The Nineteenth Byte

moonheart08

17:00

I want it to be something i can modify when the challange is ongoing for adjustments

Evan Carroll

What's the best compiled golfing language?

NieDzejkob

why not just run the program and communicate with it by a pipe or files?

moonheart08

@nie

err one sec

@NieDzejkob Speed reasons. The 'update' will be called multiple times per tick, one for each 'robot'

Some solutions could make hundreds of the things

PhiNotPi

@moonheart08 Is this some sort of long-term KOTH?

moonheart08

Yes.

Longterm, as in mostly indefinite.

Unihedron

17:05

golf.shinh.org/p.rb?add5 ??

PhiNotPi

Ever heard of screeps?

Unihedron

someone got super short brainfuck

moonheart08

@PhiNotPi Congratulations, you figured out what i'm modeling it after.

:P

I bought the game. It's quite fun

I think it'd be a awesome experiment, and KoTH, to have a similarly styled challange.

PhiNotPi

Have you heard about the KOTH Server project?

moonheart08

Nope.

I'll look into it tho

Link?

PhiNotPi

17:07

Well there's a couple of people (me included) that are interested in building a website to host KOTHs.

41

Q: Introducing a KOTH server!

Note: Due to changes with RedHat's hosting service, the website is currently down. There are some (very) early-stage plans to create a bigger, better version. That being said, if you want to see a new KOTH server be built, please voice your support so that we can gauge demand / tailor it to th...

discussion

moonheart08

I'd be glad to help. I have a server I could lend in. (But be warned, it also runs a game server, so you can't use all the CPU)

Riker

quote of teh day:

> (If you can't see the animation, then please upgrade to not Internet Explorer.)

PhiNotPi

We already have hosting with Azure.

moonheart08

Oh, neat

PhiNotPi

But the website itself is in the planning stages.

moonheart08

17:09

So, perl and Java are the supported KoTH languages?

PhiNotPi

No, that's gonna change.

moonheart08

oh

You have a list

Add rust \o/

PhiNotPi

We had a list

moonheart08

Also, give each controller it's own user, maybe?

Automating user creation and management isn't too hard

PhiNotPi

The new website is gonna be .NET-based, and should support Python, Java, etc

moonheart08

17:11

But yea, my project is to work on a screeps like KoTH, but more suitable for PPCG (Less randomization, balancing, support for things not javascript)

Screeps' server itself is opensource, but it's not suitable for these modifications.

Unihedron

a template arbiter base would be very helpful for removing redundant work between different models

moonheart08

?

You mean a template program?

PhiNotPi

17:25

Anyways, we'd love to have you on the project. As mentioned, we're basically starting from scratch right now, but I think it would be perfect for this kind of KOTH.

moonheart08

Mk. One problem: I don't know any .Net languages

I mean, a framework for general compiled programs i could make and use

Also, have you considered using Usermode Linux for sandboxing?

It's an obscure function of the kernel that lets you run linux as a usermode program

Works quite well

It supports things like memory limits, blocking networking, and quite a bit more (anything the kernel can do by default). Only issue is that it's tough to set up, and it's obscure :P

I've used it before, and getting 4.12 to compile can be a bit of a pain. I don't know if later versions actually work

So it's an option, but not the best one

Plus, i have no idea how it fares against Spectre and Meltdown

Probably fails at spectre, with meltdown fixed for programs that mess around the UML to try and access higher processes

I mean, it is proven to work (See #esoteric's hackego bot)

Lets take this to the chatroom :P

PhiNotPi

C# isn't too different than Java. (I am pretty new to C# too)

moonheart08

I know enough java to know it's horrible. :P

J. Sallé

@moonheart08 C# is less horrible Java

moonheart08

Really? That'd be awesome

J. Sallé

17:35

(or more horrible if you hate microsoft as much as anyone should)

moonheart08

(I hate microsoft. :P)

J. Sallé

Oh well >.>

moonheart08

(Mono, however, has become somewhat of a savior for me, so many games use it i couldn't play some of my favorites without it)

