Shortcut: WD:PP/GEN

Wikidata:Property proposal/Generic

From Wikidata
Jump to: navigation, search

This page is for the proposal of new properties.

Before proposing a property
  1. Check if the property already exists by looking at Wikidata:List of properties (manual list) and Special:AllPages.
  2. Check if the property is already pending or has been rejected.
  3. Check if you can give a similar label and definition as an existing Wikipedia infobox parameter, or if it can be matched to an infobox, to or from which data can be transferred automatically. See WD:WikiProject Infoboxes for suggestions.
  4. Select the right datatype for the Property.
  5. Start writing the documentation based on the preload form below and add it in the appropriate section.

Creating the property

  1. Creation can be done after 1 week by a property creator or an administrator.
  2. See steps when creating properties.

Add a request

This page is archived, currently at Archive 31.

To add a request, you should use this form:

=== {{TranslateThis | anchor = en
| de = <!-- PROPERTY NAME IN German (optional) -->
| fr = <!-- PROPERTY NAME IN French (optional) -->
<!-- |xx = property names in some other languages -->
}} ===
{{Property documentation
|status                 = <!--leave this empty-->
|description            = {{TranslateThis
  | en = ...
|subject item           =  <!-- item corresponding to the concept represented by the property, if applicable; example: item ORCID (Q51044) for property ORCID (P496) -->
|infobox parameter      = Wikipedia infobox parameters, if any; ex: "population" in [[:en:template:infobox settlement]]
|datatype               = put datatype here (item, string, media, coordinate, monolingual text, multilingual text, time, URL, number)
|domain                 = types of items that may bear this property
|allowed values         = type of linked items (Q template or text), list or range of allowed values, string pattern...
|source                 = external reference, Wikipedia list article, etc.
|example                = {{Q|1}} → {{Q|2}}
|formatter URL          = 
|filter                 = (sample: 7 digit number can be validated with edit filter [[Special:AbuseFilter/17]])
|robot and gadget jobs  = Should or are bots or gadgets doing any task with this? (Checking other properties for consistency, collecting data, etc.)


(Add your motivation for this property here.) ~~~~


For a list of infobox parameters, you might want to use table format:

{{List of properties/Header}}

{{List of properties/Row|id=
|title          = audio
|type           = media
|qualifier      =
|description    = Commons sound file
|example-subject= Q187 <!-- Il Canto degli Italiani -->
|example-object = Inno di Mameli instrumental.ogg


For blank forms, see Property documentation and List of properties/Row

Generic properties[edit]

sort key[edit]

   In progress
Description key indicating the order in which the item's label should be sorted
Data type String
Template parameter In en:Template:Persondata, de:Vorlage:Personendaten
Domain all
Allowed values strings

Source Persondata
Proposed by Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits

May be used as a qualifier for name in native language (P1559); or qualified by a language. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:46, 24 October 2014 (UTC)

Pictogram voting question.svg Question Is there an international standard how names should be sorted? As far as I know it is extremly culture-specific how names are sorted. --Pasleim (talk) 13:10, 24 October 2014 (UTC)
No idea, but we often know it, from sources, or knowledge of our own culture. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:26, 24 October 2014 (UTC)
Chinese names are for sure sorted by similarity of the signs and the number of lines in the signs (quite a difiicult task to know the exact order of 2500 signs) and not like the transcriptions according to the alphabet.--Giftzwerg 88 (talk) 17:25, 24 October 2014 (UTC)
In Europe and America usually by last name and then first name. --Crazy1880 (talk) 17:27, 24 October 2014 (UTC)
See de:Hilfe:Personendaten/Name (german Help) --Crazy1880 (talk) 17:29, 24 October 2014 (UTC)
Pictogram voting comment.svg Comment If it's a qualifier, string datatype would probably be sufficient. Otherwise, it might need to be a monolingual string. --- Jura 18:59, 24 October 2014 (UTC)

As a librarian, I have to warn you that this is a very complex matter that can't be simply reduced to "In Europe and America usually by last name and then first name". But, rejoice: There is a standard work of reference, it's called Names of Persons and issued by the IFLA. A scanned version of the 4th edition is available as a PDF at the IFLA website. For example, in Iceland the sort order is "First name - last name" (the last name in Iceland usually being not a family name, but a patronymic - most Icelanders don't have a family name). And there are languages such as Spanish with multi-part name where the sort order also isn't obvious. "Names of Persons" helps in these matters. Gestumblindi (talk) 23:31, 27 October 2014 (UTC)

Oh, and of course - I Symbol support vote.svg Support this proposal, as a meaningful sort order is very important. Gestumblindi (talk) 00:10, 28 October 2014 (UTC)

Pictogram voting comment.svg Comment Persondata DOES NOT contain sort value for names. The |name= is supposed to be surname, firstname. About 20% of the cases, it is entered wrong, usually firstname, surname. DEFAULSORT contains the sort value in all Biography articles. Sort value does not equal surname, firstname in alot of cases.

