Wikidata:Project chat

From Wikidata
(Redirected from Wikidata:Project Chat)
Jump to: navigation, search
Shortcut: WD:PC
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.
Requests for deletions and mergers can be made here.
IRC channel: #wikidata connect
Start a new discussion
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 July.




for permissions

for deletions

for deletion

for comment



Javascript devs : an entity selector for Wikidata discussion[edit]

@Tpt, Magnus Manske, Lydia Pintscher (WMDE), YMS: (@...) Hi, I begun (sketched actually) the dev of a javascript gadget on test Wikidata to insert an entity in wikitext but I can't find the time to develop it right now. It's upsetting me because I think it's important, so I through a bottle here, if someone think it's important we can make this effort collaborative :)

I thought I would replicate the search or entity selector in item pages in the wikitext edit gadget. Except validation would insert something like {{subst:Q}} in the Wikitext. One other solution is adding something to the search widget to insert in wikitext instead of going to the entity page (Lydia could this be a good query to add a ticket or a feature of the future UI ?)  – The preceding unsigned comment was added by TomT0m (talk • contribs).

Preceded by / Predecessor and Succeeded by / Successor[edit]

GZWDer, Tobias1984, Gymel, Wylve.

Micru has now created replaced (P1365) and succeeded by (P1366) to work alongside follows (P155) and followed by (P156). The plan is that P155 and P156 are for books and albums and items in a series where new items don't replace previous items while P1365 and P 1366 are for political offices and states where the previous item is replaced.

I have done some work on the labels and descriptions in English to try and make this clear (including changing the labels from those in the subjectline above) but they could probably do with some more eyes to check what I have done and to modify the labels in other languages to reflect this change. Filceolaire (talk) 13:50, 22 June 2014 (UTC)

See discussion at Property_talk:P155#new_property_P1366. Filceolaire (talk) 14:13, 22 June 2014 (UTC)

Filceolaire, I am not sure that the new labels of P155/P156 reflect well the meaning. The use of the past for "followed" implies temporal qualities that the sorting of current existing elements might not have. As an alternative, what about "previous in arrangement" and "next in arrangement"?--Micru (talk) 14:49, 22 June 2014 (UTC)
Maybe 'Follows' is a better label for this, especially where the previous item isn't replaced. Number 3 follows number 2 in the series of whole numbers.
I would strongly oppose 'next in arrangement'. Arrangement is too ill-defined a term. "next in series' works but I would still prefer it as an alias rather than a label - in English we don't use that term for the albums from a band but 'follows' works. Filceolaire (talk) 15:04, 22 June 2014 (UTC)
Ok. What about starting a deletion request for structure replaced by (P167).--Micru (talk) 15:40, 22 June 2014 (UTC)
Well, the replacement in structure replaced by (P167) necessarily goes along with complete destruction of the preceding object and well still might have another quality than replaced (P1365). So maybe it is too early to discuss deletion of structure replaced by (P167) and we should wait and see how the new properties flourish. -- Gymel (talk) 16:25, 22 June 2014 (UTC)
@Filceolaire:: I don't think that your clarification will be of any help: The normal(?) language knows about succession in office, not replacement. Even with pope Benedict you wouldn't usually say "he was so worn out that he had to be replaced". So I doubt that the new properties replaced (P1365) &c. will ever lead to a more consistent usage of follows (P155) in the sense of what might have been intended by the new properties. But my opposition has been recorded anyway and it is already documented that I don't grasp the distinction... -- Gymel (talk) 16:25, 22 June 2014 (UTC)

I am lost what the point of all this is. GerardM (talk) 20:51, 22 June 2014 (UTC)

@Filceolaire: please give a link to the formal proposal. --Succu (talk) 21:48, 22 June 2014 (UTC)

Property proposal - 'succeeded by (replaced by this person)'.
GerardM: This change was proposed by GZWDer on the basis that, in the Chinese language (he didn't say which Chinese language) - 'followed by' and 'replaced by' are different concepts and there was no easy way to combine both concepts, in Chinese, in the same property. GZWDer eventually convinced me, and enough other editors, that there is a real difference which justifies having different properties and here we are - 'Follows' and 'Followed by' for series of books, TV shows, albums. 'Replaces' and 'Replaced by' for political and hereditary offices and administrative entities. Filceolaire (talk) 00:11, 23 June 2014 (UTC)
GZWDer: which property do you think we should use for yearly sports leagues? Olympic games? days of the year (May 4 and May 5)? Filceolaire (talk) 00:11, 23 June 2014 (UTC)
When Chinese needs different words to convey a concept than that is what needs to be done for that Chinese language. It does not mean that the words need to be changed for EVERY language. Properties are functional texts, not lexical. Thanks, GerardM (talk) 07:14, 23 June 2014 (UTC)
What it means is that we need separate properties to express the distinction between "follows the previous item in an ordered sequence of items" and "succeeds the previous item in a temporal sequence of singular items", as in "Fool Moon (Q3747536) follows Storm Front (Q3974534) in The Dresden Files (Q2307373)" as opposed to "Elizabeth II (Q9682) succeeded George VI (Q280856) as Monarch of the United Kingdom (Q9134365)". Do you see the distinction, and why people might be confused if the same property were to be used for both? —Phil Boswell (talk) 07:49, 23 June 2014 (UTC)
When you say that a book "follows" another, you mean the order of publishing.. How is that not temporal? It is as clear to me as before. GerardM (talk) 09:01, 23 June 2014 (UTC)
Don't play with words. Properties are concept and the concept of the property follows is "in a list of sequential elements, this element is followed by this one". The sequential order is not specified: it can be temporal, like for a position or a publication date, but it can be another sequence. Or like for book or movie where the prequel is published/performed after the main story. For me the problem of Chinese is just a problem of translating the concept and not as they try to do it now to translate English words. Snipre (talk) 09:45, 23 June 2014 (UTC)
Now, @GerardM:, don't be silly: you know as well as I do that there is a distinction between the publishing order of a book series and the internal chronology of the overarching story. I could have used a different series: how about "Yendi (Q15032947) was published after Jhereg (Q6191037) but precedes it according to the internal chronology"? So Jhereg (Q6191037) follows Yendi (Q15032947) in the internal chronology, but Yendi (Q15032947) succeeded Jhereg (Q6191037) as the most-recently-published book in the series (and was then in turn succeeded by Teckla (Q7692712)). —Phil Boswell (talk) 10:04, 23 June 2014 (UTC)
(edit conflict)GerardM, the other day when proposing to split "part of" you said: "All this assumes that the languages the Wikidata is supporting are able to support all these "finer" points. This is very unlikely and consequently the result of implementing this will be that in many languages the labels used will be exactly the same. This is very bad and to be avoided."
But as you can see, the opposite happens. The more ambiguous a label is, the more problems it causes for other languages to translate it. The reason is that each definition might translate into different words, and we cannot assume that these words are as close together as in the source language. For that reason I would like to ask you to reconsider your position regarding finer properties. As long as broad-meaning properties don't cause distress we should prefer them, otherwise we should seek precision.
In this case there are clearly two concepts, one where all the elements in a set exist in the present, and the other one where they passed the token of existence from one to the next (only one of them was valid at a given time).--Micru (talk) 09:18, 23 June 2014 (UTC)
I agree there are two concepts involved, but the important thing seems to be the value of the property. For example Beatles for Sale (Q207336) was the 4th album of The Beatles and thus succeeded (followed) the third one A Hard Day's Night (Q182518). However Beatles for Sale (Q207336) succeeded (replaced) A Hard Day's Night (Q182518) with respect to being "Number One of the British album charts". And probably we would use neither of the concepts to formulate the relation between numbers (slots) 4 and 5 of album charts. So what property of the target value really is responsible for the distinction between the several "succession" properties? -- Gymel (talk) 12:44, 23 June 2014 (UTC)
Gymel I think in those cases it is better to use the properties followed/replaced as qualifiers to the statement "part of". In the first case you can say, "part of:beatles discography" with qualifier "followed:whatever". In the second case you can say either "part of:List of Number One of..." or "award:Number One of...", in this case with qualifier "replaced:whatever".--Micru (talk) 14:36, 23 June 2014 (UTC)
That is another freedom but not at discussion here: We can say
  • Item (followed/replaced by) another item |qualifier: (position held) item X, eg. "Pope" or "No. 1 hit")
  • Item (position held) item X |qualifier: (followed/replaced by) another item
  • Item (part of) List of items for a position |qualifier: (followed/replaced by) another item within that list
These are purely technical shuffles of the statement (given a number data type one could also play with "position in list" or "position in succession" and could defer the succession and predecession statements) and they are completely independent of the question followd vs. replaced. -- Gymel (talk) 15:32, 23 June 2014 (UTC)

First of all, it would be great if someone can define for me the terms 'replace' and 'follow'. I can't comment on the use of these properties without knowing what kind of relational concepts we are talking about here. To me, the word 'follow' implies that the preceding entity no longer exists in the status it had, as the succeeding entity obtained the aforesaid status; this status can only be occupied by one entity at one time. The word 'replace' implies that the preceding entity ceases to exist entirely and is replaced by the succeeding entity, as the succeeding entity is considered to have some characteristics pertaining to the preceding entity (e.g. building A was demolished and was replaced by building B because building B is built at the location where building A was). Please tell me if I am getting the correct idea about the two terms. Using the term 'replace' in the descriptions of replaced (P1365) and succeeded by (P1366) does not exactly help. —Wylve (talk) 16:01, 23 June 2014 (UTC)

