Open main menu

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


I need the TOC back

Where did the TOC go? Automatic TOC is one of the most important features of MediaWiki, which makes it superior with most document processing systems other than Latex. Wikidata pages tend to be extremely long, and I often visit them looking just for a specific piece of information like a sitelink: now I am no longer able to just click "Wikipedia" or "Wikisource" to reach what I need. Instead, on a page like Q44520, I'm forced to press page down 12 times. I'd rather have more things added to the TOC if possible, I can't do without a TOC. Thanks for the understanding, Nemo 09:46, 30 March 2015 (UTC)

Well, I really hated the TOC: always having to do the extra scrolling to get past it was very wearying. - Brya (talk) 10:58, 30 March 2015 (UTC)
I found it useful...
Try pressing END once and then page up fewer times. Or adding #sitelinks-wikipedia to URL.
Also there will be an office hour tommorow where you can talk to the developers. Matěj Suchánek (talk) 13:03, 30 March 2015 (UTC)
I am trying really hard to remove clutter from our pages and make them easier to understand and scan. The navigation bar is one of the victims of this and I really don't want to bring it back. What I intend to provide is a way to collapse each sitelink section and move them around. So that if you are mainly interested in Wikisource you can just move that to the top and collapse the rest. I hope that addresses most usecases. --Lydia Pintscher (WMDE) (talk) 13:10, 30 March 2015 (UTC)
But, on a small screen the sitelinks appear under the statements section. Navigation on small screens is terrible now. Sjoerd de Bruin (talk) 13:53, 30 March 2015 (UTC)
Jep that still needs fixing. There should be way fewer cases where they move below the statement section in the future (When the statement section is redesigned). Which screen size are people using who have it move below the statement section currently? --Lydia Pintscher (WMDE) (talk) 13:54, 30 March 2015 (UTC)
Well: I have a screen with a resolution of 2560 x 1440 pixels, but I mostly divide the screen in sections. So 1280 pixels wide, tat's not enough sadly. Sjoerd de Bruin (talk) 14:06, 30 March 2015 (UTC)
Shouldn't you have added collapsing before removing the ToC? And I don't see why the ToC couldn't also be an option... --SamB (talk) 00:03, 31 March 2015 (UTC)
+1 - I agree with SamB… could you put back the TOC, while the collapsing is not active, please ? or at least, put it as an option in preferences ? --Hsarrazin (talk) 17:50, 31 March 2015 (UTC)
I could now have the team spend time on getting it back for a short while but that would delay us considerably in getting the collapsing feature done and unit support. In the overall scheme of things Wikidata is helped a lot more if we spend time on those. --Lydia Pintscher (WMDE) (talk) 09:26, 1 April 2015 (UTC)

new constraints violations page

I made this page which has links to Autolist aka Wikidata Query queries of constraints violations, but specifically of value type violations for properties that point to items. I hope it's useful. The idea was to make it easy to understand, even for beginners. Plus, it has a picture of Smokey Bear. --Haplology (talk) 08:35, 1 April 2015 (UTC)

Head of Government

I've been working through the head of government (P6) violations from User:Haplology/value_type_violators, and I've found a lot from the Philippines that list (for example) head of government (P6): Barangay Captain (Q7260481). As far as I can tell, this is a case of someone misunderstanding what this property was for. Is there a different property that would fit this context or should I just delete them? Popcorndude (talk) 13:53, 1 April 2015 (UTC)

Try office held by head of government (P1313). —Wylve (talk) 15:01, 1 April 2015 (UTC)
Thanks! Unfortunately I didn't wait for an answer and deleted them. Popcorndude (talk) 15:26, 1 April 2015 (UTC)

Discontinuity between Wikidata and Wikipedia

The page Q4721535 needs some work if someone could clarify what correct policy is here.

It's currently a instance of (P31) Wikimedia disambiguation page that was created solely because en.wikipedia Alexis Taylor was at the time. That page is now a single page for a person and there is no longer a disambiguation page.

So is policy here to

  1. edit Q4721535 to turn it into the instance of (P31) person that it now is, or
  2. delete the en.wikipedia link from it and leave it orphaned (and create a new wikidata entry for the person Alexis Taylor)?

Smb1001 (talk) 14:17, 1 April 2015 (UTC)

I think the second option, but not leave it orphaned but delete it. Sjoerd de Bruin (talk) 14:39, 1 April 2015 (UTC)
Yes the second and remember to fix Rubbed Out (Q7375788) --ValterVB (talk) 16:49, 1 April 2015 (UTC)

Wikidata weekly summary #151

The "header redesign" is really "not yet there", at the moment I have to click on "edit" to see the full descriptions. How about some "show full description on mouse over" CSS hack? –Be..anyone (talk) 09:52, 30 March 2015 (UTC)
phab:T93807. Matěj Suchánek (talk) 11:49, 30 March 2015 (UTC)
 Be..anyone (talk) 03:54, 2 April 2015 (UTC)

Quality Festival on it.wikipedia

Just to inform that on it.wikipedia, for monthly Quality festival in April we are focused on pages without interlink. We hope to found a lot of good links :) More info here --ValterVB (talk) 19:23, 1 April 2015 (UTC)

Initial release of the primary sources tool

We just made an initial release on the primary sources tool to help with the Freebase migration. We are looking for editors to test it out and report issues, or developers to help improve the tools. Big shoutout to Tomayac and Sebastian Schaffert for their work so far! --Denny (talk) 19:33, 1 April 2015 (UTC)

Tool to split an item?

I know there are a couple good Merge tools.

But I need to Split an item:

I am forced to conclude there were two Cesare Baglioni, and someone *** the cat in by taking the birth date from one and the birth place from the other.

What's the easiest way to split this item in two, copy the name, and dispatch statements between the two parts as I see fit?

Currently the easiest way is to create a new item and merge into that, then remove/revert to get the result wanted. A tool would indeed be handy. --- Jura 17:34, 30 March 2015 (UTC)
Moving links is pretty trivial. I don't know that we can move properties yet (curse my recent extended but self-imposed absence). Start with creating a new item, move the applicable sitelinks, and go from there as normal. --Izno (talk) 18:50, 30 March 2015 (UTC)
Better follow what Jura suggested, otherwise you have the pain of re-creating manually references for statements:
  1. Create a new item
  2. Merge the existing one to it (uncheck "merge to oldest" in merge tool)
  3. Remove all sitelinks from the target item
  4. Modify label and/or description of target item (2 items are not allowed to be exact duplicates)
  5. Restore the original item from its history
  6. Manually move appropriate sitelinks from the old item to the new one
  7. Delete or fix incorrect statements in both items
Indeed a tool would be useful. -- LaddΩ chat ;) 19:15, 30 March 2015 (UTC)
Thanks for detailing it. I added it to Wikidata:Split items. --- Jura 22:06, 30 March 2015 (UTC) --- Jura 22:16, 30 March 2015 (UTC)
Probably are the same person. L'enciclopedia Treccani said that was born at Cremona and moved to Bologna, but some detail about the biography (ex. travel to Rome) are the same. --ValterVB (talk) 19:39, 30 March 2015 (UTC)
@ValterVB: thanks for the find! Would you mind fixing the item, since I don't read Italian. Thanks! --Vladimir Alexiev (talk) 06:53, 2 April 2015 (UTC)
@Vladimir Alexiev: For Treccani, Cesare Baglioni was born at Cremona, around the half of XVI century. I have added the source. --ValterVB (talk) 17:37, 2 April 2015 (UTC)


I remember before that there was a tool thing (can't even remember the name of it, I think it was a tool) below the main label field where several labels from other languages could be seen and they could be clicked to set it as the label on the language you're currently viewing the site with. AltLabels doesn't work now with the recent header update. --AmaryllisGardener talk 22:16, 30 March 2015 (UTC)

Have you tried the labelLister-gadget? Sjoerd de Bruin (talk) 06:38, 31 March 2015 (UTC)
It works, I have it too, they're both useful to me. --AmaryllisGardener talk 14:06, 31 March 2015 (UTC)
yes, I signalled a little above that altlabels did not work anymore :(
Labellister is not at all the same : it allows to edit ALL labels...
the interest of altlabels was that you could quickly add a name (for people mostly) from other same-alphabet labels… even those languages that you chose not to display on work window...
now, you need to :
  1. click on modify (or on labelLister - the result is almost the same)
  2. find a proper label among languages... some of them not seeable directly but only through labellister (altlabels gave you the 3 more frequent)...
  3. copy the label you wish in a very very tiny little rectangle, without accidentally copying something else
  4. (if you're in label lister : add the "new language" )
  5. click on your main language rectangle (you cannot drag-and-drop)
  6. paste the label in your main language
  7. save what you just did... (if you use beta-labellister, preview before it)
That's 6 to 8 actions, instead of 1 click !!... do you really think it's an improvement ?
could somebody please fix altLabels ? I don't know if Joern is still working on the project… if not, if a nice dev. could just adopt this orphan, it would be a great help !!!
I used to add fr labels very much on people names… if the GUI stays as it is, I won't anymore… it's much too long and difficult :( --Hsarrazin (talk) 17:06, 31 March 2015 (UTC)
Exactly. I work with labels alot too, and while the LabelLister is great for adding labels for a bunch of languages on an item, AltLabels was a way to add a label for the main language quick and easy. Can someone fix the tool? --AmaryllisGardener talk 17:17, 31 March 2015 (UTC)
I just had a look at the code (can't code, but I can understand tests and loops) : the key pb, I think, is on lines 52-55 : they check the presence of the label input box, which is precisely what has been removed from the GUI…
maybe, just restoring the main label input box would be enough, and a great thing, since it also allows drag-and-drop (from Preview or Wikidata Usefuls), and IS MUCH BIGGER... for small screens, big fingers and tired eyes ;) --Hsarrazin (talk) 17:44, 31 March 2015 (UTC)
Lydia? --AmaryllisGardener talk 18:43, 31 March 2015 (UTC)
Do I understand it correctly that the usecase is the following? You are on the item of a person for example whose name is the same in many languages. You need a way to quickly do that. Would solving phabricator:T92759 help? --Lydia Pintscher (WMDE) (talk) 09:31, 1 April 2015 (UTC)

@Lydia Pintscher (WMDE): I don't even understand what this phabricator:T92759 is about :( - something like User:Jitrixis/nameGuzzler.js ? that allows to add the same value to many languages in 1 move ? as far as I understand the issue here, it's something very different, allowing to edit other languages… not easing edition of main language :/

No, that is not what altlabels was for… - I don't think solving this issue would do the trick...

I use many tools for labels, each one having specific pros and cons… 

What altlabels did, very efficiently, and what I miss very much, was : here is this item : it's a Q5, and it is not labelled in my main work language… and there are already labels in X other languages… automatically, when I open the item it gathered the other labels, took the 3 more used, and put them just under the "main language label box" so that, if one of them is right for my main language (which is often the case for latin alphabets), I just click on the label I choose and.... whoooooom : it's added to my label…

In 1 click, just by choosing, I could add a label to an item… many items, open in as many tabs… without bothering to modify, click, type, copy, paste, save, and finally purge the item… can you do that, or restore that ability in altlabels ? --Hsarrazin (talk) 19:31, 1 April 2015 (UTC)

Ah ok. Thanks for the explanation! I'll discuss this. --Lydia Pintscher (WMDE) (talk) 09:59, 2 April 2015 (UTC)
Sorry for the slow reply on this, i'm very busy with my phd at the moment. I've fixed User:Joern/altLabels.js for now and mirrored it over on github so people can easily report issues / or help with pull requests. Joern (talk) 11:15, 12 May 2015 (UTC)

3 years!

Hey folks :)

3 years ago we started the development on Wikidata. I'm amazed at how far we've come together in those 3 years. <3 to everyone who's a part of this and helped us get to where we are. Looking forward to what we can achieve in the next 3 ;-)

Obligatory photo of toast with PHP elephants: :D

Cheers --Lydia Pintscher (WMDE) (talk) 10:51, 1 April 2015 (UTC)

\o/ (courtesy to Ricordisamoa on the mailing list for the initial cake :) TomT0m (talk) 16:00, 1 April 2015 (UTC)
+1 Snipre (talk) 20:05, 1 April 2015 (UTC)
Lydia, you opened my eyes: Wikidata (Q2013) <is a> <subclass of> April Fools' Day (Q80949). --Succu (talk) 20:28, 1 April 2015 (UTC)
:P Shhhhhh! No-one must know! --Lydia Pintscher (WMDE) (talk) 10:00, 2 April 2015 (UTC)
Years goes so quickly! --Stryn (talk) 10:12, 2 April 2015 (UTC)
Hehe no. That's the birthday of the site. We started working on the software a few months earlier :) --Lydia Pintscher (WMDE) (talk) 10:15, 2 April 2015 (UTC)

Municipality vs. footballer

For some reason, item Esporlatu (Q390704), an Italian municipality, which was created from the article of the same name in Italian Wikipedia, was described from the very beginning as an "Italian footballer". I haven't been able to find a footballer with this name so I don't know where does this confusion come from. Could somebody fix this? Thanks. --Quico (talk) 07:53, 2 April 2015 (UTC)

Human error, reverted that one. Sjoerd de Bruin (talk) 08:45, 2 April 2015 (UTC)

itwiki Quality Festival - april 2014

it:Wikipedia:Festival della qualità/Aprile 2015 is the "annual" festival on itwiki dedicated to "wikidata issues " (we had other ones in the past, we usually propose connectivity-based festivals every 3-6 months).

The theme this time are missing interlinks concerning itwiki articles, please let us know in the talk page if you have any suggestion (e.g. if there is some other working lists you would like to propose).

FYI: festivals are linked from the home page and some users are sometimes novices. --Alexmar983 (talk) 12:58, 2 April 2015 (UTC)


I made an honest effort to fix this on my own but I couldn't figure it out since I'm completely new on Wikidata and I'm not prepared to go through any more "help" pages here because none of them are helping. Maybe I will try to learn how to use this site later but later is not now. Here's the problem:

  • Q12376169 is about Tallinna Linnatranspordi AS - which is the name of a company
    • I can not add the English Wikipedia article to the Wikpedia entries because it will tell me that link is used by item Q3737251 which is about Tallinna Autobussikoondis
      • I can not remove that link from that item because it will give me this error: "Warning: The action you are about to take will remove a sitelink from this item. Sitelinks should only be removed if the page in question has been deleted, or if that link is being merged into another item. If you are trying to do neither of these, please do not submit this edit again."

And I have no bloody idea what that's supposed to mean, i.e. what I'm actually supposed to do now. Merging? What's that? Is it being merged? By who? By me? What?

On the English Wikipedia, Tallinna Linnatranspordi AS is the main article, and several different titles redirect to it, including names of former companies. I posted the list of redirects here: with explanations of what they are, i.e. English and Estonian names of this or that.

I'm somehow getting the impression that Wikidata doesn't like redirects because when I try to enter one here then it magically transforms into the target link instead. Why? I don't understand how to work with this.

The actual problem is that the interwiki links in the English Wikipedia article go to articles about Tallinna Autobussikoondis on the Estonian and Italian Wikipedias. The interwiki links should actually go here: and here:

Any help would be appreciated. — Jeraphine Gryphon (talk) 15:30, 2 April 2015 (UTC)

  Done @Jeraphine Gryphon: you are right: the warning is misleading. You can remove a sitelink when it obviously links to the wrong page, as in this case. --Csigabi (talk) 15:37, 2 April 2015 (UTC) looks so easy now ;_; Why did I not just do what you did? I don't know. — Jeraphine Gryphon (talk) 17:02, 2 April 2015 (UTC)
I have improved the warning by changing merged to moved. Matěj Suchánek (talk) 17:30, 2 April 2015 (UTC)
Oh thank you. That's better. — Jeraphine Gryphon (talk) 17:57, 2 April 2015 (UTC)

Can someone do me a favor and put a welcome template on my talk page? :D — Jeraphine Gryphon (talk) 18:14, 2 April 2015 (UTC)

The Mortar and Pestle problem

Wikipedia is an encyclopedia not a dictionary or thesaurus, so quite often articles are about several things at once, eg en:Mortar and pestle. Getty AAT (a core thesaurus in cultural heritage) has two different entries. Dewiki even has 3 entries, since it distinguishes between small and large pestles (de:Pistill vs de:Stößel).

In Wikidata there are 4 entries (correct), but the links and "part of" relations are wrong. Below I show only enwiki and dewiki, but other wikis should be considered as well:

What a mess! We need more organized ways to cope with "impedance mismatch" between wikis. --Vladimir Alexiev (talk) 06:48, 2 April 2015 (UTC)

@Vladimir Alexiev: That's the good old Bonnie and Clyde problem. AFAIK the latest solution was to wait for WD capability to links to WP redirect articles, and in the mean time create 3 items, the grouping one having 2 has part (P527), and each of the other two using part of (P361). I believe we need a Help:One vs. several sitelink-item correspondence page to clarify the suggested scheme. -- LaddΩ chat ;) 12:25, 2 April 2015 (UTC)
See Project:Be bold (Q3916099), Special:Search/Bonnie Clyde problem [1] and Wikidata:Property proposal/Generic#Redirect to. Visite fortuitement prolongée (talk) 21:28, 2 April 2015 (UTC)

special occupation

What if a person has a special occupation which is not in the list I could choose from when editing? Thank you. I know its not the issue just for occupation, its like global question. I was trying to find it out in help... --Ivanhoe2012 (talk) 10:31, 2 April 2015 (UTC)

Make a new item for it. Take a look at other occupations to see which properties are good for a occupation item. Most of them will appear if you add a instance of (P31). Sjoerd de Bruin (talk) 12:01, 2 April 2015 (UTC)
or link to a more general occupation and add a statement using "field of this occupation (P425)" to link to the item for the industry they work in or "sport (P641)" in the case of a sportsperson where there isn't an occupation item for their specific sport. Filceolaire (talk) 03:02, 3 April 2015 (UTC)

Deprecated rank or delete?

Help:Ranking tells: "The deprecated rank is used for statements that are known to include errors or that represent outdated knowledge. Marking erroneous statements as deprecated instead of simply deleting such statements has three benefits: [etc]"

But what to do about the statement: discoverer or inventor (P61):Paul Henry and Prosper Henry (Q302840) in 126 Velleda (Q137391)? The reference in the statement tells it is supported by Minor Planet Center (Q522039). But I cannot see that it does that today, and from what I know, it never has. Instead, I think the statement comes from Wikipedia. (The bot and it's owner is since long gone from this project and cannot be consulted.)

My opinion is that deprecated rank can and should be used for claims like: "Earth is flat". But I do not think deprecated rank should be used for claims that tells that the "Earth is shaped like a banana", only because a bot, by mistake, has imported that information from a Wikipedia-article.

A discussion on en.wikipedia recently stated that Wikidata is not a reliable source, and I would like to say: Wikipedia is not a source reliable enough to support deprecated claims, which are stated nowhere else! This does not mean I think statements based on Wikipedia should be banned, but that we should delete them, when they are wrong! -- Innocent bystander (talk) 07:47, 30 March 2015 (UTC)

Well, at Lucy F. Simms (Q17386627), I found it preferable to include potentially incorrect values for date of birth (P569) as any researcher will eventually come across these values and will have to figure out if they are correct or not (and where they come from). --- Jura 10:45, 30 March 2015 (UTC)
Well, Wikipedia is not a reliable source. Importing data from Wikipedia to Wikidata does not suddenly make them reliable. So Wikidata is not necessarily reliable. Judicious deletion may help improve the situation. - Brya (talk) 11:01, 30 March 2015 (UTC)
If good sources tells diffently, I think there is no problem. People very seldom are born twice.
The problem in my case is that a bot has gone wrong. MPC has never said that both of Paul and Prosper discovered "126 Velleda". en.wikipedia claim that they do, but without external source who support it.
You may argue that the brothers cooperated since they are credited for every second discovery, but that is OR, and nothing we should bother to do here. -- Innocent bystander (talk) 12:40, 30 March 2015 (UTC)
I think that if the error comes from the bot, it can be deleted (unless we want statistics on bots errors, but that's does not sound like a realistic method for doing them ;). If the error comes from Wikipedia itself, I think it can be also be deprecated, but only once we know that the error has been fixed in Wikipedia. --Zolo (talk) 15:44, 30 March 2015 (UTC)
many times, I see wrong claims, made with a wikipedia sourcing, that are really a wrong reading of the source - i.e. used the birth date of the brother, or the spouse, instead of the person's real date… - just got one this morning… while playing with Mix'n'Match… ;) - (what was ironic was that the DB item I was checking automatch for was the one used as source in the claim… and the values were different :D
In those case, I think we should just erase the wrong claim... --Hsarrazin (talk) 17:55, 31 March 2015 (UTC)
Deprecated ranks are for claims that are sourced, but the error is in the source. So the value must be sourced. The claim should not be deleted, because it is a valid claim as soon as it has a reliable source. It is in the case of reliable sources containing errors (sounds like a contradiction, but our world is not perfect). So someone might find the the source and compare it to Wikidata and that person is now able to understand that the error is not in WD, but in the source. So this prevents changing correct values into false values by good will editors. Claims that contain errors because the source is misinterpreted or claims with the vaulue is not backed up by the source should be corrected or deleted and not marked as deprecated. There is no need to establish new errors or to keep them in any ways.--Giftzwerg 88 (talk) 07:36, 4 April 2015 (UTC)
I agree 101% with Innocent Bystander. IMO Wikidata should not be a database of wikimistakes. Mistakes of a random anonymous wikipedian are not notable enough to be stored in Wikidata. Strakhov (talk) 07:58, 4 April 2015 (UTC)
The "researcher[s]" Jura1 talk about above always have the edit history of WD to look into. It's bad we cannot add good comments to our edits, so we can easily tell why we do a certain edit. -- Innocent bystander (talk) 08:12, 4 April 2015 (UTC)

Service Unavailable

I cannot add interwiki link to Wikidata items. It always shows "Service Unavailable". --XRTIER (talk) 07:56, 3 April 2015 (UTC)

Or edit labels, description, alises, or even read sometimes. -- Innocent bystander (talk) 08:20, 3 April 2015 (UTC)
There was a brief API outage, see wikitech:Incident_documentation/20150403-API.--Erik Moeller (WMF) (talk) 19:48, 3 April 2015 (UTC)

New patrol tool

Hey. To fight against vandalism, I wrote an oAuth application which is similar to Special:RecentChanges but has some more features:

  • 1-click patrol button
  • select edits by type (edited terms, sitelinks, merges...)
  • mass patrolling
  • clickable edits comments, i.e. identifiers of external databases and URLs are linked
  • hints showing you various additional information, e.g. if the format of an identifier is violated
  • 1-click translation button for labels and descriptions
  • automated description when hovering over the item's label

You find the tool on I hope with it that some more users get motivated to patrol recent changes. --Pasleim (talk) 21:46, 18 March 2015 (UTC)

This is wonderful, thank you so much!
How can we ensure this gets found even by people who don’t look at the project chat, or when this topic is archived? —DSGalaktos (talk) 22:11, 18 March 2015 (UTC)
Wow. Awesome work Pasleim! I'll distribute it over our other channels (twitter, mailing list, facebook and co) later today and will also add it to the weekly summary. In addition it should probably be linked on the pages about patroling. Could also go in the header section of recent changes? --Lydia Pintscher (WMDE) (talk) 07:03, 19 March 2015 (UTC)
This is seriously great! Is your source code on Github for pull requests? Or somewhere else? --Denny (talk) 15:06, 19 March 2015 (UTC)
Thanks for the feedback. You find now the code on --Pasleim (talk) 21:14, 19 March 2015 (UTC)
  • Interesting tool! But can I see left column in my language? Or more: a filter for viewing changes (label/desc/sitelinks) only in some set of languages (e.g. from Babel)? --Infovarius (talk) 19:49, 23 March 2015 (UTC)
Yes, it's a really great tool, yes, thanks, and yes, such a filter would be nice. Also, having the possibility to show more than 100 edits might help, though I could imagine you implemented that limit to push us in actually patrolling the edits rather than just scrolling down the list. One more thing: If I undo an edit, it disappears and I have no way (other than opening the "your edits" list) to check other edits of that user, so I'm likely to miss other vandalism. --YMS (talk) 20:16, 23 March 2015 (UTC)
@Infovarius, YMS: You can now filter by language, if you query by edited terms, i.e. label/desc/alias. Labels are now shown in your language and you can see up to 250 edits yet. The restriction to 100 edits was mainly because of performance issues. --Pasleim (talk) 21:16, 30 March 2015 (UTC)

Wikidata weekly summary #152

Financial Data

Is there any existing project group focusing on putting financial data into wikiData?

