Wikidata:Project chat/Archive/2015/08

From Wikidata
Jump to navigation Jump to search

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

Contents

Easy way to add genealogy information

I've been playing with this tool for a bit: http://tools.wmflabs.org/wikidata-todo/quick_statements.php. It works great, but I easily get confused when adding parent/child relations.

Isn't there something that would be able to do the following conversion automatically:

Anna is the mother of Petra

QofPetra P22 QofAnna QofAnna P40 QofPetra

If it doesn't exist, I'm itching to write it. So basically a converter of natural language to such statements. I guess with support for assigning Q and P numbers to the 'variables' and placeholders for the ones that are not on wikidata yet.--Polyglot (talk) 20:41, 30 July 2015 (UTC)

This doesn't actually add any information. It just adds the same information in two different forms. Please don't do it - we just end up with two statements to maintain for no benefit. Joe Filceolaire (talk) 01:25, 31 July 2015 (UTC)
@Filceolaire:I see, I was wondering about that. So which one is more important the parent or the child property? Anyway, that wasn't the original question though. I was wondering if something already existed to convert natural language statements to Q P Q statements.--Polyglot (talk) 07:40, 31 July 2015 (UTC)
@Polyglot: I had written a module about it. However you should read the module and some examples before using it. This module may contain bugs, so you should check the result before executing it.--GZWDer (talk) 05:20, 31 July 2015 (UTC)
I'll have to figure out what a module is and how to use it then.--Polyglot (talk) 07:40, 31 July 2015 (UTC)
@Filceolaire: I'd been under the impression that best practice was to have A:parent B and B:child A - I agree this is more complex than it needs to be, but it's what we're stuck with in the absence of a real symmetrical property. Is there a guideline on this somewhere? Andrew Gray (talk) 19:42, 31 July 2015 (UTC)
If you check father (P22) and mother (P25) they require an inverse property child (P40), however P40 doesn't demand an inverse P22 or P25. (Which is kind of strange to me). Mbch331 (talk) 08:08, 1 August 2015 (UTC)
The constraint report for P40 now shows missing P22,P25 directly. If added manually, the easiest is to add any of P22, P25, P40 and wait for bots to add the inverse property.--- Jura 08:14, 1 August 2015 (UTC)
I suppose one problem here is that if item A has P22:B then B must have a P40, but if A has P40:B, then B might have either P22 or P25 - it needs to have one but not required to have both, which makes for quite a complex constraint! Andrew Gray (talk) 11:31, 1 August 2015 (UTC)

Mix-up of drummer and drumkit

I don't know where I can put a request for a rather great maintenance job, it's to big for me. As a matter of fact many drummers (Q386854) are connected to drum kit (Q128309). It might be that the Wikipedia pages of different languages are not correctly linked, because for instance Slagzeuger (drummer) on German Wikipedia is a redirect to Slagzeug (drum kit). The same is happening with Nedersaksies/Dutch Lower Saxon, where dummer is a redirect to drums (drum kit). I hope someone is able to unravel them. Thanks in advance! Ymnes (talk) 07:31, 1 August 2015 (UTC)

Do you have a sample? I found one that has occupation (P106) = drummer (Q386854) and instrument (P1303) = drum kit (Q128309). This seems a bit redundant, but correct. There don't seem to be any that have occupation (P106) = drum kit (Q128309) [1]. --- Jura 07:38, 1 August 2015 (UTC)
For an overview, try TAB. I added instrument (P1303) = drum (Q11404). --- Jura 07:47, 1 August 2015 (UTC)
Well I may have been confused, because I though all these links were linked to profession, but now I look more closely, I see that they are linked to instrument. I'm sorry that I didn't notice this directly. Ymnes (talk) 16:14, 1 August 2015 (UTC)

Help with an element

Hi. I noted than no:Hemisodorcus is not linked at no label (Q18584001). I tried to fix it but it says there is a problem, but no explanation. Can someone tell me wich one is the problem? Thanks. --Ganímedes (talk) 12:18, 1 August 2015 (UTC)

In English, when saving, there is a link "details" you can click on. This displays the following error: "The link nowiki:Hemisodorcus is already used by item (Q11974967). You may remove it from (Q11974967) if it does not belong there or merge the items if they are about the exact same topic.". --- Jura 12:24, 1 August 2015 (UTC)
I think they should be merge, since Q11974967 is incomplete and had no further info. Now, How can I do this merge? Thanks. --Ganímedes (talk) 12:42, 1 August 2015 (UTC)
Help:Merge --- Jura 12:49, 1 August 2015 (UTC)
Better link: Help:Merge#Automatic_merge --- Jura 12:53, 1 August 2015 (UTC)
Mmmm, I don´t think understand how it works... Can you or someone else do it? Thanks a lot. --Ganímedes (talk) 12:57, 1 August 2015 (UTC)
I noticed the description isn't easy to read. Adding the two items on Special:MergeItems seems easiest. --- Jura 13:04, 1 August 2015 (UTC)
Ganímedes, I merged them. --Succu (talk) 16:26, 1 August 2015 (UTC)

Notes on rules of play with properties

Hi all, before it disappears into the archive, pls read Wikidata:Project_chat#Number_of_images_per_item for background on an issue regarding image (P18). This discussion is related to the discussion above at MOMA's image release, in that the property described at URL (P973) is considered a hack for when the url can't be placed anywhere else in the item. I have added that property and also described by source (P1343) to painting items where they weren't described by the painting owner and sometimes when they were, depending on how useful I found the info in the link. If I go to the discussion pages of any of these properties I don't see clear instructions, but I suppose that is where such instructions should be. Any ideas how to channel these issues? Should all comments (like this one) be put on property talk pages? Thanks for your input. Jane023 (talk) 07:45, 31 July 2015 (UTC)

On either place, it helps if you try to open a discussion about single topic (Where to discuss? or What's P973 for?) not several (like these two unrelated ones). We end up talking about our favorite things anyways. Try to do this on one place only (Don't copy posts to several places and expect people to keep it in sync). For properties, I'd start out on property talk, but I don't see what would prevent us from discussing it here. --- Jura 08:40, 31 July 2015 (UTC)
We have to discuss here. This is a generic properties, I do not see why there only should be those who watches the page and are aware of the property at all that should take all the decisions. (If there is decisions to take, as I said I'm for "use common sense" here, we should not build a strongbox of barely useful rules ...). More generally, the information spread out everyhere on property pages makes the information very hard to find. We should have an index for informations about generic properties and basic instructions on how to use them (that was one of the purpose of Help:modelling when I created it, being an entry point.) author  TomT0m / talk page 09:13, 31 July 2015 (UTC)
Prefering the English language chat over topical discussion: I think this type of reasoning is frowned upon in multilingual projects. --- Jura 07:41, 1 August 2015 (UTC)
@Jura1: The result is mainly a lot of topical discussions in english ... At least a central point is easier to translate. You argument is flawed: most topics are multilingual, and only a few WIkiProjects are translated. author  TomT0m / talk page 10:33, 1 August 2015 (UTC)
I think all wikiprojects should be translated. If not, they are only accessible to those who speak the same language as the creators (not always english-speakers) :) see discussion above --Hsarrazin (talk) 19:05, 1 August 2015 (UTC)
Indeed, but translating in 200 languages is hard, so let's limit the amount to be translated :) A few generic and solid properties usable in a vast variety of situation is better in that concern to a lot of specific properties each with their documentation if they could be merged into one. Of course, if the length of the doc of the merged property turns out to be the sum of the size of the specific properties, it would be an indication that the property should be splitted :) author  TomT0m / talk page 19:28, 1 August 2015 (UTC)

Constraint advice needed, please

Can someone please show me who to add a constraint to website account on (P553), to warn if it is used for Twitter, Instagram or Facebook? And does that automatically generate a warning to editors who try to do so, or does that need to be done separately? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:00, 1 August 2015 (UTC)

I think adding {{Constraint:Conflicts with|list = {{P|553}}: {{Q|918}} }} might work for making it show up in the constraint report (i.e. make the statement conflict with itself) which would be a bit weird but I'm not sure if there's another way to do it. It wouldn't automatically generate a warning to editors, I'm not sure how those are defined. - Nikki (talk) 18:06, 1 August 2015 (UTC)
Oh, now that I think about it, I think the warnings are generated by abuse filters. - Nikki (talk) 18:13, 1 August 2015 (UTC)
Of course. thanks. So how do we get one set up? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:36, 1 August 2015 (UTC)
No, actually, the warning that pops up is the mediawikiwiki:Wikibase Quality Extensions, which are based on our already-used constraint system. --Izno (talk) 05:21, 2 August 2015 (UTC)

Strange display...

Hello,

On Amanullah Abbasi (Q15983392), I have a strange display of name in native language (P1559) : is this caused by the right-alignment of ourdou, or is it something else.

Could someone please check it for me, if I made an error ? Thanks --Hsarrazin (talk) 18:58, 1 August 2015 (UTC)

Yes, it looks like the root cause of this issue is the right-alignment of ourdou. I think it should be reported to the development team. Tpt (talk) 19:49, 1 August 2015 (UTC)
Done --Hsarrazin (talk) 20:24, 1 August 2015 (UTC)

Suggestion: transfer redirections to wiki data

Hello.I suggest transfer redirections to wiki data Instead of repetition in all the projects , by making interwiki links and names Common conversions, and for scientific names, be in English>We can exclude some items (example) using a page in MediaWiki namespace --ديفيد عادل وهبة خليل 2 (talk) 15:26, 1 August 2015 (UTC)

You've suggested this before and were shot down then. Why is it better now? --Izno (talk) 16:23, 1 August 2015 (UTC)

I think that the proposal is useful --ديفيد عادل وهبة خليل 2 (talk) 16:49, 1 August 2015 (UTC)

Since redirections are mostly based on language, and hence, different in every language site, this would be more than pointless… counter-productive.
Moreover, as wikidata doesn't handle redirection links, it wouldn't work :) --Hsarrazin (talk) 19:01, 1 August 2015 (UTC)
I think this query (sorry can't read your name) is not about the Wikipedia redirect concept but it is about the Wikidata alias concept. Redirects on Wikipedia are often used for various common spelling of names and so forth, and it would improve Wikidata search if such aliases were propagated to more than one language. Unfortunately, though I have sometimes wanted this myself, I don't think it will work because many redirects are based on common misspellings that would only occur in the specific language. So Juan or Johann instead of John and so forth. Jane023 (talk) 10:11, 2 August 2015 (UTC)
The English wikipedia can take care of itself so lets think how wikidata could help readers of smaller wikipedias find articles in their language. Let's say a reader searches on a word in their language and wikidata has a bunch of items with that word as a label or an alias - because someone has entered labels or maybe even because some automated translator has made a best guess. Their language wikipedia doesn't have an article with exactly that name. It seems to me the best bet is to
  • show them a list of the wikidata items with that name or alias, each with a description automatically generated from the basic classification properties ('instance of' 'occupation' 'country' etc.) from which they can pick one - effectively a wikidata generated disambiguation page.
  • If they pick an item with a wikipedia article in their language show them that article
  • If they pick an item with no wikipedia article in their language then show them a Reasonator page for that item. This lists the languages with articles about that item.
  • If there are no articles in their language about that item then they can navigate to related items until they find an article in their language.

This seems, to me, to be the best way help readers on smaller language wikipedias find articles. Trying to get wikidata to do redirects has the problem that each wikipedia has a different set of articles so each needs different redirects. Joe Filceolaire (talk) 14:57, 2 August 2015 (UTC)

Merging issue

I've stumbled on two articles that may need merging, when Q20666458 came up as an article without an instance of/subclass of.

It seems Q20666458 (Butte Copper Kings) and Q5594713 (Grand Junction Rockies) should be merged, although I'm not sure. They are two baseball teams, but the second one seems to be a continuation of the first. It seems a straight merge won't work though, as they both have an article on the French Wikipedia. The Q20666458 links to a French wiki article that seems to describe the team as a former baseball team, whereas Q5594713 has English and French Wikipedia articles that describe Butte Copper Kings as a former name of the Grand Junction Rockies. I'm not sure how to proceed, as I assume we would need to remove one of the wikilinks in order to merge the wikidata articles, and it seems that the articles on the French wikipedia may need to be merged. -- – The preceding unsigned comment was added by Silverfish (talk • contribs) at 20:33, 1 August 2015 (UTC).

Have you ever tried fr:Wikipédia:Pages à fusionner? --Liuxinyu970226 (talk) 05:23, 2 August 2015 (UTC)
Either the articles can be merged, or the two items linked using said to be the same as (P460) or similar. (I think there's a "continued as" property lying around, since this issue is not new.) It's easier to do the second than the first. --Izno (talk) 05:24, 2 August 2015 (UTC)
en:Grand_Junction_Rockies#Franchise_history explains how they relate. --- Jura 05:26, 2 August 2015 (UTC)
Use replaces (P1365) and replaced by (P1366) to link between these. Joe Filceolaire (talk) 14:25, 2 August 2015 (UTC)

Email adresses for volunteers

Dear all,

A proposal to create Wikimedia email adresses for volunteers can be found at meta:Wikimedia Forum#Wikipedia / Wikimedia volunteer email adresses. Your input would be very welcome.

Sincerely, Taketa (talk) 07:12, 2 August 2015 (UTC)

Akademy awards

Hoi, the article on Akademy on en.wp includes information about three awards. I have created items for them and the people who received them. The question is, can we make the same presentation as exists in the English article? Thanks, GerardM (talk) 16:34, 2 August 2015 (UTC)

Blogged about another aspect of awards..

Wikidata weekly summary #169

Next 2 rounds of arbitrary access rollout

Hey folks :)

We'll be rolling out arbitrary access to a large number of wikis next week and the week after. See Wikidata:Arbitrary access for the full list.

Cheers --Lydia Pintscher (WMDE) (talk) 17:54, 3 August 2015 (UTC)

👍1 --AmaryllisGardener talk 21:10, 3 August 2015 (UTC)

Constraint violation which seems like it can't be resolved

When I check the constraint report for no label (Q2005111) it currently gives 2 violations. The violation for lyrics by (P676) has been resolved, but won't show up until the changed constraints have been updated in the report, but it also gives a range error on date of first performance (P1191), which I don't know how to resolve. The entered date is valid when I check the range on the property talk page. So either the range on the talk page is invalid, or the report can't handle dates very well. Mbch331 (talk) 19:24, 3 August 2015 (UTC)

Try Wikidata:Contact_the_development_team. --- Jura 08:44, 5 August 2015 (UTC)

Replace reference with pywikibot.

Hello. Is there a way to replace a reference with another one, in a claim? Xaris333 (talk) 18:47, 4 August 2015 (UTC)

Eager to learn why you would do this? Thanks, GerardM (talk) 09:06, 5 August 2015 (UTC)

Fusion problems

Can someone fusion en:Category:Extinct birds of Europe (Q20749162) with es:Categoría:Aves extintas de Europa (Q9424256) ? --Ellicrum (talk) 20:01, 4 August 2015 (UTC)

You can do that yourself, see Help:Merge. Sjoerd de Bruin (talk) 20:13, 4 August 2015 (UTC)

Properties in sports season

I am trying to add claims in sports season. I am using follows (P155), followed by (P156), sport (P641), country (P17), instance of (P31), participant (P710) with ranking (P1352), subclass of (P279) with edition number (P393), start time (P580), end time (P582), winner (P1346), organizer (P664), official website (P856). Do you know any other properties than can be added? Xaris333 (talk) 02:47, 5 August 2015 (UTC)

@Xaris333: Not an answer, but is there a wikiproject to compile all those informations ? author  TomT0m / talk page 08:36, 5 August 2015 (UTC) 
We have Wikidata:WikiProject Sports but it doesn't yet deal with sports seasons. --Pasleim (talk) 09:06, 5 August 2015 (UTC)

Instructions how to merge can't be found easily

I ran across a user today that found a very strange way to request the merger of two items. The user created a talk page with a delete request explaining the item needed to be merged with another item. I then explained the user on his talk page that he could do it himself. The user replied by saying the instructions aren't easy to be found. There is no link to Help:Merge from the community portal. In the header of this page, users are pointed to the RfD page if they want items to be merged. How can we make it easier for users to find the information on how to merge? Mbch331 (talk) 13:58, 5 August 2015 (UTC)

We could change the header so it does point to the merge page. Popcorndude (talk) 14:07, 5 August 2015 (UTC)
As I have done. Popcorndude (talk) 14:11, 5 August 2015 (UTC)
But will people be going to the Project Chat when they want to merge items? Or are the going to the Community Portal? (Not that the change wasn't good. People that do come here need to find the instructions as well.) Mbch331 (talk) 14:14, 5 August 2015 (UTC)
It would probably be a good idea to add it to the Community Portal as well, but I wasn't sure where on the page to put the link, so I haven't added it yet. Popcorndude (talk) 14:16, 5 August 2015 (UTC)
I also changed the header of the RfD page and marked it for translation. Now it only needs to be translated to other languages. Mbch331 (talk) 15:08, 5 August 2015 (UTC)

trent chif

what is ftp – The preceding unsigned comment was added by Chifural (talk • contribs).

See File Transfer Protocol (Q42283). Mbch331 (talk) 14:01, 5 August 2015 (UTC)

Need help for format constraint

I need the help of regexp expert:

Thanks. Snipre (talk) 17:11, 5 August 2015 (UTC)

I'm not entirely sure what you're asking for, do you mean "not including [@\\/]" and "must include [@\\/]" or something else entirely? Popcorndude (talk) 21:55, 5 August 2015 (UTC)
If that interpretation is correct, just remove the "@", "\\", and "/" for canonical SMILES (P233) and for isomeric SMILES (P2017) try something like "[A-Za-z0-9+\-=#$:().>[\]%]*[@\\/][A-Za-z0-9+\-=#$:().>@\\/\[\]%]+" Popcorndude (talk) 21:59, 5 August 2015 (UTC)

Pywikibot1

Hello. Look 1935–36 Cypriot First Division (Q2706388). There is participant (P710) with all the teams that took part. Is there a way, with pywikibot or with another bot, to add as a qualifier number of matches played (P1350) with number 14 to all participants? Xaris333 (talk) 19:43, 5 August 2015 (UTC)

try this:
# -*- coding: utf-8  -*-
import pywikibot
site = pywikibot.Site('wikidata', 'wikidata')
repo = site.data_repository()
item = pywikibot.ItemPage(repo, 'Q4115189')
item.get()
qualifier = pywikibot.Claim(repo, 'P1350')
value = pywikibot.WbQuantity(amount=14, error=0)
qualifier.setTarget(value)
for m in item.claims['P710']:
    m.addQualifier(qualifier)

--Pasleim (talk) 19:59, 5 August 2015 (UTC)

Thanks! Xaris333 (talk) 21:00, 5 August 2015 (UTC)

Next weeks featured item

Hoi, I blogged about Mr László Krasznahorkai who is likely to be a featured item. I haved added more awards than there just to be for Mr Krasznahorkai. My point is very much that we need to collaborate on enriching information. As I wrote, awards link people and make them of even more interest. Thanks, GerardM (talk) 06:08, 3 August 2015 (UTC)

Would you add references for the awards? As it's a featured item, its statements should be reasonably referenced. This was fine when it was promoted. --- Jura 07:00, 3 August 2015 (UTC)
Given that when I add statements I do not only do it for one specific item, sorry no. It does not fit in with my way of working. Thanks, GerardM (talk) 09:20, 3 August 2015 (UTC)
But isn't the point of a showcase item that it does not only contain a lot of information but it's well referenced too? As it is important to encourage referencing on Wikidata. Adam Harangozó (talk)
Adam: Thanks for adding them. --- Jura 05:56, 4 August 2015 (UTC)
Is the point of a featured item not that it is feature complete. This item was manifestly incomplete. Sources are by definition secondary. Thanks, GerardM (talk) 06:11, 4 August 2015 (UTC)
It was featured this week as it was promoted "showcase item". Showcase items require sources and a minimum number of statements. --- Jura 08:47, 5 August 2015 (UTC)
It proves the point that Wikidata is a work in progress. I cannot source Hungarian, I can add missing statements. Horses for courses and the prize is a fleeting moment. GerardM (talk) 09:06, 6 August 2015 (UTC)

UN volunteer project

Heading shortened from UN volunteer project: Investigate potential of English/Hindi Wikipedia as a Big Data source to inform Child Marriage Policy

Hi all

Just found this on the UN volunteering site, Investigate potential of English/Hindi Wikipedia as a Big Data source to inform Child Marriage Policy

https://www.onlinevolunteering.org/en/vol/opportunity/opportunity_form.html?id=59724

If you know anyone who would be interested please pass it on.

Cheers

Mrjohncummings (talk) 17:28, 6 August 2015 (UTC)

100.000 paintings on Wikidata

A bit of news from the WikiProject sum of all paintings: We reached the milestone of 100.000 paintings on Wikidata. Multichill (talk) 15:58, 13 August 2015 (UTC) Ps. Also mentioned on the project talk

Bot upload of some clean parts of Freebase content

Freebase contains some facts tagged as "reviewed", i.e. curated by humans. I believe that their quality is high enough for a bot upload. We may start, as a test, with 35,074 place of birth (P19) that are currently not in Wikidata. I am ready to create a bot which would add the statement if, and only if, no place of birth (P19) statement is already in the item. If there is already a different place of birth, the bot would not change anything but report the conflict on a specific page.

What do you think about it? Should I start a bot approval request? Tpt (talk) 23:17, 3 August 2015 (UTC)

Symbol support vote.svg Support This seems to me like the obvious course of action. Popcorndude (talk) 01:32, 4 August 2015 (UTC)

Tpt, can you link to an explanation at freebase what "reviewed" means (facts tagged as "reviewed")? FAQ doesn't mention this? Who reviewed? What are the sources/references for the "reviewed fact"? How many years old is the review? Did the review just "copy the truth from Wikipedia"? --Atlasowa (talk) 05:36, 4 August 2015 (UTC)

Hi @Atlasowa:. There is no external description of what it means to be "reviewed". Basically it means that these triples have gotten more scrutiny than the others and had human curation. The curation can be quite old (although I don't think it makes a difference for the place of birth), and it will often be the case that it was checked against the Web - and thus Wikipedia - by the reviewer.
No one wants to quickly force these statements into Wikidata. It is up to the community as a whole to decide whether we want these statements to be uploaded automatically, or add an additional layer of human curation through the primary sources tool. Both is fine. I think it might make sense to upload the data that was identified to have a particularly high quality directly (which you can check in the random sample and in the whole data), and focus the human curation on the data where this is not the case, but it really is up to us to decide what we are comfortable with. I understand that some of us are more impatient (Hi Gerard!) and others are more cautious, and it is good to have both (and other) voices in a respectful conversation. --Denny (talk) 17:56, 4 August 2015 (UTC)
As far as I am concerned this is sheer procrastination. :( GerardM (talk) 06:13, 4 August 2015 (UTC)
I have just opened a bot request for this dataset and created a new page to allow people to suggest new dataset to import. Tpt (talk) 21:07, 6 August 2015 (UTC)

Adding links to Spanish Wikipedia

Hello, when I try to add a link to the Spanish Wikipedia, I write "es" in the language box. But the first item that appears is Esperanto, whose initials are "eo" and has considerably less articles / items / links. Can you fix it? Thanks! -- NaBUru38 (talk) 14:28, 4 August 2015 (UTC)

Write instead "esw". --Helmoony (talk) 22:40, 6 August 2015 (UTC)

100.000 paintings on Wikidata

A bit of news from the WikiProject sum of all paintings: We reached the milestone of 100.000 paintings on Wikidata. Multichill (talk) 15:58, 13 August 2015 (UTC) Ps. Also mentioned on the project talk

Flow

Hello, if you are an user of some of the other Wikimedia projects, you may have noticed that a new discussion system called Flow is currently being deployed to replace the current one. It is still in beta and, while not yet ready for full scale deployment, it is currently possible to activate it on specific pages, provided that the project is okay with it as a global principle, and that the users of the page are ok too.

I think it would be a good idea that we authorize the following on Wikidata:

  • a user would be able to request the activation of Flow on their own talk page
  • active members of a wikiproject would be able to request the activation of Flow on the talk page of the wikiproject, after a discussion between them
  • we activate a it on a generic talk page like WD:Flow where users can test if it suits them.