I'm not a native speaker, but "follow" seems to imply some common direction. Thus "prince X" can follow his father "king Y" on the throne, however "duke Z", after having slaughtered king Y and all his kin cannot follow on the throne but only replace him (but may - after having been bitten by a rabid dog - soon follow king Y to the grave: common direction again...). The intended distinction of the properties is as follows: If some office or position can be taken only by one item at one time, the next holder of that position replaces the previous (demolition of the preceeding item is not necessarily implied, see structure replaced by (P167) for that). George H. W. Bush is still alive, but he was president of the U.S.A. and as such was replaced (succeeded) by Barak Obama. Following on the other hand is intended for positions you cannot loose, like being the 41th novel in a series: This will be eternally followed by the item constituting the 42th novel, at least from the time on when the 42th item in that series is published. However George H. W. Bush was the 41th president of the U.S.A. and in a sense still is, because noone ever can take that exact position. And in that sense Barak Obama followed him in office, occupying presidency number 42. The different properties IMHO try to reflect different attitudes of simultanity or exclusivity we have towards novels-in-series and presidents-in-offices (number 41 Bush was the president and now is a former president, whereas number 41 Book always was a novel and still is: it never turns into a "former novel"). Regarding a series "book of the month" might carry both aspects, thus perhaps the difference is not at all in the relation itself but only a nuance in emphasis. Much seems to depend on the actual wording of the list involved: Given a "series" (of whatever, e.g. all U.S. presidencies from past to present) we take a more distant point of view allowing to regard several items at once thus they "follow" each other in the order of the given series. Specializing our view to the generic item in that series (e.g. The President) it turns into something more exclusive: The item of our focus has to be "removed" from the slot before the next item can take its place and become the object of our observation. -- Gymel (talk) 21:27, 23 June 2014 (UTC)
Gymel: My understanding is that when Prince X becomes King X then King Y has to stop being King first so King X replaces King Y - there is only one king at a time.
On that basis 'Replaced by' is the appropriate property for hereditary offices, political offices, states, buildings. One has to end before the next can begin. By this logic 'sports league seasons' should use this property since each season is separate. Even if there were some odd case where the end of one season overlapped the start of the next they would still be separate items, one replacing the other.
'Follows' is the property for TV shows, books, albums, integers and other members of series where there is an order but later items do not replace earlier items. Where the publication order is different from the internal chronology e.g. the Star Wars movies, then the item can be described as 'part of:Star wars chronology' and as 'part of:star wars publication order' and each of these statements can have 'follows' and 'followed by' as qualifiers. Filceolaire (talk) 00:40, 24 June 2014 (UTC)
@Filceolaire: This seems to add more problems than it solves. Plus it seems to be tight by cultural meaning of words. Do you have an example of actual statements where the type of the items in the sequence is not enough to affect the meaning of a statement ? For example :
<X> replaced by search <Y>
could be both correct with X and Y have the same value and the relation is needed to guess what the statement means ? Otherwise it seems useless and just make harder to use Wikidata. TomT0m (talk) 16:11, 24 June 2014 (UTC)
User:TomT0m. No I don't have a case where the meaning cannot be guessed from the domain. Nevertheless I disagree that the properties are tightly linked to the meaning of the words. 'Follows (but does not replace)' and 'Replaces' are slightly different concepts. The confusion arises because these concepts can both be represented, in English and related languages, by the phrase 'preceded by'. Having two separate properties seems to me to unlink the concepts from the language. Your mileage may vary.
I agree with you that, because these concepts are so closely related in English, there may well be a lot of confusion between these properties among English speaking editors. My rewrite of the labels and descriptions for these properties was aimed at trying to reduce this confusion. I would ask anyone who is still confused to look again at those labels and descriptions and make suggestions as to how they might be improved. Filceolaire (talk) 18:27, 24 June 2014 (UTC)
@Filceolaire: It's way easier to add aliases to only one property if it is not ambiguous and adopt a more generic definition. It makes the whole thing simpler and reduce the risk of bad uses. I think we can type the whole sequences : a building can replace another building in the sequence of building that came in the exact place, there cannot be both. An album can follow another album of some author on its all time work sequence, or in the subsequence of work he played with a guitar or whatever. I think we're fine an parcimonious with a sequence concept, and two properties defining the places of items on this sequence expressing precedence and followance, or the rank, in this sequence. Then the type of the sequence gives the other properties : there is only one King of Britain at a given moment in time, there is only one building, a work do not usually replace another, except in the sequence of films that are played in a given room in a cinema. TomT0m (talk) 18:45, 24 June 2014 (UTC)
If you want to play with definition, in the case of a king position, if the previous is dead, you can't use "replaced by". You can use "replaced by" if you have something and you change it. But if the previous king is dead, you can't take him out of its place to put another person. The place is available. In that case the correct term is succession. Then if we want to go further, if a king gives up and choose its successor, you can't use *replace by* because a replacement implies an external cause (no body can say I replace myself). So if you start to distinguish between the different cases, you open the door to a lot of detailed new properties like one property for the change of a person at a position when the previous left or died (typically a king), another one for the change of a person at a position when the previous one is still alive like in election.
By the way a good example. Can we say that the movie Star Wars Episode III: Revenge of the Sith (Q42051) is followed by Star Wars Episode IV: A New Hope (Q17738) ? If we take the story or the title of the movies, yes, but if we take the "movie release", no. That's a very good example. Do we need a new property to be able to do the link between Star Wars Episode III: Revenge of the Sith (Q42051) and Star Wars Episode IV: A New Hope (Q17738) ?
We need a property defined not by a word but by a concept: "in a list of sequential elements this element is followed by this one". In that concept we do not define the link between the elements, the nature of the elements defined the relation between them: in the case of a king, this is a succession, in the case of political or professional position this is a replacement, in the case of books this is a story evolution, in the case of musician's albums this is the released date,... Again a property should be a concept and not a word. If a language can't translate the concept by an unique word, we can use a sentence or to put something like "replaced by (building)/succeed by (person)/..." for the label. The problem of the Chinese translation is not a real problem because it is a language problem and languages defines words acording to a context. The question is it is a real difference in the concept ? Then do we need to distinguish in term of data structure meaning do we need to do the difference with help of several properties to perform queries ? I speak here about logical concepts and not about words because if the concept is the same but according to context you can use different words, the concept can be described by a list of words and you don't need several properties.
So in the case of interest, we have to stop to use words but to use concepts. If I understand well we need two properties for the next concepts:
*B takes the place of A in the sequence, but A still exists
*B takes the place of A in the sequence and A doesn't exist anymore.
Is it correct ? Snipre (talk) 02:57, 25 June 2014 (UTC)
That's the most precise definition we got at this time, we just lack a usecase where using only one property would lead to an ambiguity on the nature on the relation. This is imho the principle who should guide our choices on whether or not make precise property when a generic concept actually fits. Otherwise it will add complexity and make the choice of property harder without any clear benefit. TomT0m (talk) 08:57, 25 June 2014 (UTC)
That would associate most persons in offices with the "followed by" property, contrary to what was intended ("followed" mostly for works, "replaced" for persons). With the exception of course of people who actually died while still holding that office (assassinated presidents, most but not all kings). -- Gymel (talk) 10:43, 25 June 2014 (UTC)
The major confusion here is that we are seeing "A" and "B" as "entities" as opposed to "entities in a certain state". "Entities in a certain state" does not only concern the entity itself, but its condition. In the king example, "King B" does not replace "King A", as that's a logically incomplete statement. What we should be saying is that "Person B as the King" replaces "Person A as the King", hence rendering the former statement false and the latter true. In the example of creative works, Harry Potter and the Chamber of Secrets (Q47209) does not replace Harry Potter and the Philosopher's Stone (Q43361); "Harry Potter and the Chamber of Secrets (Q47209) as the newest novel in the Harry Potter book series" replaces "Harry Potter and the Philosopher's Stone (Q43361) as the newest novel in the Harry Potter book series". The notion of "existence" should be about the existence of the condition that the entity in question is in, not the existence of the entity itself. This means that both types of concepts suggested by Snipre could be seen as one. —Wylve (talk) 12:08, 25 June 2014 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I think the propert(y|ies) we should have should allow to answer the questions: followed by what sequence, sorted according to which criterion ? There can be many submeanings to "followed by" (superseded by, replaced at the same location by, functionnally replaced by, merged into...), and I do not think it can all be solved by using more specific properties. It seems to me that we should rather try to have a consistent use of qualifiers here (I'd probably say: always use a qualifier for defining the sequence, and add one for the ordering criterion when it is not chronological). --Zolo (talk) 13:34, 25 June 2014 (UTC)

Zolo, that could be a good solution. We could just assume that for an element to be sorted it has to belong to a group (part of list/set, with qualifier previous/next), and specify in the list or set what kind of sorting the set has (sorting:chronological, non-destructive). Then we could just use one qualifier for all kind of sorting because the sorting type (if we need it) would be defined in the item that represents the set/group.--Micru (talk) 14:13, 25 June 2014 (UTC)
@Zolo: I'm OK with statements like
<King II> follows (P155) miga <King I>
        in sequence search <Kings of Country X>
, but I think the ordering criteria belongs to the <Kings of Country X> item, otherwise we would duplicate the criteria in all the statements. TomT0m (talk)
Sorry, we have a massive amount of office held.. on persons.. your notion has little bearing on what you find in Wikidata. Lists are largely reduced to indicate the office. GerardM (talk) 18:03, 29 June 2014 (UTC)
@GerardM: I'm sorry, what ? I can't understand where you are going. We've got a massive definitions problem here, it's like we do not speak the same language. TomT0m (talk) 18:55, 29 June 2014 (UTC)
@TomT0m: Have a look at Ronald Reagan.. he has predecessors and successors on the different positions he held. The notion that it should all be on the position itself flies in the face of established practice. It is therefore not a workable notion. Thanks, GerardM (talk) 19:15, 29 June 2014 (UTC)
@GerardM: Oh, OK I think I understood the problem. For me, an ordering criteria is "alphabetical order of the name", for example. But what is called "ordering criteria" here is more "the value of the name that is used to make the sort", right ? So when I say "the ordering criteria belongs to the sequence item", I mean we should have a claim
<King of Britain sequence> ordered by search <date of reign>
        type of sorting search <chronological>
for example. TomT0m (talk) 20:47, 29 June 2014 (UTC)
@TomT0m: Sorry... you speak goobledegook.. Do you REALLY think that templates make clear what you mean.. You are right, we speak a different language. Having sort order at the Wikipedia end is .... not obvious, not really sensible. Thanks, GerardM (talk) 21:20, 29 June 2014 (UTC)
To sort you must choose a way to sort. There is plenty : chronological, reverse chronological for dates, alphabetical order for strings and so on. If you want to sort a set of item, you have to chose an order and a property, property I call ordering criteria. I would think a non ambiguous terminology would be to call the property a "ordering property", and to call the value of this property in each item a "ordering value". You referred to "ordering value", I was thinking of storing the type of ordering + the sorting property. Hence the misunderstanding. TomT0m (talk) 07:07, 30 June 2014 (UTC)

Maybe as example to clarify (the way I understand it): Chapter 2 of a book follows/comes after chapter 1 (The two items belong together in an ordered fashion but coexist and the second item does not take the place of the first one, nor daoes it render the first one obsolete); edition 2 of a book replaces/succeds/takes the place of edition 1 of the same book (the second item in the sequence renders the first one deprecated/ takes the position previously held by the first item).--2A01:2A8:8401:D701:C85D:1A86:C432:F0E 21:13, 29 June 2014 (UTC)

Please sort out the mess you made[edit]

President of Venezuela and many similar instances.. GerardM (talk) 19:52, 24 June 2014 (UTC)

When you use the phrase "replaced by" for people who occupy an office, it is really disrespectful. It also has a meaning that is not the same at all as "succeeded by". Thanks, GerardM (talk) 08:37, 25 June 2014 (UTC)
User:GerardM What mess is that? Who are you talking to? If you are talking about the label for P1365 and P1366 then that is easy to fix. Please make sure the new label is chosen to minimise confusion with P155 and P156. Filceolaire (talk) 21:22, 25 June 2014 (UTC)
I have been quite clear that I am opposed to this split. So why do you expect me to fix the mess "you" insisted on. Thanks, GerardM (talk) 21:35, 25 June 2014 (UTC)

On the usage of part of (P361)[edit]

I'm still not convinced. It took a long discussion to sort out correct uses of part of, with Micru and Markus Krostch amongst others. In the context of any kind of sequences, I think a new property like in sequence would be better and less confusing than using a property used for whole/part relationship : the nature of the items is not the nature of the list. part of (P361) fits for physical objects composing a bigger physical object, I don't think it fits of a king sequence is composed of kings.

TomT0m: I agree with you, actually I wanted to write "part of set", but by mistake I put a colon where there should be none (corrected now). Regardless of how we call the property ("part of set" or "in sequence"), I would do it the other way around
<King II> in sequence search <Kings of Country X>
        follows (P155) miga <King I>
. The reason is that it is more clear to have that grouping in the cases where an item belongs to more than one set.--Micru (talk) 14:45, 25 June 2014 (UTC)
User:Micru I can see the case for a new <in sequence> or <part of ordered set> property for TV shows, books, integers, elements, football league seasons, etc. I would even use it for days of the year (common year or leap year) and state that January 1 follows December 31. I would NOT, however, use this for hereditary or elected or appointed offices. We have position held (P39) for those and it is working very well. Filceolaire (talk) 21:48, 25 June 2014 (UTC)
Filceolaire, position held (P39) is clearly a subproperty of "in sequence". For creative works we already have series (P179) which also can be qualified with the series order. So if I understood right, the proposed solution would be:
Any objection or concern? --Micru (talk) 07:49, 26 June 2014 (UTC)
Position held has NOTHING to do with "in sequence". For instance Ronald Reagan "position held" "President of the United States... WHAT? What you probably mean is that the usage of "preceded by" is in the sense of a qualifier. Arguably, this property should only be used in qualifiers. Given the class of the main property, a specific text is to be preferred. In this case "preceded by". GerardM (talk) 06:28, 27 June 2014 (UTC)
GerardM, how many people there are at any given time that are "President of the United States"? One or many? Do they happen to be "President of the United States" all at once or in sequence? That you think that a given label might be more suitable doesn't mean that the underlying concept is different... I mean, I would keep both for clarity, but let's look at things how they are and not how we would like them to be.--Micru (talk) 07:01, 27 June 2014 (UTC)
The text "replaced by" is often offensive. That is how it is. GerardM (talk) 07:14, 27 June 2014 (UTC)
GerardM, I totally understand. Above I suggested to delete the 'replace' properties, and use a more neutral label that can represent both cases in p:p155, p:p156. "previous" and "next" seem quite inoffensive to me.--Micru (talk) 08:14, 27 June 2014 (UTC)
Question on sports (I'm working on sports topics a lot lately in Wikidata, so this interests me the most :): So an item like 2013–14 NBA season (Q11144228) would then be <in sequence> NBA seasons? Probably need to create a new "NBA seasons" item then or rename List of National Basketball Association seasons (Q2622507) to that as Wikidata:Project_chat#.22is_a_list_of.22_is_used_to_build_a_query suggests in some way. I guess other sports tournaments like 2009 Dubai Tennis Championships (Q807962) could/should then also be retagged? Currently that one is instance of (P31)=Dubai Tennis Championships (Q299409), with the new property we would tag this like <in sequence>=Dubai Tennis Championships (Q299409) and change instance of (P31) to point to tennis tournament (Q13219666)? Anyway, a property to properly connects sports seasons to a "main topic" is certainly a good idea, currently there is no good way to do that as far as I see this. --Bthfan (talk) 13:09, 26 June 2014 (UTC)
I guess you could use the existing item: 2013–14 NBA season (Q11144228) <in sequence> List of National Basketball Association seasons (Q2622507). The other use in the Dubai example also seems rational. "List of" looks like a top-down approach, nothing wrong with it either, but perhaps it is harder to maintain both approaches.--Micru (talk) 13:41, 26 June 2014 (UTC)
Is there no chance to fix the problem in the lines of the current approach? I mean, stating feels rather straightforward and when trying to express only indirectly by plus we can note predecessors, successors and sequence numbers more consistently, but this leaves us with the question where to put
<> start date (P580) miga <2001-01-20>
and the other qualifiers of position held (P39): We cannot qualify the "in-sequence" property with start and end-dates, since this could be interpreted as some error: "At some time in the past, George W. Bush (Q207) was considered to be an U.S. president, now we know better..." Having to double-specify the presidency by using and not getting rid of at the same time seems rather awkward. Also we have to give some "flesh" to list of Presidents of the United States (Q35073) and therefore we must associate the individual presidents to the list anyway. What if we do not state in the George W. Bush (Q207) item but rather in the list item list of Presidents of the United States (Q35073)? This would imply creation of the inverse property to "in sequence", namely "has member in a (partially) ordered set" and we would be able to formulate . -- Gymel (talk) 15:53, 26 June 2014 (UTC)
Gymel, I would keep both approaches. I wouldn't recommend building a list on the item, sometimes there are just too many elements in a sequence to put them in a single item. You could add the position in the sequence as a qualifier, the dates can be extracted them from other statements. --Micru (talk) 07:05, 27 June 2014 (UTC)

