[01:51:48] adding sitelinks to abstractwiki is possible from Wikidata: https://www.wikidata.org/w/index.php?title=Q408&diff=2474721841&oldid=2474112934 [01:52:46] I would personally prefer that sitelinks be added automatically but this is good for now (re @u99of9: adding sitelinks to abstractwiki is possible from Wikidata: https://www.wikidata.org/w/index.php?title=Q408&diff=2474721841&oldi...) [01:54:09] Yes. I imagine it will be rare to want to link a different QID. Under what circumstances might that happen? (re @Feeglgeef: I would personally prefer that sitelinks be added automatically but this is good for now) [01:57:14] just checking, are we back on v2? (re @wmtelegram_bot: Definitely back up before April 10!! I am hoping we can get it back up tomorrow, Friday at latest.) [02:08:56] I can't think of anything. If you would want to link to a different QID you would probably be better off just moving the article (which reminds me, though not relevant to Wikifunctions, I still think there ought to be a move tool for Abstract Wikipedia articles) (re @u99of9: Yes. I imagine it will be rare to want to link a different QID. Under what circumstances [02:08:56] might that happen?) [02:16:28] T420869 [08:48:00] New function Z32431 seems to be working properly in Z32398 now. (re @Ameisenigel: Thanks for your help. I will continue this tomorrow.) [12:38:10] I'm thinking about fragment structures like this: Z32410. I'm hoping the top level call from Abstract Wikipedia can send parameters like the grammatical number of the subject, say plural, and by the time it gets to the English constructor, the English function will have the "subject is plural" boolean set. I'm not sure what the initial type should be for that [12:38:10] grammatical number, [12:38:10] because if it should eventually work in any language, we sort of need to give all possible options to the user. But most of our grammatical feature types are specific to one group of languages, instead of including "all" options. [12:45:09] Also, in English we're going to need to infer whether or not an indefinite article ("a"/"an") is needed before the subject/object. Because this is not an abstract decision, so will not be specified in the top level call. So far I can tell that plurals shouldn't have an indef article, but some singulars like "Australia" also shouldn't. Does the complete set of rules [12:45:09] seem programma [12:45:10] ble? (re @u99of9: I'm thinking about fragment structures like this: Z32410. I'm hoping the top level call from Abstract Wikipedia can send paramet...) [13:08:45] Yes, that was probably me trying to construct "Australia is the sixth-largest country in world." I agree it would be better if we could send the parameter *Q12935276* and have Q1817208 on a list of grammatical features and infer that the best lexeme form in English is L3415-F3. But I don't think we have such functional capability yet. I agree it would be worth working [13:08:45] on, because [13:08:46] other languages will have much more trouble if they have to convert from an English adjectival. (re @nyuhn: The problem is that there is already an attempt to use this function in Abstract Wikipedia, where it looks odd.) [13:09:42] Up to a point. Countries (in English) mostly have no article, but sometimes they require a definite article. (Similarly, most countries require the preposition “in” but sometimes “on” is preferred.) These are lexicographic rules, with their own general principles and exceptions. The “definiteness” of a language-neutral function shouldn’t be grammatical, since the ru [13:09:42] [13:09:43] les vary by language. Similar difficulties arise when the subject can be considered in the singular or the plural (etc, e.g. “The oboe is a woodwind instrument” vs “Oboes are woodwind instruments”, or when characteristics may be expressed using adjectives or subordinate clauses. For example, “England is a country in the United Kingdom” but “The United Kingdom is a E [13:09:43] [13:09:44] uropean and Commonwealth country”. (re @u99of9: Also, in English we're going to need to infer whether or not an indefinite article ("a"/"an") is needed before the subject/objec...) [13:21:15] [Thanks for the terminology.] For now I don't plan to be able to cover subordinate clauses at all with this group of functions. I'm not worried about your oboe example, because I think it's okay for the AW call to specify the grammatical number of the subject oboe (although, in general we should be able to ask for duals etc). I have no idea how to deal with your [13:21:15] definiteness exam [13:21:16] ple - I was hoping that would be a whole separate function, but I agree that at the AW level the user should not be expected to bifurcate according to whether a subject country needs a definite article in English!!! So should someone try to write the lexicographic rules you