J. Sallé

Jokes aside, C# is just like Java with some more stuff

Pavel

A lot more stuff

C# is awesome

moonheart08

17:36

I hated java's syntax.

so uh,

J. Sallé

Yeah the syntax is basically the same

moonheart08

I'll stick with Rust

throws a crab at C#

Pavel

@moonheart08 Mono is still Microsoft, you know.

moonheart08

Good point

Pavel

What's wrong with Java's syntax, anyway? It's basically C syntax.

moonheart08

17:39

everything is an object

everything

it gets annoying

NieDzejkob

why not use SELinux instead of user mode Linux?

moonheart08

Good point

Pavel

@moonheart08 May I introduce you to my friend F# in that case

moonheart08

looks it up

Maybe i can actually learn F#, unlike haskell lul

(Haskell just confuses me :P)

Pavel

@moonheart08 Unlike Haskell, F# isn't pure, so you can still do imperative things like loops and mutable variables.

moonheart08

17:42

Selfpings today: 1

Pavel

You can write F# without defining a class for the Main method, or as a script without a main method at all, more like python.

You can still go full C#-style OOP though

moonheart08

@NieDzejkob How might that be done?

I've never used SELinux :P

J. Sallé

Isn't F# copyrighted to Apple?

Pavel

@J.Sallé Wat

J. Sallé

I might be mixing languages up

Pavel

17:45

F# is a .NET language made my Microsoft

moonheart08

That's Swift, isn't it? :P

J. Sallé

Oh, it might be Swift

Pavel

Definitely swift

J. Sallé

Yeah it's Swift. I think I read somewhere that people were devving iOS apps in F#, that might've been it

Mr. Xcoder

Swift and Objective-C are Apple’s.

Wait I lied, Objective-C is used for Apple products, but it’s not made be Apple

@J.Sallé Probably those who attempt to use F# in Xamarin...

J. Sallé

17:52

@Mr.Xcoder maybe. It was a while ago in my college's newsletter, iirc

Adám

@Mr.Xcoder Stop hoping:

New Main Posts

0

Q: Split on first unique

Split on first unique Related. Given a printable ASCII string, split it into a list of non-empty strings with a new sub-string beginning every time a character, which has not previously been seen, occurs. Examples "mississippi" → ["m","i","ssissi","ppi"] 'P P & C G' → ["P"," P ","& ","C ","G...

code-golf string text-processing

Downgoat

@Mr.Xcoder it was originally not created by apple but NeXT got the rights IIRC and put their own guy to work on it

Mr. Xcoder

Yeah I relises that a bit later

Erik the Outgolfer

18:17

@Dennis hey, you commented after the answerer rightfully deleted their answer! :D

Dennis

One second after. :/

Stupid mod powers. :P

3

Erik the Outgolfer

powers getting uncontrollably heavy

DJMcMayhem

Who would want them? :P

NieDzejkob

@moonheart08 TIO uses SELinux and it's open source, look there :P

Erik the Outgolfer

@DJMcMayhem says the mod election candidate

moonheart08

18:20

Thanks @NieDzejkob

Erik the Outgolfer

One (1) count of moderatorial unfairness against Dennis (12012)...

Mr. Xcoder

@EriktheOutgolfer Take this, distinct sieve! :p

Erik the Outgolfer

saw that already

Adám

Is it just me, or is there something wrong with the user-script?

Mr. Xcoder

@Adám You have to refresh.

Erik the Outgolfer

18:25

@Adám the latter (widely known)

no, refreshing won't solve the issue

Adám

@Mr.Xcoder Hard (no-cache) refresh didn't solve it.

Mr. Xcoder

It does solve the issue for me.

betseg

wat

Mr. Xcoder

Adám

Never mind. Now it did.

Mr. Xcoder

18:26

Aww ninja'd. And I am sure I was on the page before it was posted, refreshed and the problem was solved.

Riker

@EriktheOutgolfer what's the problem?

is there a technical reason in teh userscript?

