Wikidata:Project chat/Archive/2016/11

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

New order on the statement sorting gadget

Hello,

For those who use MediaWiki:Gadget-statementSort.js, we recently tried to reorder the statements to have a more meaningful order. However, if you have any suggestions or will to change this order, you can discuss on the talk page. We will modify the gadget again, considering your suggestions, in a few days.

Thanks, Lea Lacroix (WMDE) (talk) 15:28, 31 October 2016 (UTC)

@Lea Lacroix (WMDE): When this was first announced I asked for a supplementary gadget to sort claims within a statement. For example population (P1082) after point in time (P585) in qualifier. Have you had any progress in that field? -- Innocent bystander (talk) 17:12, 31 October 2016 (UTC)

Thanks for the question. I don't know, I'll get some information about that. Lea Lacroix (WMDE) (talk) 09:43, 1 November 2016 (UTC)

How to link a 3rd party wiki to an item (what should be the id)?

Ff a thirdparty (nonWMF) wiki has an article on subject that has an item on wikidata, what should be the id that connects it? I am thinking pageid of the article, since the article name can change in the future and in case of moving the article without leaving a redirect pointing to it, it would break the connection. I want to propose a property for such a wiki, but I am not sure about the longtime sustainability of articlename as identificator. Thank you for your advice. --Wesalius (talk) 19:10, 31 October 2016 (UTC)

All properties linking to a 3rd party wiki do currently use the article name as identificator (P:P839, P:P1282, P:P696, P:P2000, P:P2537). But you're right that the pageids would be more stable than the article names. --Pasleim (talk) 08:49, 1 November 2016 (UTC)

I just filed a property proposal with pageid as the proposed value. --Wesalius (talk) 09:07, 1 November 2016 (UTC)

As a fallback there’s still described at URL (P973), but proposing for a specific property first is a good idea. —MisterSynergy (talk) 09:20, 1 November 2016 (UTC)

Add at least one label to newly created item based on the sitelink it was created with

Maybe there are some bots doing such a task. But sometimes I encounter items that occur in logs only under their Qid and when I examine them, they contain a sitelink to wikipedia/wikisource article which has a descriptive name. I think there should be a bot/script/service that would automatically add the item label based on the name of the sitelink article. I think we can all agree that seeing a labeled item in a log is superior to seeing "meaningless" Qid. But maybe there would be a downside to this, which I dont see, what would it be? --Wesalius (talk) 21:54, 31 October 2016 (UTC)

The label will only show up in the language for which it is added. There are plenty of items out there which have a label & description in them, but not in a language that I use, so they show up only by their Qid. Imagine what this is like for editors of Wikidata who do not read English. But besides that, there is a problem with bots adding Wikisource labels, or else we'll get labels like: The North American Review/Volume 151/Issue 406/Tolstoï and the "Kreutzer Sonata", since Wikisource makes extensive use of subpages to group works published within larger volumes. --EncycloPetey (talk) 01:33, 1 November 2016 (UTC)

The label will only show up in the language for which it is added if there is any label at all, thats the point... However I cant find the item that I wanted to give as an example. So probably I just didnt check the labels in other languages and this doesnt need to be solved. --Wesalius (talk) 05:48, 1 November 2016 (UTC)

Labels can definitely help to find items. One main reason that bot don't add them automatically is because sitelinks are always capitalized whereas labels should begin with a lowercase letter except for when uppercase is normally required or expected. Bots can't easily determine if uppercase or lowercase is expected, that's why often no label is added. However, if there is consensus that wrongly capitalized labels are better than no labels, one could change the bots' behavior. --Pasleim (talk) 08:37, 1 November 2016 (UTC)

See also related phab:T148762. Edgars2007 (talk) 09:00, 1 November 2016 (UTC)

I add labels but only for specific cases for the problem of upper/lower case. --ValterVB (talk) 10:53, 1 November 2016 (UTC)

Maybe we could set up a help/project-page to inform each other of the different rules in different languages? -- Innocent bystander (talk) 12:56, 1 November 2016 (UTC)

Birthday: Wikibase documentation and artworks

Hello all,

In development, writing documentation is as important as building new features. That's why @Ladsgroup: and @Jonas Kress (WMDE): worked on the technical documentation about Wikibase for the birthday! Both PHP and JS scripts are now described in the official Wikimedia documentation repository. Thanks also to Ashar for the release.

The story of the day is written by @Shonagon: who worked a lot on artworks, providing Crotos that displays artwork pictures from Commons when you enter a keyword, using Wikidata of course. You can read the story (in French) and discover his other tools. Thanks Shonagon!

Cheers, Lea Lacroix (WMDE) (talk) 10:23, 1 November 2016 (UTC)

Allow multiple interwiki links to the same page

In the new system for managing interwiki links, a single page can only accept one incoming link per wiki. This is probably fine in most cases, but not when there really are two or more pages that should be linked to the same page in another wiki. Here are a few examples of links from en: to ja: that should be included, but cannot under the current system:

• Omamori → お守り (disallowed because of link from Amulet)
• Japanese martial arts → 武道 (disallowed because of link from Budō)
• Inn → 旅館 (disallowed because of link from Ryokan (inn))

Are there any plans to fix this problem? -- 86.30.170.56 13:35, 1 November 2016 (UTC)

This "problem" is well-documented. Review Help:Bonnie and Clyde problem. --Izno (talk) 14:53, 1 November 2016 (UTC)

Ah, I thought it might be documented somewhere. Very reassuring to know that it's a "problem" and not a problem. -- 86.30.170.56 15:16, 1 November 2016 (UTC)

Open commenting on WMF Seeking Additional Resources for Structured Data on Commons

The Wikimedia Foundation in cooperation with Wikimedia Deutschland, has a unique opportunity to potentially secure additional resources to expedite development work on Structured Commons. We would like feedback on a 3 year plan that describes accelerated software development if these resource becomes available. We would like to invite you to participate in a conversation at: this page which provides an overview of that proposed timeline. We look forward to your comments and thoughts.

Joseph Seddon and Alex Stinson 22:30, 26 October 2016 (UTC)

With a lot of Wikidata effort seeming to go into supporting Commons and Wikidictionary, I'm wary whether there's enough effort payed to the core of Wikidata. Features like proper constraint management via properties and new data types such as the time of day take effort to be implemented. ChristianKl (talk) 09:51, 29 October 2016 (UTC)

It would be welcome when we bring support for links and red links to Wikipedia. It will bring more people to contribute, it will bring much more quality to both projects and it is a hell of a lot easier than continuously expanding what we do without maintaining and improving quality.

To be honest, Wiktionary support will not be great when we do not link it to labels. This is something that has been argued a lot and the answers are only based on feelings and not on facts. Thanks, GerardM (talk) 12:48, 29 October 2016 (UTC)

yes, not a zero sum game, by playing nice with other projects you increase value and support of wikidata. if they have outside support for this, then the fungible internal funds can be spent on other priorities, i.e. win-win. an a lot of discussion is based on hunches, not facts. need more pilots with a/b testing to increase quality. everyone is trying to hit home runs, now that they have seen a grand slam. Slowking4 (talk) 17:14, 2 November 2016 (UTC)

electromagnetic units

can someone please check the wikilinks, description and aliases of

• electromagnetic unit (Q1228244) summary articles about electromagnetic units in general
• electromagnetic unit system (Q27666180) EMU, a variant of centimeter–gram–second system of units (Q26240)
• International System of Electrical and Magnetic Units (Q1496383) differently scaled EMU that became part of MKSA system of units (Q730906) and later International System of Units (Q12457)

to see if the current structure makes sense and correct the Japanese wikilinks, labels and aliases. Maybe one should also merge International System of Electrical and Magnetic Units (Q1496383) and MKSA system of units (Q730906), but it has seperate articles in Norwegian.--Debenben (talk) 22:10, 1 November 2016 (UTC)

400,000,000th edit

We are there since this edit. Matěj Suchánek (talk) 15:03, 2 November 2016 (UTC)

Made by User:Charles Matthews. Well done! Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:54, 2 November 2016 (UTC)

Ah, an unexpected honour. Charles Matthews (talk) 16:56, 2 November 2016 (UTC)

Errors in date of birth and date of death

Hoi, many rulers have as their date of birth the date of the start of their reign. Is it possible to have a query for people who were younger than say 20 years and lived before the year 1500? With such a query we have a whole lot of issues ready to identify and solve. Thanks, GerardM (talk) 15:32, 2 November 2016 (UTC)

SELECT ?item ?itemLabel WHERE {
  ?item wdt:P31 wd:Q5 ;
      p:P569/psv:P569 [
        wikibase:timePrecision ?precb ;
         wikibase:timeValue ?birth ;
      ] ;
      p:P570/psv:P570 [
        wikibase:timePrecision ?precd;
        wikibase:timeValue ?death ;
      ] .
    FILTER(?precb >= 9). # precision of at least year
    FILTER(?precd >= 9). # precision of at least year
    FILTER(year(?death) - year(?birth) < 20) #younger than 20 years
    FILTER(year(?birth) < 1500) #lived before 1500
	SERVICE wikibase:label { bd:serviceParam wikibase:language "en" }
}

Try it!

There are indeed many dates which need to be fixed. --Pasleim (talk) 16:27, 2 November 2016 (UTC)

Wikidata import guide for muggles

Hi all

I've been working on a draft of a guide for muggles like me to learn how to prepare a spreadsheet up to the point where they can then request a bot import of the data. Here is a draft, there are still a lot of issues but I think its got most of the process mapped out. The sections on structuring data and structuring the spreadsheet still need a lot of work.

Thanks

--John Cummings (talk) 16:07, 2 November 2016 (UTC)

excellent, i would call it "proposed data import wizard". similar to bot request. may need to recruit train a team to interact with muggles. muggles' data are data too. Slowking4 (talk) 17:23, 2 November 2016 (UTC)

My little gift for Wikidata !

Hi people, I just finished a first functional version of Module:PropertyPath. It comes with a little showcase template to demonstrate how to use it, {{Show Path Items}} PropertyPath is basically an implementation of the SPARQL PropertyPath, that works in lua, extended with an operator to follow qualifiers value. An example : the grandfathers and grandmothers of a god :

• {{Show Path Items|Q105019|(P22|P25)/(P22|P25)}}

• Zeus (Q34201)   
• Hera (Q38012)   
• Zeus (Q34201)   
• Uranus (Q79999)   
• Dione (Q199923)   

(@Zolo, Izno:) author  TomT0m / talk page 18:35, 2 November 2016 (UTC)

That's pretty cool! Thank you very much :) Lea Lacroix (WMDE) (talk) 08:54, 3 November 2016 (UTC)

Birthday: Graph builder and random item

graph builder demo

Hello all,

Only two days of birthday announcements remaining, and today's a pretty cool one :) We integrated the graph builder tool into the query service. It's a lightweight tableau-style interface for visual analysis, built on Vega-lite, that's allows you to create and personalize your graph, based on the query results. Once it's done, you can also integrate it in Wikidata or Wikipedia, using the graph extension!

Example:

For now, the tool encounter some problems in Firefox, but works well with Chrome or Chromium. We'll try to fix is as soon as possible. Thanks a lot to @Jonas Kress (WMDE): and @Smalyshev (WMF): for your work!

Check out the demo video to see how it works, and try it with your favorite query ;)

The story of the day has been shared by Denny. Everything starts when you click on the "random item" link...

Check also the nice gift from @TomT0m: just above!

Cheers, Lea Lacroix (WMDE) (talk) 09:28, 3 November 2016 (UTC)

UMMC (Russian)

Uralmekhanobr (Q27577629) appears to be either the same as Ural Mining and Metallurgical Company (Q1349918) (in which case they should be merged); or related to it (in which case the relation should be made clear and the duplicate values such as Facebook username (P2013) & VK ID (P3185) resolved). Can a Russian speaker advise, please? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:56, 3 November 2016 (UTC)

They are different, mistakes in duplicate values corrected. - Kareyac (talk) 17:14, 3 November 2016 (UTC)

Birthday: 10 amazing queries, new chart visualizations and more documentation!

Hello all,

This is the second day of our birthday-celebrating week, and I have some very cool news for you.

New area chart view available on the Query Service

The Wikidata Query Service already had very nice ways to display the results, as graphs, bubble chart and map. Now we have improved this with four new types of charts: line chart, bar chart, scatter chart and area chart!

This is already deployed on the Query Service and you can try it directly by using one of the example queries : line chart, bar chart, scatter chart, area chart. Some of these views are animated, and you can play with it by moving the cursor over the chart, clicking on the legend labels to filter...

Thanks a lot to @Jonas Kress (WMDE): who worked very hard on this so these new features are perfectly ready for the birthday! If you have questions or find a bug, please ping one of us.

We also worked on documentation: have a look to the page presenting all the types of views that WDQS can provide. Don't hesitate to improve it or translate it in your language! Again, thank you Jonas for spending time on improving the doc.

Speaking about documentation, I need to present you this very cool guide to SPARQL and the Wikidata Query Service written by @TweetsFactsAndQueries:. We know that understanding how SPARQL works is not easy for beginners, and every help is welcomed to improve our actual documentation :) Feel free to give your feedbacks about this work. Thanks WikidataFacts for this birthday present!

Last but not least, we have a story about queries. @Jens Ohlig (WMDE): and @Cornelius Kibelka (WMDE): wrote a blog post called "10 cool queries for Wikidata that will blow your mind. Number 7 will shock you." (also available in German) That's not on BuzzFeed (Q5003490) but on the WMDE blog and it's a nice way to introduce newcomers to queries. Thanks for sharing!

That's all for today, I let you enjoy your Sunday and I'll be back tomorrow. By the way, Sunday is a good day to try new queries: if you want to share yours, ask for help on a difficult one or give some advices, join us on Twitter with the hashtag #SundayQuery ;)

Cheers, Lea Lacroix (WMDE) (talk) 10:17, 30 October 2016 (UTC)

That's great! Nevertheless since this new visualizations were deployed, "Image Grid" became unselectable/unavailable for me. Strakhov (talk) 21:57, 30 October 2016 (UTC)

Thanks for reporting this, we'll have a look as soon as possible. Lea Lacroix (WMDE) (talk) 10:36, 31 October 2016 (UTC)

@Lea Lacroix (WMDE): You're welcome. By the way, "Image Grid" was fixed but then "Map" went off, at least for me. Strakhov (talk) 23:05, 3 November 2016 (UTC)

@Strakhov: Indeed! Seems to be a curse with the list ;) We're working on it. Lea Lacroix (WMDE) (talk) 09:53, 4 November 2016 (UTC)

The Gupta empire and relevant Wikidata data

When you read about empires and dynasties that have come and gone, it is often hard to understand the data. Much of what Wikipedia holds does not help us understand really what is said, sometimes it needs additional data to make sense of it.

• The Gupta empire has a "Gupta year". It is mentioned in a few articles but what kind of calendar is it? What is it start point.
• The Gupta emperors had a succession, I am busy adding them at this time but they had sons and daughters that were relevant in their own right. When there is enough data I add them as Wikidata items. However, when battles and wars are mentioned, the same is possible. This may be linked to maps, this may be linked to rulers and people from other countries. All this will help understand the history of the Indian peninsula.
• Data, particularly similar data that is dated can be compared. When a war is said to have happened and a specific person is mentioned.. Does it fit his or her timeline?

The question therefore is,

can we have a tool that calculates dates from one system to the next, including Gupta.

can we have queries that show maps that include for instance the "Gupta empire", can we have them sorted by date?

with those maps, can we include the other countries that are shown on the map and given the date, can we show its rulers?

Thanks, GerardM (talk) 10:08, 1 November 2016 (UTC)

• @GerardM: not sure this is relevant, but some of the complexity regarding dates in wikidata has been reviewed at Help:Dates (which seems to be very much still in progress). Maybe something to ask the developers about? ArthurPSmith (talk) 18:03, 1 November 2016 (UTC)

Now the date is assumed to be Julian. This is probably wrong because it is not based on old Western literature but old Indian (Gupta) literature. It is important to have the tools to verify dates. Thanks, GerardM (talk) 07:04, 2 November 2016 (UTC)

• How many calendars are there that should be supported? ChristianKl (talk) 20:00, 3 November 2016 (UTC)

• There are good reasons to support each calendar if only to display them for the original data. When we have an algorithm, we can even calculate them. As to how many, I do not know. Before last week I did not even know the Gupta calendar existed. Thanks, GerardM (talk) 21:11, 3 November 2016 (UTC)
• There are 169 calendars listed at w:List of calendars, the Gupta calendar is not one of them (at least by that name) suggesting the list is not a complete set. Thryduulf (talk) 11:22, 4 November 2016 (UTC)

‎Names of living things

We seem to have a problem with Lepidodes limbulata (Q13391494), whose label is "Lepidodes limbulata". It is linked to Wikipedia articles in five languages, each with the title "Lepidodes limbulata". The item also includes four external IDs, each linking to sites that use the name "Lepidodes limbulata".

Earlier today, I added taxon name (P225) -> "Lepidodes limbulata". This was soon removed, so I reverted, with the edit summary "no reason given for removal)".

I have since been reverted, with edit summaries including "see P31" and "if you do not understand the reason ask for help at Wikidata:WikiProject Taxonomy" - none of which actually give a reason for removing the property & value. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:46, 1 November 2016 (UTC)

So what?! If there’s someone reverting your edits and they obviously have a lot of expertise in this field, as Succu (hereby informed) has in this case, don’t always run into these unnecessary edit wars and conflicts with so many different users. —MisterSynergy (talk) 20:02, 1 November 2016 (UTC)

So none of the edit summaries give a reason for removing the property/ value. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:07, 1 November 2016 (UTC)

To be honest: such things happen to me as well, also often without “useful” edit summaries. If the revert was done by a reputated editor in the field in question, I either ask on their talk page (when I am really curious), or let them do the job. There’s no need to understand virtually everything here. —MisterSynergy (talk) 20:14, 1 November 2016 (UTC)

Since I'm not trying to "understand virtually everything here", your comment appears to be a straw man. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:22, 1 November 2016 (UTC)

You were trying, and you still are doing so. My suggestion was to just accept the revert as it is, and turn attention to something else than the usefulness of a revert edit comment. —MisterSynergy (talk) 20:34, 1 November 2016 (UTC)

Ad hominem? Or do you dispute my statement that I'm not trying to "understand virtually everything here"? And thanks for the suggestion; but no. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:39, 1 November 2016 (UTC)

Your modus operandi is reverting first. The hint was refering to instance of (P31) = unavailable combination (Q17487588). The proposal was - if you object handling things this way - ask at Wikidata:WikiProject Taxonomy. --Succu (talk) 20:18, 1 November 2016 (UTC)

The "proposal" was as I stated, above. Noted that you still do not give a reason for your removal of the property/value. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:22, 1 November 2016 (UTC)

If you need more information (given by instance of (P31) = unavailable combination (Q17487588)) why do you fear a question at Wikidata:WikiProject Taxonomy? It's a question of "how do we model things right" and not of your personal feelings and a single edit of yours. --Succu (talk) 20:35, 1 November 2016 (UTC)

"Fear?" Another straw man. Still no reason. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:37, 1 November 2016 (UTC)

straw man argument (Q912820)? Why not asking or reading the information at hand? Questioning details? Are you not curious why we model things this way? --Succu (talk) 20:43, 1 November 2016 (UTC)

Based on constrains, items with taxon name (P225) must have instance of (P31)=taxon (Q16521) or subclasses but item has 2 values that aren't subclass of taxon (Q16521). I don't know why, but I think that who has modeled the taxon know why but I trust them. --ValterVB (talk) 20:48, 1 November 2016 (UTC)

Lepidodes limbulata (Q13391494) is P31 unavailable combination (Q17487588), which is P31 [sic; subclass of] botanical name (Q281801), which is part of (P361) of botanical nomenclature (Q3310776), which is subclass of (P279) of taxonomy (Q8269924). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:26, 1 November 2016 (UTC)

So what's your conclusion, Mr. Mabbett? --Succu (talk) 21:32, 1 November 2016 (UTC)

That you again refuse to give a reason for your removal of the property/value. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:38, 1 November 2016 (UTC)

Your repetitive „arguments” are boring Mr. Mabbett! Mind to be more detailed? --Succu (talk) 21:44, 1 November 2016 (UTC)

And once again, Andy always will find something that suits his opinion. How long will the community allow this? Sjoerd de Bruin (talk) 21:31, 1 November 2016 (UTC)

And once again Sjoerddebruin attacks with an adhominem. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:38, 1 November 2016 (UTC)

So there is a huge paragraph on top of my message and yet we still don't know why this was a mistake to add the said statement. Just let us know and well, we'll know! Thierry Caro (talk) 09:33, 2 November 2016 (UTC)

Indeed, it's absurd and depressing. Please don't be such dicks to each others, guys. author  TomT0m / talk page 10:17, 2 November 2016 (UTC)

Indeed it is. One of the most depressing threads I've read in recent times. What is the problem in explaining the reason for the removal of Andy's edit? Why on earth are a number of people on this thread - Succu, MisterSynergy, others - more interested in being complete dickheads than they are in sharing a scintilla of knowledge about Lepidodes limbulata w.r.t. taxon naming. This is just playground bullying, pure & simple: you should be ashamed of yourselves. --Tagishsimon (talk) 22:34, 2 November 2016 (UTC)

Do you think insulting users as „dickhead” helps a lot, Tagishsimon?! --Succu (talk) 22:44, 2 November 2016 (UTC)

The thread is all about proper naming, Succu. Call a spade a spade, eh. We'd be really happy to sit at your feet and learn what the problem is, if you'd condescend to tell us. But you seem to prefer this arseholery. Sad. --Tagishsimon (talk) 22:47, 2 November 2016 (UTC)

You need a guru, Tagishsimon? I'm the wrong person for this job. But try to be nice (arseholery). --Succu (talk) 22:52, 2 November 2016 (UTC)

No, Succu. We need an explanation, ideally from you, as to what the issue with the taxon naming is. Your conduct is wholly shameful. --Tagishsimon (talk) 22:56, 2 November 2016 (UTC)

You can play rhetorical games all you like, but you merely show arrogance and contempt by doing so. --Tagishsimon (talk) 22:58, 2 November 2016 (UTC)

The heading "Names of living things" is misleading, this should be "Database artefacts". - Brya (talk) 11:45, 2 November 2016 (UTC)

I suggest that if no reason is forthcoming, the property & value - which are supported by the linked external pages; and by the citations in the linked Wikipedia articles - should be re-added. Wikidata works by reporting what external sources state, not through argument by authority. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:01, 2 November 2016 (UTC)

Wikidata works by reporting what reliable sources say. - Brya (talk) 12:10, 2 November 2016 (UTC)

Indeed. And these are reliable sources. Furthermore, no one - not you, not Succu, who removed the property and value, nor anyone else - has made any claim - much less provided any evidence - to the contrary. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:17, 2 November 2016 (UTC)

Lepidodes Guenée (1852) non Westwood (1841) (Q5226073) is a preoccupied name. --Succu (talk) 14:38, 2 November 2016 (UTC)

And what is it that you think are reliable sources? Those five Wikipedia pages? - Brya (talk) 18:02, 2 November 2016 (UTC)

As I wrote in the post to which you previously replied (emphasis added): "...supported by the linked external pages; and by the citations in the linked Wikipedia articles". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:11, 2 November 2016 (UTC)

So you are relying on bot generated content and not on curated data? --Succu (talk) 22:18, 2 November 2016 (UTC)

I'm relying on the reliable sources I described. You are relying on... what, exactly? Your authoritah? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:08, 3 November 2016 (UTC)

While I have absolutely no special knowledge on this issue, following the link for "Unavailable combination" indicates the name is not allowed as a taxon name, so presumably the "taxon name" property can't be used for it, hence the reverts. I'm not sure why Andy is insisting it is a taxon name if it isn't? ArthurPSmith (talk) 18:34, 2 November 2016 (UTC)

I refer you to my original post in this section. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:11, 2 November 2016 (UTC)

Your post (under the misleading heading Names of living things) starts with the words „We seem to have a problem”. Looks like the only one which has a problem with my removal are you. --Succu (talk) 22:35, 2 November 2016 (UTC)

No, Succu, you know that's not true. A number of people on this thread feel that it would be better if you explained your action than play the stupid game you're engaged in. --Tagishsimon (talk) 22:51, 2 November 2016 (UTC)

The really really sad thing is that there is a person here making a perfectly reasonable request for information, and being met with a wall of contempt. Since when was that acceptable on wikidata? And by the looks there might be issues with five or so articles on various language wikipedias, but those able to provide guidence choose not to. And now I have Succu taunting me on my talk page, apparently eager for more interaction even as he/she affects to be horrified by my intemperate language. Meanwhile here's the longest thread on this page (bar one) for a long time, all arising out of a small crowd choosing to pick on Andy rather than answer a simple question. Will any of you who are engaged in this explain why you are doing this? Why this is a reasonable way of going forwards? Because really, assuredly, this appears to be the most disgusting & deplorable thread I've read in a long time. I'm genuinely shocked the attitude being shown here. --Tagishsimon (talk) 23:28, 2 November 2016 (UTC)

Why should we keep taxon name (P225) -> "Lepidodes limbulata" after all, Tagishsimon? You'll find the answer in this (elongated by you) thread. --Succu (talk) 23:42, 2 November 2016 (UTC)

It is sad. The question is what exactly it is, that is sad. A user who has a reputation, to put it mildly, to push the envelope, is asking a question at the place where he can generate the most attention. It seems to me he could have answered his own question with just a bit of quiet thought, and a few clicks (for those not familiar with this kind of situation: the name Lepidodes limbulata has once been published for a butterfly, but the generic name Lepidodes for a butterfly cannot be used because there already existed a generic name Lepidodes for a beetle. It is not possible to have a name Lepidodes limbulata in a genus Lepidodes that does not exist. As the problem has been known for a century and a half, presumably the situation has been resolved long since, but as both the supposed beetle and the supposed beetle appear obscure, the solution is not immediately apparent).

        Could Wikidata handle this kind of situation better? Obviously, yes. This kind of situation is not at all rare, in the literature there probably are tens of thousands (may well be hundreds of thousands, maybe even more) of names that may not be used for a taxon, no matter what. Two centuries worth of scientists have laboured to sweep them safely out of the way. But with everybody going into databases everywhere, and promoting every bit of jetsam and flotsam (even though clearly marked as such) into valuable content, more of it than is comfortable ends up here. Bots have created a depressing number of Wikipedia pages for fictitious taxa (that do not exist, never have existed and never will exist). It would be better if we had a property like "not-a-taxon name" (or "scientific name that may not be used as a taxon name" or "inoperable scientific name") for names that can never be the name of a taxon.

        However, the way Wikidata handles this kind of situation is clear enough that the question as asked here need not have been asked, if just a bit of quiet thought had been applied. - Brya (talk) 06:17, 3 November 2016 (UTC)

It's definitely something that is worth beeing added. The choice of Wikidata is to store those informations of obsoete sources, and there is a rank "deprecated" just dedicated to handle them. You in the project should better think of a way to use them correctly :) Do you have databases of incorrect names ? author  TomT0m / talk page 07:13, 3 November 2016 (UTC)

We have been over that, multiple times. Using deprecated status is not the solution for this problem, but, often enough, the cause of it. As it is, "taxon name" is already badly overburdened with lots of things that only very dubiously can be placed there.