I do understand that market rates would be a no go zone for legal reasons. But it should be fine for static data like

  • market places
  • financial instruments (types, instances)
  • market place calendars
  • currencies
  • maybe even published balance sheet values
  • certainly value added items like zero rates for risk free term structures

as well as some economic data like

  • inflation rates
  • cash rate

Since I am new to wiki editing in general I have some questions:

  1. In general is wikidata the right place for this category of data?
  2. How well is wikiData suited for object types like time series? Where a combination of two items ("AU CPI"+"Publish Date") would create a new item. How would this be modeled in wikiData? Is a date an item?
  3. What is the best practice to remodel a/my data model in wikiData? Are there any specific references (other than /wiki/Help:Items)?
  4. How does wikiData deal with contributions where the data may be IP protected? I just want to know, no malicious plans there.

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

We have Wikidata:WikiProject_Economics which is looking at putting economic data for countries and other administrative divisions into wikidata. Many of the properties proposed for that project could also be used for company balance sheets however most of these projects are on hold waiting for a currency datatype (or more specifically a 'quantity with dimension' datatype where the dimension is a currency) which we will probably not have for a while. When we have a currency datatype we will be able to have values for statements which specify a number and a currency associated with that number. This will permit us to compare values between different items and be sure we are comparing like with like.

This info will not be added to wikidata until we have a set of properties which can be widely used across different items. This means that it will probably be a while before we can add this info to wikidata. Filceolaire (talk) 22:26, 27 March 2015 (UTC)

For properly whacking oneself on the head with financial data models, FIBO is required reading :-) --Vladimir Alexiev (talk) 22:32, 30 March 2015 (UTC)

  • @Smartkatt: - @Mcnabber091: proposed doing a lot of what you describe at meta:Grants:IdeaLab/Global_Economic_Map. I live in New York City and some major banks and financial institutions here and in DC have attended Wikimedia community meetups and events asking how they could interface with Wikipedia and Wikidata. They uniformly seemed willing to pipe huge amounts of economic data into Wikipedia with no obvious strings, but then, none of us really understand finance and reporting, and none of us could navigate the discussion into a plan for collaboration. They would have wanted assurance that the Wikimedia community would support what they were doing, and we were unable to promise this because we did not have expertise in finance to be able to say what quality data is and how Wikidata items can be mass-piped into lots of Wikipedia articles in many languages. I think this is an interesting space. Blue Rasberry (talk) 13:42, 6 April 2015 (UTC)


Distribution of items without a claim (2015-03-30)

This statistic was made from the last dump (2015-03-30). Item numbers are subdivided into intervals of 100,000. So for instance (X,Y) = (192,73720) represents the interval [Q19200000, Q19299999] having 73,720 out of 100,000 items without a single claim. --Succu (talk) 21:27, 2 April 2015 (UTC)

@Succu: It's true that "Statements per item" has regressed according to Wikidata Stats. I'm guessing that most of these are bot-imported single-interwiki items. Can you sample a few and confirm? -- LaddΩ chat ;) 14:02, 3 April 2015 (UTC)
Yes, LaddΩ, 89 percent have only one sitelink. Most items have a sitelink to enwiki, followed by jawiki and frwiki. Wikipedia:Shortcut (Q620197) has 90 sitelinks but no claim. For the full table see here. --Succu (talk) 14:54, 3 April 2015 (UTC)

I really like the chart, and also the breakdown by Claimless items per type of sitelinks. Do we have something to mark an item as "more a Wikimedia-internal thing than an item about something in the world"? I know that Daniel has always been a proponent of this. E.g. P31 -> Wikimedia internal page, or something. --Denny (talk) 15:29, 3 April 2015 (UTC)

I think we should use Wikimedia page outside the main knowledge tree (Q17379835) or a subclass of it, but it includes "Wikimedia article", which make it a mess [3]. I once fancied to solve it through Wikimedia internal item (Q17442446) but have to admit I gave up midway. -Zolo (talk) 15:50, 3 April 2015 (UTC)
@Succu: nice statistics!
I've been working on claimless items for quite some time. I've been focusing on the Dutch Wikipedia. Infobox templates turned out to be extremely useful for adding the first claim. I bet we have many items without claims connected to an English Wikipedia article with a useful infobox. Because of the sheer size of the English Wikipedia, I haven't really worked on this one yet. The number of claimless items on the Dutch Wikipedia is now below 50.000 and so queries are relatively fast. English Wikipedia has over 600.000 items without claims so that will probably be much slower. Anyone feel like working on this to reduce this? Multichill (talk) 12:42, 4 April 2015 (UTC)
Old news. Machines are faster than human beeings. It might be usefull to use categories to add claims by bot. On the other hand is the Wikidata Game, a real help in adding claims. The number of new items will drop dramatically, if you can not save an item unless it has at least one claim. P31 or P279 are the most important claims.--Giftzwerg 88 (talk) 13:31, 4 April 2015 (UTC)
@Multichill: Have you a query to find claimless items in a language please? — Ayack (talk) 19:10, 6 April 2015 (UTC)

No statements by language / project

Is it possible to generate numbers where you find the number of items without a statement per project? These are the items that people should concentrate on. I expect that it is NOT English Wikipedia that has relatively the most issues, I expect that as a percentage the other projects are more relevant to give attention to. Thanks, GerardM (talk) 06:05, 5 April 2015 (UTC)

@succu: what are the colors in User:Succu/Statistics/NoClaims/20150330? I think this is what you're looking for Gerard. Succu, can you add a percentage column to "Sitelinks per project" so we can see the relative progress per project? Multichill (talk) 12:28, 6 April 2015 (UTC)
For the colors have a look to the middle colum, Multichill. They are the same as in Pasleims sitelink statistics. The table is only a byproduct from my statistics. Adding a percentage column is possible. --Succu (talk) 14:07, 6 April 2015 (UTC)


Hi. I´ve found a link about ¨goldsmithing¨ written spanish, but I don´t know how to link up them. I´m sending you it. Do it, if you can.

This are two different things. The Spanish article refers to Q11202319 (work of forging things out of gold, especially jewelry) and the English to Q211423 (metalworker who specializes in working with gold and other precious metals). They should not be merged. --T.seppelt (talk) 22:10, 5 April 2015 (UTC)

Merging Q2033580 with Q19366489

There was signal that fleuron (Q2033580) and aldus leaf (Q19366489) should be merged. However there are French items, which points to different categories on commons, that prevents the gadget to do the job. Could someone more competent than me verify the case and fix it? I suspect that French item in fleuron (Q2033580) needs a separate entity. Paweł Ziemian (talk) 22:16, 5 April 2015 (UTC)

Merging is not appropriate here as aldus leaf (Q19366489) is subclass of (P279) fleuron (Q2033580) (the feuille aldine is a kind of fleuron). fleuron (Q2033580) is about the general concept of flowers as typographical ornaments, while aldus leaf (Q19366489) is about a specific kind of glyph. The article en:Fleuron (typography) messes things up by being rather confused about this, describing both concepts in one article without managing to properly differentiate between them. I've checked the other interwikis and they all seem to make the distinction perfectly well (with the possible exception of ru:Флерон (типографика) which may be translated from enwiki and thus may be equally confusing). (@Coyau:) Väsk (talk) 06:11, 6 April 2015 (UTC)
Thanks for clarification. Paweł Ziemian (talk) 09:58, 6 April 2015 (UTC)

Multiple label languages at once

I noticed that the multiple label option (in more than one language) has disappeared. Is there a way to make the labels reappear? Even if it's just the languages set in my preferences, that would be great. It takes too long to have to switch language just to enter a single label in another language. The old way allowed me to enter multiple labels in one shot.

I'll add that I do have "Show labels, aliases and descriptions in all my languages on page load" checked as well, but it only shows the labels if there already is one. Thanks, Hazmat2 (talk) 22:06, 6 April 2015 (UTC)

It's probably a bug. It seems that it works on most items, but when there are enough aliases in one language it doesn't leave room to add other languages. Hazmat2 (talk) 22:13, 6 April 2015 (UTC)


My name is Tom Fish; some of you may know me by my username Guerillero. I am an Anthropologist from Washington College and I am doing my thesis work on Wikidata. I became interested in Wikidata due to the fact that it feels like an anomaly and the fact that this community has been ignored by social scientists. In the past two years, Wikidata was able to reach a level of complexity in its policies and structures that took many other communities (English Wikipedia, German Wikipedia, Commons, etc.) six or more years to build. I am looking at the following questions:

  1. How are cultures created on the internet?
  2. Do editors of Wikidata believe that they have a culture? If so, how do the community members of Wikidata define it?
  3. How does influence circulate through the Wikidata community.
  4. How does Wikidata compare to other open source communities?

If you find this interesting, I would like to talk to you for about a half-hour. This project has been approved by my intuition's Institutional Review Board and all responses will be kept confidential. You can contact me on my talk page, via email at uni, or on irc under the cloak *!*@wikipedia/Guerillero. Thank you in advance for your time. --T. Fish 05:04, 7 April 2015 (UTC)

board of governors (Q239463) and Junta Suprema Central (Q592285)

board of governors (Q239463) and Junta Suprema Central (Q592285) are two different articles in Spanish, but refer to the same article in English. This leads to board of governors (Q239463) having English link pointing to completely different article, which does not match the rest of the languages. What is the right way to handle such issue?  – The preceding unsigned comment was added by Laboramus (talk • contribs).

You may submit the issue to Wikidata:Interwiki conflicts, or, better, fix it yourself if you understand the difference between the two concepts: create a new item and move to it the incorrect links, fix labels for "moved" interwikis on the original item, add appropriate properties and labels to the new item. You may find interest in Help:Split an item. -- LaddΩ chat ;) 22:32, 7 April 2015 (UTC)

Wikidata API: Get Wikidata id

What is the recommended way using the Wikidata API to get the Wikidata id for a Wikipedia article of a certain language (so we know title and language and want to get the Wikidata id).

Example: "I Never Liked You","en" -> "Q4890860". --Jobu0101 (talk) 09:06, 7 April 2015 (UTC)
Probably /w/api.php?action=wbgetentities&format=json&sites=enwiki&titles=I Never Liked You&props=info
Or at the specific Wikipedia api --T.seppelt (talk) 09:20, 7 April 2015 (UTC)

Thanks a lot. --Jobu0101 (talk) 10:10, 7 April 2015 (UTC)

You want Special:ItemByTitle and its related API module. --Lydia Pintscher (WMDE) (talk) 14:46, 7 April 2015 (UTC)

Auto-transliterating for names of humans

Hey, I finished a system to automatically add labels for items of humans for any given pair of languages. This is RfBA. Please comment and/or suggest. Amir (talk) 19:29, 7 April 2015 (UTC)

How to connect storage media and the devices that read (and possibly write) them

For example, it seems like there should be some kind of connection between floppy disk drive (Q493576) and floppy disk (Q5293), or between optical disc drive (Q4492) and optical disc (Q234870), or Q8015040 and CD-ROM (Q7982). Or phonographs and phonograph records; no reason to assume anything digital. Are there any properties that area already appropriate, or does someone need to propose some? —SamB (talk) 23:17, 31 March 2015 (UTC)

Something like "mating component" (bidirectional?) would be appropriate. You can't express this with part of/has part, nor use, which are probably the immediate basic generic properties that I can think of. --Izno (talk) 17:22, 3 April 2015 (UTC)
Hmm, use (P366) is not explained at all well for a property as generic as it's purported to be now, and I can't help but think that "mating component" seems awfully vague as well. At the moment, I'm wondering about "storage medium" or "reads medium" and "writes medium", and I'm really not too sure what to call the inverse(s). And of course I just had to go and think "what about other kinds of formats, like file formats?"... —SamB (talk) 18:40, 8 April 2015 (UTC)

Handling subsidiary titles of nobility

Given the concept subsidiary title (Q7632005), meaning that minor titles of nobility are in practice for long periods overwritten by major ones, it would be useful to have a property and opposite relation that express that. So, with qualifiers, Duke of Abercorn (Q1264451) would stand in relation to an item "Earl of Abercorn" (to create), from 1868. A look at w:Duke of Abercorn should explain why it would be handy to code up these relationships. Charles Matthews (talk) 07:16, 7 April 2015 (UTC)

My suggestion would be as follows:
  • Each title should have a separate wikidata item.
  • Use noble title (P97) to link to all of the titles held by a person and mark the major title "preferred rank" and the minor titles "ordinary rank".
This will mean we can search for (and create a list of) the holders of each title just searching on P97.
There was a proposal to create a property which could be used as a qualifier to make properties more specific but it wasn't approved so I guess you could add the qualifier <instance of:Subsidiary title> instead. Filceolaire (talk) 16:08, 8 April 2015 (UTC)

Stupid panels

Anandhisuresh (talk) 17:16, 28 February 2018 (UTC)anandhisuresh Tobias1984
Doc James
Daniel Mietchen
Andrew Su
Projekt ANA
Pavel Dušek
Was a bee
Chris Mungall
Dr. Abhijeet Safai
Sami Mlouhi
Netha Hussain
Abhijeet Safai
Shani Evenstein
  Notified participants of WikiProject Medicine @Bluerasberry, Daniel Mietchen, DGG, Jmh649, Klortho, Hildabast: Hello, I just want to point a very bad thing which shows a deep lack of team working. Please have a look at the talp page of this item: three different panels to group properties and some properties are presented two and even three times just because people add their own panel. So do we have to stop this new mode, use category instead of panels in talk pages or do we share the properties between projects ? Snipre (talk) 19:49, 7 April 2015 (UTC)

I really don't think that the "panels" (navigation boxes?) indicate a lack of cooperation so much as an attempt to make it easier for user to find related properties. If there was an actual lack of cooperation, different groups would either be holding a flamewar about what the property meant, or [trying to] use distinct properties, but I don't see any evidence of a flamewar there, and if they were using different properties, the boxes would obviously not all be found on this one property's talk page. —SamB (talk) 18:50, 8 April 2015 (UTC)

Impossible to choose right wikipedia language when creating new item

I've been trying to create a new item however I've found it impossible to enter the right language code. Putting in the language code ko, its impossible to scroll down to Korean which sits between kg and koi. I've tried using Hangul, and a single language code turns up ko, but as soon as I try to enter the page it flips back to kg. this has happened half a dozen times, I'm editing on a tablet running, Android 4.4.2 , I've rebooted twice, and cleared my cache.--KTo288 (talk) 21:24, 7 April 2015 (UTC)

Worked around it by creating an article at enwikipedia, and linking the Korean article from wikipedia.--KTo288 (talk) 23:13, 7 April 2015 (UTC)
It is very hard to edit with a tablet. Using a keyboard and keyboard commands is much helpful for me Oursana (talk) 23:43, 7 April 2015 (UTC)
I know, I miss having a mouse, but its a sacrifice I'm willing to make. so that I can edit curled up in front of the television. The thing is the menu thingy works fine at wikipedia but not here.--KTo288 (talk) 08:04, 8 April 2015 (UTC)

Protection indicators, again

Per phab:T92384, you can now see padlock icons on semi- and fully-protected items and properties. The feature is enabled by default, but registered users can disable it in their preferences. --Ricordisamoa 01:30, 8 April 2015 (UTC)

That's great, cheers! Jared Preston (talk) 05:54, 8 April 2015 (UTC)
Thanks Ricordisamoa! :) If anyone has problems with the gadget or suggested improvements, please let me know or folks are welcome to go ahead and change the code. Aude (talk) 07:39, 8 April 2015 (UTC)

Jimbo Wales is not registered on Wikidata

For almost two years I've been watching his void user page. How is it that our beloved Jimmy didn't even care to log in and see how a random item looks like? --Ricordisamoa 11:27, 30 March 2015 (UTC)

Why should we care? Sjoerd de Bruin (talk) 11:41, 30 March 2015 (UTC)
I'm with Sjoerd here. It would be nice for him to pop in, but it doesn't really matter. --George (Talk · Contribs · CentralAuth · Log) 19:18, 30 March 2015 (UTC)
Some serious oddity on this side, a global User:Jimbo Wales page exists and works, but not here. –Be..anyone (talk) 03:40, 2 April 2015 (UTC)
Since the account isn't activated here (sulutil:Jimbo Wales) any user page should not exist. And there are hundreds of projects that account is missing on, so we are in good company. -- Innocent bystander (talk) 06:40, 2 April 2015 (UTC)
I also agree with the sentiments here. Personally I don't think Jimbo really does anything on the projects he does visit, so its really not a huge loss IMO. In fact IMO he should just cut ties and move on because even in the English Wikipedia he really doesn't do anything other than comment on the occasional discussion. Reguyla (talk) 20:03, 8 April 2015 (UTC)

Merging IP edits into Wikidata user account edits

As of 19:53, 8 April 2015 I made 342 edits today. I then created an account, named User:IP-". Is it possible that my IP edits are merged into my Wikidata user account edits? 19:55, 8 April 2015 (UTC) - Confirming the above statement. IP- (talk) 19:59, 8 April 2015 (UTC)

Nope. Sjoerd de Bruin (talk) 19:57, 8 April 2015 (UTC)
What is Nope? IP- (talk) 19:59, 8 April 2015 (UTC)
Nope means "No, this can not be done". Edoderoo (talk) 20:04, 8 April 2015 (UTC)
@IP- That is phab:T31188.--GZWDer (talk) 10:15, 9 April 2015 (UTC)
@GZWDer: - Thank you for pointing me to Phab! Nice to see this as an item there. Best regards, IP- (talk) 12:32, 9 April 2015 (UTC)

FYI: English Wikipedia blocked my account :-(. So, it was easier to edit without account. But without account, Wikidata does not let me see labels in other languages than the default ones - 50% of the default ones are absolutely useless to me. IP- (talk) 12:33, 9 April 2015 (UTC)

Wikidata:Requests for permissions/Bot/WikiGrok

May I please ask for a broader participation in this discussion? The request is stale at the moment, and I need to decide whether it should be approved or closed as unsuccessful.--Ymblanter (talk) 07:15, 9 April 2015 (UTC)

Miscategorization of letters - instance instead of subclass

All letters (letter as they exist in alphabets) that I found are miscategorized. A letter as a concept has no time or place. If an item has no time or place it is a class. The letters in a specific physical book or on monuments are instances. IP- (talk) 13:19, 9 April 2015 (UTC)

Neither are awards, materials, genre, ship classes, or taxa, but they are all considered instances. Please wait a little while before implementing your opinion as here. I have made about 1.4 million edits on this site and I still have to think hard about whether something is an instance or a subclass. --Haplology (talk) 13:41, 9 April 2015 (UTC)
I think having some official guidelines - any guidelines - on instance/subclass question would be great. It is a question which is IMO very subjective and context-dependent and each time I try to analyze it it makes my head hurt. If the guideline would have some arbitrary-ness and reflection of some taste, that's fine provided it is consistent. At least we'd have something. --Smalyshev (WMF) (talk) 18:32, 9 April 2015 (UTC)
No they are not - at least not by people that understand the difference between class and instance. And the quantity of "edits on this site" seems to not be related to the efficiency of the thinking-process. Maybe better YOU stop immediately, if you have problems understanding the concepts of class and instance. IP- (talk) 13:46, 9 April 2015 (UTC)
Dare I ask why you were blocked on Wikipedia? --Haplology (talk) 13:50, 9 April 2015 (UTC)

The user was blocked merely for his username, which does not fulfill EnWp requirements. Looking at the way the user is conducting discussions, though, other reasons seem suggestive, too.

@IP-, please either learn to behave, or leave the project.

"If an item has no time or place it is a class." is wrong given the understanding of class and instances which has formed in this project. There are plenty of items which are perfect instances without having a time or space, e.g. Harry Potter (Q3244512), red (Q3142), or the examples given by Haplology. Whereas this, and many other decisions in Wikidata, are debatable, these debates will either happen in a friendly and constructive manner, or not at all.

I very much prefer a project with a minimum of toxic behavior, even if this means loosing the potentially productive edits of contributors exhibiting such behavior. It's just not worth it. --Denny (talk) 15:29, 9 April 2015 (UTC)

Also, I would suggest to revert the changes done to letter. I do not consider the changes in the descriptions of the letters nor the switch from instance to subclass to be improvements. --Denny (talk) 15:36, 9 April 2015 (UTC)

Mistagging of letters - followed by and follows

No, the letter C does not follow the letter B. In some series that may be so, and it also may be so for some instances, but the statement is not generally true. If placed without restriction it is simply wrong. IP- (talk) 13:49, 9 April 2015 (UTC)

Mistagging of letters - part of

No, the letter Q is not part of "Latin script alphabet". It is part of the Latin script, and /some/ Latin-script alphabets. But there are Latin-script alphabets it is not part of. If the "part of" statement is made without restriction it is a false statement. IP- (talk) 13:58, 9 April 2015 (UTC)

Links between months categories before 1918

frwiki and ruwiki both have categories for months, for instance fr:Catégorie:Janvier 2015 = ru:Категория:Январь 2015 года = Category:January 2015. Between 1582 and 1918, Russian categories use the Julian calendar while French categories use the Gregorian calendar. Should they be linked to each other? Some of them are already linked (Q18191027), but some others are not (for instance fr:Catégorie:Mars 1904 and ru:Категория:Март 1904 года have no Wikidata item yet). Orlodrim (talk) 10:29, 5 April 2015 (UTC)

Tough question. We have two possibilities, obviously:
  1. Have two (or even three) items per month (Julian calendar month, Gregorian calendar month, Calendar month without specific calendar?), since they are indeed different things (different start- and end-dates etc.)
  2. Have only a single item for all of this, and e.g. the start and end date could be discerned through qualifiers.
I would prefer the latter option, have a single item, mostly because it is consistent with how we handle years (i.e. the year 1902 is only a single item, even though the Gregorian 1902 and the Julian 1902 are a week or two off).
So in short, I'd say, link them, it's consistent to what is already happening. --Denny (talk) 16:29, 6 April 2015 (UTC)
Ok, thanks for your answer. I will also link categories before 1918. Orlodrim (talk) 17:46, 10 April 2015 (UTC)


I think we can enable Flow here for testing and later to a discussion page, which in my opinion would be Wikidata:Contact the development team. Others' thought?--GZWDer (talk) 12:03, 5 April 2015 (UTC)

At the end of last year I talked to Danny (the product manager for Flow) about using Wikidata item talk pages as a next target. He wanted to finish a few other features first and fix some issues. If you want I can ping him for a status update. --Lydia Pintscher (WMDE) (talk) 12:15, 5 April 2015 (UTC)
Let's wait for the development team to signal that they're ready to deploy flow on production sites. Looking in phabricator and on mw:Flow gives me the impression that there is currently no active development going on. WMF is constantly shifting priorities and I would hate to be the guinea pig for an unfinished product. I filed phab:T95109 to update the roadmap. Multichill (talk) 12:18, 5 April 2015 (UTC)
Thanks for the nudge, Multichill. There are plans to overhaul the docs a.s.a.p. (please see this topic if you have suggestions on what other-examples it should try to follow (VE's docs? etc)).
Re: phabricator - the team does all their work in the #Collaboration-Team board and the individual 2-week-sprint boards linked there.
Re: a test page here - that could be done at any time, but Flow has recently been enabled in a number of wikis (see mw:Flow/Rollout#Done) and will start replacing LQT at mediawikiwiki this week (announcement and planned timeline), and the team has getting a ton of useful feedback recently (feature-requests and bug-reports), so I'd personally suggest delaying testing here until the team has a little bit less backlog to contend with, so that they have time to focus on any feedback that a test page here would surely result in.
Hope that helps. :) Quiddity (WMF) (talk) 16:59, 5 April 2015 (UTC)
I doubt Flow in its current state is any improvement to our current system, and way more complex to understand. Vogone (talk) 13:33, 5 April 2015 (UTC)
I used Liquid threads on the WMF development wiki a few years ago and if Flow has even half of the feautures that system had then I think it will be a major improvement on our current system. Imagine a going to a 'Your Discussions" page. It tracks discussions you have taken part in (unless you 'unfollow'). If anyone has added a comment to that discussion then the discussion gets added back to the top of this page. Project talk pages you left a query on three months ago; Project chat discussions which have gone quiet for the last few days; Property proposals you asked for clarification on a week ago; all there on one page where you can find them or better where they will find you. I want this so much. Filceolaire (talk) 18:20, 5 April 2015 (UTC)
I also used Liquid Threads a few years ago and it certainly was not an improvement. It was hard just to keep track of what was said, even more or less, not to mention its destructive tendencies. - Brya (talk) 18:56, 5 April 2015 (UTC)
I would like to see flow tested here. I think that initiatives to make interaction easier could only benefit a project like Wikidata which is very complex to begin with. Ajraddatz (talk) 21:28, 6 April 2015 (UTC)
I would be happy to put up some test pages on Wikidata. Flow is in active development by the Collaboration team 1. In the last couple weeks, we released VisualEditor integration with a simple toolbar in the entry fields, editing other people's posts, and undo functionality. Pretty soon, we'll be working on Search, a toggle for full-width viewing, and improvements to the way Flow is presented in watchlists and Echo notifications.
Would it be okay to put up a couple test pages so that you can try it out? -- DannyH (WMF) (talk) 22:28, 6 April 2015 (UTC)
Personally I think that would be a great idea for this project. Wikidata is, IMO, in a much better place to implement Flow than the English Wikipedia because it isn't as reliant on the Talk page templates and other things that are issues with flow currently there. As a fairly new user of Wikidata myself I'm not sure which ones would be a good test though, but maybe the talk page of this discussion board? Maybe the talk page of the main page. I think it should be a page with a reasonably high volume but wouldn't affect the outcome of the project itself. If it would help, feel free to use my talk page as a test page. Its not very high traffic but you have my permission to do it as a test. Reguyla (talk) 17:40, 10 April 2015 (UTC)

Star Wars U? No E.

