[00:26:20] It looks like Cory is all over this bug report already. Thanks! (re @u99of9: Help! In my talk I plan to show off the wikidata-supported-power of one of my favourite functions Z28509. But I think V2 has ca...) [02:15:37] Performance is currently dire. Things are timing out all over the place that would usually complete okay. [02:39:26] is there a function for "[name] (are|is) a [role] of (the|a)? [item]"? ik theres defining role sentence but that's "[name] is a [role]", not "the [role]", and also it doesn't account for plurals [02:42:55] Great! I see you fixed the sv indefinite article function too. Dream team! 🙌 (re @Jan_ainali: Ah, this time it was missing a lexeme, but I couldn't figure that out from the error message. Now it works! 🎉) [02:44:32] There is Z32244, but it's only for English and there isn't a corresponding global function (re @wmtelegram_bot: is there a function for "[name] (are|is) a [role] of (the|a)? [item]"? ik theres defining role sentence but that's "[na...) [02:44:36] yes and that would be really unfortunate. [02:45:08] ty Earldridge Jazzed i'll try that out [02:46:16] Al this may be related to one of the tickets you've filed already, but Z16556 is currently broken, likely V2. [02:51:09] i think keeping it simple is the way to go as much as possible. [02:51:10] I think the mul information should be dropped from the standard output. [02:51:11] Instead we could supply a link to a clear backtrace of all the steps so the user can debug each step and understand how the final output was produced. (re @Al: Yes. There are limited options while the construction is in text, rather than HTML. We could just return it as text, or with any...) [02:59:19] GrimRob has already started the work đŸ€© Z29057 (re @Npriskorn: With the current capabilities and the new copy paste UI for compositions I think we now have everything to start making more com...) [03:31:03] GrimRob contributed Z29057, but it outputs a string which is not what we need for AW [03:31:04] ChaoticV... created Z32053 which seems like a good start đŸ€© (re @Npriskorn: With the current capabilities and the new copy paste UI for compositions I think we now have everything to start making more com...) [03:32:01] Is this a stricter limit than before: "orchestrator rate limit: "1000"" ? (re @u99of9: Performance is currently dire. Things are timing out all over the place that would usually complete okay.) [03:38:55] 2456 [03:38:57] @Sannita I wonder if you have plans to support downloading AW in both source and rendered forms? [03:38:58] Should I create a ticket? [03:40:06] welcome @sunafterrainwm đŸ€© Would you like to share a little about yourself? [03:41:14] go ahead, be bold đŸ€© (re @wmtelegram_bot: (copying from the abstractwiki main page talk bc i havent gotten an answer there yet) On en:Main Page, there's a counte...) [04:04:40] He’s a user from Chinese community. I know him somewhat. (re @Npriskorn: welcome @sunafterrainwm đŸ€© Would you like to share a little about yourself?) [04:14:06] I'm not sure exactly what you mean by this but I think a year ago there was some discussion as to local Abstract Wikipedia, as in for example you could import a python library or use a program to use Abstract content locally, but I don't think it got anywhere (re @Npriskorn: @Sannita I wonder if you have plans to support downloading AW in both source and rendered [04:14:06] forms? [04:14:07] Should I create a ticket?) [04:15:03] Theoretically anybody could make such a library or program using the API at this point [04:16:57] And then this would allow you to make a "local" (processing still occurs on the WMF's end) version of articles, using the API to copy over articles that you want from the Wiki [04:17:41] It would be terrible for Wikimedia if companies have to scrape AW to get the rendered content, so providing the rendered text is essential. [04:19:09] Of course you could set up the whole pipeline and render locally, but that is probably not a great experience and very error prone. Much easier to scrape then [08:36:08] Not that I'm aware of (re @Npriskorn: i think keeping it simple is the way to go as much as possible. [08:36:08] I think the mul information should be dropped from the standard ...) [08:36:51] I think it's a bit early to discuss that, you might create a ticket but I