[16:24:43] Hallo [16:24:56] A question about the term "instantiate". [16:25:04] It doesn't appear in the glossary, and perhaps it should. [16:25:34] The definition of "content, abstract content" currently says: [16:25:44] "Technically, an instantiated constructor." [16:25:55] Could it say "Technically, an instance of a constructor"? [16:26:08] Yes, that's the same thing [16:26:09] Yes (re @amire80: Could it say "Technically, an instance of a constructor"?) [16:26:29] Constructor is something abstract [16:26:54] For example, Denny Vrandečić is an instance of human [16:27:09] OK. And the definition of "typed list" says "A typed list takes one argument, the type that all the members of the list have to instantiate." [16:27:15] Or an instantiated human [16:27:48] This means every member of the list has to be an instance of the given type (re @amire80: OK. And the definition of "typed list" says "A typed list takes one argument, the type that all the members of the list have to ...) [16:27:53] Human is abstract here [16:28:18] Have to be instantiated (re @amire80: OK. And the definition of "typed list" says "A typed list takes one argument, the type that all the members of the list have to ...) [16:28:26] Passive, please [16:28:38] It’s carried out by user [16:28:43] Could it say "A typed list takes one argument: the type that all the members of the list have to be an instance of"? [16:29:01] (I don't necessarily mean to change it in the English text. I just want to understand it better for translation.) [16:29:12] Fair enough (re @amire80: Could it say "A typed list takes one argument: the type that all the members of the list have to be an instance of"?) [16:29:32] Yes, that preserves the meaning (re @amire80: Could it say "A typed list takes one argument: the type that all the members of the list have to be an instance of"?) [16:29:38] Thanks :) [16:29:47] You're welcome [16:31:33] It's not that I dislike the idea, it's just that we don't have plans for that currently (re @Channel_Bot: @vrandecic Help me remember, please, what task from plan is related to make folksonomy / ontology of functions?) [16:49:28] It's on the list of stuff we're considering to work on for the next half year. Currently there's an internal API which can be used for experiments. (re @Channel_Bot: Do Wikifunctions have a public API? [16:49:29] [16:49:30] Like this, but from my server, not from Wikimedia project: [16:49:32] https://www.mediawiki.org/wiki/E...) [17:04:24] No current plans. I'm not convinced it's anywhere close to a high priority right now. But if someone else wants to work on that, I'm happy to help with the design and requirements. (re @Channel_Bot: Does Wikifunctions roadmap has task for provide JSON-LD (and other RDF serializations) API? [17:04:24] [17:04:26] With simple GET-request like it's i...) [17:07:08] Thanks for the discussion about structured data about functions. I read it, and believe me, I'm a big fan of structured data. But I don't see it as a particularly high priority right now. We still haven't explored what we can do within the wiki to organize the functions. Once we have explored that we will be able to recognize what kind of structured data makes sense for our use c [17:07:08] ase. Before that seems to be putting the cart in front of the horse. [17:09:46] I think it would be a good idea to change the english text. if you're having to ask (as someone who speaks english fluently and has experience with technical things), I'm sure there will be plenty of people who don't find it easy to understand (I also find the way you phrased it much easier to understand, for what it's worth) (re @amire80: (I don't necessarily mean to change it i [17:09:47] n the English text. I just want to understand it better for translation.)) [17:21:33] That is very strange indeed. I cleared the cache, and it doesn't replicate anymore. I can't see where this might have been coming from. If you or anyone else can recreate such behavior, please let us know. For now, I just can say "huh" (re @Channel_Bot: Very strange behavior with different results from call is_snake_case function (Z10324) and its implementation (Z10330) 😮 [17:21:33] [17:21:35] 1. Run...) [17:28:41] Done. (re @Nikki: I think it would be a good idea to change the english text. if you're having to ask (as someone who speaks english fluently and ...) [22:55:41] I needed structuring data and working through the API (and not through the browser UI) at the moment only to work on the current tasks of refining, testing and combining functions. No getting ahead of ourselves, for now. I understand the critical path of the project and it’s just about eliminating bottlenecks when performing work, and not about adding a bunch of different new ( [22:55:41] possibly useful) features. [22:55:42] [22:55:44] I looked into the current developments (I really believe in the project and I hope everything will work out exactly according to the implementation of the original goal) and came to the conclusion that the impossibility of programmatic work with functions and too much dependence on old approaches (which repeat the old problems of previous wikiprojects) can now be quite a serious [22:55:45] limitation (bottleneck) of the project. Perhaps I'm wrong, please point out errors in reasoning. [22:55:47] [22:55:48] 1. For example, we identify very non-obvious function bugs (like here https://www.wikifunctions.org/wiki/Talk:Z10324), but we cannot perform automated testing (a certain integration test), and we cannot even write an external a script for mass testing of functions independently and in the ways that appear in the minds of the participating developers (the way they act in their projects). [22:55:50] [22:55:51] 2ю And we are now working at the level of primitive text and logical functions, such as the anagram of a word or converting strings to snake_case / kebab-case, and what will happen next when there are thousands of them in the main pool of functions, how will we simply not drown in them? If there is no public API (and I’m not even talking about providing cloud computing resourc [22:55:53] es for running functions, only a regular JSON (or immediately JSON-LD) API for accessing the description and composition of functions), then it is impossible to create a simple table with a list of functions and their status . [22:55:54] [22:55:56] In discussions, you cannot refer to a topic (url + header anchor), the results of launching functions (and parameters for launching) cannot be shared via a link. All this points to the acumen of a standardized, simple API, on the basis of which you can work, including with the interface, and modify it. The API-first approach is sorely lacking, because now, in order to change some [22:55:57] thing in the same interface, you will need to go through many stages and additionally load the performance bottleneck of the entire project - the main team, with those things that could already be done by third-party developers, and the community use for much-needed assistance in solving current problems blocking the project (for example, a large block on proof of concept that Wi [22:55:59] kitext can in fact be replaced with functions). [22:56:00] [22:56:02] Please take my comments not as criticism of the project, but as positive criticism of some points regarding the direction of action, which is now being formed, and therefore determines its future fate and timing for the coming months (years?). I will be happy to discuss something on this topic and am ready to get involved in the development of simple, but currently missing improv [22:56:03] ements, if there is any technical opportunity to do this. (re @vrandecic: Thanks for the discussion about structured data about functions. I read it, and believe me, I'm a big fan of structured data. Bu...)