Things like [4] don't make sense to me because Obi-Wan absolutely exists in the canon Star Wars story and so putting "from fictional universe" = "Star Wars Expanded Universe" makes no sense. I guess I'll create a "Star Wars universe" item and replace all "EU" claims with "U"? I just wanted to check here before I went on editing a bunch of items. Also, I've created Wikidata:WikiProject Star Wars for anyone interested in improving Star Wars-related items. Thanks, --AmaryllisGardener talk 19:26, 9 April 2015 (UTC)

Wouldn't Obi-Wan be in both? Is it really necessary to replace "EU" claims? Dancter (talk) 20:25, 9 April 2015 (UTC)
Good point. Both. --AmaryllisGardener talk 20:49, 9 April 2015 (UTC)
I'm not very familiar with Star Wars (since I am more of a Star Trek-fan) or the use of from fictional universe (P1080) here. From what I know of Star Trek, I would regard the canonical Star Trek-universe as a subset of the non-canonical ST-universe. What is true in the canonical universe is more or less always true in the non-canonical. (But the story-line can be reinterpreted, for example if Troy really was fully unharmed when she gave birth to an alien.) There is of course always contradictions in a fictive universe, but I guess we have to live with that.
In ST we also have the problem with multiple alternatives. Should Kirk of Star Trek Into Darkness (Q171711) really use the same item as Kirk of Where No Man Has Gone Before (Q25366) or the Kirk of the Mirror-alternative? -- Innocent bystander (talk) 04:55, 10 April 2015 (UTC)
Currently, the official Expanded Universe is mostly a subset of the old Legends, but that’s about to change, since Legends is now sealed AFAIK, and new characters from episode VII and onwards won’t be part of it.
By the way, the new Wikidata:WikiProject Star Wars should be mentioned in this discussion I think :) —DSGalaktos (talk) 09:59, 10 April 2015 (UTC)

Property proposal: Moviepilot identifier

I'm not so in this formal property proposal stuff. So I'm just asking you if you think that we need a Moviepilot identifier property. Moviepilot is one of the biggest German film webpages. --Jobu0101 (talk) 21:20, 9 April 2015 (UTC)

Proper use of award received?

Was this a proper use of Property:P166? Technically speaking it wasn't Rain Man which received the award but its director. Is there a better way to deal with that issue? As you can see here also others dealed with that issue in that way. --Jobu0101 (talk) 08:40, 10 April 2015 (UTC)

Old RfCs

Just to inform you than one third of the open RfCs are more than one year old. If people involved in these RfC can have a look at them and evaluate the possibility to close the discussion, this will be a good action. Snipre (talk) 11:56, 10 April 2015 (UTC)

Stewards confirmation rules

Hello, I made a proposal on Meta to change the rules for the steward confirmations. Currently consensus to remove is required for a steward to lose his status, however I think it's fairer to the community if every steward needed the consensus to keep. As this is an issue that affects all WMF wikis, I'm sending this notification to let people know & be able to participate. Best regards, --MF-W 16:12, 10 April 2015 (UTC)

Relevance of this page

Greetings: I'm a Commons admin following a photo from there to here... and I found to contain information about a high school cheerleader of no particular notability. Is this the sort of thing acceptable in Wikidata, or is someone gaming the system? Thank you for your reply. Ellin Beltz (talk) 16:01, 9 April 2015 (UTC)

As long as there is a sitelink to Commons, Wikidata's rules allow for an item about her. If the respective Commons-Category were deleted, I think the Wikidata item could and be deleted as well. --Denny (talk) 16:37, 9 April 2015 (UTC)
Yes, seeing as there is a category on Commons, then there's going to be an item on Wikidata. Jared Preston (talk) 20:21, 9 April 2015 (UTC)
Criterion 1 of WD:N says that having only a Commons category link does not render a main article item notable. —Wylve (talk) 22:54, 9 April 2015 (UTC)
Ah, great. In that case we could probably delete Daniela Abarca González (Q15304735) (although I have been reading the point in WD:N, and I am afraid I do not completely parse the sentence about Commons) --Denny (talk) 22:57, 9 April 2015 (UTC)
Yes, IMO that sentence is badly written. I think it means that if the item is not an instance of Wikimedia category (Q4167836), and has only one sitelink that points to a Commons category, then notability is not established. However if the item is a Wikimedia category (Q4167836) and has other category site links, then criterion 1 applies. —Wylve (talk) 23:54, 9 April 2015 (UTC)
No it is not. Your citation doesn't apply with WD:N. It does not say that having only a Commons category link does not render a main article item notable. The contrary is stated there. The linking to a commons-file is not sufficient for notability, but linking to a commons-category is notable.
…An item is acceptable if and only if it fulfills at least one of these two goals, that is if it meets at least one of the criteria below:
1. It contains at least one valid site link to a page on …or Wikimedia Commons.
  • To be valid, a link must not be a… file
Oursana (talk) 00:26, 10 April 2015 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────Criterion 1 also reads, "In addition, an item with only a sitelink to a category page in Wikimedia Commons is not allowed on main article items. However, it is allowed to link Wikimedia Commons categories with categories in other Wikimedia sites in items." This prevents Daniela Abarca González (Q15304735) from being notable. —Wylve (talk) 00:52, 10 April 2015 (UTC)

I personnaly don't see anything interesting into fighting into this. I has an item ? fine. It does not ? I don't care. But more than the item itself, lets imagine possible usecases. I search images related to university or school life. The image could show up in results throw the fact that she is a student and that she's depicted into some common image by a request here.TomT0m (talk) 11:25, 10 April 2015 (UTC)
I'm not trying to retain or to fight for the deletion of this item. I'm simply trying to show what the community has decided on the notability policy. As to the use cases above, I think it is risky to make assumptions about an entity using a photo some user uploaded to Commons. The verifiability of claims that could be obtained from the photos is in question. How do we know that her name is really "Daniela Abarca González"? Her Commons category is within commons:Category:1994 births, but how do we know that she was really born in 1994? Although quite unlikely, but how do we know that she is really a student and not an actress dressed up as a student? A photo doesn't seem to give clear and trustworthy information about the entity depicted. —Wylve (talk) 21:30, 10 April 2015 (UTC)

I am not interested in this particular item, but would rather see the policy on WD:N be clarified (and then applied to this). It seems, from this discussion, that the policy is not readily understood, maybe ambiguous. --Denny (talk) 15:25, 10 April 2015 (UTC)

Agreeing with TomT0m and Denny. If there are images in Commons for an item that is an instanceOf X (here X=human), Wikidata seems the easiest Wikimedia place to store machine readable data like first name and second name. The Commons category title is not well made for this task. Maybe the file namespace e.g. File:X can be used one day to store structured information. IP- (talk) 17:34, 10 April 2015 (UTC)

You might be interested in Wikidata:Wikimedia Commons. —Wylve (talk) 21:30, 10 April 2015 (UTC)

Popups extension now also working on items and properties

Hovercards on Wikipedia. Try this on Wikidata.

Hi guys, I've just deployed a little default gadget which makes the Popups extension, also known as the Hovercards beta feature, also work with Wikidata items and properties. If you have not yet, you can enable the beta feature here. This feature is really useful on Wikidata now, because it includes the description of the item you hover and also an image if image (P18) is set. I can only recommend you to enable it. -- Bene* talk 08:59, 10 April 2015 (UTC)

PS: You may have to purge your browser cache in order to make this feature work the first time. Here are some links which work pretty well with the popups: Earth (Q2), Calocedrus decurrens (Q1399164), instance of (P31). I'm happy about your feedback. I already thought if including aliases or other information like general statements would be a good idea.
Thanks so much for making this work, Bene! :) --Lydia Pintscher (WMDE) (talk) 09:05, 10 April 2015 (UTC)
Yay, no more /*<![CDATA hovercards! Thanks! :) —DSGalaktos (talk) 09:56, 10 April 2015 (UTC)
Nice indeed! Displaying the target label and Qid would however be useful for links like this one. Any place where we can comment on that feature? -- LaddΩ chat ;) 12:05, 10 April 2015 (UTC)
\o/ TomT0m (talk) 12:31, 10 April 2015 (UTC)
@Laddo: You can add comments on the gadget's talk page for the Wikidata specific hack and on the beta feature discussion page for the Popups extension itself. Including the label and item id might indeed be useful in this case but will be redundant in other cases. I will think about how to find a user-friendly solution here. -- Bene* talk 21:06, 10 April 2015 (UTC)

Controversial deletion marked as non-controversial housekeeping

I created Help:Class or instance, which was deleted by User:John F. Lewis. The deletion broke links that I had made to that page. IP- (talk) 22:50, 10 April 2015 (UTC)

The deletion was not controversial, perhaps using 'non-controversial housekeeping' was not the best summary, 'deleting unnecessary redirect' or similar may have been better. John F. Lewis (talk) 22:54, 10 April 2015 (UTC)
John F. Lewis: 'deleting unnecessary redirect' when there were links from talk pages to it? Why delete a redirect that was in use? IP- (talk) 23:16, 10 April 2015 (UTC)
There is a difference between 'used' and 'useful'. When a page had no content on it which was actively used (e.g. around for less than a day), then it doesn't need a redirect to its new location. Instead updating the links is easier and more helpful than keeping around a page which is not helpful and provides no purpose to the knowledge base. John F. Lewis (talk) 23:20, 10 April 2015 (UTC)

Threatening with block when enforcing P279=rdfs:subClassOf and P31=rdf:type and other problematic behavior by User Denny

A) For enforcing

which is even stated on that Wikidata property pages, a user named Denny threatens to block me. [5].

B) S/he suggests to revert changes because s/he does not "consider the changes in the descriptions of the letters nor the switch from instance to subclass to be improvements". This, without giving any explanation, why the changes are not improvements.

C) S/he also writes "please either learn to behave, or leave the project." - I have learned to behave, and I do behave. Of course not in the way Denny does. My edits are based on applying scientific logic to items. I am not interested in applying what Denny can see as improvement or not. And this probably won't change.

On the Wikidata main page it says "Wikidata acts as central storage for the structured data of its Wikimedia sister projects including Wikipedia, Wikivoyage, Wikisource, and others."

It does not say "Wikidata acts as central storage of what Denny understands and wants".

IP- (talk) 16:12, 9 April 2015 (UTC)

Same old behavior? You was blocked here too. --Succu (talk) 16:25, 9 April 2015 (UTC)
Hi IP. I also think you're wrong about letters being subclasses. But let's discuss this in the thread your're reporting here (on this very page), and please in a constructive manner and not in personam. --YMS (talk) 16:23, 9 April 2015 (UTC)
Oh, a new user whose behavior is a lot like an old one ... The issue is nethertheless interesting, so I'll take it as an opportunity to push my Help:Classification page whose goal is to find way to solve this kind of issues. TomT0m (talk) 16:29, 9 April 2015 (UTC)

I never threatened you with a block. I couldn't even follow through with such a threat, which makes it rather futile to even threaten it. I was suggesting that you behave, and I am afraid that you continue to have a not very helpful approach towards discussions, which I consider toxic for the project and which I would prefer not to see.

Regarding your actual points raised: a letter like C (in particular in the sense of the abstract letter like C) is more useful to be regarded as an instance of a letter, than as a subclass of all letters, where the actual instances are the occurrences of the letter C in the real world. Because this allows us to make a query like "Give us all instances of letters", and we would get A, B, C, Đ, Д, etc. If we ask for all subsets, we would get a much bigger set, and not all of these would be letters (like, "letters for vowels" would be such a subset which would be a subset of letters, but also "Capital C" or "the letter C written with more than a single stroke" would be such subsets).

So purely pragmatically, to be able to ask for all letters it seems helpful to regard C as an instance of a letter (it certainly is not the only way to do so).

It also is useful to add further information to letters: maybe to which alphabets they belong, from what other symbol they were derived, their frequency in a given language corpus, etc. Not many of those have a sensible interpretation if regarded as sets: it sounds not intuitive to claim that the set of all occurrences of the letter A is derived from the set of all occurrences of the Greek letter Alpha. That's not a relation between the sets (as a subclass would indicate) but rather between the instances of letter (note, not of their occurrences).

Wikidata is not a project to develop an top-level ontology and to explain the world in terms of that novel ontology. It is a data backend primarily geared towards supporting the Wikipedias, but also other use cases. The data here should be potentially well referenceable. We do not try to create a taxonomy and put everything into it, but rather we want to reflect what sources state about the world. And whereas I can find plenty of sources which state "B is a letter" or "the letter B", which I will provide if so challenged by your sources (if you have children, try their first grade books), I would like to ask to see sources saying that "B is a subset of letters".

It obviously can mean that if you define letter as the set of all tokens of letters in the world, and then B as the set of all tokens of the letter B in the world. But if I define letter to be the set of all types of letters - i.e. as {A, B, C...} - then it is just as correct to have an instance of relation between that set and B.

Unlike the way you try to paint it, there is no clear true or false. But since it is both pragmatically more useful as well as better supported by literature, regarding B as an instance of letter seems to me to be more aligned with the way Wikidata is set up, and this is why I suggest to revert your changes with regards to that topic once this discussion has been resolved with sufficient consensus. --Denny (talk) 21:56, 9 April 2015 (UTC)

Set theory and type-token distinction


  1. Is set theory the right discipline here? "In disciplines such as logic, metalogic, typography, and computer programming, the type–token distinction is a distinction that separates a descriptive concept from objects that instantiate the concept, seen as particular instances of it." Source en:Type-token_distinction. And
  2. Regarding the query: If the items have no instances or subclasses, then a query can find them. But if they have, then there would be instances (e.g. physical objects) of instances (abstract objects). Instances of instances?
  3. "it sounds not intuitive to claim that the set of all occurrences of the letter A is derived from the set of all occurrences of the Greek letter Alpha." - And no single Latin A nor the class of Latin A nor the changing set of "all occurrences of the letter A" is derived from the still changing set of "all occurrences of the Greek letter Alpha".

Could it be that set theory is misleading when working with rdfs:subClassOf and rdf:type (instanceOf)? IP- (talk) 01:06, 10 April 2015 (UTC)

When dealing with rdf:type and rdfs:subClassOf, Set theory seems quite a safe bet - the 2004 RDF standard was a Model theory 'couched in the language of set theory', and the 2014 version is also using the terminology of set theory throughout [6].

Instances of instances are no problem. Clarence the lion is an instance of lion. Lion is an instance of species. Lion can be endangered, without Clarence being so.

Let's try another example: there is an interpretation where Germany is a Subclass of Country - if you regard Country as the "set of all land that that belongs to a country", and Germany as the "set of all land that belongs to Germany", then yes, Germany would be a subclass of Country. I'd still say that this is an unusual interpretation - and so I do for letter and A. I think stating that letter is the "set of all occurrences of letters" and A is the "set of all occurrences of the letter A" is an unusual interpretation. In both cases having an interpretation "Germany is an instance of country" or "A is an instance of letter" seems to be much more aligned with literature. Also, the identity criterion of sets is usually its membership. And by writing new As, I would be changing the set 'A'. I would be changing the set 'letter'. I do not think that this is what we mean when we talk about A. I am arguing that when we talk about A, we mean the abstract idea of the letter A, independent of the actual set of occurrences.

You state that 'If the items have no instances or subclasses, then a query can find them' - which is correct. The problem is rather that it would find far too many results if you ask for all subsets of letters. Let's put it this way: if we make A a subclass of letters, what is the query to get all letters (and only the letters) back, i.e. a query that results in (A, B, C, D ..., X, Y, Z) (for the common English alphabet).

What annoys me is mostly your tone earlier to Haplology and also in the discussion of your ban on English Wikipedia. The questions discussed here are not as trivial and clearcut as you seem to believe, and not everyone who believes otherwise does have 'problems understanding the concepts' or issues 'related to efficiency of the thinking-process'. In my opinion, you should apologize, independently of the result of the discussion we are having here. --Denny (talk) 15:19, 10 April 2015 (UTC)

Denny - To get all letters one just asks for all subclasses of a given class. If you want all that are in ISO 646 you query for items that are in the subclass tree of "Letter in ISO 646". Germany is not landmass but a state that has sovereignty over some landmass and some non-landmass. I would never regard Country as "set of all land that that belongs to a country". Did you read and en:Type-token_distinction? IP- (talk) 15:55, 10 April 2015 (UTC)
vowel (Q36244)     is a subclass of all letters. And it is obviously not a letter. TomT0m (talk) 16:02, 10 April 2015 (UTC)
What TomT0m says. --Denny (talk) 16:13, 10 April 2015 (UTC)
Denny: What TomT0m says is wrong. Vowel is not a subclass of "all letters". Vowel is a subclass of speech sound. IP- (talk) 16:17, 10 April 2015 (UTC)
Sigh. "letters representing vowels" or whatever you want to call it, the set {A, E, I, O, U}, is a subclass of "letter" and yet this set should not be part of the result set as I described it. Did you really not understand what was meant? --Denny (talk) 16:29, 10 April 2015 (UTC)
Classes in Wikidata are normally not named as being the plural of its members. Any letter could represent a vowel. 'w' sometimes is a pulmonic consonant, sometimes it is a semivowel. There is no one-to-one correspondence between speech sounds and letters or letter combinations. Unicode has designation "Letter" for some of the characters it encodes. So you could try to query by "has Unicode designation letter". IP- (talk) 16:59, 10 April 2015 (UTC)
As said, whatever you want to call the set {A, E, I, O, U}, that's a subclass of 'letter' which should not be returned to the query that I described, but would when using subclassing instead of instantiation. --Denny (talk) 18:38, 10 April 2015 (UTC)
There is no such set in Wikidata, so it would not be retrieved. IP- (talk) 19:07, 10 April 2015 (UTC)


Indeed, I wrote stuffs about metaclass (Q19478619)     in knowledge representation in the semantic web. The well know ontology software. Some uses of metaclass with protege (to model some "fuzzy" concepts) could even be very useful in Wikidatas context, as the person (Q215627)     whose two wikidatians of different languages are not sure they actually refers to same definitions. (see this presentation in which the author models part/whole as metaclasses, instanciated by real whole/part classes, like engine/car. TomT0m (talk) 15:46, 10 April 2015 (UTC)

TomT0m: you created:
also added
  • 'metaclass is a part of a programming language' to the third metaclass item in Wikidata [10]
The Wikipedia article lacks references for claims you inserted there. IP- (talk) 16:16, 10 April 2015 (UTC)

Behavior discussion

Denny: "What annoys me is mostly your tone earlier to Haplology and also in the discussion of your ban on English Wikipedia." - OK. And I am annoyed by people that base discussions on number of edits they made and by nonsense. Admins at enWP now admit that the original message regarding the block - not ban as you say - was misleading. IP- (talk) 15:46, 10 April 2015 (UTC)

What do you mean with "OK"?
And here you use the term 'nonsense' to discuss the contribution of other editors to this discussion. Are you seeing why this could be considered hurtful to other contributors?
What is the difference between a 'ban' and a 'block' in this case? But I freely admit that the usage of the word 'block' would have been technically more correct.
I also don't see what the admittance of the original message being misleading has to do anything with the fact that your contributions to that discussion was sprinkled with expletives.
You were claiming that I was threatening you with a block (which, by the way, is still the title of this section). Are you still standing to this claim? --Denny (talk) 16:32, 10 April 2015 (UTC)
  • With "OK" I wanted to tell you that I received that message, but I don't know what I can do about it.
  • "And here you use the term 'nonsense' to discuss the contribution of other editors to this discussion" - That is false.
  • "What is the difference between a 'ban' and a 'block' in this case?" - These are different concepts in English Wikipedia. In a thread on edit merging I openly wrote that I was blocked in English Wikipedia, which was then used by User:Haplology in a completely unrelated thread, namely about Wikidata content. Now you made a false statement in relation with my account and I prefer to be portrait correctly.
  • If you want to proceed with talking about my contributions in English Wikipedia, I would like you do so in the English Wikipedia. If you want a centralized discussion I suggest Meta.
All in all I prefer to focus on content discussions. No, idea why you re-opened this behavior stuff. IP- (talk) 16:50, 10 April 2015 (UTC)
In case you do not know what to do about it, I suggest to apologize.
What do you mean with the term 'nonsense' then, since my understanding was false?
I was not aware that bans and blocks were different concepts in the English Wikipedia. What I did mean was 'block', although I used the term 'ban'. I apologize for misusing the term.
I am not sure with what you mean with "No idea why you re-opened this behavior stuff", as you were the one creating a subheader here in order to discuss this.
In general, if someone does not behave appropriately, that person should be called out and asked to improve their behavior. If you prefer to focus on content discussion, then do so and do not attack others in person.
You did not answer my last question regarding your claim about me threatening you. --Denny (talk) 17:01, 10 April 2015 (UTC)
  • nonsense - referred to text in the blocking message(s?).
  • reopening - referred to your "What annoys me is mostly your tone earlier to Haplology and also in the discussion of your ban on English Wikipedia." in a section about "Set theory and type-token distinction"
  • "In general, if someone does not behave appropriately, that person should be called out and asked to improve their behavior." - That would not be my approach. It sounds as if there are people that think they have the knowledge to divide behavior in wanted and unwanted.
  • threat - I perceived it as such, when you said "The user was blocked merely for his username, which does not fulfill EnWp requirements. Looking at the way the user is conducting discussions, though, other reasons seem suggestive, too.".
Cheers. IP- (talk) 17:21, 10 April 2015 (UTC)
And [11] instead of discussing supports my point of view. Good that everyone can see our behavior. :-) IP- (talk) 22:19, 10 April 2015 (UTC)
Whatever you say. You are not discussing things. You continue to delete information like "P is a letter", "Hydrogen is a chemical element", "the English alphabet is an alphabet" without consensus. If you would be merely discussing these changes it would be different, but since you are not but rather continue to make potentially disruptive changes to the database, I have asked administrator's to keep an eye on you. --Denny (talk) 22:31, 10 April 2015 (UTC)
Denny: You claim I removed the information
  • "P is a letter" - I did not delete that information.
  • "Hydrogen is a chemical element" - I did not delete that information.
  • "the English alphabet is an alphabet"- I did not delete that information.
Three false and defamatory statements by you. IP- (talk) 22:33, 10 April 2015 (UTC)
[12] --Denny (talk) 22:35, 10 April 2015 (UTC)
Denny: What do you want the reader to tell with that link? Take care, all your edits are public. IP- (talk) 22:41, 10 April 2015 (UTC)
@Denny: give away, he's just an idle troll who want to maximise chaos by maximising provocations. TomT0m (talk) 16:38, 10 April 2015 (UTC)
Denny - do you think this is nice behavior? IP- (talk) 16:50, 10 April 2015 (UTC)
No, I do not. I prefer to answer such behavior as yours with kindness instead of name-calling, but since you are dealing out I am surprised by you being surprised to be receiving as well. --Denny (talk) 17:01, 10 April 2015 (UTC)
I was not surprised. But I notice that you don't ask TomT0m to "learn to behave". IP- (talk) 17:21, 10 April 2015 (UTC)

