The Nineteenth Byte

betseg

18:00

@TimmyD ಠ_ಠ

TimmyD

What's that face for?

betseg

It's actually disapproval face but I use it like O_o

TimmyD

Ah - I thought you were somewhat irritated at me for sharing the link.

CMC: Write the saddest story you can, using only four words. pop-con

betseg

I joked about that word when I was little, and it got somebody killed. Woah.

Fatalize

@mınxomaτ any news about the podcast?

@TimmyD "Daddy killed mommy, lol"

TimmyD

18:05

"Loving couple. Empty crib."

mınxomaτ

If there are news the people involved will receive an email. I'll send one later today, but there are no public news atm.

Feb 25 at 22:56, by 9cHeruinr96a

@RikerW Killing people isn't nice

TimmyD

@Fatalize Jeez, that "lol" at the end really seals it. That's moved beyond sad into horrifying. But, reminds me of a bad joke --

Little kid: Mommy, why is daddy running back and forth?
Mother: Be quiet. And hand me another box of ammo.

Which is not as good as my all-time favorite joke

So these two muffins are sitting in an oven, baking. The first one looks at the second and says "Wow, it's getting hot in here." The second one turns back and goes "Holy S---! A talking muffin!"

Anonymous

@TimmyD Buying gf, 20 gold

TimmyD

Oh jeez.

Anonymous

@Fatalize The issue is, the stats are the only exciting part of the sport. The only reason the sport is played is to generate those stats :P You could probably replace the games with the output of an RNG and it'd have the same result.

Fatalize

18:18

finally a murican that agrees with me

DrMcMoylex

@TimmyD Two monks are sitting in the monastery. The first one looks at the other and says "Wow, it sure is quiet in here." The second one turns back and goes "Holy S---! A talking monk!"

Anonymous

The only other part of the game that could be considered exciting is people in the crowd getting hit by pop flies and foul balls, but that's not guaranteed to happen

feersum

They can also get hit by flying bats, or pieces of bats.

Fatalize

And cute girls in the crowd. Do they show these in american sports?

TuxCopter

@TimmyD “This is very sad”

Anonymous

18:21

I'm pretty sure the only reasons people go to baseball games anymore are 1) They want to catch a ball, 2) They're being pressured to go, 3) They have a stake in the outcome of the game (team owners and gamblers), or 4) They want to drink beer and get into fistfights.

Helka Homba

@Mego Or the suspense, the athleticism, the camaraderie, the competitiveness, the atmosphere, the novelty, the history. There are plenty of good reasons why millions of people like baseball and other sports.

Fatalize

the athleticism

Anonymous

@HelkaHomba The athleticism in baseball is only slightly higher than in golf (mostly thanks to rampant steroid (ab)use)

Anonymous

The MLB's refusal to play hardball and crack down on steroid abuse killed the competitiveness long ago

TimmyD

The athleticism in golf comes from the discipline of repeatability, not because the competitors look like bodybuilders.

Fatalize

18:32

Yeah but in Baseball they are supposed to run and jump to get the ball yet it seems there are countless fat guys playing at the best level

Helka Homba

@Mego Well, I'm pretty sure it's higher than the athleticism in programming.

Fatalize

That's not a sport, so...?

TimmyD

Yeah, we haven't renamed our site to Code Sport yet

Kritixi Lithos

@HelkaHomba You use a lot of brain muscles in code-golfing. I consider it like exercise.

Anonymous

@HelkaHomba Nice topic change. You should run for president.

Geobits

18:35

Wait, isn't rampant steroid abuse an indicator that you do need some sort of athleticism (even if only brute strength) to play the game at the top levels?

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Binary Convolution A binary convolution is described by a number M, and is applied to a number N. For each bit in the binary representation of M, if the bit is set (1), the corresponding bit in the output is given by XORing the two bits adjacent to the corresponding bit in N (wrapping around whe...

0

A: Sandbox for Proposed Challenges

K-Means Clustering (Wikipedia) The task here is rather simple, perform a single iteration of a k-means clustering algorithm on a binary matrix. This is essentially the setup task for the main k-means algorithm, I felt the setup might be easier and entice golfing languages to give it a shot as we...

Anonymous

@Geobits Sure. But the top level of athleticism in baseball is a whole lot lower than other sports.

Geobits

Then why the steroids?