Otto von Bismark... DEFAULTSORT:Bismark, Otto   persondata |name=von Bismark, Otto.
Francisco da Costa Gomes... DEFAULTSORT:Gomez, Francisco   persondata |name=da Costa Gomes, Francisco.

Persondata name value contains names with ligatures, accents and other characters. Sort value is only to be the standard 26-letter English alphabet plus ".'.

Two people, one named José Márquez, the other Jose Marquez. If standard 26-letter alphabet is not used, the two names will be sorted in different spots.

DEFAULTSORT values DO FOLLOW IFLA guidelines. See en:WP:NAMESORT for rules. There are two exceptions. WikiProject Iceland has said to follow western sort order for DEFAULTSORT. WikiProject Brazil and WikiProject Football has said Brazilian footballers have defaultsort set to their nickname.

I have been the maintainer of DEFAULTSORT on enwiki for several years now. If you have questions, I would be the one to ask. Bgwhite (talk) 01:11, 28 October 2014 (UTC)

Yes, Persondata often contains wrong name sorting, but it does try to do name sorting - otherwise, why should the sort order in "name" be "supposed to be surname, firstname" at all? That's a way of sorting, too. I'm not that familiar with English Persondata, more with German Personendaten. Maybe there are more errors in the English variant. "von Bismarck, Otto" is wrong, it should be "Bismarck, Otto von", of course. - That said, if it turns out that the sort order in Persondata is wrong too often, maybe it would be better for Wikidata to extract it from DEFAULTSORT instead of Persondata. - There may be conflicting results when extracting from different Wikipedia language versions. For example, German Wikipedia follows the IFLA guidelines for sorting Icelandic names in DEFAULTSORT resp. (in German) SORTIERUNG (so, first name first). Gestumblindi (talk) 14:37, 28 October 2014 (UTC)

Pictogram voting comment.svg Comment as a librarian too, I can only agree to the need for a sorting value... at least for Person's names...

but I see problems, even between latin-language names...
  • the sorting habits are not the same in all countries, even in countries using the same language ; therefore, different sorting values should be made for different languages...
  • and this will be even more complicated for russian/chinese/etc. languages... Translitteration makes it very difficult to sort names... just have a look at Tchekov's name :S - and this is for a "modern" person... imagine for medieval names which could be written differently by the same person :/ - and of course, it works backwards, for our "simple" latin names, when translitterated in... just how many non-latin languages are there in wikidata ? --Hsarrazin (talk) 15:49, 28 October 2014 (UTC)
how could this be solved in wikidata... ? probably, it would be best to have each part of the name in a separate property, and then let each project assemble them to have a sortkey...
Sounds reasonable (though a bit complicated). A sort key may also be valuable for other names than names of persons - e.g. work titles - you don't want to sort all book titles beginning with "The ..." under T, e.g. English Wikipedia has DEFAULTSORT:Shining, The. Gestumblindi (talk) 20:27, 31 October 2014 (UTC)
of course not… :D — pardon me for laughing, but that is exactly what the library-catalog at my job does… :(
for titles, if you look at wikisource fr for example… the text pages are named with the correct title, and, without adding manually a DEFAULTSORT, except in very rare cases, we have a "Classement" module, automatically applied through Title or Proofreadpage header template, so that the texts are sorted according to French rules… I don't think the same rules apply in all languages… but a similar system could perhaps be set for every language :) --Hsarrazin (talk) 01:42, 1 November 2014 (UTC)
Symbol support vote.svg Support but it's important we indicate which system applies - names in Irish (Gaelic) are sorted by different rules (you ignore the Ó or Mac prefix so Ó Rourke and Mac Raeman are sorted together). Filceolaire (talk) 20:49, 31 October 2014 (UTC)
Actually no. You do not ignore Ó or Mac per University College, Dublin and National Library of Ireland. However, I've seen both systems used. But you do hit the most important thing... indicate which system applies. If supported, there should be a group who writes out the rules. Sounds like Hsarrazin and Gestumblindi should be in the group. Sounds like German and English Wikipedia's base things from IFLA guidelines, so that is probably the best starting point. I personally favour a fight to the death. :) Bgwhite (talk) 00:01, 2 November 2014 (UTC)
  • Reject as proposed. This needs a multilingual field as the sortkey will differ by language. --Izno (talk) 07:52, 21 November 2014 (UTC)
It depends. Libraries try to apply a standard that is on the one hand language-specific, on the other hand universal, such as in the already mentioned Names of Persons. Persons are entered according to the custom in the respective person's language, notwithstanding the language of the library's location. So, a library following "Names of Persons" will sort Spanish persons according to Spanish convention, and Icelandic persons according to Iceland convention, even if it's a library in Switzerland or in Poland. This would be a possible approach here IMHO. So we wouldn't need a "Spanish sorting" and a "Polish sorting" for the same person, but just one, the one according to the person's language. Gestumblindi (talk) 21:58, 25 November 2014 (UTC)
The problem with that approach is that isn't how it's done on any particular wiki--each of which will and do have different sorting conventions. --Izno (talk) 00:12, 27 November 2014 (UTC)
Symbol support vote.svg Support but using DEFAULTSORT as the source. I am speaking as someone who has used the metadata provided by Persondata. I gave up using the name parameter - it was just too random. Even using the name parameter to display the persons name was impossible! My final solution was to display the Wiki page name and use DEFAULTSORT for ordering. Periglio (talk) 02:03, 18 January 2015 (UTC)