        As to the question of whether or not such things belong here, that is quite debatable. In the short term they hinder Wikidata in achieving its purpose. In the long term they may, or may not, have some use. But anyway it seems we are stuck with them. - Brya (talk) 11:47, 3 November 2016 (UTC

Where does the name come from (and potentially return) to Wikidata? It doesn't seem to be mentioned on the Web often. Is it listed somewhere as “occupied”, or as a butterfly name? Which source says it cannot be used? Can that source be cited for the statement? --AVRS (talk) 13:20, 3 November 2016 (UTC)

It's a consequence of Principle of Homonymy (Q7245146) (=ICZN acticle 52). For the publishing info see Lepidodes (Q27670263) and Lepidodes Guenée (1852) non Westwood (1841) (Q5226073). --Succu (talk)

I think you can add the reference used for the “real” genus name as the source for unusability of the unusable species name under the unusable homonym? It alone won't help much though, so link also to what you mentioned above (with quotations if helpful). --AVRS (talk) 14:01, 3 November 2016 (UTC)

Note that Lepidodes (Q27670263) was created (and Q5226073 was modified to mention it) by Succu on 2 November, after the exchange described at the top of this section. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:16, 3 November 2016 (UTC)

Note that we have Lepidodes (Q21365698) for a year now. --Succu (talk) 14:21, 3 November 2016 (UTC)

"where he can generate the most attention" Indeed. I prefer the community to discuss such issues and reach consensus, rather than rely on appeals to authority dependent on editors who are not willing to explain their actions. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:55, 3 November 2016 (UTC)

“It is not possible to have a name Lepidodes limbulata in a genus Lepidodes that does not exist.”

(Probably a useless question) What does not exist? --AVRS (talk) 13:16, 3 November 2016 (UTC)

May the name not be used for beetles because it has been misused in an authoritative source and is now considered occupied? Or may it not be used for butterflies, because it is incorrect? --AVRS (talk) 13:17, 3 November 2016 (UTC)

There does not exist a butterfly genus that bears Lepidodes as its correct name. The fact that a name of the same spelling was published already for beetles blocks any use of a later name with this exact spelling (there is a fairly similar Lepidotes for fishes). The relevant law-book provides some particular procedures to make exceptions in such matters, but I see no indications that such procedures have been used here.

The name Lepidodes could be used for a genus of beetles, if a taxonomist circumscribes a genus so that Lepidodes is the oldest name available for it. Looking at things, this apparently is not the case, but obviously this is a guestimate only. - Brya (talk) 18:14, 3 November 2016 (UTC)

I'm getting tired of Brya changing the heading of this section. Since we're discussing the names of living things; then "‎Names of living things" seems perfectly adequate. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:08, 3 November 2016 (UTC)

The item in question isn't about a living thing but about a label of a living thing. That's how instance of (P31) is set. ChristianKl (talk) 20:45, 3 November 2016 (UTC)

It's a strange heading under which Mr. Mabbett filed his statement, but it's only a heading. After being accused not willing to answer: Are there any questions open that need a clarification from your point of view, Mr. Mabbett? --Succu (talk) 22:06, 3 November 2016 (UTC)

Birthday: improving the graph view and a leap back in time

browse properties on graph view

Hello folks,

Today I'm glad to offer you another birthday present from @Jonas Kress (WMDE): regarding the Query Service. We improved the graph view so it can now browse the properties of items. Have a look to the video to see the result, and you can try for example on this query.

We also have a new story, today @Ash_Crow: who dug deep in his personal archives to tell how his history with Wikidata started! You can read it in English or in French. Thanks Ash_Crow for this archeology exercise ;)

You also can join one of the three birthday meetups this week:

• Wikidata Hackathon in Kanpur, India, November 3rd
• Meetup in Berlin, November 4th
• Wikidathon: editing and importing data in Rennes, France, November 5th

Cheers, Lea Lacroix (WMDE) (talk) 10:50, 2 November 2016 (UTC)

In the query above, I suggest adding #defaultView:Graph if the purpose is to show the graph view, and also I suggest to replace "w" with "s" (for "spouse" instead of "wife") ;-) Cheers! Syced (talk) 08:31, 4 November 2016 (UTC)

Birthday: You know you're a Wikidata editor when... you play a Wikidata-based game!

Everything is connected

Hello all,

This is the last day of our birthday week. Today, let's play a very nice puzzle game created by Denny. It's based on Wikidata items and properties, and the goal is to organize the elements in the grid and figure out how they fit together. You will see, it's not always easy... You can play the game here, discover other levels and even create you own levels.

Thanks to the contributors: @YULdigitalpreservation, ThurnerRupert, Tobijat, Bináris, Maxlath, Incabell: for adding levels!

The story of the day is You know you're a Wikidata editor when... inspired from ideas by VIGNERON, Harmonia Amanda, Ash_Crow... and modestly drawn by me. Feel free to translate it in your own language!

Tonight is the birthday meetup in Berlin. I can't wait too meet some of you there :)

Thanks to everyone who participated by offering a present, a story, or letting a message on the birthday page!

Cheers, Lea Lacroix (WMDE) (talk) 10:31, 4 November 2016 (UTC)

Hah, like the "...you're awaiting the constraint violation reports with impatience and fear". So I can call myself Wikidatian :) --Edgars2007 (talk) 11:36, 4 November 2016 (UTC)

Sculptures items

I asked a question to WikiProject Visual Arts, Wikidata:WikiProject Visual arts/Questions#Sculptures by Canova, but I received no answer for more than two weeks: could someone have a look? Thanks, --Epìdosis 11:52, 4 November 2016 (UTC)

Upload a dataset to wikidata

Hello, I am beginner of Wikidata editing, and maybe my question about uploading data (as opposed to inserting "items" one by one) is answered somewhere, but I could not find it.

I am currently having a dataset of around 400 Latvian rockbands, their participant names and the role in the band, (musical instrument they are playing), that I have collected manually. I would like to upload that data set to Wikidata, so an embedded network graph could be made trough Wikidata query, and users could add and contribute to that graph trough Wikidata.

Is there a way by using my dataset to:

1. -upload a list of bandnames as "items", being "instance of: bands" with "country of origin: Latvia"
2. -upload a list of musician names as "items", with "instance of: human" and for example "instance of: bassist", "instance of: female"
3. -upload a list of band-musicion pairs, creating for each of the bands "has part: (the musician name)"

Thank you if someone has the time to answer.  – The preceding unsigned comment was added by LinardsLinardsLinards (talk • contribs).

• QuickStatements might help you. The pdf on the right side has an introduction.
  --- Jura 22:46, 19 October 2016 (UTC)

• Big thanks Jura for the pdf. As I see to do this I should request a volunteer for creating a bot. Am I correct?
  --- LinardsLinardsLinards 01:58, 20 October 2016 (UTC)

Apart from what Jura already said, please take care to add your source when you add the data set. ChristianKl (talk) 20:03, 20 October 2016 (UTC)

Hi @LinardsLinardsLinards: — you might be interested in checking out the Wikidata:WikiProject_Music documentation of useful properties. For instance, "has part" is not always the best answer for bands; it's usually preferable to put "member of: (band name)" on the person, instead of the other way around. Sweet kate (talk) 23:26, 26 October 2016 (UTC)

Thanks @Sweet kate: for tip, Im checking the WikiProject Music right now. My plan is to query the data-set to get the connections between bands with the same participants to create graph, as in example here: https://www.wikidata.org/wiki/User:LinardsLinardsLinards should I add both "has part" and "member of: (band name)" to be sure, that the query does not miss some connections? --LinardsLinardsLinards (talk) 22:45, 29 October 2016 (UTC)

@LinardsLinardsLinards: Definitely do "member of: (band name)." The "has part" property is trickier — it's up for debate if it's really supposed to be used for people in this way. I'm not convinced you should do it. (In a general sense, on Wikidata we don't want to have to update two reciprocal properties independently of each other; it's preferable just to do one that can then be queried for the inverse. Otherwise, if the two reciprocal properties conflict, it can be a big mess. When everything is working as it should, queries don't "miss" things, and we'd only need each data connection or fact once.) It might be worth bringing this specific question up to a wider group for feedback. Sweet kate (talk) 01:49, 5 November 2016 (UTC)

Fallback type for property to item

Hi, I'd like to discuss a proposal which aims to avoid redundant properties differing only by types. Here is an example: everybody knows existence of start time (P580) and end time (P582), which need time values as input, however I need that those properties are used with items, so I proposed to create two new similar properties (supported by ChristianKl and Thryduulf) but Yair rand said that there are many other properties that has "duplicated" version with different input types. I'd like to discuss now if worth asking developers adding a "fallback" input type to properties which need non-item input.

What do you think? -- ★ → Airon 90 10:44, 29 October 2016 (UTC)

While having fallback types might be easier, or at least more natural, to humans, it would be more cumbersome for computers, map badly to other databases or other efforts for open structurally linked data like schema.org. Personally I prefer a few "duplicated" properties instead. Wikidatas primary audience is computers like the query service or templates in wikipedia Pajn (talk) 11:42, 29 October 2016 (UTC)

Most of the difficulty is already there due to the existence of the "no value" and "unknown value" options that exist for every datatype. Some of the values are not simple, but those can be filtered out under certain conditions. --Yair rand (talk) 15:16, 30 October 2016 (UTC)

See my comment on the proposal. We actually already have a property to use an item with a specified time interval - with a few reasoning or conventional rules this means that we could "transfer" the begin and end dates of this item to the statement. In your examples, the generation of pokemon has a creation date, which would means the statement is valid after this creation date with valid in period (P1264) . This would make the query a little harder to write, but a Mediawiki template could help in fashion of templates in Category:Partial query could help (I'll think of something). author  TomT0m / talk page 11:50, 29 October 2016 (UTC)

I responded to your comment on the proposal. Would love to be convinced, but currently I'm not. Pajn (talk) 12:14, 29 October 2016 (UTC)

I don't think fallback types are a good idea because they make it harder to write queries. ChristianKl (talk) 19:43, 4 November 2016 (UTC)

Nominated for X

Is there any way to model that a person has been nominated for a prize? Concrete use case: for the Nobel prize one can get historical nomination data until 1963 (there's an enforced gap of 50 years). I think it could be interesting to have this in Wikidata as well, to answer questions like "how many times has X been nominated before winning?" or "find person who has been nominated many times without actually winning". – Jberkel (talk) 09:00, 24 October 2016 (UTC)

Use nominated for (P1411). See example at Meryl Streep (Q873). --Edgars2007 (talk) 09:03, 24 October 2016 (UTC)

Thanks. I just ran a query and only 17 results for "X nominated for Nobel Prize of Literature", I'll see if I can complete this with the official data from the Nobel archive. – Jberkel (talk) 17:19, 24 October 2016 (UTC)

I've added the nominations for literature from 1901-1965. To answer my own question: List of authors unsuccessfully nominated for Nobel prize in literature. The winner is Ramón Menéndez Pidal (Q381953) with 23 nominations. – Jberkel (talk) 09:44, 26 October 2016 (UTC)

What have you done to make sure that you linked the right people? It seems like the Nobel Prize identifies people by an ID that's currently not in Wikidata. Before important big datasets like this it would make sense to identify the people with their IDs. I create a property proposal for that ID: https://www.wikidata.org/wiki/Wikidata:Property_proposal/Nobelprize_People_Nomination_ID ChristianKl (talk) 10:14, 29 October 2016 (UTC)

I've cross-referenced the birth and death dates, so it should be fairly accurate. – Jberkel (talk) 13:13, 5 November 2016 (UTC)

Dead people playing chess

There is some discussion at Property_talk:P1087 about dead people getting ELO ratings. It seems that some find this ok.
--- Jura 08:20, 5 November 2016 (UTC)

Q6164647 and Q7003701

Q7003701 and Q6164647 are about the same object, but there is some spurious "nl language" that does not match and that prevents a merge. Jo-Jo Eumerus (talk, contributions) 11:33, 5 November 2016 (UTC)

Try the merge gadget mentioned at Help:Merge and it should work.
--- Jura 11:39, 5 November 2016 (UTC)

Petscan: find values of a parameter

Hello. [1] With this way I can find all wikidata pages (of articles of a certain wikipedia category) that are using start time (P580). Is there a way to find not only the pages that are using P580 but also the values of this parameter in each page? Xaris333 (talk) 15:54, 5 November 2016 (UTC)

On PetScan choose as output format "PagePile". After you click "Do it!" a new browser tab will be opened. There, select "Item properties". This will open the new created PagePile in TABernacle. In TABernacle only add to the property list field "P580" and click "Show". --Pasleim (talk) 16:20, 5 November 2016 (UTC)

wordlist again

Long time ago I asked for a possibility to create a wordlist. Can you help me to run this query again? Is there a new definition for connect wikidatawiki_p;? Thank you, Conny (talk) 17:48, 5 November 2016 (UTC).

replace connect wikidatawiki_p; by use wikidatawiki_p;. --Pasleim (talk) 18:46, 5 November 2016 (UTC)

adoptive child / adopted child

Hoi, one is a subclass of the other but in my understanding they are exactly the same. I often ignore classification because it is all too often confusing at best convoluted at worst but this is in my opinion silly. Can I merge this? Thanks, GerardM (talk) 08:14, 5 November 2016 (UTC)

I can't find the item "adoptive". Can you please link it? --Pasleim (talk) 16:08, 5 November 2016 (UTC)

It is adoptive child and the two are linked. Thanks, GerardM (talk) 08:23, 6 November 2016 (UTC)

@Pasleim, GerardM: adopted child (Q25858158), however I can't find "adopted child" only adopted son (Q20746725) and adopted daughter (Q20746728). Thryduulf (talk) 13:39, 6 November 2016 (UTC)

There is also adopted child (Q25858157) --Melderick (talk) 13:52, 6 November 2016 (UTC)

Only adoptive daughter needs merging? Adopted son/daughter as subclass of adoptive child makes perfect sense to me. Being bold and merging the duplicate found by Melderick. --Azertus (talk) 14:51, 6 November 2016 (UTC)

Uncertain information

I just found out a suggestion of statements and source to add through the distributed game :

Sahan Dosova Sahan Dosova (Q1076093)   
사칸 도소바
Sachan Dosov
Сахан Досава
Досова, Сахан
サーハン・ドソヴァ

Elle a joué un rôle important dans supercentenaire et centenaire.
Elle est née le 27 mars 1879 à aoul.
Elle est morte le 9 mai 2009 à aoul.
http://www.newscientist.com/gallery/lessons-in-longevity/4
lieu de naissance: Kazakhstan
NOTE: This item already has statements about lieu de naissance.
This will add a new statement and source to the item!

(note that The Game use key event @Jura1: which leads to pretty weird french sentences : she played a key role in centenary, ... what would work is to create an item for its own 100th birthday).

She would be 130 ... which is unusual and unverified, according to the source. How could we handle this ? How to handle uncertain information, in queries for example ? author  TomT0m / talk page 15:49, 6 November 2016 (UTC)

No value

Hello. How can I, with quick statement (or another tool), to add broadcast by (P3301) with "no value"? Xaris333 (talk) 17:09, 6 November 2016 (UTC)

@Xaris333: Help:Statements#Unknown_or_no_values

I know that. I need a tool to do it more easy to a list of articles. Xaris333 (talk) 20:10, 6 November 2016 (UTC)

QuickStatements doesn't currently support novalue (I spent a while trying to get this to work a few months ago...). I'm not sure any tools do at the moment. Andrew Gray (talk) 20:42, 6 November 2016 (UTC)

Tool guessing items from a list of names

Hello. Is there a tool that would return a list of items from a given list of names in a given language? This counts the gender of people in the list, for instance, but what about just letting us know to which items those names were matched first? Thierry Caro (talk) 20:07, 6 November 2016 (UTC)

Timeout issue P131/P131*

Can someone help me understand why (a) "wdt:P131*" is more expensive than (b) "wdt:P131/wdt:P131*"? I always get a timeout on the simpler notation (a).

The following query uses these:

• Items: Hamburg (Q1055)   , protected area (Q473972)   

• Properties: located in the administrative territorial entity (P131)   , instance of (P31)   , subclass of (P279)   

  SELECT ?item ?itemLabel WHERE {
    ?item wdt:P131* wd:Q1055 .
    ?item wdt:P31/wdt:P279* wd:Q473972 .
    SERVICE wikibase:label { bd:serviceParam wikibase:language "de" }
  } LIMIT 1000
  

Try it!

The following query uses these:

• Items: Hamburg (Q1055)   , protected area (Q473972)   

• Properties: located in the administrative territorial entity (P131)   , instance of (P31)   , subclass of (P279)   

  SELECT ?item ?itemLabel WHERE {
    ?item wdt:P131/wdt:P131* wd:Q1055 .
    ?item wdt:P31/wdt:P279* wd:Q473972 .
    SERVICE wikibase:label { bd:serviceParam wikibase:language "de" }
  } LIMIT 1000
  

Try it!

Secondary question is, why are a few results repeated that often or how I can limit them, if that contributes to the cost. Thanks, --Aeroid (talk) 09:34, 4 November 2016 (UTC)

I don't know why but adding hint:Query hint:optimizer "None" helps in the first query but disables the second query. Adding DISTINCT sorts out repeated results.

SELECT DISTINCT ?item ?itemLabel WHERE {
  hint:Query hint:optimizer "None" .
  ?item wdt:P131* wd:Q1055;
  wdt:P31/wdt:P279* wd:Q473972 .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "de" }
}

Try it!

--Pasleim (talk) 09:55, 4 November 2016 (UTC)

Relevant discussion: Wikidata:Contact the development team/Archive/2016/09#SPARQL: (wdt:P171)* Matěj Suchánek (talk) 21:18, 5 November 2016 (UTC)

@Smalyshev (WMF): I'm observing a drastic performance loss for queries like that using (wdt:P171)*. Earlier I could execute more than 100 queries per minute like this. Nowadays the rate is less than 10 queries per minute. --Succu (talk) 17:34, 7 November 2016 (UTC)

This seems to be known issue, reported upstream here: https://jira.blazegraph.com/browse/BLZG-2089 and hopefully fixed once we upgrade to next release (hopefully early Jan) --Smalyshev (WMF) (talk) 21:58, 15 December 2016 (UTC)

Results of RfC about Wikidata in WP:fr

One month ago the French Wikipedia started a RfC about the use of WD in WP:fr. This was the second part of a very long RfC which started one year ago. The first part dealt with the possibility to use data from WD in the different parts of an article. The results of the first part was the permission to use data from WD only in the infoboxes, in the tables and other graphs. No use in the main core of the article was permitted.

The second part was focused more on how the data can be used in the authorized parts. The results are the following: all data from WD can be used with or without sources, data from WD have to be identifiable, links to WD have to be identifiable, contributors who introduced data from WD have to check data consistency with data already present in the article, data from WD have to be imported through a template in WP and local data have the priority over WD data. So the use of WD in WP:fr is not limited by any strong policy.

However this RfC showed some strong oppositions to the principles of WD: first the lack of trace of the WD changes in the article history (this is not a problem for Commons media but in the case of WD this is a huge problem), strong fears about a possibly biased external point of view in case of disputed data choices, and complexity when handling data import using Lua code which is not so common for most contributors.

Link to the RfC here. Snipre (talk) 14:21, 5 November 2016 (UTC)

@Snipre: Merci, but I don't speak French. Can you tell me how justifiable you think these fears are? Do you share them? (Also, you may wish to post this to Wikidata:Bistro.) —Justin (koavf)❤T☮C☺M☯ 14:28, 5 November 2016 (UTC)

I don't think "lack of trace of the WD changes in the article history" is a principle of Wikidata. Whether or not a Wikipedia wants to show those changes isn't a decision that WD makes but that the Wikipedia in particular makes. Given that the French Wikipedia has their own Wikidata plugin there's nothing expect technical work stopping them from showing the changes in the article history. ChristianKl (talk) 17:28, 7 November 2016 (UTC)

Preposition property

Just checking: has preposition (Q8479428) ever been suggested as a property, such as for place names and objects? For example, English preposition in for country (Q6256) and on for island (Q23442). I searched around and couldn't find anything. It would be very useful for place names in many languages, where it could be necessary for the automatic generation of sentences such as born <date> <preposition> <place>. --Njardarlogar (talk) 20:54, 5 November 2016 (UTC)

No, there is no property for this. And problably it never will be, as it is property of word, not a property of entity. --Jklamo (talk) 10:54, 6 November 2016 (UTC)

There will probably be, but not until structured wiktionary will be up ! author  TomT0m / talk page 11:17, 6 November 2016 (UTC)

It's finally coming, so we could start think about some properties to create.
--- Jura 11:40, 6 November 2016 (UTC)

Well, if <place> is a country, then use "in", if <place> is "island", use "on". Creating items for prepositions and think it will work in different languages is to expect a lot. One of the hardest things for a Swede to learn when we read English is the prepositions. And mistakes about prepositions is the most frequent when we write and talk English. If you hear a Swede talk about "under the year 2015", he speaks perfect Swedish, but awkward English. -- Innocent bystander (talk) 13:32, 6 November 2016 (UTC)

@Jklamo: The distinction you are making does not immediately make sense to me. A preposition is a property of the entity given a language, just like a database identifier is a property of the entity given a database.

@Innocent bystander: The problem is that such rules are not fail-safe; there can be many exceptions. If I want to really trust the automatic generation, then it needs to be able to fetch the correct preposition from somewhere. --Njardarlogar (talk) 16:55, 6 November 2016 (UTC)

Wrong, it's not a property of the entity, it's a property of some denomination of the entity in some language. And there can be many, possibly with not the same linguistic rules. For example there is many work for penis (Q58)    in french. In popular language "

une bite

" has a female gender while "

un zizi

", a more childish word, has a male gender ... author  TomT0m / talk page 17:02, 6 November 2016 (UTC)

Is that any different from saying that a database ID is not a property of the entity, but a property of the database entry on the entity? --Njardarlogar (talk) 18:20, 6 November 2016 (UTC)

Yes. Usually an id by definition uniquely identifies a topic. It's not really a property of the entity by itself, it has his own section in Wikidata for example but rather a technical artifact to map wikidatas entity to their counterparts in over databases. In Wikidata, with a very few known exceptions like title (P1476) for artworks or female form of label (P2521) the names of things are done with labels. And you can put properties or qualifiers to labels. author  TomT0m / talk page 18:48, 6 November 2016 (UTC)

I (still) don't agree with the conclusion regarding semantics, but currently it is not possible to add qualifiers to regular labels. Will this change, and will it be possible to set Wiktionary items as regular labels? --Njardarlogar (talk) 09:05, 7 November 2016 (UTC)

@Njardarlogar: My guess, but you should look carefully at the proposal, is that there will be no changes in handling Wikidata items. Rather, you will be able to assign meanings (Wikidata items) to words, so eventually query the different lexeme that can have the meaning corresponding to the item. But probably not to assign a unique lexeme to an item - By the way it's not only that a word does not identify a meaning, as a meaning can have several ways to be expressed in a language, but also that a lexeme can have several meanings. So we're far from a 1/1 database id mapping. author  TomT0m / talk page 18:15, 7 November 2016 (UTC)

Where do I activate Flow ?

Tracked in Phabricator
Task T138310

Hello, Flow is not in Beta tab anymore...

After months of using it on other contributors' pages I was finally convinced that it would bring me something... and now, I can't find how to activate it ^^

Help, please ? --Hsarrazin (talk) 13:26, 6 November 2016 (UTC)

@Trizek (WMF):. Tpt (talk) 13:38, 6 November 2016 (UTC)

See Topic:Tcx28mfr2eq8ehog. --Stryn (talk) 13:41, 6 November 2016 (UTC)

Hello Hsarrazin,

As Stryn quoted, there is a problem on Flow activation. Flow is not available at the moment but the Beta function will be back in a few days: we have fixed the bug.

(Thanks Tpt for the ping!)

Trizek (WMF) (talk) 09:53, 7 November 2016 (UTC)

Entity usage exposure

Hello all,

During last months, we've been working on developing tools about Wikidata entities usage exposure on the Wikimedia projects. All the developments are now done, and you can find all the features documented here. Feel free to try and give feedbacks, and you can ping @Ladsgroup: for any question.

Cheers, Lea Lacroix (WMDE) (talk) 14:40, 7 November 2016 (UTC)

Thank you. Frwiki has just been informed about this. Thierry Caro (talk) 18:10, 7 November 2016 (UTC)

Changes to quantity precision handling

Hi all,

over the past years a lot of people have rightfully complained about how we are handling rounding and uncertainty in quantity values. Until now when entering 124 m it will be parsed, stored and displayed as 124 +/-1 m. People then often tried to change this to 124 +/-0 m in order to prevent the uncertainty from being shown. This is often incorrect. People also disagreed with the default of +/-1 instead of +/-0.5. We have sat down and discussed at length how to improve this situation and have now prepared two changes:

1. If no uncertainty is explicitly entered, we do not try to guess it and no uncertainty will be displayed. But if an uncertainty is given it will always be displayed even if it is +/-0.
2. In order to apply correct rounding (in particular for unit conversion) we still need to know the uncertainty interval. If no uncertainty was explicitly given we still will need to guess it. We have now halved the default uncertainty interval to be consistent with the rounding interval. So for example 124 would be treated as 124 +/-0.5 when applying rounding.

We now have a lot of quantity values that are marked as exact (+/-0) while their uncertainty is not explicitly known. With the above changes all of these will be shown with +/-0. We also have a lot of values that have the old default precision (+/-1) which prevents the new default from being used.

Therefor we suggest the following two bot runs:

• Remove +/-0 from quantity values.
• Remove the precision if it is equal to the old default precision.

Both should be applied to properties that represent measured quantities. We should avoid applying them to properties that represent conversions or similar exactly-defined values. We can prepare the bot for this.

This requires a breaking change to the API and JSON binding. Daniel will say more about that in a minute. We plan to make this change on the live system on November 15th. It can be tested on the beta before that. I will let you know when.

Cheers --Lydia Pintscher (WMDE) (talk) 14:37, 4 November 2016 (UTC)

Firstly a big thank you! Secondly, will asymmetric uncertainty be supported (e.g. 5.5+1.0-2.0)? Thryduulf (talk) 15:04, 4 November 2016 (UTC)

It is already supported in the data model. We have just not yet allowed entering/editing it in the user interface. It is still on the todo list to also allow it via the user interface. This change will not affect it but we're on the right track :) --Lydia Pintscher (WMDE) (talk) 15:12, 4 November 2016 (UTC)

thanks very much. accuracy of numeric data is different from least significant figure. if you have some guidelines on how to input uncertainty of numeric values that would be great. Slowking4 (talk) 18:15, 4 November 2016 (UTC)

BREAKING CHANGE: Quantity Bounds Become Optional

This is an announcement for a breaking change to the Wikidata API, JSON and RDF binding, to go live on 2016-11-15. It affects all clients that process quantity values.

As Lydia explained above, we have been working on improving our handling of quantity values. In particular, we are making upper- and lower bounds optional: When the uncertainty of a quantity measurement is not explicitly known, we no longer require the bounds to somehow be specified anyway, but allow them to be omitted.

This means that the upperBound and lowerBound fields of quantity values become optional in all API input and output, as well as the JSON dumps and the RDF mapping.

Clients that import quantities should now omit the bounds if they do not have explicit information on the uncertainty of a quantity value.

