Wikidata:Property proposal/values for this type
Jump to navigation Jump to search
values for this type
Originally proposed at Wikidata:Property proposal/Generic
- Support David (talk) 10:50, 29 April 2018 (UTC)
- Comment I think if a class item has some quantity property, you can already now infer that all instances of this class have the same quantity value. For example, credit card (Q161380) has the statement , therefore all subclasses and instances thereof have the same length and thus properties for this type (P1963) is redundant. --Pasleim (talk) 12:20, 7 May 2018 (UTC)
- Comment it is not the case the property values of classes are propery values of their instances. This is a common error in modelling that has been debunked many times over.
- Consider, for example, the USS Carter Hall (Q2468798). It has a length of 610 ft (not represented in Wikidata, though). It is an instance of Harpers Ferry-class dock landing ship (Q1586054) (also not represented in Wikidata).
- However, Harpers Ferry-class dock landing ship (Q1586054) does *not* have a length of 610 ft or indeed any other length. How can it have a length, after all - it is not a physical object. Instead it may have a design length (maybe even of 610 ft). English Wikipedia actually gets this bit right, having a general characteristics section of the InfoBox for https://en.wikipedia.org/wiki/Harpers_Ferry-class_dock_landing_ship, with a length of 609 ft 7 in.
- Harpers Ferry-class dock landing ship (Q1586054) does have some property values for properties that make sense for its instances. For example, it has a short name of Harpers Ferry, which it definitely not a short name for the USS Carter Hall (Q2468798). Peter F. Patel-Schneider (talk) 15:57, 7 May 2018 (UTC)
- @Peter F. Patel-Schneider: I thought the same but @Pasleim: can be right: I can't find examples of property values that are true for the class and false for the instances/suclasses. As regards your example, I think that if all instances of a class share some property value, then the class should have that property value, too: if all known stars have a temperature higher than 6000 C (I don't know if it's true), then also the star class should have such a property value.--Malore (talk) 00:08, 9 May 2018 (UTC)
- Oppose This proposal only applies to quantitative values; item or string or other typed values may also apply through to subclasses & instances. I think we need a more general discussion of how this ought to work in wikidata - perhaps a general qualifier: is there a technique with our current WDQS system that would facilitate this at all in some way? Pinging the ontology project ArthurPSmith (talk) 14:16, 7 May 2018 (UTC)
- Oppose This proposal needs more clarification. For example, what does "should have" mean here. Peter F. Patel-Schneider (talk) 16:01, 7 May 2018 (UTC)
- Oppose This proposal adds to the representational power of Wikidata. Either there needs to be some Wikidata process for adding these values to instances, or every consumer of Wikidata needs to be changed to look for this information when it is looking for property values. Peter F. Patel-Schneider (talk) 16:01, 7 May 2018 (UTC)
- Oppose The example could be expressed as JakobVoss (talk) 09:18, 16 June 2018 (UTC) . People and machines thinking in formal logical might have a problem with this statement but it makes more sense and is more easily to understand for anybody else. --
Notified participants of WikiProject Ontology
- Oppose. I agree with JakobVoss here. If there is a (default) value which applies to everything belonging to a class of things represented by the item, then we should just put a main statement on the class item itself. Deryck Chan (talk) 11:56, 2 July 2018 (UTC)