@Quintec :D

lol

⎕←∨/(⊢⍷⌽)

@Quintec
  ┌─┴─┐
  / ┌─┼─┐
┌─┘ ⊢ ⍷ ⌽
∨
SYNTAX ERROR

@Quintec
0

@Quintec
1

hm

⎕←(⊢≡⌽)'testset'

@jslip
1

More functions I had no idea existed xD

:)

oh Dyalogs prototype mess does more than I thought it would. Not sure whether that's useful enough to bother implementing

@dzaima I'm facing the same dilemma in k :) though I've already implemented the mess and I'm not sure how to simplify it in a consistent way

@ngn I do have the basics of prototypes already done too, it's just that everywhere I'm making arrays with new Arr(items, shape) and I need to add a prototype set to that too..

all I wanted to do was implement ↑ :p

@dzaima that sounds like overkill. You only need to make sure empty arrays have 1 hidden element.

@ngn oh that's a smart way to do that.. I just have an extra property for the prototype ಠ_ಠ

@dzaima that's duplication of information; for non-empty arrays the structure of ⊃A determines the prototype - just replace everything recursively with 0s and spaces

but even that is tedious...

@ngn it doesn't feel nice to break (×/shape) = ≢items though

@dzaima (1⌈×/shape) ≡ ≢internalItems

I don't like prototypes either

for ngn/k I'm thinking of introducing a compromise between the APL way and what's rumoured to be in k6

typed empty arrays remain as they are, with their implicit prototypes; generic empty arrays lose their prototypes (it will be something fixed, doesn't matter much what)