IP-, edit warring without consensus? --Succu (talk) 19:04, 10 April 2015 (UTC)

Not by me, since the two cancelled each other out, even before you posted here:
  • (cur | prev) 18:59, 10 April 2015‎ IP- (talk | contribs)‎ . . (29,641 bytes) (-405)‎ . . (Undid revision 209896501 by IP- (talk))
  • (cur | prev) 18:58, 10 April 2015‎ IP- (talk | contribs)‎ . . (30,046 bytes) (+405)‎ . . (Undid revision 209894848 by Denny (talk))
IP- (talk) 19:11, 10 April 2015 (UTC)
So your edits within the last hours are based on a consensus? You was asked to reach one ebfore continuing. --Succu (talk) 19:19, 10 April 2015 (UTC)

The editor has now started with removing "instance of: alphabet" from topics like English alphabet (Q754673). --Denny (talk) 19:33, 10 April 2015 (UTC)

Blocked IP-, this person has plenty of usernames and a history of being abusive. Multichill (talk) 09:50, 11 April 2015 (UTC)

Property:P1411 (nominated for)

Why do movies like Birdman or The Artist don't have the property nominated for: Academy Award for Best Picture but only award received: Academy Award for Best Picture? Is there some tacit agreement to not give the nominated for property to items which also won? --Jobu0101 (talk) 21:46, 9 April 2015 (UTC)

I don't think there is a tacit agreement but just a lack of users adding such statements. Note that nominated for (P1411) is used on 1316 items whereas award received (P166) has more than 174,000 claims.--Pasleim (talk) 22:14, 9 April 2015 (UTC)
It just looked a little bit strange to me that almost all movies which were nominated for the Academy Award for Best Picture this year got the property but the winner. But maybe it's unhappy coincidence. --Jobu0101 (talk) 22:23, 9 April 2015 (UTC)

I just checked it using a WikidataQuery: There is not a single film which received the Academy Award for Best Picture and also got nominated for it according to Wikidata. But there are 290 films in total which were nominated for the Academy Award for Best Picture according to Wikidata. This obviously shows that the people who set the nominated for property had some tacit agreement to not give the property to items which also won the Award. --Jobu0101 (talk) 17:04, 11 April 2015 (UTC)

academic degree: professor

49 items have that claim (autolist) which I think should be replaced with P106: professor because Q121594 is not an academic degree, but it is a profession. It would help a lot with value type violations for P512 overall --Haplology (talk) 03:58, 10 April 2015 (UTC)

I see that on svwp professor (Q495871) and professor (Q121594) are proposed to be merged. The first is not a disambig-article on svwiki.
sv:Professorstiteln i Sverige (professor in Sweden) maybe should have an item of it's own. It's a kind of Bonnie and Clyde-article, about Swedish professors under two different laws. -- Innocent bystander (talk) 04:10, 10 April 2015 (UTC)
Maybe so, but I think it's a separate issue from the concern that I have, which is that professor (Q121594) is not academic degree (Q189533) and therefore it is inappropriate to use Q121594 as a target value of academic degree (P512). But fortunately it is appropriate for occupation (P106) so I am proposing (a) -P512:Q121594 and (b) +P106:Q121594. It might also be acceptable for honorific prefix (P511) but only for Dutch professors (?) and it seems redundant anyway. --Haplology (talk) 06:51, 10 April 2015 (UTC)
The problem here is that the word have multiple meanings. Earlier a Swedish professor was a profession with some very special rights connected to it. (Since they were always assigned by the Government.) Today it tells more about somebody's income than anything else. -- Innocent bystander (talk) 07:26, 10 April 2015 (UTC)
In the Eastern European countries including Russia professor is both an degree (I am not sure whether it could be properly called an academic degree, since no thesis should be submitted to get it) and a position. Someone can occupy the position of a professor but not have an degree of a professor.--Ymblanter (talk) 07:49, 10 April 2015 (UTC)
I see, I didn't know that. In that case maybe a claim along those lines could be added to Q121594, by somebody knowledgeable in that area. --Haplology (talk) 11:33, 10 April 2015 (UTC)
My personal opinion is that there should just be two items then, one for the degree and one for the profession. I can't think of a any reason why they should share the same item. Hazmat2 (talk) 04:30, 11 April 2015 (UTC)
If memory serves, there might be a similar situation in Germany. I recall hearing about a court decision that declares "professor" to be a lifelong title, unconnected to whether you are employed. But I don't remember any of the details, and my memory may be wrong. WhatamIdoing (talk) 16:59, 11 April 2015 (UTC)

Unicode characters


Is there a project to import (all?) Unicode characters in Wikidata? I think mainly for arrows, bullets, pictures, emoticons... With GNOME, we now have a new application that lets you quickly and easily search for characters (like heart, which gives us ♥), but it only works in English. I am not aware of an international translation project for Unicode. To have the characters on Wikidata, would easily link to Unicode and HTML sequences (U+2605 and &#9733; for a black star), but above all, to be abble to translate official names, which would be useful to external projects (like free desktop environments, applications...) Okki (talk) 16:17, 10 April 2015 (UTC)

Would solve your problem? I'm not sure if wikidata intends to mirror about 200,000 assigned Unicode points, or even 17*65536 slots (most unassigned) covered by UTF-16 with surrogates, let alone 2^31 -1 in UCS-4 before they decided that some bits too much are almost as bad as not enough bits.
If a wiki already covers all code points in a block you are interested in with individual pages you could add it here. Otherwise, if "your" block is only covered by a list, or redirects to a list, or bitmaps on commons, knowing the rules here would help… Please post what you figure out, I'm curious. –Be..anyone (talk) 22:44, 11 April 2015 (UTC)

Linking to commons

I wanted to ask something about linking articles to commons. I have seen it done 2 different ways on here and I wondering if a certain way was preferred over the other. In some Items people use the properties for commons to link to the corresponding item such as [13]. On the same article I also linked to the category in commons under the Other wiki's section. In this example we can see them both togather, but I have seen other items with one or the other or both with no particular standard way. Which made me wonder if this was done in error or if there was even a standard way to do it. IMO, I think in the case of commons it would be better to do that under the Other wiki (in fact I think we should create one specifically for Commons in all honesty) and eliminate that from the properties but I would like to know what others think before I submit that suggestion in case this has come up before. I'm still fairly new to Wikidata so I don't know all the history of things. Reguyla (talk) 17:17, 10 April 2015 (UTC)

Thanks. Reguyla (talk) 20:14, 11 April 2015 (UTC)

Plural in English item label

For "Q867570" I changed the label "Cyrillic alphabets" to "Cyrillic alphabet" and was reverted by User:Succu [14].

S/he left a message on my talk page "Hör bitte auf... ... Labels nach deinem Gutdünken zu manipulieren. " which translates more or less as "please stop to manipulate labels at your discretion".

IP- (talk) 18:38, 10 April 2015 (UTC)

Now s/he deleted this very thread in the project chat [15].
I am looking for input by third parties. What I can see is, that items generally are written in singular form, e.g. "human", "car", "hand" and not "humans", "cars", "hands". IP- (talk) 18:46, 10 April 2015 (UTC)
Sorry this was a private message to you and not an other showact for you. BTW: You understand german very well. --Succu (talk) 18:50, 10 April 2015 (UTC)
Yes, I understand German, but you seem to have problems with English. IP- (talk) 19:53, 10 April 2015 (UTC)
@IP: I will not tolerate blatant insults like this. Consider yourself warned. --Haplology (talk) 02:34, 11 April 2015 (UTC)

There is no need to change the label, because there is no thing as "Cyrillic alphabet". Sjoerd de Bruin (talk) 18:59, 10 April 2015 (UTC)

User:Sjoerddebruin: Your claim is false, e.g. the Kazakh Cyrillic alphabet is a Cyrillic alphabet. IP- (talk) 19:52, 10 April 2015 (UTC)
It's a group of languages, not a individual language. Sjoerd de Bruin (talk) 19:55, 10 April 2015 (UTC)
User:Sjoerddebruin: Your claim is false. An alphabet is not a group of languages. IP- (talk) 20:03, 10 April 2015 (UTC)
Mind to reread the above „claim” of Sjoerddebruin, IP- --Succu (talk) 20:22, 10 April 2015 (UTC)
Succu : What is your interpretation of "it" in the claim by User:Sjoerddebruin? I read it as the subject in my sentence and that was "Kazakh Cyrillic alphabet". IP- (talk) 20:26, 10 April 2015 (UTC)
My initial „it” was here. Twisting things is not helpful. If you want to change a label move the article and have fun. --Succu (talk) 20:58, 10 April 2015 (UTC)
Succu: User:Sjoerddebruin didn't reply to your text. IP- (talk) 22:28, 10 April 2015 (UTC)

Cyrillic is a script.. It is used by many languages. It is singular. GerardM (talk) 21:09, 10 April 2015 (UTC)

Numerous alphabets are based on the Cyrillic script (Q8209)... --Succu (talk) 21:22, 10 April 2015 (UTC)
Yes there are subsets per language... GerardM (talk) 21:18, 10 April 2015 (UTC)
GerardM, Succu: The item is not about the script, but about the class of alphabets based on that script. There are various types of Cyrillic alphabets, like there are various types of aircraft. Boeing 747-400 (Q906937) is not called "Boeing 747-400s". How does Succu justify the plural? IP- (talk) 22:28, 10 April 2015 (UTC)
There is a standard for scripts; ISO 15924. There is only one script Cyrl. GerardM (talk) 04:11, 11 April 2015 (UTC)
Good for Cyrl, less clear for Latn+Latf. And figuring out the script(s) for IPA and UPA would be seriously bad. Be..anyone (talk) 23:08, 11 April 2015 (UTC)

Wikidata weekly summary #153

Fictional human or fictional character?

What's better? See here and here. --Jobu0101 (talk) 15:47, 9 April 2015 (UTC)

As "fictional human" is a subclass of "fictional character" and both Vincent Vega and Mia Wallace are fictional characters clearly intended to be humans, fictional human fits them better, as it's a more precise classification. --YMS (talk) 16:27, 9 April 2015 (UTC)
Alright. I fixed it. --Jobu0101 (talk) 17:02, 9 April 2015 (UTC)
Also sex or gender (P21) is usually applicable only to humans (including fictional ones). --Infovarius (talk) 05:10, 13 April 2015 (UTC)

Wikisource links

What to do with Wikisource linking in articles Q9372670, Q19796490, Q19796474, Q19142323. I can't understand why my edits are reverted. So far I can't keep linking other articles. --Janezdrilc (talk) 18:38, 12 April 2015 (UTC)

Fix double redirect

How can I make Q1158394 point to Q2128144? --- Jura 11:01, 12 April 2015 (UTC)

The solution seems to be "to undo the redirect" rather than attempt to edit it.
@Jura1: It is not recommanded to touch redirects, see Help:Redirects. I always found weird this custom of fixing double redirects. What if a merge is reverted in an article ? Then we have links that used to be good that may point to the wrong article when a more specific has been created. TomT0m (talk) 11:57, 12 April 2015 (UTC)
Well, something has to be done about double-redirects, otherwise people end up on blank pages. --- Jura 12:29, 12 April 2015 (UTC)
Mmm, never noticed that :/ that might be a problem for the policy. Not a big one though. I'll put something on the talk page. TomT0m (talk) 13:03, 12 April 2015 (UTC)
Don't worry. We fixed everything. --- Jura 18:08, 13 April 2015 (UTC)

Oscar award received

I tried to generate a list of all items which received an Oscar. I used autolist and claim[166:claim[31:19020]] but I didn't get a result. Why? --Jobu0101 (talk) 21:50, 12 April 2015 (UTC)

@Jobu0101: You need parenthesis to indicate requests to be executed before others: Query: claim[166:(claim[31:19020])] -- LaddΩ chat ;) 22:15, 12 April 2015 (UTC)
Thanks. By the way, what is the difference between Autolist and Autolist 2? --Jobu0101 (talk) 22:19, 12 April 2015 (UTC)
@Jobu0101: Autolist 2 is the newer version with these changes:
  1. Allow complement which is not possible in Autolist.
  2. Allow 4 source sets, which are Manual list, Category, WDQ and Find. only Category and WDQ are supported in Autolist.
  3. Allow setting namespace while doing quick intersection.
  4. Allow setting size of chunks (default is 10000). In Autolist this is hardcoded as 50.
  5. Allow adding more than one statement and removing statement from an item.
  6. Labels will be loaded on demand. In Autolist all labels are loaded before displaying the result.
  7. Items is sorted by alphabetical order (Q1, Q1001, Q11, Q2) instead of Qid (Q1, Q2, Q11, Q1001).

Note: Autolist is deprecated for editing.--GZWDer (talk) 05:06, 13 April 2015 (UTC)

Thank you. Is there some documentation for Autolist 2? I mean the most of it is self-explanatory but still a few questions remain. For example, what does "Mode WDQ NOT" mean? --Jobu0101 (talk) 06:42, 13 April 2015 (UTC)
It allows you choosing what intersection you want to do with the sets you have loaded. Eg. to show all items in a category or with some label, but not in the result of a query. Matěj Suchánek (talk) 11:32, 13 April 2015 (UTC)
Is there a template like {{query|...}} for Autolist 2? --Jobu0101 (talk) 06:56, 13 April 2015 (UTC)
@Jobu0101: I believe there isn't, but there is little need, {{Query}} does the work:
  1. Autolist (1) can show extra property values, but Autolist 2 does not
  2. You can easily cut-and-paste a query from Autolist to Autolist 2 if you want to
-- LaddΩ chat ;) 13:21, 13 April 2015 (UTC)

Using the parser function

In Wikidata weekly summary #153 I read that {{#property:P123|from=Q42}} should get parsed not only here in Wikidata but also in some Wikipedias. But as you can see () it doesn't even get parsed here in Wikidata, it just vanishes. --Jobu0101 (talk) 22:16, 12 April 2015 (UTC)

It'll come with the next code deployment. Not available yet :) --Lydia Pintscher (WMDE) (talk) 06:06, 13 April 2015 (UTC)
So what does "So far this is only Wikidata itself" mean if it doesn't even work there? --Jobu0101 (talk) 06:43, 13 April 2015 (UTC)
It means that it will come with the next deployment but since arbitrary access is only available on Wikidata itself this new feature will also only work there. The section usually contains things that are not live on yet but will be soon. --Lydia Pintscher (WMDE) (talk) 09:20, 13 April 2015 (UTC)
The big day is approaching .. put the champagne in the fridge . --- Jura 16:34, 13 April 2015 (UTC)

Linking between two properties?

Trying to make the description text from one property (which already contained the text "P###") link to another one, the preview didn't seem to understand my intent. It showed "[[P###]]", and i guess it wouldn't have saved it as a link... My second attempt was to "add qualifier" to one of the property's existing properties, specifically the "obsolete Wikidata property" statement... But i got this message when trying to save:

An error occurred while saving. Your changes could not be completed.
Details: Unexpected entity type property

Should i just add a new statement about that property, instead of trying to include the link into the description or a qualifier? Or is there some other option i'm not aware of? -- Jokes Free4Me (talk) 08:53, 13 April 2015 (UTC)

One can't link in descriptions. You can mention items or properties, see P:P21. --- Jura 11:56, 13 April 2015 (UTC)
Any particular reason for why we "can't link in descriptions"?! As for the second attempt, is there a similarly deliberate restriction that "values cannot be properties"? -- Jokes Free4Me (talk) 12:48, 13 April 2015 (UTC)
Every property has its datatype. Properties with item datatype cannot link to properties. Of course, there exist some properties with property datatype. Descriptions are used to describe, not to link. Matěj Suchánek (talk) 13:00, 13 April 2015 (UTC)
What property are you having a problem with and what are you trying to do? It is difficult to give you real help without this information. --Izno (talk) 12:54, 13 April 2015 (UTC)
Does it really matter?! Fwiw, i wanted to link p:P513 to p:P1477, or add a qualifier to the "obsolete Wikidata property" statement. -- Jokes Free4Me (talk) 17:20, 13 April 2015 (UTC)
Oh, you can use see also (P1659). Good to know. :-) -- Jokes_Free4Me (talk) 18:45, 13 April 2015 (UTC)

wikidata parser function

Please, can you tell me where can I find some information about #property parser function? I was looking for it on Italian Wikipedia but I didn't find anything. Also, there is a parser funtion to read the label of a Wikidata item? Thanks. --FRacco (talk) 10:14, 13 April 2015 (UTC)

Awkwardly enough, not in the Wikidata help namespace. This search yields a little more to dig through. --Izno (talk) 13:03, 13 April 2015 (UTC)
Try mw:Extension:Wikibase Client, mw:Help:Extension:ParserFunctions and mw:Extension:Wikibase Client/Lua. -- LaddΩ chat ;) 13:29, 13 April 2015 (UTC)

thriller subclass problem

Is it okay that thriller is subclass of film and literary work? I think that is wrong. If you use the subclass property for some item then that means that all instances belong to that class. Since in our case there are thrillers which aren't films and other thrillers which aren't literary work we should remove both subclass relations. Do you agree? --Jobu0101 (talk) 22:23, 11 April 2015 (UTC)

Yes It's wrong. It would mean with classical meaning of subclass of that the instances of this class are both films and texts. Thriller is more a genre. The subclass of film (Q11424)     should be something like thriller film, which would mean that all its instances are implied to be of genre (P136)   <thriller>.
Here we can see it: all instances of class X belong to class Y. So this is clearly wrong. If I delete this claim can I somehow refer to this discussion here? Maybe we should also tell the user who did that that he didn't understand the concept of being a subclass. --Jobu0101 (talk) 13:03, 12 April 2015 (UTC)
Use undo function and mention this discussion in the edit summary. Matěj Suchánek (talk) 17:19, 12 April 2015 (UTC)
Thanks, I didn't know that you can use the undo function to leave an edit summary. --Jobu0101 (talk) 18:51, 12 April 2015 (UTC)
@Matěj Suchánek: Oh, I misunderstood you. I thought you mean I should remove the claims and then try to undo my removal where the software instead of really undoing my removal will give me the possibility to leave an edit summary to justify my removal. But what you meant was simply to undo the edits of Infovarius (talkcontribslogs) and leave there an edit summary. --Jobu0101 (talk) 07:28, 14 April 2015 (UTC)
Or you could undo your change twice, and leave an edit summary on both undos. I agree this is not a nice thing, but it's a possibility. -- Jokes_Free4Me (talk) 08:54, 14 April 2015 (UTC)

Autolist down

How come? --Jobu0101 (talk) 19:37, 13 April 2015 (UTC)

It happens a lot. How come indeed. --Haplology (talk) 02:25, 14 April 2015 (UTC)
Interestingly, the API was still online. --Jobu0101 (talk) 07:30, 14 April 2015 (UTC)
There was a labs outage. Sjoerd de Bruin (talk) 07:33, 14 April 2015 (UTC)

Subclass is class and instance?

I just saw that Q3965271 is subclass and instance of Q130901. Is that possible or is that also a violation to Help:Basic membership properties like thriller (see above)? --Jobu0101 (talk) 18:55, 12 April 2015 (UTC)

@Jobu0101: It's not really possible to be both instance and subclass of the same item. My first impression here is that it is a subclass of relations as there is many possible subclass relationships depending of the system you work on. We could have
< Wikidata's subclass relationship > instance of (P31)   < subclass relationship >
though. But here it is labelled as a relationship of some axiomatic set theory, then it could be a concrete relationship also.
It's kind of hard to think of that example because it has a sense in Wikidata also :)
But I would say here the item is a for a kind of relations : those who respects some properties, such as the transitivity of membership of a class to its superclasses. So definitevely not an instance of relation. TomT0m (talk) 19:07, 12 April 2015 (UTC)
So if there are subclasses in two different senses, why don't we have two different items for them? --Jobu0101 (talk) 21:53, 12 April 2015 (UTC)
@Jobu0101: Because no one did the clarification job yet :) Actually this is in a mathematics area here that is close to self-reference (Q1129622)     so this might not be the easiest job in Wikidata :).
But maybe this link I also quoted in another discussion is a way to sort things out: In short, he uses a modeling trick to regroup stuffs with an abstract common pattern with class of classes, like whole/part relationships who can be seen in a lot of real world object relationship. Might be interesting to see if we can use a similar solution here. TomT0m (talk) 10:29, 13 April 2015 (UTC)
@Jobu0101, TomT0m: Actually, i am quite certain that subclass (Q3965271) is not a subclass of binary relation (Q130901) but of ordered pair (Q191290). -- Jokes_Free4Me (talk) 06:47, 14 April 2015 (UTC)
 :@Jokes Free4Me: Why do you have a "_" in the username that is displayed in your signature and not in your real username ? That could make sense if a relation is viewed as an ordered pair set, but in this sense it's equivalent to say that
< relation > subclass of (P279)   < ordered pair >
< subclass > instance of (P31)   < ordered pair >
. But I don't think this is really an advance for our problem, can you justify this a little bit, beyond provocation ? TomT0m (talk) 09:21, 14 April 2015 (UTC)
Because WikiMedia was intended to use underscores instead of spaces pretty much everywhere, including in usernames. But my underscore is a real intended character, and not a stand-in for a space. Fwiw, there was no provocation intended, i'm just trying to reach a valid categorization. As for the question, I would agree[1] with that first statement,
< relation > subclass of (P279)   < ordered pair >
, but would change the second one to
< subclass > instance of (P31)   < relation >
(which would imply
< subclass > instance of (P31)   < ordered pair >
). Anything wrong with these? -- Jokes_Free4Me (talk) 10:01, 14 April 2015 (UTC)
  1. No, actually, i had a vague notion that something was missing in my interpretation, but couldn't quite clarify my thoughts. I've just realized what it was: the 'relation' class is a subclass of the 'set of ordered pairs' class. I need to study this some more. :"> -- Jokes_Free4Me (talk) 10:01, 14 April 2015 (UTC)
  2. @TomT0m: Alright, i've done some further analysis, and there are three concepts here. I was initially using the default mindset, that presumes the "system you work on" is fixed, and in such a case there's only one "subclass relationship" for that system (i.e., the system's subclass relationship, let's call it SSR). That's why i was ready to say it's an instance of a binary and transitive relationship. However, without context, the concept of "subclass" contains, as you've said above, "many possible subclass relationships depending of the system" (all SSRs, i.e. the class of each SSR for any possible system). So it's a subclass of such relationships. And yet, opening the two wikipedia pages, they are treating it as merely the "name for the first element in any given pair in some SSR". Which is not a relationship, despite what the description here says, and thus the concept becomes a subset of "class (set theory) (Q217594)" and of "first element in a pair" (if we ever reach that level of detail). So now i'm really unsure how to proceed. Ideally, i think there should be a separate item for the "subclass relationship" we've been discussing here. -- Jokes_Free4Me (talk) 06:36, 15 April 2015 (UTC)

    Being both, class and instance

    Inspired by Subclass is class and instance and Help:Basic membership properties: Can an item be both, class and instance? I think, in programming languages it is forbidden for good reason. In Wikidata we have many (30,426) counterexamples: Query: claim[31] and claim[279]. Frankly speaking, in most cases it is obvious bullshit. --Jobu0101 (talk) 10:30, 14 April 2015 (UTC)

    @Jobu0101: In knowledge representation it is allowed : for example in the OWL language they added a feature to do this nicely. See metaclass (Q19478619)     and Help:classification. And it's really convenient in Wikidata as articles often covers both aspects. but I'm feeling tired of saying that, so I'll hope this settle the discussion @Community: can we adopt Help:classification as a guideline, please so that this discussions that last for at least 2 years be settled once and for all ? TomT0m (talk) 10:42, 14 April 2015 (UTC)
    In programming languages like Smalltalk classes are instances. In Wikidata we made the conscious decision not to have them be disjoint. --Denny (talk) 16:13, 14 April 2015 (UTC)
    But in many cases it is used wrongly. For example the first search result in Autolist 2: Q1000176. --Jobu0101 (talk) 16:21, 14 April 2015 (UTC)
    @Jobu0101: Authorize just to use subclass on class items of and you will have a bunch of stuffs wrongly marked as subclasses of each other. See the example of ship classes in Help:Classification. TomT0m (talk) 17:29, 14 April 2015 (UTC)

    I opened a RfC. TomT0m (talk) 17:26, 14 April 2015 (UTC)

    Lots of things can be classes and also instances of a particular type of class. The Ford Model T (Q182323) is a subclass of "economy car" and is an instance of a "car model". I don't think there are an individual Model Ts which have articles about them but there are individual books, trees, ships which have items in addition to the items for the editions, species and ship types that they are instances of. Filceolaire (talk) 00:17, 15 April 2015 (UTC)

    Ontology for sciences

    Hi, Is anyone aware of ontology of an ontology of sciences ? We can hardly find something consistent in this area.

