Shortcut: WD:PC

Wikidata:Project chat

From Wikidata
Jump to: navigation, search
Wikidata project chat
Place used to discuss any and all aspects of Wikidata: the project itself, policy and proposals, individual data items, technical issues, etc.
Please take a look at the frequently asked questions to see if your question has already been answered.
Also see status updates to keep up-to-date on important things around Wikidata.
Requests for deletions and merges can be made here.

IRC channel: #wikidata connect
On this page, old discussions are archived. An overview of all archives can be found at this page's archive index. The current archive is located at 2015/05.

Here is a complicated one.[edit]

The childrens book series Animal Ark ( Q4764759) is published under the collective pseudonym Lucy Daniels ( Q19872456) in the UK representing the group of authors who wrote these books under the direction of Ben M. Baglio ( Q4886094). So far so good.

When these books were republished in the USA the pseudonym was changed to Ben M. Baglio so now

  1. Ben M. Baglio is both the manager of the books and is also the collective pseudonym for the group of authors (which may or may not include Ben M. Baglio the man)
  2. the UK and the USA editions of these books have different authors.

How many item do we need to create to represent this?

This is not a trivial issue. 'Lucy Daniels" has sold more books in the UK than J. K. Rowling! Filceolaire (talk) 08:33, 9 May 2015 (UTC)

What do you want to know ? How many items for the authors group/manager or for the editions of the books ?
As far as I understand the problem, two items are necessary to represent Ben M. Baglio as group/manager and as person. Then three item for each book: one for the work without author, one for the UK edition with the item of Lucy Daniels ( Q19872456) as author and one for the US edition with the item of Ben M. Baglio (group) as author. Work item should have only the common information valid for all editions. --Snipre (talk) 18:50, 9 May 2015 (UTC)
@Snipre:, @Filceolaire:, couldn't we use qualifiers to reduce the number of items? --Denny (talk) 15:57, 16 May 2015 (UTC)
@Denny:, No we can't if we consider translations too: we can have a second edition in French and in English and then we can have several editions in different publishers in different languages for very famous books. This can't be solved by qualifiers. For simple cases yes, you can mix and use qulaifiers but you have to design a system according to the most complex cases in order to be sure to extract data using an unique pattern. Snipre (talk) 08:49, 17 May 2015 (UTC)

Roman Catholism[edit]

Hoi, according to some "catholism" and "Roman catholism" is the same. It follows that pope Francis is no longer the head of the Roman Catholic church.. I am appalled that someone merged the two. It needs urgent action in order to undo the damage. No, I am not volunteering... Thanks, GerardM (talk) 14:47, 9 May 2015 (UTC)

Not sure what you mean. Do you mean "Pre-Reformation Catholic" (Old Catholic Church (Q5169816)) vs. "Post-Reformation Catholic" (Pope in Vatican)? Jane023 (talk) 17:36, 9 May 2015 (UTC)
Ayackbot changed the statements for religion incorrectly. They need to be mass changed to their previous condition. Thanks, GerardM (talk) 18:36, 9 May 2015 (UTC)
GerardM, I don't understand your point. I replaced religion (P140): Catholic Church (Q9592) by religion (P140): catholicism (Q1841) because the second is a religion while the first isn't. What's the matter? — Ayack (talk) 10:09, 11 May 2015 (UTC)
Ayack please look at the descriptions, not just the labels. If you do you will see that "Catholic Church (Q9592)" is for the religion led by the Pope and "catholicism (Q1841)" is for a wider concept including some bits not under the Pope's authority. Please undo your change (but be careful not to move stuff that really should be under "catholicism (Q1841)") Filceolaire (talk) 13:16, 11 May 2015 (UTC)
As for sure there are different kinds of catholics. catholicism (Q1841) is some kind of religious movement. It is difficult to descripe people as a member or part of a movement. In most cases it will be better to assign a person to a specific organisation (as a church or an order). Somebody can be a member of the organisation without beeing part of the movement or vice versa. By the way Old Catholic Church (Q5169816) are not "pre-reformation catholics", they are catholics that refused to accept First Vatican Council (Q190857), while Society of St. Pius X (Q868160) refuse to accept Second Vatican Council (Q169789), but in all definitions including their own they are still catholics. Both accept Council of Trent (Q172991) which is clearly not the age of pre-reformation.--Giftzwerg 88 (talk) 17:43, 11 May 2015 (UTC)
I have asked Amir to undo these edits... Sadly the "perps" do not take responsibility :) Thanks, GerardM (talk) 12:30, 17 May 2015 (UTC)

Importing actor's height from ImDB[edit]


I was wondering if it is allowed and/or a good idea to import actors height from ImDB, e.g.


  • We need units on Wikidata first. Sjoerd de Bruin (talk) 07:03, 11 May 2015 (UTC)
  • Is ImDB a reliable source for this kind of data ? Then what is the utility of a property which can be fulfilled for most of the persons which have an element in WD ? The target of WD is not to concentrate all information of the web but to create a database. And a database can be useful only when a description model can be applied to all elements of the database. What is the use of a height property used in a very specific field like actors ? You can't do any statistics because the group is too small and not representative of the world population. Then what's about actors who played as child and never continue the actor career after this period of their life ? This property will be time dependent and this reduce again its use. --Snipre (talk) 13:09, 11 May 2015 (UTC)
  • IMDb is not a reliable source for height data, but we do need to store height data once the new datatype comes in. We probably need to restrict the use of a height property to items about persons that are 18+ years old. --AmaryllisGardener talk 13:15, 11 May 2015 (UTC)
I don't think we need to restrict a height property in this way. Instead we just give the 'height' statement a 'point in time' qualifier. Filceolaire (talk) 14:21, 11 May 2015 (UTC)
I guess that'd work fine too. --AmaryllisGardener talk 15:47, 11 May 2015 (UTC)
I also like Filceolaire's approach. I'm assuming that the goal is to not record the height of a child, and discover that he or she has grown up since then (rather than protecting someone's privacy). If so, then age 16 might have been good enough, since most kids have stopped growing by then. However, age alone couldn't be the predictor. Some kids die, or stop growing young due to medical issues; some adults shrink (especially due to osteoporosis) and rarely grow more (hormonal issues). "X height on Y date" can easily handle all of these situations. (And also turn into a growth chart for certain famous children, but we might need to accept that cost.) WhatamIdoing (talk) 06:16, 17 May 2015 (UTC)


If you are interested: m:Merchandise giveaways/Nominations. They are giving away T-shirts to wiki-users. You can candidate someone you know or "vote" for the ones who have already been proposed.--Alexmar983 (talk) 16:21, 11 May 2015 (UTC)

When they were Wikidata t-shirts I would have idea ... Wikipedia I love but it is not my thing. GerardM (talk) 15:14, 12 May 2015 (UTC)
Well, I though that if you know somewone who is active on a "content"/"local" wikipedia and here, that's still a nice gift ;)--Alexmar983 (talk) 09:17, 13 May 2015 (UTC)
Let's suggest they get some Wikidata t-shirts. Who knows, maybe they already have them ;). All the best, Taketa (talk) 17:03, 16 May 2015 (UTC)

Wikimedia category & Wikimedia category page[edit]


Are they duplicates or is it a real distinction between those entries ?


Jona (talk) 13:36, 13 May 2015 (UTC)

The first one is "Open source games" and the latter is at least in Swedish "Open source computer games". --Stryn (talk) 15:27, 13 May 2015 (UTC)
@Jona, Stryn: I guess "Open source games" and "Open source computer games" both can be translated to "Fria datorspel" in Swedish, since "open source" imply a "dator".
But "Fria" does not necessarily always means "open" in Swedish, it can also mean "gratis". But it looks like "open source" is the main content of the category, but examples of "gratis datorspel" also exists there. -- Innocent bystander (talk) 08:08, 14 May 2015 (UTC)
@Stryn: There is also "Open source video games" I keep on thinking that a clean up is necessary but I can't go further with my knowledge (is there a banner of some kind ?). Thanks Jona (talk) 21:10, 17 May 2015 (UTC)

next 2 rounds of arbitrary access coming up[edit]

Hey folks :)

The rollout of arbitrary access on Dutch Wikipedia and French Wikisource seems to be going well so we're going to continue the rollout. The next projects will be:

  • 18. May: Farsi Wikipedia, English Wikivoyage, Hebrew Wikipedia
  • 1. June: Italian Wikipedia, all remaining Wikisource

Cheers --Lydia Pintscher (WMDE) (talk) 15:20, 13 May 2015 (UTC)

Hey folks :)
Next round has been done now.
I have one request: Please do NOT go to those projects and try out arbitrary access in their articles. Leave that to the local communities. We want more projects to make use of Wikidata and not be annoyed by random people coming to their projects and messing up their articles.
Cheers --Lydia Pintscher (WMDE) (talk) 10:19, 19 May 2015 (UTC)

How to recruit translators for Wikibase[edit]

At, we are focusing on expanding language coverage for MediaWiki. The main tools are prioritisation and targeted communication. Wikibase is currently included in the "Main Wikimedia extensions", whose statistics you can see at translatewiki:Translating:Group statistics. The group is meant for translators looking for more messages to translate after MediaWiki core, without digging in the mare magnum of obscure and sometimes little-used extensions. However, I have two concerns:

  1. Wikibase has a lot of messages, over 600 (see list), making up about 20 % of the group (which contains dozens of extension);
  2. messages are often quite specialistic and always require knowledge of the Wikidata glossary.

My impression is that "general" translators have to spend a lot of effort translating these messages and that most translations come from Wikidata users anyway, at least for bigger languages (see list of translators). Do you see a value in this promotion of the extension, or can it be put in the more general bucket of "Wikimedia extensions"? --Nemo 08:13, 14 May 2015 (UTC)

I think it should be a seperated group. Some messages are hard to find. Sjoerd de Bruin (talk) 07:24, 16 May 2015 (UTC)

Subclasses of Wikimedia categories[edit]

I created a database report with the most linked category items. It turned out that we have couple of items about Wikimedia categories with subclass of (P279) claims, e.g.

