The Nineteenth Byte

New Posts

00:06

0

Q: A Fine sequence with fine interpretations

The ubiquitous Catalan numbers \$C_n\$ count the number of Dyck paths, sequences of up-steps and down-steps of length \$2n\$ that start and end on a horizontal line and never go below said line. Many other interesting sequences can be defined as the number of Dyck paths satisfying given condition...

Jo King

@RydwolfPrograms i think your brain starts boiling around 100

4

ATaco

00:27

Definitely getting heatstroke after 40

Rydwolf Programs

Maybe neuroscience uses Fahrenheit. You can at least get 180 degrees before the boiling starts, enough to totally turn your life around.

lyxal

00:42

If you want to turn your life around, just go to a bakery and get a pie

Make sure not to get 2 though, else you'll just be stuck in the bakery

lyxal

01:14

Can't believe I have to use git from the command line instead of github desktop because they added a bug to github desktop in the latest beta release

ATaco

01:49

There's deffo an untapped goldmine in Assembled/Compiled golf-langs that don't need to concern themselves with Charactersets.

Unrelated String

vyxal's currently developing a verbose mode with v3; charcoal already has one

and if you mean also forgoing the byte-command correspondence there's been various stabs at that

ATaco

Yeah but the languages themselves are still human readable, Verbose mode just allows for verbose script writing. We can go deeper...

Unrelated String

we don't grade any other languages by compiler output so you do still want an encoding for the actual submission :P

ATaco

We allow it for assembly, it's the same concept.

Unrelated String

point is there's no real difference between this and the old wave of compressed langs

though come to think of it nobody did end up getting one of those most-redundant-behavior-is-mathematically-impossible langs done

ATaco

02:05

Ultimately it doesn't matter too much, as we tend to only compete in similar language classes

Unrelated String

yeah i think part of why the fractional byte wave seems to be here to stay is it at least feels like the conventional model

Rydwolf Programs

Huffman coding, arithmetic coding, etc. has been proposed over and over again but people don't follow through because...it's just kinda boring lol

It makes the process of golfing less fun, by decoupling operations from byte count, which is the whole thing that makes SBCS/FBCS langs cool

ATaco

I also appreciate multi-character byte encodings as a neat way to represent a byte-per-instruction set

Rydwolf Programs

The one-operation-one-byte mapping that makes it feel...sorta distilled in a way

Unrelated String

i forget who did it but remember sledgehammer
iirc it was actually really competitive but since it's literally just compressed mathematica they still ended up finding it so ungodly boring to actually use that it was abandoned forevermore

ATaco

02:08

One-byte-per-instruction feels sorta… correct, to me.

Bubbler

also basically only usable by those who have bought mathematica already

Rydwolf Programs

What I plan on doing once I get around to making my golflang, which I should have finished by the time I graduate college, maybe by the time my children get to high school, I'm going to design it in such a way that I can reuse the library of built-ins for a couple of different encodings, like SBCS, catstruct, maybe something tacit, etc.

ATaco

I'd also be curious to see a 16bit golf lang with far too many instructions.

Rydwolf Programs

Maybe it could be generated by taking Vyxal, Jelly, Husk, Japt, etc. answers, translating them into some sort of golflang-assembly, then using some advanced algo stuff to pull out common combinations of operations that would make sense as discrete operators

My guess is it could actually be worse in a nontrivial number of cases

Parcly Taxel

So earlier in the chat I brought up the idea of "extended Piet"

The original Piet has 20 colours, 18 in a grid + white + black

Rydwolf Programs

02:17

Oh yeah like a full 8-bit one?

Parcly Taxel

Operations are defined by the change in colour between blocks of colour

yeah, my idea is Piet with 256 colours including black and white, which would serve as "obstacles" (or not) in an 8×8×4 grid (the colour is 8-bit colour, MSB BBRRGGG LSB)

Rydwolf Programs

@RydwolfPrograms Every 2d language I come up with is horribly cursed so I've been too scared to try making a piet-like lang myself lol

Parcly Taxel

The original Piet could only push integers onto its stack. Extended Piet should additionally be able to handle at least floating-point numbers and strings as well

Rydwolf Programs

Maybe 2-item tuples too?

