Wikidata:Property proposal/fully divised into
fully subdivised into edit
Originally proposed at Wikidata:Property proposal/Place
Description | that location or object is fully subdivised into the parts given in qualifier |
---|---|
Represents | partition (Q518554) |
Data type | Item |
Domain | places, physical objects |
Allowed values | places, physical objects |
Example |
|
Amended on 1/1/2018 This proposal needs to create one additional qualifier « into ». don’t have a better idea. of (P642) could be used as a general purpose qualifier but linguistically this does not fit. And does seem like a details. Maybe the property could be renamed into something like « fully consists of » or something like that, but maybe this would not fit in other languages as these small arbitrary words in expression tends to be quite arbitrary. Maybe the whole idea of general purpose qualifier is broken because of this simple fact, and we should create property/qualifiers pair each time to let translators freedom of translation. author TomT0m / talk page 10:46, 1 February 2018 (UTC)
- Motivation
- The administrative division process is a known and studied problem (see for example this document in french - https://publications.polymtl.ca/832/1/2012_PierreDelaPoixdeFreminville.pdf ). Also administrative territorial entities can be quite complicated and there could be several kind of divisions that fully covers a division - electoral divisions can be different from other administrative divisions of the same territory. There may also be non administrative relevant divisions of a territory. This proposal can solve this by regrouping those into several statements. Under the same rationale that lead to disjoint union Search and disjoint union of (P2738) this allows to state that the list of divisions is complete. If that information is spread into multiple « has part » statement it’s more difficult to state that we know nothing has been forgotten (Maybe with …) author TomT0m / talk page 15:39, 3 August 2017 (UTC)
- Discussion
- Why isn't this under Generic property proposals? - Brya (talk) 06:33, 5 August 2017 (UTC)
- Comment I think the first usecase should currently be handled with contains the administrative territorial entity (P150).
--- Jura 13:24, 10 August 2017 (UTC)- @Jura1: This does not seem to state that the list is full, information in Wikidata may only be partial. author TomT0m / talk page 21:33, 29 January 2018 (UTC)
- I'm aware of that. Thus I don't think it's a reason to oppose this proposal. BTW, what qualifier would "into" (in the above sample) be?
--- Jura 22:30, 29 January 2018 (UTC)
- I'm aware of that. Thus I don't think it's a reason to oppose this proposal. BTW, what qualifier would "into" (in the above sample) be?
- @Jura1: This does not seem to state that the list is full, information in Wikidata may only be partial. author TomT0m / talk page 21:33, 29 January 2018 (UTC)
- Comment @TomT0m: 1. How exactly will you represent different groupings (electoral vs administrative)? Give a specific example. 2. "Fully" is a slippery thing. If a new subregion is created, or I delete one of the statements, does that make all the "fully" statements untrue? 3. Please be precise in your descriptions and examples! You use the P' template with non-existing props, Q' template with some strings where the result just doesn't parse reasonably, etc. Second time today I see a sloppy proposal from you, and this is just not like you, Tom! --Vladimir Alexiev (talk) 09:34, 3 October 2017 (UTC)
- @Vladimir Alexiev: « How exactly will you represent different groupings (electoral vs administrative)? » with different statements if relevant, one for the electoral division, one for the administrative. The subdivisions listed as qualifier of a single statement are not supposed to overlap. Plus we know the type of the subdivisions by their instance of statements, it’s enough. the P' template is designed to handle non existing props. This one is proposed.
- « does that make all the "fully" statements untrue » : No, it just adds an end date if a new administrative subdivision is added. Just as any statement in Wikidata. A new statement with the full list of divisions has to be created.
- the P' template is designed to handle non existing props. This one is proposed.author TomT0m / talk page 21:33, 29 January 2018 (UTC)
Oppose per Vladimir Alexiev and Jura.changed to Neutral − Pintoch (talk) 08:01, 25 October 2017 (UTC)- @Pintoch: Replied to them, as they did not oppose you may want to change your vote. author TomT0m / talk page 21:40, 29 January 2018 (UTC)
- As qualifiers, that sounds more promising to me - but what would be the main value of the statement? Would you create a specific item "values in qualifiers"? That looks weird to me. Maybe that could be an item that represents the type of the parts? And also this would require creating the "into" property - creating properties with such short names is a bit dangerous because they could be used in a not-so-controlled way in many other places… So changing to "neutral" for now, because I'm not entirely convinced. − Pintoch (talk) 21:49, 29 January 2018 (UTC)
- It already exists. list of values as qualifiers (Q23766486), used for union of (P2737) and disjoint union of (P2738) author TomT0m / talk page 21:53, 29 January 2018 (UTC)
- As qualifiers, that sounds more promising to me - but what would be the main value of the statement? Would you create a specific item "values in qualifiers"? That looks weird to me. Maybe that could be an item that represents the type of the parts? And also this would require creating the "into" property - creating properties with such short names is a bit dangerous because they could be used in a not-so-controlled way in many other places… So changing to "neutral" for now, because I'm not entirely convinced. − Pintoch (talk) 21:49, 29 January 2018 (UTC)
- @Pintoch: Replied to them, as they did not oppose you may want to change your vote. author TomT0m / talk page 21:40, 29 January 2018 (UTC)
- It would be great to have a way to state that a specific number of subparts form together the whole. I'm however not sure this is the right way to go about this. It might make more sense to have a qualifier that can be used to qualify contains the administrative territorial entity (P150) in a way that indicates that all of the parts together form the whole.
