Wikidata:Property proposal/Archive/33

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


   Not done
Descriptionthe brother of one's spouse or the husband of one's sibling
Representsbrother-in-law (Q2914212)
Data typeItem
Template parameterattributs d'infobox de Wikipédia, s'il en existe ; par exemple : « population » dans fr:Modèle:Infobox Subdivision administrative
Domainhuman beings (P31:Q5)
Allowed valuesmodèle Q
ExampleJohn Tukey (Q382207)Frank Anscombe (Q3896195)
Sourceréférence externe, article de liste de Wikipédia, etc.
Robot and gadget jobsDevrait-il y avoir ou existe-t-il des bots ou des gadgets qui effectueront des tâches avec cette propriété? Par exemple: vérifier les autres propriétés afin d'être cohérent, collecter des données, automatiser un lien externe, etc.
  • Français :
     : On a des propriétés pour décrire les liens de parenté (ex Property:P7 pour les frères) mais pas de propriété pour décrire les liens par alliance qui peuvent aussi être très importants.
  • English:
     : We have properties which describe siblings (for instance Property:P7) but nothing for siblings in law.

(Ajoutez ici vos motivations pour la création de cette propriété) PAC2 (talk) 12:43, 6 July 2015 (UTC)[reply]

Q: Do all cultures use the same term for both of the relationships described? Also, what about sister in-law? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:34, 6 July 2015 (UTC)[reply]
 Oppose, déductible à l'aide des propriétés époux(se) et frère/soeur. TomT0m (talk) 18:42, 6 July 2015 (UTC)[reply]
 Oppose, deducible from existing properties. --Yair rand (talk) 20:13, 6 July 2015 (UTC)[reply]
 Oppose You may either deduce it, or use relative (P1038) with kinship to subject (P1039). Matěj Suchánek (talk) 08:02, 7 July 2015 (UTC)[reply]

 Not done; use the method described by Matěj Suchánek. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:02, 8 July 2015 (UTC)[reply]

Catalogus Professorum Halensis

Domaintypes of items that may bear this property
Sourceexternal reference
Robot and gadget jobsyes

This database has biographies of scholars who worked at the University of Halle-Wittenberg. It was originally a private website by the creator and is now hosted on the official website of the Halle university archives. It is already used in the German Wikipedia as a references for biographies (but without a template, as of now). Jonathan Groß (talk) 15:45, 19 July 2015 (UTC)[reply]


@Jonathan Groß, Magnus Manske: ✓ Done Catalogus Professorum Halensis ID (P2005) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:31, 26 July 2015 (UTC)[reply]

Broadcast Frequency

DescriptionThe frequency at which a radio station broadcasts.
Representsfrequency (Q11652)
Data typeNumber (not available yet)
Template parameter"Frequency" in en:template:infobox radio station
Domainradio station (Q14350)
ExampleKRKO (Q31451) → 1380, KTIS-FM (Q6338766) → 98.5 (FM), (AM)

Alternatively, this could have datatype String, and the examples could be "AM 1380" and "FM 98.5". Popcorndude (talk) 19:38, 3 July 2015 (UTC)[reply]


A "frequency" property, which appears to be designed for the same thing, is already on hold pending unit support at Wikidata:Property_proposal/Pending/2#frequency. - Nikki (talk) 00:16, 4 July 2015 (UTC)[reply]

In that case, I retract my proposal. Popcorndude (talk) 00:32, 4 July 2015 (UTC)[reply]

UCI code

DescriptionUCI code
Data typeString
Template parametercode UCI dans l'fr:Modèle:Infobox Équipe cycliste et dans le futur fr:Module:Infobox/Équipe cycliste
Domainsport cyclist (Q2309784)  View with Reasonator View with SQID
Allowed valueschaîne de trois caractères
ExampleSport Vlaanderen-Baloise (Q135701) → TSV
Format and edit filter validationTrois caractères, pas plus
Sourcearticle de liste de Wikipédia
Robot and gadget jobsyes

On the French Wikipedia, we are working to put on Wikidata the datas of infoboxes, it is the first step of a bigger project. The UCI gives each year (since 1995 at least) a code to the teams, see for example Liste des équipes continentales en 2015. This code is used on fr:Modèle:Infobox Équipe cycliste and will be on fr:Module:Infobox/Équipe cycliste. Because the names of teams changes, the code can change, so it will be used with start time (P580) and end time (P582). fr:Équipe cycliste Van Der Vurst-Hiko is a good example of article with UCI code. Jérémy-Günther-Heinz Jähnick (talk) 12:08, 3 July 2015 (UTC)[reply]

 Support --Pasleim (talk) 12:22, 3 July 2015 (UTC)[reply]

Facebook Places ID

Domaingeographical feature (Q618123)
SourceExterne Referenzen
Proposed byJ2hcom

Motivation. One important aspect of open data is building up a net of relations. So this should be able for facebook places relations too like it already is for e.g. open street map, wikipedia, freebase

GA candidate.svg Weak support As with the Foursquare one, I probably won't use it myself, but it seems sensible enough. I've also cleaned this one up a bit too. - Nikki (talk) 14:43, 26 June 2015 (UTC)[reply]

@J2hcom, Nikki: ✓ Done Facebook Places ID (P1997) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:53, 16 July 2015 (UTC)[reply]

ZooBank author ID

We already have ZooBank ID for name or act (P1746), but ZooBank also keeps track of authors, which can help with disambiguation. Daniel Mietchen (talk) 22:49, 15 July 2015 (UTC) OOjs UI icon error.svg WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.[reply]

 Support - I suppose. - Brya (talk) 05:37, 16 July 2015 (UTC)[reply]
 Support - A proposal I would have done myself in the future. ;) --Succu (talk) 06:09, 16 July 2015 (UTC)[reply]
 Support --Tobias1984 (talk) 12:59, 16 July 2015 (UTC)[reply]

@Daniel Mietchen, Brya, Succu, Tobias1984: ✓ Done ZooBank author ID (P2006) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:40, 26 July 2015 (UTC)[reply]

ZooBank publication ID

We already have ZooBank ID for name or act (P1746), and ZooBank author ID has been proposed above, but ZooBank also keeps track of what they call references. Since the term "reference ID" can be understood in multiple ways in both German and English, I suggest to go for "publication ID", which is more specific. Daniel Mietchen (talk) 07:03, 17 July 2015 (UTC) OOjs UI icon error.svg WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.[reply]

 Support - Sure. --Succu (talk) 07:39, 17 July 2015 (UTC)[reply]
 Support --Tobias1984 (talk) 08:08, 17 July 2015 (UTC)[reply]
 Support - Brya (talk) 10:42, 17 July 2015 (UTC)[reply]

@Daniel Mietchen, Succu, Tobias1984, Brya: ✓ Done ZooBank publication ID (P2007) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:46, 26 July 2015 (UTC)[reply]

IPNI publication ID


To make another trio complete: We already have IPNI author ID (P586) and IPNI plant ID (P961) from International Plant Names Index (Q922063) but still missing their identifier for a publication. Theses entries often have a cross reference to Taxonomic Literature II (TL-2). --Succu (talk) 11:30, 17 July 2015 (UTC) OOjs UI icon error.svg WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.[reply]

 Support --Tobias1984 (talk) 17:52, 17 July 2015 (UTC)[reply]

@Succu, Daniel Mietchen, Brya, Tobias1984: ✓ Done IPNI publication ID (P2008) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:53, 26 July 2015 (UTC)[reply]

WHMIS label

   Not done
DescriptionWorkplace Hazardous Materials Information System (Canada)
RepresentsWorkplace Hazardous Materials Information System (Q909418)
Data typeString
Domainchemical compounds
Allowed valuesA,B1,B2,B3,B4,B5,B6,D1A,D1B,D2A,D2B,D3,C,E,F
Examplemethanol (Q14982) → B2, D1B, D2A, D2B,

Similar to the GHS. Snipre (talk) 20:42, 22 June 2015 (UTC)[reply]

  • If there are only fifteen allowed values, they should be created and linked as items. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:44, 23 June 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment This property would be for SIMDUT 1988 only? My French is really bad and I'm not familiar with Canadian law, but I see only a few differences between GHS and the newest SIMDUT 2015. Probably we should create property for e.g. biological hazard pictogram from SIMDUT 2015, but the transition period (SIMDUT 1998 to SIMDUT 2015) ends by the end of 2018 and I'm not sure that we need property for SIMDUT 1988. ∼Wostr (talk) 18:11, 23 July 2015 (UTC)[reply]
    Thank you to point the new change in Canadian law. You are right the old system is not more relevant. Snipre (talk) 07:42, 30 July 2015 (UTC)[reply]


   Not done
Descriptionitem representing an eruption by this volcano
Representsvolcanic eruption (Q7692360)
Data typeItem
Domainvolcano (Q8072)
Allowed valuesinstances of volcanic eruption (Q7692360)
ExampleMount Ontake (Q1754806) => 2014 Mount Ontake eruption (Q18145950) with qualifier, start time = September 27, 2014

To connect volcanoes with eruptions, with qualifier start time (P580) ValterVB (talk) 20:09, 28 April 2015 (UTC)[reply]

  •  Support Extremely important item. 21:28, 28 April 2015 (UTC)[reply]
  •  Support why not? - Brya (talk) 05:35, 29 April 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment Copy edited. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 06:57, 29 April 2015 (UTC)[reply]
  • Pictogram voting question.svg Question would not significant event (P793) View with SQID fit ? This is typically the stuff for which we created this property. So I'll  Oppose if no good reasons is found : reason : another possibility fits, avoid multiplicity of modelling that could add confusion. TomT0m (talk) 07:08, 29 April 2015 (UTC)[reply]
    • Do you mean significant event (P793)? That seems to work, see my recent edits to 2014 Mount Ontake eruption (Q18145950). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 07:52, 29 April 2015 (UTC)[reply]
      • Yep, that's what I mean. TomT0m (talk) 09:38, 29 April 2015 (UTC)[reply]
        • Is a significant event necessarily an eruption? People killed by an avalanche might be considerated as a significant event occured on a volcano flanks but its far from being an eruption, in my point of view. Also, consider this item as soemthing strictly related to VNUM and VEL items.  SupportPaolobon140 (talk) 13:25, 29 April 2015 (UTC)[reply]
          • @Paolobon140: No it's not, but as the object will be an instance of eruption, this allows us to filter the relevant events in the infobox, for example. Or in a query. It's not hard. TomT0m (talk) 15:36, 29 April 2015 (UTC)[reply]
            • @TomT0m: I have proposed a new property because I wasn't sure that we could filter different relevant events for infobox. If is possible I'm agree to use significant event (P793). --ValterVB (talk) 15:59, 29 April 2015 (UTC)[reply]
              • @ValterVB: We need arbitrary access, not deployed yet (except in Wikidata) but the devs are working on it ATM and making progress. Should not be long, maybe in the next deployment, they seem to have made the groundwork. TomT0m (talk) 16:13, 29 April 2015 (UTC)[reply]
 Oppose per TomT0m --Pasleim (talk) 20:57, 4 May 2015 (UTC)[reply]
@ValterVB, Brya, Pigsonthewing, TomT0m, Pasleim, Paolobon140: The question which is not clear: do we have to create an item for each eruption or do we save data about eruptions as statements in the volcano item ? If we create an item for each eruption, we don't need a new property: just link the eruptions items with the volcano item using "part of" or similar properties. Snipre (talk) 23:43, 17 May 2015 (UTC)[reply]
@Snipre, ValterVB, Brya, Pigsonthewing, TomT0m, Pasleim:: I think saving data about eruptions as statements in the volcano items would be the best solution. That would originate a kind of history of known eruptions for every volcano item. VNUM number, which has been created a couple of weeks ago, would command on the eruptions list as well as official pages about volcanoes e.g. Philippine Institute of Volcanology and Seismology site for the Philippines volcanoes.Paolobon140 (talk) 09:24, 18 May 2015 (UTC)[reply]
Yes, item is the better solution, and for me to use "part of" in eruption item is better than to use significant event (P793) in volcano item. If no one oppose, we can delete this proposal. --ValterVB (talk) 16:44, 18 May 2015 (UTC)[reply]
I don't mind either way. - Brya (talk) 17:10, 18 May 2015 (UTC)[reply]
@ValterVB, Paolobon140: Using part of to mix an object with an event is a terrible idea. A leg is part of a body. A football match is a part of a competition. An event is a part of a bigger event. But an event is not a part of a physical object, or a physical object is not a part of an event. Events involves physical objects ... an eruption is an event in the life of a volcano. See the example in Help:Basic membership properties. TomT0m (talk) 17:20, 18 May 2015 (UTC)[reply]
I fully agree with TomT0m. part of (P361) must be used only to link two items of the same nature. significant event (P793) is more appropriate to link physical objects to events. Casper Tinan (talk) 20:28, 18 May 2015 (UTC)[reply]

@TomT0m: I was going to implement the reading of this data in a template (in we have already arbitrary access), but I have doubt. Now we have for example in Mount Ontake (Q1754806):

significant event      2014 Mount Ontake eruption
                       start time  27 September 2014

But AFAIK there is nothing that prevents a user to use instead:

significant event      volcanic eruption
                       start time  27 September 2014

It would be correct to use in this way? (I mean "volcanic eruption" instead of a dedicated item for an eruption). --Rotpunkt (talk) 12:15, 8 June 2015 (UTC)[reply]

@Rotpunkt: I think for consistency sake it's better if we try to limit the numbers of ways to express something. As there is items for some eruptions anyway, it's easier to keep only the item model. For data management we could have a query or a database pattern to detect the qualifier on event class pattern ... TomT0m (talk) 17:22, 8 June 2015 (UTC)[reply]
@TomT0m: Ok thanks, so everytime I find (via Lua module) an eruption written in the second way what should I do? ignore it? --Rotpunkt (talk) 17:29, 8 June 2015 (UTC)[reply]
@Rotpunkt: Or put the page in an hidden category ? That's usually done when a bad or obsolete practice is detected by template ... Then we should clean it, by hand or by bot. TomT0m (talk) 17:33, 8 June 2015 (UTC)[reply]
@TomT0m: Ok, thanks again! --Rotpunkt (talk) 18:20, 8 June 2015 (UTC)[reply]

@TomT0m, Rotpunkt, Casper Tinan, ValterVB, Pigsonthewing, Paolobon140: Can we close this proposal as not done ? Snipre (talk) 13:16, 22 June 2015 (UTC)[reply]

@Snipre: Yes Sir.Paolobon140 (talk) 08:19, 25 June 2015 (UTC)[reply]
+1 TomT0m (talk) 17:22, 4 July 2015 (UTC)[reply]

natural satellites

   Not done
DescriptionNatural satellites of the celestial body
Representsnatural satellite (Q2537)
Data typeItem
Domainastronomical object (Q6999)
Allowed valuesinstances of Q2537
ExampleEarth (Q2)Moon (Q405)
Sourcew:en:List of natural satellites

This is my first edit in here. But having read some of the docs I feel this property meets the criteria and would be useful. Chicocvenancio (talk) 04:45, 9 June 2015 (UTC)[reply]


See parent astronomical body (P397) and child astronomical body (P398). Visite fortuitement prolongée (talk) 20:11, 9 June 2015 (UTC)[reply]

Thanks. This seems more general, therefore more appropriate. 04:09, 10 June 2015 (UTC)[reply]
 Oppose Use child astronomical body (P398) / parent astronomical body (P397). -- Sergey kudryavtsev (talk) 07:00, 10 June 2015 (UTC)[reply]