You can do some cool stuff with those

Parcly Taxel

Lists could also be on the stack. Then matrices would be represented by several lists

Bubbler

02:23

In my experience, what makes Piet hard to program in is the "difference in color defines the operation" thing

It's horrible to iterate on program fragments

Jo King

a similar thing happens when golfing hexagony, adding or removing an instruction necessitates reorganising the whole codepath

ATaco

Yeah but it's an image esolang because it's about image properties. Pull too far away from that and you may aswell have bytecode.

Parcly Taxel

256 colours allows the image to be represented in SBCS

rather than any image format

Aiden Chow

02:51

@Bubbler huh, i find controlling the path of the ip optimally (as in most golfiest) to be the hardest part of piet, theres just a myriad of rules pertaining to cc and dp to keep track of, and its usually hard to find a concise structure that does what u want it to do

the difference in color thing also sucks too tho, i wish it wasnt like that lol

Bubbler

03:06

"controlling the path of the ip optimally" is hard, yes

but I mean the general process of golfing (like moving around code parts and switching one code fragment with another) is a magnitude more tedious than other langs

Aiden Chow

true, thats pretty much why i suck at golfing in piet, once i have some working code, even if the structure is ungolfy as hell, i rarely go back to change the structure in any meaningful way cuz its a pain in the ass lol. and anyways i always seem to break my working code when i try change smth lmfao

Parcly Taxel

03:26

My idea for branching in Extended Piet is through the idea of "exit points" which are pairs (cell in block, cell not in block)

Aiden Chow

would this extended piet still have the difference in color thing? like is the overall structure the same, with how the pointer moves and everything?

Parcly Taxel

Each exit point for a block is numbered according to the current direction pointer

@AidenChow Yes, except that white and black are part of the colour grid

Aiden Chow

@ParclyTaxel huh, would white and black still act the same then?

how would that work?

Parcly Taxel

Black blocks serve to inhibit exit points

When exiting a colour block the pointer tries to use the first exit point. If that is blocked by a black block, it tries the second, then the third and so on

Bubbler

Cursed idea: gray blocks act as white or black with 50% chance each

Parcly Taxel

03:33

Yeah original Piet has no builtin random function. Extended Piet needs one

Aiden Chow

im not sure if im understanding correctly. what if i wanted to access a certain builtin but the difference in color means the next color has to be black (which is what i assume u mean by black and white being in the color grid)?

Parcly Taxel

That is part of the challenge

Aiden Chow

@Bubbler that would be cool, im not gonna lie

@ParclyTaxel oh, i see...

Parcly Taxel

You can always use a white block to move the colour to a point where the next instruction doesn't per se end in a black block

By default the first exit point in the defined ordering is used when the pointer exits a block

Aiden Chow

@ParclyTaxel true, that would be a bit annoying tho

Bubbler

03:35

I bet you can get around it most of the time by selecting "good" color for the top-left cell

Parcly Taxel

Probably not very annoying since you have 256 colours and only 2 can pose an obstacle to the flow of instruction colours around the colour cuboid

Bubbler

though it does add another tedious element

Parcly Taxel

@Bubbler yes, yes

Aiden Chow

@ParclyTaxel also how are u numbering the exit points? will the dp be able to take on more values than just 0 to 3? what about the cc?

Parcly Taxel

The exit points are pairs of pixels, the first element of a pair being a pixel in the block, the second being an adjacent pixel outside the block

Aiden Chow

03:38

@Bubbler true, but tedious when it just so happen that while constructing the answer, u come across having the place a black block, then having the change all the colors as a result of that

Parcly Taxel

Rotate the program so that the direction pointer faces up. Then these pairs are lex-sorted to produce the index

The program follows the exit, and the direction from in-block pixel to out-block pixel becomes the new direction

Aiden Chow

so the dp is still 0 to 3, but the cc can range to any number in order to index which pair of pixel u exit from?

Parcly Taxel

This execution model eliminates the CC

Aiden Chow

oh, interesting

so the dp is the only thing that is determining the direction of the pointer?

Parcly Taxel

Yes

Bubbler

03:41

@ParclyTaxel Then how does one determine which index to use?