< Category:Rwanda templates (Q6509144) (View with Reasonator) > subclass of (P279) miga < Category:Country templates (Q3848) (View with Reasonator) >

In my opinion this are highly questionable claims as each wiki has its onw category system and so the claims are not in general true. Moreover, it can happen that category system get rearranged and therefore the claims are not stable. Is there any point in keeping such claims? --Pasleim (talk) 12:57, 14 May 2015 (UTC)

@Pasleim: In your example it's highly questionable indeed, as it reflects the problems of the category system that we can avoid : see Help:Classification. the Template per country item seems not a class, but a class of class. France template and USA templates for example are subclass of templates indeed, but they could be instances of class of template of a specific country. Template:Politics of Rwanda (Q13471959) is a Rwanda related template, it's certainly not a Category:Country templates (Q3848) (View with Reasonator), as it should be if the properties of subclass of and instance of are relevant here. TomT0m (talk) 13:25, 14 May 2015 (UTC)
This subclass is not depending on the category system on a wiki. On the item of templates in Rwanda, it is connected to both the category group of templates of countries and the category group of templates from Africa. Between the categories there is a relationship. As long as that relationship is preserved, even with the usage of other properties, I am fine with changes. Romaine (talk) 15:12, 14 May 2015 (UTC)
Delete these relations, as they are trying to capture the subclass of the topic of the category and not the category itself. As regards both types of categorization, there have been several recently rejected properties for capturing the "category" relationship. So it's correct simply to delete these claims. @Romaine: Wikidata:Property proposal/Archive/30#Parent category and Wikidata:Property proposal/Archive/30#category should make it pretty clear that these relationships aren't desired here. --Izno (talk) 14:07, 18 May 2015 (UTC)
@Izno: Sorry, but or you do not understand the relationship I try to describe, or you just make a false comparison. I have seen both discussions (which I agree with) and still these are different from this situation. So no, they do not make it pretty clear, those handle a different situation. And there is a relationship between the two items, that is something you ignore completely with pointing to a different discussion. So it is NOT correct to simply delete these claims. If you think this should be handled, come up with a solution to solve this without just a silly deleting. Romaine (talk) 16:33, 18 May 2015 (UTC)

what about entries in Freebase?[edit]

Once upon a time (back in 2007...) I added an entry about me in Freebase ( I never updated it, though, so it is desperately inaccurate: just to say, I actually am an author, since I published four books: but none of them is present in the entry. What will happen now that their database will be imported to Wikidata? Of course I don't care too much that the entry must be kept: but if this was the case I'd like to update it... --.mau. (talk) 13:13, 15 May 2015 (UTC)

As fare as i know this entry will die with Freebase. Freebase items will not be imported into Wikidata. The only item connected with you is the Wikinews article Q17963958 (Q17963958) ("Wiki@Home intervista ..."). --Kolja21 (talk) 16:36, 15 May 2015 (UTC)
Thanks. I read something about importation of data from Freebase (and related concerns about who used Freebase for advertising), so I was a bit puzzled. --.mau. (talk) 05:07, 19 May 2015 (UTC)
Yes, some data will be imported from Freebase but only if the item already exists in Wikidata. --Kolja21 (talk) 12:49, 19 May 2015 (UTC)

Time for local CheckUsers?[edit]

Hi all, with an increase in sockpuppets recently, as well as the ever-present need to check spambots, I think that now would be a good time for Wikidata to get some local CheckUsers. Most of the stewards who do checks are admins here and thus can't do checks here, and the response time from other stewards is rather slow. What do you all think about it? Ajraddatz (talk) 23:25, 15 May 2015 (UTC)

It would probably be used more than OS, at least. --Rschen7754 01:00, 16 May 2015 (UTC)
Good idea. --AmaryllisGardener talk 01:30, 16 May 2015 (UTC)
I guess a few more thoughts on this: I've had concerns about relegating the use of CU on this wiki to the stewards (which I won't go into here), and I believe that it would be in our best interests to handle this by ourselves. The other question is if there are candidates who would be able to get 25 votes and the right percentage (I forget what it is locally) within a 2 week period. I do have concerns about people not involved in the community opposing because they don't like how CU is used on enwiki etc., and feel that a canvassing or minimum tenure/number of edits guideline may need to be considered beforehand. Also, if these candidates are oversighters, would they be resigning their OS rights (and would we have to get more oversighters), or would they keep them?
I guess I'll be transparent and say that if I was asked if I was running today, I would probably decline; I'm enjoying the peace and quiet after stepping down from being a steward, and life is busy right now, though I might reconsider in the future. I'm also not very active as an oversighter either, but that's because the requests come in when I am at work and unable to respond to them. I'm available to respond to requests evenings and weekends in the US, but unfortunately there's not many requests then. But I'm still monitoring the OS logs, if that helps. --Rschen7754 02:10, 16 May 2015 (UTC)
I have no intention of running either, so no concerns (for now) with OS candidates. I would like to see, as part of this discussion, if anyone involved in CheckUser-y areas here would be interested in running. We should probably elect 3-5 IMO Ajraddatz (talk) 02:15, 16 May 2015 (UTC)
I support this proposal. I think it's time Wikidata has its own CUs. But the major question is, if the community wants it, who should run it? Jianhui67 talkcontribs 07:02, 16 May 2015 (UTC)
Isn't that the easiest question? At least those two users, who can get a minimum of 75%/25 support-votes. The easiest way to find one is probably to look for well-known admins with a good reputation. I am not saying that they have to be admins, but I think it is easier to find a successfull candidate among them. -- Innocent bystander (talk) 09:56, 16 May 2015 (UTC)
If any prospective candidates want to drop me an email and ask about the role or suitability, I would be happy to answer questions. --Rschen7754 15:47, 16 May 2015 (UTC)
Hi all, I don't fit the active admin envisioned above, but if we need people I am willing to help. I am not experienced as a checkuser, never had the need, but I can learn. I'm an admin and a steward so I hope I can be trusted. Active on a daily basis (always on a couple of projects if not Wikidata) and am able to respond timely to any request. Your thoughts on this are very welcome Rschen7754. All the best, Taketa (talk) 17:16, 16 May 2015 (UTC)
I'd like to go with some of the very active users here, especially in counter-vandalism. I would be glad to help train new checkusers, so experience isn't a big issue. I would personally like to see @Jasper Deng:, @Ymblanter: and @Jared Preston: consider applying. Ajraddatz (talk) 17:21, 16 May 2015 (UTC)
If there are enough people who are highly active I won't apply :). Consider me a backup if you can't get the team together. Cheers, Taketa (talk) 17:23, 16 May 2015 (UTC)
Two respected admins that have been here forever it seems, and an admin who is newer, but is active and certainly knows what he's doing. I'd support those three. --AmaryllisGardener talk 17:31, 16 May 2015 (UTC)
I'm personally okay with Ymblanter being both a CU and bureaucrat, but I suspect that others might not. --Rschen7754 17:33, 16 May 2015 (UTC)
Thanks all for support but I would be absolutely helpless as a CU since I do not know how to block a range, not even speaking about more advanced issues. If we decide to have CUs, I am not going to run, I am pretty happy in my role as admin and crat here.--Ymblanter (talk) 17:42, 16 May 2015 (UTC)
So is it time to open up to candidates? --George (Talk · Contribs · CentralAuth · Log) 05:54, 17 May 2015 (UTC)
I'd like to see if any more people are interested first, and get a feel for whether or not they would be widely supported. I don't want a repeat of the RfOS, when we had a bunch of them at the start but only one elected due to the 25 rule. Ajraddatz (talk) 06:00, 17 May 2015 (UTC)


@Ajraddatz, Ymblanter, Caliburn, AmaryllisGardener, Rschen7754: I wouldn't have any problem with the role, I'd be happy to deliver the information if required, but like Ymblanter, I don't have a clue on how to block IP ranges. I expect Rschen7754 wouldn't mind lending me a word or two of advice should we decide to go so far. Whether I am nominated or not, I think it's essential to the integrity of the project that we can conduct our own CU checks. I think we're pretty much all agreed on that note. Jared Preston (talk) 15:23, 17 May 2015 (UTC)
I am more than willing to run, as I think I have the necessary technical expertise (I would think that my knowledge of IPv6 ranges is especially useful), as long as the resultant CheckUser team handles as many checks as the stewards currently do, so that overall productivity isn't lost.--Jasper Deng (talk) 16:52, 17 May 2015 (UTC)
In my opinion, local CUs would rather be a step back for Wikidata than a step forward, considering what kind of CU requests we do get. We do not have the resources to deal with spambot checks as effectively as stewards may do who run all the cross-wiki checks, anyway. My feeling is we did not have more than 10 occasions where a non-spambot check has been necessary in over 3 years of Wikidata. This makes at max. 1 check of that non-spambot nature every 3-4 months, and this is already calculated very "optimistically". I pretty much disagree with Jared Preston's note on own CU checks are being necessary for our integrity, not for nothing we decided to let stewards continue performing spambot checks (of cross-wiki nature, which is usually the case) even if local CheckUsers would have been appointed (cf. Wikidata:Checkuser policy#Stewards), and for the rest we just have too little need to warrant own local checkusers as well as the extra burden for stewards performing cross-wiki checks, having to notify the likely small local team each time they do anything here. Vogone (talk) 19:52, 17 May 2015 (UTC)
Considering the metawiki search, it's rather one non-spambot check every 6-8 months than every 3-4 months even, which makes Ajraddatz's claim of recent "increase in sockpuppets" even more irreal. Vogone (talk) 20:00, 17 May 2015 (UTC)
I very strongly disagree. Let's look at how stewards have handled our CU requests:
m:Steward_requests/Checkuser#FreightXPress.40wikidata - Mardetanha seems to think that it is okay to publicly link an account to an IP address on a public page on Meta. In this case, it was okay, because the user self-disclosed their IP address when they created the account IP- (talkcontribslogs). The problem is, Mardetanha did not realize that this was the name of an account until after the disclosure was made.
m:Steward_requests/Checkuser/2013-12#Sanela.40wikidatawiki - stewards refused to disclose the user running unapproved bots on multiple accounts after being blocked multiple times, preventing us from enforcing our policy on multiple accounts. We even had oversighters by then, who could have handled more sensitive cases, and we were still not told who was running these sock accounts. Even after a year as a steward, I still cannot think of how publicly linking accounts would violate the privacy policy.
And as a counterpoint to the argument above, so that may be only 7-8 requests, but that means that stewards messed up 25% of them.
Combined with significant misgivings with some of the more active stewards (one of the reasons that I resigned a few months ago), I believe it would be in our better interests to elect local CUs. And much of Vogone's objections would be addressed if one of "our" stewards was a CU; they would be motivated to do the checking because this is their homewiki, and they could globally lock and block. --Rschen7754 20:07, 17 May 2015 (UTC)
Pictogram voting question.svg Question So we had two CU cases to run in the past? --Succu (talk) 21:47, 17 May 2015 (UTC)
More than that: about 7-8 public requests and other checks done by stewards for spambots or cross-wiki abusers. --Rschen7754 22:51, 17 May 2015 (UTC)
@Vogone: Hence why I said that I'd want our CU's to have the capacity to do that. I have no reason to believe that they wouldn't, given that we also have a number of steward admins here who perform checks on other wikis, but can't because they're local admins here, and because I think we have a set of capable candidates. The amount of spambots we get is now low enough such that I rarely see checks being performed here for spambots as well.--Jasper Deng (talk) 20:29, 17 May 2015 (UTC)
So when it comes down to the workload, how many cases of checkuser are there in a month or a year? I believe it would be of benefit, if there is some cross-wiki platform where positive results of all wikis are posted, so the checkusers or stewards might detect disruptive editing on a cross-wiki base. Surely there should be a routine cross-wiki check for all positive results no matter in what project the spammbot was active. The risk for the bot owner is low, so it is essentially that all edits are found quickly and appropriate response can be taken. Is there a technical possibility to redirect the edits of detected spambots to a mirrorsite instead of blocking the bot? So the spambot might do its work undisturbed on sandbox items on a mirror without harming anything. The edit of a detected bot are no threat and can be rollbacked anytime, the edits of undetected bots are more threatening. --Giftzwerg 88 (talk) 21:42, 18 May 2015 (UTC)
@Giftzwerg 88: CheckUsers have access to their own wiki but unless the case is a particularly bad one, I don't think IPs/results are stored there. They also have access to their own IRC channel and mailing list which they use for cross-wiki coordination, such as the case of cross-wiki spambots. As for redirecting edits, that would be purely a waste of time; the spam would have to be cleaned up and deleted anyways, and it is generally impossible to tell whether a given edit is spam or not using purely a technical filter; the abuse filters are not perfect for a reason. There isn't a reason to keep spam once discovered, either.--Jasper Deng (talk) 23:29, 18 May 2015 (UTC)
It would also quickly become a game of cat and mouse. Speaking from my experience as a steward: there is no true cross-wiki CheckUser functionality, though stewards have found a few ways to mitigate this which I won't go into. Usually results go to checkuser-l if cross-wiki action is needed, and they get stored on the wiki only in the case of repeated abusers where the data will likely be needed after it is no longer available using the CU function (to protect privacy, CU data is erased after it is a few months old). However, spambot IPs and accounts regularly go to checkuser-l since it affects all wikis, and because steward action is usually needed to lock accounts and globally block IPs. More spambot and cross-wiki checks would be run here if we had local CUs, simply because stewards don't have the time to run checks on every spambot or cross-wiki LTA that touches every wiki without local CUs (over 700). --Rschen7754 01:05, 19 May 2015 (UTC)
So local checkusers are dependent of the help of stewards if it comes to cross-wiki spam? It looks like this information is exchanged based on personal contacts between stewards and checkusers of different projects, which I appreciate. But I think of a more official way to communicate this informations, a kind of standard procedure to give some alert or hints to watch certain accounts or IP-adresses for their activities and check them. If there is reasonable cause for local checkuse, the same cause applies to checkuse on all other projects and there is no reason why a user with many sockpuppets won´t do the same on other projects. The cause is approved each time a local checkuser has the abuse documented on a local project. We must think in a more global way, we are talking about big data and about a multi language project, so we need a secured conection to all projects.--Giftzwerg 88 (talk) 10:26, 19 May 2015 (UTC)
@Giftzwerg 88: While these are helpful thoughts, they aren't really relevant to local CheckUsers; it's for privacy reasons that non-stewards can only have local CheckUser access, and although cross-wiki coordination is important, the vast majority of sockpuppetry cases are local (not necessarily here but a general global observation). Not every wiki can have CheckUsers because you need a sufficiently large community to vet candidates and create a CheckUser policy. To be clear, also, cross-wiki contacts between CheckUsers are not personal, they are official communication channels explicitly for the purpose of cross-wiki coordination. In short, the sharing of IP data is only done when needed. Rest assured though that CheckUsers know what are and aren't cross-wiki cases and deal with the cross-wiki cases accordingly.--Jasper Deng (talk) 16:18, 19 May 2015 (UTC)[edit]