That way, some voluntary testers would be able to get over any teething troubles that could arise with Flow, including the ones that could be specific to Wikidata, with minimal impact on users that are not eager to try it. What do you think of it? -Ash Crow (talk) 20:25, 26 July 2015 (UTC)

Symbol support vote.svg Support Popcorndude (talk) 20:34, 26 July 2015 (UTC)
Symbol support vote.svg Support Thibaut120094 (talk) 21:19, 26 July 2015 (UTC)
Symbol support vote.svg Support DSGalaktos (talk) 21:36, 26 July 2015 (UTC)
Symbol oppose vote.svg Oppose on anything but a specific test page. It would be too confusing for many editors to a find different system in use on one talk page but not another. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:39, 26 July 2015 (UTC)
I think any change will cause confusion to some people. It can also be very confusing when what you're used to suddenly vanishes, especially if you're having any issues with the thing which replaced it. Would it be better if Flow pages had a banner saying that it's a new system being tested with links to where people can find out more or report problems? That might be a good way to both indicate why the page is different from most talk pages and make it clear how to report any issues people find. - Nikki (talk) 01:08, 28 July 2015 (UTC)
My point is not that we should not change; it is that a half-and-half (or 80:20, or whatever) split is going to cause additional problems. We should test, and then switch completely. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:29, 28 July 2015 (UTC)
Symbol support vote.svg Support I believe that in order to be able to have a smooth migration to flow, testing it on more than specific test pages is required in order to be able to find most important bugs and missing features. Tpt (talk) 04:35, 27 July 2015 (UTC)
Symbol support vote.svg Support. --Stryn (talk) 05:12, 27 July 2015 (UTC)
Symbol support vote.svg Support Mbch331 (talk) 06:19, 27 July 2015 (UTC)
Symbol support vote.svg Support GerardM (talk) 09:10, 27 July 2015 (UTC)
Symbol support vote.svg Support totally agree with @Tpt: here. Seen Flow on test pages elsewhere... impossible to have an opinion on how it works :/ --Hsarrazin (talk) 10:55, 27 July 2015 (UTC)
Symbol support vote.svg Support I agree with Tpt, test pages are not the same as real usage. - Nikki (talk) 01:08, 28 July 2015 (UTC)
Symbol support vote.svg Support let's move on ! author  TomT0m / talk page 10:26, 28 July 2015 (UTC)
Symbol support vote.svg Support Let's do this already! (tJosve05a (c) 21:29, 28 July 2015 (UTC)
Comment. I still remain very negative about FLOW in general and oppose switching to it. However, I am obviously not opposed to testing it.--Ymblanter (talk) 16:05, 29 July 2015 (UTC)
Symbol support vote.svg Support I have never used it and expect it to be great! Matěj Suchánek (talk) 16:28, 29 July 2015 (UTC)
Symbol support vote.svg Support, the sooner, the better. Cdlt, VIGNERON (talk) 17:42, 29 July 2015 (UTC)
Symbol support vote.svg Support after seeing Flow expand in Catalan Wikipedia well received by new and experienced users. I have requested to enable it in Wikidata:Wiki Loves Open Data (phab:T107584) before knowing about this discussion, sorry.--Qgil-WMF (talk) 20:55, 31 July 2015 (UTC)
Symbol support vote.svg Support Ainali (talk) 19:49, 2 August 2015 (UTC)

There are a lot of people saying yes here, and I'm happy to enable some user talk pages if people want them. Is there more discussion needed? DannyH (WMF) (talk) 19:53, 30 July 2015 (UTC)

I think that we could proceed already, as this discussion has been open for a week already. Or should we spam this discussion via Wikidata:Status updates/Next to get possible more opinions?. Though still it's a bit unclear where would we enable it. At least on WD:Flow's talk page and on Wikidata:Wiki Loves Open Data. I'm fine with enabling it on my user talk page also. --Stryn (talk) 18:06, 2 August 2015 (UTC)
Consensus looks good. I'll add it to the Status Update anyway (since it has not gone out yet!) and see if anything else trickles in. John F. Lewis (talk) 18:10, 2 August 2015 (UTC)
Symbol support vote.svg Support Some testing here makes sense. --Daniel Mietchen (talk) 22:48, 2 August 2015 (UTC)
Oppose I do share Pigsonthewing's concerns and am unwilling to contact certain users through an unsufficiently tested beta feature. Vogone (talk) 23:02, 2 August 2015 (UTC)
Symbol support vote.svg Support I don't see the harm in a trial run. Kharkiv07 (T) 03:22, 3 August 2015 (UTC)
Symbol support vote.svg Support Testing for new things are good for better implementation. -Nizil Shah (talk) 18:37, 5 August 2015 (UTC)
Symbol oppose vote.svg Oppose I really don't like Flow. It is complicated and I can't see why this schould beo used instead of normal talk pages. --MGChecker (talk) 14:22, 6 August 2015 (UTC)

Since we seems to agree to have the Flow trials here, the Wikidata:Flow page is ready for your requests! --Harmonia Amanda (talk) 10:20, 7 August 2015 (UTC)

Q11415564 and Q209896

These two properties are used to indicate an "honorary degree" and an "honorary doctorate". They are in my opinion exactly the same thing. They only differ in English. I intend to merge the two. Thanks, GerardM (talk) 05:33, 2 August 2015 (UTC)

Please don't. A doctorate (Q849697) is not the same as a undergraduate degree (Q6008527). Joe Filceolaire (talk) 14:22, 2 August 2015 (UTC)
But both are subsets of academic degree (Q189533). Do any institutions explicitly award honorary undergraduate degrees? Andrew Gray (talk) 14:52, 2 August 2015 (UTC)
Does it matter? ja.wikipedia makes the distinction, so we are required to do so also. --Izno (talk) 15:26, 2 August 2015 (UTC)
@Izno: We surely are required to tell them they have a potential duplicate :) After that discussion maybe we'll have to make a distinction. author  TomT0m / talk page 15:52, 2 August 2015 (UTC)
I did a quick search and [2] and [3] were the first two results. The first one explicitly says they award two types of honorary degree, the second is an example of someone who is in Wikidata being awarded an honorary undergraduate degree. - Nikki (talk) 19:16, 2 August 2015 (UTC)
It is one thing to have a difference in Japanese. The meaning in the languages I know is EXACTLY the same. Thanks, GerardM (talk) 16:31, 2 August 2015 (UTC)
Which is irrelevant. This is not an English project, nor is it a Dutch project. --Izno (talk) 16:39, 2 August 2015 (UTC)
It is enough reason to move everything to one side. When the labels are flat out wrong, it is high time for a remedial action. Thanks, GerardM (talk) 16:53, 2 August 2015 (UTC)
Except a trivially-identified source disproves your point. --Izno (talk) 20:03, 2 August 2015 (UTC)
How? It is obvious that the information held is wrong. It may be that in Japan there is this distinction but I have added incorrectly all the time and it is not only done by me but by so many... Your point is a theory mine is not. GerardM (talk) 05:14, 3 August 2015 (UTC)
What is obvious is that there is a distinction. Period. Your problems with correctly adding a statement notwithstanding. I see no theory here. A source was provided which notes that there are in fact degrees which are awarded which are honorary degrees but not honorary doctorates--even in English.--Izno (talk) 14:33, 3 August 2015 (UTC)
Izno (talkcontribslogs) is entirely correct. My alma mater, a well-known and well-respected university in the northeastern US, routinely grants the degree of MA honoris causa to senior faculty members without an advanced degree from the university. It also grants honorary doctorates annually at the time of its commencement ceremony. Honorary doctorates are an example of/instance of/subset of honorary degrees, but are not identical. And, frankly, if there are two articles on jpwiki, then there have to be two Q-items here, right?
If Japanese is the only wiki with two separate articles, it may be a non-trivial issue as to which Japanese article should be linked to all the rest, but that's something for Japanese speakers to decide, not the rest of us. StevenJ81 (talk) 14:46, 3 August 2015 (UTC)

so how to fix this mess? GerardM (talk) 20:34, 3 August 2015 (UTC)

The best (although not easiest) solution that I can see would be to create a new item "honorary undergraduate degree", link it as a subclass of honorary degree (Q209896) and then go through all the uses of honorary degree (Q209896) to verify whether they are honorary doctorate degrees (if so, change the statements to honorary doctorate (Q11415564)), honorary undergraduate degrees (if so, change the statements to the new item) or whether it's not clear (in which case, leave the statements as they are). - Nikki (talk) 22:05, 3 August 2015 (UTC)
Just for the record: an honorary MA=Master of Arts degree is neither an honorary doctorate nor an honorary undergraduate degree. It's – wait – an honorary master's degree. So there you go.
@GerardM: Why, exactly, do you consider this a mess, anyway? StevenJ81 (talk) 22:45, 3 August 2015 (UTC)
The definition is problematic in and of itself. The data is where the stink is. That is where we have to review ALL the data. Thanks, GerardM (talk) 06:10, 4 August 2015 (UTC)
I truly do not understand what you mean. Never mind, though. You have a lot more edits on this wiki than I do. You fix it as you think best. StevenJ81 (talk) 14:26, 4 August 2015 (UTC)

Given that the data for both properties is mixed neither property provides correct data. Consequently we need a plan B. When we define all these "honorary degrees" as just that, we can qualify by indicating what the degree is about.. They would all be qualifiers, not properties. Consequently I want to empty all the "honorary doctorates" into "honorary degree". It is then left for later to have the correct degree identified. When done, the "honorary doctorate is to go the way of the Dodo.

I see no alternative as the data cannot be relied on. Thanks, GerardM (talk) 09:03, 6 August 2015 (UTC)

The consensus seems to be that no-one disagrees with you. Would you care to rephrase your "there's no alternative" statement? It is clear that there are but you seem unwilling to enact the consensus. --Izno (talk) 21:16, 6 August 2015 (UTC)
Please elaborate... We do not know if a statement of either is an honorary degree or an honorary doctorship. Consequently as I see it they are all honorary degrees and at some stage what honor they are will be added as a qualifier. From your reply I take it you have a workable alternative.. Thanks, GerardM (talk) 12:35, 9 August 2015 (UTC)
An honorary degree is the superclass of honorary doctorship. Where the source does not distinguish, use the superclass. Where it does, use the subclass. As StevenJ81 suggested, create an honorary undergraduate degree item. How is this unworkable again? --Izno (talk) 16:59, 10 August 2015 (UTC)
Right, how does this help with the current crop of data that cannot be relied on? I can make them all correct by making them an "honorary degree". Anything else will have all our data suspect. GerardM (talk) 17:37, 10 August 2015 (UTC)
Why don't you source them instead? I.e. actually improve the data? --Izno (talk) 02:05, 11 August 2015 (UTC)
Because it is too much work. In the mean time the data is wrong. GerardM (talk) 06:34, 11 August 2015 (UTC)
On the one hand you are unwilling to actually fix the data and on the other you claim it's wrong simply by fiat. You must already have a reliable source for such a statement, no? You don't get to have it both ways. --Izno (talk) 13:02, 11 August 2015 (UTC)
Proposed compromise POV. Yes, it's a database. Yes, we need to strive for accuracy. But let's understand two things. (1) Our data sources are not perfectly reliable and are not perfectly complete. (2) We do not have infinite resources here. So I have to suggest the following:
  • Going forward, let's try to categorize things correctly. If we have absolutely no idea whatsoever what the level is of the honorary degree, just make it "honorary degree" and move on.
  • Concerning data we already have, don't sweat it so much. The vast majority of honorary degrees are certainly doctorates. So I wouldn't worry about anything that is already in the system as an honorary doctorate. As people wander through, if they see one that isn't right they can change it. And in the opposite direction, don't worry about anything that is already in the system as an honorary degree, either. That entry wouldn't even be wrong, only less than complete. Let people wandering through change one if they see one that can be categorized more precisely. This is supposed to be collaborative. The individuals in this discussion cannot, and do not need to, make things perfect all by ourselves. StevenJ81 (talk) 14:28, 11 August 2015 (UTC)

Hans Christian Andersen Award

Hoi, this award is in two. It is for writing and illustration. There are two ways to solve this issue. Either split the award in two items or use qualifiers on each person. My preference is for splitting. What do you think? Thanks, GerardM (talk) 08:32, 4 August 2015 (UTC)

I would create two new items for the two categories and set they as an instance of Hans Christian Andersen Award (Q327154) --Pasleim (talk) 19:34, 4 August 2015 (UTC)
It has nothing to do with categories.. Categories are only one way to import data. Thanks, GerardM (talk) 09:08, 5 August 2015 (UTC)
I agree that we should split this into two properties. Joe Filceolaire (talk) 00:22, 9 August 2015 (UTC)

ranking (P1352)

Hello. I used claim participant of (P1344) with seasons (for example 2000–01 Cypriot First Division (Q2698250)). I want to add qualifier something to show the final place of the team in the leaque of that season. I supposed I have to used ranking (P1352). But how to used it to add the information? I try to use numbers but nothing... Xaris333 (talk) 18:41, 4 August 2015 (UTC)

For example at Ayia Napa F.C. (Q312429) you can add a qualifier "ranking: 1" under "2014–15 Cypriot First Division". --Stryn (talk) 19:00, 4 August 2015 (UTC)

Stryn It works now. But is shows 1±1. Why? Xaris333 (talk) 19:06, 4 August 2015 (UTC)

Change it to 1±0. See also Wikidata:Project_chat/Archive/2015/07#Property_P1351. --Stryn (talk) 19:13, 4 August 2015 (UTC)

Thank you very much!! Xaris333 (talk) 19:20, 4 August 2015 (UTC)

Xaris333 You can also add qualifiers to show the goals for P:P1351, goals against P:P1359, league points P:P1358, games played P:P1350, games won P:P1355, games lost P:P1366 for each participant - the whole league table. Joe Filceolaire (talk) 00:12, 9 August 2015 (UTC)
and matches/games drawn/tied (P1357). Joe Filceolaire (talk) 00:18, 9 August 2015 (UTC)

Playboy Playmates

There are a number of items on Wikidata for women who were (sometimes solely) notable for being Playboy Playmates. So Wikidata could systematise the relationship by linking each of those to the item Playboy Playmate (Q728711) (together with the qualifier point in time (P585)). But which property would be most appropriate here? occupation (P106)? award received (P166)? Something else? Gabbe (talk) 11:36, 5 August 2015 (UTC)

position held (P39)? Sjoerd de Bruin (talk) 11:39, 5 August 2015 (UTC)
position held (P39) seems to be primarily for public offices such as senators and emperors. Gabbe (talk) 13:40, 5 August 2015 (UTC)
instance of (P31) always works. --Izno (talk) 15:04, 5 August 2015 (UTC)
No. Their instance of (P31) is human (Q5). -Ash Crow (talk) 12:06, 7 August 2015 (UTC)
No. P31 can always be used for multiple values. Always. Anyone arguing otherwise has this foolish notion that it is damaging to do otherwise. True, I might not want to stuff anything-and-everything into P31 (an argument which I disagree with but I believe to be a valid argument for other reasons than that we couldn't use P31), but when there are edge cases such as this, use P31 or propose a new property. The latter for an edge case is overkill (Q2042339), and so your only recourse is to use P31. --Izno (talk) 13:48, 7 August 2015 (UTC)
Personally, I'd use P106. --- Jura 05:35, 8 August 2015 (UTC)
In my opinion there is a good case to use position held (P39) for Playmate of the Year (Q739857) and "Playmate of the month" (with start date, end date, follows, followed by etc.) but not for generic Playboy Playmate (Q728711). Joe Filceolaire (talk) 23:39, 8 August 2015 (UTC)

Quick statements

Hi! I suppose, that I can't add iw link to some Wikipedia and in the same line add label for it. Am I rigth or I'm doing something wrong? This one just adds the iw link:

Q6688886	Slvwiki	Kategorija:1098. gads	Llv	Kategorija:1098. gads

--Edgars2007 (talk) 09:52, 6 August 2015 (UTC)

I have just tested it in sandbox and it seems that you indeed have to split this expression into two lines. Also don't forget to use "quotes" when you assign a string. Matěj Suchánek (talk) 11:22, 6 August 2015 (UTC)
Thank you. --Edgars2007 (talk) 08:28, 11 August 2015 (UTC)

Translation statistics bug

Hi, I’ve translated some help pages to Russian and opened the Language statistics page. Now I’m confused: http://i.imgur.com/OjMz75c.png (the numbers don’t sum up). What should I do? — Le Loi (talk) 22:51, 6 August 2015 (UTC)

The translation progressbar is indeed buggy, there is often room for me in the bar when there is no message left to translate or update. I concur. author  TomT0m / talk page 09:20, 7 August 2015 (UTC)

Appearance of pages

Is it just me, or has something changed in the way pages are being displayed? Up till now, if I would open an item, I would see several languages displayed (English, English GB, German, and French). There would also be a preview of one of the existing WP articles and a link in the tools bar to "automatic addition". To add an article in a new language, I get a screen titled "set a sitelink", something I did not see before. I did not change any of my preferences, so I'm wondering what is going on. Any help to improve my understanding will be greatly appreciated! --Randykitty (talk) 14:09, 7 August 2015 (UTC)

The weirdest thing just happened. I opened Q2665050 and, hey!, it looked just as before: a preview, four languages, editing as usual, so I already thought this was a temporary glitch or something related to Firefox (I just installed the latest update). Guess what: after closing it, I opened the same item again a few seconds later and it has gone back to the display where I only see the English language description, no preview, etc. --Randykitty (talk) 17:24, 7 August 2015 (UTC)
Sounds like cache (Q165596) to me. -- Innocent bystander (talk) 18:16, 7 August 2015 (UTC)
I have a fast Internet connection, so I have defined my cache at 0Mb... And this happens on both my desktop and my laptop... --Randykitty (talk) 19:57, 7 August 2015 (UTC)

Today the BnF, VIAF, and LC Auth identifiers are not linking either. We have also had editing problems on Wikisource. A change to the way gadgets work appears to be the cause, so if any of this is tied to gadgets, then that may the reason it has been broken. --EncycloPetey (talk) 01:18, 8 August 2015 (UTC)

for 2 days now, I've had troubles with the GUI :
  • wrong property label for displayed value (this is obvious when you try and modify the value)
  • inactive AC links
  • wrong behaviour when trying to edit labels (opens a page to edit labels instead of opening the section
very disturbing :/ --Hsarrazin (talk) 11:53, 9 August 2015 (UTC)

(re)introducing an old friend/enemy… SORTKEY

Aubrey
Viswaprabha (talk)
Micru
Tpt
EugeneZelenko
User:Jarekt
Maximilianklein (talk)
Don-kun
VIGNERON (talk)
Jane023 (talk) 08:21, 30 May 2013 (UTC)
Alexander Doria (talk)
Ruud 23:15, 24 June 2013 (UTC)
Kolja21
arashtitan
Jayanta Nath
Yann (talk)
John Vandenberg (talk) 09:14, 30 November 2013 (UTC)
JakobVoss
Danmichaelo (talk) 19:30, 16 February 2014 (UTC)
Ravi (talk)
Mvolz (talk) 08:21, 20 July 2014 (UTC)
Hsarrazin (talk) 07:56, 9 August 2014 (UTC)
Accurimbono
Mushroom
PKM (talk) 19:58, 10 October 2014 (UTC)
Revi 16:54, 29 November 2014 (UTC)
Giftzwerg 88 (talk) 23:36, 1 January 2015 (UTC)
Almondega (talk) 00:17, 5 August 2015 (UTC)
maxlath
Jura to help sort out issues with other projects
Epìdosis
Skim (talk) 13:52, 24 June 2016 (UTC)
Marchitelli (talk) 12:29, 5 August 2016 (UTC)
BrillLyle (talk) 15:33, 26 August 2016 (UTC)
Alexmar983 (talk) 23:53, 28 August 2016 (UTC)
Finn Årup Nielsen (fnielsen) (talk) 10:44, 29 August 2016 (UTC)
Chiara (talk) 14:15, 29 August 2016 (UTC)
Thibaut120094 (talk) 20:31, 14 September 2016 (UTC)
Ivanhercaz | Discusión Plume pen w.png 15:30, 31 October 2016 (UTC)
YULdigitalpreservation (talk) 17:35, 10 November 2016 (UTC)
User:Jc3s5h
PatHadley (talk) 21:51, 15 December 2016 (UTC)
Erica (ohmyerica) (talk) 19:26, 1 January 2017 (UTC)
User:Timmy_Finnegan
Mauricio V. Genta (talk) 05:38, 12 March 2017 (UTC)
Sam Wilson 09:24, 24 May 2017 (UTC)
Sic19 (talk) 22:25, 12 July 2017 (UTC)
Andreasmperu
MartinPoulter (talk) 09:21, 20 July 2017 (UTC)
ThelmadatterThelmadatter (talk) 01:11, 13 September 2017 (UTC)
Zeroth (talk) 15:01, 16 September 2017 (UTC)
Emeritus
Ankry
Beat Estermann (talk) 20:07, 12 November 2017 (UTC)
Shilonite - specialize in cataloging Jewish & Hebrew books
Elena moz
Oa01 (talk) 10:52, 3 February 2018 (UTC)
Maria zaos (talk) 11:39, 25 March 2018 (UTC)
Wikidelo (talk) 13:07, 15 April 2018 (UTC)
Mfchris84 (talk) 10:08, 27 April 2018 (UTC)
Mlemusrojas (talk) 3:36, 30 April 2018 (UTC)
salgo60 Salgo60 (talk) 12:42, 8 May 2018 (UTC)
Dick Bos (talk) 14:35, 16 May 2018 (UTC)
Marco Chemello (BEIC) (talk) 07:26, 30 May 2018 (UTC)
Harshrathod50
 徵國單  (討論 🀄) (方孔錢 💴) 14:35, 20 July 2018 (UTC)
Alicia Fagerving (WMSE)
Louize5 (talk) 20:05, 11 September 2018 (UTC)
Pictogram voting comment.svg Notified participants of WikiProject Books Ash Crow
Dereckson
Harmonia Amanda
Hsarrazin
Jura
Чаховіч Уладзіслаў
Sascha
Joxemai
Place Clichy
Branthecan
Azertus
ToJack
Jon Harald Søby
PKM
Pmt
Sight Contamination
MaksOttoVonStirlitz
BeatrixBelibaste
ajayi adeniyi
Moebeus
Dcflyer
Looniverse
Pictogram voting comment.svg Notified participants of WikiProject Names @Magnus Manske:

Hello everybody, and especially my fellows sorting-freaks, librarians ;)

now that ListeriaBot is live and kicking, reappears an old need, that was discussed briefly among Book properties, but could not be treated, at the time : "Sortkey".

Indeed, Sortkey is not necessary in many languages, that do not use articles, and that already automatically automatically sort their human (Q5) through a reversed label (like Chinese, Russian, and probably many more that I don't know).

As a matter of fact, there is no real need for Sortkey on a lot of items, but human (Q5) and work (Q386724), especially book (Q571), need to be properly alphabetically sorted when listed, since the label sorting gives, in most languages, a sorting on given name (Q202444), like here, which is not satisfactory.

Since the sorting is completely language-dependant, this is something that cannot be properly adressed through any property, that would require an input for each concerned language.

So… what about adding a 4th language-dependant info, that could be added directly after label, description, and alias : sortkey ?

It could be directly added by contributors in their own language, when needed, and could provide reference for automatically sorting lists.

I post this here, on Project Chat, for maximum impact, but I think it could form a whole Project, so, what do you think ?