Aiden Chow

^ the cc's job is to do exactly that

Parcly Taxel

If the instruction to be executed does not branch, the program will use the first exit point in the prescribed order from the block it has just entered given its current direction

Aiden Chow

i think its better to have the dp still store right, left, up, down, while cc can choose which exit point in that direction to exit from

Parcly Taxel

(That is not blocked by a black)

Aiden Chow

at least, that makes much more sense to me

@ParclyTaxel what if i want to choose a different exit point other than the first one

Parcly Taxel

03:44

Then you have to make the block you enter correspond to a branch instruction

Aiden Chow

the entire point of the cc in the original piet is the freedom to choose which side u want to exit on based on if the cc is 0 or 1

Parcly Taxel

It pops the top number off the stack and follows the corresponding exit

Bubbler

so "change dp" and "change cc" is unified into one "branch" instruction?

Parcly Taxel

Yes

Bubbler

nice

Aiden Chow

03:46

ohh thats what u mean, i see

wait so the dp dictates which exit point to go from in each block right? what if the dp is higher than the number of exit points available?

Parcly Taxel

The DP is still only up, down, left or right

Aiden Chow

oh

so the branch instruction is choosing only from where the dp is facing or from all possible exit points?

Parcly Taxel

It is the combination of DP, the block's shape and (if branching) the popped index that determines which exit is taken

DP chooses from all unblocked exits

Aiden Chow

@ParclyTaxel each block is like this or are u describing the branch instruction?

Parcly Taxel

Each block is like this

Bubbler

03:52

I guess the branch instruction would be like "pop x; let n = count of exit points; branch to (x%n)th exit point"

Aiden Chow

so if theres branching in a block then it pops a number from the stack automatically?

Parcly Taxel

As I said, if the current instruction is not branch the program exits at the first available exit, and no number is popped (obvs)

@Bubbler Modulo seems like a good idea

Bubbler

The modulo is what dp and cc instructions do in original Piet

Aiden Chow

@ParclyTaxel i must be having a fundamental misunderstanding of ur idea. what if i want to exit the current block (which is not a branch instruction) at a different point other than the first?

Bubbler

@AidenChow You can't

Aiden Chow

03:54

oh

Parcly Taxel

You must make it so that the operation corresponding to the previous block's colour and this block's colour is branch

Aiden Chow

yea but i was just wondering if u can exit at multiple different points even without branch instruction, cuz in regular piet the cc allows u to have two different exit points with each block

Bubbler

another concern is that every reachable block will have some exit point, which makes the original Piet's "end program" condition impossible

Parcly Taxel

That's why I'm adding an explicit halt instruction

Aiden Chow

@ParclyTaxel ayyyyy alright nice

i wish piet just had a halt instruction lmfao

Parcly Taxel

03:59

Piet's termination condition in most cases requires you to make at least a 1×3 block and enter through its middle, which may make geometry hard

Aiden Chow

the l-shape construct in a 2 by 2 also works to halt piet code, but yea

L-shape, i mean

(capitalize to make more clear lol)

but halt instruction makes things so much easier

Parcly Taxel

I've slightly changed the exit point indexing so that the off-block pixels come first in the pairs of (off, on)-block pixels, the pairs of which are then lex-sorted

So if the instruction from prev to this block is not branch, exit 0 will be followed

Bubbler

That... looks honestly cursed

I'd expect all forward-pointing exits come first, all right turns come next, etc.

Parcly Taxel

This change means that in unblocked situations the exit point will be the same as in Piet with DP unchanged and CC left

@Bubbler How can you define "clockwise" and "counterclockwise" on the perimeter of a shape with holes?