Proposed conventions[edit]

For sorted sets:

  • in sequence: A and B belong to a sorted set. B(in a certain state) takes the place of A(in the same state) in the sequence, but A no longer exists in the same state (but it may exist in another state). Main use: sorted lists of only one element valid at any time (e.g. buildings in one location). Use narrower properties where appropriate (e.g. position held (P39)).
  • in series (p179): A and B belong to a sorted set. B(in a certain state) takes the place of A(in the same state) in the sequence, but A still exists (in the same state). Main use: sorted lists with all elements valid at any time (e.g. creative works and abstract concepts)

For unsorted sets (related to this RFC), no change to the labels is proposed, but I changed them just now for the sake of making easier to understand what they mean:

  • in set (p361): A and B belong to an unsorted set of their same domain (physical, event, abstract), but they are not similar to the set
  • in similar set (p279): A and B belong to an unsorted set of their same domain (physical, event, abstract), and they are similar to the set
  • in strongly similar set (p31): by their strong similarity (mostly grounded in their real life existence), A and B form an unsorted set defined by all their shared characteristics.

Shall we go ahead with this? It would need a new property "in sequence" plus two additional ones:

  • sorted according to property: which property is to be considered when sorting (start date, birth name, etc)
  • sorting criteria: according to which criteria the set is sorted (ascendent, descendent)

--Micru (talk) 10:07, 30 June 2014 (UTC)

You lost me completely. First of all, you refer to properties that have names. It helps understanding. Secondly why prescribe sorting? When something CAN be dated, it does not follow that they are dated. Look at how Reasonator does things.. It works well enough at no cost of extra overhead.. Really, you think to much but are not obvious nor clear in what you want. Thanks, GerardM (talk) 11:06, 30 June 2014 (UTC)
No. We don't need to specify the sorting criteria because even if there are several creteria, there is only one per type of sequence. You don't sort sequence of king per alphabetic order. So instead of creating a bunch of new qualifiers which won't be used in 50% of cases, just define the differrent sorting criteria in the talk page of properties followed/precessed. Stop creating complexity where good sense is enough. Snipre (talk) 11:48, 30 June 2014 (UTC)
GerardM, thanks for your feedback. I would like to see better control mechanisms (similar to "constraints" used in properties) so we don't have to worry much about human errors (or vandals), and for that we need first to have some clear definitions and then some redundant information. Yes, Reasonator does great things, but it doesn't matter to it if the information is consistent or not, because it has no control mechanisms and no intelligence. OTOH, it must not be a "prescription", just a recommendation for cases when we need (or want) to use it. An item can be in several lists, and each one has different sorting needs, sometimes it is clear, sometimes it isn't.
There were users that felt the need to make several distinctions in the meaning of each property, others that didn't feel that need, so what I am trying to do is build consensus and summarize points of view. It is an asset that we must cultivate, and it takes effort, discussions, and thinking. I remind you that much progress done here didn't come only from software features, but also from us reaching common points, whenever possible. --Micru (talk) 11:47, 30 June 2014 (UTC)
And I want to conclude that the properties followed/precessed can be a choice and not as simple to define. For me this kind of information shouldn't be put in the items as qualifiers. If I have to define the predecessor of a person to a possition, I extract the whole list of persons which held the position and then I used the start/end qualifiers to define who is the predecessor plus other criteria I want to include. An example are the popes and the antipopes. By using successor and predecessor I choose implicitly the official list of popes but if this is not what I want ? We have to let the data user selecting its own list and then its own predecessor/sucessor according to its criteria. Wikidat shouldn't propose preprocessed data but raw data and users should then work and analyze data according to their preference. Snipre (talk) 12:05, 30 June 2014 (UTC)
@Micru: If you should specify the sorting criteria this means this is not the result of a common knowledge but the result of a choice then this is not the work of Wikidata to order data according to specific criteria. So the conclusion is simple: followed/precessed should be explicit or if not, the use of properties followed/precessed is not possible because this is a choice which can be non consensual. Snipre (talk) 12:10, 30 June 2014 (UTC)
@Snipre:, while I agree with you about those properties being a result of a choice, "followed/preceded by" are useful for navigation when editing. They could be bot maintained when there are many elements because in those cases there is for sure an item representing the sequence. We also have to consider that there are sequences of just two or three elements, and in those cases it is just practical to link them (ex. Q373096 <=> Q10298666).--Micru (talk) 12:33, 30 June 2014 (UTC)
@Micru: your description of part of (P361), instance of (P31), and subclass of (P279) does not seem to make any sense to me. More precisely : I don't understand how they are related to the subset, set membership or physical composition (a car is composed of wheel and other elements). I'm not sure about your definitions of sequences and series either. As we discussed, the nature of the elements gives a lot of information about the nature of the sequence. And we can use items to type and defines the sequence. I'm not really sure we need that distinction, and we have no real usecase where this adds information. TomT0m (talk) 16:05, 30 June 2014 (UTC)
@TomT0m:, I just posted a comment on the RFC explaining my reasoning about membership properties on the RFC. About sequences and series, are you suggesting that we use only one property for both? If so, which one?--Micru (talk) 18:35, 30 June 2014 (UTC)

The terminology used is insulting[edit]

At the end of his term Ronald Reagan was succeeded by Mr Bush. He was not replaced. Replacement indicated an involuntary action and consequently is insulting. All the words used are hardly intelligible but what is unclear about us using terminology that has not even the same meaning as the old "succeeded by" ? Why persist in this ?

@GerardM: from your comment it is hard to understand to who you are addressing your comment to, or what you are replying to. --Micru (talk) 11:59, 9 July 2014 (UTC)
I address the terminology used, and this not for the first time. It concerns everyone. I changed the word and was reverted.. The use of "replaced by" is as said insulting, it means something different from what it is supposed to mean. Thanks, GerardM (talk) 12:13, 9 July 2014 (UTC)
Which word did you change? And who is supporting to use "replaced by" as you say?--Micru (talk) 12:18, 9 July 2014 (UTC)
Replaced by is the label used. I changed it back to "preceded by" this is the right and only obvious phrase in English. GerardM (talk) 13:14, 9 July 2014 (UTC)

Extension of the property "series"[edit]

Just wondering, will the property proposal for "in series" happen soon? Or is this currently part of the RFC discussion at Wikidata:Requests_for_comment/Refining_"part_of"? At least for sports seasons/events (there are a lot of items like that) "in series" would be useful to connect those to an "upper" topic. This would a good alternative to using "instance of" with an generic season/event item for this series and/or using "part of" on every season/event item. We could also think about changing the meaning/definition of series (P179) though, currently that one is for creative works only. --Bthfan (talk) 08:22, 7 July 2014 (UTC)

@Bthfan: I just started the discussion in Property_talk:P179.--Micru (talk) 12:19, 9 July 2014 (UTC)
The notion that this makes an "instance of" unncessary is something I disagree with, GerardM (talk) 13:12, 9 July 2014 (UTC)
@GerardM: you meant what I said about "instance of"? Sure, but then I could set it to instance of=sports season (or whatever seems right) instead of creating generic items like e.g. "NBA season" to use as value for "instance of". --Bthfan (talk) 15:16, 9 July 2014 (UTC)
I agree. instance of (P31) is still useful in those cases. For example in a TV series there is some episodes of different kind sometimes, like musical episode in a series in which other episode are not. TomT0m (talk) 14:14, 9 July 2014 (UTC)

Linking identifiers[edit]

We recently added BBC Your Paintings (P1367) (for instance, Reg Gammon (Q15381364) {BBC Your Paintings (P1367)  reginald-gammon). This is a reference to the URL How may we indicate that relationship., in a machine-readable manner? Or, better still, make the string a link? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:54, 23 June 2014 (UTC)

Links can be added via this script/gadget: you can enable that gadget for your account via for adding that property I would ask on the Talk page there. --Bthfan (talk) 19:20, 23 June 2014 (UTC)
Actually @Zolo already did this now :) so enable the gadget and it should work --Bthfan (talk) 19:23, 23 June 2014 (UTC)
Thank you, however I don't mean for myself, but generally. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:04, 27 June 2014 (UTC)

Anyone? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:40, 2 July 2014 (UTC)

Andy, from what you wrote I don't understand what you want to know. Usually if you want an identifier to be converted into a link, you post a request on this talk page and then some admin updates MediaWiki:Gadget-AuthorityControl.js. To get a list of correspondences between any item and their external identifier, you can use this tool by Magnus. If that is not what you want to know, please provide more details.--Micru (talk) 07:35, 4 July 2014 (UTC)
When a person who is not logged on, or who is logged in but does not have the gadget enabled, views Reg Gammon (Q15381364), the BBC Your Paintings (P1367) value "reginald-gammon" should be a blue hyperlink, to - how can we make that happen? In other words, how can we make the gadget behaviour the default for all users? I would further suggest that, rather than having MediaWiki:Gadget-AuthorityControl.js, the URL format should be part of the parameter definintion. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:50, 4 July 2014 (UTC)
I would not object to making the authoritycontrol gadget opt-out instead of opt-in. --Jakob (talk) 22:08, 4 July 2014 (UTC)

Thematic mapping of geographic statics, US and World[edit]

My interest lies in statistical phenomenology that can be shade mapped by states in the US and countries around the world. Such phenomena as demographics, crime, opinion, health, etc. -- any phenomena that has statistics by area for which the statistic can be listed, statistically summarized, then I want to convert the statistical phenomena into shade maps of the geographic areas they represent and possibly apply bi-variate statistics on the geographic distribution such as center of gravity, etc.

I am new to Wikidata and want to get an understanding of the statistical data you are assembling and if data mapping is already produced or would be amenable to adding such map graphic products to your data series.

I am having trouble getting connected into Wikidata. I contribute financially to Wikipedia.

Please help me get connected and directed appropriately.

Charles Barb

MEP information on Wikidata[edit]

I've been approched in Brussels by several organisation that spend considerable resrouces in compiling and udpating all relevant data of the Members of the European Praliament. They would all benefit from having all this information in a human and machine readable way to use in their respective projects. They would also help filling in the information. I would therefore like to create an EP MEP task force that puts all the relevant information about MEPs (including political group, committee participation and official contact information) as statements in their Wikidata item. As I am not experience with editing WD, I'd be grateful for some more experienced Wikidatian to join this effort! I guess we will need to add more statements and coordinate the effort somehow. Thanks in advance! --Dimi z (talk) 13:25, 30 June 2014 (UTC)