Clients that process quantity values must be prepared to process such values without any upper and lower bound set.

That is, instead of this

"datavalue":{
  "value":{
    "amount":"+700",
    "unit":"1",
    "upperBound":"+710",
    "lowerBound":"+690"
  },
  "type":"quantity"
}

clients may now also encounter this:

"datavalue":{
  "value":{
    "amount":"+700",
    "unit":"1"
  },
  "type":"quantity"
}

The intended semantics is that the uncertainty is unspecified if not bounds are present in the XML, JSON or RDF representation. If they are given, the interpretation is as before.

For more information, see the JSON model documentation [2]. Note that quantity bounds have been marked as optional in the documentation since August. The RDF mapping spec [3] has been adjusted accordingly.

Below are some tickets and patches for those of you who are interested in the nitty gritty of this:

• https://phabricator.wikimedia.org/T115269
• https://gerrit.wikimedia.org/r/#/c/302248
• https://github.com/DataValues/Number/commit/2e126eee1c0067c6c0f35b4fae0388ff11725307

This change is scheduled for deployment on November 15.

Please let us know if you have any comments or objections.

-- Duesentrieb (talk) 14:51, 4 November 2016 (UTC)

The sentence "Clients that import quantities should now omit the bounds if they do not have explicit information on the uncertainty of a quantity value" is confusing to me. Is a client something that imports quantities from Wikidata to someplace else? Why does it say "explicit information"? If it's coming out of Wikidata, the uncertainty bounds will be there, or they won't, right? What's the difference between explicit and non-explicit?

On the other hand, if this is meant to mean things that import quantities into Wikidata, I hope this won't dissuade editors from using the normal convention in science and engineering of using the number of significant figures as a guide to the precision; if a source says an object is 12.7 m long, it's a reasonable inference to say the object is 12.7 m with a lower bound of 12.65 m and an upper bound of 12.75 m; any reasonable scientist or engineer would come up with a similar interpretation. If this interpretation is seriously incorrect, it's the fault of the author for not explicitly stating the precision. Jc3s5h (talk) 18:38, 4 November 2016 (UTC)

"Clients that import" indeed refer to clients that import data into wikidata. And they should indeed only set the upper- and lower bound if the precision or bounds are explicitly given in the source. They should not apply the common rule of using the given number of digits to infer the significant number of digits, and thus the precision. This should be left to Wikibase, which will apply the appropriate heuristics when needed (that is, when doing unit conversion).

Clients guessing the precision during import was the old behavior, which led to the problem that we now don't know which quantities actually have a precision explicitly given in the source, and which don't. This is particularly problematic because we now want to improve our guessing heuristics. It also means that we now have statements that pretend a source had given a specific precision, while the source doesn't. Leaving it to the client to apply the heuristics for determining the precision also leaves to inconsistencies: it'S e.g. unclear if 200 should be read as 200+/-50 or 200+/-0.5. It's indeed not even entirely clear whether 1.2 should be read as 1.2+/-0.1, or 1.2+/-0.05.

So, clients should no longer guess precisions during import. If the source does not specify the precision, none should be given during import. -- Daniel Kinzler (WMDE) (talk) 17:48, 8 November 2016 (UTC)

Films using music by some composer

In some films, like Antyki (Q9158115), the music was published independently by some composer, in this case Chopin, and then used in the film. However if we use

⟨ Antyki (Q9158115)    ⟩ composer (P86) ⟨ Frédéric Chopin (Q1268)    ⟩

this is wrong as the relationship between the film and the person is not of "composer", but more like "uses music by". Are there any properties to represent this relationship or should I start a proposal?--Micru (talk) 12:58, 5 November 2016 (UTC)

I think will be good idea to have generic property for quotation. For example text works or verses could be quoted in books. --EugeneZelenko (talk) 14:07, 5 November 2016 (UTC)

How about

⟨ Antyki (Q9158115)    ⟩ has part(s) (P527) ⟨ soundtrack (Q217199)    ⟩
composer (P86) ⟨ Frédéric Chopin (Q1268)    ⟩

? Lymantria (talk) 13:56, 6 November 2016 (UTC)

@Lymantria: That method doesn't change the relationship, it just makes it more (unnecessarily) complex, but thanks anyway for taking the effort. I prefer the solution offered by EugeneZelenko, as having a property for quotation could be used for other cases as well.--Micru (talk) 08:18, 8 November 2016 (UTC)

Wouldn't it be better to indicate the composition somehow rather than the composer for non-original score? I think that would be more valuable and better structured data, while the composer is always there just one step away. And when not a whole composition is used, only themes, we have qualifier after a work by (P1877). – Máté (talk) 08:29, 8 November 2016 (UTC)

Is it possible to create a Wikidata query that outputs into a Google Sheets spreadsheet and runs regularly?

Hi all

Is it possible to create a Wikidata query that outputs into a Google Sheets spreadsheet and updates regularly? There are many external visualisation tools e.g Carto that offer amazing visualisations and which could be fed from Wikidata however I've only been able to make them work with data exports and so they are stuck at a certain point in time.

Thanks

John Cummings (talk) 15:00, 7 November 2016 (UTC)

@John Cummings: You might be able to use a tool like AutoHotkey (Q784816) to schedule a query and to post the results to Google Sheets. Their forums are pretty good, if you need help. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:13, 7 November 2016 (UTC)

Thanks very much Pigsonthewing, ideally I would like something that would run online and just update the spreadsheet either once a week + on request or something, but I think I'm imagining a tool that doesn't exist. --John Cummings (talk) 16:29, 7 November 2016 (UTC)

@John Cummings: I am not sure about the "running regularly" but this spreadsheet plugin has a WIKIQUARRY function. --Wesalius (talk) 08:26, 8 November 2016 (UTC)

Thanks very much Wesalius, I will have a look, if it can update then I this is my solution. --John Cummings (talk) 09:09, 8 November 2016 (UTC)

Wikidata weekly summary #234

Here's your quick overview of what has been happening around Wikidata over the last week.

• Fourth birthday: thank you!

Map of the birthday parties

Thanks to everyone who celebrated the birthday by organizing or attending to an event, letting a message on Wikidata or on the social networks, writing a story or creating a present!

Here's the list of the birthday presents from the community and development team:

• d:Template:User Wikidata birthday 2016 (Pigsonthewing)
• A new version of the map showing all the Wikidata items with geolocalisation (Addshore)
• Wikidata's 4th birthday logo (Incabell)
• A new release of the Wikipedia and Wikidata Tools for Google Spreadsheets add-on has made working with Wikidata a lot easier for Google Sheets users. (Tomayac)
• A guide to SPARQL and the Wikidata Query Service (TweetsFactsAndQueries)
• Four new types of charts for the Query Service: line chart, bar chart, scatter chart and area chart! Read more and try examples in the documentation (Jonas)
• {{#statements:…}}, a new parser function is currently in development with better features: try it here! (Thiemo, Wikidata dev team and volunteers)
• Search field added in Scholia, a tool to create scholarship profiles (fnielsen)
• Technical documentation about Wikibase for PHP and JS scripts (Ladsgroup and Jonas)
• new feature for the graph view in the Query Service: it can now browse the properties of items. see in the video (Jonas)
• graph builder tool for the query service to create and export graphs (demo video) (Jonas, Smalyshev)
• d:Module:PropertyPath and d:Template:Show Path Items (TomT0m)
• Everything is connected, a knowledge game based on Wikidata and Commons (list of levels) (Denny)

• Discussions
  • New request for comments: Editing descriptions from Wikipedia Android app

• Events/Press/Blogs
  • Histropedia: “The power of data visualisation combined with free knowledge” on WMF's blog
  • Presenting the Wikidata Human Gender Indicators by Max Klein
  • How the world’s first Wikidata Visiting Scholar created linked open data for five thousand works of art on WMF's blog

• Other Noteworthy Stuff
  • Music in Canada at 150 Wikipedia Project, a multifaceted campaign to increase the amount and quality of content about Canadian music in Wikipedia and Wikidata
  • 400,000,000th edit was made.
  • The average amount of statements per item became higher than five and is still growing.
  • Wikidata entities usage on Wikimedia projects

• Did you know?
  • Newest properties: DRÚSOP ID, code, encoding, BALaT image ID, DocCheck Flexikon De ID, DocCheck Flexikon En ID, biography at the Landtag of Mecklenburg-Vorpommern, Cellosaurus ID, World Spider Catalog ID, Squash Info ID, MSC ID, Yahoo! Japan Talent Database ID, Bandcamp artist ID, French National Assembly Lobbyist ID, BanQ author ID, statistical leader, Sandbox-Monolingual text II
  • Query examples:
    • Continents on Wikidata (All eleven of them) (source)
    • Sluices on the river "Vilaine" in Brittany (source)
    • Countries with most cities named after saints (source)
    • Disney characters who share name with their actors (source)
    • 204 items that cite themselves (source)
    • Statements with two references where one source cites the other (source)

• Development
  • Stable interface policy update
  • More work on federation in order to be able to use Wikidata's items and properties to describe media files on Wikimedia Commons in the future (phabricator:T149580)
  • Worked on first version of Lexeme entity type (phabricator:T148139)
  • Worked more on linking values in Lua and the property parser function instead of just returning the label (phabricator:T142940) You can test it here: https://de.wikipedia.beta.wmflabs.org/wiki/Statements
  • Improved map layer colors in the query service (phabricator:T148022)
  • Improved size of map markers in the query service (phabricator:T148496)
  • Upcoming: quantity changes
  • Worked more on automated sitelinks for Wiktionary (phabricator:T987) Test system is coming in the next days.

You can see all open tickets related to Wikidata here.

• Monthly Tasks
  • Hack on one of these.
  • Help develop the next summary here!
  • Contribute to a Showcase item
  • Help translate or proofread pages in your own language!
  • Help merge identical items across Wikimedia projects.
  • Add labels, in your own language(s), for the new properties listed above.
  • Comment on property proposals: all open proposals - proposals needing attention

Read the full report · Unsubscribe · Lea Lacroix (WMDE) 15:08, 7 November 2016 (UTC)

• Hey Lea, the "new properties" list doesn't look right, there are a bunch much newer than those. ArthurPSmith (talk) 17:07, 7 November 2016 (UTC)

• Probably should have pinged - @Lea Lacroix (WMDE): for example Bandcamp profile ID (P3283) was created October 11, not last week! ArthurPSmith (talk) 20:48, 7 November 2016 (UTC)

User:Legobot normally updated this list but it stopped working. @Legoktm: --Pasleim (talk) 16:13, 8 November 2016 (UTC)

Translate "All entered languages"

Hello. I want to translate in Greek

• All entered languages

but I can't find it in translatewiki.net. Xaris333 (talk) 22:34, 7 November 2016 (UTC)

MediaWiki:Wikibase-entitytermsforlanguagelistview-more (can be found from here) --Njardarlogar (talk) 08:24, 8 November 2016 (UTC)

Showcase items

Hi everyone. Showcase items page is abandoned? I've put some proposals, but nobody reviewed them for months. I thought that it was some kind of "good articles" for Wikidata. We need a new policy, or at least redefine and polish criteria and scope? --Escudero (talk) 15:56, 7 November 2016 (UTC)

I think Lea wanted to look into it, so I stopped doing reviews.
--- Jura 08:50, 8 November 2016 (UTC)

It is very much needed, I think. Thanks to anyone who can help review! :-) Syced (talk) 06:41, 9 November 2016 (UTC)

Hello, yes I planned to improve the design of the page in the same time as the process, but it's pending for now, so don't hesitate to continue reviewing. I'll try to suggest a new version before the end of the year. Lea Lacroix (WMDE) (talk) 09:13, 9 November 2016 (UTC)

Biographical dictionary of Iowa

Are these 424 ids/records suitable for the creation of a new external id property? --Azertus (talk) 16:42, 8 November 2016 (UTC)

• Of course. Why not? Thierry Caro (talk) 18:20, 8 November 2016 (UTC)
• I would use 1000 ids as the cut-off unless there are specific reasons why a specific ID is valuable. ChristianKl (talk) 20:16, 8 November 2016 (UTC)
• The data can of course be linked but maybe the generic property described at URL (P973) is a better solution than creating a specific property. --Pasleim (talk) 20:21, 8 November 2016 (UTC)
• Since the quantity is quite small, I also would use described at URL (P973) / described by source (P1343).--Micru (talk) 09:41, 9 November 2016 (UTC)

List namespace for listeria lists?

Currently lists are found at various places. I wonder if we wouldn't want to centralize them in their own namespace. I'm think more of reference lists than of maintenance reports (such as constraint reports).
--- Jura 11:59, 8 November 2016 (UTC)

I like the idea. That would make it more easy to use the same list in different Wiki's. It might be worthwhile to also have graphs created by the query tool in that namespace. ChristianKl (talk) 20:13, 8 November 2016 (UTC)

Yes, that would be quite useful, and they can be linked with Wikipedia wherever there is an existing hardcoded list.--Micru (talk) 09:43, 9 November 2016 (UTC)

You're basically describing the query namespace. I would wait for that and in the mean time just play around with listeria so we get our requirements for the query namespace clear. Multichill (talk) 18:26, 9 November 2016 (UTC)

I don't see any plan for that actually happening. In the meantime, I think it's worth getting things cleaned-up. Query: could work for the actual SPARQL source.
--- Jura 08:48, 10 November 2016 (UTC)

Man of letters

What's the difference between literary (Q20742892) and literary (Q18195617)? Apparently they seem the same topic. --Superchilum^{(talk to me!)} 12:09, 9 November 2016 (UTC)

@Superchilum: The difference seems to be minimal. I would merge them.--Micru (talk) 09:15, 10 November 2016 (UTC)

I would merge too. @Genium, Yamaha5: you created these two items, do you agree fro merging? Cdlt, VIGNERON (talk) 09:36, 10 November 2016 (UTC)

Unable to edit items on Firefox

Tracked in Phabricator
Task T150401

Is Wikidata:Contact the development team still working? I posted my problem there, but I see that the messages haven't been answered in more than week. So I post here again for visibility, has anyone experience trouble with Firefox? For me the edit links on the right hand side of the items don't appear...--Micru (talk) 07:34, 10 November 2016 (UTC)

Same here (latest Firefox on Windows 10), also the edit buttons for sitelinks and labels/descriptions/aliases always take me to the less comfortable special pages. – Máté (talk) 07:41, 10 November 2016 (UTC)

Thanks for you're report. Apparently it's related to the gadget PopupsFix (phab:T150401) For now you can desactivate the gadget to have the links back. We'll find a better solution soon. Lea Lacroix (WMDE) (talk) 08:56, 10 November 2016 (UTC)

The gadget has been disabled for all users for now. Sjoerd de Bruin (talk) 10:59, 10 November 2016 (UTC)

The gadget got fixed and enabled again. Details in phab:T150401. --Thiemo Mättig (WMDE) 17:50, 10 November 2016 (UTC)

Request to be unblocked in the Wikidata IRC channel

Greetings, I am requesting to be unblocked on the Wikidata IRC channel. I have always been a positive editor on this project and although I have been a bit busy on Commons lately, I would like to be able to participate in discussions about this project again. If there is a better venue to ask, please let me know.Reguyla (talk) 00:00, 4 November 2016 (UTC)

I have to admit I am a little disappointed no one has commented at all on this request. If there is a better venue than Project chat to ask this then please let me know. Reguyla (talk) 19:26, 6 November 2016 (UTC)

I would suggest pinging an IRC administrator. However, as I have never used Wikidata's IRC (and it's been months since I used any other IRC channel) I don't know who that would be. Wikidata:IRC says you can use !admin to get admin attention, but I guess that only works if you are not currently blocked? Thryduulf (talk) 21:56, 6 November 2016 (UTC)

Ah, I see that you also asked this at the administrators' noticeboard where the answer was an unequivocal "No". I consider it extremely unlikely that this request will receive a different response. Thryduulf (talk) 22:05, 6 November 2016 (UTC)

Thryduulf Yeah you're probably right. I only asked there after no one responded here for a few days thinking this just might not be the right venue. I admit I am pretty disappointed being a member in good standing in the community doesn't mean much. It's also annoying that Jasper closed it after commenting that he opposed it strongly and then being called a troll by another user but if there is no policy on Wikidata like EnWP that doing so is discouraged not much else I can do. I'll just focus my attention on commons and won't be able to help out as much here. If you want to close this go ahead. Reguyla (talk) 00:12, 7 November 2016 (UTC)

• I don't think access to IRC is needed to edit here. Why do you think this prevents you from contributing?
  --- Jura 00:18, 7 November 2016 (UTC)

You're right, it isn't required and I will continue editing without it. What it does do is make it a lot easier to ask questions and get answers. As can be seen here, I asked a question and it took several days to even get a response at all. In IRC, someone would have responded in minutes I would have the answer and or be able to collaboratively work through the problem quickly and not take a week or more of back and forth talk page discussion. For example: There are a lot of entries on Wikidata with multiple redundant claims, such as this one. Country of citizenship shows up twice. Do I approve the second one or Reject it as unsourced because the first one has to refs associated to it? Asking questions like that I usually have to wait 3 days or more for an answer but on IRC it would be much faster. There are also frequent discussions about problems that can be seen and makes it easier to know about those things in real time. Reguyla (talk) 00:51, 7 November 2016 (UTC)

Seems like you've been banned in #wikimedia-bans, that also haves a effect in other channels as #wikidata. I don't know anything about the reason for this ban at the moment (I was in Berlin celebrating this project), maybe you've got more information about it (too)? Sjoerd de Bruin (talk) 08:05, 7 November 2016 (UTC)

After some reading, I saw more discussions about this subject. You didn't mention that here, pretty disappointing. Sjoerd de Bruin (talk) 08:49, 7 November 2016 (UTC)

Sjoerddebruin Yes I am in #wikimedia-bans because AlexZ out me there. I wasn't part of the discussion to do it, but he claims he talked it over with others. My guess is that was folks from EnWP. He also contacted his friends on Wikia and had me blocked on their IRC channels as well even though I am a contributor in good standing there. So yeah, I have a serious problem with AlexZ and a couple of the other mods who play god on IRC.

Now that I have been forced to bring EnWP into the discussion, could you elaborate please on what you are disappointed about? My intention is not to hide or mislead anyone but I was trying not to bring up EnWP in my submission because in the past people have wanted me to keep silent about that so the abusive ban will not be overturned or criticized. I also admit I have had some very negative interactions with a few admins and that has been a factor in the past. I feel that admins should have to follow policy, some of them feel otherwise and I am very passionate about the projects and about the communities being treated fairly and respectfully. That has caused friction in the past by some admins who don't feel the same way.

Also, virtually all of those interactions were directly related to my ban on ENWP that I have consistently and will always believe was a complete violation of Wikipedia policy and only done as retaliation for criticizing the abusive admin culture there. I would be more than happy to explain my opinions and feelings further on whatever is addressed as long as people here actually want to discuss it. As I stated above though, I have always been a positive contributor in good standing on this project, but if the Wikidata community here feels that people from EnWP should control their IRC channel instead of the Wikidata community, then I will abide. It will just limit my ability to contribute to this site. Using Commons as a metric, when I was unblocked on their channel I was able to significantly ramp up my contributions because I was able to get live help on projects and problems. If you need to know who to bypass the #wikimedia-bans block I can tell how you what needs to be done. It's how commons was able to do it. Reguyla (talk) 15:17, 7 November 2016 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Unless someone has some desire to unblock me on the Wikidata IRC channel that I am not seeing, which seems unlikely, you can close my request to be unblocked on that channel if you want. It appears the community here is fine with allowing people on other projects to control what happens on their IRC channel with editors of the project who are in good standing here. Regardless of past interactions on EnWp, I am in a good standing here and always have been and that should count for something. Apparently it doesn't, which is unfortunate and disappointing. Reguyla (talk) 16:41, 8 November 2016 (UTC)

If no one is willing to unblock me on the Wikidata IRC channel then this discussion can be closed. Reguyla (talk) 19:46, 10 November 2016 (UTC)

When a person is famous or notable enough to be added to Wikidata?

We - The Finnish Public Broadcasting Company Yle - have been tagging our content with Wikidata-items since the beginning of April this year (for more information: http://wikimedia.fi/2016/04/15/yle-3-wikidata/) successfylly and with pleasure. We’ve also been quite active in adding data to Wikidata, mainly titles and descriptions, but also new items. We would now like to hear your opinion when a person is famous or notable enough to be added to Wikidata!

Concrete examples: We have an Internet service called Elävä arkisto http://yle.fi/aihe/elava-arkisto (Living archives, Arkivet in Swedish https://svenska.yle.fi/arkivet) where we publish content out of our archive collections. Very often there appear “common / ordinary” i.e. not famous people in these archive programs or clips. Like in this one where the life of two families in a suburb in southern Finland in the 1980s is described (sorry, only in Finnish!): http://yle.fi/aihe/artikkeli/2016/10/28/kaksi-tarinaa-koivukylan-lahiosta

Everything we know about these people, are their names, where they lived in the 1980s and their occupation, e.g. family Lahtinen lived in suburb of Koivukylä in Vantaa, the father Jorma Lahtinen worked in foundry and the mother Virva Lahtinen was a housewife taking care of their two children and neighbor's children at their home. Is there enough information to create an item for both of them in Wikidata? (Actually we already created this item https://www.wikidata.org/wiki/Q27652435 to teach the journalists of Living archives how to add data into Wikidata…) Often we are probably not (anymore) able to identify these common people unambiguously, we probably do not know the place or date of birht for instance. And when we try to find more information about them afterwards e.g. in Internet, we can not be sure, if the people we find, are really the same. But we would like to be able to tag our content with these people.

Another real need: Next spring we are going to have municipal elections in Finland, in all about 300 municipalities at the same time. There will be thousands of candidates all over the country. And we make content about them and would like to be able to tag these pieces of content consistently. For these persons we would have more metadata than in the first example: their full names, home towns, dates of birth or at least their ages and the political parties they are running for. And they can be seen as public figures because they run for a public duty. Most of these candidates disappear however from the publicity after they are not elected; only a few of them become famous nationwide. Would it be okay to add all these persons into Wikidata? (And if yes, what would be the most suitable tool for adding them?)

It would be great to hear your opinion about this subject!

With kind regards, Pia Virtanen, Metadata specialist, Yle  – The preceding unsigned comment was added by Pikrvi (talk • contribs).

• @Pla hello! Great that you are using Wikidata, both within your archives, and adding/correcting data on Wikidata itself! As to your question, the Notability policy for Wikidata is a bit vague, so I'd suggest using common sense here. For the people featuring in the 1980 archive programme, if they're listed on a public webpage about that archive programme, then I'd say they were definitely notable enough. If not, then probably not. Also relevant is perhaps how prominently they featured in the programme – were they one of the main subjects, or just an incidental feature? If you do add them, then perhaps you should also add a WikiData entry for the TV programme itself, and add a link from the person to the programme (e.g. perhaps participant in (P1344))? For the political candidates, I'd suggest that they were all notable, as they likely all have at least one reference in online sources about that election. –– Frankieroberto (talk) 14:07, 7 November 2016 (UTC)
• In general the notability police that Frankieroberto referenced goes. I don't think it's valuable to add people in the archive example that can't be identified unambiguously, because without unambiguously identification the data can't be interlinked with other Wikidata items. ChristianKl (talk) 17:59, 7 November 2016 (UTC)
• @Pikrvi: Thanks for reaching out, and thanks for your work using Wikidata. As others have suggested, use your common sense, and try to consider if those people are really that relevant for Wikidata. In my opinion they are probably not, but it is up to you to decide. A possible option could be to create items when needed for elections, and after a while delete them. Or an even better option could be to have your own Wikidata instance running, where evidently all the data would be relevant, and feed only to Wikidata the most important bits (for instance, elected candidates). It is possible to have your own repository, and apparently also federation should be possible in the near future (use Wikidata properties in your own instance). Remember that Wikidata is not a replacement for local databases, it is a complement. I am glad that you are entrusting the community with some of your data and I wish you good success.--Micru (talk) 08:31, 8 November 2016 (UTC)

I would oppose deleting items after a while. Historical election data is useful. ChristianKl (talk) 19:33, 10 November 2016 (UTC)

properties to target physical items that no longer exist

Hi, I have a terrible internet connection and it is difficult for me to do a decent research, could someone answer me about a very specific doubt here? What is the most efficient way to list items that no longer exist, possibly because of a very specific or temporally located event? Like, if I want a list of collapsed/bombed/totally rotten/burned buildings and places what should I ask for in a wikidata list quarry? Thanks.--Alexmar983 (talk) 10:14, 8 November 2016 (UTC)

@Alexmar983: The right place to ask for this is Wikidata:Request a query. But an anwser anyway

select ?building ?buildingLabel where {
  ?building wdt:P31/wdt:P279* wd:Q41176 .
  ?building wdt:P576 [] .
  
  SERVICE wikibase:label {
    bd:serviceParam wikibase:language "en" .
  }
}

Try it!

All items that don't exist anymore should have end time (P582) or a subclass of it, if Wikidata knows that they don't exist anymore. ChristianKl (talk) 20:20, 8 November 2016 (UTC)

Thank you, I didn't know about Wikidata:Request a query, I'll tell other users now.--Alexmar983 (talk) 04:19, 11 November 2016 (UTC)

IP edits on properties

Hi everyone, every once in a while I take a look at IP edits on properties. Very few edits and a high percentage of these edits are not correct or plain vandalism. I'm very much in favour of the ability of IP's editing our wiki's, but in this case it seems to cause more damage than good. Should we restrict the editing of properties to logged in users? I would prefer you comment with arguments in favour or against this. Multichill (talk) 14:23, 10 November 2016 (UTC)

I've already semi-protected some properties in the past due to vandalism. I'll be in favor of this. I only protect them at the moment when there is frequently vandalism. Sjoerd de Bruin (talk) 14:38, 10 November 2016 (UTC)

• Changes to URL formatters are already blocked by filter.
  --- Jura 14:42, 10 November 2016 (UTC)
• Ok for me if there is a clear indication of that this happens only for IPs and what the ways around are. Matěj Suchánek (talk) 20:46, 10 November 2016 (UTC)

References for "instance of human"

I added [4] to Maxima Chan Zuckerberg (Q22302383) as a reference for the statement "instance of human". My edit was, tiresomely, reverted with the edit summary "Doesn't support claim‪". I contend that it does, not only in the included image, but also in the clause "[Mark] Zuckerberg is currently one of the richest people...", a definitive statement, since "people" is the plural of "person", and "person" is a synonym for a human; and the article also states "Mark Zuckerberg holds baby daughter Maxima" (or are we now supposing that humans can father non-human children?)

Accordingly, I have restored my edit. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:00, 29 October 2016 (UTC)

Being extremely pedantic, having a child does not prove a person's humanity unless there is definitive reference for the statement that the child is human (I have not looked to see if there is such in this instance). Getting more abstract, it is not possible to infer species between generations without knowing the species of both parents in all cases as hybrids are possible (e.g. the offspring of a lion maybe a lion, liger (Q182573), tigon (Q1383362), leopon (Q956107), etc. This is not a consideration required for humans living after the extinction of Neanderthals though.

