[06:03:39] For starters yes (re @Nikki: does "We are aiming to do this in Vector 2022." mean it's only going to be for that skin?) [09:27:56] oh. that doesn't sound very helpful :( (re @Nightrose: For starters yes) [09:35:51] by which I mean I only know of two editors who use that skin, since it's not the default and there's very little reason to switch to it [10:13:21] I think it's inevitable that it'll become used mre and more and at some point default (not my decision, just what I expect currently) so if we can just fix it for one it seems natural to fix it there. But I hear your point and I'll look into what it'd take to also make it work in Vector. (re @Nikki: oh. that doesn't sound very helpful :() [11:21:47] to be honest, I can't see it being made the default any time soon, given that it's the only skin which explicitly does not support mobile devices and people on mobiles have to use the desktop site to edit wikidata [13:09:09] I'll try to get some info what the plan is. [19:13:15] When you speak about the steps, what are you referring to? Sorry, you seem to have some specific context in mind, and I lost that (re @marijke00: Where you want or ought to store the constructors also depends on what you envision or assume them to be. If they’re for step 1/...) [21:48:58] perhaps they are references to the steps shown in https://upload.wikimedia.org/wikipedia/commons/2/2d/A_proposal_of_an_NLG_architecture_for_Abstract_Wikipedia.svg ? (re @vrandecic: When you speak about the steps, what are you referring to? Sorry, you seem to have some specific context in mind, and I lost tha...) [22:24:46] If it is indeed that, then yes, agreed: morphosyntactic, phontactical, etc. functions should all be in Wikifunctions and not part of abstract content. Content is basically instantiations of the constructors. And where that should go, that's the question. But the whole realization pipeline, that turns content into text, would live as functions in Wikifunctions, and using the lexical and ontological data from Wikdiata.