Wikidata talk:WikiProject Ontology/Classes

@Peter F. Patel-Schneider: Started the page. If you want to add your contribution, don't hesitate. Snipre (talk) 21:59, 27 October 2015 (UTC)Reply

@Snipre: Will do shortly. I'm busy doing other things right now. Peter F. Patel-Schneider (talk) 22:11, 27 October 2015 (UTC)Reply

No problem. This is not a priority. Snipre (talk) 22:17, 27 October 2015 (UTC)Reply

Purpose of this page edit

What is the purpose of this page? Right now it seems to be just a list of some characteristics of classes. Peter F. Patel-Schneider (talk) 15:34, 2 December 2015 (UTC)Reply

The purpose of the page is 1) to list the general characteristics of the classes and 2) to define what will be used on WD. The main idea is to define first a glossary to be sure that everyone understand the term used in the discussions and definitions. Once we know what we use, we can start to say how we want to apply in WD. Snipre (talk) 23:05, 1 December 2015 (UTC)Reply
OK, I made some edits to clean up the terminology and provide some examples. Peter F. Patel-Schneider (talk) 15:34, 2 December 2015 (UTC)Reply
This page should better summarize best practice in Wikidata instead giving an introduction to ontology modeling in general. I added Wikidata:Classes as alias. -- JakobVoss (talk) 21:35, 28 October 2017 (UTC)Reply

What is a class? edit

class is any item? Wikidata:Glossary: Item refers to a real-world object, so, class is a real-world object? --Fractaler (talk) 14:49, 29 October 2017 (UTC)Reply

you truncated the quote from the glossary - an item is a "real-world object, concept, or event that is given an identifier " in wikidata. It may be a real-world object, or it may be an abstract thing. In general classes are abstract, so "real-world object" wouldn't apply. ArthurPSmith (talk) 14:52, 29 October 2017 (UTC)Reply
ok, an item is a real-world object, concept, or event? --Fractaler (talk) 14:56, 29 October 2017 (UTC)Reply
An Wikidata item is nothing but a Wikibase datatype. Items can be about real-world objects, concepts, and events among other things (we even have an item about Wikidata items: Wikidata item (Q16222597)). The question what an item can be about is not a ontological or philosophical question and should not be discussed at the help pages but it is only a question of notability guidelines in Wikidata -- JakobVoss (talk) 14:28, 30 October 2017 (UTC)Reply
So, an item is a real-world object, concept, or event or is not a real-world object, concept, or event? --Fractaler (talk) 14:39, 30 October 2017 (UTC)Reply
Can you explain a bit more the purpose of your question? Are you trying to improve the definition presented in the glossary? If that is the case I think I agree that it could be improved. In general maybe a better definition would be - an item is anything with a Q id - i.e. anything that meets Wikidata's notability guidelines. That means any topic for which there is a wikipedia page in any language, or anything else that may be structurally useful to describe such topics. Does that make more sense? Or are you trying to get at something else? Explain a bit more if you can thanks! ArthurPSmith (talk) 20:35, 30 October 2017 (UTC)Reply
The clarity, unambiguous definition of the term increases order. This term is one of the basic. I want to understand what "Wikidata item" is. The existing definition for "item" (real-world object, concept, or event) can clarify, classify, for example, what is (and for whom!!!):
  1. "https://www.wikidata.org/w/index.php?title=Q38791127&action=edit&redlink=1"
  2. "Q38791127"
  3. Q38791127;
  4. "https://www.wikidata.org/wiki/Q7566"
  5. "Q7566"
  6. parent (Q7566)
  7. "https://commons.wikimedia.org/wiki/File:Like_Father_like_Son.jpg"
  8. "Like_Father_like_Son.jpg"
  9. "File:Like_Father_like_Son.jpg"? Is there an item here and why? What is not a item here and why?
I think in order to understand what an item is, we need: 1) to understand Q29956845 (database (Q8513)? ontology (Q324254)? semantic wiki (Q638153)?knowledge base (Q593744)? Wikidata knowledge base?) in which 2) Q29956950 (object of database? Ontological object? object of semantic wiki? Semantic object? Wikidata semantic object?) being defined exists. Verification of the object's belonging to the space is done by removing the space. If the object did not disappear, then it was not in the given space. And vice versa. --Fractaler (talk) 06:49, 31 October 2017 (UTC)Reply
I would say that in a most literal interpretation of your question here, the "space" is Wikidata itself (or another installation of the wikibase software if referring to "items" in that context). Remove Wikidata and there are no "items" to talk about. What Wikidata items are *about*, i.e. what they conceptually represent, is potentially anything that can be represented (including real-world objects, concepts and events). Anything that exists in the universe or that could potentially be represented or conceived of in some fashion in this universe could have a Wikidata item. The things that actually have Wikidata items at any given point in time are those which the Wikidata community has considered to meet the guidelines for inclusion and which specific editors (or bots etc) have actually added to the database. Individual files (the last 3 items on your list) would generally not be considered eligible to have an item about them, but I suppose they could in certain cases. Wikidata item ID's or their URL's would not generally be considered eligible to have an item about them, but again there could be exceptions. The only thing in your list I would think is certainly always eligible to have an item as representation is "parent", as a concept representing a real-world relationship. ArthurPSmith (talk) 08:07, 31 October 2017 (UTC)Reply
but note, in an independent installation of wikibase with different rules (for example a database focused on books), "parent' might not be eligible for inclusion either. ArthurPSmith (talk) 08:10, 31 October 2017 (UTC)Reply
In in order to understand what an item is we don't need to understand any of the things you linked (Q29956845...) but only need to know how to edit Wikidata. If you want to discuss philosophy you better look out for philosophy communities instead of Wikidata. -- JakobVoss (talk) 12:43, 31 October 2017 (UTC)Reply
The database can be deleted. But what means "Remove Wikidata" (for test there are/there are no "items")? --Fractaler (talk) 08:30, 31 October 2017 (UTC)Reply
"parent", as a concept representing a real-world relationship. So, if we use removing tool, and delete parent (Q7566) - what is "concept"? It is object of what Q29956845? It is kind of Q29956845? Other? --Fractaler (talk) 08:42, 31 October 2017 (UTC)Reply
independent installation of wikibase so, "Wikidata item" is a object of Wikidata's wikibase, object of wikibase with Wikidata rules? --Fractaler (talk) 08:50, 31 October 2017 (UTC)Reply
Have I answered your question, or have I not? I'm not sure what more to say. Yes, a "Wikidata item" is an object in the wikibase system that is run by the Wikidata community, under the (far from perfectly clear) rules we have set up. It is also a representation, with an identifier (Qxxx) that stands in this system for something else which could in principle be anything, part of any universe real or imagined. ArthurPSmith (talk) 11:29, 31 October 2017 (UTC)Reply
@ArthurPSmith: It seems I understood how to formulate the question: can we make a link "class is any item" -> "class is any item"? --Fractaler (talk) 13:00, 21 November 2017 (UTC)Reply