More relevantly, I agree with you though that "[Mark] Zuckerberg is currently one of the richest people..." does support the statement "instance of human" for Maxima Chan Zuckerberg (Q22302383). Thryduulf (talk) 10:45, 30 October 2016 (UTC)

I'm not going to respond in this discussion when the other person doesn't even notify me about it. Sjoerd de Bruin (talk) 10:36, 1 November 2016 (UTC)

Being even more pedantic, the species Homo sapiens (Q15978631) is; the only member of genus Homo (Q171283), a member (along with the chimpanzees) of the tribe Hominini (Q107588), a member (along with the gorillas) of the subfamily Homininae (Q242047), and finally a member of the family Hominidae (Q635162) (or the Great Apes, which bring the orangutans into the mix). Since there are no known hybrids between humans and chimpanzees, gorillas, or urangutans (although doubtless some humans may have tried), then we can assume that the offspring of a human is a human.

More prosaically, but just as pedantically, the term "wife" (as in the reference under discussion) implies a legal relationship between a human male and a human female.

So, notwithstanding future developments in genetics and in-vitro fertilization, I think, for the moment at least, we can assume that a statement that someone is the son or daughter of a known human is sufficient to grant humanity to the someone involved. Robevans123 (talk) 21:08, 4 November 2016 (UTC)

There have been other species in the genus Homo (Q171283) but post their extinction there have been no known cases of hybridisation between Homo sapiens (Q15978631) and any other species. So, post the extinction of those other members of the genus, yes, "son or daughter of known human = human" and "mother or father of known human = human" but "son or daughter of entity who may or may not be human" does not prove the entity is human without knowing whether the son or daughter is human.

However, "the term "wife" implies a legal relationship between a human male and a human female." is not correct. Same-sex marriage, transgender people, intersex people, non-binary people, all exist and all are able to legally marry in one or more jurisdictions. There are also examples of humans marrying non-human animals (e.g. Sudanese goat marriage incident (Q11023828)) and even inanimate objects (e.g. [5][6]) although the legality and recognition of these is debateable they do show that the term "married" cannot be used to prove that all participants are human. Thryduulf (talk) 11:27, 5 November 2016 (UTC)

Well, the legal definition of wife is, at least in some jurisdictions, as I stated. I was deliberately not implying that wife = "married", and therefore that male/female was the only form of marriage or legal relationship that is available. Actually, the real problem with my thesis is that the use of "wife" in the article may not be the narrow legal definition, and that there are other, much less used, definitions of wife that apply to non-humans. The same is true for son, daughter, husband etc. Even people (eg the monkey people of the forest), and person (a self-conscious or rational being [philosophy], or a corporation, a partnership, an estate, or other legal entity recognized by law as having rights and duties [legal]) run into trouble if you look at all possible definitions.

In fact, having looked at the article again, I cannot find one word that definitely, absolutely proves that Maxima is human.

However, reading the article as a whole, it is nigh impossible to come to a conclusion other than that Maxima is human. Thinking about this reminded me of the Turing test (Q189223), and whether a similar test could be applied to articles and if the article is talking about humans. I propose the following test:

Could man on the Clapham omnibus (Q7777067), reading the article in question, come to the conclusion, beyond reasonable doubt, that the article is about humans (and not about "married dogs", unnatural behaviour with goats, strange relationships with fairground rides etc)?

I tentatively propose that such a test be called the Thryduulf-Evans Test (TET), or more colloquially, the Bleeding Obvious Test (BOT). Any article that passes TET/BOT can be assumed to be talking about humans, and is a valid source for conferring instantiation of humanity. Robevans123 (talk) 11:07, 12 November 2016 (UTC)

Naruto (Q642) and Naruto (Q26971382) the same?

Are Naruto (Q642) and Naruto (Q26971382) the same thing? The latter has no sitelinks, so I suspect so, but there might be some difference I miss. Anybody knows? --Laboramus (talk) 03:37, 5 November 2016 (UTC)

Konggaru Starry K. Erne Mogilevich Santer AldNon^Ucallinme? Thibaut120094 Shikeishu C933103 Sight Contamination -Zest Vulphere Sakretsu Jean-Frédéric Tris T7 ^{TT me}
Wallacegromit1 Jeanjung212 Bagas Chrisara ミラP CrystallineLeMonde
Nicereddy Shisma (talk) MatrosMonk Bwk24 Mickn RPI2026F1 Yirba Eniehack Wiccio IntensionalLogican

Notified participants of WikiProject Anime and Manga --Liuxinyu970226 (talk) 07:58, 5 November 2016 (UTC) Plus: @風中的刀劍:. --Liuxinyu970226 (talk) 04:57, 12 November 2016 (UTC)

The first one is, per the item's description, about the media franchise, and the latter about the manga series itself. But I think the sitelinks in the first one mostly tell only about the manga series itself, and not at all about the media franchise. --Stryn (talk) 08:54, 5 November 2016 (UTC)

Agreed with Stryn: Naruto (Q642) = w:en:Naruto the article, while Naruto (Q26971382) = w:en:Naruto#Manga ≈ list of Naruto manga volumes (Q1062669). --風中的刀劍 (talk) 13:03, 12 November 2016 (UTC)

broadcast by

Hello.

1) I want to use broadcast by (P3301) to 2016–17 Cypriot First Division (Q23756432). But the home matches of the 12 of the 14 teams are broadcast by Cytavision (Q5201296) and the other 2 (AC Omonia (Q240783), Apollon Limassol FC (Q428047)) by PrimeTel (Q7243224). How can I show that?

2) Some years before only a match was broadcast in every round. The TV channel choose one important match (of the 7 match of a round) and that was showed. Is it OK to use only

⟨ 1999–2000 Cypriot First Division (Q2698220)    ⟩ broadcast by (P3301) ⟨ Lumiere TV (Q6703234)    ⟩

 ?

Xaris333 (talk) 23:00, 5 November 2016 (UTC)

Yes, but I'd qualify it with applies to part, aspect, or form (P518). For question 1 I'd use items "home matches of <team>" (with those items given subclass of (P279) association football match (Q16466010) and participating team (P1923) <team> along with anything else relevant, maybe the venue). For question 2 use an item for the specific match. Thryduulf (talk) 13:36, 6 November 2016 (UTC)

Thryduulf

For the first one" it may be too complicate if I use items "home matches of <team>". I will need item for every team. Is that ok?

For the second one, I don't know the matches that was showed on TV. I just know (I have a source) that only one game was broadcasted by TV (on Saturdays at 19:00 o clock). Xaris333 (talk) 17:17, 6 November 2016 (UTC)

Yes, one item per team is fine - they fulfil a structural need and so are fine notability wise. For the second, you could use "applies to part, aspect, or form (P518) unknown value" qualified with "quantity (P1114) 1 association football match (Q16466010)". If we had an "applies to parts of the class" property (which I think would be rather useful anyway) that would probably be better. Thryduulf (talk) 21:52, 6 November 2016 (UTC)

@Thryduulf: Check 2016–17 Cypriot First Division (Q23756432). [7]. Xaris333 (talk) 15:13, 12 November 2016 (UTC)

That looks fine to me. Thryduulf (talk) 16:30, 12 November 2016 (UTC)

sequence of "interwikis" in Wikipedia

May you help, if there is a way to do please? Thank you, Conny (talk) 16:16, 11 November 2016 (UTC).

Mixed drinks, alcoholic or not

• Mixed drinks can be alcoholic or non-alcoholic.
• Alcoholic mixed drinks are (usually?) called cocktails.
• In Russian and some other languages, "коктейль" (from "cocktail") means "mixed drink", not sure how well-defined non-alcoholic mixed drinks are. Due to that confusion, I made some changes to the items. Now we have:
  • mixed drink (Q3246609) (all mixed drinks)
  • cocktail (Q134768) (alcoholic)
• But I just noticed both categories assigned to the items had "alcoholic" in some of their labels. There are more alcoholic categories than alcoholic items. I edited them a little, but got stuck:
  • Category:Alcoholic mixed drinks (Q7214075) — the biggest category.
  • Category:Alcoholic cocktails (Q18562959) — a smaller category that cannot be merged with the above.
  • Category:Mixed drinks (Q27832139) — I just created it for the most general stuff like mixed drink (Q3246609).
• The problem:
  • Lithuanian Wikipedia's category seem (I don't read the language) to distinguish alcoholic mixed drinks and alcoholic cocktails:
    • en:Kategorija:Maišyti alkoholiniai gėrimai
    • en:Kategorija:Alkoholiniai kokteiliai.
  • I can't distinguish between non-alcoholic mixed drinks and non-alcoholic cocktails.

--AVRS (talk) 11:41, 12 November 2016 (UTC)

I'd propose to create our own hierarchy with "clean" items and to leave up the local communities to link their articles and categories to our newly created items to make things straight.

We need a few items :

• Mixed drink / "pure" drink.
• Alcoholic drink / non alcoholic drink
• Alcoholic mix drink / non alcoholic mix drink

The properties disjoint union of (P2738) and union of (P2737)  offers a framework to properly define the items.

⟨ drink ⟩ disjoint union of (P2738) ⟨ list of values as qualifiers (Q23766486)    ⟩
list item (P11260) ⟨ Mixed drink ⟩
list item (P11260) ⟨ "pure" drink ⟩

⟨ drink ⟩ disjoint union of (P2738) ⟨ list of values as qualifiers (Q23766486)    ⟩
list item (P11260) ⟨ Alcoholic drink ⟩
list item (P11260) ⟨ non acoholic drink ⟩

⟨ Alcoholic mix drink ⟩ subclass of (P279) ⟨ Mixed drink ⟩

⟨ Alcoholic mix drink ⟩ subclass of (P279) ⟨ Alcoholic drink ⟩

...

As our item are new there is no confusion wrt. the old interwikis. Then we definitely need to actually do something with WD:XLINK to restore interwikis. author  TomT0m / talk page 11:59, 12 November 2016 (UTC)

Not all mixed alcoholic drinks are cocktails. Consider Snakebite (Q13424415) and shandy (Q13527031). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:08, 12 November 2016 (UTC)

How do you define a cocktail ? author  TomT0m / talk page 20:12, 12 November 2016 (UTC)

But English Wikipedia does not contradict calling them cocktails if I don't have a definition? --AVRS (talk) 07:11, 13 November 2016 (UTC)

I am sorry, what's the difference between drinks and beverages? Infovarius (talk) 22:28, 12 November 2016 (UTC)

@Infovarius: None. author  TomT0m / talk page 10:28, 13 November 2016 (UTC)

Sport kits

Hello. I am thinking about proposing a property about a sport kits. For example, about football kit. See w:Template:Infobox football club. The parameters are kit_alt1, pattern_la1, pattern_b1 etc (you can also see w:Template:Football kit). Usually they use colours code (like #FFFFFF, all main colors have there own items in wikidata) or specific images from commons like c:Category:Football kit body. Some Wikipedias are using or may using Wikidata so the Template:Infobox football club is automatic fill. The only problem is the football kit. We need a way to have the data about football kit in Wikipedia from Wikidata since all the data are free to use. Any help how to do that? Xaris333 (talk) 20:26, 12 November 2016 (UTC)

Fetch data from Wikidata to Wikipedia

Hello. I am trying to fetch data from Wikidata to Wikipedia, using a template. I have some problems with qualifiers and other things. Does anyone can help me? Xaris333 (talk) 10:58, 13 November 2016 (UTC)

You probably need Lua for that. Can you show us your current code? Sjoerd de Bruin (talk) 19:46, 13 November 2016 (UTC)

Hello. I am using a Greek template. el:Πρότυπο:Κουτί πληροφοριών ποδοσφαιρικού συλλόγου (Template:Infobox football club). The first problem is about the stadium. I am using home venue (P115) from club's wikidata page. Its ok. But I also want located in the administrative territorial entity (P131) and maximum capacity (P1083) from stadium's Wikidata page, not for club's Wikidata page. Xaris333 (talk) 19:58, 13 November 2016 (UTC)

Merging items

I found On Nature (Q19368481) and Anaximander (Q42458). I wondered if they are really the same, and therefore be merged into one? QZanden (talk) 21:42, 13 November 2016 (UTC)

@QZanden: The nl.wb link is about Anaximander's views on nature and the fr.wb link is about the philosopher in general and can be merged. —Justin (koavf)❤T☮C☺M☯ 21:54, 13 November 2016 (UTC)

So the fr.wb can be replaced to Q42458, and the nl.wb can stay were it is? QZanden (talk) 21:57, 13 November 2016 (UTC)

@QZanden: I believe so. There are many subpages on Wikibooks which will never have interwiki links. —Justin (koavf)❤T☮C☺M☯ 03:05, 14 November 2016 (UTC)

Password reset

I apologise that this message is in English. ⧼Centralnotice-shared-help-translate⧽

We are having a problem with attackers taking over wiki accounts with privileged user rights (for example, admins, bureaucrats, oversighters, checkusers). It appears that this may be because of weak or reused passwords.

Community members are working along with members of multiple teams at the Wikimedia Foundation to address this issue.

In the meantime, we ask that everyone takes a look at the passwords they have chosen for their wiki accounts. If you know that you've chosen a weak password, or if you've chosen a password that you are using somewhere else, please change those passwords.

Select strong passwords – eight or more characters long, and containing letters, numbers, and punctuation. Joe Sutherland (talk) / MediaWiki message delivery (talk) 23:59, 13 November 2016 (UTC)

Adding to the above section (Password reset)

Please accept my apologies - that first line should read "Help with translations!". Joe Sutherland (WMF) (talk) / MediaWiki message delivery (talk) 00:11, 14 November 2016 (UTC)

Unit conversion in Query Service

Hello all,

We recently worked on improving the way that units are matched and sorted on the Query Service, by converting the different length numbers in a normalized unit (meters).

Before, length quantities were only represented in WDQS in the unit that was used when the quantity was entered. This made it impossible to search and compare length quantities across units. See the example of ten longest things in Wikidata. We get a list of the largest numbers, with no idea of the actual length.

Now, the length quantities are also stored in meters, and are used in meters in the Query Service. See the same example. We now can get a list of the longest things on Wikidata, measured in meters.

To use this unit conversion, you need to use the prefix psn: in the query. You can find it in the menu Prefixes->Wikidata.

Length units are the first part of improving the unit conversion. Feel free to try it, ask questions or give feedbacks!

Thanks, Lea Lacroix (WMDE) (talk) 15:34, 11 November 2016 (UTC)

An explanation needed about how it works. I would have thought that it uses conversion to SI unit (P2370) as a multiplier, but then why light-year and gigaparsec are missing in psn queries, but light-second is present? --Lockal (talk) 11:19, 12 November 2016 (UTC)

• Nice. When will it be available for all quantities entered?
  --- Jura 15:33, 12 November 2016 (UTC)

@Lockal: yes, I also noticed that... reported at phab:T150656.

@Jura: it's available for all statements/properties, but not for all units. We will add the common simple units once we are sure we have sorted out any problems (like the one mentioned above). "Common" mostly means "metric or imperial", and "simple" means "conversion does not require an offset, exponent, or other fancy stuff". Some "units" will never be supported, because they have no simple rule for conversion - like currencies, or many historical units

We'll probably add offsets to allow conversion between °C and °F, and we may add exponents at some points, to support converting things like car efficiency (Europe uses l/100km, US uses miles/gallon) -- Duesentrieb (talk) 14:34, 14 November 2016 (UTC)

Wikidata weekly summary #235

Here's your quick overview of what has been happening around Wikidata over the last week.

This is the Wikidata summary of the week before 2016-11-12.

• Discussions
  • Current request for comments: Editing descriptions from Wikipedia Android app

• Events/Press/Blogs
  • Wikidata and the Chain of Death, by Karl Beecher
  • Scaling multilingual name tags with Wikidata OpenStreetMap contributor blog post
  • Wikidata and Persistent Identifiers was presented at PIDapalooza November 9th.
  • Wikidata in OpenStreetMap
  • Wikidata and the Semantic Web of Food

• Other Noteworthy Stuff
  • A new Wikidata PageRank dataset has been published. The dataset was computed in August 2016 and provides PageRank scores for 10,364,840 Wikidata entities.
  • Due to a possible hack, please consider resetting your passwords.

• Did you know?
  • Newest properties: generation time, hazard on site, ACM Digital Library event ID, ACM Digital Library citation ID, HGVS nomenclature, Supermodels.nl ID, CIViC variant ID, wurvoc.org measure ID, Réserves Naturelles de France ID, World Waterfall Database ID, Ylioppilasmatrikkeli 1853–1899 ID, ID petit-patrimoine.com, opponent during disputation, Vlindernet ID, male form of label, board member, Patrimonio Inmueble de Andalucía ID, BDI ID, ICCF ID, chesstempo ID, 365chess player ID, plan image, muscle action, SummitPost ID, lib.reviews ID, Galiciana Author ID, ICAA rating, KINENOTE person ID, NGS pumping station ID, third-party formatter URL, Open Media Database film ID, broadcast by, musical conductor, Student register of the University of Helsinki 1640–1852 ID, Belgian Senate person ID, Flemish Parliament person ID

• • Query examples:
    • Map of educational institutes from all over the world (source)
    • Current US Supreme Court justices by their date of birth (source)
    • Death dates of people with Wikidata items (source)
    • French towns that a street in Paris is name after(source)
    • Statements with “reason for deprecation” rank that aren’t deprecated(source)

• Development
  • More work on federation (phabricator:T76007)
  • Worked on first basic version of Lexeme entity type for Wiktionary support (phabricator:T148139)
  • Test system for automated language links for Wiktionary (phabricator:T987)
  • Analysed feedback prototype for editing on Wikipedia in order to refine it
  • Worked on mapping Commons workflows and needs
  • Analysed queries used with Listeria
  • More work on making ArticlePlaceholder pages indexeable by search engines (phabricator:T117693)
  • More work on enabling translations from existing articles in other languages on ArticlePlaceholder pages (phabricator:T124036)
  • Link to local articles in ArticlePlaceholders (phabricator:T113955)
  • Fix for bug that prevented editing (phabricator:T150401)
  • Experimented with improvements for property suggestions (phabricator:T132839)

You can see all open tickets related to Wikidata here.

• Monthly Tasks
  • Hack on one of these.
  • Help develop the next summary here!
  • Contribute to a Showcase item
  • Help translate or proofread pages in your own language!
  • Help merge identical items across Wikimedia projects.
  • Add labels, in your own language(s), for the new properties listed above.
  • Comment on property proposals: all open proposals - proposals needing attention

Read the full report · Unsubscribe · Lea Lacroix (WMDE) 17:21, 14 November 2016 (UTC)

Tools Labs maintenance

You might encounter some issues with tools hosted on Wikimedia Labs. This was only announced (postponed) on Labs-related mailing lists, I really wish I knew about this before. Sjoerd de Bruin (talk) 19:26, 14 November 2016 (UTC)

And Meta-Wiki page's loading didn't stop for me because of "GET https://tools-static.wmflabs.org/meta/scripts/pathoschild.ajaxtransclusiontable.js net::ERR_CONNECTION_TIMED_OUT". --Stryn (talk) 19:43, 14 November 2016 (UTC)

meta:MediaWiki_talk:Common.js#Toollabs_hosted_.js_in_Common.js.3F. Multichill (talk) 21:41, 14 November 2016 (UTC)

Generic label descritions preventing item merge

Some items can have a lot of generic label descriptions, in many languages. Those descriptions can have negligible differences in different items, e. g.:

1. "Wikimedia disambiguation page"
2. "Wikipedia disambiguation page"

Anyway, they prevent the items from simple merging by Special:MergeItems. To prepare the items for merging, it's quite tedious to clear those useless, but conflicting descriptions manually one by one. Is there some kind of automated tool for this task? --Djadjko (talk) 02:21, 15 November 2016 (UTC)

Problem on Medal of Honor (Q203535)

I think there may be a problem on Medal of Honor (Q203535) with regard to the approval of the "winners". There are so many, I think it's causing errors because everytime I approve a claim I get a {{Property-html}} for every remaining row awaiting verification of approval for the claim. I'm not sure this was designed to have a potential 3500+ rows. Reguyla (talk) 16:56, 12 November 2016 (UTC)

You shouldn't use the property "winner" on items for awards. Add them to the items for the person instead.
--- Jura 18:18, 12 November 2016 (UTC)

I didn't actually. Someone or something else added them I was just approving them. And it's like that on a lot of military awards. Maybe a bot gone astray or something. To be honest, if we are going to associate the recipients to awards I think we need to find a better way than this. The Medal of Honor is pretty rare and has only been given about 3500 times in its existence, when you get to other more common ones like the Purple heart there are tens of thousands. So if I should be removing these from the award, then let me know how I should do that, preferably en masse instead of the one at a time method. Reguyla (talk) 18:31, 12 November 2016 (UTC)

By "approving" them, you add them: https://www.wikidata.org/w/index.php?title=Q203535&action=history . If you add them, we assume you checked them.
--- Jura 18:42, 12 November 2016 (UTC)

I did check them with regards to verify they are Medal of Honor recipients. If you are now saying they should not be there at all, then the "Approve/Disapprove" functionality is flawed. I am "approving" them as Medal of Honor recipients, but the Software or something else is adding them into the pending approval disapproval state. Reguyla (talk) 19:08, 12 November 2016 (UTC)

Maybe you should read yourself into the Primary Sources tool first. It's your own responsibility to approve those claims. Sjoerd de Bruin (talk) 19:19, 12 November 2016 (UTC)

I don't think we are thinking about the same things here. The claims are correct! 100% correct. The problem is that there are too many and its crashing the page. What I am saying is that if someone believes that these claims should not be on the Medal and instread should go on the individuals pages, then there is something wrong with how Wikidata and/or the Primary Sources tool are displaying entries on those pages. Reguyla (talk) 19:26, 12 November 2016 (UTC)

For what its worth, this is exactly the kind of discussion that would be better to have on the Wikidata IRC channel, which is why I asked for it above. Reguyla (talk) 19:31, 12 November 2016 (UTC)

I have removed the data from Medal of Honor (Q203535). Would we store recipients on an award item, it wouldn't be through winner (P1346). And then we actually do not store recipients on an award item, but the other way around. Thierry Caro (talk) 19:52, 12 November 2016 (UTC)

Ok thanks. I would have been happy to do that but I appreciate the help. I agree removing them is the most appropriate thing to do and they should never have been associated to the page in the first place requiring approval, so the bot/tool looks like it needs some changes. Should I disapprove the rest that are still waiting for approval. Reguyla (talk) 19:56, 12 November 2016 (UTC)

This kind of discussion is better to have here because can be useful also for other users. --ValterVB (talk) 21:28, 12 November 2016 (UTC)

Special:ConstraintReport/Q203535 would have shown that all these claims violate constraints. There is Phab:T145930 about this problem
--- Jura 08:10, 13 November 2016 (UTC)

Thanks that's a great page to know. I just bookmarked it. I had no idea that existed. Reguyla (talk) 13:33, 13 November 2016 (UTC)

The medal of honor isn't an event as such the claim isn't correct. Even if it would be correct that doesn't mean automatically it should be approved. The primary source tool includes a lot of claims. ChristianKl (talk) 11:27, 16 November 2016 (UTC)

QuickStatements: How to add references to my statements?

I am using QuickStatements to import information from Wikivoyage.

My statements look like this:

CREATE
LAST    p:P31           wd:Q3917681
LAST    wdt:P131        Q1515
LAST    wdt:P137        Q30
LAST    wdt:P968        "mailto:consularabidja@state.gov"
LAST    wdt:P856        "http://abidjan.usembassy.gov/"
LAST    wdt:P1329       "+225 22 49 40 00"
LAST    wdt:P2900       "+225 22 49 43 23"

How to mention Wikivoyage as a reference for all of these properties? Automatically, of course. Is there a special QuickStatements syntax for this? Thanks! Syced (talk) 03:24, 15 November 2016 (UTC)

I found the answer to my own question: Just add " S248 Q373" at the end of each statement, and it will create a "Reference -> stated in -> Wikivoyage" block for it. Cheers! Syced (talk) 04:29, 15 November 2016 (UTC)

Shouldn't that be imported from Wikimedia project (P143)? Sjoerd de Bruin (talk) 08:16, 15 November 2016 (UTC)

Of course, and more specific than just Wikivoyage (Q373). Matěj Suchánek (talk) 15:06, 15 November 2016 (UTC)

New Status Indicator Icon about Relative Page Completeness

Dear Wikidata Community,

We are happy to announce today a new status indicator icon about the relative completeness of pages, available via our Recoin user script. The script adds a color-coded progress bar icon to the status bar, which expresses how extensive the information on the page is in comparison with pages about similar entities (see screenshot on the right for an example for Max Planck). Currently, the script works for all humans that have an occupation, by comparing the typical properties of entities having that occupation.

Example: The relative completeness indicator for Max Planck is at the top right.

It can be enabled by adding the following line to your common.js:

 importScript( 'User:Ls1g/recoin-core.js' );

Further information on how the values are computed is on the tool's page.

We are looking forward to your opinions and feedback, especially regarding the meaningfulness of such an indicator in general, and regarding the accuracy of the computed indicator values!

Best wishes, Ls1g (talk) 08:53, 15 November 2016 (UTC)

Great script! Some thoughts:

• I don’t like the metric used to calculate the scores. There are typically lots of properties which are rarely used in a set of a given occupation, because they are applicable to just a few persons mostly with multiple other occupations. These rarely used properties lower the score for all persons with this occupation. Example: Mahé Drysdale (Q372735), one of ~6000 persons with occupation rower (Q13382576), is ranked only as “fair” (3/5 if I see this correctly) and the script indicates 29 missing properties of which the most frequently used one is found in just 1.223% of the items of the set. The other 28 suggested properties are below 1%. To my opinion rarely used properties should not have that much impact on the score, since this makes it impossible to raise (a significant number of) items to detailed or very detailed level.
• After I added properties, how do I update the score (as a reward)? It takes a while and many page updates (CTRL+F5) until I see the new value.
• A complementing tool on Tool labs would be nice: general statistics per occupation, and rankings of items. I’m not sure whether this is as easy as it sounds on first thought…
• When using the explanations module, “Most relevant missing predicates, based on the attributes that people of same occupation have:” is displayed even on items without the occupation (P106) property.
• Use mw.loader.load('//www.wikidata.org/w/index.php?title=User:Ls1g/recoin-core.js&action=raw&ctype=text/javascript'); within your common.js instead of the old importScript() function

—MisterSynergy (talk) 10:51, 15 November 2016 (UTC)

Thanks for your comments!

• 1. Thanks for pointing out this situation. A metric that takes into account relative frequencies, not just counts, sounds indeed more appropriate. We will think about this.
• 2. For computing the score we send a SPARQL query to https://query.wikidata.org/sparql, I don't know how long it takes for changes to be reflected there. Maybe this is not the fastest way, I'll check about that.
• 3. We have this information computed offline (it backs the script), you can see it here
• 4. and 5. Will be corrected, thanks!

Ls1g (talk) 21:22, 15 November 2016 (UTC)

I have another comment related to the positioning and design of the explanations module output. The situation right now is as follows:

• the list of property suggestions is displayed below the item aliases and before the labels input list in a manner that suggests that this information actually belongs to the item, although its just a tool result
• the list uses a lot of space very much in the beginning of the item page (a valuable place, but maybe too valuable for a tool result)
• due to its late loading with JS it shifts all the information downwards after the tool result has been collected from SPARQL

My suggestion would be to make the suggestions list collapsible, load it collapsed by default, and place it somewhere else, perhaps right next to the score indicator. Regards, MisterSynergy (talk) 10:06, 16 November 2016 (UTC)

Edit summary linking

Moved here from Wikidata:Administrators' noticeboard. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:27, 15 November 2016 (UTC)

Please, can someone remove the linking of the second user in the edit summary text when we perform a rollback? I know that only the first user, whose edits are reverted, is linked to from the edit summary over on en.wiki. I don't think it's necessary to link to the user who made the last decent edit, and it's quite often a red-link simply to the userpage anyway. It's unnecessary and clutters up the contributions page of those who commit themselves to fighting vandalism.