Erik the Outgolfer

it takes too much time to reflect new winners

Riker

that's... not a coding issue

that's just caching

Erik the Outgolfer

13 mins ago, by Adám

@Mr.Xcoder Hard (no-cache) refresh didn't solve it.

and I've experienced it too

Riker

waiting did

cache != cookies

iirc the userscript uses #2 for most things

Erik the Outgolfer

18:40

it shouldn't?

I feel like cookies would be inefficient for that

Riker

/shrug

I think it just uses cookies for everything, from settings to winners

Erik the Outgolfer

for settings that's good, for winners no

maybe it does it so that it doesn't fetch from all the pages every time?

mınxomaτ

Feb 12 at 19:01, by mınxomaτ

Since I just bought one, the remarkable tablet is $100 off today (it's a digital e-ink notebook that runs linux and open source software).

It arrived today (3 days from HK to Germany, not bad). Aaaand ... it's flippin awesome. It actually does feel like paper, but more like Biro pen on paper, not pencil.

Dennis

@Riker I sincerely hope you mean local storage.

Riker

oh looks like it does use ^

J. Sallé

18:44

@mınxomaτ Kindle-like? Sounds awesome.

DJMcMayhem

@EriktheOutgolfer Yep XD

Adám

@Riker Care to explain your edit?

Riker

@Adám it looks nicer?

and the bottom thing had a weird linebreak, I think you used a tab instead of a space

Mr. Xcoder

Only 1 challenge is currently featured, and it ends tomorrow. This sounds like a good time to start handing out the bounties for Best of

totallyhuman

how many at a time?

Riker

18:46

all of them

totallyhuman

no

Adám

@Riker VTC as subjective.

Riker

@Adám this is why instant edits are a thing /shrug

Mr. Xcoder

I guess we want to hand out 2 at a time, come over and we'll discuss it there (we already did but haven't really reached a consensus)

Riker

if you feel it looks actually worse feel free to roll back

Adám

18:46

@Riker instant edits?

Riker

@Adám vs edits needing approval

Adám

@Riker Sure, if you're fixing issues etc. But just stylistic changes?

Riker

@Adám again, if it looks nicer to me I'm likely gonna edit it

does it really matter, on the grand scale of things? not really

(for either way, whether I edit or don't)

Erik the Outgolfer

sometimes such edits are important...

Riker

I felt it was worth a quick edit

Adám

18:48

@Riker But if all followed your principle then we'd end up with style-edit wars.

Riker

if you feel it's worth rolling back because you think it made teh post worse, go for it

@Adám do you think it looks worse though

or are you complaining about the matter on principle?

SE kinda assumes people are mature and aren't going to do stupid edit wars /shrug

Adám

@Riker I'm asking about the principle. You may notice that that format is one I often use.

Riker

@Adám yep

Mr. Xcoder

Personally, I would align those arrows now that they're in a single block. IMO it's the best compromise between readability and ease of usage

Adám

@Riker So every time you see one of my questions you'd want to edit the examples' format.

Riker

18:50

on that note: can you not use the unicode arrow? it doesn't format right on my older mobile phone

totallyhuman

which post?

Riker

@Adám that is also correct

I generally don't tho

@totallyhuman codegolf.stackexchange.com/questions/156170/… -

Erik the Outgolfer

hm, in defense of Adám, you're actually supposed to edit as little as possible

Riker

yep

/shrug I felt it was worth editing

Mr. Xcoder

I'd also not really use unicode arrows because... fonts... but that's your preference

Riker

18:52

I was already there fixing something else (tab did a weird lack-of-linebreak)

so I edited the codeblock while I was there

it's honestly not that huge of a deal, it's not like I'm going through every question you've posted and 'fixing' them all

Adám

@Riker Yes, I'll try to remember to use > instead of → and refrain from &emsp; in my challenges. (I'll still use them in my answers, though. If your device can't render widely supported Unicode, you'll have trouble reading APL anyway.)

Riker

@Adám emsp?

Mr. Xcoder

CMP: Which is the least active SE site? Which is the least active beta SE site?