If you want (more) predictable exit indexing use shapes with smaller perimeters and black pixels

Bubbler

04:14

true lol

@ParclyTaxel I didn't say clockwise, but one possible reordering in your image would be 0, 7 (forward), 2, 8, 5, 12 (right), 14, 13, 11, 4 (back), 10, 9, 3, 1, 6 (left)

that is, look at each direction and sort all forward exits lexicographically

Parcly Taxel

04:29

It is a good idea to sort on exit direction first – I'd go with forwards, right, left, back though

The interpreter would precompute all blocks, but only compute (and cache) exit points once it enters them

Bubbler

The reason why I ordered like that is that it agrees with Piet's DP trying order, and it's a bit simpler to code IMO

Parcly Taxel

FRBL it is then

The name Extended Piet is only a working title

By longstanding tradition every one of my Git repositories is named after a real-world location. So the final name is likely to also be a real-world location

Aiden Chow

@ParclyTaxel which is ur github?

lyxal

@ParclyTaxel name it "the server that hosts the code golf chat rooms in the stack exchange headquarters"

Very fitting given that's where it was discussed

Aiden Chow

nah even better u gotta list out all the addresses of everyone who participate in this discussion (excluding me, ofc)

lyxal

04:45

I'll start. My address is unit 69/420/3 Fake Street, New North London York, Albuquerque, New Yorkshixo

That's where I live, I decided

Parcly Taxel

04:56

@AidenChow github.com/Parcly-Taxel

gitlab.com/parclytaxel

I also have my little personal website (with pretty little ponies) at parclytaxel.art

It started with Kinross, a small Python suite for compressing SVG files (the files I work with in My Little Pony vector art)

When I needed a place to store my art I chose another git repo named Selwyn, after the Cambridge college. I had applied for Cambridge and actually got an offer, but I declined because Brexit

Later repos include Shinjuku (glider syntheses in Conway's Game of Life), Skopje (smallest oscillators im CGoL and other CAs), Malibu (accurate parametric 3D surfaces), Shibuya (unit-distance graphs and other geometric optimisation problems) and the code for my personal website itself at Portsmouth

Even the computer names of the laptops I've used are named after locations

My current laptop has the name Winterthur

Unrelated String

@ParclyTaxel whoa, nice

ATaco

Did you see when PPCG Made Tetris in GoL? @ParclyTaxel

Parcly Taxel

gitlab.com/parclytaxel/Selwyn

@ATaco I know that. I am also a member of the ConwayLife lounge, and they reacted to the Tetris-in-CGoL project with indifference

Unrelated String

sad

i guess to a gol'er the metacells are old news and the rest of the layers are kinda out of scope

Parcly Taxel

They have however shown that any Life pattern that has a synthesis in gliders, only needs 15 gliders

ATaco

05:05

Once you abstract two layers, it matters a lot less what it's built on I suppose.

Unrelated String

wow

Parcly Taxel

They have also explictly shown (recently) that there exist periodic patterns with no glider syntheses

ATaco

~~Now calculate where those 15 gliders are for a given synthesis~~

Bubbler

---

Parcly Taxel

?

Unrelated String

05:06

strikethrough formatting

before the edit they tried ~~ :P

Bubbler

yes

Parcly Taxel

no I am referring to - - -

Unrelated String

that's what i meant too :P

---strikethrough formatting---

versus ~~discord flavor strikethrough formatting~~

ATaco

<s>Html formatting save me</s>

[s]I also accept BB-Code Formatting[/s]

Sandbox Posts

07:11

0

A: Sandbox for Proposed Challenges

Making one shape out of dissections of another code-golfdecision-problem Introduction One question that I have come across recently is the possibility of dissecting a staircase of height 8 into 3 pieces, and then re-arranging those 3 pieces into a 6 by 6 square. Namely, is it possible to dissect ...

Sandbox Posts

07:25

0

A: Sandbox for Proposed Challenges

Consolidate a 6-axis Vector Typically, when we want to represent a magnitude and direction in 2D space, we use a 2-axis vector. These axes are typically called X and Y: This isn't always convenient, however. The game BattleTech is played on a hexagonal grid, and it's convenient for the axes to l...

