Wikidata:Project chat/Archive/2017/01

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

Football teams

Hello. We have association football club (Q476028) and association football team (Q15944511). I am not sure that every user understand the difference (including me). Of course,

As it is written to the article, "unqualified, the word football is understood to refer to whichever form of football is the most popular in the regional context in which the word appears. Sports commonly called 'football' in certain places include: association football (known as soccer in some countries); gridiron football (specifically American football or Canadian football); Australian rules football; rugby football (either rugby league or rugby union); and Gaelic football. These different variations of football are known as football codes."

They are is a mess with the 2 items. Some links to Wikipedia must be on the first one but are to the second one.

But, do we need association football club (Q476028)? Is association football club (Q476028) the most general article for association football team (Q15944511), gridiron football team, Australian football team, rugby team, Gaelic football team?

Furthermore, the second one is the value of the instance of (P31) to many association football teams. But, the must have the more specific association football team (Q15944511). One way to change that is by finding the items with instance of (P31) --> association football club (Q476028) and sport (P641) --> association football (Q2736) and change association football club (Q476028) with association football team (Q15944511). Anyone can help?

Xaris333 (talk) 20:34, 29 December 2016 (UTC)

Both association football team (Q15944511) and association football club (Q476028) are for association football (Q2736). The difference is that association football team (Q15944511) is for a single team, i.e. 11 people plus some reserve players. association football club (Q476028) is for an organization which can have many teams. Often the concepts of a club and main team of a club are mixed together but in cases where we have items for further teams of a club the distinction becomes clearer: FC Barcelona (Q7156) is a mix of the organization and the main male team. FC Barcelona B (Q10467), FC Barcelona C (Q2346842) and FC Barcelona Femení (Q522899) are teams of the club FC Barcelona (Q7156). --Pasleim (talk) 11:54, 30 December 2016 (UTC)

do we need association football club (Q476028)? - Yes, for eg Chelsea F.C. (Q9616) is an instance of (P31) association football club (Q476028) (organization, in which one or multiple groups of players are organized to compete as teams in association football (soccer)) - Unnited meta (talk) 12:10, 30 December 2016 (UTC)

So you are saying that FC Barcelona (Q7156) must have both instance of (P31) --> association football club (Q476028) and instance of (P31) --> association football team (Q15944511)? Then all items with instance of (P31) --> association football club (Q476028) must also have instance of (P31) --> association football team (Q15944511). Xaris333 (talk) 02:25, 31 December 2016 (UTC)

Is really confusing. All teams has also young teams etc. Lets say it again:

But almost all associations teams in the world have at least a young team. And the article for the main male team, is only for the male main team. Is a association football team (Q15944511). The others teams have their own items. I can understand a different when we have a parent club. For example, Olympiacos CFP (Q12396315) is the parent club (is a sports club (Q847017), many sports) and Olympiacos F.C. (Q19628), Olympiacos B.C. (Q847701), Olympiacos S.C. (Q2021487) are part of the parent club. But I can not understand why we have association football club (Q476028).

  • (And is more confusing thinking that there also female teams. We really need an item for that also).
  • (And a new item football team (Q28083137) is now created for the general football teams (not the association football)).
  • And after trying to translate some of the articles, not all of them are saying that a football team is only the group of 11 players+reserve players. I have left a message to all wikis just to check if their article is link to the correct item.

Xaris333 (talk) 01:40, 1 January 2017 (UTC)


What shall we do with the "has part" statements on Wikipedia:Vital articles (Q5460604) ? Currently there are only about a quarter of the 1000.

Delete? Complete? Move to some internal item? Replace with some statement on the items?
--- Jura 11:43, 31 December 2016 (UTC)

Help:Badges? --Liuxinyu970226 (talk) 12:23, 31 December 2016 (UTC)
I think the idea is that the list would be the same for all languages. Each language would probably have additional "must have" topics. These could be flagged with badges.
--- Jura 12:27, 31 December 2016 (UTC)
Good luck with that! Consensus on a single wiki will never happen, et alone among all the wikis. I think that trying to fill with "has part" is not decidedly helpful, such a moving feast, so I would suggest delete. List the interwikis, describe it, then move on. You are probably better writing a query that shows the pages, and lists the articles at the wikis, then doing some analysis to have a descriptive article. Maybe rate each top 100 with 10, top 200 with 9, ... then you may have indicative tally, get a good maths person involved!  — billinghurst sDrewth 02:05, 1 January 2017 (UTC)
Ok. I removed the P527s.
--- Jura 21:54, 1 January 2017 (UTC)

New template to replace magic words

Template:ISBN I have ported over w:Template:ISBN and w:Module:Check isxn from en.wp. Magic words as links are being phased out and although we don't have to replace all instances of them now, they will all be removed from MediaWiki in 2017. See mw:Requests_for_comment/Future_of_magic_links. We have about 150 entries in Category:Pages using ISBN magic links. —Justin (koavf)TCM 02:38, 1 January 2017 (UTC)

This requires mainly fixing the modules Module:Cite and Module:Sources, @Vlsergey, Daniel Mietchen --Pasleim (talk) 12:53, 1 January 2017 (UTC)

How to find items with no statements?

Is there an easy way, like an existing list or so, to find items on Wikidata that has not statements yet? //Mippzon (talk) 12:40, 1 January 2017 (UTC)

Wikidata:Database reports/without claims by site and the linked reports --Pasleim (talk) 12:56, 1 January 2017 (UTC)
Mippzon, if you like to clean up after Lsjbot, there is a dedicated category on svwiki for that: sv:Kategori:Robotskapade artiklar helt utan påståenden. -- Innocent bystander (talk) 12:59, 1 January 2017 (UTC)
Thanks both of you! Just what I was looking for! I'll see if I can help out in the cleaning effort after Lsjbot as well. //Mippzon (talk) 13:26, 1 January 2017 (UTC)
@Innocent bystander: Can you give a bit more information on what to actually do with Lsjbot articles? Is it to add statements to those items that will help out the cleaning process? //Mippzon (talk) 13:41, 1 January 2017 (UTC)
@Mippzon: You can add P31-claims, P625 (coordinates) and P1566 (Geonames-ID)-claims. You can also add P131 (in administrative division) and P17 (nation) where it is appropriate. It is definitely not always easy! What is a "bergsudde"? I do not know! But many other are easier. -- Innocent bystander (talk) 14:03, 1 January 2017 (UTC)
Haha, yeah, it's pretty hard sometimes when looking at those items. Some or easy though. Thanks for the guidance, I'll see what I can do! //Mippzon (talk) 14:06, 1 January 2017 (UTC)

Popular items without claims

Thanks to Pasleim we now have Wikidata:Database reports/Popular items without claims. These items have many sitelinks, but no statements at all. Please help by adding one or more statements to the items in this list. Multichill (talk) 09:20, 2 January 2017 (UTC)

Wikidata weekly summary #242

Complex merge of items? Linn and San Miguel-Linn

Linn (Q12635542) and San Manuel-Linn (Q972694) need to be merged, and any conflicting articles need to be merged too. It's because "San Miguel-Linn" and "Linn" are the same place... the federal government simply changed the name of the census-designated place.

See this map (or this one) - Same as the outline for "San Miguel-Linn" WhisperToMe (talk) 03:49, 2 January 2017 (UTC)

  Done mostly. Once the two conflicting sitelinks merged, the second item can be redirected.
--- Jura 07:32, 2 January 2017 (UTC)
Thank you so much! @Jura1:, is it OK if you also do Nurillo and Murillo. This census map shows they're the same too WhisperToMe (talk) 12:07, 2 January 2017 (UTC)
@WhisperToMe: How about trying the "Move" and/or Merge gadgets? General intro to merging is at Help:Merge.
--- Jura 12:11, 2 January 2017 (UTC)
Wow, I didn't realize that existed! I enabled it and am set to merge. Thank you! WhisperToMe (talk) 22:42, 2 January 2017 (UTC)

Request for currency data from en.voy

See here Over at voy:en:Wikivoyage:Travellers'_pub#Currencies.2C_again, we have discussed including currency conversions in our travel guides. I know these are data that have been added here and removed again in the past. Can someone give me an overview of how feasible it would be to get these data to import to this project? I'd like to help if I can but I don't have all of the technical skills. —Justin (koavf)TCM 04:10, 2 January 2017 (UTC)

The main problem would be the sheer volume - Wikidata isn't great for this sort of high-volume, purely numeric, data. United States dollar (Q4917) has spot prices for the USD on 31 December each year, just in terms of Euro, and you can see how rapidly this would grow with more than yearly resolution and using five or ten comparison currencies - which you'd need to do to make this practically useful. I wonder if the recent shared tabular data on Commons system would handle it better? Andrew Gray (talk) 13:14, 2 January 2017 (UTC)
It might be better to identify reliable sources, like national banks, who provide such data. There are a variety of ways we could model this in Wikidata. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:48, 2 January 2017 (UTC)

Domain of an equation

This isn't the domain in the sense of the set of input variables, but instead:

How do I connect Navier–Stokes equations (Q201321) to fluid dynamics (Q216320)? --Izno (talk) 22:00, 30 December 2016 (UTC)

facet of (P1269) would work. —MisterSynergy (talk) 22:14, 30 December 2016 (UTC)
I don't think this was this usage that was intended when creating this property, so this is a non answer. This is a mistake to answer this for anything "I don't know what fits" :) I'd prefer, personally, no answer at all than a non-answer. @Izno: a better answer is studies (P2578)   or studied by (P2579)   I guess. But maybe we need to create something more specific to link equations to the physical stuffs it describe or models. What do you think ? author  TomT0m / talk page 10:55, 31 December 2016 (UTC)
Yes, it was my impression that such was missing, though property "use" goes in the right direction. --Izno (talk) 13:58, 31 December 2016 (UTC)
At WD:Property proposal/models. --Izno (talk) 23:35, 3 January 2017 (UTC)

Kings of Jews.. a Wikidata perspective.

I was toying with the notion of kings at the year 0 and as a result I also had a look at the kings of the Jews. There is a problem in that historically there were at least two kingdoms; the kingdom of Judah and the kingdom of Israel. The latter came in existence after a revolt. What I intend to do is make a clear difference between the two. I will work on succession the office and I will separate out the two. This may prove controversial by some who have religious or nationalistic points of view. So I ask now for people to comment before I make these changes.

NB I will not add dates. That is a quagmire in its own right and less fundamental imho. Thanks, GerardM (talk) 08:32, 1 January 2017 (UTC)

To state the obvious first: I have some pov's of my own here, at least when it comes to the biblical kings. But I have no big problems with any opinions or claims against my pov.
I would say that "king of Jew" is a troublesome way to describe things. Not a single (biblical) Jewish king was king of only sons of the patriarch Israel. And in almost all of the biblical history, there were non-Jewish kings over parts of the Jewish nation. Some of these kings were not "real" kings, but vassals of foreign kings. Some of them co-ruled with their father or son in the beginning or end of their kingship. In some periods, there were rival kings over the same territory. Also David had a rival in the first years of his kingship.
There is even a pre-Saul Jewish king mentioned in the book of Judges. He is normally not mentioned in the lists of kings. Among the things he became known for, was being killed by a woman. -- Innocent bystander (talk) 11:59, 1 January 2017 (UTC)
The big thing is that the "kingdom of Israel" is a later separate kingdom. Originally it was the "kingdom of Judah". Thanks, GerardM (⧼TalkpageΑlinktext⧽) 12:29, 1 January 2017 (UTC)
Yes, you have to the define which country/realm they were kings of. The name/label of those kingdoms could be a large issue of debate, but to list them is probably less problematic (but not always easy). Not even the bible itself has only one name for these nations. Israel, Ephraim, Judah, Judah and Benjamin, house of David, house of Jerobeam etc... -- Innocent bystander (talk) 12:54, 1 January 2017 (UTC)

I have restricted myself to the kingdom of Israel and the kingdom of Judah. The kingdom of Judah is the oldest continuing one so I preserve its name to the past. The kingdom of Israel is later and has its own line. thanks, GerardM (talk) 15:50, 2 January 2017 (UTC)

  1. You are right to avoid dates. That's a huge quagmire that it is most definitely not worth addressing here.
  2. I would point out that the combined kingdom (that of David [after the suppression of Saul's line], Solomon and Rehoboam [until the breakaway of Jeroboam]) is conventionally called "Israel" rather than "Judah". I'm not being fussy about how you want to deal with it here—only that I don't want anything missed because it is called "Israel" and not "Judah" during that period. And people could legitimately search "Kings of Israel" looking for Solomon, for example, so that needs to be able to come out correctly.
  3. FWIW, while "King of Israel" was certainly a title used at various points, and in fact so was "King of Judah", "King of the Jews" was not ever used by Jews. That kind of language appears in the Christian Bible around the time of the execution of Jesus. It has nothing to do with the Kingdoms of Israel and Judah of the First Temple era. StevenJ81 (talk) 16:58, 2 January 2017 (UTC)
The rule I have applied for any country is that as long as there is a continuance in space and/or dynasty we keep the name of a country. Wikidata does not have the luxury that it supports a different name depending on a date. The name "kingdom of Israel" is taken as it is the name of a split. Thanks, GerardM (talk) 19:33, 2 January 2017 (UTC)

You could use "official name" with qualifiers "starts" and "ends". I think hat should capture it, no? --Denny (talk) 17:07, 3 January 2017 (UTC)

They could. But it does not make a real difference. The fact is that Wikidata functionality and as far as I know even the proposed changes for Wiktionary support are deficient. Thanks, GerardM (talk) 18:43, 3 January 2017 (UTC)

Are political crisis (Q15887996) and political crisis (Q3002772) the same?

To me it seems as political crisis (Q15887996) and political crisis (Q3002772) are the same. Are they and can they in that case be merged? //Mippzon (talk) 15:22, 3 January 2017 (UTC)

cs, fr, it and ru-speakers wanted! -- Innocent bystander (talk) 15:30, 3 January 2017 (UTC)
All those aricles are about the same topic: explanation of inner and international (exept fr) crises between political sides with examples and no one has a source - Kareyac (talk) 15:59, 3 January 2017 (UTC)
  Support (cs-N) Matěj Suchánek (talk) 19:30, 3 January 2017 (UTC)
If all agree, please perform the merge as I don't remember how to do it :) //Mippzon (talk) 21:11, 3 January 2017 (UTC)
  Done Lymantria (talk) 21:41, 3 January 2017 (UTC)
Just to confirm as ru-N speaker that the merge is fine.--Ymblanter (talk) 21:43, 3 January 2017 (UTC)

Protected areas

Hello. Wikidata:WikiProject Protected areas of France has just been renamed to Wikidata:WikiProject Protected areas. With such a new name, it gets a much wider scope and this is why I'm leaving this message here. We can now recruit participants interested in protected areas situated in any country. If you are one of those people, you can join here. Don't hesitate to do so. Green data might be the future of Wikidata! Thierry Caro (talk) 07:27, 4 January 2017 (UTC)

Stolen identities

Vaclav Jelinek (Q28122388), a spy, used the identity of Erwin van Haarlem (Q28122549), a missing person. How can this relationship best be expressed? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:28, 4 January 2017 (UTC)

Combine pseudonym (P742) with named after (P138) perhaps? ArthurPSmith (talk) 14:54, 4 January 2017 (UTC)
Good idea. I've done that, also qualified as instance of identity theft (Q471880). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:00, 4 January 2017 (UTC)

Guidance on wikimedia disambiguation with regard to interwikis?

Do we have guidance on how to do interwikis for disambiguation pages? Is the disambiguation based on the hard word or the concept/meaning of the word. I see both approaches taken to the point that there is discrepancy, and with our restrictions on label and description, it is causing difficulties. For example Memory (littoral meaning) (Q228636) and Memory (Q20660661) cannot be resolved in English.

I would have thought that it has to be on the hard word, as within a language a word can presumably have different contexts, so concepts have to be disregarded.

I will also through into that mix the process where a surname has the same spelling, I cannot see good guidance on how these should be set up, though I have seen numbers of them and there is a logic loop that may be better described with some guidance.

Noting that at the moment "Wikidata:Disambiguation" goes to Help:Labels, and we also have Help:Statements with a disambiguation section. Thanks.  — billinghurst sDrewth 01:55, 1 January 2017 (UTC)

Here is another example of a frustration Henry V (Q3133121) and Heinrich V (Q161527), and there are others. Maybe this is only an issue in the English language, whichever, it still needs a resolution.  — billinghurst sDrewth 07:16, 1 January 2017 (UTC)
Wikidata:WikiProject Disambiguation pages/guidelines may help to understand. --Stryn (talk) 11:25, 1 January 2017 (UTC)
Thanks Stryn. The page could do with some typical examples of the additional properties that can/should be added. I see people making convention of what is expected to happen, and it would be great to capture such conventions, even use of said to be the same as (P460) (one way/two way?)  — billinghurst sDrewth 05:45, 2 January 2017 (UTC)
Ideally, one wouldn't need to edit them manually at all. Phab:T139912
--- Jura 06:51, 2 January 2017 (UTC)
P460 can't be treated automatically. --Infovarius (talk) 20:51, 4 January 2017 (UTC)

"employer" with qualifier "as"

My students edit entries about scholars: Ivan Gorsky and Vladimir Ermakov.

Does it correct to write in the field "employer" the qualifier "as", for example "university teacher" in the entry Ivan Gorsky? -- Andrew Krizhanovsky (talk) 15:04, 5 January 2017 (UTC)

I qualify with subject has role (P2868) in such cases. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:35, 5 January 2017 (UTC)

Disambiguation page linked from film item

Once in a while, I come across problems like this. How can we improve detection and fixing of such things?
--- Jura 22:35, 3 January 2017 (UTC)

  Comment from a little exploration of a few they look to be long established issues that we imported with wikilinks. Best that I can recommend is to build a maintenance page for the problem, and link the queries. For maintenance queries it would be nice if we could generate a count of a petscan query listed on a hosting page so people know whether to drill down to fix things. Is that something that could be coordinated with listeriabot?  — billinghurst sDrewth 22:59, 4 January 2017 (UTC)
And in terms of the other tools, at this stage we have the "move a link" gadget tool, and it would be great to have a gadget tool that allows the moving of a link to a new item. That would save the step forward, create, and step back situation, which does get rather tedious.  — billinghurst sDrewth 23:22, 4 January 2017 (UTC)
  • Given their number, is it worth doing that manually? It's essentially non-info added to an important item. If it's a disambiguation page at Wikipedia, we could simply disconnect it. Connection to a new item can be done in a separate step.
    --- Jura 09:07, 6 January 2017 (UTC)

Season of a league and a cup

I need a clear answer. I have asked so many times but I haven't a clear answer. Two examples:

2015–16 Cypriot Cup (Q20645820) instance of (P31) Cypriot Cup (Q245970) or

2015–16 Cypriot Cup (Q20645820) subclass of (P279) Cypriot Cup (Q245970) or

2015–16 Cypriot Cup (Q20645820) part of (P361) Cypriot Cup (Q245970)

and of course there is 2015–16 Cypriot Cup (Q20645820) instance of (P31) sports season of a sports club (Q1539532) which just describe the item as a season. Should I use both items for P31?

2015–16 Cypriot First Division (Q19906304) instance of (P31) Cypriot First Division (Q155965) or

2015–16 Cypriot First Division (Q19906304) subclass of (P279) Cypriot First Division (Q155965) or

2015–16 Cypriot First Division (Q19906304) part of (P361) Cypriot First Division (Q155965)

and of course there is 2015–16 Cypriot First Division (Q19906304) instance of (P31) sports season of a sports club (Q1539532) which just describe the item as a season. Should I use both items for P31?

Xaris333 (talk) 06:33, 5 January 2017 (UTC)

part of (P361) is definitely non appropriate. author  TomT0m / talk page 07:48, 5 January 2017 (UTC)
I am seriously thinking to propose a property call "season of". Xaris333 (talk) 08:26, 5 January 2017 (UTC)
For award ceremonies, I saw (and now use) facet of (P1269), which is more general than part of (P361). For example : 27th European Film Awards (Q15131260) instance of (P31) award ceremony (Q4504495) and 27th European Film Awards (Q15131260) facet of (P1269) European Film Awards (Q223740). Koxinga (talk) 11:30, 5 January 2017 (UTC)
"topic of which this item is an aspect, item that offers a broader perspective on the same topic". A season in not that... Xaris333 (talk) 13:23, 5 January 2017 (UTC)
  • There is no clear answer because there is no clear consensus how to model recurring events. At the moment there are basically two systems around
System 1 cf. 2016 Tour de France (Q18574623) Sytem 2 cf. 2016 Cannes Film Festival (Q21061237)
2015–16 Cypriot First Division (Q19906304) instance of (P31) Cypriot First Division (Q155965) 2015–16 Cypriot First Division (Q19906304) instance of (P31) sports season (Q27020041)
2015–16 Cypriot First Division (Q19906304) part of the series (P179) Cypriot First Division (Q155965)
Cypriot First Division (Q155965) subclass of (P279) event (Q1656682)
Cypriot First Division (Q155965) instance of (P31) recurring event (Q15275719)
Cypriot First Division (Q155965) instance of (P31) event (Q1656682)
--Pasleim (talk) 17:49, 5 January 2017 (UTC)
The first model is better, both because it minimize statements - only one statement in actual championship instance systems versus several in each, which leads to "number of seasons" + 2 statements in the first model, versus "number of seasons" * 2 + 1 in the second. Second because clearly it's a very clear case of a type/instance relationship. There is clearly several instances of the "Ligue 1", one each year. Each of this instances shares a lot of common characteristics : each are association football competition occuring in france between french teams. Each team has two matches against every other competitor teams, ... All of those are clearly its own event, so they are instances. The class of all those is clearly a class of recurring event. On the other hand, the "season" item has been questioned : what is a "sport season" for example ? I think no clear answer has emerged. A sport season can be composed of several competitions and not only a championship. author  TomT0m / talk page 18:30, 5 January 2017 (UTC)
Very difficult situation. I was hoping for a final decision. Every one is using different properties in each item. Xaris333 (talk) 23:54, 5 January 2017 (UTC)
  • Not sure if a sports season is comparable to a more punctual event. For film film festival, there were at least four of five different approaches initially and items about the same weren't necessarily connected. The current model simplifies maintenance in the sense that all "film festival edition"-items can be identified in one step and have similar properties and constraints. This can appeal to those with more abstract interest or may not. Maybe for sports seasons, the same applies. Maybe not.
    --- Jura 09:33, 6 January 2017 (UTC)

Storing File formats and their rendering Softwares in Wikidata for Data Preservation archives

I am Sharmeela Ashwin, a Masters student at 'Albert Luedwigs University',Germany. I am working on a Master project where in I would like to store the file formats and their rendering softwares in Wikidata through a framework named EaaS. The wikidata is going to be a database storing all the softwares that will help in opening the different file formats. The wikidata page would contain the PRONOM ID for the specific file formats and their rendering softwares. I would like to continuously save this data each time in wikidata through EaaS whenever the user stores or updates in EaaS. I see there are options like through Sparql queries or through APIs.Kindly suggest me the best way to achieve the above tasks.

Best Regards, Sharmeelaashwin (talk) 15:57, 5 January 2017 (UTC)

@Sharmeelaashwin: Welcome. Could you give, or link to, some example data, please? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:19, 5 January 2017 (UTC)
@Sharmeelaashwin: Hi there, there are two properties for PRONOM: PRONOM file format identifier (P2748) and PRONOM software identifier (P2749). You can currently see an example of writable/readable file format on Adobe Acrobat (Q207902), and you can find more by seeing which items use the PRONOM properties. There was a similar request for writing from a website not long ago, and they will be using a bot for writing to Wikidata. To run a bot on Wikidata you need to request a bot flag, you can read more about the approval process here, a nice manual for creating a bot, and apply for permission here. In the request page and the archives you can see some source code examples from other users, probably the easiest is to use Pywikibot.--Micru (talk) 09:00, 6 January 2017 (UTC)
@Andy:, @Micru:Thank you Andy and Micru for your prompt responses. Please find the details of EaaS in here. I will follow the steps mentioned by Micru and get back if I have any more queries.
@Sharmeelaashwin: Welcome. WikiProject_Informatics may be of interest. I took a first pass at mapping PRONOM to Wikidata properties here. This blog post provides additional details about some related efforts. It is great to see lots of interest in this topic. YULdigitalpreservation (talk) 13:06, 6 January 2017 (UTC)

Guide for making Listeria (Q24045615) lists?

Are there any simple guides or tutorials on how to create those Listeria (Q24045615) lists from scratch? //Mippzon (talk) 18:07, 5 January 2017 (UTC)

@Mippzon: Jag brukar låna idéer från Wikidata:Request a query. -- Innocent bystander (talk) 05:02, 6 January 2017 (UTC)
You can also just copy a list made by someone else and adjust the properties for your needs. I make a lot of lists about art and artists, so I just reuse the list and swap out the Q numbers for the artist (for list of their artworks) or museum (for list of artworks in their collection). The basic thing is that Wikipedia must allow ListeriaBot and have granted bot rights to the bot. In your case you are on Swedish Wikipedia so this is not a problem. I have published lists in my userspace on Swedish Wikipedia so you can look at those. The key is the query which is then fed into the listeria templates "Wikidata list" and then you close with "Wikidata list end". After saving the page you should see the option to "update manually". Click that and the list will be built by the bot. Jane023 (talk) 06:30, 6 January 2017 (UTC)
ListeriaBot does not have bot-flag on svwiki, but it is not required there when editing in low speed and outside article namespace. This far, only one article is using ListeriaBot, but it is widely used to maintain articles and different subjects. -- Innocent bystander (talk) 06:39, 6 January 2017 (UTC)
That's too bad - it used to have the bot flag on svwiki! I guess that means no listeria lists will update on Swedish Wikipedia anymore. BTW that has nothing to do with the article namespace as far as I know. Jane023 (talk) 09:17, 6 January 2017 (UTC)
Template:Replyto:Jane023 No, it has never had any local botflag on svwiki. As I said, you do not have to have a bot-flag for slow bot-editing on svwiki. Au contraire, seeing this bot in the RC and Watchlist, can be a good feature. When people see in RC that somebody has recently died, an item have a new picture or anything else. It can give people new ideas of what they can edit, both here at Wikidata and at Wikipedia. Something that has made it hard for the bot to be accepted in ns:0 is probably that it does not provide very fotogenique tables and does not provide any sources. -- Innocent bystander (talk) 09:30, 6 January 2017 (UTC)
Yes I saw that my list is still updating here. I agree about having lists in namespace 0 - at first I thought it was a good thing but today I am not so sure. The best part of auto updates are things like changes to death dates as you say, but also just links suddenly going blue, which is something most people don't think of when they make a new article. It is handy if objects named in lists are auto-linked when an article about them is created. Jane023 (talk) 09:34, 6 January 2017 (UTC)
I think it would be a good idea to start collecting sample use cases of Listeria.
--- Jura 07:31, 6 January 2017 (UTC)
I used Listeria as the driving force behind my statistics for the June 2016 TED writing challenge. In one month there were updated articles in most of the languages for which I set up listeria lists in my userspace. The nice thing about it is that listeria never gets tired of updating, so I can now see that since the challenge ended there have been even more updates than during the challenge, which means that this is potentially a great tool to use for writing challenges in general. I wish there was some way to go back in time to see whether the impact of having lists helps more than just "normal organic Wikipedia article growth" whatever that is. Jane023 (talk) 09:27, 6 January 2017 (UTC)

How to change a property type?

I like to propose the type change of the property Property:P969 (located at street address) from “String” to “Monolingual text”. The are are several countries with more than one official language, at least locally. For instance, besides German in Germany there are other official languages in several states like Sorbian, Danish of Frisian. The street signs are usually bilingually labelled. In countries with non-latin writing the postal authorities often allow to write addresses in latin languages, mostly English or French. Therefore, we should be able to distinguish multilingual addresses and to specify the language used.

Where and how can I make this proposal? --RolandUnger (talk) 08:22, 6 January 2017 (UTC)

It's already at Wikidata:Properties_for_deletion#located_at_street_address_.28P969.29. In the meantime, you can specify the language with a qualifier.
--- Jura 08:28, 6 January 2017 (UTC)

is dead? Why data is not loading into a map? Even for examples. --Infovarius (talk) 23:36, 4 January 2017 (UTC)

@Magnus Manske:, is it yours? --Infovarius (talk) 22:44, 6 January 2017 (UTC)

Bot edits descriptions

As we lack edit description and we have a lot of bot edits I propose to adjust descriptions from popular tools at least. For example, PetScan saves each query and each bot job under unique PSID, so we can have this ID in each edit made by this specific bot job. May be something for other Widar tools too. User:Magnus Manske, please? --Infovarius (talk) 22:40, 6 January 2017 (UTC)

Adding Property:date of death to an existing item

I'm fairly new to Wikidata editing so am performing certain edits for the first time. For a biographical item whose subject is recently deceased - how do I add the Property:date of death to the item so I can supply the information including a web reference. AND where might I have found an explanation of this on my own? (I navigated through and read numerous pages about Properties but wound up here to ask directly.) Thank you -- Deborahjay (talk) 12:28, 7 January 2017 (UTC)

Add a new statement and just start typing date of death in the first field. If there is a matching property it will show up during typing. Mbch331 (talk) 12:34, 7 January 2017 (UTC)
Thanks - I see now that at the bottom of the Statements there's an option of + add that displays a drop-down menu of existing unused Properties. Among these are not only the one I need but other suggestions relevant to a biography that I might glean from content in the mainspace articles previously added by other editors.Thank you - and I'll be back! Cheers, Deborahjay (talk) 12:51, 7 January 2017 (UTC)
@Deborahjay: Take a look to Help:Contents where you will find an introduction to Wikidata plus some tours on how to how to add statements. For your specific question, go to the item page, scroll to the end of the page, click on "add", and then start typing the property name or select one of the suggested properties.--Micru (talk) 12:43, 7 January 2017 (UTC)
Worthy suggestions, @Micru:, that I'll take to heart since I've matured from my original attitude as a hacker type from the 1980s scorning User Manuals (even as I wrote them). -- Cheers, Deborahjay (talk) 12:51, 7 January 2017 (UTC)

importing data on companies

Hello, my proposal is to add data on companies, including stakeholders as one of their properties.

I am inspired by this talk: their result:

and the dataset they used, from Bureau Van Dijk. Bureau van dijk collect public information on companies, and resells this information as business to business services, with high premium fees.

I would like to contribute to public factual knowledge by adding data on companies, so allow also people not affiliated with research institutes or corporations to have a basic access to innovation sector and a worldwide map of economics.

I am proposing this topic to ask for feedback. Also, I am asking if automation of import with a bot is a viable process for your policy.

Is there anybody interested like me in this topic?

Gg4u (talk) 18:43, 14 January 2017 (UTC)Gg4u


Can someone convince User:Nikkimaria to migrate the Findagrave ID numbers here before she deletes them at Wikipedia. She is on a tear deleting them, she is over 500 deletions now. --Richard Arthur Norton (1958- ) (talk) 04:21, 5 January 2017 (UTC)

@Nikkimaria: - although you really should post at her en.Wikipedia talk page. (Pity they are not in templates; then they could be easily imported.) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:53, 5 January 2017 (UTC)
[I see that Nikkimaria is now posting on this page.] Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:06, 7 January 2017 (UTC)
You are welcome to do so if you feel there is some advantage to it; it ought not to be used as a source for other details, either here or at en-wiki. Nikkimaria (talk) 21:11, 7 January 2017 (UTC)

VGMdb Intergration

Is it possible to integrate VGMdb into Wikidata? --YoumuWarm (talk) 15:09, 5 January 2017 (UTC)

@YoumuWarm: Yes. I've created Wikidata:Property proposal/VGMDb artist ID to start with (pleas ecoemnt there). We could also have properties for IDs for albums and labels, if desired. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:31, 5 January 2017 (UTC)
Thank you both, I also created Wikidata:Property proposal/VGMDb album ID, if it goes well, I'm also going to create one for labels. Regards. --Thibaut120094 (talk) 22:19, 7 January 2017 (UTC)

Quora donates data

I'm pleased to announce a massive data donation (our biggest yet?) from Quora (Q51711). At my request, they have generously given us over 1.5 million topic IDs, which Magnus has kindly loaded into Mix'n'match (Q28054658) for checking and import as Quora topic ID (P3417). The data is in catalogue 319.