Symbol support vote.svg Support But I suggest to link it to some language rules. As mentioned above, there are different rules for different languages, so we need a name of type monolinugal (original name as a person would describe itself in its mother tongue) and a set of rules for a defined group of some languages and may be some other sets of rules for some other languages. We need this property if we want to replace the templates by properties of Wikidata.--Giftzwerg 88 (talk) 00:39, 16 February 2015 (UTC)

Pictogram voting comment.svg Doubts sv.wikisource and sv.wikipedia do not have the same sorting order, so I find it difficult to find a good solution here, even if we have one order for each language. And the sorting order of Swedish names depend on their age, so the language does not give enough information. -- Innocent bystander (talk) 09:31, 23 March 2015 (UTC)

language, except for works or persons[edit]

   In progress
Description language of item. Use more specific properties language of work (or name) (P407) or original language of work (P364) for works and native language (P103) or languages spoken or published (P1412) for persons.
Data type Item
Domain any item that has or uses a language except works or persons, like names, words, phrases, proverbs.
Allowed values items for languages

Jean (Q4160311), female given name → English (Q1860)
Jean (Q7521081), male given name → French (Q150)
Robot and gadget jobs no persons and no works of any kind allowed.
  • Pictogram voting comment.svg Comment P:P364 used to have the label "language". To match more closely its description, it's now labelled "language of the original work". This leaves a gap for cases like the above. --- Jura 04:52, 12 December 2014 (UTC)
  • Symbol support vote.svg Support --- Jura 04:52, 12 December 2014 (UTC)
  • Symbol support vote.svg SupportPictogram voting comment.svg Comment One day we will have also wictionary linked, so every word must have a language. Not only names, but also radio/tv stations broadcast in languages. However names are a bad example, because names tend to move between languages, sometimes unchanged over centuries and sometimes more or less modified.--Giftzwerg 88 (talk) 05:14, 12 December 2014 (UTC)
Pictogram voting comment.svg Comment I changed the domain to "any". --- Jura 05:16, 12 December 2014 (UTC)
In that discussion User:Snipre came up with the idea to merge language of work (or name) (P407) and original language of work (P364) into a "language" property which would no longer be restricted to works. --Pasleim (talk) 12:56, 18 April 2015 (UTC)
I don't see that discussion going anywhere, nor is the merge proposal formulated that way. The result is that we still haven't sorted out this issue. --- Jura 13:03, 18 April 2015 (UTC)
@Jura1, Pigsonthewing: Agree, it might take some time until there is a consensus so I won't oppose the creation of this property. --Pasleim (talk) 09:21, 19 April 2015 (UTC)

Time2wait.svg On hold per Pasleim. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:02, 18 April 2015 (UTC)

  • That other two properties have just no impact on this. --- Jura 12:05, 18 April 2015 (UTC)
    • @Pigsonthewing: I removed "on hold" as Pasleim changed is comment. --- Jura 15:48, 20 April 2015 (UTC)
    @Jura: Please try to think to that possibility: language of work (or name) (P407) and original language of work (P364) merged into a new property called "language" ? Why do we have to have one property "language of work" and one property "language of name" ? Snipre (talk) 13:57, 20 April 2015 (UTC)
    The reason for this request is that someone changed the label of the P364. --- Jura 15:46, 20 April 2015 (UTC)

different from[edit]

   Done: P1889
different from
an item that is different from another item, but they are often confused
Description This item is different from that one, and they are often confused. In contrast to said to be the same as (P460) that expresses uncertainty "... the statement is disputed", different from expresses a strong negation
Data type Item
Template parameter not that I know of
Domain any item
Allowed values any item
Example Philip A. Goodwin (Q7183094) (a politician) is different from Philip R. Goodwin (Q7184251) (a painter). Many pairs in P245 (ULAN) Unique value are examples of "different from" (and the rest should be merged).
Source owl:differentFrom (OWL) and gvp:aat2100_distinguished_from (Getty's LOD) (note: this should be moved to the property statements)
Robot and gadget jobs

  • Check that the two items do not have the same authority identifier (AAT, TGN, ULAN, VIAF, GND, etc). The "Unique value constraint" on these identifiers already checks this (eg see P245 (ULAN) Unique value), but together with an explicit claim "different from", we can split that section into "Delete authority identifier from one of the items" vs "Merge the two items".
  • Check that the two items are not subject of the same statement: if Q1 different from Q2, there should be no statements Q3 P1 Q1 and Q3 P1 Q2. Of course, that's a warning not a certain mistake.
Proposed by Vladimir Alexiev (talk)

Automated jobs (and less often people) make mistakes and identify two things that are in fact different. It's useful to document the difference, so others won't repeat the same mistake. It can also be used to improve authority identifier management (see the first "robot job" above).