 Not done; Use child astronomical body (P398) / parent astronomical body (P397). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:11, 11 June 2015 (UTC)[reply]

printed as

Descriptionuse as qualifier to indicate used spelling or how a certain information was printed in the source
Data typeString
Allowed valuesall
Example<Before the embers (Q17356113)> author <Léon Barracand (Q1571850)> with qualifier printed as <Léon Grandet>
Format and edit filter validationnone
Proposed byMicru (talk)
  • Necessary to show how a name was printed on a certain book (old spellings), or how a certain statement was represented on print (ex. "mid 18th to early 19th century").--Micru (talk) 14:39, 15 August 2014 (UTC)[reply]
  •  Support I've actually just suggested the need for something similar for dates. A lot of images I was uploading for the British Library recently had dates like "Mid 17th century to early 18th century". We can express that currently as 1780 (1730-1830), but we're losing information, and somebody else with specialist knowledge might have chosen different date ranges. So it's information that it would be a good thing to have a mechanism to store. I don't know whether "Printed as" is better or worse than "in orginal source", but I think it is functionality that we should have. Jheald (talk) 16:28, 15 August 2014 (UTC)[reply]
  • Pictogram voting comment.svg Comment Similarly to P357 (P357) we should use "By Statement" like Open Library does it. Example The duel, and other stories, "by Anton Chekhov, from the Russian by Constance Garnett". (Otherwise we would need to add a qualifier for the translator, 2nd author etc.) --Kolja21 (talk) 11:01, 20 August 2014 (UTC)[reply]
  • Pictogram voting comment.svg Comment Datatype: "Monolingual-text" might be better. --Kolja21 (talk) 16:29, 12 September 2014 (UTC)[reply]
  • Pictogram voting comment.svg Comment musicbrainz calles this an artist attribution which would be a better lable since it then would also apply to other kinds of creative works. and it, of course shouldn't be a multilingual variable since one issue of a printed book should only state a single language. if not there could be multiple modifiers.--Shisma (talk) 11:15, 25 September 2014 (UTC)[reply]
  • Pictogram voting comment.svg Comment I think {P387 (P387) can be used for this as a qualifier for date statements etc. to preserve the exact wording but still have the ability to add info to timelines etc. Filceolaire (talk) 21:43, 31 October 2014 (UTC)[reply]
  •  Support but rename to "original wording" and add aliases "printed as", "originally printed as", "author statement". I haven't seen the term "By Statement" in cataloging practice. "Attribution" could be added as alias, but it involves more scholarly activities of trying to figure out who created a certain work, with various outcomes (eg "formerly attributed to", "mis-attributed to", "in the style of", "following", "studio of", "workshop of" etc) --Vladimir Alexiev (talk) 10:04, 25 March 2015 (UTC)[reply]
  •  Support - still not created, after more than a year… I remember discussing it (otherwhere), and was sure it already existed… please, please, create it… it's really needed for a lot of works where the name of the author is not indicated on the work, (like initials)... --Hsarrazin (talk) 09:39, 23 May 2015 (UTC)[reply]
  • ✓ Done, following Vladimir Alexiev's suggested wording: object stated as (P1932) MSGJ (talk) 20:21, 10 June 2015 (UTC)[reply]

lyrics derived from

   Not done
DescriptionWork the lyrics were copied from, derived from, or based on
Representslyrics (Q602446)
Data typeItem
Domainmusical works Special values (having specialized validation schemes): Persons, Taxons
Allowed valuesall
ExampleTwinkle, Twinkle, Little Star (Q1058356) => The Star (Q18336412)
Proposed byMvolz (talk)

These two properties are needed for pieces of music/ songs where either the lyrics or the melody or both comes from another work. For instance, the children's songs Twinkle, Twinkle, Little Star (Q1058356), Baa, Baa, Black Sheep (Q552300), Altijd is Kortjakje ziek (Q18336199) and Alphabet song (Q931920) all have the same melody as Ah! vous dirais-je, Maman (Q18336185). And the melody was also the theme of several works of classical music, such as Variations on a Nursery Tune (Q7915823) and Twelve Variations on "Ah vous dirai-je, Maman" (Q2603975). The same is true for lyrics as well; it is relatively common for a song to be produced with lyrics from a previously published poem. For instance, the lyrics of Twinkle, Twinkle, Little Star (Q1058356) are from the poem The Star (Q18336412).Mvolz (talk) 22:45, 21 October 2014 (UTC)[reply]

 Support Seems reasonable to me. Would the alternative would be to use based on (P144) or not express the relationship in wikidata? --Mike Linksvayer (talk) 01:05, 22 October 2014 (UTC)[reply]

  • I think based on (P144) is better for this as it also covers songs where the poem has had minor alterations so I will  Oppose this. Filceolaire (talk) 21:58, 31 October 2014 (UTC)[reply]
@Filceolaire:: The problem is that based on (P144) doesn't semantically distinguish between the elements of the song it's based on (i.e. melody or lyrics, which are distinct components). I agree though that most aren't completely intact, although for the most part this involves repetition of lines as a refrain. I've changed it to "lyrics derived from"- @Mike Linksvayer:, does that sound reasonable to you? Mvolz (talk) 22:50, 2 November 2014 (UTC)[reply]
I'm afraid not. You can always use applies to part (P518) as a qualifier to based on (P144) if need be. I really don't think we need another property. Filceolaire (talk) 20:27, 3 November 2014 (UTC)[reply]
@Mvolz:, @Filceolaire: it seems useful to semantically distinguish which elements of another work a song is based on. Whether to do this with new properties or with qualifiers, I don't know. Are there general guidelines for wikidata on the use of qualifiers instead of new properties? And to make sure I understand use of applies to part (P518) as qualifier for these cases, value would be lyrics (Q602446) or melody (Q170412)? Mike Linksvayer (talk) 17:52, 5 November 2014 (UTC)[reply]
I am actually in favour of a dedicated property for cases where a melody is reused. I would only use based on (P144) for applies to part (P518):lyrics (Q602446). Filceolaire (talk) 19:14, 5 November 2014 (UTC)[reply]
Can you help me understand the distinction? Is it that when a melody is used, it would be incorrect to state that the work is based on the melody, but when lyrics are used, it is correct to state that the work is based on the lyrics? Or some other reason a new property is best for one, qualifier for other? Mike Linksvayer (talk) 03:58, 6 November 2014 (UTC)[reply]
Mike Linksvayer: When a new song is TTTO it usually reuses the tune without changing it so Danny Boy::TTTO:Derry Air. The tune of Danny Boy is not based on it is the actual Derry Air. Stories or songs which are based on other stories or songs will change things a lot more than that so I  Oppose a lyrics derived from property. Filceolaire (talk) 15:22, 15 November 2014 (UTC)[reply]
  •  Not done There doesn't seem to be consensus for this property. MSGJ (talk) 20:56, 10 June 2015 (UTC)[reply]

WorldArt identifier

   Not done
Descriptioncode number assigned to the subject on
Data typeString
Template parameteruk:Шаблон:Extref-anime, ru:Шаблон:Extref-anime, ru:Шаблон:Extref-manga, ru:Шаблон:Extref-people, ru:Шаблон:Extref-company waid (wamid)
Domainanime (Q1107), manga (Q8274), human (Q5), organization (Q43229); can potentially be used for video game (Q7889), film (Q11424), television series (Q5398426)
Allowed values (anime|manga|people|company|game|cinema)\/\d+
ExampleElfen Lied (Q214771) => anime/2746 =>
Elfen Lied (Q214771) => manga/4 =>
Masamune Shirow (Q442602) => people/227 =>
Clamp (Q334471) => company/21 =>
Robot and gadget jobslink autogeneration by a gadget
Proposed by4th-otaku (talk)

One of the main anime databases recognized by ru-wiki Anime and manga project; also referred to by Ukrainian Wikipedia. --4th-otaku (talk) 17:35, 8 December 2014 (UTC)[reply]

screenshot image

   Not done
DescriptionScreenshot of the software
Data typeCommons media file
Template parameteren:Template:Infobox software: "screenshot"
Domainsoftware (Q7397)
ExampleUbuntu (Q381)
Robot and gadget jobsReplace image (P18) with this property on every item which belongs to software (Q7397). Some might not be screenshots; need to involve humans in this process.
Proposed byRezonansowy (talk)

If we have coat of arms image, locator map, flag image, signature image, logo image instead of generic use of image (P18) then why not create a screenshot prop insted of it. Rezonansowy (talk) 17:35, 10 February 2015 (UTC)[reply]

This would be useful as a more specific-purpose property. Same like these given above. --Rezonansowy (talk) 17:50, 10 February 2015 (UTC)[reply]
The proposed bot-job, "Replace image (P18) with this property on every item which belongs to software (Q7397)" should not run, as some of the existing images may be of logos, or product packaging. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:00, 15 February 2015 (UTC)[reply]
Hmm, good observation. --Rezonansowy (talk) 11:52, 16 February 2015 (UTC)[reply]
Andy, would you support? --Rezonansowy (talk) 13:36, 25 February 2015 (UTC)[reply]
Not while that job is still listed; otherwise, yes. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:45, 26 February 2015 (UTC)[reply]
@Andy, I've taken the liberty of striking that bot job out in its present form. —SamB (talk) 00:22, 16 April 2015 (UTC)[reply]
Can you please explain what added functionality creating a separate property adds? What can you do with a new property what is hard or impossible to do with the current generic property?
We created the other properties because we wanted to add multiple specific images to an item. In line with that I'm still considering proposing "portrait image" so we can use that on persons. Anyway, how would screenshot help? Multichill (talk) 19:44, 8 March 2015 (UTC)[reply]
@Multichill: I did it for the consistency. See my first comment; if we have locator map image (P242), flag image (P41), signature (P109), logo image (P154) then why not to create a screenshot image? --Rezonansowy (talk) 13:44, 22 March 2015 (UTC)[reply]
 Oppose because image works just fine. No reason besides consistency. Multichill (talk) 14:55, 22 March 2015 (UTC)[reply]
@Multichill, Casper Tinan: What's wrong with the consistency? It's fine reason. --Rezonansowy (talk) 14:57, 25 March 2015 (UTC)[reply]

 Oppose Per Multichill. Casper Tinan (talk) 08:33, 23 March 2015 (UTC)[reply]

 Oppose. The reason why we have logo image, flag image, etc is because they are used on items about another entity besides the logo and flag. Will screenshot image be used on other items besides software items? If not, then there is no reason to create it. —Wylve (talk) 21:56, 4 April 2015 (UTC)[reply]

@Wylve: Well, there can be (and often are) other images besides screenshots on software items. Images that don't seem to fit any of the other image properties. (Though it's not exactly as though there is an easy way to get a readable list of all those properties at once to check for sure.) See, for instance, Netscape Navigator (Q235419). —SamB (talk) 00:22, 16 April 2015 (UTC)[reply]
@SamB: I'm not sure why a browser war usage graph would be a relevant image for Netscape Navigator (Q235419). The most relevant would be a software screenshot, which is after all a graphical representation of the software itself, not an aspect of it (like market share). —Wylve (talk) 06:34, 16 April 2015 (UTC)[reply]
@Wylve: Note that it can be also used for websites, so there's a reason to create it. --Rezonansowy (talk) 18:39, 6 May 2015 (UTC)[reply]
@Rezonansowy: Why can't websites use image (P18)? —Wylve (talk) 20:18, 6 May 2015 (UTC)[reply]
Because screenshot is more specific for this purpose, same like the other properties mentioned by me in the first comment. --Rezonansowy (talk) 20:09, 7 May 2015 (UTC)[reply]
I don't see the need for specificity. image (P18) is used to show a graphical representation of the entity in question. As I have mentioned above, the other specific image properties were created to show graphical representations of an aspect of the entity in question (e.g. logo of the item, map of the item, etc). Being a website or a piece of software, a graphical representation of the entity would naturally be a screenshot of it. Therefore specificity is redundant in these cases. —Wylve (talk) 21:48, 7 May 2015 (UTC)[reply]

 SupportSamB (talk) 00:22, 16 April 2015 (UTC)[reply]

Pictogram voting comment.svg Comment I changed it a little, from screenshot to screenshot image (for consistency as well). --Rezonansowy (talk) 20:12, 7 May 2015 (UTC)[reply]

  •  Not done There doesn't appear to be consensus for this property. Just use image (P18). MSGJ (talk) 21:29, 10 June 2015 (UTC)[reply]

MobyGames ID

DomainVideo games and video game series
Allowed valuesgamename, or platform/gamename

Decent database, good reference for age ratings and release dates. There would be a value per-platform, with one entry referring to all platforms, and then platform specific entries, with the appropriate qualifiers. --George (Talk · Contribs · CentralAuth · Log) 09:56, 26 May 2015 (UTC)[reply]

Discussion ID

DescriptionID of animated film in database (Q4066284)
Data typeString
Template parameter|anim_id= in ru:Шаблон:Мультсериал & ru:Шаблон:Мультфильм
Domainanimated film (Q202866)
Allowed values\d+
ExampleHedgehog in the Fog (Q260138)2684
Formatter URL$1
Robot and gadget jobsPBot can import values from Russian Wikipedia is the largest database about Soviet and Russian animation films. Links to it used in the infoboxes in the Russian Wikipedia. —putnik 10:49, 28 May 2015 (UTC)[reply]

 Support, one more high quality data source --Ghuron (talk) 12:51, 28 May 2015 (UTC)[reply]
 Support -- Sergey kudryavtsev (talk) 18:37, 29 May 2015 (UTC)[reply]
✓ Done: film ID (P1934) MSGJ (talk) 22:17, 10 June 2015 (UTC)[reply]

MovieMeter movie identifier

Template parameter"moviemeter" in nl:sjabloon:Infobox_film
Allowed valuesrange of numbers
Format and edit filter validationnumbers only
Robot and gadget jobsbot import from nlwiki

Popular Dutch website, similar to Rotten Tomatoes (Q105584). Sjoerd de Bruin (talk) 08:16, 18 June 2015 (UTC)[reply]


@Sjoerddebruin: ✓ Done MovieMeter film ID (P1970) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:08, 27 June 2015 (UTC)[reply]

MovieMeter director identifier

   Done: no label (P1969) (Talk and documentation)
Allowed valuesrange of numbers
Format and edit filter validationnumbers only

Popular Dutch website, similar to Rotten Tomatoes (Q105584). Sjoerd de Bruin (talk) 08:16, 18 June 2015 (UTC)[reply]


@Sjoerddebruin: ✓ Done P1969 (P1969) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:01, 27 June 2015 (UTC)[reply]

Open Hub identifier

   Done: Open Hub ID (P1972) (Talk and documentation)

Useful, community-curated, and freely-licensed resource on free software. Jean-Fred (talk) 12:55, 21 June 2015 (UTC)[reply]


@Jean-Frédéric: ✓ Done Open Hub ID (P1972) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:04, 30 June 2015 (UTC)[reply]

Thanks Andy! Jean-Fred (talk) 18:23, 30 June 2015 (UTC)[reply]

chapter of

   Not done
Descriptiondescribe a relation between two Wikisource-items
Data typeItem
DomainWikisource books
Allowed valuesmain page of the Wikisource edition
ExampleGospel of Matthew, translated by Helge Åkeson (Q20203636) chapter_of: the Bible, translated by Helge Åkeson (Q19524142)
Robot and gadget jobsonly carefully

I propose that this property should help to describe the relation between edition-items on Wikisource. WD:N tells it is still undetermined if these pages should have WD-items, but I think it is time to discuss how they should be included. Observe that not all chapter-pages on Wikisource can be found in subpages and not all subpages are chapter in a book. -- Innocent bystander (talk) 17:13, 23 June 2015 (UTC)[reply]

