[00:32:25] What are the other benefits of enumeration? To be worth making hundreds of persistent objects, they need to do something better than validation of an in-range natural number would achieve. (re @Al: It’s only one of the points… We didn’t have dropdowns until after [[Z16098]]. It is odd, though, that specifying a type does not...) [02:33:30] I've also failed in Python: Z22256 (re @Al: Yes. It didn’t seem to work for me in Python but that was probably just bad debugging 🙄) [02:48:38] I improved it, and it's now giving the correct result, but now it's double wrapping in a Z22 for some reason which is causing an error (re @u99of9: I've also failed in Python: Z22256 (after U:Mohanad mentioned difficulty on project chat)) [03:17:14] Keep trying if you want, but otherwise I'd prefer to revert to the pure JSON so we learn how to do it in parallel with JS. (re @Feeglgeef: I improved it, and it's now giving the correct result, but now it's double wrapping in a Z22 for some reason which is causing an...) [03:37:20] That boils down to [[Wikifunctions:Representing identity]] (or, if it doesn’t, it should). I explored a lightweight alternative in [[Wikifunctions:Type proposals/Typed string]]. “Typed number” could be something similar, but that would be less efficient because of the type conversion. (re @u99of9: What are the other benefits of enumeration? To be worth making hundreds of [03:37:20] pe [03:37:21] rsistent objects, they need to do something better ...) [03:38:48] It most definitely should not be called "Typed string?" (re @Al: That boils down to [[Wikifunctions:Representing identity]] (or, if it doesn’t, it should). I explored a lightweight alternative ...) [03:38:54] That doesn't even mean anything [03:39:24] All strings have type string [03:48:47] Yeah, so a “typed string” is a sub-type of string 😏 (re @Feeglgeef: All strings have type string) [03:49:58] Right, but a brick house is not a sub-brick of house, it's a house made of bricks. (re @Al: Yeah, so a “typed string” is a sub-type of string 😏) [03:50:37] A typed list is a list with a type [03:52:28] And a typed string is a string with a type… but I’m not attached to the label. (re @Feeglgeef: A typed list is a list with a type) [03:53:11] Not in the Wikifunctions sense of the type, which is way too ambiguous (re @Al: And a typed string is a string with a type… but I’m not attached to the label.) [04:03:56] Back to this. Proton counting is not some external authority constraining us, unless we are worried that God will change his mind. So the remaining benefit seems to be that WF could then have callable labels and descriptions for every separate element in every separate language. While true, I expect that kind of data should be hosted and called from Wikidata (unless [04:03:56] we are in dis [04:03:57] pute about what languages we serve). (re @Al: That boils down to [[Wikifunctions:Representing identity]] (or, if it doesn’t, it should). I explored a lightweight alternative ...) [04:18:47] I agree. We could just go with object + reference + function (as in “Typed string”, with Z1 instead of Z6) where reference can be a Z2 or Wikidata ID. It doesn’t matter to me whether Oxygen is represented by “O” or 8 or Q629, they are all references to the same thing (in most contexts). (re @u99of9: Back to this. Proton counting is not some external authority constraini [04:18:47] [04:18:48] ng our options, unless we are worried that God will chang...) [04:23:16] I see a constrained Wikidata box to be more useful here (re @u99of9: Back to this. Proton counting is not some external authority constraining our options, unless we are worried that God will chang...) [04:23:56] For example, if you typed 8 in that constrained box, Oxygen would be the first thing to pop up [04:24:35] T385895 is the task [04:27:44] Not very relevant but won't this certainly kill everyone on earth? Then it also wouldn't be a problem we have to solve (re @u99of9: Back to this. Proton counting is not some external authority constraining our options, unless we are worried that God will chang...) [09:01:53] What do we expect will happen in the future when the community decides we've made a bad decision about a Type? [09:08:48] It depends on what’s bad about it. Worst case scenario is that we define a better type and perhaps deprecate the original. A huge pain, potentially, but we would have to take it on a case by case basis. (re @u99of9: What do we expect will happen in the future when the community decides we've made a bad decision about a Type? I fear some of ou...) [09:14:44] Yes I suppose we can't know until it happens and we build some experience at that. [09:24:35] Not too much experience, I hope! We got a taste with Z1-typed lists and lists with type-converters, but I think wrapping deprecated types, function by function, would be more straightforward than that, because the existing implementations should not start behaving differently. (re @u99of9: Yes I suppose we can't know until it happens and we build some experience at that.) [11:06:08] It's a nice reminder of what Identity is meant to help with. But that document doesn't say which of the models we ended up going with! (re @wikilinksbot: [[Wikifunctions:Representing identity]] [11:06:09] [[Wikifunctions:Type proposals/Typed string]]) [11:36:34] We can't see all tests on the function page anymore: : https://tools-static.wmflabs.org/bridgebot/8868042c/file_68496.jpg [11:50:29] We can see the first fifty 🤷‍♂️ (re @u99of9: We can't see all tests on the function page anymore:) [11:51:43] Oh yeah... ok that's enough :) (re @Al: We can see the first fifty 🤷‍♂️) [12:39:23] If anyone wants to do similar for a different language, I found [[:w:List_of_numeral_systems]] very helpful (re @u99of9: I just switched the natural number reader Z14290 to a language configuration (first to help Malayalam). It quadruples the call d...) [13:59:58] None of the above 😏 (re @u99of9: It's a nice reminder of what Identity is meant to help with. But that document doesn't say which of the models we ended up going...) [15:15:31] [[m:Wikifunctions]] is now available for translation. [18:58:44] Could Wikilambda define a function that handles Wikidata value parsing, then a Wikifunctions Staff member could make it a composition of a community function? [18:58:56] @vrandecic [19:00:38] I generally agree with most of the type decisions made while I was registered, the only one of note would be the floating point special values. (re @u99of9: What do we expect will happen in the future when the community decides we've made a bad decision about a Type? I fear some of ou...)