    Is algebra a subclass of math, for example ? That's OK if we consider that maths are a set of knowledge. Then Algebra is a subset of that knowledge.

    We could also have a mathematics main subdiscipline item with a

    < Algebra > instance of (P31)   < mathematics main subdiscipline >

    . Any idea ? As it is archaeology (Q23498)     is both a subclass and an instance of "social sciences".

    by the way, the property proposal page for sciences seems a little dead ... Natural science property proposal, there is a property study of that could apply to math disciplines I proposed who seems blocked for months ...

    TomT0m (talk) 20:04, 14 April 2015 (UTC)

    Ice cream

    Hi, I've been looking into different kinds of ice creams, but not sure if they should be instances of ice cream or subgroup of ice cream. Example: soft serve (Q1397211) is set to be instance of (P31) ice cream (Q13233). Is this really correct? Another example: chocolate ice cream (Q3013199) is subclass of (P279) to ice cream (Q13233). Which one is correct when we talk about types of ice creams, not specific brands or products? //Mippzon (talk) 17:11, 12 April 2015 (UTC)

    @Mippzon: Maybe Help:Classification could be of any help :) The subclass form is correct according to the Token/type distinction, as children eats a lot of italian chocolate icecream. But if we regards in terms of class of classes, it's also possible if the icecream type is branded to include a
    < Whatever icecream brand > instance of (P31)   < icecream brand >
    both with
    < Whatever icecream brand > subclass of (P279)   < chocolate icecream >
    . Icecream brand can in this case be regarded as a class of classes. TomT0m (talk) 17:17, 12 April 2015 (UTC)
    Haha, this twisted my head a bit! So, am I following this correct: chocolate ice cream (Q3013199) could be both a subgroup to ice cream (Q13233) AND an instance of ice cream (Q13233)? Since we both means the abstract concept and the physical actual chocolate ice cream that some of us eat? :) //Mippzon (talk) 17:28, 12 April 2015 (UTC)
    The guideline I follow is that a class is a lot of instances and there generally aren't a lot of statements you can make about a class. An instance of an ice cream, on the other hand, is a particular recipe/brand and you can make statements about it - recipe, start date, brand owner, etc. Based on this guideline both soft serve and chocolate are probably subclasses of ice cream since there are various different recipes/brands of soft serve and of chocolate ice cream.
    I don't think ice cream can be both an instance and a class but some other things can. For instance a model T Ford is an instance of a brand/model with an introduction date and a cancellation date and is also a class of millions of individual cars, each of which is an instance of a Model T and each of which has a construction date and a destruction date, which are different from the dates for the brand/class/model. At least that is my opinion. Filceolaire (talk) 17:42, 12 April 2015 (UTC)
    Thanks for making it a bit more concrete! I think I got my answer there. //Mippzon (talk) 17:46, 12 April 2015 (UTC)
    @Mippzon: If you can help to write Help:Classification in a more accessible fashion, you're welcome :) because what Joe just said and what Joe said about car model is what I wanted to say, just replacing "Car model" with "Ice cream brand" :) TomT0m (talk) 18:27, 12 April 2015 (UTC)
    Treating a model of a car as an instance of a model makes sense to me too, but in practice it leads to thousands of constraint violations of P176, at least, maybe on others too. An instance of a model of a car is conceptually an instance of a product, but it's not connected to product (Q2424752), so how is a machine to know? We humans know, but it can't be queried. Meanwhile lots of other things like Canon Cat (Q1033523) are conceptually the same yet are treated differently: a Canon Cat (Q1033523) is an instance of computer, but a Volkswagen Taro II (Q10392384) is an instance of a car model, not a car, and is not connect to Q2424752/product anywhere. In truth there are two schools of thought about "ice cream" versus "ice cream brand" operating at the same time with no agreement or coordination. --Haplology (talk) 03:17, 14 April 2015 (UTC)
    I said something, not that it matters. --Haplology (talk) 12:43, 15 April 2015 (UTC)
    @Haplology: I'll ask you to ping me because it's easy to miss answers sometimes. The notification system is a great thing. Maybe you are right for products if they are a generalization of the car model. I took product in the sense of any instance of object that went out of some factory (and the english article says A manufacturer usually provides an identifier for each particular type of product they make, known as a model, which in consistent with that. We should look at the articles to sort out definitions or to see if there is a ambiguity. For the consistency, That's exactly why I proposed a RfC. TomT0m (talk) 13:43, 15 April 2015 (UTC)


    Could we please have some resolution about the domain of product or material produced (P1056)? It was created to be about products of mines, such as ores or natural gas, but now, more items make claims of P1056 about finished products etc. (161 out of 221 total). Please leave input on the talk page. Thanks --Haplology (talk) 04:48, 13 April 2015 (UTC)

    Thanks to those who responded. It looks like the consensus is to broaden property to match current actual usage. Would somebody else please do that when the time comes. Thanks --Haplology (talk) 12:42, 15 April 2015 (UTC)

    Wikidata report in GLAM newsletter

    I've started a new Wikidata report, as part of the "the Month in GLAM" newsletter. The first edition is in draft. I'll try to write something each month, but please feel free to add your own contributions. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:31, 15 April 2015 (UTC)

    Date written (not published)

    How do people put the date that something is written, but no published, such as in a letter? For now, I put the date under publication date (P577) for Letter from Louis XIV to Count Tallard (Q19096740) but it's not really appropriate as it's a letter and "published" implies something different. Thanks, Hazmat2 (talk) 12:36, 14 April 2015 (UTC)

    You could use significant event (P793) with qualifiers point in time (P585) and location (P276) --Pasleim (talk) 16:41, 14 April 2015 (UTC)
    Sounds like a plan. Thanks, Hazmat2 (talk) 17:00, 14 April 2015 (UTC)
    Come to think of it, what would significant event (P793) go under? Or could you edit the item above to give an example? Thanks, Hazmat2 (talk) 17:07, 14 April 2015 (UTC)
    I think inception (P571) could work. It's described as "date or point in time when the organization/object was founded/created". –Hardwigg (talk) 18:21, 14 April 2015 (UTC)
    Second that. Sjoerd de Bruin (talk) 18:22, 14 April 2015 (UTC)
    That sounds a bit better. Thanks, Hazmat2 (talk) 18:51, 14 April 2015 (UTC)
    It´s not that difficult: inception (P571) is for the date of creating a written work, publication date (P577) is for the first and any other edition. So we usually have multiple items, one for the work, the other for a distinct edition. Somtimes a work is written down centuries or even thousands of years before it is published in print like Gospel of Thomas (Q131546) or Story of Wenamun (Q1579541). Sometimes there are editions in different langugages, different editions in different years in the original language etc, each needs an own item to identify it clearly.--Giftzwerg 88 (talk) 23:58, 15 April 2015 (UTC)


    Greetings, I've been away from Wikidata and now that I came back I would like to be informed about the state of reference. Most of the items are unsourced I those who aren't, normally the properties used are disputed. The documentation I've found doesn't really help, since it's to ambiguous. There is any Wikiproject to reference items or any page that suggests what is a reliable source in Wikidata, maybe even a sourced item that can serve as an example? Thank you in advance. - Sarilho1 (talk) 21:52, 15 April 2015 (UTC)

    Help:Sources. You may check one of Wikidata:Showcase items. -- LaddΩ chat ;) 21:59, 15 April 2015 (UTC)

    Head of Government

    I've been fixing constraint violations on head of government (P6) and I was wondering if there was a property that could be used for "Head of government corresponding to this item", since head of government (P6) is a violation. Popcorndude (talk) 19:20, 14 April 2015 (UTC)

    Sorry, that was really vague. I'm specifically talking about cabinet (Q640506) Popcorndude (talk) 19:21, 14 April 2015 (UTC)
    As I understand it, cabinets are listed using executive body (P208), while head of government (P6) is just for people. Do you have an example of where this doesn't work? --Arctic.gnome (talk) 20:23, 14 April 2015 (UTC)
    This is an instance of the cabinet itself having head of government (P6), e.g. 28th Canadian Ministry (Q220542): head of government (P6) => Stephen Harper (Q206) Popcorndude (talk) 21:51, 14 April 2015 (UTC)
    Interesting. I don't think Harper is head of government OF the cabinet. A more accurate property might be chairperson (P488). Even better would be something like: item exists pursuant to incumbency of => Stephen Harper (Q206) => P794 (P794) => Prime Minister of Canada (Q839078). – The preceding unsigned comment was added by Arctic.gnome (talk • contribs).
    Might officeholder (P1308) work for this? Or should I propose a new property? Popcorndude (talk) 19:09, 15 April 2015 (UTC)
    I had some doubts of head of government (P6) when it was introduced, and I translated it more fuzzy as "ledare" (leader) in Swedish. I didn't thought "Head of Government" fit to the corresponding to Mayor in Sweden. -- Innocent bystander (talk) 19:47, 15 April 2015 (UTC)
    A new property might be useful. It could have a slightly wider application than just ministries:
    --Arctic.gnome (talk) 06:48, 16 April 2015 (UTC)
    I have proposed such a property: Wikidata:Property_proposal/Unsorted#corresponding_incumbent Popcorndude (talk) 11:47, 16 April 2015 (UTC)


    Is someone in contact with VIAF (the Virtual International Authority File)? Apparently, they're now also using data from Wikidata, but the links to Wikidata they generate are malformed. For example the entry for Friedrich Schiller at - under "Authority/Source Record" there's a link to - which results in an "Invalid ID: Friedrich_Schiller" message. I suppose it should link to instead. Gestumblindi (talk) 20:53, 15 April 2015 (UTC)

    I'll poke someone there. --LydiaPintscher (talk) 10:46, 16 April 2015 (UTC)
    We also have an on-going dialogue with them, at en:Wikipedia talk:Authority control/VIAF. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:14, 16 April 2015 (UTC)

    Sorry, this is a result of the change for Wikidata going in before we switched the underlying data. Should be fixed today or tomorrow ThomasBHickey (talk) 16:40, 16 April 2015 (UTC)

    Thanks, appears to be fixed. Out of curiosity: Why does the link point to the anchor "sitelinks-wikipedia" (e.g. at ) instead of just to the start of the entry? Gestumblindi (talk) 19:42, 16 April 2015 (UTC)

    Lack of easy edit summary

    By the way, that's something I miss on Wikidata... it's not possible to give a short reason for an edit you make, an "edit summary", except through a less than evident detour, as mentioned in a recent discussion here. For example, I'd like to remove the German Wikipedia article from Q885115 as the topic is different from the English article (and the supposedly Marathi one, which actually is in English as well). The English article is about "limitations and exceptions to copyright" in general, whereas the German article deals specifically only with German law (it's not an article needing globalization, as it's defined as an article about this issue in German law - a general article like in en-WP could be written separately). So, the interwiki is wrong. However, it's not self-evident that it's wrong, and in such cases I would very much like to leave a comment with my edit. But if I click on "remove", it's just removed without giving me a chance to comment. If I gathered it correctly, I could leave a comment by undoing my edit, leaving a comment there, and then re-reverting. Not intuitive at all - and it needs three edits (remove, revert with comment, revert again) instead of one... Gestumblindi (talk) 21:37, 15 April 2015 (UTC)

    @Gestumblindi: phab:T47224.--GZWDer (talk) 04:59, 16 April 2015 (UTC)
    Mockup Wikidata sitelink edit summary
    User:Ricordisamoa made a mockup for a possible interface to allow specifying an edit summary on the Wikibase extension, in 2013. --Atlasowa (talk) 13:04, 16 April 2015 (UTC)
      Support --Jobu0101 (talk) 13:31, 16 April 2015 (UTC)
      Support, good plan. When I removed de:Vorlage:Vorlage from Template:Tlx (Q5896717) I "obviously" (not) added it to Template:Tl (Q3926051), where it fits better, if at all. Nobody interested in {{tl|tlx}} or {{tlx|tl}} has a chance to figure out why I did that. –Be..anyone (talk) 15:08, 17 April 2015 (UTC)

    Is there any chance that this IMHO quite important and basic functionality of the interface can be added in the near future? When I see that a mockup was already made back in 2013 and nothing happened, I wonder about the priorities here... People should be able to make edits that are transparent and comprehensible instead of being forced to remove/add stuff without any explanation. Gestumblindi (talk) 18:25, 17 April 2015 (UTC)

    For the item I mentioned, I did now use the complicated detour with three edits... Gestumblindi (talk) 18:32, 17 April 2015 (UTC)

    WikidataQuery question: Items which use title (P1476)

    How can I find all items which use title (P1476)? Query: claim[1476] does now work. --Jobu0101 (talk) 11:45, 16 April 2015 (UTC)

    Special:WhatLinksHere/Property:P1476. It is mostly used in references. I dunno how to query references, though. -- LaddΩ chat ;) 12:45, 16 April 2015 (UTC)
    I'm looking for a WikidataQuery. --Jobu0101 (talk) 13:47, 16 April 2015 (UTC)
    Unfortunately quering references is not implemented. --Succu (talk) 14:11, 16 April 2015 (UTC)
    @Succu: No, it's not about quering references. I want to find all items which use title (P1476) as property. --Jobu0101 (talk) 14:30, 16 April 2015 (UTC)
    @Jobu0101: Query: claim[1476] does work, there's just only one item using it as a property... You might have more success with Query: claim[357], though it's been made obsolete recently. -- LaddΩ chat ;) 01:38, 17 April 2015 (UTC)
    @Laddo: No, there are many items using this property. See for example Q15648198 and Q172241. --Jobu0101 (talk) 08:02, 17 April 2015 (UTC)
    @Jobu0101: Wow, I though they had fixed those issues with WDQ... According to Wikidata:Database reports/Constraint violations/P1476, there are indeed 4210 items using it as a property. WDQ is an external tool managed outside WD (code at BitBucket), and WD does not yet have such a service (see Wikidata:Development_plan#Simple_queries and Wikidata:Development_plan#Complex_queries). Last time there was such a glitch, they had to reset the tool manually. I asked the dev team if they can do something about it. -- LaddΩ chat ;) 12:14, 17 April 2015 (UTC)

    Confused about where some entries came from

    I apologize if this has been asked before but I have noticed a number of entries that look like this one with virtually no data and zero links to other Wiki's. I was wondering if this was an error of some kind or if this was done for a reason. Reguyla (talk) 20:14, 16 April 2015 (UTC)

    There is data and it's not a error. I think a bot added items for every town in China... Sjoerd de Bruin (talk) 21:05, 16 April 2015 (UTC)
    Thanks, I can't help but wonder where it came from though since it doesn't link to any wiki's. Was it based on data from another site? I'm just asking so when I see it I can link to the site or fill in info so its more useful. I can't really do that if I cant figure out where the data came from.Reguyla (talk) 22:47, 16 April 2015 (UTC)
    I believe this is the bot you are asking about, and yes, another site is listed a the source, just not on the items. Popcorndude (talk) 22:58, 16 April 2015 (UTC)
    @Reguyla: To make it clear: The are three links to Chinese Wikipedia. But there are not side links (= links to articles). The links are "hidden" in the sources ("imported from"). --Kolja21 (talk) 23:56, 16 April 2015 (UTC)
    Ok great thanks for your patience in helping me understand that. Still new to Wikidata so trying to figure out how to do stuff here. Reguyla (talk) 13:45, 17 April 2015 (UTC)

    WikiProjects (Simple English Wikipedia)

    As many of you probably know, all WikiProjects at Simple English Wikipedia are hosted in user space, not in main space. At one point, at least, Wikidata was permitting links to these project pages, notwithstanding the general prohibition against links to user space here. Does that exception still exist? StevenJ81 (talk) 17:00, 17 April 2015 (UTC)

    Gender Vandalism (and a tag for gender changes?)

    Coming from En.Wiki, this place seems almost like a vandalism free wonderland, but I immediately stumbled upon some person who had been labelled as transgender for over six months. This got me wondering whether WikiData had recent changes tags, which you do, and why there isn't one for changing the gender field. When you think about it, once set the gender field shouldn't have to be changed, and any changes to it should be watched as it is likely they are vandalism. What's everyone else think? Keep in mind I have no idea how tags work, but I'm sure it would be possible to have a tag for changing Property:P21. EoRdE6 (talk) 01:32, 14 April 2015 (UTC)

    No we don't as far as I know at least, and I agree that it would be helpful to pay more attention to changes in existing data because in my experience, new or anonymous users who change values are more likely to be vandalizing them than users who add new data. I know it's not quite the same but the closest thing for P21 is abuse filter 13. --Haplology (talk) 02:17, 14 April 2015 (UTC)
    I would note that people who have been long known as male or female may become publicly known as transgender, and thus would need a change made in Wikidata. Happens often enough. - PKM (talk) 18:27, 14 April 2015 (UTC)
    I think it is reasonable to require a reference for any use of a gender other than 'male' or 'female'. Could we create a tool to flag such cases? Filceolaire (talk) 22:20, 15 April 2015 (UTC)
    At least this one was tagged as a problem. EoRdE6 (talk) 02:20, 16 April 2015 (UTC)

    It's so much worse

    It's worse than just that. I was in the pub yesterday with a friend who looked something up on Wikipedia. At the bottom of the article, thanks to the new mobile version of Wikipedia, there were three links to related people. One of these was a link to Thomas McCarthy (en.wp page), who was helpfully described as an "American actress". That data was pulled straight from Wikidata and has been sitting there unhelpfully since 31 July 2014. If the Foundation are going to use Wikidata as the source for descriptions to annotate the new mobile UI, then the Wikidata project needs to up its game about handling vandalism.

    Vandalising Wikipedia affects only one version of Wikipedia. Vandalising Wikidata affects all Wikipedias (and maybe Wikivoyages, Wikibooks, Wikisources, Wikinewses and so on). We have no limitation on preventing IP editors from editing the data items that represent living people.

    In the past I tried to monitor recent changes, but the sheer amount of changes that were coming in, combined with the seeming lack of anyone else doing much RC patrolling, made the exercise feel a little like poor Sisyphus pushing the rock up the hill.

    The current situation seems to be this: any random IP address user, with no history on Wikipedia or any other WMF project, can turn up, change the name, gender, profession or description of a living person, and that data will seep straight into Wikipedia without anyone noticing. There are less eyes on Wikidata: there aren't so many recent changes patrollers, and Wikipedians who might care about some topic are probably not monitoring for changes to the equivalent data item on Wikidata. We should probably find a way to solve this problem before it gets out of hand. That might mean (heresy warning!) turning off IP editing. Or now that we have SUL, having some way of prioritising checking claims that are added to data items by users who have no discernible experience on any other WMF project. Most legitimate editors to Wikidata will have some editing history on, say, Wikipedia.

    If someone comes over from English Wikipedia and they've got 20,000 edits and are an administrator, we should probably trust their edits slightly more than we do an IP with no editing experience anywhere on any WMF wiki. For a lot of Wikipedians, Wikidata is just plumbing that powers inter-wiki links and maybe infoboxes and now the mobile UI. They don't want to have to think too much about the plumbing, and when the plumbing breaks, they don't want a catastrophic reversal of misinformation squirted all over their wiki (if only because there's enough there already that needs hunting down). —Tom Morris (talk) 07:38, 17 April 2015 (UTC)

    @Tom Morris: We definitely have to change the way Wikipedians look at Wikidata. Wikidata is not (just) plumbing, it's a powerful tool for cooperation, interproject and interlanguages. For example for fighting vandalism, when the issues of watchlist integration will be solved at least, this means any WIkipedian or Wikimedian that watch any article linked to an item will be able to fight vandalism on every other project through Wikidata. If Wikipedians don't get that and continue thinking Wikidata as an external project, and don't play a cooperative game, this could turn into a fight. No-one wants that. TomT0m (talk) 08:01, 17 April 2015 (UTC)
    @TomT0m: Wikidata is a central clearinghouse for structured content. There are four ways to relate it to the rest of the sisterhood — two separate binary decisions. You can have either instant propagation, or when a change takes place on one side of a connection you can offer local editors on the other end of the connection with an opportunity to adopt or not adopt the change at their end. And, you can have this propagation be only from Wikidata to other projects, or you can have it go both ways. Instantaneous propagation would be a vandal's paradise, so it's really sad to see that being pushed; on the other hand, tying different projects together by offering local editors the opportunity to consider adopting changes at another project seems to me a really marvelous way to maximize local control, local morale, and cooperative spirit (all of which instant propagation minimizes). With instant propagation, two-way propagation only magnifies its destructive potential. However, once one embraces semi-automated optional propagation, two-way propagation becomes a great idea; somebody at any local project makes an edit, and editors at Wikidata are notified of it and offered an easy way to accept it if that seems wise, then once they do so, editors across the sisterhood are offered the same choice; and that one user on a local project is able to share their knowledge with the entire sisterhood through human cooperation without the likelihood of blatant vandalism, and with a huge morale win all around. --Pi zero (talk) 14:47, 17 April 2015 (UTC)
    @Pi zero: I don't know, it seems to me a way to separate communities. I would better have a situation in which everyone feels concerned, not a one where some can reproach the others stuffs like "this is nonsense, I spend my time refusing your bullshit offers, please get YOUR community in order" (I would nether talk like that, but it's a possibility ;). This would put a lot of pressure on Wikidata editors, and the others would feel as clients of the company, like authorised to complain because of whatever reason they will find. I'd better have a situation in which everyone feels concerned and everyone contributes in Wikidata. Think that the people who contributes in infoboxes right now would just have to contribute to Wikidata instead. Otherwise they could decide, "they have bad datas, I'll contribute to my infoboxes manually", and we are stuck into this situation. The only change for them is that instead of tracking vandalism on infoboxes they would have to track vandalism in Wikidata, seems not really like a big deal. TomT0m (talk) 15:08, 17 April 2015 (UTC)
    @TomT0m: If stuff is being pushed in from outside, people won't be interested in contributing. (Imho that's demonstrated by the community decline that befell Serbian Wikinews, but trying to make the case for that interpretation of the decline seems rather complicated as I contemplate it.) Of course it's possible to set up a system that offers in an annoying way, or that offers in a helpful way; if it pushes its offer on you, or if there's a lot of crying wolf, then the system making the offer has not been well designed.