As you can see, there are over 259K automated matches. They include some duplicates, typos or non-English alternative names. Those of you who use Quora may choose to resolve (or report) them there. Otherwise, you can may simply "remove" them in Mix'n'Match.

Now you can't complain that you have nothing to do over the holiday period ;-)

Merry Christmas! Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:46, 24 December 2016 (UTC)

Given that's a massive donation with might have errors, how about using the Primary Sources tool to add them?  – The preceding unsigned comment was added by ChristianKl (talk • contribs) at 11:04, 24 December 2016‎ (UTC).
I've not used PS, so will let others comment on that; but what "errors" do you anticipate, and why? Why do you suppose M'n'M is not adequate to trap them? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:30, 24 December 2016 (UTC)
You spoke yourself about typos or non-English alternative names. There might be further issues with different concepts that have the same name. ChristianKl (talk) 11:45, 24 December 2016 (UTC)
Indeed. But they are not "errors". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:44, 24 December 2016 (UTC)
Agree with Christian: Sounds like a good use of Primary Sources. Especially as the initial analysis has shown problems with the data.
--- Jura 11:37, 24 December 2016 (UTC)
The Primary Sources Tool is for data where we already have the connection between the other database and Wikidata via an identifier. What Quora has given is a list of all their topics. Those are now being matched up with Wikidata IDs in Mix'n'Match. I'd say this is exactly what should happen and the PST won't help here. --LydiaPintscher (talk) 13:31, 24 December 2016 (UTC)
I don't mean to be a buzzkill, but in trying Mix 'n' Match for just 10 minutes on Quora's donated topics, it seems it's very messy on their side. It brings me to either 1) question the usefulness of this exercise right now or 2) what our experience is with external databases for which not much care has been put into its maintenance. For example, Quora's Caledonia topic [1] is both about the font en:Caledonia (typeface), and the Latin name for Scotland, en:Caledonia. Another Quora topic, on The Garden of Earthly Delights (2004 movie) [2] contains, erroneously, questions all about the Hieronymus Bosch painting. Do we link to The Garden of Earthly Delights (Q321303) to the wrong Quora topic, even though it has relevant questions for the Wikidata item? Seeking the best practices or advice from the more experienced Wikidata folks here. -- Fuzheado (talk) 16:07, 24 December 2016 (UTC)
I would never use this „valuable“ site as a reference for anything. My first check was Carl Linnaeus (Q1043). One of the both questions available at the moment is „What was Linnaeus’s contribution to evolution?“ Both start with a jokingly hint to the number of his children. The public available answer attributs to him 24 children. The other one needs a login and counts 7 children. The public available answer is wrong, the hidden one is true. The site is a complete waste of time and not a „great“ contribution to knowledge. --Succu (talk) 22:02, 24 December 2016 (UTC)
You seem to be labouring under the mistaken impression that the purpose of adding external-IDs is for "referencing". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:49, 24 December 2016 (UTC)
Maybe you should reread the whole thread?! --Succu (talk) 23:09, 24 December 2016 (UTC)
@Pigsonthewing: What do you consider the purpose of Quora external ID's to be? ChristianKl (talk) 13:11, 25 December 2016 (UTC)
That was a question for the property creation discussion; which you opposed; and in which you were in a minority of one. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:11, 26 December 2016 (UTC)
Why do you think that question doesn't matter when it comes to importanting data and setting standards for the importation process? ChristianKl (talk) 15:15, 26 December 2016 (UTC)
Where did I say it doesn't matter"? Stop trying to put words in my mouth. And stop trying to re-litigate a process where you were clearly out of step with consensus the first time around. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:38, 26 December 2016 (UTC)
Let me rephrase the question of ChristianKl: Does it matter when it comes to importing data and setting standards for the importation process? What do you think Andy Mabbett? --Succu (talk) 21:58, 26 December 2016 (UTC)

Thank you Andy Mabbett for the work you have put into this. Do not get upset by the nagging of the local grouches. I use quora often and even though it might not be great source for some branches of natural sciences it is valuable in lot of other topics. It has its place here among other external ids. Cheers. Wesalius (talk) 07:42, 25 December 2016 (UTC)

Thank you; I won't. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:11, 26 December 2016 (UTC)
@Wesalius: - Please don't frame reasonable questioning of the effort as "nagging." If there are concerns with the approach, better to discuss them now before it gets too far into the process. Thanks. -- Fuzheado (talk) 13:51, 26 December 2016 (UTC)
There is a difference between reasonable questioning and nagging. I experienced a very edgy discussions with the kinds of Jura1 and ChristianKI in the past, so I know what I am talking about when I call it nagging. They do not assume good faith and I sense the same "hate everything from the start" approach here. Wesalius (talk) 19:39, 26 December 2016 (UTC)
@Wesalius: am I wrong or are your judging my above contribution as „nagging“ too? --Succu (talk) 21:49, 26 December 2016 (UTC)
Nah, yours is just a little too agressive (but that is just how I subjectively feel it, no offense intended ;-), I guess thats just the "linus torwalds way of communication"), but at least you stated actual example of what you think is wrong (which is constructive, therefore not nagging). Wesalius (talk) 08:46, 27 December 2016 (UTC)

I have already posted concerns about this dataset on the Facebook group, Wikipedia Weekly, but am going to paste them here as well.
I am concerned about both the non-tertiary nature of this data as well as its cleanliness and usefulness. This data is crowdsourced information with no overarching editing or monitoring. Why is this data useful to either Wikidata or more importantly Wikipedia? It seems like original research at best, garbage at minimum. I believe there are so many sources of metadata that have a higher importance and usefulness and value. I don't understand why this information is being pulled into Wikidata at all. It's not suitable as a citation on Wikipedia. It's just fundamentally unhelpful data.
I am even more concerned with the methodology of this effort. I have had such a negative interaction with Andy Mabbett in the past when I have questioned his past efforts that quickly devolved into an emotional and angry space (on his part). I would like to avoid that if possible here but fear even asking basic questions will trigger a similar response. But I refuse to be bullied and cowed so I am just bracing myself here for the response.
The fact is that Andy has undertaken radical efforts using bots on Wikipedia without doing due diligence or consensus building. While I think the efforts are not intended to be negative this approach ends up having the effect of a hidden or sneaky agenda to very impactful -- often negatively impactful -- work. It comes off poorly and undercuts the potential upside of the efforts. -- Erika aka BrillLyle (talk) 13:39, 26 December 2016 (UTC)


  • "quickly devolved into an emotional and angry space" that ad-hominem attack is an utter mischaracterisation of the time I asked you for evidence to back up an accusation against me and another editor, that you stated was true, and you had none.
  • "crowdsourced information with no overarching editing or monitoring" and that is again you making statements with no evidence to back them up.
  • "The fact is that Andy has undertaken radical efforts using bots on Wikipedia..." Aside from a single run of exactly six Wikidata edits in a 2013 training session, I have never used a bot on any WMF project. So much for "fact".
  • "a hidden or sneaky agenda" It's clearly not me who is trying to bully and cow.

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:50, 26 December 2016 (UTC)

  • I think we should be able to determine the merits of the dataset independently of Facebook, Wikipedia or its proposer. The (in-)accessibility of the website or the suitability of the dataset for authority control purposes or any other, doesn't necessarily impact its suitability for Wikidata. We do a property for Twitter hashtags, so we might as well have a property for other discussion website. One of the problems that occurred to me and hadn't been discussed when the property was first proposed, is that datatype isn't consistent with some of others that have string-datatype. Accordingly, I listed the property for conversion at Wikidata:Properties_for_deletion#Property:P3417.
    --- Jura 15:16, 26 December 2016 (UTC)
I am mystified. As I understand the concept "dataset", Quora is not a dataset, as it does not contain data. Quora seems to be a platform hosting comments. There may be users who find it useful to look at these comments, but there are dangers too: all comments I have seen so far had factual errors in them. Social media like this are notorious for echoing rumours, urban myths, deliberate falshoods, etc. I don't see at all why Wikimedia should compromise its "vision" by promoting Quora. - Brya (talk) 04:22, 27 December 2016 (UTC)
Technically, the mapping itself is a dataset. And I don't think we got a policy about a mapping to imply pointing to a structured dataset to be stored here, or any judgement about the data it points to have been a requirement to store a mapping up to know. In my understanding, one of the keys about wikidata is making links between datas on the internet and the web thanks to the stuffs they are about. author  TomT0m / talk page 09:44, 29 December 2016 (UTC)
Well, if you feel that way you should replace the
"Wikidata acts as central storage for the structured data of its Wikimedia sister projects including Wikipedia, Wikivoyage, Wikisource, and others."
on the main page by
"Wikidata is a mapping of places on the web holding dubious stuff that is to be imported into Wikimedia sister projects including Wikipedia, Wikivoyage, Wikisource, and others." - Brya (talk) 12:59, 29 December 2016 (UTC)
Who are you addressing here? Your edit has broken the threading and divorced User:GerardM's comment (below, under a sub-heading which you inserted) from that to which it was a reply. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:21, 30 December 2016 (UTC)


What is relevant is that when Wikidata aims to provide service. As it is we have an awful track record of linking to other crowd sourced projects. It only is accepted when it is squeaky clean and has a reliability higher than our own. We do not reach out and cooperate because of this stance. For "professional" and "official" sources meanwhile we tolerate error rates of 25%.. I am totally happy that when people seek information they find Wikipedia or OpenStreetMap and when they seek questions Quora. It adds value. Thanks, GerardM (talk) 08:25, 27 December 2016 (UTC)
Which source do you mean with 25% error rates? ChristianKl (talk) 11:29, 1 January 2017 (UTC)
I have answered that before and it was pooh poohed away. The point is that many of the crowd sourced projects are in a much better position and are actively maintained. Like Wikidata they are not absolutely good nor are they absolutely bad. The fact is that we do not cooperate and it is a shame because as a result both projects are worse of. Thanks, GerardM (talk) 08:17, 8 January 2017 (UTC)

Using the interface with keyboard

Hello all, and happy new year :)

We are currently working on improving the user interface, especially regarding the workflow with keyboard. If you are used to navigate on websites using more keyboard than mouse, could you tell us how does it works for you on Wikidata, which issues do you encounter while navigating with keyboard? What would you like to see improved?

Thanks, Lea Lacroix (WMDE) (talk) 10:29, 3 January 2017 (UTC)

This matter is of particular importance for people with sight issues, who hear pages read to them by software. I'll ping some users in that situation, on en.WP. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:23, 3 January 2017 (UTC)
I add statements sometimes just using the keyboard. There are a few little annoying issues, like that the property suggestions box doesn't disappear when switching to the next input. Also, if you press Enter on "Remove" for a qualifier, it will (sometimes?) save the statement instead of just removing the qualifier. Also, the "More"s in the suggestion boxes pile up quite a bit. --Yair rand (talk) 23:00, 3 January 2017 (UTC)
Is phab:T149798 what you mean by the property suggestions not disappearing? There's also phab:T88804 for multiple "More"s. - Nikki (talk) 20:06, 6 January 2017 (UTC)
I've also created phab:T154869 for removing qualifiers/references. - Nikki (talk) 14:54, 8 January 2017 (UTC)
Thanks, Andy, for pinging me. I mostly change interwiki links here (either deliberately or as a side effect of my history-merging work), but I sometimes add/modify properties. As we discussed in person, I don't have any problems using Wikidata with my screen reader JAWS beyond those experienced by sighted people (like trying to figure out why an interwiki change isn't going through). However, I'm happy to test out any improvements made in this area. Graham87 (talk) 08:42, 4 January 2017 (UTC)

Alias for name in English, written in a non-roman alphabet

In the "alias" field for English-language named entities, previous edits have included names in non-roman alphabets. For example: the data item for Plato has aliases for English in the following alphabets: Arabic Greek, Hebrew, Korean, and Chinese (ZH). What's the rule for this? And when I encounter these, may I manually remove them in an edit? -- Deborahjay (talk) 18:24, 7 January 2017 (UTC)

Yes they must be removed. --Thibaut120094 (talk) 19:40, 7 January 2017 (UTC)
This edit by User:BotMultichill added the non-Roman alphabets, together with other non-English aliases for Plato. I complained about such edits here, here, and here to no avail.
My view of the situation is that the source BotMultichillT is relying on, ULAN, in effect lumps all alias together. The source has a capability to separate sources by language, but seldom exercises that capability. Wikidata, on the other hand, separates sources by language.
I view this bot as making the defacto claim that English is the universal language and every word in every langugage is an English word.
The bot is very persistent about restoring non-Englsh aliases. Thibaut120094's edit to Plato won't stick unless something is done about the bot. Jc3s5h (talk) 20:41, 7 January 2017 (UTC)
Pinging Multichill. --Thibaut120094 (talk) 20:42, 7 January 2017 (UTC)
Hey, there we have Jc3s5h. The person who starts nasty topics and disappears. I see you added spreading lies to your repertoire too. As I told you before: The bot won't add non-Latin1/2 aliases (actual code for the people who do understand). So no, these won't be added again. Multichill (talk) 21:20, 7 January 2017 (UTC)
Apparently the bot will no longer add non-English aliases which are written in Latin alphabets. But as of 5 January it was still adding non-English aliases. Jc3s5h (talk) 22:21, 7 January 2017 (UTC)
That's intentional as explained by Jane. Multichill (talk) 23:28, 7 January 2017 (UTC)
I can certainly understand that an English-speaking person or English-language work might refer to a person by the name the person used in his/her naive language, as well as an anglicized name (both Frederick the Great and Konig von Preussen Friedrich II might be found in English works). But I really don't understand why we would list Francesco Bacone as an English alias for Francis Bacon (Q37388), particularly since we have an Italian language entry under the name Francesco Bacone. ULAN seems to only have a preferred name and variants without distinction as to which language the variant is in. But since Wikidata is providing multiple lists of aliases, in a number of languages, why are we putting every variant that is written in a Latin alphabet in the list of English aliases? Jc3s5h (talk) 01:39, 8 January 2017 (UTC)
This is clearly a difficult concept for you to understand, but as I tried to explain in the other thread, the spelling is not just a factor of what the actual artist called himself/herself or how the artist signed his/her works in whatever country he/she was working in at the time. It also has to do with historical collections and how the artist name is spelled in those collections. Specifically, museum holdings that collaborate with the Getty, and yes, English is their basic language and thus the "lingua franca" of the Getty. I think you will find that most museums are willing to defer to their expertise in name-spelling for most artists in Western collections, and these days for quite a few Asian artists as well. We still are looking for a "Getty of Russia/Finland", "Getty of Africa", "Getty of Japan" and "Getty of China". Suggestions are welcome. Jane023 (talk) 08:54, 8 January 2017 (UTC)

How can we solve the "mystery deletion" issue?

Three items have just popped up on my watchlist, as deleted. There is no way for me to see what these items are.

I can ask the deleting admin, or another, to find out for me, but that's time consuming for both of us and may well be a waste of time if I agree with the deletion, and should not be necessary.

How can we stop this from happening?

@Lydia Pintscher (WMDE): so that this is on the Dev. team radar. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:42, 7 January 2017 (UTC)

Mind to give at least some examples Mr. Mabbett, or should the developers solve riddles too? --Succu (talk) 22:38, 7 January 2017 (UTC)
Is the concept of a deleted item too enigmatic for you, Succu, or are you indulging in some gratuitous trolling? Here's an example of a deleted item: Q27684138. What was it? Per Andy, who knows. hth --Tagishsimon (talk) 00:30, 8 January 2017 (UTC)
See phab:T148279. - Nikki (talk) 09:26, 8 January 2017 (UTC)
It has been noted and previously requested at Wikidata:AN[3] that deleting administrators put a more informative deletion, or retain some of the labels. DR archives become useless without informative labels, read and weep Wikidata:Requests for deletions/Archive/2016/12/30  — billinghurst sDrewth 09:46, 8 January 2017 (UTC)

Yiddish language in Hebrew alphabet

In several items whose subject is a person, I've noticed that

  • The Yiddish-language spelling is given as an alias of the Hebrew language label
  • Yiddish does not appear as a language option when I pull down "all entered languages."

This is the case even for an item such as Szmul Zygielbojm with an entry in the Yiddish Wikipedia. Yiddish is a language written in the Hebrew alphabet, but with its own orthography that in most cases differs from the Hebrew. Hebrew transcription of foreign name (e.g. from the French) already varies from a transliteration of roman-alphabet orthography, thus requiring actual aliases in Hebrew reflecting this common discrepancy (e.g. initial letters, affecting the item label). Kindly explain this, as it impacts my conducting future edits. -- Deborahjay (talk) 21:19, 7 January 2017 (UTC)

@Deborahjay: Yiddish has it's own language code "yi". I did the babel trick on your user page. You should now see these languages when you try to edit an item. Multichill (talk) 23:27, 7 January 2017 (UTC)
For people and places, it's quite common to add aliases for native names, particularly if they're written with the same script, e.g. Warsaw (Q270) has "Warszawa" as an English alias. "All entered languages" only shows languages which currently have a label, description or alias, not all available languages. You can change your default language by changing the user interface language (there's a language switcher at the top of the page) and you can change which other languages are shown by default by changing the Babel box on your user page like multichill did for you above. - Nikki (talk) 09:17, 8 January 2017 (UTC)
Thank you, @Nikki:, I see that now. I suppose the initial set, which appeared prior to the Babel boxes, was offered based on my IP address coming from Israel: English-Hebrew-Arabic-Russian but not Amharic or the languages of other large immigrant communities. An interesting reflection of sociolinguistics in my adoptive country. -- Deborahjay (talk) 10:10, 8 January 2017 (UTC)

High Mass (Q5755916) - disambiguation page or not?

High Mass (Q5755916) is stated as a disambiguation page here on Wikidata, which is correct for the connected English Wikipedia article, but not for the Swedish one. How to handle this? //Mippzon (talk) 00:17, 8 January 2017 (UTC)

Each article needs a distinct wikidata item. EN is a disambig page. SE is about a particular form of mass mainly associated with the Swedish church. So delink one of the wikilinks, and make a new wikidata item to which it can be linked. --Tagishsimon (talk) 00:27, 8 January 2017 (UTC)
  • The svwiki article shouldn't be linked from that item as it's not a disambiguation page. We might have a significant number of articles lost in items for disambiguation pages.
    --- Jura 06:34, 8 January 2017 (UTC)
@Jura1, Mippzon: The svwiki article is not limited to only the Swedish lutheran church. It mention that it is applied to several churches, and name Catholic church and Swedish lutheran church as examples. Thereafter an example from in unnamed church is described in detail. I agree that the enwiki and svwiki-articles should not be present in the same item. The Swedish article is so ambiguous to its nature, that I would call it a semi-disambig. It looks like it need some hard love, to be useful! --- Innocent bystander (talk) 09:21, 8 January 2017 (UTC)

Is there any notability to this?

MechQuester (talk) 04:07, 8 January 2017 (UTC)

  • Why not? --- Jura 06:35, 8 January 2017 (UTC)
I was asking as it seems to have 10 different external identifiers, but looking at some of them, I'm not really sure if they are about the same person.
--- Jura 07:06, 8 January 2017 (UTC)
Appears to be about the mother of this not-very-notable person - Q24248446, a 21-year-old anti-Ahmadiyya activist, and has been created along with Wikidata items on all his other immediate family members. Follow the trail at English Wikipedia, probably all has to be cleaned up.--Pharos (talk) 07:36, 8 January 2017 (UTC)
You are kidding me. This person is notable in her own right if only because of the award she received. Another reason is that she is the mother of another item we hold. In my opinion the discussion of notability is different from Wikipedia. Much has been added to Wikidata where a Wikipedia has nothing. Thanks, GerardM (talk) 09:05, 8 January 2017 (UTC)
I am extremely skeptical that any members of this family have actually received The 500 Most Influential Muslims (Q16823995), there are no reference for this and a search of the award's website finds nothing. It is perhaps a sign of maturity of Wikidata that it is now attracting Wikipedia-like vanity spam.--Pharos (talk) 13:54, 8 January 2017 (UTC)
The trail of sockpuppets makes me doubt all of the statements. Someone who behaves like that is not acting in good faith and is not editing with a neutral point of view. It wouldn't be the first time someone has linked incorrect IDs and added completely false statements to try and appear notable. - Nikki (talk) 14:09, 8 January 2017 (UTC)
I removed all the identifiers since they were clearly about other people. I also checked all the issues of The 500 Most Influential Muslims (Q16823995) and I didn't find any member of her familiy to be mentioned in there. The only reason why Q28147796 could be notable is because she is the mother of Q24248446, another dubious item. --Pasleim (talk) 18:43, 8 January 2017 (UTC)
I decided to delete Q28147796 because there were too many suspicious statements on the item itself, on the linked items and on the items edited by the same users. --Pasleim (talk) 20:30, 8 January 2017 (UTC)

Some people has neutral sex by body modification

Norrie May-Welby (Q2184751) and Mao Sugiyama (Q26702590) have neutral sex. These people should have neutral sex as value of instance of (P31)sex or gender (P21). But Wikidata does not have this sex.Lava03 (talk) 12:58, 8 January 2017 (UTC)

They have instance of (P31)=human (Q5) and human (Q5) is neutral, or you mean something of different? --ValterVB (talk) 13:52, 8 January 2017 (UTC)
sex or gender (P21).--Lava03 (talk) 14:06, 8 January 2017 (UTC)
Anyone - you included - may create an item called, say "neutral sex", and use that. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:59, 8 January 2017 (UTC)

Property for reverse side

Is there suitable P for images of reverse side of paintings?--Stolbovsky (talk) 20:59, 8 January 2017 (UTC)

There are is recto of (P2681) and is verso of (P2682)
--- Jura 23:30, 8 January 2017 (UTC)

Item without sitelink with imported from Wikimedia project (P143)= wikisomething...

We have a lot of item that don't have sitelink, but have like source: imported from Wikimedia project (P143)=some wiki (ex. Q27048749). It's correct keep this source or is more correct deleted this kind of source because not more verifiable? --ValterVB (talk) 14:08, 8 January 2017 (UTC)

I would personally prefer to remove them. They aren't valid sources (according to Help:Sources) and if the page has been deleted, we can't even use them as a way to find a proper source. - Nikki (talk) 14:36, 8 January 2017 (UTC)
I would never remove imported from Wikimedia project (P143), I would always replace it with real sources. Just removing imported from Wikimedia project (P143) is a step backwards because you don't know any more where the data came from. Multichill (talk) 15:44, 8 January 2017 (UTC)
Adding P143 without specifying which page the information comes from, is not much better than "P248:Google search engine". -- Innocent bystander (talk) 16:17, 8 January 2017 (UTC)
I often add statements without sources while adding information from a source that I apply to one of the statements. This is more a factor of shorthand (and not being able to easily copy references, something I requested in the wishlist survey). When I do this on existing items, I tend to remove the "imported from W..." references if they match my source, but I don't add a reference except to one statement (often "described at url" or some authority control property). I agree that these are fairly useless anyway. What does it matter if someone added the item by hand or imported it? Without a useful reference the "imported from" hasn't ever helped me find anything (yet). Jane023 (talk) 08:43, 9 January 2017 (UTC)
I have occasionally used the "imported from" references when investigating bad values to see whether the original page was wrong or whether something went wrong when importing the value, or when trying to find a proper reference for something but neither of those are possible if the page no longer exists. I can see some advantages in the "imported from" statements when the page is still available (although I think the disadvantages outweigh the advantages) but I really don't know what I would use an "imported from some Wikipedia" reference for if the page no longer exists. - Nikki (talk) 09:31, 9 January 2017 (UTC)
"If the page no longer exists"? Yea, but that statements does not tell WHICH page?! These imports are often done from categories and lists. Those pages cannot easily be tracked by the present or deleted sitelinks in the item. A category on Wikipedia can never have a source attached to it, so it is a very poor way to import things to Wikidata.
One of our users on svwiki used a good source and added "Category:psychiatrist (Q211346)" to a number of Wikipedia pages. Later she learned that she interpreted the old Swedish term "hospitalsläkare" wrong, and removed the category. When she tried to remove "occucation:psychiatrist" here at Wikidata, she was reverted with the argument that the statement was sourced. -- Innocent bystander (talk) 10:52, 9 January 2017 (UTC)

Wikidata weekly summary #242

Butterflies that are also flowering plants

Here is a query to find species of (flowering) plants with Butterflies and Moths of North America ID (P3398). The query is quite ugly due to this long term unsolved issue. It seems like there are numerous organisms with the same scientific name and at the moment those items are conglomerates of properties and IDs of multiple organisms, like this "Spiny oak slug (Q7577673): species of insect". I wonder if it would be possible to improve my query and/or create a list of other such items so we could fix them and in the future pay extra attention to them. Pinging some of the users that were editing those items: @Succu:, @ Magnus Manske:, @Termininja:, @Daniel Mietchen:. --Jarekt (talk) 20:19, 9 January 2017 (UTC)

Problems like this are well known. --Succu (talk) 20:57, 9 January 2017 (UTC)
Succu (and possibly others) , thanks for fixing those cases. I was planning on looking into them this evening. --Jarekt (talk) 05:05, 10 January 2017 (UTC)
Jarekt, I fixed around 160 additional cases. --Succu (talk) 16:55, 10 January 2017 (UTC)

Grant to fund work on Structured Data on Commons

Hi all, the WMF and WMDE just announced funding for work on Structured data on Commons via a grant from the en:Sloan Foundation. You can find the announcement at the Wikimedia blog. More information about the grant is at commons:Commons:Structured data/Sloan Grant. If you have questions, please join us at the structured data Commons talk page, Astinson (WMF) (talk) 20:40, 9 January 2017 (UTC)

Distinction between صواب and ثواب in Arabic

Could anyone who reads Arabic confirm if correctness (Q22688571) is a duplicate of Sawab (Q7428523) or something distinct? If they are distinct, could you please explain what the difference is, and propose a suitable English label for correctness (Q22688571)? If they are the same, would you mind going to the Arabic Wikipedia and merging the corresponding articles so that the items can be merged? Thanks, SJK (talk) 09:00, 10 January 2017 (UTC)

Item for subject of newly created article in a non-Latin alphabet WP

A biographical article created in the Hebrew Wikipedia gives the subject's name in Hebrew as the page name. The lead sentence provides the romanized (Latin alphabet) form of the name in parentheses just prior to DOB-DOD. How and when is an item created here in Wikidata' - and how do I locate it for the purpose of adding more content? E.g. I tried שילה ליפשיץ and Shilo Lifschutz in the "Search Wikidata" field which yielded nothing. -- Deborahjay (talk) 11:40, 9 January 2017 (UTC)

An item can be created as soon as there is a Wikipedia article in any language. If you have thoroughly searched for existing items and did not find anything, feel free to go ahead and create one (there’s a “Create a new item” link below the Wikidata logo). In case of an article in a non-Latin alphabet Wikimedia project it is particularly advisable to add an English label as well, to help others to find this item. If you add authority control properties (with external-ID type) to the new item, we might be able to spot duplicates (in case there are any) and merge both items later to a single one. —MisterSynergy (talk) 12:11, 9 January 2017 (UTC)
@MisterSynergy:The only search I conducted is described above (last sentence before my sig and your reply). Part of my activity here is exactly about adding a label in English. I performed a Merge once, but now am interested in avoiding workarounds if I can cultivating optimal techniques, with alternatives for exceptional cases. -- Deborahjay (talk) 13:29, 9 January 2017 (UTC)
There is little you can do other than thorough searching. If you don’t find anything, don’t hesitate to create a new item. It is not at all a problem if it needs to be merged later—we definitely don’t run out of Q-IDs ;-). The authority control properties help a lot when one wants to identify potential duplicates, because external IDs typically map 1:1 to Wikidata items and are thus expected to be unique in Wikidata. —MisterSynergy (talk) 13:41, 9 January 2017 (UTC)
If there are interwikis present in the article, it might be more efficient to have a bot add the article links to existing items.
Once done, Duplicity (Q23751912) might be a good tool.
There are also a few bots that attempt to import Latin script labels from non-Latin WP, at least in Russian and some Asian languages.
--- Jura 10:07, 10 January 2017 (UTC)
@Deborahjay: If the subject has an identifier, like VIAF ID (P214), IMDb ID (P345), ORCID iD (P496), or many others, you an do a look-up using that; see this post for details. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:44, 10 January 2017 (UTC)
I have added some info in this ticket including VIAF, ISNI and NLI. Geagea (talk) 03:13, 11 January 2017 (UTC)

Why are coordinates rounded unneccessarily?

