UTC)

It shouldn't be that hard. I think I fixed it, but hopefully the page Help:Label can help you. FallingGravity (talk) 02:47, 14 March 2016 (UTC)
Thanks for doing that; Help:Label doesn't actually say how to do a rename. On a related point, someone has added an "alias" to that item which is the name without diacritics, while over on w:Wikipedia talk:Persondata we reached the conclusion that a diacritic-free version is usually a mis-spelling, not an alias (there are many artefacts in the Persondata merge because of the way Persondata was populated with default sort keys). DavidBrooks (talk) 23:38, 14 March 2016 (UTC)

Chart of nuclides tool available

Ricordisamoa and I have updated the Wikidata Periodic Table app to add a new "chart of the nuclides" feature, displaying some of the data from the over 4000 isotopes listed in Wikidata. You can view the nuclides color-coded by half-life or by decay mode. The chart is zoomable and you can pan around to focus on particular groups of isotopes. Each isotope has a label provided as a tooltip, and clicking on the box brings you to the associated Wikidata page. I think it's a nice illustration of the power of Wikidata as a datasource. It also gives a quick check for any data that's out of place (for example when carbon-14 (Q840660) was misplaced for a little while recently it was very noticeable). Thanks to Pamputt and Tobias1984 for suggestions. Unfortunately the label and click-through feature may not work on touch-screen mobile devices. The nuclides chart is available at https://tools.wmflabs.org/ptable/nuclides - enjoy! ArthurPSmith (talk) 14:38, 14 March 2016 (UTC)

Thank you very much for this work. Note that it also help to fix some errors on half-lif such as gold-190 or americium-244. Pamputt (talk) 06:45, 15 March 2016 (UTC)

Looking for a simple tool to generate Q numbers

Is there a tool which will generate Qcodes from a long list of Latin names of birds, please? I can then upload the Welsh corresponding name using 'Names as Labels'. Llywelyn2000 (talk) 10:56, 14 March 2016 (UTC)

Look at that but before doing any creation we should check if existing items about birds don't correspond to birds in your list to avoid duplicate. We have already a large number of duplicates so it is better to perform a check of existing items before any massive creation of items. Snipre (talk) 11:14, 14 March 2016 (UTC)
He's not looking to create new items but to add labels in the current ones. --Izno (talk) 12:40, 14 March 2016 (UTC)
Ok, I misunderstood the term Qcode. Snipre (talk) 15:53, 14 March 2016 (UTC)
@Llywelyn2000: If you have a list of Wikipedia article titles, Magnus' Linked Items tool will give you the equivalent Q numbers. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:00, 14 March 2016 (UTC)
Thanks for these two links. I should have said 'items' or Q numbers. What if they don't exist on wiki? Llywelyn2000 (talk) 22:29, 14 March 2016 (UTC)
@Pigsonthewing: It's a csv file. Just tried 'Asplenium trichomanes' on Linked items, and it came up with 50 items; I'm looking for that one unique code. Llywelyn2000 (talk) 10:55, 15 March 2016 (UTC)
@Llywelyn2000: I'd be very surprised if there is no en.WP article for each UK bird species, though scientific names will often redirect to common names - Linked Items handles that. Simply wrap your list of names in double square brackets and paste, one per line, into Linked Items, selecting enwiki. A list of Q numbers should be returned. I've just tried [[Asplenium trichomanes]] and it returned Q902211. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:47, 15 March 2016 (UTC)
Brilliant! Many thanks Andy! Llywelyn2000 (talk) 12:38, 15 March 2016 (UTC)

Names of types of county subdivisions

So I missed Wikidata:Project_chat/Archive/2016/01#Labels:_A_general_term_vs._a_specific_one, but it seems things were not settled very clerly there.

Should county of the United States (Q47168) be called "county of "county of the United States" ?

I am strongly in favour of short "county" form. A label is supposed to be the "most common name that the item would be known by" and in just about every context, you say that Los Angeles is a county, not "county of the United States" or a "county of California". Beside, when there can be an ambiguity about the location, clients can easily solve that using located in the administrative territorial entity (P131) or country (P17) while is it much more complex to remove the unneeded "of the United States" part. This "of the United States" already causes problems in templates like fr:Modèle Localité and will do so in all sorts of Wikidata-based templates or lists. Not that Wikipedia has to be built around Wikipedia's needs, just that Wikipedia and all other clients should be able to rely on the assumption that labels in Wikidata match common usage. --Zolo (talk) 15:51, 14 March 2016 (UTC)

Making a label the shortest description that something is known by is fine if you're just going to name the item and that's the end of it. But if you are going to fill in data about the item, like we do at WikiData, the label needs to be long enough to distinguish the item from other similar items that have different kinds of things that you can add to it. For example, I think counties in the UK are only used for ceremonial purposes, so it would make no sense to add a county legislature or county chief executive officer. I know Connecticut abolished all county government function in the 1960s. So I think it does make sense to call county of the United States (Q47168) "county of the United States". It's also helpful that we have county of Connecticut (Q13414354) since they lack any government functions. Jc3s5h (talk) 22:50, 14 March 2016 (UTC)
I thought these cases were what "aliases" and "descriptions" are for. I strongly favour the "bare" name of the item instead of disambiguation names in disguise for the label. Additionally to aliases and description, "of the United States" is just more data to fill (maybe through applies to jurisdiction (P1001)?). --Asqueladd (talk) 09:18, 15 March 2016 (UTC)
@Jc3s5h I totally agree that we need different items when the meaning is different, but that does not mean to that the label should be different. I cannot imagine a situation where you are going to add a chief executive officer to an item, but then realize that it is written "instance of couty of Connecticut" and so refrain from doing so. Disambiguation is supposed to be done in the description, not the label. -Zolo (talk) 15:05, 15 March 2016 (UTC)
@Zolo:, I suppose the most convenient label depends on the context in which you are looking at the label. If I'm using the user interface to view Fairfield County (Q54231) and I were to see it was an instance of "county", and I was aware of the different kinds of counties that exist around the world, I might wonder what kind of county it was, or whether the best possible parent class had been chosen. I could hover my mouse over the word "county" and see that it is an instance of Q13414354, but that number won't be meaningful to a person, so I would actually have to click on the link to find out that it was really a county of Connecticut. So in the context I just described, "county of Connecticut" is a more convenient label. Jc3s5h (talk) 16:39, 15 March 2016 (UTC)
@Jc3s5h: yes, many labels are not really fit for all contextes. But I think on balance things are clearly in favor of the shorter format here. The longer form is only needed when there is no context explaining that we are referring to counties of Connecticu, and that is not very frequent. Even when reading directly on Wikidata, you generally don't reach an item by chance, and even when you do the description or located in the administrative territorial entity (P131) will provide you with the needed geographic context. For maintenance purposes, it is indeed more convenient to see directly whether the item is tagged as an instance of county of Connecticut or another, less accurate value, but given the number of items about administrative divisions, serious maintenance is only feasible through constraint violation reports and other mass listing/editing tools, where the label is not that important. More generally I don't think that readability on Wikidata should be given precedence over usability for clients. --Zolo (talk) 17:16, 15 March 2016 (UTC)

Wikidata weekly summary #200

Wikidata weekly summary #200

Wrong datatype for 'World Heritage criteria (2005) (P2609)' property

P2609 was recently created, with the datatype of "string", The usage instructions say "Use one of the following (i), (ii), (iii), (iv), (v), (vi), (vii), (viii), (ix), (x)". Those ten values should be items, with the datatype changed accordingly. There is currently only one item using the property, so a prompt change won't cause disruption. CC: @Amadalvarez, Micru:. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:27, 12 March 2016 (UTC)

I've now created the ten properties. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:34, 12 March 2016 (UTC)
I would agree with a change to item datatype. --Izno (talk) 16:05, 12 March 2016 (UTC)
My mistake. Thanks for sorting it out! Btw, the data can be imported from en-wp templates.--Micru (talk) 21:27, 12 March 2016 (UTC)
FYI, I actually have a complete list of criteria in a spreadsheet (thanks to User:John Cummings, Wikimedian in residence at UNESCO) so will organise a batch edit referencing the source data now that we have the property sorted :) NavinoEvans (talk) 00:49, 13 March 2016 (UTC)
I agree with the change. Sorry if I was wrong when I asked for the property. @NavinoEvans:, please ping to me when you upload the values. Thanks, --Amadalvarez (talk) 18:00, 13 March 2016 (UTC)
@Amadalvarez: will do. NavinoEvans (talk) 10:57, 14 March 2016 (UTC)

Sorry. Should selection criterion for World Heritage (Q23038930) be merged with selection criterion for World Heritage (Q22809610) ?. What would be the correct Statements ?. Thanks, --Amadalvarez (talk) 18:10, 13 March 2016 (UTC)

@Amadalvarez: Done. Please take a look. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:21, 16 March 2016 (UTC)

OK, now recreated at World Heritage criteria (P2614). P2609 has been deleted. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:30, 16 March 2016 (UTC)

2x duplicate - Nive River

The duplicates having Q21.. are from sv + ceb, probably created by Swedish bot man. 91.9.109.141 20:02, 15 March 2016 (UTC)

merged them with the gadget "merge" which you can activate in your preferences. --Pasleim (talk) 09:06, 16 March 2016 (UTC)

bug in watchkist?

In my watchlist in he-wikipedia, i see any change on a wikidata item of higher taxonom of an item i have on my watchlist.

For example, in my watchlist i have 17 article of flower in angiosperms. two days ago gl-wikipedia change the name of the article and i got 17 notation, for each article, about this. and angiosperms are not in my watchlist.

It append again with 2 change in animal kingdom is the last days. all my watch list is full of the same change, when Q729 (or the Hebrew article) is not in my watch list. 3 Days ago in append with Viridiplantae, that don't exist in hebrew wikipedia.

Why it append? In Hebrew-wikipedia we take taxonomial data from wikidata, can it create this problem? - yona b (talk) 09:42, 16 March 2016 (UTC)

@יונה בנדלאק: Hi, I'm not sure I fully got your description, so please clarify if my answer is not useful to you. Yes, this is expected, all relevant Wikidata changes (these are changes to items directly linked to the pages you watch and the changes to further entities used on these pages) can been seen on the local watchlist. These are basically all changes you can see in the page history of the entities in question. - Hoo man (talk) 11:46, 16 March 2016 (UTC)
I will give an example. in my watch list i have he:טמרין שחור. it's relate to Q672895. someone did a change in Q7380 so i see:
(diff | hist) . . D טמרין שחור (Q7380); 17:10 . . Pigsonthewing-bot (‏נוצרה טענה: Property:P2612: primates, #quickstatements)
why i see it is not a change on Q672895. If it was only 1 line it ok. but i write alot article of speaces from primates so it apear several times. - yona b (talk) 12:01, 16 March 2016 (UTC)
It's a known bug phab:T45578. thanks for the answer Hoo man. - yona b (talk) 05:57, 17 March 2016 (UTC)

SineBot

I propose to ask the operator of en:User:SineBot to have it operate on this project. It will add a sig to unsigned talk page posts, and to unsigned posts on discussion pages like this one. Here is a sample edit. Any objections? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:54, 16 March 2016 (UTC)

Sounds like a good idea. --Srittau (talk) 11:02, 16 March 2016 (UTC)
Fine by me too. It can be useful for engaging newbies in discussions and the like. Ajraddatz (talk) 22:06, 16 March 2016 (UTC)

Open call for Individual Engagement Grants

 

Hey folks! The Individual Engagement Grants (IEG) program is accepting proposals from March 14th to April 12th to fund new tools, research, outreach efforts, and other experiments that enhance the work of Wikimedia volunteers. Whether you need a small or large amount of funds (up to $30,000 USD), IEGs can support you and your team’s project development time in addition to project expenses such as materials, travel, and rental space.

Also accepting candidates to join the IEG Committee through March 25th.

With thanks, I JethroBT (WMF) 23:01, 16 March 2016 (UTC)

is Northern Ireland a country?

In the past week there have been many revisions like this, and I wonder if there is a better way to accommodate the diversity of opinions over the status of Northern Ireland as a country, or if there is simply a strict interpretation of P17. --Haplology (talk) 01:46, 12 March 2016 (UTC)

I think it should all be reverted. We have located in the administrative territorial entity (P131) for this.--Ymblanter (talk) 14:15, 12 March 2016 (UTC)
The addition of "country" is and has been mandatory for locations for a while now. I would suggest an RFC if you think we should change that. --Izno (talk) 16:02, 12 March 2016 (UTC)
As far as I am concerned, the country in terms of P17 is the UK, not Northern Ireland.--Ymblanter (talk) 12:25, 13 March 2016 (UTC)
This is something Wikidata is built for. You need to provide sources to show who thinks what about its status and make sure the Help:Ranks are set appropriately. --Izno (talk) 16:01, 12 March 2016 (UTC)
It looks like rank is not appropriate for claiming that NI is a country because it says deprecated rank is for "errors or outdated knowledge" rather than a very current and controversial issue of current times. --Haplology (talk) 02:04, 16 March 2016 (UTC)
Isn't this a matter of how the language is used? In Swedish we sometimes use the same word for NI/Scotland/Wales/England as we do for Götaland (Q201694). And in English you do not use the word "country" for Götaland. -- Innocent bystander (talk) 07:11, 16 March 2016 (UTC)
@Haplology: Rank is absolutely appropriate. There are three ranks, not two: Preferred, normal, and deprecated. In this instance, since most of the world observes the UK's stance on the issue, NI is preferred to be a subdivision of the UK. But you can still add a second, 'normal' claim with North Ireland. --Izno (talk) 11:34, 16 March 2016 (UTC)
I see, thank you. --Haplology (talk) 00:05, 17 March 2016 (UTC)
country (P17) has always been problematic. You only have to look at the description and label (in English) to see how muddled up it is. The description says "sovereign state of this item", but a sovereign state is not the exact same thing as country. Calling it one thing, and describing it as something else is going to lead to problems. Northern Ireland is a country (a constituent country of the UK), but is not a sovereign state. As country (P17) stands it is of little use, however we can make sense of it all by using and following ISO 3166 (Q106487) related properties. By simply following the ISO "rules" we can avoid endless debates and arguments. Danrok (talk) 01:52, 17 March 2016 (UTC)
Currently, P17 is described as the sovereign state for an item (which in this case is the UK), with some exceptions allowed for partially recognised states and some constituent parts with very high levels of autonomy (but Northern Ireland isn't on that list), so I agree with Ymblanter that the edits should be reverted and P131 used instead, unless someone can find a good reason to add Northern Ireland to the list of exceptions. - Nikki (talk) 11:52, 17 March 2016 (UTC)

API change rank and order

I've added a new code value settlements. I would like to change rank of old value on "deprecated" or, may be, to change the order by moving the new value in the first place. Advise how to act in this case, and examples in the API request for 1. change of rank and 2. change of the order, please. Игорь Темиров (talk) 11:39, 15 March 2016 (UTC)

You should not change the rank of the old value to deprecated but the rank of the new value to preferred, see Help:Ranking. With the API this can be done with action=wbeditentity. Changing the order of statements is generally not possible. --Pasleim (talk) 09:14, 16 March 2016 (UTC)

Qualifier for "last update"