Pictogram voting question.svg Question Would this be a rather generic property which would take the logical part-of relation and leaving part of (P361) to a physical relation? (a chapter can end in the middle of an odd page and apart from paper-splitting there is no physical operation which could divide it from the next) -- Gymel (talk) 17:55, 23 June 2015 (UTC)[reply]
The problem I see with "part of", is that property has become to generic. This property should have a single-value-constraint, and I see no need to describe the opposite direction. A single poem can be "part of" several books, but a single Wikisource-edition of a poem is a "chapter_of" one and only one book. -- Innocent bystander (talk) 18:14, 23 June 2015 (UTC)[reply]
book (Q571) has no place when we differentiate between works and editions, it's coming from a completely different approach. I have a hard time trying to imagine in what way a poem can be part of one or several "book", except as in "the (or more precisely an expression like a version or a translation of) the poem is incorporated in that book", but that's pretty close to the property you are proposing.
Thus a range constraint could be set to work (Q386724)? However, also works may be logical parts of works, think of Gulliver's Travels (Q181488) or entries in a biographic dictionary. The items standing for wikisource entries are certainly of edition type, as you state, although depending on the situation they may also stand for the work (i.e. that logical part of the work covered by the item). Thus normally one would have to create a work-type item (as logical part-of the entire work) and connect the wikisource edition-item via edition or translation of (P629) with that. I see that in many situations this intermediate work-type item does not gain much (it would have to attract author and other vital information from the wikisource edition-item and leave that less versatile) so your proposed property tries a shortcut by extending "logical part-of" from the domain works to that of version, edition, or translation (Q3331189). Or should it be exclusively reserved for this "shortcut" case, meaning that only editions should be equipped with it and "normal" work-work relations should continue using part of (P361)?
But I see that the property value the Bible, translated by Helge Åkeson (Q19524142) in your example is already a certain editition (FRBR manifestation), which in turn is declared to be a specific edition of a specific translation, also declared as edition (FRBR expression) of the bible (FRBR work). In this case the proposed property is IMHO trying to state the only "natural" partitive relation for editions (edition-item for a chapter relating to an edition-item for the whole) and I can't see how to encourage users to abstain from using the generic part of (P361) (following your example Gospel of Matthew (Q392302) is not yet related to New Testament (Q18813), but the generic way to formulate this partitive work-work relation would be by means of part of (P361) again) (OTOH I also don't see how users can be convinced to not overuse part of (P361) for all kinds of cross-domain relations turning it into some too unspecific relation).
So how to express that the claim for part-of relation is "particulary clean", as in your case "between wikisource items"? Would qualifying edition or translation of (P629) be a solution, like ("applies to part" often stands for "applies to aspect", replicating instance of (P31) of the items here)? I'd rather not like to see since it's not a "private" relation. -- Gymel (talk) 08:00, 24 June 2015 (UTC)[reply]
The main reason I need this property, is to fix the Interwiki on Wikisource. On s:sv:Bibeln (Åkesons översättning) you can see that the page have interwiki to many different editions of the Bible. The Interwiki is not based on that several pages have sitelinks in the same item. That is the way interwiki is made on Wikipedia and other places. We want to have Interwiki-links to all editions of Q1845 even if there is several in one language. Therefor, this is solved by the help of a module, who follows the "edition of" and "editions"-properties here on Wikidata. Since there is more than one version of the Bible in English, you can see separators like "English (Geneva)" to describe that it is a part of the Geneva-bible, and not King James or anything else. The separators are hardcoded into this module today. But to be able to identify that the Gospel of Matthew in the Geneva-bible should use the same separator, we need to describe a relation between the chapter of "Matthew in the Geneva-bible" and the "main page of the Geneva-bible". Otherwise the hardcoding would be devastating, and it would be easier to add the interwiki manually, as we have done this far, and failed. "Part of" tends to be used in so many other ways. A psalm could be found in several places, but on Wikisource, the online-edition can only be a copy of one of them.
The main issue for Svwikisource is therefor not what is of (P642):Swedish Wikisource (Q16515025), but rather the pages in any other WS-project, including the text-namespaces in Wikipedia-projects like alswiki.
-- Innocent bystander (talk) 09:23, 24 June 2015 (UTC)[reply]
So you really only want a property as precaution or protection of the necessarily exclusive part-whole relation as expressed in the namespace-hierarchy between s:sv:Bibeln (Åkesons översättning)/Enligt Matteus and s:sv:Bibeln (Åkesons översättning) against users introducing some other part-whole relation into the wikidata items. But we can't imagine what kind of fact they would be aiming to express by that, so it remains purely hypothetic. Certainly, part of (P361) is extremely generic in scope, but you are operating on very special items so I assume the presence of part of (P361) on one of these items will almost inevitably have exactly that meaning you are relying on. O.k., the Bible, translated by Helge Åkeson (Q19524142)) could be part-of an hypothetic item "Åkesons complete works" and this item would perhaps no sitelink to swedish Wikisource and therefore things might break for you. Do you need a qualifier "this relation corresponds to / is backed by a site relation on" with value "Svedish Wikisource". Or even more specialized "this relation corresponds to a namespace relation on ...". Unlike the possible qualifiers above this would not try to refine the kind of relation (between edition-type items here) but rather point out usability for special cases without leaking details of the actual use in s:sv which would not be intelligible here? -- Gymel (talk) 23:20, 27 June 2015 (UTC)[reply]
As I said above, this is not something exclusivly for sv.wikisource, it's for all Wikisource-projects, and a handfull of Wikipedias with a Text-namespace. In fact, to make this work on sv.wikisource, it's more interesting to add such relations to items related to the other projects, than to the sv-project. But since I hope many of the other projects, are interested in Wikidata supported Interwiki, I guess they are interested in having such relation even for the sv-items.
This is also about inheritance of properties. A chapter of an edition, normally have the same Author, Publication-date, etc as the main item in the book. If there are more than one "part of", how do I know which item have the right Publisher, Editor etc?
A "Åkesons complete works"-edition do not have more than one publication date, the work-item may have, but not the edition item. -- Innocent bystander (talk) 06:14, 28 June 2015 (UTC)[reply]
If something is an edition and part-of another edition-item, then the usual expectation is that some edition-typic properties are inherited unless also set in the more confined item. Same with a part-of work-to-work relation. But I'd say its quite common for any part-of relation: A red fish can have blue fins, and if one doesn't specify the fin color one would assume it to be red, too. For the rest you have unfortunately lost me: s:sv:Bibeln (Åkesons översättning)/Enligt Matteus wants to link to all edition-type articles on certain projects, i.e. all "children" with . And it wants to label that link by the label of the edition they belong to (and group them "by inheritance" according to the language noted in the edition item). This becomes coherent if a different path confirms that this "edition they belong to" indeed is an edition of a bible translation: But there everything becomes muddy because it could be a partial translation (i.e. one of the N.T. only) and on s:sv you need a double P629 (from edition to translation to work) but cannot influence how other Wikisources will deal with that. Or some other Wikisource will not have the book Matthew as a complete text but the single chapters therein: There could be an issue with the granularity of the part-whole relation and your proposed "chapter-of" relation would suffer the same, "chapter" not being a concept which can exactly be defined and still kept flexible enough to server more than one book: You cannot prevent a "Gospel of Matthew in the Luther translation" item to be up-linked to a "N.T. in the Luther translation" item reflecting the (hypothetic) situation in s:de and only in the next step up-linked to the "Luther translation of the bible" which s:sv would like to use as a label. -- Gymel (talk) 22:31, 29 June 2015 (UTC)[reply]
You rise some really good concerns here. One Swedish bible we have, was published three times. (1526, 1536 and 1541) The 1526-edition was only about NT, the 1536 only included a few books in the OT and some Apocrypha. The 1541 inlcuded NT, OT, Apochrypha and the letter of Manasse. These books also includeded some non-biblical texts. All these three editions are regarded as an edition of The Gustav Vasa bible, and the Gustav Vasa-bible is normally regarded as a bible-translation, even if all editions are not about all of the bible and also includes some non-biblical texts. And of course, on Wikisource we do not have complete editions of any of these bibles. We only have fractions, even if we have a complete facsimile of the 1536-edition. (Well, some pages are missing, the 1536-editon was published in few numbers and only the Psalms exists in more than one copy.)
The Swedish "Normalupplagan" also exists in several different editions with different orthography (Q43091). It was published several times in the end of the 19th and the beginning of the 20th century. Not one single edition have both OT and NT, but together they have both parts.
We also have "The middle age"-bible. The texts looks more like Danish, but it's typical for middle age-Swedish. These texts are not published together, or not even at the same time and do not have the same translators. These translations were hand-written. They were republished in the 19th century, and it is that book we have a copy of on Wikisource. These texts only include fractions of the bible: Josue, Judicum, Jwdit, Liber Hester, Ruth, Libri Machabeorum, Apocalipsis Sancti Iohannis apostoli and Ewangelium Nichodemi. The last book is a NT-apochrypha, and normally never regarded as a part of the bible.
Since editions normally never have exactly the same set of texts, I see no problem with having a translation of only NT as an edition of Bible (Q1845).
Even if I do not regard the apochrypa as a rightful part of the holy scriptures, I see no big problem with a translation of only those books as an "edition of" the bible. The bible maybe have it's own problems here. But not even all editions of "Raven by Poe" have exactly the same number of verses. Also StarTrek have non-canonical texts. -- Innocent bystander (talk) 07:23, 1 July 2015 (UTC)[reply]

 Not done Obviously -- Innocent bystander (talk) 17:21, 16 July 2015 (UTC)[reply]

Internet Archive ID

   Not done
DescriptionThe Internet Archive stores work that are public domain and make them available to the public.
RepresentsInternet Archive (Q461)
Data typeString
Domainfilm (Q11424) and short film (Q24862), single (Q134556), album (Q482994), book (Q571), video game (Q7889)
ExampleGaslight (Q5526486) could link to
Formatter URL$1
Robot and gadget jobsSeveral Wikipedia articles link to items on the Internet Archive either as a regular external link or by using templates like en:Template:Internet Archive film (2183 transclusions), en:Template:Internet Archive short film (818 transclusions) and en:Template:Internet Archive game (256 transclusions). A bot could collect Internet Archive IDs from these links and templates and add the links to corresponding entry on Wikidata.)

The Internet Archive stores work that are public domain and make them available to the public. To eventually be able to search Wikidata for intersection of properties where one such property is availability on the Internet archive would be great. Example: "List all books available on the Internet Archive by male author with female protagonist set in 17th century" or "List all films available on the Internet Archive that are animated and with subject matter money". --Bensin (talk) 18:54, 27 June 2015 (UTC)[reply]

Proposal withdrawn per talk. But if someone with bot skills would like to look into extracting data from the templates that would be good. --Bensin (talk) 20:31, 27 June 2015 (UTC)[reply]
Wikidata:Bot requests is probably the best place to request that. :) - Nikki (talk) 21:12, 27 June 2015 (UTC)[reply]

Isn't this the same as Internet Archive ID (P724)? - Nikki (talk) 20:04, 27 June 2015 (UTC)[reply]

It's exactly the same. Weird I didn't find it. I'll link to IA from the property's talk page. Thanks! --Bensin (talk) 20:31, 27 June 2015 (UTC)[reply]

 Not done Withdrawn by proponent as duplicate of Internet Archive ID (P724). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:35, 27 June 2015 (UTC)[reply]

FSK film rating

Template parameterFSK in de:Vorlage:Infobox Film
Domainfilm (Q11424)
Allowed valuesFSK 0, FSK 6, FSK 12, FSK 16, FSK 18 (The appropiate Wikidata Q items will be created at proposal acceptance)
ExampleReservoir Dogs (Q72962) => FSK 18
Sourcede:Freiwillige_Selbstkontrolle_der_Filmwirtschaft#Altersfreigaben, en:Freiwillige_Selbstkontrolle_der_Filmwirtschaft#Overview

This proposal is motivated to create an property for the movie rating system used in Germany equivalent to existing properties like MPAA film rating (P1657), as well as ESRB rating (P852), PEGI rating (P908), USK rating (P914) for games ratings. The source of this proposal is a discussion at the wiki project for movies on the German Wikipedia. Mps (talk) 13:47, 30 June 2015 (UTC)[reply]


Pictogram voting comment.svg Comment This might be a special case of the proposal #intended public/target above. -- Gymel (talk) 18:27, 30 June 2015 (UTC)[reply]

1) e.g. the film de:Die große Stille has no restiction, but ist is surely not the intention of the creator, that the film is viewed by children, even though it will not harm to them.
2) If people who are not in the focus of the creator are watching the film, it has no consequences, but if the FSK-release is violated, it can have legal consequences e.g. §131 StGB.
--Färber (talk) 08:27, 1 July 2015 (UTC)[reply]
Well, @TomT0m: above is a bit vague about his intentions for "his" proposal. But instead of having dozends of rating properties for each country and each kind of medium (movies, computer games, ...) one could at least think of a more general rated for audience property which would have specific item values for e.g. FSK16 as value and would be qualified as (don't we have a better property to express that an agency assigned something, implying that only that agency is entitled to do so?) -- Gymel (talk) 12:35, 1 July 2015 (UTC)[reply]
 Support. Intended for audiences is too subjective. Sometimes even the age limits are different (under 15's only in one country, under 16's in another). We should just show all the actual ratings received from all the rating authorities even if we do end up with lots or properties. Filceolaire (talk) 18:56, 1 July 2015 (UTC)[reply]
If an author says he wrote a book for children (or a work. or …) or if there is a «from 7 to 77 years» on a boardgame box, it's no subjective and i has little to do with agency ratings. TomT0m (talk) 19:06, 1 July 2015 (UTC)[reply]

@Mps, Gymel, Färber, Filceolaire, TomT0m: ✓ Done FSK film rating (P1981) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:14, 6 July 2015 (UTC)[reply]


   Done: CPDL ID (P2000) (Talk and documentation)
RepresentsChoral Public Domain Library (Q519635)
Domaincreative work/musical compositions
ExampleAntonio Vivaldi (Q1340) -> Antonio Vivaldi (composer) // William Shakespeare (Q692) -> William Shakespeare (lyricist) // Cantigas de Santa Maria (Q1033824) -> Cantigas de Santa Maria (work)
Robot and gadget jobsimport w:Template:ChoralWiki
Proposed byCapmo (talk) 05:36, 10 July 2015 (UTC)[reply]

@Capmo, Filceolaire, Nikki: ✓ Done CPDL ID (P2000) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:16, 17 July 2015 (UTC)[reply]

Thanks! Capmo (talk) 14:26, 23 July 2015 (UTC)[reply]

Endangered languages

Please redirect this if it's being posted in the wrong place. I went to Wikimedia (Wikimedia topic Endangered Languages) with the idea and as far as I can tell this is where I am supposed to bring it.

I'd like to propose that all Wikipedia articles on languages include their conservation status, in a format nearly identical to that used for animals. While articles on animals get their citations from the IUCN Red List, the conservation status of languages would be cited from the UNESCO Red Book on world languages. Since this is a rather broad idea affecting a large number of articles, I wanted to bring it up somewhere I thought it would be heard rather than on an individual article. User:PiRSquared17 suggested it could also include conservation data from Ethnologue. At any rate, let me know what your thoughts are, everyone. Interlaker (talk) 22:18, 7 February 2015 (UTC)[reply]

  • I personally  Support the creation of the property, which could be called "endangered language status" or something like that. Is there an official range of values? Presumably we can just adopt UNESCO's. However, note, this is similar to the case of the property cultural heritage site/monument; I can't remember whether that one was created or instead the information added within instance of (P31). --Nemo 16:32, 8 February 2015 (UTC)[reply]

UNESCO language status

Template parameter"vitality" in w:ja:Template:Infobox Language
Domainlanguoid (Q17376908)
Allowed valueslisted here
ExampleJapanese (Q5287) => 1 safe, Ainu (Q27969) => 5 critically endangered
Proposed byVisite fortuitement prolongée (talk) 22:04, 9 February 2015 (UTC)[reply]

Motivation: Suggested by @Interlaker, Nemo bis:.

@Visite fortuitement prolongée, Interlaker, Filceolaire, TomT0m, Pamputt, Pasleim: ✓ Done UNESCO language status (P1999). Please create the necessary items. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 08:59, 17 July 2015 (UTC)[reply]

✓ Done. Visite fortuitement prolongée (talk) 19:50, 17 July 2015 (UTC)[reply]

EGIDS language status

   Not done
DescriptionEGIDS language status
Data typeString
Domainlanguoid (Q17376908)
Allowed valueslisted here
ExampleJapanese (Q5287) => 1 national, Ainu (Q27969) => 8b nearly extinct
SourceEthnologue (Q14790),
Proposed byVisite fortuitement prolongée (talk) 22:04, 9 February 2015 (UTC)[reply]

Motivation: Suggested by @Interlaker, Nemo bis:.