I have a question that worries me somewhat although it doesn't concern my own work. I made some edits where i corrected a couple of coordinates for churches in the Baltimore area that had their longitude given as East instead of West, so they turned up in China (e. g. [4]). Now I find the coordinates have been rounded by this edit to the nearest arcsecond which means the precision has been degraded. (1 arcsecond is about 30 m). Why is this? In the example above, the church's position had been spot on inside the building polygon in openstreetmap (except for the East/West mistake), now it is several meters outside. --SevenGrader (talk) 20:50, 10 January 2017 (UTC)

You should be able to set the precision of the coordinates when you edit them. It can go up to 1/1000th of an arcsecond. Just for fun, here is a query showing all churches with a position and an explicit precision. We can see that the position of the Wooden churches of Southern Lesser Poland (Q854267) is only known with as precision of ten degrees ... Koxinga (talk) 21:49, 10 January 2017 (UTC)
Thanks you, I think setting the precision intentionally low is nice when you want it (like in your example), but my problem is: the value was rounded automatically (and unneccessarily) and now I can't get the lost precision back (I tried setting the precision to 1/1000 of an arcsecond) unless I dig into the history of the item. The original value as entered in the English WP was this: [5]. But now on Wikidata we only have this: [6]. Think, for example, someone makes a list with all the positions of the graves in a cemetery. With the procedure in place now, the graves would be automatically placed on a 30m grid, with some different graves likely assigned to the same position. --SevenGrader (talk) 22:14, 10 January 2017 (UTC)
the problem being that map coords can be wrong. or conflict. sources can be single point per city, and hard to tell which are good. we need a reliable process to reflect the precision given and then review it. need a GPS data team. Slowking4 (talk) 03:02, 11 January 2017 (UTC)
In my opinion, the problem is that digits were being thrown away by an opaque process, programmatically, where they should have been kept. --SevenGrader (talk) 09:38, 11 January 2017 (UTC)

Example query broken: Children of Genghis Khan

Click the button "Examples" at then find the example query called "Children of Genghis Khan" and run it: It displays many empty bubbles. I guess a label should be in each white bubble. Cheers! Syced (talk) 08:10, 11 January 2017 (UTC)

Also, less important but still worth fixing: that query generates duplicates, that can be fixed like this. Syced (talk) 08:17, 11 January 2017 (UTC)
Same with the "Music genres" query. I guess it is a bug of the graph view itself? I am using Firefox on Linux. Syced (talk) 08:23, 11 January 2017 (UTC)

Teams of countries of the United Kingdom

Hello. United Kingdom has four constituent parts: England, Wales, Scotland and Northern Ireland. Each part has it own association football federation. (The Football Association (Q9500), Irish Football Association (Q620280), Football Association of Wales (Q275173), Scottish Football Association (Q478320)). Should an English team like Chelsea F.C. (Q9616) has country (P17) fill with England (Q21) or United Kingdom (Q145)? Should a Scottish team like Celtic F.C. (Q19593) has country (P17) fill with Scotland (Q22) or United Kingdom (Q145)? Xaris333 (talk) 16:17, 29 December 2016 (UTC)

I've been using United Kingdom for institutions in any of those, but maybe that's not the right answer in this case (or others)? ArthurPSmith (talk) 16:41, 29 December 2016 (UTC)
country (P17) should only be used with sovereign state, i.e. not with England (Q21) or Scotland (Q22). For sports teams, however, country for sport (P1532) might anyway be a better choice. All administrative territorial entity are allowed values of country for sport (P1532). --Pasleim (talk) 19:32, 29 December 2016 (UTC)
country (P17) sovereign state of this item
country for sport (P1532) country a person or a team represents when playing a sport (not all teams represents their country, they do not all playing international matches)
country of origin (P495) country of origin of the creative work or subject item
Maybe the last one is the perfect choice.
And now I have just noticed that Centro Social Deportivo Barber (Q1940547) which is a football club in Curaçao (Q25279) (Atlantic Ocean) has country (P17) the Netherlands (Q55). Its obvious that we have a problem here. We are not giving the "correct" information. Xaris333 (talk) 20:21, 29 December 2016 (UTC)
Then the English label of P17 needs to be changed: both England and Scotland are countries, and indeed we describe them as such and list them as instances of Q3336843 (labelled in English as "country of the United Kingdom"). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:55, 29 December 2016 (UTC)
@Xaris333: I believe (and constraint approves) that country (P17), country of origin (P495) and country for sport (P1532) should have values of sovereign state (Q3624078). The problem is that in non-English languages England (Q21) is not country (Q6256). You can use qualifier located in the administrative territorial entity (P131) for England (Q21). --Infovarius (talk) 21:01, 4 January 2017 (UTC)
The football at England is famous as the English Football and the teams as the English teams. Not as the UK Football and as UK teams. I think we are doing a big mistake... Xaris333 (talk) 00:19, 5 January 2017 (UTC)
And just see what we have now for England national team... and . England national team is not represents UK, only a part of it, England... Am I the only one who see the problems? Xaris333 (talk) 13:28, 5 January 2017 (UTC)
I think the problem here is that we associates sport teams with nations, and not with what they represents, an organisation. The organisation in the case of English football team is The Football Association (Q9500). And from what I know, UK has more than four teams in UEFA. Also Gibraltar Football Association (Q508376) have a team. And not all football teams in the national leagues of England are located in England. -- Innocent bystander (talk) 13:36, 5 January 2017 (UTC)
Yes. Swansea City A.F.C. (Q18659) is a team in Wales (Q25) but is taking part to the national leagues of England. And other example: Derry City F.C. (Q459622) is a team in Northern Ireland (Q26). But the team is taking part to competitions in Ireland (Q27). Xaris333 (talk) 13:57, 5 January 2017 (UTC)
And the teams of Liechtenstein (Q347) plays in the Suisse leagues. -- Innocent bystander (talk) 15:22, 5 January 2017 (UTC)
So maybe to change the scope of country for sport (P1532) to "country/organisation"? --Infovarius (talk) 23:03, 6 January 2017 (UTC)
You mean to add the associations? Not the countries? Xaris333 (talk) 05:08, 7 January 2017 (UTC)
Something alike, yes. --Infovarius (talk) 10:51, 12 January 2017 (UTC)
Gibraltar is a "British Overseas Territory"; it is not part of the United Kingdom. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 00:07, 7 January 2017 (UTC)

Petscan should have an option to add description

AFter doing a bunch of Petscans, I think the application should be able to add descriptions as well, as an option. MechQuester (talk) 15:11, 9 January 2017 (UTC)

I agree, maybe quick statements could be merged to PetScan? Now you can add descriptions using the quick statements tool. --Stryn (talk) 16:33, 9 January 2017 (UTC)
Maybe you could try this new Descriptioner tool which uses SPARQL. Matěj Suchánek (talk) 17:24, 9 January 2017 (UTC)
I had used quick statements and that it could be integrated into petscan. It would be a bit convienient. MechQuester (talk) 17:34, 9 January 2017 (UTC)
When you compare this to the existing tool by Magnus to create automated descriptions, it is pathetic. Try this for instance. Just add some statements like what it is and it will change... PERFECT. I wonder all these years why people bother adding descriptions. Static descriptions suck big time. Thanks, GerardM (talk) 17:43, 9 January 2017 (UTC)
Reasonator does not even edit descriptions. MechQuester (talk) 18:56, 9 January 2017 (UTC)
The descriptions it creates are not saved but they are at least useful and they change for the better when you add relevant statements. They are also available in other languages. Thanks, GerardM (talk) 19:28, 9 January 2017 (UTC)
Descriptions are useful as for the stuff upon which I am working that there is often insufficient information to identify the object or to suitably differentiate. I alos find them lag significantly compared to static descriptions.  — billinghurst sDrewth 11:35, 10 January 2017 (UTC)
I use automated descriptions to disambiguate; I add statements and thereby understand what is what. The problem with static descriptions is that they are often wrong and do not get updated. This is where automated descriptions shine. They work in any language. Thanks, GerardM (talk) 14:27, 10 January 2017 (UTC)
What I don't like with automated descriptions is, that they don't work in any language. For example, the German automated description of May Andersen (Q436251) is "Dänische Modelin (*1982) ♀" [7]. That are three grammar mistakes in this short phrase. Automated descriptions are okay for reasonator but they are not ready to be displayed on Wikidata. --Pasleim (talk) 15:10, 10 January 2017 (UTC)
It is a matter of improving the scripts for automated descriptions. Perhaps with the the Wiktionary support it will be even easier to conjugate. When you state that it is not ready for Wikidata, do understand that the quality of manual descriptions remains substandard. The question is what is more important. Thanks, GerardM (talk) 16:00, 10 January 2017 (UTC)
Then can we make the personal display type preferential through gadgets?  — billinghurst sDrewth 01:40, 11 January 2017 (UTC)
@GerardM: Try this, for example. Do you still feel happy with automated descriptions? --Infovarius (talk) 13:49, 12 January 2017 (UTC)
I cannot read Cyrillic but it is much better than a number. The descriptions are shown to me in English and there is little for me in there. That is not a problem. When someone starts adding values to these items the descriptions will change. So I am happy with it. Thanks, GerardM (talk) 16:15, 12 January 2017 (UTC)

Would the creators of Petscan work on integrating Descriptioner? Both together would be nice MechQuester (talk) 20:05, 9 January 2017 (UTC)

+1 to MechQuesters suggestion! @Matěj Suchánek: except that some of us just crap at SPARQL! Whereas Petscan allows the reasonable search and results, plus it is easy to modify a query and to have it permanent. For example, when I am adding a series of biographical articles from a biographical dictionary at English Wikisource, Petscan is perfect to generate a list of works, however, at that point I am stymied. Now if Petscan gave me a query to plug into Descriptioner then it may be somewhat less painful (and a learning exercise), however, at this point NA-AH!  — billinghurst sDrewth 12:57, 10 January 2017 (UTC)

Maybe it would be the least amount of coding for the tool maintainers if Descriptioner could take as an input a pagepile (which is a possible output of petscan) besides SPARQL Wesalius (talk) 21:47, 11 January 2017 (UTC)

Or @Wesalius: Petscan generates a SPARQL query to plug into descriptioner.  — billinghurst sDrewth 03:09, 12 January 2017 (UTC)
Yeah SPARQL is like programming and needs a bit of learning curve and not all can learn code. It would be great if both could work together. MechQuester (talk) 14:52, 11 January 2017 (UTC)

P155 and P156

Why follows (P155) and followed by (P156) can only can only be used as a qualifier (Exceptions are possible as rare values may exist). They are properties that can be use as properties. Xaris333 (talk) 21:19, 9 January 2017 (UTC)

(Relevant discussion.) Can you give an example where non-qualifier use would make sense? --Yair rand (talk) 03:13, 10 January 2017 (UTC)
Years, such as 2016 (Q25245). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:33, 10 January 2017 (UTC)
2016 (Q25245) is an instance of year (Q577), leap year (Q19828) and leap year starting on Friday and ending on Saturday (Q217036). If the property follows (P155) is not used as qualifier, the meaning is ambiguous. Does it refer to the prior year, prior leap year or prior leap year starting on a Friday? --Pasleim (talk) 12:49, 10 January 2017 (UTC)
Fair point. So how should 2016 (Q25245) be changed? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:15, 10 January 2017 (UTC)
Pasleim's changes to the item are appropriate, although calendar year (Q3186692) seems more appropriate than year (Q577). I do wonder, though, how or whether the Julian and the Gregorian years should be accounted for in the same item (for example, 2016 (Q25245) having a different start time (P580) of '1 Jan 2016' in the Gregorian and '14 Jan 2016' in the Julian). Mahir256 (talk) 14:37, 10 January 2017 (UTC)
@Multichill: I was trying to make this clear in the case of Gregorian calendar days. Mahir256 (talk) 14:14, 10 January 2017 (UTC)
2016–17 Premier League (Q23009701) --> follows (P155) --> 2015–16 Premier League (Q19346732) Xaris333 (talk) 12:19, 10 January 2017 (UTC)
@Xaris333: which is only true within the _scope_ of "premier league season", but not in terms soccer tournaments. To follow it has to have some scope, and if it has scope than it is a qualifier. 2017 Australian Open tournament follows ... 1) 2016 Australian Open; 2) the 2017 Sydney International; 3) 2016 US Open.  — billinghurst sDrewth 13:02, 10 January 2017 (UTC)
How do you suggest to show the next item (the next season) of an item? Xaris333 (talk) 14:14, 10 January 2017 (UTC)
instance of (P31) sports season of a sports club (Q1539532) with qualifiers (part of (P361), maybe?) Premier League (Q9448) and appropriate follows (P155) and followed by (P156) qualifiers. (You may wish to create an item "Premier League season" and use this for instance of (P31) instead.) Mahir256 (talk) 14:37, 10 January 2017 (UTC)
instance of (P31) is the identity of an item. Why do you want to add things about the identity of an item as a qualifier instead of a statement? And why this special case for these?
If you do insist on having it as a qualifier, I don't think instance of (P31) is the right location to put it. I would say 2014: part of the series (P179) -> "The sequence of all years" and add it as a qualifier to that one. Multichill (talk) 15:53, 10 January 2017 (UTC)
Short answer: because "next" and "previous" have sense only in some sequence, and each entity can belong to several sequences. instance of (P31) marks those sequences. P.S. There was some time ago an idea to mark sequences as a P31-qualifier to follows (P155)/followed by (P156) which can be regarded as an alternative. Infovarius (talk) 13:53, 12 January 2017 (UTC)

So many properties with incorrect use, so many similar items which different properties for the same topic... :( Xaris333 (talk) 14:18, 11 January 2017 (UTC)

Let's improve them together! :) Infovarius (talk) 13:53, 12 January 2017 (UTC)
I am trying to find the correct properties for football (and sports items in general) and is really difficult. We must decide if items like Premier League season is the solution. I have a similar discussion Wikidata:Project chat/Archive/2016/12#Which one. I have created series of Ligue 1 (Q27927008). Is that the solution? Can you give me an example how to use P155 and P156? It will be better for everyone if we decided what statements a sport season can have. Maybe is good to created page-guides from different things per subject. For example, a guide for a country item etc Xaris333 (talk) 17:29, 12 January 2017 (UTC)
+1 to building some guiding help pages that cover these matters. I had hoped that the Projects were going to do them. There has been some work on the showcases, so maybe building a good list of showcases (AND MAKING IT FINDABLE) might also be useful.  — billinghurst sDrewth 23:30, 12 January 2017 (UTC)

Region of origin

The cheese mozzarella (Q14088)'s country of origin (P495) is Italy (Q38).

But how to express that the cheese Sarda (Q19621602)'s region of origin is Sardinia (Q1462)?

Is location of creation (P1071) the way to go?

Thanks! Syced (talk) 05:16, 11 January 2017 (UTC)

@Syced: It seems that it would be appropriate for you to propose a more general property "place of origin". After that we can see if we keep or migrate "country of origin".--Micru (talk) 08:46, 11 January 2017 (UTC)
Sarda (Q19621602) is currently listed as an instance of both a cheese and a mammal. It appears that the latter was the intention of its creator. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:52, 11 January 2017 (UTC)
would place of origin be a subset or on par with country of origin? MechQuester (talk) 14:53, 11 January 2017 (UTC)
country of origin (P495) is used for creative works, and if it were migrated to "P:Region of origin" I think that would not be appropiated. With cheese and stuff like that I see no that problem with regionalising. So, if a new property is really needed... I would migrate the part "or subject item" on the English description "country of origin of the creative work or subject item" and I'd keep P495 only for creative works (i. e. country where that book/edition/newspaper was published). Strakhov (talk) 15:02, 11 January 2017 (UTC)
@MechQuester: "country of origin" would be a subproperty of "place of origin".
@Strakhov: "country of origin" has labels in around 87 languages and it is used thousands of items, it wouldn't be easy to change the scope of the property. Besides there has been complaints about its lack of usefulness in the context of publications.--Micru (talk) 20:35, 11 January 2017 (UTC)
How about indigenous to (P2341)? --Infovarius (talk) 13:55, 12 January 2017 (UTC)

Property for MPs

I point out Talk:Q20113710. We are discussing if we must use Property:P2124 or Property:P1410 to indicate the number of MPs of a european parliamentary group.--Caarl 95 (talk) 21:45, 11 January 2017 (UTC)

number of representatives in an organization/legislature or won in elections (P1410) seems better to me. As well as presumably existing exactly for this case, other parliamentary groups have seats within more than one legislature, so it's useful to scope it like this. --Oravrattas (talk) 12:28, 12 January 2017 (UTC)

Petscan interwiki now set for developed queries

To let people know that I have reconfigured the interwiki for Petscan so that the default includes ?psid=, so something like petscan:550574 can be used as permanent shortcuts.  — billinghurst sDrewth 06:36, 12 January 2017 (UTC)

how would this affect Petscan? MechQuester (talk) 16:33, 12 January 2017 (UTC)
Affect? It cannot. Special:Interwiki is a a means to provide standard and easy/ier links. In this case it works in the limited case of psid, though that can be valuable in numbers of situations; as with all the other use of interwikis.  — billinghurst sDrewth 19:31, 12 January 2017 (UTC)


How do I indicate using a qualifier that an award has been awarded posthumously. Instance of what I find as an example is imho wrong. Thanks, GerardM (talk) 09:06, 12 January 2017 (UTC)

I would have still gone for "point in time", and ultimately with a reference.  — billinghurst sDrewth 09:45, 12 January 2017 (UTC)

Petscan not working on some entries

What prevents pet scan from working on entries? w:Devon Hughes for example is a redirect that keeps popping up and doesn't process at all. What gives? Is this a bug? or something? MechQuester (talk) 00:26, 13 January 2017 (UTC)

@MechQuester: Old data would be my guess. There was a wikidata item with a redirect to that page, and I removed the item in the past week or so by merging. See if it is the old item, or the actual item with the interwiki. Without seeing your data/query/... you are otherwise making us guess and wave our hands at the undefined. That is to say, don't use Devon Hughes (Q12333334), instead use Devon Hughes (Q1209452)  — billinghurst sDrewth 03:27, 13 January 2017 (UTC)
Makes sense. I thank you Billinghurst (talkcontribslogs). MechQuester (talk) 05:14, 13 January 2017 (UTC)

German Wikipedia duplicate

The items magician (Q15855449) and illusionist (Q1658894) seem to be equivalent, however in the German wikipedia there are two pages about the same topic, on of them is a list. Could somebody take a look and see if it is possible to merge both? --Micru (talk) 10:22, 6 January 2017 (UTC)

The topic of illusionist (Q1658894) in German wikipedia seems to me broader than the one of magician (Q15855449), as it includes also other artists (e.g. painters of trompe-l'œil (Q468930)). But in the other language versions magician (Q15855449) and illusionist (Q1658894) seem to be equivalent to me, too. Maybe we should move the other language versions from illusionist (Q1658894) to magician (Q15855449) and make magician (Q15855449) a subclass of illusionist (Q1658894)? Valentina.Anitnelav (talk) 10:39, 6 January 2017 (UTC)
What to do with "links here" persons? --Infovarius (talk) 22:43, 6 January 2017 (UTC)
Why are you asking about conflict in de-articles at en-chat and not at Wikidata:Interwiki conflicts? Infovarius (talk) 22:43, 6 January 2017 (UTC)
I had completely forgotten about that page... sorry. You can move the discussion there if you wish.--Micru (talk) 10:57, 7 January 2017 (UTC)
@Micru: as one of the pages linked from this item is a disambiguation page, there is a similar problem as at Wikidata:Project_chat#Disambiguation_page_linked_from_film_item.
--- Jura 06:44, 8 January 2017 (UTC)
Indeed. So is everyone OK, if I move all article sitelinks from illusionist (Q1658894) to magician (Q15855449) and change P106:Q1658894->P106:Q15855449? --Infovarius (talk) 13:36, 12 January 2017 (UTC)
I am doing. --Infovarius (talk) 19:36, 13 January 2017 (UTC)

Importing lists of repositories from Wikispecies

What do folk consider the best way to match, or import, the lists of repositories at:

We need to avoid creating duplicates of existing items, add the short codes as aliases, and check that we're linking to the correct repository, not its parent institution. The Wikidata IDs then need to be written back to the tables in Wikidata, adding the column where required.

Note that a parent category of individual repository-categories, also exists. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's editsery 15:34, 12 January 2017 (UTC)

@Circeus: Could you say something to this --Succu (talk) 15:43, 12 January 2017 (UTC)
Magnus has kindly loaded these into Mix'n'Match: [8]. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:50, 12 January 2017 (UTC)
Another way to help is to have a look at Wikidata:Database_reports/items_without_claims_categories/specieswiki. --Succu (talk) 22:59, 13 January 2017 (UTC)
The problem of duplicates and "collections" not matching "institutions" remain as omnipresent as they were when I commented on this issue back in september. If anything, they are worse with the old list, which contains even more duplicate or obsolete acronyms than the category/individual pages. Circeus (talk)

AutoHotkey for Windows

AutoHotkey (AHK) is a free, open utility for Windows, that automates actions such as typing a particular string; or opening a programme or website. We've started to compile some example scripts for using it with Wikipedia and sister projects, at en:Wikipedia:AutoHotkey. If you have any AHK scripts that are useful when editing Wikidata, please share them there, or in a comment here. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:46, 13 January 2017 (UTC)

Use of P560 for things other than cardinal directions

I've seen direction (P560) used on a lot of railway station items as a qualifier to indicate the terminus of the line or service for adjacent station (P197). Shouldn't those use destination point (P1444) instead? (If yes, would it be possible for a bot to modify all uses of P560 used as a qualifier (for P197), with the value being an item which is an instance of station (Q719456) or its subclasses, to use this property?) Jc86035 (talk) 11:33, 9 January 2017 (UTC)

Unless there is only one more station/terminal in a certain direction, each passenger will have his or her own destination in mind. Jc3s5h (talk) 15:18, 9 January 2017 (UTC)
@Jc3s5h: I was assuming that the destination point would be referring to the vehicle's/line's and not the passengers'. Maybe it might be better to create a new property? Jc86035 (talk) 16:03, 9 January 2017 (UTC)
@Jc3s5h: Or should the cardinal directions use direction relative to location (P654)? Jc86035 (talk) 15:50, 10 January 2017 (UTC)
Looking at the property proposal for direction relative to location (P654), I get the idea that the situation where direction relative to location (P654) might be used is if an adjacent station property is provided, but the name of the adjacent station is obscure, so it is desired to state its relationship to a better-known landmark. So one might enter in the Rutland station (Q2373537) Amtrack station that the adjacent station is Castleton station (Q5050630), and add a direction relative to location (P654) qualifier to indicate that the Castleton station (Q5050630) station is east of Castleton State College (Q1049687). Jc3s5h (talk) 17:14, 10 January 2017 (UTC)
@Jc3s5h: Ah, I see. I think I misunderstood the meaning of the property. direction (P560) is still only supposed to be used for the 16 cardinal and intercardinal directions (some violations), so I think it could be replaced by terminus (P559) for values which are railway stations? Jc86035 (talk) 01:34, 11 January 2017 (UTC)
I can't really understand what you're getting at without a complete example, but I have edited Ethan Allen Express (Q254597) to add the two termini using terminus (P559). I believe Ethan Allen Express (Q254597) now illustrates the correct use of terminus (P559). Jc3s5h (talk) 12:07, 11 January 2017 (UTC)
@Jc3s5h: For clarity:
  • Item Hung Hom station (Q996591), an intercity/metro station, has several uses of the adjacent station property.
  • Before several edits in November 2016 by Mahir256, the "direction" qualifier for each of the adjacent stations was "north", "west" or another cardinal direction.
  • The edits changed the qualifier to use values such as "Lo Wu Station" and "Tuen Mun Station", which are violations of the constraints of the "direction" property as they're not cardinal or intercardinal directions.
  • Both actual directions and railway stations are therefore used inconsistently in the property as qualifier, and it would probably be better to use another property for the railway stations.
Hope this helps. Jc86035 (talk) 06:39, 12 January 2017 (UTC)
My apologies for the problems that have been caused by my edits: direction (P560) was the closest property that I could find for this purpose, especially since the way I used it matched the existing infoboxes for those stations most closely and no constraint violations occurred while using it. Although it seems most expedient to me to repurpose destination point (P1444) for non-event items, I think a new property 'towards' is definitely in order (one which could also apply for highway exits and interchanges). Mahir256 (talk) 15:53, 12 January 2017 (UTC)
@Mahir256: I originally planned to submit a property proposal (I have it on my computer), but I thought that one of the existing properties would suffice. Should I post the proposal? Jc86035 (talk) 03:38, 13 January 2017 (UTC)
You may as well submit the proposal to see whether there's any consensus to formally change the scope of another existing property. (I also recognize, judging from the results of your query, that I am far from the first person to use direction (P560) the way I have.) Mahir256 (talk) 03:42, 13 January 2017 (UTC)
Added at Wikidata:Property proposal/towards. Jc86035 (talk) 15:32, 14 January 2017 (UTC)

Importing "Data licence Germany" Licensed Data

Am I allowed to import data from reports licensed under the Data licence Germany? While the license is obviously not compatible with CC0 I am unsure as to how much the facts themselves are actually protected by copyright law.

My idea would be to add population data from this report. Assuming that none of the introduction texts and other details of the presentation are reused can this legally be imported into wikidata?  – The preceding unsigned comment was added by $DATA (talk • contribs) at 00:43, 13 January 2017‎ (UTC).

I'm not sure that anyone can tell you for sure. The main problem is that it is potentially covered by EU database rights (see meta:Wikilegal/Database Rights and en:Database Directive) because it is a German source. The license they use does say that you can merge their data with other datasets (which I would interpret as waiving the database rights) but it also says that you have to attribute them - there is nothing to suggest that the database rights have been waived if the terms of the license haven't been met. On the other hand, it's not clear if database rights would apply (the second page I linked says that there needs to be "substantial investment") and it's also not clear whether you would be using a "substantial part" of the data. (Of course, what I say is my opinion and should not be taken as legal advice, I have no real understanding of how all this works) - Nikki (talk) 10:52, 14 January 2017 (UTC)


I help to create an new item please help me --White Gold AJ Gaspar (talk) 11:58, 14 January 2017 (UTC)

I've left a note on your talk page, with links to tutorials and help pages. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:40, 14 January 2017 (UTC)

Use of P560 for things other than cardinal directions

I've seen direction (P560) used on a lot of railway station items as a qualifier to indicate the terminus of the line or service for adjacent station (P197). Shouldn't those use destination point (P1444) instead? (If yes, would it be possible for a bot to modify all uses of P560 used as a qualifier (for P197), with the value being an item which is an instance of station (Q719456) or its subclasses, to use this property?) Jc86035 (talk) 11:33, 9 January 2017 (UTC)

Unless there is only one more station/terminal in a certain direction, each passenger will have his or her own destination in mind. Jc3s5h (talk) 15:18, 9 January 2017 (UTC)
@Jc3s5h: I was assuming that the destination point would be referring to the vehicle's/line's and not the passengers'. Maybe it might be better to create a new property? Jc86035 (talk) 16:03, 9 January 2017 (UTC)
@Jc3s5h: Or should the cardinal directions use direction relative to location (P654)? Jc86035 (talk) 15:50, 10 January 2017 (UTC)
Looking at the property proposal for direction relative to location (P654), I get the idea that the situation where direction relative to location (P654) might be used is if an adjacent station property is provided, but the name of the adjacent station is obscure, so it is desired to state its relationship to a better-known landmark. So one might enter in the Rutland station (Q2373537) Amtrack station that the adjacent station is Castleton station (Q5050630), and add a direction relative to location (P654) qualifier to indicate that the Castleton station (Q5050630) station is east of Castleton State College (Q1049687). Jc3s5h (talk) 17:14, 10 January 2017 (UTC)
@Jc3s5h: Ah, I see. I think I misunderstood the meaning of the property. direction (P560) is still only supposed to be used for the 16 cardinal and intercardinal directions (some violations), so I think it could be replaced by terminus (P559) for values which are railway stations? Jc86035 (talk) 01:34, 11 January 2017 (UTC)
I can't really understand what you're getting at without a complete example, but I have edited Ethan Allen Express (Q254597) to add the two termini using terminus (P559). I believe Ethan Allen Express (Q254597) now illustrates the correct use of terminus (P559). Jc3s5h (talk) 12:07, 11 January 2017 (UTC)
@Jc3s5h: For clarity:
  • Item Hung Hom station (Q996591), an intercity/metro station, has several uses of the adjacent station property.
  • Before several edits in November 2016 by Mahir256, the "direction" qualifier for each of the adjacent stations was "north", "west" or another cardinal direction.
  • The edits changed the qualifier to use values such as "Lo Wu Station" and "Tuen Mun Station", which are violations of the constraints of the "direction" property as they're not cardinal or intercardinal directions.
  • Both actual directions and railway stations are therefore used inconsistently in the property as qualifier, and it would probably be better to use another property for the railway stations.
Hope this helps. Jc86035 (talk) 06:39, 12 January 2017 (UTC)
My apologies for the problems that have been caused by my edits: direction (P560) was the closest property that I could find for this purpose, especially since the way I used it matched the existing infoboxes for those stations most closely and no constraint violations occurred while using it. Although it seems most expedient to me to repurpose destination point (P1444) for non-event items, I think a new property 'towards' is definitely in order (one which could also apply for highway exits and interchanges). Mahir256 (talk) 15:53, 12 January 2017 (UTC)
@Mahir256: I originally planned to submit a property proposal (I have it on my computer), but I thought that one of the existing properties would suffice. Should I post the proposal? Jc86035 (talk) 03:38, 13 January 2017 (UTC)
You may as well submit the proposal to see whether there's any consensus to formally change the scope of another existing property. (I also recognize, judging from the results of your query, that I am far from the first person to use direction (P560) the way I have.) Mahir256 (talk) 03:42, 13 January 2017 (UTC)
Added at Wikidata:Property proposal/towards. Jc86035 (talk) 15:32, 14 January 2017 (UTC)


Hi everybody. We works on the development of the Module:Cycling race in around twenty languages. This permits to display tables and infobox in different Wikipedias always with the same line of code. I already know these properties ISO 9:1995 (P2183) and Revised Hepburn romanization (P2125), they work very well and they work in one sense. My questions are : have we similar properties that permits to translate an alphabet in another ? If we lack of properties, can somebody explain me names or norms to launch properties proposal ?

We have started development in russian, japanese, macedonian and latvian. My wish is to continue with chinese and hindi. For functions as generalclassification and listofteams, we have an interest to use translitterated team names. Jérémy-Günther-Heinz Jähnick (talk) 10:25, 12 January 2017 (UTC)

We have a generic propriety for transliterations: transliteration (P2440), but more precise proprieties like Revised Hepburn romanization (P2125) would be more than welcome. --Thibaut120094 (talk) 10:58, 12 January 2017 (UTC)
transliteration (P2440) has the subproperties pinyin transliteration (P1721), McCune-Reischauer romanization (P1942), Revised Romanization (P2001), Revised Hepburn romanization (P2125), Georgian national system of romanization (P2126), Hungarian-style transcription (P2719) --Pasleim (talk) 11:11, 12 January 2017 (UTC)
Transliteration is problematic. They are often language specific. This means that it is not possible to have rules that are always useable. Thanks, GerardM (talk) 16:18, 12 January 2017 (UTC)
Ok, thank you. I finally have found en:List of ISO romanizations but I want the contrary of romanizations. Jérémy-Günther-Heinz Jähnick (talk) 17:15, 12 January 2017 (UTC)
@Jérémy-Günther-Heinz Jähnick: I don`t quite understand your question, but in Macedonian doesn`t exist transliteration, and we have to transcribe every foreign name with specific rules for every language according our Institute for Macedonian Language (Regulatory language body). On example your user name on Macedonian will be „Жереми-Гинтер-Хајнц-Јеник“, and will include both rules for French and German language. --Ehrlich91 (talk) 17:51, 12 January 2017 (UTC)
@Ehrlich91: so can we imagine a property transcription according the Institute for Macedonian Language (similar in its principles with Hungarian-style transcription (P2719)) would be correct ? Jérémy-Günther-Heinz Jähnick (talk) 17:59, 12 January 2017 (UTC)
@Jérémy-Günther-Heinz Jähnick: But, name of the item is already written according that, I don`t see why to use specific property. --Ehrlich91 (talk) 18:03, 12 January 2017 (UTC)
Names of the teams change along the seasons, see UAE Team Emirates (Q135837). Sometimes, the name change during a season like Mitchelton-Scott (Q266770). So we need to associate the transcription at values of official name (P1448). When a team change its name, we add in latin caracters a new value (with date). You on your Wikipedia will have this value until you add a transcription. And because we will have a day numerous languages to translitterate/transcript I hope, we need to make the difference between them to permit the algorithm to give you good values. Jérémy-Günther-Heinz Jähnick (talk) 18:13, 12 January 2017 (UTC) @Ehrlich91: Jérémy-Günther-Heinz Jähnick (talk) 21:34, 12 January 2017 (UTC)
@Jérémy-Günther-Heinz Jähnick: I know that, but still I don`t quite understand why to use transliteration and to add separate properties, when you should make just to update your articles. Personally, I tend to update every single article that I created, so your proposal for me is not adding something new, instead it gives me more work to do. Additionally, transliteration don`t exist in my language. --Ehrlich91 (talk) 19:03, 14 January 2017 (UTC)
@Ehrlich91: It is difficult to explain for me : Ineos Grenadier (Q200009) is Скај for you since its creation in 2010. But for Mitchelton-Scott (Q266770) the name change during seasons. So to display correctly your functions about infoboxes, classifications and listofteams, with the transcription in macedonian, we need to have a property to add as qualifier. If the team change again its name, we enter a value in latin language, a start time (P580), and you and other users that use other alphabets you will add a value as qualifier when you see on your articles you have a team name in latin alphabet. It is a step forward in the adaptation of the algorithm in different languages. Jérémy-Günther-Heinz Jähnick (talk) 19:44, 14 January 2017 (UTC)

Importing "Data licence Germany" Licensed Data

Am I allowed to import data from reports licensed under the Data licence Germany? While the license is obviously not compatible with CC0 I am unsure as to how much the facts themselves are actually protected by copyright law.

My idea would be to add population data from this report. Assuming that none of the introduction texts and other details of the presentation are reused can this legally be imported into wikidata?  – The preceding unsigned comment was added by $DATA (talk • contribs) at 00:43, 13 January 2017‎ (UTC).

I'm not sure that anyone can tell you for sure. The main problem is that it is potentially covered by EU database rights (see meta:Wikilegal/Database Rights and en:Database Directive) because it is a German source. The license they use does say that you can merge their data with other datasets (which I would interpret as waiving the database rights) but it also says that you have to attribute them - there is nothing to suggest that the database rights have been waived if the terms of the license haven't been met. On the other hand, it's not clear if database rights would apply (the second page I linked says that there needs to be "substantial investment") and it's also not clear whether you would be using a "substantial part" of the data. (Of course, what I say is my opinion and should not be taken as legal advice, I have no real understanding of how all this works) - Nikki (talk) 10:52, 14 January 2017 (UTC)

Season of a league and a cup (2)

Continue from Wikidata:Project chat/Archive/2017/01#Season of a league and a cup.

Now we have sports season of league or competition (P3450) that solve the problem. So we have

2015–16 Cypriot First Division (Q19906304) sports season of league or competition (P3450) Cypriot First Division (Q155965)

My question is about instance of (P31) and subclass of (P279). Sould we use

2015–16 Cypriot First Division (Q19906304) instance of (P31) sports season of a sports club (Q1539532) and

2015–16 Cypriot First Division (Q19906304) instance of (P31) Cypriot First Division (Q155965)

I think that the first is enough.

And what about

2015–16 Cypriot First Division (Q19906304) subclass of (P279) Cypriot First Division (Q155965) ? Do we need that property to these kind of item?