For some data that are respect to frequent changes, we need a method to display the last update of the data because it might by outdated. A concrete example is the performance data of soccer players. We have number of matches played/races/starts (P1350) and number of points/goals/set scored (P1351) for the number of games and goals of a player for a certain team. Experimentally, I retrieve these data for the players' infoboxes in the German Wikipedia. However, this data is only meaningful together with a "last update" information because the displayed data is most likely outdated. I used point in time (P585) for this in the past (see for example Q6451050#P54, qualifier in the member of sports team (P54)CF Montréal (Q167615) statement). However, I feel like this is not the perfect property to use. Is there any better or, if not, what do you think of creating one for this purpose? Yellowcard (talk) 13:19, 16 March 2016 (UTC)

I was also thinking about that in relation with owned by (P127) and minority owners of companies traded on stock exchange. My conclusion was that point in time (P585) is the most suitable solution. --Jklamo (talk) 14:21, 16 March 2016 (UTC)
I've also used point in time (P585) to qualify employees (P1128) as that will remain true, even if the number changes (which I would add as a second statement) and doesn't imply that the data is or is not presently correct. Thryduulf (talk: local | en.wp | en.wikt) 13:12, 17 March 2016 (UTC)
Personnally, I'm fine with this specific use of point in time (P585). --Casper Tinan (talk) 14:29, 17 March 2016 (UTC)
point in time (P585) is the correct qualifier for those cases that Yellowcard brings up. Better yet, a new claim and associated qualifier should be added when new goals are scored or matches played. --Izno (talk) 15:54, 17 March 2016 (UTC)

Constraints and people

Can anyone explain why we have constraints like "items with this property should also have sex or gender (P21)", "...should also have place of birth (P19)" and "...should also have date of birth (P569)" on properties, for example ORCID iD (P496), which are meant to apply to people? It means we have long lists of violations on each such property, which are nothing to do with that property. Can we remove them, or move them somewhere central? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:48, 16 March 2016 (UTC)

I think (but I am not sure) that this is the case, because there is no way of having constraints of the form "instances of X must have a property Y". This is a kind of a workaround. That said, I also dislike this very much, since it artificially bloat constraint violations of unrelated properties and makes it harder to solve those violations. --Srittau (talk) 14:10, 16 March 2016 (UTC)
I think the best solution would be to be able to add constraints to items. These constraints would apply to all instances of the item or sub-classes thereof. --Srittau (talk) 14:13, 16 March 2016 (UTC)
+1. All person should have a place of birth, date of birth, given name etc independent if they have an ORCID or any other identifier. So the constraint should be on human (Q5) and not on the property site. --Pasleim (talk) 14:20, 16 March 2016 (UTC)
Constraints to items does sound like the best solution, and makes more sense than the status quo which seems to be a work-around itself. Ajraddatz (talk) 19:03, 16 March 2016 (UTC)
Before some starts mass removing constraints from items: Please don't. I (and probably other) use these constraints to slowly improve the quality of certain subsets of items. You add a bunch of constraints, over time you clean up all the violations, you add some more constraints, etc. This way you can raise the quality and also make sure that it's not declining again (those would show up in the constraint report). Of course this only works if the number of restraint violations are reasonable to fix and it's reasonable to assume the data is available. I remember having a constraint for field of work (P101) on a lot of person properties. List of violations was huge and nobody fixing them. Makes no sense and was removed.
In case of ORCID iD (P496) I would remove the place of birth (P19) constraint (595 violations on a total of 885 items). I'm not sure about date of birth (P569), clicked some random items and about half of them at a date of birth on Wikipedia. 180 items so you could just try to fix it. Multichill (talk) 21:11, 16 March 2016 (UTC)
If the constraint for P21 fails and can't be fixed, it's likely that either there is some other issue: P31 is incorrect, main property was applied to an incorrect item, or the main property isn't really suitable for Wikidata or don't have as much information as we should on the topic. Obviously, users who like to work through a long list of items lacking this or that property can easily set this up on their user talk pages. Somehow I doubt the usefulness of trying to find places of birth of people who died ages ago.
--- Jura 21:29, 16 March 2016 (UTC)
@Multichill: We have to avoid to use general tools for personal purposes: we have the Wikidata query tool to perform that kind of analysis so we don't need to load the bot with unnecessary checks. Snipre (talk) 09:35, 17 March 2016 (UTC)
I fully agree with Snipre regarding this issue. --Casper Tinan (talk) 16:55, 17 March 2016 (UTC)
I'm describing a use case how I use it, this is not a personal purpose. These properties have constraints that should be improved, that's a general use case. Multichill (talk) 18:12, 17 March 2016 (UTC)
That reports of missing, but likely to exist, values are needed is not in dispute. But this is a clumsy and unhelpful way to achieve that.Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:25, 17 March 2016 (UTC)

Wikipedia article link from item

Hi. Is there any way to convert item to wikipedia article link? (e.g. Q13052182hy:Արիս Շաքլյան)--ԱշոտՏՆՂ (talk) 08:01, 17 March 2016 (UTC)

It depends what you are looking for. For a lot of items, PagePile can do that with its "filters", but it takes (or it took me) some time to figure out its options.
If you are just interested in one or two items, try {{#invoke:Wikidata|getLabel|entity=Q13052182|link=wikipedia|lang=hy}}.
--- Jura 09:01, 17 March 2016 (UTC)
Thanks--ԱշոտՏՆՂ (talk) 09:20, 17 March 2016 (UTC)

Draft glossary

Joe Filceolaire, who sadly died recently, had a Draft:Glossary in his user space. Would anyone like to finish and publish it? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:08, 14 March 2016 (UTC)

I'm not sure what there is to finish? I know I reviewed the glossary when he posted it, gave him some suggestions then, and I thought generally it was definitely an improvement. It will need translation work when it's posted. Is there a consensus that it's ready to be put in place? If I don't hear objections and it hasn't already been done I'll probably go ahead and make this change Thursday (17 March). ArthurPSmith (talk) 11:52, 15 March 2016 (UTC)
I am afraid that my concern is still valid. If we introduce this brand new text, we will lose 7 mostly complete translations a even more incomplete translations. --Jklamo (talk) 15:01, 15 March 2016 (UTC)
@Jklamo, Pigsonthewing: I've been going through the page comparing the text with the original - there is a lot new but there is a lot that is just moved around (and some not even moved, though often somewhat reworded) - I think the new text is at least 50% the same (other than reordering) as the original. Joe also kept most of the translation tags in the page in the right places, though new ones need to be added. Anyway, I think while replacing the text would mangle some of the existing translations it wouldn't completely lose them. I'm going to work on getting it closer to suitable for a drop-in replacement with better language tagging and ask you to look at it then and see what you think. I do think the organization and most of the content is an improvement so it would be a shame to lose it. Thanks ArthurPSmith (talk) 21:06, 17 March 2016 (UTC)

P217 - should it be broadened?

At Wikidata:Property proposal/Transportation#fleet or registration number there is a fundamental disagreement regarding whether inventory number (P217) is a suitable property to use to store the fleet and/or registration numbers of vehicles. Pigsonthewing believes that it is, despite that such use would breach almost every constraint on that property; Danrok and myself believe that p217 should remain for the accession and inventory numbers of GLAM and similar institutions' collections only and a new property should be created for fleet and registration numbers of vehicles. This is not going to be sorted without additional input and pinging the other supporters of the property proposal and leaving a message on the property talk page has not been successful at bringing that. Thryduulf (talk: local | en.wp | en.wikt) 12:25, 16 March 2016 (UTC)

To me this inventory number is meaningless without the collection or inventory it belongs to, so I don't think this is a real issue, actually they are all identifiers. If you just want to know if the collection is a GLAM collection, just add it in the item of the collection. A complex constraint then will do whatever you want. But again, this would be simpler if each of these collections had their properties. author  TomT0m / talk page 12:33, 16 March 2016 (UTC)
That is precisely what I did in the example I gave to Thryduulf: [3], in the discussion cited (albeit using the issuing body property; that can be changed if needed). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:39, 16 March 2016 (UTC)
[ec] Your subheading is misleading. This is not about broadening the definition of P217, but about applying the existing definition: "identifier for a physical object... in a collection", which is precisely what a fleet number is. There is currently nothing restricting it, as you imagine, to "the accession and inventory numbers of GLAM and similar institutions' collections only". As I have pointed out, constraints can be changed. Furthermore, I have already said, and repeated for clarity, "No prejudice against a separate property, if that is not deemed appropriate, for registration numbers", since registration numbers are a somewhat different beast (for reasons I have already given you, in the linked discussion. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:37, 16 March 2016 (UTC)
When every constraint on the existing property needs to be removed to allow the new values to fit, that is very much broadening the scope. The constraints and all the discussion around the property make it clear to everyone who has commented other than you that it is about a very specific type of collection. Thryduulf (talk: local | en.wp | en.wikt) 15:34, 16 March 2016 (UTC)
"clear to everyone who has commented other than you" See TomTom's comment, disagreeing with you, just above. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:37, 16 March 2016 (UTC)

(edit conflict) It's worth pointing out that pigsonthewing is now edit warring (see [4]) about this. Thryduulf (talk: local | en.wp | en.wikt) 12:40, 16 March 2016 (UTC)

I've been applying the decision reached at Wikidata:Property proposal/Transportation#fleet or registration number. You have been edit warring in an attempt to subvert that. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:43, 16 March 2016 (UTC)
(Directed at both of you:) Accusations of edit warring are not helping anyone. Let's try to focus on the content of the proposal, instead. --14:21, 16 March 2016 (UTC)
No, you have been applying your opinion in a discussion that is contrary to the consensus of those commenting before you and disagreed with by everybody who has commented since. You may wish it to be a decision that has the force of consensus, but it is not. Thryduulf (talk: local | en.wp | en.wikt) 15:34, 16 March 2016 (UTC)
"disagreed with by everybody who has commented since" I refuted that allegation the last time you made it, just above. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:20, 16 March 2016 (UTC)
I am supporting the creation of a property for a "registration number" (as issued by some external authority). Fleet number is a bit problematic, there is no concrete definition for it. Suffice to say it is normally issued by the owner of the item. Danrok (talk) 14:57, 16 March 2016 (UTC)
A fleet number is a unique identifier for a vehicle in a fleet of vehicles, issued by the owner or operator of the fleet or by the owner or operator of the system the vehicle is used on. A registration number is a unique identifier for a vehicle issued by a (usually external) authority on behalf of an administrative territorial entity. I see a large overlap between them but I'm happy with separate properties for the two. I do not see either as sharing any overlap with GALM inventory or accession numbers. Thryduulf (talk: local | en.wp | en.wikt) 15:34, 16 March 2016 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── At the danger of (necessarily) repeating myself:

A [fleet/inventory] number is a unique identifier for a [vehicle/object] in a [fleet/collection] of [vehicles/objects], issued by the owner or operator of the [fleet/collection]

HTH. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:33, 16 March 2016 (UTC)

Hi Andy, I've reverted your edit on that property and am making a new one for fleet number, which (as argued in that section) is clearly a different thing. Please stop edit warring now, thanks. Ajraddatz (talk) 16:41, 16 March 2016 (UTC)

User:Ajraddatz has now created P2616 (P2616), overruling my closure and while this discussion is still taking place. That's surely unacceptable? It's the Property creator equivalent of wheel-warring. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:48, 16 March 2016 (UTC)

Andy, I appreciate all of the good work you do here. But sometimes you take it way too far, when you have a particular view that you decide to push on everyone else. This is one of those cases. Discussion on that new property has been taking place for almost a month now, and you are the only person to express a dissenting view. You are clearly involved in the discussion, and shouldn't have been closing it in the first place. I could remove your property creator rights, but I think overall you still use them well. Please, take a step back, and maybe focus on different things instead of the continual arguing of these tiny points. Thanks, Ajraddatz (talk) 16:52, 16 March 2016 (UTC)
"you are the only person to express a dissenting view" Now you're making things up - and even if you were not, it would not make you right, nor what you did acceptable. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:57, 16 March 2016 (UTC)

@Ajraddatz: It is not acceptable to create the property at this moment in time. User:Thryduulf opened this topic to get more user input as there is a fundamental disagreement on Wikidata:Property proposal/Transportation. Five hours after this request many users have not even seen this topic so creating then the property is inappropriate. --Pasleim (talk) 17:24, 16 March 2016 (UTC)

Hi Pasleim, thanks for the second opinion. You're probably correct that I closed it too soon (as I assume you also think Pigsonthewing did), and I am more than willing to delete the property and allow for more public input. My concern here is that there is no actual discussion happening - there is one user forcing his opinion on everyone else, and that itself is preventing discussion, not promoting it. I've noticed a trend here recently of one loud user (not necessarily the same one) pushing an opinion until everyone else gives up. That isn't something that I like to see, and would prefer if Pigsonthewing (talkcontribslogs) would refrain from further textwall comments on this topic - he has made his view perfectly clear, and now it's time to step back and let others comment. Ajraddatz (talk) 18:56, 16 March 2016 (UTC)
You didn't "close it too soon"; it was already closed, and you overturned that. Yes, you should delete the property; and restore my "not done" closure. Your "only one user" accusation remains false. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:04, 16 March 2016 (UTC)
As I said above, and as you chose to ignore above, you had no business closing a discussion that you obviously have a strong opinion on. You also have not demonstrated a single other person on that discussion who shares your opinion. I repeat: please, stop forcing your opinion on others, take a step back, and let discussion happen. Expanded: T0mtom has generally agreed here, but not on the discussion page. There is still consensus there for creating the item, and your closure as "not done" is totally inappropriate. Ajraddatz (talk) 19:10, 16 March 2016 (UTC)
Poppycock. The only reason I explained - succinctly - why I closed the proposal as I did is because I was asked to. And noted you're back-pedalling on your bogus "only person to express a dissenting view" claim. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:15, 16 March 2016 (UTC)
I note that P2616 has now been deleted, but that the proposal has been marked as "In progress", not "not done". This is highly improper; doubly so while there is a discussion here. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:23, 16 March 2016 (UTC)
There is ongoing discussion. Kindly don't mark it as not done again, and refrain from further use of your PC flag to close discussions based on your own personal preference. Ajraddatz (talk) 22:35, 16 March 2016 (UTC)
The ongoing discussion is here. I cannot "refrain from closing discussions based on my own personal preference", because I did not do so in the first place; I acted in my capacity (but not using a "flag") as a PC to prevent a duplicate property from being created. Your reversal of that closure is what is ultra vires. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:27, 16 March 2016 (UTC)


On the main question here - given the other identifier properties we have, I'm surprised it hasn't been proposed to create separate identifier properties for each type of vehicle identification/registration system out there. For example vehicle identification number (Q304948) is a unique registration string for cars and other personal motor vehicles, wouldn't a property specifically for that be better than a general "registration number" property? ArthurPSmith (talk) 18:04, 16 March 2016 (UTC)

Just add my two cents as the problem is starting to reach a certain level of mess in different pages like Project chat, Request for deletion and Property proposal.
The proposed property is a mess because it is too general and can be used to covert items which have more specific properties: we have IMO ship number (P458) for boat and aircraft registration (P426) for planes, and the new property which can be used for boat, planes, cars, ... Definitively a general property should not be used because we started to use specific properties depending on the type of vehicles so we should keep the same philosophy in order to avoid bad use of the general property for items where a specific property exists. So we can keep this property but we have to define its use to one type of vehicule. Snipre (talk) 19:15, 16 March 2016 (UTC)
@Snipre: are you suggesting that as we have specific properties for boats and planes, we should have a specific property for rail vehciles, a specific property for road vehicles (etc?) rather than a generic property for vehicles that do not have a specific type available? If so, then I have no objection to that. Thryduulf (talk: local | en.wp | en.wikt) 13:08, 17 March 2016 (UTC)
@Snipre:I don't get your point. IMO ship number (P458) and aircraft registration (P426) are registration numbers issued by international external authorities. Fleet numbers are typically issued by the owner of the object. The fact we use distinct properties for the registration numbers issued by the different authorities that supervise transportation matters doesn't mean we can't have a general number for fleet numbers. --Casper Tinan (talk) 13:53, 17 March 2016 (UTC)
Casper Tinan Question: Are there other registration systems than the international ones for boat and airplanes ? Snipre (talk) 15:02, 17 March 2016 (UTC)
@Snipre, Casper Tinan: I don't know about aircraft but there certainly are for boats. this Excel spreadsheet published by Marine Management Organisation (Q6764168) gives "Port letters and numbers", "Registry of Shipping and Seamen number" and "Licence number" for each vessel on a list relating to the UK fishing fleet (I don't know any more about any of them though) and Canal & River Trust (Q5030960) issue registration numbers for vessels using the inland waterways in England and Wales (I would be extremely surprised if there wasn't an equivalent also in Scotland). Thryduulf (talk: local | en.wp | en.wikt) 21:44, 17 March 2016 (UTC)

Hi, there are many lang-links (interwikis) at local wikis which aren't removed. for finding them you should type insource:/\[\[xxx:/ (xxx is langlink like en,fr, de, fa,...) at local wiki search box like below:

At en.wikipedia

...