To confirm, instead of:

(Reverted edits by User:User (talk) to last revision by XXX)

I'd like to see:

(Reverted edits by User:User (talk) to last revision by XXX)

I don't know if an administrator can do this, or if it has to be someone from the development team. Lydia? Jared Preston (talk) 20:32, 14 November 2016 (UTC)

I as a vandal fighter found(*) these links particularly helpful, as they allowed(*) me to quickly see in my contribution list if it was worth checking the previous revision, too (e.g. the previous edit was by an IP: yes; it was by another vandal fighter: no).

(*): Now I'm using the great User:Yair rand/DiffLists.js script and cannot imagine doing vandalism patrolling without it any more. But as it rewrites the summary, I hardly get to see this part of the summary anymore anyway, so I would probably not even notice if this is changed. --YMS (talk) 06:56, 15 November 2016 (UTC)

Yes, but had I been the second name, I needn't have been linked. You know me anyway. If an IP address had done the previous edits, the link is to their user page. 99% of IP addresses don't have a user page, therefore I don't see the need to have a link there. I'm not saying delete the first link, but the last one. Jared Preston (talk) 13:17, 15 November 2016 (UTC)

What about changing the link to the contributions page? I think we do that on nlwiki too. Sjoerd de Bruin (talk) 15:08, 15 November 2016 (UTC)

Well, Sjoerd, that's not exactly what I'd like to see, but that's just a question of taste. As I say, I may be biased with my previous experience at en.wiki. I don't know what nl.wiki does. However, linking to the contributions page of both users would be preferential over what we have now (which I see as no use whatsoever). And thank you, Matěj, I didn't know that! Jared Preston (talk) 15:20, 15 November 2016 (UTC)

I can be done at MediaWiki:Revertpage, by the way. Matěj Suchánek (talk) 15:04, 15 November 2016 (UTC)

Moved here, as this is not a decision for admins alone. Once consensus is reached, a notice can be made requesting an admin to enact the decision. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:27, 15 November 2016 (UTC)

So would anyone actually mind this change being done??? Jared Preston (talk) 10:51, 16 November 2016 (UTC)

wikidata-externalid-url

Who is tracking users at Wikipedia and why? The url https://tools.wmflabs.org/wikidata-externalid-url is used for a lot of lookups, why? Can someone point me to where this is discussed and who is in change of the service? Who is ArthurPSmith, he now has access to information that even the checkusers don't have. Jeblad (talk) 20:30, 15 November 2016 (UTC)

I don't see any tracking software in the code. The tool is used for complicated identifiers, like IMDb. Sjoerd de Bruin (talk) 20:49, 15 November 2016 (UTC)

This service will expose the readers, whether there are any explicit tracking done is not important. The user operating this service will have access to the readers IP-adress, and any logs will identify the clients. As far as I know Wikimedia Foundation goes to great length to hide or remove all identifiable information, and you guys do this? Seriously? (phab:T150803) Jeblad (talk) 21:02, 15 November 2016 (UTC)

Hmm, this hadn't come up before but it's a good point. This was a workaround for problems with complex (parametrized) formatter URL's and external id's. Suggestions on doing better welcome. I just wiped out the logs on the tools server and honestly I haven't looked at them except to see how much it was being used. The original discussion that resulted in this was here. ArthurPSmith (talk) 22:10, 15 November 2016 (UTC)

The core problem is recreating an URL from an identifier, or to quote a previous coworker "to recreate the cow from the ground meat". The problem arise because the identifier, even if it is a valid one according to RDF, it is not necessarily a valid component in an URL. There exists although a valid transform from one form to the other, and it is defined by the pair format as a regular expression (P1793) and formatter URL (P1630). The former is used for validating the value, but also to break the value apart so it can be used in the later. When we refer to $1 it is the overall pattern, while $2 is the first parenthesis in the pattern. This is the usual interpretation, please correct me if you guys are doing something different.

Partial solution: The problem is how to handle ISNI (P213). This has a pattern \d\d\d\d \d\d\d\d \d\d\d\d \d\d\d[\dX] and a format https://tools.wmflabs.org/wikidata-externalid-url/?p=213&url_prefix=http://isni.org/&id=$1. If the pattern is changed to (\d\d\d\d)\s?(\d\d\d\d)\s?(\d\d\d\d)\s?(\d\d\d[\dX]) and the format to http://isni.org/$2$3$4$5 then it should give a valid URL.

Full solution: The previous solution does not work for IMDb ID (P345) because small differences in the identifiers, the initial two letters, indicates dispatching to several different destination URLs, but the present scheme only works for a single URL. Assume that we make a property "alternate id substitution" for use in the property definition that holds both the pattern and the format. It would be like a Perl substitution pattern, and would hold something like [nm\d+][8]. (The first characters are used as delimiter.) The interpretation of this property is that it creates some alternate form of the identifier, in this case an URL. It can be any type of identifier, so it should be qualified with a type. We can have several such "alternate id substitutions", but only the ones that match will have a valid interpretation when instantiated for the specific identifier. The item For Your Eyes Only (Q332330) does not have a valid identifier for this substitution, but Roger Moore (Q134333) has a valid identifier. Any valid substitution can be used, but usually there will be one or zero valid substitutions.

There are two other qualifiers that might be necessary for the full solution. One is "reject if pattern matches" and the other is "accept if pattern matches". If a substitution has a qualifier for reject, then it should not be used if the pattern matches. If a qualifier says accept, then it should be used only if the pattern matches.

The actual pattern should be limited to primitives that can be easily mapped into Lua patterns, and not the full Perl galore. That would make the substitutions nearly impossible to implement in Lua. This will although impose a limitation on alternations, that is [(nm|tt)\d+][$1] will not be a valid substitution, but it can be rewritten as two substitutions.

I'm pretty sure the full solution is a sufficient to recreate all of the URLs in use on this project. Not only URLs can be created, but also other types of identifiers. The only type of identifiers that are troublesome are those where some kind of recalculation is necessary. This can be handled in Perl substitutions, but it would imply using the execute syntax which creates security issues. A solution could be to define some kind of limited language, but I'm pretty sure this problem will newer arise at all. Jeblad (talk) 09:54, 16 November 2016 (UTC)

Something along those lines could work as a more general solution. However for now the only thing the formatter URL code within wikibase knows anything about is the full id string ($1). On the other hand, it turns out anything running on the tools server already hides external IP addresses via a proxy in front of everything, so there is not an actual problem in this case (the log files contain only a 10.68.* internal IP address). ArthurPSmith (talk) 15:31, 16 November 2016 (UTC)

QuickStatements ignores statements

Is it just me, or currently QuickStatements has severe bugs?

In particular, it does not correctly set P31 properties, despite showing the log pretending it succeeds.

I reported these bugs on Bitbucket, but I guess Magnus is busy right now? Or am I the only one experiencing these problems? Any similar tool I could use? Thanks! Syced (talk) 09:41, 16 November 2016 (UTC)

Maybe you should read the instructions, you're using unsupported SPARQL markup. Sjoerd de Bruin (talk) 09:49, 16 November 2016 (UTC)

Our encyclopedia as data contributor

Hi everyone,

I would like to contribute data from the project "1914-1918-online. International Encyclopedia of the First World War" (link: http://encyclopedia.1914-1918-online.net), which is an English-language online reference work on World War One. Launched in October 2014, the multi-perspective, public-access encyclopedia is a collaborative project by the largest network of WW1 researchers worldwide, spanning more than 50 countries. 1914-1918-online features innovative navigation and search functions based on semantic wiki technology and will eventually contain more than 1,600 articles.

All metadata available in the encyclopedia is licensed under the CC0 Public Domain Dedication. The Creative Commons (CC) License (“No Rights Reserved”) allows you to enhance and reuse the metadata for any purposes without restriction. The datasets can be downloaded in the RDF/XML format, each article within the encyclopedia features an RDF/XML download. Moreover, a BEACON index is also provided: http://encyclopedia.1914-1918-online.net/beacon/1914-1918-Online.beacon.txt If this data is of interest for Wikidata, what will be the next step to contribute?

Thank you very much!

Kind regards,

ChrisOES (talk) 11:18, 16 November 2016 (UTC)

Wikidata:Data donation has some useful links and guidelines on this. It's possible your encyclopedia uses an identifier for encyclopedia items, so one useful step might be requesting a specific property to link wikidata items to your articles for example. Then the actual linking could be done via the Mix n Match tool. Or perhaps there are existing properties that you have information for that could be imported that way or via QuickStatements. Good luck! ArthurPSmith (talk) 15:21, 16 November 2016 (UTC)

What is a link was made from "A" to B?

I've gotten numerous notifications about it.  – The preceding unsigned comment was added by MechQuester (talk • contribs) at 10. 11. 2016, 19:19 (UTC).

Someone used an item you had created in a statement. You will get more information after clicking on them. If you don't want to get these notifications (I recommend to keep it), you can opt-out at Special:Preferences#mw-prefsection-echo. Matěj Suchánek (talk) 20:42, 10 November 2016 (UTC)

ah ok. Thank you very much. MechQuester (talk) 03:23, 13 November 2016 (UTC)

@MechQuester: “I recommend to keep it”: it's more annoying than useful if you have created an item for something like a Creative Commons license. :( --AVRS (talk) 13:43, 14 November 2016 (UTC)

• @AVRS: hopefully Phab:T46787 will lead to the ability to choose not to receive notifications for some items (in my case lap (Q26484625)). Thryduulf (talk) 21:45, 16 November 2016 (UTC)

Language support

Wikidata has English (en), "British English" (en-gb; superfluous, same as en), and Canadian English (en-ca). But it does not support other variants of English, notably American (en-us), but also Australian (en-au), Indian (en-in), New Zealand (en-nz), South African (en-za) and perhaps others. Many plants and animals have different names in one or more of these languages, but until they are enabled, it is not possible to indicate usage. I already added a note about this on the 'Contact the development team' page, but it is getting bogged down in semantics with no action to create the required language codes. Could these be added, please? Thanks! MPF (talk) 10:51, 16 November 2016 (UTC)

The discussion referred to above is at Wikidata:Contact the development team#Language support. The first reply there includes "here's the process to add new language codes". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:08, 16 November 2016 (UTC)

Yep, I know about the guidance on adding new language codes; unfortunately, it's well beyond my technological abilities, so I'm asking someone else with the know-how to do it, please ;-) MPF (talk) 15:19, 16 November 2016 (UTC)

I doubt your assumption en-gb is the same as en is correct. --Succu (talk) 21:53, 16 November 2016 (UTC)

@MPF: It' easy, really, you fill the request via the link Pigsonthewing give you. Like I just asked for Abenaki (Q197936) (see. phab:T150633) two days ago. --Fralambert (talk) 04:58, 17 November 2016 (UTC)

So, there's a few concerns I have here. First, as Succu notes, en is not necessarily en-gb - "en" is effectively the consensus mixture of all the variants, much like the "messy en" we use on Wikipedia. But that's a small issue. More generally, I think adding more variants for English is a bad idea. On the face of it, it's a great idea. It's good to be able to define that P462 is colour and not color, that Q682 is sulphur and not sulfur, etc. But let's look at the effect of them just now...

As a demonstration, picked arbitrarily because I was working on it earlier today, there are 16981 items with Encyclopædia Britannica Online ID (P1417). All but one, 16980, have an en label; 3705 have en-ca labels; 3879 have en-gb labels. 167 have a description in en-ca; 541 in en-gb; 15084 in en proper. So in 'en', virtually all items have labels, and 89% have a meaningful description. In en-gb, it's 23% with labels and 3% descriptions; in en-ca, 22% and 1% respectively. As a result, these items are much less usable for someone using en-XX language codes. The same issue applies to properties - 3218 properties have en labels, but just 931 en-gb and 249 en-ca. Users are told they can use en-gb or en-ca, but get a much worse experience if they do.

We can get round this, to a degree, with language fallbacks ("display en if en-XX not known"), but there's also a linked problem of fragmentation. If someone corrects a label/description in en, it's very unlikely they'll remember to look for and update en-gb or en-ca (or vice versa). This means that there's an increased chance for the labels and descriptions to drift apart - which is especially a problem for properties. sex or gender (P21) is given as having five standard values for people in 'general' and British English, but only three in Canadian, for a minor example. member of (P463) has an important caveat in en, but none in en-gb or en-ca. These are relatively small issues, but they could easily get complicated.

To me, these are signs that we simply don't have the capacity to maintain the two variants we have at the moment, much less maintain four or five more. We'd be trading a small benefit in a minority of contexts (fixing occasional spelling issues) - and believe me, I personally value that benefit, I am sitting in London and I like seeing colour and petrol - for a much bigger long-term maintenance problem across the board.

So, are there better ways we could do this? For example... could we find a way to tag a specific alias as belonging to a certain dialect? eg/ sulfur (Q682) could have the alias 'sulphur' explicitly tagged as en-gb, plus one for 'sulfur' tagged en-us. This would require a bit of coding work, but it might allow us to resolve the problem without adding a lot of language codes. Thoughts? Andrew Gray (talk) 22:56, 16 November 2016 (UTC)

The Wikidictionary integration is in the works and provide a new way. ChristianKl (talk) 23:15, 16 November 2016 (UTC)

Friendliness ≈ kindness?

How is best to deal with friendliness and kindness? At present, there are three items friendliness (Q16519308), kindness (Q488085) and friendliness (Q17249291). Either they belong together, or not – de.wikiquote has entries for both of the latter two items, so I don't know what's best to do with the first. If someone with better language abilities than myself could have a proper look at this, or is simply more courageous, I'd be very grateful! Jared Preston (talk) 15:26, 16 November 2016 (UTC)

At least as far as the English meanings go, kindness has a moral quality that is not necessarily there for friendliness; friendliness has a comradeship quality that is not necessarily there for kindness. I.e. I would argue they are different concepts. How you map other language versions onto them I do not know, but I'm guessing friendliness (Q16519308) and friendliness (Q17249291) should be merged, and some of the wikilinks from kindness (Q488085) moved. ArthurPSmith (talk) 15:38, 16 November 2016 (UTC)

I think the English word "Kindness" is nearer to the German word "Freundlichkeit" than "Friendliness" is. But I don't know the way the words in the other languages work. ChristianKl (talk) 22:46, 16 November 2016 (UTC)

I think friendliness is more related to benevolence (Q16509717). --Succu (talk) 23:15, 16 November 2016 (UTC)

Wikidata Content Schema

I have a proposal for a Wikidata Content Schema, a concept for representing different data models on Wikidata. Please comment on the linked page. Harej (talk) 00:56, 13 November 2016 (UTC)

• It reminds me of Wikidata:Property_proposal/Archive/40.
  --- Jura 08:30, 13 November 2016 (UTC)
  • Jura, the main difference is that those are constraints on the use of properties. This is more like constraints on items. For example, if something is instance of (P31) human (Q5), then this system could be used to describe the required properties for the item. It is not always appropriate to use property constraints, since properties can be used in multiple domains of knowledge. Harej (talk) 22:46, 13 November 2016 (UTC)

@Harej: Perhaps next time could you start a RfC before creating a new project page: we have plenty of project pages which stay at a proposal level and this is just confusing for contributors looking for "official" guidelines. Snipre (talk) 22:24, 13 November 2016 (UTC)

• What you are suggesting is mainly done by Wikiprojects. These provide guidance on how to use properties for items in specific fields.
  --- Jura 11:01, 17 November 2016 (UTC)

P27 Country of citizenship

Hi, can anyone help out with an area of confusion for me as I'm looking to make sure Wikidata items for Scottish subjects are correctly identified as such. Case in point: I'm trying to create a timeline of female Scottish architects and Wikidata I've always thought could cope with multiple values for a property (e.g. where there are multiple occupations for instance) but the field for nationality/country of citizenship (P27) where I have added a few edits to add 'Scotland' underneath 'UK' have been reverted. My thinking being that it is perfectly legitimate for someone to be both British and Scottish at the same time especially as P27 covers nationality as well as citizenship. I realise the key word 'citizenship' is more of a legal thing but the nationality you identify as is where I'm seeking clarity and the UK setup is perhaps a little unusual in this respect. There are at least 1200 other items where P27 has Scotland as a value -are they all wrong? If P27 is not the right property then please let me know which one is... if it's P172 (Ethnic group) the fact that there are only 45 items with Scotland as that value should tell you where the weight of opinion currently lies. Any clarity would be appreciated. Cheers. Stinglehammer (talk) 11:34, 17 November 2016 (UTC)

Since the Acts of Union 1707 (Q193515) in 1706-1707 were enacted there is no such think as a Scottish citizenship (or an English citizenship for that matter). You can belong to one of 4 constituent country of the United Kingdom (Q3336843) ethnically and nationally but only one citizenship exists. DGtal (talk) 14:07, 17 November 2016 (UTC)

I agree that there's no Scottish citizenship. Ethicity (P172) seems more fitting. As the description suggests it's a property that needs sources. The fact that it's not widely used just shows that nobody was interested in adding data. Maybe a new property is also warranted. ChristianKl (talk) 10:40, 18 November 2016 (UTC)

Historical names / Aliases

Probably I'm not the first with this problem, but I cannot find a solution for this problem. Organizations, teams, people change sometimes their names during their existence. Is it possible to display the correct name for the time this name was the valid one in Wikidata statements? Example: In Q705485 on rank 7 is listed KFC Uerdingen 05, but during this time the correct name for the team was FC Bayer 05 Uerdingen. The same problem occurs for all women when changing their name after marriage. --Florentyna (talk) 11:52, 17 November 2016 (UTC)

How about using official name (P1448) and adding start and end dates? DGtal (talk) 14:09, 17 November 2016 (UTC)

A good starting point, but I'm afraid that it will not work like expected. Two major goals I think must be reached: First, the displayed information on Wikidata should be correct (in the mentioned example there was no KFC Uerdingen 05 during this time), and second, if somebody wants to create an article based on the Wikidata item, the article will contain these wrong information - and this should be avoided. --Florentyna (talk) 16:42, 17 November 2016 (UTC)

One problem with "official name" is that some names are widely used, without being "official". The capital of Austria is named one thing in German, another thing in English. But the English name is, as far as I know, not "official". Very few places outside of Scandinavia, Finland and the Baltics have Swedish names, but German names are widely used in Sweden about places in large parts of Europe. Places in Finland often have an "official" Swedish name, but they have sometimes fallen out of use. -- Innocent bystander (talk) 19:35, 17 November 2016 (UTC)

"native label" is an alternative, but I think DGtal suggestion can work out. One just needs to learn how to make use of the statements. Obviously, it can't work if the statements aren't there. If it's not merely a change of name, separate items for previous organizations should be made.
--- Jura 09:35, 18 November 2016 (UTC)

Soon we will have a Wikidictionary integration for how concepts are named. It might be worthwhile to bring this issue up in that discussion. ChristianKl (talk) 10:42, 18 November 2016 (UTC)

FIFA country codes

Do we have a property for FIFA country codes? Xaris333 (talk) 20:17, 17 November 2016 (UTC)

Here's a search for the string "P:FIFA" (without quotes), which is how you can query property labels. It suggests not. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:57, 18 November 2016 (UTC)

Ok. I will suggest it. Xaris333 (talk) 14:24, 18 November 2016 (UTC)

P2656

P2656 (P2656) is for the current ranking. How can we show the highest and the lowest ranking a national team had? Xaris333 (talk) 20:35, 17 November 2016 (UTC)

MIN() and MAX() ?
--- Jura 09:36, 18 November 2016 (UTC)

How to use it? For example, how can I use it for the lowest (25th, March 1998) rank for Spain men's national football team (Q42267). Xaris333 (talk) 14:23, 18 November 2016 (UTC)

Add the value for that specific (min/max)date, and let a template/module in the client compare all numbers and tell which was lowest/highest and at which date. - Innocent bystander (talk) 17:47, 18 November 2016 (UTC)

Use of Q7901319 in P364

• Q7901319 is Kanbun in most languages, however what the wikidata item and the corresponding wikipedia article is describing is actually "kanbun converted into literary Japanese", i.e. a conversion mechanism, or a converted result
• The actual wikidata code/wikipedia article for Kanbun text is Q37041, i.e. lzh.
• P364 is "original language of work", And there are a number of old Japanese works written in lzh being tagged as Q7901319, for instance Q813935, while those Japanese words are written in lzh without any conversion. Those conversion are applied by later readers.
• I think the P364 of those articles should be changed to Q37041 but I cannot determine how many wikidata item would be influenced and I don't think I can review each of them manually...
• C933103 (talk) 14:09, 18 November 2016 (UTC)

Donation of thesis data from University of Oxford

Hi, in my role as Wikimedian In Residence at the University of Oxford, I've been given a data table of 8,243 theses from the Oxford University Research Archive (ORA), dating from the 1960s to the present day. More than 3,000 of these have the full text available openly online. I'm thinking that this could advance the goals of the Wikicite project. Some of these are theses by notable people (such as en:Richard Gombrich and en:Ronald Hutton) but of course many of the authors are not notable in their own right. All entities are of type doctoral thesis (Q187685). The data table includes title (label), abstract (which I realise is not suitable for WD), author name (sometimes as a string, sometimes initials, sometimes separate first and last names), publication year =publication date (P577), link to entry in ORA (which about half the time has the full text) =archive URL (P1065)?. There are no identifiers associated with the author name, but to disambiguate names we have the fact that they studied for a doctoral degree at Oxford, and their degree was accepted in a particular year. I assume this is valuable as bibliographic data, and welcome advice on the best way to go about importing it. Thanks in advance for any help. MartinPoulter (talk) 16:02, 17 November 2016 (UTC)

is there some sort of persistent identifier for the doctoral theses that falls under our existing external ids in wikidata, such as a DOI? I would suggest to import each thesis as an item with instance of (P31) doctoral thesis (Q187685) and include publication date (P577), title (P1476) and archive URL (P1065) (and a DOI or other persistent ID if one exists), and then attempt to link with author (P50) to items if those individuals exist in wikidata, or else leave just author name string (P2093). ArthurPSmith (talk) 20:45, 17 November 2016 (UTC)

@MartinPoulter: If the authors have a Wikipedia entry (or any authority control data) they qualify for an entry in Wikidata. May you please dump the raw list of names in English Wikipedia? We could collaboratively do the disambiguation work there.--Kopiersperre (talk) 09:24, 18 November 2016 (UTC)

Thanks User:ArthurPSmith and User:Kopiersperre for valuable suggestions that I will act on. I take it QuickStatements is the best way to do this kind of import? MartinPoulter (talk) 14:29, 18 November 2016 (UTC)

@MartinPoulter: Are you planning on putting digital copies of the theses themselves on s:? —Justin (koavf)❤T☮C☺M☯ 21:04, 18 November 2016 (UTC)

@Koavf: Ideally I would, but I've no time to do that. Structured links between ORA, WD and WP will enable the whole process, though. Cheers, MartinPoulter (talk) 14:55, 7 December 2016 (UTC)

P118

league or competition (P118). I am using this properties in sports teams to show the league they are playing this season. For example,

⟨ Leicester City F.C. (Q19481)    ⟩ league or competition (P118) ⟨ 2016–17 Premier League (Q23009701)    ⟩

Is that correct? I have found many sport team's items that are using P118 for the general league. For example,

⟨ Leicester City F.C. (Q19481)    ⟩ league or competition (P118) ⟨ Premier League (Q9448)    ⟩

Which of these are correct use of P118? Is there another way to show the current season?

Xaris333 (talk) 19:11, 17 November 2016 (UTC)

I'd guess the league is the persistent organisation through the year, and that a season is a specific competition. But I don't know if anyone can actually tell ... Is there a corresponding WikiProject for sport ? We should create it : WikiProject Sport. author  TomT0m / talk page 10:14, 18 November 2016 (UTC)

There is Wikidata:WikiProject Sports. However, we can learn from the constraints on Property talk:P118 that season items are not allowed as values, but league items are. —MisterSynergy (talk) 10:23, 18 November 2016 (UTC)

And how can we add the information about current season? Is an information that is usually used in Wikipedia templates and is really useful. About P118 why to only use it with league items? If I read item Leicester City F.C. (Q19481), I preffer to learn that is taking part specific in 2016–17 Premier League (Q23009701) than in general Premier League (Q9448). Xaris333 (talk) 14:29, 18 November 2016 (UTC)

You could use participant in (P1344) line on Arsenal F.C. (Q9617). --Pasleim (talk) 15:00, 18 November 2016 (UTC)

participant in (P1344) is for "event a person or an organization was a participant in". For the past. Not current season. You can add all previous seasons. I am speaking about current season. Xaris333 (talk) 15:14, 18 November 2016 (UTC)

participant in (P1344) should be ok. Contrasts nicely with participating team (P1923) used in 2016–17 Premier League (Q23009701) . If its a previous or current event will be determined by the start time (P580) and end time (P582) of the event - 2016–17 Premier League (Q23009701) . league or competition (P118) can be used to represent the overall league with start time (P580) and end time (P582) to qualify the duration they were active in that league. Unnited meta (talk) 18:55, 18 November 2016 (UTC)

But with participant in (P1344) you could also add the cup season and the the international cup season. Too complicate... Xaris333 (talk) 19:27, 18 November 2016 (UTC)

YouTube channel ID (P2397)

YouTube channel ID (P2397) is for channel with formatter url https://www.youtube.com/channel/$1. What about user with formatter url https://www.youtube.com/user/$1? Xaris333 (talk) 13:31, 19 November 2016 (UTC)

If you view any video on the channel and click on the channel name below it, the URL will be in the /channel/ format. --Entbert (talk) 14:03, 19 November 2016 (UTC)

Thanks! Xaris333 (talk) 16:07, 19 November 2016 (UTC)

Properties for a template

Hello. Using parameters of en:Template:Infobox football league: how can we show with a property on an item page:

• first  : The first season
• levels  : The level that the league sits within that country's football pyramid.
• domest_cup  : Names of domestic cups
• league_cup  : Names of league cups
• champions  : Name of most recent champion
• season  : Years of most recent champion

You can see Cypriot Second Division (Q2186582) as an example. I have added as many properties I can.

Xaris333 (talk) 00:23, 20 November 2016 (UTC)

Why is VE hidden in the editing bar using an unlabelled pencil icon

Sorry if I've missed something but why has Wikidata switched to having the visual editor button as unlabelled pencil icon in the editing window? I doesn't seem to be in any of the documentation and is super hard to find.

Thanks

--John Cummings (talk) 16:15, 9 November 2016 (UTC)

What is the "Editing bar" you're talking about? Do you mean the editing bar what is shown here project chat as edit mode? Or at item pages? --Stryn (talk) 18:28, 9 November 2016 (UTC)

Hello @John Cummings:, if I understand correctly, when you edit a talk page, the button to switch to Visual Editor has changed to an icon without label. That's what I see too. I have to admit I don't remember how it used to be before, let me check the recent changes about the user interface. Lea Lacroix (WMDE) (talk) 08:50, 10 November 2016 (UTC)

Hi John,

That pencil button is labelled on all pages I've checked, on all wikis: title="Switch to visual editing". That feature exists since a while now. Maybe you haven't noticed it yet? Is it the same button?