- Maybe something like:
- ⟨ Nantes (Q12191) ⟩ contains the administrative territorial entity (P150) [[Special:Search/Property:contains the administrative territorial entity (P150)|Search]] ⟨ canton of Nantes-1 (Q1726454) ⟩
part of full partition Search ⟨ "A" ⟩ - ⟨ Nantes (Q12191) ⟩ contains the administrative territorial entity (P150) [[Special:Search/Property:contains the administrative territorial entity (P150)|Search]] ⟨ canton of Nantes-2 (Q969444) ⟩
part of full partition Search ⟨ "A" ⟩ - ⟨ Nantes (Q12191) ⟩ contains the administrative territorial entity (P150) [[Special:Search/Property:contains the administrative territorial entity (P150)|Search]] ⟨ canton of Nantes-3 (Q599224) ⟩
part of full partition Search ⟨ "A" ⟩ - ChristianKl (✉) 21:59, 17 December 2017 (UTC)
- @ChristianKl: I don’t get it, "A" seem to be Nantes (Q12191) actually. However the number of parts could be stated with ⟨ Nantes (Q12191) ⟩ has part(s) of the class (P2670) ⟨ canton of France (Q18524218) ⟩for example. author TomT0m / talk page 21:40, 29 January 2018 (UTC)
quantity (P1114) ⟨ 10 ⟩- Thinking about it, It’s also possible tu use some class items as the main value of the statement. Assuming that « Canton de Nantes » is the class of all Nantes’ canton, it may be possible to state . The number of Nantes’ canton can then be stated by . Then we could check Wikidata is complete by counting the instances : Interesting stuff is that there is already a few results of known complete class ! Also @Pintoch: as he seem not to like the idea of having a dummy main statement value (a note on this, I initially proposed to use unknown value Help_in the proposal for « (disjoint) union of » — that would be perfectly meaningful here as this would mean « there is an entity like « Canton de Nantes » but nobody bothered to create it (or the creator of the statement did not found it » — someone suggested that using an item would be more informative for the lambda user as it allows to have a textual description). So we would have two ways of using the property. author TomT0m / talk page 11:24, 30 January 2018 (UTC)Try it!
select (count(?instances) as ?count) ?class ?number_of_instances { # this query computes all the classes Wikidata seems to know all the instances ?instances wdt:P31/wdt:P279* ?class . ?class wdt:P1114 ?number_of_instances . } group by ?class ?number_of_instances having (?number_of_instances = ?count)
- A follow up of this discussion : Wikidata:Project_chat#Instance_completeness (permalink : https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat&oldid=624744034#Instance_completeness ) author TomT0m / talk page 12:55, 30 January 2018 (UTC)
- Thinking about it, It’s also possible tu use some class items as the main value of the statement. Assuming that « Canton de Nantes » is the class of all Nantes’ canton, it may be possible to state . The number of Nantes’ canton can then be stated by . Then we could check Wikidata is complete by counting the instances :
- @ChristianKl: I don’t get it, "A" seem to be Nantes (Q12191) actually. However the number of parts could be stated with
- Interesting proposal. I think the inherent limitation at the moment is that qualifiers can only be given for statements, not for properties, as ideally, this would be captured by a generic qualifier on the respective property. Generic approach, because many other things than contains the administrative territorial entity (P150) can be complete, for instance child (P40), spouse (P26), languages spoken, written or signed (P1412), ... But no matter how this is modelled, maintenance is a challenge, when data changes, ideally, the user should be prompted to confirm that completeness still holds. Ls1g (talk) 22:20, 1 February 2018 (UTC)
- Some concerns: (1) what happens if it is not really full; (2) data duplication -- e.g., data values for fully_divised_into = data values for contains administrative territorial entity; (3) it is not generic, as also @Ls1g said -- e.g., what if we want to say "complete for all children". Fadirra (talk) 23:23, 1 February 2018 (UTC)
- Support at least Peninsular Malaysia (Q1973345) would need it. --Liuxinyu970226 (talk) 06:25, 4 February 2018 (UTC)
Not done Not enough support. Micru (talk) 08:30, 15 March 2018 (UTC)