[05:57:24] Links showed in the stream: [05:57:25] https://www.wikidata.org/wiki/Wikidata:Lexicographical_data/How_to_help/WordNet [05:57:26] https://www.wikidata.org/wiki/Wikidata:Lexicographical_data/How_to_help/Concepticon (re @amire80: On one hand, thanks. On the other hand... there are people who are good at learning from two-hour videos, but I'm not one of the...) [12:47:22] I wouldn't say there are conclusions, but this is a work in progress [12:47:23] The question at hand is one that broaches multiple issues though - while agree it is useful to work out how to model senses like 'recently' around linked items, the semantics of prepositions and other grammatical words are challenging to make direct translations with between languages which aren't closely related. It would be more effective to work out how to express [12:47:23] the same inf [12:47:25] ormation without requiring equivalent grammatical words in different languages. (See also the problem with the discouraged 'of' property in Wikidata - many languages don't have an equivalent to 'of') (re @amire80: I think that @mahir256 and @bgo_eiu thought about this, but I don't remember their conclusions.) [12:54:03] I haven't been following the progress with Abstract Wikipedia as closely as I'd like to, but I would say broadly discussions regarding the use of lexemes for it could benefit from a 'zoomed out' approach to the problems at hand. For example, it is easier to talk about how to make a statement to the effect that "Dublin is a city in Ireland" in both Arabic and Irish [12:54:03] than to try and [12:54:04] work out how to translate "is" into Arabic and Irish [13:11:29] New type implemented from the Type proposals list: Complex numbers ( https://www.wikifunctions.org/wiki/Wikifunctions:Catalogue/Number_operations#Complex_numbers ) — If you play around with it before the announcement this week, let me know about any errors [13:32:23] I'm working on a display function (re @vrandecic: New type implemented from the Type proposals list: Complex numbers ( https://www.wikifunctions.org/wiki/Wikifunctions:Catalogue/...) [13:33:57] Z33219 [14:32:09] I've boldly changed the name to "complex128," encouraged by my rename to "float64" which is still in place 15 months later. As always, everybody is welcome to revert and discuss, as if anybody objects we should decide on a name sooner rather than later. [14:35:31] I think that’s broadly the plan. We don’t “translate” prepositions (in particular), we interpret the relationship between Wikidata items and choose how to express such a relationship in a given language. Even in English, locating a place “within” a place may or must have different surface forms depending on the context. Dublin is a city… *in* Ireland, but *on* the i [14:35:31] [14:35:32] sland of Ireland, a Europe*an* city, *on* earth; Washington, D.C is the capital *of* the United States and the country*’s* capital. A “within” relationship may be expressed with different Wikidata property types, but the extent to which these should be unified by sharing a common set of functions is, to my mind, very much an open question. (re @bgo_eiu: I haven't been [14:35:32] follo [14:35:34] wing the progress with Abstract Wikipedia as closely as I'd like to, but I would say broadly discussions reg...) [14:56:59] 🤔 Isn’t it more like “Complex number (float64)”? The number is “∈ ℂ” and approximated by two float64 instances. (re @Feeglgeef: I've boldly changed the name to "complex128," encouraged by my rename to "float64" which is still in place 15 months later. As a...) [15:34:52] I suppose (re @Al: 🤔 Isn’t it more like “Complex number (float64)”? The number is “∈ ℂ” and approximated by two float64 instances.) [15:35:19] I've started [[WF:CLANKER]] as a draft for an AI policy [15:51:44] I'm not entirely sold on using a "derogatory" word as an alias for the guideline [15:52:40] Would [[WF:AI]] be ok? I didn't want to take it in case anyone else wanted to use it. [15:53:35] it would be better, yes [15:54:29] I think the emphasis is misplaced: instead of focusing on dissuading people from using a tool (dissuasion unlikely to work on people excited about those tools), I suggest it would be more effective to focus on the human contributor's ultimate responsibility for every edit, and on repercussions of repeatedly making unconstructive edits or creating a lot of cleanup [15:54:29] work for the community. [15:58:47] Perhaps I should rename it to "Responsible use of Artificial Intelligence" or something along those lines? (re @abartov: I think the emphasis is misplaced: instead of focusing on dissuading people from using a tool (dissuasion unlikely to work on pe...) [15:58:47] And thus, maybe the "AI" page should actually more simply express the notion (I am not taking to time to phrase it nicely just now) that it's acceptable to use AI code generators in one's contribution process, but YOU ARE RESPONSIBLE etc. [15:58:49] In other words, start with a positive acceptance of people's desire to use AI. But then explain what we consider responsible use. [15:58:50] Detailing the failure modes of AI can come later, for those truly failing to imagine what could possibly go wrong. (And they are legion...) [16:27:25] It should really build on, clarify or extend our existing policies and guidelines. For example, if you use natural language text in a test case, you should very openly declare when your fluency in that language is limited. If you adapt code from another source, you should identify that source, whether or not intellectual property rights apply… (just off the top of my [16:27:26] head). (re [16:27:26] @abartov: And thus, maybe the "AI" page should actually more simply express the notion (I am not taking to time to phrase it nicely just n...) [17:23:56] I suppose we could then just unify everything into an "Editing Guidelines" page? (re @Al: It should really build on, clarify or extend our existing policies and guidelines. For example, if you use natural language text...) [19:57:37] Might as well start unified and keep them tightly focused. Maybe also “Content policy and guidelines”. [19:57:37] For natural-language content, however short, verifiability (loosely interpreted) needs careful consideration. [19:57:38] Real-life examples from the public domain are preferable, and appropriate attribution is recommended. [19:57:40] Original content is also permitted (in which case CC0 applies). [19:57:41] In either case, editors must be sure that the example is idiomatic in the specified language and are encouraged to indicate (in the edit summary, for example) whether it is in any way atypical. [19:57:43] Content must be eligible for release under CC0. Material subject to copyright must not be included. The concept of “fair use” does not apply, as all contributions are dedicated to the public domain. [19:57:44] (To clarify this, although copyright material can, in general, be included within content dedicated under CC0, it must be clearly marked and attributed, so that licensees “can reuse the material as expected. Licensors should clearly mark any material not subject to the license. This includes other CC-licensed material, or material used under an exception or limitation to [19:57:44] copyri [19:57:46] ght.” [excerpt from https://creativecommons.org/publicdomain/zero/1.0/legalcode.en]) (re @Feeglgeef: I suppose we could then just unify everything into an "Editing Guidelines" page?) [19:58:41] Thinking about the language, do we have babel boxes yet? (re @Al: It should really build on, clarify or extend our existing policies and guidelines. For example, if you use natural language text...) [20:04:26] We do, I have some on mine (re @Jan_ainali: Thinking about the language, do we have babel boxes yet?) [20:06:25] I've created [[WF:Editing guidelines]] [20:06:34] Feel free to add