At fr.wikipedia

most of these local pages have interwiki-conflict and should be solved by human.

for bot

if you have bot at local wiki, you can run this code to remove interwikis of without-conflict pages (you should bot-permission at wikidata and localwiki):

#This command will check and remove interwikis from en.wikipedia's pages which have fr: links
python pywikibot/pwb.py interwikidata -clean -langs:en -lang:en "-search:insource:/\[\[fr:/"
#Only categories
python pywikibot/pwb.py interwikidata -clean -langs:en -lang:en "-search:category:insource:/\[\[fr:/"

Yamaha5 (talk) 06:22, 17 March 2016 (UTC)

Note that meta:User:YiFeiBot already removes unnecessary interwiki links from a lot of wikis. - Nikki (talk) 08:33, 17 March 2016 (UTC)
@Nikki: at fa.wikipedia it works but every week when I check enlinks we have many pages to work on! may be it should check more pagesYamaha5 (talk) 09:14, 17 March 2016 (UTC)
For anyone's attention, see also WD:Bot requests#ceb.wikipedia language link import. If I had a global bot, this could actually be a great fun, so I would ask their operators instead. Matěj Suchánek (talk) 19:30, 17 March 2016 (UTC)

merge problem

hello,

I would like to merge Q16975819 (Husab mine) with Q17028067 (Husab Uranium Project). When I have tried to use the gadget "Merge.js", it answer : "Error while "Please wait...": A conflict detected on enwiki: Q16975819 with enwiki:Husab mine, Q17028067 with enwiki:Husab Mine" could you help me please? --BalaiFaubert (talk) 09:48, 17 March 2016 (UTC)

As en:Husab mine redirects to en:Husab Mine, you could simply delete the redirect on Q16975819 and then merge the two.
--- Jura 09:56, 17 March 2016 (UTC)

Thanks Jura! (merci :-)--BalaiFaubert (talk) 19:51, 17 March 2016 (UTC)

Planned maintenance will result in downtime during April

The Wikimedia Technical Operations department is planning an important test of the new "full" data center in Texas. The test will result in about 30 minutes of downtime for all the wikis, including Meta, on two days that week. This work was originally scheduled for this coming week, but has been postponed until the week of 18 April 2016. The official schedule is kept on Wikitech; more information is at m:Tech/Server switch 2016. More announcements and notifications for editors are planned.

If you experienced problems with the five-minute read-only test on Tuesday, 15 March around 07:05 UTC, or if you have suggestions for places to announce this, then please contact me directly at w:en:User talk:Whatamidoing (WMF). Thanks, Whatamidoing (WMF) (talk) 01:56, 18 March 2016 (UTC)

Help for clean up

Could someone (ideally somebody who understands Chinese) clean up the contributions of 36.225.99.141? Items maybe need to be merged or at least need some claims. Thanks, -- T.seppelt (talk) 14:56, 18 March 2016 (UTC)

As this is English project chat, you might want to try Wikidata:互助客棧.
--- Jura 15:00, 18 March 2016 (UTC)

Recording property use in templates in sister projects

Please see discussion at Template talk:ExternalUse#Better values; or a property?. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:27, 18 March 2016 (UTC)

Image suggestions for paintings

Hi people, I made a little tool to suggestion images to be added to painting items. It seems to be only a tiny bit addictive. Please help to get more items illustrated here! Some instructions are at the top of the page. Thank you, Multichill (talk) 20:56, 16 March 2016 (UTC)

Hi Maarten. Sounds great, but the Antwerpen museum links are all broken for whatever reason.--Ymblanter (talk) 19:42, 18 March 2016 (UTC)

Deletion of P1887

Please see discussion of P1887, for vice counties of the United Kingdom (used in biological recording), at Wikidata:Administrators' noticeboard#P1887. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:06, 18 March 2016 (UTC)

Please have a look at Wikidata:Administrators'_noticeboard#Property creator rights of Pigsonthewing. --Succu (talk) 22:16, 18 March 2016 (UTC)
indeed, more eyes, familiar with property creation requests, would be useful there. Thank you. It has no bearing on the deletion of P1887 though. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:25, 18 March 2016 (UTC)

OTRS anyone?

Is there any accepted way to link to tickets in OTRS, or do we use reference URL (P854)? It's about a formal query to a government body and the reply about official written form of Norway (Q20) in other languages, Q20#P1448. (Ticket#2016031010010694) Jeblad (talk) 14:37, 17 March 2016 (UTC)

There was a discussion about OTRS and Wikidata on this page not so long ago, I proposed to create a property back then... Sjoerd de Bruin (talk) 17:04, 17 March 2016 (UTC)
I think an OTRS ticket id property is the best way to store it. Mbch331 (talk) 17:08, 17 March 2016 (UTC)
@Lydia Pintscher (WMDE): Any news about OTRS ? We had a discussion in January about the system to implement for OTRS tickets (see Wikidata:Project_chat/Archive/2016/01#CC-0 confirmation via OTRS but no planning was defined. Snipre (talk) 07:58, 18 March 2016 (UTC)
It is still waiting for someone to draft a default text similar to the one Commons has so I can take it to legal for review. --Lydia Pintscher (WMDE) (talk) 10:33, 19 March 2016 (UTC)
@Lydia Pintscher (WMDE), Jeblad, Sjoerddebruin, Mbch331: I started a property proposal there. Please provide for feedback in order to go ahead with that topic.
The second step id the creation of the page Wikidata:OTRS. Snipre (talk) 08:15, 18 March 2016 (UTC)
I added a comment. I think a property is the best solution, but I'm not sure if there is any use for this except in a source statement. Jeblad (talk) 19:16, 18 March 2016 (UTC)

Disney XD confusion

I'm confused about Disney XD (Q1858139) and Disney XD (Nordic) (Q3030487). The English description implies the former is about the international franchise and the latter is about the American channel, but actually, the latter seems to be used for Scandinavia as well (the enwiki and nnwiki sitelinks point to that one). nyuszika7h (talk) 17:51, 18 March 2016 (UTC)

Disney XD (Q1858139) is the Disney XD channel in the USA (US TV providers only in the info box). Whereas, Disney XD (Q1858139) is the Disney XD channel in Scandinavia and the Baltics. The Wikipedia links need to be fixed, some are linked to the wrong item, some are redirects. Danrok (talk) 00:49, 20 March 2016 (UTC)

p767 or p1706

At [5] it is noted that National Transportation Safety Board (Q1053145) will be assisting Interstate Aviation Committee (Q245076) with the investigation into Flydubai Flight 981 (Q23253749). I'm not sure whether it is best to record that with a qualifier of contributor to the creative work or subject (P767) or together with (P1706)? The latter might indicate the two organisations being equal partners working together, which isn't really the case here, but I'm not sure if the former is intended for this sort of usage? Thryduulf (talk: local | en.wp | en.wikt) 01:04, 20 March 2016 (UTC)

Religions genre?

Working on the WikiProject Religions, I'm searching for a way to tag generic properties of religions.

At the moment, the project enforces the use of subclass of (P279) for class of religions (for instance: Baptists (Q93191) subclass of (P279) Protestantism (Q23540)). This works well for religions that are identified as a derivative of broader religions.

But I don't know how to link religion to generic ideas like monotheism (Q9159). Very often, there is not a single origin for related religions (think of Abrahamic religion (Q47280) and atenism (Q146107)) and thus the use of subclass of (P279) seems misleading. If it's the correct property, should we use subclass of (P279) {Q|9159}} for every religion or only for the highest religion in the hierarchy that match it? If so, how to handle religions where such a generic concept change (e.g. atenism (Q146107))?

Video games may have the same kind of problems and have solved it with the use of genre (P136), should we use it or create a similar one for philosophical ideas?

Or should we use facet of (P1269)? It seems to be in the wrong direction: monotheism (Q9159) facet of (P1269) atenism (Q146107) (and of a lot more) instead of atenism (Q146107) "have for aspect" monotheism (Q9159) (which is cleaner). Maybe has part(s) (P527) or manifestation of (P1557)? But those ones seems to be too related to physical objects…

TL;DR:

Any advices, ideas or comments highly welcomed. — nojhan () 17:43, 13 March 2016 (UTC)

  Notified participants of WikiProject Religions

I think its a pretty bad risky idea to use subclass of (P279). I'd use
⟨ religion ⟩ subclass of (P279)   ⟨ belief ⟩
and instance of (P31) or something like that - defining a religion as a set of belief, and using subclass of only if the religion is really a subset of belief. I'd use based on (P144) (the french label seems to have moved in a bad direction) for religions who are schism of each other, this property is intended for example for "software forks". To classify religions I'd create subclass of religion, I'd more likely go on the metaclass level (see Help:Classification) and create classes like Polytheistic religion, with something like
⟨ Polytheistic religion ⟩ subclass of (P279)   ⟨ religion ⟩
, or
⟨ Polytheistic religion ⟩ subclass of (P279)   ⟨ religion ⟩
.
More words about subclassing religions: I'd see a religion as a set of beliefs. To be a subclass of (specialization) some religions would be a religion that shares all the properties of the parent religions, in particular all of the dogma in the case of dogmatic religion, because an instance of a subclass must share all the properties of the instances of the superclass. So, a superclass of Christianism must be the least common subset of all Christianistic religion like Roman Catolism or Anglicanism.
The "tokens" of Christianism (again see Help:Classification) would be events/periods similar to "The pope believes in Christianism". author  TomT0m / talk page 08:07, 14 March 2016 (UTC)
The French for based on (P144) did not move in a bad direction. See talk page for relevant constraints. separated from (P807) is the property you're looking for. I still think subclass of (P279) is the correct statement to make, but P807 is also useful (for representing something different e.g. the Puritans schismed from the Church of England).

I don't agree that "the pope believes in Christianity" would be a valid token; it would only be so in the narrow worldview where Christianity represents a set of beliefs, when it is much more than that. --Izno (talk) 12:36, 14 March 2016 (UTC)

Well, consider a religion you don't believe in then :) I think we're looking for something that can actually can be objectivated. Last time I check, the Christian creed (Q183770)      was a part of the cult, and its a proclamation of belief. So "the pope believes" is not wrong, as he probably said it itself many times. But nothing says that Wikidata has to be complete wrt. the subjects it represents, and that we can really represents fully what it is for a Christian to be a Christian. author  TomT0m / talk page 13:34, 14 March 2016 (UTC)

As it happens, I'm an atheist, so I am considering religions here that I don't believe in. #woops

The statement '"creed" = "set of beliefs" part of "religion" implies "belief of the Pop"e subclass of "religion"/"Christianity"/"Catholicism"' is not one that is true. --Izno (talk) 14:18, 14 March 2016 (UTC)

woops indeed, then what did you mean by "this is so much than that" ?
Good point, but I don't think that's correct per this reasoning, I'll refine my thoughts : I meant that someone who believes in a religion instanciate the religion as a whole, not each of the statements that belongs to the religion (like the creed). That a religion is a class of beliefs would imply that a subclass of that religion would be ... a subset of the beliefs and that the tokens would be individual beliefs, which I don't think is a good way to model things. Now define the class "christianism" as "someone (declare he) believes in the resurection at some point (and act more or less according to this belief)". Then a catholics is a christian (at that point), who share more beliefs related to catolicism specifically. If we define "someone (declare he) believes in the resurection at some point (and act more or less according to this belief) and that Maria is its mother and maria is a virgin and ..." then Catholicics are a subclass of Christian, and the class reflects the beliefs of catholics, and is clearly a specialization of the other class. The formulation "set of belief" was ambiguous, thanks for pointing it. I'll be a little more correct and precise then : "a religion is instanciated when someone (declares) believing in each beliefs that defines the religion". This beliefs forms a set that defines (partly?) a religion. We may add other properties to the definition. author  TomT0m / talk page 14:39, 14 March 2016 (UTC)
"someone who believes in a religion instantiate the religion as a whole, not each of the statements that belongs to the religion" ← This statement is utterly wrong in the case of religions. A religion is neither described by the literature in that way, it's far more complex and subjective. Even for simple religions like Catholicism, where a clear authority decide what the credo is, it is impossible to draw a clear line between beliefs that are part or not of the religion. Now think of Buddhism, were some studies have shown that the only common idea between all of its forms is four words (and I don't mean four ideas, I mean four words). Still, the literature keep calling all those religion "Buddhism", simply because a religion is more than a set of belief.
@Nojhan: Then we're entering subjectivity and the kingdom of individual PoVs. The only thing we can know is if someone claims to be of some religion, and how people define those religion, or to each definition they claim to belong. But in Wikidata I'm convinced that we must rely on known definitions, and that each significant concept and definition deserve its own item. Not sure we disagree :) In the end at least there seem to be an agreement than there is something all Buddhist have in common, despite their claims to be Buddhists. author  TomT0m / talk page
@TomT0m: I'm searching for an approach that is feasible in practice in order to be used (here, I'm trying to organize data to create a graphical timeline). If some ways are better than other in the sense that they will ease more use, I would be happy to use them. But if the way is so complicated that I failed to grasp it or if it demands tremendous effort to do something, then I will just do something else. So far, your explanations, while interesting from a philosophical perspective, fall in the second category (as far as I'm concerned). I would suggest you to work hard on a popularization tutoriel. — nojhan () 15:11, 16 March 2016 (UTC)
My point is that I cannot decide myself of what are the important classification details and, even if I could, it's infeasible in practice. Thus, I'm better not deciding and just using sources to build a simple hierarchy of classes, which is indeed a set of inconsistent subjective views but take into account a lot more of information than just the belief systems.
About the use of based on (P144) and separated from (P807), I have considered their use at first, but then came with a lot of example that do not match. For instance, Manichaeism (Q131165) claims to be based on (P144) Buddhism, but many scholars claim that one can hardly found any similarity with Buddhism in practice. Although it's commonly described as being separated from (P807) Zoroastrianism (Q9601), it rather claim being a syncretism (Q172904) that gather at least three religions. And this is just a simple example. In face of the difficulty of deciding whether a relationship is based on, inspired by (P941) (which are for works), separated from or even an school of (P1780); and the risk of having one of each relationship for each source, I've found it really more generic, simpler and easier to maintain to use influenced by (P737). — nojhan () 09:26, 15 March 2016 (UTC)
@Nojhan: P1310)
So I don't actually see a problem in how I want to model things. author  TomT0m / talk page 12:39, 15 March 2016 (UTC)
EDIT : Oh, and we have statement disputed by (P1310)   for modelling disputes on claims, so we even can add that some sources disputes the claims of the religion itself. author  TomT0m / talk page 13:16, 15 March 2016 (UTC)r
That's interesting, one more property I failed to find (and I've tried hard, I'm starting to think there is a property organization problem, here). I will see if I could make a recommendation with all those relation properties in the project. — nojhan () 15:11, 16 March 2016 (UTC)
I think that there is a problem since almost the very beginning :) author  TomT0m / talk page 11:52, 17 March 2016 (UTC)
Not all religions are beliefs or belief systems, or even involve them. --Yair rand (talk) 22:19, 14 March 2016 (UTC)
@Yair rand: This can be discussed, but that would need more details. author  TomT0m / talk page 12:39, 15 March 2016 (UTC)
Consider has characteristic (P1552) also. I think Christianity subclass of (P279) monotheism isn't that bad a statement.
I was looking for it, thanks. — nojhan () 09:26, 15 March 2016 (UTC)