Fatalize

Too lazy to exercise

Geobits

Pretty sure that's not how it works :P

DrMcMoylex

18:37

@Mego I disagree. Can you throw an object at 95 mph?

Anonymous

Because every sport has players roiding up. The MLB is just especially bad because it doesn't have frequent enough tests or harsh enough penalties for roiding.

Anonymous

@DrMcMoylex Pitchers are the sole position where there is a significant amount of athleticism. That's 1/9 of the team.

New Main Posts

1

Q: Draw some mountain peaks

Write a program or function that draws a mountain range, where each subsequent larger mountain peak is "behind" the ones in front, and alternates which side is visible. This is a mountain range of size 1 /\ This is a mountain range of size 2 /\ /\ \ This is a mountain range of size 3 /...

DrMcMoylex

OK fine. Can you hit an object moving at 95 mph?

Anonymous

Sometimes

Geobits

18:39

How big is the object? :P

Anonymous

Which is about all you can say about MLB hitters

Geobits

Yeah, every level of sometimes is equivalent ;)

Sometimes I stub my toe. Sometimes stars explode.

Fatalize

Not saying that Baseball doesn't require athelticism, just that I don't understand why teams, fans and players are completely cool with players being way too fat when being more athletic would make them better runners/hitters

TimmyD

Well, but would it?

Fatalize

Are you arguing that being more athletic makes you run slower?

Anonymous

18:41

The best batting average from this past season was .367 - that's roughly one in three balls hits

Geobits

@Fatalize If it would do that, I'm sure at least one team would try that and win every year, or until others caught on. You think nobody's thought of that?

@Mego I'm fairly sure my average against an MLB pitcher would be much closer to zero.

Helka Homba

@Fatalize I'm just saying that until you can hit a home run or throw a 100mph pitch, you can't complain about lack of athleticism. I'm not even a baseball fan, I just don't get it when people try to make it sound like a thing enjoyed by millions of people around the works is not worthwhile.

TimmyD

@Mego No, that's roughly one in three times at bat the batter put the ball into play and made base safely.

Fatalize

@HelkaHomba So you can't complain about trump because you know less than him about politics and money by the same argument I guess?

TimmyD

@Fatalize No, I'm saying that "athletic" =/= "looking like a bodybuilder"

Anonymous

18:44

@TimmyD True. Unfortunately, pure hits isn't a tracked statistic by the MLB

TimmyD

@Mego AB - K - BB

Adnan

@Dennis Could you please pull 05AB1E?

Anonymous

@TimmyD I don't see a K

Dennis

@Adnan Done.

Adnan

Thanks :)

TimmyD

18:45

@Mego SO

(at bats) - (strikeouts) - (base on balls / walks)

That equals times that the batter has struck the ball and put it into play.

Helka Homba

@Fatalize That'd be more like complaining about/insulting a team. Gripe at the system, but don't write it off completely.

DrMcMoylex

Anonymous

@HelkaHomba So to better fit the analogy, we shouldn't gripe about the American political system and its many flaws (FPTP voting, indirect election of presidents, lack of term limits for Congress, corruption) unless we have degrees in political science?

confusedandamused

Can anyone give a few examples when you would want to implement a sorting algo - or would make use of one?

Anonymous

@confusedandamused When you're doing homework? :P

confusedandamused

18:56

I mean outside of that lol

Or interview question - like actually in a project

betseg

When you need to sort stuff?

Anonymous

That was heavily-loaded sarcasm to mean "we're not going to do your homework for you"

TimmyD

@confusedandamused Arranging data for human consumption. We humans like names in alphabetical order.

confusedandamused

Oh it isn't homework I'm literally just asking - I'm not even in school (wow is that weird to say)

Anonymous

To give you a nudge in the right direction: there are many algorithms that perform better with (or even require) sorted input.

confusedandamused

18:58

They require the input to be sorted in a specific way for example?

feersum

Bubble sort performs better with sorted input, for example.

5

betseg

:O

confusedandamused

Are most of the sorting algos used with integers as opposed to strings?

