The Nineteenth Byte

OldSandboxPosts

01:00

Sandbox posts last active a week ago: Next digit of rational number [codegolf] [string]

pxeger

07:23

@Neil week.golf/challenge.php?id=1, click the light bulb, click the language logo to switch to Python, then click "Solution №1", it will shrink, then click it again and it will show you the code

Neil

... I don't see a light bulb. I see a header, a title, a description, an area to test and submit a solution, a leaderboard, and a footer

mousetail

@Neil It's only on last weeks challenges

Neil

???

emanresu A

I call this the pxeger theme

mousetail

Based

Looks better than the green IMHO

emanresu A

07:31

Okay. userstyle:

#bulb_content, #time > div.vertical-align, #run, #delete_code, #reset_code{
  background-color: #fc774d !important;
}
#green {
  color: #fc774d !important;
}

Currently only works on that page, will try and make one that works globally

mousetail

I've been trying to shorten my python solution there for days but I keep changing everything but miraculously ending up at exactly the same length

pxeger

07:51

Neil

08:43

@pxeger yeah I couldn't be bothered to create an account

mousetail

@Neil You can log in with your SO account

pxeger

I don't know why you should have to tbh

emanresu A

^

@mousetail I finished this: gist.github.com/chunkybanana/1568836b25134553fff5198df4d1becb

It's as a userscript but you can copy the css in the middle if you want a userstyl

mousetail

TY

Looks great on the graph page too

emanresu A

Went through all the pages, although I might've missed some

mousetail

08:52

Dots are still green but that's not a big deal

emanresu A

Might be able to change that

The only thing I haven't figured out how to change is the upvote image, although that actually might be possible with some JS

mousetail

You need .a_point{background-color for the dots

pxeger

@emanresuA CSS hue filter?

emanresu A

Huh, that's a thing?

@mousetail Ah, missed that because I hadn't done anything so there was nothing

Also, does this person know that canvases exist? That's such an impractical way to do it lol

mousetail

CSS filter is the way to go yes, but hue isn't the right one since the icon is white

emanresu A

08:55

I mean when it's green

mousetail

@emanresuA You can't as easily add click and hover effects to canvases

emanresu A

I guess :P

mousetail

I always use SVG for graphs

Divs works too but it's hard to get to behave consistently between browsers. Canvas looks bad when zoomed in or you need to write a lot of code to fix the resolution. Plus the click and hover effects thing

emanresu A

Good point :P

lyxal

Anyone else agree that more than one language a month should be added to week.golf?

pxeger

08:59

yes

add as many languages as you can

emanresu A

Updated the gist

^^

mousetail

No, there will be too many languages too quickly

pxeger

artificially limiting it may be fun but seems very pointless

@mousetail how is that a problem?

mousetail

The site is most fun if there are multiple people competing in a single language

More languages=less people per langauge

pxeger

People can use more than one language

If Zsh gets added, that doesn't mean I'll stop golfing in Python

mousetail

09:01

Yes, but most people do only 2 or 3

lyxal

I disagree - seeing everyone answering in a single language makes me feel like I've missed the opportunity to use that language

emanresu A

^^^, and we just want to golf in specific langs

Also ^^

lyxal

Like I see all the answers in python and think "well, that's already been golfed to oblivion, no point trying that language"

emanresu A

^

mousetail

I think they could use a few more languages, but if they hope to stay interesting long term limiting the number is key

lyxal

09:03

Maybe 5 new languages a month?

emanresu A

Also:

in Vyxal, 7 hours ago, by emanresu A

Honestly I dislike the challenges.

in Vyxal, 7 hours ago, by emanresu A

They're all chameleon challenges and they'd recieve a bunch of downvotes here

in Vyxal, 7 hours ago, by lyxal

Also I bet SBCS languages will be scored as utf-8 instead of SBCS

in Vyxal, 7 hours ago, by emanresu A

Perhaps we should just leave this alone. I'm not sure this is a place we want to be.

