[00:20:34] Al I was pondering your comment in Phabricator about choosing output formats, and realised that we should ask for something similar to a lightweight enum, but where the items are Wikifunctions items, not from Wikidata. For example, we could flexibly enumerate all float64 formatting functions. Then a Wikipedia usage could be forced to choose from the list of valid [00:20:34] options. [01:00:20] There’s a task for the ZID equivalent. I think David raised it. I’ll track it down in the morning. 🎃 (re @u99of9: Al I was pondering your comment in Phabricator about choosing output formats, and realised that we should ask for something simi...) [08:47:25] It’s all done, apparently 🤔 *T394116* (re @Al: There’s a task for the ZID equivalent. I think David raised it. I’ll track it down in the morning. 🎃) [08:49:37] Oh! That explains why my scan didn't find it. (re @Al: It’s all done, apparently 🤔 T394116) [08:50:11] `Z884` (re @u99of9: Oh! That explains why my scan didn't find it.) [08:56:06] There should probably be an open task to make it operational 🤷‍♂️ (re @u99of9: Oh! That explains why my scan didn't find it.) [10:52:53] I don’t know why Z28375 isn’t working. I’ve duplicated it as Z28377 (?) and I’ll try regressing the function definition back towards what it should be when my reserve of patience has returned to an acceptable level 😏 [10:53:32] "We added a display function (https://www.wikifunctions.org/wiki/Z21583) that was created earlier this year to the monolingual text Type (https://www.wikifunctions.org/wiki/Z11). This doesn’t change the display in Wikifunctions, as monolingual texts have a bespoke display component anyway, but it enables Functions that have this Type as a return type to be used as [10:53:32] embedded Wiki [10:53:33] functions calls." .... Ahh, so we need to make displays for a few more types then. Readers too? [10:56:28] Not duplicate. I'll show a pic: (re @Al: I don’t know why Z28375 isn’t working. I’ve duplicated it as Z28377 (?) and I’ll try regressing the function definition back tow...) [10:56:47] We’ll know when the table is complete! Typed lists are something of a priority, but we need a technical steer for those. (re @u99of9: "We added a display function that was created earlier this year to the monolingual text Type. This doesn’t change the display in...) [10:57:48] https://tools-static.wmflabs.org/bridgebot/7f627934/file_73936.jpg [10:57:49] https://tools-static.wmflabs.org/bridgebot/fd97300c/file_73937.jpg [10:59:34] Why are they a priority? I haven't even thought about them. I guess maybe they should use html? (re @Al: We’ll know when the table is complete! Typed lists are something of a priority, but we need a technical steer for those.) [10:59:34] Thanks! Same as your issue the other day? Must be usability 😉 (re @u99of9: ) [11:01:03] Yes. Interestingly it looks like @genocation has a way of auto-detecting this. https://www.wikifunctions.org/w/index.php?title=Wikifunctions:Request_for_cleanup&curid=65735&diff=219054&oldid=218807 (re @Al: Thanks! Same as your issue the other day? Must be usability 😉) [11:04:20] I wish I had one! I searched for the substring `"Z7K1": {` in the content, and looked into all function call function keys that were something other than a reference. [11:38:42] Just for Z7K1 being an object, I get seven results with [[Special:Search/: "Z1K1 Z7 Z7K1 Z1K1"]]. In two cases, the object is a [[Z18]], the rest are a [[Z7]]. We’ll look into those. (re @genocation: I wish I had one! I searched for the substring "Z7K1": { in the content, and looked into all function call function keys that we...) [11:41:17] I suppose Boolean, then? (re @u99of9: Why are they a priority? I haven't even thought about them. I guess maybe they should use html?) [11:59:24] Search it is then. Can we put that search link at the top of the cleanup section so it's easy to maintain? (re @Al: Just for Z7K1 being an object, I get seven results with [[Special:Search/: "Z1K1 Z7 Z7K1 Z1K1"]]. In two cases, the object is a ...) [12:00:50] I'm thinking everybody cares more about Chemical elements right? 😁 (re @Al: I suppose Boolean, then?) [12:01:59] I thought lightweight enumerations would just work! (re @u99of9: I'm thinking everybody cares more about Chemical elements. Right? 😁) [12:03:48] But more importantly, I looked at David's video and realised how tough it will be to read/display/localise the Wikidata time object. Even more complicated than Wd quantity. That could be an epic. [12:04:33] What does working mean for their display? Do they print the label as a string in the configured language? (re @Al: I thought lightweight enumerations would just work!) [12:08:48] Should we wait for Julian before doing this? Do we need a timezone enum first to convert back and forwards from integer offset minutes? (re @u99of9: But more importantly, I looked at David's video and realised how tough it will be to read/display/localise the Wikidata time obj...) [12:10:29] Or maybe I should push on with quantities and do arbitrary unit conversion then a convert function with options? [12:10:40] So many options! [12:23:01] 🤔 They’re just a Z6091, right? What’s the plan there? (re @u99of9: What does working mean for their display? Do they print the label as a string in the configured language?) [12:24:41] We do need an agreed way forward on non-Gregorian dates 🤷‍♂️ (re @u99of9: Should we wait for Julian before doing this? Do we need a timezone enum first to convert back and forwards from integer offset m...) [12:25:20] ✅ Done. (re @u99of9: Search it is then. Can we put that search link at the top of the cleanup section so it's easy to maintain?) [12:25:35] QIDs aren't exactly a user friendly output on wiki. I'd advocate for the WD label, but obviously we won't be able to read that back in (except within limited enum types where there's a chance to check against all the labels). (re @Al: 🤔 They’re just a Z6091, right? What’s the plan there?) [12:31:14] I think the reading is just the bespoke component. Display is the open question, but I agree: it can only be the Wikidata label… with language tag if not in the requested language? (re @u99of9: QIDs aren't exactly a user friendly output on wiki. I'd advocate for the WD label in the language [defaulting to QID or fallback...) [12:35:41] I haven't really looked at how the component works on-wiki. All I see is the Wikitext where it would be nice (although not essential) to also allow for {{#function:Z28261|hydrogen}} (re @Al: I think the reading is just the bespoke component. Display is the open question, but I agree: it can only be the Wikidata label…...) [12:45:28] The latest newsletter has the following passage: [12:45:28] If you ever need to check whether a specific Wikimedia wiki has Wikifunctions enabled, you can look at the bottom of your wiki's Special:Version (https://www.wikifunctions.org/wiki/Special:Version) page to see if {{#function}} is listed, or more generally you can refer to the official source of truth. [12:45:30] Was "the official source of truth" supposed to be a link? [12:52:08] I hear you 🤔 David might have some thoughts. It seems to me that we would need the more targeted fetches first. (re @u99of9: I haven't really looked at how the component works on-wiki. All I see is the Wikitext where it would be nice (although not essen...) [12:55:27] The (?targetted) QID reader will be easy to write, but perhaps difficult to connect like the equality function seems to be. But matching to a list of enums mapped to labels also seems reasonably possible. (re @Al: I hear you 🤔 David might have some thoughts. It seems to me that we would need the more targeted fetches first.) [20:20:02] 6962 [20:20:13] thank you