Wikidata:Property proposal/identifies

identifies

Originally proposed at Wikidata:Property proposal/Generic

   Done: identifies (P10476) (Talk and documentation)

Description	class of entity this identifier identifies
Represents	identifier (Q853614)
Data type	Item
Domain	identifier (Q853614) (generally instances)
Example 1	YouTube video ID (Q110851517) → YouTube video (Q63412991)
Example 2	flight number (Q133663) → scheduled air service (Q15817877)
Example 3	International Standard Recording Code (Q1148336) → audio recording (Q3302947)
Example 4	VIAF ID (Q19832964) → <novalue>
Single-value constraint	yes

Motivation

I had no way of expressing the relationship between an item I made for an identifier and the thing that it identifies. Lectrician1 (talk) 02:52, 8 February 2022 (UTC)

Discussion

•  Support Mahir256 (talk) 03:02, 8 February 2022 (UTC)
•  Support Moebeus (talk) 03:27, 8 February 2022 (UTC)
•  Comment If nothing else I would think main subject (P921) works for this purpose? But I suppose a dedicated property isn't a bad thing here. ArthurPSmith (talk) 19:36, 8 February 2022 (UTC)
  • Agree. Domain should probably include property namespace (with samples from there). --- Jura 12:59, 9 February 2022 (UTC)

    @Jura1: I don't think this is necessary. The reason for this is that Wikidata properties can be related to a Wikidata item through Wikidata item of this property (P1629) and the Wikidata item is the one that would "identifies", if created. For example, you might want to use "identifies" in YouTube video ID (P1651), but I think that this could be structured by doing YouTube video ID (P1651)Wikidata item of this property (P1629)YouTube video ID (Q110851517) and YouTube video ID (Q110851517)identifiesYouTube video (Q63412991).

    In general, each property can have a Wikidata item counterpart and the property can be related to the Wikidata item through Wikidata item of this property (P1629), so no need to have the "Property" namespace in the domain of "identifies".

    Rdrg109 (talk) 15:45, 9 February 2022 (UTC)

    It may have been the plan, but the scope of the property is too vague. Accordingly, we ended up having applicable 'stated in' value (P9073) in addition.

    Only 15% of external-id properties seem to have a value that is an instance or subclass of identifier (Q853614). --- Jura 17:28, 9 February 2022 (UTC)

@ArthurPSmith: I don't find that main subject (P921) works as an alternative for this proposed property. --Gymnicus (talk) 10:15, 10 February 2022 (UTC)

•  Support There is currently no property for clearly storing this information -- Rdrg109 (talk) 15:48, 9 February 2022 (UTC)
•  Question what would be value for VIAF ID (Q19832964) (or VIAF ID (P214))? It currently doesn't have type constraint. --- Jura 17:39, 9 February 2022 (UTC)

@Jura1: In my opinion, this property can not or not meaningfully be used for all identifiers. It really depends on what can be proven with the identifier. To take your example, the VIAF ID (Q19832964) is an identifier of an authority file. (Perhaps this should also be indicated in the data object. Because, according to the data object, it is currently an online database (Q7094076)) In principle, almost everything that can be seen from the GND ID (Q54506313), for example, can have an entry in an authority file. For example, persons, subject terms, organizations and events are provided with entries in the Integrated Authority File (Q36578). What I mean to say is that when it comes to authority files, you either have to find an umbrella term or you have to list all the possibilities. The latter could sometimes be quite long and finding a generic term can also be difficult. With regard to the VIAF identification, the only possibility that I could think of was the statement VIAF ID (Q19832964)identifiesentity (Q35120). But does this really make sense? I would say that identifier for authority file are outside the project scope for this property, as are identifiers for lexicons, for example. --Gymnicus (talk) 10:15, 10 February 2022 (UTC)

• I think using <novalue> could do for VIAF (unless someone creates an item for equivalent class it identifies). Also, I think we would want a single value constraint on this. --- Jura 17:22, 3 March 2022 (UTC)

•  Support while I still have a question about using this property, I still consider it useful and support its creation. --Gymnicus (talk) 10:15, 10 February 2022 (UTC)
•  Question As Jura has already written above, there is currently only a data object for a small part of the identifiers. Is your goal now, Lectrician1, to also create a data object for each identifier, or would it perhaps make more sense to expand the area of application of the property? So that you can also use the property in the property area. I would definitely favor the second option. --Gymnicus (talk) 10:15, 10 February 2022 (UTC)
•  Oppose given the open questions about scope and values. --- Jura 07:09, 14 February 2022 (UTC)
  • fixed the proposal above. Should solve most open questions.
•  Support There are MANY things to say about an identifier itself !!!, which dictates that an item/concept for the identifier has to be created in Wikidata. Further than, we need this proposed dedicated property to link the "item/concept of an identifier" to the object it represents. Both (identifier and object of identifier) are then free to have many statements applied (issued by, length, type, official website, etc.) to accurately describe and provide disambiguation. --Thadguidry (talk) 17:38, 14 February 2022 (UTC)
•  Support —MasterRus21thCentury (talk) 18:16, 4 March 2022 (UTC)

@Lectrician1, Mahir256, Moebeus, ArthurPSmith, Jura1, Rdrg109: @Gymnicus, Thadguidry, MasterRus21thCentury:  Done identifies (P10476) Pamputt (talk) 10:35, 10 March 2022 (UTC)