    Cooperating consciously with people on other projects, on the other hand, helps raise awareness of other projects, increase cooperation with other projects, and not lose local control — that last being vital both to local morale and to preventing rampant autopropagation of vandalism. (Okay, if Wikidata has a rampant vandalism problem, then other projects may get annoyed — but on the other hand, if the other projects have the option of sending a counter-offer back to Wikidata to treat it as vandalism, and the other projects are also given an easy way to respond by coming to Wikidata themselves ot fix it, that could further foster collaborative attitude and enable the other projects to help combat vandalism at Wikidata.)

    I feel I should point out, when your say "they would have to track vandalism in Wikidata, seems not really like a big deal", that's wildly contrary to my experience. At Wikinews, now that we're getting interwikis piped from Wikidata (which is clearly a disaster impatiently waiting to happen, but I digress), I've been trying to follow related Wikidata changes. It's impossible. I'd estimate less than one in a thousand edits that show up are in any way related to Wikinews, and the proportion continues to drop as Wikidata's structure is filled out. So, if I want to be aware when Wikinews's interwikis are vandalized, I have to sift through thousands upon thousands of irrelevant edits to Wikidata, hoping to not miss that incredibly rare edit that has anything at all to do with Wikinews. --Pi zero (talk) 19:24, 17 April 2015 (UTC)
    That's why I put a word about the quality of the watchlist integration, beeing worked on at Wikidata:Development_plan#Improved_watchlist_integration, they're looking for input somewhere by the way on that point :) TomT0m (talk) 07:58, 18 April 2015 (UTC)
    @TomT0m:, the communities are separate. They are separate by language, by policies, by different sets of institutional memory, different community processes and a different set of hierarchy running things (as in, different administrators, bureaucrats, CheckUsers, Oversighters etc.). The separate communities are able to collaborate together—Commons works at least some of the time. The way to make that run smoothly isn't to just pretend that any friction between the communities can be resolved by pretending it isn't there, but treating the concerns of different communities as worth investigating.
    All I'm saying is that for me, Wikidata transitioned from a fun experiment in data to something that affects ordinary members of the Wikipedia reading public (rather than the editing communities) when I saw that a male actor was described as an actress in the Foundation's new mobile related pages functionality. Once Wikidata transitions to something that affects the ordinary reader of articles, then it starts mattering that we get it right. Hence this thread: rather than being a way to incite inter-community bickering, I was hoping that Wikidata users might consider whether it should be possible for a IP editor with few contributions to cause vandalism that directly affects Wikipedia and whether or not Wikidata's practices to catch such vandalism are up to scratch? And if they aren't, whether we need to consider, say, requiring basic competence on another WMF wiki before granting the ability to affect all wikis in such a drastic way? —Tom Morris (talk) 20:08, 17 April 2015 (UTC)
    @Tom Morris: Hi, I think you just illustrated quite well what I'm afraid of :/ I think that if we keep to see projects as separated and put to much on Wikidata's community, we'll nether have the manpower to get things done. And Wikidatans will have a lot on their plate. I think that to make things keeping advancing we must push the sentiment that every Wikimedian has to be vigilant. TomT0m (talk) 07:58, 18 April 2015 (UTC)
    To illustrate that, I say that your problem will be solved when, instead of watching if a vandal did an edit on an infobox in an article he watches on his local wiki, which can already affect a great deal of person, the infobox will take his datas on Wikidata and he watches the relevant data. It's no different from the situation before, except he will benefit from the cooperation of other people using the same data on a different wiki. So the vandalism, all in all, will be catched earlier. Of course we really need feedback on how to improve that process, some key discussion is taking place on the wikidata development plan page. Wikidata:Development_plan#Improved_watchlist_integration Of course we won't lose anything if wikipedians can take a look on wikidata items as well, that's why imho its important to raise their conscieness of the project. For example people who wrote an article about a famous book could be very happy to see the integration of Wikibook ... and as such fell concerned. TomT0m (talk) 08:25, 18 April 2015 (UTC)
    Yes, if Wikidata is given a high profile at Wikipedia's and made accessible from those Wikipedia's it is no more than logical that vandalism and careless edits will get more and more frequent. Also, the original mission of Wikidata is to hold structured data. What happens now, offering the disambiguating descriptons of Wikidata to Wikipedia readers as short descriptions of Wikipedia topics is a departure from that mission. - Brya (talk) 06:04, 18 April 2015 (UTC)
    It is sad that they abuse Wikidata for this. It is much better for automated descriptions to be used. It benefits from new statements and it works in any language not just English. GerardM (talk) 10:09, 18 April 2015 (UTC)

    how to get id such as Q42 from an english name such as 'Douglas Adams'?

    Basically the opposite of I'd like to go from english names of people and organizations to their unique alpha-numeric identifiers. If possible I'd like to use Java. Is this part of the Wikidata Toolkit API? I've gone through the documentation but I've not seen how it can be done.  – The preceding unsigned comment was added by S.Matthew English (talk • contribs).

    There is some special pages : Special:ItemByTitle, using with a simple webapi, for example Special:ItemByTitle/en/Douglas Adams, who uses the name of an article of the english wikipedia. Otherwise if your query could return several items with a similar identifier, you will have to use the search API, I guess. ItemByTitle should be a method available in your framework. I never played with Wikidata Toolkit yet, so I can't help you more :) TomT0m (talk) 09:48, 17 April 2015 (UTC)

    Hi Tom, that's excellent. do you know how I could pack that into a java program, like the one, or as you said you've not used that API, maybe with another API. you see, I'm trying to write some code wherein that functionality is generalizable, such that i could throw any person name or org name at it and it would tell me the alpha-numeric id, do you know what I mean?  – The preceding unsigned comment was added by S.Matthew English (talk • contribs).

    I suppose it depends what result you'd expect for "Peter Miller". --- Jura 10:21, 17 April 2015 (UTC)

    Just use the WikidataAPI (in case you know the title of the English article): [16]. --Jobu0101 (talk) 10:28, 17 April 2015 (UTC)

    @Jobu0101 I don't understand how to use that though, can you show me an example, from following that link it looks like that request is using the ID and not the english language title.

    @Peter_Miller from the English language title I want to retrieve the alpha-numeric identifier, i.e. Q007 or soemthing  – The preceding unsigned comment was added by S.Matthew English (talk • contribs).

    Try WikidataToolkit or Pywikibot. I guess if you can't figure out how to proceed yourself, you'll just have to get a little more experience and try to express yourself a little more clearly about the context and what you want. Otherwise nobody can help you (depending of your exact purpose here). TomT0m (talk) 11:09, 17 April 2015 (UTC)

    As Jura pointed out, many names occur multiple times. There are lots of items with the English language title "Peter Miller" and there are lots of identifiers for different people with the name "Peter Miller". By the way, that's something I have wondered about for some time: Wikidata follows the philosophy that items are differentiated by their description, the title itself doesn't contain differentiating data. Sometimes this leads to confusion and I have wondered whether it might not be a good idea to use some kind of additional specifications in the titles themselves - but I suppose that would be seen as muddling the data ;-) Gestumblindi (talk) 11:14, 17 April 2015 (UTC)

    Look at the url of [17]. In the end you specify the title: &titles=Douglas%20Adams. And please ping my properly next time using {{ping|jobu0101}}. --Jobu0101 (talk) 12:58, 17 April 2015 (UTC)
    @Jobu0101: I suppose you're referring not to my post, but to the unsigned one (I'll fix that now for him) by S.Matthew English, above TomT0m's reply. @S.Matthew English: it would be helpful if you signed your posts :-) Gestumblindi (talk) 18:20, 17 April 2015 (UTC)
    Yeah, I was referring to the unsigned post. --Jobu0101 (talk) 18:40, 17 April 2015 (UTC)

    Hi guys, I did it like this:

    How do I sign posts? I'm quite new to this community.  – The preceding unsigned comment was added by (talk • contribs).

    You can sign with ~~~~ - that's converted automatically to your signature and timestamp. Or you use the icon for "signature and timestamp" at the top of the editing box, that also inserts your signature. And if you're S.Matthew English, you need to be logged in for that to work, the preceding comment was posted by someone not logged in :-) Gestumblindi (talk) 14:25, 18 April 2015 (UTC)

    Strange property

    P1152 (P1152) has datatype Item but it seems it should store identifiers. This is maybe the reason why it doesn't appear on any items. Matěj Suchánek (talk) 19:36, 17 April 2015 (UTC)

    similar case genomic assembly (P659) --Pasleim (talk) 19:50, 17 April 2015 (UTC)
    P1152 (P1152) is linked to from 63 pages: 9 User, 35 User talk, 16 Wikidata:Database Reports, the remaining 3 are this discussion, the property proposal, and the talk page for statement is subject of (P805). The use of the property appears to have been intended to be cartoon (Q627603) => P1152 (P1152): 20000004, which is indeed, not an item. Popcorndude (talk) 00:39, 18 April 2015 (UTC)
    If you look at wikitext of the proposal, "string" seems to have been requested, but the actual datatype of the item seems to override that.
    Anyways, two options: fix it or delete it as there isn't any interest in the property. --- Jura 06:07, 18 April 2015 (UTC)

    change property but keep value

    Is there any good way to do this? For example, there are 13320 radio station (Q14350) with located in the administrative territorial entity (P131) where they should have headquarters location (P159). Popcorndude (talk) 15:36, 18 April 2015 (UTC)

    Someone at Wikidata:Bot requests can probably help with that. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:46, 18 April 2015 (UTC)

    programatically determine relationship between two ID values

    for instance, if I have Terrell Buckley (Q5571382) and Miami Dolphins (Q223243), which are related on Wikidata by the statement "member of sports team", can I use, for instance WikidataToolkit to query Wikidata, feed in those two entities and get back "member of sports team"? And if yes, how? S.Matthew English (talk) 05:24, 19 April 2015 (UTC)

    Wikidata weekly summary #154

    Using nickname as a qualifier for someone's nickname in an organisation

    Hi all, I want to use nickname (P1449) as a qualifier for the property member of (P463) but now it is just for entities? Can this be expanded to include people? I placed the same question on the talkpage of the property. Thx Jane023 (talk) 12:43, 19 April 2015 (UTC)

    As the property has a statement instance of (P31) Wikidata property for items about people (Q18608871) I assume it can be used for people as well. -- Bene* talk 12:45, 19 April 2015 (UTC)
    Thanks! Jane023 (talk) 13:02, 19 April 2015 (UTC)
    Jane023 Can you give a specific example of where you want to use it? If you are using it as a qualifier to member of (P463) I am guessing it is a nickname for members of that organisation rather than a nickname for the individual the statement is about. If that is the case then I think you shouldn't use nickname (P1449) as a qualifier for the property member of (P463) - it isn't specific to that person. The statement using nickname (P1449) should be on the item about the organisation instead. at least that is my opinion. Filceolaire (talk) 15:57, 19 April 2015 (UTC)
    I put the example on the talk page of nickname (P1449) along with an example and I think we should continue discussion there. Jane023 (talk) 16:18, 19 April 2015 (UTC)

    candidate (P726)

    candidate (P726) has a 96% constraint violation rate. It was created to be a relation of election=>person/party, but literally ~90% of the uses are person=>election. This bothers me. Is it time to propose an inverse property? --Haplology (talk) 07:14, 15 April 2015 (UTC)

    Probably better to redefine P726 to match its usage and create a new property for election=>person/party. Filceolaire (talk) 22:02, 15 April 2015 (UTC)
    Ok, sounds reasonable. Still this situation with P726 is just too much for me to handle. There is no community, just imperious users and chaos. I can't handle this anymore.
    @GerardM: you broke this property. At least take five minutes to at least note your changes, proposed and approved by the law unto himself who is Gerard, on the talk page. It's been more than a year since somebody else asked much more politely than I could have. Five minutes. Honestly. --Haplology (talk) 04:58, 16 April 2015 (UTC)
    I disagree that it should be redefined. The current definition is something that it is useful to have property for, while the way the property is currently used is useless. I suggest having a bot mass-delete all constraint-violating statements. --Yair rand (talk) 05:05, 16 April 2015 (UTC)
    @Haplology: If this is the way you want to approach me, you just failed completely except that I am annoyed so not interested. A minimum of politeness is the least you can do. GerardM (talk) 05:19, 16 April 2015 (UTC)
    In English, the meaning would be clearer if election=>person was called "had candidate(s)" and person=>election was called "candidate for". --Arctic.gnome (talk) 06:52, 16 April 2015 (UTC)
    I've cleaned up all the constraints violations. --Yair rand (talk) 22:16, 20 April 2015 (UTC)

    subclass issues

    Similarly to the #thriller subclass problem above, female (Q6581072) claims to be a subclass of both female organism (Q43445) (!) and human (Q5), which is impossible since Q43445 and Q5 should not intersect. IMO, the real hierarchy here is female organism (Q43445) is a subclass of female (Q6581072) (not the other way around!), but then some "non-fictional female" as subclass of both female (Q6581072) and human (Q5), and then a "fictional female" as the subclass of female (Q6581072) that is the complement of "non-fictional female". Currently, this incorrect categorization allows the claim that all Muses (Q66016) (and any similar concepts) are humans... (one of the reasons for the existence of the issue discussed at #WikidataQuery question: Items which are humans) -- Jokes_Free4Me (talk) 09:44, 15 April 2015 (UTC)

    Item labels tend to be changed in some languages (but not all) and hierarchies transformed. Things that may have made sense at some point get lost. In general, it's probably not worth bothering too much with the subclass stuff. --- Jura 11:32, 15 April 2015 (UTC)
    Rather the opposite: I believe that the subclass hierarchy is essential, because it is language-independent. Build a strong hierarchy and keep an eye on it afterwards, as users have a tendency to "fix it" to fit labels from their own language. It would be useful to either warn or automatically keep track of changes to the values of subclass of (P279) or instance of (P31) on items - users should be made aware of the impact of such changes, that, better yet, should be approved by other users. -- LaddΩ chat ;) 13:09, 15 April 2015 (UTC)
    I'm not sure where you got the idea that "Q43445 and Q5 should not intersect". While female (Q6581072) is used for human women and female organism (Q43445) is only used for non-humans that doesn't mean that Q43445 doesn't include humans; it just means that female (Q6581072) is a more specific subclass of female organism (Q43445) - specifically female creatures who are human. If female (Q6581072) is defined like that then it is a subclass of both female organism (Q43445) (!) and human (Q5). Women are, after all, female animals. Filceolaire (talk) 21:58, 15 April 2015 (UTC)

    I am afraid that we are overusing and misusing subclass of (P279) quite frequently. For example, in female (Q6581072) the description states "human who is female", but there is a claim "instance of sex of humans" and a claim "subclass of gender" and a claim "subclass of human". We don't say "Marie Curie instance of Female", we say "Marie Curie sex or gender Female". female (Q6581072) conflates (at least) two notions: female as a human gender or sex, and female as the set of all humans that are regarded to have this gender. The latter should be represented by a query, and we should clean up the identity of the item. --Denny (talk) 18:50, 19 April 2015 (UTC)

    Also, again: we should not use subclass of (P279) so often. Using this property is almost always an information loss, and we should use it only in the use cases where it actually makes sense. We don't say "Polish subclass of Human and Marie Curie instance of Polish". We have more appropriate properties to express that. --Denny (talk) 18:50, 19 April 2015 (UTC)

    Denny we don't say 'Polish' at all. We say 'country of citizenship:Poland'. If we did have an item for 'Polish' representing the class of people with 'country of citizenship:Poland' then the appropriate statements for that item would be 'instance of:nationality' and 'subclass of:human'. There is not much else you could say about that item.
    There are a number of strange wikipedia articles about groups of things or people. If we have Wikidata items for those Wikipedia articles then about the only useful statement we can make about that item is that it is a subclass of something - even though that item is unlikely to be used as the target for an 'instance of' statement. Are you saying we shouldn't make that 'subclass of' statement and just leave those items with no statements? Filceolaire (talk) 19:40, 19 April 2015 (UTC)
    A more general issue is that we have lots of wikipedia articles called 'list of mayors of Foo' which we sitelink to wikidata items called 'mayor of Foo' with statements 'instance of:political office', 'instance of:wikimedia list article' 'subclass of:mayor' and which we use as the target of 'position held:mayor of Foo' statements on other items. If we did not have these 'list of' articles then we might well have changed these statements to 'position held:mayor' qualifier 'of:Foo' but we do have these items so we use them. At some point in the future we may well be able to replace the wikipedia 'list of mayors of Foo' articles with lists automatically generated by wikidata queries but we are not there yet. Filceolaire (talk) 19:54, 19 April 2015 (UTC)
    @Denny: Filceolaire's right, our item are often blurry. But I think a feature like, when queries are ready, on an item page showing the, let's call them inferred classes, as for example the list of queries for which the item is in the set of result when this is possible, could help a lot to reduce the redundancy while helping us to clear the concepts. After all, classes or sets with intensional or extensional definitions are all in all just ... sets or classes. On the other hand, we may know that an object is an instance of some set while Wikidata do not have enough information yet to infer this. I think it could be very useful to manually asserting this fact in those cases to give a hint that something is probably missing in this case ... TomT0m (talk) 08:17, 20 April 2015 (UTC)

    @Filceolaire:, @TomT0m: That's my point. We don't say Polish at all! That is why I am surprised why we would want to say Marie Curie instance of Female. We have more appropriate properties for that, namely sex or gender (P21). --Denny (talk) 15:23, 20 April 2015 (UTC)

    Nominations are being accepted for 2015 Wikimedia Foundation elections

    This is a message from the 2015 Wikimedia Foundation Elections Committee. Translations are available.


    I am pleased to announce that nominations are now being accepted for the 2015 Wikimedia Foundation Elections. This year the Board and the FDC Staff are looking for a diverse set of candidates from regions and projects that are traditionally under-represented on the board and in the movement as well as candidates with experience in technology, product or finance. To this end they have published letters describing what they think is needed and, recognizing that those who know the community the best are the community themselves, the election committee is accepting nominations for community members you think should run and will reach out to those nominated to provide them with information about the job and the election process.

    This year, elections are being held for the following roles:

    Board of Trustees
    The Board of Trustees is the decision-making body that is ultimately responsible for the long term sustainability of the Foundation, so we value wide input into its selection. There are three positions being filled. More information about this role can be found at the board elections page.

    Funds Dissemination Committee (FDC)
    The Funds Dissemination Committee (FDC) makes recommendations about how to allocate Wikimedia movement funds to eligible entities. There are five positions being filled. More information about this role can be found at the FDC elections page.

    Funds Dissemination Committee (FDC) Ombud
    The FDC Ombud receives complaints and feedback about the FDC process, investigates complaints at the request of the Board of Trustees, and summarizes the investigations and feedback for the Board of Trustees on an annual basis. One position is being filled. More information about this role can be found at the FDC Ombudsperson elections page.

    The candidacy submission phase lasts from 00:00 UTC April 20 to 23:59 UTC May 5 for the Board and from 00:00 UTCApril 20 to 23:59 UTC April 30 for the FDC and FDC Ombudsperson. This year, we are accepting both self-nominations and nominations of others. More information on this election and the nomination process can be found on the 2015 Wikimedia elections page on Meta-Wiki.

    Please feel free to post a note about the election on your project's village pump. Any questions related to the election can be posted on the talk page on Meta, or sent to the election committee's mailing list, board-elections -at-

    On behalf of the Elections Committee,
    -Gregory Varnum (User:Varnent)
    Coordinator, 2015 Wikimedia Foundation Elections Committee

    Posted by the MediaWiki message delivery on behalf of the 2015 Wikimedia Foundation Elections Committee, 05:03, 21 April 2015 (UTC) • TranslateGet help

    WikidataQuery question: Items which are humans

    Using how can I say that I'm only looking for items which are humans? CLAIM[31:5] does not work since it covers no subclasses of humans. --Jobu0101 (talk) 20:28, 11 April 2015 (UTC)

    claim[31:5] shouldn't have any subclasses. --- Jura 20:37, 11 April 2015 (UTC)
    I mean in Wikidata there are many humans which don't have claim[31:5] because they have claim[31:?] where Q? is a subclass of Q5. Since thouse items are also humans I want to cover them with my query but I don't get them covered with claim[31:5] because is not concerned about subclasses. --Jobu0101 (talk) 20:42, 11 April 2015 (UTC)
    Do you have a sample item? --- Jura 20:51, 11 April 2015 (UTC)
    You're looking for claim[31:(tree[5][][279])]. However, the difference to claim[31:5] is only marginal: 2,765,488 vs 2,767,718 items. --Pasleim (talk) 20:54, 11 April 2015 (UTC)
    The difference should be zero. --- Jura 21:03, 11 April 2015 (UTC)
    Looking at the items on this, there might actually be no humans among those items. --- Jura 21:07, 11 April 2015 (UTC)

    Thanks. What about them:[31:14943515]%20and%20claim[31:5]? I think this list should be empty because Q14943515 is a subclass of Q5. --Jobu0101 (talk) 21:38, 11 April 2015 (UTC)

    No, because Q5 shouldn't have any subclasses. Besides, we don't want religious wars either. --- Jura 22:00, 11 April 2015 (UTC)
    So you mean we should remove all subclasses of Q5? --Jobu0101 (talk) 22:17, 11 April 2015 (UTC)
    I think for the purpose of your query, you can assume that all humans have P31:Q5. At least, that's what many other tools do. --- Jura 10:46, 12 April 2015 (UTC)

    By the way, this is the marginal difference of 764 items. --Jobu0101 (talk) 21:44, 11 April 2015 (UTC)

    Check it out with my link to Autolist above. --- Jura 22:00, 11 April 2015 (UTC)
    322 of them are Q14943515's, many of the others listed appear to be positions or figures from Greek Mythology. Popcorndude (talk) 23:18, 11 April 2015 (UTC)
    @Jobu0101: I wrote a template to write these queries parts : CLAIM[31:(TREE[5][][279279])] ({{WDQ/instances|Q5}} gives this. It's usable with {{WDQ}} to generate a link to the WDQ corresponding query : {{WDQ|{{WDQ/instances|Q5}}|human instances}} gives human instances. TomT0m (talk) 10:15, 12 April 2015 (UTC)

    @Jura1: I don't see why there should not have any subclass. There is no reason and we are technically ready to allow this. It's just a matter of queries and WDQ and the future query engine of Wikidata will perfectly be able to handle this. This is a limittion with no real reasons. And absolutely not a community consensus. TomT0m (talk) 10:33, 12 April 2015 (UTC)