doubt it will get attention in this phase (re @Npriskorn: @Sannita I wonder if you have plans to support downloading AW in both source and rendered forms (finished sentences)? [08:36:52] The AI com...) [09:24:21] I understand, yes it is very early and not much content of value yet. (re @Sannita: I think it's a bit early to discuss that, you might create a ticket but I doubt it will get attention in this phase) [10:55:24] @Jan_ainali https://abstract.wikipedia.org/wiki/Q711 [10:56:35] https://tools-static.wmflabs.org/bridgebot/60a9c72e/file_79051.jpg [10:58:01] It's exactly the same function but one instance gave incorrect output 😅 : https://tools-static.wmflabs.org/bridgebot/f5287275/file_79052.jpg [10:58:18] Missing an â€œĂ€r”? (re @Npriskorn: ) [10:58:32] I don't understand why and here a backtrace would be super helpful I think. [10:59:06] It’s not the same function đŸ€·â€â™‚ïž (re @Npriskorn: It's exactly the same function but one instance gave incorrect output 😅) [10:59:20] No, it's different functions. Z26570 vs Z28016 (re @Npriskorn: It's exactly the same function but one instance gave incorrect output 😅) [11:00:42] correct: Ulan Bator _Ă€r_ huvudstad_en_ _i_ Mongoliet (re @Al: Missing an â€œĂ€r”?) [11:01:19] oh , I only looked at the outermost one 😅 (re @Jan_ainali: No, it's different functions. Z26570 vs Z28016) [11:02:02] Looking at the test cases of Z28026, it will be more complicated in Swedish. (re @Npriskorn: correct: Ulan Bator Ă€r huvudstaden i Mongoliet) [11:03:44] The last test case, with ownership, would need a total reconstruction of the sentence. [11:06:49] It’s more complicated for English too. The last test case is unidiomatic in English. I even put that comment in the description! (re @Jan_ainali: Looking at the test cases of Z28026, it will be more complicated in Swedish.) [11:09:47] Or perhaps all would be flipped, that would be smoother in all cases and avoid the complication of different prepositions. Like this: [11:09:47] (Determined genitive of [dependency]) [role] "Ă€r" [subject] (re @Jan_ainali: The last test case, with ownership, would need a total reconstruction of the sentence.) [11:10:25] Maybe 2 versions of the function would be better, one with genitive and one without? [11:10:26] without: Ulan Bator Ă€r huvudstaden i Mongoliet [11:10:28] with: Pluto Ă€r Mickey Mouses hund [11:10:29] I'm not a native and the second one might work better in Danish (re @Jan_ainali: The last test case, with ownership, would need a total reconstruction of the sentence.) [11:11:43] example? (re @Jan_ainali: Or perhaps all would be flipped, that would be smoother in all cases and avoid the complication of different prepositions. Like ...) [11:13:40] New Yorks huvudstad Ă€r Albany. Storbritanniens huvudstad Ă€r London. Blodets fĂ€rg Ă€r röd. Musse Piggs hund Ă€r Pluto. [11:13:40] But thinking about it, shifting the subject to be the object shifts the focus in the sentence. Hmmm.. [11:18:41] In English too. I think in the general case it depends on the nature of the complement or predicate. If we think of it as a core P31-type representation, it depends on the nature of the subject too. But I think that’s language specific; the article call just establishes the relation and language-specific functions have to navigate the complexities. (re @Npriskorn: Maybe 2 [11:18:42] versi [11:18:43] ons of the function would be better, one with genitive and one without? [11:18:44] without: Ulan Bator Ă€r huvudstaden i Mongol...) [11:19:40] I guess one could say [subject] Ă€r (Genitive of [dependency]) [role]. [11:19:41] Albany Ă€r New Yorks huvudstad. London Ă€r Storbritanniens huvudstad. Röd Ă€r blods fĂ€rg. Pluto Ă€r Musse Piggs hund. [11:19:43] Although the third one will sound super weird with most nouns. (re @Jan_ainali: New Yorks huvudstad Ă€r Albany. Storbritanniens huvudstad Ă€r London. Blodets fĂ€rg Ă€r röd. Musse Piggs hund Ă€r Pluto. [11:19:44] But thinking...) [11:24:05] Perhaps the third example is just weird in English too? [11:24:11] Yes. The same in English: “Red is the colour of blood” rather than *“Red is blood’s colour”. Sometimes either construction is idiomatic, sometimes not. (re @Jan_ainali: I guess one could say [subject] Ă€r (Genitive of [dependency]) [role]. [11:24:11] Albany Ă€r New Yorks huvudstad. London Ă€r Storbritanniens h...) [11:25:38] But what is it meant to mean? Is blood defining the color red? That's just weird without specifying blood from what organism. [11:26:37] It might be better to pick encyclopedic test cases rather than poetic. :) [11:28:56] this is a valid point, let's change the test cases to make our life simpler and the functions suitable for our purposes. [11:28:56] noone has come here to write poetry I assume (re @Jan_ainali: It might be better to pick encyclopedic test cases rather than poetic. :)) [11:29:31] “Red is the colour of blood đŸ©ž, tomatoes 🍅 and London buses.” 