[00:07:15] In some languages you may even need to call for their marital status! (re @NicolasVIGNERON: oh, like a first function would retrieve "label : Albert Einstein, gender : male" and pass it to some other function, interresti...) [07:07:35] Note that in Wikidata already you don't have to use "noun". "verb", "adjective" as the lexical category, but you could use "root" as the category, if that helps. (re @acanthamoeba_castellanii: Truthfully, I still think that the current lexicographical data on Wikidata might need to be improved further, because the lexic...) [07:08:49] I had disconnected it because I wanted to see if the compositions work. The compositions seem to work, it is just the test that doesn't work, it seems. It's good that you connected it again, I got distracted mid-debugging. Thanks! (re @u99of9: Oh, I unknowingly reverted @vrandecic who disconnected Z34047. My simplistic view is that it's the only implementation that [07:08:50] pass...) [07:10:36] *Abstract Wikipedia Newsletter #245* [07:10:36] * The Foundation's search for the perfect language: Review of an academic paper critically assessing the Abstract Wikipedia project [07:10:38] * News in Types: ‘Tis the seasons [07:10:39] * Fresh Functions weekly: 61 new Functions [07:10:41] https://www.wikifunctions.org/wiki/Wikifunctions:Status_updates/2026-04-25 [07:11:31] But if the top category becomes root, then is it possible for each lexical form under it to be labelled as "verb", "noun" etc? (re @vrandecic: Note that in Wikidata already you don't have to use "noun". "verb", "adjective" as the lexical category, but you could use "root...) [07:26:04] It is [07:28:28] but here's a thing I haven't understood: do senses refer to the root, or do they refer to nouns, verbs, etc.? If the latter, that makes it more difficult, because a Lexeme is basically a bundle of forms and senses, assuming that all senses are the same across all forms of the given lexeme, more or less. [07:32:42] I don't think it us a must for all lexical form to share the same sense. Each lexical form should be able to define its own sense, or better, each lexical form could be matched with corresponding lexical form from other languages [07:33:55] But even if the current system is able to link between lexical forms of various languages, I am not sure how it works. [07:34:56] I thought that if if not all forms are shared by all senses, it's a good signal to create two separate lexemes. [07:43:54] That's how the current Wikidata lexicological data works. Let me take Malay language lexemes as examples, "kerja" (noun, "work") and "kerja" (verb, "work") are separated due to different lexical category. Furthermore, "pekerjaan" (noun, "career") is separated from "kerja" (noun, "work") due to different sense. Although all words share the same root [07:43:54] "kerja". (re @Jan_ainali: I tho [07:43:54] ught that if if not all forms are shared by all senses, it's a good signal to create two separate lexemes.) [07:59:00] I filed now T424415 (re @vrandecic: I had disconnected it because I wanted to see if the compositions work. The compositions seem to work, it is just the test that ...) [08:02:46] Thanks, Denny, but it’s already reported as *T423853*. (re @vrandecic: I filed now T424415) [08:04:36] Oh, but the interesting part in the new one is that it only happens in the test, not when doing a function call [08:04:46] I'll still point to it. [08:07:45] Compositions generally, I think 🤷‍♂️ (re @vrandecic: Oh, but the interesting part in the new one is that it only happens in the test, not when doing a function call) [08:08:59] Aaaaah! Why would Z34148 work but Z33761 not?? [08:24:32] It’s the indirection, I think. The problem affects only a few implementations where the argument reference is not resolved up front. In v1, we could use Echo to force the evaluation, but this no longer has that effect. This is generally good, I guess, but I don’t see any way to avoid the problem currently, other than crossing over into code. (re @vrandecic: Aaaaah! Why [08:24:32] would [08:24:33] Z34148 work but Z33761 not??)