@ArthurPSmith, JakobVoss:

  • I don’t agree that class is an item indeed, and I agree with @Fractaler:. The goal of Wikidata is not to talk about its items, it is rather to talk about the objects those items describe. This is known as « ontological realism ». I’d rather replace « class » in this document by « class-item ». A class-item is an item to refers to a class. Also the definition given here is a non definition of an instance relationship, it just says « an instance is an item who have an « instance of » statement. This is a problem as if I add a statement anywhere, even if obviously incorrect, this fulfills this definition who does not give any criteria to judge what is correct or not. Not any help at all for the user. A class in general is also not defined with its Wikidata instances, it’s defined by some definition of what its instances are. And its instances are not items … but objects, events or abstract concepts. Lets call « Wikidata instances » of a class the set of objects, concrete or abstract, who have an item on Wikidata for which there is a statement instance of (P31) with a class-item (or a wikidata-subclass of it) referring to that class. Wikidata-subclass of a class refers to any class for which there is a class-item with a chain of subclass of (P279) statements from the subclass-item to the class-item. This implies, as Wikidata is not complete, that a class may have many more instances than its Wikidata instances, this could be called the « class completeness problem ». The introduction seem to me an example of a use mention problem that we should avoid.
    • I’d rather have an introduction similar to what I put in User:TomT0m/Classification referring to the general purpose of classification and ontological taxonomy. Classification in Wikidata is just applied classification, it does not defines the field.
    But anyway, I’ll be bold and make right now corrections of the introduction. author  TomT0m / talk page 11:00, 3 February 2018 (UTC)Reply
  • I'd avoid rather academic jargon such as "We will define" and artificial words such as "class-item" and "instance-item" unless they are regularly used in discussion. Nevermind unless the page becomes too technical and philosophical. Better document how classification in Wikidata actually works instead of how we would like to have it in an indeal world. -- JakobVoss (talk) 17:56, 5 February 2018 (UTC)Reply

Value edit

Which? value from Wikidata (Q40218570) ? --Fractaler (talk) 13:23, 31 October 2017 (UTC)Reply

Good spotted. This item needs to be deleted. It is only used in items such as Category:Pages using Wikidata property P2632 (Q23306280) but gives no additional information. Use instance of (P31) and Wikidata property usage tracking category (Q24514938) instead. -- JakobVoss (talk) 13:13, 21 November 2017 (UTC)Reply

Class naming in singular edit

This page says that "Wikidata labels for classes should be singular, at least in English". I don't think this is codified anywhere yet, but there's a discussion on it at Help_talk:Label#Plural/singular. --Yair rand (talk) 05:05, 18 April 2019 (UTC)Reply

Major part of class tree partially detached? edit

I just happened to notice that structure (Q517966) is not a subclass of (P279) anything (and has never been, according to its revision history), just an instance of (P31) concept (Q151885); yet it was semi-protected in 2019 for being "one of the most used Wikidata entities" (which seems plausible, given that it's a parent class of group of living things (Q16334298), though the latter is also a subclass of (P279) group of physical objects (Q61961344), so the detachment perhaps isn't that significant). At the same time, it's said to be the same as (P460) social structure (Q211606) and has very few WP articles (in Dutch and German only).

I couldn't find a sociology or general social sciences wikiproject, otherwise I would probably have posted there. As sociology isn't really within my field of expertise, I'm hesitant to do anything about this myself, but I'd like to refer the matter somewhere for it to be resolved. Since there are German and Dutch articles for both, my guess is that it's distinct from social structure (Q211606) but still related to it somehow, such as a subclass of (P279) in one direction or another. In any case, I hope another parent class than entity (Q35120) can be identified. --SM5POR (talk) 10:08, 17 July 2020 (UTC)Reply

Return to the project page "WikiProject Ontology/Classes".