[01:53:27] David I just hit my first case of really needing a statement that doesn't get fetched because it has a qualifier. The nearest test is Z26798, but I need this for Z25999. [03:45:18] Thanks for mentioning; I remember the previous exchange about this and that was one of the motivations for some current work - We are currently working on adding qualifiers, which of course means we'll start importing those statements that are currently omitted. I expect this work will be deployed in the next few weeks, and will have more to say about it soon. (re@u99of [03:45:18] 9: David [03:45:19] I just hit my first case of really needing a statement that doesn't get fetched because it has a qualifier. The nearest te...) [05:00:07] Even if our display functions do fail to complete quickly enough, I think the message "Please try again." is not the best. Instead it would help to suggest that the user expand the ">" so they can see the underlying value structure. : https://tools-static.wmflabs.org/bridgebot/6ef50c41/file_72213.jpg [07:19:16] I did some scratch programming some 15 years ago, and one thing it did better than the compositions do is how easy it was to see in which block you were operating in. Clear colors and sharper borders. In Wikifunctions, it is all the same background, no horizontal borders, with just some very thin lines showing how many levels in you are. Which really isn't easy to [07:19:16] see at a glance when you're more than three layers deep. I'm also not fond of the narrow view, screenshot form my small 14" laptop with 1920x1080 resolution. : https://tools-static.wmflabs.org/bridgebot/e8595cbd/file_72217.jpg [07:20:23] Even more fun on my iPhone mini 😏 (re @Jan_ainali: I did some scratch programming some 15 years ago, and one thing it did better than the compositions do is how easy it was to see...) [07:21:10] Oh, I found the button for wide screen. A bit better, but still narrow. : https://tools-static.wmflabs.org/bridgebot/8d6b98a5/file_72218.jpg [07:21:54] *T301418* (re @Jan_ainali: I did some scratch programming some 15 years ago, and one thing it did better than the compositions do is how easy it was to see...) [07:33:31] “Mobile first”: it’s better now we have shorter labels, but a “use ZIDs” option is worth considering. : https://tools-static.wmflabs.org/bridgebot/8bec7ce3/file_72219.jpg [12:36:21] I think I would hate ZIDs. How about a "don't use more than 50% of the space on vertical lines" strategy. (re @Al: “Mobile first”: it’s better now we have shorter labels, but a “use ZIDs” option is worth considering.) [12:42:51] 🤔 You wouldn’t have to use the option! Are you thinking of indenting only the first line or two of the label’s “paragraph”? (re @u99of9: I think I would hate ZIDs. How about a "don't use more than 50% of the space on vertical lines" strategy?) [12:44:48] My first instinct would be that the width of the indents should be ~10x thinner. (re @Al: 🤔 You wouldn’t have to use the option! Are you thinking of indenting only the first line or two of the label’s “paragraph”?) [12:50:28] Or the most deeply nested object might be the least indented (because that’s where the interesting action generally is). (re @u99of9: My first instinct would be that the width of the indents should be ~10x thinner.) [18:04:14] I was just looking at Q319 to see if it could be a nice example, but of course it turns out that the instance of that is much more complicated than just Q634 [18:15:08] "Stockholm är en storstad." but "Paris är en turistort." 😏 We could begin to define an order of precedence for returning statements, but those are treacherous waters. [18:15:56] it's actually a good example, the direct instance is not always the best one, we may need to find the "best" superclass (re @Jan_ainali: I was just looking at Q319 to see if it could be a nice example, but of course it turns out that the instance of that is much mo...) [18:16:44] Agreed. Also useful would be to fallback to P279 if there are no p31 (which is quite common). [18:18:20] but not impossible, it could be something simple to "go to the first superclass with a local interwiki link" (re @Al: "Stockholm är en storstad." but "Paris är en turistort." 😏 We could begin to define an order of precedence for returning stateme...) [18:18:49] That’s a really easy change to make. (re @Jan_ainali: Agreed. Also useful would be to fallback to P279 if there are no p31 (which is quite common).) [18:20:29] I was about to use Q255208 as a soon upcoming local example :) [18:23:31] "Kräftskiva är en fest." (re @Jan_ainali: I was about to use Q255208 as a soon upcoming local example :)) [18:24:30] Unfortunately, "Banan är ett bär." 🤔 [18:25:23] Technically more correct? (re @Al: Unfortunately, "Banan är ett bär." 🤔) [18:26:40] Feel free to amend your test case! (re @Jan_ainali: Technically more correct? [18:26:41] A banana is an elongated, edible fruit—botanically a berry[1]) [18:26:43] But if P279 is only the fallback, shouldn't banana stay a fruit? [18:31:14] Maybe. That’s what I was thinking about… 🤔 Yeah, it’s currently just a simple filter by property. I suppose that’s not what I intended, even though I called it “filter…” 😎 (re @Jan_ainali: But if P279 is only the fallback, shouldn't banana stay a fruit?) [18:49:50] ✅ fixed (re @Jan_ainali: But if P279 is only the fallback, shouldn't banana stay a fruit?) [19:01:37] I think this is slightly amazing! 🚀 [19:04:19] We need a fallback for when the class has no corresponding lexeme, but it’s a fun window on Wikidata content. (re @Jan_ainali: I think this is slightly amazing! 🚀) [19:05:45] Yes it is! This might also be an eye-opener for some Wikidatians of why sprinkling instanceofs and subclasses rather than just adding the most narrow one causes issues. (re @Al: We need a fallback for when the class has no corresponding lexeme, but it’s a fun window on Wikidata content.) [19:47:14] Q164 is an interesting example. The first P279 is more salient than the P31. That applies to instances of Q16521 too. 🤷‍♂️ (re @Jan_ainali: Yes it is! This might also be an eye-opener for some Wikidatians of why sprinkling instanceofs and subclasses rather than just a...) [19:50:24] Oh yes, taxons are special beasts. Luckily, they are plenty enough that specific functions for them are worth the time investment. [19:54:50] Tricky is the type of p31 of the square. It is two words, where I am not sure that it would be obvious to have a lexeme for the combination rather than two separately (L579835 L53934) [20:28:33] Is it possible to send someone a link to a function, with the input prefilled (so that the person who gets the link only need to click run to see the function in action)?