pxeger

@emanresuA This is the problem with "golfing servers" (also like code.golf and golf.shinh.org) as opposed to fora like ours. In order to work in the system, all programs have to use the same IO method, which means way too many bytes wasted on reading STDIN or whatever

it's just not fun

And so far they've all been bizarre two-in-one problems for no good reason

Ulam numbers alone would be fine, but then applying it n times is, IMO, a worthless addition

emanresu A

Agreed

in Vyxal, 7 hours ago, by Unrelated String

yeah honestly i would try not to do anything that prolongs the life of that site

@pxeger I think there's probably a way to do it well, although highly configurable I/O comes with the annoyance of having to configure it

mousetail

Even if you don't want the site to live long you have to understand the author probably has different intentions

Zion mycelia adamancy

09:30

@pxeger Honestly, I don't even think I/O is the problem. The real problem is that the challenge writing isn't freely open to all

emanresu A

That's a really smart perspective that I never thought about

Zion mycelia adamancy

Writing a challenge is just as much, if not more, fun as answering one. Any site that prevents its users from doing that is fundamentally flawed

emanresu A

Also, tfw I use fricking root access and an operation's still not permitted???

pxeger

man user_namespaces

also "operation not permitted" (EPERM) normally means a technical limitation, not exactly a permissions thing (which would be "permission denied" (EACCESS))

Zion mycelia adamancy

@Zionmyceliaadamancy Especially as it isn't hard to add some kind of wrapper when adding a language, so that I/O is done in that language's most natural way. It isn't perfect, but it works most of the time

pxeger

10:18

Opinions on this as a 5-bit codepage:

abcdefghijklmnopqrstuvwxyz012.-'

emanresu A

10:30

I'd advise just assigning stuff as you go

pxeger

Everything's gonna be either 10 or 15 bits, so I don't think I can really do that

emanresu A

Wait what

Is this for newspeak or something else?

pxeger

only digraphs and trigraphs, no single glyphs

this is something else

I want to do a 10 bit language but can't be bothered to make a 1024 character codepage

emanresu A

You're gonna have to implement 1024 elements

pxeger

and by splitting it into 5 bit chunks, you get 25% shorter "digraphs" (by using trigraphs instead of quadgraphs)

@emanresuA I think I'll manage :)

emanresu A

10:33

That's four times a SBCS golflang

And even those take months-years

pxeger

Well we'll see

emanresu A

Good luck, and I'll see you in 2026.

Wheat Wizard

hgl has well of 1024 elements already.

pxeger

that's exactly what made me think 1024 was feasible

emanresu A

Yeah, but you have space for all of that

pxeger

10:34

also things like MATL

actually I think MATL doesn't have quite that many but still

Wheat Wizard

I mean obviously there's a bit of a difference because hgl has so much space that there's basically no filter. If something could possibly be helpful might as well add it.

Although something I realized more and more is that my time with hgl is a finite resource.

pxeger

and now I think about it I'll probably have a few single-char elements (only 5 bits) so it might be only like 900

emanresu A

@WheatWizard Wait what?

pxeger

but I won't need to finish all 1024 to get started

@emanresuA The time you can spend developing it is finite

Wheat Wizard

@emanresuA It takes time to add things to hgl, so it's worth considering if adding certain things are worth my time.

Or rather it's worth trying to figure out what the most valuable holes are right now and work on filling those first.

lyxal

10:51

@emanresuA the owner of week.golf has said that SBCS will be used for languages that have them

emanresu A

That's good

Bubbler

I keep thinking about various langs but all of them are based on pure functional programming and I need to write a runner for that anyway ;_;

Wheat Wizard

Like what ideas?

Command Master

Hello! Is there a reason most golfing languages (that I'm aware of) use single byte characters, instead of ones with a variable amount of bits? Some builtins are a lot more commons then others, so it would make sense to have those be shorter, and you can just split the final program to bytes

Bubbler

a 2d lang on a hexagonal grid with semantics somewhat close to flobnar or cascade, for example

emanresu A

11:03

@CommandMaster my upcoming golfing language has that. It's also probably useless :P

Command Master

@emanresuA The language, or having shorter-then-byte characters?

emanresu A

Having shorter-than-byte chars

Although there's also a human-readable encoding

Command Master

@emanresuA Why? Because usually you don't use enough common characters to get 8 bits?

emanresu A

That, and because I don't want to implement a full SBCS golflang

lyxal

Hey guys is anyone here spamming account creation on week.golf? Because someone is creating a whole bunch of troll accounts for some reason

If it is anyone here, don't. It's really unfunny and doesn't accomplish much

Wheat Wizard

11:09

@CommandMaster A lot of them do have two byte commands. fwiw until recently it was extremely difficult to get a fractional byte count, so 9 bit commands would usually cost just as much as 16 bit commands.