mention into a function? What inputs/outputs should it have? (re @Al: Up to a point. Countries [13:21:16] (in Englis [13:21:17] h) mostly have no article, but sometimes they require a definite article. (Similarly, most co...) [13:26:30] See my thread with @nyuhn about the possibility of adjectival modifiers ("woodwind"). I'm hopeful on that front that we can send a list of adjectival QIDs and grammatical feature QIDs, along with the object noun's QID, to get back the string that gets put into "object string" in Z32410. So, hopefully it's already set up for this situation (e.g. hopefully we could [13:26:30] infer the input [13:26:31] strings for Z32417). (re @Al: Up to a point. Countries (in English) mostly have no article, but sometimes they require a definite article. (Similarly, most co...) [13:36:01] I suppose this (https://www.wikifunctions.org/view/en/Z6830?call=%7B%22Z1K1%22%3A%22Z7%22%2C%22Z7K1%22%3A%22Z6830%22%2C%22Z6830K1%22%3A%7B%22Z1K1%22%3A%22Z6091%22%2C%22Z6091K1%22%3A%22Q12935276%22%7D%2C%22Z6830K2%22%3A%7B%22Z1K1%22%3A%22Z6092%22%2C%22Z6092K1%22%3A%22P5137%22%7D%2C%22Z6830K3%22%3A%22Z1002%22%7D) is a starting point, but we need to filter for the best [13:36:01] form. @dvd_cc [13:36:02] c27919 was doing good work like Z27327, but as far as I recall it didn't quite cover this particular requirement. (re @u99of9: Yes, that was probably me trying to construct "Australia is the sixth-largest country in world." I agree it would be better if w...) [14:07:08] It’s just important to allow grammatical rules to apply per language. The higher level function says, “here is the topic of this sentence and here is what we’re saying about it”, which is the basic structure of Wikidata. We probably need basic classification according to whether the subject is a person, place, event etc, whatever it is that language-level functions will t [14:07:08] [14:07:10] ypically use to determine whether to use which articles or forms of expression. Some languages prefer subordinate clauses where others might have compound nouns. But the abstract content just says that Australia is a country and we leave the French function to work out that it needs to be “L’Australie” and that she is feminine and singular, unlike “Les États-Unis”, who [14:07:10] [14:07:11] are masculine and plural. At this stage, though, I would leave the classification to the language-level functions and avoid the whole “article-less” or “article-ful” nonsense altogether. (re @u99of9: [Thanks for the terminology.] For now I don't plan to be able to cover subordinate clauses at all with this group of functions. ...) [14:19:11] I guess. This is basically just subject–predicate, so different languages can render it with different representations broadly equivalent to: [14:19:13] “Australia has a highly developed economy” [14:19:14] “Australia’s economy is highly developed” [14:19:16] “Australia has an economy that is highly developed” [14:19:17] “Australia is a country with a highly developed economy” [14:19:19] “Australia is a highly-developed-economy-country” [14:19:20] “As for Australia’s economy, it is highly developed” and so forth. (re @u99of9: See my thread with @nyuhn about the possibility of adjectival modifiers ("woodwind"). I'm hopeful on that front that we can send...) [18:09:15] Hi! Can a functioneer connect this implementation? My functioneer rights request has not been reviewed yet. [18:09:16] [[Z32395]] [18:09:17] Thanks! ❤️ [18:11:34] Done (also, you don't have to wikilink w/ the bot) (re @poslovitch: Hi! Can a functioneer connect this implementation? My functioneer rights request has not been reviewed yet. [18:11:34] [[Z32395]] [18:11:35] Thanks! ❤...) [19:01:20] I had a bit of a poke around and it seems that we might be able to rely on item aliases to determine whether the English label for a country requires the definite article. It would also work for the universe and the sun but not (currently) the solar system. Checked UK, USA, UAE, the Philippines, and confirmed that neither Ukraine nor Congo has an alias that uses “the” [19:01:20] (dated [19:01:20] and deprecated). (re @u99of9: [Thanks for the terminology.] For now I don't plan to be able to cover subordinate clauses at all with this group of functions. ...) [19:30:45] That's correct; it still hasn't been enabled. I'm thinking I should take time to figure out some simple performance monitoring (for example, how often are we getting orchestrator timeouts), and check that before and after enabling the import of references. Since I'm extremely busy through end of March, I'd be happy to get this done during the first half of April. But [19:30:45] please let [19:30:46] me know if there are objectives that would make this more urgent. The relevant ticket is T404652. (re @u99of9: David Am I right that we still can't get the references on Wikidata statements? Is there