Important: Obviously it's crucial to add a reference to such a statement, lest endless discussion ensues. But what to add? I found that Philip A. Goodwin (Q7183094) is different from Philip R. Goodwin (Q7184251) by inspecting their properties (different years, different professions), and ULAN for the latter. Ideally, I'd like to add a comment to that effect and my user name: but I can't do that in a reference. --Vladimir Alexiev (talk) 21:52, 11 January 2015 (UTC)

  • Symbol support vote.svg Support --- Jura 16:17, 7 February 2015 (UTC)
  • I think this should be placed in talk pages instead of item, as clients don't need it.--GZWDer (talk) 06:14, 16 February 2015 (UTC)
         select * {?x gvp:aat2100_distinguished_from ?y}
  • Symbol support vote.svg Support agree with Vladimir TomT0m (talk) 14:01, 22 March 2015 (UTC)
  • Pictogram voting question.svg Question - Is the intent that this essentially work as a "not to be confused with" statement? If so, I support it as useful externally as well as internally. Josh Baumgartner (talk) 18:01, 27 February 2015 (UTC)
    @Joshbaumgartner: It certainly is a useful usecase. TomT0m (talk) 14:07, 22 March 2015 (UTC)
  • Isn't the fact that we have two objects in our database always going to imply this property (except where that property contains a "same as")? Objects are different concepts always. --Izno (talk) 17:28, 9 March 2015 (UTC)
    @Izno: This depends on whether or not we make the Unique name assumption (Q7886954) (View with Reasonator). The OWL language, from which this property proposal is inspired, does not make it, so ... maybe in Wikidata in general it is not possible to make it, although it will be possible safely in subsets of Wikidata's dataset. TomT0m (talk) 14:07, 22 March 2015 (UTC)
The purpose is exactly "not to be confused with" (see the description). This will be used sparingly on two items that are likely to, or already have been, confused. OWL has a similar statement: owl:differentFrom. --Vladimir Alexiev (talk) 09:37, 1 April 2015 (UTC)
@Vladimir Alexiev, GZWDer, Jura1, Joshbaumgartner, TomT0m, Izno:@Filceolaire, T.seppelt: ✓ Done, see different from (P1889). --George (Talk · Contribs · CentralAuth · Log) 16:39, 19 May 2015 (UTC)

number of entries/articles[edit]

   In progress
Description number of entries/articles of encyclopedia/database
Data type Number
Domain encyclopedia/database
Allowed values number>0
Example Wikidata (Q2013) => 13,967,954
Proposed by GZWDer (talk)

GZWDer (talk) 06:10, 21 January 2015 (UTC)