So how do we proceed here? Interlaker (talk) 22:11, 9 February 2015 (UTC)[reply]

Interlaker We wait for users to support or oppose and when an admin thinks there is a quorum they either reject the proposal or they create the property. This can take up to a month or even longer. If the property is approved then you can start using it to add the status to the wikidata item for each language. Once the wikidata item has the status it can be added to infoboxes in any of the 280 language wikipedias which have articles about this language. Start by adding your support below. Filceolaire (talk) 23:49, 1 March 2015 (UTC)[reply]
  •  Support but change datatype to 'item' to match IUCN conservation status (P141) and create 11 items to act as targets for this property. Filceolaire (talk) 23:49, 1 March 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment @Interlaker, Visite fortuitement prolongée, Filceolaire: I too suggest creating an item for each grade, and switching to datatype=item. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:25, 23 March 2015 (UTC)[reply]
  •  Support Pamputt (talk) 08:55, 19 April 2015 (UTC)[reply]
  •  Support with datatype "item" --Pasleim (talk) 19:16, 26 April 2015 (UTC)[reply]
  •  Oppose Ethnologue is not a RS for this information. They just added the category, and it's clear that in many cases they just made a guess, based on the demographic data they have, which they are also not a RS for. (It may be ethnic rather than speaker data, or may be half a century old.) There was recently a Rutger's class project on endangered languages, and the students often linked to, which suggests that many of the evaluations have little reliability. For example, Ethnologue rates Kiong as 8a moribund.[1] The other site has it as "severely endangered", but with only a 20% certaintly.[2] It looks at first glance that Ethn. has a credible source for it's assessment (Dimmendaal 2007), but that is likely only a ref for "Shifting to Efik", which does not mean the language is moribund or even severely endangered, and not for "Spoken only by people of sixty years or older" (i.e., moribund), which we must assume is unreferenced and therefore unreliable. Ethnologue would see that comment and therefore assign a ranking of 8a to the language, but without verifying that the statement is current or reliable. I don't know there assessment of this language is wrong (I just picked an article I edited yesterday), but often their assessments clearly are wrong, and unlike, they give no indication of the credibility of the ranking. I think we should have a discussion at the language wikiproject as to which sources are reliable before we start rolling out claims like these. Kwamikagami (talk) 21:13, 21 May 2015 (UTC)[reply]
  • Pictogram voting question.svg Question Is there an other source than for EGIDS language status ? Visite fortuitement prolongée (talk) 21:24, 21 May 2015 (UTC)[reply]
    • I can not find any source outside and SIL ( use and some other sources, but their scale is different from EGIDS, with only 6 levels like UNESCO scale). And the authors of EGIDS scale are members of SIL. Visite fortuitement prolongée (talk) 21:49, 21 May 2015 (UTC)[reply]
  •  Oppose per Kwami. The fact that give "0 international" level to no language in the world, even English and Spanish, should have gave me a hint. Visite fortuitement prolongée (talk) 21:49, 21 May 2015 (UTC)[reply]
  • @Interlaker, Nemo bis, Filceolaire, TomT0m, Pamputt, Pasleim: Visite fortuitement prolongée (talk) 14:34, 23 May 2015 (UTC)[reply]
    • I am not a linguist but should we discuss the reliability of a source. As far as I know, ethnologue is a very well known publication. Thus, it does not matter to Wikidata whether the data are reliable; the EGIDS code exists and we just have to decide if we want to import it into Wikidata. This import could allow people to reuse this EGIDS code correlated to other data of an item. It is enough to import these data IMO. Pamputt (talk) 15:47, 23 May 2015 (UTC)[reply]
  • I  Oppose using data from an unreliable source. I am happy for us to have links out to unreliable sources like IMDB to help them coordinate info with us and import info from us. Similarly I would vote to add an egids language reference code (if such a thing exists) to language items on wikidata but that is different from having a property in wikidata based on a particular datum in egids which we don't think is reliable. Once the info is in wikidata it will get added to wikipedias and allsorts of other sites because we have endorsed it. Filceolaire (talk) 02:50, 27 June 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment There seems to be no exclusive link between the classification and the suggested source. If this can be sourced from somewhere else, why not? --- Jura 17:59, 3 July 2015 (UTC)[reply]
    • Indeed. However, as I have wrote above (21 May 2015), I have searched (½ hour) an other source using or showing EGIDS scale, and I have found none. Visite fortuitement prolongée (talk) 19:13, 3 July 2015 (UTC)[reply]
      • Actually, in the meantime, I found a place that says it was developed for them. --- Jura 15:11, 4 July 2015 (UTC)[reply]

 Not done No consensus.--Micru (talk) 14:34, 27 February 2016 (UTC)[reply]

  • Pictogram voting comment.svg Comment If we talk about reliable source, UN Atlas of the World's Languages in Danger also has some flaws. They're using old resources and not showing the recent status. Reliable sources are hard to find for defining language status, as we encounter so many contradictory resources.--Beeyan (talk) 08:52, 3 November 2016 (UTC)[reply]


   Done: cuisine (P2012) (Talk and documentation)
DescriptionType of food served by a restaurant or restaurant chain
Data typeItem
Template parameter|food-type= in en:Template: Infobox restaurant
Domainrestaurants or restaurant chains
Allowed valuessubclasses (or instances) of cuisine (Q1778821)
ExampleHindoostane Coffee House (Q5766176)Indian cuisine (Q192087)

Primarily for use as a qualifier of instance of (P31). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:30, 9 May 2015 (UTC)[reply]


Anime News Network ID

Anime News Network person ID

Domainhuman (Q5)
Allowed values[0-9]+
Sourceparameter "annid" in ru:Шаблон:Extref-people, "id" in ru:Шаблон:Ann-people, "id" in en:Template:Ann where "type" is "people"

From the discussion on the original proposal (@Vlsergey, Hausratte, Micru: ping), the existing P1361 (P1361) was originally proposed as five separate properties but was added as a single one because they were not considered different enough. I think that was a mistake so I'm now proposing separate properties. See [3] for how the proposals looked before they were combined into a single one.

Main reasons:

  • The identifiers we currently use are made up, there is no such identifier "person/15", only a "person" entry with ID "15". There were other proposed formats in the original discussion ("anime.php?id=49", "anime49" and "a49") which seems to support the argument that there's no such thing as a single Anime News Network ID. It also makes it far less intuitive to enter and easier to get wrong.
  • As far as I understand it, the URL formatter does not support turning something like "people/15" into but that would be easy with separate properties. Other things which want to create the URL also can't without first splitting the "ID" into usable components.
  • With separate properties, we can add more specific constraints. An item with an Anime News Network person ID should be instance of (P31) human (Q5) and an item with a Anime News Network anime ID should be instance of (P31) anime (Q1107). Right now, a person could have an anime ID or an anime could have a person ID and we wouldn't be able to easily detect it.

I haven't included releases in this proposal because I couldn't find any examples of where it would be used.

- Nikki (talk) 01:24, 22 June 2015 (UTC)[reply]

  •  Support as yet another bunch of authority control properties. Be careful of the constraints however - some anime author names are pseudonyms used by a "group of humans". Some of these have enough backstory to make them fictional characters (e.g. Tharg the Mighty (Q7710764) the fictional editor of 2000AD). Filceolaire (talk) 16:59, 28 June 2015 (UTC)[reply]
  •  Support per Filceolaire. Pikolas (talk) 13:53, 29 June 2015 (UTC)[reply]

@Nikki, Filceolaire, Pikolas: ✓ Done Anime News Network person ID (P1982) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:23, 6 July 2015 (UTC)[reply]

Anime News Network company ID

Domaincompany (Q783794)
Allowed values[0-9]+
Sourceparameter "annid" in ru:Шаблон:Extref-company, "id" in ru:Шаблон:Ann-company, "id" in en:Template:Ann where "type" is "company"

See above. - Nikki (talk) 01:24, 22 June 2015 (UTC)[reply]


@Nikki: ✓ Done Anime News Network company ID (P1983) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:28, 6 July 2015 (UTC)[reply]

Anime News Network anime ID

Domainanime (Q1107)
Allowed values[0-9]+
Sourceparameter "annid" in ru:Шаблон:Extref-anime, "id" in ru:Шаблон:Ann-anime, "id" in en:Template:Ann where "type" is "anime", "ova" or "movie"

See above. - Nikki (talk) 01:24, 22 June 2015 (UTC)[reply]

@Nikki: Please check the example is valid. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:29, 6 July 2015 (UTC)[reply]
@Pigsonthewing: Fixed. - Nikki (talk) 04:36, 7 July 2015 (UTC)[reply]

@Nikki: ✓ Done Anime News Network anime ID (P1985) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:08, 7 July 2015 (UTC)[reply]

Anime News Network manga ID

Domainmanga (Q8274)
Allowed values[0-9]+
Sourceparameter "annid" in ru:Шаблон:Extref-manga, "id" in ru:Шаблон:Ann-manga, "id" in en:Template:Ann where "type" is "manga", "manhwa", "manhua" or "magazine"

See above. - Nikki (talk) 01:24, 22 June 2015 (UTC)[reply]


@Nikki: ✓ Done Anime News Network manga ID (P1984) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:34, 6 July 2015 (UTC)[reply]

ship length

   Not done
Descriptionthe length of a ship in meters
Data typeNumber (not available yet)
Template parameter"Ship length" in en:template:Infobox ship characteristics
ExampleBismarck (Q150857) → 251

Antrocent (talk) 17:37, 21 July 2015 (UTC)[reply]

There is already a proposed "length" property (Wikidata:Property_proposal/Pending/2#Length_2) waiting for support for numbers with units to be finished. - Nikki (talk) 17:56, 21 July 2015 (UTC)[reply]
  • Oppose creation per Nikki. --Izno (talk) 19:19, 21 July 2015 (UTC)[reply]

 Not done - duplicates earlier proposal. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:56, 26 July 2015 (UTC)[reply]

Wikisource index page

Descriptionqualifier of scan file (P996) to indicate the corresponding Wikisource Index page
Data typeURL
Allowed valuesWikisource Index pages
ExampleThe Kiss and Other Stories by Anton Tchekhoff (Q15839163) =>,_1908.pdf
Q19186576 =>Индекс:Дмитрий Минаев - На перепутьи, 1871.pdf]
Robot and gadget jobsimport from wikisource
Proposed by--Micru (talk) 11:57, 19 October 2014 (UTC)[reply]
  • Property proposed to link the scan file with its index page. There can be links to different Wikisources by using a different language.--Micru (talk) 11:57, 19 October 2014 (UTC)[reply]

@Sergey kudryavtsev, Ernest-Mtl, Aubrey, Billinghurst: Done.--Micru (talk) 09:22, 23 June 2015 (UTC)[reply]

Commons Creator

   Not done
Data typelink-invalid datatype (not in Module:i18n/datatype)
Example 1MISSING
Example 2MISSING
Example 3MISSING
Proposed byGerardM (talk) 13:26, 18 January 2014 (UTC)[reply]

The objective is create a machine readable link from Wikidata to someone who is known as a "Creator" on Commons. In this way it is possible to find images attributes to creators and present them for instance in the "Reasonator" eg Michelangelo and Michelangelo as a "Creator". This is what it links to. GerardM (talk) 13:26, 18 January 2014 (UTC)[reply]

That would give us items that should have a one-on-one relation to another item without any obvious benefit. The only argument you provide is "this is how we did it last time".. GerardM (talk) 13:17, 21 January 2014 (UTC)[reply]
Pictogram voting question.svg Question - Wouldn't it be easier to wait for the "F-items" suggestion, so we can add to the images on commons "creator = michelangelo"? This looks a little bit like a hack rather than a semantic-web solution. --Tobias1984 (talk) 13:24, 21 January 2014 (UTC)[reply]
  •  Support Creator use link to Wikidata, but in Wikidata not exist link to creator. Very long search, and data in Wikidata and Creator is not =.--Пробегающий (talk) 19:02, 31 January 2014 (UTC)[reply]
  •  Oppose This is a property of the Commons file and as such it should be recorded there. If it needs to be machine readable (and it does) then Commons needs to make it so. Filceolaire (talk) 06:22, 25 February 2014 (UTC)[reply]
@GerardM: couldn't this just be done using a sitelink to the wikidata item, pointing to the Commons page for the Creator? This would make it easier for the Creator page on Commons to import info from Wikidata but it would mean that we would have to either
  1. move the existing Commons sitelink on these wikidata items (it points to the CommonsCategory for this creator) or
  2. convince Lydia to allow multiple sitelinks from an item to a client wiki (which would also let us merge all the Category items on Wikidata into the corresponding non-Category items).
Filceolaire (talk) 17:13, 26 February 2014 (UTC)[reply]
Pictogram voting question.svg Question @GerardM: What is link datatype and in which sense will it be different from Commons Creator page (P1472)? --Pasleim (talk) 21:57, 21 January 2015 (UTC)[reply]
It is the same.. This is the older request.. GerardM (talk) 08:31, 22 January 2015 (UTC)[reply]

 Not done Duplicates Commons Creator page (P1472) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:40, 23 June 2015 (UTC)[reply]

Commons Institution

   Not done
Data typelink-invalid datatype (not in Module:i18n/datatype)
Example 1MISSING
Example 2MISSING
Example 3MISSING
Proposed byGerardM (talk) 13:26, 18 January 2014 (UTC)[reply]

The objective is create a machine readable link from Wikidata to what is known as an "Institution" on Commons. In this way it is possible to find images attributes to creators and present them for instance in the "Reasonator" eg Tropenmuseum and as a "Institution". This is what it links to. GerardM (talk) 13:31, 18 January 2014 (UTC)[reply]

No real argument provided above.. GerardM (talk) 13:17, 21 January 2014 (UTC)[reply]

 Not done Duplicates Commons Institution page (P1612). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:38, 23 June 2015 (UTC)[reply]


   Done: demonym (P1549) (Talk and documentation)
DescriptionUsed to replace deleted P49.
Data typeMultilingual text (not available yet)
ExampleRussia (Q159)=>Russian (in English)
Proposed byGZWDer (talk)

Motivation. GZWDer (talk) 10:14, 18 October 2013 (UTC)[reply]

I don't know how it works in other languages, but in the Russian language demonyms for the nation, men and women are different. Which of them offered to write, only for nation? --putnik 11:20, 18 October 2013 (UTC)[reply]
Yes, this is a very complicated area. It sounds more like something Wiktionary would handle.  Oppose creating this property, at least while Wiktionary support isn't up yet. --Yair rand (talk) 20:30, 22 October 2013 (UTC)[reply]

source website

DescriptionMain page of website/ online database from which this property is obtained
Data typeURL
Template parameter|source=, in Template:Property documentation
ExampleORCID iD (P496) =>

Motivation: To move such data from {{Property documentation}} to a property. Proposed by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:51, 8 May 2015 (UTC)[reply]

Which data item are you proposing to move that's in {{Property documentation}}? --Izno (talk) 20:35, 8 May 2015 (UTC)[reply]
|source=, where the value is a URL. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:46, 9 May 2015 (UTC)[reply]

@Izno: ✓ Done source website for the property (P1896) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:57, 21 May 2015 (UTC)[reply]

URI pattern

DomainWikidata property (Q18616576) - Wikidata:Glossary#Property
Allowed valuesURIs which contains an $1 as a wildcard for the statement value
ExampleIdRef ID (P269) ==>$1/id
Robot and gadget jobsno
Proposed byGuenthermi 16:44, 8 Mai 2015 (UTC)

Many properties in Wikidata represent ids that refer to external datasets, and many of these external datasets are also available in RDF (Resource Description Framework (Q54872)). The RDF version of a dataset uses URIs as identifiers that are different from the plain id. There already is a property formatter URL (P1630), which makes it possible to provide templates to create URLs for linking to Web resources based on the statement value. However, this Web page URL is usually different from the URI used in RDF, and it is therefore necessary to have a similar property for RDF URIs.

Having a "formatter URI" property would make it possible to automatically include links to external datasets into the Wikidata RDF export. This in turn will make it possible for users to combine these exports with the exports of other databases to ask queries over the combined data. Links to other RDF resources will also improve the integration of Wikidata into the Linked Data cloud, if the URIs will be included into Wikidata's live exports. With RDF being the main format for Wikidata's official query facility currently under development, and many ongoing activities for consolidating RDF support, this property therefore plays an important role. The current RDF exports of Wikidata Toolkit already include hardcoded URI formats for some important datasets: examples include ISO 639-3 code (P220), GND ID (P227), VIAF ID (P214), OCLC control number (P243), Library of Congress authority ID (P244), IdRef ID (P269), National Diet Library ID (P349), MusicBrainz artist ID (P434), MusicBrainz work ID (P435), MusicBrainz release group ID (P436), ChEMBL ID (P592), ChemSpider ID (P661), PubChem CID (P662), KEGG ID (P665), Gene Ontology ID (P686), MusicBrainz label ID (P966), MusicBrainz area ID (P982), MusicBrainz place ID (P1004). It would be much better if the Wikidata community could control which URIs are used for all of these (and for many others that are not currently supported in RDF at all).