Helka Homba

@confusedandamused The language you're using might not have the algorithm you need

Anonymous

Any sorting algorithm works with any sortable input

Maltysen

19:01

@feersum doesn't it always take the same amount of steps?

feersum

@Maltysen No.

confusedandamused

@HelkaHomba Well in c# if you're sorting elements of an array you could use Array.sort() are some algos built under the hood into a language> '

?*

Anonymous

@Maltysen A smart bubble sort implementation will stop once it no longer modifies the list

Anonymous

Linear search performs better with sorted data because you can stop at a certain point and know for sure that you have either found what you're looking for, or it doesn't exist. With unsorted data, you have to search through the entire list. With binary search, the data is required to be sorted.

Anonymous

@confusedandamused Different languages use different algorithms. Lots use quicksort. Some use more modern sorting algorithms that work on heuristics and combine parts of multiple algorithms (like timsort, which is used in Python).

TimmyD

19:06

@Mego No relation.

Anonymous

@TimmyD :P What algo does PowerShell use?

TimmyD

@Mego Probably quicksort, since I think that's what .NET uses.

Anonymous

Partial credit

Anonymous

.NET uses insertion sort, heapsort, or quicksort, depending on the size of the data

TimmyD

msdn.microsoft.com/en-us/library/afwbytk2(v=vs.110).aspx

Literally just found that

Anonymous

19:11

Yep - introsort was the name I was trying to think of but failed to

feersum

That page claims quicksort is n log n in the worst case.

confusedandamused

Is it ever necessary to implement these sortings algos yourself then?

Shebang

Usually no.

MetaEd

@feersum That's not nearly the worst case.

Shebang

Unless it's an interview question ;)

Anonymous

19:15

Usually no, but it's not a bad idea to learn them

TimmyD

@feersum Hah, that's funny.

confusedandamused

@Mego that's my goal currently - that and compiler and CLR theory

or just basics to learn more

MetaEd

16

A: Are there any worse sorting algorithms than Bogosort (a.k.a Monkey Sort)?

You should do some research into the exciting field of Pessimal Algorithms and Simplexity Analysis. These authors work on the problem of developing a sort with a pessimal best-case (your bogosort's best case is Omega(n), while slowsort (see paper) has a non-polynomial best-case time complexity).

feersum

It's right above the section heading "Notes to callers".

Anonymous

@feersum Yeah, Microsoft is real bad at stuff like that

Anonymous

19:16

@MetaEd That's not relevant to that comment...

mınxomaτ

@TimmyD Implementation details: github.com/dotnet/coreclr/blob/…

TimmyD

Yeah. I don't go to Microsoft for non-technical data. They tell me what argument such-and-such to such-and-such command does? Perfect. They say why it does that? Usually at least somewhat wrong.

MetaEd

@Mego That depends. If you're looking for a worst case sort, it's relevant.

Anonymous

@MetaEd The comment was literally "MS says quicksort's worst case time complexity is n log n"

TimmyD

@MetaEd Err, no ... that comment specifically referenced quicksort, not sorting in general.

Anonymous

19:19

Tsk tsk, poor reading comprehension skills from an English.SE mod :P

Anonymous

We're going to need your diamond back

MetaEd

@Mego Probably because not enough coffee.

confusedandamused

Anyone have recommendations on reading to learn about compilers?

mınxomaτ

@Mego The implementation switches between many different modified quick-sort implementations (see above link).

confusedandamused

I was reading: compilers.iecc.com/crenshaw

Anonymous

19:20

@MetaEd If you require coffee to function, you fit right in here :)

MetaEd

@Mego You're saying coffee is a required argument?

TimmyD

@confusedandamused Have you read anything by Donald Knuth?

confusedandamused

@TimmyD I don't believe I have - but I think he is mentioned more than once in "Coders At Work"

Anonymous

@mınxomaτ Sure, but a bad choice of pivots each time will still get you n^2

MetaEd

@Mego I'm sure there's a Java pun in here somewhere but I'm too dense to find it.