  • Symbol support vote.svg Support --- Jura 12:29, 22 February 2015 (UTC)
  • @GZWDer: I'm amazed that it doesn't seem there are any properties yet for size, extent, height, width (tried autocomplete on some item, maybe I didn't search right?), etc. So definitely some dimension properties are needed, but I think should be defined in a more general way. Eg in this case "size" or "extent" with qualifier "unit" that can be left out (in this case), or be specified as an item "entries", "records", "pages", or whatever needed. --Vladimir Alexiev (talk) 11:48, 3 April 2015 (UTC)
    Quantity with unit data type is still in development. (Happens to be one of the top priorities.) --Izno (talk) 17:16, 3 April 2015 (UTC)
  • Pictogram voting question.svg Question Wouldn't this be better as a more generic "number of parts", to be used with a qualifier ("articles", "pages", "bricks", or whatever)? Unit measurements such as length and width would then be for other properties. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:37, 24 April 2015 (UTC)
    support for Andy's idea. --Pasleim (talk) 16:33, 27 April 2015 (UTC)
    @Pigsonthewing, Izno, Pasleim: Why not the opposite ? We already have properties for whole/part, has part (P527) and part of (P361) (although it might be a good idea to split them according to the corresponding RfC) why not something like
    < Wikipedia > composed of search < articles >
            quantity (P1114) miga < 100000 >
     ? We already have the properties. TomT0m (talk) 16:47, 27 April 2015 (UTC)
    @TomT0m: I see a problem with has part (P527) if we want to extend it to lets say "number of users". Saying
    < Wikipedia > has part (P527) miga < users >
            quantity (P1114) miga < 100000 >
    looks strange to me. However, if we turn it around and say
    < Wikipedia > number search < 100000 >
            of (P642) miga < users >
    it's in my opinion more natural. --Pasleim (talk) 21:17, 12 May 2015 (UTC)
    I on the other end of the line have a big problem with of (P642) miga personally, because it can be used everywhere and is not really well defined, plus a big problem with a main snack number: 10000 because it's meaningless as is, and it's exported alone in some Wikidata exports. Wikipedia is made of article, on the other hand, makes much more sense. As a car is made of 4 wheels, a car is made of wheels, and a car is made of 4. I actually already used this expression in Wikidada. But for the size of the community, it deserves a more specialized property anyway. Something like contributors or contributes, as contributors builds and manages the project. But we would nether say that Shakespeare is a part of Romeo and Juliet :) TomT0m (talk) 06:35, 13 May 2015 (UTC)
    @Pigsonthewing: Per TomT0m, once you go that direction I think we already have the properties (I'm not sure those properties would correctly be used in that fashion however. It's lingering around my head as a potentially good idea but I'm not sure about specifics. --Izno (talk) 15:45, 28 April 2015 (UTC)
  • Tentative support. Still musing on Pigs's suggestion. --Izno (talk) 15:45, 28 April 2015 (UTC)
  • Symbol support vote.svg Support. Filceolaire (talk) 23:21, 28 April 2015 (UTC)
  • Symbol oppose vote.svg Oppose while nobody has a good argument against the
    < Wikipedia > composed of search < articles >
            quantity (P1114) miga < 100000 >
    solution. TomT0m (talk) 10:34, 29 April 2015 (UTC)
    To give an example, and generalisation, I used this into the integer couple (Q19841009) (View with Reasonator) to express that an ordered pair of integers is made of two integers. This is a generic solution I did not really found a problem with. TomT0m (talk) 10:37, 29 April 2015 (UTC)
  • Symbol oppose vote.svg Oppose - has part (P527) with the quantity (P1114) qualifier is a cleaner solution. It also is easy to cover a work that has multiple types of parts (321 entries, 12 appendices, 3 indices, 42 tables, etc.) Josh Baumgartner (talk) 18:55, 5 May 2015 (UTC)

source language of given name[edit]

   In progress
Description language this spelling of a first name or given name comes from. Of use for Wikidata:WikiProject Names.
Data type Item
Domain given name items
Allowed values items for languages

Ash Crow
Harmonia Amanda
Чаховіч Уладзіслаў
Pictogram voting comment.svg Notified participants of Wikiproject Names

  • Pictogram voting comment.svg Comment This is to replace P:P364 on items for given names. P364 used to have the label "language", but is now labelled "language of the original work", so it can't be used anymore and we need to find a replacement. The more general proposal above (#language.2C_except_for_works_or_persons) didn't seem to fly. --- Jura 11:36, 7 February 2015 (UTC)
  • Symbol support vote.svg Support per proposal. --- Jura 11:36, 7 February 2015 (UTC)
  • Pictogram voting question.svg Question @Jura1: What's wrong with the generic "language" properties in this case ? Does not seem right to create one property for name, one other for whatever else, etc. A gender statement or/and make a statement like seems useful, however
    TomT0m (talk) 16:39, 23 February 2015 (UTC)
    • I don't know. It's just that I need some sort of a property for these items. Given the new (English) label of P:P364, I can't use that anymore. --- Jura 18:10, 23 February 2015 (UTC)
    • BTW, I improved the presentation of the example above. --- Jura 17:43, 24 February 2015 (UTC)
  • Pictogram voting question.svg Question What about native label (P1705). It can be used in any supported language instead of label, and consists both the source language and original writing form. That method is best if infobox want to show the name in original form in Wikipedia article. However, there is problem for spoken only languages, but do we need it? If yes I would prefer generic property that point for the source language of any name, no only the given one, for etymological purposes. Paweł Ziemian (talk) 14:49, 11 April 2015 (UTC)
  • Symbol oppose vote.svg Oppose. A property that is only used for given names is way too narrow. We do need a language but this isn't it. Filceolaire (talk) 17:58, 11 April 2015 (UTC)

The historical name of this[edit]

   Not done
Description Object for this historical name
Data type Item
Domain historical names, items without links to other Wikimedia projects
Allowed values geographical object (Q618123), organization (Q43229) and others.
Example Belarusian State Polytechnic Institute (Q17126687)Belarusian National Technical University (Q815268)
Belarusian State Polytechnic Academy (Q15626860)Belarusian National Technical University (Q815268)
Leningrad → Saint Petersburg (Q656)