  •  Support hopefully an easy decision since we already have a role model for this in the form of formatter URL (P1630) --Markus Krötzsch (talk) 14:49, 9 May 2015 (UTC)[reply]
  • Should it be named "RDF formatter URI"? Visite fortuitement prolongée (talk) 14:42, 9 May 2015 (UTC)[reply]
    • RDF is just one of several formats that uses URIs to identify objects. OWL is another one. In general, the property does not depend on which context the URI is used in. The difference to "formatter URL" is that the URI really identifies the object, whereas the URL just points to some related page. But "identify" is already part of "URI" so I don't know how to emphasize this further. --Markus Krötzsch (talk) 14:49, 9 May 2015 (UTC)[reply]
    • I now realize that "formatter" in "formatter URL" may actually refer to the Javascript formatter that displays links in HTML. Maybe a better name for the new property would be "URI pattern" to emphasize that it is a pattern (not a URI). --Markus Krötzsch (talk) 15:00, 9 May 2015 (UTC)[reply]
  • I would suggest to also use IRI instead of URI, to be compatible with RDF 1.1, so "IRI pattern". I am pondering about bringing up to also put this into context with the predicate to make the link out... skos:exactMatch, skos:closeMatch, rdf:seeAlso, owl:sameAs, ... see the work done on "scientific lenses" by the Open PHACTS consortium. Egon Willighagen (talk) 15:28, 9 May 2015 (UTC)[reply]
    • My default assumption would be that rdfs:seeAlso is used, but you are right that one could introduce a qualifier property that specifies another RDF property to use for linking. Btw, don't forget to add a support template if you support the property creation ;-) --Markus Krötzsch (talk) 15:36, 9 May 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment "How to name this property?" Summing up the above discussions on name in one place.
    • After the comments of Egon Willighagen and Visite fortuitement prolongée, my preferred label was IRI pattern, but I agree with Smalyshev that URI pattern would be more understandable, so this is my favourite now. --Markus Krötzsch (talk) 10:02, 15 May 2015 (UTC)[reply]
  •  Support, IRI pattern sound goods to me. --Succu (talk) 20:05, 9 May 2015 (UTC)[reply]
  •  Support I think it's great and good replacement for ad-hoc handling like mentioned above. It would be useful not only for RDF exports but for any tool that wants to make link between Wikidata stored data and outside repositories it refers to. I'm not sure however about the name - I think something like "URI pattern" would be better. IRI sounds a bit RDF-centric, as it looks like it can be only used for RDF. --Smalyshev (WMF) (talk) 02:23, 10 May 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment If accepted we'd add an option to Wikibase to make use of this property in the RDF output for example to better link Wikidata with the rest of the Linked Data Web. --Lydia Pintscher (WMDE) (talk) 15:53, 12 May 2015 (UTC)[reply]
  •  Support --Pasleim (talk) 20:49, 12 May 2015 (UTC)[reply]
  •  Oppose current name proposal.  Support under another name. Most editors won't understand the distinction between "formater URL" and "URI pattern", and we don't know how those will translate into other languages. Something like "Linked data URI pattern" would be less ambiguous. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:42, 15 May 2015 (UTC)[reply]
    • Pictogram voting comment.svg Comment I do not share this concern, for the following reasons:
      1. The usage of the proposed property is limited to property pages (a small set of pages edited mostly by expert users), and the values of the property will rarely change after property creation.
      2. By moving from "formatter URI" to "URI pattern", we have already decreased the similarity of both property names.
      3. The description will provide additional clarification to avoid confusion.
      4. As "URI" is the official name of a Web technology standard, it should be preserved in translations. It seems much more difficult to translate the informal term "linked data" in an understandable way (for example, the literal German translation of this term would be largely incomprehensible even to experts). In general, I don't think it is a good basis for discussing property proposals to speculate about their unknown translations: any property could be translated in an ambiguous or inappropriate way, but this should not prevent us from creating properties ;-).
--Markus Krötzsch (talk) 17:32, 15 May 2015 (UTC)[reply]
  •  Support Property can be renamed later, if necessary. No need to stall on this. --Denny (talk) 16:21, 20 May 2015 (UTC)[reply]

✓ Done formatter URI for RDF resource (P1921). Hope that label suits. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:43, 30 May 2015 (UTC)[reply]

Teuchos entry

   Done: Teuchos ID (P2018) (Talk and documentation)
Descriptionbiographical entry in the prosopography of the Teuchos centre
RepresentsTeuchos Center (Q2406946)
Data typeString
Domaintypes of items that may bear this property
Allowed valuestext
ExampleWilhelm Kroll (Q73314)Kroll.Wilhelm
Sourceexternal reference
Formatter URL$1
Robot and gadget jobsyes

This database has detailed biographical and bibliographical entries for scholars (mainly philologists) who contributed to textual criticism of ancient Latin and Greek authors in the 19th and 20th century. It was created and is maintained by Wilt Aden Schröder and hosted by the University of Hamburg. As of now, the database has 110 detailed biographical entries (which are a great reference) plus about 3500 entries from an annotated edition of Wilhelm Pökel’s 1881 lexicon, which don't have much data but can be useful.

This database is already in use on the German Wikipedia (w:de:Vorlage:Teuchos-Prosopographie) and will be a great addition for Wikidata. Jonathan Groß (talk) 13:13, 20 July 2015 (UTC)[reply]

@Marcus Cyron, DerHexer: Könnt ihr bitte pro forma euer OK geben, damit endlich jemand das Ding einrichten kann? Jonathan Groß (talk) 20:49, 28 July 2015 (UTC)[reply]

@Marcus Cyron, DerHexer, Jonathan Groß:

work period (start) and word period (end)

Descriptionperiod during which a person flourished (fl. = "floruit") in their professional activity
Data typePoint in time
Template parameter"workperiod" in c:template:creator
Domainpersons (including pseudonyms, house names etc.)
Allowed valuesany date
ExampleVincent van Gogh (Q5582) => between circa 1880 and circa July 1890
Format and edit filter validation(sample: 7 digit number can be validated with edit filter Special:AbuseFilter/17)
Sourceeg Commons:Creator:Vincent van Gogh, "fl." in biographies, VIAF, museum sites, etc
Robot and gadget jobsShould or are bots or gadgets doing any task with this? (Checking other properties for consistency, collecting data, etc.)
Proposed byJheald (talk) ammended by Filceolaire (talk)
Descriptionperiod during which a person flourished (fl. = "floruit") in their professional activity
Data typePoint in time
Template parameter"workperiod" in c:template:creator
Domainpersons (including pseudonyms, house names etc.)
Allowed valuesany date
ExampleVincent van Gogh (Q5582) => circa July 1890
Format and edit filter validation(sample: 7 digit number can be validated with edit filter Special:AbuseFilter/17)
Sourceeg Commons:Creator:Vincent van Gogh, "fl." in biographies, VIAF, museum sites, etc
Robot and gadget jobsShould or are bots or gadgets doing any task with this? (Checking other properties for consistency, collecting data, etc.)
Proposed byJheald (talk) ammended by Filceolaire (talk)

Proposal amended as the discussion below Filceolaire (talk) 23:00, 10 May 2015 (UTC)
Standard property for artists, creators etc: it's not unusual that we may not know when they were born, when they died, etc, but know quite well when they were active and their works were produced.[reply]

I'm also looking to migrate content from the Commons Creator template to WD, with a view to ultimately dematerialising it from Commons altogether (see Template:Creator/wrapper/test for some early tests). But I don't want to lose any functionality or data in the process. I've done my best to fill out the birdcage above, but need some help to know how best to code what are often likely to be probably quite approximate date ranges. Jheald (talk) 23:23, 20 September 2014 (UTC)[reply]

@Jheald: Would floruit (P1317) work? --Jakob (talk) 23:59, 20 September 2014 (UTC)[reply]
@Jakec: it's kinda different. For usual "life" "floruit" floruit (P1317) should be used, but we still need new property for years of active sport career or art career. -- Vlsergey (talk) 01:02, 21 September 2014 (UTC)[reply]
@Jakec:, @Vlsergey: Thanks, I had missed floruit (P1317). I think P1317 probably should work, but would need some adaptation to its existing rules and regulations. In particular, as per the Vincent van Gogh example above, one should be able to use it when one does know the date of birth and date of death, to indicate the period of activity. Secondly, we'd need to check that standard library routines for extracting and presenting such dates (eg Module:Wikidata) cope well with a range that can be actually quite precise. Per mw:Wikibase/DataModel#Dates_and_times, one could try to code this 1885-06-00 / precision:10 / before:5 / after 5.1, but I suspect that would probably decode to "about June 1885" (@Zolo: ?); in particular, there would be no way to specify that the two ends of the range have different precisions. Alternatively, does one make two entries, with separate start date and end date qualifiers? But then does the software know how to pick that up, and to interpret it as a range? Also, with two entries, it's hard to link them to show that the range has been specified by the source. Perhaps what's needed is an adjustment to the time datatype, to make it more possible to specify ranges with really quite well-known start and end points. Jheald (talk) 06:56, 21 September 2014 (UTC)[reply]
I think it's better to have two properties. Usually only the start point is specified and end point is assumed. -- Vlsergey (talk) 14:40, 28 September 2014 (UTC)[reply]
On further thought, having looked for typical examples of how this is used on c:Template:Creator templates, I do agree with User:Vlsergey that there is a case for the two different properties.
floruit (P1317) is good for asserting a particular moment in time when the artist worked.
But often what existing c:Template:Creator templates specify is much more precise -- an actual range, with well-identified start and end points.
That's a different thing. But the question remains, how best to specify the property. start time (P580) and end time (P582) are useful as qualifiers, but what assertion should they qualify? And could one still specify end cause (P1534) as a qualifer to a qualifier? Or if one wants to be able to use properties like end cause (P1534), does that mean a new first-rank property is needed? Jheald (talk) 13:46, 30 September 2014 (UTC)[reply]
Jheald if end cause (P1534) is used as a qualifier and the qualifiers subsequently all get moved out of order it is still clear what it means so there isn't a 'qualifier of a qualifier proplem. It can be a qualifier just like start time (P580) or end time (P582). Filceolbaire (talk) 22:34, 31 October 2014 (UTC)[reply]
Maybe two new properties. 'Floruit (start date)' and 'Floruit (end date)'?. Note that the span for a 'point in time' date - like 'Floruit' uses - is an uncertainty. It should never be be used for a range. Filceolaire (talk) 22:34, 31 October 2014 (UTC)[reply]
I think you're right. I think that's a good way forward.  Support. Jheald (talk) 01:34, 2 November 2014 (UTC)[reply]
  • Based on the comments above, I  Support. Eurodyne (talk) 23:43, 11 December 2014 (UTC)[reply]
 Support, though I like work period better than fluorit for these start and end dates. Still, we can have aliases. - - PKM (talk) 02:54, 15 December 2014 (UTC)[reply]
 Support based on above comments and because I think it could be useful for writers/artists/musicians, whose exactly birth date (or death date) is unknown but we know the period of their activity. --Nastoshka (talk) 23:40, 16 February 2015 (UTC)[reply]

I don't understand what exactly is the proposal here: to replace "floruit" with two properties 'Floruit (start date)' and 'Floruit (end date)'? But then, sometimes we just know that the author was active in a particular year, but we are not sure how many years he worked before and after that moment. How do you model that with two properties? BTW, I think "work period" is a better definition than "floruit": floruit makes sense when talking about writers, artists and the like, and doesn't sound good with sportsmen or politicians or others; while "work period" is in fact more general and can cover all cases. Candalua (talk) 08:59, 17 February 2015 (UTC)[reply]

Candalua The proposal is to add a new property. The discussion has raised an alternative proposal - to replace "floruit" with two new properties giving the start and end date. This is appropriate even if the person was only active in a particular year since quoting a year in wikidata technically refers to an instant in time which we are uncertain about exactly when. It does not refer to a period in time. To refer to a period - even one year - you need a start and end date, even if these appear to be the same because all we know is that both are sometime in the same year.

Pictogram voting question.svg Question. Why not just use start time (P580) and end time (P582) on occupation (P106)? Mushroom (talk) 00:31, 13 March 2015 (UTC)[reply]

Mushroom This is a special property to use when we don't have a date of birth or death. Filceolaire (talk) 13:41, 13 March 2015 (UTC)[reply]
  •  Oppose Instead we should replace "Floruit" with two new properties - 'Floruit (start date)' and 'Floruit (end date)'. Filceolaire (talk) 13:41, 13 March 2015 (UTC)[reply]
  •  Support Jheald Mushroom Candalua Nastoshka Eurodyne I have rewritten the property proposal and added the proposal below as the discussion above. I want to use this property for collective pseudonyms as well - these have their own items, separate from the individuals using them. Filceolaire (talk) 23:00, 10 May 2015 (UTC)[reply]
  • BA candidate.svg Weak oppose I'm getting the impression that Wikidata support for time intervals (start - end) as a data type is not under pursuit. The split into two properties has slight advantages over the current one for life dates floruit (P1317) (start and end date can be qualified individually, however one usually does not want to provide individual references) but will become problematic for persons with several professional careers: The proposed properties would have to be qualified by their corresponding occupation and I see no gain over using e.g. field of work (P101) or occupation (P106) qualified by the usual start and end date properties start time (P580) / end time (P582) or perhaps in this case more often time of earliest written record (P1249) or earliest date (P1319) / latest date (P1326). (And for persons with two careers in the same field both approaches are equally messy).
Furthermore I cannot see how floruit (P1317) and the proposed properties could coexist: Except for rare cases where a person is mentioned in the context of a specific point of time the mor common usage for P1317 is to record a work period (manifest in works still extant or mentioned in some written records). -- Gymel (talk) 06:25, 28 May 2015 (UTC)[reply]
  • I havge merged the discussion on these two properties because it was starting to get disjointed. Gymel seems to raise valid points. Would you like to address them Filceolaire or anyone else? MSGJ (talk) 08:32, 11 June 2015 (UTC)[reply]

 Not done No consensus.--Micru (talk) 14:34, 26 June 2015 (UTC)[reply]

Micru can we hold back on closing this discussion while I answer Gymel's weak Oppose.
Filceolaire, reverted. If flourit is deleted I guess this properties could take its place.--Micru (talk) 07:58, 27 June 2015 (UTC)[reply]
Gymel; These properties are for cases where we don't know birth or death dates. They should also be of use for collective pseudonyms and double acts as well since a pair of people don't have a birth or death date. I don't think start time (P580) / end time (P582) quite work for these cases.
I agree that floruit (P1317) doesn't coexist easily with these properties and I would propose that floruit (P1317) be deleted if these properties are created since, as you have said, floruit should refer to a period, not a single date. Filceolaire (talk) 02:32, 27 June 2015 (UTC)[reply]
Did I say so? Sometimes only one painting or one mentioning is known, and even that date is so unsecure that it deserves qualifying by sourcing circumstances (P1480). But I agree, one can certainly view upon this as kind of degenerate interval. Having two properties as proposed introduces an artificial obstacle, one would have to decide whether this single date is of "start" or "end" kind. OTOH several values for floruit (P1317) are hard to understand, and qualifying them with start time (P580) and end time (P582) is just a kludge. Going with three properties (floruit begin, floruit end, floriuit point in time) seems overkill: Right now the connection to date of birth (P569) and date of death (P570) is not close enough to be detected by many applications I suspect (floruit seen as best approximation for birth and death dates which are not known). Thus perhaps it would be wise to introduce two Q-items "earliest/latest indication of activity" and allow them as values for sourcing circumstances (P1480) to be used with birth/death dates or the current floruit (e.g.: floruit 17th century, qualified by 1680 earliest / 1695 latest)? Ooops, no: for floruit that would be almost completely redundant to using start time (P580) and end time (P582).
So sorry, I simply can't think of a solution: Having one property seems best, and having several properties seems to be calling for trouble. But lacking "compound" data tyües the only way to keep "related" dates together seems to lie in keeping them within one property and not spreading them. -- Gymel (talk) 08:38, 27 June 2015 (UTC)[reply]
  •  Oppose "floruit" should do. --- Jura 06:01, 7 July 2015 (UTC)[reply]
    • and yet Jura "floruit" doesn't do in practice for many cases. The current "floruit" property has a date datatype so it refers to a point in time but in so may cases we want to refer to a period of time.
    • Even for those cases where we only know one approximate date (the date of publication of a book for instance) we still know that the person (or group of persons) was active for some period in producing that work so we can use "work period (end)" for these cases.. Joe Filceolaire (talk) 20:41, 23 July 2015 (UTC)[reply]
@Micru: have we done enough to get these properties approved now? If they are approved then I will propose the deletion of "floruit" the next day. Joe Filceolaire (talk) 20:46, 23 July 2015 (UTC)[reply]
@Micru:? Joe Filceolaire (talk) 23:17, 11 August 2015 (UTC)[reply]
@Filceolaire, Jura1, Gymel, Jakec, MSGJ: Jheald Mushroom Candalua Nastoshka Eurodyne. Done. --Micru (talk) 21:27, 20 August 2015 (UTC)[reply]

AllMovie artist ID

Domainhuman (Q5)
Format and edit filter validation(sample: 7 digit number can be validated with edit filter Special:AbuseFilter/17)
Robot and gadget jobsimport from Template:AllMovie name (Q6057473)

For some reason we have AllMovie title ID (P1562) for films at but don't have a corresponding property for people. Template:AllMovie name (Q6057473) has been used over 4000 times on the English Wikipedia and similar templates exist on quite a few other projects, so this seems like it would be useful to have.