Xaris333 (talk) 10:33, 14 January 2017 (UTC)

  • Now that there is a dedicated property, let's use that (first one, in your numbering). No P279 either.
    --- Jura 10:48, 14 January 2017 (UTC)

Great. Thanks. Xaris333 (talk) 11:03, 14 January 2017 (UTC)

You're kidding ? I did no even had the time to notice this proposition and it's already done ? author  TomT0m / talk page 11:44, 14 January 2017 (UTC)
Considering I participated in a discussion on project chat recently considering this question and that I don't remember beeing even notified - correct me if that wrong - that there was a creation, that I was in favor of another solution, I totally feel bypassed here and I'm utterly unsatisfied of what happened here. author  TomT0m / talk page 11:49, 14 January 2017 (UTC)
@TomT0m: Ok. You are right. I am really sorry. Lets discuss it. Xaris333 (talk) 12:26, 14 January 2017 (UTC)
@Xaris333: I guess one fundamental problem is "what is a league" ? Is it an organisation ? a type of competition ? author  TomT0m / talk page 12:33, 14 January 2017 (UTC)
@TomT0m: According to [9] Wiktionary is an organization. Wikipedia says that sometimes is consider as an organization and sometimes as a competition. w:Sports league#Terminology. In my opinion sport league is either a competition or a group of competitions. At w:Premier League says that "The Premier League is a corporation in which the 20 member clubs act as shareholders. Seasons run from August to May." Xaris333 (talk) 12:40, 14 January 2017 (UTC)
@Xaris333: Then my view would be : there is a class of competitions like "top level championship" and that "premier league championship" is a subclass of it. All instances of "premier league championship" are organized by the "Premier league" organization. I think this model definitely semantically works. On the other hand "season of" is a lot more fuzzy and do not really explain the nature of the relationships. This does not play well with the organization and competition concept imho. Could we express the relationships beetween the more basics properties "organized by" and "subclass of" ? I don't really thinks so. This hides a lot of informations that are left implicit and are only found in wiktionary or in the wikipedia articles as a consequence ... imho this is bad. author  TomT0m / talk page 13:44, 14 January 2017 (UTC)
So, what do you suggest to use? Xaris333 (talk) 13:48, 14 January 2017 (UTC)
@Xaris333: I'd propose to drop the "sport season" concept for a start, which is ill defined - if at all - and problematic, as I said previously.
I'd also propose to make clear that leagues items are about organizations and to create items for the types of competitions it organizes : have one items for "premier league championship" and one for "premier league organisation" adding the following claims :
  • ⟨ "Premier league championship" ⟩ subclass of (P279)   ⟨ top level national football championship ⟩
    ... subtype of "football competition" and whats necessary for what we want to do
  • ⟨ "Premier league championship" ⟩ has quality (P1552)   ⟨ values in qualifier ⟩
    organized by search ⟨ "Premier league organization" ⟩
  • and just for the specific competitions :
    ⟨ Premier league 2016 2017 ⟩ instance of (P31)   ⟨ Premier league championship ⟩
What do you think ? author  TomT0m / talk page 14:22, 14 January 2017 (UTC)
@TomT0m: But some championships are organized by associations like Cyprus Football Association (Q473248). And what about the cup seasons? Xaris333 (talk) 15:03, 14 January 2017 (UTC)
<quote>But some championships are organized by associations like Cyprus Football Association (Q473248)</quote>
  • Is then just the league a part of the association ? This seems like a non problem for me. What exactly would be the problem to model this my way ?
  • The cup seasons are competitions as well, aren't they ? No problem for my model, just create items for those classes of competitions. Actually it seems they already exists, see Coupe de France (Q212412)     . This class of competitions is a subclass of football cup as well, and has an organizer as well : French Football Federation (Q244750)     . Cups have specific instances : 2012–13 Coupe de France (Q1137518)      is for example a concrete instance of french football cup. We don't need the cumbersome concept of a "season" for stuffs that are essentially competitions. My model works well for all of those. author  TomT0m / talk page 15:43, 14 January 2017 (UTC)
The new property doesn't imply anything about the organizers. As for the rest, it seems you suggested the same earlier (before the property proposal, here on project chat as well). Besides, if you look at the sample, you will notice that there is much less implicit than in your version. This makes it easier to maintain (at least for those actually doing it).
--- Jura 16:02, 14 January 2017 (UTC)
The only thing I can answer to this is "I don't agree". author  TomT0m / talk page 16:37, 14 January 2017 (UTC)

I give up. We will never find a solution. Everyone will continue add any statement he/she wants. The mess will continue. Xaris333 (talk) 18:10, 14 January 2017 (UTC)

It can be hard sometime to reach a consensus, for sure. I can give you my POV on what is a good model and why I propose this : it's good to have a few concepts and principle on top of which we can reason to define what Wikidata model should be. Mine are : don't create a property when it's not necessary - it's easier to learn and reuse consistently a few existing concepts than to reinvent the wheel on any field in the universe. This allows not to redo the same discussion over and over and repeating the same mistakes everywhere. Base ourself on elsewhere widely used concepts like classes - collection of objects sharing the same properties - use existing concepts and try to fit with external definitions. Use the concepts that are conceptually clear.
That's the spirit of my proposal above. author  TomT0m / talk page 18:24, 14 January 2017 (UTC)
Maybe you could provide a few diffs to relevant edits of yours identifying seasons and applying the relevant league. This way we can check how your wheels turn. Merely by spirit, it isn't going to happen. As far as I'm concerned, the property was created following the our process and unless one decides to delete it, it's there to be used.
--- Jura 18:55, 14 January 2017 (UTC)
I'm unconfortable with your comment since we quited the reasoning part of the discussion and we entered a more personal and/or politics side : you're in favor of a do-ocraty, I know this. This does not make my arguments wrong. Plus if I'm correct you were the one creating this property so maybe it's a conflict ... author  TomT0m / talk page 19:02, 14 January 2017 (UTC)

What does sports season of league or competition (P3450) actually achieve? It would be good if we design and agree upon the full thing and then propose new properties if required. Also since there is no tagging system it would be a good idea to put a note on football related discussions happening in Project Chat in Wikidata:WikiProject Association football.

Previously I had made a suggestion - A Football Season structure - instance/subclass/part of which didn't get consensus but didn't see any disagreements either -- Unnited meta (talk) 04:06, 15 January 2017 (UTC)


I help to create an new item please help me --White Gold AJ Gaspar (talk) 11:58, 14 January 2017 (UTC)

I've left a note on your talk page, with links to tutorials and help pages. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:40, 14 January 2017 (UTC)

How many edits per day?

We are planning an import of 27 000 items to Wikidata. Each item will have several properties. What is the bot edit limit per day, which we can use to calculate the duration of the import? – Susanna Ånäs (Susannaanas) (talk) 12:34, 5 January 2017 (UTC)

There is no hard limit. You can edit as fast as you want until you see that the system gets untable by observing the numbers on Special:DispatchStats. If you aren't doing more than one edit per second you are on the safe site. Using the API, you can create an item and add statements in the same edit, so you can basically complete your import in 27 000s. --Pasleim (talk) 13:20, 5 January 2017 (UTC)

@Susannaanas: You should first review WD:Bots in case you have not.

@Pasleim: Perhaps that should be added to WD:Bots. --Izno (talk) 14:02, 5 January 2017 (UTC)

@Susannaanas: that's a rather large amount of items. What kind of things are we talking about here? Be sure to discus the import itself to make sure it doesn't collide with Wikidata:Notability. Multichill (talk) 17:20, 5 January 2017 (UTC)
The topic has been addressed in this discussion, in the second part regarding elections. We will establish a page for the import and a WikiProject for elections in general also. – Susanna Ånäs (Susannaanas) (talk) 17:30, 5 January 2017 (UTC) Edited 17:40, 5 January 2017 (UTC)
I completely missed that discussion Susanna.
So you want to create items for 27.000 people who might get elected. That's roughly 0,5% of the entire population (of 5,4M), right? How many of them will actually be elected?
Wikidata is not a data last resting place, data should be used. I doubt that would be the case for these 27K items. I looked at the enwp notability criteria for some guidance. I would limit it to the politicians who got elected. Would that work for you? Multichill (talk) 10:48, 11 January 2017 (UTC)
As we see it this is an Open Democracy project. The individual candidates are so to speak nodes in the data set that describes the Finnish Municipal Election 2017. This gives a huge value to the data as all candidates will be queried, and thus create a very interesting graph of data.
We plan to query the data, and in those cases it is very important that we have a complete data set. So we can answer questions like the age distribution of candidates in relation to the economic condition of the municipality and political party they represent.
This is linked data rather than encyclopaedic data. And in future election we can see to how large extent candidates are running again, and so forth.
Additionally we at Yle (Finnish Broadcasting Company) use this data for tagging. We will produce web pages for most of the candidates to assist the people’s choice in the election. And in the same spirit we hope this data set can serve the public as open democratic data. And we hope that Wikidata would serve this purpose. Mickhinds (talk) 11:31, 11 January 2017 (UTC)
  • I don't think citing enwp notability makes any sense. That policy is irrelevant for Wikidata. I think it's good to have high-quality data added to Wikidata. Our criteria is the description by a serious source. The government source about which candidates run fulfills that criteria. The same is true for the Finnish Broadcasting Company.
Wikidata items aren't equivalent to Wikipedia pages. There are Wikipedia article might list a hundred people. If Wikidata wants to store those 100 people every person needs his own item if Wikidata wants to represent the same information as Wikipedia. In this case an election is important enough that the Finish Wikipedia could hosts a link of candidates for various municipal elections.
In the spirit of Open Democracy I also consider this data to be valuable. Wikidata doesn't have anything to gain to be exclusionist in a case like this. ChristianKl (talk) 15:15, 16 January 2017 (UTC)

Internal link fixes

I just fixed an internal link, which had been entered on Wikidata:Administrators' noticeboard as:

and which I changed to:


On most WMF Wikis, this would be regarded as routine housekeeping.

However, I have twice been reverted.

I seek confirmation from the community that such edits, where they in no way change the meaning of the post or section concerned, are acceptable. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:03, 7 January 2017 (UTC)

Personal opinion:
  • In cases like this one, the style of linking by the original editor could have indicated their level of experience
  • The link was clickable before your “routine housekeeping”, so technically there wasn’t even problem
  • “Fixing” it creates extra load on ~350 watchlists (per info page)
I would recommend to leave these things as they are. —MisterSynergy (talk) 14:12, 7 January 2017 (UTC)
@MisterSynergy: Interesting points. Note that my edit was marked as "minor", so would not effect all watchlists. That aside, would you agree that reverting such an edit also creates extra load on watchlists? What if such edits are made at the same time as another change (like, say, a reply)? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:26, 7 January 2017 (UTC)
  • Reverts of these fixes aren’t any better—in terms of unnecessary watchlist load. There can be other reasons which sometimes justify a revert, but in this specific case I do not want to judge about the behavior of @Sjoerddebruin. You initially asked about the housekeeping.
  • “Minor edits” can be filtered, but since each user uses this flag a little differently, many users typically also look at minor edits as well. Particularly on important pages such as the Administrator’s noticeboard.
Regards, MisterSynergy (talk) 14:41, 7 January 2017 (UTC)
Also: marking your edit as minor would hide the page from watchlists during specific settings (like: only show latest change and don't show minor changes). Sjoerd de Bruin (talk) 15:04, 7 January 2017 (UTC)
And again I wasn't pinged in the initial post, this is unacceptable. Sjoerd de Bruin (talk) 15:02, 7 January 2017 (UTC)
Editing other users comments is almost always a bad idea (the revert is just about automatic). The only user who can judge if such an edit is "acceptable", is the user who made the comment in the first place, and instead of going into harm's way, it easier and more helpful to simply post a suggestion on his Talk page. - Brya (talk) 15:41, 7 January 2017 (UTC)
As I noted above, "On most WMF Wikis, this would be regarded as routine housekeeping". "Simple and helpful suggestions" are often ignored; like those I've given you multiple times about your broken indenting, which I've once again fixed here. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:33, 7 January 2017 (UTC)
Personally I prefer that nobody touches edits of other user in talk pages. You can warn the user but don't modify what the user has wrote (except for vandalism or privacy). --ValterVB (talk) 18:53, 7 January 2017 (UTC)
„On most WMF Wikis, this would be regarded as routine housekeeping.“ Would be interesting to know on which projects your editing experience is based. --Succu (talk) 22:14, 7 January 2017 (UTC)

This reminds of previous discussions with the same user where they had also been asked not to touch other users comments and proposals prior to archiving them. See notably Wikidata:Administrators'_noticeboard/Archive/2016/03#Property_creator_rights_of_Pigsonthewing. Apparently they still persist with their problematic approach.
--- Jura 06:41, 8 January 2017 (UTC)

No, Jura. What happened there was that you made a screed of baseless and unsubstantiated (as indeed they still are) ad hominem assertions, including that I routinely removed a certain type of post from my talk page, and my reply to you there was "No Jura, I only remove posts from you on my talk page, and I do so whatever the subject. The reason for that is left as an exercise for the reader. " Nothing, in that regard, has changed. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:51, 8 January 2017 (UTC)

There are often many advantages with converting full urls to internal wikilinks (though that is a little less with Special:Contributions). The existing tools like popups has good versatility with internal wikilinks, so in this case hovering over a Special:Contributions link allows you to view the contributions without having to load the page, and often follow links down the popups. That said, unless there is a good reason to change a link, then the courtesy is to leave it; and if reverted for changing, why pick the fight over who is right. At the same time it surprises me those who wish to continue their arguments and pick fights, surely there are better things to do.  — billinghurst sDrewth 13:27, 11 January 2017 (UTC)

Thank you for acknowledging the benefits of fixing such inks, and raising the popups use-case, which I'd overlooked (despite using that tool!). discussing such cases - on evidence, rather than emotion - is how we reach consensus, and develop polices and practices which best serve our mission. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:11, 16 January 2017 (UTC)


Returning to the subject of fixing links, a section currently near the bottom of this page has a link formatted like:

which renders as:

It serves us far better if the link is formatted as:


so as to render as:

Douglas Adams (Q42)

A link formatted as:

[ Foo]

renders as:


and misleads the reader into thinking it is an external link. Formatting it as:


does not:


Furthermore, bare links do not turn red when the linked item is deleted. Compare:



In none of these cases does fixing such a link (where the fix does not alter the context of a discussion, such as, say, one like this, about bare links) do anything other than making life easier for our colleagues and readers, and thus improving the project. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:27, 8 January 2017 (UTC)

More excuses for bad manners? --Succu (talk) 21:58, 8 January 2017 (UTC)
It seems hopeless. How many users have to revert them? Wikidata:Administrators'_noticeboard/Archive/2016/03#Property_creator_rights_of_Pigsonthewing illustrated that Pigsonthewing edited proposals by other users prior to them being archived leading to the removed parts being replaced with elements that were added by Pigsonthewing on properties themselves. This despite Pigsonthewing being asked not to do so.
At [10] an administrator asked Pigsonthewing: "Please don't edit others messages.". Despite this they reverted the edit once more.
The same reminder had been given to Pigsonthewing before.
--- Jura 22:44, 8 January 2017 (UTC)
You seem intent on derailing this discussion with irrelevances; but I'm sure our colleagues are literate enough to read my refutation of your false allegations, in the original dicussion, to which you conventiently link. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:44, 8 January 2017 (UTC)
Maybe you could answer the question above: How many users have to revert you until you stop editing others users comments?
--- Jura 23:49, 8 January 2017 (UTC)
Nobody has to revert such changes. The majority of times I've made such changes on this wiki, nobody has reverted me, and nobody has complained. The reasons for this, and why a very small number of people chose to make an issue of it in a minority of cases, are again left as an exercise for the reader. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:10, 9 January 2017 (UTC)
Well, if people revert you, why do you keep reverting them? How many times will you be doing that? How would you summarize the feedback received above?
--- Jura 10:03, 10 January 2017 (UTC)
Is this an alpha male fight? 1) Don't change links unless there is true value. 2) Don't revert unless there is true value. If you are playing the change and revert game to prove a point, there is no point it is facile and a pointless pissing competition.  — billinghurst sDrewth 13:33, 11 January 2017 (UTC)

Hidden and open reverts

At Panarctic Flora ID (P2434) starts his next series of hidden and open reverts. He is declaring that our property part of (P361) the database Panarctic Flora (Q28064236) what looks doubtful to me. --Succu (talk) 19:41, 12 January 2017 (UTC)

I haven't "started" anything of the kind. Get the beam out of your own eye, and stop trying to derail this discussion of internal link fixes with irrelevancies. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:02, 12 January 2017 (UTC)
It's related to your manners to fight a "discussion" and not argue at it. --Succu (talk) 20:13, 12 January 2017 (UTC)
Like I said: Get the beam out of your own eye. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:53, 12 January 2017 (UTC)
Citing Matthew 7:5 (Q16146509)? Wow. --Succu (talk) 21:00, 12 January 2017 (UTC)


It became clear from discussions at enwp that many Wikidata pages are giving a person's occupation as linguist when this is unlikely to be appropriate.[11][12][13] Here is one example. Is this simply a problem of the term "linguist" in English being somewhat ambiguous or Wikidata's definition of "occupation" being unclear? Continuing with Kazimiera Zawistowska as an example, how did the classification arise? (I have now removed it). Can I tell its source? In en:Kazimiera Zawistowska the term "linguist" does not and did not occur at all. Maybe in other languages? There seem to have been thousands of such edits in a huge batch in December 2013 and many of them may be wrong. Should they be reviewed or undone? It may be a case where no information is better than wrong information. In a few cases where the enwp article had a (mistaken) category of "linguist" I have removed it.[14] Would some (semi-)automated process have updated Wikidata accordingly? Finally, is this the right forum to be raising issues like this? Thincat (talk) 13:13, 13 January 2017 (UTC)

I think the best way to resolve this is to start adding references to claims and removing those without references. —Wylve (talk) 13:22, 13 January 2017 (UTC)
Mmm. There are thousands without references and personally I have no ability to remove, still less to reference, these. Thincat (talk) 13:26, 13 January 2017 (UTC)
I haven't reviewed your previous discussions on this in detail, but my guess from a brief overview is you probably need a bot to do some cleanup here, so it might be good to post a specific detailed request of what needs to be done at Wikidata:Bot requests. Wikidata items should describe single concepts independent of language, and linguist (Q14467526) and translator (Q333634) are definitely two different concept; wrong statements should be corrected or at least removed. If the occupation (P106) statements for linguist (Q14467526) have a source other than "imported from * wikipedia" they should be preserved or at least examined by a person rather than modified by a bot. And yes this is a great place to ask a question like this. ArthurPSmith (talk) 13:28, 13 January 2017 (UTC)
Can we tell whether diff imported from enwp (or any other wp)? To me it simply appears out of the blue. Thincat (talk) 13:40, 13 January 2017 (UTC)
You could ask @GerardM: but there was apparently no reference attached to the claim, so it had no documented source - feel free to remove such statements (as you just did). ArthurPSmith (talk) 13:43, 13 January 2017 (UTC)
These are some more items with occupation (P106)=linguist (Q14467526), no source and a link to English Wikipedia without categorized in en:Category:Linguists --Pasleim (talk) 13:51, 13 January 2017 (UTC)

I have tried to get an idea of the error rate. I considered 264 women claimed on Wikidata to have an occupation of linguist. 112 had no enwp article so I did no further assessment. For 70 the Wikidata description claimed they were a linguist, philologist, etc so I accepted this as confirming their status. I checked the enwp articles for the remaining 82 people and decided that in 60 cases they were not in fact linguists. I did not go to original sources or look at other language wikipedias so I would not have realised if an enwp article was simply failing to report that they were linguists. In three cases enwp had been placing them in a category linguists but I considered (based on the text) this was incorrect and so removed the category. In these cases there was no transfer error over to Wikidata but there was nonetheless a classification error. So, I suggest that 57 out of 152 people were wrongly classified. Very roughly 37% Thincat (talk) 20:22, 13 January 2017 (UTC)

Usual for GerardM :) As he was importing bulk from some trash Wikipedias. --Infovarius (talk) 13:26, 16 January 2017 (UTC)

Date conversion (regex?) needed

Please could somebody provide a pair of regexes (for a find and replace operation) to any and all dates in the format 27 October 1942 (of which I have many, in a TSV), and to return them in the format used by QuickStatements (i.e. +1942-10-27T00:00:00Z/9. A formula for LibreOffice spreadsheets, or Google Sheets, would also work. Or, alternatively, is there an online tool to do that, in bulk?

Perhaps we could start a page with that and other useful conversions? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:38, 14 January 2017 (UTC)

Use of regexes strongly depends on the software in which they are used and the data which they are applied on, so it is quite difficult to provide a perfect answer. I would just do the following: open the TSV file in a powerful plain text editor (such as Notepad++ in case of a Windows machine), replace all month names by their numerical representation individually (October => 10, no regexes needed if performed twelve times), and then do a simple regex replacement such as (\d\d) (\d\d) (\d\d\d\d)+\3-\2-\1T00:00:00Z\\9. This entire reshaping process takes less than two minutes if performed correctly, so if you need (considerably) more time you’d need to ask for instructions again… ;-) —MisterSynergy (talk) 14:39, 14 January 2017 (UTC)
@MisterSynergy: Thank you; that's the kind of thing I meant, but it doesn't catch single-digit dates like 8 December 2016 (8 12 2016). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:12, 14 January 2017 (UTC)
Oh I see, this was not clear to me. Alter the search string of the regex-replacement to (\d?\d) (\d\d) (\d\d\d\d) (just one extra ?) and do the conversion with the same replace-string as before. The single-digit days need to be fixed now by an additional step: regex-replace (\+\d\d\d\d\-\d\d\-)(\dT00:00:00Z\\9) with \10\2 (all Notepad++-notation). This just adds the missing 0 at the right place. —MisterSynergy (talk) 00:10, 15 January 2017 (UTC)
If you have a set or a range of digits, often a regex like \d{4} for four in a row and \d{1,2} will find single or double. Often it makes regexes simpler to read without having to count all the \d. I like David Child's regex cheat sheet.  — billinghurst sDrewth 10:40, 15 January 2017 (UTC)

Good answers here already, I would like to add that for constructings regexes there are very useful services such as which can show you live on your dataset what will be captured. If you can fit with your dataset in their limit, you can directly use the service for the formatting. Wesalius (talk) 07:27, 16 January 2017 (UTC)

Importing data from Wikipedia template


1) A template in Wikipedia has the official name (P1448) of an item. How can I import that data to Wikidata item? It is a monolingual text and you have to put the language.

2) A template in Wikipedia has the color (P462) of an item. I know how to import that data if is only one color in the template [15]. If the colors are more that one, is importing only the first one. Furthermore, if the parameter has an image in front of the colors, it is not importing anything.

Xaris333 (talk) 20:58, 14 January 2017 (UTC)

1) You have to know the language if you want to import the text. If there is no way to fetch it, then there is no way to import the whole datavalue. None of the automated importing tools that I know can import monolingual text.
2) This is a common limitation for most automated importing tools.
Matěj Suchánek (talk) 21:10, 15 January 2017 (UTC)

I know the language :) So, no tool to help me... Thanks for your response. Xaris333 (talk) 17:29, 16 January 2017 (UTC)

Merge request

Can the entries Template:Plume (Q20736171) and Template:Plume (Q13455439) be merged? They represent the same template.--Auric (talk) 00:03, 15 January 2017 (UTC)

Confirmed and   Done. --Liuxinyu970226 (talk) 00:14, 15 January 2017 (UTC)
See also: Help:Merge.
--- Jura 09:46, 15 January 2017 (UTC)

English Wikipedia RfC on Wikidata planned

Colleagues my wish to read, or participate in discussion at, en:Wikipedia:Wikidata/2017 State of affairs.

It opens:

This page is intended as a preparation for a site-wide RfC about the role of Wikidata on enwiki. Before such an RfC can be had, it seems like a good idea to list a few things here. The section on "uses" should be pretty straightforward: the sections on benefits and disadvantages should be somewhat factual (no "I love it" or "I hate it"), but not necessarily 100% objective

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:53, 12 January 2017 (UTC)

well they fully protected the page, so i guess they do not want my input, but they can discuss among themselves. yet more drama, after the translation extension. Slowking4 (talk) 00:13, 13 January 2017 (UTC)
What saddens me about this kind of debate is that there is actually a bias : there is an elite that takes things in hand, but for them the opinion of the contributors who are discreet and do a huge job in their corner does not matter. The correct course of action would be to see property by property what can be done and make it work perfectly well. And when they are shown examples to support that they do not have the manpower needed to make their updates, they evade the remark ... Jérémy-Günther-Heinz Jähnick (talk) 19:58, 14 January 2017 (UTC)
The worst part, imho, is the "community fork" that seem to be implied by some proposition. For example the impression that "they" made efforts to be welcoming and that "we" did not is kind of depressing. author  TomT0m / talk page
the lack of self awareness is troubling. one is the same admin who said "basically we don't care what they do on other wikis" (except now when it visibly impacts english wiki). hard to know how to get them to work backlogs with teams, rather than blocks. Slowking4 (talk) 02:53, 17 January 2017 (UTC)
About interwikilinking : if someone can add the project WD:XLINK as an alternative to allowing redirects, I'd be forever beholden. author  TomT0m / talk page 09:44, 15 January 2017 (UTC)

  Comment Seen this. Do you know why it pop-ups know ? We're finally done with the frwiki decision and ... yet another round. author  TomT0m / talk page 09:53, 15 January 2017 (UTC)


Hi, I notice that entries for Wikipedia pages (say, atom (Q9121)) don't have a link to Wiktionary definition pages, but there are links to projects such as Wikibooks and Wikiquote. Is there a reason for this? Thanks, Icebob99 (talk) 17:53, 16 January 2017 (UTC)

Hi, it is quite difficult to explain the reason. The simplest explanation I can think of is that the basic Wikidata unit is an entity whereas the main Wiktionary unit is a word. One word can name many entites and one entity can have many names, so the relations are not 1:1. On the other hand, one Wikipedia article (usually) describes one entity and one entity can have one Wikipedia article (per language version, of course). There is an ongoing development of the system Wiktionaries will share the data with. You can find more information at Wikidata:Wiktionary. Matěj Suchánek (talk) 19:17, 16 January 2017 (UTC)

Wikia: property for IDs or create items?

We're discussing the best way to store details of Wikia wikis, at Wikidata:Property proposal/Wikia wiki ID. Further constructive comments be appreciated. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:27, 16 January 2017 (UTC)

What World University and School would like to add to Wikidata

Hi All, (Lydia and Lea),

Nice to chat in the Wikidata office hour just now - .

To begin, CC World University and School would like to add CC MIT OpenCourseWare in its 7 languages and CC Yale OYC to Wikidata.

CC WUaS seeks to accredit on CC MIT OCW in its 7 languages & & to create major universities in all countries' main languages to offer free CC Bachelor, Ph.D., law, M.D. and I.B. high school degrees in all ~200 countries' main and official languages - the "Harvards/MITs/Stanfords of the web". WUaS also seeks to develop in all 7,097 living languages as wiki schools for open teaching and learning.

World University and School's main areas for growth for Wikidata, furthermore, lie in the ~ 10 main areas here -

And WUaS also has another planned wing, which should benefit Wikidata pragmatically, around WUaS's 14 planned revenue streams

How best to begin this process?

Thank you!

All the best, Scott Scott WUaS (talk) 19:37, 5 January 2017 (UTC)

I don't understand what you are proposing - you want to add "opencourseware" to wikidata? You mean add items for each course? I suppose that's reasonable. Please bear in mind wikidata is CC-0 (other wikimedia sites have different rules) so in particular no copyrighted material should be placed in wikidata. ArthurPSmith (talk) 20:06, 5 January 2017 (UTC)
Yes and great - WUaS would like to add CC "opencourseware" to wikidata, and probably items from each course, each course to become a wiki subject page at WUaS, and eventually for translation into each of all 200 countries' main languages, building upon the open licensing of MIT OCW Translated Courses ( . Thanks for the heads' up about CC licensing consistency: MIT OCW is CC 4.0 - ( - so we'd have to work out the licensing further. Scott WUaS (talk) 20:41, 5 January 2017 (UTC)
@Scott WUaS: What do you mean by "adding opencourseware"? Do you mean adding links to the videos/materials? It would be easier if you would create some test items on , there you can create all the properties that you need and show an example. After that you can propose the missing properties on WD:PP.--Micru (talk) 10:03, 6 January 2017 (UTC)
@Micru: In addition to adding the main OCW URL, particularly from MIT OCW in its 7 languages (but also Yale OYC), to Wikidata so that each course can appear in each of the Subjects here - - (and also so that any given course can be listed in the top 1/3 section of, for example, the Economics' wiki subject pages (both macro and micro) - - emerging newly in probably MediaWiki), WUaS would also like to explore creating a wiki page for each MIT OCW course, based on the SUBJECT TEMPLATE - - which informs all ~720 WUaS Wikia pages, and also anticipate/plan for translation of MIT OCW into all countries' main languages for credit. This may include adding all the links from any given MIT OCW course, for example, newly to an individual WUaS MIT OCW course page in Wikidata/Mediawiki. Thanks very much for letting me know about which I didn't know about. Could we possibly please develop this together in, since I don't know coding in Wikidata very well? Thank you. Scott WUaS (talk) 00:20, 7 January 2017 (UTC)
@Scott WUaS: Perhaps it is better that you take a look first at how Wikidata works here: Help:Contents. Then you can see other items and properties to see how it is done. And finally use test wikidata to practice with creating items and properties to model your data.--Micru (talk) 10:56, 7 January 2017 (UTC)
@Scott WUaS: We don't understand what you want to do and how Wikidata could benefit from this.
I looked around (mail and other place) and you seem to be quite active talking with different people, but that doesn't make clear what you want to achieve.
Let's put it bluntly (Dutch directness): Everyone seems to be very friendly and patient with you, but it seems they're just being polite.
You seem to be wasting our time. You seem to be of good faith, but lacking the wiki competence.
Maybe you should start editing Wikidata so you can experience it and see what is all about. Multichill (talk) 12:06, 7 January 2017 (UTC)
- The mission of Wikidata is all knowledge, and multichill: World University donated itself to Wikidata in October 2015 per Lydia and other Wikidatans. (See, too and In World University's knowledge generating mission, and having donated it to Wikidata, and in our CC MIT OCW in 7 languages' centricity, WUaS seeks to create knowledge generating universities and schools in all 8k languages. Building on beginning WUaS in MediaWiki on January 6, 2016, in both English and German (but not in Wikidata at the time), CC WUaS seeks to accredit on MIT OCW in its 7 languages & and Yale OYC to create major wiki universities, and offer free CC degrees, in all countries' main languages. WUaS also seeks to become a growth story for Wikidata. WUaS also has another planned wing, which should benefit Wikidata pragmatically, around WUaS's 14 planned revenue streams - Wikidata, which seeks to be a platform of all knowledge, will further this with World University. WUaS also dovetails with Wikidata's "Welcome" - "Wikidata is a free and open knowledge base that can be read and edited by both humans and machines.Wikidata acts as central storage for the structured data of its Wikimedia sister projects including Wikipedia, Wikivoyage, Wikisource, and others. Wikidata also provides support to many other sites and services beyond just Wikimedia projects! The content of Wikidata is available under a free license, exported using standard formats, and can be interlinked to other open data sets on the linked data web" - for one.
@Micru: Thanks very much. I'll seek to begin building World University and School in to practice creating items and properties to model WUaS's, MIT OCW's and Yale OYC's data. Scott WUaS (talk) 17:57, 7 January 2017 (UTC)
@Micru: Tested some WUaS ideas in yesterday, with this course - - for example, and other MIT OCW and Yale OYC. Out of this, reasons furthermore to explore adding MIT OCW in 7 languages and Yale OYC to Wikidata in the future include for translation (e.g. re Content Translation and Google's Zero-Shot) and for other possible reasons. Thanks for letting me know about Scott WUaS (talk) 17:10, 8 January 2017 (UTC)
@Scott WUaS:Could you link to the test cases you created? ChristianKl (talk) 09:17, 17 January 2017 (UTC)
  • Personally, I figured eventually I'd find samples of what WU is so enthusiastic about. Somehow I thought that the impact for Wikidata would be to create items for courses, e.g. elementary typing (Q28148153), except that such course items would be linking to pages outside WMF sites?
    --- Jura 06:53, 8 January 2017 (UTC)

After testing some WUaS ideas in yesterday, I think that developing MIT OCW with a "front end" in Wikiversity/MediaWiki (with the MIT OCW course description part embedded - in a Wikiversity page) with WUaS's SUBJECT TEMPLATE - 's - bottom two sections using a new kind of WUaS COURSE TEMPLATE newly in Wikiversity/MediaWiki is a good idea (that can also lead to a "course catalog" and "You at WUaS"). By developing MIT OCW-centric World University also in Wikiversity/MediaWiki, there will be many ways to develop items for courses in Wikidata and for Wikiversity/MediaWiki (within the Wikimedia ecosystem/ WMF sites). Scott WUaS (talk) 17:15, 8 January 2017 (UTC)

@Scott WUaS: WUaS seems to just be a collection of your bookmarks, plus descriptions that you copied from the targets, plus a lot of repeated content on every page, that you hope will eventually turn into something more. For example:
What exactly is it that you think could be added to, e.g. Spanish (Q1321)? Please be very specific. Multichill (talk) 10:52, 11 January 2017 (UTC)
@Multichil: Think of these WUaS wiki pages as CC MIT OCW-centric and CC Yale OYC-centric school templates/blueprints/design plans (planning to develop in Wikiversity/MediaWiki from Wikidata), different from Wikipedia articles.