I am interested in helping out. I have been involved in members of parliament for many countries.. including United Kingdom, India, South Africa.. Thanks, GerardM (talk) 13:53, 30 June 2014 (UTC)
Nice idea. I added some more constraints so that Wikidata:Database reports/Constraint violations/P1186 can be used to improve existing items. Quite a few existing items to improve! Could probably add even more constraints (country and language related) after we trimmed down this list.
If we would have a list of id's and names, User:Magnus Manske could probably add it to mix'n'match. That way we can find the MEPs who don't have an item here yet. Multichill (talk) 05:12, 4 July 2014 (UTC)
I have imported the current ~750 MEPs into mix'n'match. Since we have a dedicated property, I could auto-match >50% of them through that identifier. The rest is up for manual matching here. --Magnus Manske (talk) 09:55, 4 July 2014 (UTC)
Update: Have also imported all people with a MEP directory identifier (P1186) into mix'n'match. >2000 people in the MEP list there now, >80% have a Wikidata item. Many show up in the pre-filled Google search; we do have Wikipedia articles, and thus Wikidata items, for most of the remaining ~20%! --Magnus Manske (talk) 10:51, 4 July 2014 (UTC)
Dimi z as you can see we already started editing quite a bit. Do you know if all new members already have an entry? Take for example Theodoros Zagorakis (Q296452), he seems to be missing.
Thanks Magnus for adding these! I'm matching them and fixing constraints, maybe you want to help too Gerard? Multichill (talk) 10:15, 5 July 2014 (UTC)
I had already added many MP's based on them being in a category.. Thanks, GerardM (talk) 11:36, 5 July 2014 (UTC)
Hi, everyone! Thanks for the awesome help. I admit I am sometimes somewhat confused when it comes to Wikidata. Still learning and - more importantly - learning the ropes. Taking this MEP as an example, I'd like to be able to add the following statements: which Parliaments's he's be sitting in (i.e. 7th, 8th legislative period), which committees he's been sitting on (as active/substitute member), official mailing addres, official phone number, official email (all public data) and political group within the EP (not same national political party). Are there such statments already and if not, where can I request them? I'd like to be able to explain and show to newcomers how to edit Wikidata. Sorry for the perhaps stupid questions, but I really appreciate all help! Thanks! --Dimi z (talk) 12:10, 5 July 2014 (UTC)
Dimi z, I tracked down/created the items for the different elections (Elections to the European Parliament (Q1128324)) and sessions ((legislative term (Q15238777)):
  1. European Parliament election, 1979 (Q1376068) -> First European Parliament (Q17315702) - en:Category:MEPs 1979–84
  2. European Parliament election, 1984 (Q1376075) -> Second European Parliament (Q17315703) - en:Category:MEPs 1984–89
  3. European Parliament election, 1989 (Q1376076) -> Third European Parliament (Q17315704) - en:Category:MEPs 1989–94
  4. European Parliament election, 1994 (Q1376071) -> Fourth European Parliament (Q17315706) - en:Category:MEPs 1994–99
  5. European Parliament election in 1999 (Q1851815) -> Fifth European Parliament (Q16836722) - en:Category:MEPs 1999–2004
  6. European Parliament election in 2004 (Q1331918) -> Sixth European Parliament (Q4642661) - en:Category:MEPs 2004–09
  7. European Parliament election, 2009 (Q210152) -> Seventh European Parliament (Q4644021) - en:Category:MEPs 2009–14
  8. European Parliament election, 2014 (Q1376095) -> Eighth European Parliament (Q17315694) - en:Category:MEPs 2014–19
You can use part of (P361) Eighth European Parliament (Q17315694) on the members to indicate what parliaments they are sitting in. The category can be used as input for autolist2 to add it to all the people (or maybe Gerard can help out here?)
I used member of (P463) for the committees, but I'm not 100% sure if that's the right way to go
mailing addres, official phone number, official email is probably a bit out of scope
For the political parties you can use member of political party (P102) and add the one or more parties he or she is member of (national or European)
You can see all of this in practice at Marietje Schaake (Q433561). Multichill (talk) 14:50, 5 July 2014 (UTC)
Personally I have an trouble using part of (P361) in this case (hard to draw the line,
<some person> part of (P361) miga <some musical group>
is ok for me) and couldn't the individual parliaments be used as qualifiers for , e.g. with ? Start and end dates of the terms served then often could be derived as an additional benefit. -- Gymel (talk) 15:14, 5 July 2014 (UTC)
I like adding it directly. It seems to be common for parliaments, see for example Barack Obama (Q76).
Start and end date might be different, take for example Jens-Peter Bonde (Q1264754)
Magnus everything is matched. Can you flush it out? Multichill (talk) 16:50, 5 July 2014 (UTC)

Wikidata is now updated with all mix'n'match data. I found ~800 MEPs not on marked Wikidata or listed in mix'n'match; the latter requires names (easy to get from Wikidata) and MEP IDs (could also get from Wikidata if entered there, but that would defeat the purpose of the tool...). --Magnus Manske (talk) 21:01, 5 July 2014 (UTC)

Magnus, I'm a bit confused. Are you sure you imported everyone from the website? Take for example Abel Matutes (Q318474). It's not linked, it doesn't show up in mix'n'match, but he does have a page with id 1988. On the other hand we have people like Alain Savary (Q477121) who were in the European Parliament prior to it being elected. People like that don't seem to be documented on their website. Multichill (talk) 09:34, 6 July 2014 (UTC)
After a bit of clicking around I found the full list (xml). It contains 3593 items, mix'n'match only 2056 items so we still have to find about 1500 missing items. Multichill (talk) 11:13, 6 July 2014 (UTC)
Magnus updated it, so about 1500 more items in mix'n'match to work on! Multichill (talk) 21:08, 6 July 2014 (UTC)

Time for some selfies![edit]

Hey folks!

Wikimania is coming up and I need your help! I will be giving a keynote about Wikidata and would like to show the human side of our awesome project. Wikidata wouldn't be what it is without all of you and I see Wikimania not only as an opportunity to highlight all of the exciting software developments around Wikidata, but to also show off what an awesome, creative and diverse community we have here!

To do this, I would like as many of you as possible to send me a photo of yourself - a selfie (Q12068677) would be perfect. Bonus points if it either demonstrates a connection you have with Wikidata or includes the Wikidata logo somehow. You can be subtle and sneaky with the logo or go big and bold—the only limit is your creativity! All submissions should be licensed CC-BY and you must have the right to upload and use the image. Additionally it’d be awesome if you could send me a sentence or two about what Wikidata means to you. (Either use my user talk page or send me an email.) Submission deadline is July 20th. When you're ready, upload your file to Wikimedia Commons in the category “Wikidata selfies”. Need some inspiration? Check the category.

I’m sure we can be just as famous as the Oscar selfie :P

Cheers --Lydia Pintscher (WMDE) (talk) 15:44, 30 June 2014 (UTC)

The category is starting to fill up \o/ But we still need quite a few more for this stunt to work. I just uploaded a picture for the dev team. Please consider joining. You can also send the image to me if you want. --Lydia Pintscher (WMDE) (talk) 13:00, 4 July 2014 (UTC)

Wikidata just got 10 times easier to use[edit]

Hey folks :)

We have just deployed the entity suggester. This helps you with suggesting properties. So when you now add a new statement to an item it will suggest what should most likely be added to that item. One example: You are on an item about a person but it doesn't have a date of birth yet. Since a lot of other items about persons have a date of birth it will suggest you also add one to this item. This will make it a lot easier for you to figure out what the hell is missing on an item and which property to use.

Thank you so much to the student team who worked on this as part of their bachelor thesis over the last months as well as everyone who gave feedback and helped them along the way.

I'm really happy to see this huge improvement towards making Wikidata easier to use. I hope so are you.

Cheers --Lydia Pintscher (WMDE) (talk) 19:20, 1 July 2014 (UTC)

That's nice! --AmaryllisGardener talk 19:41, 1 July 2014 (UTC)
That's great :) Well done to the developers. Delsion23 (talk) 20:22, 2 July 2014 (UTC)
Woooooooo! ·addshore· talk to me! 13:15, 5 July 2014 (UTC)

Gender redundancy[edit]

Why do we have male (Q6581097)/female (Q6581072) and male animal (Q44148)/female animal (Q43445)? Surely the latter pair are redundant, and we can rely in instance of to tell us whether the subject is human, a horse, or whatever? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:28, 2 July 2014 (UTC)

Andy, see Property_talk:P21#Values_not_making_sense. I agree with your position, though. The current redundancy conflates two properties: biological taxonomic classification and sex/gender. This is a bad separation of concerns. I'm aware of the awkwardness in some languages of having a property for only sex/gender, but given that Wikidata is more of an ontology than a linguistic datastore, I think applying male and female to non-human animals makes more sense. Emw (talk) 22:42, 2 July 2014 (UTC)
Andy: At one stage we did use the same item for both these concepts. It was changed because various users said that, in their language, there was no word that covered both male humans and male animals and the same for females. After some discussion the current arrangement was agreed.
There was another big discussion about whether we should have separate properties for sex and gender (You didn't ask about this but I thought others new users might be interested in the history). I advocated against separate properties on the grounds that in pretty much every case we have no way of knowing what the biological sex of a person is. We can tell what gender they publicly express and from that we can guess at their biology. After a long discussion we agreed to change the label for P21 from 'sex' to 'sex or gender'. The en:intersex article lists five different factors that control or express a persons 'biological sex' and there are a number of other factors related to gender. In the cases where we have specific information on these we can have multiple values for sex or gender (P21) miga with qualifiers to specify which factor each value refers to. Filceolaire (talk) 21:09, 3 July 2014 (UTC)
Several in-depth threads of the discussion Filceolaire refers to have a jumping-off point at Property_talk:P21#Issues_list. An archived proposal for a separate 'gender' property is at Wikidata:Property_proposal/Archive/20#gender. Tying that discussion about sex/gender into this one about male/male animal, we see that there is little appetite in the Wikidata community, at least with this property, for a clear separation of concerns. Not only does P21 conflate sex and gender, it also conflates sex and biological classification. Emw (talk) 11:11, 4 July 2014 (UTC)

Full list of properties[edit]

I extracted the full list of properties here. I can do it for other languages if the concept is accepted, but I will have to convert the unicode characters into ascii characters first. Put comment in the talk page. Snipre (talk) 22:55, 2 July 2014 (UTC)

Snipre - For better loading time of Wikidata:List of properties, I moved the table to Wikidata:List of properties/all in one table. There is also Wikidata:List of properties/all. Tamawashi (talk) 22:20, 5 July 2014 (UTC)
For computing purposes, there is also, who is usable with software libraries like catlib in pywikibot to work on all properties in a really few lines of code (like 2 lines to iterate on all properties). TomT0m (talk) 14:49, 6 July 2014 (UTC)


Maybe this template [1] must be replaced with human (Q5) instead person (Q215627) ? --Rippitippi (talk) 01:02, 4 July 2014 (UTC)

Not in all cases we want to restrict ourselves to humans. There are several properties which can be used for humans and fictional characters. For properties that are only for humans we should probably create a new template Constraint:Human. However, what's more important is that User:KrBot (@Ivan A. Krestinin:) is supporting such templates. --Pasleim (talk) 03:48, 4 July 2014 (UTC)
@Pasleim: I say it should be called "constraint:domain" and extended to any classes. I did not find a matching constraint in Template:Constraint. TomT0m (talk) 11:05, 4 July 2014 (UTC)

What do you wish you had known when you first started using and/or contributing to Wikidata?[edit]

Hi all,

I recently made some edits to Help:FAQ as part of a larger sitewide documentation overhaul (more info on this here).

To compare my edits with the previous version of Help:FAQ please see the diffs at

As you'll see, my changes mostly consist of deleting a few questions that were either duplicates of others or did not seem like they would be questions that were frequently asked, for example, "According to the history of an item I changed, the number of bytes went up or down significantly, even though I only made a minor change. Why did this happen?" I also added in more links to our other documentation.

I'd like to update the page a bit more and re-arrange the questions thematically but first would love your help with improving the content we have. Could people please have a look at the current FAQ and let me know what is missing? If you're not sure, maybe consider the following as jumping off points:

  • What questions did you first have when you started contributing to Wikidata?
  • What did you wish you had known from the start and what concepts or task took some time and perhaps trial n error to learn and understand?
  • If you often help newbies and/or hang out on IRC, what types of questions do you frequently hear from other users?

Please also let me know if you have any concerns about the changes I've made (for example, if I've removed a question that you think is absolutely essential to have on the page). If you see something wrong (like a typo) or something that could be easily improved upon (like an example), please just go ahead and fix it—no need to comment here first!

Thank everyone!. -Thepwnco (talk) 22:52, 4 July 2014 (UTC)

Thepwnco, some things that should appear in the FAQ:
  • Babel templates
  • Reasonator, Autolists2, and other tools
  • Wikiprojects
--Micru (talk) 10:37, 5 July 2014 (UTC)
Thepwnco - great initiative!
  • Babel extension, mentioned somewhere above, so that at the item editing page one could see labels and descriptions in languages one is interested in.
  • Differences between WikiProjects and Task Forces, some seem to have been renamed from TF to WP
Tamawashi (talk) 13:21, 5 July 2014 (UTC)
List of properties, of course, and their respective Property documentations, and the method to search for a property ("P:" prefix to search keywords). LaddΩ chat ;) 14:07, 5 July 2014 (UTC)
  • WD:showcase items. These are really valuable examples of what properties to use on different types of item and how to use them and will get more valuable as more showcase items are added. Filceolaire (talk) 21:25, 5 July 2014 (UTC)
I am normaly not at wikidata, but I saw this in the newsletter. If I saw in the history that the change in bytes differed more than the edit - I would require an explanation. Christian75 (talk) 21:53, 5 July 2014 (UTC)
@Thepwnco: I feel too old in Wikidata to answer this question. Is this project chat the right target ? I think a better place to ask would be a real newb place, like the newbies welcome and patronage projects on Wikipedia for example. Did you talk to Wikipedia people who helps newbies ? Maybe a cool thing to do for you would be to help a newb to its first steps into Wikidata :) TomT0m (talk) 14:42, 6 July 2014 (UTC)
@TomT0m: don't feel too old - your experience with the project is what makes you an expert on what's useful to know when working with Wikidata! There will be other efforts to include newbies and their perspectives (including looking at what materials are available on other Wikimedia projects and what questions are coming into the info e-mail account) so please feel free to add anything you can think of. -Thepwnco (talk) 17:27, 7 July 2014 (UTC)

categories that include info so that Reasonator shows its content[edit]

These categories include a statement of "is a list of" "human" with qualifiers that make it as precise as necessary. When you look at them with Reasonator you will be shown the result of a query defined by these qualifiers.