@Pamputt, ValterVB, Eurodyne: Pinging you because you proposed/supported AllMovie title ID (P1562).

- Nikki (talk) 01:41, 21 July 2015 (UTC)[reply]

 Support Pamputt (talk) 05:05, 21 July 2015 (UTC)[reply]
 Support --Marek Koudelka (talk) 17:42, 21 July 2015 (UTC)[reply]
 Support --ValterVB (talk) 19:48, 22 July 2015 (UTC)[reply]

@Nikki, Pamputt, Marek Koudelka, ValterVB: ✓ Done AllMovie person ID (P2019) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:27, 6 August 2015 (UTC)[reply]

United States Armed Forces service number

Domainhuman (Q5)
Allowed valuesopen
Format and edit filter validationnone
SourceService number (United States Armed Forces)
Robot and gadget jobsWe may be able to crawl the NARA database of military service histories and match them to wikidata entries.

This will link to service records maintained by NARA. Electronic Army Serial Number Merged File, 1938-1946 contains the Enlistment Records for 8,706,394 people who served in WWII. Here is a typical record Richard Arthur Norton (1958- ) (talk) 06:12, 22 July 2015 (UTC)[reply]

  • If this number can be used for different countries, it needs a qualifier to indicate what country system ;) --Hsarrazin (talk) 06:37, 22 July 2015 (UTC)[reply]
  • If there's ever likely to be a formatter URL for these, we might be better with separate properties for US military service number; UK military service number, German military service number, etc. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:57, 23 July 2015 (UTC)[reply]
  • I *think* it might be better to leave this as a generic field for all countries/services (and using military branch (P241) as a qualifier) rather than relying on specific properties for each, which are likely to get very complex - most countries will have at least half a dozen different numbering schemes used historically and very few of these will have a suitable online resource for the formatter URL to point to in any case. Andrew Gray (talk) 11:39, 29 July 2015 (UTC)[reply]
  •  Oppose no generic field for social security like identifiers. Potentially support a property for an US only "service number" per "Service numbers are public information available under the Freedom of Information Act, unlike social security numbers which are protected by the Privacy Act of 1974." on en:Service number (United States Armed Forces). --- Jura 05:14, 8 August 2015 (UTC)[reply]
    • Ok, given the changed scope:  Support --- Jura 20:58, 10 August 2015 (UTC)[reply]

@Richard Arthur Norton (1958- ), Hsarrazin, Andrew Gray: ✓ Done United States Armed Forces service number (P2028) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:50, 15 August 2015 (UTC)[reply] ID

Motivation resp. is a large database of soccer players which contains useful information like the club career of each single soccer player. If this property gets created, I plan to write a tool which does crosschecks between Wikidata and as we have a lot of vandalism on items about soccer players. Pasleim (talk) 15:02, 23 July 2015 (UTC)[reply]


@Pasleim: ✓ Done player ID (P2020) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:45, 6 August 2015 (UTC)[reply]

Erdős number

   Done: Erdős number (P2021) (Talk and documentation)
Allowed valuesinteger ≥0
Robot and gadget jobs

An obviously useful property. GZWDer (talk) 15:10, 30 July 2015 (UTC)[reply]


See also Six Degrees of Kevin Bacon (Q117054). Visite fortuitement prolongée (talk) 20:06, 31 July 2015 (UTC)[reply]

 Support. Joe Filceolaire (talk) 15:42, 2 August 2015 (UTC)[reply]

@GZWDer, Visite fortuitement prolongée, Filceolaire: ✓ Done Erdős number (P2021) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:54, 6 August 2015 (UTC)[reply]

Dictionary of Ulster Biography identifier

DescriptionA link to an individual's page on the Dictionary of Ulster Biography
Sourcemany articles already link to this as an external reference, and it is already within the mix 'n' match tool

Useful external source, such as Australian Dictionary of Biography ID (P1907) and Dictionary of Welsh Biography ID (P1648). Rock drum (talk) 16:10, 4 August 2015 (UTC)[reply]


@Andrew Gray: May be of interest. Rock drum (talk) 16:10, 4 August 2015 (UTC)[reply]

@Rock drum, Andrew Gray, Kharkiv07, Jonathan Groß: ✓ Done Dictionary of Ulster Biography ID (P2029) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:58, 15 August 2015 (UTC)[reply]

Linkedin Profile

   Done: no label (P2035) (Talk and documentation)
Descriptionpublic autobiographic profile of the person at
RepresentsLinkedIn (Q213660)
Data typeURL
Allowed valuesurls starting with or
ExampleBarack Obama (Q76)
Joan Rivers (Q240933)
Format and edit filter validationadd to ensure urls per allowed values
Sourceexternal reference, Wikipedia list article, etc.
  • Pictogram voting comment.svg Comment this to replace the current use of "website account on" and "website account name" as it doesn't work well with linkedin addresses. @Nikki: --- Jura 08:09, 8 August 2015 (UTC)[reply]
  •  Support Jonathan Groß (talk) 06:32, 10 August 2015 (UTC)[reply]

@Jonathan Groß, Jura1: Done.--Micru (talk) 18:57, 23 August 2015 (UTC)[reply]


   Not done
DescriptionBiographer, person or group of persons that produced a biography of the item
Representsbiographer (Q864380)
Data typeItem
ExampleAdolf Hitler (Q352)Ian Kershaw (Q312532) or Marcel Pagnol (Q236630)Marcel Pagnol (Q236630)

Any biography of an important person would include a mention of those who previously wrote his or her biography, whether in the footnotes or before. As such, the relation between someone and his or her biographers should be easily documented here. Thierry Caro (talk) 23:54, 8 August 2015 (UTC)[reply]


How would we define "biographer"? Countless people wrote about Adolf Hitler (including himself), which ones can be called biographers? Jonathan Groß (talk) 07:43, 9 August 2015 (UTC)[reply]

To be called someone's biographer, one usually has to produce an important biography (Q36279) of that person, whether because it is extensive or because it is short but contains most of the things we can know about the subject. This generally excludes people who produce, write or direct a biographical film (Q645928). Authors of an biographical article (Q19389637) can be listed if this work is one of the rare available sources of data about the person's life. I think common sense can prevail here just like it does elsewhere with other properties. Thierry Caro (talk) 08:50, 9 August 2015 (UTC)[reply]
 Oppose. Redundant to author (P50)/main subject (P921) data. --Yair rand (talk) 20:58, 9 August 2015 (UTC)[reply]
 Oppose. Use described by source (P1343) to link to the biographies instead. The items for the biographies can have an "author" statement.  – The preceding unsigned comment was added by Filceolaire (talk • contribs).

 Not done No support. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:44, 15 August 2015 (UTC)[reply]


   Done: length (P2043) (Talk and documentation)
Descriptionmeasured dimension of an object
Data typeNumber (not available yet)
Template parameterlength
Allowed valuestype of linked items (Q template or text), list or range of allowed values, string pattern...
ExampleTemplate:Q1 => 43,5 km
SourceExterne Referenzen, Listen in der Wikipedia (entweder Infobox oder Quelle)
Proposed byJcornelius (talk)

I'm really sorry, I'm new to Wikidata and still didn't get all the details. However, I'm currently adding a lot of information about underground stations, systems and and so on and the property "length" is essentially for this topic. Like, measuring railway lines and so. I really tried to find a property for this, but I didn't achieve. So I thought it may no exist, so here's the proposal. Correct me please, if I did something wrong. Jcornelius (talk) 15:28, 13 March 2014 (UTC)[reply]

Isomeric SMILES

Data typeString
Domainchemical compounds
SourceMain databases about chemical xompounds

Jasper Deng
Egon Willighagen
Denise Slenter
Daniel Mietchen
Emily Temple-Wood
Pablo Busatto (Almondega)
Antony Williams (EPA)
Devon Fyson
Samuel Clark
Tris T7
Robert Giessmann
Cord Wiljes
Adriano Rutz
Jonathan Bisson
Charles Tapley Hoyt
Peter Murray-Rust
Pictogram voting comment.svg Notified participants of WikiProject Chemistry

There is alredy a property for the SMILES notation of chemical compounds, see canonical SMILES (P233) but this property is not specific and mixes canonical and isomeric SMILES. This disturbs our constraints system which requires unique and single value for the use of p233. So the idea is to convert the current SMILES property into a canonical SMILES property and to create a new property for the isomeric SMILES notation. Snipre (talk) 14:11, 19 July 2015 (UTC)[reply]

@Snipre, Filceolaire: ✓ Done isomeric SMILES (P2017) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:13, 4 August 2015 (UTC)[reply]

Colour Index International Constitution Number

Allowed valuesNumbers (CI 15510 not allowed, 15510 allowed)

(Add your motivation for this property here.) Teolemon (talk) 18:21, 12 July 2015 (UTC)[reply]

 Support Quite useful to look for chemical compounds responsible of colors. Snipre (talk) 22:17, 12 July 2015 (UTC)[reply]
There is a C.I. Generic Name and the consecutive C.I. Constitution Number. For example indigo (Q422662) is C.I. 73000 but also C.I. Pigment Blue 66. Probably only the Constitution Number is meant here.--Kopiersperre (talk) 06:02, 13 July 2015 (UTC)[reply]
 Support. author  TomT0m / talk page 18:07, 30 July 2015 (UTC)[reply]

@Teolemon, Snipre, Kopiersperre, TomT0m: ✓ Done Colour Index International constitution ID (P2027), With datatype="string". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:37, 15 August 2015 (UTC)[reply]

Colour Index International Generic Name

   Not done
DescriptionColour Index International Generic Name for manufactured colour products
RepresentsColour Index International (Q868387)
Data typeString
Allowed valuesText : C.I. Pigment Blue 66
ExampleC.I. Pigment Blue 66 → indigo (Q422662)
SourceColour Index International
  •  Oppose This name should be an aliase. Snipre (talk) 07:46, 10 August 2015 (UTC)[reply]

 Not done No support. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:18, 15 August 2015 (UTC)[reply]

Name Etymology

   Not done
DescriptionAn account of the origin and historical development of a name.
Data typeMonolingual text
Template parameter"meaning" in en:template:infobox family name
DomainGiven Name (given name (Q202444))
Allowed valuestext
ExampleSophia (Q2302787) → Wisdom, Madison (Q1424362) → Son of Matthew, Gabriel (Q4925914) → "Able-bodied one of God"; or "messenger of God"

Many people when looking for a name for a baby are interested in the meaning of the name.

  •  Oppose This is a description, which should be defined in Wikipedia or Wiktionary. It might consists of whole paragraph of text. What about support for other languages? Paweł Ziemian (talk) 17:23, 10 July 2015 (UTC)[reply]
    @Paweł Ziemian: Most meaning infobox properties for given names contain very short meaning, usualy 10 words max, many times 1 or two words, and this is why this property is useful instead of the description, that can contain multiple paragraphs with a lot more info than the etymology LeonardoBueno (talk) 19:13 , 10 July 2015 (UTC)
    I think the property proposal is good, however, we should wait for Wiktionary support in Wikidata to implement it. The data type here could be an array of items, which are words (one or more) used to create the new name. These words might come from foreign languages such as Greek, Latin or Hebrew. Machines cannot understand what the description mean, so forget about any query for it or analyze it in automatic way. You don't need to read everything from Wikidata to put it infobox. Some information might be still provided in old way as direct parameter in Wikipedia article wikicode. Paweł Ziemian (talk) 19:28, 10 July 2015 (UTC)[reply]
  • Wait for Wiktionary integration. Etymology is clearly something Wiktionary handles and if we should ever handle Wiktionary in some fashion, the etymology of a phrase will be important. Until then, we don't have the mechanisms to multilingually use this data item. --Izno (talk) 19:34, 10 July 2015 (UTC)[reply]
  •  Oppose. If we do etymologies in structured data then we will link to the item for the word that the current word is derived from. This is inherently a wikitionary thing as wikidata doesn't have items for words - it has items for concepts, each of which has multiple labels. For now the place to put the monolingual text paragraph with the etymology of a name is in the wikipedia article about the name. (If the wikipedia article is a disambiguation list of people with that name then add the etymological paragraph at the top and convert the wikipedia DAB to a List.) Filceolaire (talk) 02:01, 12 July 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment I tend to agree that it's a bad idea to refer to linguistic entites in Wikidata. However, I think if the entity is actually a name, this is possible without too much problems. There is several preceding examples : human names, title names, the named after property … knowing that Wikidata wiktionary is not here until an unknown amount of time, this could be a temporary solution. I guess stuffs will have to be re-exanimated when that finally happens. (Side note OOjs UI icon error.svg WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. the nature of taxon and taxon names might be to sort out :) ) TomT0m (talk) 09:41, 12 July 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment this is already in the scope of Wikipedia. Even with Wiktionary support (we don't know if/when we get it for sure), would the structure be different? It seems to me that this would go into a monolingual string property anyways. Is this correct? --- Jura 09:53, 12 July 2015 (UTC)[reply]
    • Like I said above, I think etymology should be a link to a word item - the item for the word from which the current word is derived, so it's not a monolingual text item. A name could be derived from the Irish word for Blacksmith (like Magowan (Q1755217)) or from the English word for Blacksmith (like Smith (Q1158446)) but wikidata only has one item for Blacksmith.
    • We already have named after (P138). You can, for instance, use that to link personal names derived from the bible to the biblical characters they are named after. Filceolaire (talk) 13:13, 12 July 2015 (UTC)[reply]
      • enwiki seems to do mostly without such links. Wiktionary has them, but links to pages that also cover other things - they don't seem to be targeted. In conclusion, I don't think they are actually needed and monolingual string type could do for this type of information. --- Jura 09:53, 17 July 2015 (UTC)[reply]
      • As for P138: Sometimes, it's a tricky question to figure out if one was named after the other or if the other was just the best known early bearer of a name. If it's clear, I suppose people already use them. Still, the information is different from the one in this proposed property. --- Jura 09:53, 17 July 2015 (UTC)[reply]
  •  Oppose. Why only English? We should create 200+ (or 6000+) similar properties for other languages? --Infovarius (talk) 22:30, 12 July 2015 (UTC)[reply]
    • It's monolingual text, so you can define the language of text. --- Jura 09:53, 17 July 2015 (UTC)[reply]
  • Pictogram voting comment.svg Comment the main issue I see is that Wikidata (and probably also Wikipedia) at some point got flooded with a large number of etymologies that are (let's say it this way:) difficult to assess. As Wikipedia is better equipped to sort out such things, I tend to concentrate on the other aspects of given names and family names for now. --- Jura 09:53, 17 July 2015 (UTC)[reply]
  •  Oppose - this is for wiktionary or wp, not wikidata :/ --Hsarrazin (talk) 18:05, 1 August 2015 (UTC)[reply]

 Not done Widely opposed. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:59, 15 August 2015 (UTC)[reply]

Github username

Descriptionthis item's username on Github
Data typeString
Domainsoftware developer (Q183888)
Allowed valuestype of linked items (Q template or text), list or range of allowed values, string pattern...
ExampleMichael Niedermayer (Q20757848)michaelni
Format and edit filter validationcheck that it's a real account
Formatter URL$1

Proposing with the aim of splitting website account on (P553) (see Wikidata:Project chat#Website user names)).--Kopiersperre (talk) 11:09, 7 August 2015 (UTC)[reply]

  • Don't know if it's really useful to split this. Currently there are 18 items in Wikidata that have an account on GitHub. Mbch331 (talk) 11:28, 7 August 2015 (UTC)[reply]
  •  Support The English Wikipedia alone has over 130 links to GitHub profiles. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:51, 7 August 2015 (UTC)[reply]
  •  Support Not only are there a reasonable amount now, but this is a cateogry that will continue to grow exponentially. Kharkiv07 (T) 02:36, 8 August 2015 (UTC)[reply]
✓ Done GitHub username (P2037) Sannita - not just another sysop 15:36, 10 September 2015 (UTC)[reply]

From, To

Descriptionqualifier to indicate the start point of any linear geographical item -
Data typeItem
Template parameterWikipedia infobox parameters, if any; ex: "population" in en:template:infobox settlement
Domainevents, linear geographical items
Allowed valuesplaces
ExampleIrish immigration to Mexico (Q3043325):instance of (P31):human migration (Q177626) -> Republic of Ireland (Q27)
Format and edit filter validation(sample: 7 digit number can be validated with edit filter Special:AbuseFilter/17)
Sourceexternal reference, Wikipedia list article, etc.
Robot and gadget jobsShould or are bots or gadgets doing any task with this? (Checking other properties for consistency, collecting data, etc.)
Descriptionqualifier to indicate the end point of any linear geographical item -
Data typeItem
Template parameterWikipedia infobox parameters, if any; ex: "population" in en:template:infobox settlement
Domainevents, linear geographical items
Allowed valuesplaces
ExampleIrish immigration to Mexico (Q3043325):instance of (P31):human migration (Q177626) -> Mexico (Q96)
Format and edit filter validation(sample: 7 digit number can be validated with edit filter Special:AbuseFilter/17)
Sourceexternal reference, Wikipedia list article, etc.
Robot and gadget jobsShould or are bots or gadgets doing any task with this? (Checking other properties for consistency, collecting data, etc.)

