Wikidata talk:WikiProject Ontology

From Wikidata
Jump to navigation Jump to search
SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 7 days. For the archive overview, see Archive/. The latest archive is located at Archive for 2024.

metasubclass of (P2445) misuse

[edit]

See https://www.wikidata.org/wiki/Property_talk:P2445#misuse_of_this_property for an analysis of problems with metasubclass of (P2445). As stated there, I'm going to remove the incorrect links in the near future. Peter F. Patel-Schneider (talk) 16:24, 21 July 2024 (UTC)[reply]

Done near the end of July 2024. Peter F. Patel-Schneider (talk) 08:35, 12 August 2024 (UTC)[reply]

instances of instances of physical object (Q223557)

[edit]

I created Ontology Problems/instances of instances of physical object to note problems with instances of instances of physical object and motivate an upgrade to the Wikidata constraint system. Peter F. Patel-Schneider (talk) 08:49, 12 August 2024 (UTC)[reply]

I'd support a constraint to prevent instances of instances -- where's the best way yo propose it? Fgnievinski (talk) 20:04, 6 September 2024 (UTC)[reply]
@Fgnievinski Instances of instances of physical object (Q223557)? That would be great but I don't know what constraint could be used for that. I'm not sure where to propose new kinds of constraints. Peter F. Patel-Schneider (talk) 16:38, 25 September 2024 (UTC)[reply]
I've just found the place to propose property constraints is the Phabricator, creating a new task and adding the following task as a parent: https://phabricator.wikimedia.org/T213803 -- see also Help:Property constraints portal. I'd suggest running the draft through an RFC to gain constructive feedback for improvements before submitting the proposal, though. Fgnievinski (talk) 17:31, 25 September 2024 (UTC)[reply]

aircraft engine and similar classes that are supposed to contain physical objects but instead contain models

[edit]

I am trying to get some activity on fixing problems with classes like Q743004, which is a subclass of Q223557 but has engine models as instances. See https://www.wikidata.org/wiki/Wikidata:Project_chat#aircraft_engine_(Q743004)_and_models_(also_many_classes_similar_to_aircraft_engine_(Q743004)) for more information. Peter F. Patel-Schneider (talk) 14:36, 19 September 2024 (UTC)[reply]

I was thinking that Wikidata SPARQL query equivalent (P3921) could be used to retrieve instances of classes like plumber (Q252924). For more information and discussion please see and use Property_talk:P3921#use_on_classes_like_plumber_(Q252924). Peter F. Patel-Schneider (talk) 16:39, 25 September 2024 (UTC)[reply]

very high number of anti-patterns in chemistry domain

[edit]

Out of 2,232,196 subclasses of chemical entity (Q43460564) and 1,173,130 instances of chemical entity (Q43460564), 815,715 are both instances and subclasses of chemical entity (Q43460564). 812,338 of these have no superclass that is both an instance and a subclass of chemical entity (Q43460564).

This is a very high rate of anti-patterns. What can be done to reduce this number and keep it down? Peter F. Patel-Schneider (talk) 19:25, 27 September 2024 (UTC)[reply]

Get a consensus from the Wikidata:WikiProject Chemistry people? I think the bulk of these have been entered by bots, so maybe finding those bot authors and discussing the right way to do this with them would also work. There may be some third party users of this data however who depend on how it is now for some reason, so they should also be helped to fix things. ArthurPSmith (talk) 21:04, 27 September 2024 (UTC)[reply]

order bounds for instances of variable-order class (Q23958852)

[edit]

Some potential instances of variable-order class (Q23958852) have bounds on their order, e.g., disease (Q12136), which has (or should have) instances that are non-classes or first-order classes. It would be useful to be able to document these bounds, i.e., saying that disease (Q12136) is a variable-order class (Q23958852) of order 1 to 2. Is there a property that can be used for this purpose as a qualifier on the instance of (P31) statement? I took a look and there doesn't seem to be a property that is an ideal fit, but lower limit (P5447) and upper limit (P5448) look like they might be usable. Does this seem permissible or should I try to create new properties for this purpose? Peter F. Patel-Schneider (talk) 12:51, 2 October 2024 (UTC)[reply]

I think this is a good related example: Wikidata:Property_proposal/exponent_of_base_unit. Fgnievinski (talk) 23:42, 2 October 2024 (UTC)[reply]

ontology queries wanted for benchmarking project

[edit]

I am starting work in small project to benchmark query engine performance on useful Wikidata queries. I am asking for queries from the Wikidata user community to potentially be part of the benchmark. If you are a user of any Wikidata SPARQL service please send queries that you find useful to my email. Say what you used the query for and whether you would like to be noted as the source of the query. Queries that take considerable time or time out or query parts of the Wikidata ontology are especially welcome, particularly if the query caused you to switch from the official Wikidata Query Service to some other service. More information about the project is available in Wikidata:Scaling_Wikidata/Benchmarking. Peter F. Patel-Schneider (talk) 13:27, 1 November 2024 (UTC)[reply]

is there really no better way of marking this? Instance of building, instance of "civil registration"

[edit]

Is https://www.wikidata.org/wiki/Q3824913 really classified well?