Maybe it's more valuable to setup a parent class and use "has quality" e.g. "monotheistic religion" has quality "monotheism"; Christianity subclass of monotheistic religion). --Izno (talk) 12:36, 14 March 2016 (UTC)

⟨ Christianity ⟩ subclass of (P279)   ⟨ monotheistic religion ⟩
If christianity is a class of religion I would agree, but I don't think so. I think Christianity is a set of minimal belief shared by all Christian religion, including the belief in the resurection and in the old testament. Then Cathololicism is also a set of belief, and is just a special case of Christianity. So I'd note
⟨ Cathololicism ⟩ subclass of (P279)   ⟨ Christianity ⟩
and
⟨ Catholicism ⟩ instance of (P31)   ⟨ monotheistic religion ⟩
author  TomT0m / talk page 13:34, 14 March 2016 (UTC)

That seems to be your opinion. I don't think anyone shares it, seeing as those statements on those items have been fairly stable.

There are classes of Catholicism as well. --Izno (talk) 13:49, 14 March 2016 (UTC)

Its much more than an opinion, it is an opinion and the rationale behind it. I think the statements are stable because nobody really challenged them because of ... the lack of rationale behind it. I gave all the basis behind this reasoning : A religion is a set of belief/cult(?), marked as instance of (P31) religion. A set of belief should only be a subclass of other set of beliefs, because of the transitivity of the subclass relationship. Hence the subclassing beetween Catholicism and Christianity. Catolicism is a special kind of Chtistianity, right ? Only certain set of beliefs are religions, and something that is a special kind of christianity might not be a religion, so "religion" should be at the metaclass level, as explained in Help:Classification. What's your rationale ? author  TomT0m / talk page 14:14, 14 March 2016 (UTC)
First, a religion is not just a set of beliefs, it's more than that (think rituals, culture, history, etc.). Some religions don't even have a minimal set of common beliefs (e.g. Buddhism(s)). Second, the transitivity of the subclass relationship is a HUGE problem in the case of religions, because a LOT of religions commonly described as inheriting from other ones just wipe out some of their characteristics. For instance, atenism (Q146107) clearly is a subclass of (P279) religion of ancient Egypt (Q447131) but should clearly NOT inherit polytheism (Q9163), because it's a monotheism (Q9159)! And this kind of problems are LEGIONS among religions (using Catholicism as an example to derive a rationale is clearly not a good idea).
I have no idea whether an element should or should not be a class and/xor an instance, and I've seen countless debate around on that matter (man, you have to read this RFC about colors). But, after a lot of discussions, I've came to the conclusion that there is no such thing as an instance of a religion, except in the head of a single instance of a human. — nojhan () 09:26, 15 March 2016 (UTC
I'm one of the few who wrote on that RfC so I know what we are talking about. I also wrote Help:Classification which has very strong basis (philosophical like the token/type distinction, Punning in OWL2 to allow metaclassing, ... I agree an instance of religion is in the head of someone. This make a religion a class of token, actually, and the religion concept itself a metaclass in the sense of Help:Classification, in a loose sense, if a religion can be defined as the class of all of the tokens mentioned, and that we're aware of the tokens in Wikidata if people says publicly they do, and that all people that said it agrees that they have the same religion, then the extension of the concept of "religion" is the class of all these classes. author  TomT0m / talk page 13:16, 15 March 2016 (UTC)
@TomT0m: I could not praise (haha) more the need to have an informative and clear Help:Classification. But, it took me three readings of that page to start understanding what you just said, despite having read your comment five times. To me, this does not look like a feasible approach in practice. But let's pretend it's just a matter of having the approach written down more clearly. Again, I have no idea whether Wikidata should allow classes to be instances and to be honest, I don't really care, just let me know what the community decided and I will do my best to comply with it. However, it should be something that is feasible in practice, not a complicated set of beautiful-but-useless rules. At the moment, I'm still not convinced that your approach is simple enough to be used in my case. — nojhan () 15:11, 16 March 2016 (UTC)
@Nojhan: I'm all for someone helping me to make this page more clear. But from what I could observe, the reality is that every time someone asks something related to classification, every approach that do not follows those principles lead to problems. Trasitivity of subclassing often allows Paris to become an instance of "type of administrative division", which is absurd, when not using those principle. Every system should ba as simple as possible ... but not simpler. Here in Wikidata we have both to classify Paris and to classify concepts of the french administrative division like "commune", "region", the same for all countries, way more abstract. Having a system that allows absurdities like the one above leads to make more complex to write useful complex and meaningful constraint that allows to find (real) mistakes, besides very simple constraints that exists just because someone decided it has to be like this and actually try to enforce it at the risks of edit warring with others, but does not serve any other purpose (or very peculiar one). That's why I insist on having "meaningful" constraint and principles. And imho these principles are very meaningful. Even if it needs a (little) bit of thought at first to grasp, they gives answers and consistency. And a common vocabulary for everyone here that goes beyond the sum of all peculiarities. author  TomT0m / talk page 12:05, 17 March 2016 (UTC)

The English description of subclass of (P279) is "all instances of these items are instances of those items; this item is a class of that item." Applying a less controversial analogy, a Ford Mustang is a subclass of automobile model. It's really a matter of general perception of whether a Ford Mustang is an automobile model or not. Some cases can be controversial. For example, some US states classify the Can-Am Spyder as a motorcycle, and require that drivers have a motorcycle endorsement on their license, while other states classify it as an automobile (which means a much greater number of drivers are allowed to drive it). Jc3s5h (talk) 15:16, 14 March 2016 (UTC)

Mmm not sure I follow you. What do you mean by a "Ford Mustang" ? I'd say that the car on this photo   is an instance of the model Ford Mustang 1. "Ford Mustang" as a whole can then be considered either as the class of whole cars Ford sold under the trademark "Ford Mustang", then
⟨   ⟩ instance of (P31)   ⟨ Ford Mustang ⟩
and
⟨ Ford Mustang 1 ⟩ subclass of (P279)   ⟨ Ford Mustang ⟩
, or as the class of all Ford Mustang models in history, but then I'd rather label it Ford Mustang model with
⟨ Ford Mustang 1 ⟩ instance of (P31)   ⟨ Ford Mustang model ⟩
.
The controversy you mention can esily be solved by creating one class for each classifications, for example specific classes motorcycle (as defined in the US state whatever1) and motorcycle (as defined in the US state whatever2). author  TomT0m / talk page 15:45, 14 March 2016 (UTC)
EDIT: and of course
⟨ Ford Mustang model ⟩ subclass of (P279)   ⟨ car model ⟩
because each ford Mustang model is more generally a car model. author  TomT0m / talk page 15:55, 14 March 2016 (UTC)
I found all those arguments really interesting (I really do), but I fail to see how to match religions in that framework. Religions are very complex, subjective, vague, mixed set of ideas, that cannot be really objectivized. We have to rely on what sources are telling rather than on an imaginary perfect classification beyond reach. That's why I simplified the ontology to just use subclass of (P279), heavily supported with sources. — nojhan () 09:26, 15 March 2016 (UTC)
@nojhan: I doubt that sources actually use the subclass formalism and that their use matches the definition of subclass of (P279) in building a set of relationships/hierarchies of religions, what makes you believes that ? An example of such sources ? author  TomT0m / talk page 13:16, 15 March 2016 (UTC)
@TomT0m: Sources use that formalism every time they say that "Rinzai is a mahayana buddhism", in which you can spot three different categories. Note that, as I said before, the idea of a hierarchy is not suited for religions, but if it's about relationships, then it might be OK. Nevertheless, it's still ours to decide how to proceed. I think the key point of this whole debate is: should we stay close to sources (simple classes, like Buddhism > Mahayana > Rinzai) or decide ourselves of a better way (like your approach with metaclasses). So far, I personally think that the first option is easier to maintain, more neutral and less error-prone. — nojhan () 15:11, 16 March 2016 (UTC)
@Nojhan: less error-prone. I don't think so. Different sources might use different basis for their classifications. And, by the way, in the case of classifications in an extended sense of directed acyclic graph (Q1195339), not in a strict sense, and this is imposed actually by the definition of subclass of (P279)  . Because of the definition we use, a loop in the religion subclass relationship, like "religion1" subclass of "religion2" subclass of "religion 3" subclass of "religion 1" would imply that religion1 = religion2 = religion3. And that could happen in case of conflicted sources for example, which make very important to be able to specify the classification used by the source, and it's a feature of metaclassing to be able to do that. You mark a class as class instance of the classes used in this (or that) classification, and that allows us to know that this is due to conflicting PoVs amongst sources, and to filter the classification we want to use in a query quite easily while using the generic properties.
Secondly because you can't just import a classification in Wikidata without checking first if the criteria used for this classification is consistent with our properties, and if it is not it ight be necessary to create new properties to stay consistent overall. I'
In short : i'm all for staying close to the sources, but I don't think it makes unecessary to have the discussion we currently have. However, by experience checking the definitions used in the RIs of french wikipedia articles, they are very very often compatible with the generic classification principles I highlight here because these are deeply rooted and that modern philosophical and scientifical thoughts integrated those. It's less the case for enwiki one as far as my experience can see. Plus they are pretty compatible with our very day use of vocabulary, I would very much not be surprised to hear christiansm is a monotheism (class level using subclass) on one hand and christianism is a monotheistic religion (metaclass level using instance of) on the other in a regular conversation about religion. I very much doubt all of this is original work. It just need to be dugged out.
@TomT0m: I find it difficult to argue more on the abstract concept of an universal ontology (about which I must confess I'm kind of a disbeliever, most likely because of a professional bias). Now that I better understand your point and you better understand the specific context of religions, may I suggest you to consider rapidly sorting out a draft of an ontology about religions as you envisaged it? I think the project's discussion page would be a better place than this chatroom. We can discuss in french, also, if you prefer. — nojhan () 18:34, 17 March 2016 (UTC)

Monotheism

Answering the original question : let a Monotheism be a(n instance of) religion that includes in its set of beliefs I believe that there is a god, and only one. Then Judaism and Christianism are both subclasses of Monotheism, because every Christian and every Jew believes this by definition. {Ping|Izno|Nojhan|Yair Rand}} would not that make sense ? author  TomT0m / talk page 10:45, 16 March 2016 (UTC)

@TomT0m: It would not, IMHO. For instance, Judaism can also be a subclass of Canaan religion (from which it derivates), which is an henotheism. Catholicism is also described as a polytheism (with three gods) by islamic sources. See also the example of the Atenism (a monotheism derivating from a polytheism) above. Or Zurvanism (a polytheism derivating from a monotheism). Transitivity of inheritance cannot be easily applied, I'm afraid. — nojhan () 15:11, 16 March 2016 (UTC)
@Nojhan: Transitivity of inheritance cannot be easily applied I agree, that's why I advised in my first post to be careful about subclass of (P279) because, by definition, there is transitivity involved in such relationships. In your examples, some kind of forked from approach seem way more practical, apart some very generic belief classes like monotheism, theism, non-theistic religion ... Especially, I adivise against using subclass of for just the reason of based on. A is a sentence in a source deserves examination however. author  TomT0m / talk page 16:19, 16 March 2016 (UTC)
I came to the same conclusion and that's why the project ask to use subclass as a way to translate the global (history+philosophy+rituals+culture+…) classification (most commonly) used among scholars. For the based on relationships, I will try to build something consistent with the properties mentionned above. For the generic characteristics themselves, I think has characteristic (P1552) seems to be a very good option. — nojhan () 08:54, 17 March 2016 (UTC)
@TomT0m: Judaism and Christianity are both monotheistic religions, but not every adherent of those religions (or person who identifies as having them as "their religion") is monotheistic. It's complicated. --Yair rand (talk) 23:53, 16 March 2016 (UTC)
@Yair Rand: That does not matter, it's a PoV issue. Two points:
  1. We cannot say
    ⟨ someone ⟩ instance of (P31)   ⟨ monotheism ⟩
    in the model I propose, because the instances are not human by themselves.
  2. What we say by
    ⟨ someone ⟩ religion Search ⟨ Catholicism ⟩
    is that someone claims to be a catholics for example, that's We can't know if deep down he actually believes the catholics dogma. And if he claims he believes there is several gods, he's likely to be challenge as a catholics by someone else (as far as we can be aware of).
I think the cultural or community appartenance are actually a different topics, and actually we have a more relevant property for this now. author  TomT0m / talk page 10:38, 17 March 2016 (UTC)
I don't think this is a PoV issue. Your premise is incorrect; the statements "Bob's religion is X", "X is a monotheistic religion", and "Bob is not a monotheist" are not always contradictory. For example, according to Pew, 15% of the population of Israel are simultaneously Jewish by religion and atheistic. Granted, certain religions probably have internal rules for determining who is part of that religion, which can include certain beliefs, which could result in PoV conflicts where different groups would have different points of view regarding a person's religious affiliation, but that is not inherent in being a monotheistic religion.
Some religions, such as (to the best of my knowledge) all of the Abrahamic religions, have associated belief systems. Alternatively, one could say that they are belief systems with associated practices, values, rituals, affiliation, rules, scripture, history, etc. Modeling any of these components as the "core" component with which subclass of (P279) can be assigned as proposed comes with a lot of problems. Religions are complicated. --Yair rand (talk) 00:37, 18 March 2016 (UTC)
Upon further reflection, I'm not sure my previous comment was accurate. Requiring the belief set as a prerequisite to being part of a religious group may indeed be a property inherent to any monotheistic religion. Or not. I'm not sure... :/ --Yair rand (talk) 12:41, 20 March 2016 (UTC)

About Skin Diseases

Hello we should have more information about Dermatology. in this,there are less and concise. Sunn Moh Orange (talk) 11:50, 20 March 2016 (UTC)

See the above RFC for some proposals to reform the property creation process. MediaWiki message delivery (talk) 06:29, 20 March 2016 (UTC) (on behalf of Rschen7754)