Aiden Chow

07:41

@SandboxPosts bruh i just make two falsey test case for my sandboxed challenge and bubbler just come along and prove both of them are in fact truthy like literally fml

Bubbler

~~don't forget, I have 10k rep on puzzling~~ hehe

Aiden Chow

ok well i just make two new one if they are actually truthy im actually done lmfao

yo one of my meta feedback is suggested test cases, maybe u could, u know, lend a helping hand, ~~considering you have 10k rep in puzzling~~

also like think about how bad i am at puzzles :P

Bubbler

accb
aaab
aabb
aaab

cb
cb
bb
ab
aaa
aa
aaa

Aiden Chow

fuck

ok well im just taking that one out completely i give up on that one

well.... i have a severe lack of test cases now

Bubbler

proving something is solvable is easy, the opposite is shit hard

I guess I gotta whip up a solution for that

Aiden Chow

07:52

ok the other false test case is false right

dont tell theres a solution to that

Bubbler

almost surely yes

Aiden Chow

@Bubbler errr cant u just come up with some falsey test cases real quick without a program?

also how tf would u even solve this, im not big brained enough to even think of a solution to my own challenge lmfao

Bubbler

xxxx
xxxx
xxxx
xxxx

xxxxx
x   x
x   x
x   x
xxxxx

is almost surely unsolvable in 3 pieces

Aiden Chow

k cool im putting that one in, thx lol

might as well put it in truthy as well but with 4 piece lmao, i really need those test cases

@Bubbler also how did u solve these so fast.... u seen similar problems to this before or what????

Bubbler

I just fiddled with polyomino-type puzzles for years

pxeger

09:07

Under what circumstances can /usr/../ not be the same as /?

mousetail

Would symlinks maybe effect this?

pxeger

I think so, but /usr is not a symlink

Unrelated String

until the edit i was about to ask, is this a puzzle or is this something utterly inexplicable that you found in an actual script

mousetail

Is it a mount point?

pxeger

/ is mounted as an overlay filesystem

@UnrelatedString real script. I'm debugging ato.pxeger.com/…

Unrelated String

09:11

nice

pxeger

/usr/local/bin/hops is a symlink to ../../../opt/cabal/ghc-9.4.4/hops-0.8.5-xyz/bin/hops

and /opt/cabal/ghc-9.4.4/hops-0.8.5-xyz/bin/hops exists

but trying to read /usr/local/bin/hops gives ENOENT

You can play with this behaviour more at ato.pxeger.com/…

I think I see why it's happening, but I have no idea how to fix it

mousetail

Unrelated String

LMAO

nice

mousetail

09:34

@pxeger Can you set the symlink to use a absolute path instead?

pxeger

09:52

Yes that will work but it doesn't fix the underlying issue with the filesystem layout

Sandbox Posts

0

A: Sandbox for Proposed Challenges

Convert a Haskell do-expression to its Rust equivalent In hope that Rust will support the ? operator for every monad. Objective Given a simple Haskell do-expression, output its Rust equivalent. Input The input is a single string consisting of multiple lines. The first line shall be do. The other ...

Sandbox Posts

10:31

0

A: Sandbox for Proposed Challenges

Floating-point numbers have errors in addition, oh my! Given two IEEE double-precision floating-point numbers \$x\$ and \$y\$, output the most accurate IEEE double-precision floating-point approximation of the floating-point error that would be caused by performing the floating-point addition \$x...

Parcly Taxel

11:09

1

A: A Fine sequence with fine interpretations

Jelly, 12 bytes +Żc¹NÐeƤ§ÆmA Try it online! Port of emanresu A's golf to mathcat's Vyxal solution. +ŻcµUNÐeḋJ:L is a more direct port for the same length, but this is funnier. Jelly, 18 bytes o2µḤœcµṬ-*ÄAƑ×ṂḂ)S Try it online! Haven't actually tried any closed form or recursive formulae, but ign...

I don't get why +Żc¹NÐeƤ§ÆmA is funnier

Unrelated String

explanation added :P

Parcly Taxel

so it's the no-op that makes it funnier?

Unrelated String

using prefixes to replace the reverse-and-multiply-by-index, resulting in the need for an absolute value at the end :P

emanresu A

11:29

Clever, but it's easier to negate every other element, cumsum then sum (at least in vyxal)

*cumsum then mean

Unrelated String

oh so vyxal does have a mean builtin

that's a relief lmao

also now this is reminding me of that challenge where i wrote a solution that two others ported but then i ported their original version back

*versions

emanresu A

I feel like there's gotta be a 10 in Jelly

Unrelated String

yeah

trying to re-port your vyxal 10 comes back up to 11 having to absolute value at the end and/or keep the range reversed then re-reverse it for the cumsum

does uninterleave just happen to go "backwards" somehow or am i missing something

emanresu A

Yeah, it pushes a[0::2] then a[1::2]