Anonymous

19:22

@MetaEd Always

mınxomaτ

@confusedandamused Well, the standard book for compiler design and compiler toolchain frontends is the "dragon book". In-depth books about compiler backends, specifically optimization are Agner's manuals: agner.org/optimize

Anonymous

@confusedandamused That's a good starting point, but compiler theory has progressed a lot in the last 21 years. Aaaand I've been ninja'd to the dragon book recommendation.

confusedandamused

@Mego Do you like the Dragon Book thus far? Most posts/books seem vastly out of date as you mentioned

Anonymous

I don't own a copy of the dragon book, and I haven't read through Crenshaw's tutorial fully

Anonymous

So I can't offer an opinion :P

Anonymous

19:24

However, I highly doubt either of them go into JIT compilation, which is a very neat new-ish technique that a lot of compilers are leveraging

mınxomaτ

Another way to learn is just to do it. I found that you don't have to know anything about compiler to make a decent one :D

confusedandamused

Yeah in my time at school they didnt offer a compiler course which I would have taken

@mınxomaτ That's what Crenshaw said too :)

But it seems his is written/written for Turbo Pascal

mınxomaτ

@confusedandamused That course would have probably used the dragon book as it's basis. But as mentioned before, it doesn't really go in depth in terms of backend (LLVM (or IR in general), optimization) etc.

Anonymous

My uni offered a compiler theory class, taught by the software engineering professor. He usually taught it by making the first project (due about 1/3 of the way through the course) "make a compiler". The rest of the class went over various techniques and optimizations to be applied to those rudimentary compilers.

mınxomaτ

If you are brave, you can go through the source of my first compiler here: github.com/turbo/Perseus. It's a single-pass non-optimizing compiler with an integrated linker and assembler.

It's even neatly categorized: github.com/turbo/Perseus/tree/master/Source/src, which is very unusual for me.

confusedandamused

19:33

Is it mostly written in vb?

mınxomaτ

It's entirely written in VB. You can compile it and decompile with ILSpy to get a C# version.

I'd like to think that VB is very easy to understand though. It also uses no .NET components.

confusedandamused

I just just confused when viewing the languages in the repo it shows 36% C++ 21% SourcePawn and some other stuff

mınxomaτ

Yeah, the code lens on this repo is wrong.

It gets confused by all the example code.

confusedandamused

:D

I'll start looking through this for fun then thanks for the link

quartata

SourcePawn uses .inc for header files I believe which is why it gets confused on that one

mınxomaτ

19:46

Well, Perseus is Cish like Pawn, too. So that adds to the confusion with C++.

Anonymous

Worst sorting algorithm I can come up with:

Anonymous

import random

def bogosort(data):
    while not sorted(data):
        chunk = random.shuffle(data)
    return data

def sort(data):
    for i in range(1, len(data)+1):
        data[:i] = bogosort(data[:i])
    return data

Anonymous

The algorithm is O(n^2 * n!)

Anonymous

That's the worst worst-case performance I can come up with that actually has a decent chance of terminating

feersum

I think 1! + 2! + 3! + ... + n! would be O(n!)

So the complexity is actually the same as bogosort.

O(n!) average

Anonymous

19:53

bogosort is O(n * n!) (since you have to check each iteration if the list is sorted, which is linear). You perform bogosort n times, giving n^2 * n!

feersum

Sorry, O(n * n!)

confusedandamused

How do you figure?

xnor

it's only O(n!) because it's at most 2*n!

feersum

Do you think sum(i! for i in 1...n) is not O(n!) ?

Anonymous

Oh wait you're right

Anonymous

19:54

I was thinking sum(n! for i in 1...n)

Anonymous

So in the worst case it's no worse than normal bogosort, but for basically every other case it will take longer

TimmyD

Sep 12 at 14:12, by TimmyD

Bubblesort is best sort. Cuz it's got bubbles.

feersum

We should keep track of how many times we have shuffled, to ensure we don't try the same permutations again.

After 2^19937 - 1 failures, look for a new Mersenne prime and start over with a larger PRNG.

Anonymous

@feersum That seems like it would improve the running time. status-declined

flawr

@TimmyD Spaghetti-sort is better, because it's got spaghetti.

Anonymous

20:01

Spaghetti > bubbles

Anonymous

(comparing taste, of course)

flawr

Also comparing calories.

Anonymous

And number of religions

Anonymous

Now we just wait until people jokingly start "worshipping" the Purple Bubble Monster and declare "war" on Pastafarians

Anonymous

@feersum Bonus points if we don't keep track of which Mersenne primes have been tried, to allow for repeats :P

Nathan Merrill

20:05

I have a bunch of checks to prevent the user from doing he shouldn't in my KoTH. I think I'm going to throw a StopCheatingException in those cases

feersum

We could flip a coin until we get heads. Then test 2^(number of times we got tails in a row) - 1 for primality using the regex engine.