WUaS is seeking too to develop a WUaS COURSE TEMPLATE (based on in Wikiversity/MediaWiki as "front end" - with 1) "WUaS Idea- and Academic Resources" and 2) "WUaS Navigation" below each MIT OCW course - (e.g. course's main section EMBEDDED in a WUaS Wikiversity/MediaWiki page - and with Wikiversity courses - and also "embedding" CC Yale OYC courses, as well as free MITx/edX courses, if possible), and with each MIT OCW course (in each of all 7 languages) in Wikidata, each with a Wikidata Q-item number as a kind of "back end" (linking all of the sub-URLs on each page).

So, to "be very specific" re your Spanish Wikidata entity question, e.g. Spanish (Q1321), WUaS would not add MIT OCW courses to this page - neither in English - e.g. - nor in Spanish - - but WUaS will likely create a new Q-item entry (if one doesn't already exist) in Wikidata re (with its sub-links) to be added to this school / university, newly in WUaS Spanish language wiki page in Wikiversity/MediaWiki, as well as to an individual course wiki page, for example, to "wikify" both Yale OYC and MIT OCW, both in Wikiversity/MediaWiki and Wikidata, in WUaS wiki subjects as schools/universities. Scott WUaS (talk) 20:14, 11 January 2017 (UTC)

Related CC WUaS developing in CC Wikiversity colloquium emerging here - Scott WUaS (talk) 03:47, 12 January 2017 (UTC)

Could we find a way please to embed "just the course part" from the main, middle "frame" or "cell" from each CC MIT OCW course in 7 languages and CC Yale OYC course in a new WUaS MediaWiki template with these SUBJECT_TEMPLATE resources/section headings ( also in the process creating a new WUaS MediaWiiki COURSE TEMPLATE? Here are two courses as examples - - and - . WUaS would also plan to begin a new Q-item entity for each MIT OCW course and Yale OYC in Wikidata, with each course's sub-links to be added in its Q-item.

MIT OCW's addition to and value for Wikidata, in terms of education, knowledge generation, and technically too, would partly lie with CC WUaS offering free CC university and high school degrees (accrediting on CC MIT OCW and CC Yale OYC) in multiple languages. 

In developing CC WUaS with CC MIT OCW and CC Yale OYC in CC Wikidata, we might also better be able to anticipate further translation of MIT OCW into countries' main languages for accrediting WUaS Bachelor, Ph.D., law and M.D., as well as I.B. high school, degrees in CC WUAS in MediaWiki (rather than with CC Wikiversity). 

I'm writing further about embedding free, highest quality, OpenCourseWare, in general, and particularly about edX, HarvardX and MITx courses. While some edX courses charge money (while offering their courses free at the same time), and WUaS's planned COURSE TEMPLATE probably in MediaWiki wouldn't want to post this, the following HarvardX JuryX course is archived and is free: 

Here's the archived Harvard Law JuryX course with Professor Charles Nesson (which I took this past autumn) - ; its videos with a history of the jury process plus other valuable instructional interviews and resources are still available. How best to plan a COURSE TEMPLATE (again based on, but anticipating similar anomalous OCW courses in terms of free OpenCourseWare in many countries' languages. How to embed this JuryX course, and plan for embedding many different web sites' OCW in many different languages (beyond the CC MIT OCW in 7 languages and CC Yale OYC)?   (And here's the beginning CC World University and School online law school - - planned in all countries' main languages (since the 10 other beginning law schools part way down), accrediting on CC MIT OCW for Bachelor and Ph.D. degrees (before we move into a new wiki emerging from Wikipedia's database, Wikidata in 358 languages). This JuryX course in a WUaS planned COURSE TEMPLATE will probably emerge in these law schools in various languages, as well.  Scott WUaS (talk)

Social Networks Archival Context data donation

I have secured another bulk data donation, courtesy of the University of Virginia: 128K of Social Networks Archival Context IDs, matched to Wikipedia articles, for SNAC Ark ID (P3430). Magnus is kindly uploading these as I speak. This will obviously take a while.

I've also enquired about any unmatched IDs that they might have, that we can work on in Mix'n'Match. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:17, 10 January 2017 (UTC)

@YULdigitalpreservation: FYI. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:41, 11 January 2017 (UTC)
@Pigsonthewing: Thank you for getting the data donation. YULdigitalpreservation (talk) 13:26, 17 January 2017 (UTC)

Looks like these are all now imported. Thanks, Magnus! Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:43, 17 January 2017 (UTC)

Properties for naming married women

Just seeking feedback on the process for naming married women, I cannot find a good example in our help text. So what I have been doing is ...

  1. add birth name
  2. add first name(s), qualify with series ordinal as required, and preferred rank, if not first used is not preference
  3. add birth family name, qualify series ordinal: 1
  4. add married family name, qualify series ordinal: 2
  5. depending on how known after marriage, add preferred rank (usually to married name for period where I am working.)

Example at Georgia Wood Pangborn (Q20739165). Improvements for suggestion? Anyone feel that it is inappropriate?  — billinghurst sDrewth 04:36, 12 January 2017 (UTC)

Qualifers can be start time (P580)/end time (P582). --Infovarius (talk) 13:57, 12 January 2017 (UTC)
@Billinghurst:, like @Infovarius: said, we usually add (when we can) start time (P580)/end time (P582) as qualifiers for married names. I also use has cause (P828): marriage (Q8445) as a qualifier. --Harmonia Amanda (talk) 08:17, 17 January 2017 (UTC)

How to model physical, literary and other remains?

We have place of burial (P119) to indicate where people have been buried, but sometimes, their physical remains may not all have been buried together. Should we model that on the person's item or should we start separate items for these remains (think Albert Einstein's brain (Q2464312))?

Plus, some people have literary and other non-physical remains that find their way into various collections, e.g. at GLAM institutions. Should we have a separate property for that on the person's item or again another item? Finally, in some cases, remains of "notable" non-human organisms (e.g. Ham (Q28270)) or even machinery (e.g. Atlantis (Q54381)) may need similar treatment.

Any pointers and suggestions welcome. --Daniel Mietchen (talk) 01:44, 16 January 2017 (UTC)

And also sometimes the physical remains moves to different location. To new burial place. Geagea (talk) 01:57, 16 January 2017 (UTC)
@Geagea: I think that could probably be modeled by way of multiple place of burial (P119) statements, each with start time (P580) and end time (P582) qualifiers. --Daniel Mietchen (talk) 02:12, 16 January 2017 (UTC)
For Jacques-Louis David (Q83155), I've used applies to part (P518) : heart (Q1072). Pyb (talk) 08:29, 16 January 2017 (UTC)
A new property would make sense. ChristianKl (talk) 09:02, 17 January 2017 (UTC)

Redirect in

François Pithou (Q1451161). Article in is a redirect to other person: Pierre Pithou (Q2093849). Is it ok? not sure what to do. 12:48, 16 January 2017 (UTC)

no, they are two different man. - yona b (talk) 13:41, 16 January 2017 (UTC)
I've restored the en.Wikipedia article and added some sources there. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:07, 16 January 2017 (UTC)
here is a better image c:File:François Pithou par Gérard Edelinck.jpg curious no author page at french wikisource although there are a lot of google book hits [16] Slowking4 (talk) 02:45, 17 January 2017 (UTC)
This is no Wikidata issue but a en.Wikipedia issue. Whether or not François Pithou (Q1451161) is notable enough to warrant his own article is the decision of en.Wikipedia. If he isn't notable enough than it makes sense to redirect to the page of his brother where he's also mentioned. ChristianKl (talk) 09:00, 17 January 2017 (UTC)
I have added a father for the four brothers. It has them well connected :) Thanks, GerardM (talk) 10:02, 17 January 2017 (UTC)

Possible vandalism in undesirable edits to en labels in non-en language

Hi I have reverted a few edits that I thought were vandalism, but I am not sure what is going on, so I am just going to report here a pattern I noticed just today - a user is changing English labels into Spanish (or variant thereof). This might even be a good faith edit if the person doesn't know they are changing the English label but are trying to add a label in their own language. Unfortunately it's an ip address. See this diff. Is there any way to warn about changing labels in general? I am thinking about something like "You are changing an English label, do you want to proceed?" or something like that. Jane023 (talk) 10:50, 17 January 2017 (UTC)

That is not a good faith edit. I see no need in the notice that you mention as there is already a clear indication that the label is written in a given language.--Micru (talk) 13:20, 17 January 2017 (UTC)
OK. I will only pick a few of these up probably, as I have no way of checking the work of IP adresses. This showed up on my enwiki watchlist for a list of people by occupation. The other one was for a depicts field in a painting here: diff. Jane023 (talk) 14:55, 17 January 2017 (UTC)

Badge for templates that can work with Wikidata only?

To make it easier to find such templates, how about creating a badge that could be added to templates that can work with Wikidata only?
--- Jura 09:46, 15 January 2017 (UTC)

What do you mean by "Wikidata only"? It has become a practise on svwiki that all Wikidata-supported templates have to have an option to override the content that can be found on Wikidata. -- Innocent bystander (talk) 15:30, 15 January 2017 (UTC)
I mean templates that can be used without the need to fill-in data locally. Obviously, they still could do that, but don't have to.
--- Jura 15:36, 15 January 2017 (UTC)
Just to give an example: Q27999440. Matěj Suchánek (talk) 21:05, 15 January 2017 (UTC)
Yes, the badge could be on the following sitelinks: cs:Šablona:Části české obce, cs:Šablona:Infobox - osoba, sv:Mall:Faktamall film WD.
--- Jura 21:46, 15 January 2017 (UTC)
Good idea, I support this. Wesalius (talk) 07:32, 16 January 2017 (UTC)
Templates in hyWP, that can be used without the need to fill, but usually include a local non-free image or filled line(s), if it poorly or not supported or if poorly filled in WD: Կաղապար:Տեղեկաքարտ Գիտնական, Կաղապար:Տեղեկաքարտ Թեյ, Կաղապար:Տեղեկաքարտ Ծրագիր, Կաղապար:Տեղեկաքարտ Ծրագրավորման լեզու, Կաղապար:Տեղեկաքարտ Ծրագրավորման լեզու, Կաղապար:Տեղեկաքարտ Դպրոց, Կաղապար:Տեղեկաքարտ Դրոշ and many others. - Kareyac (talk) 19:28, 17 January 2017 (UTC)
  •   Comment As Innocent bystander intimated, the name "Wikidata only" badge could be confusing<->misleading<->misinsterpreted, if we are talking about templates that do require not addition of parameters. So if you are asking about templates that are Wikidata-ready, Wikidata-configured, Wikidata-driven, Wikidata-full for "all" parameters of the template, like Template:Authority control which can be populated completely from WD, then YES, that is fine. If we are talking about partial population, ie. some WD, some manual, then not with that criteria; though we may wish to consider "Wikidata-partial" after we have evaluated the initial proposal.  — billinghurst sDrewth 00:13, 18 January 2017 (UTC)

History of first settlers and of first hotel built in the town of Nyngan

I believe that the Nyngan Hotel opposite the railway crossing from the east was built as the first hotel in Nyngan. I also believe that the person who built the hotel was by the name of Laurence or Lawrence Galvin from Ireland. This hotel contained 12 bedrooms and was much later about 1975 ??? sold to a family by the name of BIRT and was deemed to be unsuitable to be used as a a hotel due to firre regulations. Could someone advise me if the above information is correct. Also if a George McClealland was on the shire council in the 1950's also a family by the name of Antill lived within Nyngan. Any help would be greatful.

Roger Edwards  – The preceding unsigned comment was added by (talk • contribs) at 02:39, 17 January 2017‎ (UTC).

Though this seems to be about the place in Australia which we represent at Nyngan (Q917556), your query is not related to Wikidata specifically. You may do better to ask at en:Wikipedia:Reference desk. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:39, 17 January 2017 (UTC)

Text to wikiData conversion

Given a text, is the conversion to wikiData format (i.e text to graph) conversion done manually or some software/algorithm/rules used to do it automatically ?  – The preceding unsigned comment was added by 2606:a000:4dcd:af00:6ca8:a4a6:88b:b3fa (talk • contribs) at 18:50, 17 January 2017‎ (UTC).

I've left some links to tutorials and help pages, on your talk page. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:34, 17 January 2017 (UTC)
We don’t have any magic algorithms that can extract data from an arbitrary, fully unstructured text. However, since many available inputs contain some degree of structure, we do indeed use quite a lot of automation to import data to Wikidata. Mostly from Wikipedias, whose unstructured sources contain structured and therefore machine-readable elements such as categories or templates. In general it is technically also possible to (semi-)automatically import from external databases with a public interface or highly structured HTML pages (via scraping). —MisterSynergy (talk) 20:00, 17 January 2017 (UTC)

Traffic Signs Regulations and General Directions (Q7832471)

This should be more specfic, but there wasn't a property for document version, or current document date.

There are multiple versions of this so, I'm not sure how to proceed

Also I got a note as to "repealed by" and "ammended by" which I proposed a long time ago, but no-one seems to be using.

Does Wikidata have the means to do Outreach through GLAM? I ask because my main reason for wanting the parameters is so that WikiData could import a portion of the OGL based database holds (it's more than just the legislation text) , which would make it nearly possible to auto generate a legislative history infobox purely from WikiData held information :) ShakespeareFan00 (talk) 21:09, 17 January 2017 (UTC)


Demanding Add new Languages on Wikipedia: Amharic, Alconquinn, Dakota, Blackfoot, Syrian, Ainu, Okinawa, Ural Altai, Aboriginal, Tasmanian, Shelta, Ulster Scot, Cumbric, Old Irish, Quechan, Assyrian, Newfoundland Scottish, Zuni, and the etc. What do you think?  – The preceding unsigned comment was added by (talk • contribs) at 05:50, 18 January 2017‎ (UTC).

@ What do you mean by "Add new Languages on Wikipedia"? Some of these languages are already available in Wikipedia. Pamputt (talk) 06:22, 18 January 2017 (UTC)
There is a process for adding languages and that is taken with crosswiki approach. m:Language proposal policy  — billinghurst sDrewth 06:34, 18 January 2017 (UTC)

Petscan: collecting redirects only?

I am looking to identify any English Wikisource author page (select a category: +2 down & author: ns) which is redirect = yes (tab 2) and has an item (tab 5), so my query is petscan:677381. I thought that the redirect tick would limit but it does not seem to do so. Can anyone suggest an alternate? Thanks.  — billinghurst sDrewth 02:41, 18 January 2017 (UTC)

@Billinghurst: You might try WD:Request a query. --Izno (talk) 13:27, 18 January 2017 (UTC)
The option works for smaller queries, maybe the system cannot handle so many authors. I would try to subtract some pages through "Negative categories", "Has none of these templates" etc.
A workaround could be an SQL query which doesn't handle well category trees, though. Matěj Suchánek (talk) 14:17, 18 January 2017 (UTC)
Thanks to you both. I will just work through the next layer-down (or two) categories. It is not me misinterpreting the instructions and the switches.  — billinghurst sDrewth 21:49, 18 January 2017 (UTC)

Bug in merge?

Merge added word "redirected" to en label and description. Don't see anything suspicious in MediaWiki:Gadget-Merge.js (which hasn't been edited for some months now), so there is problem somewhere else? --Edgars2007 (talk) 13:40, 18 January 2017 (UTC)

No. Someone add it manually. It is not from merging items. - yona b (talk) 14:07, 18 January 2017 (UTC)
(edit conflict) Do Special:Diff/429962627 and Special:Diff/429962639 explain? Matěj Suchánek (talk) 14:09, 18 January 2017 (UTC)
I feel so stupid now. Thanks. --Edgars2007 (talk) 14:30, 18 January 2017 (UTC)

Wikidata weekly summary #243

I "solved" majority of the 2500 people without gender, but with photo task. There are now 232 are left (mostly multiple people on the photo/sculptures/drawings/my missclicks). I went forward and assumed their gender from the photo, but there might be some errors since it is 2017 and gender questions are rather complicated :-) Cheers Wesalius (talk) 07:31, 19 January 2017 (UTC)

Instance of person definition

Hi, I'm going through the persons category on the Wikidata Game, and I am popping across some fictional and mythological characters. Do these entities count as people in the eyes of Wikidata? Thanks, Icebob99 (talk) 03:47, 19 January 2017 (UTC)

They have their own classes, e.g. fictional character (Q95074) and mythical character (Q4271324). Matěj Suchánek (talk) 10:42, 19 January 2017 (UTC)

Property proposals needing comments!

As can be seen at Wikidata:Property proposal/Overview there are a vast number of newly proposed properties (I count about 60 in the last day or two). Many of these are from GZWDer and seem to be pulled from a wide variety of wikipedia templates where we still seem to be missing corresponding properties here. Please take a look and add supporting comments where you feel appropriate, or otherwise. Thanks! ArthurPSmith (talk) 19:49, 17 January 2017 (UTC)

I am not happy with GZWDer flooding wd:pp. If he is semi-retired as he states on his userpage, what is the point of proposing so many properties that he is not going to use? Besides, many of his recently created properties have not been notified to the participants in the discussion, and they are lacking a connection through "instance of" to the property tree. @GZWDer: Care to comment? --Micru (talk) 20:17, 17 January 2017 (UTC)
After sabbatical this edit GZWDer starts directly with property creations... --Succu (talk) 21:47, 17 January 2017 (UTC)
There is no policy requiring people to maintain a certain level of editing before contributing property proposals. If that's all an editor does, they're still helping to develop the project, and the proposals should be judged on their merits, regardless of who made them. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:12, 17 January 2017 (UTC)
There is no policy is one of your beloved phrases. - So what? Did his/her property creations reflect our comunity sense of doing things right= --Succu (talk) 22:25, 17 January 2017 (UTC)
Very sorry as I have not add instance of (P31) to newly created properties. For "not going to use": I'm going to populate existing properties (currently doing competition class (P2094)) once I have completed all my proposal. I think it's better to have properties in advance as properties can not be created on demand immediately without discussion.--GZWDer (talk) 03:47, 18 January 2017 (UTC)
@GZWDer: When creating properties you should also add the property to the example items and add relevant constraints, see the steps on Wikidata:Property creators. - Nikki (talk) 08:36, 18 January 2017 (UTC)
Done.--GZWDer (talk) 09:06, 18 January 2017 (UTC)
@GZWDer: Can you please add "instance of" to the properties that you have created? But please don't do like ID (P3495), where you just added "instance of:Wikidata property", use more specific items, for instance for that case it would be Wikidata property for identification in the film industry (Q22964274).--Micru (talk) 10:38, 18 January 2017 (UTC)

I doubt it is healthy when the project is flushed with so many new proposals. During the last months and years there have been many complaints that it takes too long till a property is created, there have been complaints that not enough users are commenting on property proposals and there were complaints that there are not enough active property creators. Both the number of active users [17] and property creators [18] did not strongly changed during the last year so I predict that it will now take even longer until a property is created. Moreover, I doubt it is healthy if users are proposing properties when they are not interested in the topic. If a proposal gets accepted, the property also needs to be maintained. Somebody must check permanently that the property is not misused, otherwise it is worthless. Property proposer are not obliged to do it but who else should do it? Note that there are not more users around just because there is more work to do. --Pasleim (talk) 23:07, 18 January 2017 (UTC)

Given the magnitude of the flooding, and the low quality of the proposals, I propose that if any of those property proposals gets at least one support vote in the next two weeks we keep them, otherwise we mark them as "not done".
Regarding the unused properties, I think we should have a RFC about what to do with them. Maybe we can come up with a process less cumbersome than WD:PFD.--Micru (talk) 19:47, 19 January 2017 (UTC)
I don't think unused properties have any damage to Wikidata. Having some properties just indicates information Wikidata may provide (though far from complete). They are just like stubs and anyone can populate it. Deleting them does not make Wikidata better.--GZWDer (talk) 06:07, 20 January 2017 (UTC)

Identifier property for Worldcat Identities

I am trying to determine how to input a Worldcat Identity property for an organization. For people, Wikidata & Worldcat use the LCCN number, like n80126289 for Isaac Asimov (Q34981). Worldcat then derives a value of lccn-n80-126289 for their identifier for Isaac Asmiov. For organizations, Worldcat uses a different format. For the Python Software Foundation (Q83818) for example, Worldcat uses nc-python software foundation. What Identifiers property can I use for this? Peaceray (talk) 21:45, 19 January 2017 (UTC)

Merge not quite completed

I am very inexperienced hereabouts. To the best of my ability, I followed the instructions in Help:Merge to merge unneeded Q22262517 with much more informative Q12035751. All worked well, AFAIK, until the stage of turning the former into a redirect. Special:RedirectEntity objects, saying that it isn't empty. Could somebody please tidy up after me and tell me what silly mistake I made? Thanks! -- Hoary (talk) 01:29, 20 January 2017 (UTC)

I used the Help:Merge#Gadget, and the merge worked. I'm not sure what frustrated your merge. --Tagishsimon (talk) 04:17, 20 January 2017 (UTC)
Many thanks, Tagishsimon. -- Hoary (talk) 04:49, 20 January 2017 (UTC)

Multiple versions of the same publication


This page is explaining the history of the periodical currently known as International Review of the Red Cross. In Wikidata we have two items Q6052761 and Q27722028 (I merged Q24238387). In my opinion, the remaining two items should still be merged given that they both relate to the 'mother title' and the issues numbering is uninterrupted. The new item would then have multiple ISSN (P236) and title (P1476) with qualifiers start time (P580) and end time (P582). Beside of that, it would be necessary to create another item for the review published between 1869 and 1918 (under two different titles) and 2 items for the exclusively English (1961-1997) and Spanish (1976-1998) versions. Is it the right approach ? Louperivois (talk) 21:32, 17 January 2017 (UTC)

I agree with that. The content it seems would be overlapping, provided the differences are minor. They are sufficient. Thats my 2¢. MechQuester (talk) 23:24, 17 January 2017 (UTC)
The proposal is somewhat problematic. The translations are their own separate works, and would be "translations of". While the original authors may be the same, the attributes of the translations are different. The editors may be different, etc. So while there may be a common parent for all the works in a conceptual sense, there are numbers of cascading subsidiary items as we drop down to editions.

I still don't think that we have a good _shared_ understanding of "books" <-> "literary works" <-> "editions" <-> "translation".  — billinghurst sDrewth 02:51, 18 January 2017 (UTC)

Addendum. If the same plant species exist in various items for various names, then books in different forms probably do too if they have different provenance, or other clear differences. That the Cambridge needs to split them into four should give us a good indicator.  — billinghurst sDrewth
Let me precise that in my idea, Wikidata would ultimately have 4 different items corresponding to the 4 core versions that have their own issues numbering (IRC, BRC, RCE and RCS). But there would not be ~9 items corresponding to each different ISSN. Louperivois (talk) 03:54, 18 January 2017 (UTC)
We need a different item for each difference that matters to a data consumer who wishes to check the correctness of a claim in Wikidata that is supported by a reference to the journal. So everything that matters in looking up and understanding the reference must be the same: the title (so the data consumer knows he/she has found the right publication; maybe it's enough if an alternate title is given as an alias), the language (so the data consumer can understand it), the volume and issue numbering, and the page numbering (so the correct page can be found to verify the Wikidata claim). Jc3s5h (talk) 14:12, 20 January 2017 (UTC)

How can a merge be undone

The item for shah was merged. It is now specific for Iran. There were shahs of other countries so it is wrong. Thanks, GerardM (talk) 23:51, 20 January 2017 (UTC)

Rollback (or undo) both items.  — billinghurst sDrewth 00:32, 21 January 2017 (UTC)

Watchlist rendering

Malformed rendering

Is anyone else experiencing weird rendering on their watchlist? I filed a Phabricator bug but no one seems to have noticed. Jc86035 (talk) 12:24, 20 January 2017 (UTC)

Visually it’s all fine here (up-to-date Firefox). However there is a ul-tag nested directy into another ul-tag, which isn’t correct as far as I know (I’m not an expert, though). —MisterSynergy (talk) 12:31, 20 January 2017 (UTC)
The top components are transcluded files Mediawiki:watchlist-summary, Mediawiki:watchlist-details, and Mediawiki:wlheader-showupdated and if they are playing up is more likely a local issue than particular to the watchlist functionality. That being the case, you need to ping an administrator. To look at a watchlist without transcluded parts showing look at it without the active local configuration elements ...  — billinghurst sDrewth 00:42, 21 January 2017 (UTC)
@Billinghurst: Apparently it's because I set my language to British English and the HTML wasn't updated properly. I've fixed the messages. Jc86035 (talk) 04:10, 21 January 2017 (UTC)
Grrr, the dreaded and problematic -gb and -ca on default English language wikis. Such a wonderful innovation.  — billinghurst sDrewth 04:32, 21 January 2017 (UTC)
No no, this is just another innovation. Matěj Suchánek (talk) 09:02, 21 January 2017 (UTC)

Kartographer extension on Russian Wikipedia

From a post on the OSM-talk mailing list, FYI:

Russian Wikipedia just replaced all of their map links in the upper right corner (geohack) with the <maplink> Kartographer extension! Moreover, when clicking the link, it also shows the location outline, if that object exists in OpenStreetMap with a corresponding Wikidata ID (ways and relations only, no nodes). My deepest respect to my former Interactive Team colleagues and volunteers who have made it possible! (This was community wishlist #21) Example - city of Salzburg (click coordinates in the upper right corner, or in the infobox on the side): ru:Зальцбург

If you create a Wikidata item about an administrative area, building or other physical object which appears on maps, then please add the Wikidata ID to the relevant object in OpenStreetMap, using key:wikidata= (here's how to contribute). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 07:45, 21 January 2017 (UTC)

Never married persons

Today there was an edit war on Franz Kafka (Q905) who is known for not having got married. Users Rodejong and Villy Fink Isaksen repeatedly changed spouse (P26): no value to unmarried (Q28341938). The page is now protected to stop warring and to make clear whether the use of no value is correct for people who never got married. Note that unmarried (Q28341938) is propably duplicate of Q22101595 which does have several uses. Matěj Suchánek (talk) 20:43, 15 January 2017 (UTC)

I can add that the removing of the claim spouse (P26): no value probably was because of an infobox template at Danish Wikipedia which cannot handle the value no value. But nevertheless this seems like a case of exactly what that value was made for. Best regards, Dipsacus fullonum (talk) 20:55, 15 January 2017 (UTC)
The matter of the infobox is irrelevant for this discussion though Dipsacus fullonum, and the issue has been solved. However "no value" is needlesly ambiguous, using the "unmarried"/"never married" is a more logical choice. If a property to an item is added it stands to reason it must be filled, otherwise it has no purpose other than to create confusion. -- Vrenak (talk) 21:02, 15 January 2017 (UTC)
Unless you can tell us the date Kafka married an entity called "unmarried", and source that, then "no value" is correct. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:17, 15 January 2017 (UTC)
@Vrenak: Yes, a claim must have a value, but no value and also unknown value are perfectly fine values. unmarried (Q28341938) or Q22101595 aren't good values because you would think that if the value is an item, it will represent an actual spouse. Best regards, Dipsacus fullonum (talk) 21:25, 15 January 2017 (UTC)
Kafka has to marry another person for the property to make sense. Just look at the constraints of spouse (P26) and see if they make sense for Q22101595 :
Allows a start time (P580), end time (P582), place of marriage (P2842). How can we fill those?
Koxinga (talk) 21:31, 15 January 2017 (UTC)
Bad constraints are never a reason to fudge data. Fix the constraints. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:46, 15 January 2017 (UTC)
Constraints also illustrate what kind of data is expected. Anyway, you agree with me above, so I am not sure what is your logic here. Koxinga (talk) 22:28, 15 January 2017 (UTC)
In this case, personally I'd favor "novalue" as well. I added a few when working on Wikidata:Database_reports/Constraint_violations/P26.
Maybe we should find a way to store the item that contains the textual description of novalue.
--- Jura 21:42, 15 January 2017 (UTC)
Dipsacus fullonum, "no value" is not a value, your logic is flawed, "no value" literally means there is no value, not even a zero. So if you have a claim you need to fill it out with a value, if there is no regular value to fill in you have a zero value for the claim, for spouses this can be unmarried or widowed, depending on context, similarly you can't have a location with "no value" inhabitants, it can have "0" but it can't have "no value". logically if you want to have a claim it must be filled, or you remove the claim itself. -- Vrenak (talk) 22:13, 15 January 2017 (UTC)
If you count the number of spouses per person, using novalue gets you a count of 0 for items using that. That's what we want for people who never married. They never married, they didn't have a spouse nor did they have an entity called "unmarried" as spouse. Using "novalue" works fine for this property. Now we just have to find a way to fix the infobox. The problem seems to be at da:Skabelon:Infoboks person.
--- Jura 22:22, 15 January 2017 (UTC)
Your logic would be correct if the property was the number of spouses. However, the property is the spouse item. If there is no spouse and we know it, a novalue is a way of saying that this property can not be filled: it is kept empty not because we don't know but because we know there is nothing to enter. Koxinga (talk) 22:25, 15 January 2017 (UTC)
In view of the description of "no value" at mediawikiwiki:Wikibase/DataModel#PropertyNoValueSnak "no value" seems appropriate in the case of a dead person who we know was never married, when taken together with our practice of listing all known spouses, whether the subject of the item was married at the time of death or not. In a more general context, outside of Wikidata, "unmarried" just means not married at this moment, and "never married" means not married up to this point in time. Whenever we document these terms we should be mindful of how these terms are used within Wikidata versus how they are used in a more general setting, and how the shades of meaning are different for living vs. dead people. Jc3s5h (talk) 23:05, 15 January 2017 (UTC)
The other means to manage this is to do a count; ie. a property "number of marriages", where the count can be 0 .. n. This does actually have an advantage, as 1) there are numbers of people who we know married, though never know the spouse; 2) there is a lot of reluctance or inactivity to create a spouse of another otherwise non-notable person. From Dictionary of National Biography there are many mentions of spouses, and I know that I pass adding that detail, sometimes for lack of helpful detail, other through the effort required.  — billinghurst sDrewth 03:54, 16 January 2017 (UTC)
I sometimes use "no value" to prevent people from adding "Coat Of Arms" and "Sister city" without thought. -- Innocent bystander (talk) 16:52, 16 January 2017 (UTC)
But please add only "no value" to coat of arms if you know that the entity does not have coat of arms. Just a missing file on Commons does not eligible the use of "no value" --Pasleim (talk) 07:17, 17 January 2017 (UTC)
@Pasleim: I see no point in adding COA:"no value" to every single thing that does not have a COA. But when people repeatedly add a COA to wrong item, for example into Malmö (Q2211) instead of Malmö Municipality (Q503361), then it could be useful to stop these "games" and other "automatic imports" that causes us problems. -- Innocent bystander (talk) 08:01, 19 January 2017 (UTC)
No one is married to unmarried (Q28341938) or Q22101595. Neither are appropriate as a value, and both should be deleted. The entire point of novalue is for cases like this. --Yair rand (talk) 22:42, 16 January 2017 (UTC)

If I am allowed to add to the confusion: A few times, I have run into the case where a work by an anonymous has been claimed to be by anonymous (Q4233718). It seems to me, that like no value is better for unmarried, then unknown value is better for these cases? — Finn Årup Nielsen (fnielsen) (talk) 23:58, 16 January 2017 (UTC)

Yes, unknown value would be appropriate for this case. Otherwise we have problems because two books who's authors are both "anonymous" would be treated as being by the same author. ChristianKl (talk) 09:08, 17 January 2017 (UTC)

If anyone like to hear what I say, I think it is better to write "unmarried" rather than "no value". Unmarried is best used for people who's marital status is definitely known as unmmarired. "No value" is open ended and suitable for people who's marital status is unknown. In this case, its best to have unmarried because we know he's unmarried. MechQuester (talk) 05:57, 17 January 2017 (UTC)

"no value" is not open ended. If the marital status is unknown, one has to use "unknown value". "No value" should only be used if one definitely knows that the person is unmarried. In case of Franz Kafka (Q905) this is given.
The property isn't "marital status" but "spouse". Unmarried might be a reasonable value for a marital status property but it's not a spouse. ChristianKl (talk) 08:46, 17 January 2017 (UTC)
and unknown value could be used for not knowing "if" married, or "to whom" married. Though the guidance on whether to populate that way is interesting as in other places the absence of data is used for "unknown". On similar note, do we differentiate between a marriage and just co-habitation?
We have unmarried partner (P451) for unmarried partners. It's currently not possible to say that a person is definitely married but the partner is unknown. At least not directly, you could still use "unknown value" + the start qualifier and link a source as a reference. ChristianKl (talk) 16:38, 22 January 2017 (UTC)

The intention of "no value" in the data model is indeed to express that for this property, there is no value that would be a valid filler. For the property spouse, using "no value" means that the person has no spouse, i.e. is not married. Using "unmarried" is bad, because then, when you ask for people that are married to the same person, you get that all these people are that are married to "unmarried" are married to the same person. "No value" is meant exactly for this use case. "Unknown value" is meant for the anonymous use case. Otherwise, again, it is not the same entity that has written all books by anonymous. For some things, having a "no value" makes not much sense. For example, a town cannot have a population of "no value", since 0 is a valid value. Also, a (non-fictional) human cannot have no mother, for example (as of the current state of technology). For a spouse, though, "no value" makes absolutely sense and is exactly what it is intended for. Maybe the display of "no value" is causing confusion, and if there is a better way to phrase it, that would certainly be welcome, but that's how the data model was defined and how it is implemented. --Denny (talk) 22:09, 20 January 2017 (UTC)

Problems with the administrative divisions in Indonesia

Hi everyone, but specially to Beeyan and Fexpr, whose I think could help me so much.