502 Bad Gateway. --Jobu0101 (talk) 08:54, 16 May 2015 (UTC)

It works for me at the moment but remember that his endpoint is still beta and not really an official query service. There is no guarantee that it works all the time. -- Bene* talk 09:55, 16 May 2015 (UTC)
It's back up, assuming this is caused by this problem, but I'm not sure. Multichill (talk) 10:26, 16 May 2015 (UTC)

Merging two items[edit]

Not sure if there is any discussion site which only deal such issues but I want to suggest merging two items: Q6331763 (Politics templates) and Q6421011 (Politics and government templates). I can understand why it has not yet been done, so that is why I want to bring it up for discussion. Any thoughts? Kristian Vangen 21:23, 16 May 2015 (UTC)

There's a language conflict for ar and vi. So merging won't be possible. Lymantria (talk) 07:17, 17 May 2015 (UTC)

FWD: the shape of a stroopwafel[edit]

User talk:Edoderoo#Can real objects be circles? --Ricordisamoa 00:00, 17 May 2015 (UTC)


@1776004789风不止: Does anyone see reason why Airborne early warning and control (Q408110) was split? (Check history.) I have undone links which were not readded. Matěj Suchánek (talk) 08:16, 17 May 2015 (UTC)

This kind of edits happen everyday. It seems like they don't understand our project, so they remove ALL sitelinks. Sjoerd de Bruin (talk) 09:00, 17 May 2015 (UTC)
I've merged a few items which that user seems to have jumbled up. Revert if I've misunderstood any differences between them. Prefall (talk) 14:40, 17 May 2015 (UTC)

Menu drop-down broken?[edit]

Hi, on item pages I no longer see the drop down option "More", with the menu selections "Merge" and "Move". Is there something wrong with the java script tools? Thx. Jane023 (talk) 10:29, 17 May 2015 (UTC)

The tool for advanced label support has been missing for some time as well... Sadly... It means that I have not fixed many labels I know to be wrong. GerardM (talk) 12:28, 17 May 2015 (UTC)
Solved! Thanks to whoever fixed it. Jane023 (talk) 07:33, 18 May 2015 (UTC)

Information to all the users of Wikidata[edit]

Wikimedia Foundation Board of Trustees Elections 2015 has been started. Please Follow the link to vote for Trustees of Wikimedia Foundation Board. -- Eagle.svg.pngTulsi Bhagat (Talk) 14:23, 17 May 2015 (UTC)

I am running as one of the candidates. The candidate presentations and the questions provide plenty of reading material for making an informed choice. I would be happy to see Wikidata well represented among the voters! --Denny (talk) 17:12, 17 May 2015 (UTC)

WikiNews links[edit]

Pirate Party Germany (Q13129) should be an item of it's own for the four wikinews articles, that are linked by main subject to this item? In this way we can only have one WikiNews-article about any subject, which is wrong anyways. Edoderoo (talk) 16:27, 17 May 2015 (UTC)

The way Wikinews articles are currently handled is a mess. Wikinews articles are about events, but are getting linked to items about partly-associated entities and then treated both as items about the associated topic, and as items about the Wikinews article. Much of the contents of Special:WhatLinksHere/Q17633526 are just wrong. In my opinion, we need to rethink how Wikinews articles are handled in general, and then do some major repair work on the existing items and links. I don't think that any of these items should link to Wikinews article (Q17633526) or use language of work (or name) (P407), really. --Yair rand (talk) 16:58, 17 May 2015 (UTC)
The way I treated them until now: a Wikinews article has it's own wikidata-entry, always. It is not linked to any of the other projects, to my idea. Then the language makes more sense, as all the listed languages do have a wikinews-article in that particular language. It now goes wrong where a Wikinews article is linked on a Wikidata entry of a person, a building, a city, etc. We indeed need to do (quite) some repair work, but I believe that 99% of the items are actually OK, but the 1% is causing quite some headache. Edoderoo (talk) 18:10, 17 May 2015 (UTC)
as an example: Jamaican Usain Bolt breaks world record 100m to 9.72 sec (Q19931054) is now taken out of Usain Bolt (Q1189) (see history for my edits). I see that the Polish language article is a news article on the same subject, but a different occasion, so I will create a new entry (Usain Bolt sets world record 200m (Q19931072)) for that as well. To my opinion, that is how we should treat Wikinews articles. Edoderoo (talk) 18:36, 17 May 2015 (UTC)
The use of language of work (or name) (P407) in Wikinews items is wrong because it is not corresponding to the general way in WD. We don't add language of work (or name) (P407) to all others items when they are linked to a specific WP (we don't add language of work (or name) (P407): english to an item linked to WP:en). ::: An item about one Wikinews article or about an WP article if linked to english Wikinews or WP:en is ALWAYS in English. This is just a redundant information which is a useless maintenance work to maintain as all redundant informations. Snipre (talk) 19:40, 17 May 2015 (UTC)
Right, that is about the smallest issue. What about the big issue, that a WikiNews article is linked on the WD-entry of it's main subject? Edoderoo (talk) 20:07, 17 May 2015 (UTC)
@Edoderoo: This is a structural problem. Acording to Wikinews structure, everything should be an item but this is not the case in WD. For example if there is a Wikinews article about the birth of a person, we should have an item about the birth of that person. But in the current structure of WD, birth data are added as statements. More we are performing some experiments with WD and more I thing Wikinews can't be included in the current WD. This is not the only issue we have with a database which is used to store data and to store links between projects. We have two different objects and the problem is we try to keep them together and something to create interactions inside the same system. Typical examples are the categories items. Snipre (talk) 21:10, 17 May 2015 (UTC)
The articles you linked to are about events. As such, they should have instance of (P31) event (Q1190554), and they should be linked to articles on other wikis about the event. If there were a Wikipedia article on "Establishment of the German Pirate Party", it could be linked to the Wikinews articles you referred to earlier. These items can have normal relations with other items. For example, the item on Pirate Party Germany (Q13129) could link to the item about its establishment using subject of (P805) as a qualifier of inception (P571). --Yair rand (talk) 20:14, 17 May 2015 (UTC)
@Yair rand: The problem is that WD doesn't store each event as an item. Just think about birth and death of people: most of data about these kind of events are stored as statements. According to Wikinews structure these should be items in order to create appropriated links. This is the problem of mixing data in a structured way with site-links which suppose that everything can be an item in order to create an unique link. Or we have to allow multiples links from the same site to an unique item, but in that case we create a phone book and not more a structured database. Snipre (talk) 21:18, 17 May 2015 (UTC)
Wikidata has plenty of items on events. If you have an item regarding a relation between two topics or an topic and a time/location/etc, and that relation is primarily stored as a statement, use subject of (P805) as a qualifier to the statement to associate the other article. For example, if there was an item "birth of X", then you would add "subject of (P805): birth of X" as a qualifier to X's date of birth (P569). --Yair rand (talk) 21:24, 17 May 2015 (UTC)
The problem is not the fact that some events are represented by items. The problem is that for some events which are defined as statements we need to create empty or data redundant items just to be able to create a link to the Wikines article, the problem is in that mixture of models (sometimes event as item, sometimes event as statement). We have to choose between two models: a all-events-as-item model or we keep the statement for describing some events and we have to find a different way to link the statement directly to the site via a sitelink. Snipre (talk) 23:25, 17 May 2015 (UTC)
Most of the connections can be linked by main subject (P921) out of the wikidata-entry, especially when the news is about persons. In case of events, like accidents, it might be that there is no related article, and there is nothing to be linked. To my idea, the good thing when using main subject (P921) is that the wikinews articles will show up when using Reasonator, without the link they will simply never show up. Edoderoo (talk) 22:09, 17 May 2015 (UTC)
Why wouldn't they show up if the items' topics were considered the article topics instead of the articles themselves, like we do with Wikipedia articles? They could even be linked with significant event (P793) in the otherwise statement-less cases. --Yair rand (talk) 22:46, 17 May 2015 (UTC)
You mean in the way it was in Usain Bolt before I took the Wikinews articles out? Because you can then only link ONE Wikinews article on every Wikipedia-article. Where many politicians have multiple Wikinews articles about them. Major events (Tour de France, Olympic Games, etc) probably have more then one item too. In database language, there is a one-to-many relation, where it is now linked as a one-to-one relationship. Edoderoo (talk) 22:53, 17 May 2015 (UTC)
No, that's not what I mean. I mean having a separate item for the event, with the Wikinews articles linked, but not treating it as an item for a Wikinews article, but rather as an item for the event that is the topic of the linked articles. There should not be a linked Wikipedia article unless Wikipedia also has an article specifically on the event. --Yair rand (talk) 23:17, 17 May 2015 (UTC)
I start to believe that we need some clear examples, on how it's wrong, and how it's correct. Preferably with perma-links, else the items might change over time. I'm afraid that we're not all interpreting the terms in the same way (which might be due to not having all English as primary language, which is as well the case for me) and clear examples might bypass that problem. Edoderoo (talk) 07:09, 18 May 2015 (UTC)
First version, with the topics combined, which is currently the state of many items. This obviously have many problems, and I think we can all agree that it shouldn't be the standard.
Second version, with an item that has the Wikinews article itself as its topic. If I have understood correctly, this is the system supported by Edoderoo. Since it's being considered as a published piece of news, it has main subject (P921) linking to the other item.
Third version, with an item that has an event as its topic, with links to Wikinews articles that cover that particular event. As an ordinary event, it fits into the data model in the ordinary manner, and is linked to from the other item as shown. Should Wikipedia also have in article on the event (which is not true in the example given), then it would also have pages linked to that event's item. This is the system that I support.
--Yair rand (talk) 15:58, 18 May 2015 (UTC)
Given the case there are one or multiple wikinews article about an event, lets say a sports final, it makes no sense to assign language of work (or name) (P407) to the item. Everyone, or even a stupid bot can detect the language(s) of the article(s) based on the wikilink(s). So if there is a link on Jamaican Usain Bolt breaks world record 100m to 9.72 sec (Q19931054) to de.wikinews we can safely assume that the language is German. We also do not assign language of work (or name) (P407) with value German (Q188) to Usain Bolt (Q1189) because there is a German article linked to the item.--Giftzwerg 88 (talk) 21:57, 18 May 2015 (UTC)
I am actually in favour of the second hypotesis: it saves the links between WN versions, but we also can define what the article is about, and the item itself can be used as a qualifier or even as a source. --Sannita - not just another sysop 16:33, 19 May 2015 (UTC)
@Sannita: Can you give an example of when such an item would be useful as a qualifier? --Yair rand (talk) 20:11, 19 May 2015 (UTC)
@Yair rand: I'm sure the idea of the qualifier made sense when I wrote that, but I literally can't remember what I thought. So for the time being, you might want to scratch the qualifier part and keep just the source part. Sannita - not just another sysop 15:13, 21 May 2015 (UTC)