Dennis

@Mego Slowsort has best-case performance n^(log2(n)/(2+eps)) because it uses a multiply and surrender approach.

MetaEd

@Dennis Yeah. That's the one I linked to.

Dennis

Oops, didn't scroll up that far.

TimmyD

@NathanMerrill try{cheat}catch(cheater){throw Exception.LanceArmstrong}

flawr

20:10

@Dennis I like that paradigm

MetaEd

However my personal favorite, though it isn't guaranteed to complete, was the miraclesort on the same page. Basically, repeat (null) until (list is sorted). Cosmic rays flipping bits on the list should eventually result in a sorted list and the test will succeed.

flawr

subtract and defeat

Nathan Merrill

@MetaEd nah, statistically, what'll happen is the bit flip will occur on the boolean returned by the "list is sorted" function

so, you'd end up waiting for a miracle and still not get what you want

TimmyD

@NathanMerrill Kinda like Duke Nukem Forever.

MetaEd

@NathanMerrill How about keeping code and data separate and running the code on a CRC stick.

mınxomaτ

20:13

This appeared after closing Atom. I can't understand how you can create a program so damn resource-wasting and inefficient as Atom. And then GitHub calls this app "high-performance" with a straight face.

Nathan Merrill

atom not responding was the reason I stopped using it

I've never had that problem with Notepad++

flawr

Yep, <3 NP++

mınxomaτ

Notepad++ is just SciTe made non-crossplatform and with added bloat.

Nathan Merrill

depends on what you consider bloat. I actually like the features it provides

mınxomaτ

Like what? Compared to a pure SciTe I mean.

MetaEd

20:18

Oh, hey, here's your problem. You're not using vi.

DrMcMoylex

Oh dear, here we go again

(Even though I agree)

Geobits

yesterday, by Geobits

Watching people argue about text editors is kinda like watching people fight over whose poison ivy rash is better.

MetaEd

I'm sure I should be using something else, but I'm an old dog.

El'endia Starman

Jeez, I just might consider switching to emacs because some vi/vim users are so snobby.

quartata

emacs users are a lot snobbier actually

flawr

20:21

@mınxomaτ Do you use SciTE? It's the first time I heard of that editor (currently downloading=)

Nathan Merrill

@mınxomaτ I like dialogs to set options, as well as several plugins I use

mınxomaτ

What NP++ provides via the GUI in terms of configuration is a proprietary abstraction (and thus subset) of what SciTe provides on it's own with the integrated Lua scripting. I don't claim that any editor is better here, but I strongly dislike what NP++ does to an editor that was meant to be lightweight, modular and cross-platform.

Geobits

@El'endiaStarman use ms word if you want to get to all of them

El'endia Starman

@quartata There's a strange lack of them in this room though.

MetaEd

@El'endiaStarman I tried, really I tried. As an old TECO user I really, really tried.

Shebang

20:22

Woo, 5,000 :)

3

Shebang, Canada

5k 14 40

code-golf string math kolmogorov-complexity ascii-art hello-world number

El'endia Starman

@Geobits Muahahaha...

mınxomaτ

@flawr Use it for basically everything now. But again, use what's better for you.

DrMcMoylex

Noice!

feersum

I downloaded TECO for lolz and made a couple answers here with it.

It actually was the shortest of all languages on one question.

flawr

@mınxomaτ That is why I'm gonna test it:)

MetaEd

20:24

@feersum Runtime or source code?

mınxomaτ

@flawr When you get vanilla SciTe, you have to configure it to your liking first. I'm using the AutoIt fork of SciTe, which doesn't add any major features, but comes pre-configured with settings I like.

feersum

The thing I downloaded had source also, but I just used the binary.

MetaEd

@feersum I mean, was your runtime short or your sourcecode short?

feersum

Source code of course.

Its runtime is quite slow... (in general)

MetaEd

@feersum Right, that's what I would expect.

Also, like APL it's write-only.

mınxomaτ

20:27

@flawr Important reading: scintilla.org/SciTEDoc.html

MetaEd

I see y'all do Euler Project here.

Anonymous

@El'endiaStarman We try to be reasonable about text editors in this room. Indentation, on the other hand...