    If you say so. --- Jura 10:42, 12 April 2015 (UTC)
    @Jura1: This ... does not seem like an answer, what do you mean exactly ? TomT0m (talk) 12:02, 12 April 2015 (UTC)
    A human is a human is a human. When it is not possible to say that someone is a human he is not. There is absolutely no consensus to think otherwise. Thanks GerardM (talk) 12:28, 12 April 2015 (UTC)
    @GerardM: There is consensus on Help:Basic membership properties, that's all I know. And instances of human is totally possible, I know that too. TomT0m (talk) 12:37, 12 April 2015 (UTC)
    You confuse the rule with possibly the exception. Either a human is a human or he is not. Thanks, GerardM (talk) 12:41, 12 April 2015 (UTC)
    @GerardM: I'm sorry, what ? TomT0m (talk) 12:45, 12 April 2015 (UTC)
    The rule is that on {{Item documentation}} there is a link to generate all instances of a class, and that if you look on Talk:Q5 there is a call, and that's true for any other class, to get all instances of that class. The exception is that to get all elements in a class you just query the P31/Qitem pair. TomT0m (talk) 12:58, 12 April 2015 (UTC)
    GerardM "A human is a human unless he (or she) is not" or unless we are not sure. Q16934977 for instance is an entire class of people who might be historical or might be fictional or might be a bit of both. Filceolaire (talk) 18:57, 12 April 2015 (UTC)
    How is a legendary person human. As far as I am concerned I do not care ... they certainly do not have a story that is verifiable in any way. So no. GerardM (talk) 20:23, 12 April 2015 (UTC)
    GerardM, the point of that class is that the story is verifiable in some ways, but not in others, and so scholars sincerely disagree about whether the person actually existed or was fictional. Human history does not divide neatly into 100% agreement that a person named in an old source actually existed, and 100% agreement that the person never existed. How would you count the ones for which scholars disagree? WhatamIdoing (talk) 07:33, 21 April 2015 (UTC)
    Wilhelm Tell and James Bond are as far as I am concerned not human, they are fictional. They may share attributes of a human but that is it. I do not really like the argument "scholars do this or that"; either you do adopt an argument or I can safely disregard it. It is your argument or it is not yours to champion. GerardM (talk) 08:35, 21 April 2015 (UTC)
    @WhatamIdoing: In that case, we can create two items, linked with fictional analog of (P1074)  . One for the obviously fictional character that appears in fil we could hardly call biopics, and one for the disputed human, maybe with the
    < the disputed human > instance of (P31)   < human >
    qualified with a statement disputed by (P1310)   qualifier. TomT0m (talk) 09:01, 21 April 2015 (UTC)
    This is not an argument your giving.. So not really. An instance of should not be disputed. Thanks, GerardM (talk) 09:42, 21 April 2015 (UTC)
    @GerardM: I'm not trying to argument anything, I'm trying to help him solve his problem. As far as I know, the existence of some object is desputed in the real world, and we still have a NPOV policy. This is exactly what qualifiers are for, and even if the semantics is usually not standard in typing properties, I don't see in the name of what An instance of should not be disputed. Except your opinion. TomT0m (talk) 09:52, 21 April 2015 (UTC)
    It is a matter of making an qualification.. a human is "the animal" and a person (a higher level thingie) is whatever.. Now that makes sense to me.. suggesting that something mythical is an animal does not. GerardM (talk) 10:36, 21 April 2015 (UTC)
    @GerardM: The point is that sometimes a person inspires legends. Then sometimes historian wonder if there is an historical figure who existed and inspired the fiction. They are separate entities, and the historical person may indeed have existed. If there is real arguments for this, I'm not opposed to give an item to the hypotetical historical figure, and use qualifier in the instance of (P31) statement as they are supposed to be used : to modify the sense of the claim. Significantly probably. <joke>And THIS IS NOT GIBBERISH.</joke> TomT0m (talk) 10:59, 21 April 2015 (UTC)
    GerardM, I'm not thinking of William Tell. AFAICT, there are zero scholars who believe that William Tell existed. I'm thinking of cases more like Mother Shipton, whose purported posthumous publications are widely regarded as a hoax, but whose name may have been taken (was probably, but not definitely) from a real 16th century woman, with a birth and death date, who lived in a known place and did undertake the profession ascribed to her. What do you do with her? Is she a "human" or a "fictional character"? I like the idea of putting her in a subclass that identifies the uncertainty around her existence. WhatamIdoing (talk) 16:00, 21 April 2015 (UTC)
    • I would agree with the comments above that we should not use subclasses of human; P31:Q5 works well and avoids the complexity of having to manage a hierarchy of subclasses to ensure we avoid any problems with animals, fictional people, etc being included. The approach we've settled on is simple, reliable, and easy to understand - let's not overcomplicate it if we don't have an absolutely pressing need to. Andrew Gray (talk) 20:04, 15 April 2015 (UTC)

    "Lua error: not enough memory"

    When I try to view Wikidata:Property proposal/all (all property proposals), the items after 16.12 display "Lua error: not enough memory" all over the place... Probably there's a better place to report this, but I'm still not really familiar with the bug tracking platforms, so I hope someone will forward this to the right place :-) Gestumblindi (talk) 22:44, 20 April 2015 (UTC)

    That's usual on such large pages, no quick fix for that. Sjoerd de Bruin (talk) 06:57, 21 April 2015 (UTC)

    An award celebration is NOT an award

    Hoi, there are many items like "1992 Laurence Olivier Awards". They are in and off themselves not awards, they are celebrations. I want to aggressively change these.. Thanks, GerardM (talk) 17:22, 14 April 2015 (UTC)

    @GerardM: Q16154017 is still instance of award. Did you start already with your changes? --Jobu0101 (talk) 13:41, 16 April 2015 (UTC)
    Look also here. Maybe that helps you to find them (has to be done for each year). --Jobu0101 (talk) 13:45, 16 April 2015 (UTC)
    I started making changes one year at a time starting with 2015. GerardM (talk) 15:47, 16 April 2015 (UTC)
    Yes, should be corrected. For some of them it might be useful to create a seperate item as a container for more properties that fit a whole series (e.g. Golden Globe Award ceremony (Q19311591) ) Michiel1972 (talk) 23:00, 21 April 2015 (UTC)


    Hoi, People like Razan Zaitouneh disappeared. They are probably dead. I have no idea how to do justice to such a situation. When you consider Iraq and Syria, there are probably many people who disappeared without a trace.. What to do ? Thanks, GerardM (talk) 14:20, 20 April 2015 (UTC)

    Maybe significant event (P793) unexplained disappearance (Q7884274)? Sjoerd de Bruin (talk) 14:49, 20 April 2015 (UTC)
      Support, and qualified with point in time (P585) ---- LaddΩ chat ;) 21:52, 20 April 2015 (UTC)
    There is date of disappearance (P746) for this. --- Jura 01:18, 21 April 2015 (UTC)
    Thanks :) GerardM (talk) 06:27, 21 April 2015 (UTC)
    Ah, it's such a shame that we've got a lot of properties, but some of them are not being used a lot. Sjoerd de Bruin (talk) 06:55, 21 April 2015 (UTC)
    It's hard to search/find properties, mostly I "discover" suited properties on related items or by brute force. Better search suggestions welcome. –Be..anyone (talk) 22:19, 21 April 2015 (UTC)
    Indeed. I would actually go further, and say property/class relationships should not only be accessible at the property end, in the talk page with constraint and documentation templates, or opaquely through a search or suggestion algorithm. You should be able to go to missing person (Q388505) and be able to find date of disappearance (P746) through there. Perhaps with the work being done to express property constraints as statements, this could also be addressed. Dancter (talk) 23:30, 21 April 2015 (UTC)

    Original title

    Moved to Property talk:P1476. --Jobu0101 (talk) 15:49, 22 April 2015 (UTC)

    Some notes

    Hello Wikidata editors,

    these days, I look a bit more into Wikidata since I always come here for the former interwiki links. There are some issues I noticed. Probably, most of them are already discussed and I dunno where.

    1. For filmographies: If you add an actor to a film object, the movie is not immediately listet at the actors page. So, the movie needs to be added to the filmography of the actor, too, separately. However, it would be better if you can add an actor to the movie and then it will also appear on the actors page so that it is connected somehow. Same when you add a film to an actor object.
    2. I think blood type would be suitable property for people. In Western countries, it might be not relevant and hard to figure out, however, in Asia, nearly all celebrities feature the blood type on their official websites.
    3. Maybe, birthday by the lunar calender could be added. I just know, that for Korean actors, sometimes the birth dates differ from different websites. Often it is because on site show the lunar birthday, the other site the usual gregorian calender birthday.
    4. On Wikipedia, there are some romanization templates (see Song Kang-ho for Korean). It would be good, if there was also a possibility to make a property for romanizations like McCune-Reischaer, Pinyin, Romaji etc.
    5. For German dates, they usually follow the "rule" that there is a dot after the day, e.g. 1. Januar 2015. However, here it is shown without dot: 1 Januar 2015.
    6. I read that Wikidata will import Freebase. But isn't there a problem that Freebase has more properties for persons?
    7. Perhaps, a property like "original title" would be good for films.
    8. If you look at Q102098, Moon Geun-young, when your layout language is English, it shows for the property "surname" Wen. This can be quite confusing for some people, since there is no "Wen" in Moon Geun-young. Of course, Korean family names are based on Chinese characters, which is 文 here and spelled 문 in Korean (Mun), while in Pinyin (Chinese to latin), it is Wen. Maybe, a solution can be found to make it less confusing.

    That's all. --Christian140 (talk) 12:04, 19 April 2015 (UTC)

    @Christian140: Hello! I will try to answer you:
    1. We will propably have to wait until querries are enabled. (There is no "starred in a film" property.)
    2. You can propose the property at WD:Property proposal (person).
    3. I believe the developers are working on this. Everything related to time datatype is tracked at phab:T87764.
    4. As 2. (pinyin transliteration (P1721) exists.)
    5. As 3.
    6. I'm not sure here, something might have been discussed at Wikidata talk:WikiProject Freebase#Property mappings.
    7. title (P1476)
    Any more questions or additional explanation? Matěj Suchánek (talk) 15:24, 19 April 2015 (UTC)
    Thank you for your comments. I will look into it. I guess I will make a few property proposals then. And ah, yes, I think the cast of a film should also have something like a ranking, for example, lead role, supporting role... I will also check out the "title" property. Thx. --Christian140 (talk) 17:40, 19 April 2015 (UTC)
    For the blood type we already had a property proposal: Wikidata:Property proposal/Archive/22#Blood Type --Pasleim (talk) 17:43, 19 April 2015 (UTC)
    Nooooo... Ahhh... Thank you :( I just made the proposal unfortunately. However, I think blood type will be proposed as long till it gets approval. When Wikidata becomes more relevant for Wikipedia and the editors of the Korean, Japanese or Chinese Wikipedia spend more time here, some will propose it again, I guess. --Christian140 (talk) 17:57, 19 April 2015 (UTC)

    @Christian140: By the way, there are some issues with the title (P1476) property. See Wikidata:Project_chat#Original_title. Maybe you can help us there to find a solution. --Jobu0101 (talk) 00:16, 20 April 2015 (UTC)

    @Filceolaire, Jobu0101: There is no issues if the general rule of "one identified object = one item" is followed. The main problem is that people are lazy and don't provide the necessary information to create and identify each object.
    • Use title (P1476) only for the title of the specific object described by the item.
    • Several uses of title (P1476) means that the object was edited/distributed from the beginning in different versions.
    • Each time a translation is done and a specific edition is released a new item has to be created.
    This rule solves most of the cases. I still wait for the description of the special cases which can't follow the above situation. And better use the talk page of the property to discuss its use: the project chat doesn't allow a good storage and easy access to future questions and solutions. Snipre (talk) 12:46, 20 April 2015 (UTC)
    I'd guess that blood type would violate privacy laws in a lot of places. StevenJ81 (talk) 14:38, 20 April 2015 (UTC)
    @Snipre: I think it is more complicated. Read for example Christian140's comment there. --Jobu0101 (talk) 16:04, 22 April 2015 (UTC)

    Technical specifications for movies - How to manage them?

    Someone can help here?  – The preceding unsigned comment was added by ValterVB (talk • contribs). <span class="mw-code" style="text-align:right; display: inline-block; padding: 0.1em;"TomT0m (talk) 17:21, 21 April 2015 (UTC)

    Merge problems

    Can someone merge Rozin (Q19299407) with German Rozin (Q7375377) ? 18:22, 22 April 2015 (UTC)

      Merged--Pasleim (talk) 18:46, 22 April 2015 (UTC)

    Page status indicator

    Full revert of a recent change for the page status indicators, s.v.p., the remaining width is far too small for ICBM coordinates working everywhere incl. here before the change. –Be..anyone (talk) 00:14, 22 April 2015 (UTC)

    We don't have geohack here, I thought... Sjoerd de Bruin (talk) 06:53, 22 April 2015 (UTC)
    Can you please link to some item or page where they are a problem? Aude (talk) 08:19, 22 April 2015 (UTC)
    Hi, it was in the upper right of my user page here. Oddly, it is okay now. Could something I don't see with an AdBlocker push the page status indicator to the right? I had a line beak in what you can see now. –Be..anyone (talk) 06:38, 23 April 2015 (UTC)

    Wikidata:Requests for comment/Opting out of Global sysops 2

    A RFC has been started in regards to allowing global sysops to continue to perform admin actions on Wikidata. --MediaWiki message delivery (talk) 01:14, 23 April 2015 (UTC)

    French-Engish translation needed

    Please would someone kindly translate the short French description at Wikidata:Property proposal/References#catholic-hierarchy diocese id into English? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:27, 19 April 2015 (UTC)

    Actually it isnot a description, but a motiv. Visite fortuitement prolongée (talk) 21:40, 19 April 2015 (UTC)
    Until your recent edit it was in the |description= parameter. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:50, 20 April 2015 (UTC)
    Indeed. Visite fortuitement prolongée (talk) 19:35, 20 April 2015 (UTC)

    Can anyone help with translation, please? Google Translate is no use, in this case. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:28, 23 April 2015 (UTC)

    Mess with japanese foster parents

    Yesterday someone merged Q11666901 (adoptive parents) with parent-in-law (Q15977638) probably fooled by the "in-law" in the english description. Now User:KrBot has "fixed" this redirect when used as qualifier for type of kinship (P1039) (obviously for japanese people only, the actual number of items is quite manageable). Unfortunately I cannot understand what relation relly was intendend in the japanese items (foster parent/father/mother? adoptive parent/father/mother?) and the perhaps the items in question would have to be fixed anyway to match the correct gender of the target person. -- Gymel (talk) 08:23, 20 April 2015 (UTC)

    I'm not sure about this particular case, but I left a note on the bot's talk page to stop "solving" redirects as that makes it impossible to undo a wrong merge+redirect. -- Bene* talk 13:08, 20 April 2015 (UTC)
    For the record [18] currently yields 31 items which should exactly be those in question. 27 of them are male, 4 femaleAll of the target items are of male sex thus I suspect the users redirect was not only to the wrong term (but we don't have items yet for adoptive parents/father/mother?) but also the original usage has been categorically wrong (should have been father/mother instead of parents) or the japanese term is indifferent of gender). Can someone forward this request to some japanese-language corner in order to find out what the original Q11666901 really was about (almost all items seem to concern members of the nobility thus a concept may be involved which cannot be mapped at all to adoption or fostering...) -- Gymel (talk) 14:59, 20 April 2015 (UTC)
    The Japanese concept of ja:義親 is described in the Chinese Wikipedia article (zh:乾親) thus:-
    Basically for the Japanese 義親 are all kin not directly related by blood, this includes parents in law and brothers and sisters in law, adopted/foster children, adoptive/foster parents and siblings, and oath/sworn brothers.--KTo288 (talk) 21:37, 22 April 2015 (UTC)
    So if the person doing the merge was Japanese, according to his/her worldview, adoptive parents are exactly the same as parents in law and the merge is correct.--KTo288 (talk) 21:45, 22 April 2015 (UTC)
    @KTo288: That means would hold (or is 義親 always a single person)? -- Gymel (talk) 23:12, 22 April 2015 (UTC)
    @Gymel: as I understand it, yes would be true in general all, but of course in specific cases the ties would be between individuals.--KTo288 (talk) 04:36, 23 April 2015 (UTC)

    According to the discussion I reverted the merge of nominal kinship (Q11666901) with parent-in-law (Q15977638) and the subsequent changes in qualifiers which now again can be found by WDQ with their old qualifier item. -- Gymel (talk) 18:34, 23 April 2015 (UTC)

    gender question (message moved)

    Please unify the subclassing of "male" and "female"

    (moved from wd:dev)

    "Female" (as a gender) is a subclass of "human". "Male" is not. The correctness of this subclassing is debated, but what it is urgent is that "male" and "female" get the same treatment. Presently, if you look for pages with gender human (where the gender is closed transitively by superclass) you get only women.  – The preceding unsigned comment was added by Vigna (talk • contribs).

    I removed "subclass: human" from female. It makes more sense to list male and female as instances of "gender", and for "gender" to hold the relationship with human, IMO. In theory almost any property that can be applied to humans could be listed as a subclass of human: nationalities, occupations, races, genders, etc. I think that would be overloading the intended use of "subclass" though. Kaldari (talk) 01:02, 22 April 2015 (UTC)
    Vigna, Kaldari I have added 'subclass of:human' to both 'male' and 'female' as these two items are both restricted to be used for describing humans and are not used for male and female non-humans (use 'male animal' and 'female animal' instead for non-humans. See this discussion). 'Gender' is not a subclass of human. A 'gender' is not a human. OK? Filceolaire (talk) 08:02, 22 April 2015 (UTC)
    @Filceolaire: Not the good solution: we have to have a similar solution for male human and male animal. Then we have a problem with fictional character: Sarek (Q2712069), Vulcan father of Spock in Star Trek universe, is defined as male but this is not a human.
    We have to have a general way to classify these items. Snipre (talk) 16:31, 22 April 2015 (UTC)
    Property Sarek (Q2712069) Bugs Bunny (Q183102) James T. Kirk (Q16311) Barack Obama (Q76)
    sex or gender (P21) male (Q6581097) male (Q6581097) male (Q6581097) male (Q6581097)
    Race/Species Vulcan (Q6497384) Rabbit? human (Q5) human (Q5)
    Real/fictional fictional fictional fictional real
    ethnic group (P172) - - - African Americans (Q49085),Irish Americans (Q1075293)
    Snipre (talk) 16:43, 22 April 2015 (UTC)

    Here's the way that I see the question. We have these animals in the zoo:

    Animals in the zoo
    Animal Male Female
    Elephants Male elephants Female elephants
    Tigers Male tigers Female tigers
    Bears Male bears Female bears

    We want to be able to subdivide the animals, because sex matters to the zoo's breeding program. We also want to be able to make bigger groups (all the males, regardless of species; all the tigers, regardless of sex).

    How would we represent this? Do we need to have add "elephant", "male" and "male elephant" to any given instance of a male elephant? WhatamIdoing (talk) 16:49, 22 April 2015 (UTC)

    @Filceolaire: I would understand a
    < man > subclass of (P279)   < human >
    < woman > subclass of (P279)   < human >
    with the man and woman defined as human with a female sex or genre, for example, either by a constraint or a query, I don't care. On the spirit of the Wikidata:Requests_for_comment/Adopt_Help:Classification_as_an_official_help_page, I'm not sure how to define the range of the sex or genre property thought :) it seems that the facts here are the person has a XY sexual chromosom pair and feel like a man. Then the genre would be a feeling also ... Maybe a property feels like could be an alias for genre. And the domain would be man or woman, and
    < man > instance of (P31)   < sexual identity >
    < woman > instance of (P31)   < sexual identity >
    . This is the best I can do /o\. TomT0m (talk) 16:52, 22 April 2015 (UTC)
    Addition: with of course the range of the feel like property would be <sexual identity>, and its domain human. TomT0m (talk)
    TomT0m There was a long discussion (see Property_talk:P21) about whether we need separate properties for sex and for gender and what exactly is meant by saying that someone has male or female sex. The conclusion was that for the vast majority of the people on wikidata we have no idea what their biological sex is (in any of the 5 different things that make up biological sex - see en:intersex. We don't even know their gender identity. All we know is their gender expression - how they present themselves. It was agreed that is sufficient for property sex or gender (P21) since if we require anything more specific then we would have to leave this statement blank for most humans on wikidata. In the few cases where we do know more then this additional knowledge can be expressed in qualifiers.
    Apart from using wikidata as general playground and blurring the fragile distinctions between sex and gender, human and homo sapiens, organism and person, being and character and whatsnot there is also another serious problem in connection with statements : Many wikipedia articles linked to female organism (Q43445) deal with female organisms, i.e. also include female plants. Probably also with the male counterpart... -- Gymel (talk) 16:50, 22 April 2015 (UTC)
    Gymel I agree that sex or gender (P21) blurs the distinction between sex and gender. It is designed that way to reflect our lack of knowledge. For most people on wikidata we know that their gender presentation (how they dress etc.) probably reflects their biological sex. We also know that for some people their gender presentation does not match their biological sex but we have no idea which people this is true for. sex or gender (P21) reflects that ambiguity.
    I agree that female organism (Q43445) should probably be renamed "female organism" to reflect how it is used. Filceolaire (talk) 22:38, 22 April 2015 (UTC)
    Exactly. Having a dedicated property sex or gender (P21) in addition to instance of (P31) is a way to deal with ambiguities. But IMHO not connecting everything subclass of (P279)-wise ad ultimo is the more important way: "Male" is an appropriate attribute for Spock's father even if we do not know anything about Biology and Society of Vulcan to make more specific claims like "Vulcan Male" one may be tempted to use with instance of (P31). And to keep it this ambiguous way we must be very conservative and especially refrain making male (Q6581097) a subclass of anything crossing the bridge from the appropriately abstract person (Q215627) to human (Q5) or real terrestrial life forms. I know that subclasses in wikipedia are not prescriptive but making any male a subclass of human (Q5) may not make them humans but at least spills terrestrical biology onto poor Sarek and Bugs Bunny. Therefore I regard male (Q6581097) having subclass of (P279)s at all as a high risk for future problems. -- Gymel (talk) 23:08, 22 April 2015 (UTC)
    In other words, male (Q6581097) is not a class of organism, it's a property we attach to organisms. Woman, on the other hand, is a class of human who have a property of calling their sex "female". Of course male and female properties make sense only for organisms who reproducts sexually. TomT0m (talk) 07:16, 23 April 2015 (UTC)
    male and female can be attributed by extension to any individual (probably because our "wiring" has an antropomorphic or antorpocentric bias). Thus even the smalles hint (me creating a story about the adventures of two peculiarly shaped stones named "Peter" and "Mary", becoming internet celebrities and rising to wikidata notability within weeks...) may result in an (certainly justified!) assignment of sex or gender (P21). It very much depends on your definition of "organism" whether I can agree with your remark. I would say anything classified as individual (Q795052) is a possible candidate for carrying sex or gender (P21) (not always male or female of course), but I'm not sure wether every plant in my garden involved in sexual reproduction (Q182353) qualifies as individual (Q795052) (but it is an instance of organism (Q7239) or - since I just invented my garden - a fictional analogue of that). -- Gymel (talk) 08:07, 23 April 2015 (UTC)
    @Gymel: Yes, I'm the one who proposed the fictional analog of (P1074)   property, and the whole point was to separate the fictional and real classes :) Of course, an organism here is a real one, it is pretty much well defined I guess. An organism with sexual reproduction is also, maybe with weird exceptions as always in biology (the rule in biology is that there is exceptions to the rule). If we apply the sex or gender (P21)   property to a fictional character, of course this is a whole different thing. Way less important though :) There is a lot more possibilities and less constraints in fiction :) What we need in Wikidata is to keep the hierarchies separate to be able to apply constraints only to real stuffs, and live more freedom in fictional entities. Of course man and woman are subclasses of real world humans item. This does the trick. TomT0m (talk) 08:17, 23 April 2015 (UTC)
    @TomT0m: sex or gender (P21)   is expanded to genre (P136) but I take it you did mean sex or gender (P21) (same word in french...)? My impression of the biological nomenclature in wikidata is that we have the (a) complete taxonomic tree here but it is quite insulated, e.g. not using subclass of (P279) but the dedicated parent taxon (P171) to model the hierarchical relation. And - most important - these items do not break out of the system they belong to by using excessive instance of (P31) or subclass of (P279). A comparable practice should be implemented for physical objects with some reality, this will be of course much harder to achieve and I'm not sure wether the border should be drawn between "real" and "fictional" or rather between "intrinsic" and "extrinsic" attributes or whatever distinction may be appropriate to put e.g. the biological and antropological aspects of human beings on one side and the social ones of "personality" on the other. Properties pertaining (mostly) to one or the other side then should have a range of items which in turn usually do not cross the border again with their P31/P279 chain. As a consequence one would have either a companion attribute for sex or gender (P21) reserved for objects where you cannot derive sex or gender from inspection or interview, or fictional analougues of male and female for use in sex or gender (P21) or - this I would prefer - simply accept that sex or gender (P21) applies to both sides of the divide and therefore the range elements male (Q6581097) and female (Q6581072) should only be linked via P31/P279 to very abstract concepts, avoiding to create a closer connection with items having a definite location on one side of the divide. For me is probably all what can be said about female in the context of wikidata, and all other current values of P31 and P279 for that item I deem questionable. -- Gymel (talk) 09:16, 23 April 2015 (UTC)
    @Gymel: My vision of instance of (P31) and subclass of (P279) is exposed in Help:Classification. In this vision, there is no problem into classifying species and other clades with subclass of (P279) and instance of (P31), as a clade describes a set of organism, and a clade who is a child taxon of this clade is a subset of this set of organisms. clades like Kingdom can be marked
    < animal > instance of (P31)   < kingdom >
    , kingdom beeing a metaclass. Other examples and pitfalls of not using metaclasses are describes in the RfC Wikidata:Requests_for_comment/Adopt_Help:Classification_as_an_official_help_page.
    Of course the fictional analog of solution and disjoint tree allows to use the sex or gender (P21)   on fictional item, that's the whole point. We know it has a real world semantic only when applied to an object who is an instance of a subclass of the real object class.
    For me is probably all what can be said about female in the context of wikidata Always in the spirit of Help:Classification, I try to find tokens in the real world to make things precise, at least as an exercise :) And I think that someone saying I feel like a woman, meaning it, is a manifestation of a real feeling, and that (of course Wikidata will not be complete enough to be able to query for this) this validates the model, we could even say somethings like feeling like a woman is a subclass of feeling and an instance of Q48264, which validates the model and make me concur with you. And tools like Help:Classification helps us think (if you agree, please say that on the RfC :) TomT0m (talk) 09:49, 23 April 2015 (UTC)
    @TomT0m: Does your classification give some rules about the use of classes like fictional human (Q15632617) instead of using 2 "instance of" with human (Q5) and fictional character (Q95074) ? The examples above indicate that if we want to use fictional human (Q15632617) we have to create similar classes like "fictional Vulcan" and "fictional rabbit" to keep a consistency. Snipre (talk) 11:20, 23 April 2015 (UTC)
    @Snipre: Fiction Vulcan is useless, Vulcan is fictionnal by essence :) fictional human (Q15632617) should not be linked with human (Q5)     with instance of (P31), but with fictional analog of (P1074)  . Otherwise if we query something like this generic query to retrieve all instances of a class, James Bond would show up. For me, James Bond is a fictional character, not a human, this is very different. But yes, Fiction Rabbit make sense, for Bugs Bunny for example. This does not mean that every class will have to have its counterpart, just when it is relevant. See Wikidata:WikiProject Fictional universes. TomT0m (talk) 11:47, 23 April 2015 (UTC)
    Addition For the Vulcan,
    < Vucan > subclass of (P279)   < fiction person >
    make sense. It's then in the subclass tree of fictional entities. TomT0m (talk) 12:05, 23 April 2015 (UTC)
    "fictional human" is a subclass of "fictional character" is a subclass of "fictional entity". Mr Spock should in my opinion be "instance of:fictional character". Just 'cause we have "fictional human" that doesn't mean we have to create "fictional vulcan" as well.
    Similarly "biblical character" is a subclass of "religious character" is a subclass of "legendary character" is a subclass of "person" (since these are all characters who at least claim to be based on historical persons so they may be fictional or part fictional or not fictional at all). Filceolaire (talk) 06:03, 11 May 2015 (UTC)