Or is may be related to your configuration. Do you have the same issue on all wikis? If yes, do you use a recent browser? If that's only on Wikidata, you may have an unexpected issue with a script or a gadget; that will require more investigation.

Best, Trizek (WMF) (talk) 10:12, 10 November 2016 (UTC)

Hi all, sorry for the confusion, I'm using Windows 10 with with latest version of Chrome and Firefox and see the both on each. What I mean is I only see one Edit button at the top of any non Q or P page I go to. When I click on that I get Source editor and there is small pencil icon in the toolbar that switches to VE. I haven't changed any skins or done any plug ins on my account or anything, plain vanilla. What I'm confused about is why this is VE hidden away like this and not an 'edit' and 'edit source' button to press to edit a page like I see when I edit other Wikimedia sites. I seems very strange to have VE which will lower the technical bar for new contributors (and make it quicker to edit for non new contributors for many tasks) but then hide it and not document where it can be found. I just wanted to understand why it was like that.

Thanks

--John Cummings (talk) 12:48, 10 November 2016 (UTC)

Ah, I see; that's different. :)

The visual editor is still considered as a Beta feature on Wikidata. That's why click on "edit" launches the wikitext editor. There is then a shortcut to use VE.

If the community asks to have the visual editor on Wikidata as a default editor (in parallel of the wikitext one, that will stay), users will be able to select their preferred editor in their preferences and switch between VE and wikitext as it is already possible with that pencil icon.

That's a community decision to make! Trizek (WMF) (talk) 13:35, 10 November 2016 (UTC)

i made some similar icon comments in VE feedback. depending on the talk page it will default to wikitext editor requiring the toggle over. we spend a lot of time tutoring newbies about the magic pencil icon. very un-intuitive. they also changed the default to toggling between editors. you can change to 2 buttons in your preferences, editing tab. leave your feedback here https://www.mediawiki.org/wiki/VisualEditor/Feedback . Slowking4 (talk) 00:06, 17 November 2016 (UTC)

HiSlowking4, I'm not sure to understand your comment. Preferences on editing tab are not available on every wiki. If you don't enable the visual editor on Wikidata (in the Beta tab), you will not have that two-tabs setting. And what do you mean about talk pages? Trizek (WMF) (talk) 16:50, 17 November 2016 (UTC)

the magic pencil icon is the same bug / feature across all wikis. VE does not work on talk pages other than this one (and your indent does not work) - this is a major problem / how do you do to newbies at editathons - we are perpetually telling newbies to sign up at the editathon / required to use wikitext, and then switch to VE. i would just as soon stop. when might that be? see also [9] Slowking4 (talk) 03:40, 21 November 2016 (UTC)

Language of a url source

Using reference URL (P854) and archive URL (P1065) for sources, shouldn't be able to show the language of the content of the url page. We can't use a qualifier (like language of work or name (P407)) for a source. Xaris333 (talk) 13:45, 19 November 2016 (UTC)

Who says we can't? Sjoerd de Bruin (talk) 12:59, 20 November 2016 (UTC)

I agree. I add it all the time as a seperate field lik I add retrieved (P813), publication date (P577), title (P1476) etc. DGtal (talk) 13:02, 20 November 2016 (UTC)

Ok. I asked because I thought that qualifier was only about the object of the property, not about another qualifier. Xaris333 (talk) 15:16, 20 November 2016 (UTC)

Princely states Q1336152

Hoi, at this time princely states are associated in relation to the British Raj. This is wrong on several levels. They are former countries and once they became part of the Raj, they kept much of their ways intact. I think at most they all may be qualifiers for this to be the "instance of" and have them as country. Thanks, GerardM (talk) 06:38, 21 November 2016 (UTC)

Encyclopædia Britannica's 'empty-ish' concepts in Mix'n'Match

Encyclopædia Britannica is one of the datasets in Mix'n'Match, where it helps fill the Encyclopædia Britannica Online ID (P1417) property. However, the dataset in Mix'n'Match contains many links to concepts that don't have a full page on EB yet, just what they call a 'directory page' (a more or less empty-ish page referring to the concept and how it is used in other articles - example 1, example 2, example 3). I have seen several users who removed these specific matches from Mix'n'Match and even remove those statements from Wikidata items. What do people think of that? I think we should keep them, as it's IMHO quite useful to have as many links as possible to EB's concept cloud, even if they don't have full articles, but that's my POV. Ideas? Spinster 💬 23:13, 13 November 2016 (UTC)

Such behaviour is disruptive, and should cease. P1417 is "Encyclopædia Britannica Online ID", not "Encyclopædia Britannica articles over a certain length". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:46, 14 November 2016 (UTC)

Thanks Andy, good to hear you agree with me. Anyone else who has an opinion on this? I'd ideally like to make a note on the property's talk page based on consensus, so that we can then point people to that. Spinster 💬 17:43, 20 November 2016 (UTC)

I think we should not link to EB "stubs".--Kopiersperre (talk) 09:56, 21 November 2016 (UTC)

Why not? They are not different from topics in other concept databases we link to, for example the Art and Architecture Thesaurus. Identifiers are about linking topics, not articles. Spinster 💬 10:07, 21 November 2016 (UTC)

Quite a few of these 'stubs' are not much better than spam, certainly not an asset. Unfortunately there are quite a few 'databases' out there (and linked to in Wikidata) that I have not been able to find any useful content in, at all. - Brya (talk) 11:45, 21 November 2016 (UTC)

The "asset" is the identifier (the property is labelled "Encyclopædia Britannica Online ID" for a reason), not the content of the linked page. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:06, 21 November 2016 (UTC)

• I think the identifier is meant for encyclopedia articles. If there isn't any, the property shouldn't be used.
  --- Jura 11:59, 21 November 2016 (UTC)

Such thoughts are neither supported by the property label (see above), nor the discussion of the property's proposal, which noted its use as a form of authority control. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:06, 21 November 2016 (UTC)

Just ignore the property description, but you might want to add that it isn't supported by what someone said on 18:46, 14 November 2016 here. You would be perfectly right about that.
--- Jura 13:17, 21 November 2016 (UTC)

Tomatoes

I found two items about tomatoes and i wondered if they are the same. It is about: tomato (Q20638126) and tomato (Q23501). Are they really the same? QZanden (talk) 13:50, 14 November 2016 (UTC)

tomato (Q20638126) is the fruit of the plant tomato (Q23501). --AVRS (talk) 14:32, 14 November 2016 (UTC)

Thanks, I clarified it in my own language (dutch). Q.Zanden questions? 13:26, 21 November 2016 (UTC)

Copy references tool still broken

Tracked in Phabricator
Task T151027

The copy references gadget is still broken after the last bug fix was deployed on it. Can you please either revert the last "bug fix" or fix the gadget? ChristianKl (talk) 10:08, 17 November 2016 (UTC)

Can you please specify what exactly is "broken"? Sjoerd de Bruin (talk) 10:29, 17 November 2016 (UTC)

Hello @ChristianKl:, I'll investigate on this, but I need more information about what is working wrong, what did you expect and how it's behaving differently from expected. Thanks, Lea Lacroix (WMDE) (talk) 11:26, 17 November 2016 (UTC)

• The behavior of fixing on bug by introducing another, I would advocate that the person who looks at the tool again tries to use it in different ways to make sure it works and doesn't just focus on the issue noted. Having unit tests might also make sense.

But what's the problem I'm having? I add one statement A. Then I add a reference R to it. Next I add another statement B. Then I click on copy of R and try to paste it to B. That doesn't work but produces an error. ChristianKl (talk) 11:38, 17 November 2016 (UTC)

I think you have to reload the page to be able to "paste". I do not think we ever have been able to paste to a new statement.

One problem I see now is that the "copy"-links are not present every time I load a page. -- Innocent bystander (talk) 11:45, 17 November 2016 (UTC)

Hm, I think it has to do with other scripts (thus blocking the loading of further Javascript), the following error appears in my console: TypeError: undefined is not an object (evaluating 'transData[ lang ]'). Sjoerd de Bruin (talk) 15:06, 17 November 2016 (UTC)

I remember the ability to do this without reloading the page before the changes broke the tool 1-2 months ago. ChristianKl (talk) 15:30, 21 November 2016 (UTC)

It was never possible to paste references to new statements without reloading. Sjoerd de Bruin (talk) 18:08, 21 November 2016 (UTC)

WikiProject as a namespace

Maybe it has already passed by here, but why is there no namespace for the WikiProjects? It would be much easier and the links would be shorter... Q.Zanden questions? 14:49, 21 November 2016 (UTC)

Wikidata weekly summary #236

Here's your quick overview of what has been happening around Wikidata over the last week.

• Discussions
  • New request for comments: How to make new languages enabled on Wikidata

• Events/Press/Blogs
  • Problematizing and Addressing the Article as Concept Assumption in Wikipedia (Q27827559)
  • schema.org, Wikidata, Knowledge Graph: strands of the modern semantic web by Dan Scott

• Other Noteworthy Stuff
  • Parlement & Politiek ID (P1749) has been completed in Mix'n'match. Over 5000 politicians from the Netherlands (and also Belgium) now have a link to this resource
  • Two policies about abandoned tools on Tool Labs are currently voted on
  • The Community Wishlist Survey is now closed and the vote phase will start on November 28. Here's the category for Wikidata
  • Data import hub and Data import guide: feel free to give feedbacks on these documents!

• Did you know?
  • Newest properties: musipedia tune ID, Adult Film Database person ID, WHO international non-proprietary names ID, designed to carry, National Library of Greece ID, PermID, HKMDb person ID, RxNorm CUI, Vote Smart ID, legislation.gov.uk ID, significant person, Kvikmyndir person ID, Kvikmyndir film ID, World Surf League ID, Encyclopedia of Surfing ID
  • Query examples:
    • Occupations of women, with French female-form labels (source, database report)
    • Documents with most signatories (source)
    • Filming locations of James Bond films (source)
    • United States National Historic Landmarks not actually located in the United States (source)
    • Items with VIAF, but no P31/279 (source)
  • Newest database reports: List of Italian language films without articles in Italian Wikipedia

• Development
  • Unit conversion has been rolled out for the first dimension for a short while now.
  • Undo diffs now utilize language fallbacks for property and item labels (gerrit:315291). Thanks to Matěj Suchánek!
  • Reworked database schema for Cognate (the extension that will do automatic interwiki links for Wiktionary). It is still waiting for security review before it can be deployed.
  • Worked more on basic version of Lexeme entity type
  • Fixed a bug where the links on action=info where wrong (phabricator:T149598)
  • Getting close to finishing the groundwork for Federation so that in the future you'll be able to use Wikidata's items and properties to describe files on Commons.
  • Added Innu-aimun as a language for monolingual text values (phabricator:T151129)
  • Worked on linking to Wikipedia articles in the statements on an ArticlePlaceholder (phabricator:T113955)
  • ArticlePlaceholder got a button to translate articles via the ContentTranslation extension (phabricator:T124036)
  • Prepared a bot to fix quantity values after the recent changes to precisions (phabricator:T142087)
  • Did user interviews for use cases and workflows on Commons
  • Analyzed more Listeria queries in order to figure out steps forward for easier query writing

You can see all open tickets related to Wikidata here.

• Monthly Tasks
  • Hack on one of these.
  • Help write the next summary here!
  • Contribute to a Showcase item
  • Help translate or proofread pages in your own language!
  • Help merge identical items across Wikimedia projects.
  • Add labels, in your own language(s), for the new properties listed above.
  • Comment on property proposals: all open proposals - proposals needing attention

Read the full report · Unsubscribe · Lea Lacroix (WMDE) 15:08, 21 November 2016 (UTC)

Polymers, fibers and fabrics

I am looking at the list of fabrics and was looking at improving their classification. That leads to https://en.wikipedia.org/wiki/Synthetic_fiber so I am wondering if Wikidata items are needed for each or some of these:

• polymers, e.g. Nylon
• fibers, e.g. Nylon fiber (I have added this but wanted to discuss before doing more)
• fabric

Also seeing a lot of these treating a specific fabric type as an "instsance of" textile. IMHO they should be subclasses, so I have been changing some of those.

Thoughts? Pauljmackay (talk) 09:25, 21 November 2016 (UTC)

An item for "fabric" would seem to be a good idea, and at least some cleanup on related items. - Brya (talk) 11:51, 21 November 2016 (UTC)

There is already an item for textile (same as fabric really) for the general item about fabric. I was asking if there should be new items for each type of fabric for different material types, particularly the synthentic ones? Pauljmackay (talk) 10:54, 22 November 2016 (UTC)

Well, fabric is not necessarily the same as textile, which is why a separate item would be useful. - Brya (talk) 12:14, 22 November 2016 (UTC)

How closely related should images included at P18 be?

I had a discussion here about how closely related should images included at P:18 be. I think that some of the articles would never have an free alternative, like song covers, so the next best alternative should be using an image of an singer while performing a specific song or a free image that resembles the cover. Some of the images also have the costumes which are used in the video clip of the song or on the cover, because there is no way you can use the cover of the song since it's copyrighted. Also, in the case for lists, where collages cannot be made due to copyright issues or missing images, an image related to that list should suffice. At Romanian Wikipedia, where I am an administrator, we use https://ro.wikipedia.org/wiki/Modul:Pagina_principal%C4%83 which takes it's images from Wikidata.Ionutzmovie (talk) 14:37, 22 November 2016 (UTC)

Another question of mine is: what could be used as a country's image? A collage with the most important cities?Ionutzmovie (talk) 14:37, 22 November 2016 (UTC)

• There is a specific property for collages, generally more suitable for those looking for that. As for lists, in general, I wouldn't add an image to an item about a list. An exception might be File:Liste des maires de Bordeaux2.jpg suitable for list of mayors of Bordeaux (Q3251116). Reasonator can compile a gallery on the fly if needed: https://tools.wmflabs.org/reasonator/?&q=27898073
  --- Jura 14:57, 22 November 2016 (UTC)

Data import hub, data preperation instructions and import workflow for muggles

Hi all

Myself and NavinoEvans have been working on a bare bone as possible workflow and instructions for making importing data into Wikidata available to muggles like me. We have written instructions up to the point where people would make a request on the 'bot requests' page.

Please take a look and share your thoughts

https://www.wikidata.org/wiki/User:John_Cummings/Dataimporthub

https://www.wikidata.org/wiki/User:John_Cummings/wikidataimport_guide

Thanks very much

--John Cummings (talk) 18:19, 20 November 2016 (UTC)

Quite timely since i've just been poking around for some documentation of how one would need to go about an import. This is looking great, nice work! The data I have is a spreadsheet of place label translations, do you think this might be a good candidate to test this workflow with? --Planemad (talk) 17:19, 22 November 2016 (UTC)

Planemad sorry to miss your message before, sure please go ahead, it will be very helpful to surface any issues. Probably best to ask here as a new topic if you have any questions so that multiple people can join in. Thanks very much --John Cummings (talk) 13:41, 23 November 2016 (UTC)

Wikidata Logo - so very close to a barcode, but not quite

I was playing around with a Code-93 barcode (Q2981417) generator and I realized that it would only take a little bit more space to have the WikiData logo be a Code-93 barcode for "W".

Could the Wikidata logo be changed to a valid barcode? I think this would be very cool. 144.62.220.200 00:36, 21 November 2016 (UTC)

I don't know enough about Code-93 barcodes to know what that would look like, but the current logo is "WIKI" in morse code, so if making it a barcode would change that, I don't think it would be a good idea. --Yair rand (talk) 04:13, 21 November 2016 (UTC)

Every code made up from bars is a bar code. Various standard codes exist, with Code 93 being just one of them. Using the morse alphabet as a bar code is completely uncommon, but historically, the morse alphabet is where the idea for bar codes comes from, and the very first bar code ever to be drawn actually was a morse bar code. So the Wikidata logo is a bar code. It's an uncommon one, but it's a nice historic reference. Changing it to any other code would not improve it in my eyes, as it could never be the one correct code. Also note that in the original voting for the Wikidata logo, its morse code nature was clearly mentioned. --YMS (talk) 12:15, 23 November 2016 (UTC)

Do you know the wikidata-morse-code-renderer? --Atlasowa (talk) 21:43, 21 November 2016 (UTC)

• 
  Wikidata logo: "WIKI" in Morse code
• 
  The word "Metadata" in Wikidata Morse code
• 
  Wikidata Birthday Creating Morse Code

Infinity

You address the concept of larger than large in relation to infinity. However, no reference is made to smaller than small which is infinity in reverse. Both concepts appear to beyond human comprehension.

Hmm, there is maths formulas and definitions that defines this - for infinitey big set theory is an help (I know definitions like "an infinite set in a set for which there exists a map from a strict subset of himself with himself) But it's not so easy to model in Wikidata. And there also exist way to model mathematically "infinitely small" numbers (see infinitesimal (Q193885)   ) But it's the same, these are not really easy to define in Wikidata. author  TomT0m / talk page 15:00, 22 November 2016 (UTC)

not in reverse, same cardinality. see also w:Controversy over Cantor's theory. Slowking4 (talk) 21:50, 23 November 2016 (UTC)

Press coverage

Wikidata:Press coverage stops at December 2014. Is there a more up-to-date resource? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:22, 22 November 2016 (UTC)

Issues of WD:Status updates. Matěj Suchánek (talk) 13:18, 23 November 2016 (UTC)

Freebase IDs in the wild at Google Arts & Culture

Today I stumbled upon the Google Arts & Culture (Q27905940) website, created by Google Cultural Institute (Q16927659) (which, as an aside, has in itself subsumed most of Google Arts & Culture (Q369089)). I noticed the website is (re-)using Freebase IDs for their URLs. E.g. Edward Hopper (Q203401) with Freebase ID /m/0hc3t has a page at entity/m0hc3t (notice the slash has been dropped). There are pages for events as well, e.g. World War II and (partner) institutions (e.g. https://www.google.com/culturalinstitute/beta/partner/stichting-europeana).

Now this seems like a worthy resource to link to (if they manage to keep it running for a few years longer)? Any ideas of how we might leverage our collection of Freebase IDs? --Azertus (talk) 20:43, 22 November 2016 (UTC)

Hmm .. "403. That’s an error. We're sorry, but you do not have access to this page. That’s all we know."
--- Jura 21:07, 22 November 2016 (UTC)

Got none for entity/m0hc3t, World War II and Europeana... --Succu (talk) 21:18, 22 November 2016 (UTC)

Hmm, both of you? That is weird! I just checked all three links in a private tab (so not logged into any Google account) and reached them without a problem. I have no idea what's happening... --Azertus (talk) 21:36, 22 November 2016 (UTC)

I got no 403 Forbidden (Q1138586) for the above mentioned URLs. --Succu (talk) 21:40, 22 November 2016 (UTC)

I got that by navigating through the site. As it's in Beta, not really a problem. Just maybe not suitable for linking.
--- Jura 21:43, 22 November 2016 (UTC)

Q uses licensed music

How would I include songs used in a movie or video game? Examples: Riot (Q17029444) in NHL 14 (Q13462549) or Riot (Q7335432) in WWE SmackDown vs. Raw 2007 (Q927012) or the list at w:Music of Dance Dance Revolution (1998 video game). Dispenser (talk) 21:24, 22 November 2016 (UTC)

As far as I can see, currently, there's only the property soundtrack release (P406) to link to a dedicated soundtrack album. If there is no such album, one could argue whether using tracklist (P658) would be mis-using it. For your "Music of Dance Dance Revolution" example, for me this actually sounds okay, while for e.g. some computer game item (or some movie item, even some event item, ...), this feeld a bit over the top. --YMS (talk) 11:59, 23 November 2016 (UTC)

Soundtracks sometimes have extra songs or are missing songs from the main work. Numbers have been cut from Chicago (Q189889) for pacing, while video game soundtracks sometimes cut songs to fit on a single CD. Finally, GTA: San Andreas patched out songs that Rockstar lost the rights to (list). Dispenser (talk) 12:01, 24 November 2016 (UTC)

Merge into the older item should be the (only) default choice ?

Has there ever been a discussion about this topic? The merge wizard should show 'Always merge into the older item' as default (as it is showing to create a redirect with every merge). After merging into a new item the old history is also buried deep inside with the old item. ex. Q6434202 (created on 7 March 2013) merged into Q25719804 (created on 10 july 2016) and with that edit, all the previous 156 edit history is gone with the old item.--Mrutyunjaya Kar (talk) 13:41, 23 November 2016 (UTC)

Before the redirects were implemented, there used to be some discussions whether the items should be merged to the older one or the one which was more complete/containing more important sitelinks/was probably used more widely. I was a big advocate of the latter solution, especially as this promised to lose less external links (but also like you say the history, etc.). But now that the other item isn't deleted any more and thus no links are actually lost, I don't see a reason any more why the user should be able to decide which one to keep. --YMS (talk) 14:04, 23 November 2016 (UTC)

Agreed. It is so annoying when I find articles that have been merged into newer ones instead of the opposite, and I don't understand why people keep doing this. These merges can cause problems to external applications interfacing with Wikidata, because the older and more established items suddenly become redirects, plus you lose history for no benefit. @Ebrahim: what do you think? Mushroom (talk) 23:19, 23 November 2016 (UTC)

@MKar, YMS, Mushroom: Well, my experience of handling items about villages in Sweden (Q34) is that we have had many items with only one single svwiki sitelink or one single nlwiki sitelink. When these finally have been merged, they often have been merged into the older item, which in most cases is the ones with the nlwiki-sitelink. These merges often looks awful, since the nlwiki-items have had very little of Tender Loving Care (Q7699601) from the community here. While the Swedish items have been filled with a lot of information and have probably been visited and used much more frequent than the Dutch items. Why in such cases merge into the older item, who only is older because we have a larger Dutch community here at Wikidata than we have a Swedish one? -- Innocent bystander (talk) 08:41, 24 November 2016 (UTC)

@Innocent bystander, MKar, YMS: The main problem I see is that an external application using the Wikidata Query Service will save the ID of the older item, and then suddenly when querying for that ID it will receive no results and get confused (there is actually a way to resolve redirects through SPARQL, but people may not know that; I discovered that just now). I also don't like the inconsistency, for instance most items from the Tagalog Wikipedia are empty, link to very short articles, and need to be merged. Many of them have in fact been merged, but some have kept the Tagalog ID and some the other one, in a totally random way. I would rather have a clear general rule. Mushroom (talk) 15:57, 24 November 2016 (UTC)

I would prefer to have a rule that says that the best item should be preserved. In the case of the Dutch/Swedish items, that would be the Swedish items. A third party application takes probably more interest in an item with statements like "population", "area", "name", "SCB-id" than one only telling: "instance of:a human settlement", "country:Sweden". -- Innocent bystander (talk) 17:00, 24 November 2016 (UTC)

I would agree with such a rule, i.e. keep the one with the highest number of statements, or keep the one with most sitelinks, or keep the one with the most complete associated article. It would be great if the merge gadget could select that automatically. Mushroom (talk) 17:14, 24 November 2016 (UTC)

There are still items with fairly low QIDs that don't happen to have any statements. Once in while, people come across them and merge them with items with thousands of links .. Fortunately, a bot fixes the links, but still, it would be preferable to merge these items into the item with the higher ID. Anyways, I think the first time I merged with QuickStatements, I did the entire list the other way round ..
--- Jura 17:17, 24 November 2016 (UTC)

RfC on family relation properties

There is currently an RfC on making family relation properties gender neutral. There is also a relevant property proposal on creating a P:sibling. As family relation properties are few of the more important properties here and any changes might have a significant effect on our schema, I am posting this here so that the RfC gets more community attention and input. —Wylve (talk) 09:41, 24 November 2016 (UTC)

Parliamentary terms

Each term of the UK Parliament has an item, for example: 53rd Parliament of the United Kingdom (Q21084470). Members of Parliament are member of (P463) each parliamentary term item. These parliamentary term items are instance of (P31):parliamentary term in the United Kingdom (Q21094819), which is ultimately a temporal entity (Q26907166) from its class tree. An entity cannot be a member of a temporal entity (Q26907166); they can only be a member of a class (Q18844919) or any of its subclasses. Should we duplicate every parliamentary term items into ones describing the temporal session of parliament and the organization that exists within that temporal session? —Wylve (talk) 09:47, 24 November 2016 (UTC)

Does it actually exist as a separate organization? I'm not sure if either are necessary, given that the items already have position held (P39) Member of Parliament (Q16707842) (which should have start/end dates). --Yair rand (talk) 17:46, 24 November 2016 (UTC)

• There is a qualifier for position held (P39) that should eventually hold them (parliamentary term (P2937)), but @Andrew Gray: wanted to think it over. Maybe we can move them now?
  --- Jura 17:51, 24 November 2016 (UTC)

Sorry for neglecting this - the identifiers work got in the way of working on this. Ideally yes, start-end dates seem a better idea, but we only have those on a small fraction of items at the moment. I'll have another look at where we stand on backfilling those (should be able to get good data from HoP/Hansard). Until those are in place, leaving the 'member of nth Parliament' links seems safest, as I think at least one external reuser might be using this data to group them. Will look into this over the weekend if I can & report back. Andrew Gray (talk) 18:09, 24 November 2016 (UTC)

Pseudo or alternative identity

Hi, some writer, for example, write under another identity, a french famous example is :

⟨ livre ⟩ author (P50) ⟨ Romain Gary (Q157322)    ⟩
as Search ⟨ https://www.wikidata.org/wiki/Q13218867 ⟩

.

I would want to discuss guidelines on this here, as it may affect several wikiprojects. Do we have guidelines on this ? I propose to create an help page Help:Pseudonyms to write Wikidata guidelines to provide global instructions on how to handle cases where people used an alternate identity for their work.

The idea, if we can get an agreement on this, is that we get a consistent handling on such cases so that the data will be easier to process later, and that we don't get WikiProject Books who comes up with a solution, and an incompatible solution on another project (say WikiProject Musicals) ...

On a programmer perspective - say a Wikipedia infobox coder - this would mean that the same code will be usable for both Musical and Books infoboxes.

If we can agree on these preliminar, here would be my proposal :

• Create a class for "Alternative identity" - we can see that Romain Gary (Q13218867)    has currently no instance of (P31) statement, classify those kind of items into it. we can later discuss on the classes those class is a subclass of
• Use the item of the real person in statements - Or unknown value Help if unknown
• use the as Search qualifier with the item for the alternate identity used by the person in the relevant statement.

 – The preceding unsigned comment was added by TomT0m (talk • contribs).