Riker

oh, emphasis space? TIL taht's a thing

@Mr.Xcoder well tech writers isn't making it past private beta

that may take the cake right now

@Adám it doesn't do APL, nope, jelly is a 50/50 shot too :p

Adám

@Riker No, M-space. A space with the width of a capital letter M.

Riker

18:57

huh

TIL

Mr. Xcoder

I find Vegetarianism remarkably inactive (I am not talking about those that will get closed into account)

Adám

@Riker SE apparently treats tabs as four spaces, so you can use leading tabs to markdown code. So I use &emsp;s to create indentation.

Erik the Outgolfer

is that the case anymore?

Adám

@EriktheOutgolfer Please ping.

Erik the Outgolfer

@Adám wasn't directed specifically to you...

Adám

19:02

@EriktheOutgolfer I meant "please ping what you're responding to".

totallyhuman

hehe i'm using the forall character for a foreach loop

Erik the Outgolfer

@Adám not everybody appreciates that in here...

Adám

@EriktheOutgolfer They surely do when it is completely unclear what is being responded to, no?

Erik the Outgolfer

@Adám if they did, then I would have, yes, but unfortunately that's not always true

Adám

@EriktheOutgolfer "is that the case anymore [that tabs=4 spaces]?"; "is that the case anymore [that Vegetarianism is remarkably inactive]?"; "s that the case anymore [that tech writers isn't making it past private beta]?"; etc.

Erik the Outgolfer

19:06

@Adám we tend to assume the reply goes to the latest possible message, and generally pings are avoided due to making a "ping" sound towards the owner of the message that gets pinged to, or it makes notifications

and some people don't like that when it happens many times in a short period

Οurous

@EriktheOutgolfer pings make a sound?

Erik the Outgolfer

yes

Riker

@Adám it doesn't actually allow tabs normally

it replaces them with 4 spaces

I assumed you broke it and pasted a raw tab /shrug

@Mr.Xcoder yep it is quite inactive

I was gonna say that next but I took a snack break

Adám

@EriktheOutgolfer Uh, silencing chat pings (without silencing your device completely) is two clicks away.

Riker

@Οurous if you have your sound on yes :p

Οurous

19:09

@Riker I do... I just... Oh the tab is muted. TIO pings make noise.

Erik the Outgolfer

@Adám the situation isn't so simple here; yes, sometimes such a sound is needed if somebody wants somebody else's attention, but hearing ping ping ping ping... annoys them a tad bit

Οurous

TIL my phone autocorrects TIL to TIO

Riker

@Οurous lol

moonheart08

TIL people's phones can actually autocorrect TIL to TIO

Οurous

@moonheart08 it gets better. Sometimes it corrects words to emoji.

Erik the Outgolfer

19:13

most probably Οurous just added "TIO" in the dictionary and slang terms like "TIL" are unrecognized

Riker

@Οurous that's horrifying

moonheart08

Someone is bad at designing autocorrect. It is terrifying

Οurous

@Riker it's 🦆 annoying, that's what it is.

Riker

LOL

moonheart08

LOL

Blue

19:14

:P

Οurous

._.

Erik the Outgolfer

can we stop please...

moonheart08

~~never~~

Adám

@Οurous I have just one thing to say to emoji: 🔥

Blue

Mine seems to have learned what I type

It's scary

Erik the Outgolfer

19:15

the government is watching you, beware...

moonheart08

Mine (On my tablet) thinks i'm a normal person, and trys to autocorrect, for example, C++ to see

Blue

Mine just knows all the ways possible to hug someone

4

J. Sallé

I've learned to actually use autocorrect to type instead of fighting it. I used to triple-check everything I typed on my phone, now I'm mostly used to it.

moonheart08

I just disable it

Unihedron

now it's O(n log n) instead of O(n^3)?