Note: I've restored this section, which was archived by User:Hazard-Bot, less than half an hour after it was posted. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:07, 20 March 2016 (UTC)

Russian speaker needed

Could someone who reads Russian please translate the one-line description at Wikidata:Property proposal/Person#place of detention? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:27, 19 March 2016 (UTC)

This is just "place of detention"--Ymblanter (talk) 20:31, 19 March 2016 (UTC)
@Ymblanter: Thank you. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:15, 20 March 2016 (UTC)

Hazard-bot issue

Per my note in the preceding section, User:Hazard-Bot is misbehaving. Has it archived the RfC from other pages? Should it be blocked, until fixed? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:14, 20 March 2016 (UTC)

RfC from other pages? Don't know what do you mean but I don't think so. And should not be blocked. I think the bot couldn't read the date correctly because there was text (on behalf...) after the signature. --Stryn (talk) 14:24, 20 March 2016 (UTC)
Sorry: "...RfC notice from..." - again, per my note in the preceding section. thanks for spotting the cause. If the bot is prematurely archiving such notices, which are clearly important, it shoudl not be allowed to continue. Though perhaps the bot issuing the notices should also be tweaked? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:49, 20 March 2016 (UTC)
Ugh, why can't you just talk at one place? Please see User talk:Hazard-SJ for more. Sjoerd de Bruin (talk) 19:04, 20 March 2016 (UTC)

see Ron James (Q7363976) and Ron James (Q23071382) pointing to the same page en:Ron James (footballer, born 1970). Bug or Feature? -- Gymel (talk) 14:29, 20 March 2016 (UTC)

My understanding is it shouldn't be possible to have two items with the same sitelink. It looks like when the page moved on enwiki, it changed the sitelink on wd, but somehow made it look like the page didn't have a corresponding wd item, so Jura1 created a new item for it. I've merged the items, as they are clearly about the same person. I'm sure there is someone who can fill us in on the technical details. There's a bug report here that seems relevant: [6] Silverfish (talk) 17:00, 20 March 2016 (UTC)
Not sure if it's the same bug. Somehow the pagemove didn't happen as it should. Maybe @Lydia Pintscher (WMDE): can help making sure it gets sorted out.
--- Jura 17:32, 20 March 2016 (UTC)
Thanks for the note. These weird things sometimes happen. The underlying issue seems to be that the database table we're using to make the check if a sitelink already exists is overloaded. We're doing too much with it. We're working on migrating away from it. That will hopefully solve it. --Lydia Pintscher (WMDE) (talk) 08:59, 21 March 2016 (UTC)

Why P and Q?

Can anyone direct me to a link saying why "Q" and "P" were chosen as Wikidata prefixes for items and properties?

  1. Does "P", the prefix for properties, stand for "property"?
  2. Does "Q", the prefix for items, stand for anything?

I have heard the rumor that Denny was influential in naming these and might have made the decision. There is another rumor that "Q" stands for "qittens" and reflects the Wikimedia movement's longstanding commitment to uplift the profile of cats. Who has a story? What has been published on the matter?

Thanks. Blue Rasberry (talk) 16:53, 21 March 2016 (UTC)

Read A personal note, and a secret from Denny. ;) --Succu (talk) 17:14, 21 March 2016 (UTC)
Amazing, great story, thanks. Blue Rasberry (talk) 18:06, 21 March 2016 (UTC)
Agree about great story. Good story to tell to your grandchildren, when Wikidata will be very, very cool :) --Edgars2007 (talk) 18:19, 21 March 2016 (UTC)

Hi, I just used the lua API call the devteam introduced to lookup a property by label in Module:Properties , (thanks to them), so in templates {{P'}} and {{C}} we can now use the enlgish label of the property and the corresponding links will appear.

author  TomT0m / talk page 18:55, 21 March 2016 (UTC)

Get multilingual Description of a Property

Hi, How can I get the Description of a Property in other language? For example, in the attached snapshot the Property Brother has Arabic Description 'أخ". From schema point of view, how is this Property linked to Description?

Thanks, Abed.  – The preceding unsigned comment was added by Fattahsafa (talk • contribs) at 08:52, 12 March 2016‎ (UTC).

Descriptions are entered largely with a bot. As they are stuck, not much happens to them and often the information is based on too little. Descriptions are awful in this way. Compare it to automated descriptions, they are dynamically generated and are available in any language giving sufficient labels. You can wonder how they fit in from a schema point of view but the content of descriptions is flawed in too many ways to take them seriously. GerardM (talk) 09:32, 22 March 2016 (UTC)

Recent changes

Recent changes (https://en.wikipedia.org/wiki/Special:RecentChanges) provides the last 500 changes and changes made in the last 30 days. Is it possible to go back thousands of changes and months or years? Thank you very much in advance!

Yes. If you select the 500 and 30 options, you'll notice that those numbers appear in the URL. Simply change them to however many changes / however many days you want :) Ajraddatz (talk) 21:44, 16 January 2016 (UTC)
Thank you, I already tried before what you suggest. But I got the impression that this does not work. If it worked, there would be 1.000.000 edits visible here https://en.wikipedia.org/w/index.php?title=Special:RecentChanges&days=1000&from=&limit=1000000. Any idea how to fix this?
Do you have a specific task or project behind your question? --Voll (talk) 07:51, 18 January 2016 (UTC)
Afaik, the limit for recent changes is 5000 changes, as for the search results pages. I personally feel this as a bit limiting, too. Maybe result pages of more than 5000 entries would actually be getting a bit too large, but a pagination which would allow more than 5000 entries in total (e.g. show 500 entries by default, but offer a "Next 500" link, like it is done for e.g. the search results, and allow that for more than 10 times) would be very helpful for patrolling vandalism etc. --YMS (talk) 08:25, 18 January 2016 (UTC)
No, it is limited to only 30 days, number of changes isn't limited. At least, it should be like that (about number of changes). --Edgars2007 (talk) 09:01, 18 January 2016 (UTC)
The recent_changes database table (or whatever its exact name is) contains all changes of the last 30 days, but Special:Recent Changes will deliver 5000 entries max. This is no limitation within the available data (the special page just reads said table), this is just a limit hardcoded in the special page itself. As there is no pagination, there is no way to see the next 5000 entries. If you want to see any changes older than thevnewest 5000 there, you will have to set up some more restrictive filters (or query the database with some external tools instead). --YMS (talk) 06:59, 23 March 2016 (UTC)
@Voll: I am a journalist and need as much data as possible to conduct data journalism. @all: Thank you. Who is able to change the 30-day-limit? What must I do to get more data?
Wikidata:Data access --Molarus 17:58, 15 March 2016 (UTC)
The 30-days-limit affects, among other things, CheckUser and therefor also our privacy-policy. It will therefor not easily be changed. But as the Data access-page linked above tells, there are other ways to get the information you are looking for. -- Innocent bystander (talk) 08:03, 16 March 2016 (UTC)

Just to notify that @Matěj Suchánek: have created a script to check if sitelink are disambiguation or redirect. To activate the script add in yours commons this row:

mw.loader.load( '//www.wikidata.org/w/index.php?title=User:Matěj Suchánek/checkSitelinks.js&action=raw&ctype=text/javascript', 'text/javascript' );

so you add a link called "Check sitelinks!" in sitelink section. If you click on it the script shows with icons if are disambiguation or redirect. Very useful. Thanks to Matěj --ValterVB (talk) 16:41, 20 March 2016 (UTC)

As I understand it redirects are allowed.. So what is the point? Thanks, GerardM (talk) 10:17, 23 March 2016 (UTC)

Help Please on wikidata page Harmonices Mundi

Hello you all wizards, the Authority control on pages related to Harmonices Mundi , and Mysterium Cosmographicum by (Johannes Kepler) adds a blank/dead Worldcat link. Cannot understand where the problem comes from. Can you help?--DDupard (talk) 12:35, 21 March 2016 (UTC)

@DDupard: Where are these pages? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:33, 21 March 2016 (UTC)
en:Harmonices Mundi and en:Mysterium Cosmographicum, see the "authority control" at the bottom of those sites and click "WorldCat". The links are not working. I don't know where's the problem, because I don't know anything about those sites how they should work. --Stryn (talk) 17:13, 21 March 2016 (UTC)
Hello @Pigsonthewing: , pages are: Harmonices Mundi (Q1066757) , Mysterium Cosmographicum (Q721138), it does not work on either the English or French Pages....--DDupard (talk) 18:10, 21 March 2016 (UTC)
@DDupard: Ah, we usually refer to those as "items". You can link them like {{Q|1066757}}, which renders as: Harmonices Mundi (Q1066757) (showing the label for your preferred language). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:22, 21 March 2016 (UTC)
@Matěj Suchánek:, Can you help?--DDupard (talk) 21:50, 22 March 2016 (UTC)
@Mbch331: Thanks for looking into it. Other question, this one about Infobox Biographie2: Sometimes it calls for illustration, some time it does not; for example on WP.fr pages Joep Franssens (Q2735633) and William G. Whittaker (Q2798389) . What is the reason for the difference? thanks again--DDupard (talk) 05:54, 23 March 2016 (UTC)
I don't know the template itself, but there are 2 possibilities: 1) The pages where there's an image, the image is added to the infobox itself as a parameter. 2) The template queries for image (P18) on Wikidata and if that's missing returns a general placeholder (as on the mentioned pages). The infobox can't guess for an image, it has to be defined somewhere what image to use, either as a parameter on the template, or as a statement on Wikidata. Mbch331 (talk) 06:51, 23 March 2016 (UTC)
@DDupard: What do you want me to help with and why are you asking me? Matěj Suchánek (talk) 14:54, 23 March 2016 (UTC)
@Matěj Suchánek:: Questions are stated above, and since you are an active participant on wikidata, thought you might be able to solve the problem.--DDupard (talk) 15:20, 23 March 2016 (UTC)
See Jura1 and Mbch331's investigation. Matěj Suchánek (talk) 15:28, 23 March 2016 (UTC)

What's the best way to manage uncertain family links ?
By uncertain, I mean the following use cases I have seen in history books:

  1. unsupported relationship, now regarded as plainly wrong
  2. dubious relationship, but still regarded as possible
  3. hypothetical relationship reconstructed by historians

Case 1 is easily managed with the deprecated rank, but I am not sure about other cases.
When working with wikidata, I have started to use sourcing circumstances (P1480) as a qualifier with the following values :

At the same time, I have seen an alternate way being used, e.g. Richard (Q4394904). Uncertain children and parents are all put as relative (P1038), with a specific type of kinship :

I think we should decide upon one solution or the other before they get too heavily used.
What do you all think ? Which is the best way ? --Melderick (talk) 18:02, 23 March 2016 (UTC)

Thanks for posting! I think you should set up a Wikidata:Family relationships page in which these are listed, because I often wonder what to use and can't remember how some of these are supposed to work. The property talk pages should point to that page as a reference. Jane023 (talk) 18:18, 23 March 2016 (UTC)
I created Wikidata:WikiProject Parenthood recently. This topic might be interested for the project. Paweł Ziemian (talk) 16:52, 24 March 2016 (UTC)
may be father (Q21152551) etc don't make sense, in my opinion. sourcing circumstances (P1480) seems more workable. --Yair rand (talk) 18:15, 23 March 2016 (UTC)
If you set father (P22) with sourcing circumstances (P1480) = hypothetically (Q18603603), you can have a person that have two fathers (example : Arnulf (Q337189)). It doesn't agree with Wikidata:Database reports/Constraint violations/P22 ("Single value" violations). It's why I prefer using relative (P1038), with a specific type of kinship. Odejea (talk) 20:21, 23 March 2016 (UTC)
Using relative (P1038) has some drawbacks:
* It makes it harder for automatic tools to retrieve data to build family trees
* Hypothetic children are separed from known children, especially in items with a lot of statements. Someone could forget checking relative (P1038) and add hypothetic child twice.
* Sure it doesn't violate Father's "single value" rule, but they can't have a chance to violate "male sex gender". The "inverse" rule is partly checked (can't verify the type of kinship)
Melderick (talk) 08:20, 24 March 2016 (UTC)

Is there a property for a "blurb"?

Books and movies have blurbs written about them; short summaries that are generally promotional by nature. Is there a property for this? The text amount is about 50-100 words. Can these go on Wikisource? Thx. --Jane023 (talk) 18:22, 23 March 2016 (UTC)

Those are copyrighted, aren't they? – Máté (talk) 18:31, 23 March 2016 (UTC)
These are copyrighted and not licensed compatibly for Wikidata, much less Wikimedia. --Izno (talk) 19:15, 23 March 2016 (UTC)
Generally yes, these would normally be copyrighted if they are more recent than 70 years after an author's death etc., but if they are out-of-copyright, we can use them. I am specifically asking here in the case of the TED talks, which have descriptive summaries that have been donated as part of their data donation (and have been translated into several languages). I would like to use these, but they are not suitable for our short descriptions. These are the short summaries that appear on their website when you follow a link to a ted talk. --Jane023 (talk) 20:01, 23 March 2016 (UTC)
Aside from the specific case mentioned by Jane: No, they are not necessarily copyrighted. They may, for example, be copyright-expired, ineligible for copyright (work of the US federal government, for instance) or under an open licence. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:24, 23 March 2016 (UTC)

An example is Manu Prakash: A 50-cent microscope that folds like origami (Q22938728): "Perhaps you’ve punched out a paper doll or folded an origami swan? TED Fellow Manu Prakash and his team have created a microscope made of paper that's just as easy to fold and use. A sparkling demo that shows how this invention could revolutionize healthcare in developing countries … and turn almost anything into a fun, hands-on science experiment." It is used here and is available in 12 languages. There is also a paper about it: Foldscope (Q15935848). --Jane023 (talk) 08:14, 24 March 2016 (UTC)

They don't need to, as this is part of a massive data donation and they fully understand the consequences. We were hoping to get the full translated transcripts (if you notice the entire talk is translated in the same languages) but this they could not do. We also don't have the stills from the talks. But the blurbs are good I think. I just don't know where to put them. --Jane023 (talk) 08:44, 24 March 2016 (UTC)
Where does it say that is part thereof? Is there a ticket?
--- Jura 08:49, 24 March 2016 (UTC)
See Wikidata:TED. What do you mean by a ticket? --Jane023 (talk) 08:52, 24 March 2016 (UTC)

Property for recipient of correspondence?

In Letter regarding indictment and departure of Devyani Khobragade to Judge Scheindlin (Q23308035) I would like to add the name of the judge who received the correspondence, though cannot fathom the property to use? Any ideas? Not even sure which of the projects would have this under their ægis.  — billinghurst sDrewth 23:37, 20 March 2016 (UTC)

The best I can think of is
although that is contrived and I don't like it! Thryduulf (talk: local | en.wp | en.wikt) 02:15, 21 March 2016 (UTC)
In lieu of any other suggestions or recommendations of where to refer the matter, that will be the means.  — billinghurst sDrewth 23:16, 21 March 2016 (UTC)
You might wish to propose a new property, but suggesting this means as an alternative, and see which gets the consensus - see Wikidata:Property proposal/Transportation#route map where I did something similar. Thryduulf (talk: local | en.wp | en.wikt) 00:15, 22 March 2016 (UTC)
@Thryduulf: At this point of time, the politics of this place has me at the point of ambivalence, and the apparent stratification of effort, the approach, mememe-ism, etc. will not have me adding to the bureaucracy. I will continue to add properties to WS items, and to add authors, I am refusing to enter the fray. If someone does create it, then hopefully someone will let me know that we have a correspondence recipient term.  — billinghurst sDrewth 10:36, 25 March 2016 (UTC)

Conversion of datatype to external-id: process is stuck

The process for deciding which properties to convert to the "external-id" datatype seems to be bogged down. It beggars belief that the community has yet to agree - after several weeks - that ISNI (P213) (an ISO identifier!), CERL Thesaurus ID (P1871) and P1946 (P1946) are to converted.