Wikimedia Foundation Board of Trustees elections 2015[edit]

Wmf logo vert pms.svg

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

Voting has begun for eligible voters in the 2015 elections for the Wikimedia Foundation Board of Trustees. Questions and discussion with the candidates for the Board will continue during the voting.

The Wikimedia Foundation Board of Trustees is the ultimate governing authority of the Wikimedia Foundation, a 501(c)(3) non-profit organization registered in the United States. The Wikimedia Foundation manages many diverse projects such as Wikipedia and Commons.

The voting phase lasts from 00:00 UTC May 17 to 23:59 UTC May 31. Click here to vote. More information on the candidates and the elections can be found on the 2015 Board election page on Meta-Wiki.

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

Posted by the MediaWiki message delivery 17:20, 17 May 2015 (UTC) • TranslateGet help

Wikidata weekly summary #158[edit]

Open for SI-unit properties[edit]

Progress on implementing numeric properties with units - such as length/distance, area, time and speed - is slow and only rarely appears on the weekly update. A number of properties are on hold due to this.

Would it be possible to allow properties, where an SI-unit exists. The description of the unit could include the SI-unit in use. Examples:

  • Height - the vertical length stated in m.
  • Area - stated in m2.
  • Speed - stated in m/s.

When in the future the units are here, a bot would be able to append them by property.

Off course this will annoy authors used to feet and °C to have to use m and K - and prefixed units (eg. km) are ruled out as well. But it's just for now.

Lua-modules will still be able to display the numbers as the language of the wiki in question might dictate.

Poul G (talk) 17:01, 18 May 2015 (UTC)

I mean, what I think matters is that if people are willing to use a temporary solution, they must be ready and willing to do a mass conversion once units are implemented.--Jasper Deng (talk) 00:26, 19 May 2015 (UTC)
It does sound tempting given that we could easily loose another year or two waiting the for the first implementation of units. Conversion from the temporary solution should be quite easy.
The main issue I see is that the quantity datatype still needs improvement, see: Wikidata:Project_chat/Archive/2015/05#population.2B-0. --- Jura 06:00, 19 May 2015 (UTC)
We are not years but only a few months away from having unit support. In the end it is an editorial decision but please do keep in mind that anyone who uses this data will have to make significant adaptions again later. --Lydia Pintscher (WMDE) (talk) 10:05, 19 May 2015 (UTC)
+1 with Lydia Pintscher (WMDE). If you take the time to follow the work progress of the development team you can have a good estimation of what will happen in the next 2-4 months. Better waiting than doing partial job which will be a mess to handle later. We have already enough things to modify because initial data import or structure was not correct. Instead of forcing to be able to add some hundreds of values better prepare data in an external file and be ready to import all in once with a bot when the datatype is ready. Snipre (talk) 11:37, 19 May 2015 (UTC)
I did check the developers ticket-list, as the number-property rarely surfaces in the weekly reports. And found a few with low activity and relatively few followers. Where should I have looked to see "few months" or "2-4 months"? Poul G (talk) 21:51, 19 May 2015 (UTC)

Not sure how to add something "safely"[edit]

I didn't have much time to look through the help pages, and I couldn't quickly see how best to do this; could someone else do it and leave a tutorial (or link to one) on how best to do it? I was afraid I'd break the metadata by introducing something with wrong formatting or wrong content if I just clicked the "Add reference".

Hilbert space, Q190056, can be linked to OCLC ARN 2072176. Could you please add it? Nyttend (talk) 17:32, 18 May 2015 (UTC)

@Nyttend: I took a shot at it but apparently fail at such things since I could not get a link to show up that directed me to the appropriate location. Do you have a URL that would be associated with that number, where$1 is the format of the url and $1 is the ID? It should be numerical only. You can add it yourself to Q190056 in similar fashion to what I just failed to do... (don't forget, this is a wiki, so trying things out yourself is okay).

Otherwise, we may not have the property to make the claim you are trying to make. --Izno (talk) 19:45, 18 May 2015 (UTC)

No, I don't have a URL, unfortunately. When you're using OCLC Connexion Client and browse the LCSH authority file for "Hilbert space", it shows you a subject header for "Hilbert space", and the OCLC record number for this SH is 2072176. It's not the title of a work, so it doesn't have an OCLC number per se. This is precisely why I was hesitant — I didn't even know if we were attempting to link LCSH entries with other data, let alone how to do it if we should. Nyttend (talk) 22:14, 18 May 2015 (UTC)
Two questions: Is the ID you're providing a unique, permanent ID, and, do you know if it can be derived from some other ID (OCLC or other closely related ID)? If yes and no respectively, it's probably a good candidate for WD:Property proposal/Authority control. --Izno (talk) 15:45, 19 May 2015 (UTC)
WD:WikiProject Authority control might be able to help too. --Izno (talk) 16:01, 19 May 2015 (UTC)
Izno, I expect that it's permanent, because the only way to assign this LCSH a new number would be to delete the authority file and create a new one, and that would mess up every single linked OCLC record of every type, whether ordinary bibliographic entries or other things. However, I was proposing using the OCLC number just because it was conveniently available for the LCSH. Is it possible to link LC subject headings to other kinds of records for the same topic? Q190056 currently links to an NDL entry that looks like it might be comparable to the LC topical subject heading that's listed in the print edition of LCSH and the free online edition of LCSH. Is my request now clear? Nyttend (talk) 03:18, 20 May 2015 (UTC)
Your request was cleaer to begin with--I just don't have the domain knowledge. --Izno (talk) 03:25, 20 May 2015 (UTC)
@Multichill, Kolja21: Maybe one of you can help. --Izno (talk) 03:29, 20 May 2015 (UTC)
Okay, thanks. I've only been working in cataloging since January (I didn't have the chance to take a cataloging class in library school; they don't even offer it anymore!), so I'm not as confident on this kind of thing as I wish I were. Not to mention my near-total unfamiliarity with how Wikidata typically operates. Nyttend (talk) 03:32, 20 May 2015 (UTC)
The LCSH can be looked up here. Hilbert space (Q190056): LCAuth identifier (P244) = sh85060803. P244 can be used for Library of Congress Name Authority File (Q18912790) and Library of Congress Subject Headings (Q1823134) (more infos by Gymel). --Kolja21 (talk) 04:17, 20 May 2015 (UTC)
Huh? At Property talk:P244#LCSH I merely stated my observation that this property apparently is also used for LCSH numbers, contrary to the definition at the time of the proposal. Changing the description of P244 might be an easy resolution, however at en:Template_talk:Authority control#At the moment, it is used almost exclusively in biographical articles there are complaints that this extended usage breaks the Authority Control template at en:WP: They also use the P244 value to provide links to OCLC WorldCat and this appears to work for the "n"-valued ids of the LCNAF entries but not for the "s"-valued of LCSH. AFAIK OCLC has developed its own system of FAST headings (Faceted Application of Subject Terminology, cf. en:Faceted Application of Subject Terminology but for the Hilbert Space example (LCCN sh 85060803) the identifier would be aka fst956785 (but this does not provide links to the resources related to that).
As for User:Nyttend's example "ARN 2072176": Keyword search for "2072176" yields mostly titles related to Hilbert Spaces in WorldCat, and searches for something with 060803 or 85060803 from the LCCN or 956785 from FAST result in nothing helpful. Also for subject headings there does not exist an interface like for persons, organizations &c. So I tend to back the claim that "ARN 2072176" is or corresponds to some internal OCLC number for their copy of LCSH for the benefit of their members but there seems no way to research, verify or even utilize this from the outside, i.e. with the means provided by -- Gymel (talk) 00:14, 21 May 2015 (UTC)

1980's-90's exercise shows on PBS or KVCR[edit]

Does anyone remember the 1980's-90's exercise shows on PBS or KVCR if so, do you remember the names?