any update on when this might be coming...) [20:15:18] Thanks for reporting this! I've created T420998, and next week should be a good time for me to look at this. (re @u99of9: David I think I found a problem with Z6839 when trying to get https://en.wikipedia.org/wiki/Victoria_(state) from Q36687. I've a...) [21:06:09] Okay. Performance is a super high priority indeed. But verifiability is too, so we will eventually need references before serious editors want to start helping us out. (re @David: That's correct; it still hasn't been enabled. I'm thinking I should take time to figure out some simple performance monitoring (...) [21:10:32] Why do we need Wikidata references? Can we not just make our own references? (re @u99of9: Okay. Performance is a super high priority. But verifiability is too, so we will eventually need references before serious edito...) [21:14:27] We could, but that would be duplicated effort and storage. (re @Feeglgeef: Why do we need Wikidata references? Can we not just make our own references?) [21:19:25] Yes, this is what we will have to do for now. The one I made for Q408 was more fiddly and annoying than most non-technical editors will stomach. And I have a general concern that duplication may become a big concern for AW, not just in references. (re @Feeglgeef: Why do we need Wikidata references? Can we not just make our own references?) [21:21:27] Who is in charge of wikilinksbot? Can you make us a syntax so we can send links to AW easily? (re @wikilinksbot: Q408 🇦🇺 Australia) [21:24:45] A volunteer I believe (re @u99of9: Who is in charge of wikilinksbot? Can you make us a syntax so we can send links to AW easily?) [21:24:46] It's @jhsoby. We don't have an official prefix yet, right? (re @u99of9: Who is in charge of wikilinksbot? Can you make us a syntax so we can send links to AW easily?) [21:24:52] No (re @Jan_ainali: It's @jhsoby. We don't have an official prefix yet, right?) [21:25:32] T420654 [21:26:04] Isn't it abstract:, according to the meta Interwiki map page? [21:27:11] It doesn't seem to actually work (re @wmtelegram_bot: Isn't it abstract:, according to the meta Interwiki map page?) [21:27:18] [[abstract:Q1]] [21:27:27] Nope [21:31:18] See this earlier discussion on the topic of the link bot (re @Jan_ainali: But if you only can get one (considering it is a volunteer needing to change the bot) would you still rather have the default li...) [21:33:37] It might be because interwiki.php hasn't been updated yet while the meta page has, so will probably have to wait and see for when the update script for that runs. [21:40:11] It's ancient history, but I'm curious about it: what are all the changes that were done in https://www.wikifunctions.org/wiki/Z8010?uselang=en&diff=prev&oldid=51389 ? One is adding a label in German. But what is the other change? Z2K5 is description, but as far as I can tell, no actual description is added. Does the system add Z2K5 even if there's no content to write [21:40:11] in it? [22:09:04] 0010 [22:09:26] Welcome Asaf \o/ [22:10:53] I think it should. But there appear to be these five exceptions. (https://www.wikifunctions.org/w/index.php?search=z2k1+%21z2k5&title=Special%3ASearch&profile=advanced&fulltext=1&ns0=1) (re @amire80: It's ancient history, but I'm curious about it: what are all the changes that were done in https://www.wikifunctions.org/wiki/Z8...) [22:31:35] could a functioneer connect the [[Z32486]] & [[Z32474]] to [[Z32473]] [22:35:55] nvm on Z32486 gotta fix [22:37:26] Done [22:38:17] thx [22:43:02] I'd prefer a QID to generate two lines, one linking WD, one AW. (re @Jan_ainali: This is an interesting conundrum. Do we want the default link to be to Wikidata still or should we change it to Abstract Wikiped...) [22:48:25] That's an interesting approach I hadn't imagined. Tbh I'm a bit surprised that outdated aliases have been purged, because I thought the timespan of an alias was not restricted to modern terms. (re @Al: I had a bit of a poke around and it seems that we might be able to rely on item aliases to determine whether the English label f...) [22:53:56] I never said they’d been purged. Maybe we just got lucky? 🤷‍♂️ (re @u99of9: That's an interesting approach I hadn't imagined. I'm a bit surprised that outdated aliases have been purged, because I thought ...) [22:55:00] I can't believe "The Ukraine" was never present. (re @Al: I never said they’d been purged. Maybe we just got lucky? 🤷‍♂️) [23:01:46] Here is one from "a Brit"/"the Brits" (sometimes important meaning is encoded in the singular). https://www.wikidata.org/w/index.php?title=Q212&diff=343100&oldid=343097 (re @u99of9: I can't believe "The Ukraine" was never present.) [23:07:08] We don’t only whinge 😉 (re @u99of9: Here is one from "a Brit"/"the Brits" (sometimes important meaning is encoded in the singular). https://www.wikidata.org/w/index...)