Blue

19:19

I use swipe typing and never see autocorrect anymore

J. Sallé

Swipe typing is such a foreign concept to me, the first time I tried to use it my brain just couldn't form the words

moonheart08

@Unihedron pretty sure it's O(n^5 log !n)

Completely accurate description of time wasted by autocorrect

Completely and totally offtopic from PPCG itself, but the factorio devs TOTALLY got their new custom rendering engine working first try.
https://factorio.com/blog/post/fff-230

DJMcMayhem

@moonheart08 I saw that, it cracked me up lol

BTW, if you're a fan of factorio, we've got a PPCG gaming room where factorio is a pretty popular topic. You're welcome there :)

moonheart08

:O

\o/

New Sandboxed Posts

19:47

0

A: Sandbox for Proposed Challenges

The Euler Masheroni Constant The Euler Masheroni constant is a very interesting number. It is defined to be and appears seemingly everywhere in number theory. Famously, it is unknown whether it is rational, irrational, or transcendental. The Challenge: Pretty simple one here: Given N, your pr...

DLosc

20:23

I think this might make a good meta post if your intention is to see if you can ask about programming languages here (not their usage (SO) but other stuff), but otherwise I think this challenge is a bit too opinion-based to be on-topic for this site, unfortunately. It's a very interesting topic though! — HyperNeutrino Feb 11 at 4:26

> this challenge

I think you've been spending too much time on PPCG, @HyperNeutrino

;)

Downgoat

Should this code work:

let a = 1.414
let b: Float = pow(a, 2)

i.e. not throw an error?

Mr. Xcoder

Yeah it should work. But if you’re using Swift... It won’t work :P

Downgoat

wait really

Οurous

@Downgoat can you implicitly cast Int to Float, or is 2 a Float literal or does pow have a version with pow :: Float Int -> Float

Potato44

why would it throw an error? Or is 1.414 a double?

Οurous

20:32

If any, yes.

Mr. Xcoder

@Downgoat Try it online!

Downgoat

@Οurous sorry, assume I wrote 2.0

Οurous

Then yes, it should work.

Downgoat

@Mr.Xcoder huh, I was referencing an example from swift doc page about type deduction :|

Οurous

Actually, do you have Double and Float?

Downgoat

20:34

yes

Mr. Xcoder

pow returns a Double so you must either make b a double or convert the result to a Double

Downgoat

assuming pow works for both float and double

Οurous

Then yes.

Mr. Xcoder

It works for both float and double, but it always returns doubles :/

Οurous

And I recommend literals defaulting to Double, and having an implicit downcast to Float only when you need one. But only explicit upcasting.

Mr. Xcoder

20:35

Also, why are you asking?

Downgoat

this is why you should make a modern language based on a 32+ year old framework

@Mr.Xcoder for VSL

Mr. Xcoder

Oh a new programming language you’re making?

totallyhuman

it's been there for a while

Mr. Xcoder

Ah yeah

Downgoat

yeah: github.com/vsl-lang/VSL/tree/add-backend

Mr. Xcoder

20:36

I do recall seeing it now... Ninja’d kinda

Downgoat

originally I was taking a detour by making LLIR but that's another huge project so for now directly compiling to LLVM

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

RGB to Xterm Color Converter code-golf mapping 256-color Xterm-compatible terminals add 240 colors on top of the usual 16 system colors. Colors 16-231 use 6 levels (0, 95, 135, 175, 215, 255) of red, green, and blue, ordered lexicographically. Colors 232-255 are simply 24 levels of gray (8......

totallyhuman

stack-based language with jelly-like functions: good or nah?

Mr. Xcoder

Good if you pop the elements in reverse order.

I cannot recall how many times I needed this in 05ab1e

totallyhuman

@Mr.Xcoder what do you mean?

Mr. Xcoder

20:44

ab/ ≠ a/b, but ab/ == b/a.

Of course it's up to you :P

I am completely biased

totallyhuman

oh you misunderstood my original message

jelly's newline separated links

Mr. Xcoder

Oh...