Articles mixed with disambiguation pages or categories[edit]

Looking at Wikidata:Database reports/Constraint violations there are a lot of cases where disambiguation pages or categories are merged with articles. Often the property instance of (P31) for disambiguation page or category then gets deleted and everything looks fine. (But of cause it isn't: The different items are still mixed and the descriptions are mixed as well.) Why not block the merging for articles with disambiguation pages and categories? --Kolja21 (talk) 21:09, 18 May 2015 (UTC)

More usually it's the case that a wiki is causing these changes. Two main ways this happens:
  • A wiki takes a disambiguation page and changes it to an actual topic page.
  • A wiki takes a topic page and changes it to a disambiguation page.
Some wikis do this I imagine without moving the pages in question, which Wikidata would otherwise catch. From what I recall, Wikidata probably doesn't catch all of the page moves because there's some lag on the Wikibase client script running, so that might be a 3rd way we're not catching it.

Otherwise, the 4th way is simply that they're extant problems from prior to using Wikidata. I would imagine we've caught most of those by now... In other words, there's nothing we can really do except let the constraints catch the offending items. --Izno (talk) 21:14, 18 May 2015 (UTC)

Sometimes it is really an error where there are disambig pages mixed with some real articles. In some cases the article is a mix between a disambig and a stub. I´ve also found some of these.--Giftzwerg 88 (talk) 22:06, 18 May 2015 (UTC)
Some may be leftovers before everyone has a SUL. If somebody without a SUL moved a page and created a disambiguation page on the former page, the sitelinks would stay the same and would be incorrect. You should compare the maintance category with the statements on Wikidata. Sjoerd de Bruin (talk) 06:41, 19 May 2015 (UTC)
There are articles which are categorised as Disambiguation pages which correspond to real world topics which need wikidata items. A lot of the 'Surname' pages are this. In some languages the article just lists people with that surname, in other languages the article has an introductory sentence about the surname before the list. In my opinion these articles should be linked to a wikidata item which has statements 'instance of:wikimedia disambiguation page' and 'instance of:surname'. These items are then used as the targets of 'surname' statements. The alternative is to create separate items for the disambiguation articles and for the surname, where the item for the surname has no articles linked to it. this means we have twice the number of items to maintain and we have statements linking to items with no sitelinks. Extra maintenance and less usefulness. Filceolaire (talk) 06:36, 22 May 2015 (UTC)

Using Wikidata on small wikis[edit]

So, I was thinking that we could use start using Wikidata (besides a few isolated uses) in small wikis (with {{#property:P#}}), for things starting with something like the population and head of government/state of countries. For example, by making this edit (except with wikilinks around it). With help, we could make sure all Wikidata items about countries could be manually verified to have the correct and up to date population and head of state/goverment properties, with the correct rank of course. Then we could make that edit to all of the country articles on scowiki (and other small wikis where someone involved with this project would also be an active editor at). We also have to make sure that the items on the politicians have labels, which might make population listings easier, but the problem with the populations is that you couldn't list a date or reference on the wiki. Thoughts or volunteers? --AmaryllisGardener talk 01:58, 19 May 2015 (UTC)

I'm not sure of the schedule, but aren't we only a couple of months away from having arbitrary access on the rest of the Wikipedias? If I understand correctly, after that's available even automatically showing sources and dates would be possible. I recommend waiting and then making a bunch of changes to the infobox templates so that the pages don't need to be edited manually. --Yair rand (talk) 02:47, 19 May 2015 (UTC)
Just be careful - a few years ago ocwiki tried this but in a horrible implementation (bot-created stubs populated with magic infoboxes that broke). [1] is the result. --Rschen7754 03:01, 19 May 2015 (UTC)
@Rschen7754: That does look like something we'd have to be careful not to do, but it looks like they tried to start out with too much. @Yair rand: Well, it can wait a while. --AmaryllisGardener talk 03:09, 19 May 2015 (UTC)
I've also followed the oc experiment. Like Rschen7754 said it's a total mess with empty templates (oc:Modèl:Infobox) or templates that have been changed so often that it's hard to understand how they are linked with each other (oc:Modèl:Debuta Infobox V2). It would be good to start an initiative to clean up the mess and then we see what is the best way Wikidata can help (or destruct) small Wikis. --Kolja21 (talk) 13:00, 19 May 2015 (UTC)
BTW: Is there an admin that can help to delete the pages marked for deletion? oc:Categoria:Wikipèdia:Supression has more than 800 entries. --Kolja21 (talk) 13:15, 19 May 2015 (UTC)
Dropped a request at m:SRM, since I suspect that it may be too much for the 2 active admins. --Rschen7754 13:33, 19 May 2015 (UTC)
I am in pronciple positive but we also need to think about how we guarantee that Wikidata items used across several hundred projects are vandalism-free. I have a dozen of country entries on my watchlist, and they get vandalized on a regular basis.--Ymblanter (talk) 15:53, 19 May 2015 (UTC)
Yeah, vandalism's a problem. I was thinking that those involved in this project could watch a dozen or two country items. That way, we could make sure that any vandalism is reverted. BTW, for heads of state/government, to list that using Wikidata is very simple, and doesn't require dates or refs, just the name. --AmaryllisGardener talk 16:53, 19 May 2015 (UTC)
RE: "for heads of state/government ... is very simple" Maybe it is simple, but to be useful, you sometimes need the title of the person(s). Andorra have two heads of states, with two different titles AFIK. And in the Commonwelth there is both a Governor and a Queen. And the city of Stockholm have 12 mayors, each with a very specific title. -- Innocent bystander (talk) 10:10, 20 May 2015 (UTC)
Isn't this pretty much the same as I did two years ago on ru-wiki? See here for the examples. Same story on es-wiki on tennis related stuff. Edoderoo (talk) 12:16, 21 May 2015 (UTC)
@Edoderoo: Yes. Did your thing work? --AmaryllisGardener talk 17:25, 21 May 2015 (UTC)
Sure, on quite some wiki's that was implemented in the corresponding tennis infobox. Some Wiki's reverted due to several known or unknown reasons. The German wiki because the wikidata entry uses a dash instead of a semi-colon in the statistic numbers, the Dutch because they don't want to use Wikidata at all, the others I don't remember or just don't know. In the mean time those tennis statistics seem to be pretty well updated accross the communities, and someone even tried to get allowance to update part of it by an automated bot, but this gave issues with database rights. Still, for me it's a good show case of how Wikidata can be used to keep often changed data up to date on several wiki's. Some of the smaller wiki's can keep their tennis statistics pretty much updated without extra (or any) effort from their own community, as they benefit from the changes done by the larger foreign communities. Edoderoo (talk) 18:59, 21 May 2015 (UTC)
I might try changing the tennis infoboxes on scowiki, and if anyone complains (which is very unlikely, the whole community's me and two other admins) I can revert. I'll switch the country article's infoboxes individually (as in article by article like this, not changing all articles at the same time by changing the infobox) to insure that it's good. I'll hold off with changing the country infobox for a while. Thanks for the input everyone! :) --AmaryllisGardener talk 19:15, 21 May 2015 (UTC)

Wikidata, cycling and list of participants[edit]

Bonjour à tous. Je travaille sur la version francophone de Wikipédia, mais de manière plus importante à l'illustration du cyclisme (exemple). Entre deux reportages sur le terrain, je prépare la transition vers Wikidata, un projet qui avance petit à petit, mais qui progresse.
Aujourd'hui, ma question va porter sur les équipes et les coureurs participant à une course cycliste, au travers de l'article Grand Prix de Denain 2015 comme exemple. Comment faire (simple) avec Wikidata pour entrer une liste d'équipes, tout en sachant qu'il faut ensuite aller rechercher leur code UCI en trois lettre et leur pays de provenance, et qu'elles ont différentes places dans le tableau selon leur catégorie (UCI WorldTeams, Équipes continentales professionnelles, Équipes continentales...) ? Selon moi, un élément participating teams pourrait être créé. Mais ces mêmes équipes sont reprises dans la liste des partants, où je dois associer à un coureur son numéro de dossart et sa place, voire éventuellement un abandon au cours d'une étape (lorsqu'il s'agit d'une course à étapes comme le Tour de France ; Liste des coureurs du Tour de France 2014). Il faut également mentionner les directeurs sportifs, voire les maillots spécifiques que peuvent revêtir les coureurs. Tout ceci est encore assez compliqué pour moi. je souhaiterais quelque chose de simple à remplir, ce qui permettrait à d'autres contributeurs de générer les listes des partants, et d'effectuer sans trop de mal de nombreuses traductions. Cordialement, JGHJ.