Discussion

 Question Do someone know about a project that already have guidelines on this issue  Question Is the proposed solution good ? WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.  Comment by @billinghurst: (on source) I would have thought that pseudonyms (to known persons) would all be handled on the author page

This would not allow to express that a specific pseudonym has been used to sign a specific book. But it seems right to have a "identity of" property for example, something like

⟨ Romain Gary (Q13218867)    ⟩ identity of Search ⟨ Q157322 ⟩

also, the permalink is supposed to be the help page that will (I hope) handle the result of this discussion. author  TomT0m / talk page 14:11, 20 November 2016 (UTC)

We have sooooo many of similar examples where the name assigned is not the official description of the item. For example, assignation of publishers where the current name is different from the publisher at the point of publication. So it seems that what you are looking to have is the ability to pick one of the alternate descriptive names, rather than the primary item name. This also happens for authors in many cases outside of pseudonyms, eg. authors with titles, women marrying, etc. So solution needs to be broader than this pocket of example.  — billinghurst sDrewth 04:23, 21 November 2016 (UTC)

That could be an option, and there is also object named as (P1932). Generally the pseudonyms should go in the author item itself, there is no need to create separate items like in the example you used (which is empty, and the award should go in the author page). Btw, I removed the section for voting, because so far there is nothing to vote about.--Micru (talk) 07:57, 21 November 2016 (UTC)

It makes sense to have seperate items for pseudonyms, as some have their own identifiers. Sjoerd de Bruin (talk) 10:03, 22 November 2016 (UTC)

I support TomT0ms proposition(s). For the moment I see three possibilities to handle alternative identities:

1) Just ignore them and credit the person with her legal name for a work (maybe use object named as (P1932) subject named as (P1810) and add the pseudonym as a literal).

2) Create an own item for each alternative identity and attribute the works to the identity under which it was published. Thereby the alternative identities should be connected to the "natural person", for example by the property TomT0m proposed (identity of).

3) Attribute the work to the "natural person", but make clear, that it was published using the alternative identity via a qualifier (e.g. P794 (P794)). (Which means that we need an item for each pseudonym)

I don't like the first possibility that much for the same reason TomT0m hinted at: Pseudonyms are often different from birth names of married persons or different spellings as they're often used to separate different contexts of writing; see Peter Bieri (Q115630), who uses the Name "Peter Bieri" for his philosophical writing and the Pseudonym "Pascal Mercier" for his literary works.

As to the second and the third possibility I don't have a strong preference (I think both would work). In both possibilities would be clusters of works belonging to a certain identity / queries with all works belonging to a single identity.

There is a FRBR-draft (page 61f) with some thoughts which could be helpful.Valentina.Anitnelav (talk) 14:24, 22 November 2016 (UTC)

There is also subject named as (P1810) which is used as a qualifier to handle pseudonyms. --Harmonia Amanda (talk) 07:21, 24 November 2016 (UTC)

Sorry, I chose the wrong property (corrected) - but is it then possible to query the works a person published under a certain pseudonym? Valentina.Anitnelav (talk) 08:28, 24 November 2016 (UTC)

The string datatype makes it hard to do - we would have to do fuzzy search because there may be variant of the crediting string (let's say he's credited as "P. Mercier") - and it's not really consistent with the names which are now monolingual string datatype. There is this way also no real mapping between the items of the pseudonym/identity and the string. author  TomT0m / talk page 18:00, 24 November 2016 (UTC)

Number 3 would be my preference. Number 1 apparently seems simpler but in fact, as TomT0m says, all these string remain separate and querying is harder. Of course you could write regular expressions to cover all possible spellings of each specific pseudonym, but why go through the hassle? Create an item and that's it. Why not number 2? Because the pseudonym is not a person. Furthermore, If I'm looking for all films that feature a specific actor, I want the simplest query to return them all (pseudonym or no pseudonym). Same for books, etc. Mushroom (talk) 20:48, 24 November 2016 (UTC)

Is it just me?

For some time I try to filter PetScan results from a certain category, and this works with really basic queries (like, only show people with:

select ?item where {?item wdt:P31 wd:Q5}

. But when I make my query a bit smarter, like all people withouth a profession/P106, I never get any results. Either 0, or a bad gateway error. Am I doing something wrong, is this a known bug, or is this just too much asked? Edoderoo (talk) 17:28, 22 November 2016 (UTC)

Would you provide permalinks? Matěj Suchánek (talk) 18:50, 22 November 2016 (UTC)

Unfortunately, you do not get a link until a query has results. But SELECT ?item WHERE { ?item wdt:P31 wd:Q5 . OPTIONAL {?item wdt:P106 ?profession} FILTER(!bound(?profession)) } didn't work as sparql part next to a category with people. It said 502 Bad Gateway. Edoderoo (talk) 22:42, 22 November 2016 (UTC)

It's just you ;)

A claim[31:5] equivalent doesn't work.
--- Jura 19:04, 22 November 2016 (UTC)

To clarify, the underlying SPARQL query times out. You might get very lucky once if the query is cached or so, but in general, "all humans" is hard enough. You can use the "Uses items/props" (with "none") setting in the "Wikidata" tab. I put that there precisely because SPARQL won't cut it, but if you have categories, these work as a subsequent filter rather than a separate query to subset against. --Magnus Manske (talk) 22:59, 22 November 2016 (UTC)

Hi, Edoderoo, Have you tried an approach like this one (see the wikidata tab)? Lymantria (talk) 08:18, 23 November 2016 (UTC)

Nice, a pretty alternative, although partly outside SparQL. PetScan is a powerful tool in the end, thanks Magnus for creating that for us! Edoderoo (talk) 16:25, 23 November 2016 (UTC)

The other way round is slightly quicker [10]. Still, I can't search labels across [31:5] anymore.
--- Jura 20:16, 24 November 2016 (UTC)

Not starting the fire

I found that English Wikipedia has an impressive list of things mentioned in We didn't start the fire. On Twitter, someone asked is this could be encoded into Wikidata, e.g., for "songs that mention the JFK shooting" type queries. There are too many of them in this example to warrant main subject (P921), and not all things mentioned are events, so significant event (P793) can't be applied either (and the events don't really relate to the song that way anyway). What would be a good property (with qualifiers?) for this? --Magnus Manske (talk) 21:35, 22 November 2016 (UTC)

Some kind of "work refers to" or "work mentions" property would be what's needed here. I think that (other than main subject (P921), or for people characters (P674)) we don't really have anything suitable at the moment. This would also have the side-effect of allowing Wikidata to replicate the "in popular culture" sections of Wikipedia articles. Whether this is a good thing or a bad thing is up to the reader :-) Andrew Gray (talk) 21:43, 22 November 2016 (UTC)

More than replicating "In popular culture". "Work mentions" or similar would be highly arbitrary property. A book of several hundred pages, a two-hour movie or a database mentions hundreds, thousands or millions of things that have Wikidata items. And once there would be such a property, there'd be no way to stop people adding statements like "mentions grass", "mentions the sky", "mentions New York City", "mentions <any given name>", ... to pretty much all creative works, with no chance of ever reaching something like completeness of these kinds of statements for pretty much every item affected. --YMS (talk) 11:53, 23 November 2016 (UTC)

Exactly. This would be something similar to IMDb keywords (especially reference to...), which I sometimes find useful but are also severely limited by incompleteness. They also suffer from bad standardization, which may be less of a problem here since we already have so many entities to reference and they are mostly properly classified. But still I'm worried about the quality, completeness, sourcing, and maintainability of such a system. Mushroom (talk) 23:12, 23 November 2016 (UTC)

Which maintenance problem do you envision ? It seems to be a quite simple relationship : either a work mentions some event or stuff or it does not. There is no evolving in this, it rarely will stop to mention something else, or it's a highly dynamic work. author  TomT0m / talk page 17:55, 24 November 2016 (UTC)

I was not referring to evolution but verifiability, e.g. it is pretty hard to verify whether little-known (and pretty bad) 1997 Korean film Lament (Q6481742) actually makes a reference to The Terminator (Q162255). I think it does, but I might be confusing it with a different one. If people start mass-adding these from their own memories, or copying them from disputable sources, it will require significant maintenance to verify them all. Mushroom (talk) 20:21, 24 November 2016 (UTC)

That's included in quality and sourcing then. The completeness in absolute can't really be acquired if there is not a reference for every work, so completeness can be defined only as completeness wrt. sources. But anyway, this does not seem a reason not to try anything at all considering the relative innocence of those datas - the worst we risk is that we get a superset of the works that have references to another, not really a big deal to me. It's always possible to filter out the not sourced claims and/or to whitelist some known sourced to reduced the size if that goes out of control. The "named after" property does not seems overused. author  TomT0m / talk page 20:43, 24 November 2016 (UTC)

Sure, I was just highlighting these issues that worry me. I wouldn't vote against creating the property, I'm open to trying and see how it goes. If it gets out of control we can always reconsider. Mushroom (talk) 21:07, 24 November 2016 (UTC)

Streets in Wikidata?

I was wondering if it would be possible to query how many streets are named after men or after women in a certain country, but the amount of streets is certainly not very complete. The total number of instances of streets seems to be 266038, but the number of streets with "located in country: Spain" is just 606. From this I was wondering:

• Would it make sense to create items for ALL the streets/roads/ways in the world? Some estimates consider that there are about 3-4 million street. Digestible?
• Is there any database from where the data could be freely imported? OSM?
• Is this a project that could interest the wikipedias to display automatic lists of streets in a certain city?

--Micru (talk) 14:30, 23 November 2016 (UTC)

Most of the streets you found, were imported a year ago, and are located in Q55/The Netherlands. They come from a freely licensed database, provided by the Dutch government. I'm not sure if OpenStreetMap will be a reliable enough source, and if the data is available in the right format. If it is, it would be possible to upload it, to my idea. Edoderoo (talk) 16:08, 23 November 2016 (UTC)

I think the estimate of 3-4 million streets is far to low. We already have 230,000 items for streets in the Netherlands. An on OSM the highway tag is used more than 100 million times [11]. --Pasleim (talk) 17:38, 23 November 2016 (UTC)

"Quality process on Wikidata": Let the bots run on Wikidata and: Just force Wikipedians to the assembly line of Wikidata for maintenance...

See also Wikidata:Project_chat/Archive/2015/03#An item for every street? and Wikidata:Project_chat/Archive/2015/03#Redux. Welcome to the wiki-data-dumping-ground!

You want your bot to dump millions of street-items into wikidata? Go ahead, nobody at wikidata cares.

Your bot wont maintain or update the data? Go ahead, nobody cares at wikidata.

You wont look out for vandalism of those millions of streets? Well, nobody cares either.

Just dump it all on wikidata. It wont be deleted. Because at wikidata, nobody cares.

Then, flood all the crap automatically onto Wikipedias (via bots, or article placeholders). Just leave the dirty work on your datadump to Wikipedians. They will surely, absolutely care for your bot-data-dump. Or maybe not. Watch smaller Wikipedias first getting swamped with wikidata-boticles (like occitan Wikipedia). Then watch how Wikipedians will stop caring. Because nobody cares for bot-created garbage-data-maintenance. --Atlasowa (talk) 17:50, 23 November 2016 (UTC)

• Thanks Atlasowa for the link to the previous discussion, it is very helpful. I think Denny's comment in the middle captured a good philosophy on this: "I would suggest that we selectively allow for such datasets to be added, whenever the person who wants to do the upload also at least sketches how to aim for a sustainable data quality. So, I would very much support to gather datasets like all streets, etc., and to integrate them. There is no rush, but our vision should be to actually have this data in some day, not too far in the future. Especially since we already have a well-established data set for Paris streets, New York streets, etc." - the main criterion there being some idea of how the effort will "aim for a sustainable data quality". How do we get there for these potentially large datasets? ArthurPSmith (talk) 19:45, 23 November 2016 (UTC)
• Atlasowa, thanks for the links to past conversations and for sharing your perspective. I understand your frustration with permissiveness, as there cannot be fixed rules for this or for that. That is why I ask, to find out the boundaries. I find it unfair that those items exist for the Netherlands, but not for other countries. If they should not exist, then they should be deleted. My main motivation was to answer a question that I had, but if it is not possible then it is not be attempted. However, there are also many items that there are not used, and they are not connected with anything, like Maltose ABC transporter substrate-binding protein BMA2129 (Q23543378). Should we have a separate wikidata project for data dumps like these?--Micru (talk) 07:26, 24 November 2016 (UTC)

Perhaps we should focus on what is already listed at Wikidata:WikiProject Roads (the most important highways) before adding more low-quality street items. --Rschen7754 07:43, 24 November 2016 (UTC)

• @Rschen7754, Atlasowa: are there any quality issues with the NL dataset? If not, I don't quite understand your comments.
  --- Jura 08:07, 24 November 2016 (UTC)
  • See the complaints made the last few times this came up, in the project chat archive links above. --Rschen7754 08:09, 24 November 2016 (UTC)
    • I recall people being overwhelmed by their number, but I don't really recall any actual quality issues: so can you just mention one?
      --- Jura 08:21, 24 November 2016 (UTC)

Wikidata:Project_chat/Archive/2015/10#Remember the bot-import of a quarter million dutch streets? Now guess what...

Wikidata:Bot_requests/Archive/2015/12#VIAF Identifiers for Dutch streets --Atlasowa (talk) 19:54, 24 November 2016 (UTC)

VIAF identifiers are not part of the dataset. They were added by a VIAF bot.
--- Jura 21:25, 24 November 2016 (UTC)

• Atlosowa has one valid point: if the goal is to just dump data into Wikidata, it is not a good idea. But start with thinking in smaller projects, like the roads in the Netherlands. This data is used (on Wikidata and Wikidata-items), is maintained up to some level, and therefor people care. Often people from the Netherlands. If there is a good data set for another country, it might be an idae to import it. With good data and a dedicated project, most of Atlasowa's isses will be tackled. With just a dump of a sjidload of data most issues of Atlasowa will become a headache. Edoderoo (talk) 08:13, 24 November 2016 (UTC)
• I'm not an expert on licenses, but as I understand it, we can't import data from OSM because they don't release their data as CC0. Rather than trying to duplicate all streets from OSM in Wikidata (and then having to maintain them and keep them in sync), I think it would be a much better idea to tag streets in OSM using the existing key name:etymology:wikidata. I'm not sure if there's a simple way to query both datasets right now, but if there's a demand for it, I can't see why it couldn't be made possible. - Nikki (talk) 12:08, 24 November 2016 (UTC)

Translate

Hello. I want to translate to Greek the phrases "Item by sitelinks" and "Improved search". Where can I find them? Xaris333 (talk) 13:32, 24 November 2016 (UTC)

Both appear to come from MediaWiki:Gadget-Search.js. - Nikki (talk) 16:02, 24 November 2016 (UTC)

Thanks. Xaris333 (talk) 20:06, 24 November 2016 (UTC)

Two point in time qualifiers for single population value for Fanipaĺ (Q201439)

I found that for Fanipaĺ (Q201439) the value for population (P1082) has two point in time (P585) qualifiers,with values 2014 and 2015 respectively.Should one population reading be allowed to have two separate point in time values. Even if the values were same for the two years it would make more sense to have separate population value for the two years. Thanks, Kaaro01 (talk) 19:07, 24 November 2016 (UTC)

Previously, there was only 2014, the new date was added by an IP. I agree with you, these are two separate pieces of information which for instance may have different reference. Matěj Suchánek (talk) 19:56, 24 November 2016 (UTC)

'Humans with missing claims' report has Lua errors

The table on page at Wikidata:Database reports/Humans with missing claims is failing with multiple "Lua error: too many expensive function calls." errors. Can it be subdivided, or is there another work-around? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:42, 24 November 2016 (UTC)

Recently, there were some changes to the software. Now it's possible to access data per user language, ie. we have no longer to load the whole entity (which is an "expensive function call") and find the correct label. AFAIK loading only label is not expensive. So if more agree on this, we could change templates and modules to work more efficiently. Matěj Suchánek (talk) 20:07, 24 November 2016 (UTC)

Sounds really good to me. Thx. --Succu (talk) 21:32, 24 November 2016 (UTC)

Lists of items on Wikidata

Hi, I'd like to insert some lists of items made of the output of queries on Wikidata pages (i.e. WD:PKMN in order to check data quality). How can I do that? -- ★ → Airon 90 13:59, 24 November 2016 (UTC)

What Wikipedia? Thanks, GerardM (talk) 14:26, 24 November 2016 (UTC)

I meant Wikidata, sorry --★ → Airon 90 14:30, 24 November 2016 (UTC)

Try {{Wikidata list}}. Matěj Suchánek (talk) 20:08, 24 November 2016 (UTC)

Thank you very much, Matěj Suchánek, but I think that {{Wikidata list}} is too restrictive. By now I just put links to WQS ;) --★ → Airon 90 11:25, 25 November 2016 (UTC)

How to express an attempt?

I was transfering data to 2003 São Tomé and Príncipe coup d'état attempt (Q27860366) and that got me thinking. Making it an instance of coup d'état (Q45382) may be misleading because it was not successful. How can I express that? Is there a qualifier I can put? --Luk3 (talk) 14:14, 25 November 2016 (UTC)

use attempted coup d'état (Q25906438) --Pasleim (talk) 14:22, 25 November 2016 (UTC)

Thank you! Sorry, I couldn't find this item before... --Luk3 (talk) 16:57, 25 November 2016 (UTC)

Data import hub and guidance on preparing data for import for non technical people now available

Hi all

After quite a lot of time and effort myself and NavinoEvans have put together both a data import hub for people to work together to import data from external sources and a guide for non technical people to learn how to import data. These resources have been designed to work with the existing bot requests page, data donation page and partnerships and data imports chat.

Data Import Hub
Why import data into Wikidata.
Learn how to import data
Bot requests
Ask a data import question

Thanks very much

--John Cummings (talk) 16:05, 25 November 2016 (UTC)

Notability of Book-Namespace

On Wikidata:Notability, I could not find a rule to include or not to include pages from the Book:-Namespace. So e.g. for en:Book:Mars, is it allowed to create an item? Steak (talk) 08:55, 24 November 2016 (UTC)

• With the goals of Wikicite I see no reason not to create items like this. ChristianKl (talk) 11:21, 24 November 2016 (UTC)
• It seems that en:Book:Mars should be linked with Mars (Q111), but there is no property for that. Is en-wp the only wikipedia using the Book namespace?--Micru (talk) 13:19, 24 November 2016 (UTC)
  • No, enwiki is not the only one. There are many others: Book:Harry Potter (Q14271905). --Stryn (talk) 18:16, 24 November 2016 (UTC)
• A Wikipedia-Book: is a bit like an ordered MediaWiki:Category.
  --- Jura 14:06, 24 November 2016 (UTC)
• No rule not to include it but it doesn't seem a good idea to create such items. Why would you need items about these pages? Wouldn't a property as suggested by Micru suffice, yes? On fr.wiki, we have the same system but without a specific separate namespace (eg. fr:Wikipédia:Livres/Constellations). Cdlt, VIGNERON (talk) 16:00, 24 November 2016 (UTC)

From my point of view, I dont like items linking to the book namespace. But there are already a few (e.g. Book:Africa (Q26869052)), so clearifying this and stating the consensus in WD:Notability seems useful. @ChristianKl:, I dont see a connection to Wikicite. Pages in the Book namespace are not valid citation resources. Steak (talk) 18:09, 24 November 2016 (UTC)

@Steak: do you have a list? Are all items empty like Book:Africa (Q26869052)? (there is 5,625 books on en:Category:Wikipedia books (community books) but apparently most of them doesn't have an item) Did you ask the creator (JMGM for this item) why they created it? After seeing this example, I can see one legit use for these items: interwiki links (and I learned that there is a Livro: namespace on pt.wiki). Cdlt, VIGNERON (talk) 09:11, 25 November 2016 (UTC)

A list with all items with links to the Book-Namespace in en:wp can be found here. 183 items. In my opinion, they should be deleted and the Book-Namespace declared as not notable. Steak (talk) 21:23, 25 November 2016 (UTC)

Yes the items would be needed for interwiki links. We could be treating Book namespace like template namespace: you need 2 sitelinks. I also think we should have properties linking to Book pages. Do we need an item to link to those or can properties link directly to pages on various language wikipedias?--Jarekt (talk) 22:51, 25 November 2016 (UTC)

Lua

I need some one to help me with Lua about taking data from wikidata to wikipedia through template. Xaris333 (talk) 13:12, 24 November 2016 (UTC)

What Wikipedia? What template? What are you not able to handle? Please always provide as much information as you can, so that you can get help faster. Matěj Suchánek (talk) 19:59, 24 November 2016 (UTC)

There are a lot of things. About football templates. Xaris333 (talk) 20:58, 24 November 2016 (UTC)

Please check c:Module:Authority control, c:Module:Creator or c:Module:Coordinates. They might be already doing what you need. I can help with questions about them if something is unclear. --Jarekt (talk) 22:54, 25 November 2016 (UTC)

Translating WikiData to Tunisian Arabic dialect

Dear Users, I have a matter with the name of the Tunisian Arabic dialect in the WikiData system. The name of aeb-arab is التونسي and not تونسي and the name of aeb-latin is Et-tounsi and not Tûnsî. I ask about how to efficiently adjust this in WikiData and in other Wikimedia projects. Please reply me soon. Thank you in advance. --Csisc (talk) 13:46, 23 November 2016 (UTC)

@Csisc: I think you need to open a ticket on Phabricator. Pamputt (talk) 17:59, 23 November 2016 (UTC)

@Pamputt: I thank you for your answer. I ask if you can do that for me. --Csisc (talk) 11:14, 26 November 2016 (UTC)

Language support

Where's it gone? Started a section on this a few days ago and it seems to have been deleted? Any progress on getting it sorted? - MPF (talk) 19:17, 25 November 2016 (UTC)

The section was archived. --Succu (talk) 19:34, 25 November 2016 (UTC)

Why, when it hadn't been concluded yet? - MPF (talk) 22:56, 25 November 2016 (UTC)

If nobody is answering for 7 days, a topic gets automatically archived. --Pasleim (talk) 11:19, 26 November 2016 (UTC)

Data donation of translated place names

Hi everyone, i'm a long time contributor of maps to Wikipedia, and am excited to support a data donation to Wikidata as part of my work at the OpenStreetMap powered mapping company Mapbox. We have recently been improving OSM map data with Wikidata links to enable localising the maps to virtually any language that is available using Wikidata labels. For a few important places that lacked sufficient language covereage we have been able to get them manually translated and verified by professional translators.

A total of 331 places translated into 8 languages. 1508 translations ended up perfectly matching the existing labels on Wikidata, and 1148 are the potential new or updated labels that need to be imported. The diff, all translations, and existing Wikidata labels are on separate sheets.

Data: Google Spreadsheet.

A few notes about this data:

• All the translations were done by professionals in their native language and can be determined to be of the best quality
• Translations were also spot checked by internally at Mapbox by language speakers against online sources and were found to have no major discrepancies
• Some of the translations differ from the existing ones on Wikidata. It would be valuable to have a community review of these to determine which is more appropriate, or if this should just be added as an alias
• The zh-tw translations need to be reviewed if it is the appropriate for Taiwanese speakers and can be skipped if it overlaps with existing translations in other zh variants

Since the translators are native speakers is it ok for them to contribute them directly from their account, or does this have to be done with a separate import account with specialized tools after another round of review? Has there been any similar attempt previously that could help understand the workflow better? --Planemad (talk) 12:58, 23 November 2016 (UTC)

I think having the translators contribute the updates directly would be a great idea - they get credit for the improvements and are also directly accountable if there's any kind of dispute. You might want to look for or organize a WikiProject on this to invite other wikidata editors who might have an interest in improving this data - Category:Geographical WikiProjects has a list of geographical wikiprojsects in Wikidata. ArthurPSmith (talk) 15:15, 23 November 2016 (UTC)

Thanks for the links, i'm not sure if this warrants a creation of a new WikiProject, but i'll definitely hit up those interested. Also going to post on the mailing list to get some more feedback --Planemad (talk) 11:03, 24 November 2016 (UTC)

• How do you translate place names? en:"Cambridge" > fr:"pont de Cam"?
  --- Jura 15:17, 23 November 2016 (UTC)
  • Not sure if you are joking or serious: many places do have different names in different languages, especially for languages not using the latin alphabet. Cambridge is still Cambridge in French, but London is Londres for example (and 倫敦 in Mandarin). Koxinga (talk) 20:59, 23 November 2016 (UTC)
  • Let's hear about the methodology actually used.
    --- Jura 21:19, 23 November 2016 (UTC)
    • The methodology is how one would usually translate proper nouns, use the localized name in that language if it exists using an existing source, else use the spelling that would be the best match for the pronunciation --Planemad (talk) 08:50, 24 November 2016 (UTC)

• If these are signed off by a professional translator, then I'm really happy for us to take them - thanks for putting this together! Individual accounts seem fine by me. Andrew Gray (talk) 13:42, 24 November 2016 (UTC)
• I'm a bit worried about forms that are different from what Wikipedia collected in the last 15 years (and what is now on Wikidata). How can these be explained? Other than compilation of names from various sources (which can be referenced if needed), can we be sure that the remaining ones are merely transcriptions and transliterations?
  --- Jura 14:03, 24 November 2016 (UTC)

Some bs translations are problematic since they seem to be simply copies from English transcription with no regard for bs transcription rules. Nikola (talk) 15:19, 24 November 2016 (UTC)

That's really, really cool, @Planemad:! Thank you and everyone involved for the effort. I would be very happy to see it uploaded to Wikidata. Do you know how many you are adding to Wikidata that are currently missing?

I think that either, a single centralised upload as well as each contributor doing their changes, is fine. The discussion seem to point to a preference for individual contributions, in order to keep a simpler provenance. Whatever is easier for you and your merry band of translators is fine for me. Again, thank you! --Denny (talk) 02:44, 25 November 2016 (UTC)

Thank you everyone for the positive responses! We'll work on coordinating the individual translators to upload them. @ Denny: We have around 20 languages at the moment, each done by one translator, so thats as many accounts. We'll prepare a list and have the whole process documented for to help similar work in future --Planemad (talk) 18:25, 27 November 2016 (UTC)

Collectible automobiles

I have data on individual "notable" (collectibles, usually expensive) automobiles (by VIN/serial number), owner history. Would be interested in donating these to wikidata, if there is interest. I suppose each automobile would be an instance of (object) a particular model (class). This would be like donating data on other notable/famous objects (paintings, old pottery) and their history, somewhat like that Flemish art collection.Ketil3 (talk) 04:30, 25 November 2016 (UTC)

@Ketil3: Please see Wikidata:Data Import Guide. Matěj Suchánek (talk) 08:20, 27 November 2016 (UTC)

Statements vs qualifiers

I started editing Caminhos Longos (Q25840841) and adding some statements and qualifiers and I wondered what the exact difference is between a statement and a qualifier. Because I could add a statement with the same content as a qualifier. What is more appreciated:

1. More statements and less qualifiers
2. More qualifiers in one statement and less statements?

I'm looking forward to hear the answer! Q.Zanden questions? 18:43, 26 November 2016 (UTC)

On an item you are using statements that are descrying the item ("A statement is how the information we know about an item—the data we have about it—gets recorded in Wikidata."). Some of the statements may have qualifiers. "Qualifiers are used in order to further describe or refine the value of a property given in a statement." In you are example, place of publication (P291) is using as a qualifier to publication date (P577). That's wrong. publication date (P577) must use as a statement because is an information directly for Caminhos Longos (Q25840841). Xaris333 (talk) 21:37, 26 November 2016 (UTC)

But how to show that the movie was in 1981 in portugal released and in 1983 (for example) in England? Do I have to make a new statement for the new information? Q.Zanden questions? 23:26, 26 November 2016 (UTC)

I think

⟨ Caminhos Longos (Q25840841)    ⟩ place of publication (P291) ⟨ Portugal (Q45)    ⟩
point in time (P585) ⟨ "1981" ⟩

and

⟨ Caminhos Longos (Q25840841)    ⟩ place of publication (P291) ⟨ England (Q21)    ⟩
point in time (P585) ⟨ "1983" ⟩

Xaris333 (talk) 00:07, 27 November 2016 (UTC)

According to Wikidata:WikiProject Movies/Properties, the recommended way is:

⟨ Caminhos Longos (Q25840841)    ⟩ publication date (P577) ⟨ "1981" ⟩
place of publication (P291) ⟨ Portugal (Q45)    ⟩

Mushroom (talk) 00:33, 27 November 2016 (UTC)

"FAST" vs "FAST Authority File"

We seem to have a problem with Faceted Application of Subject Terminology (Q3294867). It is stated as an instance of authority file (Q36524), and edits like this are quoting it as a source, but it is described as a "syntax" in both its English Wikidata description, and its English Wikipedia article. The linked website is labelled "FAST Authority File", for which there appears to be no Wikidata item. Ping User:Multichill, whose bot made the edits. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:53, 26 November 2016 (UTC)

What is the problem exactly? I'm just syncing some data based on FAST RDF. Multichill (talk) 21:24, 26 November 2016 (UTC)

The problem is that the source you're citing is not the syntax you claim it to be. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:31, 26 November 2016 (UTC)

Ok? This is the one linked from FAST ID (P2163). Looks fine to me. If you have another item, I can use that one for future edits. Multichill (talk) 21:58, 26 November 2016 (UTC)

By the way, bot seems to be done. 70.000 missing Library of Congress authority ID (P244) links have been added based on FAST ID (P2163). Just 50 items left to check manually. The list at Wikidata:Database_reports/Constraint_violations/P244#.22Unique_value.22_violations grew quite a bit in case anyone feel like helping out. Some are duplicates and can be merged. Some of them are cases where the ID was added to the wrong person and now this becomes visible. Multichill (talk) 22:57, 26 November 2016 (UTC)

"OK"? No, it isn't. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:08, 27 November 2016 (UTC)

Save is sometimes failed because of "OK"?

