Topic on User talk:TomT0m

I noticed on Property talk:P2738 that you proposed a "disjoint" property. What happened to that proposal?

EDIT: Sorry it wasn't you who proposed the property.

I still think it's a not so good idea :) This property should do the trick in most if not all cases with less redundancy risks.

There are cases that disjoint union of does not cover. For example, how can one use it to state that dog and human are disjoint? There seems to be a need for some other way of stating disjointness. I will probably put together a request for comments, as an improvement of disjointness in Wikidata is more than just a property proposal.

It definitely should work if you use inferences, it applies transitively on the classes. The actual problem is that the taxonomy project does not want to use classes :/

It should be enough to say somewhere that their smallest common ancestor taxon is Boreoeutheria (Q132666) if I searched right. It has two child taxons that should be asserted disjoint. The human taxon is subclass of one, and the other on the other, see this query.

So dogs and humans can be asserted disjoint by a "disjoint union of". The solution could be to manually handle the taxonomy case in the gadget, for example. There also lacks a way to link items like human (Q5) and https://w.wiki/8Tcm to their taxon to do this.

If we don’t want this link, we can rely on a more casual life classification on Wikidata with the (casual) "animal" class, make (disjoint) subclasses for it like "humans" "wild animal" "domestic animal", make "dog" a subclass of "domestic animal" an we are done. We get that cattle is also disjoint with humans on the process.

Actually … See The disjoint union statement on the « animal » item. It’s already stated that human are disjoint with non human animals. So it’s enough that "dog" is a subclass of "non-human animal" directly or indirectly for my gadget to catch any problem with any irl werewolf :)

I only had to add one statement to solve this problem to state that domesticated animals are non-human animals : https://www.wikidata.org/w/index.php?title=Q622852&diff=prev&oldid=2028039516 

You can check with this templated query wether or there exist a "disjoint union of" statement that makes two classes anywhere in the tree disjoints : https://w.wiki/8TiA you can check on it that "dog" and "human" are disjoint and click on the link on the right to go to the statement.

You definitely not need to add a statement that "dog" and "human" are disjoint, then. And likely you can check that cows are also disjoint with humans with the same query.