    Marking adiministrative categories

    I think that all Hidden_categories should be made instance of (P31) Wikimedia administration category (Q15647814)

    --Vladimir Alexiev (talk) 12:57, 22 April 2015 (UTC)

    Novalue in qualifiers/references

    I'd like to discuss the usage of novalue in qualifiers or references. I've made a post to the list on it here: so I won't repeat it all here, the TLDR version is that I don't think it makes sense to use novalue in either. I would like to hear what people think about it. --Smalyshev (WMF) (talk) 18:37, 24 April 2015 (UTC)

    Badge needed for wikisource items... (Bis)


    Last month, I posted this subject on the project Chat.

    these are needed to easily see texts that have been completed from those that are still not "validated", since a bot imported thousands of links without bothering to ask about it on wikisources. The question of "how they would look locally is irrelevant, since these status are automatically managed by the system on wikisources. The important thing is to see them here on wikidata.

    Wikisource participants seemed ok with the idea, could the badges please be created on Wikidata ? --Hsarrazin (talk) 20:34, 22 April 2015 (UTC)

    I have created a task for this on phabricator. phab:T97014 We can probably use and for the badge icons, if that's ok. Aude (talk) 10:57, 23 April 2015 (UTC)
    Or is something else desired as the default on Wikidata? On each Wikisource, these can be customized (different icons used), but we need some default list of badges ("validated" and "corrected"?) and icons for those. Aude (talk) 11:00, 23 April 2015 (UTC)
    those will be ok for me, but if other people prefer other wikisource icons, it will be ok, as long as they are symbolic of the "completeness" of the text. Thanks --Hsarrazin (talk) 19:01, 24 April 2015 (UTC)

    Deletion request of a bunch of pages

    Hi all, today I tried to request to delete around 2000 items on Wikidata:Requests for deletions, But it doesn't show all of them. I tried to make a list (each with 90 items) but in this list it shows only 9 sections out of 21. Can you please correct it? All those items are belonging to pages which were deleted on hiwiki.☆★Sanjeev Kumar (talk) 22:10, 24 April 2015 (UTC)

    Merge problems

    Can someone merge de:Kategorie:Universität in Ankara (Q9150290) with english en:Category:Universities and colleges in Ankara (Q10176265) 00:34, 25 April 2015 (UTC)

      Done Pamputt (talk) 05:50, 25 April 2015 (UTC)

    Combining Q112865 & Q15149663 (Mexican States)

    Can these items (Q112865 and Q15149663) combine in some way? Maybe they are totally different items but there is something wrong. Both are marked in Finnish with the same item "Meksikon osavaltiot" (in English States of Mexico). Q112865: en:Administrative_divisions_of_Mexico is pretty much the same as Q15149663: fi:Meksikon_osavaltiot. I am sorry fragmentation of discussion in many places: Talk:Q112865 and Requests for comment/Adopt Help:Classification as an official help page (Deleted comment). --Raid5 (talk) 11:48, 20 April 2015 (UTC) edit. --Raid5 (talk) 14:35, 20 April 2015 (UTC)

    @Raid5: Depending on how much an individual article at administrative territorial entity of Mexico (Q112865) covers not only estados but also municipios as second-order subdivisions it could or could not be moved over to state of Mexico (Q15149663). Looking at the associated "main categories" Category:States of Mexico (Q7146004) and Category:Subdivisions of Mexico (Q6418549) shows a clear overlap, i.e. many wikipedia instances have both categories. I think in the sense of "first-level-subdivisions only" vs. "any subdivision" holds and the items should not be merged. -- Gymel (talk) 05:48, 21 April 2015 (UTC)
    I think the current Finnish article has now fewer things. It would be difficult to understand if it would be made an another article and then the Finnish Wikipedia article would be equal to current English article. The result would be two new almost the same kind of articles. But you may be right, because you know better structure of Wikidata. Thank you. --Raid5 (talk) 13:16, 21 April 2015 (UTC)
    @Raid5: Perhaps I did not correctly understand your question, but I let me rephrase my answer:
    1. There is no wikipedia which has two articles (one belonging to administrative territorial entity of Mexico (Q112865) and the other to state of Mexico (Q15149663)), so technically the items could be merged
    2. however the different (and definitively non mergeable) "main categories" for the two items stress the fact, that "primary subdivisions" (i.e. states) and "any subdivision" (i.e. states and municipalities) are distinct topics and therefore the wikidata items should not be merged.
    3. current assignment of individual wikipedia articles to the two items seems to be in alignment with the article titles: "states" here, "subdivisions" there. Deep inspection of each single wikipedia article might yield a better assignment (e.g. an article named "subdivisions" which in fact only deals with the primary subdivisions should be transferred to the "states" item. Or the other way round, an article named "states" which also explains in length how the states are subdivided should move to the "subdivisions" item).
    4. Please feel free to move fi:Meksikon osavaltiot to that item where it fits best based on its content (Hm. I gather that it is not an article at all but just a list or table...). It would be nice then if you also could set the finnish labels for the two items in such a way, that confusion is reduced (thus the acutal article title "Meksikon osavaltiot" should be kept in mind). -- Gymel (talk) 15:17, 21 April 2015 (UTC)
    @Gymel: Hi, we're having a word in the RfC about administrative division classification (For example, an item type of administrative division in Mexico, with
    < Mexican state > instance of (P31)   < type of administrative division in Mexico >
    is beein discussed, consistently with the framework I propose in Help:Classification. This could help sorting this kind of items out correctly, so if we could read and leave a comment on Wikidata:Requests for comment/Adopt Help:Classification as an official help page I would be pleased :) TomT0m (talk) 15:24, 21 April 2015 (UTC)
    @Gymel: Yes I know there is no two articles about that. I meant if, then they would be almost the same. I will not argue the matter (because I don't know much about structure of Wikidata). I am trying to study the things of which you wrote and perhaps then I understand better. --Raid5 (talk) 17:24, 21 April 2015 (UTC)
    Raid5 No. 17:25, 25 April 2015 (UTC)

    Slow creation of properties

    Hello, there are many property proposals, which have had some supporting and no opposing votes since months ago, like the input set. Why haven't they been created yet? Petr Matas 12:38, 25 April 2015 (UTC)

    I've been working through the backlog, and have now created input set (P1851). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:27, 25 April 2015 (UTC)
    @Pigsonthewing: Thank you, Andy! I would like to bring Similar item and its qualifiers into focus as well. In December you wrote that the proposal is unclear, but the details are described in the linked RfC. Have you seen that? Petr Matas 14:50, 25 April 2015 (UTC)
    @Petr Matas: I've replied there. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:56, 25 April 2015 (UTC)


    I want really want to start integrating the Wiktionary into Wikidata. With whom do I have to get in contact and what pages do I have to know for this? --Impériale (talk) 08:20, 25 April 2015 (UTC)

    See Wikidata:Development plan#Wiktionary support and Wikidata:Wiktionary. -- Bene* talk 12:33, 25 April 2015 (UTC)
    Thank you :) --Impériale (talk) 09:22, 26 April 2015 (UTC)

    Strange invisible change by KrBot


    KrBot has very recently (a few minutes ago) modified all scanned file on Wikimedia Commons (P996) claims that I recently did, like

    I really cannot see what is the modification. Could someone pleaseexplain it to me, so that when I add thousands of other such claims (for all articles of the same origin in wikisource), I do not make an error that would need such a bot-correction ?

    Thank you very much --Hsarrazin (talk) 07:35, 26 April 2015 (UTC)

    Your file name had underscores instead of spaces. --Pasleim (talk) 07:42, 26 April 2015 (UTC)
    oh, thank you Pasleim, sorry for the accidental copy/paste - the links were working, so I did not know it mattered :) --Hsarrazin (talk) 07:46, 26 April 2015 (UTC)

    Wikidata weekly summary #155

    How to add "A has exactly one B, which is C

    How can I tell Wikidata that astronomical year numbering (Q751976) has a 0 (Q23104), which is 1 BC (Q25299)?
    And when I remove three improper properties from 0 (Q23104) (1, 2, 3), how can I add edit summaries, so other users won't fix it back? Watchduck (quack) 20:05, 24 April 2015 (UTC)

    Dunno, the wikidata editor is a hopeless piece of junk from script hell. Otherwise you could try see also (P1659) 1 BC (Q25299) on 0 (Q23104) and vice versa as a clue. But as is this cannot be saved without any clue what's wrong or missing. Unbreak now: Be..anyone (talk) 22:19, 25 April 2015 (UTC)
    Perhaps that would work? It's the best that I could come up with. I admit to not be clear about anything about this, though. Neither the items nor the properties involved are seem clearly defined to me. Am I understanding correctly that 0 (Q23104) is a year-numbering concept, while 1 BC (Q25299) (existing aliases: 1 BCE, 753 AUC, 0) refers to a fixed period of time, independent of year-numbering system? Dancter (talk) 00:44, 26 April 2015 (UTC)
    Yes, "year zero" is a concept used in some year numbering systems (like a.y.n.), and 1 BC is a specific year in history: "1 BC is the year zero of astronomical year numbering."
    Actually this isn't more complex than "Olympus Mons is the highest point of Mars." or "Bloemfontein is a capital of South Africa.".
    Probably the correct way is to create the property "has year zero". Unfortunately there seems to be no semantic way to make clear that this is a one-to-one relationship. Watchduck (quack) 20:13, 26 April 2015 (UTC)

    Redirect after (trivial) manual merge

    Hi. I just emptied this page after moving the only link to Q10927834. I don't know if this counts as a merge, and if so, how to make a redirect. --Rinaku (t · c) 19:07, 26 April 2015 (UTC)

    You can read about merging at Help:Merge. Moving a link is okay, but you could also move the label for Russian. Redirecting is possible by API or the gadget (see the help). Matěj Suchánek (talk) 19:27, 26 April 2015 (UTC)
    There was no label. I was unsure about merging the page after emptying it. Anyway, done. --Rinaku (t · c) 19:31, 26 April 2015 (UTC)

    wd-condensed.css: Condensed Wikidata UI CSS

    I've been working on some CSS to make Wikidata's UI more condensed and it is now stable enough for other people to try.

    Add @import url('//'); to your common.css file (Special:Preferences#mw-prefsection-rendering, and click "Shared CSS/JavaScript for all skins: Custom CSS").

    Here are some screenshots:

    Notable improvements:

    • Responsive layout; layout changes depending on the window width to make viewing/editing easier
    • Columned view: statements in a statementgroup are now inlined to take less space
    • edit/add buttons are invisible until you hover over a statementgroup/statement

    Known issues:

    • Columned view has holes when statements have inconsistent heights
    • Width of references list is sometimes small

    I've used it for the last few weeks or so, and I believe I've found/fixed all the critical bugs. Only tested on the latest browsers (~ Chrome 40+, Firefox 36+, IE11).

    Hope this is useful. –Hardwigg (talk) 08:46, 25 April 2015 (UTC)

    wow, nice, will certainly try it :)
    the way you manage inline display of statements, while keeping them in column for edition is exactly what I would like for language management ;) - now, it's very difficult to add a long label to an item, even worse for descriptions and aliases :( --Hsarrazin (talk) 09:26, 25 April 2015 (UTC)
    @Hardwigg: could one have subClassOf and instanceOf always at the top? In many cases both act similar to the description. FreightXPress (talk) 23:35, 25 April 2015 (UTC)
    @FreightXPress: Not with plain CSS. If you want that, add the following to your common.js page:
    if ( mw.config.get('wgNamespaceNumber') === 0 ) { // On an entity page
    	$('.wikibase-statementgrouplistview > .wikibase-listview')
    		.prepend($('#P31, #P279'));
    Hardwigg (talk) 02:59, 26 April 2015 (UTC)
    Thanks for sharing this. :) It's great to see people trying to reduce the amount of unnecessary whitespace. That's one of the things which bothers me most about the current layout. I won't be using this as it is though. :( My comments:
    • The hover effect is actually a big hindrance to me, not an improvement. It means I can't see where I'm aiming for on my main computer until I'm already close. It also seems to make it completely impossible to edit statements on touch screen devices like my tablet, since they don't really do hovering.
    • I don't really like the position of the edit link. It feels a bit lost amongst all the qualifiers and references and I keep wanting to interpret it as a link for editing the references, not for editing the statement.
    • The "add qualifier" link when adding a statement appears on the very right of my screen, a long way from the input field and save links.
    • The font sizes for the edit/save links and reference toggle are too small for me and it also makes them harder to click because they're even smaller than they were and I have to be even more accurate with my aiming. It's also problematic on touch screen devices where accurate clicking is difficult.
    • The headings for the "In more languages" section don't have enough padding and the letters are touching or even hanging out of the grey bar.
    • The responsive design doesn't seem to kick in on my mobile or tablet, not sure if it's possible to make that work.
    • This also seems to move the sitelinks (which are currently on the right for me on my main computer) back to the bottom. I'm not sure if that was intended or not, but most of the time there is only one statement for a particular property, so it means the right side of my screen is mostly wasted again, like it used to be before the sitelinks moved there.
    Thanks for the piece of JS above! The statements having no consistent ordering is the other thing which bothers me but I hadn't managed to figure out how to reorder them using JS. - Nikki (talk) 09:21, 26 April 2015 (UTC)
    @Nikki: Thanks for the great feedback, Nikki! I'm working on these issues, and will post an update sometime in the next few days. -Hardwigg (talk) 03:27, 28 April 2015 (UTC)

    Hardwigg - thanks for the JS. I extended it, see below #wd-sorted-statements.js. I agree with Nikki that hiding edit links and requiring hover is a hinderance. FreightXPress (talk) 17:13, 26 April 2015 (UTC)

    Two posts for vietnamese-american author Nha Ca

    Q112180187 and Q10800987 refers to the same person. I do not know how to correct this. Boberger (talk) 12:22, 27 April 2015 (UTC)

    Could you please correct the first link.--Ymblanter (talk) 13:07, 27 April 2015 (UTC)
    Did you mean Q19629357? I have merged these. --Yair rand (talk) 14:06, 27 April 2015 (UTC)

    Merge problems (2)

    Can someone merge englisch en:Category:Biotechnologists (Q7626947) with German de:Kategorie:Biotechnologe (Q8908992) ? 17:33, 27 April 2015 (UTC)

      Merged --Pasleim (talk) 17:41, 27 April 2015 (UTC)

    Can someone merge english en:Category:Bion satellites (Q7021297) with German de:Kategorie:Forschungssatellit (Biologie und Medizin) (Q8957715) ? 17:40, 27 April 2015 (UTC)

    Can someone merge en:Category:Hunting legislation with German de:Kategorie:Jagdrecht (Q8978391) ? 17:50, 27 April 2015 (UTC)

    Can someone merge en:Category:Biotechnology law (Q7487449) with German de:Kategorie:Gentechnikrecht ((Q16863875)) ? 17:54, 27 April 2015 (UTC)

    Done. --Marek Koudelka (talk) 18:00, 27 April 2015 (UTC)

    Can someone merge en:Category:Cell culture (Q7305817) with German de:Kategorie:Zellkultur ((Q19617161)) ? 18:14, 27 April 2015 (UTC)

    Can someone merge en:Category:Marine reserves (Q7145269) with German de:Kategorie:Meeresschutzgebiet (Q9017452) ? 18:25, 27 April 2015 (UTC)

      Merged, I think it's better if you register here and activate the merge gadget. These requests will become annoying after some times. Sjoerd de Bruin (talk) 18:26, 27 April 2015 (UTC)
    Even without an user account it's possible to merge items with Special:MergeItems --Pasleim (talk) 18:40, 27 April 2015 (UTC)

    Match available Wikipedia data with empty Wikidata data (From infoboxes)

    Hello Wikidatorians,

    After discovering this project today I noticed a potential improvement on the current system of Wikidata but i'm not sure if this is possible or not, so it will be great if someone can review this comment and let me know how we can do this. After viewing Anzac Day which was a public holiday recently observed in Australia and New Zealand I noticed that the date observed (which is reoccurring annually on the same day) is April 25th, this data was not entered into Wikidata until the 15th Sept 2013 however the data was showing on Wikipedia in the infobox ever since 25th Nov 2006 is it possible to programmatically connect data available in Wikipedia from infobox's, to data available in Wikidata that has not been entered? is this even a good idea?

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

    First of all: Wikidata didn't exist in November 2006. Wikidata was started on October 29, 2012, statements weren't enabled until February 2013, and data for what day of the year holidays occur could not be added to Wikidata until September 3, 2013.
    Lots of data from Wikipedia is frequently mass-imported to Wikidata, and of course many Wikipedias are using data from Wikidata in their infoboxes. Importing data can be problematic as the infoboxes are sometimes unsourced, or sometimes importing the sources automatically can be difficult. --Yair rand (talk) 08:38, 28 April 2015 (UTC)
    Thank you for clarifying the details around timings. Since Wikipedia infobox data is problematic do you then think it is a better solution for someone like me or anyone else to manually update data in Wikidata and find sources on government websites for things like public holiday dates if they reoccur on the same day each year. --Deepheater (talk) 03:23, 29 April 2015 (UTC)
    Yes, that would be exactly right. --Denny (talk) 16:21, 30 April 2015 (UTC)
    Is there a commonly used method to enter the date of a Wikidata event such as a Public holiday that is not a specific date of the year like April 25th every year. How would someone enter the date of an event if that date was a unique date to this year like on the 26th of April this year but on the 27th of April next year for example. Also is there a method to enter data such as this event occurs on the First monday before a specific date every year. --Deepheater (talk) 22:19, 30 April 2015 (UTC)
    Use day in year for periodic occurrence (P837) to link to the item for the particular day (we have items for each day - even February 30 (Q37096).) To search for a property put "P:" in front of the search term in the search box. Filceolaire (talk) 05:47, 11 May 2015 (UTC)

    Merge Problems

    Can someone merge en:Category:Fishing museums (Q16780536) with German de:Kategorie:Fischereimuseum (Q8954440) ? 18:45, 28 April 2015 (UTC)

      Done see Q8954440. Delsion23 (talk) 23:58, 28 April 2015 (UTC)

    Generating items from Wikipedia categories

    In the category cy:Categori:Eisteddfodau yn ôl blwyddyn on, the articles for eisteddfodau (plural of eisteddfod) up to 1881 have Wikidata items, but not the ones for 1882 to the present day. Can the remaining items be generated automatically from the Wikipedia category? Ham II (talk) 05:17, 29 April 2015 (UTC)

    With you can create them directly. --- Jura 06:34, 29 April 2015 (UTC)
    @Jura1:   Thank you.—starting to get the hang of it now! Ham II (talk) 06:52, 30 April 2015 (UTC)
    To add English labels you can edit the result from [19] and add it with [20]. Adding "point in time" is a bit more complicated. --- Jura 08:23, 30 April 2015 (UTC)

    How to Create company database in Wikidata

    Hi there,

    It would be great if you help me about how to save company database into wiki data. As we know Freebase is a medium to get fetched data as a knowledge graph into google searches. Since Freebase stopped providing services and its entire data is migrating to So how we will come to know or able to get database as knowledge graph for google search engine. Please suggest the steps to create company page as simply we had to create in freebase. --VarunNegiUr (talk) 08:38, 29 April 2015 (UTC)

    Please read this first. Also, your company needs to meet our notability policy. Sjoerd de Bruin (talk) 08:47, 29 April 2015 (UTC)

    extract items ordered by value on a certain property

    I would like to (continually) be given a list of all items that have a certain property, ordered by the value of that property. Specifically, I would like all the items that have the Perry Index (P1852) property, ordered numerically by their value. So, say, a table the first column listed the P1852, the second column listed the (possibly more than one) items that have that specific value.

    What is, at the moment, the easiest way to that? I know of Wikidata:Database reports/Constraint violations, but I don't mean to limit the above query to values that are erroneous. Gabbe (talk) 09:48, 29 April 2015 (UTC)

    I also know how to construct this, but not how to sort the table by its rightmost column. Gabbe (talk) 10:26, 29 April 2015 (UTC)
    I would try downloading it. Spreadsheet software should be able to sort it. Popcorndude (talk) 11:38, 29 April 2015 (UTC)
    You might want to try TAB --- Jura 11:46, 29 April 2015 (UTC)

    Commons category cannot be linked anymore

    commons category can no longer be linked at Other Sites, e.g. Otto Modersohn (Q71425) c:Category:Otto Modersohn --Oursana (talk) 12:28, 29 April 2015 (UTC)

    Seems working here. Sjoerd de Bruin (talk) 12:30, 29 April 2015 (UTC)
    Thanks for your help, today it finally works again--Oursana (talk) 13:45, 30 April 2015 (UTC)

    Should usage instructions be included in descriptions?

    Please join the discussion at Help talk:Description#Putting usage instructions in descriptions. Cheers! Kaldari (talk) 21:35, 29 April 2015 (UTC)

    Why not use appropriate properties to describe? And reduce "description" to comment for rare cases. FreightXPress (talk) 14:21, 30 April 2015 (UTC)

    Merge problems: Eyal

    Can someone merge German article de:Eyal (Q1385639) with english article en:Eyal (disambiguation) (Q5422549) ? 23:26, 29 April 2015 (UTC)

      Merged -- LaddΩ chat ;) 02:29, 30 April 2015 (UTC)