The APL Orchard

12:47 AM

@arcfide yes I'd say I"m familiar with the basic APL syntax, what I'm unfamiliar with is certain operators, and how operators are used. I can read a train just fine. I just often don't know this operator or that operator, like how right tack is used, or the various quad operators are used.

What's more is that I don't understand number theory to the extent that it would seem required to write APL effectively, because as it appears from attempting to write "vector oriented" solutions, as opposed to functional, recursive, iterative, or object oriented solutions, one would need a firm grasp in number theory or some other form of category theory or linear algebra

dzaima

7:49 AM

@nathanrogers nope. Until recently I hadn't gotten formally taught vectors (still haven't gotten taught matrices) but I still understood APL.

I mean sure, I have no idea what the hell a cross product is, but that doesn't seem to be required to write APL efficiently.

Adám

9:04 AM

@nathanrogers I suppose you by operator mean glyph (or in APL terms: primitive), and not APL's specific use of the word. Obviously, you'll have to get familiar with the most common primitives, but that comes fairly quickly for the most common ones. ⊢ Is actually the very simplest of all functions. It ignores any (though optional) left argument and returns its right argument; equivalent to {⍵}. It is use to access the right argument in a tacit function, and to separate arrays.

dzaima

to add to the above - ⊢ is equivalent to {⍵}. It is. There's nothing special about it (nor about {⍵}, it's a regular simple function). It is that simple

Adám

@dzaima CMC: Given a black box function which is either ⊢ or {⍵}, determine if it is ⊢ (i.e. return 1 or 0).

dzaima

@Adám o.O {⍵}⍣¯1⊢1 doesn't error?!?

Adám

@dzaima Hm, that may be because {⍵} and ⊢ are actually the same internally. I'll log the issue. Thanks!

@dzaima I'm down to 13.

dzaima

9:29 AM

no idea what the hell you're doing, I have 18

Adám

@dzaima My 13 is a tradfn. I've also got a dop (which needs a dummy right arg) at 16. What have you got?

ngn

@Adám does "black box" mean no ⎕cr?

Adám

@ngn No, but maybe it should. If I can't use ⎕CR then I need 16 for both tradfn and dop.

ngn

1=≢∊⎕cr'f'

Adám

@ngn You need f←⎕⋄ before that, no?

ngn

9:35 AM

@Adám idk, you posted the cmc :)

Adám

@ngn You can save one byte: ∊⎕CR → ⎕VR

Ok, and how about if you can't read the code?

ngn

right

Adám

Got 15 (including getting the input).

ngn

{0::0⋄⎕⍣¯1⊢1}0 ?

Adám

@ngn dzaima found that {⍵}⍣¯1 is allowed. I logged that.

ngn

9:41 AM

@Adám in the version i have it errors

Adám

@ngn What version is that?

ngn

@Adám 15.0.29007

i know... to old, never mind, I'll try to think of something else

Adám

@ngn {⍵}⍣¯1⊢0 doesn't error in my 15.0.29412 or 14.1.33725

ngn

@Adám oh wait... it behaves differently in the session?

Adám

@ngn … than in …? (love your chat animations)

ngn

9:45 AM

@Adám as a program, with "-script"

@Adám :) markdown hates me

Adám

@ngn What exactly are you running?

ngn

#!/bin/bash
(echo ∇M;tail -n+3 $0;echo -e '∇\nM\n⎕off')|dyalog -script;exit $?
⎕io←0

f←{⍵}
{0::0⋄f⍣¯1⊢1}0

Adám

@ngn OK, can repro, but I don't think it is -script that makes the difference.

ngn

huh! i've just discovered that that works in the repl too, but when I replace f with ⎕, it behaves differently

Adám

@ngn Exactly, I was just about to post: Try it online!

Anyway, let's see if you can figure out my 15 or 16 without relying on this…

Hint: I rely on queryable characteristics which are not source code.

ngn

10:01 AM

@Adám ⎕nc?

Adám

@ngn Yup, that gave me my 16er.

ngn

f←⎕⋄3.3=⎕nc⊂,'f' ⍝ but that's too long

yet golfable

Adám

No, that's the exact 16 bytes I found.

@ngn Really?

dzaima

15

Adám

@dzaima Nice. That is what I got too.

dzaima

10:05 AM

14 with swapped 0 and 1

ngn

10:36 AM

@Adám 0⍴# is a nonce error. Are we supposed to be able to do ⍬⊢¨#?

Adám

@ngn 0⍴# isn't (by me) and shouldn't be a nonce error.

ngn

strange... 0⍴# is not a nonce error on tio

grr... life's too short for this

Adám

@ngn I'd say upgrade, but then again, it works in my 15 too.

ngn

it's my mistake - what i do to 0⍴# after that causes an error; I always get into this trap with namespaces...

H.PWiz

1:49 PM

@dzaima Why should it error?

dzaima

@H.PWiz I don't think it should error, but it's just that every other dfn errors.

H.PWiz

Yeah, there is special code for {⍵}⍣¯1 in the interpreter. Presumably because its an idiom. It does seem wrong that f←{⍵} ⋄ f⍣¯1⊢0 fails though

dzaima

@H.PWiz I see no reason for that to be an idiom as it's equivalent to ⊢⍣¯1 which is equivalent to ⊢ though..

