[04:17:32] FYI I had some success at writing the original function you were trying for in the Volunteers' corner: Z23604 [04:21:21] The key insight was to use Z12676 to simplify the lists of lists. We've had that function for a long time, but this is only the second use in a composition. [08:01:04] I’ll probably add a comment to *T390226*. The “variadic” functions should avoid the need for typing and untyping lists, if possible. This looks like a persuasive use case, when we consider languages with multiple inflections. I made a start with [[User:GrounderUK/Tagging implementations]] and we should probably tag Z22491 as “pseudo variadic” (using something more intel [08:01:04] [08:01:04] ligible like “contains Wikidata reference constant Q146786” as an alias). [No time this morning.] (re @u99of9: FYI I had some success at writing the original function you were trying for in the Volunteers' corner: Z23604) [11:39:04] It looks like I didn't need quite as much typing and untyping as was in the version you commented on. But I agree it's worth being careful about the return typing of anything that generates lists. (re @Al: I’ll probably add a comment to T390226. The “variadic” functions should avoid the need for typing and untyping lists, if possibl...) [12:16:26] Yes… *T367005 *should go without saying (though I’ll say it) but *T370028 *will need some thought. We can only untype a list currently because map does not do so. We should consider inferring the resultant list type from the return type of the mapped function, but whether we should type the list heuristically when the function’s return type is Z1, I’m not sure. But I’ll [12:16:26] [12:16:27] mention that on the ticket… (re @u99of9: It looks like I didn't need quite as much typing and untyping as was in the version you commented on. But I agree it's worth bei...) [14:39:44] 1GT95