Despite good-faith best efforts by a number of editors, I think the format of User:Addshore/Identifiers and its sub-pages is partly to blame; it's not easy to see the status of an individual property, nor understand what some of the objections, if any are.

I suggest we hat those pages, and list all the properties that remain to be converted in a new page, with each as a sub-heading, sorted numerically. Comments can then be made, on each individually, and each section archived once consensus is reached (and enacted if necessary). I'm open to alternative suggestions. CC User:Addshore as a courtesy. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:09, 22 March 2016 (UTC)


Redux

I suggest we hat those pages, and list all the properties that remain to be converted in a new page, with each as a sub-heading, sorted numerically. Comments can then be made, on each individually, and each section archived once consensus is reached (and enacted if necessary). I'm open to alternative suggestions. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:04, 25 March 2016 (UTC)

There are still some 600 properties with string datatype. It's quite unlikely that you get more user input on a page with 600 sub-headings than you get now. There are also many unchecked properties on User:Addshore/Identifiers/1. I think we can apply your approach as soon as we are down at lets say 100 disputed items. --Pasleim (talk) 13:27, 25 March 2016 (UTC)
Indeed, 600 would be too many for that approach. I hadn't realised the scale of the problem. How else might we speed things up? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:33, 25 March 2016 (UTC)
Start by checking properties you proposed and/or created.
--- Jura 13:46, 25 March 2016 (UTC)

Hello everyone,

at least since a couple of weeks IPv6 user (e.g. 2003:66:8F38:2C28:2596:DA18:2364:DD08, 2003:66:8F38:2C28:21BA:F695:E00A:CB32) are adding film related information (e.g. Danish National Filmography person ID (P2626), ČSFD person ID (P2605), Filmportal ID (P2639)). The edits happen at frequent intervals and from several accounts. I didn't come across any wrong information. I decided to mass-patrol the edits which is probably not the best solution. When we block the accounts due to bot policies we would maybe loose the potential. Talking to the user doesn't seem to work. I left a message on a talk page with a request to create an account but nothing happened. Would do you think should we do? -- T.seppelt (talk) 08:56, 24 March 2016 (UTC)

To me, those look like regular edits (not using some automated tools). If you have good system (way to add statements), the frequency of edits can be like that. --Edgars2007 (talk) 09:02, 24 March 2016 (UTC)
Found one more, 2003:66:8F55:EA78:3C5E:14CF:EAA8:DE37. Check the location. In Bavaria (Q980), only movie-related. 2min interval, 1 edit / item and in the UTC+2 night. It's just hard for me to believe that this edits happen manually. -- T.seppelt (talk) 09:24, 24 March 2016 (UTC)
I'm not saying, that it's absoluterly not some bot, what I'm saying, is that those can be manual edits. My method (in big steps): create some wikitable with many, many links (to WD Qx item and Google search "name site:site"), then open some 100 tabs or whatever and have some fun. Of course, this method is not for IT people :) And UTC+2 isn't so important. You can't forbid people from editing Wikidata at (their local) night :) --Edgars2007 (talk) 09:45, 24 March 2016 (UTC)
If it's a pure bot it wouldn't add descriptions for new created properties [7]. And as long as there are no bad edits, and the editing speed is kept quite low (around one edit per minute) I don't see a reason to block this IP user. --Pasleim (talk) 10:17, 24 March 2016 (UTC)
This isn't a bot. This is an anonymous German speaking user who doesn't wish to create an account – we can't force him to, either. Jared Preston (talk) 10:35, 24 March 2016 (UTC)
No, it does not look like a bot or even an automated tool, then the edit comments would have looked otherwise and the use of Properties been more consistent. Maybe we can set a bot to patrol the edits as soon as the ip is recognised? We had a system for that on svwiki. The username/ip was added to an edit protected page and a bot was patrolling according the usernames/ips in that list. -- Innocent bystander (talk) 10:48, 24 March 2016 (UTC)
Okay, you convinced me. Such a system would be good I think. It would be good to write a Wikidata equivalent for Wikipedia:New pages patrol/patrolled pages (Q11302303). -- T.seppelt (talk) 11:43, 24 March 2016 (UTC)
When I looked into these IPs late last year, they edited every single hour of every single day for 9 days straight. Maybe some of the edits are done by a human, but it seems like a lot are done by a bot. I really think that huge numbers of edits like these should be done under a user account, so that we have someone to contact when there are issues (I have seen people report issues on talk pages several times, but who knows which IP address they were using at the time). - Nikki (talk) 10:29, 25 March 2016 (UTC)

multilingual discussions project

How about a project maintaining a list of volunteer translators in some pair of languages that could help people of different mother language in a discussion beetween contributors ?

maybe we could create a template "call for translators" {{Cft|ru|fr}} {{Cft|ru|en}} ? that would work like {{Ping project}} but would use pages like WikiProjectTranslators/fr/en for example ... Any support/ideas ? author  TomT0m / talk page 15:30, 24 March 2016 (UTC)

Interesting idea. Is there already something like that for Commons? Is so, maybe we could re-use that. --Jane023 (talk) 06:50, 25 March 2016 (UTC)

Described by Armenian Soviet Encyclopedia

I've seen lots of statements like the one (in this case by User:ԱշոտՏՆՂ, but I haven't looked who added who made the rest): this edit, which adds described by source (P1343) > Armenian Soviet Encyclopedia (Q2657718), with no article title, page number or other ID. I'm not convinced that's useful. Is it? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:35, 17 March 2016 (UTC)

I don't think it's useful, no. Wikidata has quite the referencing problem, and I think part of it is the lack of adequate documentation and easy ways of adding references. A possible fix would be a built-in feature like what is used to cite on the Wikipedias, where it then autofills in the proper values on the item, but I lack the technical knowledge of how to implement that. Ajraddatz (talk) 22:12, 17 March 2016 (UTC)
This has nothing to do with referencing,. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:02, 18 March 2016 (UTC)
Fascinating. My comment on referencing has... nothing to do with referencing. Perhaps I misread your initial comment. Ajraddatz (talk) 07:06, 21 March 2016 (UTC)
Imagine how many encyclopedias may be added in P1343 of items like United States of America (Q30), Napoleon Bonaparte (Q517) or Winston Churchill (Q8016), etc. Hundreds and hundreds, most of them - minor and almost unknown to the general public, like is Armenian Soviet Encyclopedia (Q2657718).
Not useful. P1343 needs some restrictions in use (i.e. to not use in items about people and countries). --XXN, 12:31, 18 March 2016 (UTC)
Hi, Armenian Soviet Encyclopedia is the biggest Armenian language encyclopedia. It has 30000 articles. I also added volume[8] and can add page too. Also see Wikidata:Project_chat/Archive/2015/12#Spam.3F_P1343:_Armenian_Encyclopedia_.28Q2657718.29--ԱշոտՏՆՂ (talk) 13:46, 18 March 2016 (UTC)
It may be reasonable to allow to show/hide references depending on user languages. Same with language-specific identifiers. ---EugeneZelenko (talk) 14:18, 18 March 2016 (UTC)

I did ask this question some months ago (calling it spam). I am still not convinced it is useful. Maybe only when at least edition number is added, so we can check if it the claim is true. Michiel1972 (talk) 16:58, 18 March 2016 (UTC)

I am not happy with this property either; it may well swamp many items completely, while offering little value. This has nothing to do with data; perhaps time for a Wikibibliography-project? - Brya (talk) 17:27, 18 March 2016 (UTC)
I can add volume and page number in almost 12000 items. It is also possible to add wikisource url[9]. Wikidata is multilangual project and Armenian Soviet Encyclopedia is one of the most interesting sources for Armenian language users--ԱշոտՏՆՂ (talk) 18:47, 18 March 2016 (UTC)
The Wikisource URL would make this far more beneficial. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:52, 18 March 2016 (UTC)
So what must I do? remove this statement or add page, url? I think there is no problem with notability. Page number and url will make this much more helpful--ԱշոտՏՆՂ (talk) 07:58, 20 March 2016 (UTC)

@Wikidatans, So how many entries can we add in P1343?
User0001 will add P1343 > Armenian Soviet Encyclopedia (Q2657718)
User0002 will add P1343 > Azerbaijani Soviet Encyclopedia (Q2634919)
User0003 will add P1343 > Kyrgyz Soviet Encyclopedia (Q1188287)
...
User0203 will add P1343 > Tajik Soviet Encyclopedia (Q3567092)
User0251 will add P1343 > Uzbek Soviet Encyclopedia (Q2607783)
and so on.
Are we ready to accept right now several tens > to hundreds of entries in P1343?--XXN, 11:16, 20 March 2016 (UTC)

So what are you suggest? Delete P1343? Add only big encyclopedia? how many articles must the encyclopedia have to become big?--ԱշոտՏՆՂ (talk) 12:26, 20 March 2016 (UTC)

As I believe I stated in the property proposal for this property, it should never have been created, precisely for the issues being brought up now. --Izno (talk) 16:51, 20 March 2016 (UTC)

@Izno: You may well be right in your view, but in fact it seems there were no objections to its proposal. You could always nominate it for deletion. An inverse property may be the solution. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:01, 20 March 2016 (UTC)
Being this web a database we shouldn't have problems with... data. What is that about swamping items? The interface should be developed in order to hide/fold properties with many statements? ok. But deleting data/properties? This property allows some pretty cool lists and, of course, provides useful data, what it is all about, not about having pretty items. Strakhov (talk) 19:29, 20 March 2016 (UTC)
The problem is that a lot of claims makes the pages slow to load. On entities about encyclopedias, this is probably okay since there probably won't be many other claims. On anything else? --Izno (talk) 21:19, 20 March 2016 (UTC)
This is a general issue; one that we shall have to grapple with sooner or later, no matter what happens in this particular case. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:57, 20 March 2016 (UTC)
Yes, the slowness-to-load of pages is something that almost certainly is going to get worse, accompanied by a decrease in ease-to-edit. But that does not mean we should go full steam ahead to swamping (and sinking) items with 'data' which adds no information on the subject of the item, but rather belongs in a separate project with bibliographic material. - Brya (talk) 06:26, 21 March 2016 (UTC)
It is also noteworthy that the Encyclopædia Britannica Online has found a way to spam itself, by having lots of URL's on topics where it does not have an actual entry, but where it has just gathered a few sentences or phrases that it hopes do bear on the topic, somehow. - Brya (talk) 06:36, 21 March 2016 (UTC)

Will it be possible to make item like a book? So we can thumb the pages of an item :)--ԱշոտՏՆՂ (talk) 16:43, 21 March 2016 (UTC)

It is possible to have a new space (adding an extra page to each item), but this will be a big undertaking. - Brya (talk) 06:55, 22 March 2016 (UTC)

The number of sources can be grown up to ∞ by user A and fallen to zero by user B. Approval of community to any desition is OK to me, but extreme solutions are very doubtful. The significance of Armenian Soviet Encyclopedia (Q2657718) as source of information (Q3523102)/historical source (Q15622636) in any page can decided one by one. We have reliable source in hy.wikisource under CC 3.0 license. Adding page/volume/article title can improve verifiability. I would ask ԱշոտՏՆՂ to do it. - Kareyac (talk) 18:17, 26 March 2016 (UTC)

I am only vaguely familiar with what Wikidata does or how it operates, so please bear with me. I am trying to add an Interwiki language link to the English Wikipedia "Template:OTRS ticket" page and am failing. The corresponding page on the French Wikipedia is "Modèle:OTRS". I clicked on the "Add a link" on the English template page and was directed to a "Create a new item" page where I then tried to enter correct entries in the boxes shown. I was not allowed to change the entry for "Site of the first linked page" (which is a grayed-out "enwiki") or the entry for "Name of the first linked page" (which is a grayed-out "Template:OTRS ticket"). After filling in the blanks that I could enter, I clicked "Create" and was directed to an error page. Here is what I know to be true: the corresponding page on the French Wikipedia for the English language page "Template:OTRS ticket" is "Modèle:OTRS" (https://fr.wikipedia.org/wiki/Mod%C3%A8le:OTRS). I also know that there is currently no link on the English page to the French one, and it looks like there should be one (no?). I also know that I have followed what appear to the directions for creating such a link, and became lost immediately. Can someone please create such a link? Also, and this may not be the correct place for this discussion, the German "equivalent" page which is linked to from the English "Template:OTRS ticket" page is not the correct German page— the English page links to the German equivalent of the English "Template:Permission OTRS" page, which is NOT an equivalent page (my German is too rusty to know what IS the correct page, but I am certain that that is not it). Getting that fixed would be great, too! Thank you! KDS4444 (talk) 03:18, 26 March 2016 (UTC)

I believe I've fixed it for you. I don't know why you were experiencing problems with editing the Wikidata item. Apparently you did create it successfully. The odd German link was coming from an old-style interwiki link hiding on the template's documentation page, which I removed. Cheers, Bovlb (talk) 05:58, 26 March 2016 (UTC)

Rank of Rolling Stone's "The 500 Greatest Songs of All Time".