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

@Hsarrazin: Are you aware of no label (P1964). - Innocent bystander (talk) 11:17, 8 August 2015 (UTC)
@Innocent bystander:
Thanks, I wasn't, but this can't work in all languages. In French, Vincent van Gogh (Q5582) would be sorted at Van Gogh, Vincent - certainly not Gogh, Vincent van :(( — just look at the wide range of sortkeys on top viaf item for him.
The choice of a property to store sortkey is illogical, especially if you cannot define in the property the language to which it applies (like monolingual does, but this seems to be string).
I agree with your point on PDD :sortkey requires 1 per language : that's why I suggest to add it to language related fields, not as a property. -> to show this, I just added a stupidly qualified value to Vincent van Gogh (Q5582) [4]
I just cannot see how no label (P1964) could work with cross-alphabet items :/ — where is the discussion for creating this property ? --Hsarrazin (talk) 15:58, 8 August 2015 (UTC)
In Dutch we would sort Vincent van Gogh as Gogh, Vincent van (we don't consider van as a part of the last name, nor as a part of the first name. Tussenvoegsels (don't know a good English word) are placed after the first name when sorting). Mbch331 (talk) 16:08, 8 August 2015 (UTC)
errr… I finally found the creation discussion and I remember very well what I wrote there ;
I am amazed that with so few Symbol support vote.svg Support and so many concerned comments, the property was finally created.
I professionally work daily on sortkey for people names and work titles, and even in 1 language (fr) the rules to apply are so complicated that they require pages to find how to sort names. How could this be applied through a string property ? --Hsarrazin (talk) 16:17, 8 August 2015 (UTC)
Well, to many proposals would stall if every doubt would halt the property-creation-process. I therefor understand that it was created. But I doubt that the users involved fully read the linked IFLA-publication. As I wrote in the discussion, IFLA depends on both the location and language of the library. The nationality of the person is of help to identify if Andersson is a patronymicon or not, but it does not give full support to create fail safe sorting rules. -- Innocent bystander (talk) 16:25, 8 August 2015 (UTC)
Sounds to me like no label (P1964) needs to be replaced with a property with monolingual datatype. Note that the language indicated can be one of three varieties of english (en, en-GB or en-Ca) but only one variety of french though surely this could be fixed. To have the datatype changed make a proposal at Wikidata:Properties for deletion. Joe Filceolaire (talk) 16:58, 8 August 2015 (UTC)
I'm not sure a property would really be a solution : it really should be language-linked in obvious place, directly linked to the label. If the label was modified in any language, without revision of the sortkey, the result would be… hazardous. --Hsarrazin (talk) 18:07, 8 August 2015 (UTC)

Help ! weirdest behaviour in editing, since yesterday…

While trying to edit instance of (P31), I ended up https://www.wikidata.org/w/index.php?title=Q4023906&action=history…

When editing, I clicked of P31, but the result was quite different, which appeared only after refreshing. This is really annoying, since one can accidentally change the wrong property without noticing. In fact, I think it happened 3 or 4 times yesterday evening, while cleaning up some items, but hey if you can find them, you're better than me :/

Did someone have the same problems ? --Hsarrazin (talk) 07:27, 8 August 2015 (UTC)

Q7605775 --> Q1368652

Hello, I have moved the only sitelink (plus english label) of Q7605775 to Q1368652, as this is the older and more accurate item. In my opinion Q7605775 could be deleted (or merged), but I do not really understand the process. Will please somebody else do the remaining tasks? --Telford (talk) 10:48, 8 August 2015 (UTC)

Done. Help:Merge describes how you can merge items, by the way. - Nikki (talk) 11:36, 8 August 2015 (UTC)
Thank you! --Telford (talk) 13:36, 8 August 2015 (UTC)
P.S. What I do not understand: why is an item without any content (and with no links to it) kept as redirect and not just deleted? --Telford (talk) 15:51, 8 August 2015 (UTC)
Because there are possible external incoming links. And those need to land on the correct page. And there is a need to preserve the history of the item according to the license of Wikidata. Mbch331 (talk) 15:56, 8 August 2015 (UTC)
The license https://creativecommons.org/publicdomain/zero/1.0/ does not require to keep track of the history, however it is useful to understand the changes made in data structure and to undo changes or merges if necessary. It also helps to identify possible vandalism, possible vandalism then can be traced to the exact edit an is not charged on the one who merges the items without seening the errors.--Giftzwerg 88 (talk) 11:21, 9 August 2015 (UTC)

Mae (Q2684993)

Hello, I tried to add the German label "Mae" to Mae (Q2684993) and failed. The error report said that MAE (Q355740) was already using the same label. The first thing that bothers me with this is that multiple items should be allowed to have the same label. The second thing is that in this case, the software is wrong, because MAE and Mae aren't the same. The software should at least be case-sensitive. Jonathan Groß (talk) 11:58, 8 August 2015 (UTC)

Case-sensitive and case-insensitive is more for Wikidata:Contact the development team. @Lea Lacroix (WMDE): could you look at this? Mbch331 (talk) 12:13, 8 August 2015 (UTC)
Jonathan Groß, Add a description as well. The only problem is if the label and the description both match those of another item. Joe Filceolaire (talk) 16:49, 8 August 2015 (UTC)
Just to know... (Problem on Upper/lower case) I don't like but developers decided :( --ValterVB (talk) 07:53, 9 August 2015 (UTC)
The labels and descriptions is to discern items with the same label in a search. The search function is not case sensitve.--Giftzwerg 88 (talk) 11:42, 9 August 2015 (UTC)

List with Q numbers

Hello. Is there a way to have a list with Q numbers of all articles of a wikipedia category? Xaris333 (talk) 22:10, 8 August 2015 (UTC)

Yes. There are two ways to look at it. Are you interested what should be listed in a category or do you want to know what is in the category. Issues are in the first you will get all items never mind in what Wikipedia they are. The second requires that all articles have an item. Thanks, GerardM (talk) 22:20, 8 August 2015 (UTC)

I want to find all Q numbers for the articles in en:Category:Cypriot Cup seasons. If an article doesn't have an item, then it will not be listed. I just find a way to do that with [5]. But, is there a way to have the list sort by articles name in Wikipedeia (or by label in wikidata) and not by Q number? Xaris333 (talk) 22:48, 8 August 2015 (UTC)

I don't think autolist does sort by Q-number - this is the order it uses. It doesn't seem to be sorted by enwiki article age, either. Andrew Gray (talk) 00:05, 9 August 2015 (UTC)

Yes. But you can easily sort the list by Q number using word for example. Xaris333 (talk) 00:32, 9 August 2015 (UTC)

@Xaris333:
in fact, Autolist sorts by Q numbers, on alphabetical basis (and not numerical) ;)
this tool will allow you to see all items in a category that are not yet on wikidata (and create them if you like). It is very useful to check if a category is complete on wikidata :) --Hsarrazin (talk) 11:43, 9 August 2015 (UTC)
What do you mean with "Autolist sorts by Q numbers, on alphabetical basis"? Xaris333 (talk) 14:16, 9 August 2015 (UTC)
I mean that the sorting is made using figures (like letters), not considering the total numerical value. Qids are considered as words not as numbers, like this :
  • Q1
  • Q10
  • Q100
  • Q101
  • Q1022
  • Q105
  • Q15, etc.
every Qids beginning with 1, the 2, then 3, etc. :)
This is generally what you get in, in most apps, when you sort numbers with a letter before… :) --Hsarrazin (talk) 22:26, 12 August 2015 (UTC)

Links are gone

Claims for VIAF, BnF, and LC Auth used to link to the sites, as recently as three days ago. Now the external links do not work. How do I reactivate these links? --EncycloPetey (talk) 00:54, 9 August 2015 (UTC)

Do you have items where it no longer works? Mbch331 (talk) 05:48, 9 August 2015 (UTC)
Works for me, e.g. Q2001. --Stryn (talk) 08:57, 9 August 2015 (UTC)
I've had the problem for 2 days. — generally, refreshing the page fixes it.
Seems the AC gadget is sometimes not very up ;) --Hsarrazin (talk) 11:45, 9 August 2015 (UTC)
Example: Q934597
Until two days ago, refreshing would correct the problem for me (if it happened), but now refreshing does not. The VIAF, BnF, etc., are displayed only as black, unlinked text for me. --EncycloPetey (talk) 13:21, 9 August 2015 (UTC)
They turn up as links to me. Mbch331 (talk) 13:31, 9 August 2015 (UTC)
I've just noticed that links to Commons categories aren't working for me either, although they work from other projects. Likewise, the VIAF (etc) links work from Wikipedia and Wikisource templates that use Wikidata information, but they are not working here. --EncycloPetey (talk) 14:09, 9 August 2015 (UTC)
Further information: The problem seems to be specific to Firefox. Using Safari displays the links. Is anyone else using Firefox, who could let me know if they have the same problem? --EncycloPetey (talk) 03:05, 10 August 2015 (UTC)
I use Firefox and don't have any problems. Mbch331 (talk) 06:44, 10 August 2015 (UTC)
MacOS or PC? It's odd that I have this problem in Firefox (but did not 2 days ago), but do not have this problem in Safari. --EncycloPetey (talk) 15:09, 10 August 2015 (UTC)
PC with Windows 7. Mbch331 (talk) 15:40, 10 August 2015 (UTC)

Other sites

The links to other Wikiprojects has "Other sites". I am only aware of adding "commons" for the link to Wikicommons, but what other values can be added? Again, why doesn't Wikicommons just have an entry like all the other interwikilinks ... --Richard Arthur Norton (1958- ) (talk) 02:22, 9 August 2015 (UTC)

There are a handful of oddballs that aren't yet integrated, and then some which there are only one-of (i.e. no other languages but the one wiki's). Commons is one. Wikidata is a second one. Wikispecies is a third. Meta wiki is probably a fourth. --Izno (talk) 03:41, 9 August 2015 (UTC)
@Richard Arthur Norton (1958- ): Currectly only Commons and Wikidata are supported.--GZWDer (talk) 04:30, 9 August 2015 (UTC)
The sections are there to group all the different language links. Commons doesn't have its own section because there is nothing to group together. - Nikki (talk) 07:12, 9 August 2015 (UTC)
Is it a idea to show all possible sites as suggestions? Sjoerd de Bruin (talk) 08:13, 9 August 2015 (UTC)

VIAF at Wikipedia

When I add a VIAF and LCCN number at Wikipedia to an old article, do I have to add it here too, or will a bot migrate the data over for me. --Richard Arthur Norton (1958- ) (talk) 04:36, 9 August 2015 (UTC)

I think it depends on the project, if it has active bots related to it and if there are any conflicts or not related to the VIAF and to the page. -- Innocent bystander (talk) 06:09, 9 August 2015 (UTC)
I would say, add at least the VIAF value to wd item, or better : add it directly to wikidata : normally the AC template should display it on your project ;) --Hsarrazin (talk) 11:48, 9 August 2015 (UTC)

How to indicate doubt when a comparison with an external source does not match

Wikidata is a project that is essentially curated by people. We have added large amounts of data and depending on all kinds of variables a given percentage of our data is wrong. That is not that much of a problem when there are good ways of handling these errors. There is also a given percentage of our data that does not match other external sources. This is different in that we do not KNOW if Wikidata or the external source is correct. It does not matter really if Wikidata has a source associated with a statement; sources are known to be in error as well.

My proposal is that we can flag statements that are in doubt. To keep it simple, I want us to concentrate on differences with sources. Sources as in online databases. The idea is that we can always read a source for comparison purposes. When there is a difference, we add a status and a reference to the external source. This is needed to create a workflow. When our data is wrong, it can be amended and the comparison for the record can be repeated. The status for that statement does not go to blank, it indicates that the value of the statement was changed. The least this does is add confidence in the resulting quality.

When the external source is wrong, there is a different status. This status requires a source for a statement because just stating "we are right" is NOT adding confidence to our quality. When both Wikidata and the external data are wrong, the statement is changed AND a source is provided for the new value for the statement.

The reason to concentrate on comparison with databases is because of the numbers involved. An external source easily has millions of records and a percentage will differ. The result is thousands and thousands of differing values. We want attention to these statements by our community and we therefore need a workflow. Having qualifiers for quality is easy and obvious. We do not need to know when everything is fine; that is implicit because we always show the identifiers to external sources.

When there is an agreement for mutual support with the external source, we can because of this mechanism automate quality control things in both ways. When one source (external of Wikidata) has info the other does NOT have, we can add the missing data and use the other source as a reference. As Wikidata has a workflow for differences, it may be that some mechanisms to improve quality can be automated. For instance, when an item is used on a Wikipedia, it may be flagged there as well that attention on underlying data is requested.

What do you think. Does this help us improve quality?

Thanks, GerardM (talk) 08:17, 9 August 2015 (UTC)

As is my habit, I blogged about things. GerardM (talk) 09:00, 9 August 2015 (UTC)
@GerardM: How can Wikidata be correct ? Wikidat can't be correct because WD is not a source so WD can never be correct.
Before starting a complex system we have to analyze the different cases:
  1. Difference between a WD statement without source and an external DB/source. → Delete the current statement and replace the value with the one from the external DB/source.
  2. Difference between a WD statement with wikipedia as source and an external DB/source. → The best is to look for the source details in wikipedia and to check the source. The minimal way is to delete the current statement and replace the value with the one from the external DB/source.
  3. Difference between a WD statement with an external reference and another external DB/source. → Both values should be added as statement with their sources informations, according to the principle that WD can't decide what is the truth.
If only one value can exist, just use the constraints system to detect the statements with several values. When several values can exist there is no way for WD to decide which is correct. The only way to distinguidh between 2 different values is to use the ranking. So the only tool we need is a tool which spots statements with several values and no "preferred" ranking. But even with that system we still can have statements without preferred ranking to due impossibility to choose a value: this is the work of the data user to create filter from its side using its own ranking of sources (or other filters like source languages) to select the data he wants.
This will be a completely new way to work for wikipedia and discussions will necessary to define startegy to choose one value when several are available. Snipre (talk) 21:00, 9 August 2015 (UTC)
This is a real problem for some works of literature. I have been adding VIAF for ancient Greek drama, and there are often duplicate VIAF identifiers because the VIAF database has redundant data. There is no "preferred" value because both values are correct, each for a different set of national systems. Sometimes there are even three different VIAF identifiers, and all are correct. --EncycloPetey (talk) 22:32, 9 August 2015 (UTC)
as regards VIAF, all values are correct (they exist on VIAF and are linked to specific national library databases)... until those are clustered. Some VIAF ID will then disappear, which requires to periodically check the VIAF values (because VIAF is nearly as much beta as wikidata and has not finished to merge the different files).
When there are more than 1 VIAF record for the same author, I generally consider the one with ISNI as the master one, but ISNI keeps records for all different identities (pen-names), so there may be multiple ISNI too ;)
As VIAF regularly improves their data, and cleans them up, and adds new IDs, would it be possible to have a periodic automated clean-up, that would check the existence of VIAF-IDs, and import new linked-IDs (GND, BNF, etc.) from the known VIAF ? it would be quite useful :)
I agree with you Snipre, wikidata cannot be correct in the sense wd is not a source, but it can be correct in the sense all statements are sourced, and if not sourced, checked against sources.
I understand GerardM's question as a need for a tool that could automatically retrieve statements that require checking (unsourced) and also, perhaps to compare with recognized sources (like VIAF for ex.) and mark for manual checking those with different values.
GerardM, can you confirm that's what you meant ? if so, yes it would be useful to have such a tool.
Also, somebody mentioned, a few days ago, the possibility to archive online consulted pages (like frwiki does), so that if the page disappears (404), we could still consult it as it was on the retrieved (P813). If technically possible, that would be great… --Hsarrazin (talk) 23:55, 9 August 2015 (UTC)
There is a difference between "correct" and "authorative". When VIAF has muliple records for the same person, it is correct. This is not to be denied. When a statement is unsourced, it is not sourced because another resource says the same thing. That is a fallacy. What I am on about is that there are DIFFERENCES. When facts do not match, it is something we want to take action on. Every source has its errors so we should never assume that one or the other is right. This is where research, sources make a difference.
Wikidata is correct in whatever percentage of all cases. The same is true for any other resource. Wikidata is not the definitive resource for anything and the same is true for any other resource. Differences need research, effective action on differences is the only way in which we improve quality. Everything else is window dressing. Having sources for all statements is a noble thing to aim for. It is not realistic that we will have that situation anytime soon. Working on the differences is an attainable goal and it will help us and the other resources to improve quality in a meaningful way. Thanks, GerardM (talk) 09:13, 10 August 2015 (UTC)
@GerardM: Again, can you clarify about which difference you are speaking because as I mentioned above there are several cases and only one need a special treatment. Then you have to distinguish between cases where several values can exist for a single property depending on precision, date, methods,...
If you want to focus on identifiers, just use the constraint system with its reports. No need of flag for that. And for me the flag system is a bad idea: if you see a problem, solve it or leave it, don't start to flag and to let others people solving the problem. Snipre (talk) 11:21, 10 August 2015 (UTC)
@Snipre: When we prefer only one value and reality has it that multiple values are possible, contstraints are artificial and wrong. When a bot finds thousands of errors, flagging them makes sense. It invites people to participate and, we can measure how well we are doing in improving quality. Flagging is of particular value for bots, they can easily find thousands of issues. They cannot resolve it. So your argument does not wash. When a problem is to be solved, it needs to be known as problematic.
The point is to improve quality. To do that in a practical way, we need to think about what is practical and forget about current dogma. GerardM (talk) 13:29, 10 August 2015 (UTC)
Sorry but you don't answer my question. Can you clarify about which cases you speak about ? I agree with you about quality improvement and I will support any action helping to increase data quality. But I don't see a lot of cases where your flag system can help:
* the case of identifiers is not concerned by your flag system because identifiers by definition should be unique or at least single. This can be managed by the current constraint system. So we have already the tool to correct that.
* the case of values which are not sourced on WD is not concerned too because WD is not a reference: if there is a difference between an unsourced data on WD and a value from an external DB, no need of deep reflexion or long search, just delete the value coming from nowhere by a value perhaps wrong but sourced.
With these two cases we cover most of the current data in WD.
Your flag system is relevant only for one case, the case where two values exist with some source data but are different. And here even if an check is perhaps necessary to insure that data were correctly imported it is really difficult to say what is correct or not unless you find a source which clearly defined that one of your values is wrong.
I don't know if you understand but take this two examples and explain how your system will improve quality:
* One statement provides several values for the population of a city.
* One statement gives three boiling temperatures using different determination methods.
If you don't flag these statements, please provide an example where your system is relevant. Snipre (talk) 14:20, 10 August 2015 (UTC)
This is about any type of value where Wikidata does not concur with data in another source. It seems obvious that you miss that point. Wikidata may be right, the other source may be right. We do not know, what we care about is flagging a statement where there is a disparity. We should never assume anything here. This is the point where quality makes a difference. Once we have researched one statement, we will provide a genuine source to reflect on what is good. Assumption is the mother of all fuck-ups.
As VIAF has potentially multiple records for one instance, our software should allow for this. It is factually correct when we identify this. It is for VIAF to remedy the situation. It is for us to associate all VIAF identifiers with our identifiers.
The idea that we should remove statements because we have an issue is imho "not good". Research is needed. When we find that OUR value is wrong, it is soon enough to remove it. Thanks, GerardM (talk) 16:31, 10 August 2015 (UTC)
Again how can you assume that values in Wikidata can be correct ? You assume that Wikidata values can be correct and this is in contradiction with your point of view "We should never assume anything here".
And I find strange that you never distinguish between value with source from value without source. How can you speak about quality without mentioning these different cases ?
If you want to speak about quality in a DB, please first define whats is data quality. Please analyze the differences between values with sources, values from WP, values without sources in terms of quality in order to provide practical elements.
Reality is simple: data quality is decreasing acoording to that ordre values with sources > values from WP > values without sources. This order is based on one point of view, the reliability. Is reliable a data which can be checked at any time, meaning a value hith all necessary details to find its origin and by the way the possibility to compare the value with the original value from its original document.
Your idea is good but you have to leave the conceptual design and to define the details of your idea. Snipre (talk) 21:20, 11 August 2015 (UTC)

„Wikidata is a project that is essentially curated by people.“ this is a lie, GerardM. Allmost all contributions are added without a reference. --Succu (talk) 22:07, 11 August 2015 (UTC)

We can label a statement as 'preferred rank', 'normal rank', or 'deprecated rank'. We need to be able to add a 'dubious rank'. This is not for every statement without a reference. It is for statements which bots have an actual reason to believe may not be true - usually because a statement has two different values (imported from different sources) where constraints say it should have one unique value. There may be a case for statements about ethnicity and sexual orientation which don't have references as these should always have references.
The 'dubious rank' can be cleared by any editor but will be reset by the bot unless all but one of the values is deleted or marked 'deprecated rank' or an exception to the constraints is recorded (for unique values) or references are added (not just 'imported from'). Anyone agree? If you agree then I will create a bug on maniphest for this. Joe Filceolaire (talk) 22:44, 11 August 2015 (UTC)
Symbol support vote.svg Support Popcorndude (talk) 23:24, 11 August 2015 (UTC)
Contrary to @Snipre: above, I would strongly suggest not routinely removing an unsourced value without further investigation just because another different value is sourced.
Over the last several months, I have been using the 'search' function on Mix'n'match to go through painters that have a match to the Your Paintings database, and looking to see whether Mix'n'match can find corresponding matches in ULAN, RKD Artists, various biographical dictionaries, Library of Congress, British Museum etc. As part of the process, Mix'n'match shows me the dates of birth and death for any items on Wikidata that have been matched, as well as corresponding dates where available from other databases with potential name matches.
Having done several thousand of these searches now, I have a lot of respect for eg RKD artists. But the Dutch database is not perfect, and there is a significant percentage of cases where it has a date which appears to be wrong, sometimes very wrong, and does not match what is given in eg a detailed biography from the gold-standard UK Dictionary of National Biography. (Sometimes also the headline date turns out to be only one of several potential dates given in the full text, or turns out to be an activity-range date, rather than an actual death date). Whilst probably not the majority case, nevertheless it is not that uncommon for the existing Wikidata date to be more accurate, even if it is unsourced.
Where there are conflicting values for the same property, my strong recommendation would therefore be to leave both on the item, as a signal that further investigation is appropriate, rather than to automatically delete one or the other without investigation, merely to suppress the contradiction.
If we care about data quality, we should be trying to preserve data from as many sources as possible, including unsourced Wikipedias, to maximise the chance of potential errors being revealed. Which is why, incidentally, we should also be trying to assimilate as much of Freebase as we can as soon as we can, as the best way to try to reveal asquickly as we can more statements that may be uncertain or conflicted. Jheald (talk) 22:43, 11 August 2015 (UTC)
Symbol support vote.svg Support Popcorndude (talk) 23:24, 11 August 2015 (UTC)
@Jheald: What I don't understand is the need to let an unsourced data when we have another source which can be different: the quality of the second value will be always better than the first one because it is more complete and reliable.
So my question is always unanswered: what is data quality ? If we agree that WD doesn't look for truth, the only quality criterion is sourced data and sourced data from sources considered as authority in their field. If you have other definition for quality please explain it because here is the problem.
If you want to spot unsourced data as candidate for better investigation you can tag 80-90% of the current statements because they have no source so your tag is useless because not specific. A tag is necessary to spot some items among a large groups, if most of the group is tagged, the tag is not necessary.
My reasoning is very simple: why do you want to spend so much on a small number of statements with a source when a so large number of statements don't have any source ? What is the purpose to have 5% of data with very good quality after several cross-checking and 95% of data with no quality because they have no source ? Snipre (talk) 08:55, 12 August 2015 (UTC)
@Snipre: What your analysis ignores is the possibility of further investigation. We don't have to just make a choice just on the basis of the values we have -- there is also the option of further research. Preserving dissonant values helps indicate when that option might be a good idea -- and therefore makes for a higher quality database.
I think we do aspire to more nearly approach truth, when we talk about quality. "Quality" is not just a box-ticking exercise of whether statements have sources. As I've written above, from practical experience, even good sources can be flawed, and/or reflect incomplete scholarship.
I don't have a position on the tag proposal. But what I do object to is the view that just because a value has been imported from Wikipedia or from Freebase it should be ignored. That fails to recognise how much of Wikipedia and Freebase is of high accuracy and practical value - and how valuable even as flawed sources they can be, for corroboration and/or to suggest potential errors (or at least uncertainty) in sourced statements, and in the building of a basic web of information that is then much more open to checking for consistency. Jheald (talk) 11:53, 12 August 2015 (UTC)
I agree with Jheald,
many unsourced wd statements were made from wp articles that are sourced, but the source was not transferred... too many items to ad claims to, to few people, tools that allow to see the info, but not to add the source when editing.
yes, Magnus Manske, that's your brilliant tools I speak about (Game, WDuseful, WDFIS, Autolist, etc.), that do not allow to auto-retrieve the name of the wp project the info is collected from, which makes it very painful to source, afterwards…
but in most cases, and with the notable exception of category-based Autolist claims, that can be very tricky (because of unstructured sub-cats), a very large part of claims have sources deeply buried in the wikilinks that were used to edit :/
I don't think erasing them because temporarily unsourced (they will finally be, once wp contributors will see the interest of adding infos to the wikidata items, with source :)
@Lydia Pintscher (WMDE): Also, I think, it would be very interesting to have access to the source of claims when retrieving info on wp templates (so that, in an Infobox, for instance, sourced infos would give access to the source, and unsourced info would be recognizable as such... italic, maybe ?) — Don't know if it is feasible, for now…  but I'm sure, it would be quite an incentive for wp contributors to transfer info AND sources on wd ;) --Hsarrazin (talk) 22:42, 12 August 2015 (UTC)
That's already possible with Lua. Can be made nicer still though. --Lydia Pintscher (WMDE) (talk) 07:38, 13 August 2015 (UTC)

Commons Categories

When adding a Commons Category to an item that links to Wikipedia articles, some people add it an a link in Other Sites and some people add it as a Commons category property statement. Just wondering which is best and if there is a report that shows the incorrectly added Commons Categories? -- WOSlinker (talk) 11:15, 9 August 2015 (UTC)

Officially a Commonscat should only be entered to the other sites section when the item is an item about a cat. In other cases it should be added as a statement. However not everyone follows those rules. Don't think there is a report to show incorrectly added commonscats. In my opinion it isn't really wrong when the rules isn't followed. Mbch331 (talk) 11:21, 9 August 2015 (UTC)
As regards the interwiki links, the rule is to link the category on Commons to the category on The Other Wikis. Where there doesn't exist a category on The Other Wikis, it's common but not correct to link the Commons to the item about the actual entity.

As for Commons category (P373), according to the property documentation, it's supposed to be added to the item about the entity and the item for the category for the entity.

However, given that arbitrary access has rolled out to even the largest wikis, I think we should move to deprecate/delete P373 actually. --Izno (talk) 16:31, 9 August 2015 (UTC)

Deprecate/delete? There are persons with a commonscat that have no cat on any Wikipedia. E.g. Category:Mark van Veen only exists on Commons. Not anywhere else. If we aren't allowed to link the commonscat in the other wiki's section, then how do we link him to the Commonscat (if P373 is deleted)? Mbch331 (talk) 16:50, 9 August 2015 (UTC)
It's the other way around. With arbitrary access we have a lot less restrictions. A template can just do Haarlem (Q9920) -> topic's main category (P910) -> Category:Haarlem (Q7427769) -> sitelink to Commons:Category:Haarlem and if that's not there, fallback to Commons category (P373). If the templates are implemented properly, all cross namespace links can be eliminated without losing a single link to Commons on Wikipedia. Multichill (talk) 16:24, 11 August 2015 (UTC)
We don't actually need the fallback to 373 regardless. What I'm suggesting is that if an item doesn't have a topic's main cat or if the topic's main cat item doesn't have the sitelink, then the template should look for a sitelink on the current item. So the fallback is actually to the 'local' item's Commons interwiki link. (Order of logic could be reversed as desired i.e. look for local sitelink first, then go look for sitelink at topic's main category.) --Izno (talk) 16:45, 11 August 2015 (UTC)
There are thousands and thousands of categories like this. Many people have Commons categories, for instance, to group their associated media, but there would not be a single-person category on a Wikipedia project. What I am hearing is that a secondary way to indicate Commons category (P373) is to actually create a wikidata item for each category on Commons that does not have a corresponding category anywhere else. Then use topic's main category (P910) to link to that. Right? In the mean time please don't delete Commons category (P373), as there's lots of useful info stored in there in lots of Wikidata items. Sweet kate (talk) 21:11, 11 August 2015 (UTC)
Please can everybody go on filling in Commons category (P373) wherever appropriate. For one thing, the script c:User:TheDJ/wdcat.js depends on it, that adds a link to Reasonator whenever a Commons category page is displayed that has a P373 pointing to it.
Also, as I understand it, Wikidata:Notability currently recommends not to create a category item on Wikidata, if the only cat it would link to is on Commons. Jheald (talk) 22:10, 11 August 2015 (UTC)

I haven't looked at WD:N in a while, though that seems to be the practice for items without a gallery and without a corresponding non-Commons category item.

