The APL Orchard

« first day (2355 days earlier) ← previous day next day → last day (306 days later) »

Schiphol

8:04 AM

@RubenVerg Right, because ⍺ is already to the left of the whole thing, right?

2 hours later…

Schiphol

9:35 AM

If I want to divide some array in groups of n members I can do

(1=n|⍳≢test)⊂test

I've been trying to figure out how not to repeat "test" here. Some combination of jot and right tack or something, but have been unable to. How's it done? And perhaps some general lessons for applying combinators in this kind of situation?

5 hours later…

RubenVerg

2:49 PM

⊢⊂⍨1=|∘(⍳≢)

Schiphol

@RubenVerg Thanks. How does that work? taking ⍺=n and ⍵= test does not seem to.

RubenVerg

I usually write trains starting from the inner part of the function - first I had ⊣|(⍳⍤≢⍤⊢), which trivially converts to |∘(⍳≢), then I added the 1=, then made it into a fork (1=|∘(⍳≢))⊂⊢ and then decided the commuted version looked nicer with one less set of parens

      3 (⊢⊂⍨1=|∘(⍳≢)) ⎕a
┌───┬───┬───┬───┬───┬───┬───┬───┬──┐
│ABC│DEF│GHI│JKL│MNO│PQR│STU│VWX│YZ│
└───┴───┴───┴───┴───┴───┴───┴───┴──┘

did I misunderstand what you want?

Schiphol

oh, parens around the whole thing, of course.

no, no, you perfectly understood. I just slapped my inputs to both sides and ofc it didn't work

RubenVerg

ye sorry I'm used to doing F gets whatever and then calling F, would've had parens in the code otherwise :)

Schiphol

RubenVerg

2:59 PM

I think there's a way to do away with the parents around the 2-train to the right - might work on that later

oh I'm dumb

⊢⊂⍨1=|∘⍳∘≢ works

lmk if my explanation for how I came up with it isn't clear enough - I can try to be more detailed if you wish

I remember being in your place except I didn't know this place was a thing :)

okay saying this sounds like I'm a train god - I've just had slightly more time to learn than you

Schiphol

I'll work on your explanation for a while and let you know, thanks :)

RubenVerg

Schiphol

this feels distinctly like train divine intervention to me, yep

I feel that I try to write stuff in a right-to-left, first do this then do that fashion and the nicer versions, like yours there, do not quite work like that

RubenVerg

lmao give me a dfn and I shall tacitify it

I'm basically that bot in the Haskell IRC that converts lambdas to combinators

Schiphol

hahaha

6 hours later…

dzaima

9:02 PM

@RubenVerg as a "hint" that it should be stable - if it weren't, there'd be no need to have both ⍋ and ⍒, as ⍒x would just be ⌽⍋x. And of course also that grade being stable is very useful

2 hours later…

RubenVerg

11:05 PM

you could say the same about > and <⍨ /s

I know that they are surely stable everywhere, this was just a long winded way of saying "this hint requires some assumptions the docs don't state"

I was also curious to see if any other dialects mentioned stability in the documentation

so I just made that list in case anyone else cared

« first day (2355 days earlier) ← previous day next day → last day (306 days later) »

all rooms