😏 (re @Jan_ainali: But what is it meant to mean? Is blood defining the color red? That's just weird without specifying blood from what organism.) [11:53:10] Edge cases are important, so I wouldn’t argue for changing them “to make our life simpler”. But they should be aligned to the desired results rather than what the function happens to produce. With that criterion in mind, the third case is correct for English and the fourth is incorrect. That probably just means that the English function needs two branches or sub-functions [11:53:10] ( [11:53:11] maybe more). That’s probably true for all Germanic languages, and the distinction is relevant in the non-defining case as well. We talk about the capital city *of* a country, but the largest city *in* a country and the largest settlement *on* an island, for example. (re @Npriskorn: this is a valid point, let's change the test cases to make our life simpler and the [11:53:11] functions sui [11:53:13] table for our purposes. [11:53:14] noone h...) [12:01:14] Yes. In Swedish, it still is a capital in a country (huvudstad i [land]), and that's easy to build. The third one is more tricky. "Röd Ă€r fĂ€rgen i blod." using in would be technically correct, but kind of describing what color you would see if you are in the blood. "Röd Ă€r fĂ€rgen pĂ„ blod." using on is also correct, but is more describing the property of the blood [12:01:14] than the [12:01:16] role of the color. "Röd Ă€r fĂ€rgen av blod." using of is also correct, but, at least to me, implies what color something smeared by blood gets. "Röd Ă€r blods fĂ€rg." using genitive defines the role of the color but just sound super weird to me and you would only find it on poetry or horror novels. [12:03:40] /me is laughing/crying in Breton (where almost every case is an edge case...) [12:08:26] "Ruz eo liv ar gwad." (Red is the color of the blood) [12:08:28] "Liv ar gwad (a) zo ruz" (the color of the blood is red) [12:08:29] And many more variations... [12:08:31] Where you choose depending on where you want to emphasis and depending on the previous sentences/context [12:11:35] Yeah, I think it will be increasingly more important to in detail define in the description of the function what it is meant to do (and perhaps even what it should not be used for!). [12:11:42] My little suburb article has a few features people may be interested in. Unfortunately sunset calculations are still too slow to render. https://abstract.wikipedia.org/view/en/Q55771891 [12:12:55] And use lots of test cases to demonstrate it. (re @Jan_ainali: Yeah, I think it will be increasingly more important to in detail define in the description of the function what it is meant to ...) [12:15:09] The function is supposed to be defining, with qualification. Red is a colour; it is the colour of blood (for example). So, red is the colour of blood. The blood of all mammals is red, but that doesn’t define red; it characterises the blood of mammals. Perhaps “Blue is the colour of the sky” would be a better test case? (re @Jan_ainali: Yes. In Swedish, it still is a [12:15:09] capital [12:15:10] in a country (huvudstad i [land]), and that's easy to build. The third one is more tricky...) [12:17:49] (Why not "Blue is the color of sky."?) Yes, that would make it better. (Just adding "the" before "blood" would be just as good in my opinion.) (re @Al: The function is supposed to be defining, with qualification. Red is a colour; it is the colour of blood (for example). So, red i...) [12:18:05] But on the other hand "X has Y" is turning into a monster. Today's craziness led it to Z32935 which may be hard. [12:18:30] Yes. I believe it contains the first Wikipedia sentence in the wiki