Scripts can be modified. Tools serve us, not the other way around. --Izno (talk) 03:01, 12 August 2015 (UTC)

Yes, you could do it that way (create a lot of items--I actually support this though my point of view seems to be on the fringe), but I would suggest avoiding doing so unless the item in question did not also have a link to a gallery of the topic.

Yes, we don't delete anything without deprecating it first, usually by reintroducing the data in another form. That's just normal. --Izno (talk) 03:01, 12 August 2015 (UTC)

Wikidata weekly summary #170

difference between field of work (P101) and field of this occupation (P425)

Hello,

I don't see the difference between field of work (P101) and field of this occupation (P425).

Could you please explain ? --Hsarrazin (talk) 19:10, 9 August 2015 (UTC)

In case of a profession use field of this occupation (P425) in all other cases use field of work (P101). Mbch331 (talk) 19:19, 9 August 2015 (UTC)
arf, I think I got it wrong on a bunch of items today… the property auto-selection proposed field of work (P101), and I did not see field of this occupation (P425). Probably a label-translation problem :/ --Hsarrazin (talk) 00:05, 10 August 2015 (UTC)
The difference between work and a profession is not clear. Is there a list of things that are professions as opposed to just work? Is a ditch digger a worker and a plumber a profession. Is it a matter of one being a member of a union? If the workers are organized into a union does that make it a profession? Does a profession require a license issued by a governing body? Does a profession require specialized training that must be completed before employment? --Richard Arthur Norton (1958- ) (talk) 01:57, 11 August 2015 (UTC)
In my opinion there is no clear criterion which can be used to decide what is a profession and what is not. A ditch-digger is not; a plumber probably is not; an American licensed professional engineer in the civil engineering field who specializes in hydraulic engineering is a professional. But in the UK I understand the word "engineer" is often used to describe a repair person, who would usually not be considered a professional. Jc3s5h (talk) 13:02, 11 August 2015 (UTC)
The difference is very simple. field of work (P101) aims at linking human beings (and more specifically researchers, teachers, lecturers and professors) to their field of work whereas field of this occupation (P425) aims at linking a profession to the corresponding field of expertise. Let's take the example of Mr. X, a plumber.
Mr. X > field of work (P101) > 'Plumbing'
or Mr. X > occupation (P106) > 'Plumber' and 'Plumber' > field of this occupation (P425) > 'Plumbing'.
--Casper Tinan (talk) 14:04, 11 August 2015 (UTC)
Note that the practice on wikidata (not sure if this is documented anywhere) is that pretty much every human on wikidata should have a statement using occupation (P106). Note that 'occupation' is wider than 'job' or 'profession' as it can include things a person is known for which they don't get paid for (amature sportspersons etc.). field of this occupation (P425) or sport (P641) can then be used to make this more specific so we don't need a separate item for every single specialist occupation. Joe Filceolaire (talk) 22:29, 11 August 2015 (UTC)
there is a strong difficulty with the fact that the words don't have exactly the same meaning in different languages, so trying to define profession/craft/job/occupation too strictly would probably lead to different practices because of different languages.
as for Mr. X > occupation (P106) > 'Plumber' and 'Plumber' > field of this occupation (P425) > 'Plumbing', I would say this is totally redundant and has no interest at all.
my point was considering for example historians, and wanting to give their specialty (Medieval history, WW2, Roman empire, etc.) - I still don't know for sure wich property to use… --Hsarrazin (talk) 22:50, 12 August 2015 (UTC)
Hsarrazin that one, in my opinion, is definitely field of this occupation (P425). Joe Filceolaire (talk) 15:52, 13 August 2015 (UTC)
Thanks Joe Filceolaire. I believe my confusion could be caused by very similar translations ;)
I think I did the contrary on some items - will try to find them. --Hsarrazin (talk) 19:28, 13 August 2015 (UTC)
@Filceolaire: It's field of work (P101) which should be used to specify occupation (P106), not field of this occupation (P425). To clarify, another example :
Or in short, P101 and P106 are for persons, P425 for jobs/occupations/professions. --Pasleim (talk) 13:58, 15 August 2015 (UTC)

National symbols and its manifestations

Hello, today I met another case about (German) national symbols. National colours of Germany (Q2253357) links the German Wikipedia article to "Black red gold", a specific colour pattern (in use 1848, 1919-1933, 1949-), with the English Wikipedia article "National colours of Germany" (describing also the periods in German history when Black-white-red were in use).

Earlier, I recognized that we have WD items on National Anthem of Germany (Q4122341) (the national anthem of Germany) but also on Das Lied der Deutschen (Q44042) (the historical song on which the actual German anthem is based).

The problem is that sometimes a Wikipedia language version treats the national symbol as such (which can change during time), and sometimes with a specific manifestation of that symbol (which exists also before or after it served as national symbol).

How to deal with that? We can try to connect in WD only what really relates to each other, but actually is this a more general problem of Wikipedia. Z. (talk) 21:50, 9 August 2015 (UTC)

@Ziko: Each flag like each anthem has to have a specific item. Items like National colours of Germany (Q2253357) is more a compilation of information than an unique concept as the colours changed with the time. And the problem of National colours of Germany (Q2253357) is the concept of Germany which is a modern concept and is difficult to identify before 1900. Are the German empires part of this item ?
Most of the time you can create items for each possible national symbol and don't focus on items which are compilation because they aren't really clear about which topic they cover. Snipre (talk) 07:38, 10 August 2015 (UTC)
And every aspect of a relation is difficult to describe by simple statements. Sweden do not have any official national anthem, but there is one song de facto used as national anthem. Describing every aspect of this relation is better done in a WP-article than here on WD. -- Innocent bystander (talk) 08:31, 10 August 2015 (UTC)
So it seems to be primarily a problem of the English article that should have a different name or not be linked to its German counterpart. Should "Black red gold" get an WD item of its own? Z. (talk) 13:55, 10 August 2015 (UTC)
@Ziko: Yes. Snipre (talk) 06:56, 11 August 2015 (UTC)

Multilingual project !

WikiProject Names is not translated. It's although a featured Wiki Project ! Am I the only one to think we should take more care about the multilingual aspect of Wikidata ? author  TomT0m / talk page 10:12, 10 August 2015 (UTC)

You are probably not alone but...
The problem with these translations here is that translated pages makes it more difficult to find the updated original texts. The main page is maybe not the most important page here, but the version I see every time I get here, looks like it has not been updated for years. Sofixit, you may say, but I will not. No!, the time I have for this project, I prefer not to spend on translating meta-pages that will very soon be outdated. And since the number of those here who speaks my language but not read English, is very low, the gain of such translations is extremly low. These translations also tends to create a nomenclature that are never used outside the meta-pages. -- Innocent bystander (talk) 16:38, 10 August 2015 (UTC)
@Innocent bystander: Maybe speakers of your language are like that, but it's not the case in france :) Although of course a lot of people, for example whose who works in computing, can read and for a lot of them write english, this is not the case for everyone. Worse : there is a cultural aversion against english, and some people react not well, they here "multilingual project = english project" which triggers old cultural best enemy reactions :) So when advocating for Wikidata we have at least to show them proofs of goodwill, reassure them that no, it's not because the discussion are in english that the project is totally under anglo-saxon culture and that all culture are took into account. The french "Bistro" is an active place, and some people confirms they are lost in the help pages in the first place (I think we're not good at it) and that the fact everything is in english does not help. That's not even mentioning whose who don't even want to quit frwiki. But no one is asking you to translate if you don't want to :) I think the nomenclature problem is a false one, if really the terms used in the glossaries are not used then we should change the glossaries, but maybe using of the wrong term is a sign the notions behind are not well understood ... in that case what is needed is an explanation, we need to stay strict on this for having a good basis for intercultural and language communication. I tend to add sometime the english original term in my fr translations in glossaries, or Wikidata items (even better). As Wikidata data model has been thought a lot, and is proving to be useful, it would be a shame we don't spread its correct use because of bad understanding and approximative knowledge of it. author  TomT0m / talk page 17:01, 10 August 2015 (UTC)
You are of course welcome to translate to francaise, but the bot who came to svwiki and encouraged people to translate Wikidata-pages to Swedish caused us more problem than it gained. People not active in the WD-project was the most interested in translating, while the active WD-users had their discussions in English. When a Swedish glossary was introduced, it mainly caused us trouble, since no discussions are made in Swedish. And if you wonder this is about a translation problem about archeology, not about Wikidata. Here, we mainly embrace anglo-saxon culture, while we can have some stigma against anglo-american culture. But that does not affect computer science. -- Innocent bystander (talk) 17:25, 10 August 2015 (UTC)
@Innocent bystander: Yes, when new people come to a project, they can make mistakes :) But whose to blame, whose who knows and did not make the translation effort, or whose who are a little too much enthousiast ? /o\ Seriously, being multilingual is in the root DNA of this project, my opinion is that took that as an opportunity to understand how Wikidata can help multinlanguages project, starting by himself :) Just using items for concepts as terms is great ... author  TomT0m / talk page 17:47, 10 August 2015 (UTC)
My two cents as an American and an anglophone (who is active on a few other-language Wikipedias too, including fr and he):
For better or for worse what we really need to do, I'd wager, is somewhere in between the two extremes. I can affirm that WD:Bistro is a very active place. And we all know that there's this little cultural rivalry between francophones and anglophones ... At the same time, WD:מזנון, the Hebrew equivalent of this page or Bistro, has had exactly one (non-automated) posting since November 2013. The overwhelming majority of native Hebrew speakers, no doubt, do their Wikidata business mainly on this page.
So I suspect that there are a handful of languages where a full effort should be made to translate project pages as completely as possible. French and German are certainly two. I could make a pretty well-educated guess on a few more, but don't want to start a heated discussion on that now; ultimately, if anyone really feels her/his language should be included in that handful of languages, no one will stop him/her from working on it.
However, I think that for most other languages, we should perhaps think of a much more limited translation effort. I'm thinking of things like the following:
  • The most important help pages, or at least an embassy where people needing help in a given language can find help from volunteers.
  • Templates people are likely to see or want on user pages. (Extra benefit: some of that effort can be ported over to Meta-Wiki for use in templates on transcludable global user pages.)
  • Maybe a list of languages where more complete translation efforts were undertaken. What I mean is: Someone who is a native Swedish speaker comes looking for help. Help is limited in Swedish. But the individual's German is far stronger than her/his English. So let's make sure that person knows that a full set of help is available not only in English but also in German.
  • Clear instructions as to how to set the default interface language for this wiki, whether to the native language or to the "core" language the user is most comfortable with.
I fear being taken as a cultural imperialist in saying this, but I'm really trying not to be. I promise. I'm trying to reflect what Innocent bystander said above: Where limited translation resources are available to put into the project, what is the most efficient way to make use of those resources? If someone has a better idea, or can better focus my suggestions, I would more than welcome that.
PS to TomT0m: Ma nouvelle belle-fille est Lyonnaise. Je dois commencer à pratiquer mon français, tout de suite. StevenJ81 (talk) 18:25, 10 August 2015 (UTC)
Héhé :) Tu es le bienvenu sur le bistro :)
Ça m'a aidé, Tom. Vraiment! Quand j'ai appris à parler français, il n'y avait pas de Internet. (Je pense que DARPAnet existait. Je pense. Mais pas pour moi.) Donc je ne connais pas: "sur" le bistro? "à" WhatsApp? "en" Facebook? Qu'est-ce qui sont les prépositions correctes? (;-) (Tu peux répondre chez moi.) StevenJ81 (talk) 14:37, 11 August 2015 (UTC)
I'd say the first priority is to make pages translatable. That's the technical step. After that, translating something can be very easy, it's a few clicks away to anybody (who speaks english and another language) :) author  TomT0m / talk page 18:44, 10 August 2015 (UTC)
Also I want to mention, that current translation instrument has a number of usability issues. As the result I often add translation to {{LangSwitch}}, but rarely to <translate>. — Ivan A. Krestinin (talk) 19:38, 10 August 2015 (UTC)

Wikipedia entry to Wikidata entry

When does a new entry at Wikipedia get an automated entry at Wikidata? --Richard Arthur Norton (1958- ) (talk) 01:50, 11 August 2015 (UTC)

It doesn't. You can create one yourself or wait for another user (or bot) to do it. What article are we talking about, exactly? Jonathan Groß (talk) 05:49, 11 August 2015 (UTC)
@Richard Arthur Norton (1958- ): Sometimes Dexbot will create it. You can also create items from a list of pages via https://tools.wmflabs.org/wikidata-todo/creator.html --GZWDer (talk) 06:00, 11 August 2015 (UTC)
I create a new biography at Wikipedia at least once a week based on images from the Bain collection at the Library of Congress. See wikipedia:Olive Kline, as one of the newest. I always assumed a bot would assign a Wikidata number to the entry. --Richard Arthur Norton (1958- ) (talk) 18:08, 11 August 2015 (UTC)
We sorely need a tool that takes the URL of an en.WP biography, reads data from the infobox (probably making use of microformat markup; I'm happy to advise) and categories, and starts a Wikidata site. Obviously, it could be re-purposed for other languages and other topics (perhaps by reading from a config file), but that use case alone would justify some effort. When's the next hackathon? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:33, 11 August 2015 (UTC)
well, one of our Russian contributors, Vlsergey I believe, has made quite an effort in this sense, but, unhappily, it doesn't seem to work anymore, or, more exactly, all data are retrieved and organized, but the updating of wikidata fails in the last step. If something could be done to fix it, that would be very helpful.
the scripts can be found on Books Project, since they were aimed as aid for bibliography. But the "Persons" script suits almost any human (Q5). --Hsarrazin (talk) 23:06, 12 August 2015 (UTC)
@Hsarrazin: it's a different purpose script -- it does not take current infobox into account. But i've updated URL's, the main one is still supported. In addition I do have a bot code to find dicrepancies between article infobox and wikidata data, so i can create a list of "proposed changes" (discrepancies and missing data list). But those change can't be applied because my bot is blocked on Wikidata. -- Vlsergey (talk) 07:52, 13 August 2015 (UTC)
@Vlsergey:, sorry, I made a small confusion :) - well, if the infobox is fed from wikidata, it does take it into account, but there are no discrepancies.
The main purpose is to edit wikidata without going in wikidata, isn't it ? Thanks, anyway, this is a major first step to cross-wiki edition :) --Hsarrazin (talk) 08:29, 13 August 2015 (UTC)

Collect all discussion about units

Hello. I started Wikidata:Units because I wanted there to be some place where I could read about how units work in Wikidata. Is there some other central place for discussing this? If not, please link to discussions on the topic there.

I think this page should be friendly for visitors from outside Wikidata because this is a topic of broad interest. Blue Rasberry (talk) 15:47, 11 August 2015 (UTC)

@Bluerasberry: There is nothing to say about units because ... the datatype is not implemented. The page is a false good idea because it is just too much specific and a help page will be more convenient.
Unit is only a small part of the problem. We have to include uncertainty and different questions about scientific notation so better create a help with "Numeric datatype" as title. All future information about this datatype will added in an unique place.
Currently the discussions about units are here. Snipre (talk) 13:41, 12 August 2015 (UTC)
In case other places where the topic of numeric datatype is discussed:
* Official dataModel
* Help:Datatype
Snipre (talk) 14:23, 12 August 2015 (UTC)
@Snipre: Thanks, I added what you shared.
Dead ends are useful too. Wikimedia projects have an aversion to saying when information does not exist, or when people cannot do things. This usually works, but in some cases, a lot of people go looking for the same thing and need to land on a page which tells them to not look anymore because nothing is there to find.
Wikidata:WikiProject Medicine wants a lot of drug values inserted with units and Wikidata:Menu Challenge wants recipes here and it would be nice to have an indication which says "these things are not possible now". Saying this somewhere will help visitors to Wikidata know that they should hold their interest and be ready to comment when the project develops. I would like for there to be a place where people from other Wikimedia projects (like Wikipedias) can comment if they like. Phabricator is intimidating - it is mostly for developers and not for Wikimedia readers or content creators. If it is not trouble, I would like for this page (it could be renamed) to be somewhere here. Blue Rasberry (talk) 16:42, 12 August 2015 (UTC)

Template imports

Who can help with template imports? My request at Wikidata:Administrators' noticeboard#Template imports, made on 30 July, is unanswered. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:18, 11 August 2015 (UTC)

What to do where things are not clearly within notability

What is the recommendation to do with items such as William Hayes (Q18603380)? It has no wikilink which pushes it outside of Wikidata:Notability, and depending on one's point of view of mushing it may or may not have general notability. If it is defined as notable, and not requiring wikilinks, then the ability to know that it falls within acceptance and have that notated on the page itself would be useful (how? totally unsure, just know that it would be helpful).  — billinghurst sDrewth 01:27, 12 August 2015 (UTC)

Criterion 3: "It fulfills some structural need, for example: it is needed to make statements made in other items more useful." William Hayes (Q18603380) seems to pass. --Yair rand (talk) 02:19, 12 August 2015 (UTC)
this object is linked from 1983 Iditarod (Q18577442), thus shall remains. -- Vlsergey (talk) 07:25, 12 August 2015 (UTC)
@Harmonia Amanda: nice overview, must have been quite a bit of work. Maybe you can tell us a bit about the background? Multichill (talk) 07:29, 12 August 2015 (UTC)
Here we have a problem according to the need to put all participants of a race just because we want to have the ranking of the race. Just think about the New York City Marathon with 50'000 persons each year. Do we want to add all those persons in WD ? I think we have to put a limit like to add race ranking only for professional competitions. Snipre (talk) 09:00, 12 August 2015 (UTC)
In my opinion the item is clearly notable as it's used by another item. I have proposed a new version of notability policy, the discussion is continued at Wikidata:Administrators' noticeboard#Wikidata:Requests for comment/Wikidata:Notability overhaul.--GZWDer (talk) 09:21, 12 August 2015 (UTC)
I think WD:Use common sense applies there. - Nikki (talk) 09:27, 12 August 2015 (UTC)
For an overview: claim106:500097--- Jura 09:43, 12 August 2015 (UTC)
@@Snipre: and others probably because the new York City Marathon doesn't list all its participants on the official website in a comprehensive list (you can search their database but the ranking isn't the primordial information), when all important sled dog races do it. And I started with the Iditarod (and the Yukon Quest) but it's not the only sled dog race William Hayes participated to. The item will be linked to more when I'll be working at those others races (right now I'm working to import information from some European races, not Alaskan ones). And I'm also working on a script to facilitate the creation of Wikipedia articles about each races and about the winners (for whom sources exists but it seems like I'm alone on liking sled dog races on WP). There is 1414 mushers on wikidata right now, and I created only those who are participating in a race I treated. Meaning, all items without a race had been created because they have a wikipedia link. I estimate when I'm finished with notable races, it'll be around 3000 mushers (which are the people the most known in this sport). Way less than the participants to one New York Marathon. And probably there will be the occasional musher who participated in only one great race and didn't do anything else, but right now, only with my working notes, I could create at least 800 stubs which would be notable on wikipedia (if I had the patience to do that only by hand). If you're interested you can always go on the Wikidata:WikiProject Sled dog racing. --Harmonia Amanda (talk) 13:18, 12 August 2015 (UTC)
@Harmonia Amanda: This was not a criticism of your work and I thought that you were starting with the most famous races so we only have the main mushers at the moment who are professional and if not who they are quite famous in their discipline so no problem of notability will occur.
I was thinking for the next steps, after the major races were added, when some people will start to add regional or local races with people without any notability in their sport. Snipre (talk) 13:31, 12 August 2015 (UTC)
@Snipre:. Wow you're optimistic. I'm the only one working on this and I only listed on my todo-list races which have a Wikipedia article. And very few mushers are "professional". Most of them have another work (ok, they're many veterinarians, handlers for others mushers, owners of kennels who are also doing some tourism tricks, etc.). Actually, I don't know any musher who are living only from the races, even the most famous ones. It's probably pretty different in others sports? --Harmonia Amanda (talk) 13:42, 12 August 2015 (UTC)
The advantage of sport like Sled dog racing is their focus on a limited number of fans. So most of the time participants are the same from one race to another one. But if you take sport like trail running or marathon we arrive in another dimension with some stars and a lot of amateurs. Here we have a problem considering ranking in WD. Just the example of the New York Marathon: there is a ranking, the data are online and even in an electronic format so someone with some programing skills can easily import all the ranking into WD using a bot and creating automatically thousands of new items.
I think we have to be careful and perhaps the corresponding projects should define some rules about which races can have their ranking on WD. Snipre (talk) 14:06, 12 August 2015 (UTC)
@Nikki: Common sense is nothing once people finished to edit on main subjects and start to work on subjects with lower importance. This is a well known trend in some WPs where current new articles are often about subject with a very small impact. What is common sense to judge about importance of regional or local race ? The number of participants ? The coverage in media ? Snipre (talk) 14:06, 12 August 2015 (UTC)
What is considered acceptable on wikidata in five years time (or even one years time) is likely to be wider than what s acceptable today - I won't be surprised if we hit a hundred million items in a few years. Just get your head down and attend to what needs doing to improve the items we have now and bear in mind that at some stage in the future we will have more. Joe Filceolaire (talk) 15:48, 13 August 2015 (UTC)

Wikidata needs your vote

Hey folks,

As you know Wikidata is one of the 100 winners of the "Germany - Land der Ideen" competition 2015. We now have the chance to win the Public’s Choice Award as well! At the moment, Wikidata is ranked 15th. We need to be among the first 10 to enter the next round and win the Public’s Choice Award. Voting is open until 23 August.

The information about the competition and projects is available in English (https://www.land-der-ideen.de/en/projects-germany/landmarks-land-ideas/competition-2015-urban-space-rural-space-cyberspace and https://www.land-der-ideen.de/en/projects-germany/landmarks-land-ideas/2015-award-recipients/category-science/wikidata) but the voting process only in German. But it’s pretty easy to vote anyway:

You can repeat this daily until 23 August. You have one vote every day.

Let's win this! :D

Cheers --Lydia Pintscher (WMDE) (talk) 12:46, 12 August 2015 (UTC)

Largest arbitrary access rollout done \o/

Hey folks :)

Just wanted to let you know that we've just finished rolling out arbitrary access to a large number of wikis. Complete list is at Wikidata:Arbitrary access. --Lydia Pintscher (WMDE) (talk) 13:39, 12 August 2015 (UTC)

Congrats! --- Jura 14:08, 12 August 2015 (UTC)
Yay \o/ --AmaryllisGardener talk 14:46, 12 August 2015 (UTC)
And the beast shall come forth surrounded by a roiling cloud of items. The house of the unbelievers shall be razed and they shall be scorched to the earth. Their properties shall blink until the end of days.
from The Book of Atadikiw, 08:15
author  TomT0m / talk page 14:55, 12 August 2015 (UTC)
Speaking for Ladino Wikipedia: Thank you. (I tested it in its most basic form, and it works.)
Now here's a question: What it would probably be nice to do with this is to be able to nest these calls. Thus:
Prime Minister = P6. Party of Prime Minister = P102|from whatever Q item is returned from P6. But it didn't seem to want to let me do that, unless I'm missing something. StevenJ81 (talk) 14:57, 12 August 2015 (UTC)
I believe that is only possible with Lua. --Lydia Pintscher (WMDE) (talk) 15:02, 12 August 2015 (UTC)
Thank you, Lydia. I think we have very limited Lua on ladwiki just yet, and I am hardly one who can change that readily. StevenJ81 (talk) 15:30, 12 August 2015 (UTC)

Should I merge these?