Adding statements based on categories using Autolist2 works fine when you restrict it by requiring "instance of" "human". Thanks, GerardM (talk) 07:48, 5 July 2014 (UTC)

Shouldn't these use category's main topic (P301) miga or category combines topics (P971) miga instead of is a list of (P360) miga? Filceolaire (talk) 21:31, 5 July 2014 (UTC)
No, that is not how it works. The reason is that it also works on "Wikimedia list article". GerardM (talk) 07:18, 6 July 2014 (UTC)

Pages do not load properly when using Firefox[edit]

I am mainly using Firefox 30, and I find out that data item pages do not load properly when using Firefox, including:

  • Not showing "(xx entries)" which should be next to "Wikipedia pages linked to this item"
  • When clicking [edit], instead of showing a field for inputing new link and the options [remove] [cancel] on the smae page, it links to another page in Special:SetSiteLink

However, such problems do not happen when using Google Chrome.--Itkit (talk) 08:06, 5 July 2014 (UTC)

I experienced similar problem for English language version. I also use Firefox 30. --EugeneZelenko (talk) 14:05, 5 July 2014 (UTC)
I can't reproduce your problems at the moment. If the problems persist, can you please give us more details about them? Do these problems also appear if you append ?debug=true to the item url? And can you look into Firefox's web console (Ctrl + Shift + K) and tell us what errors (if any) appear there? - Hoo man (talk) 14:59, 6 July 2014 (UTC)
Somehow Q17305033 with ?debug=true works. Error messages from web console for plain Q17305033 are:

"JQMIGRATE: Logging is active" load.php:150 "Exception thrown by jquery.wikibase.listview" load.php:161 "TypeError: base._childConstructors is undefined" TypeError: base._childConstructors is undefined Stack trace: $.widget@ line 4 > eval:3:652 @*:43:1 @*:43:1 runScript@ execute@ mw.loader</<.implement@ @*:45:1


"Exception thrown by jquery.wikibase.snakview" load.php:161 "TypeError: base._childConstructors is undefined" TypeError: base._childConstructors is undefined Stack trace: $.widget@ line 4 > eval:3:652 @*:78:1 @*:78:1 runScript@ execute@ mw.loader</<.implement@ @*:92:1


EugeneZelenko (talk) 14:20, 7 July 2014 (UTC)

occupations and art media[edit]

Hi I was linking some painting images used as illustrations on Wikipedia to their associated Wikidata biography items and one of them wasn't a painting but an engraving after a painting. So I linked it as an "engraving" with creator=engraver. While doing this, I realized that in English the word for painting (and the enwiki article) is used for the art of painting and the physical object. Same for engraving, so my qualifier "instance of engraving" is actually an "instance of the art of engraving". Don't we need "medium" and then "art of creating medium" for each art form? Jane023 (talk) 12:14, 5 July 2014 (UTC)

There are distinct items between arts and their respective media, such as art of painting (Q11629) (View with Reasonator) vs. painting (Q3305213) (View with Reasonator) , and both get listed if you search for "painting"; I believe many others art similarly have distinct items. Items that are "art subclasses" should not be pointed to by instance of (P31) of any other items, thus all these items need to be fixed, for example. LaddΩ chat ;) 14:29, 5 July 2014 (UTC)
Thanks, I was afraid of that. I will try to rename the titles of all the items that I see (so tapestry-making vs tapestry, art of engraving vs engraving, art of sculpting vs sculpture, etc). When I am in the middle of an "instance of" I can't see what I am selecting otherwise. Jane023 (talk) 15:27, 5 July 2014 (UTC)
We have field of work (P101) miga which links to the 'art of' item. This is mostly used with occupation (P106) miga so a person can have 'occupation:artist'; 'field of work:engraving' and we don't need to create a separate item for each specialised field. That works for Occupation but from what you say it looks like we will need to create separate items for each type of image so we can classify each image on Commons once Commonsdata is launched. When Commonsdata happens then the qualifiers you are creating can be moved there as properties of the image. Filceolaire (talk) 20:24, 5 July 2014 (UTC)

This is actually a bigger job than I expected. It turns out several other languages have combined their articles for "medium" and "art of creating medium", such as "stipple engraving", "steel engraving", and so forth. This means that the Wikidata items need to be separated out. When in doubt I will create new items for the object vs. the creative process (such as en:Painting (object). Jane023 (talk) 23:25, 5 July 2014 (UTC)

Maybe we should have a property for 'medium' with properties such as 'oil painting', 'water colour', 'computer graphics', 'metalwork, 'whittling', etc. Filceolaire (talk) 21:24, 6 July 2014 (UTC)
No that won't solve the problem - at some point you need to split between the noun and the verb. If I am holding an engraving, then I don't want to say that I am holding an instance of the art of engraving. Perhaps we need another thing like "result", as in, I am holding an instance of the result of the art of engraving? Jane023 (talk) 21:32, 6 July 2014 (UTC)
This is a problem common to all kind of processes, also algorithms, computer programs and so on. I don't think we already got the properties, but I fell we need properties for input and output or product, or result of a process. TomT0m (talk) 05:14, 7 July 2014 (UTC)

Instances of something else than entity[edit]

Ca 1.8 million items, details: Talk:Q35120#Instances of something else than entity. Tamawashi (talk) 12:59, 5 July 2014 (UTC)

See my comments on that talk page re the wikidata High Level Ontology. TLDR = yes we have not yet completed the process of linking all the 'subclass of:class item' statements into one hierarchy connecting all class items to a single root item but your statistics show we have already some made progress towards this. Filceolaire (talk) 20:01, 5 July 2014 (UTC)

Widar warning[edit]

This may cause creation of many false claims via Widar. Tamawashi (talk) 13:47, 6 July 2014 (UTC)

Here is my selfie[edit]

Wikidatans at work.jpg

Sometimes things do not go without any hitches on Wikidata :) TomT0m (talk) 14:17, 6 July 2014 (UTC)

Duplicate information with occupation and subclasses[edit]

Hello there,

I've noticed that over the last few days/weeks some of the items on my watchlist have gotten new claims with property occupation and value musician although they already have a claim that they're a composer which is a subclass of musician. Examples of this include Michael Giacchino and Jerry Goldsmith. I do think this is just duplicate information and the musician claim should be removed. I did try to find earlier discussions about this but failed and I'm a bit reluctant to remove this information without knowing it's the right thing to do. So should I go ahead and remove the musician claim again or is this valid information? --Mineo (talk) 17:08, 6 July 2014 (UTC)

@Mineo: This is not a problem. A class regroup individuals who have same characteristics. Musicians are people who plays music. Perhaps even musicians might not have music as a main occupation. In languages such as OWL2, classes can be defined with a so called Class expression, for example the class Musician can be defined as Human or animal who has music as an occupation or hobby. In those languages, and OWL is an inspiration and a precedent for us all to build Wikidata. Classes are also useful for defining guidelines for properties. For example if we have a property plays instrument, we can define the constraint that usually item who will have this property are instances of the class <musician>.
Of course then there might be an abuse and we could define a lot of classes. As Emw likes to state, in OWL classes might not be explicitely stated, and can be inferred by a program who understans class expressions. Wikibase itself can not atm and it's not in the roadmap atm.
In the case of Musician though, I don't think it's a problem to use this class as it's obviously a class we use all the times. When we say Musicians make life happier for example. We refer implicitely to people who plays music.
We should add words about this in Help:Classification. TomT0m (talk) 19:38, 6 July 2014 (UTC)
  • Mineo, good point, I agree that claims like "Michael Giacchino occupation musician" can be removed given that the claims "Michael Giacchino occupation composer" and "composer subclass of musician" exist. TomT0m's explanation of class expressions in OWL is interesting but seems unrelated to your particular question. Emw (talk) 19:57, 6 July 2014 (UTC)
    Oh, you're right, I've been confused. Actually maybe this highlight something : there is a possible confusion beetween occupations as a class of occupations and classes of all people who have this occupation. As they are both classes, there is no way we use the same item here. TomT0m (talk) 20:38, 6 July 2014 (UTC)
  • @Emw: Thanks, I've reverted the edits on the items in my watchlist, but those were added with Widar, so there might be more of them. --Mineo (talk) 21:21, 6 July 2014 (UTC)
Arguably, a musician is not restricted in the same manner as a more defined "occupation". Was Frederick the great a musician or both a flautist and a composer?? Or is it best to indicate that he was a musician and dabbled in composition and playing the flute?? Given that we use occupation it is clearly wrong in either case.
An other argument to use musician is that many of our sources do not differentiate between the many roles. As at this stage the lack of information is out biggest problem, it is to be preferred to harvest such information now and once we have decided what is the best approach we can easily modify in any which way. We need data first. That should be our game plan for now. GerardM (talk) 07:42, 7 July 2014 (UTC)
  • You are right, using the occupation property for Frederick would be wrong, but that's an entirely different problem :-) I'm also not sure how the vagueness of sources relates to my question because I was specifically asking about cases where the sources are *not* vague (= we know what kind of musician the person is).--Mineo (talk) 08:49, 8 July 2014 (UTC)
Many Wikipedia do not make that distinction and consequently you will have both "occupation"s. GerardM (talk) 10:52, 9 July 2014 (UTC)

@Emw: any tool or query that could detect instances that have a "instance of A" and "instance of B" where "A" is a subclass of "B"? Tamawashi (talk) 11:26, 7 July 2014 (UTC)

P131, P172 for languages[edit]

Hi. Is it allowed to use these properties for language items? Also, may coordinates be added? Regards, πr2 (t • c) 20:23, 6 July 2014 (UTC)

In many cases languages are defined related to countries and administrative territorial entities for political reasons (Why are Galician (Q9307) and Portuguese (Q5146) considered to be different languages?) so I guess is in the administrative territorial entity (P131) could be used for languages.
'Ethnic group' very often refers to people with a common language. It may the principal thing distinguishing that group from others so I suppose ethnic group (P172) miga is appropriate for some languages.
range map image (P181) is meant for animals but I think it could be used for languages as well. Filceolaire (talk) 21:15, 6 July 2014 (UTC)
Is P181 the same as P242? πr2 (t • c) 23:44, 6 July 2014 (UTC)
What you indicate about Galician and Portuguese (as being the same) indicates that it is best that you do not touch languages at all. The notion that you can identify languages as being spoken in a specific "administrative unit" gives often the notion that it is spoken exclusively which is often plain wrong. When you indicate that only a specific ethnic group speaks a language it is typically dead wrong as well. Please do not do this. GerardM (talk) 07:35, 7 July 2014 (UTC)
@GerardM: What are you talking about? As far as I know Galician and Portuguese are distinct, but closely related, and have been considered either separate languages or dialects (classification/relation). Are you saying we should not list the regions languages/dialects/language families are mainly spoken in for political regions? Obviously languages are not always spoken by a single ethnic group, but IMO there should be some way to indicate a link between e.g. Pitjantjatjara (Q3249806) and Pitjantjatjara language (Q2982063). I don't know which property to use and whether it would be better to indicate this on the language item, the ethnic group item, or both. πr2 (t • c) 12:31, 7 July 2014 (UTC)
Making languages that are recognised as languages no more than dialects is highly political and ultimately pointless. You only get yourself into a territory where you / we should not want to go. WMF accepts ISO-639-3 for this reason; we are still suffering from the consequences of well intentioned but ultimately destructive opinions. Thanks, GerardM (talk) 12:43, 7 July 2014 (UTC)
I think you are confusing what I wrote with what Filceolaire wrote. I'm just wondering how to associate languages/dialects/etc. with locations. I don't care whether they are considered languages or dialects. You're the one who brought politics into this. πr2 (t • c) 13:01, 7 July 2014 (UTC)
For example, it would be useful to somehow connect Patagonian Welsh (Q3798706) to Chubut Province (Q45007) or Y Wladfa (Q478236). This would be useful for languages and dialects, no? Ethnologue already lists locations where languages are mostly spoken, as do enwiki infoboxes. Could you please explain why you think adding this information would be destructive? πr2 (t • c) 13:17, 7 July 2014 (UTC)
DYK that some languages are spoken more in New York than in the area they originated ? GerardM (talk) 10:46, 9 July 2014 (UTC)
I don't see the problem. If there are significant communities of speakers in New York (or wherever), it can be indicated with the property. Perhaps which areas are the place(s) of origin of the language could be indicated using a qualifier. Already sources like MultiTree contain this kind of information, for example indicating that Garifuna is spoken in the United States. Ethnologue has information for each country a language is spoken in, example ("Also Spoken In"). Why would this be problematic? I think collecting this information would be useful. It could also help to make maps like this, if we allow coordinates for languages spoken only in one region (for example, many dialects are only used in one village/town). πr2 (t • c) 19:14, 10 July 2014 (UTC)