 where a “Wikipedia sentence” is defined as containing a link to a Wikipedia article. (re @u99of9: My little suburb article has a few features people may be interested in. Unfortunately sunset calculations are still too slow to...) [12:20:27] Sounds a little circular ... (re @Al: Yes. I believe it contains the first Wikipedia sentence in the wiki
 where a “Wikipedia sentence” is defined as containing a lin...) [12:21:42] ... but yeah. I remember when the www was all about links. (re @u99of9: Sounds a little circular ...) [12:24:07] language is awfully complicated, I wonder why humans bother at all with it đŸ€Ł (re @Al: Edge cases are important, so I wouldn’t argue for changing them “to make our life simpler”. But they should be aligned to the de...) [12:25:42] We wouldn’t use an article there. It’s a mass noun used generically, like “beauty is truth”. (re @Jan_ainali: (Why not "Blue is the color of sky."?) Yes, that would make it better. (Just adding "the" before "blood" would be just as good i...) [12:27:58] As Abstract Wikipedia shall be! But selectively suppressing the links to avoid repetition is a little challenging. (re @u99of9: ... but yeah. I remember when the www was all about links.) [12:27:59] "Blood is red" would be better in most cases IMO. Could be qualified with a prefix. Human blood is red. (re @Jan_ainali: Yes. In Swedish, it still is a capital in a country (huvudstad i [land]), and that's easy to build. The third one is more tricky...) [12:28:50] /me [12:31:00] Monster functions đŸ€Ł (re @u99of9: But on the other hand "X has Y" is turning into a monster. Today's craziness led it to Z32935 which may be hard.) [12:31:45] Wikifunctions is the largest rabbi hole ever đŸ€Ł [12:33:23] Yes, but I’m afraid that misses the point. The function is defining its given subject, which happens to be “red” in this case. “Red is the colour of visible light with the longest wavelength” sort of thing? (re @Npriskorn: "Blood is red" would be better in most cases IMO. Could be qualified with a prefix. Human blood is red. Or the blood of humans i...) [12:35:09] All the knowledge in every language 😎 (re @Npriskorn: Wikifunctions is the largest rabbit hole ever đŸ€Ł) [12:39:18] đŸ€” 
although we could consider it a special case of anaphora, like the use of pronouns, dropped in conjunctions. (re @Al: As Abstract Wikipedia shall be! But selectively suppressing the links to avoid repetition is a little challenging.) [12:55:39] Maybe, but doesn’t it just function as some sort of Kleenean complement to Z32645? I’m not convinced that they are really separate questions. Putting it another way, we hardly ever have to think about whether the context requires definite, indefinite or zero. The hard part is codifying the selection; I don’t think we need to do that three times. Sorry, I haven’t given thi [12:55:39] [12:55:40] s much thought yet đŸ€·â€â™‚ïž (re @u99of9: But on the other hand "X has Y" is turning into a monster. Today's craziness led it to Z32935 which may be hard.) [13:03:19] I agree it's a choice between three options. Yes that can ultimately be decided in one function. But for now I'm trying to divide the task until I understand it more. Sometimes I think the indications will have to be weighed against each other. For example, the sky item you said is a mass noun could be used non-generically "colour of *the* sky". So I think some of the [13:03:19] sub functio [13:03:20] ns will often come back with "maybe"... (re @Al: Maybe, but doesn’t it just function as some sort of Kleenean complement to Z32645? I’m not convinced that they are really separa...) [13:03:44] I think so, but *T420021* is resolved đŸ€” (re @u99of9: Al this may be related to one of the tickets you've filed already, but Z16556 is currently broken, likely V2.) [13:09:26] I also realised today that we may first have to choose the lexeme, then call for its grammar, before deciding the articles. Probably obvious. But it means maybe I should structure a whole extra layer between the abstract, the grammatical and the constructor. (re @u99of9: I agree it's a choice between three options. Yes that can ultimately be decided in one function. [13:09:26] But for now I [13:09:26] 'm trying to divid...) [13:11:27] It’s “blood” that is the mass noun (used generically). “The sky” is just singular. “Rain is wet and the sky is blue; clouds are white or grey” (who said we weren’t here to write poetry?) Hmm