Command Master

Yes, I'm aware. What had changed recently?

lyxal

Yeah, the fractional byte meta was only recently experimented with once fractional byte counts were ruled valid

You used to have to round byte counts iirc

But now you can go as far as having irrational byte counts

emanresu A

Worse, you had to have an encoding that fitted into full bytes

Wheat Wizard

Or build a computer that actually used fractional bytes? The old consensus was honestly not super clear.

Arguably the new consensus and the old consensus are the same thing said with different emphasis.

pxeger

@CommandMaster I think it's because it's hard: both an inconvenient encoding to implement, and more importantly, very hard to know which builtins warrant which bit encodings.

mousetail

11:21

Seems pretty easy. Copy commands from a existing language then scrape existing programs to know which are most common

pxeger

but if you give a builtin a longer encoding, it won't be used much because it's not golfy, which will make you think it deserves a longer encoding, even if it doesn't

it's a complex self-referential optimisation game

mousetail

That's why you copy a existing language

So there's plenty of data that can't be poluted

Wheat Wizard

One thing I want to do is collect a corpus of Jelly answers and unleash a NN compression algo on them. It could be very useful in telling what sort of atom combinations are most useful.

pxeger

and also, variable-encoding languages are hard to use because you can't immediately tell what program is going to be shorter just by looking at it

mousetail

Need a dedicated editor that changes the physical width of a character depending on it's bit size

Maybe a font could do that

Wheat Wizard

11:23

@pxeger You could have the atoms be variable width but the characters be fixed width.

pxeger

I mean, ultimately, you could write plain ones and zeroes, which would directly correspond to the lengths of the builtins

I guess you could do it by using elements as "words", where the number of chars in the word corresponds to the bit length. Is that what you meant @WheatWizard

Wheat Wizard

Yes.

pxeger

I once thought of an idea for a symbolic encoding for a language where every command is a word of up to 7 letters, and you separate words with tabs so they all have the same visual width

that's kind of the opposite

@pxeger I decided it wasn't that useful, especially because of string literals

Wheat Wizard

It might be interesting for a 2D language.

Since those really care about alignment.

pxeger

it would be a pretty wonky 2D language though; everything would be much wider than it is tall

mousetail

11:27

Make every word a 7x7 block

pxeger

make a programming language out of lego

mousetail

That would be cool

emanresu A

And impractical

Wheat Wizard

I want to make a programming language where the character width is variable as determined by the thue morse sequence.

mousetail

Impractical languages are the most fun

pxeger

11:29

@emanresuA not much more impractical than scratch, and people use that

emanresu A

Make a language like Cascade, but with gravity

mousetail

For a lego lang all blocks would be different sizes which would limit how you can attach them together in 3d

pxeger

you could use the shape of the blocks as a type system

Wheat Wizard

In a normal language you can write the same function and put it anywhere in a file. However in this language it needs to align with the grain of the file to fit so functions cannot simply be translated around.

mousetail

Yes

In lego there are also a few special blocks that allow half-stud offsets that would be interesting in a language

pxeger

11:34

to make a lego-like language more practical, you could do it more like a typical 2D language but with polyominoes instead of individual chars

mousetail

Lego should be 3d at least, 2d would defeat the point

pxeger

but 3d would be a massive pain

Command Master

I was thinking about a language where you can get the value of any node in the AST, even results in the future

mousetail

@emanresuA On the activity page disabled buttons are still orange

emanresu A

Oops

pxeger

12:18

CMC: given a string, return the same string but with the second character removed

abcdeb -> acdeb

input will always be at least length 2

mousetail

@pxeger lambda x:x[0]+x[2:] (19 byes)

Command Master

@pxeger ʒN≠

Bubbler

2d lego reminds me of a game called Junkbot

Wheat Wizard

@pxeger f(a:_:b)=a:b

Command Master

@WheatWizard What language is this?