Another API question[edit]

How do I get the ID of the wikidata item if I only know the name of the corresponding article in the English Wikipedia using the Wikidata API? --Jobu0101 (talk) 15:20, 7 July 2014 (UTC)

You can't. Snipre (talk) 15:30, 7 July 2014 (UTC)
But you can get it from the article of WP:en. Snipre (talk) 15:32, 7 July 2014 (UTC)
Using the API there? How is that possbile? --Jobu0101 (talk) 15:36, 7 July 2014 (UTC)
This request somehow gives a strange result. Always $1 instead of the respective title. For example "url": "$1". Is that a bug? --Jobu0101 (talk) 15:42, 7 July 2014 (UTC) ; -- Vlsergey (talk) 15:47, 7 July 2014 (UTC)
Indeed, there is away and you have found it above Vlsergey! The only other way to do this is using the api of the site that you know the name of the article on, within the page info you will be given the item id. ·addshore· talk to me! 15:51, 7 July 2014 (UTC)
Or you just use a bot framework that solves this all for you. Multichill (talk) 16:11, 7 July 2014 (UTC)

Thanks to all of you. --Jobu0101 (talk) 17:35, 7 July 2014 (UTC)

Sorry I didn't think to use the sitelinks. But definitively the labels can't be used. Snipre (talk) 08:20, 8 July 2014 (UTC)

Deletion warnings, deletion handling via user-subpages, professionalism[edit]

User:Delusion23 informed me that the following three items appear on a user-subpage of User:Pasleim and wrote "Please use them in items or provide sources or they are likely to be deleted." [2]:

  • canton of Gran Colombia (Q17305515) - a class for the cantons of Gran Colombia - if no canton has an item, how could I use that on other items?
  • Chuda State (Q17318475) - a princely state of the British Raj - I have no idea on which item I could use that, it's just one of hundreds of princely states.
  • island group of Kiribati (Q17305516) - first level administrative territorial entities as defined by FIPS and ISO - if the island groups have no item yet, I don't know where I could use that item.

Alternatively for all three - How do I add a source?

Instead of tracking via a user-subpage could that please happen in the Wikidata-namespace? Deletions are a serious point of controversy that did let many people leave individual Wikipedias. So there should be a high level of professionalism when handling these. Wikidata:Notability looks good. Maybe a subpage Wikidata:Notability/List of likely unnotable items could be created? And then a bot automatically informs users if an item that they had created appears on that list? Or, even easier, a link to the creator is added in the list and then the MediaWiki-software informs the user automatically. Tamawashi (talk) 07:58, 8 July 2014 (UTC)

I agree that it would be good if a system like that described by Tamawashi were put in place. Automated messages to talk pages when a page is still showing signs of not being notable after, say, two days would be good. Then if the items are not improved to pass WD:N after a couple more days the items could be deleted. Unless of course the item is an obvious test page or vandalism in which case it can be deleted straight away. It would also be good if that particular page hosted by Pasleim could be at a more official page. This way more people would be aware of its existance, both item creators and admins.
The issue with sourcing is an ongoing Wikidata development problem. I'm still amazed the project has had to go on for almost 2 years without any simple method for adding refs and sources being initiated. Delsion23 (talk) 18:37, 8 July 2014 (UTC)
Sorry but we have a system for sourcing since 1 year: Help:Sources. After if you want to have an automatic system able to manage books, scientific articles, websites and other medias, I can just encourage you to code one or to find someone to do it. Snipre (talk) 21:27, 8 July 2014 (UTC)
I can move User:Pasleim/notability to Wikidata namespace. But we have first to agree, which items should show up on the list. At the moment, all items with
  • no sitelinks
  • no backlinks
  • older than 20 hours
  • less than 3 statements
are listed. Especially the last two points have to be discussed. --Pasleim (talk) 02:47, 9 July 2014 (UTC)
@Pasleim: I suggest to use your setup. Any further changes can be done later. Tamawashi (talk) 08:56, 9 July 2014 (UTC)
BTW, canton of Gran Colombia (Q17305515) could be de-orphaned, there is at least one item on Wikipedia: Zulia Canton (Q8075149). For the administrative types it is not that difficult to add them without get warnings - as the actual units are notable as well, one can easily create items for them, and then link these with is in the administrative territorial entity (P131) and contains administrative territorial entity (P150). Ahoerstemeier (talk) 05:57, 9 July 2014 (UTC)
Great, thank you. Tamawashi (talk) 08:56, 9 July 2014 (UTC)

@Pasleim: It looks as if most reports are under "Wikidata:Database reports", so I suggest to place it there. Maybe Wikidata:Database reports/Notability. Tamawashi (talk) 16:32, 9 July 2014 (UTC)

Catalog property[edit]

I have used catalog code (P528) (with qualifier catalog (P972), as stated on the property page) to add external catalog identifiers to Wikidata, where the catalog has no property of its own, e.g. to add Christoph von Scheurl (Q70425) to the Dictionary of Art Historians (Q17166797). Only later I discovered that "catalog code" apparently means "astronomic catalog code".

  • Should there be such a generic catalog property, with a qualifier to specify the catalog, for catalogs that do not have their own property? (I think: yes)
  • Should we use catalog code (P528) for this? (I think:yes)

As a side note, you can already get all items in a specific catalog annotated in this format, e.g. here for Dictionary of Art Historians (Q17166797). --Magnus Manske (talk) 10:56, 8 July 2014 (UTC)

I agree, this is good flexibility. Just a detail though : I think it would be better to do it the other way around : just a catalog number by itself is not meaningful, it's not useful for example in the RDF export with just the main snak of statements as triples. It would make imho more sense to have a main snak meaningful by itself, this would give statements like
<Mars> in catalog search <astronomical catalog>
        identifier search <XXZSF...>
. But it's just a detail and make the migration a little bit less straightforward. TomT0m (talk) 11:09, 8 July 2014 (UTC)
@Magnus Manske: where do you see that catalog code (P528) is just for astronomical objects? That property was extended long ago to cover all catalogs, if there is any trace left from the former restriction it should be removed.--Micru (talk) 12:26, 8 July 2014 (UTC)
@Micru: I noticed the alias "astronomical catalog", then on the talk page "the catalogue name of an astronomical object". If it's already all catalogs, then my question is solved :-) --Magnus Manske (talk) 14:12, 8 July 2014 (UTC)
@Magnus Manske: I don't see the need of a generic property. Creating a specific property "Dictionary of Art Historians ID" is easy and allow MediaWiki:Gadget-AuthorityControl.js to provide a direct link.
Another problem is that the word "catalog" doesn't seem appropriate to qualify a "dictionary" (at least in French but maybe also in English). — Ayack (talk) 13:43, 8 July 2014 (UTC)
@Ayack: The gadget could be altered; shouldn't be too hard. Creating properties for all possible catalogs seems ... inflationary. Oh, and "catalog" seems appropriate to me in this context; "authority control ID" might be more precise, but sounds somewhat bureaucratic. --Magnus Manske (talk) 14:15, 8 July 2014 (UTC)
@Magnus Manske: Both methods are OK for me, but only if we use one, including Freebase, VIAF, etc. ... — Ayack (talk) 14:53, 8 July 2014 (UTC)
@Magnus Manske: I corrected the description of catalog code, but TBH I think the property you need here is described by source (P1343), it fits perfectly with a dictionary, or an encyclopedia describing the item. If you want a direct link, then perhaps it is better to have a property so we can have some constraints in place.--Micru (talk) 15:05, 8 July 2014 (UTC)
@Micru: OK, but how do I set the actual identifier in a source for described by source (P1343)? --Magnus Manske (talk) 15:35, 8 July 2014 (UTC)
@Magnus Manske: That's a very good question. I have seen that Vlsergey is using title (P357), but that is a poor solution because it disturbs the constraints of the property. I have proposed a generic label property for such cases.--Micru (talk) 17:17, 8 July 2014 (UTC)
I don't see why p1343 should be used here. catalog code (P528)/catalog (P972) seems to fit well and p1343 is not as precise: it just states that the value talks about the item, not that it is its main topic. --Zolo (talk) 17:20, 8 July 2014 (UTC)
@Zolo: a catalog is just a list of elements with an identifier, it doesn't need to say anything about the item. P1343 describes the item in general terms (dictionaries, and encyclopedias).--Micru (talk) 17:48, 8 July 2014 (UTC)
@Micru:. I don't follow you. described by source (P1343): X. Just means that the item is mentionned in X. "catalogue code : X catalogue: Y" means: this item corresponds to entry X in reference Y. Sometimes the entry will be long sometimes very short, but it does not make much difference, and often an approximate length of the text can be guessed from the nature of Y. That said I would agree that the word catalogue may sound too restrictive for that and I would be fine with changing the labels to something like "identifier" and "scope of the ID" (though the second one is slightly freakish). --Zolo (talk) 17:57, 8 July 2014 (UTC)
@Zolo: the description of p1343 says: "dictionary or encyclopaedia where this concept is described". Magnus wants to add a dictionary entry to each item, so that fits with the description of the property. Of course it is needed another property to narrow the scope. P972 as it is now refers exclusively to catalogs, and not to anything else. I am not sure it is a good idea to expand it further for all identifiers, since we can create more properties as suited.--Micru (talk) 18:12, 8 July 2014 (UTC)
@Micru:. I had misread p1343... But I do not think it is a good idea to have a separate property just because one reference is a catalogue and another a dictionary. In some contexts, it may help with displaying things nicely, but on the whole that seems unwieldy to have to manage several properties, that have the same structural meaning (item is entry X in reference Y). --Zolo (talk) 18:22, 8 July 2014 (UTC)
@Zolo: I'm also ok merging both, since it is more or less the same. I also think that changing "catalog code" to a more generic "identifier" is also positive (in that case there won't be needed another "label" property).--Micru (talk) 11:49, 9 July 2014 (UTC)

Hmm seeing this discussion I think my proposition above is not that really more complicated to implement. And it make more sense : if we know an item is catalogued somewhere but don't know the id we can just create a statement and a constraint report can list all the catalogued item missing an id. An id without knowing the catalog is pretty useless. If we are gonna make this change now is the time. I had no answer, which mean no concrete opposition ... TomT0m (talk) 11:28, 9 July 2014 (UTC)

So do you want to use "in catalog" (p:p972) as the main property and "identifier" (p:p528) as the qualifier? Yep, it makes sense to invert it and to put constraints in place.--Micru (talk) 11:49, 9 July 2014 (UTC)
I'm opposed to this. Infoboxes cannot easily access qualifiers so it make more sense to give them the specific info related to the item - especially as it is possible to tell the catalog from the context since infoboxes are used within a specific domain. E.g. station code for railway stations.Looking at a statement in isolation you need the catalog it comes from. In context however the code is more useful since you can usually tell very easily what network the station is part of. Filceolaire (talk) 07:40, 10 July 2014 (UTC)
@Filceolaire: Any infobox should use Lua method to include property value -- to correctly handle deprecated and preferred values, formats and other issues. Therefore, from infobox point of view it doesn't matter if it's property value or qualifier value, it will be just a small addition to method arguments. -- Vlsergey (talk) 08:04, 10 July 2014 (UTC)
@Filceolaire: It's not a problem in the software engeneering point of view, it will push Lua and infoboxes coders to solve the problem and make easy to access qualifiers. It would be a really bad idea to base a decision on a non circonstancial issue like this one. TomT0m (talk) 08:22, 10 July 2014 (UTC)
@Filceolaire:. It is pretty easy to access qualifiers from Lua (see fr:Module:Wikidata). And anyway, it will often be necessary to access the qualifier, otherwise to make sure that we are talking about the right catalogue
@TomT0m, Micru: I do not understand why it is better to have make the catalogue number a qualifier. A catalogue name without a catalogue number is about as uselesss as a catalogue numbe without a catalogue name. In specialized publications, catalogue numbers are sometimes given without the catalogue name, because it appears to be obious from the context or from the identifier format. I think we should always try to make the catalogue name explicit, but there are certainly cases where we can import the identifier now, and add the catalogue name later. At least that seems more plausible to me than knowing that an object is in a catalogue/database without knowing it ID there. --Zolo (talk) 08:47, 10 July 2014 (UTC)

Accessing statements from Wikipedia[edit]

Hi, I am trying to figure out how to access Wikidata statements from the Wikipedias. I have read mw:Extension:Wikibase_Client/Lua. It describes how to get a list of properties and the property values. But nothing about qualifiers, sources or rank are mentioned. Does that mean that it is not possible to access qualifiers, sources and rank of statements from Wikipedia pages? If so, is it planned for the future? Thank you for any advice, 11:02, 8 July 2014 (UTC)