 “all the colours of *the* rainbow” đŸ€·â€â™‚ïž (re @u99of9: I agree it's a choice between three options. Yes that can ultimately be decided in one function. But for n [13:11:28] ow I'm trying to divid...) [13:13:57] Items and lexemes, yes. Selectively, per language, they could usefully be bundled together in a compound fetch. (re @u99of9: I also realised today that we may first have to choose the lexeme, then call for its grammar, before deciding the articles. Prob...) [13:14:43] Oh yes, blood is simpler. But Jan was also leaving a zero article on sky "colour of sky", which imo is a "maybe"... (re @Al: It’s “blood” that is the mass noun (used generically). “The sky” is just singular. “Rain is wet and the sky is blue; clouds are ...) [13:16:55] I'm not sure what you mean here. As I understand it the AW call will not specify the lexeme, so it has to be the second level call after the item? (re @Al: Items and lexemes, yes. Selectively, per language, they could usefully be bundled together in a compound fetch.) [13:21:48] That is what I mean (or this is): when fetching an item for a language (specified by AW) a selective fetch could also fetch lexemes for that item and language, if lexemes were one of the “parts” (which is what makes it a “compound” fetch). (re @u99of9: I'm not sure what you mean here. As I understand it the AW call will not specify the lexeme, so it has to be the second [13:21:48] level ca...) [13:25:21] đŸ€” 
we could build it ourselves, I suppose, but I was thinking of a built-in. (re @Al: That is what I mean (or this is): when fetching an item for a language (specified by AW) a selective fetch could also fetch lexe...) [13:57:49] I am planning to make a global function out of Z32244, but it has a redundant language input (Z32244K4) that needs to be removed. BTW I don't have functioneer rights [14:09:20] I don’t think the K4 is redundant; it’s how we distinguish between language variants like en-us and en-gb. Here, for example. (https [14:09:20] //www.wikifunctions.org/view/en/Z32244?call=%7B%22Z1K1%22%3A%22Z7%22%2C%22Z7K1%22%3A%22Z32244%22%2C%22Z32244K1%22%3A%7B%22Z1K1%22%3A%22Z6091%22%2C%22Z6091K1%22%3A%22Q1088%22%7D%2C%22Z32244K2%22%3A%7B%22Z1K1%22%3A%22Z6091%22%2C%22Z6091K1%22%3A%22 [14:09:20] Q22006653%22%7D%2C%22Z32244K3%22%3A%7B%22Z1K1%22%3A%22Z6091%22%2C%22Z6091K1%22%3A%22Q1052%22%7D%2C%22Z32244K4%22%3A%22Z1199%22%7D) (It should be “the rainbow”, of course). (re @Earldridge Jazzed: I am planning to make a global function out of Z32244, but it has a redundant language input (Z32244K4) that needs to be removed...) [14:57:14] If you’re planning to lift this function to the global level, it’s worth considering the specificity of ‘a [Y] of [Z]’ (assuming [X] is fixed). The function is really completing an assertion about that X: X is fixed as (semantic) subject; the rest is the predicate — role Y applied to Z. How that surfaces linguistically — as a noun phrase, a copula, or anything else [14:57:14] [14:57:14] — is a language decision. There was some discussion about this earlier. (re @Earldridge Jazzed: I am planning to make a global function out of Z32244, but it has a redundant language input (Z32244K4) that needs to be removed...) [15:00:36] [in Jan’s reply, for example, to tie in to my latest thought] (re @Jan_ainali: Yes. In Swedish, it still is a capital in a country (huvudstad i [land]), and that's easy to build. The third one is more tricky...) [15:45:51] Is there a Phabricator ticket already for showing a label in the notifications? Now I have to click through to fully understand what was appreciated. : https://tools-static.wmflabs.org/bridgebot/1de2a38f/file_79056.jpg [16:15:04] I don’t recall seeing one. Good idea. We could also link Phabricator tickets from a page on-wiki. I found this one *T420466 đŸ€·â€â™‚ïž* (re @Jan_ainali: Is there a Phabricator ticket already for showing a label in the notifications? Now I have to click through to fully understand ...) [16:33:50] Ah, we have a new umbrella [16:33:50] https://phabricator.wikimedia.org/project/profile/8604/ (re @Jan_ainali: Is there a Phabricator ticket already for showing a label in the notifications? Now I have to click through to fully understand ...) [18:23:46] Just thinking about Z32822, and realising that while in