How do we store the rank of a song in Rolling Stone's The 500 Greatest Songs of All Time? Note that there are two ranks. One from 2004 and one from 2010. Both should be stored in Wikidata for all songs which are in at least one of those lists. --Jobu0101 (talk) 19:11, 26 March 2016 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── This should be something like:

  • P:[Listed in ranking] -> Q[Rolling Stone's 500 Greatest Songs of All Time]
    • position -> 21
    • point in time -> 2004
  • P:[Listed in ranking] -> Q[Rolling Stone's 500 Greatest Songs of All Time]
    • position -> 34
    • point in time -> 2010

I don't think we have anything suitable for the property right now; but it's worth double-checking, before proposing one. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:17, 26 March 2016 (UTC)

Sounds cool. I think we definitely need such a property. It has so many applications! --Jobu0101 (talk) 20:07, 26 March 2016 (UTC)
charted in (P2291), maybe? Sjoerd de Bruin (talk) 20:21, 26 March 2016 (UTC)
more like ranking (P1352)  . But I'd suggest to create an item for each year's ranking, and linking the item to the rank with charted in (P2291)  . I think what's that WikiProject Cycling does. author  TomT0m / talk page 20:26, 26 March 2016 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────

So:

or:

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:40, 26 March 2016 (UTC)

Project for coins, banknotes and stamps

I've been approached to help start a "WikiProject Numismatics", which I'm happy to do. Should we include banknotes and/ or stamps, with a suitable rename, or keep it narrow? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:49, 26 March 2016 (UTC)

Constraints and people

There was recently a discussion about constraints and people. The conclusion was that we need to track missing claims like sex or gender (P21), place of birth (P19) and date of birth (P569) on people but that constraint violations are not the right place to do so. I've now created a new database report for this tracking. If you are working on a specific property, you can add it to Wikidata:Database reports/Humans with missing claims/input and a report will be created and regularly updated for this property. --Pasleim (talk) 13:40, 25 March 2016 (UTC)

Great idea! Could you perhaps also add occupation (P106) which is a quite common constraint in many reports I'm watching? -- Gymel (talk) 23:55, 25 March 2016 (UTC)
  • Wasn't one argument to reduce bot activity? As Krbot already does a lot of data crunching anyways, a new system would actually duplicate most of it. Given that a new approach can always gain us new experience, it might still be a good idea. In any case, I'd keep P21 on property talk pages.
    --- Jura 09:16, 26 March 2016 (UTC)
Yes, idea is great, but wouldn't it be more better (from user PoV) to allow something like this for input: Pr#Px#Py#Pz? X, y, z is needed properties for Pr report. --Edgars2007 (talk) 10:04, 26 March 2016 (UTC)
Keeping sex or gender (P21) on property talk pages would not solve the original problems: Spurious constraint violations. Also, I am more concerned about editor workload than on bot workload. --Srittau (talk) 15:27, 26 March 2016 (UTC)
Do you have a few samples of what you call "spurious constraint violations"? Preferably some you solved yourself? How would such violations be a problem?
--- Jura 15:42, 26 March 2016 (UTC)
"Spurious" is probably not the right word. Correct would be "the huge amount of constraint violations that are duplicated and unrelated to the property". --Srittau (talk) 17:29, 26 March 2016 (UTC)
"huge amount" of missing sex or gender (P21)? Where is that?
--- Jura 17:32, 26 March 2016 (UTC)
In various constraint violation pages that I have come across. Look for yourself. --Srittau (talk) 18:21, 26 March 2016 (UTC)
So it's just theoretical.
--- Jura 18:23, 26 March 2016 (UTC)
As I said: no. Stop being annoying. --Srittau (talk) 18:42, 26 March 2016 (UTC)
We can't really discuss the issue you see and that may be there if you don't recall where you saw it. If it annoys you to discuss it, ok, let's forget what you brought up.
--- Jura 18:47, 26 March 2016 (UTC)
Some people are trying to get work done, and can't be bothered to look up "proofs", just to jump through the hoops of others. But since you insist: Wikidata:Database reports/Constraint violations/P344 is just one of many examples. --Srittau (talk) 19:15, 26 March 2016 (UTC)
You were talking about "huge numbers" and "editor workload". This suggests that it's definitely worth looking into it. Unfortunately, the sample you gave us isn't covered by Pasleim's report.
--- Jura 08:02, 27 March 2016 (UTC)
I created the lists not to duplicate the work of KrBot but to sort out the spurious constraints. I mean with that if a person is missing a sex or gender (P21) claim this is not an error in our data but just a missing piece of information. On the other hand, if we have a format, single value, unique value etc. constraint violation these are in most cases really erros, i.e. mistakes that happend by entering the data or vandalism. As there are user who prefer to resolve errors and other who prefer to make our data more complete I think it's justify to have two sytems. --Pasleim (talk) 16:46, 26 March 2016 (UTC)
How does this compare to mandatory and non-mandatory constraints? Ivan A. Krestinin implemented these to do more or less the same.
--- Jura 16:57, 26 March 2016 (UTC)
Mandatory constraints are to prevent data degradation and was not made to sort out 'nice to have claims' from mistakes. --Pasleim (talk) 17:15, 26 March 2016 (UTC)
You can change display style any/or labels of {{Constraint:Item}} and {{Constraint:Target required claim}} if its confuse somebody. Or create additional parameter, something like "missing_data=true" to separate strong constraints and missing data constraints. Report duplication is good point. My bot or another bot can fail or have some bugs. Cross-check makes whole system more stable. — Ivan A. Krestinin (talk) 17:28, 26 March 2016 (UTC)

P31 values for Wikiversity pages

Just wondering, what are good values to use for pages of Wikiversity?

For Wikinews, there is Wikinews article (Q17633526).

Wikidata:Wikiversity doesn't seem to mention it.

@Abc82: maybe you want to comment
--- Jura 10:12, 26 March 2016 (UTC)

Inspire campain and Wikidata

Hi, only a few days left to comment/propose your help/endorse stuffs in the inspire campain. There is proposals but very few participation and comments inside. It's a pity. There is a lot of (more or less) Wikidata related proposals (please complete if I missed some) that could benefit inputs from us:

a lot of stuffs for Wikidata, hot topic ... too bad there is so few participation ... yet ? .

author  TomT0m / talk page 14:32, 27 March 2016 (UTC)

Hotels

I was looking at some hotel items like Plaza Hotel (Q1066676), Waldorf-Astoria Hotel (Q1123997) and Four Seasons Hotel New York (Q992616). Two questions:

  • Should the building and the hotel (the organization that resides in the building) be two separate items?
  • Are there any properties to add the number of rooms to a hotel or does that need a new suggestion?

//Mippzon (talk) 18:32, 24 March 2016 (UTC)

I'd like to answer your first question with a clear "maybe": Are the building and the organization closely linked? I would not bother with separate items. Had the building multiple uses during its existence or did the organization move buildings? Use two items. Has the organization multiple hotels? Definitely multiple items. --Srittau (talk) 19:19, 24 March 2016 (UTC)
What Srittau said, and note this is also true for other organizations that are linked to important buildings, such as museums, libraries, (former) town halls or churches, etc. For the important cultural heritage objects there are often two items on Wikidata already because someone wrote a Wikipedia article about the building and someone else wrote an article about the organization. --Jane023 (talk) 06:54, 25 March 2016 (UTC)
And about second question - didn't find anything suitable. So you could propose it. --Edgars2007 (talk) 07:01, 25 March 2016 (UTC)
I yesterday added Folkets Park (Q23498066): an instance of "public recreation space, usually featuring large grassed areas, trees, children's play facilities, etc" in Sundsvall, Sweden. Those activities ended in 2015, but now it is used by a school. Tricky... -- Innocent bystander (talk) 07:47, 25 March 2016 (UTC)
Thanks! I'll see if I can try out the propose process! I think I've stumbled upon a similar case: Hotell Triangeln (Q531394) is a scrysraper like building built as a hotel buiding. The first one to use the building was Sheraton Hotels and Resorts (Q634831), then later Scandic Hotels (Q129391), then Hilton Hotels & Resorts (Q598884), then Scandic Hotels (Q129391) again. The building has not changed its use as in the example above regarding Folkets Park (Q23498066) which is even harder to figure out how to handle. Is it possible to somehow state the usage of a building/place? Like from year xxxx to yyyy it was a hotel, then from year zzzz to aaaa it was a museum etc? //Mippzon (talk) 08:11, 25 March 2016 (UTC)

Another thing about hotels: I got some tips that one could use has part(s) (P527) for defining what and how many of a type of room a hotel consists of. But I'm having a hard time figuring out the object structure, since there are not many items for hotel rooms and so on. Does someone have any suggestions? Should I create a new item called "hotel room" or should it be "double room", "single room", "suite" etc? What is a "hotel room" instance of? //Mippzon (talk) 16:00, 25 March 2016 (UTC)

I've created hotel room (Q23541360) if someone wants to help out :) //Mippzon (talk) 18:25, 25 March 2016 (UTC)
A hotel room is a subclass of (P279) location. There may be more definite types of location than that. --Izno (talk) 23:02, 27 March 2016 (UTC)

Destructive edits?

There seems to be some destructive edits on Horacio Rodríguez Larreta (Q2843714) that I don't know how to roll back. Someone that knows maybe should take a look? //Mippzon (talk) 21:46, 27 March 2016 (UTC)

  Done --Srittau (talk) 22:31, 27 March 2016 (UTC)

Nuclear power plant map

I saw this map listing all power plants in the world, and I was wondering if is there any way to do that also using Wikidata? Is our data comparable?--Micru (talk) 08:09, 25 March 2016 (UTC)

@Micru: this should be a starting point. We have 247 nuclear power plant (Q134447) hat the moment. I saw examples for automatic maps but I can't find them at the moment. Maybe someone else could help. -- T.seppelt (talk) 12:40, 25 March 2016 (UTC)
Extension:Graph/Demo/Sparql this is the demo page with similar examples for creating a map. -- T.seppelt (talk) 12:48, 25 March 2016 (UTC)
BTW, there is such map for museums. --Edgars2007 (talk) 12:55, 25 March 2016 (UTC)
I just noticed that you can display a map in the Query Service (Display → Map). This should be a solution. -- T.seppelt (talk) 12:58, 25 March 2016 (UTC)
@T.seppelt: If I would add the type of every APP (e.g. Q1783549 or CANDU reactor (Q558181)) would you do a map with <​graph>?--Kopiersperre (talk) 21:24, 25 March 2016 (UTC)
Yes, I would be interested to do that. Let me know how you connect the items about the APPs with their types. --T.seppelt (talk) 23:31, 25 March 2016 (UTC)

@T.seppelt: Hallo T.seppelt, da es keine Property für „Baureihe“ gibt, werde ich wohl die einzelnen KKWs als instance of (P31) oder subclass of (P279) zu ihren Baureihen auffassen.--Kopiersperre (talk) 08:03, 26 March 2016 (UTC)

@Kopiersperre: This sounds reasonable. A functional class hierarchy for nuclear power plant (Q134447) is anyways missing. I think it is clear that types of power plants should have only subclass of (P279) and single power plants instance of (P31) = < one of the types >. All types should be linked as chain of subclass of (P279) to nuclear power plant (Q134447). Wikidata Graph shows how far we got until know. There seem to be confusions with nuclear reactor (Q80877). Look at [10]. We need to define the difference between those and have a look on every single power plant in order to classify it properly. -- T.seppelt (talk) 09:02, 26 March 2016 (UTC)
How do you properly classify R4 nuclear reactor (Q4357537)? It was designed to produce weapons plutonium and energy but the Swedish nuclear program was closed before it started and the plant is today powered with oil. -- Innocent bystander (talk) 10:35, 26 March 2016 (UTC)
Two options: Either two instance of (P31) claims (nuclear reactor (Q80877)/nuclear power plant (Q134447)??? and oil-fired power station (Q261604)) with qualifiers or one general instance of (P31)power station (Q159719). I would prefer the first one. -- T.seppelt (talk)
Me  Like -- Innocent bystander (talk) 12:18, 26 March 2016 (UTC)

@T.seppelt: Difficult: Reactor types are subclasses of both nuclear reactor (Q80877) and nuclear power plant (Q134447).--Kopiersperre (talk) 11:10, 27 March 2016 (UTC)

@Kopiersperre: A reactor is a part of a nuclear power plant (Q134447). By the way a plant can have several reactors and perhaps have different types of reactors. So the subclass relation is not appropriate. The data structure of a plant should be modified in order to include that feature. Snipre (talk) 20:08, 27 March 2016 (UTC)
I meant that most modern designs also include the non-reactor parts.--Kopiersperre (talk) 21:23, 27 March 2016 (UTC)
@Snipre: So you want a separate item for every block of a NPP?--Kopiersperre (talk) 21:25, 27 March 2016 (UTC)
entity instance of (P31) nuclear plant has part(s) (P527) nuclear reactor <qualified> number <number> --Izno (talk) 23:01, 27 March 2016 (UTC)
⟨ entity ⟩ has parts of the class Search ⟨ nuclear reactor ⟩
quantity (P1114)   ⟨ the number ⟩
now that we have a property specific for that kind of relationship. author  TomT0m / talk page 09:52, 28 March 2016 (UTC)

Merge Q20853628 into Q19635877

I think we can Merge Q20853628 into Q19635877.

Keeping Q19635877 as the main entry for this topic.

Then the Wikimedia Commons sister-link can remain at the main data page at Q19635877.

Thank you,

Cirt (talk) 03:52, 28 March 2016 (UTC)

This is not how we link Commons; see WD:Commons. --Izno (talk) 11:39, 28 March 2016 (UTC)

Q5687624 & Q3497127 (join them together)

Hello. I am trying to append this entry: https://en.wikipedia.org/wiki/Photographic_lighting and this one https://es.wikipedia.org/wiki/Iluminaci%C3%B3n_en_fotograf%C3%ADa because they are the same. But I am not allowed or whatever. The previous system was so easy, why on earth they had to change it?

Thank you. -- El Mono Español. 8:31 - 28/03/2016

Was so easy, easier even than before Wikidata, yay! Help:Merge could help you. --Stryn (talk) 07:36, 28 March 2016 (UTC)

Next office hour on IRC

Hey everyone :)

We'll be doing another office hour on IRC for all things Wikidata on April 12th at 4PM UTC. Check here for your timezone. I'll give an overview of what's been happening over the past 3 months and give an update on what's coming up. We'll have time for questions as well. If you have any topics you'd like to bring up please let me know. As always there will be logs for people who can't attend. Hope to see many of you there.

Cheers --Lydia Pintscher (WMDE) (talk) 12:15, 28 March 2016 (UTC)

Wikidata weekly summary #202

Eugène d'Harcourt

There are two Eugène d'Harcourt: composer and politician. Should they be renamed and how? Ardomlank (talk) 16:59, 28 March 2016 (UTC)

They shouldn't be renamed. We use descriptions to distinguish things with the same name. - Nikki (talk) 17:25, 28 March 2016 (UTC)

Linked items

Hi, is it possible to use Linked items tool for categories?--ԱշոտՏՆՂ (talk) 21:56, 26 March 2016 (UTC)

@ԱշոտՏՆՂ: You could export from Autolist to PagePile and then import there. CatScan2 might get you directly what you are looking for.
--- Jura 07:21, 27 March 2016 (UTC)
Thanks Jura, but what if I have wikiquote page list and it is impossible to generate that list with CatScan or Autolist?--ԱշոտՏՆՂ (talk) 08:10, 27 March 2016 (UTC)
@ԱշոտՏՆՂ:: Maybe like this (for https://en.wikiquote.org/wiki/Category:Authors )?
--- Jura 08:16, 27 March 2016 (UTC)
Jura, unfortunately that categories are not in one place. I generate it with AWB List compare and used different project pages in each sides (hywikiquote-hywikipedia). I want to marge common categories with Quickstatements.--ԱշոտՏՆՂ (talk) 08:27, 27 March 2016 (UTC)

I see. There are some limitations with categories (compared to articles in these categories).

(1) Maybe the following could work for you with Quickstatements if you are comfortable about the category names (I'm using enwiki/enwikiquote just as a sample):

First column are articles from: "enwiki"

Category:Screenwriters	Senwikiquote	Category:Screenwriters

It adds the Wikiquote category to the category with the same name at Wikipedia. If the category doesn't have an item, it skips it. If the category has another wikiquote link, it overwrites it.

(2) Autolist could generate a list of items and get_item_names has an option "Wikipedia list". This might be the category at hywiki if you use "hy" as exclusive language code.