All the item content is returned by wikibase.getEntityObject(). The returned content is a Lua table that may be manipulated directly. So it's possible to access qualifiers, sources and ranks right now. Tpt (talk) 12:06, 8 July 2014 (UTC)
Look at Module:Wikidata (Q12069631) on your WP. If it doesn't exists copy/paste the code from another wikipedia. Be careful there is no an unique version of that code, each WP uses a different one with different options. Snipre (talk) 12:41, 8 July 2014 (UTC)
Thank you, Tpt and Snipre. I will look into that. Is the format of the table returned by mw.wikibase.getEntityObject documented anywhere? And, not least, is the current format stable, or could the format be changed, so any pages which relies on it could be broken? 15:36, 8 July 2014 (UTC)
Not sure about the API, the Wikibase client extension is labeled as "beta version". I think they would make sure not to break existing Lua code at this stage (after all some Wikipedia already use Wikidata), but not really sure about this. Wikibase client documentation is not really good. For fetching qualifiers and other properties, I would take a look at the German Wikidata module at it has more functionality than the one from the English Wikipedia. For getting the sources, take a look at the "getReferences" function. --Bthfan (talk) 08:05, 9 July 2014 (UTC)
Thank you, Bthfan. I asked about the stableness of the table format because when doing computer programming it is often considered a bad thing to directly use internal structures in objects. Generally it is better only to use the documented interface methods in order to not restrict the possibilities for further development of the structure. However the problem here is that there are only documented interface methods to get labels, links, properties and property values. So I will have to make my own interface methods (or reuse from e.g. German Wikipedia). I think the Wikidata developers may have a future problem by not delivering their own access methods so direct access to table can be avoided. 10:48, 9 July 2014 (UTC)
It is, I don't know why we don't have a common library do do that yet to be a little more robust and make easier to code. We should maker this a guideline for lua coders who are for a lot of them not really software engeeniers or professional coders. I proposed to the Pywikibot dev team to align a Wikidata API to the one in python of pywikibot for example, but had unfortunately no answers. @Markus Kroëtch: is also supervising the coding a java API to work with Wikidata datas, maybe it would be cool to share experiences of all those coding teams. And of Lua and infoboxes coders on Wikipedia, but all of this world is a little bit too fragmented in my opinion. I guess the best answer is to code and communicate around that code, write Help pages in several languages. TomT0m (talk) 11:36, 9 July 2014 (UTC)

Linking to restricted access databases ?[edit]

Should we accept identifier-properties for linking to websites that are not fully accessible for free ? I was thinking of Artprice or Deezer, that have some great resources, but mostly behind a paywall. --Zolo (talk) 13:04, 8 July 2014 (UTC)

Is Wikidata aiming to become ScienceDirect or PubMed?[edit]

This is a question of Wikidata notability and how we plan on referencing and sourcing information here. Will there be any future developments that allow for the easy linking to science papers, books, websites etc. Or do we start making items for every single reference in the world as I've seen here and then link to these items? We have to reference somehow, but what is the best way to do it? Cheers. Delsion23 (talk) 18:46, 8 July 2014 (UTC)

reference URL (P854) makes far more sense. --Jakob (talk) 19:28, 8 July 2014 (UTC)
True, it would make more sense if the items listed here instead used that property and linked directly to the ducument here. Delsion23 (talk) 19:44, 8 July 2014 (UTC)
Creating items for individual articles, especially those cited in several items, is recommended in Help:Sources (written after a long discussion).
using just an url is not enough, for about the same reasons that we do not just provide an URL when citing a source in Wikipedia.
Of course gadget or any other thing that would make the whole process smoother would be more than welcome. --Zolo (talk) 19:52, 8 July 2014 (UTC)
@Delusion23: How can you source the imported data in WP with a link ? What happens when links are broken ? Weblinks are the worse idea in referencing because nobody can ensure the validity of the link in the future. Wikidata is not ScienceDirect or PubMed but just a normal database which provides the values and the references because values and references are two parts of an unique entity. Snipre (talk) 20:50, 8 July 2014 (UTC)
The sourcing system needs to be overhauled so that you can add multiple parameters to a source (url, author, title, etc.) That seems like a better solution than creating an item every time you want to cite something. --Jakob (talk)
I agree with Jakob. Creating an item every time you want to reference something is not the best way for Wikidata to work. It would be better if there was an inbuilt functionality for us to add sources and information about the sources like page numbers, journal names etc. Delsion23 (talk) 21:13, 8 July 2014 (UTC)
You can already do that. We need to make sourcing significantly easier and that is on my plan for 2015. --Lydia Pintscher (WMDE) (talk) 21:59, 8 July 2014 (UTC)
I see that now. @Snipre, Zolo: Why not just do that then? In addition to a url, you can add an author, publisher, date, etc. into a single reference. --Jakob (talk) 22:09, 8 July 2014 (UTC)
@Lydia Pintscher (WMDE): This assumes url stability. Unfortunately there is a lot of dead urls on Wikipedia, and it's not so easy to change each occurences of them if the document is cited several times, sometimes even with a different URL. Metadata below just url are useful to be more robust on URL change, and creating an item makes easier to find if a document has already been cited. It seems a more practicable in the real world than having unfortunately url ids who have poor robustness guaranties in a lot of cases. Creating an item is not memory costfull compared to this and can be made really easy with a good UI. It seem more pragmatic to me. TomT0m (talk) 11:44, 9 July 2014 (UTC)
@Jakec: The current system is appropriate if you consider Wikisource and Wikiquote: they need items for most of the books and for their different editions. So if you want to do the job twice. The current system is the most appropriate if you consider that some references can be used hundred times: why do we need to enter even automatically hundred times the same references data ? To increase the memory of servers ? The solution you proposed was what I wanted to see in wikidata before we decided to use the current system. I was proposing that solution and after long discussions I was convinced by the current system because it is the most appropriate for multiple use of the same reference. Just open a reference book or article: the normal system is to define the reference data in one place and to use an ID in the rest of the document. Snipre (talk) 22:47, 8 July 2014 (UTC)
@Delusion23: Sorry to ask you that but did you already try to add reference data ? What you request is already available: you enter all the parameters of a reference in the section Reference present in each statement (see an example here). If you are doing that several times you will see why creating one item if simpler. And you can use the API to do that automatically. Snipre (talk) 22:47, 8 July 2014 (UTC)

I agree with Zolo and Snipre on this. A source should ideally have its own Wikidata item. This fulfills a structural need and thus meets our notability threshold per WD:N criterion 3. For the impatient, referencing via reference URL (P854) is better than nothing.

Jakob says "The sourcing system needs to be overhauled so that you can add multiple parameters to a source (url, author, title, etc.) That seems like a better solution than creating an item every time you want to cite something." This need is already met by items: those "parameters" already exist properties as like reference URL (P854), author (P50), title (P357), and an ecosystem of other properties described in Help:Sources, as Zolo says. I agree with Jakob that it's inconvenient to do that, especially for one-time references, but I don't think it's significantly more work than finding and entering data on 'URL', 'title', 'author' etc. and entering them as qualifiers on a source. However, having that information centralized in an makes it easier to query in interesting ways, and reduces duplicate data entry (e.g. you don't need to list the author each time a paper is referenced).

If Lydia does allocate developer resources to this, I hope it will be for features like automatic extraction and population of author (P50), title (P357), date of publication (P577), etc. given a URL or URN -- kind of like the game-changing Diberri template filler. Emphasizing precision in references, by providing input fields for properties like quote (P387) and page (P304) in such a tool, would also be great. Emw (talk) 01:31, 9 July 2014 (UTC)

Historical note: minerals are natural substances with a chemical formula range. Nowadays, the definition of each mineral is done under the patronage of the International Mineralogical Association (IMA). The last accepted modification is summarized in the last IMA Master List, and on scientific papers about each case. The definition changes with improvements in technology and science. Mineral definitions are conventions, and changes in these conventions are small revolutions (Wikidata:WikiProject Mineralogy/IMA number references#List of IMA numbers of other special procedures, for instance). The professors and curators that decide these changes are country representatives at the IMA and are notable. These people published a pile of scientific publications and Wikimedia projects don't have enough scientific editors. Many professors and curators don't have an article on a Wikimedia project, and it doesn't prove that these people aren't notable enough. Their names will still be in a scientific encyclopedia in 300 hundred years time, many professional sportspersons/ gladiators will lose their notability in 300 years time. In my opinion, CC0-compatibility rule doesn't apply for conventions (agreed upon vocabulary and mineral definitions). No communication is possible without an agreed upon convention. We don't need all scientific references as items but we need the most important ones as items. We need to be able to express changes on Wikidata as 'start date' and 'end date'. --Chris.urs-o (talk) 05:40, 9 July 2014 (UTC)
That would be one thing we can do yeah and it makes a lot of sense to me. --Lydia Pintscher (WMDE) (talk) 06:15, 9 July 2014 (UTC)
@Lydia Pintscher (WMDE): there is an OPW project to extract sources metadata for VE, I think it should be reused here too. Also Thepwnco proposed to have a special page (or a gadget) that selecting which kind of source the user wants to create it would display some property fields or others. I think both ideas combined with a "create and link new source" option in the references section would help greatly to reduce the effort when adding a source. It was also proposed that a list with used sources in the item with the option to drag&drop them, or copy&paste whole source, would also improve the efficiency. --Micru (talk) 07:55, 9 July 2014 (UTC)

Label/description problem in "In other languages" section[edit]

Since earlier today, when typing a label or description in the "In other languages" section, there is no longer a "save | cancel" box as there is for your first language. What is causing this problem? Jared Preston (talk) 21:44, 8 July 2014 (UTC)

I can reproduce it. I filed bugzilla:67696. --Lydia Pintscher (WMDE) (talk) 22:04, 8 July 2014 (UTC)
Danke Lydia. Jared Preston (talk) 22:06, 8 July 2014 (UTC)
It was fixed last night. Great! Jared Preston (talk) 10:38, 10 July 2014 (UTC)

Proposing item merge and merging[edit]


I've nominated/proposed two Wikipedia articles which describes the same person for merging. I know Wikidata Game can be used for merging. But how do I nominate items for merging? Also merging without the game. Items involved Q17323735 and Q17320640. --Enock4seth (talk) 12:05, 8 July 2014 (UTC)

I would enable the "Merge" gadget at in your settings. Then you get a "Merge with..." link/button at the top of the Wikidata website under the "More..." menu. With that one you can easily merge two items. The problem in your case though is that those items have interwiki links that conflict with each other. Both have a link to the English Wikipedia, a Wikidata item can only store one link per Wikipedia (language). So this one also needs to be resolved at the English Wikipedia, those two articles probably should be merged into one single article. BTW: Did you manually insert the time stamp in your signature? Somehow looks wrong ;) --Bthfan (talk) 07:46, 9 July 2014 (UTC)
Bthfan, the problem is the merge on Wikipedia, Wikidata behaves fine.

I was wondering the same thing, and since I do tend to merge items now and then on English Wikipedia after noticing this with two Italian articels I went ahead and merged them (gasp!) the Italian Wikipedia. So far I received no backlash, but the guy was very dead and the info looked 100% covered. This one is going to take some time, because it's a recent death and they are both new articles. The same rules hold as for here: oldest article trumps newer one, so I would say "Castro de Destroyer" is where the item should point to, but it's best to just wait it out a week or two - I see a merge template is already in place. Jane023 (talk) 07:54, 9 July 2014 (UTC)
@Ebraminio: ideally, I think that when the two items to be merged have sitelinks to the same language, MediaWiki:Gadget-Merge.js should display a message asking if the two articles are really about the same topic, and if we click yes, it should send a message or add a Template:Merge (Q6919004) to the articles. --Zolo (talk) 09:54, 9 July 2014 (UTC)
What a really great and creative idea but it would be subject of another gadget IMO. I wrote merge.js when there were no merge API and anyone was doing merge (moving each of claims and sitelinks) manually. IMO it should just do things that would be easy review and revert-able by other users and a cross-wiki request Template:Merge adding is outside of its goal however I will not stop anyone else to implement such ability on merge.js itself. Just IMHO requesting content merge is a prestep of an actual item merge, you see a lot of specially lengthy articles on enwiki that is not merged after years of requesting merge and this essentially could be related on activity of that wiki community so it is not an action that after a request and just a short time would be possible. This can be subject of a bot that automatically request merge for Wikipedia communities and check if community done the content merge then do the Wikidata items merge –ebraminiotalk 15:07, 9 July 2014 (UTC)

I have encountered duplicate articles in the same language in the past. If I feel like trying, I look for the templates in that language, to propose mergers:

I assume it's the same in most languages as on English Wikipedia:

  • On the article with the better name: {{mergefrom|Some worse name}}
  • On the article with the worse name: {{mergeto|Some better name}}

or if you don't know which is better:

  • {{merge|The other name}} on each article

The help page for each project is Wikipedia:Merging (Q11773605). --Closeapple (talk) 21:08, 10 July 2014 (UTC)