Wheat Wizard

12:28

Haskell

Also Curry

lyxal

@pxeger vyxal: 1⋎

PyGamer0

feebcak for elements for flax?

also i change list syntax to (1 2 3(3 4 5))

Sandbox Posts

13:01

0

A: Sandbox for Proposed Challenges

Infinite Fibonacci word code-golf sequence fibonacci string The famous Fibonacci sequence of integers is defined as follows: \$ F_0 = 0 \\ F_1 = 1 \\ F_n = F_{n-1} + F_{n-2} \$ But what if we use this same recurrence relation to produce an infinite sequence of strings? Instead of addition, we'll ...

mathcat

13:25

I am 65 minutes too early and my internet connection is E :/

I feel like being in the void would've been less boring

lyxal

65 minutes to what?

and the void?

y'all I need context here

pxeger

> y'all

this is your brain on American English

mathcat

I have some appointment in 64 minutes

pxeger

you start using y'all when you're talking to only one person

lyxal

@pxeger bold of y'all to assume I was only asking one person

mathcat

13:27

^

pxeger

@mathcat hang on, a minute ago you said 65, and now it's 64?! Get your story straight!

mathcat

weird, it's in 63 minutes now

guis I think time changes

lyxal

no, I don't think it do

mathcat

well what do the time does Australia now?

pxeger

23:31:59

dammit I'm being rate limited

lyxal

13:32

mathcat

wait a sec

lyxal

that's the time

it's Now o'clock

and that clock is always correct

pxeger

NameError: name 'x' is not defined

mousetail

Why is Rick 11?

lyxal

@pxeger yeah because you don't have the prelude for the clock

mathcat

13:33

wow lyxal just proved time is constant

pxeger

Rick is 56

lyxal

@mousetail *10

bold of you to assume my clock has 12 items

mousetail

I've seen a similar meme clock where all the math expressions actually refer to the time at that position. This is defiantly not that.

pxeger

It doesn't seem very defiant to me. It's just a clock.

mousetail

I can't spelll

mathcat

13:35

me niether

mousetail

I just type random gibberish then use a AI trained to make it look like a sentence

pxeger

> One word in this sentence is misspelled

lyxal

okay so maybe I lied

it's not now o'clock

Dec 23, 2021 at 12:46, by lyxal

it's raze time

mathcat

@pxeger woah that's clever

@lyxal it ain't loading for me

I still have 65 minutes, I can wait

pxeger

Hang on, it's back to 65 now??

lyxal

13:37

yup

we just went back in time

mathcat

Time is constant again now, right?

mousetail

What is the countdown for?

pxeger

11 mins ago, by mathcat

I have some appointment in 64 minutes

mousetail

ok same

lyxal

@mathcat which void?

there's a whole bunch of voids I can think of

so you need to be a little more specific

mathcat

13:41

The void where 0 divided by 0 is a complex number.

lyxal

fair enough

Adám

mousetail

@Adám How does this work? Im isn't 90 degrees of from Re

Adám

No, but in APL 11○X is Im(x)

mathcat

"[...] 0 divided by 0 is a complex number"
lyxal: fair enough

quotation power

NoHaxJustRadvylf

15:27

@emanresuA Funnily enough I have a userstyle that does the opposite: turns pxeger's orange pfp into green :p

pxeger

Sacrilege!

PyGamer0

books.google.com/ngrams/…

NoHaxJustRadvylf

More useful if you include "y'all" since that's the main way it's spelled

PyGamer0

16:06

@NoHaxJustRadvylf books.google.com/ngrams/…

NoHaxJustRadvylf

Only works for me after focusing the search bar and hitting enter

pxeger

Same

weird bug?

Adám

@WheatWizard @AncientSwordRage Can you give chat.stackexchange.com/users/538234/user110569 explicit write access to chat.stackexchange.com/rooms/52405/the-apl-orchard?

mathcat

yay finally wifi!

pxeger

how was your appointment

were you appointed?

Adám

16:18

~~were~~ what was it you

Neil

16:36

@pxeger Charcoal, 4 bytes: ΦＳ⊖κ

thejonymyster

Tragic: we have only one challenge which awards swag

PyGamer0