Could someone who speaks German help me figure out if these should be merged, and if not, what the second actually represents? Popcorndude (talk) 21:46, 12 August 2015 (UTC)

I wouldn't say it's about phonemes and they both have German Wikipedia pages, so you can't merge them anyway. "Laut" is "sound", it basically says that a "Laut" is a noise made by a human or animal voice and goes on to talk about the idea of speech sounds. The introduction is quite general, then the "Sprachlaute" section mentions phones, but there's another page for those, so it couldn't be merged with that either. - Nikki (talk) 21:59, 12 August 2015 (UTC)
Don't merge — phoneme (Q8183) is language-specific, while phone (Q202064) is cross-linguistic but still limited to speech, whereas no label (Q20080358) is broader and encompasses human vocalizations like vocable (Q7939174) as well as animal vocalizations, e.g. bird vocalization (Q1126556). --Daniel Mietchen (talk) 22:33, 12 August 2015 (UTC)
I changed a bit the classification of all of these (statements stays, discussions are burried :) ) Autodesc should be able to help anyone to understand how items are differents. Hope that helps. author  TomT0m / talk page 09:21, 13 August 2015 (UTC)

COI interwiki links

Hi, I'm only now getting into the meat of Wikidata, so forgive the basic question. This is in regards to the COI interwiki links at Q4663309#sitelinks-wikipedia. As Hebrew Wikipedia doesn't have a COI page per se, I'd like to link to the paid editing guidelines, but I get a message "The link hewiki:ויקיפדיה:עריכה בתשלום is already used by item Wikipedia:Paid editing (policy) (Q16186099). You may remove it from Wikipedia:Paid editing (policy) (Q16186099) if it does not belong there or merge the items if they are about the exact same topic." What's the accepted remedy for this? Thanks. -- Fuzheado (talk) 22:15, 12 August 2015 (UTC)\

The issue here stems from enwiki, it has both a conflict of interest policy and a paid editing policy. Because two exist, you're going to need to chose which one is more appropriate for hewiki. I don't know the slightest bit of Hebrew, so somebody who's more well-versed in it should make that call. If you decide that it belongs at the COI policy, then click "edit" in the links, remove it, and add it back in the other page. Kharkiv07 (T) 22:24, 12 August 2015 (UTC)
I suppose that someone in the grand design of interwiki in Wikidata, it was decided that only one exact page in en.wp corresponds with one in he.wp. But I'm not sure why this should always be the case. Why can't multiple pages in en.wp map onto the same he.wp page? As for this specific instance, it's a bit more complicated in that the page is, in fact, a failed proposal and not policy at all! But it would be useful for folks seeing it in en.wp to know that he.wp has a policy page. Similarly, it's useful to point from directly to the he.wp page on paid editing. I can't imagine I'm the first person to point out how many can map to one, or is the expected remedy to have all this redirected in one language edition, and have one and only one interwiki link to the other language? (I see now that I probably should have posted this in Wikidata:Interwiki_conflicts.) -- Fuzheado (talk) 22:38, 12 August 2015 (UTC)
That's probably where you should have posted, and it's a long story here, believe me. (I'm temporarily serving as an administrator at Ladino Wikipedia, and I have, for example, a page on Astrolojiya and a page on אסטרולוגיה. But only one of them can link ... )
There are two solutions. One is to go to as many of the pages on COI as you can and insert manual interwiki links. The other is to go to hewiki, create a COI page (?איך אומרים), and put a filler sentence or two in. Then go to wikidata and link that page up. Then go back to your hewiki conflict-of-interest page and convert it into a redirect to the paid-editing page. But you have to take that interim step of creating a page that is not a redirect page first, or you can't link it up through Wikidata. ...בהצלחה רבה StevenJ81 (talk) 02:38, 13 August 2015 (UTC)
I would add: The main issue driving this one-to-one rule (I think) is not so much the question of whether more than one page on enwiki can have a link to that specific page on hewiki. There are workarounds for that; they are not always ideal, but they work. The issue is more that there is one set of interwiki links on the hewiki page; where in enwiki (and elsewhere) is that set of links going to point? There are a lot of good reasons not to have multiple full sets of interwiki links in a single article. If the hewiki article appears in more than one Q-item, which set of links would appear on the hewiki page? You see? There are challenges, and it's almost impossible for all of them to have perfect solutions. StevenJ81 (talk) 17:53, 13 August 2015 (UTC)

Israel (Q2366978)

Should Israel Q2366978 have two new ones added, one for family name and one for given name? We generally break down any name into the two distinct types of names. Wikipedia has combined them into one entry. --Richard Arthur Norton (1958- ) (talk) 04:59, 13 August 2015 (UTC)

The ones for given name already exist as Izrael (Q20087961) and Israel (Q19819746). Unless the one for surname exists as well, a new item should be created and this item should never be linked as name but as metaitem to either types. See Wikidata:Wikiproject Names.
Since I consider having an item for every variant of every name a weird idea, I don't participate in devolping this system and I can't tell you, what should or shouldn't be created. Anyway, there are further forms of the name, like Jisrael, Jisroel, Yisrael, Yisra'el, Yisroel, Israél, Isráél, Yiśraʾel, Israël, İsrail and maybe more. This would be only the Latin script; There will be much more in Hebrew, Arabic, Cyrilic and other scripts.--Shlomo (talk) 13:48, 13 August 2015 (UTC)
And it seems that you have found an interwiki conflict as the hrwiki sitelink is about different thing than name. Matěj Suchánek (talk) 05:35, 13 August 2015 (UTC)

label filter

At this time it is not possible to use same label for different items. How to handle disambigation pages now? F.e. for making mores (Q1115221) in de correct, I had to add redundancy here (in look of description)... Thank you, Conny (talk) 09:49, 13 August 2015 (UTC).

This is an interwiki conflict before beeing a labelling problem. The item with a statement referring to "muslim law" is labelled "moeurs" in french and is in frwiki a disambig page. This need to be solved before good labels can be found :) And it will always be possible that two items have the same name in Wikidata as two different things can share a name and it's the item who must be used to identify precisely which one we are referring to. author  TomT0m / talk page 09:56, 13 August 2015 (UTC)

New requirements for bot work

I started a discussion about new constraints for bot in order to avoid damage to contributors work. Some bots are still replacing existing statements without any reason, just because they are not analyzing the existing data to adopt the appropriate action. I proposed a list of checks with corresponding actions to perform before any data import/modification in order to respect the work of other contributors. Please comment there. Snipre (talk) 11:18, 13 August 2015 (UTC)

Does someone know if there is a ping template to contact all bots or their bot operator ? Snipre (talk) 11:39, 13 August 2015 (UTC)

100.000 paintings on Wikidata

A bit of news from the WikiProject sum of all paintings: We reached the milestone of 100.000 paintings on Wikidata. Multichill (talk) 15:58, 13 August 2015 (UTC) Ps. Also mentioned on the project talk

Hmm, for some reason the archive bot insists on archiving this topic right away. Multichill (talk) 07:56, 14 August 2015 (UTC)

Property:P1104

Property:P1104 for number of pages displays as + /- 1, is this intentional? See Q20816986

It is a flaw in this datatype, write 480+-0 when you add such statements. -- Innocent bystander (talk) 16:01, 13 August 2015 (UTC)
Can't it be easily fixed so that the field only contains whole number? Seems like someone with admin rights could do it right now. Do we not have enough people with those rights? --Richard Arthur Norton (1958- ) (talk) 21:38, 13 August 2015 (UTC)
Its not a flaw in the datatype itself, having the possibility to indicate a value not known to the very last digit is intentionally. Its the user interface which has this flaw, and this can only be fixed by the developers - they have this bug in their todo list for quite some time already. Ahoerstemeier (talk) 21:59, 13 August 2015 (UTC)
This is basically phab:T105623 and phab:T72341. --Izno (talk) 00:40, 14 August 2015 (UTC)

VIAF synonyms

When we find multiple VIAF entries for the same person in the VIAF database, should we add them all? That way the people that are in charge of VIAF can use Wikidata to search for entries that need to be merged. --Richard Arthur Norton (1958- ) (talk) 21:38, 13 August 2015 (UTC)

I think we should but I don't think there are any guidelines about it. I'm fairly certain there are constraints on the property for One Value (as well as Unique Value), so it'll show up, but that's okay in this case because that's not our fault. --Izno (talk) 00:09, 14 August 2015 (UTC)
No, you can add multiple values. --Richard Arthur Norton (1958- ) (talk) 04:14, 14 August 2015 (UTC)
When I say constraint I mean the use of {{Constraint}} on the property's talk page and subsequent reports at Wikidata:Database reports/Constraint violations. (Have fun. The constraint reports can last you many hours of work if you so desire.) --Izno (talk) 04:48, 14 August 2015 (UTC)
AFAIK, ISNI (P213) and VIAF ID (P214) should not have the "single value" constraint, as this does not depend on us.
  • ISNI (P213) : an ISNI is created for each identity of a creator : some authors with many pen-names can have many more than 1. If this constraint is kept, it should be authorized to add other values, with pseudonym (P742) qualifier, and modify reports so that the ones with qualifiers would not appear in automatic constraint reports… see Paul Bérato (Q3370783) who has 10 VIAF and 2 ISNI (maybe more…)
  • VIAF ID (P214) : VIAF, like wd, is still working on clustering many of its records. A lot of authors have 2, 3, sometimes more ID, as long as VIAF has not joined them. We should not put the burden on wd contributors. :) --Hsarrazin (talk) 10:01, 15 August 2015 (UTC)
It is important that all VIAF for a given person are included. Wikidata:Database_reports/Constraint_violations/P214#Single_value can be used to check potential issues, but the report on individual pages aren't of much use. Thus, for VIAF ID (P214), I moved the single value constraint further down on the property talk page. After the weekly update, the violations on individual items wont appear anymore. --- Jura 10:57, 15 August 2015 (UTC)
My impression is that as VIAF is interfaced with Wikidata, they will notice if one Wikidata item links to two or more VIAF items and merge them accordingly. In that sense, Wikidata also helps feed the VIAF database. Here's one item where that happened. Hopefully, VIAF has some review process in place to prevent bad data from multiplying. Väsk (talk) 16:29, 16 August 2015 (UTC)

Conolly's Folly (Q5162139)

I do not understand much about Wikidata, so perhaps I am asking a dumb question. I am a bit puzzled about the entries to Conolly's Folly (Q5162139) and what the different Wikipedias make of it. A Wikipedia article exists in four Wikipedias: de, en, es, and ga. All four are linked on the Wikidata page. However, when I go to the English Wikipedia article only two (es and ga) are listed unter "Languages", the link to the German article ist missing. In all other Wikipedias the three links to all the other languages are provided. What is going wrong? --ChoG (talk) 22:33, 13 August 2015 (UTC)

Maybe a cache problem? I do see the link to the German article on the English Wikipedia article! --Diwas (talk) 22:56, 13 August 2015 (UTC)
Purge the en.Wikipedia page. (Append ?action=purge to the URL and hit Enter.) --Izno (talk) 00:04, 14 August 2015 (UTC)
Great, thank you! That solved the problem! --ChoG (talk) 09:45, 15 August 2015 (UTC)

Merge problem

I wanted to merge "Kategorie" (Q17236800) with "Category" (Q224414) but Special:MergeItems tells me: "Failed to merge items, please resolve any conflicts first. Error: Conflicting descriptions for language bn." I don't see there any entry for language bn.--Avron (talk) 07:29, 14 August 2015 (UTC)

I just merged them using the merge gadget. Mbch331 (talk) 07:45, 14 August 2015 (UTC)

Vicente

Do we need three items for Vicente (Vicente (Q16741111) / no label (Q20819601) / no label (Q20819594)) because the name can be transliterated in Cyrillic languages in different ways? It's a Spanish/Portuguese name so I would say only 1 item and the Cyrillic transliterations can be aliasses. Mbch331 (talk) 09:52, 14 August 2015 (UTC)

  • It closely related to two questions. 1. Usage of those entries as values of given name (P735). Should they display the correct "translation" of the name into cyrillic languages? If yes, we will need multiple name entries per person, because same name can be translated differently for different persons. Contrariwise would one omit given name (P735) / family name (P734) (who said delete properties? did i said delete properties? no way...), or just not use them for any purpose except statistics, we can have one item per name per name language. I.e. say that all three elements is the same element and it depicts one single name from one single language... (or actually two items -- one for Spanish and second for Portuguese) it became more logical and the whole name system starts to make sense. 2. The second question is interwiki. For this particular case (single Spanish/Portuguese name and no links to, say, ruwiki) it's quite simple. Just imagine that all three items have they own different interwiki links... -- Vlsergey (talk) 10:51, 14 August 2015 (UTC)

Template:Sports properties

Is a new template about sports properties! Thierry Caro (talk) 00:52, 15 August 2015 (UTC)

Yay! Thibaut120094 (talk) 02:43, 15 August 2015 (UTC)

Properties for lower or upper level?

Are their any properties for lower or upper level? For example, 2015–16 Football League Championship (Q19776263) is the lower level of 2015–16 Premier League (Q19346732). Xaris333 (talk) 02:37, 16 August 2015 (UTC)

part of (P361) and has part (P527) ( and inverse ) Mbch331 (talk) 08:02, 16 August 2015 (UTC)
@Mbch331: No, he meant for example Ligue 1 (Q13394) View with Reasonator View with SQID > Ligue 2 (Q217374) View with Reasonator View with SQID > Championnat National (Q864298) View with Reasonator View with SQID, which you can't solely do with part of.
You can use, as far as I know, either series ordinal (P1545) View with SQID but it's string datatype so it's not automatically ordered, which is bad for a ranking property, or follows (P155) View with SQID followed by (P156) View with SQID to qualify part of (P361), like
< Ligue 1 > part of (P361) View with SQID < French Football pro league >
followed by (P156) View with SQID < Ligue 2 >
. author  TomT0m / talk page 08:49, 16 August 2015 (UTC)


None of those are useful for lower or upper level.

TomT0m, followed by (P156) can be used for 2014–15 Ligue 1 (Q16335962) > 2015–16 Ligue 1 (Q19820211) or for 2014–15 Ligue 2 (Q16537500) > 2015–16 Ligue 2 (Q19952019). So as follows (P155) (the opposite order). It can't be used to show also the lower level (for example 2014–15 Ligue 1 (Q16335962) > 2014–15 Ligue 2 (Q16537500)). Xaris333 (talk) 19:15, 16 August 2015 (UTC)

@Xaris333: If they are used as qualifier of part of (P361) or instance of (P31) the nature of the sequence it is a part of can be used to know the meaning of the qualifier. For example, the sequence of Ligue 1 championships is a temporal sequence of championships we can use
< Ligue 1 2005 > instance of (P31) View with SQID < Ligue 1 >
.
< Ligue 1 2005 > part of (P361) View with SQID < FFF championships 2005 >
might do the trick ... I proposed a in sequence qualifier in the past, but it was refused. author  TomT0m / talk page 12:16, 17 August 2015 (UTC)

Wikidata weekly summary #171

How to obtain title and description with https://wdq.wmflabs.org/?

How to obtain title and description with https://wdq.wmflabs.org/? For example, for https://www.wikidata.org/wiki/Q937 these are "Albert Einstein" and "German-American physicist and founder of the theory of relativity". How to obtain these values for a list of entites?

The database of WDQ doesn't contain those. Sjoerd de Bruin (talk) 20:27, 6 August 2015 (UTC)
Is it possible to fix this? Or there is some reason for this? Dims (talk) 14:33, 7 August 2015 (UTC)
If you really need the labels you could use SPARQL instead. If you plug your WDQ query into https://tools.wmflabs.org/wdq2sparql/w2s.php it will output the corresponding SPARQL query. That could then be manipulated to get the labels as well. The examples provided here should help get you started. Gabbe (talk) 18:32, 7 August 2015 (UTC)
Thanks a lot, this is probably great! Unfortunately, I can't find (yet) hot to pass query and get response directly, without a provided web-form http://wdqs-beta.wmflabs.org/ Dims (talk) 20:06, 7 August 2015 (UTC)
@Dims: you can query the endpoint directly at http://wdqs-beta.wmflabs.org/bigdata/namespace/wdq/sparql?query=..., e.g. tinyurl.com/nvnvrph. Note that if you want JSON response, you'll need to pass the header 'Accept' : 'application/sparql-results+json' --Smalyshev (WMF) (talk) 18:08, 18 August 2015 (UTC)
Corrections: easier way to get JSON response is just to add format=json query parameter in the URL. --Smalyshev (WMF) (talk) 18:13, 18 August 2015 (UTC)

Draggable sitelinks not working...

Hi,

Today, Draggable sitelink does not work anymore for me. Am I the only one ? (before I make a bug report).--Hsarrazin (talk) 09:09, 15 August 2015 (UTC)

Does not work for me neither since the Mediawiki update. I left a message on the gadget talkpage. Thibaut120094 (talk) 12:42, 15 August 2015 (UTC)
Race condition? Devs wanted! --Ricordisamoa 07:20, 18 August 2015 (UTC)

Fusion problems

Can someone fusion de:Kategorie:Hinduistische Kunst (Q19478380) with en:Category.Hindu art (Q8514359)? 188.96.228.156 21:15, 17 August 2015 (UTC)

Can someone fusion de:Kategorie:Hinduistische Kultur with en:Category:Hindu culture (Q8514371) ? 188.96.228.156 21:19, 17 August 2015 (UTC)

✓ Done Popcorndude (talk) 21:45, 17 August 2015 (UTC)

quick autolist q for MAC users

How do I import an autolist query file wikidata_query_result.tsv on a MAC without losing all of my umlauts? I tried importing as Windows Ansi, Macintosh, UTF 6.0 UTF-8 and so forth, but no umlauts yet. Any ideas? `thx Jane023 (talk) 08:57, 18 August 2015 (UTC)

There was quite related problem with SQL quarry. You could try open the file with Notepad or Notepad++ (if you have it or something similar on Mac). At least, for .csv files it helps. --Edgars2007 (talk) 09:20, 18 August 2015 (UTC)
Thanks I have TextEdit so will try that. I now have umlauts but can't get the Spanish backwards accent (like for Velasquez), so making some progress. Jane023 (talk) 09:26, 18 August 2015 (UTC)

Vandalism

Hi! Wanted to ask, if there is some good way with dealing with such vandalism: see edits made by Мұхамеджан Амангелді here. I think there should be made some script or database report for such cases (yes, I understand, that there would be some false positives). And, more important - how to revert such edits not screwing up everything (so - I can't press "restore")? Not by adding one sitelink by one - it is too booring :) --Edgars2007 (talk) 09:16, 18 August 2015 (UTC)

I see this behaviour every week. Thought we've had a abusefilter for it... Sjoerd de Bruin (talk) 09:19, 18 August 2015 (UTC)
We have Special:AbuseFilter/10 and Special:AbuseFilter/71. If you cannot restore, you may group wrong revisions and undo these groups. [6][7] Matěj Suchánek (talk) 09:29, 18 August 2015 (UTC)
Doesn't look quite simple (take a look at my first post - not screwing up everything), but OK, will try. --Edgars2007 (talk) 09:58, 18 August 2015 (UTC)

Bot request

Where do I make a bot request? I don't want to make the bot, I want somebody else to make and run the bot. Thx. Jane023 (talk) 11:02, 18 August 2015 (UTC)

Wikidata:Bot requests. Matěj Suchánek (talk) 11:06, 18 August 2015 (UTC)
Thx Posted there. Jane023 (talk) 11:48, 18 August 2015 (UTC)

Best way to get sitelinks for lots of items at once ?

Is there a good way to get the sitelinks to a particular wiki for lots of items at once? Specifically, I am interested in the ~8,500 items that have Art UK artist ID (P1367), which I can obtain from WDQ; I would then like then to be able to get the sitelinks of all corresponding articles on en-wiki, to update these tracking pages.

Up until now I have been using Autolist, and then pressing the 'Download' button by hand to get the en-wiki links. But I'd like to make the process more automated. Presumably there is a good way for my script to get all ~6,800 en-wiki links in one go (as I imagine Autolist is doing), without having to call the item API 8,500 times? What's the best way to achieve this? Jheald (talk) 23:00, 17 August 2015 (UTC)

here is all of 6837 sitelinks for the 8678 items. I ran "claim[1367]" on Autolist and set the output method to "PagePile", then ran this on the pile, selecting the filter "From WikiData". The whole process took about 30 seconds. Is this what you were looking for? Popcorndude (talk) 23:52, 17 August 2015 (UTC)
@Popcorndude: Interesting. Ideally I'm looking to make a script that can run automatically as a cron process, say once a week, without any manual intervention at all. It looks like that might be possible with WDQ and PagePile, though I'm not 100% clear on how to do it. The existing data merging and page generation is written in Perl, so it would be nice to have a call I can just add to the Perl program to find out which of the Q-numbers from WDQ have en-wiki links and which don't, and what those en-wiki links are. Jheald (talk) 00:53, 18 August 2015 (UTC)
Well, you can automatically download the results with something like https://tools.wmflabs.org/pagepile/api.php?id=378&action=get_data&format=text&doit
PagePile has an API, so you could probably automate that part. Unfortunately, this is the closest thing I could find to documentation of said API.
At this point, you have sitelinks, but not items. You could probably feed those into a simple pywikibot script pretty easily. Popcorndude (talk) 01:58, 18 August 2015 (UTC)
Looks like the following is the call I need, from the wikidata API, with the Q-numbers batched up into groups of 50 or 500 at a time.
https://www.wikidata.org/w/api.php?action=wbgetentities&ids=Q1%7CQ42%7CQ2823073&props=sitelinks&sitefilter=enwiki
(Not sure whether I need a bot flag to get 500 at a time, or whether it's enough for the account just to be a bot, but that should become clear soon enough). Jheald (talk) 08:23, 18 August 2015 (UTC)
Magnus put together a list of enwp sitelinks a couple of weeks ago here - it may be a little out of date but probably good enough. You could just search for the relevant Q-numbers in that and match across. Andrew Gray (talk) 11:24, 18 August 2015 (UTC)
Some interesting stuff mentioned above. Similarly, I would like to find a way to auto-generate and populate pages like Wikisource:Authors-A which are currently done manually, and as such are not kept up to date, or are complete. So really looking forward to see a tool that can either run those, or can be more easily set up to run cron jobs from Tools Lab.  — billinghurst sDrewth 01:38, 19 August 2015 (UTC)

Central login

I get the message "You are centrally logged in". When reloading the page I am still not logged in. When clicking the "Log in" link in the top right corner i get to the Log in template but are automatically sent back to Wikidatapage I was clicking from. I guess this is known or temporary, but I also want to see what happens with signatures when I save.--95.199.140.177 05:21, 18 August 2015 (UTC)

OK, not centrally logged in then. And it is not possible for me to log in.--95.199.140.177 05:22, 18 August 2015 (UTC)
Delete your cookies. Sjoerd de Bruin (talk) 06:40, 18 August 2015 (UTC)
This should have been resolved in T109038. Mbch331 (talk) 06:52, 18 August 2015 (UTC)
Thank you, I forgot "Delete your cookies" is the new "Restart your computer":). I did, and was logged out also from Wikipedia. When logging in at Wikipedia I was centrally logged in also to Wikidata. So it worked, but I guess it is a temporary work-around.--LittleGun (talk) 07:09, 18 August 2015 (UTC)
@LittleGun: There was a change last week to separate the cookies for main wikidata and the testbed wikidata, and it was an imperfect change. <shrug> If you just clear all your cookies, rather than delete the specific and related cookies, then yes it will have a wider impact on your other logins on this occasion. As such it is a situational change, it should have no permanent repercussions.  — billinghurst sDrewth 01:32, 19 August 2015 (UTC)

Trends in links from items to Commons

I've updated some searches I first ran a year ago, to try to get a handle on how people are linking from items to Commons categories and galleries. Here is some summarisation.

(1) Links from article-items to Commons categories

We now have 884,439 Commons categories identifiable with a Wikidata article-like item by either sitelink or Commons category (P373), up from 708,057 a year ago. This compares to about 3.4 million Commons categories in total, though given the number of maintenance and intersection categories, it is only a finite proportion that could ever be expected to be relateable to article-like items. Of these,

  • 807,776 (up from 697,049) are pointed to by a Commons category (P373) statement on an article like item.
  • 207,494 (up from 94,899) are connected by a sitelink from an article-item to a Commonscat.
Comment In the past year the number of direct sitelinks from articles to Commonscats has gone up by over 110,000 and more than doubled -- despite being deprecated. This includes the number of such sitelinks without corresponding Commons category (P373) properties, which has gone up from 31,112 to 76,663 in this period.
In my opinion, this needs to be addressed, if Commons category (P373) is to be underlined as the way users should expect to find Commons categories connected to article items, and vice-versa. New P373 statements should urgently be created for the 76,663 sitelinks which don't have them; after which all 207,494 sitelinks from article items to Commonscats should be removed. There needs to be a single clear system in place, that users can understand and software can rely on. Article <-> commonscat sitelinks do not provide that, because of the expectation that such sitelinks should lead to a gallery on Commons, leading to the inability to always be able to create Article <-> commonscat sitelinks if Article <-> gallery sitelinks are also in play. Having taken the decision that P373 is the preferred way to indicate the Article <-> commonscat relationship, we should ensure that that is in fact the actuality and the consistent expectation.

(2) Links from category-items to Commons categories

323,825 Commons categories are now identifiable with a Wikidata category-like item, up from 262,041 a year ago. Of these,

  • 246,578 (up from 239,152) are connected by both a sitelink and a Commons category (P373)
  • 64,688 (up from 11,651) are connected by only a sitelink
  • 12,586 (up from 11,238) are connected by only a P373.
Comment It is striking that of the 61,784 increase in connections over the last year, a net 53,037 (ie 85%) appear to be accounted for by addition of a sitelink without addition of a P373.
This is despite an RfC to create only sitelinks in such cases, not P373s being definitely rejected. (Though it was perhaps a little unclear whether the proposal being rejected was to delete only P373s from category-items in such cases, or P373s on all items). If we think P373s are worth having on category-items, it is perhaps worth making sure that they are in fact being created.
The P373s may appear redundant, so their actual value may be worth discussing; but they do represent a way that can always be used to specify a cat-item <-> commons cat relationship, which can always be used, even if an article-item is sitting on the commons cat sitelink. So that may be some reason to preserve them (and/or some software may expect them). But in this case they should not be missing from almost 65,000 items.