Could not save due to an error.

OK.

Can someone tell me what means this? --Liuxinyu970226 (talk) 06:34, 27 November 2016 (UTC)

This could be whatever. What did you attempt to save? Matěj Suchánek (talk) 08:11, 27 November 2016 (UTC)

Rethinking the way of describing units

I have started a discussion here: Wikidata talk:WikiProject Physics#Rethinking the way of describing units. We have not found a good solution yet, so I would be glad for further thoughts and ideas.--Debenben (talk) 23:36, 27 November 2016 (UTC)

OpenStreetMap relation ID (P402)

Would it be feasible to run a bot which automatically checks if Wikidata IDs were added/changed/deleted in OpenStreetMap changesets and then updates the Wikidata items with the OSM object identifiers, similarly to the pre-Wikidata interwiki bots? See Wikidata:Properties for deletion § OpenStreetMap Relation identifier (P402). Jc86035 (talk) 02:58, 28 November 2016 (UTC)

Some translations missing in Catalan, unable to find them

Hi,

when I look at the discussion of any property (ie. date of birth (P569), in the header of section 1 I see something like this (or please see this image):

Documentación (es) / নথি (bn) / dokumentáció (hu) / Дакумэнтацыя (be-tarask) / Документация (ru) / Dokumentation (de) / dokumentācija (lv) / dokumentation (da) / Dokumentado (eo) / 解説 (ja) / dokumentace (cs) / Documentazione (it) / Dokumentation (sv) / Dokumentacja (pl) / dokumentasjon (nb) / documentatie (nl) / 說明文件 (zh-hant) / documentation (fr) / документација (mk) / 이 속성의 세부정보 (ko) / documentation (en) / توثيق (ar) / 文档 (zh-hans) / dokumentasjon (nn) – (Ajudeu a traduir això en català.)

If I switch my language to, say, Spanish, I only see

Documentación

So it seems that the translation for Catalan (ca) is missing. This poses three questions:

• 1) Where can I find the string to translate?
• 2) Why is it so difficult to find that string?
• 3) Why, if the translation is non-existing, it shows such a long string with all the translation which difficults reading the page?

This happens in more places around Wikidata, and probably it happens to more languages which are not as used as English, Spanish, etc.

Thanks for your help.--Arnaugir (talk) 17:56, 26 November 2016 (UTC)

@Arnaugir: Are you looking for Module:I18n/property documentation? Multichill (talk) 21:27, 26 November 2016 (UTC)

That string seems to be coming from Template:Int documentation. I don't know why it's so hard to find (I had trouble too). I assume it shows all of them because it doesn't know which one you can understand. - Nikki (talk) 21:42, 26 November 2016 (UTC)

Tip: open edit window, preview without changes and open "Templates used in this preview:" below. You will see all templates transcluded into that page.

The template that is responsible for generating "such a long string" is {{TranslateThis}}. There is also {{LangSwitch}} which only shows the best language. Matěj Suchánek (talk) 08:17, 27 November 2016 (UTC)

Arguably templates are not appropriate if you need to be really clever to find a string and localise it. When a functionality is agreed on it should be translatable at translatewiki.net. Thanks, GerardM (talk) 09:23, 27 November 2016 (UTC)

Thank you all. I agree with GerardM that the string should be moved to translatewiki. Translation is missing for a lot of languages so I guess many people is not seeing it well, like me.--Arnaugir (talk) 09:34, 27 November 2016 (UTC)

Hi people, I'm the creator of the category Category:Expression_translation_template and all the templates in it. I don't know how to transfer something in translatewiki.net, my understanding is that only medawiki software messages are translated there, so having translatable translates seemed to a more immidiate and efficient compromise that allowed easy reuses for terms that are frequently used in WikiProjects and so on. author  TomT0m / talk page 12:57, 27 November 2016 (UTC)

translatewiki.net can only be used for system messages. For all other messages we need to use either translatable templates or the translate extension, though the second option is more suitable if a whole pages has to be translated and not just a single phrase. To improve the situation we could show only one translation of a fallback language and make the text "Ajudeu a traduir això en català." clickable. --Pasleim (talk) 13:59, 27 November 2016 (UTC)

... which is likely to cause problems when it's already inside another link. Matěj Suchánek (talk) 13:41, 28 November 2016 (UTC)

NASA Payload Specialist biographies?

NASA biographical ID (P2030) has the formatter URL http://www.jsc.nasa.gov/Bios/htmlbios/$1.html, but the URL of Payload Specialist bios begins with http://www.jsc.nasa.gov/Bios/PS/ (e.g., Charles D. Walker). What to do about them? -- IvanP (talk) 14:54, 28 November 2016 (UTC)

@IvanP: You could ask NASA to set up redirects. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:32, 28 November 2016 (UTC)

Human Development Index

The Human Development Index (P1081) of Czech Republic (Q213), Denmark (Q35), Iceland (Q189), Ireland (Q27), Mexico (Q96), Netherlands (Q55), New Zealand (Q664), Singapore (Q334), Spain (Q29), Sweden (Q34), Switzerland (Q39) and Yemen (Q805) is given with “±0.001” at the end, that of Afghanistan (Q889), Israel (Q801), Malaysia (Q833), Norway (Q20), South Africa (Q258) and South Korea (Q884) with “±0”, Canada (Q16), United Kingdom (Q145) and United States of America (Q30) have a mixed version, Slovakia (Q214) has “±0.01” and those that I have entered (see, e.g., Niger (Q1032)) have no “±”. “±0.01” can be discarded since the HDI is rounded to three decimals, but which version of the remaining is to be preferred? I would choose the one without “±” since the non-rounded figure is itself only an estimation.

I also think that Monte Castelo (Q21328578), Aguascalientes City (Q200805), Center of Fortaleza (Q18483386), Fortaleza (Q43463), Île-de-France (Q13917), Recreio dos Bandeirantes (Q1580587), Rio de Janeiro (Q41428) and Tehuacán (Q842261) should have no HDI. -- IvanP (talk) 14:54, 28 November 2016 (UTC)

I would prefer, at least for sovereign states, to remove the "±" from the HDIs given since they are given with no expressed uncertainty by the UNDP. For subdivisions, or for values not calculated by an agency of the UN, I am less certain about the issue, but they should at least be kept. Here, for example, is a map with the Brazilian values. Mahir256 (talk) 14:59, 29 November 2016 (UTC)

Galapagos tortoises

We have a rather difficult discussion (perhaps due to errors in mutual understanding) about what to make with Chelonoidis nigra nigra (Q486672) – please see Talk:Q486672#This is a mess. Your input would be welcomed. Thanks. --Vachovec1 (talk) 20:07, 28 November 2016 (UTC)

Do we, Vachovec1? Please provide references for your POV. --Succu (talk) 20:40, 28 November 2016 (UTC)

Community Wishlist Voting is Open

The voting for the Community Wishlist 2016 is open: https://meta.wikimedia.org/wiki/2016_Community_Wishlist_Survey/Categories/Wikidata ChristianKl (talk) 16:17, 29 November 2016 (UTC)

Plural

Hello. Using height (P2048) with unit metre (Q11573) in Greek Language (the same problem maybe exist in some other languages), one can read that a person is 2,05 metre. But the correct expression is 2,05 metres. What can we do? Xaris333 (talk) 20:41, 23 November 2016 (UTC)

Physical quantities such as height (P2048) just have a value and a unit, thus it is formally correct. I suggest to ignore the problems arising in natural languages with plural forms, since claims on item pages are typically not read by humans in a natural language (apart from some Wikidata editors). —MisterSynergy (talk) 20:57, 23 November 2016 (UTC)

It is only because of the user interface that this is the case. I do no read Wikidata items because they are totally unintelligible. The use of Reasonator is what makes sense for a human to use Wikidata. Wikidata UI is at best good for data entry. Thanks, GerardM (talk) 06:02, 30 November 2016 (UTC)

But if a Wikipedia article takes this information for a Wikidata item, then there is a problem. Xaris333 (talk) 21:10, 23 November 2016 (UTC)

Wikipedia's can use a special tool {{#plural}} that can adjust the plural form of a word. See also on mediawiki. Q.Zanden questions? 21:36, 23 November 2016 (UTC)

In ruwiki we use P558 (P558) instead of unit label, it solves most of problems. --Lockal (talk) 08:06, 24 November 2016 (UTC)

• There are Phab:T77983 "[Story] Use localized unit symbols" and Phab:T86528 "[RFC] Unit Localization" about this.
  --- Jura 08:15, 24 November 2016 (UTC)

My experience is that the plural of units is not handled consistent in my language, {{#plural}} therefor does not solve this in a simple manner. We may have irregular or regular plural depending on how they are used. Using P558 (P558) solve all problems I am aware of, except for money, where we use other methods. -- Innocent bystander (talk) 08:29, 24 November 2016 (UTC)

One example: We say "1 meter" in Swedish, but we also say "7 meter", while we say "1 sekund/7 sekunder". "meter" has a plural form, "metrar" but is not used here. If somebody says/writes "7 metrar", I expect him not to talk about something 7 meter long, but rather 7 distances, each being 1 meter long. -- Innocent bystander (talk) 12:54, 25 November 2016 (UTC)

Languages suggested on the login page

Hello,

We noticed that on the login page of Wikidata, there are only a few language links suggested. For example, on Commons, there are many more. This can be modified here.

Obviously we can't have all the languages displayed, but why not adding some more of your choice? I let you decide and do it :) Lea Lacroix (WMDE) (talk) 15:46, 29 November 2016 (UTC)

Login I definitely agree that we should revise this. No disrespect to the Dutch speakers in our community but why is this here and Chinese (varieties) is not? Huge world languages with hundreds of millions of speakers like Arabic and Russian are not on here, nor very large regional languages which surpass German in speakers such as Bengali or Japanese as well. I don't think it would be overly cumbersome to have 20 languages here--this is a small list on a page. —Justin (koavf)❤T☮C☺M☯ 19:05, 29 November 2016 (UTC)

 Support. I don't think there is any disadvantages, and there really are some huge languages out. --Luk3 (talk) 20:38, 29 November 2016 (UTC)

Perhaps we should just copy the Commons list, and add any others that anyone specifically requests. --Yair rand (talk) 23:37, 29 November 2016 (UTC)

Aliases for Homo sapiens

Homo sapiens is the taxonomic name for a human, for mankind and womankind, for men and women; people have been known as Homo sapiens since the time of Linnaeus.

And yet my addition of the English-language aliases, on Homo sapiens (Q15978631), of "human", "mankind", "womankind" , "man", "woman", and "people" has been reverted, apparently because those aliases are also used on human (Q5).

So far as I am aware, aliases are not required to be unique; those listed above meet the guidelines at Help:Aliases and should be restored to Q15978631. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:04, 16 November 2016 (UTC)

And yet again you fail to inform the other party about this "discussion" that should take place on the talk page of the said item. Sjoerd de Bruin (talk) 15:15, 16 November 2016 (UTC)

And yet again you criticise while neglecting to address at all the issue at hand. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:23, 16 November 2016 (UTC)

P.S. Help:Aliases specifically says "Use the criteria below to determine appropriate aliases for items and, when unsure, ask on the Project Chat." (Emboldening added). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:59, 17 November 2016 (UTC)

„my addition of [...] has been reverted, apparently because those aliases are also used on human (Q5)”. Probably not, because human (Q5) lacks such aliasses. At e.g. Womankind there is no indication both terms refer to the same thing. --Succu (talk) 23:02, 16 November 2016 (UTC)

The edit summaries in question were "these are mismatched; belong in Q5" and "Q5 and Q15978631 are different items, for different concepts, each with their own aliases". No other reason for their removal was given. But note my use of the word "apparently". Wiktionary is not a reliable source; nor is it complete. But even so it says "womankind" is "Women, taken collectively"; and woman is is defined as "literally... female human being". "Human being" is, of course, defined there as a "primate... of the species Homo sapiens". And you too fail to address the key point: these aliases meet the guidelines at Help:Aliases and should be restored to Q15978631. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:39, 17 November 2016 (UTC)

en:Womankind guides me to the nearest meaning and not to a loosely related term. The more common en:Mankind directs me to disambiguation page with a short explanation. Aliases are „alternative names“ for the same thing, intended to make WD more searchable. Do you really think Womankind and Homo sapiens are interchangeable? BTW: Aliases in a nutshell states: „Aliases are alternative names, like nicknames, acronyms, or translations.“ --Succu (talk) 21:33, 17 November 2016 (UTC)

It's unwise to rely on a "nutshell" summary, when the full text is available. It includes: "All of the other common names that an entry might go by, including alternative names; acronyms and abbreviations; and alternative translations and transliterations, should be recorded as aliases." There is no requirement that aliases be "interchangable"; but do you think "human" and "Homo sapiens" are not? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:58, 17 November 2016 (UTC)

It's „unwise“? Maybe... Skiping answers is what? „Dumb“? A pointed answer to your (unrelated) question is: Not all instances of Homo sapiens behave themself as human. You can ignore the summary Aliases in a nutshell and ignore common practice. Why, beyond this guidance page, should we add this „alternative names“? Nonformal arguments are wellcome. I gave some why we shouldn't. --Succu (talk) 22:20, 17 November 2016 (UTC)

I suppose I should feel gratified that my earlier remarks about seeking the biggest possible audience and devoting all energy into arguing rather than into a little thought, are borne out again, but I could have done without this. Aliases are there to help the reader find what he is looking for, and as Q5 and Q15978631 are linked (if you find the one, you have found the other), any kind of argument about copying aliases into the other item is frivoluus.

        As repeated above, Q5 and Q15978631 are different items, for different concepts, Q5 is "human", a component of human society, having properties like an occupation, family name, age, sexual orientation, etc. On the other hand, Q15978631 is the species Homo sapiens, having properties like a taxonomic position, being able to be identified by using a a field guide, etc. Words like human, man, mankind, woman, womankind, people are not aliases for the animal species Homo sapiens; on the contrary, they all presuppose an ability to speak a language, a certain degree of education, etc. Using aliases such as these only has the effect of creating confusion for the reader, or, all too often, expanding confusion for some readers. - Brya (talk) 18:03, 17 November 2016 (UTC)

You are correct in that this is a topic deserving a wide audience and input from additional colleagues (and indeed, in the previous case, it was only by drawing in a wider audience that the explanations that were initially refused were eventually forthcoming); and that "aliases are there to help the reader find what he [or indeed she] is looking for". You are incorrect in your various ad homnem comments, and in your assumption - or rather assertion, made with no supporting evidence - that the aliases in question will be confusing to our colleagues. To suggest that they are not aliases for Homo sapiens is simply false. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:40, 17 November 2016 (UTC)

Ah, the "I know what I know, don't ask me to think about it"-attitude. - Brya (talk) 06:43, 18 November 2016 (UTC)

These are not different concepts, this is two viewpoints on the same concept. The human species described scientifically versus the species and its members described by themselves. There is now consensus that this two objects are the same, just faceted. Homo sapiens' forms a society for sure. But I don't see how we could not tell

⟨ human (Q5)    ⟩ part of (P361) ⟨ human society ⟩

 just as like

⟨ homo sapiens ⟩ part of (P361) ⟨ human society ⟩

, there is absolutely no contradiction. author  TomT0m / talk page 11:12, 18 November 2016 (UTC)

concept : "an idea of what something is or how it works". - Brya (talk) 11:51, 18 November 2016 (UTC)

@Brya: : A citation from your link : concept may apply to the idea formed by consideration of instances of a species : do we agree that every instance of homo-sapiens is an instance of human and conversely ? Or would you say that "Human" and "homo-sapiens" are distinct concepts describing the same object ? If so maybe this puts a spotlight on a conceptual problem here : how can we express that two concepts are supposed to describe the same object ? I'd tend to say then that Taxonomy is a scientific theory and that taxons are descriptions of objects of the real world. author  TomT0m / talk page 12:01, 18 November 2016 (UTC)

A person is an absolute, an animal is an absolute. A taxonomic concept of what animals together constitute a species is a concept. Human society may be studied by various scientific disciplines, which all will create their own concepts. - Brya (talk) 12:15, 18 November 2016 (UTC)

@Brya: Do you understand this do not answer the problems I raised in a previous post ? How to state two concepts are related to the same "absoute" as you say ? Clearly modern science would have no problem to say that a human is an animal. One natural solution would be to state that

⟨ human (Q5)    ⟩ subclass of (P279) ⟨ animal (Q729)    ⟩

. Can this be expressed as

⟨ human (Q5)    ⟩ subclass of (P279) ⟨ animal (Q729)    ⟩

 ? author  TomT0m / talk page 12:54, 18 November 2016 (UTC)

The problem you raised in your previous post was that these two items do not reflect different concepts. But as shown, they do reflect two different concepts, so pragmatically there is no problem.

        But, sure, these are not finely tuned relationships. In particular, Q5 is a "fuzzy" concept, as every user (bots excepted) is human and knows what that entails, and expects Q5 to reflect that. That is a lot of users and a lot of viewpoints, to be contained in a single item.

        On the other hand, Q15978631 has the same issue as the other two million items that deals with a taxon. The only available, widely accepted way of referring to a taxon is by scientific name, and there is no 1:1 mapping between scientific names and taxon concepts. A scientific name is a formal entity in its own right, sharply delimited. However, there may be any number of taxon concepts that it can be applied to. - Brya (talk) 18:01, 18 November 2016 (UTC)

I don't agree. First if there is a lot of (major) concepts for a human we could not agree on, each would deserve its own item, and ... we don't have a lot of items for this. We don't talk of a person (Q215627)    here which is philosophically much fuzzier. The problem is that ... there is a mainstream modern viewpoint of this that identifies humanity with the homo sapiens taxon that we could not reflect here and we would have no way to reflect this on wikidata ? (quote : fr:Humanité L'humanité est à la fois l'ensemble des individus appartenant à l'espèce humaine (Homo sapiens) mais aussi les caractéristiques particulières qui définissent l'appartenance à cet ensemble. - translation : humanity is both the set of all individuals belonging to human species (homo sapiens) but also the specificities that make someone belonging to that set) On the other hand, you claim that there is a lot of taxon concept with one names. But in that case we have only one item for all of them ? So, pragmatically, this is very much unsatisfying if we listen to your arguments : we have two items for many concepts, not two items for two concepts ... If the taxons items actually are items for taxon names, we should have actually a relationship between taxon concepts and taxon names like a "taxon name" property with item data type ... So this does not make any sense both pragmatically and theorically to me, really. For the sake of beeing able to refect several viewpoints ... we actually don't reflect any viewpoint. author  TomT0m / talk page 09:27, 19 November 2016 (UTC)

Well, pragmatically speaking, for the foreseeable future only in exceptional cases will individual taxon concepts get their own items. Instead, we will be hard put to eliminate most of the fictitious taxa.

        As to concepts for human, who knows how many items we will end up with. Indeed, Q215627 is another item closely linked with Q5, and in this case there is no formal relation. And there is no clear demarcation between these. Very fuzzy. - Brya (talk) 11:18, 19 November 2016 (UTC)

for the foreseeable future only in exceptional cases will individual taxon concepts get their own items => and you just don't provide a way to model this ... The problem, as I feel it, here is that you not only don't provide a convincing model to express the relationships between all these items correctly, but that you seem to be so focused into your own notion of data quality that you almost forbid anyone else to try to build a correct model ... Although you acknowledge there is weaknesses everywhere and the result of your hard work is doomed to have some weaknesses itself as it just can't guarantee that the species are well described besides their name ??? Seriously, I don't understand. There is probably a better way to organize stuffs which could let you do what you do with the datas you want ...

Very fuzzy. What about

⟨ human (Q5)    ⟩ subclass of (P279) ⟨ person (Q215627)    ⟩

 ? This does not seems like a problem to me ... Why do you see problems where I see none ? author  TomT0m / talk page 11:40, 19 November 2016 (UTC)

I suppose Q5 could be a subclass of Q215627: this depends on the definition of "human" and the definition of "person".

        We already have items for individual taxon concepts, so this is not academic. There just are not many of them.

        You seem to be so focused into your own notion of what constitutes a correct model that you are impervious to reality, and see no problems anywhere. - Brya (talk) 06:15, 20 November 2016 (UTC)

@Brya: My belief if that if we do things correctly, we will be able to express the relationships between two models. I don't think that Wikidata is a set of disjoint unrelated viewpoints that should have sanity barriers between them not to contaminate each other. I also think that if there is really that much definitions of "human" each deserves its own item. You seem to argue there is many ... you did not prove this is true. author  TomT0m / talk page 09:53, 20 November 2016 (UTC)

Linnés concept of Homo sapiens, provided in his 10th edition of Systema naturae (1758), is very different from our modern view. In 1735 the key quality he related with our species was „Nosce te ipsum“. For the first time after Aristoteles someone placed „humans“ within the animal kingdom. In 1758 Linnés viewpoint changed. Homo sapiens was place in the order primate (Q7380). Additional he distinguished two species: Homo diurnus (we) and Homo nocturnus (= Orangutan (Q41050)). Not that bad for a start. Probably you can imagine how the fossil findings of e.g. Neanderthal (Q40171) later changed the concept of Homo sapiens multiple times. --Succu (talk) 21:41, 20 November 2016 (UTC)

Isn't that ... deprecated ? author  TomT0m / talk page 21:52, 20 November 2016 (UTC)

Deprecating what exactly? --Succu (talk) 21:59, 20 November 2016 (UTC)

Well, if Line's viewpoint is considered obsolete now, pretty much any statement you could make by using this source that is not backed up with modern one has to be deprecated. If a taxon is considered obsolete now, it should only be used as object of deprecated statement, should'nt it ? author  TomT0m / talk page 07:44, 21 November 2016 (UTC)

Given we had a property „Taxon concepts” pointing to an item labeled „Homo sapiens sec. Linnaeus, 1758” this statement could be marked as deprecated. You can not simply rejcect all statements based on this specific reference. --Succu (talk) 18:45, 21 November 2016 (UTC)

Hi, sorry to barge in - I don't get the difference between Human and Homo Sapiens, and this discussion here didn't help me. Is there another place that explains it? --Denny (talk) 05:17, 22 November 2016 (UTC)

I wonder if anyone will answer you? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:22, 26 November 2016 (UTC)

• I would also prefer to merge the two items. Most of the times human is used on Wikidata it's used in a way that means a member of homo sapiens sapiens. Given that this seems to be a deeper change to our ontology, how about having a RfC to merge the two?

If we for example have a class of a human specific disease, it's easier to set the constraints when these aren't two different items. ChristianKl (talk) 10:29, 30 November 2016 (UTC)

Wikidata weekly summary #237

−

−

Here's your quick overview of what has been happening around Wikidata over the last week.

−

Discussions

• Closed request for comments: Semi-automatic Addition of References to Wikidata Statements

Events/Press/Blogs

• Past: Wikidata introduction by Pigsonthewing at DJUK 2016

Other Noteworthy Stuff

• Items without statements is down to 5% for enwiki (source)
• Wikidata:WikiProject Biographical Identifiers

Did you know?

• Newest properties: Auckland Art Gallery artist ID, Observatoire du Patrimoine Religieux ID, Géopatronyme ID, Prabook ID, GECD film ID, GECD person ID, Enciclopedia Treccani, stereoisomer of, Tennishof ID, operating income, PictoRight ID code, Nobel Prize People Nomination ID, negative prognostic predictor, positive prognostic predictor, negative diagnostic predictor, positive diagnostic predictor, negative therapeutic predictor, positive therapeutic predictor, National Recreation Trails Database ID
• Query example : List of sovereign state flags by proportion (source)
• Newest tool: Display the link to Wikiversity in the title of Wikipedia using Wikidata
• Newest database report: Unique films: film items with a link to a single Wikipedia

Development

• Adding new wikis to interwiki sort order and move sort orders to WMF config. phab:T111023

You can see all open tickets related to Wikidata here.

Monthly Tasks

• Hack on one of these.
• Help develop the next summary here!
• Contribute to a Showcase item
• Help translate or proofread pages in your own language!
• Help merge identical items across Wikimedia projects.
• Add labels, in your own language(s), for the new properties listed above.
• Comment on property proposals: all open proposals - proposals needing attention

Read the full report · Unsubscribe ·

 – The preceding unsigned comment was added by Kareyac (talk • contribs). 07:19, 30 November 2016‎ (UTC)

Reference to death notice

For copyright clearing for work of artists, it's useful to reference to the death notice, where there often a name to the heirs can be found. What is the best way to link to a death notice like for example this one http://www.delpher.nl/nl/kranten/view?coll=ddd&identifier=KBNRC01:000027223:mpeg21:a0121? I have added it here Mari Andriessen (Q844964) as a source for date of death (P570), with a qualifier instance of (P31) death notice (Q2438528) and reference URL (P854) but could this be better a property like image of grave (P1442) ? --Hannolans (talk) 14:17, 30 November 2016 (UTC)

WikidataCon, October 28th-29th

Hello all,

I'm very glad to give you some piece of information about an event we will organize in 2017: the WikidataCon :)

This event will take place to celebrate Wikidata's 5th birthday, but most of all, to celebrate you, the editors who build Wikidata everyday.

We will try to make this event as open and contributive as possible, especially in the content: you will be able to choose what you want to hear, do, participate to, and you will be the first providers of the content.

For any question, feedback, suggestion, please use the talk page.

I'm looking forward to organizing this event with you!

Cheers, Lea Lacroix (WMDE) (talk) 14:23, 30 November 2016 (UTC)