[Google Translate] Hello everyone. I work on the French version of Wikipedia, but more importantly the cycling illustration (example). Between two stories from the field, I prepare the transition to Wikidata, a project advancing gradually, but progress.
Now my question is about the teams and riders participating in a bicycle race, through article Grand Prix de Denain 2015 as an example. How (simple) with Wikidata to enter a list of teams while knowing that they must then go seek their three letter code UCI and their country of origin, and have different places in the table according to their category (UCI WorldTeams, Professional Continental Teams, Continental Teams ...)? I think something participante teams could be created. But these same teams are included in the list of starters, where I have to associate with a runner dossart its number and its place, or possibly a discontinuation during a step (in the case of a stage race like the Tour de France; Liste des coureurs du Tour de France 2014). Mention must also be the team managers or specific jerseys that can take the riders. All this is further complicated enough for me. I would like something simple to fill, which would allow other contributors to generate the lists of participants, and perform without too much difficulty many translations. Sincerely, Jérémy-Günther-Heinz Jähnick (talk) 12:35, 19 May 2015 (UTC)

Teams shouldn't be used in definition of the participants list of a competition but only individual participants. Teams can be deduce from the "CV" of each runner defined in his item. Mainly because teams have often some substitutes which don't take part to the competition. If I know that a runner took part to a competition and at the same time I know he is part of a team at that moment of his career, I can deduce that the team took part to that competition. Snipre (talk) 14:39, 19 May 2015 (UTC)
Oui, je le savais déjà qu'il allait être possible de déduire les équipes à partir des coureurs (tout comme leur nationalité), mais j'ai besoin que Wikidata fasse la différence suivant les compétitions qui alignent 6, 7, 8, 9 voire 10 coureurs sur une course pour que le tableau tel que présenté dans mon exemple soit correctement rempli, tout en sachant qu'il arrive qu'il y ait des trous dans les effectifs présentés. Et enfin, pour aller plus loin, je souhaite que le nom de l'équipe pointe vers sa saison 2015, et qu'il y a ce problème récurrent que les équipes changent régulièrement de nom au fil des saisons. Bref, c'est quelque chose de plus complexe qu'il n'y paraît, d'autant plus que je souhaiterais que tout soit ensuite uniformiser sur les différentes Wikipédia puisque je souhaite pouvoir potentiellement créer des articles de courses dans une vingtaine de langues. Je vais te contacter pour d'autres questions et des points précis. Jérémy-Günther-Heinz Jähnick (talk) 15:51, 19 May 2015 (UTC)
Pas compris ton problème d'équipe avec 6, 7, 8, 9 voire 10 coureurs: c'est la combinaison entre le fait que qu'il existe un lien entre tous les coureurs qui ont participé à une course et le lien entre une équipe et un coureur à un moment précis qui te permet de construire l'équipe qui a participé à une course. Ensuite, le problème de nome est simple: il suffit d'utiliser la propriété "nom officiel" qui doit être définit pour un intervale de temps donné et récupérer le nom correspondant à la date de la compétition. Vu la complexité du sujet, il faudra envisager de faire l'extraction de données hors ligne et de sauver les données en dur dans les pages d'articles et non par avoir un système de connexion à WD à chaque ouverture de page. Snipre (talk) 17:35, 19 May 2015 (UTC)
Conflit d'édition @Jérémy-Günther-Heinz Jähnick: Utilises des qualificateurs (je bosse sur les courses de chiens de traîneau, et j'ai la même question compelxe de gestion des participants à gérer). Moi par exemple j'utilise participant (P710) pour les individus et en qualificateur de P710 sport number (P1618), ranking (P1352), point in time (P585) (mais pas trop utilisable pour le cyclisme où tout le monde arrive le même jour). Pour les abandons j'utilise significant event (P793) avec scratched (Q18595374), withdrawal (Q18609099) ou disqualification (Q18595902) (les 3 possibilités dans ce sport) et j'ajoute alors location (P276) avec le checkpoint (possible d'indiquer l'étape d'abandon à la place puisque je crois qu'il y a un item par étape sur WD pour le Tour). Tu peux sans doute ajouter en qualificateur member of (P463) avec l'item de l'équipe… Concernant le nombre de membres dans l'équipe, je n'ai pas actuellement de solution, vu que la situation n'est pas non plus rélgée pour les chiens de traîneau. Voir 2015 Iditarod (Q19455277) pour un exemple et Wikidata:WikiProject Sled dog racing/Todo pour ma méthode. En espérant avoir aidé… --Harmonia Amanda (talk) 17:46, 19 May 2015 (UTC)
@Harmonia Amanda: merci pour ta réponse, je vois que tu as le même problème que moi concernant la demande de création de propriété où des contributeurs cherchent à faire des économies et à en créer le moins possible au lieu d'écouter ce que demande le proposant, tout en sachant que chaque sport a bel et bien ses spécificités. Grâce à toi, je sais maintenant comment indiquer les abandons et autres.
@Snipre: : ça reste encore trop compliqué pour des contributeurs lambda. Il me faudrait la possibilité d'avoir des qualificatif de qualificatifs.
participant (P710) => Topsport Vlaanderen-Baloise 2015 (Q18746658), avec en qualificatif le listage de six à dix coureurs, et ces coureurs là se verraient qualifier d'un numéro de dossard, et au fil de la course d'abandons, puis de places (Harmonia Amanda : toutes les courses par étapes auront un élément par étape, même les moins importantes, vu qu'ils gèrent également autre chose). Je pense même qu'en qualificatif des équipes on devrait avoir d'abord le numéro de dossart, puis le coureur, et enfin sa place ou son abandon, parce qu'il arrive parfois comme à ma course de dimanche que l'équipe ne vienne pas au dernier moment, et que les dossarts lui aient déjà été réservés. À mon sans, il ne faut pas se contenter d'un système unique pour tous les sports, mais bien tenir compte des spécificités. Jérémy-Günther-Heinz Jähnick (talk) 07:35, 20 May 2015 (UTC)
=> autre équipe, et ainsi de suite.
Ma solution est beaucoup plus simple à gérer pour des contributeurs comme celui du projet cyclisme et reste assez logique et instinctive. Avec ce système, on n'est pas embêté avec le nom des équipes, puisque ça revient à suivre la solution actuelle où le lien renvoit vers la saison 2015 (les articles de saison sont les articles principaux, ceux des équipes sont des articles de fond, et non plus l'inverse). Par ce système, en plus de générer le tableau de la liste des participants, on peut également extraire les données de l'autre tableau listant les équipes participantes, ce qui fait d'une pierre deux coups.
@Jérémy-Günther-Heinz Jähnick: On ne cherche pas à faire des économies de propriétés, on veut simplement s'assurer que l'on une structure aussi générale que possible pour faciliter l'utilisation des données notamment les recherches: si pour chaque sport ou theme il faut appliquer une recherché en function d'une structure différente, on ne pourra jamais fournir d'outils généraux pour l'utilisation sur WP. Imagine que pour chaque type de liste tu doives créer une requête particulière pour cause de stockage des données selon un format spécifique et tu verras que tu auras du mal à trouver des personnes pour te donner un coup de main sur le travail de maintenance. Il faut donc penser global, mais global sportif: trouver les différents cas pour gérer les résultats sportifs (pas en function du sport, mais du type de données pour représenter tous les détails pertinents du classement final) et ensuite demander les propriétés adéquates. Snipre (talk) 16:24, 22 May 2015 (UTC)
@Harmonia Amanda: Ta manière de gérer les données est intéressante: mais au lieu d'utiliser point in time (P585) qui n'est pas comprehensible et surtout inutile pour les performances qui ne sont pas des dates, mieux vaudrait créer une nouvelle propriété "performance" qui accepte des valeurs avec unite et qui puisse être utilisée pour tous les sports: durée de la course pour le cyclisme et les course de chiens de traineau (je ne pense pas que les résultat sont donnés en jour), de même que pour toutes les courses de l'athlétisme, mais aussi longueur pour le saut en longueur, le lancer du javelot ou du poids,... significant event (P793) pour les abandons est un choix un peu difficile à faire passer, mieux vaudrait une propriété spécifique et voir si on peut la généraliser à d'autres sports pour favoriser son utilisation. Et pour l'équipe, je ne le noterait pas dans le classement, pour éviter les doublons avec les données indiquées sous l'élément du sportif, il faut trouver l'endroit le plus approprié entre le classement des courses ou les éléments des sportifs voire les éléments des equips elles-memes. Le plus important est de bien définir ce paramètre, car sinon chacun va créer son système et on ne pourra jamais avoir un moyen unique de récupérer cette information (il faudra 3 codes différents en fonction de l'endroit où se trouve la donnée équipe). Snipre (talk) 16:37, 22 May 2015 (UTC)

Property for "commander"/"commanded by"/"command of"[edit]

Need a property to link Detlev Krankenhagen (Q4237593) to German submarine U-549 (Q563394). Anyone know of an existing one? --Izno (talk) 15:58, 19 May 2015 (UTC)

Izno: You could use commander of (P598). — Ayack (talk) 16:42, 20 May 2015 (UTC)
Thanks. --Izno (talk) 19:12, 20 May 2015 (UTC)

Merge problems[edit]

Can someone merge en:Category:Federico Fellini (Q16812095) with de:Kategorie:Federico Fellini (Q9156497) ? 22:56, 19 May 2015 (UTC)

→ ← Merged --Pasleim (talk) 22:58, 19 May 2015 (UTC)

of (P642) miga[edit]

this diff just popped up a claim

< band (Q215380) (View with Reasonator) > subclass of (P279) miga < group of humans (Q16334295) (View with Reasonator) >
        of (P642) miga < musician (Q639669) (View with Reasonator) >


This is a big problem to me, as I don't know what this means and the of qualifier is highly unspecified. We should either precise the scope and the intend uses of it or find a better solution. (@Emw, X meta, zolo: Pinging probably interested wikidatans.

For example, I'd write this as

< band (Q215380) (View with Reasonator) > composed of search < musician (Q639669) (View with Reasonator) >

, which is a solution with less statements and a more precise meaning. Clearly a music band has musician members, for sure. TomT0m (talk) 09:50, 20 May 2015 (UTC)

I agree that "of" does not seem to be a very clear property in general, and that it should not be applied here in particular. However,

musician is classified as a profession, not as a subclass of humans, and don't think a group of humans should be Composed of search musicians. I would rather use occupation (P106) either as a standalone statement or as a qualifier "instance of group of humans". --Zolo (talk) 09:59, 20 May 2015 (UTC)

@Zolo: By definition, profession (Q28640) (View with Reasonator) is an activity that someone does for money. So it's not a really good solution to say every music band earns money for it, it's pretty false in the general case. I prefer the more less specific field of work search property. Qualifying instance of seems pretty weird and not really well specified to me. Formally maybe it could be used to precise some parameters of a metaclass, like template (Q1411845) (View with Reasonator) ? Like a group of things could be used as a metaclass (Q19478619) (View with Reasonator) in Protégé (Q2066865) (View with Reasonator) (cf. la section sur Protégé dans l'article Wikipédia), and the of qualifier used to instanciate the things parameter ? I'll ping maybe @Denny, Markus Krötzsch: on this :) TomT0m (talk) 10:45, 20 May 2015 (UTC)
@TomT0m: The French label of occupation (P106) is inconsistent with other languages. The subject item of occupation (P106) is occupation (Q13516667) and not profession (Q28640) so we may have to change the French label from profession to occupation --Pasleim (talk) 12:12, 20 May 2015 (UTC)
@Pasleim: Yet another use of of : . I'd prefer an instance of social sciences concept, but we need a clarification on concept anyway. Help:Classification could help, but I'm not sure how. Clearly when I make a hat, it's an event who can be classified in the hatmaking class of events. That someone often makes hat is possible using the occupation property for sure to link the person to those kind of classes. I think then that of course
< hatmaking > instance of (P31) miga < occupation >
makes sense. I'm not sure how the concept concept fits here. I would be happy with a metaclass social science class with a kind of statement
< social science class > classes used by search < social sciences >
< social science class > subclass of (P279) miga < social science class >
. TomT0m (talk) 12:55, 20 May 2015 (UTC)
I forgot
< occupation > subclass of (P279) miga < social science class >
. This would make occupations a subset of all the classes used by social sciences. TomT0m (talk) 12:58, 20 May 2015 (UTC)