So we can identify where things start from and where they go to. Joe Filceolaire (talk) 19:25, 12 August 2015 (UTC)[reply]

 Support But this needs a bit of thinking :) The domain has to be precised. Geographical feature do not have to have usually a from and a to. Processes have an inital situation and an a stop situation. Stream are places where water moves. There is a point from which the water flows and a point water reaches, but the stream (actually the stream bed stream bed (Q1429491)) itself do not moves :). (group) Migration is a process in which people in a group move one by one (or several by several) to another point. I guess a group
⟨ group migration ⟩ has part or parts (P527) View with SQID ⟨ individual migration ⟩
. So we should have a way to say also «what moves» in the process I guess … I love these exercises.
We already have origin of the watercourse (P885) and mouth of the watercourse (P403) for watercourses like streams. - Nikki (talk) 21:35, 12 August 2015 (UTC)[reply]
Streaming would be the process. The bed is the way. A stream may be a bit of both ? author  TomT0m / talk page 20:49, 12 August 2015 (UTC)[reply]
Thinking about it, as Nikki commented, we don't need this for rivers and streams. It's not appropriate for roads and highways either - coordinates for <endpoints> are better - but it is good for all sorts of journeys e.g. space probes (from:earth, to:Mars); fictional journeys ("Canterbury Tales" main subject:pilgrimage, from:London, to:Canterbury) and I am sure there are lots more uses. Joe Filceolaire (talk) 16:00, 14 August 2015 (UTC)[reply]
For geographical features use terminus (P559), for journeys we have start point (P1427) and destination point (P1444). --Pasleim (talk) 07:39, 21 August 2015 (UTC)[reply]
Withdrawn per Pasleim. start point (P1427) and destination point (P1444) do the job. Joe Filceolaire (talk) 09:30, 21 August 2015 (UTC)[reply]

Catalogus Professorum Academiae Groninganae identifier

Domainuniversity professors of the University of Groningen (Q850730)
Allowed valuesa row of numbers
Robot and gadget jobsmix-n-match and can be matched with Nationale Thesaurus voor Auteurs ID (P1006) as the page contains that ID

Contains at this moment 2340 (former) university professors of the University of Groningen (Q850730), since 1614. Useful as source as it contains valuable information. Sjoerd de Bruin (talk) 17:48, 25 July 2015 (UTC)[reply]


@Sjoerddebruin: ✓ Done Catalogus Professorum Academiae Groninganae ID (P2016) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:48, 3 August 2015 (UTC)[reply]

Hansard identifier

Domaintypes of items that may bear this property
Allowed valuesText with numbers and hyphens
SourceCould be imported into mix-and-match (easily scrapable index). About 8500 pages on enwiki use the "hansard-contribs" template which will provide probable matches.

This provides a fairly robust data source for British politicians over the past two centuries. It neatly bridges the gap between History of Parliament ID (P1614), which runs up to 1832 (Commons only, some gaps, for now), and biography pages (BEING DELETED) (P1996), which covers only recent years, and allows us to have at least one identifier for every member of the Commons and most of the Lords. Andrew Gray (talk) 11:46, 26 July 2015 (UTC)[reply]

@Rock drum, Jdforrester: - possibly of interest. Andrew Gray (talk) 11:49, 26 July 2015 (UTC)[reply]

@Andrew Gray, Rock drum, Jdforrester, Magnus Manske: ✓ Done Hansard (1803–2005) ID (P2015) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:41, 3 August 2015 (UTC)[reply]

German cattle breed ID

Template parameterde
Domainmostly cattle breed (Q20011627) (Definition of this item is "handmade", subject cattle is not as easy as cats or dogs; (domestic) yaks, (domestic) water buffalos, and bisons are cattle, too)
Allowed values01 – 99, vgl. ViehVerkV
Format and edit filter validationsubclass of Bovini (Q192205); two-digit number
SourceBGBl. I 2010, 227 (Umsetzung der Verordnung (EG) Nr. 1760/2000)
Robot and gadget jobsplease check, if it is a positiv two-digit number (no. 1 is "01") + not "00" + item is a (sub)subclass of Bovini (Q192205) + instance of (P31) breed (Q38829) (or subclass of it, like list of cattle breeds (Q1501517))
  • amtl. Bezeichnung, "identifier"
  • zusätzliche Überprüfbarkeit der Eigenschaft "Rinderrasse", da das de-Verständnis von "Rasse" sich nicht zwingend mit dem UK/US-Verständnis deckt PigeonIP (talk) 12:22, 1 August 2015 (UTC)[reply]
  •  Support --- Jura 05:43, 8 August 2015 (UTC)[reply]

@PigeonIP: ✓ Done German cattle breed ID (P2024) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:43, 9 August 2015 (UTC)[reply]

Findagrave cemetery number

Allowed values
Robot and gadget jobsA bot can be created to match Findagrave cemetery entries with existing Wikidata entries and create a new entry for cemeteries without Wikidata entries. Alternate can be to use the link in Wikipedia articles to supply the Findagrave cemetery number.

Currently there are no links to the massive cemetery database at Findagrave, we link to graves (Find a Grave memorial ID (P535)), not cemeteries. The cemetery numbers are formatted differently. Graves are formatted "GSid" and cemeteries are formatted as "CRid" followed by a number. We need a universal authority control for cemeteries. Geonames ID has over 500 cemeteries called Hill Cemetery. Richard Arthur Norton (1958- ) (talk) 19:05, 4 August 2015 (UTC)[reply]

 Support. Joe Filceolaire (talk) 16:30, 5 August 2015 (UTC)[reply]
 Support. Thierry Caro (talk) 02:40, 9 August 2015 (UTC)[reply]

@Richard Arthur Norton (1958- ), Filceolaire, Thierry Caro: ✓ Done Find A Grave cemetery ID (P2025) Datatype changed to string. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:02, 11 August 2015 (UTC)[reply]

NASA Biographical ID

Proposed by--Adert (talk) 12:41, 7 August 2015 (UTC)[reply]

@Adert: ✓ Done NASA biographical ID (P2030) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:59, 16 August 2015 (UTC)[reply]

ResearchGate ID

Allowed valuesvalid ResearchGate URL slugs
SourceWikipedia external links
Robot and gadget jobsPossibly Mix'n'Match

Good source of lists of works. Will be useful for cross-linking with other identifiers. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:56, 17 August 2015 (UTC)[reply]


@Filceolaire: ✓ Done ResearchGate profile ID (P2038) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:23, 26 August 2015 (UTC)[reply]

National Gallery of Victoria artist id

DescriptionIdentifier assigned to an artist by the National Gallery of Victoria in Australia
Data typeString
Domainhumans, mostly painters (a lot of them from Australia)
Allowed valuesa number
ExampleAnsel Adams (Q60809) → 2372 (
Format and edit filter validation\d+
Formatter URL$1/
Robot and gadget jobsmix'n'match will probably be used to collect the data

I don't want to create a property for every artist list in a museum out there. This case stands out because they seem to have people in their collection (mostly Australians) who don't show up in our usual sources of artist information. Multichill (talk) 19:28, 23 August 2015 (UTC)[reply]



   Not done
DescriptionHead of village within municipality.
Data typeString
Template parameterстароста
ExampleRunguri (Q4400111) → Фенин Василь Миколайович
Source[6], [7]

According new administrative reform in Ukraine each village will have each head of village - starosta, so we need infobox parameter for it.Ahonc (talk) 14:55, 2 August 2015 (UTC)[reply]


OpenStreetMap Way identifier

   Not done
DescriptionThe ID of the way of this place/object in OpenStreetMap (Q936)
Data typeString
Proposed byAgamitsudo (talk)


Today there is this OSM property [8] based on "relation". It would be great to have same properties for "node" and "way":

Thank you, --Agamitsudo (talk) 11:44, 22 January 2015 (UTC)[reply]

  •  Oppose OSM identifiers are not stable. Tag the way in OSM, with the Wikdiata ID (See Key:wikidata), and use coordinate location (P625) to find it. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:53, 6 March 2015 (UTC)[reply]
    • Andy how does that work? Which coordinate links to a 'way' when the way is an area or a river or a highway? Filceolaire (talk) 18:32, 4 May 2015 (UTC)[reply]
      • Coordinates never "link" to ways; find the object or relation within a suitable radius of the coordinates, which has the Wikdiata ID. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:08, 4 May 2015 (UTC)[reply]
        • Andy is that really the best we can do? OSM has ways defined for borders, rivers, highways, railways. We really, really want to have properties to link to those 'ways' from the corresponding wikidata items but we can't link to the OSM way. Really? Surely there some way to do this? Using a coordinate and searching won't help - remember that a lot of borders follow other borders - the border of a country is also the border of an adjacent country and of various subdivisions of both of those countries. Is there a chance to ask OSM to think about how this might be done? Filceolaire (talk) 00:39, 8 July 2015 (UTC)[reply]
  • I understand the appeal of wanting to link to specific objects in OSM, but I have to agree with Andy here. A way has no stable identity, OSM does not try to preserve IDs or prevent them from changing their meaning, plus the way OSM models things means that many things must be represented by multiple (often large numbers of) ways - an item in Wikidata is very often not represented by a single way (the border of Nordrhein-Westfalen is 502 ways, the Danube is 243 ways, the M25 in London is 527 ways, even the straightforward Jungfraubahn route from Kleine Scheidegg to Jungfraujoch is 12 ways, to list the first examples of borders, rivers, highways and railways that came to mind for me). I think the only good solutions for linking from Wikidata to OSM would have to come from OSM itself, e.g. making it easier to find everything which has a specific Wikidata ID via a search like or even introducing more integrated Wikidata support, like having a dedicated page like with the map highlighting all the objects and the sidebar listing the objects and showing some extra information pulled from Wikidata. - Nikki (talk) 06:55, 8 July 2015 (UTC)[reply]

 Not done No support. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:01, 15 August 2015 (UTC)[reply]

watershed area

   Done: watershed area (P2053) (Talk and documentation)
DescriptionThe area of a stream's watershed
Data typenumber with unit-invalid datatype (not in Module:i18n/datatype)
Template parameter"watershed" in w:template:infobox river
Allowed valuesnumbers with unit of area
ExampleHeberly Run (Q17544815) --> 6.42 [unit:square miles]
Sourcevarious, there is no one centralized dataset for all the world's streams
Proposed by--Jakob (talk)

One of the most important properties of streams and rivers. Different from the proposed area property because this is the area of the watershed, not the stream itself. --Jakob (talk) 14:34, 30 August 2014 (UTC)[reply]

Wait until the 'number with dimension' datatype is created so areas can be compared and converted to different units. Filceolaire (talk) 21:25, 31 October 2014 (UTC)[reply]

Time2wait.svg On hold pending data type. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:51, 22 March 2015 (UTC)[reply]

@Jakec: ✓ Done watershed area (P2053) --Rippitippi (talk) 10:14, 11 September 2015 (UTC)[reply]


   Done: wingspan (P2050) (Talk and documentation)
Data typeNumber (not available yet)


   Done: density (P2054) (Talk and documentation)
Descriptiondensity of a substance with phase of matter and temperature as qualifiers
Data typeNumber (not available yet)
Template parameteren, fr and de infoboxes about chemicals
Examplesee Wikidata:Chemistry_task_force/Properties.
Proposed byGZWDer (talk) 03:07, 7 June 2013 (UTC)[reply]
 Support Snipre (talk) 12:09, 7 June 2013 (UTC)[reply]
 Support --Kaligula (talk) 07:06, 19 June 2013 (UTC)[reply]
 Support--Saehrimnir (talk) 15:46, 24 June 2013 (UTC)[reply]
 Support --Cvf-ps (talk) 11:02, 26 June 2013 (UTC)[reply]
  •  Support it is necessary for infobox. Sunpriat (talk) 19:44, 6 October 2014 (UTC)[reply]

✓ Done --Tobias1984 (talk) 18:35, 11 September 2015 (UTC)[reply]

Electrical conductivity

Descriptionelectrical conductivity of a substance with phase of matter and temperature as qualifiers
Data typeNumber (not available yet)
Template parameteren, fr and de infoboxes about chemicals
Examplesee Wikidata:Chemistry_task_force/Properties.
Proposed byGZWDer (talk) 03:07, 7 June 2013 (UTC)[reply]
 Support Snipre (talk) 12:09, 7 June 2013 (UTC)[reply]
 Support --Kaligula (talk) 07:06, 19 June 2013 (UTC)[reply]
 Support--Saehrimnir (talk) 15:46, 24 June 2013 (UTC)[reply]

✓ Done --Tobias1984 (talk) 19:14, 11 September 2015 (UTC)[reply]

Heat capacity (en)

Descriptionheat capacity of a substance with phase of matter and temperature as qualifiers
Data typeNumber (not available yet)
Template parameteren, fr and de infoboxes about chemicals
Examplesee Wikidata:Chemistry_task_force/Properties.
Proposed byGZWDer (talk) 03:07, 7 June 2013 (UTC)[reply]
 Support Snipre (talk) 12:09, 7 June 2013 (UTC)[reply]
 Support --Kaligula (talk) 07:06, 19 June 2013 (UTC)[reply]
 Support--Saehrimnir (talk) 15:46, 24 June 2013 (UTC)[reply]
 Support --Cvf-ps (talk) 11:07, 26 June 2013 (UTC)[reply]

✓ Done --Tobias1984 (talk) 19:43, 11 September 2015 (UTC)[reply]


   Done: vehicle range (P2073) (Talk and documentation)
Descriptionmaximum distance equipment can reach or be operated within (note conditions as qualifications)
Data typeNumber (not available yet)
Example<Boeing 737> range <10,200 km>
Proposed byJoshbaumgartner (talk)
 Oppose 10,200km for Boeing 737 would be maximum theoretical range in optimal environment. Also, 737 series have many variants and range varies depend on type of aircraft. For this reason, I think that this property could be misused. Kwj2772 (talk) 15:28, 6 May 2013 (UTC)[reply]
 Support --Tobias1984 (talk) 14:01, 16 May 2013 (UTC)[reply]
 Support It can also be applied to weapons, missiles, etc.--Micru (talk) 22:01, 22 August 2013 (UTC)[reply]
Pictogram voting comment.svg Comment I suppose just about any kind of vehicle could use this as well. Description has been updated to reflect wider usage and proposal moved to generic transportation. Joshbaumgartner (talk) 00:45, 27 October 2013 (UTC)[reply]
 Support Macadamia1472 (talk) 22:57, 18 October 2013 (UTC)[reply]
✓ Done as vehicle range (P2073). Josh Baumgartner (talk) 05:49, 16 September 2015 (UTC)[reply]

Sales certified by

   Not done
DescriptionSales certification status for creative works album and single by certifying authority and level
Data typeItem
Template parameter"certification" in en:template:infobox single
Domainalbum (Q482994) album, single (Q134556) single
Allowed valuesgold, silver, platinum, multi-platinum--must have qualifiers for certifying authorities
ExampleA Love Supreme (Q5106) => RIAA certification (Q2503026)
Sourceexternal reference, Wikipedia list article (either infobox or source)
Robot and gadget jobsBots can import from en:Category:Albums by certification, etc.
Proposed byJustin (koavf)TCM

Common and definable structured data for music releases. Also actually stored as a database from most certifying authorities. —Justin (koavf)TCM 16:43, 18 September 2014 (UTC)[reply]

Changed datatype to item.--Micru (talk) 10:13, 19 October 2014 (UTC)[reply]
Justin can you complete the example by adding the item for the Sales Certification as I don't understand this. Should it be Sales Certification Authority? Filceolaire (talk) 21:48, 31 October 2014 (UTC)[reply]
@Filceolaire: Sorry. I don't understand. Can you reword this? —Justin (koavf)TCM 05:49, 14 November 2014 (UTC)[reply]
Justin I've editted the Example above. "=>" represents your proposed 'Sales Certification' property. Qnnnn represents whatever Qitem this property would link to in this example. Can you change Qnnnn to the correct Q number for the Sales Certification for A Love Supreme. Hope this helps. Filceolaire (talk) 19:18, 14 November 2014 (UTC)[reply]
@Koavf: Please respond. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:59, 6 March 2015 (UTC)[reply]

@Pigsonthewing: and @Filceolaire: I don't understand still. I guess I'm stupid. —Justin (koavf)TCM 06:44, 7 March 2015 (UTC)[reply]

@Koavf: ;-) Please see the example given in the "lyrics derived from" section, below, and emulate that. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 07:56, 7 March 2015 (UTC)[reply]
@Pigsonthewing: I guess this is it. —Justin (koavf)TCM 16:34, 7 March 2015 (UTC)[reply]
  • Thanks @Koavf:.  Support. Note that I have changed the property name to better match my understanding of what this property does. Filceolaire (talk) 20:29, 7 March 2015 (UTC)[reply]
  • What about making this a more generic "certified by", so it can be used for things like newspaper circulation and TV viewing figures? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:29, 22 March 2015 (UTC)[reply]
  •  Oppose @Koavf: this is not properly modeled.