Суть заключается в том, чтобы ссылаясь в статье на элемент, описывающий историческое название объекта, использовалось его название, но при этом гиперссылка вела на статью про объект. Сейчас же, если человек, например, родился в Ленинграде, то в карточке всё равно выдаст Санкт-Петербург, если создать отдельную страницу для названия Ленинград, то в статье вовсе не будет гиперссылки на Санкт-Петербург, хоть и будет выводится правильное название. (ru) – (Please translate this into English.) Чаховіч Уладзіслаў (talk) 12:24, 4 May 2015 (UTC)

@Чаховіч Уладзіслаў: Symbol oppose vote.svg Oppose, use
< Belarusian State Polytechnic Institute (Q17126687) (View with Reasonator) > official name search < Belarusian National Technical University (Q815268) (View with Reasonator) >
        valid in period search < the period >
or follows (P155) and co. if you're not referring to a name but a former organization. Note that the datatype of official name is not item but string (@Infovarius, Ivan A. Krestinin: could one of you please translate ?) TomT0m (talk) 18:08, 7 May 2015 (UTC)
Возможно создание отдельного элемента для имени действительно избыточно и достаточно воспользоваться свойством official name (P1448) с квалификаторами start time (P580), end time (P582)? Хотя если у названия будет существенно больше свойств, чем просто дата начала/дата окончания действия, то может быть имеет смысл и отдельный элемент создать. — Ivan A. Krestinin (talk) 20:56, 7 May 2015 (UTC)
Квалификаторы создают много проблем, например, как карточка определит, что человек учился именно в период, когда БНТУ носил название Belarusian State Polytechnic Academy (Q15626860)? Получается, что нужно искать и ручками вводить дату окончания (или поступления?), а если это так, то какой в этом смысл, если можно просто написать Belarusian State Polytechnic Academy (Q15626860)?--Чаховіч Уладзіслаў (talk) 08:44, 9 May 2015 (UTC)
Соглашусь, предлагаемое вами решение проще варианта с квалификаторами, но есть у него и проблемы. Например, когда кто-то захочет составить список всех людей учившихся в этом университете, то запрос сильно усложнится. Пожалуй не возьмусь судить какой же из подходов лучше, Symbol neutral vote.svg Neutral. — Ivan A. Krestinin (talk) 19:45, 18 May 2015 (UTC)

 Not done No support. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:40, 21 May 2015 (UTC)

(Disjoint)UnionOf (or any better name)[edit]

   In progress
Description every instance of this class is an instance of at least/exactly one class in that list of classes
Data type Item
Domain list, or metaclass
Allowed values class list, or metaclass list (resp)
< nucleon (Q102165) (View with Reasonator) > Disjoint union of search < proton (Q2294) (View with Reasonator) >
        together with search < neutron (Q2348) (View with Reasonator) >
Format and edit filter validation (sample: 7 digit number can be validated with edit filter Special:AbuseFilter/17)
Robot and gadget jobs check consistency of the instantiation relation wrt. these statements
This proposal is a proposal for Two properties:
  • UnionOf
  • DisjointUnionOf
(to get the definition of the first, take the text in green, resp in red for the second) plus one new qualifier :
  • Together With.

Sometime users use has part (P527) as a kind of inverse properties of subclass of (P279). This is semantically wrong as has part if for composition relationship of physical objects (or classes of those type) that are parts of other (bigger) physical objects (or classes of those whole objects), per Help:BMP, and NOT to give a list of the subclasses, like in the example give who was modeled as has part.

There would not be a case if we were just talking of an inverse property of subclass of (P279), but we're not. Here we want to say that the set of all instances of a set of classes (here proton and neutrons), when regrouped, is exactly the set of instances of the superclass (here nucleon). No instance of nucleon is neither a proton nor a neutron.

Example : let Academics be the set of person who teaches, studies or research in a faculty.
let teacher, student, researcher (be the class of persons who teaches, rep. studies and research in a faculty.
< Academics > Union of search < Teacher >
        Together with search < Students >
        together with search < Researcher >
means that any student, researcher and teacher is an academic, that no person who is neither student, researcher nor teacher is an academic. But a student can also teach, as a researcher.

With the disjoint variant (in green), the constraint is added that an instance of nucleon is an instance of one and only one of the subclasses. Said differently, that the set of subclasses is a Partition of a set (Q381060) (View with Reasonator) of the subject class.

< nucleon (Q102165) (View with Reasonator) > Disjoint union of search < proton (Q2294) (View with Reasonator) >
        together with search < neutron (Q2348) (View with Reasonator) >
as a nucleon is a proton or a neutron, but not both.

So the motivation is double :

  • give an alternative to those who want to put a list of subclasses of a class (don't do that, this is incorrect, do that instead)
  • improve our modeling expression power.