H.PWiz

I mean that {⍵} is an idiom

dzaima

@H.PWiz ah. Why is it an idiom though? ._.

I understand {⍺} being an idiom since that's not ⊣, but {⍵} literally is ⊢

H.PWiz

2:06 PM

No idea tbh

Adám

{⍺} and {⍵} were made idioms before ⊣ and ⊢ were added to the language. It was done to speed up by bypassing parsing the dfns' code.

dzaima

ah, that makes sense

I understand wanting to do that as my APLs performance of 1+1+1+1+1+1 is almost the same as 10000⍴123456 :P

Quintec

@dzaima Huh, why?

Do you mean that as in “both are too fast to measure” or what

dzaima

2:22 PM

@Quintec both are equally slow, tested by timing each evaluated 1000 times. I didn't bother making the parser fast :p

Quintec

I’m still confused as why that is

dzaima

@Quintec why 1+1+1+1+1+1 is so slow? because I didn't make it fast

Quintec

But 5 operations should be faster than 10000, no?

dzaima

@Quintec the 1s are jumping around this whole file for each character of that, and instantiating objects left and right. ⍴'s just got these 4 lines to worry about.

also this too

it's not the math that's slowing everything down. It's literally everything else that has to happen.

for each character it goes trough this while loop at least once (every other char twice I think), and each is call is this thing which goes trough the given string and inefficiently goes trough a LinkedList.

ngn

2:40 PM

sidenote: k has a system command \t:1000 expr to time 1000 evaluations of an expression. I think all languages should have that, it's very handy.

Adám

]runtime -compare 1+1+1+1+1+1 10000⍴123456

Dyalog APL

@Adám

  1+1+1+1+1+1  → 1.7E¯7 |     0% ⎕⎕
* 10000⍴123456 → 3.5E¯6 | +1944% ⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕

ngn

dyalog has "cmpx" (if I remember correctly) which can compare the performance of multiple expressions and it's smart - it increases the number of iterations exponentially until the time goes over a certain threshold

dzaima

@ngn my APL: 100 ⎕htime 'expr'

Adám

@dzaima h?

ngn

2:43 PM

@Adám good to know, i never learnt to use those ucmds

dzaima

@Adám human. It prettifies to milliseconds/seconds/nanos. Just ⎕TIME gives out a nanosecond number

Adám

@dzaima Yikes, quad-name proliferation…

Erik the Outgolfer

@ngn hmm...yeah

ngn

@EriktheOutgolfer yeah, yeah, but it's a habit

also, ucmds are awkward to use from a script file

Erik the Outgolfer

ucmds are useless from a script file :P

ngn

2:45 PM

"⎕SE.something something..." I can never remember that

Adám

⎕SE.UCMD'…'

And if you want to avoid doubling quotes:

result←⎕SE.UCMD⍞
…

ngn

@Adám hm... i remember seeing much longer paths, with at least two dots after ⎕se

Adám

@ngn There are a lot of things in ⎕SE. Maybe you're thinking of ⎕SE.Dyalog.Utils.repObj?

ngn

unfortunately ⍞ can't be used in a script

@Adám yeah

so, I learnt something new today... let me try it

VALUE ERROR
M[6] ⎕SE.UCMD']runtime -compare 1+1+1+1+1+1 10000⍴123456'
    ∧

(this is from a file which is piped through dyalog -script)

Adám

@ngn Works by me, but try ⎕SE.UCMD'result←runtime …

ngn

2:52 PM

@Adám same thing - value error

Adám

@ngn OK, so you're trying to capture a result. I'll add ⍬⊤⍬ to 17.1.

ngn

@Adám how can I not try to capture a result?

I understand my setup is unusual, by the way, this is not to rant against ucmds

dzaima

@ngn this mess, maybe?

ngn

@dzaima that worked! :)

⎕io←1 ⍝ doesn't work with 0
(⎕NS⍬).⍎'⎕CY''salt''⋄⎕SE.UCMD''←box on -trains=tree''⊣enableSALT⋄⍬'
(⎕NS⍬).⍎'⎕SE.UCMD'']runtime -compare 1+1+1+1+1+1 10000⍴123456'''

dzaima

that's in my "important.txt" file :p

ngn

3:01 PM

I guess the "box on" part was not essential, only the ⎕cy'salt'

Adám

@ngn Btw, you can golf the first . and the ] and ompare

ngn

@Adám first: make it work, second: golf it :)

I don't understand why, but the "box on" part is essential - it fails without it

oh, hang on... it's the "enableSALT" that matters

sneaky :)

@dzaima what else is there that you are willing to reveal? :)

@Adám golfed: (⎕NS⍬)⍎'⎕io←1⋄⎕CY''salt''⋄enableSALT⋄⎕io←0⋄⎕SE.UCMD'']runtime -c 1+1+1+1+1+1 10000⍴123456'''

dzaima

@ngn an old version of my fds, JS APLifiying functions, a python program to transpile BF to C++, some tio links to stuff for homework, and more TIO links. I think that's really the only important thing there :p