@Uriel True, but dzaima's solution is also O(n²). Doesn't matter in this case. However, it wasn't the solution I was leading up to in the preceding chat.

@HyperNeutrino Very nice. (⍵∊∘.×⍨⍵)/⍵ is kind of a no-op. You are getting all the values that are in the multiplication table, but you already have them – in the multiplication table. The fact that there are too many in there doesn't matter, as they are both composite and too large, so {⍵~∘.×⍨⍵}∘{1↓⍳⍵} is enough.

It can be golfed further by going tact.

@Adám happy new year! 😊

@Uriel You too!

@Adám Ah okay, thanks! By going tacit, do you mean not having so many weird w's or something? I don't quite get the difference between tacit APL and (non-tacit ?) APL yet :P

@HyperNeutrino Yes. I don't have time to teach now and until next week, but have a look at:

Ah okay, thanks! :D

@Adám as this was my first attempt at doing something useful in APL (or any language like this) I tried to not look at clues in the chat as I wanted to see what I can do. If I knew what ~ does as a dyad, then it'd have been the obvious solution, but I had fun doing this without it :D

my first attempt was creating a modulo table, but I couldn't find the modulo operator quickly (I didn't think residue could be it), and decided that the way I did it is more fun