English there is only one form of the demonym, in other languages there can be differences in gender and in number. Should this function take a gender and plural parameter or should it be up to the calling function to turn it into the right form? [19:12:53] I’d expect function that chooses the form to be responsible for its inflection and orthography. Initially, this function has no data, only a reference. Ideally we’d fetch the item (or from the item) just once. For some languages, we may also need to fetch from the corresponding lexeme. That probably applies to English demonyms too, since some noun forms differ from the [19:12:53] adject [19:12:53] ival form (Spaniard/Spanish, for example). [19:12:55] In practice, that means fetching (at least) a selection of statements and labels from the item and then selecting from among the linked lexemes (separate fetch of references) by looking at each lexeme in turn (fetch) to determine its lexical category (and relevant forms). [19:12:56] I’m not sure that answers your question, but a function implementation like Z32398 may give you a feeling for what I have in mind. In this case there are separate fetches for the sex or gender and the item’s label, and the feminine form is constructed formulaically in one case whereas in the other case, the linked lexeme determines the inflection of the article. [19:12:58] As has been mentioned before, this is all experimental! 😎 (re @Gears: Just thinking about Z32822, and realising that while in English there is only one form of the demonym, in other languages there ...) [19:22:42] Is there any reason why I can switch to `mul` on Wikifunctions but not on Wikidata, when selecting the UI language? Also, it seems like it cannot be changed permanently but simply adds `?uselang=mul`? [19:34:45] I don’t know the reason, I’m afraid, but continuity of uselang on Wikifunctions and Abstract Wikipedia was supposedly fixed by T420253. (re @adafede: Is there any reason why I can switch to mul on Wikifunctions but not on Wikidata, when selecting the UI language? Also, it seems...) [20:12:33] I took the liberty of disconnecting an implementation here because none of the tests passed. [20:12:34] Z32473 [20:13:18] This function should be changed to return html if we want it to be used in AW [20:26:20] New suggestion for display of fallback langcode https://www.wikifunctions.org/w/index.php?title=Talk:Z29749&dtrepliedto=new%7CTalk%3AZ29749 [20:26:50] We now have a new reference function with a dedicated class and it is added in the end. I think the language code should be similarly added as a sup span with a suitable class to avoid confusion for users. So9q (talk) 20:25, 31 March 2026 (UTC) [20:28:36] @Jan_ainali wdyt? Then it would hopefully be clear that the langcode is not part of the output. [20:28:37] We could further add a link to a help page to the spsn to help users understand what happened [20:32:27] I am not sure. There will be superscripts in other places, how would one deduce this one is not part of the output? (re @Npriskorn: @Jan_ainali wdyt? Then it would hopefully be clear that the langcode is not part of the output. [20:32:28] We could further add a link to a...) [20:36:49] by clicking on it? (re @Jan_ainali: I am not sure. There will be superscripts in other places, how would one deduce this one is not part of the output?) [20:36:56] All the fragment functions would need to be converted to HTML first, I think. I don’t think you can have a span in the middle of the string that you wrap as a Z89. (I did try, but maybe I just got it wrong). (re @Npriskorn: @Jan_ainali wdyt? Then it would hopefully be clear that the langcode is not part of the output. [20:36:58] We could further add a link to a...) [20:37:01] https://www.wikifunctions.org/w/index.php?title=Talk:Z32473&dtrepliedto=new%7CTalk%3AZ32473 Al GrounderUK [20:37:37] I think I have too poor imagination this late in the evening :) (re @Npriskorn: by clicking on it?) [20:39:13] we could have a single helper function like for references that add the fallback html sup thing and use that everywhere 😀 (re @Al: All the fragment functions would need to be converted to HTML first, I think. I don’t think you can have a span in the middle of...) [20:54:02] Plausibly