(3) P301 / P910 relationships for items with Commons categories

The number of category's main topic (P301) / topic's main category (P910) relationships has almost doubled in the last 12 months.

Comment This is still a long way short of the 884,439 total number of article items with a commonscat -- but probably always will be, as for many of these items there may be many pictures to justify a commonscat, but few related articles to justify a wikidata category. P373 is therefore likely to remain the standard way to link an article with a commonscat (and vice-versa), rather than P910 and then a cat <-> cat sitelink.
But given that there are 323,825 cat-items with commonscats, still a lot more than the 166,764 cat items with a P301, there may still be scope to identify new P301 / P910 pairs, by looking for unconnected cat-items and article-items that share a P373 linking to the same commonscat.

(4) Links from article-items to Commons galleries

87,890 Commons galleries can be connected with article items (out of about 112,000 galleries in total)

  • 87,109 (up from 84,167) are connected by sitelinks
  • An additional 781 are connected by Commons gallery (P935) without a sitelink
Comment (i) P935 is still very little used; (ii) Currently, more than twice as many article-items sitelink to Commonscats (deprecated) as to galleries (207,494 vs 87,109), up from roughly parity just a year ago. Action is needed, if we want to keep the expectation that article-items sitelink to galleries (only), not to commonscats.

Look forward to people's thoughts on any of this. Jheald (talk) 17:48, 18 August 2015 (UTC)

The RfC you mention in (2) was closed rather quickly because it was created by a sockpuppet of an indefinitely blocked user. I don't think you can draw any useful conclusions from that one. - Nikki (talk) 18:39, 18 August 2015 (UTC)
Looks like big fail of category-category model. 2.4 millions categories are not connected to corresponding Wikipedia articles. Navigation problem is not resolved for a year or more... — Ivan A. Krestinin (talk) 20:22, 18 August 2015 (UTC)
Thanks for the numbers! But one set of numbers is not up-to date in your report, number of galleries and categories on Commons (as stats page on Commons is not updated anymore). But we have some rough number from stats.wikimedia.org, so the number are 4.2 Commons categories as of June 2015 (up from 3.5 as of June 2014, +20%) and 204 k of articles (including galleries, anybody is able to extract current number of galleries on Commons?) as of June 2015 (up from 191 k as of June 2014, +7%). Simply categories, not galleries are lifeblood of Commons.
One reason of these numbers is that people on Commons are simply using WD (Phase 1) to replace interwikis. So when commons category was connected to articles using old interwiki links (common case), they add commons category sitelink to article item and deleting old interwiki links.
Another reason is simply that even on WD there is no consensus about category-category model. And old incorrectly closed RfC (Wikidata:Requests for comment/Commons links) is just demonstration of that missing consensus. --Jklamo (talk) 21:50, 18 August 2015 (UTC)

How can we improve Wikimedia grants to support you better?

My apologies for posting this message in English. Please help translate it if you can.

Hello,

The Wikimedia Foundation would like your feedback about how we can reimagine Wikimedia Foundation grants, to better support people and ideas in your Wikimedia project. Ways to participate:

Feedback is welcome in any language.

With thanks,

I JethroBT (WMF), Community Resources, Wikimedia Foundation.

(Opt-out Instructions) This message was sent by I JethroBT (WMF) (talkcontribslogs) through MediaWiki message delivery. 23:09, 18 August 2015 (UTC)

This message's links were all broken. Fixed to direct to Meta. --Yair rand (talk) 00:00, 19 August 2015 (UTC)

Books with IMDb identifier

I think there should be no results: Query: claim[31:571] and claim[345]. --Jobu0101 (talk) 08:52, 15 August 2015 (UTC)

@Jobu0101: be careful with that. A lot of films are actually adaptations of a book. Take for example P.S. Your Cat Is Dead (Q1131202). If you look at en:P.S. Your Cat Is Dead, you'll see that the article is about the play, the book and the film. On Wikidata you should create three items and interconnect these. It's better to keep the book so people know that this is one of the items that need to be untangled. Multichill (talk) 10:20, 15 August 2015 (UTC)
So who wants to work on that? --Jobu0101 (talk) 13:31, 17 August 2015 (UTC)
@Multichill: ... and maybe a fourth item corresponding to the union of all three, which would directly connect to the Wikipedia article, and which the items for the book, the play and the film would each have the relation facet of (P1269) this fourth (union) item? This union item would presumably have the instance of (P31) = creative work (Q17537576) -- or should a new sub-class be created, for "creative works that include manifestations in multiple mediums". This would make it possible to translate general statements discussing such works, without specifically limiting them to be statements about the book / play / film. Jheald (talk) 16:07, 17 August 2015 (UTC)
Adaptations are typically based on (P144) an original work that was originally available in a particular medium (in most of these cases, movies based on books) and it is typically best to link the adaptations directly to the source material without intermediaries. Unless there is a special case where the original story resided beyond any event or medium, that solution sounds unnecessarily cumbersome. Väsk (talk) 17:10, 17 August 2015 (UTC)
The Japanese Wikipedia, for example, routinely only has one article for a book and its (possibly numerous) film adaptations. How should the item corresponding to such an article be handled in your opinion Jobu0101? Gabbe (talk) 07:00, 19 August 2015 (UTC)
@Gabbe: They should be split into several items. Main about book and other linked items -- about film adaptations. jawiki can use arbitrary access to display infoboxes or other linked film information. -- Vlsergey (talk) 10:08, 19 August 2015 (UTC)

Inaccurate info

Why can't I edit inaccurate info on Q2818875 who died on 4 March 1790, not in 1824?! 62.233.53.231 19:37, 19 August 2015 (UTC)

@62.233.53.231: You did fix it. 19:50, 19 August 2015 (UTC)

Famous animals and constraint violations ?

Hello, when completing the items of famous animals such as Hanno (Q1479200), Nils Olav (Q1144606) or Choupette (Q5105465), their instance of (P31) should be their specie, right ? However, instance of (P31) requires the target to have subclass of (P279), which is incompatible with taxon name (P225), a mandatory property for items with parent taxon (P171). How can I solve this issue ? Should the constraint on instance of (P31) be changed to allow for both subclass of (P279) and parent taxon (P171) ?

Thanks in advance for your help. -- Whidou (talk) 10:00, 7 August 2015 (UTC)

99of9
Abbe98
Achim Raschka (talk)
Brya (talk)
Dan Koehl (talk)
Daniel Mietchen (talk)
Faendalimas
FelixReimann (talk)
Infovarius (talk)
Jean-Marc Vanel
Joel Sachs
Josve05a (talk)
Klortho (talk)
Lymantria (talk)
MargaretRDonald
Mellis (talk)
Michael Goodyear
MPF
Mr. Fulano (talk)
Nis Jørgensen
Peter Coxhead
PhiLiP
Andy Mabbett (talk)
Plantdrew
Prot D
pvmoutside
Rod Page
Soulkeeper (talk)
Strobilomyces (talk)
Tinm
Tom.Reding
Tommy Kronkvist (talk)
TomT0m
Tubezlob
RaboKarbakian
Circeus Pictogram voting comment.svg Notified participants of WikiProject Taxonomy

If you want my opinion, we should get rid of parent taxon (P171) and follow the principles of Help:Classification and Help:BMP. It's actually a duplicate of instance of (P31). author  TomT0m / talk page 10:44, 7 August 2015 (UTC)
I don't know why taxon name (P225) even has that constraint to not use subclass of (P279). Wikidata has the ability to have multiple-ontologies and I don't see how a subclass of (P279) statement conflicts with anything WikiProject Taxonomy does. --Tobias1984 (talk) 11:17, 7 August 2015 (UTC)
Every species item should be <instance of:taxon>. It should be possible to update the constraints to accommodate this. I know this doesn't line up with how other stuff is done but I disagree with Tobias1984. I don't think we should replace "parent taxon" with "subclass of". Joe Filceolaire (talk) 17:04, 8 August 2015 (UTC)
@Filceolaire: I didn't say that I want to replace "subclass of" with "parent taxon", but I don't see how these statements can collide with each other. "parent taxon" has a more strict taxonomic sense, while "subclass of" is the generic and more flexible property. One example why we need both is: The top-most taxon of any taxonomy, does not have a "parent taxon", but it is a sublcass of something (e.g. matter). There are similar examples at the lower end of the taxonomy tree where we need p279. I think p31= taxon does also not collide with any of the above. --Tobias1984 (talk) 18:28, 8 August 2015 (UTC)
@Filceolaire, Tobias1984:. I still wait for something objective that we would loss by deleting "parent taxon" (seriously). A species should be
< the species > instance of (P31) View with SQID < species >
with
< species > subclass of (P279) View with SQID < taxon >
imho.
I fully agree that this constraint is a nonsense and should be deleted. I understand that WikiProject Taxonomy want to keep parent taxon (although I'm not sure why) but I don't understand why this should block any other initiative. author  TomT0m / talk page 16:40, 9 August 2015 (UTC)

Whidou: What makes an animal „famous“? --Succu (talk) 21:53, 9 August 2015 (UTC)

That's a serious question ? author  TomT0m / talk page 07:18, 10 August 2015 (UTC)
Sure TomT0m. Maybe a question of relevance? Just mocking... I was asking Whidou, not you. --Succu (talk) 22:00, 10 August 2015 (UTC)
Mock as much as you want, this has no influence on relevance. It's a question for everyone, not for a single user. And as someone who raised the problem before in front of you, and was ignored, I feel like I'm relevant. So, an answer : Wikidata as as many famous animals that Wikipedias can have, and even more potentially. So what make them famous is not really a relevant question here. author  TomT0m / talk page 07:55, 11 August 2015 (UTC)

A related problem: Find A Grave memorial ID (P535) can be used on humans and animals, but I can't find a way to say "human or animal" in the constraints, because most animals I looked at (e.g. dog (Q144)) aren't a subclass of anything. - Nikki (talk) 08:36, 11 August 2015 (UTC)

@Nikki: I added two values to p279. I actually think that q144 is a good example of why coexistance of p279 and parent-taxon makes sense. One is stricter and the other more flexible. And the small redundancy in a few statements is not important in my opinion. --Tobias1984 (talk) 12:03, 11 August 2015 (UTC)
@Tobias1984: A few statements, multiplied by the numbers of taxons does not remain a few statements. All this is because Wikidata is property based and not class based. I see reverted P279 statements on taxons very often, so don't expect it to stay that way ... author  TomT0m / talk page 12:08, 11 August 2015 (UTC)
I don't have an issue with both existing, but that doesn't work when people keep removing the subclass statements. subclass of (P279) pet (Q39201) has been removed from dog (Q144) before (see [8]). subclass of (P279) Canis lupus (Q18498) has been removed three times before (see [9], [10] and [11]). - Nikki (talk) 12:33, 11 August 2015 (UTC)
@TomT0m: We should probably speak to those people that revert such edits and ask them for their rational. Only vandalism should be reverted without discussion. Everything else needs to follow the "assume good faith" rule. - And yes multiplying by the number of taxons we get a huge total of redundant statements. But I also think we should not try to prematurely optimize Wikidata. I believe that we are in a phase where we can try things out and then choose the things that work best. --Tobias1984 (talk) 12:39, 11 August 2015 (UTC)
Indeed. author  TomT0m / talk page 12:57, 11 August 2015 (UTC)
@Tobias1984: I think those edits are reverted because of the constraint violations : if we remove subclass of (P279) from taxon name (P225)'s incompatibility list, everything should be fine. -- Whidou (talk) 12:52, 11 August 2015 (UTC)
Done. -- Whidou (talk) 12:57, 11 August 2015 (UTC)
The constraint was added to find errors, e.g. diseases intermixed with taxa. That possibility is gone now. All subclass of (P279) statements are unsourced and a little bit of original research. This is especially true if you try to construct a monohierarchy because otherwise the resulting ontology is meaningless. Wikispecies does this. But this is not the way we try to reflect the diversity of taxonomic options. --Succu (talk) 16:03, 11 August 2015 (UTC)
I'm sure we cam find solutions: To express that no desease is an organism, we can use disjoint with statements to say that if A and B are disjoint, no sub(sub…)class of A should be a subclass of B and vice versa, or Disjoint union of statements. WDQ queries can be generated from that, with (TREE[279][A][] AND TREE[279][B][]) I don't see why subclass of should be more unsourced than parent taxon on taxons either. And what do you mean by monohierarchy ? A directed acyclic graph (Q1195339) View with Reasonator View with SQID of classes would be a polyhierarchy whereas arborescence (Q1520825) View with Reasonator View with SQID would be a monohierarchy ? then Wikidata can't be a monohierarchy because it's supposed to reflect neutrally any relevant taxonomic viewpoints. author  TomT0m / talk page 16:34, 11 August 2015 (UTC)
This had nothing todo with disjoint with & co. There where simply some misplaced statements. I doubt there is a lot of literature which states taxon A is subclass of (P279) of taxon B. It's easy to find them for parent taxon (P171) (e.g. User:Succu/Animalia) Yes, „Wikidata can't be a monohierarchy because it's supposed to reflect neutrally“ all points of view, a central point of WD self-conception. But this another topic. Maybe for the forgotten Wikidata:Lounge. --Succu (talk) 19:37, 11 August 2015 (UTC)
I'm trying to find solutions to the problems you mention. author  TomT0m / talk page 09:16, 12 August 2015 (UTC)
Well, subclass of pet, and subclass of Gray wolf are redundant, saying the same thing twice, only more crookedly. Is this going to be one of those cases where Wikimedians are going to tear things apart, just for the sake of tearing things apart? - Brya (talk) 16:50, 11 August 2015 (UTC)
@Succu: And for the sourcing problem, as parent taxon is an hypo/hyponym (linguistics) (Q680042) relation, there is actually no problem to source subclass of with the same sources (actually if it's a subproperty this should be obvious). author  TomT0m / talk page 18:46, 12 August 2015 (UTC)
This linguistic term is intended to reflect monohierarchies. Legerdemains are not suitable as reliable references. --Succu (talk) 19:28, 12 August 2015 (UTC)
@Succu: We can actually reconstruct the tree with these properties : Wikidata:Property proposal/Generic § subclass for example if a taxon A has three immediate subtaxon B, C, D we could say that Template:A to express that any A organism is exactly one amongst B, C or D. This is equivalent.
For the sourcing part, this is no trick, as logically being a parent taxon means being a subclass. So a source that says it's a parent taxon can as well source the subclassing. author  TomT0m / talk page 19:56, 12 August 2015 (UTC)
Sorry, but are you living in some kind of virtual reality (Q170519) of your own? --Succu (talk) 21:29, 13 August 2015 (UTC)
«subclass of pet, and subclass of Gray wolf are redundant» Mmm no they are not. Some wolves or dogs ae not pets, and some pets are not dogs, so certainly not Gray Wolves.
«just for the sake of tearing things apart» There is reasons for this, and it does not tears appart anything.

But to answer the original question: "parent taxon" is a subproperty of "subclass of" so there is no constraint violation in "instance of gorilla". - Brya (talk) 05:54, 12 August 2015 (UTC)

@Brya: The problem is that I'm not aware of any tool who can deal properly with that. It's tricky … And is known to be a problem in some tools, those who are OWL based (statements on properties with special meaning liki subclass of is often forgotten) author  TomT0m / talk page 09:16, 12 August 2015 (UTC)
If OWL-based tools give you such trouble, you would be well-advised to stay away from them. - Brya (talk) 10:51, 12 August 2015 (UTC)
@Brya: It's not as this was only an example. Are you aware of a tool that actually handles this, to be more constructive ? author  TomT0m / talk page 10:59, 12 August 2015 (UTC)
As an example of the tools we're speaking of, see this thread on the Wikidata mailing list. author  TomT0m / talk page 11:50, 12 August 2015 (UTC)
I am not much for tools, and here I am not even aware of what problem you may have found? - Brya (talk) 17:11, 12 August 2015 (UTC)
"IRIs from the reserved vocabulary other than owl:topObjectProperty and owl:bottomObjectProperty MUST NOT be used to identify object properties in an OWL 2 DL ontology." (from http://www.w3.org/TR/owl2-syntax/#Object_Properties ). As subclass of is an rdfs property, it is a part of the reserved vocabulary. (The exceptions are the top and bottom properties of course because any property is a property of the top one an the bottom one is a subproperty of all properties.
This is to guarante some logical properties (some queries on an OWL ontology might have no answer and make the reasoner freeze and take all the memory - this can make the ontology undecidable, see https://www.cs.ox.ac.uk/boris.motik/pubs/motik07metamodeling-journal.pdf to be complete).
As far as I'm aware of WDQ for example can't even use subproperty statements. author  TomT0m / talk page 17:50, 12 August 2015 (UTC)
TomT0m, Maybe I missed some discussions: WD is obligated to follow this? --Succu (talk) 21:48, 13 August 2015 (UTC)
@Succu: No, you did not miss anything. But as Wikidata is a part of the semantic and Linked Open Data world, it would make sense to make an effort to stay compatible. It would make tools like protege available to work with. As I mentioned to Brya (talkcontribslogs) there is not much tools that supports subproperties, think of Wikidata constraints for example who makes a very poor use of them, if they do any at all. subclass of (P279) and instance of (P31) are proving to be useful anyway in the whole project in a vast variety of fields and community use them widely. I provided ways to make anyway a tree (monohierarchy) using standard properties even if we don't use parent taxon (even if we still can use it, I'm just saying we can also do without). The unique parent constraint for a taxon B, for example, can be achieved by checking that there is a unique other taxon A such that
< B > subclass of (P279) View with SQID < A >
and
< A > disjoint union of (P2738) View with SQID < ... >
together with (P1706) View with SQID < B >
... search < ... >
. That way we can directly import the ontology in an OWL tool, and there is advanced one, and work with it. As those tool are aware of the meaning of disjoint union, subclass of and so on, they can make more advanced consistency checks. And ontology editors could become a good way to edit Wikidata (think that several items are shown on the same screen in an ontology editors while in Wikidata itself you need one screen per items ...) I think we have a lot to win by following standards when possible and let Wikidata ecosystem grow with other ecosystems in harmony. Anyway both models can live together without problems in Wikidata. The parent taxon model can live in and OWL ontology as well, as long as it's not a subproperty :) We can probably even model things with implicit classes taking into account parent taxon. The class of all taxons can be defined as the class of all items with a parent taxon statement. Maybe more powerful things can be expressed, such as checking that a species never has an other species as parent taxon. author  TomT0m / talk page 22:58, 13 August 2015 (UTC)
Going back to original question. I believe that this should be not problem because parent taxon (P171) is a subproperty of subclass of (P279) as Brya has mentioned above. So this relation should be supported by main tools if it isn't now. By constraints (@Ivan A. Krestinin:), by Reasonator/WDQ (@Magnus Manske:), what else? --Infovarius (talk) 19:17, 20 August 2015 (UTC)
Famous animals set has interesting structure: the most instances of this set are instances of pet (Q39201). Sometimes it contains instances of elephant (Q7378), but never contains instances of Escherichia coli (Q25419). So domain organism (Q7239) is logically correct for properties like place of death (P20), but is too wide. Proposal is define domain like human (Q5), pet (Q39201), horse (Q726), elephant (Q7378) and some other. This definition is incomplete and will be extended many times, but it is more strong than domain organism (Q7239). And this definition is free from originaly discussed issue. — Ivan A. Krestinin (talk) 20:25, 20 August 2015 (UTC)
@Ivan A. Krestinin:. OK for them being subclass of pets, although they can also be beasts of burden or animal raised for fooding humans, but not OK with being free from the original issue. human (Q5) is definitely a taxon, has a taxon rank and so on ... author  TomT0m / talk page 06:37, 21 August 2015 (UTC)
 ;@Infovarius: : Please read the part when we discuss with Brya of subproperties of subclass of (P279). They are disallowed on OWL2 DL for avoiding big issues that would render the language undecidable, that means that some "queries" on an OWL ontologies can become just unanswerable and that the query engine (reasoner in this case) would try to find one and lost himself filling the memory before failing ... OWL is a important standard, and there is tools such Protégé (Q2066865) and its web version who already have been proposed to help collaborative work on Wikidata). So there is good reasons not to do this. Not to mention after some time we still have no support of subproperties in any of our tools. By contrast it's possible to do a lot of things just with subclass of (P279), see the {{PropertyForThisType}} template I just coded yesterday. (Although the SPARQL engine can probably use subproperties). author  TomT0m / talk page 06:52, 21 August 2015 (UTC)

Any progress on property metadata?

I was looking at the discussion as to whether to delete input set (P1851) and got to wondering what is happening to the proposals for "range" and "domain" properties for properties (discussion here). In the absence of these properties I see people have started using domain (P1568) and codomain (P1571) to describe properties.

  • Should we be using these properties for statements about properties?
  • Are we ready to create metadata properties for 'property domain' and 'property range'? Joe Filceolaire (talk) 16:09, 19 August 2015 (UTC)
Specific properties are better. Snipre (talk) 08:49, 20 August 2015 (UTC)
@Snipre: You're cryptic today, yet another message I can't find a meaning for :) author  TomT0m / talk page 09:16, 20 August 2015 (UTC)
Just read the questions: do we want to use a general property domain or metadata properties for properties. Snipre (talk) 14:01, 20 August 2015 (UTC)
@Snipre: It's not really a matter of specific/generic and the fact that the property applies on properties or on items but a matter of level of representation. Yes, indeed wikidata properties are used to represent mathematical relations, but it's very different to say things about abstract maths functions like sinus, than to say in Wikidata itself things about Wikidata model. This is tricky, and can trigger modelling like problem of self-reference (Q1129622) View with Reasonator View with SQID. There is restrictions on this about this on ontology languages (cf. the discussion we had in Brya on project chat about subproperties on subclass of (P279) recently on this chat). In no way domains of variable has to be treated as a specific case of domain of an abstract math function. One should not be a specific case of the other, even if on the purely conceptual level we can think of statements as defining a relation in Wikidata, imho. author  TomT0m / talk page 14:14, 20 August 2015 (UTC)
@Filceolaire: : We have properties for this type (P1963) which assigns a property to a domain, maybe property domain would be a bit redundant … But you are right, the properties for maths functions amd relations should not be used for Wikidata properties. author  TomT0m / talk page 09:16, 20 August 2015 (UTC)
@Filceolaire: Just created a template to make this useful : {{PropertyForThisType}}. This generates a WDQS query to generate a list of properties appliable to some class of items, for example With the class of all country (Q6256) View with Reasonator View with SQID we can see that coordinates can be entered, because a city is a location. author  TomT0m / talk page 10:14, 20 August 2015
properties for this type (P1963) is not very explicit and if you look at the property proposal, this one is the inverse of this proposal. Then properties for this type (P1963) is a very bad property because each element of WD can used it and finally you can have twice each property in an element as a normal property to store data and as a value of this property. Better to store the information about the use of a property in the property page than saying in each element the list of all properties which can be used with the element. Snipre (talk) 14:18, 20 August 2015 (UTC)
@Snipre: What ? No ! It's a property for classes items like human (Q5) (a class is type) to say that it applies to instances of human. No human instance item will have properties for this type (P1963), that would not make any sense. author  TomT0m / talk page 14:32, 20 August 2015 (UTC)
@TomT0m: So provide use information on the property page because there is nothing indicating that. A constraint should be added too in order to track misuse of the property. Then I think that using this property in the main space in a bad practice because we mix properties for element description with properties for WD organization. We create meta properties for properties for this purpose and here I don't think we have a clear distinction between the different kind of properties. Why do we store the constraints on the properties side and not on the class elements side ? Why do we store the application domains for properties on the class elements side and not on the properties side ? I think there is a lack of reflexion. Snipre (talk) 12:46, 21 August 2015 (UTC)
@Snipre: We will, I guess we forgot about this since Zolo's proposal :) and the discussion below explain part of this forgetting.
"Why do we store the constraints on the properties side and not on the class elements side ? Why do we store the application domains for properties on the class elements side and not on the properties side ?" These are good questions and you'll be interested in this thread on about dbpedia experience on wikidata mailing list. Everything is centered aroud properties here because of the property proposal process. I think it leads to dispersion of informations on a lot of places when documenting classes would be more efficient as properties are often used together ... Basically : we store the constraint on properties because Ivan is the one who builds the constraints report, and community followed him. I tried to talk with him of more class oriented constraint like one based on the (disjoint) union of properties, he simply replied this is class based not property based. This explains that. author  TomT0m / talk page 13:04, 21 August 2015 (UTC)

Some questions

Have some stupid questions.