I think the example looks rather understandable to me. I wonder if it works in other languages, too. That would be for me the much more important measure whether it makes sense or not. Another important measure are the use cases for such a statement in the other Wikimedia projects. That's what we should be measuring against. Considerations about Metaclasses, Protege, Reasoning and Class subsumption are for me secondary concerns, to be honest. --Denny (talk) 16:28, 20 May 2015 (UTC)

@Denny: That it make sense in natural language is indeed a nice feature, but it's like it is more a gut oriented usecase than a structured one. This implies the problems with several ways to say one thing, language interpretation dependancies, misunderstanding and so on, cultural differences problems, ... . As far as I'm concerned, this defeats a lot of the purposes of structuring datas and mimics a lot of problems with natural languages. The problem here is mainly that we have a multi purpose qualifier with no idea on how it is used by consumer, how it can be used to model things and a really poor documentation on the talkpage. As far as I can tell, it usually used to say something similar that the subclass statements says, I do not always see the added value. TomT0m (talk) 16:41, 20 May 2015 (UTC)
@TomT0m:, as I said, if it works in the other languages too then I assume it is not a natural language thing. Natural languages have so few universals that I would be very much surprised if we stumbled upon one here. So, if it works not only in English and German and related languages, but also in Chinese, Turkic, Arabic, etc., then I am inclined to say that the property has some reasonable conceptual interpretation. --Denny (talk) 16:58, 20 May 2015 (UTC)
P.S.: Actually, thinking about my limited understanding of Turkic languages, I would be rather surprised if it worked there, but I would like to want a more advanced speaker to weigh in. --Denny (talk) 17:00, 20 May 2015 (UTC)
I'm actually trying to find a good interpretation :). Not convinced yet there is some (that is not redundant), for example in the two examples I provided it seems that there is a lot of work to do to know, for example if the qualifier makes the members of the bands musicians, if the group has musician as an occupation ... whether in the second it seems pretty clear that a concept has a domain of relevance. In one case it's used on an instance of statement, in the second it's used on a subclass of statement. Clearly instance of and subclass of have a lot of literature written on them ... but with this qualifier it is a whole new world. TomT0m (talk) 17:46, 20 May 2015 (UTC)

Full names and non-country nationalities[edit]

I've been participating off-and-on for a while, and there are two things that I'm not too sure about when I edit items about people and they keep coming up.

First, what do we do with full names? Many people go by shorter variations of their official birth name, including both people who sometimes include their initials or middle names and individuals with multiple last names (as in Spanish and Portuguese naming customs, where usually only one is needed for day-to-day life). Do they all get put in aliases? This seems like it would not scale very well, considering most languages would have the same value but we have to add them separately for each. Part of me has always wanted to use birth name (P1477), but that's typically use for name changes, not for people who just have a longer "official" name. Do we have an existing procedure for this?

Second, how do we handle non-country nationalities, like Galician, Catalan, Scottish, English, etc.? This goes for both in the item description and statements. I typically use the more specific nationality as an item descriptor (e.g., Scottish politician, Catalan businessman), but then put the actual country in the statement country of citizenship (P27). I don't see any places where we could add "Catalunya" or "Scotland" in the actual statements list for that item, but it feels weird to leave it out, especially when a few Wikipedians actually put that information in the "nationality" spot of infoboxes.

How do other users typically handle these? Cbrown1023 talk 18:18, 20 May 2015 (UTC)

"Nationality" is a very ambiguous concept. To add where a person was born, use place of birth (P19). To add their ethnicity, use ethnic group (P172). To add where they have lived, use residence (P551), preferably with start time (P580)/end time (P582). To add what country or region they've been a politician of, add it to the item about their positions. Is there any other data that determines what nationality a person is? --Yair rand (talk) 20:40, 20 May 2015 (UTC)
I'm mostly referring to autonomous regions like Catalonia (Q5705) and Scotland (Q22) which are technically part of a larger country, but have some degree of autonomy and regional identification. (Those don't really seem the same as an ethnic group (P172) to me, but others may disagree.) There is no Scottish or Catalan passports, but most people identify as being from those regions rather than the broader countries like United Kingdom (Q145) or Spain (Q29). Look at w:ca:Artur Mas i Gavarró, for example, the nationality is indicated as "Catalan" rather than "Spanish". If we pulled Wikidata information for infoboxes on cawiki, we would lose that distinction.
On a similar note, how do we handle country of citizenship (P27) for British citizens? Should it be the broader United Kingdom (Q145) or something more specific like England (Q21), Scotland (Q22), or Wales (Q25)? Technically the latter options are more specific and are still countries, but "citizenship" applies more to the United Kingdom than its constituent countries. Cbrown1023 talk 15:07, 21 May 2015 (UTC)
What determines whether a person is of Scottish or Catalan nationality? Just being born there? Ancestry? Self-identification? Is there anything that can clearly show that a person is of a particular non-country nationality, or are the infoboxes just populated by editorial decisions based on a variety of factors?
Re country of citizenship (P27), presumably United Kingdom (Q145), unless there actually is a formal citizenship for UK constituent countries. --Yair rand (talk) 17:52, 21 May 2015 (UTC)
Re aliases: All used versions of an individual's name should be aliases, IMO. That's what aliases are for: a list of anything that someone would reasonably use to refer to the subject. (Sorry if I've misunderstood the question.) --Yair rand (talk) 20:44, 20 May 2015 (UTC)
birth name (P1477) is the correct property for the full name (at birth). To make it searchable, the same would need to be added as alias. --- Jura 04:45, 21 May 2015 (UTC)
Thanks to both you, that's mostly in line with what I typically do. Even though birth name (P1477) says it's only supposed to be used when the birth name is different (i.e., when there was a name change), I sometimes ignore that. Wikidata should really show the birth name results in search results too though, not just aliases. Cbrown1023 talk 15:07, 21 May 2015 (UTC)

put the same refference to multiple properties at once[edit]

Hello. Is there a way to put the same refference to multiple properties at once? For example in Q17442744 I want to put the same reference for all teams in p:P710 at once. Xaris333 (talk) 22:29, 20 May 2015 (UTC)

Not yet. The developers are working on it: phab:T76233. --Yair rand (talk) 22:55, 20 May 2015 (UTC)
@Xaris333: you can use described by source (P1343) and describe your source with authors/urls/pages once, then only specify stated in (P248) as source for all entity properties. Another option is to create separate Wikidata Item for your source, put all your source properties into this new item, and make reference to this item as value of stated in (P248) for statements in original item. -- VlSergey (трёп) 09:29, 21 May 2015 (UTC)
@Vlsergey: Why do you use described by source (P1343) if there is a reference section ? Use the reference section of the declaration and point to the item containing the data of the source using stated in (P248) and then add the specific data like page under the reference section. Snipre (talk) 12:26, 21 May 2015 (UTC)
@Snipre: sometimes it is more convenient. have a look at Arthur Conan Doyle (Q35610) for example. There are duplicate sources at date of birth (P569) and date of death (P570), but there are only a "reference" mention of those sources using stated in (P248). Full "addresses" (title, page, url) of those sources are placed in described by source (P1343) value. We will describe those sources in described by source (P1343) anyway (like "Literature" part of Wikipedia article), so why not to use it? Thus we removing duplication and we don't need to introduce new wikidata items. Well, it works nicely (only?) when encyclopedic articles is used as a source, because single item is described in encyclopedia once. For more complicated cases like article in the book or magazine, it's better to create separate item and link to it using stated in (P248) in reference, as you (and me -- second option in my message above) described. -- VlSergey (gab) 12:44, 21 May 2015 (UTC)
@Xaris333: Please have a look at Help:Sources and if you have other questions please use the talk page. Snipre (talk) 12:26, 21 May 2015 (UTC)

Constraint Irreflexive[edit]

I propose a new Template:Constraint "Irreflexive". Applies to properties that can't connect an item to itself.

Eg it would be very useful for A different from (P1889) B: such claim is recorded when A and B are often confused, so I DO get confused when entering it for A, and have made the mistake of picking up A.

Could also be useful for familial relations.

--Vladimir Alexiev (talk) 07:54, 21 May 2015 (UTC)

@Vladimir Alexiev: All properties (whether actually irreflexive or not) are considered irreflexive in constraint violations report. See Wikidata:Database reports/Constraint violations/P17#"Self link" violations (although it is not real violations). In addition, there're Special:AbuseFilter/49 plus self-referencing tag.--GZWDer (talk) 09:44, 21 May 2015 (UTC)

Al Rayyan: municipality AND town in Qatar, three different data items here, Persian language difficulty[edit]

Hallo, many language versions of WP do not differentiate the municipality of Al Rayyan from the town of the same name. Until now, Wikidata has had three different items for this disentangled set:

  1. Q311272 which I have used for all articles that are concerned with the municipality only
  2. Q18576873 which I have used for all articles that are concerned with the town only
  3. Q11686604 which used to contain links to the Polish and the Macedonian language versions of articles for the municipality (I have moved those to item no. 1).

Now, I think I have identified the articles (in Latin and Cyrillic alphabets which I can read) for either item correctly but I am unsure about the articles in arabic languages. Especially, the items in Persian are a riddle to me:

Q311272 links to the Persian Wikipedia article fa:الریان and to the Persian Wikivoyage article voy:fa:الریان, Q11686604 links to the Persian Wikipedia article fa:ریان and to the Persian Wikivoyage article voy:fa:الریان (استان). I have no idea what either of them is concerned with.