Gaia does that already and it works pretty well

totallyhuman

i'm essentially gonna number each function, using superscript numbers, ⁵ will call the function on the 5th line, ¹² will call the function on the 12th line (if you every really need it)

@Mr.Xcoder hmm, i haven't given that much thought

makin strawpoll

Οurous

Just have three stacks. One for left args, right args, and monadic args.

totallyhuman

strawpoll.me/15108311

@Οurous but what about functions with more than two arguments?

dzaima

20:49

@Mr.Xcoder for SOGL for the most often use cases of that I just made another function popping in the reverse order

Οurous

@totallyhuman uhm. I just didn't have any in Dirty. Your desires may vary.

dzaima

@Οurous does that help if the arguments are on the wrong stacks? :p

Οurous

@dzaima it's preferential. So it grabs from left if there's something on left. Otherwise from center. Otherwise from right. Etc.

dzaima

@Οurous but if items are on every stack you'd still have to waste bytes swapping the items around (or maybe your language does something smarter, idk)

Οurous

@dzaima I have like 80 commands just for moving stacks and items on them

dzaima

21:00

@Οurous well that's 80 instructions gone (and you still have to waste a byte every time one has to be used :p)

Οurous

@dzaima It's not solely a golfing language, so I can live with that :P

dzaima

@Οurous it's a bad suggestion for a golfing language is that I'm saying

HyperNeutrino

@DLosc uh no I don't that was... uh... intentional

:P

totallyhuman

i.imgur.com/UNLvyYm.png great

Adám

21:16

I think I came up with a great argument for APL's "right-to-left" (or right-associativity, as some prefer to call it):

Imagine a programming language that uses the syntax foo(goo(hoo(x))). Nothing controversial there, right?

Mr. Xcoder

@totallyhuman ;-;

DJMcMayhem

@Adám Right...

Adám

Now imagine that the language auto-closes parentheses for you: foo(goo(hoo(x (like TI-Basic does, and like Excel will suggest). Still good, right?

dzaima

@Adám good - no, understandable - yeah :p

Adám

Then maybe we could just omit the parentheses altogether: foo goo hoo x

DJMcMayhem

21:19

But that is made much more difficult due to variable arity

Mr. Xcoder

^

Adám

OK, but I thought of that. So here we go again:

Imagine a programming language that uses the syntax A.foo(B.goo(C.hoo(x))). Nothing controversial there, right?

totallyhuman

ok...

Adám

Now imagine that the language auto-closes parentheses for you: A.foo(B.goo(C.hoo(x Still good, right?

totallyhuman

yup

Mr. Xcoder

21:20

Yeah

Adám

Then maybe we could just omit the dots and parentheses altogether: A foo B goo C hoo x

totallyhuman

but what if C is a function

Mr. Xcoder

Also 0/10 should be foo((goo ∘ hoo)(x)) :P

Adám

@totallyhuman Not an issue, then it is A.foo(B.goo(C(hoo(x or A.foo(B.goo(C(hoo(x)))) when written in full.

totallyhuman

right but then C hoo x is ambiguous

Adám

21:22

@totallyhuman Yes, you need to know the classes to parse it.

totallyhuman

ok

Zgarb

@Adám For me, the best argument was always "you don't need to memorize precedence tables".

Adám

@totallyhuman Many APLers will capitalise functions and not arrays (or the vice versa). Syntax colouring helps too.

totallyhuman

welp it's settled

dzaima

@Zgarb but then the question of RTL or LTR arises.

H.PWiz

21:24

@Zgarb Except for some binding stuff for operators, Although that is usually what you would expect

Adám

@Zgarb That's given, but I've been struggling with why APL needs to be "backwards".

Enter the APL ninjas…

dzaima

I've been thinking about how useful would RTL APL expression entering be

H.PWiz

"Right to left" makes totally more sense to me

Adám

@dzaima Yeah, be problem is that you'd want to enter identifiers LTR, no? I just use the Home (and End) keys a lot.