  1. Is this edit OK? So basically for all people I have to put P31:Q5? And in this case Q215380 should be for P106?
  2. At Russian Wikipedia there is category Biographies. I would like to get those persons, who hasn't P31:Q5, but I see, that I can't put noclaim[31:5] alone in WDQ. Some ideas? Edgars2007 (talk) 23:11, 19 August 2015 (UTC)
Hello!
Yes, all humans should have P31:Q5. band (Q215380) isn't a profession. If he's part of a specific groupe then you use P463 with the item of this specific group as value. For P106 the value is probably musician (Q639669).
In Autolist for exemple you can search in a specific category and then use NOT for the query with claim[31:5]: [31%3A5&statementlist=&run=Run&mode_manual=or&mode_cat=and&mode_wdq=not&mode_find=or&chunk_size=10000 like that]. But beware of false positive: verify the list before adding automatically the claims. It seems like many entries on this list shouldn't have claim31:5: there a group, couple, brothers, etc. --Harmonia Amanda (talk) 00:03, 20 August 2015 (UTC)
Q928822: It's not clear as the article in ruwiki and frwiki are about a person, in the other two wikis about a "project" by that person. --- Jura 02:40, 20 August 2015 (UTC)
Oh, I don't read Czech and German so I didn't know. Then the solution is to separate the links, for the person in an item and the project in another. --Harmonia Amanda (talk) 06:41, 20 August 2015 (UTC)
Thanks. Yes, I noticed, that there could be some false positives in that category. --Edgars2007 (talk) 07:43, 20 August 2015 (UTC)
The important thing is to be clear if the statements in the item are about the person or about their one-person project/band. Statements about one will not neccessarily be true about the other. This is why we need two items. If you are ever in doubt about whether or not to split an item just let the statements guide you. Joe Filceolaire (talk) 18:13, 21 August 2015 (UTC)

Double display

Moin Moin, if I would add a property to an item, I get a double display, now. Is that right or am I doing something wrong? Regards --Crazy1880 (talk) 14:56, 20 August 2015 (UTC)

I have it too. Probably the result of the latest update. Didn't notice it until this morning, when I also noticed it now also shows aliasses. Mbch331 (talk) 15:07, 20 August 2015 (UTC)
It looks like it yeah :( I've opened phabricator:T109697 for it this morning. --Lydia Pintscher (WMDE) (talk) 15:59, 20 August 2015 (UTC)
Is this really new? I think I have seen such behavior since Gorm the Old (Q314490) was king! -- Innocent bystander (talk) 12:56, 22 August 2015 (UTC)
Yes it is. I had it in the past, then it was gone and 2 days ago it returned. Mbch331 (talk) 13:18, 22 August 2015 (UTC)
I don't like it how the aliases are shown just as bold and as prominent as the property name, it's too much clutter for my field of vision. Jared Preston (talk) 16:30, 20 August 2015 (UTC)

Property for 'medium of instruction'

I was editting Coláiste Lurgan (Q20819105). This is an school where Irish (Q9142) is the medium of instruction - Irish is used to teach other subjects. Any suggestions as to what is the best property for the 'medium of instruction'; there are a number of schools around the world where this is an important property.@Sjoerddebruin: - any suggestions? Joe Filceolaire (talk) 18:07, 21 August 2015 (UTC)

I think you can easily make a proposal for organisations for this. A medium would be defined as a tool or topics used to teach something, and would apply to any teaching structure, at least. author  TomT0m / talk page 10:28, 22 August 2015 (UTC)
Like Irish and Montessori education (Q190218) using the same property? -- Innocent bystander (talk) 12:54, 22 August 2015 (UTC)
Innocent bystander I think Montessori education (Q190218) is more a teaching method (Q1813494) or pedagogy (Q7922) than a medium of instruction (Q6807178). Joe Filceolaire (talk) 16:42, 22 August 2015 (UTC)
Could we use this for Radio and TV stations and newspapers too? Joe Filceolaire (talk) 16:42, 22 August 2015 (UTC)

Group of cities and "human settlement" in "instance of" property

IS it OK to delete "instance of" = "human settlement" statement from Four Holy Cities item? It is a group, not a single human settlement. --Voll (talk) 11:52, 12 August 2015 (UTC)

In this case, it should be subclass of (P279) something or another. --Izno (talk) 13:22, 12 August 2015 (UTC)
@Izno: I don't think so, it's a similar case to class of people and music band. As a music band is a group of people, a group of cities is … ag group of cities.
< Holy cities > has part (P527) View with SQID < Jerusalem >
applies. We could if we wanted to however define a class «city of the 4 Holy cities», which could be a subclass of «city» or «Holy city» with Jerusalem as an instance, but this would be another item.
To answer Voll (talkcontribslogs) then, if this is an instance of «group of cities», then if a group of city is a special case of settlement then we should have
< group of cities > subclass of (P279) View with SQID < settlement >
and the statement is redundant because any group of city will become a settlement according to subclass of (P279) meaning an definition. What's to do then is to search the defintion of the «settlement» item. author  TomT0m / talk page 13:38, 12 August 2015 (UTC)
In my opinion a group is a class so 'Four Holy Cities is a subclass of 'city' and each of the four is an 'instance of:Four holy Cities" though that would flow better if the wikidata item for 'Four Holy Cities" was renamed "holy cities'. Joe Filceolaire (talk) 20:58, 23 August 2015 (UTC)

Syndic / Syndikus

The German Wikipedia article de:Syndikus and English Wikipedia's en:Syndic are interwiki-linked through Wikidata item Q1339249. However, they define their subject quite differently: The English article is about a term applied in certain countries to an officer of government with varying powers, and secondly to a representative or delegate of a university, institution or other corporation, entrusted with special functions or powers. German Wikipedia's introduction defines a "Syndikus" as ein Rechtsanwalt, der im Rahmen eines dauerhaften Beschäftigungsverhältnisses seine Arbeitszeit und Arbeitskraft einem nichtanwaltlichen Arbeitgeber wie einem Unternehmen, Verband oder einer Berufsständischen Körperschaft sowie Stiftung zur Verfügung stellt, that is, roughly translated into English: "a lawyer who, as a permanent employee, provides his working time and labor to a a non-lawyer employer, such as a company, association, a professional organization or foundation." This is similar to the second definition in the English article, but German Wikipedia strictly defines a "Syndikus" as a lawyer acting in the context of German law, although the section "Geschichte" (history) also describes some historical offices using this term. So, my question is: Do you think that these articles still describe essentially the same thing and should be linked through a single Wikidata item - or is the focus too different for that, and we should create another item? If the latter, we would also need to assign the other interwiki links to the appropriate item. Some seem to define a "syndic" similar to German Wikipedia, however, of course without the focus on German law. Gestumblindi (talk) 22:27, 21 August 2015 (UTC)

Look at the statements. Do these relate to both meanings or do they just relate to one meaning? If we had two wikidata items - one for each of the meanings mentioned in the en article - then how would their statements differ? If one set of statements covers both meanings then we don't need two items. If the two meanings do need different statements then create separate items for these meanings and their statements and keep the existing wikidata item with links to the new items using "has part".
Whenever the question is "to split or not to split?" then let the statements guide you. Hope that helps. Joe Filceolaire (talk) 16:35, 22 August 2015 (UTC)
Thank you, but... well, have you looked at the statements of Q1339249? Not very helpful in this case, I'd say. Because the only statement is a Freebase identifier, and this in turn only contains content from English Wikipedia. Gestumblindi (talk) 16:45, 22 August 2015 (UTC)
Very true but I was Not just referring to the statements on the item but also the statements you want to add to the item, the statements you need to describe the item (or items). Those are the statements that will guide you as to whether or a not an item needs to be split, or perhaps I should say at what point the item needs to be split. Joe Filceolaire (talk) 21:08, 23 August 2015 (UTC)

Pywikibot-Add reference

Hello. Let say that we have in a wikidata item a property with its claims. But with no reference. I try, giving the Q number and the P number to add the source (I wanted to add as a source P143 with Q11918). The code was

# -*- coding: utf-8  -*-
import pywikibot
site = pywikibot.Site('wikidata', 'wikidata')
repo = site.data_repository()
item = pywikibot.ItemPage(repo, 'Q3561900')
item.get()
references = pywikibot.Claim(repo, 'P143')
value = pywikibot.ItemPage(repo, 'Q11918')
references.setTarget(value)
for m in item.claims['P710']:
    m.addSources(references)

It works. But if the property has more than one item, it add reference only to the first one. Any ideas to solve this? To add the same reference to all items? Xaris333 (talk) 04:17, 24 August 2015 (UTC)

It is a little stupid, but yes every statement should have a reference. But then better avoid to source with greek WP because WP is not a source. P143 was used only by bot when large data import. The greek article has sources so use them in WD too in order to allow all WP to cite the original source. Snipre (talk) 07:31, 24 August 2015 (UTC)
Yes, I know. This is just an example. I can use a url as a reference. But the problem still exist. How to add the same reference to all items? Xaris333 (talk) 13:53, 24 August 2015 (UTC)
@Xaris333: Install the gadget DuplicateReferences in your Special:MyPage/common.js by copy pasting this code line:
importScript( 'User:Bene*/DuplicateReferences.js' ); // [[User:Bene*/DuplicateReferences.js]]
After you create one reference, your reload the page and you can copy-paste your reference several times. Snipre (talk) 16:15, 24 August 2015 (UTC)
Snipre, I can see and select the copy but how to paste? Xaris333 (talk) 16:56, 24 August 2015 (UTC)
@Xaris333: Clic the reference section under the statement you already add, you should have a copy beside the "edit" button, clic on it then go the statement where you want to add the reference, clic the reference section and you should see a "insert reference" beside the "add reference" button. Clic on it and refresh the page. Snipre (talk) 17:19, 24 August 2015 (UTC)
Οκ, I did it. Many thanks! Xaris333 (talk) 17:20, 24 August 2015 (UTC)

Wikidata weekly summary #172

Property for date of announcement

Do we have property for that? "inception" doesn't seem ideal. --- Jura 15:58, 24 August 2015 (UTC)

significant event (P793) View with SQID. author  TomT0m / talk page 16:18, 24 August 2015 (UTC)
There don't seem to be any. Would you have a sample? --- Jura 18:18, 24 August 2015 (UTC)
Any what ? Just add one. author  TomT0m / talk page 18:46, 24 August 2015 (UTC)
Sample (or use case). Ok, then there is currently no use of that property for this. --- Jura 08:35, 25 August 2015 (UTC)
Nope, but it's quite similar to the other use cases, just add one item "announcement date" and a line to the table and it is done. author  TomT0m / talk page 08:47, 25 August 2015 (UTC)

Quantities, and properties as target?

I saw last week's update mentioned the "quantities" data type may be out for testing soon - any updates on where and when? I did see it was already an option on test.wikidata.org (I created a "half life" property with Quantity as the data type) but trying to use it there didn't seem to be any way to specify the units, which I thought was a key feature for quantities. I would like to see how they really work in action!

Also, an unrelated question - with properties that have items as data types, is there any way to look at an item and see all the other items that point to it via their properties? I'm pretty new here so if there's an obvious way to do that I've missed it. For example, for a given class, can we somehow find all items that are instances of that class? ArthurPSmith (talk) 19:49, 24 August 2015 (UTC)

That last part should be possible with WDQ/Autolist. Mbch331 (talk) 19:59, 24 August 2015 (UTC)
Thanks! Well, it partly solves the problem - if I know what property I want to look at (eg. instance of or subclass). But I can't figure out a way to get a list across all properties. For example, what items have "organization" as the value of a claim? If I put in a query of the form "claim[*: 43229]" I get no matches. I've tried the sparql endpoints but they seem to always fail with a timeout... ArthurPSmith (talk) 20:36, 24 August 2015 (UTC)
I don't think WDQ accepts wildcards. You need to fill in a Property. Most common form is which would be in WDQ CLAIM[31:43229]. Mbch331 (talk) 20:53, 24 August 2015 (UTC)
Try "What links here" and limit to main namespace, that should do what you want (unless you want to exclude qualifiers). Popcorndude (talk) 22:41, 24 August 2015 (UTC)
Ah, that works. Good suggestion, thanks! ArthurPSmith (talk) 13:38, 25 August 2015 (UTC)
Quantities as a datatype is available already here. What's coming is support for units in the quantity datatype :) I'll post on this page as soon as we have it up for testing. --Lydia Pintscher (WMDE) (talk) 09:02, 25 August 2015 (UTC)
Looking forward to it! ArthurPSmith (talk) 13:38, 25 August 2015 (UTC)

Merging relationship properties

Time and technology having moved on, I've reopened an old discussion (from May 2013!) at Wikidata:Requests for comment/Merging relationship properties - should we go to a single "parent" and "sibling" property? Comments appreciated. Andrew Gray (talk) 20:45, 24 August 2015 (UTC)

Looking at 2013 New York Yankees season (Q276815)

This is <instance of (P31):sports season of a sports club (Q1539532)> but what property should we use to link this item to the New York Yankees (Q213417)? Joe Filceolaire (talk) 02:31, 25 August 2015 (UTC)

Some items are using part of (P361), e.g. F.C. Atlas season Apertura 2009 (Q5423671) but not sure if this is optimal. --Pasleim (talk) 09:38, 25 August 2015 (UTC)

Local SPARQL endpoint

Hoi, I missed the announcement of an endpoint for use with SPARQL. It is mentioned on the Data access article here on Wikidata. As I understand it, it is a query engine and as such it can be used in many ways. What I am looking for is examples of this functionality that I understand is making use of live Wikidata data.

We desperately need to understand what it is and what it does. For me the best part is that it can use WDQ statements, the obvious question is therefore how to use it in the toolchain developed by Magnus. I also heard about all kinds of sparkly demos. I would like to see them run on our local data.. Who can show us what this tool is there for and how it can be used. I am happy to accept blogposts on my blog. Maybe it may be posted on the official blog..

Thanks in advance! GerardM (talk) 10:36, 25 August 2015 (UTC)

this might be what you're looking for. Popcorndude (talk) 12:34, 25 August 2015 (UTC)
It's already used by {{PropertyForThisType}}. The idea is that if a property applies to the instances of a class, like human (Q5), listed by statements using the "property for this type" Wikidata property, or one of its superclass (for example animals and objects might have a "height" property), any human is an animal, so the height property applies to human as well), then it will be listed in the results of the query. author  TomT0m / talk page 12:52, 25 August 2015 (UTC)
The property has few examples and I get on the one where I can find it that LUA runs out of memory. Thanks, GerardM (talk) 13:26, 25 August 2015 (UTC)
LUA What ??? What are you talking about ? author  TomT0m / talk page 13:30, 25 August 2015 (UTC)
The sample queries on https://wdqs-beta.wmflabs.org are returning ERROR: 502 Bad Gateway. It's not LUA but shouldn't be so. --Pasleim (talk) 15:05, 25 August 2015 (UTC)
It seems the instance had a hickup. Stas gave it a kick and it hopefully works again now. --Lydia Pintscher (WMDE) (talk) 16:26, 25 August 2015 (UTC)
It's beta, so glitches are possible. Please ping me on my user talk page if it happens again. --Smalyshev (WMF) (talk) 16:27, 25 August 2015 (UTC)
this says LUA runs out of memory for me. Thanks, GerardM (talk) 17:36, 25 August 2015 (UTC)

Whither Wikispecies?

It has always struck me that Wikispecies is the closest project to Wikidata in terms of having structured data that could easily be imported, and whose functionality could most easily be recreated using Wikidata content. Yet it's low down the list of projects to be worked on. Why is that? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:43, 15 August 2015 (UTC)

Last time we asked Lydia, the answer was "politics". --Izno (talk) 18:38, 15 August 2015 (UTC)
Aside from politics, there's also the issue of integrating the classification scheme at Wikispecies with the many and disparate classification schemes in use on all the various Wikipedias. It's not a simple matter to do that, when the same group of plants may have two separate data items on Wikidata because the Wikipedias choose to apply different names and use different classification systems. Add to that the fact that Wikidata's data items on those organisms are being heavily edited by people banned from Wikispecies (and from several Wikipedias and Wiktionaries), and you get more politics. --EncycloPetey (talk) 00:30, 16 August 2015 (UTC)
The lack of active users is probably not helping either. Multichill (talk) 09:52, 16 August 2015 (UTC)
Wikispecies had a dominating sysop who drove many people away before finally being stripped of his adminship. That's kept the numbers down. Wikispecies editors also need a very particular set of skills and specialist taxonomic knowledge to do more than the most basic editing. Such people are not very common, so the numbers will only grow slowly. Even then, many of those users will not have the dual knowledge it would take to interface with Wikidata. Many potential users who do how the requisite knowledge are working on other species database projects instead. Wikispecies is a much more narrow niche for editing than Wikipedia or Wikidata. --EncycloPetey (talk) 15:47, 16 August 2015 (UTC)
@Pigsonthewing: You bring up a good question. In addition to the good points raised above, there are users there who are also concerned about the project essentially becoming redundant and who don't want that because it's either 1.) a personal pet project, 2.) they think it can provide something more and greater than Wikidata, or 3.) some combination of the two. —Justin (koavf)TCM 15:47, 19 August 2015 (UTC)
Down in their heart the people of Wikispecies feel that this project will be redundant within a few years. The German Wikipedia for example never accepted links to Wikispecies because the German Wiki is based on different data and the quality problems of Wikispecies were never solved. Wikispecies is also limited in usablitiy as a monolingual project, that expects all possible users to be able to contribute in English and only in English.--Giftzwerg 88 (talk) 18:08, 26 August 2015 (UTC)

Central login (2)

Central login does not appear to work with google chrome anymore. And it does not allow one to login either. Not sure what the issue is. Central login with firefox however still works. I get a notice that says "Central login You are centrally logged in. Reload the page to apply your user settings" But when I release with shift F5 nothing happens. User:Doc James 24.66.183.25 03:58, 25 August 2015 (UTC)

There is a section on this page with the exact same problem. Please search before you ask. Sjoerd de Bruin (talk) 06:49, 25 August 2015 (UTC)
Thanks yes here [[12]] 24.66.183.25 18:35, 25 August 2015 (UTC)
Okay worked Doc James (talk · contribs · email) (if I write on your page reply on mine) 18:38, 25 August 2015 (UTC)

If you can still reproduce this, we need to know your cookies for wikidata.org and subdomains to be able to fix this (before you delete them). (At least the cookie names/keys, don't publicly publish the values. You can send me a private message through Phabricator or on wiki if you want to be sure to not accidentally publish something private.) - Jan Zerebecki 17:42, 26 August 2015 (UTC)

PLbot disaster

etc EfrinEfrin (talk) 19:53, 25 August 2015 (UTC)

I miss some information. Why is this a disaster? Please, assume good faith. Sjoerd de Bruin (talk) 20:00, 25 August 2015 (UTC)
Hi EfrinEfrin or should I just say Tamawashi? I think it's time to lock this account down. Multichill (talk) 20:40, 25 August 2015 (UTC)

EfrinEfrin, Where exactly Pasleims Bot failed and created a disaster (Q3839081)? --Succu (talk) 21:22, 25 August 2015 (UTC)

For the first diff the German WP reads "ist eine Eparchie der mit der römisch-katholischen Kirche unierten Syro-Malabarischen Kirche mit Sitz in Cicero in den Vereinigten Staaten." If it is only "uniert" it is not "römisch-katholisch" itself. I know, other viewpoints may exist. From German WP correct would be "Eparchie der Syro-Malabarischen Kirche".
The other failure is, that the bot does not show by which rule it worked. So, one cannot check where else these claims might have been added. Same with Widar - one has a rule in Widar, run it, but nobody can see what rule one did apply. So, EfrinEfrin (talk) 21:35, 25 August 2015 (UTC)
Thank you Tamawashi. Multichill block him. --Succu (talk) 21:44, 25 August 2015 (UTC)

German-Dutch censorship collaboration. Blockwarts ahoi! EfrinEfrin (talk) 01:40, 26 August 2015 (UTC)

  • Socking confirmed through my enwp CU bit. Blocked. Courcelles (talk) 12:04, 26 August 2015 (UTC)

Proposal for rewriting WD:Glossary

I have posted a proposal for rearranging and rewriting WD:Glossary at Wikidata talk:Glossary#Proposed reorganisation. I'll wait a day before starting in case anyone has comments. Joe Filceolaire (talk) 17:25, 26 August 2015 (UTC)

is everybody dead here ?

Hello, (echo echo echooooooooo.....)

hey, 20 hours with no modif' - where is everybody gone ? :D --Hsarrazin (talk) 12:50, 23 August 2015 (UTC)

I seriously wonder if there is a critical mass to get things done here. Lots of issues get brought up, and are just left unresolved. Jc3s5h (talk) 13:14, 23 August 2015 (UTC)
Everybody is busy contributing .. so there is no time for chat .. --- Jura 13:16, 23 August 2015 (UTC)
Yup, working very hard. Same applies to the developers, who are doing the best they can. Complaining about them doesn't help. Sjoerd de Bruin (talk) 13:30, 23 August 2015 (UTC)
@Hsarrazin, Jc3s5h, Jura1: Per en:project:Deceased Wikipedians (Deceased Wikimedians (Q4655446) View with Reasonator View with SQID on your wiki if available), these deceased users were logged in Wikidata (but what? What thing should be done here?):
Deceased users on Wikidata (links to their CentralAuth information)

--Liuxinyu970226 (talk) 00:11, 27 August 2015 (UTC)

How to treat 'daughter articles'

This is a bit of a follow on to Wikidata:Project_chat#Entity, article, item above but on a related and broader question.

On wikipedia, when an article gets too long, some of the content may be split off into 'daughter articles' covering things like 'History', 'economics', lists, etc. Articles about the same subject in other languages and other similar but shorter articles, will include all this information in the one article. How should we treat the wikidata items associated with these 'daughter articles'?

  1. Use instance of (P31) or main subject (P921) or facet of (P1269) to link the daughter article to the main article?
  2. Use part of (P361)has part (P527) to link the main item to the daughters?
  3. Where should we put statements? If the city of Foo has a daughter article about the 'economy of Foo' then should we put all the statements about the economy of Foo on the item for the daughter article even though there are only three languages which have that daughter article?
  4. Should we put the statements about the economy of Foo on the item about Foo because those are statements about Foo and wikidata doesn't tell wikipedia how to organise it's articles and wikipedia doesn't tell wikidata how to organise it's items.
  5. What about the economy of the city of Bar which doesn't have an 'economy of Bar' daughter item? Should we create such an item just to hold the statements about the economy?

Personally I say: 1 Use P1269. 2 Yes. Use 'part ofhas part'. 3, 4, 5 Put the statements in the main item. Joe Filceolaire (talk) 17:15, 26 August 2015 (UTC)

I'm not sure what you mean by point 2. Would you have Q954 part of (P361) Q1468532? --Yair rand (talk) 18:29, 26 August 2015 (UTC)
User:Filceolaire, are you sure you don't mean has part (P527)? I quite agree with the rest, though we could concievably put the statements on both, or maybe some of them. Popcorndude (talk) 19:43, 26 August 2015 (UTC)
Ooops. Fixed. Joe Filceolaire (talk) 20:35, 26 August 2015 (UTC)
Unless we have a clear consensus of dividing a topic into subtopics among all wikipedias I think WD shouldn't take care of these subarticles. WD is a database meaning we need a structure which can be applied to all elements of the same group in order to perform querys. Snipre (talk) 20:56, 26 August 2015 (UTC)
Snipre: We have a clear consensus that wikidata will handle sitelinks and language links for all wikipedia articles even if the different language wikipedias themselves have no consensus as to how they organise articles. That means we will have items for all these daughter articles so that we have somewhere to put their sitelinks. Now we need a consensus on what statements to add to those items. That is what my proposal above is about. Joe Filceolaire (talk) 21:40, 26 August 2015 (UTC)
Point 1: I agree with using facet of (P1269). Point 2: Not sure. has part (P527) is rather ambiguous. Maybe one-way links is enough? Point 3: I think we might need to deal with this issue on a more case-by-case basis. I don't recommend having GDP or other similar markers in the economy items, but I don't think all data should be barred from all daughter items. --Yair rand (talk) 22:34, 26 August 2015 (UTC)
ontology speaking 
I'm not sure part of (P361) / has part (P527) are not really appropriate here, it's pretty clear an organisation has organisation parts, but a city is something somewhat more blurry : geographical features, buildings, population, activities of the populations, administration and organisation, all of these interacting, almost an organism by itself :) ... economy informations seems to me informations about the human activities (usage of the resources, transaction beetween human) ... (Searching briefly I see there is whole books related to those kind of questions : https://books.google.fr/books?id=Du_uKN1yAqwC&pg=PA168&lpg=PA168&dq=%22part+of%22+ontology&source=bl&ots=Tv_EEuOSQC&sig=lw5Mz2WAQMTW39AuG2u-8NHCHzA&hl=fr&sa=X&ved=0CGsQ6AEwCWoVChMI1Or9tt3IxwIVhNIaCh2nhQhK#v=onepage&q=%22part%20of%22%20ontology&f=false ) author  TomT0m / talk page 07:34, 27 August 2015 (UTC)
  • Supposedly Wikipedia would attempt to divide subjects into meaningful units. "Economy of bar" might be a subject or field of study by itself. As such it would need appropriate statements for P31. Lazy solution is to add "named after" "bar". --- Jura 07:40, 27 August 2015 (UTC)

Anyone understand how Open Street Map WIWOSM works?

OSM has started tagging OSM node, way, area, and relation with "wikidata=Q########" to link these to their associated wikidata item and (as far as I can tell) their proposal is that we link to these using a search as described on WIWOSM.

