Wikidata:Property proposal/possibly reproduction of

image may represent

Originally proposed at Wikidata:Property proposal/Commons

   Withdrawn

Description	image should be using "digital representation of" (P6243), but it's not entirely clear if the value used here is the correct one
Data type	Item
Example 1	File:'Quarry, Pontoise' by Camille Pissarro, c. 1874.jpg] → The Quarry, Pontoise (Q21673134) (likely yes)
Example 2	File:'Elizabeth Boott Duveneck' by Frank Duveneck, Cincinnati Art Museum.JPG → Elizabeth Boott Duveneck (Q46956757), Portrait of a Lady (Elizabeth Boott Duveneck) (Q28788745), Elizabeth Boott Duveneck (Q46959040) (not clear which one and if there might not be duplicate items)
Example 3	File:'Ducks and Turkeys' by Edvard Munch, 1913.JPG → Ducks and Turkeys in Snow (Q18889772) (possibly part of a series)
Example 4	File:L'empereur Napoleon III de Franz-Xaver Winterhalter.jpg → <new item for "L'empereur Napoleon III"> (one of several known copies, original thought to be lost according to c:Category:Emperor Napoleon III in Coronation Robes (Winterhalter))
Example 5	File:'Cloud Study', by John Constable, 1822, Tate Britain.JPG → some of Special:Search/Cloud Study painting by John Constable -Yale or Special:Search/Cloud Study painting by John Constable
Planned use	for c:Category:Paintings without Wikidata item: temporary statement
See also	digital representation of (P6243), Wikidata:Property_proposal/detail_of_painting

Motivation

To sort through the 88000 files of c:Category:Paintings without Wikidata item, it should be possible to determine that we might already have an item, but it's not sufficiently certain to add a statement. Accordingly, a temporary statement could be helpful. Above a property for that. A specialist might more easily determine the correct statement. Please add more samples above. If the samples are resolved in the meantime, please don't remove them. (Add your motivation for this property here.) --- Jura 10:18, 4 August 2019 (UTC)

Discussion

•  Oppose We need a more generic mechanism for items that have been identified (possibly by machine) as "best guesses", but require confirmation or assessment. Rather than to create a bespoke property for the present niche application, it would be good to agree a more general way to indicate this. One approach used on wikidata is to add the statement but with deprecated rank and reason for deprecated rank (P2241) = unconfirmed (Q28831311). Other ways of signalling a provisional nature of a statement could also be envisioned. Jheald (talk) 18:25, 5 August 2019 (UTC)
  • Interesting thoughts, though I don't quite see the disadvantage of solving the problem now, even if the solution may only be temporary (it's meant to be an temporary statement anyways). Otherwise we might end up like the guys with some other database field: they keep discussing the optimal solution, but x years down the road, their field is still a mess. --- Jura 11:18, 7 August 2019 (UTC)
•  Oppose Having a bespoke solution for this use-case seems problematic and would invite a lot of other bespoke solutions for similar cases. Using the depricated rank as Jheald suggests seems to be the best with the tools we have currently, but it's still cumbersome. Maybe we need a 4th rank type for this (depricated/uncertain/normal/preferred)? ChristianKl ❪✉❫ 13:21, 8 August 2019 (UTC)
  • Why would this be problematic? There is no numeric limit on the number of properties we can create. --- Jura 13:30, 8 August 2019 (UTC)
    • While there's no numeric limit on the number of properties, users have to learn about properties to be able to use them appropriately. Increased complexity makes it harder to interact with Wikidata. ChristianKl ❪✉❫ 16:15, 26 August 2019 (UTC)
      • I don't think many users understand easily how the cumbersome alternative works. Even the ones that get seem to find it cumbersome. Anyways, personally, I recall which images I skipped .. if it isn't stored, others might just need to re-do the same. --- Jura 07:43, 9 September 2019 (UTC)
•  Oppose I am slightly against the solution. I am slightly against the fact of covering uncertainty by structured data. For this, we have categories on commons and I think they are enough. And to cope with the uncertainty of which Wikidata item is the right one, it might be better to have this as a qualifier for digital representation of. --Juandev (talk) 20:08, 29 September 2019 (UTC)
•  Comment withdrawing this. If Commons users prefer to continue with categories, we don't want to prevent them from doing that. --- Jura 16:46, 6 November 2019 (UTC)