Assuming that sales level designations are comparable across certification authorities, you need to:

    • create items to represent the different levels, eg Double_platinum currently redirects
    • create a property "Sales level" (alias "Sales certification") that will use the different levels
    • create a property "Certified by" to be used as qualifier on "Sales level" and hold the certifying agency

--Vladimir Alexiev (talk) 10:15, 25 March 2015 (UTC)[reply]

Pictogram voting question.svg Question is not it just a question of sourcing ? It would not be cool to duplicate the referencing property. It could also be a question of kind of award : there could be several items gold (description, given by ...) or use a property delivered by like any other award. TomT0m (talk) 20:57, 10 June 2015 (UTC)[reply]

 Not done Josh Baumgartner (talk) 07:22, 20 September 2015 (UTC)[reply]

first publication / last publication

   Not done
Descriptiondate or point in time when the first issue, instalment or episode in this series was published, broadcast or otherwise made available for the first time / date or point in time when the last issue...
Data typePoint in time
Template parameterfirst_aired and last_aired in en:Template:Infobox television and several others
Domainserialised work (Q386724)
ExampleDallas (Q168708) → April 2, 1978 and May 3, 1991
New York Herald Tribune (Q532494) → March 19, 1924 and April 24, 1966
Sourceexternal reference, Wikipedia list article, etc.

This would be a property pair that would work as an equivalent of publication date (P577) for serialised works, i.e. newspapers, periodicals, television series etcetera.

At the moment we are using a mix of start time (P580) and inception (P571) and their end date equivalents as well as publication date (P577), neither of which is satisfactory. P580 is too general and better suited as a qualifier. P571 is more appropriate for organisations, as works are generally not "founded" at the date of first publication. In the case of newspapers, it is likely that they are founded before they publish their first issue. Using P577 would imply that the entire series was published at once.

This will be very useful for all sorts of lists, infoboxes and queries. It is also a crucial quality that is used in many catalogues for periodicals etcetera, so we need clearly defined properties for this purpose. Väsk (talk) 13:47, 30 May 2015 (UTC)[reply]

Help me understand why start time (P580) cannot be used for this purpose? MSGJ (talk) 22:24, 10 June 2015 (UTC)[reply]
Basically the same reason we use publication date (P577) instead of point in time (P585). start time (P580) is useful, but not ideal, as it isn't specific enough and somewhat wrong. The description says it expresses when "an item begins to exist", although most works begin to exist before they are published. Conversely, end time (P582) expresses when items cease to exist, but works typically don't cease to exist after the last instalment is published. These properties are also used as qualifiers to express when a statement begins and ceases to be true, which is problematic. Väsk (talk) 22:24, 12 June 2015 (UTC)[reply]
 Oppose: publication date (P577) is inherently the date of first publication for most applications. In cases where final publication/release/etc. is important, use significant event (P793) with an appropriate item ("Final episode broadcast" or what have you) with date qualifiers. Josh Baumgartner (talk) 16:45, 12 June 2015 (UTC)[reply]
Using publication date (P577) to mean two different things sounds like a bad idea. It is not clear that the date the first instalment was published should be regarded as the main date of publication, either from general reasoning or that property's description. In fact, back when many books (including major reference works) were first published as individual booklets, the publication date would often be the date when the last booklet was published and they became available as bound volumes.
@Väsk: If a book was published once as an individual booklet, and again later as part of a larger work, each of those editions should have its own item, and the publication date (P577) for each would indicate the date of publication for that specific edition. Josh Baumgartner (talk) 18:07, 14 September 2015 (UTC)[reply]
As for significant event (P793), that is fine for values that are unusual or hard to generalise and prone to ambiguities. In this case specific properties are preferable if we want to maintain a controlled and useful vocabulary. In addition, solutions that depend on qualifiers will limit the potential of queries for the foreseeable future. Väsk (talk) 22:24, 12 June 2015 (UTC)[reply]
I agree that significant event (P793) is a bit of a catch-all for situations that are not otherwise cleanly handled by existing properties, and I also understand that for the time being, accessing that data may not be supported by the interfaces currently available. The software will get better and more comprehensive though. Josh Baumgartner (talk) 18:07, 14 September 2015 (UTC)[reply]
  •  Support but I prefer the name "publication date of first episode/installment" and "publication date of last installment/episode". "Last publication" would get used by people for the date of the latest edition of a book. Filceolaire (talk) 22:10, 18 June 2015 (UTC)[reply]
 Oppose agree with Josh Baumgartner here. By definition, publication date (P577) is date when a work was first published (became public). There is a publication date (P577) for a work, and one for each of its editions, which allows for every case. :) --Hsarrazin (talk) 12:54, 9 September 2015 (UTC)[reply]

 Not done Josh Baumgartner (talk) 07:22, 20 September 2015 (UTC)[reply]


   Not done
Descriptionindicates that the Creative Work contains a reference to, but is not necessarily about a concept. It can be used for cited authors and books. Useful especially for Wikisources.
Same as property "mentions":
Data typemonolingual-invalid datatype (not in Module:i18n/datatype)
Allowed valuesCreative Works, Persons
ExampleStoria della letteratura italiana (Q3974134) => Dante Alighieri (Q1067)
" => Brunetto Latini (Q366328)
" => The Divine Comedy (Q40185) ...
Robot and gadget jobsOn Italian Wikisource there are templates for Author and Texts Cited. They could either be linked to Wikidata or we could run a bot.
Proposed byAubrey (talk) 11:17, 8 June 2015 (UTC)[reply]


This proposal was probably made two years ago, but I can't find it in the archive. I place it here, but could be easily moved to Sister Project Property proposal, as my case is Italian Wikisource (but easily scalable). Wikilinks in Wikisource (see example: s:it:Storia della letteratura italiana (De Sanctis)/I) are cited authors and texts that are already present in the same Wikisource. In Italian Wikisource, for example, we use dedicated templates: s:it:template:AutoreCitato e s:it:TestoCitato. With those, we know which Authors and Texts are mentioned by a certain chapter/page/book. But it also goes the other way: these templates generate automatic categories that count all occurences in which an Author or a Text has been mentioned (see example: s:it:Categoria:Testi_in_cui_è_citato_Dante_Alighieri. Aubrey (talk) 11:17, 8 June 2015 (UTC)[reply]

The example you provide seems to be one for a "subject" relationship and we already have main subject (P921) (depending on the language the label might discourage multiple values or not) which is used e.g. to link biography article items like Ellis, John (1643?-1738) (DNB00) (Q19094241) to the item John Ellis (Q15072636) for the person. How would the proposed property differ from that? If it is intended to (be able to) record all references contained in the object, wouldn't the value items then always be of class "work"? -- Gymel (talk) 19:23, 8 June 2015 (UTC)[reply]
It's not a "subject" because it's more about "references". But references can be simply in the bibliography, or in the footnotes, but on Wikisource we use it tagging Works and Authors directly in the text. So it's not just "works", but also people. Of course, we can discuss it and maybe decide to have multiple properties (one for people, one for works), but the idea is that a CreativeWork can mention other works or other people. I mostly think about books, articles, essays, where I find this property to be very useful. But I understand that if you intend "mention" in the broadest meaning it's also a "rabbit hole" and you could concepts, ideas, things, etc :-) . --Aubrey (talk) 10:42, 10 June 2015 (UTC)[reply]
We have based on (P144) and inspired by (P941) which might be useful in this context (you refer to Wikisource, so the domain of your proposed property would be a specific edition of a certain work, and P144 and P941 relate this work to other works). A work or edition may contain bibliographic references (general ones or for specific quotes in the text) which could be modeled as Wikidata items for the article, "book" or edition cited, so the range type of your property would be wikidata items of a certain class, qualified with the usual properties for "references" in the lines of Wikidata:List of properties/Others#Source properties and probably one needs a new property to specify the location in the source item where the reference can be found (stating "contains reference" to "Odyssée" somewhere in an 500 page text will yield further questions like "why?" or "in what context?" the wikidata item cannot answer so it will have to point to a more specific point like page or chapter in the source). This narrow definition would be quite interesting in itself, allowing citation analysis in a Google Scholar style or performed by the commercial venues calculating impact factors. But we still would have to create zillions of items first before becoming a substitute for that. For certain Wikisource editions, depending on how explicit the references are already marked, it will still be much work.
You are probably thinking in the lines of a (TEI encoded) "scientific edition" of a text (travel journals, diaries, letters, ...), where the editors tag "noteworthy" places, concepts, persons, works and whatever to build an expensive index, providing short explanations to these at a central place. The terms are manifest in the work, but the choice of those used for indexing is done "subjectively" by the editors with the needs of the specific edition in mind (and locating these mentions on the work level is necessarily coarse, whe could say "Chapter 1" or "1797-12-01" but cannot provide page numbers). I think it would be a perfectly good idea to provide the indexing terms in a specific electronic edition with their corresponding wikidata id's (TEI-P5 allows creating thesaurus-like data islands for concept detached from the actual wording and location in the text and allows linking to quite general "Authority files" IIRC). If I imagine now that these then are "slurped in" to the hundreds into a single property of the relevant Wikidata item for the work or edition: This would be an incoherent swamp of "everything is related to everything". E.g. Darwin probably desrcibed thousands of animals and places on his journeys and an edition of his journals can cope with that in a meaningful way by keeping the connection between the animal, the place the animal was spotted, the day this happened and so on. You really would need all that data to make something useful of it. To have that in Wikidata one probably would have to clip the text into the smallest possible units, each of them represented by an individual wikidata item and additionally provide any instance of the proposed property with additional qualifiers depending on the specific editorial circumstances of the work. But I really doubt that you can completely import all data created in the process of a scientific edition into wikidata. And I doubt that it is really necessary: For a specific bird it is probably easy to state that it was first described by Darwin on a certain day at a certain place and give the edition as a usual reference. -- Gymel (talk) 08:00, 11 June 2015 (UTC)[reply]
Wow, this is a really good feedback: thank you for your time in writing this.
You make really good points: yes, the "mentions" statement on Wikidata would be full of claims, and would be probably not sustainable. I was actually thinking (from my (Italian) Wikisource point of view) of mentioning just "authors" and "works", as we already are tagging texts with the specific templates I mentioned.
Maybe a "bibliography" property would be even more narrow, but at the moment Wikidata is only accepting "works" as books, not scientific articles (as far as I know: at least, there are not many on Wikidata, as there are not many books).
Now, I'm a bit confused: I still think that this kind of statements would be very interesting and useful (we could, potentially, create en:Project Xanadu and reconstruct the hypertext of literature), but it's true that if we do not set boundaries we'd flood Wikidata with items and statements.
I started from Wikisource because these kind of works are easily linkable, and we can match them with proper "authors" and "works" present on Wikidata. Would that be too much too? Aubrey (talk) 20:50, 11 June 2015 (UTC)[reply]
See my first long paragraph above: Wikisource items from e.g. Dictionary of National Biography, 1885–1900 (Q15987216) or Allgemeine Deutsche Biographie (Q590208) are already represented here as articles (to be precise as editions/variants of the articles as work since they also exist as print and online in the context of digitization projects independent of Wikisource). Any other citation needed in the reference data for any claim for any item here should be represented by an item for the book, article, dictionary entry whatever. For an example of an journal article cf. Cinquième complément à l’analysis situs (Q18198988) (which is used as a reference for a claim in Poincaré conjecture (Q203586)). I just edited it heavily, and still have doubts about the double usage of published in (P1433), and that the whole Wiki univese does know nothing about the reporting year (? I mean the one associated to the volume number and which more often than not differs from the acutal year of publication) is simply outrageous...
To provide meaningful page numbers or chapter names or whatever is necessary to provide a meaningful reference the item should represent a specific edition (linked to the scientific work as more abstract concept), but this may be implicit if there is only one edition or for citation purposes a printed version and a parallel online version can be combined in one reference statement. The item for the book or article then again is connected with an item for its author(s), so your hypothetical "bibliography" property would have just one Wikidata item, namely that of the referenced article as value. Like any other "reference" the property would be accompanied by a constraint of volume (P478), page(s) (P304), section, verse, paragraph, or clause (P958) and/or reference URL (P854) and perhaps edition number (P393) as a not really clean shortcut to avoid creating specific edition items (different editions may have different number of pages, different publication years, different publishers, different ISBNs, National Bibliography numbers and what not - to put them all into a single item for the "book" will create something unintelligble). Your property in the general case will additionally need parallel versions of the some of these constraint properties to indicate where in the source item the reference can be found (on a certain page within a certain volume, or in connection with a specific footnote number, within or at the end of a certain chapter / entry, at the end of the whole volume or work). In your specialized case of smaller articles in Wikisource I suppose that references are concentrated at the end of each article and can easily be spotted and therefore these hypothetical extra qualifying properties won't really be needed. So in the simplest case it would boil down to
⟨ Q19141228  View with Reasonator View with SQID ⟩ references Search ⟨ Q18198988 ⟩
⟨ Q19141228  View with Reasonator View with SQID ⟩ references Search ⟨ Q18198988 ⟩
page(s) (P304) View with SQID ⟨ "100" ⟩
. -- Gymel (talk) 23:00, 11 June 2015 (UTC)[reply]
I'm afraid I too  Oppose this property. There is however some discussion of using wikidata to work with references on other wikis. Enter a book in a 'reference' template and it would automagically link to the corresponding wikidata item, fetching the author names, dates etc. leaving you to just fill in the chapter page and verse. This would mean that all references like this could be machine readable and someone smarter than me could analyse them. This would, of course, only pick up the references. Mentions would be identified by wikilinks (at least for the first mention). It is possible that redlinks (no local wiki-article) could link to wikidata items making these analysable too. None of this, as I see it, needs new properties on wikidata. It is all about clever things happening on wikisource and will likely take another year or more before it works. Hope this helps. Filceolaire (talk) 22:04, 18 June 2015 (UTC)[reply]

 Oppose use main subject (P921). --Succu (talk) 22:11, 6 August 2015 (UTC)[reply]

 Not doneJosh Baumgartner (talk) 07:22, 20 September 2015 (UTC)[reply]

BLDAM object ID

DescriptionObject ID to identify cultural monuments in Brandenburg, Germany
Data typeString
Domaincan be used for cultural heritage monument in Germany (Q11691318)
Allowed values00000000 - 99999999, see,,;sort=ort,