Do we need a "WIWOSM" property with a string value which happens to match the item Q##### number and a 'formatter url' which turns this into the WIWOSM query? Or have I misunderstood? Andy Mabbett? Sorry if this is a bit incoherent. Joe Filceolaire (talk) 01:08, 27 August 2015 (UTC)

@Filceolaire: i may be wrong, but "wikidata=Q########" looks like OSM query need to use our own already existed entity ID as search key. I.e. no new property neither any mapping is required. The proposal/possibility is to get rid of dozens "wikipedia=lang:article" links pointed to single object and replaces them with single link using existing wikidata ID. -- Vlsergey (talk) 08:04, 27 August 2015 (UTC)
No we don't need to do anything. Well, we can help by showing appreciation to OSM mappers, and welcoming them here ;-) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:45, 27 August 2015 (UTC)
I've just made Wikidata:OpenStreetMap (shortcut: WD:OSM]) to facilitate cooperation. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:00, 27 August 2015 (UTC)

importing Project Gutenberg ebook IDs

(I originally wanted to send this message on the mailing list, I tried both wikidata@lists.wikimedia.org and wikidata-l@lists.wikimedia.org, but it never worked while never returning 'bad recipient' error or anything, so here it is)

Hi! Since last Sunday, there is a property to map ebooks ids from gutenberg.org: P2034 Given gutenberg.org doesn't have an API and doesn't like robots, has anyone elaborated a strategy on how to import all those ids? How was it done for authors? P1938

Here is what I was thinking to do if no better alternative emerge:

Gutenberg as mirrors including ftp directories such as this one, on which text files serving as indexes reference the books like this:

Boyhood, by Leo Tolstoy                                                   2450
  [Translator: C.J. Hogarth] (Note alt. author spelling: Tolstoi)
Dec 2000 The Common Law, by Oliver Wendell Holmes, Jr. [#3][cmnlwxxx.xxx] 2449
Dec 2000 The Colored Cadet at West Point, by Henry Flipper [ccawpxxx.xxx] 2448
Dec 2000 Eminent Victorians, by Lytton Strachey            [mnvctxxx.xxx] 2447

with the end of line number being the expected P2034 id.

It doesn't look like any standard format I know, so I was thinking to write a custom parser, extract the titles and authors, search the title on Wikidata API and then do a sort of custom Wikidata Game (?)

context: I'm very interested in this property as it can potentially allows inventaire.io to display a nice "download ebook" button on some 49000 books pages with a snap of the fingers \o/ any suggestion welcome! Bests,

Zorglub27 (talk) 10:16, 27 August 2015 (UTC)

The matching is already being done in Mix-n-Match. --Lydia Pintscher (WMDE) (talk) 10:21, 27 August 2015 (UTC)
nice!! I see "Project Gutenberg Authors" but not ebooks though -- Zorglub27 (talk) 10:36, 27 August 2015 (UTC)

Limitation for Wikidata Query Service Beta

Does somene know if there is a size limitation for query using Wikidata Query Service Beta ? I am looking for data with a potential of 15000 items and I want to extract for each of them 5 values. Is it too much ? Snipre (talk) 18:18, 27 August 2015 (UTC)

API breaking change (9th September)

Hi all! Please note there is an API breaking change (mainly for XML) scheduled for 9th September 2015. To read about it please see User:Addshore/API_Break_September_2015. This has also been announced to the relevant mailing lists and will be announced in the weekly summary. Any questions just ask! ·addshore· talk to me! 14:39, 27 August 2015 (UTC)

@addshore:: hi, I'm not sure to catch what is new in the Sample JSON changes https://www.wikidata.org/wiki/User:Addshore/API_Break_September_2015 (out of the fact that "Douglas Adams" will be renamed "Death Star" of course). Does it has do to with claims, aliases, and descriptions not being in the "before" and being in the "after"? -- Zorglub27 (talk) 15:45, 27 August 2015 (UTC)
Yes, basically before the change there would have been attempts to remove empty parts of entities (in some modules). Now no such attempt to hide these empty parts will be made. ·addshore· talk to me! 07:57, 28 August 2015 (UTC)

Constraint, please

Please could someone add constraints to ORCID iD (P496) (which was founded in 2010, and is only available to living people), that throws a warning if the subject's deathdate is before 2010; or their birthdate is before, say, 1875. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:42, 27 August 2015 (UTC)

Dharmasiri Gamage (Q18750235) died 2004 but has an ORCID. --Pasleim (talk) 20:55, 27 August 2015 (UTC)
Another example that ORCID is not a valid site for authority control. Sjoerd de Bruin (talk) 12:04, 28 August 2015 (UTC)
What utter tommyrot. There are more VIAF errors on Wikipedia than ORCID IDs. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:01, 28 August 2015 (UTC)

Joseph Reagle not an author?

According to some, "author" or "auteur" in French is very broad. My understanding is that "auteur" is more like "creator". The problem is that because of this point in the French language, Mr Reagle and many others is no longer an author. Obviously such language issues endanger the core of Wikidata. When because of the understanding or the use of a particular label Wikidata content is to be changed, it is obvious that every language will pose us with challenges.

The whole notion of the use of labels is very much not unique. This example is just another practical issue why Wikidata does not have a proper strategy for dealing with semantics in multiple languages. What to do? Thanks, GerardM (talk) 05:31, 28 August 2015 (UTC)

Would writer (Q36180) have too narrow a meaning in French? In some languages like English "writer" is a generalization of "technical writer", in others it might be more along "homme des lettres" and emphasize literary quality. Also, there are things to broad to be considered a profession: Would you state
< ... > occupation (P106) View with SQID < creator (Q2500638) View with Reasonator View with SQID >
in your favorite language?
< ... > occupation (P106) View with SQID < author (Q482980) View with Reasonator View with SQID >
??
< ... > occupation (P106) View with SQID < writer (Q36180) View with Reasonator View with SQID >
??? It's not only the problem of the proper object to choose but also one with slightly diverging meanings of the property they are used for. Worst of all, the labels for the objects are choosen to help differentiating them and this sometimes seems to preclude the use in certain contexts (looking at the synonyms most languages cannot keep up the three-tiered distinction of your example or rather they do it differently and therefore Wikidata will never be able to keep that distinction clear). -- Gymel (talk) 07:05, 28 August 2015 (UTC)
@GerardM: My understanding is that the item "author" is related to the relation between a text and the person who created it, like Victor Hugo wrote "Les misérable" (the item corresponding to the "author" property), and not about person who writes text for a living. A relationship is not an occupation. author  TomT0m / talk page 10:22, 28 August 2015 (UTC)
@TomT0m: You will find that many people have author as an occupation. In the mean time, an "essayist" is used on Mr Reagle and he is more than that. Thanks, GerardM (talk) 11:10, 28 August 2015 (UTC)
@GerardM: This is not about Mr Reagle being an author at all, this is about "is the item used an item about the occupation of being an author". And I don't think it was. author  TomT0m / talk page 11:15, 28 August 2015 (UTC)
Mmm Looking at this item it seems it was an occupation actually. I don't understand the issue then. author  TomT0m / talk page 15:25, 28 August 2015 (UTC)

m:Requests for comment/Global ban for Tobias Conradi

Per m:Global bans, I am notifying the project of this proposal. Everyone is welcome to go and voice their opinion of the proposal and about the user in general. This user is also known as Tamawashi here.--GZWDer (talk) 09:38, 28 August 2015 (UTC)

Relationships between academic and research organizations - just use part of (P361)?

I'd like to see the relationships between academic and research organizations encoded more systematically in wikidata, but I'm not sure this requires new properties or just applying "part of" (P361) perhaps with qualifiers. Right now this seems very ad hoc. For example, the colleges of the University of Cambridge and University of Oxford have their own special class (for example college of the University of Cambridge) of which each individual college is an instance, and the class is declared to be "part of" University of Cambridge. So there is no direct "part of" relationship between for example Trinity College and University of Cambridge but at least an indirect relationship through the class. For University of California on the other hand, there is at least one campus with a direct "part of" relationship (Davis), but others use "parent company" (P749) - for example University of California, Los Angeles - and that isn't even in a sub-property relationship with "part of" ("parent company" is a sub-property of "owned by" which has no further sub-property relationship defined). Meanwhile UC Berkeley has no relationship defined with University of California at all right now. There is also complexity in relationships with research institutes at universities - for example the Beckman Institute for Advanced Science and Technology has no defined relationship with the University of Illinois, Urbana Champaign in wikidata, despite being a research unit of that university. My general inclination on this is to recommend new properties - "academic unit of" and "research unit of" which are sub-properties of "part of", and to recommend to consistently apply them for these sorts of relationships. There are a large number of templates in wikipedia now that include detailed relationships of this sort - for example the Template:University of Illinois at Urbana–Champaign campus page for English wikipedia - though they include some even finer subdivisions of the relations. Anyway, the data is there in wikipedia and elsewhere, if we could figure out a way to usefully record it here. If I do any work on these soon I expect I'll stick with "part of" for now, but I would like some advice or suggestions on if there's a better way to do this. Thanks! ArthurPSmith (talk) 17:45, 26 August 2015 (UTC)

@ArthurPSmith: A class should not be a part of an instance, it's the college who should be. A class can be part of another class say
< Ca > part of (P361) View with SQID < Ca >
, meaning that a Ca instance is a part of Cb instance. Otherwise part of (P361) would be ambiguous ... Here the every college should be a part of the university. As far as I remember this was discussed in WD:RfC/Refining part of for example ... but you're right there should be a better help page as Help:BMP is limited. author  TomT0m / talk page 18:44, 26 August 2015 (UTC)
ArthurPSmith, TomT0m: A class can be a 'subclass of' another class but it would be unusual for it to be a 'part of' a bigger class. A 'College' is not a 'research institute' is not a 'university' so a 'college' or a 'research institute' is not a 'subclass of:university'. Each individual oxford college is 'part of:Oxford University' and is an 'instance of:Oxford College'. I think UCB is 'part of:University of California'. For Research institutes I suspect their relationships with the Universities are more varied with only some being owned by the University. There is nothing to stop you having both 'part of:named University' and 'parent company:named University' so it can be found either way. Maybe 'parent company' should be a sub-property of 'part of' as well as 'owned by'. Hope that helps. Joe Filceolaire (talk) 20:53, 26 August 2015 (UTC)
TomT0m Joe Filceolaire so should I go ahead and remove the "part of" claim for college of the University of Cambridge (a class) relative to University of Cambridge (an instance of university) and add "part of" claims for all the individual colleges and other relationships of that sort? Is there some other property that would could still indicate the relationship between the class of Cambridge colleges and the instance as a university? And is this a case where it might be worth looking into recommending new properties for these relationships? ArthurPSmith (talk) 14:22, 27 August 2015 (UTC)
ArthurPSmith: Help:Basic_membership_properties says that if every 'instance of:class A' is a 'part of' an 'instance of:class B' then you can say 'class A:part of:class B' but I've never really agreed with that. YMMV. Joe Filceolaire (talk) 20:03, 27 August 2015 (UTC)
By the way, here's the correct link for WD:RfC/Refining part of. I think I need to read a bit. I had read Help:BMP and I think it's pretty good actually. Though perhaps the hypothetical examples should be replaced by real ones to improve clarity. I found the class examples particularly confusing - if something is sometimes part of another class but sometimes not, should it be indicated by "part of"? ArthurPSmith (talk) 15:16, 27 August 2015 (UTC)
I've created a modified version of the Help:BMP page here: User:ArthurPSmith/Test/Help:Basic membership properties - comments welcome, I expect to continue working on this next week, it's definitely not final. I also made some notes on the discussion on TomT0m's RfC for that page. One thing that came to me though while looking into this more is a similar case to the <college of the University of Cambridge> (class) - "part of" <University of Cambridge> (non-class) case. The second example along those lines is the relationship between inner planets (a class) and Solar System (an individual object, not a class). There is a "part of" claim there, and it seems entirely appropriate to me. Similarly the asteroid belt, and other classes of solar system objects. I think this is a common case, where you have a collection of entities that form a class X, but if all of those individual entities are also "part of" another individual entity Y, it just makes sense to consider the class X to be "part of" the non-class Y. There definitely are at least these 4 or 5 cases in wikidata now - would it be worth digging up more examples? ArthurPSmith (talk) 21:05, 28 August 2015 (UTC)

Fusion problems

Can someone fusion en:Category:People from Kłodzko County (Q13283590) with German de:Kategorie:Person (Glatzer Land) ? MarinaPoltaswa (talk) 23:10, 28 August 2015 (UTC)

✓ Done - Mbch331 (talk) 05:37, 29 August 2015 (UTC)

Isidro A. T. Savillo (Q18207526)

I don't know what to say... just have a look. Matěj Suchánek (talk) 18:48, 28 August 2015 (UTC)

+w:Talk:Isidro A. T. Savillo. --- Jura 09:02, 29 August 2015 (UTC)
Interesting use of the "reference" elements: Q18207526#P106. --- Jura 06:31, 30 August 2015 (UTC)

Offer to serve as an administrator

I am offering to serve as an administrator again. Please see Pigsonthewing 2. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:24, 29 August 2015 (UTC)

Q404

user:Addshore has created Q404 as a redirect to mathematics (Q395), which has caused Pywikibot tests to fail. https://phabricator.wikimedia.org/T110819 . I thought it was intentional (and cute) that 404 was missing, symbolising w:HTTP 404. Could it be deleted again? John Vandenberg (talk) 01:31, 30 August 2015 (UTC)

Well, I guess I will leave this up to the community in general! Technically this should be a redirect, it was created to represent one concept, and that concept is now stored at another location. However I'm sure we can let one slide ;) ·addshore· talk to me! 06:17, 30 August 2015 (UTC)
But the concept it was originally intended to represent has always been a duplicate. If any duplicate can be restored, how can we assume that any item wont be restored? i.e. if we change this test to use a different deleted item, is it only a matter of time before it is also undeleted? John Vandenberg (talk) 07:43, 30 August 2015 (UTC)
You can always make a test item and let a sysop delete it. Sjoerd de Bruin (talk) 07:54, 30 August 2015 (UTC)
@John Vandenberg: if you need an item that does not exist: Q7. Or do you need an item that is deleted? Can look up an item with a very low id if you need one. Multichill (talk) 08:49, 30 August 2015 (UTC)
That works as a substitute. Thanks, John Vandenberg (talk) 08:58, 30 August 2015 (UTC)

Q1538552

This may no be the right place for a request like this, but after searching the help pages, I couldn't find an appropraiate page. I don't have the proper permission to do this myself, so could someone please fix Q1538552. Right now, it says "No label defined", sp it doesn't show up properly in WP-biography boxes. Cheers! Asav (talk) 09:53, 30 August 2015 (UTC)

You should have the rights to add a label. Just added the English label, please verify it's correct. (And if other languages are needed, please provide language and label in that language.) Mbch331 (talk) 09:59, 30 August 2015 (UTC)
Thanks! I wasn't aware that the English label is the one that defines the overall entry. I thought it was a different entity altogether. But now I know better... Asav (talk) 10:19, 30 August 2015 (UTC)
@Asav: it's not the neglish label who defines the entity :) it's probably the language the more people can translate from/to though, that's why it is important. author  TomT0m / talk page 14:55, 30 August 2015 (UTC)

Proposal for notation

There have been a number of proposals for how to show statements in discussions and I think I have tried most of them. The one I like best is:

  • <property:value> (property:value) {property:value}
Where <property:value> is a property value pair applied to the item; (property:value) is a qualifier; {property:value} is a reference. All are on the same line to signify which <property:value> the qualifier and reference apply to.
Where an item has a property with two values then have <property:value> twice on different lines, each with it's own qualifiers and references.
This is what I will be using from now on - unless anybody else has a better idea?
Anyone want to make three templates for these? Joe Filceolaire (talk) 16:27, 30 August 2015 (UTC)
Well, we already have {{St}} (simple: no qualifiers and references) and {{C}} (more complex: qualifiers, but still no references). I haven’t yet found the need to embed a statement with references. —Galaktos (talk) 16:46, 30 August 2015 (UTC)

Best practice?

Hi, I am curious: what is the best way to include information about a book? As you can see in Wikipedia: The Missing Manual (1. ed) (Q428213), I have added the publication date, main subject, ISBN number, etc. in two different ways: as a property of the main item, and also as a qualifier of the "book" property. Which is the correct way? -Pete F (talk) 17:19, 17 August 2015 (UTC)

I believe putting properties directly on the item is prefered. Popcorndude (talk) 17:26, 17 August 2015 (UTC)
The second structure can be used for books which won't be used in the future anymore. But as you can forseen if this will be the case better use the first structure with properties. Snipre (talk) 19:09, 17 August 2015 (UTC)
OK, thank you both for the quick response. Snipre (talkcontribslogs), I don't really understand what you mean by "books which won't be used in the future anymore" though -- used for what? What is an example of such a book? -Pete F (talk) 22:06, 17 August 2015 (UTC)
The second version violates quite a few constraints, and, in my opinion, doesn't make any sort of sense. I recommend never using it or anything similar to it. Qualifiers don't do that kind of thing. --Yair rand (talk) 22:10, 17 August 2015 (UTC)
Pete F Note that there is some discussion as to whether this is an instance of a 'book'. Isn't it a bunch of books i.e. a subclass of 'book' and an instance of user guide (Q1057179)? Joe Filceolaire (talk) 15:14, 18 August 2015 (UTC)
Thanks Filceolaire (talkcontribslogs), but I'm not sure I understand. Broughton's book was published (by O'Reilly) on dead trees. It was subsequently given a free license and uploaded to Wikipedia so that it could be kept up to date etc. Is there really any doubt that the main thing referred to is a "book"? -Pete F (talk) 01:52, 26 August 2015 (UTC)
No. The statements on this item describe a "literary work". An "instance of:book" is the copy I have on my bookshelf and would have properties like "owner", "location". Wikipedia does actually have some articles about individual copies of books - the "Lincoln Bible (Q1816474)" for example. I do agree with you that there are an awful lot of 'literary works' on wikidata with an "instance of:book" statement but (in my opinion) most of these should be "instance of:literary work" or one of the many subclasses of "literary work" such as "user guide". They can be "subclass of:book" as well I suppose - unless they are ebooks. At least that is my opinion. Joe Filceolaire (talk) 16:34, 26 August 2015 (UTC)
"book" has multiple meanings (wikt:en:book#Noun), you can say that something is a book (meaning 2) and that a particular copy of it is also a book (meaning 1). Trying to use "literary work" (which I think is broader in meaning anyway) instead is unlikely to stop people from continuing to use the existing "book" item, because it is an instance of a book (meaning 2) and people will naturally try to say that. I think a better option would be to have two "book" items, one for each meaning. - Nikki (talk) 17:46, 26 August 2015 (UTC)
Yes Nikki (talkcontribslogs), that does seem like the approach that would help other users avoid the mistake I made -- and also learn something about Wikidata's way of organizing things in the process. -Pete F (talk) 18:36, 31 August 2015 (UTC)
@Peteforsyth: "instance of: book" can be used only when you refer to an existing copy of a book. The element you create is refering to all copies and should be defined as "instance of: edition". Then when several editions exists you can create another level to link the different editions with "instance of: work". This is the FRBR classification.
So the element you created refers to one edition an dnot to one specific copy so I correct the instance of with edition Snipre (talk) 17:13, 26 August 2015 (UTC)
book (Q571): "medium for a collection of words and/or pictures to represent knowledge, often manifested in bound paper and ink, or in e-books". Some ~57,000 books (not the physical object) are currently listed as instance of (P31) book (Q571). A single edition of a book should be linked using has edition (P747)/edition or translation of (P629), and should have instance of (P31) version, edition, or translation (Q3331189). --Yair rand (talk) 18:00, 26 August 2015 (UTC)
Hello, here are the preconisations of Books project : it says that items for works should have book (Q571) as P31 - but I agree with Snipre here : a book (Q571) is a medium therefore an object (or an e-object) : thus it is the level of an version, edition, or translation (Q3331189), not of a literary work (Q7725634), that can have many editions… a medium can have only 1 publisher, 1 date, 1 isbn => it is an edition, not a work…
honestly, I've ceased to work on those, for now, because book (Q571) is not satisfactory for work level. — this does not match FRBR… --Hsarrazin (talk) 18:57, 26 August 2015 (UTC)

This is all very helpful -- thank you all for helping me understand the way Wikidata "thinks" about these matters. I see there has been some work on the example I initially brought up: Special:Diff/243715381/245655153 However, if there is only going to be one Wikidata item for Wikipedia: The Missing Manual (even in the short term), wouldn't it be better to make that item about the literary work, and not the edition? It seems that as you have worked to resolve the artifacts of my confusion on that, you have gravitated toward the less useful choice. Or am I missing something? -Pete F (talk) 18:34, 31 August 2015 (UTC)

HTML markup in descriptions

Just encountered & quot; in couple of English descriptions what may be mistakes of early import. Will be good idea to run bot. --EugeneZelenko (talk) 02:38, 30 August 2015 (UTC)

Interesting. On which items? Can you identify the edit that added them? ·addshore· talk to me! 06:15, 30 August 2015 (UTC)
Also take a look at phab:T108533. That seems to be a source of urlencoded junk too.
So I expected a lot of items being affected by this, but it's not that much.
MariaDB [wikidatawiki_p]> SELECT COUNT(*) FROM wb_terms WHERE term_text LIKE "%&quot;%" LIMIT 1;
returns 580 results. If this is a bot related problem I would expect way more results or someone already did a clean up run and just missed a couple. Probably best to copy this request over to Wikidata:Bot requests. Multichill (talk) 08:45, 30 August 2015 (UTC)
I had cleaned-up a few en-labels some time ago .. --- Jura 11:18, 31 August 2015 (UTC)

Wikidata is like

Wikidata is like Father of Wikimedia family.--Sonia Sevilla (talk) 07:13, 31 August 2015 (UTC)

Given his birthdate, I would rather say the whizz kid. _Zolo (talk) 09:39, 31 August 2015 (UTC)

Wikidata weekly summary #173

Shooting range register

Hello, I want a functionality (for instance on Wikipedia or an external site) where you can search for the nearest en:shooting range. I had in mind a search function where you can filter by country, counties and so on until you get close enough, with an interactive map showing the alternatives. Therefore I've laid a foundation by collecting information about many shooting ranges in Norway, see no:Skytebaner i Norge or en:Shooting ranges in Norway. The information is now stored in lists in the articles, but is there a way to move the information to Wikidata? And how can the data be used for such a search funtion I had in mind? I have little knowledge of how Wikidata works, but I've heard "arbitrary access" or "query" might be some appropriate tools. Thanks, Sauer202 (talk) 19:51, 29 August 2015 (UTC)

Hell no. The notion that Wikidata is abused for firearms aficionados is appalling. Please do not even think about it. Thanks GerardM (talk) 19:55, 29 August 2015 (UTC)
I'm sorry, can you please explain what you mean?Sauer202 (talk) 19:59, 29 August 2015 (UTC)
The notion that Wikidata would be useful for people who care about guns and firing them is not something I would consider worthwhile. It is the worst proposal I have heard in a long time. GerardM (talk) 20:53, 29 August 2015 (UTC)
I am sorry that we disagree, as I don't see the problem of listing sport venues on Wikipedia. Sauer202 (talk) 21:03, 29 August 2015 (UTC)
GerardM, please stop pushing your own opinion here. Sjoerd de Bruin (talk) 21:08, 29 August 2015 (UTC)
If all the shooting ranges have items and all of those items have coordinate location (P625) and instance of (P31): shooting range (Q521839), then you could query them on AutoList with "claim[31:521839] and around[625,lat,lon,rad]", which should find all ranges within rad km of (lat, lon). Popcorndude (talk) 20:08, 29 August 2015 (UTC)
Thanks! So it's just a matter of plotting the correct information into the Query editor? I tried to use it with no luck, but I'll try to read up on the Wikidata FAQ. Will every range listed with coordinates show up, even though there's no article for each range? Sauer202 (talk) 21:03, 29 August 2015 (UTC)
Probably nothing came up because none of the items within the specified radius have enough data. I entered just the "claim[]" part and got 11 items. Popcorndude (talk) 21:12, 29 August 2015 (UTC)
I also get 11 items when I only paste the "claim[]" part. All the hits have their own Wikipedia article, for instance en:Løvenskiold shooting range. Is there a way to include those ranges listed which don't have their own article? Sauer202 (talk) 21:21, 29 August 2015 (UTC)
Here are some maps from the Wikipedia lists, most of which doesn't have their own article. How to get them into Wikidata?
Sauer202 (talk) 21:52, 29 August 2015 (UTC)
I would consult Wikidata:Notability, just to be safe, then use Special:NewItem to create each of them or possibly QuickStatements to do all of them at once. Popcorndude (talk) 22:52, 29 August 2015 (UTC)
@Sauer202: This is something that may be better suited to using the data (already) in OpenStreetMap; not a Wikimedia Foundation project, but often described as "the Wikipedia of Maps". It has nearly 7K items tagged Tag:sport=shooting Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 07:22, 30 August 2015 (UTC)
Yes. What Andy said. If you add thousands of non-notable shooting ranges to wikidata they will almost certainly get deleted. Wikidata is not a place to have items for each individual shooting range but that is exactly in the scope of Open Street Map and they have the tags already in place.. Joe Filceolaire (talk) 09:55, 30 August 2015 (UTC)
Thanks a lot, I will look into OSM. Is there a way to give each range a unique ID that can be used for instance in both articles on Wikipedia and OSM? I was hoping Wikidata maybe could provide such ID's, which then could be connected to associated clubs, opening hours, etc. like found in no:Mal:Infoboks_skytebane. Sauer202 (talk) 11:59, 30 August 2015 (UTC)
Only by creating a Wikidata item for each, and tag the item in OSM with the Q number. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:17, 31 August 2015 (UTC)