Note : the qualifier Together With is here because the set of subclasses must exist in the same statement. This is because if we gave the list of classes in a list of statement, we could not express that there is no other subclass needed to cover all the instances of the superclass, as Wikidata follows the open world assumption (Q851949) (View with Reasonator) (This means we could even use two statement to model two alternative divisions of the class in subclasses for all of its instances). For example let's just add a subclass to nucleon, I don't know highly energetic nucleon. It's a subclass of nucleon, but it would not make sense to add it into

< nucleon (Q102165) (View with Reasonator) > Disjoint union of search < proton (Q2294) (View with Reasonator) >
        together with search < neutron (Q2348) (View with Reasonator) >

because either a proton or a neutron could be highly energetical. It's a different classification axis. TomT0m (talk) 15:26, 9 May 2015 (UTC)

Note2 : DO NOT USE if the class list is not close, of course, that is if there is instances of the superclass that is not an instance of any of the child classes, or if it's not sure whether or not.

They are inspired by owl:unionOf and owl:unionOf and owl:disjointunionof (and alldisjointclasses but only because I did not realize there was actually disjointunionof /o\). TomT0m (talk) 19:08, 7 May 2015 (UTC)



  • "(sample: 7..."
  • "qnd"
  • line break before "wol:disjointunionof"
  • How many properties do you propose?
  • I dont see why "Disjoint" is mentionned at all.
  • "of this class .... of that list of classes" Which and which?
  • In everyday language, "UnionOf" seems (to me) related with has part (P527), not with classes.
  • If this proposal is about class partition, why not name it "class partition" or "subclass partition"?
  • If this proposal is not about class partition, why mentionning Partition of a set (Q381060)?

Visite fortuitement prolongée (talk) 21:36, 8 May 2015 (UTC)

@Visite fortuitement prolongée: should be better now. this and that is a common way to speak of resp the subject and the object of the claim. The names are the name given in owl as the links can show, but other are OK if people find this more clear. Union in math is the operation on set, for sure. I don't know about common english

There was a bug in Template:Color. I fixed it. Visite fortuitement prolongée (talk) 14:34, 9 May 2015 (UTC)


  • No slash inside "au moins une une et une seule".
  • Why "list, or metaclass" and not "Classes", like in Property talk:P279?
  • Why "class list, or metaclass list (resp)" and not "Classes", like in Property talk:P279?
  • "(sample: 7..."
  • "instanèiation"
  • Why not a semantic list after "This proposal is a proposal for"?
  • "too statement"
  • What is the difference between subclass property and "UnionOf" property?
    • Please give examples.
  • What is the difference between "UnionOf" property and "DisjointUnionOf" property?
    • Please give examples.
  • Please give an example where "Together With" qualifier property is required.

Visite fortuitement prolongée (talk) 14:34, 9 May 2015 (UTC)