Unrelated String

huh, so yNY shouldn't act any different from NÐe

emanresu A

11:40

The direct equivalent of NÐe is ⁽Nẇ (although that doesn't save any bytes)

Unrelated String

lol

yNY is also just niced

*r

has that under feel to it

lyxal

@UnrelatedString I think you mean laurel

emanresu A

Hmm, I'm starting to wonder if the only way a jelly 10 is possible is by somehow arranging it to not use chain separators

Unrelated String

strictly speaking the Chain Separator is wholly unnecessary but yeah something has to break up the chain there

maybe some kind of convolution power thing could work but that probably comes out much longer

emanresu A

I tried some cursed stuff with caird's fork, didn't really get anywhere

Unrelated String

11:45

lol nice

mathcat

@UnrelatedString I'm doing some big brain ascii-only coding here

but like, how do you guys come up with this stuff?

emanresu A

I'm gonna go to sleep now, but _\ could be interesting... I haven't got it to work but something along those lines might.

Unrelated String

i've tried various forms of that too and it's been finicky but it REALLY feels like it should work

and yeah i need to sleep now too

_@/ kinda looks like a wave :P

UndoneStudios

12:22

Today I updated the Python 3 turtle docs

mathcat

~~weird flex but okay~~

UndoneStudios

Somehow the doc was showing wrong info for about 11 years but no-one noticed

mousetail

Happens more often than you think.

Glad it's fixed

UndoneStudios

You're welcome

Turtle is pretty much not heard about in these days

With a little smarts you could adapt it for everyday coding

But in its current form not so useful

mathcat

@UndoneStudios eh I disagree

UndoneStudios

12:33

@mathcat Why not?

mousetail

It's used a lot for teaching programing

I see a lot of questions on it on SO

mathcat

^

I don't see any practical uses here.

Even stuff as simple as drawing letters is awfully difficult.

mousetail

It's not intended for use in production, thus it's no surprise that it isn't good for that purpose

It's basically like scratch in that regard

UndoneStudios

@mathcat Can't you use the write() method

mathcat

huh I was pretty sure that didn't exist

one sec

I meant drawing letters, not writing them. :P

UndoneStudios

12:38

@mousetail Yeah that's the only reason I still use turtle. Personally, I'm dissatisfied with Scratch, as it is, so I'm trying to move on from using it (for 3 years I've been using it!). However, until pygame finally does download on my system, the only graphic system similar to Scratch that I can use is turtle.

@mathcat What's the difference?

mousetail

Pygame is much better, but still barely used in real games

mathcat

yeah, definetely.

mousetail

And too hard to get responsive for non-games

UndoneStudios

So what do y'all use?

mathcat

Java for game dev

mousetail

12:39

Rust + Bevy

UndoneStudios

I mean using Python for games

mathcat

Honest opinion: Don't.

mousetail

^ pygame is probably the best option

but don't is more realistic

For non-game GUIs QT

UndoneStudios

Well I haven't learned a new language (and it usually takes 2 years) so I'm going to still have to use Python for some time

Other than C, which is too verbose to do literally anything useful

mousetail

Learning more languages will help you understand python better

UndoneStudios

12:41

Like which?

mathcat

e.g. scala gave me some useful insights about fold (python's reduce).

lyxal

@mathcat sounds like you now have lots of FP Options :p

UndoneStudios

@mathcat Python's reduce?

mousetail

@UndoneStudios Some opinionated recomendations: Typescript, Rust, Kotlin, C#

mathcat

@lyxal I should probably learn Haskell

lyxal

12:44

@mathcat good be a good Option too.

mathcat

@UndoneStudios If you are interested in game development, as I said, I'd recommend coding in Processing to learn Java.

mousetail

For games I'd recomend C#+Unity engine

mathcat

I've always been afraid of FP and its obscure functions, but today I shall overcome my fear.