Could anyone please help me disentangle these? Or could you direct me to a more specific help desk? Thanks, --Andropov (talk) 13:43, 21 May 2015 (UTC)

done Michiel1972 (talk) 14:27, 21 May 2015 (UTC)
Thank you! --Andropov (talk) 14:29, 21 May 2015 (UTC)

Aliases in other-language scripts[edit]

Hello. I'm new to Wikidata but have been trying to clean up some of the items for Korean music artists, primarily by adding descriptions. However, I've also been cleaning up the English- and Korean-language "also known as" fields by, for instance, removing the names of individual members from group entries, which apparently were pulled in by accident. My question comes from the following: I've also removed names in the alias field of English entries in Korean and Chinese scripts (they are already present in those languages' entries, so there was no actual loss of information). Additionally, I've removed copious amounts of alternative romanizations from these fields, as these aren't usually included in the subjects' Wikipedia entries unless they are especially prominent or confusing. I figured all these things had been pulled in automatically when information was pulled into Wikidata. However, I belatedly checked the article histories and found that the info had been added manually by another user. I left a message for that user but it appears he/she is only active here sporadically. I've read the guideline for aliases but found no help. Could anyone provide clarification about inclusion of non-roman scripts and alternate romanizations as aliases in the English-language Wikidata? Thank you so much for any help! Shinyang-i (talk) 17:35, 21 May 2015 (UTC)

@Shinyang-i: Aliases are intended to help users find an item; if the user types one of those other romanisations, they will only find the item if it is present as an alias, Please restore those you have removed. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:43, 21 May 2015 (UTC)
@Pigsonthewing:Can you provide a guideline as to how many romanizations should be included? There can be sometimes 30 or 40 for a single name. It will go far, far beyond what were originally present. Also, are aliases providing the same functionality as redirects do on Wikipedia? Shinyang-i (talk) 17:45, 21 May 2015 (UTC)
@Shinyang-i: See Help:Aliases. --Yair rand (talk) 17:55, 21 May 2015 (UTC)
@Yair rand: Already did that, as stated above. Can anyone actually engage with me in a discussion of this topic instead of admonishing and pointing? This is worse than Wikipedia for biting the newbies. Shinyang-i (talk) 17:58, 21 May 2015 (UTC)
This seems to get high rather quickly here. There is no limit to the number of alias an item can have, and this is not a priority to set one. Aliases are just here to help, nothing more. I see no reason to set a limit. Do you have only aesthetics concerns ? If it ain't broken, don't fix it. TomT0m (talk) 18:09, 21 May 2015 (UTC)
Of course I don't have "aesthetics concerns", I was merely trying to get information; I don't know why you're endowing me with emotional motivations on this. I have zero problem with including dozens of romanizations. But I was told to restore what I'd removed, yet what I'd removed was only a drop in the bucket of what apparently was supposed to be there, and thus the answer wasn't exactly clear. The documentation on Wikidata is pretty sparse and there aren't a lot of historic discussions to refer to. Navigation is a little more challenging than at Wikipedia. A lot of the existing items are not "clean" at the moment, and there are no such things as "Good Articles" to look to for best practices. It's impossible to know if changes previously made to items are "correct" or not without asking. Thus, "stupid questions" like mine are inevitable and I think should be treated with a tad more respect than what I feel I've received. I'm pretty blunt myself but this has been a bit much. Not very motivating. Shinyang-i (talk) 18:17, 21 May 2015 (UTC)
Is it clear now ? TomT0m (talk) 18:35, 21 May 2015 (UTC)
Well, no. Further questions:
  • The "fuzzy" spelling issue mentioned in the Aliases documentation - is it live yet and what kind of "fuzzy" spelling does it recognize? For instance, will "nyeo" and "nyuh" in romanization of Korean be seen as "fuzzy" spellings?
  • Should other-language-script names be included, and if so, how many languages? Those from all Wikipedias? Previously only Korean and Chinese were added. How about hanja (subject's name in Korean Chinese-origin characters, which may or may not be the same as their name in Chinese)? Thai? Russian? Etc.
  • Clarify capitalization - in my experiences on Wikipedia, capitalization of a single word makes no difference but it does for all-caps words; i.e. btb = Btb but btb or Btb =/= BTB. Is that correct?
  • "Alias" implies it's an alias for the item in question, but on Wikipedia redirects are often used to send a user who types in the name of a non-notable band member to the band's article. So should non-notable band member names be included in aliases of the band, as well? (Aliases documentation indicates "no", yet they were there and I was told to restore what I'd deleted...)
  • If a person's entry name is their full name but they are often called by just one of those names, does the single name need to be added as an alias? Example: Item name: John Smith. Is "John" needed as an alias? Again, it seems like "no", but I was told to restore what'd I'd removed, and I removed a lot of those...
These are the major issues I've run into so far. I'm sure more will pop up. Thank you for your time. Shinyang-i (talk) 18:51, 21 May 2015 (UTC)
This is not major issues from my point of view, and has a very low impact. So my answer will be I don't really care. And most of Wikidatans as well. That's why you can't find a lot of documentation on this, and as far as I can tell it's a good thing :) So my advice : don't take too much time on this and just focus on something else. TomT0m (talk) 19:09, 21 May 2015 (UTC)
Except maybe for the band members one : it should be clear that an item about a band must not be used to speak on one of its member who should have its own item. So if there is no item for the band member, remove the alias so that this do not happens. On the other hand if the band member item exists and that it is far better know than the name of the band, it could help to find the band item. TomT0m (talk) 19:13, 21 May 2015 (UTC)
@Shinyang-i: From what I can see, the aliases are there to make it easier to find the items. There is no sematic statements based on them. There is no way to add a source to them and there is nothing who tells that they always have to be correctly spelled. I would prefer to say that there is no definite truth about what the aliases should include, except that they should include things that will help the users and the search-engins to find the correct item. My opinion is that you should make a judgement of your own, and not try to identify an exact policy for every issue. Of course, there are aliases that should not be there. If you search for Antichrist, you should normally not be linked to Barack Obama. -- Innocent bystander (talk) 19:18, 21 May 2015 (UTC)
@Innocent bystander: Thank you for the informative and polite answer, and for not telling me no one cares about my contributions. It's much appreciated. As it is, I will just do what I think makes sense and not interact with other uses. First time I screw up and get yelled at, I'll be gone. Thanks again. Shinyang-i (talk) 19:42, 21 May 2015 (UTC)
Hi @Shinyang-i: :) I'm a bit late to the discussion, so I hope you're still around.
I think you have a very good question. The impression I get from both Help:Aliases and the prominence of aliases in the interface is that aliases are only designed for commonly used things which could be considered alternative labels, yet the impression I get from quite a few users is that aliases are just a place to put as many hints for the search as you feel like adding. Those are actually two separate concepts, so it's not a surprise to me that people are disagreeing.
I think that if aliases are not actually important, are just hints for the search and are not really of interest to humans, they should be made a lot less prominent in the interface. When the aliases take up so much space on the screen (above all of the interesting data even!), a direct consequence of too many aliases is that it makes the site harder to use: They get in the way of viewing and editing statements and they make editing labels slower.
If aliases are supposed to be alternate labels which are of interest to humans, then there is the question of how people can add data which is only for helping the search.
Either way, Help:Aliases does not seem to match how many people are currently using the field, which is a problem, because it will lead to more scenarios like this one.
- Nikki (talk) 01:17, 22 May 2015 (UTC)
Hi Shinyang-i. Welcome to Wikidata. You mention you have been looking at Korean and English language labels and aliases so you are familiar with the labels in other languages. If you have a Babel Box on your home page then that can help control which other language labels you see. See User_language for details. Feel free to drop by my page if you have questions you don't feel comfortable about asking here. Hope this helps. Filceolaire (talk) 06:09, 22 May 2015 (UTC)

Time to ping @Denny: here :) And also some of the major writer of Help:aliases, @Thepwnco, Sven Manguard, Bene*: (from the numbers on this history, who else ? ) to check community consensus. TomT0m (talk) 06:23, 22 May 2015 (UTC)

@Shinyang-i: Sorry for the not so great experience you had so far with the project. I hope it will get better. I usually have a really fun time on Wikidata. Thanks, TomT0m for pinging me in. Aliases were designed to support the search, this is also the reason why they don't have qualifiers and references, e.g. "Pope Francis" would be a good alias for "Francis" and the other way around, when referring to the current pope. It is merely meant as a technicality to help people find entities easier. They should not be considered to be a proper part of the knowledge base, but more meta-data about the topic that helps with searching.

Ideally search should get better over time with new software features. Romanizations or other transliterations, fuzzy, soundex-like search terms, etc. should automatically be created and enrich the search for entities, as well as some of the monolingual properties, like ISO codes, etc., translations, etc. But we are unfortunately not there yet. This would make great volunteer contributions, or summer of code-like projects!

I would, in general, try to use common sense in the absence of data. The more prominent a topic, the more care we should give to make sure it is found. The UI is indeed a bit misleading (also, I never been entirely happy with the "also known as" wording). But also less prominent topics deserve to be treated with the Aliases that will help surfacing them. Basically what I do is when I use Wikidata, and something does not show up with the search term I have used when I expected it, I go and add it. It would be great if we could actually gather the data and see where we are missing out: what are the search terms that do not deliver even though we could? @Lydia Pintscher (WMDE):, any chance of getting that data? This would allow us to make further rational decisions on the usage of aliases. --Denny (talk) 18:08, 22 May 2015 (UTC)

Items for Wikisource[edit]

We have Q19063070 which is a document about a chap called James Green, but he also has an item at Q6134904. Can we not just have one item for him, and link all these together in the sitelinks? If this is not possible, then what should the label and description be for the Wikisource item? MSGJ (talk) 08:22, 22 May 2015 (UTC)

These items should be/could be linked together, but not necessarily by the site-project links. If one items is about a person, and another about an encyclopeadic article about that person, it's not the same thing, but they can be linked. This is pretty much analoge to the problems I brought up earlier this week on Wikinews. A news article about a person is not the same as that person, and should have different Q-items in Wikidata that are linked. Edoderoo (talk) 11:23, 22 May 2015 (UTC)
Vlsergey: thanks for your detailed answer. I have a few comments / suggestions:
  • Is this "convention" documented anywhere? It is official?
  • I have relabelled a few of these by hand. Perhaps you could look at my recent contributions and tell me if they look okay?
  • I think it is the responsibility of those creating / importing items to do so correctly, and to make sure they have a proper label and description. The current arrangement is polluting the database making it hard to find other items.
  • Would it be a good task for a bot to relabel all of these items accordingly?
MSGJ (talk) 11:31, 22 May 2015 (UTC)
    • @MSGJ: 1. this convention is not documented yet, it's only implemented in some JavaScript code used by ruwiki and ruwikisource. We are expecting the formal process of discussion to start after arbitrary access is enabled on ruwiki & ruwikisource, because we will need bot work to handle ~300k links like above. 2. Actually, title of item in Wikidata doesn't matter a lot. Current labels looks good (and correct according to convntions above), but the most important point that we have prefix ("DNB00" or "DNB12"). Thus noone will made a mistake linking those items instead of original topics items. 3. Well, may be. But my bot is blocked :-) I assume one need to make those conventions "official", find some consensus among all users of Wikidata, may be open some RFC, after this bot can do his work. -- VlSergey (gab) 11:39, 22 May 2015 (UTC)

interwiki text[edit]

How can I access a specific interwiki link? Lets say I want to get the link to for Q1524 (Athens). If I could get just a text string I would be happy aswell. Thnx in advance -- Spiros790 (talk) 18:55, 22 May 2015 (UTC)