Destructible Watermelon

YO

flawr

@mınxomaτ Thank you for all the info. For what I'm using it, both npp and scite seem to do about the same.

mınxomaτ

Well, the editor text window inside NPP is just SciTe wrapped. You can actually detach the editor from NPP and get a standalone SciTe instance :D

TuxCopter

20:35

how

mınxomaτ

@TuxCopter If you know how native windows in Windows work, you can just extract the child window, give it a dummy parent and delete the old NPP shell

Scintilla, the editor engine powering SciTe, NPP, Geany etc. is always just a child window. The API is basically just window messages sent either way.

flawr

@mınxomaτ Well I really just use it for the syntaxhighlighting, regex find&replace, and "rectangle selecting"

TimmyD

@flawr I love that feature of NPP

mınxomaτ

@TimmyD That's all provided by scintilla, not NPP

The entire editor field is a separate program.

feersum

rectangle selecting?

mınxomaτ

20:38

@feersum Alt-Select

mbomb007

This might make a good inspiration for a challenge:

wired.com/2016/11/weird-globe-folding-map-isnt-perfect-close

flawr

TimmyD

@mınxomaτ Huh. TIL.

mbomb007

mınxomaτ

@flawr Also, Ctrl-Click to place multiple cursors in the code at once.

@TimmyD That's why NPP doesn't really add anything apart from the plugin-system (which is tied to Windows).

flawr

20:39

@mınxomaτ npp can't do that!

@mınxomaτ really neat feature!

mınxomaτ

@flawr Because it dumbs down the scintilla protocol

Nathan Merrill

@flawr nah, it can. its just a setting

mınxomaτ

Oh, that's fine then

TuxCopter

Sublime has multiple cursors too

flawr

@NathanMerrill where can you find that?

Nathan Merrill

20:41

Preferences -> Editing

flawr

Oh great, thank you! <3

PhiNotPi

@mbomb007 I've seen this image posted in a ton of places, and I honestly don't see what's impressive about this projection compared to other projections.

flawr

@NathanMerrill I wonder why they do not activate that by default

Nathan Merrill

dunno

TimmyD

@mınxomaτ Ah. I do use several plugins decently often (mainly PowerShell Lexer and HTMLTag).

And the hex editor.

mınxomaτ

20:50

@TimmyD What's the PowerShell Lexer?

El'endia Starman

@PhiNotPi It's a rectangular map where the areas are proportioned correctly.

TimmyD

@mınxomaτ Pretty much what it says. Syntax highlighting and whatnot for PowerShell.

mınxomaτ

@TimmyD BTW, as mentioned before, SciTe relies on LUA as an open-source extension API. There are many plugins/extensions for SciTe. An example for hex editing would be sl-hexedit.luaforge.net

TimmyD

I don't use it as much, now that the PowerShell ISE (Integrated Scripting Environment) has gotten better.

mınxomaτ

@TimmyD SciTe has full support for PowerShell, including folding. Also, it supports running PS with the integrated console

That shouldn't be something you need to add to NPP. Technically, it should be supported out of the box.

feersum

20:54

Get cracking

PhiNotPi

^ one of the coolest map projections

Dennis

@feersum >:U

Geobits

I prefer accurate directions over accurate areas for most purposes

Destructible Watermelon

@feersum but, what is the sequence

mınxomaτ

@TimmyD This leaves HTMLTag. As far as I can tell from their SF page, SciTe supports all these features without extensions.

feersum

20:55

Oh right, I should wirte the sequence...

it's primes.

Destructible Watermelon

oh

TimmyD

@mınxomaτ Neat. I'll take a look at SciTe.

mınxomaτ

Read the docs first to get a better idea about what might be missing for you: scintilla.org/SciTEDoc.html scintilla.org/SciTEFAQ.html

Destructible Watermelon

we don't even know if the generated befunge program is longer or shorter

Geobits

Shorter

betseg

20:58

Yay Vim turned 25 today

Destructible Watermelon

@Geobits -_- how do you know

Oh

I was referring to the one generated by the shorter code

TuxCopter

github.com/westerp/ebf-compiler/blob/… lolwut

Geobits

@DestructibleWatermelon Who says I do? I haven't even looked at what you're talking about :P

Destructible Watermelon

:p

MetaEd

@betseg vi itself is 40 years old, right?

Destructible Watermelon

20:59

Uh, I had an idea for a koth

Transcript for