(and look up a haskell tutorial online :d)

lyxal

@mousetail tkinter is where it's at :p

mousetail

I still have nightmares about tkinter, trying to make it look not shit

The styling is so limited, it will always look like a windows 98 program

mathcat

12:52

oof true

mousetail

Maybe that style will make a comeback and tkinter will be hip once again

lyxal

@mousetail wanna know what else makes everything look like a windows 98 program? VB.Net

if there's something ever to have trauma about it's that

mousetail

True

lyxal

Me when I tkinter

mousetail

The alignment and background color could be fixed, but not the borders

Also one good thing: the dropdown menu from <select> is actually easier to style in TK than in HTML

lyxal

13:03

the repo that screenshot is from

fun times

mathcat

ew single-quote docstrings

lmao is that a reference to the oreo meme?

mousetail

The ><> program would just create a infinite loop

if the stack started with enough elements

lyxal

@mathcat knowyourmeme.com/memes/reeeeeee

mousetail

Real pros would not comment it out and actually write reeeeeeeeeeeeeeeeeeeeeee everywhere

Seggan

gotta be fun counting es

mousetail

13:11

Have 2 different modules that differ only by number of es

mathcat

that's ... evil

mousetail

import re as reeeeeeeeeeeeeeee
import regex as reeeeeeeeeeeee

Seggan

@UndoneStudios used to use pygame, LibGDX for Java/Kotlin, Unity is gud too (tho c# is not the best)

mousetail

C# is a slightly better Java

mathcat

from IO import BytesIO
from lO import BytesIO

Seggan

13:13

@UndoneStudios no.

@mousetail yep, but still bad

lyxal

@mathcat from 10 import 8ytes10

mousetail

If your other option is to use unreal with C++ I'd prefer C#

Seggan

i heard unreal it a bit more low level than unity

@mousetail c++ is one of the worst langs

maybe rust could be adapted for it

mousetail

It's a bit odd, very high level in some areas but very low level in others

Seggan

i like unity

mostly high level like scratch, gives low level when needed

lyxal

13:16

Gamemaker is cool for game making

Seggan

if it wasnt for c# id totally embrace it

mousetail

Gamemaker is actually really good too

Seggan

i was actually considering making a kotlin -> .net compiler

@mousetail github.com/MaikKlein/unreal-rust

lyxal

@mousetail not only do you get a PC/mac version, you also get an accidental android version :p

mousetail

My issue with unity is not as much C# but the way they keep changing things to fast before actually properly finishing a system. When I last used it there was a choice between a depricated old pathfinding system and a incomplete new one.

Seggan

13:22

~~roll your own~~

mousetail

Vulkan made that unachievable

Bevy is good though if you want a low level graphics enginge

Seggan

yeah vulkan is pain

opengl is too, but not as much

mousetail

OpenGL is pretty easy to use

Especially older versions

mathcat

13:36

@lyxal That's not even valid :d

New Posts

0

Q: Fibonacci Binary Squares

I was playing with the Fibonacci sequence in binary like so (note that the binary representations are written here from smallest bit to largest bit): 1 1 1 1 01 2 11 3 101 5 00...

code-golf sequence fibonacci

Ginger

13:54

so I've been doing some thinking about how my ideal interpreted language would handle packaging and distribution of source files

and I've come up with something I like

Seggan

@mousetail its easy to use, the basic stuff

but once you get to rotation matrices etc its all math beyond calculus

mousetail

Yea you need a good foundation in linear algebra, but you can't really avoid that no matter what library or enginge you use

Seggan

not necessarily

mousetail

I tutored linear algebra at university for years so it's not really a problem for me

Seggan

ofc vectors are a must for any game dev, but most non-rendering stuff is ~algebra 2 - trig level

mousetail

13:58

You really need a solid understanding of matricies and quaternions even in Unity

Ginger

in a nutshell, a collection of source files that import each other gets turned into a module, which can be grouped with other modules into one file to form a bundle, which can be run directly by the interpreter or turned into an "application bundle", a natively executable file which contains the bundle data and when run downloads and installs the appropriate interpreter version if one doesn't already exist and uses that to run the bundle code

Seggan

@mousetail i havent encountered them anywhere except maybe custom shaders

mousetail

When doing some custom physics things I needed them a ton. I once made a game with a plane that needed different friction values depending on it's orientation relative to it's direction of movement

Transcript for