 but while almost all AW content is HTML wrapping strings, we don’t appear to have that option available. The superscript could be a stylised Q or L to indicate whether it comes from an item’s label or a lexeme. Which it is, and in which language, should be clear from the link text, but we could give it more or less prominence according to the item’s “styling [20:54:02] [20:54:02] ”, which is ultimately a function reference. (We need this anyway so that the result text for an item can be specified as linked, bold, italics and so forth.) (re @Npriskorn: we could have a single helper function like for references that add the fallback html sup thing and use that everywhere 😀) [22:31:47] https://tools-static.wmflabs.org/bridgebot/71c64123/file_79078.jpg [22:32:08] Please use screenshots 🙏 (re @Csisc1994: ) [22:34:07] https://tools-static.wmflabs.org/bridgebot/3bf6ebd3/file_79079.jpg [22:34:25] There is a problem with the Orchestrator. [22:35:32] This always occurs although I only use the "statement with reference" function and most of what exists in the Page is hard-coded in English and does not rely on heavy function calls. [22:37:07] A regular non-stub Wikipedia article should include around 35 statements. [22:37:54] There has been a resource issue but it seems to have eased now. I think each of the calls is sent straight to a queue, so it doesn’t care how simple they are. (re @Csisc1994: There is a problem with the Orchestrator.) [22:38:12] I see. [22:38:23] I propose to implement caching. [22:38:49] It is expensive to render function calls every time. [22:42:26] Yes, there is caching, but it is set to expire quite quickly while in beta. On a smaller page that is rendering correctly, if you change languages and then go back to the original, it should quickly give you the results from the cache. But I think the set-up is changing this week anyway. [22:45:04] Also, clicking on functions now redirect to the main page of Abstract Wikipedia and not to Wikifunctions. [22:48:24] Yes, that’s “ready to deploy” *T420670.* (re @Csisc1994: Also, clicking on functions now redirect to the main page of Abstract Wikipedia and not to Wikifunctions.) [22:49:31] Excellent. Another point. When I try to enter a QID, it does not return what I am looking for. [22:50:51] https://tools-static.wmflabs.org/bridgebot/bdc94180/file_79080.jpg [22:51:02] A huge problem [22:53:42] You can just type in what it is or alternatively expand and input the ID [22:54:24] Same problem. I type "research". No results found. [22:55:19] When I write research on ElasticSearch in Wikidata, I find it as Q42240. [22:56:06] This means somehow that QID in function calls is not linked to ElasticSearch for some reasons. [22:58:23] It’s working for me on Abstract Wikipedia đŸ€·â€â™‚ïž (re @Csisc1994: This means somehow that QID in function calls is not linked to ElasticSearch for some reasons.) [22:59:04] Does it work now from your side. (re @Al: It’s working for me on Abstract Wikipedia đŸ€·â€â™‚ïž) [22:59:46] From my side, the Orchestrator, the Abstract Wikipedia search bar and the QID in function calls do not work. [23:00:07] https://tools-static.wmflabs.org/bridgebot/499d456e/file_79081.jpg [23:00:24] Where is this? Sometimes it was a bit slow for me, and I sometimes had to re-type it to get it to try again. (re @Csisc1994: Excellent. Another point. When I try to enter a QID, it does not return what I am looking for.) [23:00:47] I will try to refresh it and see. [23:01:17] I works now. [23:02:20] I just implemented my first multilingual section title. [23:03:16] I will be playing around and tell you what will happen. Thank you for your support. [23:04:06] Thank you for sticking with it! Good night. (re @Csisc1994: I will be playing around and tell you what will happen. Thank you for your support.) [23:04:18] You can use Z32145 (re @Csisc1994: I just implemented my first multilingual section title.) [23:06:06] Ah, but that has fallbacks 😏🎃 (re @u99of9: You can use Z32145) [23:06:50] Yes! It's a good thing for section titles. (re @Al: Ah, but that has fallbacks 😏🎃) [23:08:19] Perhaps they should be annotated, but it's much better than getting an error message. (re @u99of9: Yes! It's a good thing for section titles.) [23:09:50] Yes
 I’ll leave you to catch up on the suggestions 👋 (re @u99of9: Perhaps they should be annotated, but it's much better than getting an error message.)