Currently this building ends being classified as an object that exists outside physical reality. Is it really correct?

en: civil registration (system of registration of basic vital records such as birth, marriage and death) [1]

en: catalogue (enumeration of items systematically arranged for a specific purpose, usually with brief descriptive information included in each entry) [2]
en: knowledge organization system (generic term used in knowledge organization about authority files, classification schemes, thesauri, topic maps, ontologies etc.) [3]
en: conceptual system (system composed of non-physical objects, i.e. ideas or concepts) [4]
en: abstract entity (entity that exists outside physical reality, including abstract objects and properties) [5] this was unexpected here as it indicates an object that exists outside physical reality !!!!!!!!!!!

I initially listed it at User:Mateusz Konieczny/failing testcases - though at least @Swpb: (who helped to fix a loooooooot of bogus ontologies, thanks for enormous help) has no idea how to fix it - see https://www.wikidata.org/w/index.php?title=User:Mateusz_Konieczny/failing_testcases&diff=prev&oldid=2232222077

For me it looks clearly wrong. Is it possible to at least mark it as a case of entry mixing different topics?

Mateusz Konieczny (talk) 18:39, 7 November 2024 (UTC)[reply]

You are correct. Antiguo hospital de Almoradí (Q3824913) is a building (Q41176). It is not a civil registration (Q83708009), nor is it a juez de paz (Q16586086). Both these statements should be removed. Perhaps there is some other relationship that could replace them, but I leave that up to others. Peter F. Patel-Schneider (talk) 19:00, 7 November 2024 (UTC)[reply]
Yes, you and Peter are correct. I mis-categorized this article. All fixed. Swpb (talk) 20:00, 7 November 2024 (UTC)[reply]

A former state both a material and immaterial entity

[edit]

see the classification.js report

The upper class common issue is geographic region (Q82794) (see the far more succint query report, which is both material and immaterial, and mostly at the end of the day I'm not sure if region of space (Q26713767) is a class of immaterial stuff ? author  TomT0m / talk page 15:22, 24 November 2024 (UTC)[reply]

I don't think geographic region (Q82794) should be a subclass of region of space (Q26713767). A geographic region is tied to a globe (the Earth) which is not in a fixed position in "space" (there really isn't such a thing as fixed positions anyway) so it wouldn't be identifiable as a "region of space". ArthurPSmith (talk) 18:01, 25 November 2024 (UTC)[reply]
If a geographic region doesn't qualify as a region of space, then what would qualify - only outer space, void of any masses? What's needed is a specification of a frame of reference. A geographic region is a region of physical space in an Earth-centered, co-rotating frame. Fgnievinski (talk) 04:57, 26 November 2024 (UTC)[reply]
@Fgnievinski: That's a good point. I guess that was User:TomT0m's second question - whether region of space (Q26713767) was immaterial (as it is now classified). If it's defined in a mathematical or theoretical physics sense - a range of coordinate values in some conceptual space then it would be immaterial. Certainly given that region of space (Q26713767) is claimed to have boundary (Q875399) as a part, that's clearly a mathematical (immaterial) concept. region of space (Q26713767) in that sense wouldn't have any instances that were material objects, or subclasses that have material instances. However there clearly is a separate meaning which is a definable region of the actual universe in relation to things within that universe, and I think that would be, even without considering geographic regions, a material entity. And it looks like most of the usage of region of space (Q26713767) refers to that second meaning. The main problem in the chain is I think spatial region (Q124711484) which comes from BFO and is declared to be immaterial. Perhaps that shouldn't be in the subclass tree? ArthurPSmith (talk) 14:33, 26 November 2024 (UTC)[reply]
An abstract mathematical or geometrical region (Q130249688) has manifestation as a physical or real-world region of space (Q26713767). The materialization, realization, or reification is achieved by means of a frame of reference (Q184876). That was my rationale when I last edited these items. I've now made a few tweaks to make it more explicit Q130249688 is abstract and Q26713767 implies a Q184876. I didn't touch spatial region (Q124711484) because I didn't want to interfere with the BFO. As for the boundary issue, I've replaced it with spatial boundary (Q125176223). Fgnievinski (talk) 15:20, 26 November 2024 (UTC)[reply]
The situation of a former state could be resolved defining its period of existence. An analogy is a dead person: it was material when it was alive, now the memory of it is immaterial. Fgnievinski (talk) 05:01, 26 November 2024 (UTC)[reply]