Wikidata:Subclasses and instances[edit]

As a result of Talk:Q35120#Instances of something else than entity, I created Wikidata:Subclasses and instances. Tamawashi (talk) 09:06, 9 July 2014 (UTC)

@Tamawashi: I created Help:Classification which seem to have the same purposes. I renamed the page to reflect it's a statistical page Wikidata:Statistics:Subclasses and instances TomT0m (talk) 11:55, 9 July 2014 (UTC)
@TomT0m: Your edits gave me some inspiration. I expanded the page but moved it back. I created a link to a future "Wikidata:Database reports/..." page. I think that is the place where most of the reports are. Tamawashi (talk) 16:22, 9 July 2014 (UTC)

Vandalism at Wikidata:Subclasses and instances[edit]

I asked User:Succu to provide evidence for his previous claim so that everyone using Wikidata:Subclasses and instances could see it. Instead he now is vandalizing one the page, e.g. removing a sourced statement about the output of a WikiData-Query. Tamawashi (talk) 17:46, 9 July 2014 (UTC)

Tamawashi refused to discuss. See his/her discussion page. Vandalism is a realy strong accusation! --Succu (talk) 17:50, 9 July 2014 (UTC)
User:Succu - Yes, I reverted some of your contributions to my talk page. The talk page shows that I thanked you for your preceding preceding edit and that I asked you to back up your claims at the relevant page. That could benefit more users than only me. See a similar initiative of mine further above, related to notability reports [3]. You still have not backed up your claims. Tamawashi (talk) 11:06, 10 July 2014 (UTC)

@Succu, Tamawashi: Hi guys, I don't want to enter this troll, but I'll just say I'm preparing a variation of {{Superclass tree}} who will show something like that, I'll take a taxon example, based on the instance of (P31) informations. It will generate, using just instance of (P31) and subclass of (P279) claims, tables like this for an item :

class type of class
... ...
animal reign
... ...
xerus genera
xerus xinauris species

The advantage of this is that this will demonstrate how powerful we can be with keeping genericity : this will not only work for taxas themselves, but also for any class of items for which the superclasses have instance of (P31) statements. The table will be ordered by class inclusion, naturally.

I'll post updates as soon as I have something to show on TomT0m (talk) 08:13, 10 July 2014 (UTC)

@TomT0m: A documentation for superclasses that have a claim "instance of" would be great! Please link from Wikidata:Subclasses and instances when ready. Tamawashi (talk) 11:13, 10 July 2014 (UTC)

Vandalism 2 at Wikidata:Subclasses and instances[edit]

User:Succu continues [4]. Tamawashi (talk) 21:10, 10 July 2014 (UTC)

(edit conflict) I moved Wikidata talk:Subclasses and instances to the users userspace, because I see no justification being in the main space. I think, it reflects a only his/her personal point of view. --Succu (talk) 21:14, 10 July 2014 (UTC)

I have to agree with Succu. This page should be either in userspace or part of a WikiProject. The same applies for Wikidata:Administrative_territorial_entity. --Pasleim (talk) 21:59, 10 July 2014 (UTC)
Wikidata:Subclasses and instances is not a page stating containing personal point of views. If a page should be moved to the user space than a better candidate is Wikidata:Project chat which you both contribute too write "I think" and "I have to agree" statements, without any reference to a policy. Is this an "I like" game here, or do people work based on rules and reasoning? Tamawashi (talk) 22:42, 10 July 2014 (UTC)

Commons category links in items[edit]

Is this the right way to link Wikipedia articles and Commons categories? I thought Commons categories are only linked to other Wikimedia categories? (See Wikidata:Notability part 1, bullet point 4) 09:18, 9 July 2014 (UTC)

Also see on this, it looks like for article items Commons category (P373) should be used instead of a sitelink when linking to common categories. So yes, I think those edits are wrong. --Bthfan (talk) 10:26, 9 July 2014 (UTC)
Though when I see the discussion at Wikidata:Requests_for_comment/Commons_links I'm no longer sure about right/wrong...I'm confused now :S --Bthfan (talk) 10:35, 9 July 2014 (UTC)
@Bthfan: There are basically two currents of thought: "mergists" and "splittists". The mergists would have to have together the sitelinks to categories and sitelinks to articles in one item, whereas the splittists prefer not to mix them, since they represent different concepts. Additionally, there is the problem that Commons has organized mostly around Categories and not around content pages (understandably so, given the limitations that content pages had to handle dynamic content), but some users don't think that is enough reason to create items for their categories. And yet additionally there is Bugzilla47930, which prevents us to access sitelinks from unconnected items. For now better to link to Commons galleries in the sitelinks, and link to Commons categories with Commons category (P373). Everyone is waiting to see what happens with the "Wikidata-for-Commons" and to see gz47930 solved, then some changes might be proposed.--Micru (talk) 11:10, 9 July 2014 (UTC)

Query all items that have coordinates but no image[edit]

Can somebody help me with a query for all items that have coordinate location (P625), but no image (P18)? I would like to create a map showing where images are missing. Possibly compile a few maps for the next Wiki loves [...] contest. A Comma-separated values (Q935809)-file would be ideal for importing into QGIS.
On a related note we should lobby the Commons-app-people to get Wikidata-data into the Commons-app ( A mobile-app that leads people to the places where images are missing would be really cool. Gamification of the image-hunt (similar to the tools by @Magnus Manske:) would probably be a big boost to Commons. Too bad we don't have a global requested image database. Currently all the requests are spread across all Wikimedia-sites Wikipedia:Requested pictures (Q6750881). -Tobias1984 (talk) 10:22, 10 July 2014 (UTC)

You can see the results for your query here, and download data here (add "&props=625" to the URL to also get coordinates for each item). But, I also have the tool you want! You can even specify an "administrative region"! Netherlands example. --Magnus Manske (talk) 10:33, 10 July 2014 (UTC)
@Magnus Manske: Thanks. Will look at it later. Seems like the server is down at the moment. -Tobias1984 (talk) 11:34, 10 July 2014 (UTC)
@Magnus Manske: - The interactive maps are really great. I am just having some trouble with the query syntax. How can I get the different language labels as columns? -Tobias1984 (talk) 16:54, 10 July 2014 (UTC)
You don't. You get the item IDs, which you can then use to get the labels, descriptions, Wikipedia pages etc. from Wikidata. No need to duplicate things. You can get one language here (under "download", after you ran the query). --Magnus Manske (talk) 19:30, 10 July 2014 (UTC)

N.B.: Your results might end up including broadcast stations in Canada and the USA. There could be 10000+ of them, but I don't know how many have coordinates. Coordinates on broadcast stations in North America are usually for the transmitter location: Often there is nothing to see there but an antenna in a field. --Closeapple (talk) 20:42, 10 July 2014 (UTC)

Groningen in Groningen and Groningen[edit]

On the page Groningen (Q749) for is in the administrative territorial entity (P131) (alias "is located in") it says 1) Groningen 2) Groningen. So it comes down to Groningen is located in Groningen and Groningen. I have seen things like this on items of non-English-speaking countries several times. I am not aware of examples from English-speaking countries, but would welcome to hear. Also it would be nice if some disambiguaiton information would be shown, e.g. the description or at least the Q-number. Tamawashi (talk) 12:41, 10 July 2014 (UTC)

You mean to show the description when looking at the claims on a Wikidata page? --Bthfan (talk) 12:51, 10 July 2014 (UTC)
Do you have another solution to make the page human readable? If identification of the linked item is not important, then why show it labels of linked items at all? Tamawashi (talk) 13:13, 10 July 2014 (UTC)
To me it was not clear what you meant with your post, so I asked... --Bthfan (talk) 13:18, 10 July 2014 (UTC)
I am sorry. After posting I already saw I read more into it than what was written. And I made a new reply, which ended in an edit conflict with you: "Yes, either that or show at least the Q-number as is done when using Template:Q, e.g. Groningen (Q892526). Then one could at least see that two different items are linked and that none is a self-reference. There are hundreds of items that are located in the area of an item with the same English label. This will increase if editors reduce names to base names, e.g. Foo Region to Foo, containing Foo Province (reduced to Foo) and the capital Foo." Tamawashi (talk) 13:24, 10 July 2014 (UTC)
What? Forum shopping again? Come on Tamawashi I already asked you at Talk:Q749 to please stop doing that! Multichill (talk) 12:52, 10 July 2014 (UTC)
What? Claim of forum shopping again? I already asked you to stop making false claims Talk:Q749. Tamawashi (talk) 13:13, 10 July 2014 (UTC)

P931place served (by airport)[edit]

The English label of place served by airport (P931) has just been changed from "place served" to "place served by airport". There has been a proposal on its talk page, since May, for it to serve a more general purpose (places are served by radio station, police forces, etc). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:52, 10 July 2014 (UTC)

Software skipping top result in drop-down suggestion[edit]

For some reason the software is not including the top result (e.g., the country for Brazil, Q33999 for actor) in the drop down suggestions in the search bar or when adding a new statement. Cbrown1023 talk 19:07, 10 July 2014 (UTC)

We're looking into it right now. Hopefully fixed soon. --Lydia Pintscher (WMDE) (talk) 10:06, 11 July 2014 (UTC)

ttwiki duplicate Barack Obama[edit]

The Tatar Wikipedia (ttwiki, татарча/tatarça) has two articles for Barack Obama. First, for Barack Obama (Q76), there is tt:Барак Обама, which I guess is language tt-Cyrl. Then for Barack Obama (Q13202704) there is tt:Barak Obama, which I guess is tt or tt-Latn; it has a template that says "Bu mäqälä Tatar (Latin) Wikipediäseneñ saylangan mäqälälär rätenä kerä." I don't understand that language. Is that allowed on ttwiki? What does that template mean? (@Dobroknig: because that's the only active user with Wikidata:Babel en + tt.) --Closeapple (talk) 20:28, 10 July 2014 (UTC)


The new page Wikidata:Properties lists pages that are related to properties but not to specific ones. Additions welcome.

I manually expanded the outdated list at Wikidata:List of properties/Geographical feature#Administrative territorial entity identifier. Very helpful for expansion was Wikidata:List of properties/all in one table created by User:Snipre. I would suggest a bot creates such a complete list of existing properties, maybe even including one column showing the number of items that use the property. The suggested place would be Wikidata:Database reports/Properties/All, similar to Wikidata:Database reports/Popular properties. Tamawashi (talk) 21:17, 10 July 2014 (UTC)

Missing Oscar nominated movies: User:Jobu0101/Oscars[edit]

I've created a list with all movies which received a nomination for an Academy Award but aren't linked here in Wikidata. Maybe you want to help me to get that list smaller. --Jobu0101 (talk) 23:31, 10 July 2014 (UTC)

I checked someone, but already exist, A lot of movie are created on, maybe is necessary to correct the list..Not so real I'm working on it. --ValterVB (talk) 09:35, 11 July 2014 (UTC)

I have checked 1978 to 2014 and removed ones that had items after I added the IMDb code. The ones that are left from those years did not appear to have items. New items will be needed. 09:39, 11 July 2014 (UTC)

Checked 1943, 1970, 1971 and 1972. 10:04, 11 July 2014 (UTC)
Thanks for your help. --Jobu0101 (talk) 11:24, 11 July 2014 (UTC)

Number of search results[edit]

It seems that the number of search results by default is brought back from 50 to 20. I prefer 50. Is it possible to set another default somewhere? I didn't find it. Kind regards, Lymantria (talk) 06:07, 11 July 2014 (UTC)

API: simplest way to obtain most used values for property[edit]

Hi. I'm in progress of creating new gadget for person infobox edit, and there is a non-urgent question. I have a field related to occupation (P106). By default such field allow to enter reference to any wikidata item. But it will be much better to provide user the list of most used values here, so he can quickly select of them (or enter his own, if he don't like the TOP 10 one). So the question is: how to obtain the most used values for any specific property? It is not need to be direct API access -- i can ask bot to query some other page (from toolserver) and update gadget JavaScript "dictionary" once per day. -- Vlsergey (talk) 08:26, 11 July 2014 (UTC)

The team of students that did the entity suggester are working on making it also suggest values at the moment. That should be accessible via the API then. That should cover it :) --Lydia Pintscher (WMDE) (talk) 09:01, 11 July 2014 (UTC)

"Subclass of" and "instance of" with the same value[edit]

Is there any way to query all items that contain both "subclass of" and "instance of" with the same value? I have seen some Widar edits lately that add "subclass of:value" without checking if there is already an "instance of:same value" present in the same item.--Micru (talk) 08:32, 11 July 2014 (UTC)

Autolist2 is your friend :) Thanks, GerardM (talk) 10:46, 11 July 2014 (UTC)
Thanks, I corrected the linked query (it should be p279, not p278) unfortunately that query returns 9200 items, but it doesn't really show the items that have the same value of "instance of" and "subclass of".--Micru (talk) 11:01, 11 July 2014 (UTC)