I set the Spanish label and description for 59950 items with the instance of (P31) and fourth-level administrative division in Indonesia (Q2225692). The description that I used in Spanish was "pueblo de Indonesia" but then, when my bot made all the changes, I think deeply and check that fourth-level administrative division in Indonesia (Q2225692) in Spanish is "aldea de Indonesia", so I begin to change the description from "pueblo de Indonesia" to "aldea de Indonesia" with QuickStatements. But now, checking more items I have discover that many items have fourth-level administrative division in Indonesia (Q2225692) and desa (Q26211545) and I have new doubts: are this two items the same administrative division or different?

I understand that in Indonesia exists fourth administrative divisions: the smallest, a desa which in English is a village and in Spanish is pueblo and then, inside a desa, Indonesia has another one, a kampung, which in English is a hamlet and in Spanish is aldea. Checking the two items that I said in the previous paragraph, I have two possible theories: 1. fourth-level administrative division in Indonesia (Q2225692) and desa (Q26211545) are the same item; or 2. the first is for desa and the second for kampung.

In the other hand I think it is a bit confused because there are many items with the two items in the instance of, but I imagine that it could be only one of them, and then, one a subclass of the other, or some structure in that way. What do you think about it?

Excuse me for all the wrong edits and my mistakes. My intention in my first task with CanaryBot was help to have the Indonesian data in Spanish too, but I didn't it in the right way. Depending of the correct answer to my questions, could be necessary to add pueblo de to desa in Indonesia and aldea de to kampung. If anyone think that I have to revert my 1000 approximate edits with QuickStatements, please told me and I revert it to the last descriptions before my edits.

I await your answer.

Regards, Ivanhercaz   (Talk) 01:27, 19 January 2017 (UTC)


Raisha Rachmat04

  Notified participants of WikiProject Administrative Units in Indonesia. Regards, Ivanhercaz   (Talk) 01:14, 20 January 2017 (UTC)

Hello @Ivanhercaz:, fourth-level administrative division in Indonesia (Q2225692) is the lowest formal government administration in Indonesia; the lowest informal administration is Rukun Tetangga (Q12509020). The naming convention for fourth-level administrative division in Indonesia (Q2225692) is different, based on their provinces or area. So, at least there're 10 different terms, kelurahan (Q965568) or desa (Q26211545) or gampong (Q4285979) or nagari (Q882149) or dusun (Q23308490) or kampung (Q12488911) (in Lampung) or kampung of Papua (Q12488913) (in Papua) or kampung of East Kalimantan Province (Q24659756) (in Kalimantan) or pekon (Q19944049) or lembang (Q12494403). Beeyan (talk) 06:34, 20 January 2017 (UTC)
fyi, dusun in Central Java (Q3557) will be the fifth informal administration, but dusun (Q23308490) in Bungo (Q7373) is fourth-level administrative division in Indonesia (Q2225692). There reason why there're so many usage of words for different concept because Indonesia has a lot of regional languages. Beeyan (talk) 06:44, 20 January 2017 (UTC)
@Beeyan: Hi! Thank you for your answer. It seems harder than I thought... Well, I think that I am going to revert my latest editions made with QuickStatements and I am going to set, again, "pueblo de Indonesia" in the Spanish description. Once more time, thank you for your help. I am going to fix it in Spanish. If some time you need a bot to set labels and descriptions, I can try to help you.
Regards, Ivanhercaz   (Talk) 11:23, 20 January 2017 (UTC)
@Ivanhercaz: denada, encantado di conoscerti! Yup, even if you're asking randomly to any Indonesians, I'm sure that they don't even know another name is exist if they're not coming from this spesific area. It's very complex even for Indonesian people too. The best thing is by adding the definition of fourth level administration in Indonesia. --Beeyan (talk) 04:31, 23 January 2017 (UTC)

Daily (or every two days?) reference drives

Hi all, after reading w:Wikipedia:Wikidata/2017 State of affairs, it is quite clear that Wikidata is not well-received on enwiki. Aside from the many claims that seem to stem from a misunderstanding of Wikidata's purpose and integration with Wikipedias, I think one criticism is valid. Wikidata has many claims that are not sourced, and these include claims that are imported from Wikimedia project (P143) various Wikipedias. This opens us to criticisms for violating w:WP:V and w:WP:BLP. I suggest that we have a daily (or every two days) reference drives where we focus on improving a single item by adding the references from various Wikipedias as well as finding new ones. The end product doesn't need to be showcase material, but at least it will be well-referenced and Wikipedia usable in that regard. Any thoughts? —Wylve (talk) 12:18, 20 January 2017 (UTC)

The entire relationship between Wikidata and Wikipedias has to be put into greater perspective. Improving the references situation at Wikidata is an important and noble goal, but I don’t think that it significantly improves how Wikipedia communities think about our project here. If one wants to point to situations that went wrong with Wikidata, one can do that in any situation — no matter how good Wikidata actually is. My impression after reading the enwiki discussion is that there are quite a bunch of users over there that do exactly that. As an experienced dewiki community member I remember plenty similar discussions from that project as well.
It is always important to be aware of what we are doing here. Wikidata is Wikimedia’s approach to drastically streamline the way it manages its valuable knowledge. Like practically any other large organization in this world whose core asset is unstructured data (such as wikitext), Wikimedia and its communities spend effort to separate structured from unstructured data these days. The actual implementation of Wikimedias structured data is Wikibase, which is an extremely powerful solution to my experience. But Wikidata was just the first step to structure Wikipedia knowledge (the most important asset), Wikimedia Commons and Wiktionary will follow soon and other fields such as WikiCite are likely about to be developed in future as well. Structured data will be the backbone of Wikimedia in future, even if most Wikipedia communities do not understand this fact right now.
To my opinion we don’t have to worry about their current reluctance to include Wikidata into their articles. Wikipedias grew a low in the past years, possibly more than anybody imagined when this journey started some 15 years ago. However, there is a vast amount of information to maintain in good shape, and if you browse to the corners of Wikipedia’s knowledge (no matter of which language), you can easily spot the limits of the communities’ ability to properly do this job. Even today in 2017 there is so much content which would profit from any kind of automation (I’d estimate that this applies to much more than 50% of all articles). Yet it is clear that the more you walk to the central content, the less automation is necessary or useful. These central articles have typically been written by experienced, influential editors.
There are a couple of things to take care of at the same time:
  • The separation of structured data from unstructured wikitext significantly adds complexity, which is a problem for many users that are not so tech-savvy. We have to make sure that there is really good software which envelopes the internals of this data model for users do not want to see it — without suspending them from editing Wikimedia’s projects. The VisualEditor is important for that, but by far not enough to deal with this issue.
  • Structured data is efficiently maintained, but in encyclopedias it needs to be put into context (unstructured wikitext can do that much better than structured data). This is something only human editors can do, and this is something which drastically limits the degree of automation we should strive for. Since Wikipedias as well as Wikidata itself are quite diverse in depth and quality (both in form and content), it is not at all wise to take a binary approach such as Automation everywhere or Forbid all Wikidata usage/automation.
  • In many situations, “data usage from Wikidata” can simply mean to compare a local value with a Wikidata value and add the Wikipedia page to a maintenance category if there is a difference. Which of the two values is actually displayed then is of minor importance. However, there is (to my knowledge) no coordinated effort to provide powerful templates and modules in Wikipedias right now which enable users to decide on different levels (per-article, per-template, per-Wikidata property, per-Wikidata datatype, per-Wikipedia project, etc. …) which data to display by default.
  • Please always remember as well that there is much testing and trial-and-error involved in this procedure of structuring Wikimedia’s knowledge. But we don’t have another option than to use structured data in the future, given the fact that Wikipedia’s rate of deterioration of quality is ever growing.
Ping @Fram as the initiator of this enwiki discussion. —MisterSynergy (talk) 13:25, 20 January 2017 (UTC)
Different Wikidata editors have different reasons for editing. Declaring reference drives as a specific project ignores the motivations for which an editor comes. Wikidata should focus on being inviting to different people who want to come and participate no matter what kind of data they want to contribute.
In general unreferenced data shouldn't matter to en.Wikipedia. It's easy for en.Wiki to simply ignore all unreferenced data in Wikidata. Having bad references might be more problematic. ChristianKl (talk) 07:20, 22 January 2017 (UTC)
Agree, if Wikidata wasn't attempted as a tool for Wikipedia and Wikisource, my interest in this project would be zero (Q204). The inclusion of the data here to Wikipedia is a good benchmark to test if our models here works.
Besides the "imported from"-issues and the encouragement of editing without knowing anything about the subject, I see one large threat. Phase 1. The non-Wikidatatians too often tends to see Wikidata only as a place to fetch Interwiki. When I manually as an IP adds interwiki to a page in Wikipedia, I am notified that "We do not add interwiki by the wikicode anylonger, instead we use Wikidata". Poorly matching subjects are therefor merged. Articles about families, names, disambigs and lists of things tends to be very mixed up. -- Innocent bystander (talk) 08:14, 22 January 2017 (UTC)
I think that integration with Wikipedia and Wikisource is very valuable but I don't think it's the only purpose of Wikidata. If everything goes well with WikiFactMine, scientists might go directly to Wikidata to look up facts (and their references) for an item that comes up in their research.
It's valuable for Wikidata if people who have an interest in open structured data participate even when they aren't directly interested in Wikipedia. Think big tent. Not everything on Wikidata has to be useful for Wikipedia, but Wikipedia should be able to decide to only important the data that's useful for it. When unsourced statements aren't useful Wikipedia can simply decide to only import sourced claims for a property. ChristianKl (talk) 18:59, 22 January 2017 (UTC)
in my experience the reference problem will not be improvedm until there is some leadership to lead a quality circle to add references. and not to propose a policy to block people who add un-referenced statements. the team need not conflict with the others in the big tent. until the leadership is provided, the problem will not be improved, despite all the hand-wringing elsewhere. Slowking4 (talk) 23:30, 22 January 2017 (UTC)

Adding primary sources as references

I will reflect again for the umpteenth time, we do not make it easy to add primary sources to the system. I can find a baptismal record for many people in history, and to add it here is a PITA. It is not just the addition of the reference data, it is the requirement to create the whole record to reference it. If I am adding (and creating) primary/secondary records for a birth, a death, a marriage every time I have to create a person, the likelihood is slim. On the occasions that it is on the web then often they are behind pay firewalls, with complex urls, not designed for ready referencing. All a problem! So, I will continue to add the primary research undertaken to the author talk page at the Wikisource interwiki link. We still require a better methodology to reference. Even then will the Wikipedias accept a primary source as a reference anyway?

Secondarily, we have many published secondary resources at the Wikisources for adding this referential data, add it is still painful to add a reference from there to here. I know that we have a project for the addition of data from other wikis, and one can hope that such a project will allow for not only editing, but the pushing of referential statements. For example, 63 volumes of Dictionary of National Biography to their constituent items.  — billinghurst sDrewth 04:49, 21 January 2017 (UTC)

My experience is that primary sources are not allowed to prove the "notability" of a subject on Wikipedia. But when the notability is proofed in some other way, primary sources can be used. But one problem is that reading baptismal records can be very difficult, and can be regarded as "Original Research". Laws can also be difficult to interpret. -- Innocent bystander (talk) 08:20, 22 January 2017 (UTC)
do not know why you are raising ENWP issues here. we need a wizard / process improvement. maybe a wishlist item? Slowking4 (talk) 23:24, 22 January 2017 (UTC)

Brute force creation makes other work

I am not sure why we continue with the brute force creation by some people here. Where we end up having to them merge duplicates, which is more work for others. Surely there is now a better means to identify matches, or potential matches, and create only those that need creating then separating out matches and potential matches. -- – The preceding unsigned comment was added by billinghurst (talk • contribs) at 00:31, 21 January 2017 (UTC).

I am afraid that it is matter of perspective. We are on the right track in that we find more and more reason to connect things at the start. For instance, I added a person because he was awarded a prize and he was incorrectly linked on the English Wikipedia. I added his VIAF indicator because he is extremely likely to be linked through other sources. Your point that we should create only those that need creating has a powerful problem; what needs creating? What to do when for whatever reason another person is already there because she is already there and has no article yet.
It would be great when we have better tooling to curate links on the Wikipedias because there is a percentage that is wrong. I have argued and will continue to do so that we would do better at sharing the sum of all knowledge when Wikipedia and Wikidata cooperate. That will only be possible when we start to think in terms of what do we have to offer each other. At this time the import of a lot of data has not finished and as we lack the tools, we do not curate Wikipedia but do improve Wikidata. Thanks, GerardM (talk) 06:45, 21 January 2017 (UTC)
@billinghurst: "Surely there is now a better means to identify matches, or potential matches" That's great to hear. Please post details. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 07:47, 21 January 2017 (UTC)
@GerardM: I am not talking about that sort of edit/creation. I am talking about those who run bots through petscan and other tools and create items in their hundreds <-> thousands with no effort or consideration for existing items that may be matches. Be they plant articles or disambiguation pages, running a fill from enwiki that creates duplicates in such cases is tantamount to non-considerate editing. I know it is my choice to do merges to clean up, however, there is nothing quite as annoying as attempting a diligent cleanup to find that a bot creation run has gone through and whack whack whack'd more duplicates into play. UNHELPFUL! I am not wishing to point fingers, I am hoping that the community can say that we have reached a point of maturity that something more elegant can be put into place.  — billinghurst sDrewth 08:15, 21 January 2017 (UTC)
@billinghurst: The nlwiki system seems to work quite well. A couple of users keep an eye on the new articles and try to connect them or create new items.
If articles are not connected, are at least 4 weeks old and haven't been edited for at least 3 weeks, a bot will create a new (empty) item. This is in place to limit the backlog (it's the broom wagon (Q14823)).
The new empty items show up at Wikidata:Database reports/without claims by site/nlwiki. People keep an eye on that, bottom article is May 2014.
The result is quite good. We did have some annoying encounters with well willing users who ran some automated jobs blindly creating empty items for everything messing up this whole process. We kindly asked them to never run on nlwiki again, they didn't speak the language anyway.
Maybe set up the same system for other languages like enwiki? I can enable the broom wagon for enwp too or you can run it yourself. Multichill (talk) 09:24, 21 January 2017 (UTC)
Multichill, I am sure that there are numbers of means that it can be done, and I am glad that nlWP is showing some leadership in having a system. The users doing this are certainly well-meaning, I just think that it is short-sighted that they do it, and that the tools are so (easily) configured that way rather than it being either an advanced function. Even if there is a ready means to run a duplicity type check so that where the label exists it pushes it to a duplicity check, or report that needs an override tick to push such bot runs through. There are more knowledgeable here about the tools/processes/solutions, I am just here venting about my (negative) experiences with the hope that there is a little consensus and the ability to move to a less duplicate generating plan.  — billinghurst sDrewth 10:37, 21 January 2017 (UTC)
  • Personally, I think it takes more time to merge items than to run Duplicity (Q23751912). Except if one has some plan to merge duplicates oneself, I don't think one should run PetScan with the option "select all".
    --- Jura 10:29, 21 January 2017 (UTC)
    • I don't think so. Almost all toolses and reports (petscan, harvesttemplate, projectmerge, without claims by site, Wikidata Games, deaths at Wikipedia, etc) does not work for unconnected pages. Therefore, it's better to have many stub items and improve them later.--GZWDer (talk) 11:41, 21 January 2017 (UTC)
      • The idea is to limit the number of items for the same topics. Of course we can use those tools on two items and add twice dates of birth, twice date of death, twice occupation, etc, but we don't need to do that on two separate items. If you use the filter PetScan provides to avoid duplicates, you should be fine for most wikis.
        --- Jura 11:46, 21 January 2017 (UTC)
      • Duplicity (Q23751912) is a much more convenient tool than projectmerge. So no, it's a bad idea to create items only for them to end up on projectmerge.
        --- Jura 11:57, 21 January 2017 (UTC)
        • No data will be lost when items got merged, but no data can be added if the page is unconnected. Making new items with data also helps finding merge candidates.
        • Duplicity is indeed a good tool, but we don't have enough people to check every new pages. PetScan already allows skipping pages whose terms already exists in Wikidata, which may be link candidates.

--GZWDer (talk) 13:59, 21 January 2017 (UTC)

Can you confirm that you (your flood account) uses that option?
--- Jura 14:02, 21 January 2017 (UTC)
+1 with billinghurst. I am curating one dataset after someone creates several thousands of new items 4 months ago: there were around 800 potential duplicates. Now around 700 are still here: it seems that I am the only one who does the curation job.
@Pigsonthewing: Not difficult: when you have a dataset, instead of importing the data by creating new items, keep the datasetin your computer, match the identifiers of your dataset with the one available in WD and at the end create only items for the values without correspondance with WD. And you can be creative: instead of using identifers, you can use combinaison of properties like for persons date of birth and date of death. Extract the date of birth and of death from WD and compare with the ones in your datasets. If you find a correspondance (same date for the birth and the death), it is worth to check before any item creation if the item is relevant for one entry in your dataset. Some kind of preprocessing job. Snipre (talk) 07:05, 22 January 2017 (UTC)
It's worse if a bot considers two items falsely to be a match then if he creates duplicate items when the items aren't duplicates. In cases like important people where different people have the same name I don't see the problem with creating more new items. If someone later has a problem with the fact that duplicate items exist they can merge. ChristianKl (talk) 07:43, 22 January 2017 (UTC)
"they can merge".. interesting attitude. I noticed you haven't done so in the last three month.
--- Jura 09:34, 22 January 2017 (UTC)
It's not true that I didn't use merge in the last three months. But even if that would be true it just suggests that the items with whom I have dealt have no need for merging. ChristianKl (talk) 18:33, 22 January 2017 (UTC)
@ChristianKl: I didn't say do not use bots, nor that bots were not helpful, nor that the occasional duplicates are a problem, and not one word about using bots to do matches and merges. So an off-hand dismissal of the issue is not helpful.

I asked that we don't do brute force creations, that means consider the prospect for duplicates prior to running a bot through, especiall as there are already bots that do matching of similar data. I still think that the ability to have a dummy run through to exclude potential matches to a separate list, then review those is doable.

The reason why duplicates are a problem is the wasted resources in trying to identify differences from a WD item to another WD item that both attach to other sources. Far easier to have that diligence applied at the time of creation.  — billinghurst sDrewth 03:28, 23 January 2017 (UTC)

Having just done a five-way merge (Q21542964/ Q21454103 / Q21555944 / Q21546110 / Q21288402), can I please encourage people to clean up as they go along. Unmerged items really mess up our cross-referencing between different databases, and they make life far more difficult for tools like the auto-matcher in Mix'n'match. Of course, don't do bad merges, because they are also a nightmare. But please, don't leave potential unmerged items around for more than one or two weeks maximum without checking. If you can't check for potential merges within that sort of timescale, please then throttle your uploads and don't upload so much data at once -- don't upload more than you can check. Jheald (talk) 18:00, 22 January 2017 (UTC)


There is a problem on the German and English Wikipedia pages. The English page w:Cavalier does not have an link entry for the German page w:de:Kavalier. The German page does have a link to the English page, but it is missing a lot that appear on the English page.

  1. Can someone please fix it.
  2. Can someone please leave me a link on my talk page to advise on how to fix such problems in the future (teach a man to fish).

-- PBS (talk) 09:18, 21 January 2017 (UTC)

@PBS: They are different items cavalier (Q2284765) and Chevalier (Q354421) with some overlap in languages. Someone is going to need to sort out whether they are specifically different, or there is a conflict to resolve.  — billinghurst sDrewth 10:40, 21 January 2017 (UTC)
Noting that Chevalier (Q354421) is a disambiguation page and it looks like an article linking to it, so that seems in error.  — billinghurst sDrewth 10:42, 21 January 2017 (UTC)
de:Cavalier (a dab page) and en:Cavalier (disambiguation) link to another page Cavalier (Q420534), so it appears that the pages that link to Chevalier (Q354421) should either be linked to cavalier (Q2284765) if they are an article about cavaliers, or Cavalier (Q420534) if they are a dab page or some other page if they are are about just about a horse rider (equestrianism (Q179226), or to cavalry (cavalry (Q47315)), but the link Chevalier (Q354421) should probably be removed as a duplicate. -- PBS (talk) 11:22, 21 January 2017 (UTC)
this is a special case (aren't they all) where the english has an historical connotation, versus generic term. i guess we will have to hand fix these. don't know if there is a mix and match tool to provide a list, or a flag. Slowking4 (talk) 23:21, 22 January 2017 (UTC)

What if duplicate items are created?

Hi! I imagine that some duplicates of items are created every now and then. A user may not know that the item exists and create a new one and goes on with life. Are such duplicates being automatically detected somehow? Or are there some tools out there to help finding them? I guess the Wikidata game had such capabilities at some point but are there other options? //Mippzon (talk) 16:54, 22 January 2017 (UTC)

  • Even we one takes reasonable care, it's likely to happen. Help:Merge provides an introduction.
    --- Jura 17:03, 22 January 2017 (UTC)
There are projects around looking at duplicates, have a poke at User:Pasleim/projectmerge. There are also tools like Wikidata-todo's Duplicity that allows you to work from a list to look for potential matches.  — billinghurst sDrewth 00:34, 23 January 2017 (UTC)

Property for members of parliamentary groups

There seems to be no consensus on which property is the right for stating the number of MPs of a parliamentary group. I argue to use member count (P2124), because the group exists of MPs and hence the MPs are the members of the group. @Caarl 95: argues to use number of representatives in an organization/legislature or won in elections (P1410), which is usually used for the number of MPs of a political parties (see Talk:Q20113710 for full argumentation - Please notice that the english description of P1410 has been changed yesterday to include groups, the German and French still only describe a use for political parties). My argument is, that a party has seats in a legislature (number of representatives in an organization/legislature or won in elections (P1410)). However, a parliamentary group does not have seats in the legislature, but is composed of members of the legislature. Therefore member count (P2124) is the correct property for the number of members of a parliamentary group. @Oravrattas: argument (here) is also not valid: a parliamentary group is defined by its parliament and therefore has only seats in one legislature. For example, Group of the Alliance of Liberals and Democrats for Europe (Q839097) and Alliance of Liberals and Democrats for Europe in the Parliamentary Assembly of the Council of Europe (Q4732455) are different entities, different groups with the same name. A party with the same name exists (Alliance of Liberals and Democrats for Europe Party (Q25079)), which has seats in different legislatures. My conclusion is: number of representatives in an organization/legislature or won in elections (P1410) is correct for political parties, member count (P2124) is correct for parliamentary groups. --ElTres (talk) 14:25, 20 January 2017 (UTC)

@ElTres: There seems to be a lot of confusion around elections, parliaments and things like that. The main reason is probably because it looks so differently in different nations and on different levels. I guess The EU-parliament is probably extra complicated, since the EU-groups are not very homogeneous. Two Swedish parties are representing EPP in EU. That did not mean that they cooperated in any way in the last EU-election. The EU groups/parties were not in any way visible here. When a Swedish MEP is interviewed on TV, they are mentioned as representing a Swedish group of MEP's not a EU-group. It is only when leaders of such a groups are mentioned, they are mentioned as representing a EU-group. Otherwise they are mentioned as representing a national party.
-- Innocent bystander (talk) 08:38, 21 January 2017 (UTC)
I would agree this is something that does vary a lot - certainly ElTres's summary is not how I would think of it in the UK, where (for example) the parliamentary group is treated as effectively the same as the party. So I'm not sure we can say that one approach is obviously "correct". Andrew Gray (talk) 11:07, 21 January 2017 (UTC)
In the Swedish riksdag the parliament group and the party could be more or less in conflict with each other. The group and the party often have different leaders, since you have to be an MP to be a part of the group and be its leader. If you are a member of the Government, you normally resign your seat in the parliament. The governing part(y|ies) therefor always have other group-leaders than the party itself. And a party can have two or more leaders, while a group always have only one. -- Innocent bystander (talk) 11:17, 21 January 2017 (UTC)
This split between leaders does happen in the UK Parliament as well, particularly with smaller parties. And the distinction between the Parliamentary Labour Party (Q14472169) and the Labour Party (Q9630) has been in the news a lot over the last year or so. However, I think that it's precisely because things work differently in different legislatures that it's useful to have a common property to say that there are N seats are held by members of X body, whether X is a party, faction, club, political group, or whatever. I see no value in using a different property for one of these cases, especially if that property is simply a broad member count (P2124), and number of representatives in an organization/legislature or won in elections (P1410) was created for this very purpose. --Oravrattas (talk) 20:24, 22 January 2017 (UTC)
@ElTres: Note that the original proposal for this property was explicitly for "number of seats hold by political group or party in legislature" (emphasis mine). This was never meant to be restricted to parties only. It is a relatively simple matter to change the descriptions in German, French, or any other languages if they currently say that this is only for parties. --Oravrattas (talk) 20:29, 22 January 2017 (UTC)
It's not that easy. Take the US congress. Berny Sanders was for a while member of the Democratic Caucus despite being elected as an independent and not being a member of the democratic party. Using member count (P2124) is good because it means that we can say when a person leaves a group like the Democratic Caucus or joins it. ChristianKl (talk) 10:32, 23 January 2017 (UTC)

Thanks for all contributions. I can accept the overall conclusion to use number of representatives in an organization/legislature or won in elections (P1410). --ElTres (talk) 19:37, 23 January 2017 (UTC)

not applicable

Hello. Fetching data from wikidata to Greek Wikipedia, when the property has "no value", we get "not applicable". Do you know where to translate this? Xaris333 (talk) 14:20, 21 January 2017 (UTC)

el:Module:I18n/wikidata--GZWDer (talk) 14:52, 21 January 2017 (UTC)

Thanks! Xaris333 (talk) 15:37, 21 January 2017 (UTC)

@GZWDer: When a property has as date "June 2016" we have a problem in Greek Language (we have w:Grammatical case). In wikidata is correct: "Ιούνιος 2016". But in Wikipedia is shows "Ιουνίου 2016" (that will be correct if we have the date, e.x. 25 Ιουνίου 2016). Do you know how to correct this? Xaris333 (talk) 20:04, 21 January 2017 (UTC)

This is an issue of local Module:Date. @Jarekt:.--GZWDer (talk) 07:43, 22 January 2017 (UTC)
I do not know how Module:Date is used on Wikidata, but {{ISOdate}} calling this module seems to work correctly:
  • {{ISOdate|2016-06|lang=el}} gives ""
  • {{ISOdate|2016-06-25|lang=el}} gives ""
@Zolo: maybe you know how to fix this. --Jarekt (talk) 16:36, 22 January 2017 (UTC)
@Xaris333: using w:el:Template:ISOdate works for me. So maybe elWP needs to update their templates??? You haven't provided examples for us to explore at that end, though it seems that it may simply be how elWP is calling the data from WD.  — billinghurst sDrewth 00:46, 23 January 2017 (UTC)
@billinghurst: @Jarekt: See w:el:Ολυμπιακός Λευκωσίας (ποδόσφαιρο). Is correct in wikidata (Ιούνιος 1931) but wrong in the article (Ιουνίου 1931). Xaris333 (talk) 15:22, 23 January 2017 (UTC)
This seems to be an error in the #time parser function. Calling {{#time:F Y|1931-06|el}} on Wikidata returns Ιούνιος 1931 but on elwiki Ιουνίου 1931. --Pasleim (talk) 16:22, 23 January 2017 (UTC)
@Pasleim: We are using "Ιουνίου" when we have the full date. For example, 26 Ιουνίου 1932. But if we have only month and year, we must have Ϊούνιος 1932. The difference is a w:Grammatical case. Xaris333 (talk) 17:26, 23 January 2017 (UTC)
Xaris333, we understand that, but it seems like this issue has nothing to do with wikidata. As user:Pasleim pointed out the same call to the basic parser function gives different results on Wikidata and el-wiki and you identified the el-wiki answer as the wrong one. I think you will need to alert the technical community on el-wiki about this or file bug report. --Jarekt (talk) 20:32, 23 January 2017 (UTC)

Commons cat -> Wikidata script now working again

This script adds a small box on a Commons category page, to let you know if there is a corresponding article-like item on Wikidata with a Commons category (P373) pointing to the Commons category page.

The script runs whenever you're browsing Commons categories. If the Commons cat page doesn't already include a Wikidata link on the page, it's well worth adding one, using e.g.:

I find it quite useful to spot when P373s are missing, on Commons categories that really ought to have them; and also, to stop me adding a P373 to a Commonscat, if there's one already I didn't know about from an existing item -- a sign that, instead, the two items should probably be merged.

To give it a go, simply add the line


to your common.js on Commons.

It had stopped working because it was previously relying on WDQ for its lookups; I've now tweaked it to use SPARQL instead.

I think I got the changes correct, but do give it a try & let me know if anything doesn't work.

All best, Jheald (talk) 21:31, 21 January 2017 (UTC)

@Jheald: Useful, thank you. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:40, 22 January 2017 (UTC)
@Jheald: thanks from me as well. --Jarekt (talk) 20:34, 23 January 2017 (UTC)

Irish-American - is it an instance of ethnic group?

Im looking at the edits of Special:Contributions/ and s/he is adding "ethnic group" to items such as Irish American, Cuban America, Jamaican American. Would they be P31 of "ethnic group"? Strictly speaking, Irish is an ethnicity but Irish American? To me, its more of an identity, not really ethnic group. And I see identity as related but distinct from ethnic group. Any opinions?