Thank you, the proposal is much better now. Visite fortuitement prolongée (talk) 17:46, 9 May 2015 (UTC)
« Why "list, or metaclass" and not "Classes" » → I'll edit the Property documentation myself if the properties are created. Visite fortuitement prolongée (talk) 20:20, 11 May 2015 (UTC)
Per Help:metaclass_for an explanation. A class is a set of indvidual. two classes unionned gives another set of individuals, another class. A class of class unioned with a class of class gives a class of class, similarly. But a class should not be unioned with a metaclass to avoid some ontological problems. TomT0m (talk) 20:38, 11 May 2015 (UTC)
  • I'm not sure if I get the "Together with" qualifier point. Qualifiers are also open to be extended, so I don't see why having all values within a statement makes it more close than having one statement for each value. Qualifiers are extendable as statements are, there is no way to prohibit that. Therefore, I'm also not sure if we should actually introduce such limited classes, as that is contradictory to the open system of Wikidata. -- Bene* talk 21:08, 9 May 2015 (UTC)
    Good question. I take the opinion of Markus here : see this post on the ML. I think this makes sense for closed lists, it's an easy way to express that, way easier than for example having a sequence of items with followed by (P156)/follows (P155) terminated by
    <  > followed by search < no value Help >
    . TomT0m (talk) 09:23, 10 May 2015 (UTC)
  • Symbol support vote.svg Support "Disjoint Union of" property
  • Symbol support vote.svg Support "Union of" property
  • Symbol oppose vote.svg Oppose "together with" qualifier property. Instead of listing some subclasses with a qualifier you should just add more values to the main property.
Disjoint union of:proton
OK? Filceolaire (talk) 18:43, 11 May 2015 (UTC)
What you are suggesting, in the writing style that I prefer, is:
  • nucleon => proton, neutron
TomT0m want to allow several, distinct, subclass lists, like having in the same item both:
This way, it can not work. TomT0m need a change of Wikidata data structure, or a trick. "together with" is a trick. Of course, since TomT0m has not show an actual example of several, distinct, subclass lists, I do no support "together with", and you are free to oppose it. But I hope that with my explanation, you understand the need. Visite fortuitement prolongée (talk) 19:43, 11 May 2015 (UTC)
It's not a trick, it's a use of qualifiers to express close lists. TomT0m (talk)
@Visite fortuitement prolongée: A real life example : elements can be divided by for example : Monoisotopic element (Q3588104) (View with Reasonator) and non monoisotopic one. They can be divided by element groups such as noble gas (Q19609) (View with Reasonator)metalloid (Q19596) (View with Reasonator) some of these covers the set of chemical elements taken together. Some overlaps with others. TomT0m (talk) 20:30, 11 May 2015 (UTC)
@Filceolaire: No. This must be expressed in one statement. It make sense to say that Paul is a son of Jack. It also make sense that Jenny is a daughter of Jack. But from these two satements, we can't say if Jack have more children per open world assumption (Q851949) (View with Reasonator) which is reasonable in Wikidatas case. With disjoint union this does not make sense to say « nucleon union of proton » and « nucleon union of neutron », no more than in english. This has to be « nucleon union of proton and neutron ». Plus this allows us to say the list is close (see the discussion with Bene* above and the linked mail for this.) Without a qualifier this is as good as «has subclass». TomT0m (talk) 19:58, 11 May 2015 (UTC)
« Without a qualifier this is as good as «has subclass». » → Not exactly. "nucleon => UnionOf => proton, neutron, antiproton, antineutron" say more than "nucleon => subclass => proton, neutron, antiproton, antineutron"; because it say that the classes list is complete ("closed"). The difference is that without "Together With" qualifier, an item can have only 1 classes list, so we could no tell that... wait, what was this real example of you? Visite fortuitement prolongée (talk) 20:20, 11 May 2015 (UTC)
No, the meaning argument that a set of two sentence is not the same as one sentence still holds. But you're kind of tiring me. If you really don't want to be convinced follow the links to the owl properties, you'll see they uses one compound statements, not several statements with the same properties, for good reasons. TomT0m (talk) 20:30, 11 May 2015 (UTC)
@Visite fortuitement prolongée, Filceolaire: Any new question or vote ? It's been a while. TomT0m (talk) 19:49, 19 May 2015 (UTC)
Maybe you should move "(to get the definition [...] the second)" just after "a proposal for Two properties" ? Visite fortuitement prolongée (talk) 20:08, 19 May 2015 (UTC)
  • Symbol oppose vote.svg Oppose I never thought I would O in a proposal of mine the solution without the qualifier. Without it it's as good as has_subclass, as explained above. Of course I Symbol support vote.svg Support the original proposal :) TomT0m (talk) 20:14, 19 May 2015 (UTC)

Properties for this type[edit]

   In progress
Data type Property
Domain items with a subclass of (P279) statement
Example human (Q5) => sex or gender (P21)
Robot and gadget jobs Create constraint violation report

Could help us for documentation, conceptual clarity and maintenance. The most obvious use I see would be the creation of new kinds of "constraint violation" reports. The current system is organized by property, like Wikidata:Database reports/Constraint violations/P31. It is really useful for some purposes, but leaves holes and would usefully be supplemented by a system based on the item instance of (P31) statements. For instance, if human (Q5) (or a superclass of it) has a "properties for this type: sex or gender (P21)" statement, a constraint violation report would warn us about items that are instance of humans but do not have a P21. It is not easy to do this sort of things using a property-based method like Wikidata:Database reports/Constraint violations/P21.

@TomT0m, emw: for class issues, @Ivan A. Krestinin: for constraint reports. Zolo (talk) 18:16, 18 May 2015 (UTC)


Pictogram voting question.svg Question Will optional properties be linked using this property? For example human (Q5) => date of death (P570). — Ivan A. Krestinin (talk) 19:53, 18 May 2015 (UTC)

I don't really know, but I think it should start with the most straight-forward non-optional properties. For optional properties, maybe we should have a second property, or maybe it should use qualifiers to have more flexibility. --Zolo (talk) 06:21, 19 May 2015 (UTC)

Pictogram voting comment.svg Comment seems to be similar to this proposal --Pasleim (talk) 19:58, 18 May 2015 (UTC)

@Pasleim: actually I am not sure to get how Wikidata:Property_proposal/Property_metadata#property is meant to work but afaik it is supposed to be used in properties, while this proposal is for a property to be used in items. --Zolo (talk) 06:21, 19 May 2015 (UTC)
Yes that proposal focuses on constraints on properties but there isn't a big difference between the constraint date of birth (P569) => sex or gender (P21) and human (Q5) => sex or gender (P21) so using the same property could simplify it. --Pasleim (talk) 16:00, 20 May 2015 (UTC)

Pictogram voting comment.svg Comment From the OWL point of view, this is (as the preceding proposal) a part of a class expression definition, more precisely an existential quantification. A class can be defined such that its instances have this property with some value. TomT0m (talk) 19:47, 19 May 2015 (UTC)

  • Symbol support vote.svg Support with an alias to the OWL formulation. TomT0m (talk) 19:47, 19 May 2015 (UTC)