Both approaches have some limits. Maybe there is a better way to do this.
--- Jura 08:58, 27 March 2016 (UTC)

@ԱշոտՏՆՂ: it can be done, using SQL. I currently have some problems there, but most probably later can give you the list. --Edgars2007 (talk) 09:28, 27 March 2016 (UTC)
@ԱշոտՏՆՂ: so here are categories, which are not connected to Wikidata. --Edgars2007 (talk) 20:05, 27 March 2016 (UTC)

Jura, Edgars2007 thanks for help. We have new tool and it is possible to do this with its Manual list option[11]. Great tool! ^_^--ԱշոտՏՆՂ (talk) 21:00, 28 March 2016 (UTC)

P1840 - problem with constraint violations report

investigated by (P1840) should only be used on items that are an instance or subclass of event (Q1656682). However, the constraint violation report currently shows all 47 uses are violations. This is clearly not correct as e.g. . Thryduulf (talk: local | en.wp | en.wikt) 20:03, 27 March 2016 (UTC)
It looks to me that the constraint references the wrong event item. It should probably reference occurrence (Q1190554), instead of event (Q1656682). --Srittau (talk) 20:13, 27 March 2016 (UTC)
You're right Srittau. I've corrected the constraint. event (Q1656682) doesn't need an investigation, but occurrence (Q1190554) does in some cases. Mbch331 (talk) 21:12, 27 March 2016 (UTC)
Thank you both. Thryduulf (talk: local | en.wp | en.wikt) 20:14, 28 March 2016 (UTC)

How to indicate that building was demolished at a certain date?

How can I indicate that Kempehallen (Q6387054) was intentionally demolished at 2012-08-01? //Mippzon (talk) 21:40, 27 March 2016 (UTC)

@Mippzon: This is done using significant event (P793) = demolition (Q331483), qualified with point in time (P585). You may have a look at rules for reporting significant events at Property_talk:P793#Rules. -- LaddΩ chat ;) 22:51, 27 March 2016 (UTC)
Thanks @Laddo:! Just what I was looking for and then some more :) //Mippzon (talk) 07:47, 28 March 2016 (UTC)
@Mippzon: You could also consider using dissolved, abolished or demolished date (P576) which lists building demolition in its description. Bovlb (talk) 15:53, 28 March 2016 (UTC)
@Bovlb: Not a good thing to have two different ways to store data: nobody can then extract all data by a single query. I propose to start a discussion about the need to focus on only one property to describe demolition of building there. Snipre (talk) 22:03, 28 March 2016 (UTC)

Telescope infobox parameters

Hi all. As some of you may know, I've been converting en:Template:Infobox telescope to use Wikidata, as a prototype/case study for entirely wikidata-driven enwp infoboxes (e.g. see en:South Pole Telescope, but also a lot of infoboxes in other telescopes now either fully or partially). I'd appreciate some help/feedback with getting the last few parameters wikidata-ified, as I think some of them need new properties, while others probably just need to be removed from the infobox.

The ones that aren't covered by Wikidata at the moment are:

  • "caption" - the description of the image used in the infobox. I think this is the most problematic infobox parameter to support through Wikidata, as it needs language-specific captions. I suspect this might have to stay locally-defined, or maybe removed.
  • "weather" - this is currently used inconsistently, e.g. for the number of clear nights at the telescope, or the typical weather at the telescope. I think this should be removed from the infobox, unless there are any obvious properties that could be used here.
  • Wavelength - this is typically a range of wavelengths (or sometimes frequencies or energies) that the telescope observes over. We have astronomical filter (P1227), but that's a different thing (it's more aimed at astronomical cameras/instruments than telescopes). So I think this needs a new wikidata property, although given that sometimes this might be better expressed as a frequency, and other times as an energy, I'm not quite sure how this should be defined.
  • "diameter2" and "diameter3" - these are diameters of the secondary and tertiary mirrors (where these exist on the telescope). diameter (P2386) is used for the main dish diameter, so I guess qualifiers could be used here.
  • "mounting" - how the telescope is mounted, e.g. "alt-azimuth", or "equatorial" or similar. I suspect this needs a new wikidata property, unless a related one already exists?
  • "dome" - the dome structure surrounding the telescope (although sometimes this is 'None'). Again, this probably needs a new Wikidata property, perhaps as "enclosure"?

Does anyone have any thoughts on these? I'd also appreciate any feedback on how Wikidata properties are currently used in the template - have I used the right properties, or are there better ways of structuring the information? You can see a list of these, and also the ones that aren't yet covered, at en:Template:Infobox_telescope#Parameters. (@RexxS, Lydia Pintscher (WMDE)) Thanks. Mike Peel (talk) 17:55, 28 March 2016 (UTC)


  • The coordinates have gone live in the English Wikipedia without ever deciding if the precision of the coordinates should reflect the size of the object, the precision to which some appropriate part of the object is known, or the precision stated in the source. For example, the radio telescope in Hancock, New Hampshire, (which does not seem to have an entry yet) has a position of x = 1446374.819 m, y = -4447939.678 m, and z = 4322306.196 m in the International Terrestrial Reference Frame (2008, epoch January 1, 2016). Obviously the telescope is much larger than 1 mm in size. (Of course the x, y, z position could easily be converted to latitude and longitude, if only we knew what precision it should be converted to.) Jc3s5h (talk) 18:12, 29 March 2016 (UTC)

New section with "popular" items on main page

I'd like to suggest a new section on the main page, similar in spirit to the recent events section on many language Wikipedias - but instead of an actual, manually written recent event section, this is a list of the 3-5 top items, and a picture linked from those items, based on something like the Wikipedia Live Monitor (see on Twitter). My suggestion would be a bot that all 3-24 hours edits a template, and the template can then also be manually changed in case something embarrassing pops up, and the template is then transcluded on the Main page.

Not sure if better discussed here, or on the main page talk page. --Denny (talk) 21:09, 29 March 2016 (UTC)

Hi, is there any way to find all items that have ru label but don't have ruwiki link?--ԱշոտՏՆՂ (talk) 15:27, 30 March 2016 (UTC)

select ?item where { 
     
     ?item rdfs:label ?itemLabelRu filter (lang(?itemLabelRu) = "ru"&& contains(?item) 
     
      OPTIONAL { ?wartruitem schema:about ?item . ?wartruitem schema:inLanguage "ru" }
     FILTER(!bound(?wartruitem)) 
} limit 100
Try it!
Jack Bauer shows up in first result for me. author  TomT0m / talk page 15:41, 30 March 2016 (UTC)
TomT0m, thanks a lot! I didn't know about those templates. Very useful--ԱշոտՏՆՂ (talk) 16:15, 30 March 2016 (UTC)

Merging Q1500726 and Q22956438

My Danish is far from perfect but I am pretty sure Q1500726 and Q22956438 refer to the same object. Could someone merge them? Kanorp (talk) 16:06, 30 March 2016 (UTC)

Done--ԱշոտՏՆՂ (talk) 16:13, 30 March 2016 (UTC)

General properties vs constraints

In an ongoing property proposal for the "type location" of a mineral (in short: the place where it was first found) I suggested suing one property for that and the type location of a species. It was counter suggested that we should use separate properties "to have domain specific constraints".

I don't wish to fork that specific discussion, but I would like a general consensus (or a pointer to past consensus) on whether our constraint system requires us to have more specific properties, or allows us to make them more general. For instance, can we have a constraint that says "if this is an instance of Foo, check for X, but if it is an instance of Bar, check for Z"." If we cannot, should such functionality be developed? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:51, 29 March 2016 (UTC)

I can not speak for this discussion, but in general - and especially in property proposal discussions - I have the feeling that there is sometimes an over-reliance on qualifiers when we should use specialized sub-properties. To me, qualifiers should be used to express additional information about the value of a property. Using them to add additional information about the used property is a stop-gap. Specialized properties have three major advantages in my opinion:
  • Sub-properties are far more easy to discover and use correctly and consistently than a more generic property, which requires a qualifier with a specific object as value.
  • Properties are easier and faster to query than a property/qualifier combination.
  • Specialized properties can have tighter constraints and can be better customized for certain situations.
--Srittau (talk) 01:18, 30 March 2016 (UTC)
Srittau I disagree about the fact that "Sub-properties are far more easy to discover and use correctly and consistently". When you have several thousand of properties, finding the good property among several ones with small differences in definition is quite challenging. Currently this is no up-to-date lists of properties for the general item classes like persons, animals, cars, boats, chemical elements, ... Some projects maintain their own list but there are currently no tools which allow you to list all properties for a class of items. Snipre (talk) 08:10, 30 March 2016 (UTC)
It is much easier to find a matching property with a good name, though, than to find the matching, more generic (and more generically named) property and then to find out that you have to use certain qualifier with it. --Srittau (talk) 10:44, 30 March 2016 (UTC)
From the different cases where I was involved, this was not the case: people use the first property they find even if it doesn't match exactly the original definition of the property. An example is duration (P2047) (I let you read the talk page to understand the problem). The translation is often a critical point so when you say "property with a good name" I think you don't translate properties labels frequently because this is not easy to have a good name in all languages.
My conclusion is the next one: having more properties is a good thing from use point of view but you need to offer an excellent way to choose the properties or at least to access to the whole list of similar properties. We don't have that system now, we only rely on property labels which are not reliable because of the translation problem. Before promoting more specific properties we have to find a better way to look for properties. Snipre (talk) 12:05, 30 March 2016 (UTC)
Our constraint system is - initially - property centric. I argue for a while that it would be better to be "class centric". If only for one reason because it would make easier to regroup, sort and discuss informations in a more consistent way.I think we will all agree it's not so easy to find the right property on wikidata, globally documentation about properties is spread out on each property pages and we fail to have a place where we can efficiently recover say, all properties that could be applied to a person, even if there is a number of initiatives to do such things other time - infobox-centric documentation had a moderate success, I just know about the cycling race who is really successful thanks to JGHJ, there is some classes of properties which we should actually use - Property for a person,: Wikidata property for items about people (Q18608871)      WikiProjects are domain centric and maintains list of properties which could or could not be regrouped by the class of items they applies to.
After this general beforetalk, thanks to the query service our constraint system has been now greatly extended thanks to {{Complex constraint}} which actually allows to write constraints wich filters the property according to the class of items. This is good news and gives a lot of air to us ;) Future is bright. author  TomT0m / talk page 11:08, 30 March 2016 (UTC)
discoverer or inventor (P61) is mixing different kinds of events e.g. (Big Bang (Q323)). --Succu (talk) 20:59, 30 March 2016 (UTC)
Is or not an event?! --Succu (talk) 22:01, 30 March 2016 (UTC)

Recent changes

Recent changes (https://en.wikipedia.org/wiki/Special:RecentChanges) provides the last 500 changes and changes made in the last 30 days. Is it possible to go back thousands of changes and months or years? Thank you very much in advance!

Yes. If you select the 500 and 30 options, you'll notice that those numbers appear in the URL. Simply change them to however many changes / however many days you want :) Ajraddatz (talk) 21:44, 16 January 2016 (UTC)
Thank you, I already tried before what you suggest. But I got the impression that this does not work. If it worked, there would be 1.000.000 edits visible here https://en.wikipedia.org/w/index.php?title=Special:RecentChanges&days=1000&from=&limit=1000000. Any idea how to fix this?
Do you have a specific task or project behind your question? --Voll (talk) 07:51, 18 January 2016 (UTC)
Afaik, the limit for recent changes is 5000 changes, as for the search results pages. I personally feel this as a bit limiting, too. Maybe result pages of more than 5000 entries would actually be getting a bit too large, but a pagination which would allow more than 5000 entries in total (e.g. show 500 entries by default, but offer a "Next 500" link, like it is done for e.g. the search results, and allow that for more than 10 times) would be very helpful for patrolling vandalism etc. --YMS (talk) 08:25, 18 January 2016 (UTC)
No, it is limited to only 30 days, number of changes isn't limited. At least, it should be like that (about number of changes). --Edgars2007 (talk) 09:01, 18 January 2016 (UTC)
The recent_changes database table (or whatever its exact name is) contains all changes of the last 30 days, but Special:Recent Changes will deliver 5000 entries max. This is no limitation within the available data (the special page just reads said table), this is just a limit hardcoded in the special page itself. As there is no pagination, there is no way to see the next 5000 entries. If you want to see any changes older than thevnewest 5000 there, you will have to set up some more restrictive filters (or query the database with some external tools instead). --YMS (talk) 06:59, 23 March 2016 (UTC)
@Voll: I am a journalist and need as much data as possible to conduct data journalism. @all: Thank you. Who can change the 30-day-limit? What must I do to get more data?
Wikidata:Data access --Molarus 17:58, 15 March 2016 (UTC)
The 30-days-limit affects, among other things, CheckUser and therefor also our privacy-policy. It will therefor not easily be changed. But as the Data access-page linked above tells, there are other ways to get the information you are looking for. -- Innocent bystander (talk) 08:03, 16 March 2016 (UTC)

Bug in 'NamesAsLabels' tool

Just a heads up to users of the 'NamesAsLabels' tool. I've found some errors in data it has just returned. For example, the return from the query:

https://tools.wmflabs.org/wikidata-todo/add_name_labels.php?from=fr&to=id&doit=Do+it!&wdq=

wrongly includes:

  • Q651 Lid "Identifiant BPN"
  • Q845 Lid "Identifiant SRHP"
  • Q1294 Lid "Code eco-région WWF"
  • Q1332 Lid "Coordonnées du point le plus au nord"
  • Q1333 Lid "Coordonnées du point le plus au sud"
  • Q1544 Lid "Federal Register Document Number"
  • Q1395 Lid "Identifiant NCI"

among many valid lines. I've notified User:Magnus Manske. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:56, 19 March 2016 (UTC)

Has anyone identified (or fixed) the cause of this bug? Is it a tool issue, or an API issue? User:Magnus Manske? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:51, 23 March 2016 (UTC)

User:Magnus Manske tells me this is fixed; unfortunately I've run out of time to check, and I'll be mostly offline for a few days. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:15, 31 March 2016 (UTC)

Establish a company entry - Complete novice

Hello,

I am a member of a startup company based out of India. I wish to create an entry for our business in Wikidata. The name of our company is CREO Technologies. The website is https://www.creosense.com/

We are a consumer electronics company. In the past, we were known by a different name - MangoMan Consumer Electronics. We have released two products in the Indian market - Teewe and Teewe 2. These are both media streaming devices somewhat similar to Google Chromecast

Now, with a new brand name CREO, we are launching a smartphone - Mark 1. Mark 1 is a flagship phone with competitive current gen hardware. It runs on FUEL OS - An Android fork developed in-house by the CREO team. The phone launch will be done on April 13th.

CREO Mark 1 is a phone that comes with a refreshing and ambitious promise of "A new phone every month". The company aims to give monthly OS updates that keep the phone updated and running like new. At the same time, CREO aims to add functionality to complement those already on Android to give the user a great experience.

The primary goal of creating a Wikidata entry is for SEO purposes. But not far behind is the fact that as a startup, we want people to know about us. We are more than happy to follow the community guidelines and create an entry that is objective and serves the users well.

Personally, I am not the most technically oriented person, but I do have developers I can take help from if needed. This happens to be my first time doing something like this. Would appreciate any help as to how should I go about doing this.

Thanks for the time all.

In general, we only accept entries on Wikidata for entities that already have an article on Wikipedia (in any language). If your company already has such an article, and that article is not threatened to be deleted, please point us to it, and we can create the corresponding item here on Wikidata, if it does not exist yet. Otherwise we will not accept a new item. --Srittau (talk) 11:58, 31 March 2016 (UTC)

New search syntax at the BNF since january

Please, update the Special page Special:BookSources - in the Key National libraries section - and in National "France" section

the correct syntax for BNF search (French national Library) is now http://catalogue.bnf.fr/rechercher.do?motRecherche=9782321000679

Thanks to anyone who can do it :) --Hsarrazin (talk) 15:44, 31 March 2016 (UTC)

  Done at Wikidata:Book sources. Matěj Suchánek (talk) 16:12, 31 March 2016 (UTC)