Zgarb

@H.PWiz At least in J, operators follow the same rules but left to right. I don't remember if APL is exactly the same.

dzaima

21:27

@H.PWiz it does for me too, but it really didn't while starting learning APL. It took until I wanted to write a spec for a LTR APL for me to understand how useful it is :p

H.PWiz

I think there is dome weird stuff with . although, I'm might just be confused because it is also used for REFs

Adám

@Zgarb Yes, APL and J are almost identical in execution. J is basically APL spelled differently, and with somewhat different vocabulary.

totallyhuman

CMC: figure out how to achieve ` in GFM

J. Sallé

@Adám APL has by far the coolest glyphs though.

Zgarb

I learned J after APL, and I remember that some syntax was different. Maybe it was just the weird function-operator duality of / and friends.

Adám

21:29

@Zgarb In fact, J was originally called APL/?. Compare to the various APL flavours APL\360, APLX, APL+, APL#,…

@J.Sallé That's true. I can barely parse J. I have to insert spaces to deal with the bi- and tri-glyphs.

J. Sallé

@Adám btw, any good resources for learning APL Object Orientation on the fly?

Adám

@Zgarb Yes, most APLs have hybrid (or schizophrenic, as we prefer to call them) slashes.

@J.Sallé Other than the last two lessons and the next one or two?

J. Sallé

@Adám I was looking for something in a pdf-like format, but the lessons will help, of course.

Adám

@J.Sallé Yes, there are two excellent pdfs. Hang on…

@H.PWiz Yes, . is way overloaded, but mostly does what you'd expect. That being said, J's . is even worse.

J. Sallé

@Adám Yay! ~~My Kindle and~~ I appreciate it :D

Adám

21:34

@J.Sallé Object Oriented Programming for APL Programmers and Object Oriented Programming for Impatient APL Programmers.

@Zgarb Luckily we may be close to a cure for slashy schizophrenia: The atop operator. If ⍤ is atop, then ⊢⍤/ cures the / and makes it be a normal function.

dzaima

I have never understood the deal with / :p

J. Sallé

@Adám Thanks a lot!

Zgarb

@Adám Looks like a person carrying a suitcase yelling for a taxi. :P

J. Sallé

@Zgarb Aaaaaand now I can't unsee that.

Adám

@Zgarb Also 2-trains are different: In APL (f g) Y is f g Y and X (f g) Y is f X g Y. In J (f g) Y is Y f (Y g Y) and X (f g) Y is X f (Y g Y).

@Zgarb ⍨

@dzaima In the early days, the concept of operators was not fully developed. Thus +/3 1 4 (gives 8) and 1 0 1/3 1 4 (gives 3 4) were established. Later it was recognised that the first / is an operator (it takes + as operand) and the second is a function (it takes an array as left argument, and at the time, array operands had not been considered yet, hence the special bracket-axis syntax +[1] etc.)

dzaima

21:50

@Adám but now that operands being able to be arrays is normal, where's the problem?

Adám

@dzaima However, this didn't cause much ambiguity, only impurity (the slashes couldn't be classified properly), until tacit APL programming became a thing in J. What was the role of / in (f/g) Y ? Was it an f-reduction of g Y or was it the f Y compression of g Y ? Some decided that slashes should prefer to be operators, others made more elaborate rules. The issue still stands.

@dzaima Indeed, SAX (and thus APL/? a.k.a. J) redefined compress/replicate and expand to be monadic operators. However, this makes them awkward to use in trains.

@dzaima One thing is certain (for now at least): An operator cannot take an operator as operand (just like a function cannot take a function as argument), so you can never have DyadicOperator Operator and thus ∘. is unambiguous, and ⊢{⍺←⊢ ⋄ ⍺⍺ ⍺ ⍵⍵ ⍵}/ forces / to be a function.

dzaima

ah so / decides whether it's an operator or a function based on context? The nicknames start making sense :p

H.PWiz

@Adám Huh

Adám

@dzaima Yup, they are indeed a very sick family that needs a cure.

dzaima

@Adám why can't it just be an operator always?

Adám

21:58

@H.PWiz Yeah :-( APL is super-lazy when parsing. Parsing is pushed off until the very last possible moment. For laughs, try parsing ⎕⎕.⎕⎕

@dzaima Because operators cannot take operators as operands. How would you do a replicate-reduction // ?

Transcript for