Two issues I want to bring up with this user.

  1. 2 items. After I looked at her IMDB profile, their roles are extremely minor. Q28445406 and Q28469401.
  2. Also, their edits to some of the articles are just head scratchers.

-- MechQuester (talk) 06:49, 22 January 2017 (UTC)

All context really. It might be an ethnic group in a third country, eg. Irish-American community in Brazil, however, it is not really an ethnic group in US. I would agree with you in a second country it is more an identity.  — billinghurst sDrewth 07:46, 22 January 2017 (UTC)
I think we need a specific class for those concept such as American administration "ethnic group". This might be valuable information for Wikidata but administrative concepts might not really qualifies for beeing scientific truth and might more be a political artifact. This would give something like :
⟨ Irish American ⟩ instance of (P31)   ⟨ USA administration "ethnic group" ⟩
. author  TomT0m / talk page 09:10, 22 January 2017 (UTC)
I'm not sure there is a useful distinction between the Irish ethnic group and the Irish American ethnic group. But when deciding the question, it would probably be more useful to look at the period 1845 to around 1930 in the US, when many more distinctions were made between Irish Americans and other Americans of European extraction. Of course today hardly any distinctions are made. Jc3s5h (talk) 14:53, 22 January 2017 (UTC)
The point is not to make a difference, it's to note that the concept is typically american, and more precisely is used by the US administration. Whether or not the concept has any relevance at all is a matter of philosophical debate. But it's should be clear it's not a universally accepted classification. author  TomT0m / talk page 07:53, 23 January 2017 (UTC)
This is an recurring editor that keeps adding non-notable items and non-sourced statements. See User talk: and User talk: Sjoerd de Bruin (talk) 15:50, 22 January 2017 (UTC)
@Sjoerddebruin, he is back. MechQuester (talk) 20:51, 22 January 2017 (UTC)
Please use {{Ping}} next time. Blocked. Sjoerd de Bruin (talk) 19:01, 23 January 2017 (UTC)

Wikinews sitelinks

What are wikinews sitelinks used for? Should all news of a person be in their wikidata item? Chicocvenancio (talk) 18:31, 22 January 2017 (UTC)

Oh, no! Every wiki news item has it's own WikiData-item, and the sitelink is linking to one or several language-versions of the item. If a person is the main subject, you can use the property "main subject" to link it. It can't be connected directly, as one person can have many different news articles connected. Edoderoo (talk) 22:58, 22 January 2017 (UTC)
PS, see d:Q28473965 for an example. Edoderoo (talk)
i do not see any reason to have a data item on every story. you do not have an item to every nytimes article.[19] i note russia wikinews has a link to article subject, but english wikinews does not. maybe you should have a category to link to. Slowking4 (talk) 23:15, 22 January 2017 (UTC)
@Slowking4: It's definitely true that there are very few interwiki links for individual news stories, so one of the functions of Wikidata isn't really applicable there. But we can have items for individual stories and they can have meaningful statements about their main subject, for instance and someone performing searches could get more useful information from that. —Justin (koavf)TCM 18:27, 23 January 2017 (UTC)
Individual news stories seem like a particularly awkward case for wikidata-based interwikis (although these same problems occur with all wikidata-based interwikis, the problems are particularly blatant with news articles). Wikidata tries to provide separate items whenever there is any distinction between things, it tries hard to not conflate distinct things, whereas interwikis should be provided as often as possible whenever there is a most-nearly-analogous page on another project. But with news stories, it's rather routine for two different news stories, supposedly "about the same event", to be really two different news stories. A news article is a snapshot in time, and under ordinary circumstances no two snapshots of something are quite the same; they're taken at slightly different moments, they're taken from slightly different angles, the lighting is different. If there's some big disaster (an earthquake, or tsunami, or airplane crash, or whatever), and French Wikinews publishes an article about it on Tuesday, English Wikinews publishes an article about it on Wednesday (containing some information on development of the story after the French article was published), German Wikinews publishes one Thursday, and French publishes another on Friday, probably no two of those four articles are alike. It'll be a problem for English or German Wikinews to decide which of the two French articles to interwiki to, but whatever they decide Wikidata would be wise to position itself so that it can avoid the politics of the question. It can get to be even more fun (so to speak) if French and German and English Wikinews all publish on Tuesday but choose to focus on very different aspects of the disaster. --Pi zero (talk) 21:13, 23 January 2017 (UTC)
it's all good, i will just go a create a WD item for each EB1911 article, and pubmed article. and especially every article used as a reference. much easier to find them. Slowking4 (talk) 02:55, 24 January 2017 (UTC)

Should we use part of (P361) or location (P276)?

Should we use part of (P361) or location (P276) when stating that a roller coaster (Q204832) is in a specific amusement park? //Mippzon (talk) 08:07, 23 January 2017 (UTC)

I think location makes more sense. ChristianKl (talk) 10:44, 23 January 2017 (UTC)
May be both?--Ymblanter (talk) 16:25, 23 January 2017 (UTC)
I noticed when browsing roller coaster (Q204832) I noticed many of the used part of (P361). But maybe as you say, we should use both? //Mippzon (talk) 16:48, 23 January 2017 (UTC)

Find all articles from specific category (recursively) that do not have WikiData entry

Do you know a way to find all articles that have specific category or any if it's subcategories assigned, in specific local wikipedia, that do not have entry in WikiData. I would like to add such entries, but it will take a lot of time to find such articles in the category that I need, since there are a lot of them. --StanProg (talk) 10:50, 23 January 2017 (UTC)

You can use PetScan. For example this is all articles from "Страницы разрешения неоднозначностей по алфавиту" category of ruwiki that do not have Wikidata item.--ԱշոտՏՆՂ (talk) 12:03, 23 January 2017 (UTC)
Duplicity also provides such functionality. It offers possible matches on Wikidata as well. —MisterSynergy (talk) 20:19, 23 January 2017 (UTC)

Belated-welcome template

I've just created {{Welcome-belated}} and {{Welcome-belated/text}}. Please can someone mark them for translation? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:58, 30 January 2017 (UTC)

I wonder if there is way we could avoid duplicatin with {{Welcome}}. They look almost same, except for the initinal paragraph. Matěj Suchánek (talk) 15:13, 30 January 2017 (UTC)
I expect that that will change, in time. Compare en:Template:Welcome and en:Template:Welcome-belated . Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:24, 30 January 2017 (UTC)
  Done Matěj Suchánek (talk) 20:04, 30 January 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 20:04, 30 January 2017 (UTC)

New nomination

Hello.Please participate here.Thank you --ديفيد عادل وهبة خليل 2 (talk) 08:28, 24 January 2017 (UTC)

Property documentation template question

Would it be possible for someone to have a look at the question I raised here two weeks ago? Thanks. Carcharoth (talk) 11:50, 24 January 2017 (UTC)

How does one edit pages like Module:Property documentation? It looks really difficult to do. Carcharoth (talk) 11:57, 24 January 2017 (UTC)
Yeah, they are called modules and you have to know Lua (Q207316) in order to maintain them. Matěj Suchánek (talk) 12:58, 24 January 2017 (UTC)
@Jura1: You added this in Special:Diff/330807441. Matěj Suchánek (talk) 12:58, 24 January 2017 (UTC)
What I want to do is change "linkText = "+"" to something like "linkText = "run query to view all instances of this property"", which is more informative than "+". I worked out that you have to click the 'paragraph' button 'P' in the toolbar at the top (the 'toggle invisible characters' bit) to be able to see that and edit that. Very opaque and difficult for new editors to understand and access. Wikidata doesn't make it easy for people to understand how to edit Carcharoth (talk) 13:09, 24 January 2017 (UTC) Made the change I wanted to make, see here. Carcharoth (talk) 13:11, 24 January 2017 (UTC)

It is still not quite right. If you look at Property talk:P1920 you see that the links at the top link to running a query. So if you want to find all instances of that property, you click once to bring up the query, and then again to bring up the results. Wouldn't it make more sense to link people in one click to a list of all instances of a property? It would also make more sense to link it from the 'Current uses: 192' bit on that page. It would also make even more sense to link from the main property page, rather than the talk page. Where was the decision taken to put stuff like this on the talk pages? Carcharoth (talk) 13:18, 24 January 2017 (UTC)

Fallback languages for labels in clients

Fallback languages as of 2015

The Lua functions mw.wikibase.label and mw.wikibase.getLabelWithLang in Wikibase clients use fallback languages according to their descriptions. Where and how is the used sequence of fallback languages defined? Thank you for your help. Dipsacus fullonum (talk) 13:24, 24 January 2017 (UTC)

It's hard-coded inside MediaWiki. The rules are:
  1. the last language is always English,
  2. variants like Swiss German fall back to the main (parent) language,
  3. some languages may also fall back to a language which majority of its speakers understands; you can find this information for each language inside this folder, eg. here you can find that Czech falls back to Slovak.
Matěj Suchánek (talk) 14:46, 24 January 2017 (UTC)
I think if fallbacks other than these chains are also needed, but shouldn't be hardcoded (e.g. we finally cancelled uk -> ru fallback, but some non-WMF users (i.e. (Q9376349)) may still want it; and at least I wanna fallback yue to zh-hant or zh-hk/zh-mo but @Shinjiman: opposed it?), then this Phabricator task may help you. --Liuxinyu970226 (talk) 15:07, 24 January 2017 (UTC)
Also, regarding this svg file, I suggest both zh-hans and zh-hant be fallbacked each other, like pt <-> pt-br and cdo <-> nan. --Liuxinyu970226 (talk) 15:11, 24 January 2017 (UTC)
Thank you all for the answers. It was not what I hoped for, but new functions to get labels with configurable fallback languages fortunately can be made in Lua. Best regards, Dipsacus fullonum (talk) 21:15, 24 January 2017 (UTC)

Can Descriptioner override the existing description?

Im just wondering. MechQuester (talk) 16:17, 24 January 2017 (UTC)

no --Pasleim (talk) 17:07, 24 January 2017 (UTC)

Leader not head of a country

When a person is recognised as a leader to his people but is not the head of government. How do you do this in Wikidata? Thanks, GerardM (talk) 15:58, 24 January 2017 (UTC)

Still a leader of the country, in my opinion. He's the leader of the state known as the country, not the government, thus a representative. MechQuester (talk) 16:07, 24 January 2017 (UTC)
We are talking about a function of a people who has no formal connection to a government.. nor of a country as that people may live in multiple countries. Thanks, GerardM (talk) 05:05, 25 January 2017 (UTC)
A person, "who manages any kind of group", here population of a country / multiple countries is director / manager (P1037). - Kareyac (talk) 06:45, 25 January 2017 (UTC)
Lets make the assumption that there is a (kind of) leader of all Elfdalian (Q254950)-speaking people, then I guess it's wrong to put that claim in the item about the "country of Elfdalia" or in the item about the language of Elfdalian. It is probably better to have such information in an item dedicated for the "the (informal) group of Elfdalian people and speakers of Elfdalian". That item could be described as a sort of "informal network", without government. Such a network can have a (sort of) leader, a founder or father/mother-figure. -- Innocent bystander (talk) 08:37, 25 January 2017 (UTC)
For founder we have founded by (P112). Innocent bystander, sorry for misleading typo. - Kareyac (talk) 09:33, 25 January 2017 (UTC)
A country can have a head of state, they can have a head of government; they can have viceroy type of position. But what are you meaning by a "leader"? They have an official position or they do not. If they are an influencer, a spokesman, an intellectual, ... So I think that the loose use of the word leader, needs to be better qualified as there is some reason or purpose behind their rise in the area of influence.  — billinghurst sDrewth 09:47, 25 January 2017 (UTC)

can't enter values for "inventory number"

At Cloudy Sky - Mediterranean Sea (Q18627381) [20] I can't enter the value "84.XM.1388" for Getty Museum. I can't even re-enter the existing value "1971.577.2" for Art Institute of Chicago. I checked the regex for this prop and it seems correct. So what gives? --Vladimir Alexiev (talk) 08:37, 25 January 2017 (UTC)

Lets take the stupid answers first?! Have you tried to reload the page? I experience problems like those you describe from time to time, but they are often solved by reloading the page (sometimes more than once). -- Innocent bystander (talk) 09:50, 25 January 2017 (UTC)
I sometimes encounter that problem if I copy values from a source that includes hidden characters. Have you tried typing the value manually? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:27, 25 January 2017 (UTC)
WORKSFORME? Multichill (talk) 17:34, 25 January 2017 (UTC)

How does patrolling in languages like Arabic and Chinese work in practice?

If I look at the recent changes list to check for vandalism, what am I supposed to do when I see a label in a language which I don't speak and that uses a different alphabet so that I have no real ability to see whether it's right. What do other editors do? ChristianKl (talk) 11:47, 25 January 2017 (UTC)

More than in any other Wikimedia project, filtering is important for RC work here at Wikidata. There are at least two extremely useful tools to do that: User:Yair rand/DiffLists.js allows filtering of RC, watchlist and Special:Contributions in the Wikidata frontend; reCh by User:Pasleim provides similar functionality in an external application, including a very powerful batch-patrol functionality. By using one or both of these tools, one can filter RC in different manners and make sure that things one definitely does not understand (such as zh/ar terms, etc) do not overwhelm during RC work. —MisterSynergy (talk) 11:57, 25 January 2017 (UTC)
The reCh tool provides a translation feature. Behind terms, there is a small icon. If you click it, the English translation is displayed. --Pasleim (talk) 12:18, 25 January 2017 (UTC)
  • Is there a reason why the script by Yair rand isn't in the gadget list? ChristianKl (talk) 12:31, 25 January 2017 (UTC)
  • No idea, but I would support a proposal to make it a gadget. To my opinion it would even be worth to implement similar functionality directly in Mediawiki. The classical layout of RC, watchlist and Special:Contributions works nicely for unstructured text, but this project is fundamentally different. —MisterSynergy (talk) 12:39, 25 January 2017 (UTC)
  • I speak Chinese decently and for the msot part, they are good. MechQuester (talk) 17:03, 25 January 2017 (UTC)

Merge request: Q20967764 and Q5108886

Hi! Can someone merge Voore (Q20967764) and Voore (Q5108886)? //Mippzon (talk) 16:52, 30 January 2017 (UTC)

  Done. - Kareyac (talk) 16:59, 30 January 2017 (UTC)
Thanks! //Mippzon (talk) 17:03, 30 January 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 12:24, 31 January 2017 (UTC)

URL to diff

{{URL to diff|}} isn't working, but I can't see why. Can anyone fix it, please? IIRC, it was imported from, or modelled on, the version on en.Wikipedia, which does work. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:58, 19 January 2017 (UTC)

Fixed, it was the page move that broke it. Matěj Suchánek (talk) 11:34, 20 January 2017 (UTC)
@Matěj Suchánek: Thank you. I hadn't expected that when I moved it, and am surprised to see a hard-coded reference to a template name, in the Lua module. Is here not a way of avoiding that, or of making the module recognise redirects? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 07:53, 21 January 2017 (UTC)
From the documentation, I assume this hard-coding seems to be necessary. Even though there are ways to work with redirects in Lua, I don't think the module should attempt to fix them. Matěj Suchánek (talk) 08:57, 21 January 2017 (UTC)

I would seem sensible for templates affectced by the above issue to be move-protected. Is there a counter argument? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:55, 24 January 2017 (UTC)

Maybe we should just make sure that users that can edit these templates make sure that they know what they are doing. If problems are detected, just leave a note on their talk page.
--- Jura 17:27, 26 January 2017 (UTC)

We don't need a sandbox item.

Given that we have I don't see the point of Wikidata sandbox (Q4115189). I think it would be useful if the main website would link to for the preferred sandbox.

Having the Sandbox item on the main Wikidata means that it might influence queries. ChristianKl (talk) 11:38, 24 January 2017 (UTC)

Can you test modules on client projects through I know that the sandbox item is used a lot for that on svwiki. Ainali (talk) 12:03, 24 January 2017 (UTC)
I don't think can be considered a replacement for the sandbox item. It doesn't have the same properties and items, it doesn't have the same gadgets and doesn't always behave the same. I'm also not aware of any tools which allow you to pick which server to edit against (e.g. what if you want to test your QuickStatements commands?). - Nikki (talk) 13:02, 24 January 2017 (UTC)
There is a whole group of sandbox items, not only one. I think they are very useful as Ainali and Nikki describes above. If a bot once a week empty these items, they only cause temporary problems in the queries. -- Innocent bystander (talk) 13:49, 24 January 2017 (UTC)
@Ainali: We can import (or just copy em since all non-Main/Property namespaces are still CC BY-SA?) modules from en/fr/de... wikis, like what a number of Indic Wikipedias are doing. --Liuxinyu970226 (talk) 15:14, 24 January 2017 (UTC)
@Liuxinyu970226: How would that help the users on those projects? They most probably want to test the modules in their real environment since they in turn probably rely on other modules and templates. Ainali (talk) 15:19, 24 January 2017 (UTC)
One reason we have not imported these modules, is that we have locally discovered demands that these (en/fr/de-)modules have not fullfilled yet. -- Innocent bystander (talk) 17:09, 26 January 2017 (UTC)
Maybe ChristianKl can outline how he does testing?
--- Jura 17:19, 26 January 2017 (UTC)

Logical OR in the SPARQL templates

I'm using the SPARQL and SPARQL2 templates. My latest SPARQL query contains a logical OR, indicated in SPARQL by two pipes:

FILTER (!BOUND(?lang) || ?lang = wd:Q1860)

And the template cuts it off (See here). What do I need to do to escape the double pipe? Or is there a change that can be made to the templates? Thanks in advance for any help, MartinPoulter (talk) 20:23, 24 January 2017 (UTC)

Use {{!}} → |. Matěj Suchánek (talk) 20:33, 24 January 2017 (UTC)
I coded something to not to escape stuffs using lua. The idea is that it's easy in lua to concatenate all the numeric arguments of a template call with pipes between them and that it become unecessary to escape the pipes. I'll re-find this, please stand by. author  TomT0m / talk page 11:51, 25 January 2017 (UTC)
It's Module:ConcatArgs and actually it's used in the SPARQL template. Maybe there is a bug concerning the double pipe, I'll investigate. author  TomT0m / talk page 11:56, 25 January 2017 (UTC)
Found the problem : the "=" symbol. Mediawiki thinks "?lang" is a named parameter. It works without escaping by saying it's supposed to be the 2nd numeric parameter :
{{SPARQL|query=SELECT DISTINCT ?person ?name ?language ?death (URI(CONCAT("", ?gutenberg)) AS ?gberglink) WHERE {
  ?person wdt:P1938 ?gutenberg.
  ?person wdt:P570 ?death. # Dead people only
    ?enws schema:about ?person.
    ?enws schema:isPartOf <>
  OPTIONAL {?person wdt:P1412 ?lang}.
  FILTER (!BOUND(?lang) ||2= ?lang = wd:Q1860) # Language: English or absent
  BIND(IF(BOUND(?lang),"English","Not specified") AS ?language
)  ?person rdfs:label ?name.
  FILTER((LANG(?name)) = "en")
ORDER BY ?death
I don't think it's fixable, so no better solution than the proposed one. author  TomT0m / talk page 21:34, 25 January 2017 (UTC)

Is there an easy way to get a list of all items with a specific type of claim, as well as the claim itself?

Hi, I've been looking for a while for an easy way to do this, but drawn a blank. What I'd like is to get a list (or page-able subset) of all items with claim Universal Decimal Classification (P1190) in them, along with the actual value of the claim made. For instance, a single record for religion in China (Q1482612) might read "Q1482612|221".

It's easy enough to get the list using "what links here", but then I'd have to look individually at every item to get the value which would brutalise the server I suspect. Thanks in advance. Lankiveil (talk) 02:10, 26 January 2017 (UTC).

This should do it. Only 122 items. --Tagishsimon (talk) 03:00, 26 January 2017 (UTC)
That is exactly what I wanted, thanks @Tagishsimon:! Lankiveil (talk) 08:54, 26 January 2017 (UTC).

P1617 regex

I added the regex constraint {{Constraint:Format |pattern=[0-9a-f\-]+ |mandatory=true}} to Property talk:P1617. The constraint report now lists exceptions including cbeab979-c95b-432e-a3bb-2b1d502f4db5 and 1e655b90-b289-4762-9bae-ee980eeae9f9. According to the regex tester I use, these should be valid. What have I missed? I do note that all the unexpected exceptions have repeated, adjacent characters. Is that a coincidence? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:12, 26 January 2017 (UTC)

I guess it is the space at the end of the pattern causing the problem. --Pasleim (talk) 14:33, 26 January 2017 (UTC)
It is best to enclose the pattern by <nowiki></nowiki>. Perhaps the problem is that the final space is (sometimes?) included. Lymantria (talk) 14:38, 26 January 2017 (UTC)

Wikidata weekly summary #244

I've noticed that all wind mills/tide mills/fire stations in these queries has no sitelinks. Are they notable? Can I create such elements for the same objects in my city? --Infovarius (talk) 15:29, 27 January 2017 (UTC)


In English, we have two lists of his compositions, en:List of compositions by Francis Poulenc and en:FP (Poulenc). The former corresponds to Italian and Japanese, the latter to the French fr:Liste des œuvres de Francis Poulenc. Help? --Gerda Arendt (talk) 17:29, 24 January 2017 (UTC)

I do not think we can help unless you decide to merge the two entries on the English Wikipedia. If you think that Italian, Japanese, and French should be linked to the same English item this can be easily done.--Ymblanter (talk) 20:00, 24 January 2017 (UTC)
A merge was done for Max Reger, but Poulenc was written by an editor who retired, - so for respect I wouldn't want to touch it. Also: some argue that people who can't sort would still need the bulleted list. I could imagine one entry for list, the other for complete catalogue. --Gerda Arendt (talk) 12:47, 25 January 2017 (UTC)
@Gerda Arendt:, on Wikidata all three items (fr, it, ja) are connected to en:List of compositions by Francis Poulenc, and en:FP (Poulenc) is not connected to anything. The link you see at en:FP (Poulenc) is because someone added by hand an interwiki link to the English article. If you want, you can also add there Italian and Japanese links.--Ymblanter (talk) 13:26, 25 January 2017 (UTC)
I was the someone. What I miss would be a link from French to that article, because FP is a translation of the French, while the English list (translated to Italian and Japanese) is only a subset. --Gerda Arendt (talk) 13:29, 25 January 2017 (UTC)
We have a mediawiki restriction that one wikidata item can only link to one Wikipedia (in a given language) article, and one Wikipedia article can only be linked from one Wikidata item. This means that in the given configuration on the Wikidata end we can only shift links around (for example, move the French link from one item to another one). We can not have the same French article references from two Wikidata items. It is technically not possible.--Ymblanter (talk) 13:48, 25 January 2017 (UTC)

en:FP (Poulenc) is about a specific catalog of Poulenc's works, whereas en:List of compositions by Francis Poulenc is a list of compositions by Poulenc. Whereas these items are highly correlated, they are quite different. It might be a bit easier to see for Mozart, where the same difference is between en:Köchel catalogue and en:List of compositions by Wolfgang Amadeus Mozart. Or, put differently, the list of compositions are Wikipedia list articles, the catalogues are "real world" entities that exist outside of Wikipedia. This is particularly important where there are several catalogues for a single composer. --Denny (talk) 17:40, 25 January 2017 (UTC)

  • I completed the two items (Q1469914, Q28441364). From its title, I think it's correct to link the French list from the first one.
    --- Jura 08:21, 26 January 2017 (UTC)
  • I've renamed english article from FP (Poulenc) to FP (catalogue) as "Poulenc" is quite synonim but not description of the title. And the author is Schmidt after all. --Infovarius (talk) 14:42, 27 January 2017 (UTC)

Bug at langlinks API

HI, I made an article in fa.wikipedia when i check the api it doesn't purge interwiki.

I purge many time at wikidata, fawikipedia, enwikipedia but still API doesn't show new interwiki linkYamaha5 (talk) 21:51, 26 January 2017 (UTC)
@Yamaha5: It is there now
"lang": "fa",
"*": "\u06f1\u06f0\u06f0 \u0645\u0627\u06cc\u0644 \u0647\u0627\u0648\u0633"
 — billinghurst sDrewth 03:10, 27 January 2017 (UTC)
More than 2 hours it doesn't show. the api has lag and it should be solve Yamaha5 (talk) 07:18, 27 January 2017 (UTC)

Query regarding the descriptions for wikidata entities

Hi all,

We have been using wikidata json dump and labels, wiki links and external ids are very useful. Thanks a lot for such a good work. I was trying to look at the descriptions of wikidata entities and looks like there's not much consistency.

For eg: Brad Pitt: American actor

            Angelina Jolie: American actress, film director, and screenwriter
            Jennifer Aniston: television and film actress from the United States

           Even for cities:
           Tokyo: capital of Japan with 13 million inhabitants, and one of 47 prefectures of Japan
           Seattle: major city in state of Washington, United States; county seat of King County, Washington
           Mumbai: capital and district of Maharashtra, India
           Washington D.C: capital city of the United States

I was thinking descriptions for wikidata entities is equivalent to notable_for attribute in freebase dump. Can you please let me know if there's any other attribute in wikidata entity which I could use to map as short Description?

Thanks in advance.


you could auto-generate descriptions using properties like instance of (P31), subclass of (P279), country of citizenship (P27), occupation (P106) etc. What properties are useful depends on what kind of entity you are looking at. ArthurPSmith (talk) 15:25, 27 January 2017 (UTC)


Can some experienced user(s) check recent contributions of user GiorgiXIII? There are several items related to army and armed forces affected, including these ones [21], [22], [23], with tens of sitelinks and/or labels changed/removed. There may be some lost sitelinks and/or altered tree classification. Right now I can't spend more time looking into this case. XXN, 00:15, 27 January 2017 (UTC)

I have reverted those edits and assuming good faith at the moment. It seems to me that the user thought some sitelinks were misplaced and was trying to fix it. —Wylve (talk) 02:52, 27 January 2017 (UTC)
I fixed some sitelink issues that I think User:GiorgiXIII had some issues with. The issue remains that w:lv:Armija is a more comprehensive disambiguation page for various senses of the word "army" (meaning either the military or only land troops). We also have ground forces (Q3505278) which I am unable to determine what it is supposed to represent. Maybe some speakers of Slavic languages can help. —Wylve (talk) 03:20, 27 January 2017 (UTC)
Added some Slavic descriptions, hope it helps. - Kareyac (talk) 05:06, 27 January 2017 (UTC)
@Kareyac: What is the difference between ground forces (Q3505278) and army (Q37726), if any? —Wylve (talk) 09:12, 27 January 2017 (UTC)
@Wylve: please try to use dictionary or some online translator. I tried: ru, be, cs, sr and uk WP articles describe ground forces (Q3505278) as all ground forces as class (not Navy or AF) and army (Q37726) as all ground forces or its unit/part/division (eg. Army № ...). I didnt research army (Q37726) in WPs if they have no ground forces (Q3505278) to compare. - Kareyac (talk) 09:55, 27 January 2017 (UTC)
Helo. I am Georgian man, I very good speac in georgian and russian. Georgian შეიარაღებული ძალები — russian Вооруженные силы, Georgian არმია — russian Армия, Georgian სახმელეთო ძალები — russian Сухопутные войска. My agge 54 yar. I live in Tbilisi (republic Georgia).GiorgiXIII (talk) 13:19, 27 January 2017 (UTC)
@GiorgiXIII: Welcome to Wikidata! You can use Wikidata:ფორუმი or Wikidata:Форум (or this page, of course). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:44, 27 January 2017 (UTC)

Language fallback

fallback chains (might be out of data)

I'm not clear how language fallback works. If we have a label and description in "en", is there any benefit in providing identical text in "en-GB", for example? If not, it seems to me that doing so increases the maintenance burden when one of those values is changed. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:25, 25 January 2017 (UTC)

The way it works is that if user sets his preferred language to "en-GB" then the software will check if we have message in that language first before moving on into fallback "en" language. So Andy you are right there is no benefit (that I can think of) to adding "en-GB" message which is the same as "en". --Jarekt (talk) 18:43, 25 January 2017 (UTC)
Can non-Wikimedia users of data use the same fallback mechanism? —Wylve (talk) 19:07, 25 January 2017 (UTC)
You can edit your modules even in WMF-wikis so that you can make a home-made "fallback". For example, having English as fallback for Swedish in words originally written with Cyrillic script (Q8209) is not a very good idea. Norwegian and Danish and even German are in those cases considered better. -- Innocent bystander (talk) 19:18, 25 January 2017 (UTC)
Module:Fallback implements the fallback chain in the image. c:Module:Date's langSwitch function calls mw.language.getFallbacksFor Lua function to do the job. The results are almost identical. --Jarekt (talk) 21:07, 25 January 2017 (UTC)
@Pigsonthewing: There is virtually no value in en-CA/en-GB in the first place, especially since they only exist for those two varieties (no -US, -IN, etc.) Simply put, we should have a bot with a good dictionary automatically generate them anytime an en label is made at all. —Justin (koavf)TCM 03:42, 30 January 2017 (UTC)

Need a second opinion in a discussion about a role of Wikidata

I am engaged in a lively debate with user:Brya at WikiProject_Taxonomy about role of the Wikidata projects as related to taxonomy. We seem to be unable to convince each other and nobody else from WikiProject_Taxonomy seems to be joining the discussion, so I am seeking more perspectives on the issue. We are discussing Q21440769 an item for a family of flies called Heterocheilidae (McAlpine, 1991). Unfortunately McAlpine who named the family 26 years ago made a mistake and used the same name as a family of worms ( Heterocheilidae (Q5320961) ) was using since 1915. It does not seem like any biologist published anything about this naming issue. Both names are beeing used at the same time, sometimes even in the same publication, like in this 2011 publication (pages 80 and 227), despite the fact that ICZN convention states that the newer name is incorrect. We agree about those facts, but we do not agree about how to reflect them in the database. My position is that per Wikidata:Verifiability we report most up to date facts from literature, so we should report "Heterocheilidae" (McAlpine, 1991) as the current name and use the item to store data about this family of flies. We already have this item tagged as an instance of later homonym (Q17276484) and I think that is all we need to do, since we only suppose to report published facts, not do our own investigation. user:Brya, who reverted my edits, has a firm belief that since the name is wrong we can not use the item to store data about this family, until I "find[] a taxonomist to fix this permanently, for the whole world". Can someone help us find some middle ground here? --Jarekt (talk) 20:34, 25 January 2017 (UTC)

Technically, we can have two items, one for the flies, another for the wormies. They can have the same name/label (with a different description). That's all technically. If there are sources that the grouping is as you describe, we should register this, unless there are better sources that this grouping is outdated (then we might use qualifiers) or wrong (than we should forget about it, and name it in the article, but not in WikiData). But my knowledge of taxonomy is pretty superficial. Edoderoo (talk) 21:08, 25 January 2017 (UTC)
@Jarekt: What kind of „facts“ do you want to add to this item? I've made some minor improvements around the genus Heterocheila (Q14604579) (2 species) including Relationships of the genus Heterocheila (Diptera: Sciomyzoidea) with description of a new family (Q28528221) by David K. McAlpine (Q21502932). --Succu (talk) 22:26, 25 January 2017 (UTC)
How about the information that was removed: taxon name (P225), taxon rank (P105), parent taxon (P171), and that is is a taxon (Q16521). That would be a good start. Some of the queries I was using rely on those properties. --Jarekt (talk) 02:38, 26 January 2017 (UTC)
Jarekt has his facts wrong: these two items do not use the same name. These are two different scientific names, that is, two different formal entities, that have in common that they share the same spelling. In all other respects they are different (like author, date, type). The later of these two names may not be used as the scientific name for a taxon (Article 52.2). At some point, a taxonomist will (hopefully) take action to amend the situation.
        However, we have an item for the later name, and properties can be stored there (and indeed are there), so the problem is limited. If desired, the name can be added using name (P2561) as a qualifier. What we really should have is a separate property "scientific name, that may not be used a taxon name", but we don't have one. - Brya (talk) 04:40, 26 January 2017 (UTC)
Ok, different name with the same spelling. And the problem with the item is not limited, if my attempts to place this family in the taxonomy tree or provide currently used scientific name is reverted --Jarekt (talk) 14:01, 26 January 2017 (UTC)
I think there is little we can do at the moment. I wouldn't even recommend deprecating the official name of the family of flies since de facto it has become the official name. Not everyone actually follows all the rules to the letter even in academia. The maximum I can recommend is using different from (P1889) on both items and adding a note in the talk page. DGtal (talk) 06:54, 26 January 2017 (UTC)
But my issue is that it is not our job to be the spelling police for the biology community, correcting their errors. Our job is to make published information available in the database form. If there is a published source saying not to use this scientific name than we should change it, but at the moment that is the name used by scientific community. Wikipedia:No original research (Q4656524) policy prohibits this kind of behavior on Wikipedias, but I guess Wikidata is different. --Jarekt (talk) 14:01, 26 January 2017 (UTC)
It seems to me Jarekt is correct here - we should be reflecting what has been published by reliable sources, with proper referencing etc - the statement is not a statement by wikidata, it is a documentation by wikidata of what the source states. ArthurPSmith (talk) 15:19, 27 January 2017 (UTC)
But we are recording that the name has been published, and we could record by who the name is used. It is just that the fact that the name has been used, or even that the name is being used, does not mean that it is a valid name of a family: the rules say it can not be. - Brya (talk) 18:53, 27 January 2017 (UTC)
But Brya our role is not to be interpreters of scientific rules, only to store published facts. So, if name of this family is being used by scientific community than we should place it in Q21440769: taxon name (P225), taxon rank (P105), etc. If the taxon name is not in taxon name (P225) property, it is not being recorded, since any other placement will be ignored by the queries. --Jarekt (talk) 19:23, 27 January 2017 (UTC)

Achim Raschka (talk)
Andrawaag (talk)
Brya (talk)
CanadianCodhead (talk)
Dan Koehl (talk)
Daniel Mietchen (talk)
FelixReimann (talk)
Hyperik (talk)
Infomuse (talk)
Infovarius (talk)
Jean-Marc Vanel
Joel Sachs
Klortho (talk)
Lymantria (talk)
Magnefl (talk)
Mellis (talk)
Michael Goodyear
Mr. Fulano (talk)
Nis Jørgensen
Peter Coxhead
Andy Mabbett (talk)
Prot D
Rod Page
Strobilomyces (talk)
Stuchka (talk)
Succu (talk)
TiagoLubiana (talk)
Tommy Kronkvist (talk)
Tris T7 TT me
William Avery
Mike Krüger
GoEThe (talk)
  Notified participants of WikiProject Taxonomy Can we have some more viewpoints from Taxonomy project? --Jarekt (talk) 19:29, 27 January 2017 (UTC)

Before having an opinion, I checked up at WS and there they are separated through using (Diptera) resp. (Nematoda); Heterocheilidae_(Diptera) and Heterocheilidae_(Nematoda) Dan Koehl (talk) 20:11, 27 January 2017 (UTC)
Reg the Diptera: The name of this taxon appears to be invalid under the relevant nomenclatural code, as it is a junior homonym of Heterocheilidae Railliet & Henry, 1915. Dan Koehl (talk) 20:15, 27 January 2017 (UTC)
The question is: should wikidata store published scientific name of Q21440769 in taxon name (P225), even if our interpretation of naming rules suggest that the name scientists have been using for all those decades does not meet naming rules? That is the essence of this discussion. The note on WS, just alerts people about the naming issue. --Jarekt (talk) 20:31, 27 January 2017 (UTC)
The question is: how to model this. --Succu (talk) 21:02, 27 January 2017 (UTC)
BTW: It's not our interpretation of ICZN. It's one of the easier facts and a common rule called Principle of Priority (Q2110868). So how many times was this name used to denote the taxon concept of David K. McAlpine (Q21502932). You cited one. --Succu (talk) 21:36, 27 January 2017 (UTC)
I would not call the period from 1991 to now "all those decades". I am mystified by the "our interpretation of naming rules" (who is this we, where do "naming rules" come in, and what is this "interpretation"?). Also, the "scientific rules" is bewildering, science does not come into it.
        What we are dealing with here is the International Code of Zoological Nomenclature, one of a number of Codices, lawbooks, that have worldwide support, to the degree that there are no competing alternatives (that have any kind of real support). These lawbooks govern (and, to a real degree, create) their own autonymous nomenclatural universes.
        And in this particular case it is crystal clear that this name can not be used as the correct name of a taxon in this particular universe. The name has nevertheless been published, and is being used here and there (and we can record this usage). But probably it is not being used a lot, or it would have been prominent enough for somebody to take appropriate action.
        I find this desire to cover up errors and create an alternate reality quite disturbing. - Brya (talk) 05:24, 28 January 2017 (UTC)
Despite of the fact that both taxa have been named more than 20 years ago, they are still used in publications. If we draw conclusions that a used taxon name (P225) is in fact not valid under the ICZN, however obvious in this case, that is an act of original research. We should not. Lymantria (talk) 07:29, 28 January 2017 (UTC)

<grin> we choose for a specific understanding of taxonomy and it is wrong </grin>. The scientific name is a combination of several parts. For a species it is a Genus, a species name, an author, a publication and a date. A species is defined by a type. The same type can be used for many scientifically correct genera at or below the level of species and they are roughly the same. Each entry is correct from a taxonomic point of view. Given that we are not in the business of having it right, we have a mess and all kinds of claims can be made. Wikipedia is an encyclopaedia and its aim is to present a view, the current understanding. Wikidata is not Wikipedia, it does not need to be. It is a project in its own right. Thanks, GerardM (talk) 08:26, 28 January 2017 (UTC)

@GerardM: Actually, the "Wikipedia is an encyclopaedia and its aim is to present a view, the current understanding" belongs in "What Wikipedia is not: Wikipedia_is_not_a_soapbox". The aim of Wikipedia is to represent "fairly, proportionately, and, as far as possible, without editorial bias, all of the significant views that have been published by reliable sources on a topic." [emphasis added] Other projects like Wikispecies and Commons have chosen to adopt instead a Single-Point-of-View stance, so NPoV is not a universal WMF feature. However, Wikidata is intended to serve Wikipedia, so it should be compatible with NPoV.
@Lymantria: I think a policy of just copy-and-paste of whatever is found in sources is untenable. Just as any user who writes a Wikipedia page should place content in context, in a user should make sure that what is entered into a Wikidata item is "structured data" (per the "Wikidata acts as central storage for the structured data" of the main page). - Brya (talk) 09:10, 28 January 2017 (UTC)
@Brya: The tool we have for that is our Notability criteria. In this case, several publications have relied on the supposedly unvalid name and not just a not very reliable database. I think we should refrain of passing by publications that can be mentioned a "reliable" source. The fact that the name is incorrect, does not make the data unstructured. One could even argue that taking away taxon name (P225) etc. removes the available structure. Lymantria (talk) 09:39, 28 January 2017 (UTC)
Wikidata has as one of its functions that it serves Wikipedia. If that is all it does I would not participate. Thanks, GerardM (talk) 15:02, 28 January 2017 (UTC)
@Brya: : Please create an "ICZN name" property. This code is supposed to provide a normalized identifier of a concept - a taxon in our case. This is pretty a common situation in Wikidata, and in such cases it's clear which code should be used. This way you will have stuff organized the way you want while having a chance for others to adopt another viewpoint as it's standard in Wikidata. Wikidata is a platform that permits web scale data crossing between several databases. It's one of its obvious usecase as we as many many identifiers stored. The ICZN one, on this viewpoint, is an important one, but not an uncommon one. This does not mean it has to eat and shade any other. author  TomT0m / talk page 10:32, 28 January 2017 (UTC)

Achim Raschka (talk)
Andrawaag (talk)
Brya (talk)
CanadianCodhead (talk)
Dan Koehl (talk)
Daniel Mietchen (talk)
FelixReimann (talk)
Hyperik (talk)
Infomuse (talk)
Infovarius (talk)
Jean-Marc Vanel
Joel Sachs
Klortho (talk)
Lymantria (talk)
Magnefl (talk)
Mellis (talk)
Michael Goodyear
Mr. Fulano (talk)
Nis Jørgensen
Peter Coxhead
Andy Mabbett (talk)
Prot D
Rod Page
Strobilomyces (talk)
Stuchka (talk)
Succu (talk)
TiagoLubiana (talk)
Tommy Kronkvist (talk)
Tris T7 TT me
William Avery
Mike Krüger
GoEThe (talk)
  Notified participants of WikiProject Taxonomy : I went bold and I sketched a proposal, please read this page. I did not included it yet as an official proposal as it lacks all the technical details. I hope this can put an end to this controversy. author  TomT0m / talk page 10:52, 28 January 2017 (UTC)

Just wanted to make a point as a nomenclatural taxonomist. Under the code a name is supposed to be followed as per usage until such time as a valid nomenclatural act has been published to declare it unavailable for whatever reason. This nomenclatural act is governed by the rules of publication (arts 8-9) and WS does not meet these articles hence can only mention an issue but cannot change the nomenclature. It does not matter how problematic a name is you must use it until demonstrated to be at issue. The reason for this is to avoid multiple nomanclatures, anyone can read the code and possibly determine a name has an issue and should be unavailable, but others may disagree, if everyone did what they felt was right, we end up with multiple names. So to determine that a name is a junior homonym on Wikimedia you must cite the nomenclatural act that declared it a homonym, if you cannot then all you can say is it may be a homonym under the code but then use it anyway. I know not everyone follows this, and hence we have multiple names on many taxa. But this is how it is supposed to work. Cheers Scott Thomson (Faendalimas) talk 11:14, 28 January 2017 (UTC)
Yup. Brya needs to get off his high horse and if he feels that strongly about this issue, then he can go ahead and publish a new name himself. If a nobody like me can affect nomenclature, so can he. In the meantime he should respect the principle of no original research. Circeus (talk) 12:57, 28 January 2017 (UTC)
@Lymantria: the Notability policy determines if an entity deserves to have an item (which has not been called in doubt, in this case). Also, we have lots of items for homonyms which appear to fail these Notability criteria, but which have a page on some Wikipedia, so we are stuck with them. We need a way to structure items on homonyms, and it seems weird to invent new criteria of our own (NOR?) when there is a lawbook (accepted worldwide) which was set up for this purpose. - Brya (talk) 12:00, 28 January 2017 (UTC)
@Scott Thomson: I agree that Wikidata should not try to create nomenclatural acts (like creating a new spelling for one of these homonyms), and that we can't anyway. In this case I see no reason to assume that either name should not be "available" (in the sense of the Code). There are lots of provisions in the Code which factor in "prevailing usage" (not always clear in their application), but I see no indication in the Code that it applies to homonyms. It is mentioned in Article 23.9, but there it is a guide to a taxonomist in declaring one of them a nomen protectum and the other a nomen oblitum, and it is this declaration which would be the nomenclatural act. On the other hand, there are ways a taxonomist can deal with homonymy, explicitly set out (Article 52.4, 52.5 and the appeal to the Commission). I see no reason to put Article 52.1 out of commission. - Brya (talk) 12:21, 28 January 2017 (UTC)
@Brya, a nomenclatural act is any declaration that changes the validity, availability or usage of a name. So apart from new taxa and spelling corrections etc, it also includes placing names in synonymy or homonymy. Like many "rule books" the code is written for those it is intended. I am not saying whether this is good or bad but it is what it is. It takes a lot of understanding, discussion and patience to understand it properly. I think it is admirable that you try to apply the code here. Just be wary of the implications of what you do. Always be explicit that you cannot make a nomenclatural act on here. This prevents people thinking you have, when you cannot. One of my objectives is a stable nomenclature, like all members of the ICZN. For a nomenclatural system to be usable it must be stable and clear in its meaning. cheers Scott Thomson (Faendalimas) talk 13:37, 28 January 2017 (UTC)
As you say a nomenclatural act is any formal declaration that establishes a name or changes the application of a name. Placing a name in synonymy is mostly a matter of taxonomy. Whether a name is a homonym is not determined by a nomenclatural act, but by the facts. It is the resolution of a problem with a homonym that takes a nomenclatural act. I see nothing in the zoological Code that says its rules are dependent on action by a zoologist for them to take effect: that would be really weird, anyway. - Brya (talk) 16:32, 28 January 2017 (UTC)
@Brya: "Whether a name is a homonym is not determined by a nomenclatural act, but by the facts." That of course is not entirely correct. The pure homonymy may be a fact, judging the validity of the homonymous names is not. Lymantria (talk) 18:05, 28 January 2017 (UTC)
@Brya: yes it is the correction of the issue that is the act, I agree, which is also what I said, placing two names in homonymy is the act of recognizing the senior and junior homonym rendering the junior name invalid under the code. That is if it goes by priority, it does not always for a variety of reasons. The point is we cannot here designate one name as the senior homonym and one as the junior in an official by the code way, all we can do is effectively state the facts and leave it be. Obviously sending this information to specialists for those taxa is a good step towards resolving it as mentioned later by Neferkheperre in his post on the topic. Cheers Scott Thomson (Faendalimas) talk 00:40, 29 January 2017 (UTC)
I don't think TomT0ms bold approach is very helpful to solve the complexity of the problem. The problem we (an others) face is sketched in an article called Good and Bad Names published at the website of the Global Names Architecture (GNA). If you want to learn mor about GNA please read Towards a Global Names Architecture: The future of indexing scientific names (Q22117529). --Succu (talk) 17:04, 28 January 2017 (UTC)
A draft on a nomenclatural ontology for biological names is NOMEN. --Succu (talk) 18:33, 28 January 2017 (UTC)
Here is what we have: We at Wikispecies or Wikidata cannot generate nomenclatural acts here, because of the very nature of our copyright. Publications must be secure, and meet certain conditions. As we build our taxon data bases, we are going to find homonyms, since we are nearly universal in our mission. In my own field, Cirripedia, I have found three, one of which is family-group. As I have been entering reference citations for Zootaxa and Zookeys, I have discovered several more. By ICZN rules, new replacement names of family-group homonyms require application to ICZN for rulings. Genus and species do not. I checked Heterocheilidae in ICZN lists and indexes, and no results.
What to do: For listing in Wikispecies and Wikidata, until new replacement names are published, is to differentiate them, as NAME (Author), or NAME (Higher taxon group), or both. I try to find appropriate specialists and notify them. This has had success, and has helped build Wikispecies status. Neferkheperre (talk) 21:26, 28 January 2017 (UTC)
  1. As agreed on from the first, but raised again any number of times, Wikidata cannot publish nomenclatural acts. The problem cannot be fixed here.
  2. As also agreed on from the first, but raised again several number of times, it is possible, and recommended, to contact experts on the group in question and draw attention to the problems found. Hopefully, they will take formal action (by my count, I am up to three replacement names that are in press).
  3. @Scott Thomson: whether homonymy exists or not, and what is the senior homonym and what is the junior homonym, each is a matter of objective fact, ruled by the Code. It helps if somebody puts it in print, but that does not alter the facts. Hopefully, when somebody does go into print, he will at the same time also take steps to resolve the issue, which would be a formal act.
  4. @Lymantria, it is indeed important that both names are formally established: if the elder name is a nomen nudum or otherwise not formally established, there is no homonymy. But determining if a name is formally established does not (normally) require a formal act, but again is a matter of objective fact. Indexing centres make such decisions routinely all the time: "this is not a formally established name, as it fails the [...] requirement"; "that one is all right" (and, no, such decisions by indexing centres are not formal actions). Admittedly, there are grey areas, especially with old names recently dug up, that are dubious and have to be referred to higher authority, but these are very much the exception.
  5. As pointed out by Neferkheperre, homonymy in family-group names in zoology is special. Article 55.3.1 requires that such cases are submitted to the Commission for "a ruling to remove homonymy". If such a case is submitted, the Commission is likely to alter, "emend" the spelling of one of these names (likely the least used name), so as to remove the homonymy (example of case). This means that HETEROCEILIDAE is not like other homonyms, but rather is like Schroedinger's cat: one of these names is going to have a different spelling, but we can't know which, and we cannot provide a definitive spelling for either of these names. - Brya (talk) 06:29, 29 January 2017 (UTC)
Brya, the problem is that you refuse to realise that from the point of view of Wikimedia projects, whether or not these names are correct according to the code is completely irrelevant. As far as I am concerned, an edit such as this basically qualifies as nothing short of vandalism ("Blanking: The removal of most or all of a legitimate piece of information", information which is this case is pretty much completely unaffected by the fact the name is not correct), and if you did anything remotely similar on Wikispecies you'd land yourself with warnings, and I must assume you know this because you ave not attempted to remove any of that information from the Wikispecies page. Circeus (talk) 11:04, 29 January 2017 (UTC)
I am sorry to hear you feel that stuff copied off the internet automatically is "a legitimate piece of information". Any database that takes itself seriously is concerned about data quality. - Brya (talk) 17:35, 29 January 2017 (UTC)
Circeus, I think your judgment as vandalism is a little bit to harsh. Wikidata is full of automated processes (bots, WIDAR users, ...) and homonymy is the source of multiple errors done in the past by interwiki bots and by careless users done in the present. It's easier to detect hemihomonyms (i.e names handled by different Codes) than homonyms related to the same Code. In a highly automated environment like GBIF it's not unusual the wrong parent taxon is selected as in Heterocheilus Diesing, 1839. Bryas approach is not perfect, but I think he tries to reolve that kind of errors. --Succu (talk) 20:21, 29 January 2017 (UTC)
My apologies. I was assuming the data was referring to the correct higher taxon when Brya made the change (and of that were the case I would still consider such an edit vandalism, FWIW). his approach still tends to be ham-handed (I just stumbled on an edit on Wikispecies that is just wrong owing either to incorrect terminology or misunderstanding of the relevant code).
Fighting off the occasional bad data existing in improted database can be a hassle. Speaking of, is tehre a property to mark that information from a widely used database (GBIF, ITIS, plantlist, WoRMS...) is incorrect and should not be used to "fix" a wikidata property? Circeus (talk) 00:05, 30 January 2017 (UTC)
Circeus, I do not know why Succu is explaining Brya's action as revert of bad edit done by a bot or careless user. The information is correct and it was added by hand. On Commons Where I usually work at that kind of edit would also be considered vandalism. --Jarekt (talk) 04:09, 30 January 2017 (UTC)
Circeus, if you want to make false accusations about what happens at Wikispecies, you should do so at Wikispecies, not here. - Brya (talk) 05:08, 30 January 2017 (UTC)

Recent changes for claims involving a particular property

Is there any way to extract a list of recent changes in claims that involve a particular property

-- eg all recent changes to claims involving Cooper-Hewitt Person ID (P2011)

but without including any other changes on items that include a P2011 ?

I feel this is something that could help projects to keep a closer eye on changes involving their key properties, to monitor who has been changing what, and why. Jheald (talk) 14:12, 27 January 2017 (UTC)

User:Yair rand/DiffLists.js could be useful. However, it is a user script which needs to be installed on a per-user base by adding mw.loader.load('//'); to Special:MyPage/common.js. After “installation” it provides additional filter functionality in Special:RecentChanges, Special:Watchlist and Special:Contributions. —MisterSynergy (talk) 15:00, 27 January 2017 (UTC)
User:Pasleim/Lost Values could be useful. Matěj Suchánek (talk) 15:39, 27 January 2017 (UTC)
These look great, really useful. Unfortunately, Pasleim's page seems not to have been updated since April last year. @Yair rand:'s script looks to be exactly what I was looking for; but I can't seem to get it to show me changes earlier than the latest minute -- I was really looking for changes over the last month, or even longer. Is anybody else having this problem? Jheald (talk) 14:39, 29 January 2017 (UTC)
I'm guessing that what's happening is that Yair rand's script can only show me (some of) the 500 most recent edits, which it then filters. Whereas to do that kind of filtration over all the edits in the last month (eg to answer @Multichill:'s question above about who has been adding sitelinks to Commons), may be beyond what can be achieved with a client-side script, and would need something running server side. Jheald (talk) 14:45, 29 January 2017 (UTC)
@Jheald: Recent additions and changes of P2011, using Special:RelatedChanges. Unfortunately, it does not include edits that completely remove P2011 statements from the item, though. You might be able to use a {{Wikidata list}} to track removals. --Yair rand (talk) 21:17, 29 January 2017 (UTC)
@Yair rand: Very, very nice. Thank you. So what I needed to do was to check the "pages linked to the given page instead" box, when I clicked on "related changes"; and also to limit on Namespace:Main to get further back in time. Also, with the explicit link, I change from asking for the last 500 edits to eg the last 5000 edits to filter, eg for a property that has seen more updates recently. Plus the colour-coding is very nice. All in all a very nice tool. Thank you very much.
And, in answer to @Multichill:'s question, about sitelinks being added to Commons, it does seem that at least for sitelinks added in the last hour or so (most recent 10,000 edits in the Recent Changes list), they do seem to have been being added by a variety of editors rather than a bot. (Though of course there might have been a bot adding sitelinks at some time in the past). They do seem to be being added reasonably sanely: Category <-> Category links for the most part where the article-item is a class, Commonscat <-> Article item links where it is an instance. Jheald (talk) 22:14, 29 January 2017 (UTC)

Cleveland State University

Historical community information should be included for Cleveland State University (OH). Viking Hall and other dormitories were important living communities for students prior to the school's massive renovation. These buildings are a part of CSU's rich campus history as well.  – The preceding unsigned comment was added by 2601:548:c100:499b:242b:c054:333e:817 (talk • contribs) at 16:20, 27 January 2017‎ (UTC).

You are feel free to edit Q1100801. --Liuxinyu970226 (talk) 12:02, 28 January 2017 (UTC)

Property-like items

There are these four strange items: PictoRight registry (Q27827683), Q27163421, Q26465959, Q24575428. They look like a property, but they are items in the main namespace. Do we need them for a particular application, or should they be deleted? —MisterSynergy (talk) 12:02, 29 January 2017 (UTC)

I presume the persons who created them just didn't know the proper process for getting a property created. If they are not being used (and I don't think any of them are being used in any claims of other items) I say delete them. (It would be nice if the Wikidata software enabled certain properties to be declared as metaproperties only, i.e. not usable on normal items, and then if it blocked any attempts to use such properties on normal items...) SJK (talk) 12:10, 29 January 2017 (UTC)
How about pinging the creators? ChristianKl (talk) 14:13, 29 January 2017 (UTC)
@Hannolans, Спасимир, Mozel W., Ping08:. --Epìdosis 14:24, 29 January 2017 (UTC)
Maybe PictoRight registry (Q27827683) could be linked in PictoRight ID code (P3361) through subject item of this property (P1629)   Done Not very useful for now, though. The rest, well... Strakhov (talk) 14:30, 29 January 2017 (UTC)
You can delete this item Q24575428. He's not needed. — Ping08 (talk) 14:38, 29 January 2017 (UTC)
Yes, can be deleted I was not aware. --Hannolans (talk) 15:06, 29 January 2017 (UTC)
Yes, Q27163421 is to be deleted. Sorry for the inconvenience and thank you!--Spasimir (talk) 15:21, 29 January 2017 (UTC)

Thanks for all your replies. The first item is now linked to the property which has meanwhile been created, the other ones are proposed for deletion. If you need real properties for these databases, please visit Wikidata:Property proposal and request it there. —MisterSynergy (talk) 17:33, 29 January 2017 (UTC)

Institution of Mechanical Engineers: Chinese Wikipedia article

Could a Chinese speaker please check whether Institute of Mechanical Engineers (Q1569225) is the same as Institution of Mechanical Engineers (Q15051986)? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:49, 27 January 2017 (UTC)

@GZWDer:. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:57, 27 January 2017 (UTC)
The Chinese article seems to talks about a different organisation with headquarters in Hong-Kong. I could not access the official website given, but the version is here:
Koxinga (talk) 23:28, 27 January 2017 (UTC)
@Pigsonthewing:, unrelated. MechQuester (talk) 05:22, 28 January 2017 (UTC)
Since the zh label of Institution of Mechanical Engineers (Q15051986) was added by Cewbot, @Kanashimi: ^^ --Liuxinyu970226 (talk) 07:01, 28 January 2017 (UTC)


Thank, all. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:43, 28 January 2017 (UTC)

w:zh:食物浪費 modified. --Kanashimi (talk) 10:26, 30 January 2017 (UTC)

Great Officers of State

Should the Great Officers of State (Lord Chancellor (Q217217), Lord High Steward (Q510373), Lord Privy Seal (Q910308), Lord High Treasurer (Q944583), Lord Great Chamberlain (Q1798290), Lord President of the Council (Q943379), Lord High Constable of England (Q955642), Earl Marshal (Q1265164), Lord High Admiral (Q16153574)) be instance of (P31) or subclass of (P279) to Great Officer of State (Q1544356) ? SJK (talk) 06:56, 29 January 2017 (UTC)

Since Great Officer of State (Q1544356) is defined currently as subclass of (P279) position (Q4164871) and these are instances of government positions, they should also be instance of (P31) Great Officer of State (Q1544356). ArthurPSmith (talk) 13:54, 30 January 2017 (UTC)


The English label on The Hitchhiker's Guide to the Galaxy pentalogy (Q25169), as well as its description ("1979-1992 series of five books by Douglas Adams") and topic's main category (P910) (in English, "Category:Novel adaptations of The Hitchhiker's Guide to the Galaxy") make clear that it is about the series of books, but the en.Wikipedia article is about the whole franchise, including radio plays, television and film. I suspect (but cannot determine categorically) that the latter is true for the other interwiki links. How should this be resolved? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:36, 29 January 2017 (UTC)

Split the item in two? One for the book series and one for the franchise? --Yair rand (talk) 21:20, 29 January 2017 (UTC)
Well, of course. But which should the current item represent? How should the individual interwiki links be distributed, given that none of us can read all of the languages involved? Where should the existing fifteen inbound Wikidata links be pointed? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:54, 29 January 2017 (UTC)
This is the task of WP to decide which item represents the best the content of their articles. We just have to provide the items for the different individual concepts. Snipre (talk) 08:04, 30 January 2017 (UTC)
Perhaps. But on the day we split the item, what should we do with those interwiki links (they're not just for Wikipedias, BTW)? And how should we notify all the sister projects affected that we have split the item? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:32, 30 January 2017 (UTC)

P407 or P364 ?

Hi. This is me wishing to add a language in which a newspaper was written. Should I use language of work or name (P407) or original language of film or TV show (P364)? According to the label I'd favor the first one ("language of work or name") against the second one (original language of work -> "the original language a work/edition (...a translation) was translated from"). But reading documentation info of both properties... it's stated completely the opposite. Making life easier... both properties are currently proposed for deletion. Strakhov (talk) 03:00, 30 January 2017 (UTC)

language of work or name (P407) and original language of film or TV show (P364) are not proposed for deletion but are proposed to be merged into each other. --Pasleim (talk) 09:13, 30 January 2017 (UTC)
Yes, I supposed something like that. I guess the place for filing complaints and/or remarking inconsistencies and that stuff is the property-for-deletion/merging-request itself. Thanks, Strakhov (talk) 10:18, 30 January 2017 (UTC)

What are notable products?

I would appreciate some thoughts one what products could be considered notable and worthy of inclusion in Wikidata. I could imagine it would be beneficial to add many products but what are the community limits? If a product does not have a page on Wikipedia then the other criteria is whether it can be publically referenced, but that is true of most products through sites like Amazon, etc. Thanks! Pauljmackay (talk) 12:13, 30 January 2017 (UTC)

"I could imagine it would be beneficial". If you don't have any clear objective for the data you want to import, I think this is useless to import them. Snipre (talk) 12:42, 30 January 2017 (UTC)

Wikidata weekly summary #245

Internal identifiers

Are Wikimedia page outside the main knowledge tree (Q17379835) and Wikimedia internal item (Q17442446) the same? Jc86035 (talk) 11:20, 20 January 2017 (UTC)

Just say, my opinion is   Support merging both. --Liuxinyu970226 (talk) 04:16, 21 January 2017 (UTC)
  On hold We should take a look at the subclasses first. Matěj Suchánek (talk) 09:20, 21 January 2017 (UTC)
I think Wikimedia page outside the main knowledge tree (Q17379835) should be a subclass of Wikimedia internal item (Q17442446). Not every internal Wikimedia thing is a page, is it? --Yair rand (talk) 01:46, 23 January 2017 (UTC)
@Yair rand: Not every internal Wikimedia thing is a page [citation needed (Q3544030)] --Liuxinyu970226 (talk) 15:18, 24 January 2017 (UTC)
What about single user group as a concept? That's not a page but a group of users, isn't it? Matěj Suchánek (talk) 15:45, 24 January 2017 (UTC)
@Matěj Suchánek, Liuxinyu970226, Yair rand: Tried repairing them by adding subclass of (P279)Wikimedia internal item (Q17442446) to Wikimedia page outside the main knowledge tree (Q17379835) and part of (P361)Wikimedia Movement (Q3568028) to Wikimedia internal item (Q17442446); not sure if subclass of MediaWiki page should remain for the latter item. Jc86035 (talk) 15:55, 31 January 2017 (UTC)


Eye of the Storm (Q28549977)

There is a link to a commons category. Is it notable enough to not send to RFD? MechQuester (talk) 21:50, 30 January 2017 (UTC)

There is no roller coaster of this name in this park. Jérémy-Günther-Heinz Jähnick (talk) 21:59, 30 January 2017 (UTC)
@MechQuester: usually, I would say yes, but here the link is not valid as there is no category on Commons, so no.
@Jérémy-Günther-Heinz Jähnick: Indeed, there seems to be a confusion between Six Flags New England (Q846548) and Kentucky Kingdom (Q369789) (also known as « Six Flags Kentucky Kingdom » and where « For the 2017 season Kentucky Kingdom has announced Eye of the Storm » as stated on the en.wp article).
Cdlt, VIGNERON (talk) 23:16, 30 January 2017 (UTC)
I do more searchs, and I found this. This attraction is not considered as a roller coaster by RCDB. I follow the new about amusement park, and it is a new model duplicated in many parks. Jérémy-Günther-Heinz Jähnick (talk) 09:53, 31 January 2017 (UTC)

Property cemetery

Have we any place/page for names/titles of deleted properties? Sometimes reading of old discussions turns to a problem with several unknown. - Kareyac (talk) 09:00, 31 January 2017 (UTC)

Special:Tags appears busted

Can anyone get the page Special:Tags to appear? For me I get a 405 timeout.  — billinghurst sDrewth 10:49, 31 January 2017 (UTC)

This issue is reported on phabricator. --Pasleim (talk) 11:05, 31 January 2017 (UTC)

Please check property talk pages before mass importing data

I've just spent a considerable amount of time preparing to import values for AFL Tables player ID (P3547), as I said I would do, yesterday, in Property talk:P3547#Import, only to find another editor has already embarked on the same task.

I encourage all data importers to check property talk pages before embarking on a mass import; and to use those pages to announce their own intentions. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:49, 31 January 2017 (UTC)

More opinions needed for proposed property

Hi. We could use some more opinions over at this proposed property discussion. We seem to have arrived at the stage where we are repeating ourselves. Ijon (talk) 18:41, 31 January 2017 (UTC)


I suppose that all P31: musical term (Q20202269) should have some other class/superclass (too, or instead). How can we organize this? May be even to remove all P31:Q20202269 because they are too vague? --Infovarius (talk) 14:21, 27 January 2017 (UTC)

The very first one I checked had interwiki-links, and therefor languages connected. Those can not be removed for sure. Edoderoo (talk) 14:43, 27 January 2017 (UTC)
May be I was not very clear. I meant to remove the statement "P31:Q20202269" in such items, not sitelinks or items or anything else. --Infovarius (talk) 16:02, 27 January 2017 (UTC)
I would not remove it, as it is not incorrect. If you have a better specification, then feel free to replace them, but destroying info that is not wrong is not going to make it better. Edoderoo (talk) 17:33, 27 January 2017 (UTC)
The problem is that this "not incorrect" but useless value prevents most users from filling the more useful value. --Infovarius (talk) 18:58, 31 January 2017 (UTC)
None of the items using that statement that I've seen are using it accurately. We rarely actually have items about terms, as opposed to the actual things/concepts themselves. All or almost all of these statements should be removed. --Yair rand (talk) 21:24, 29 January 2017 (UTC)