The Nineteenth Byte

12:00 AM

I thought I'd create the help-center-rewrite tag, just to make it easier for people to spin their proposals out when necessary. Saves flagging for status-tags, and pushing stuff to your Jira board :P

New Posts

12:13 AM

0

Q: Finite time, infinite independent randomnesses

For each positive integer \$i\$, output it for \$\frac 1{i^2}\$ probable. All these events are independent. Your program should halt almost always in finite time. Notice that \$\sum_{i=1}^{\infty}\frac 1{i^2}=\frac{\pi^2}6<\infty\$, so there's almost always only finite outputs. Shortest code wins.

Dude coinheringaahing

@NewPosts l4m2 really needs to start using the Sandbox

Bubbler

They do use the sandbox, but only occasionally

and multiple attempts to use it have failed afaik (resulted in negative score and closure), so I guess they just decided not to use it

Dude coinheringaahing

They're at 29/5 for Curious, that's not a good indicator

They seem to have a habit of giving up on questions that get closed, which is unfortunate when they refuse to improve their questions in the first place

@BrowncatPrograms New Posts (userscript, not bot) is still broken with the header id changes from a couple of months back

Bubbler

12:31 AM

@Dudecoinheringaahing Often the topic they bring up are either very hard to grasp (so most of the CGCC audience see it as unclear) or not backed up by research at all (which means we don't know whether some problems are even solvable or not)

See, the newest one is 100% clear to me but unclear to four users already

Dude coinheringaahing

> For each positive integer \$i\$, output it for \$\frac 1 {i^2}\$ probable

That's the entire challenge spec, and I have no idea what it means. Everything else is just fluff

If they defined what " for \$\frac 1 {i^2}\$ probable" means, I'll happily retract my close vote, but challenges should make a best effort to be as widely understood as possible, and that involves defining terms as well as you can

Bubbler

@Dudecoinheringaahing Better now?

Dude coinheringaahing

Much; CV retracted

emanresu A

They've edited it

Dude coinheringaahing

Bubbler edited it

emanresu A

12:45 AM

O

Thanks Bubbler

Worth answering?

Dude coinheringaahing

If I can figure out how :P

Ugh, handling 1 is annoying

exedraj

> However, your program should halt almost always in finite time.

but the python program provided doesn't halt

Bubbler

What, are you doing it in Jelly or something?

exedraj

so do we print until the condition is no longer true?

Dude coinheringaahing

@Bubbler ...I mean, what did you expect :P

hyper-neutrino

12:51 AM

wait what, so the idea is to halt it at some random position that makes the theoretical probabilities equivalent?

how does this even work

how is it possible to output infinitely many numbers in finite time

is it possible to learn this power

exedraj

1 min ago, by exedraj

but the python program provided doesn't halt

1 min ago, by exedraj

so do we print until the condition is no longer true?

already asked.

Bubbler

I think the intended approach is to pick the maximum n using the probability formula I added, and then do random generation within that range

emanresu A

Oh. Wait, so implementing the pseudocode's not valid?

hyper-neutrino

wack

Bubbler

That way you get the equivalent probability distribution within (almost always) finite time

emanresu A

12:53 AM

@Bubbler Is implementing the pseudocode invalid?

exedraj

^

Bubbler

Yes, it's invalid

Dude coinheringaahing

@Bubbler Wait, but how do you pick \$n\$?

hyper-neutrino

wait so can i just do like n = random(); for x in range(int(1 / (n * (n + 1)))) lol?

Dude coinheringaahing

Frick, I can;t un-retract my VTC

exedraj

12:55 AM

neither

hyper-neutrino

if you two both think this should be closed then i would be happy to supply the fifth theoretical close vote since i also don't really know what i'm meant to do here

Bubbler

@hyper-neutrino I need to think about it

exedraj

@hyper-neutrino I'd re-cast my vtc

Dude coinheringaahing

Same here

hyper-neutrino

i think i kinda get the idea behind it but i don't know if my interpretation would be correct because there is a chance i am not understanding bubbler who also may be misunderstanding l4m2

emanresu A

12:56 AM

I've cast a VTC

Dude coinheringaahing

If the pseudo code isn't correct, I have no idea what the challenge is asking for

Bubbler

It is almost correct, it just doesn't halt even when there is nothing to output further

Okay, guys all wait until I supply a correct answer

This is the only way to make it clear I think

exedraj

Clarity shouldn't come from answers though

Dude coinheringaahing

ಠ_ಠ It's 2am, why do people keep posting challenges at 2am for me to FGITW, only to lose out to hyper???

hyper-neutrino

a reference implementation would definitely help but more so i'd rather not look through that math equation just to figure out what "output infinite numbers in finite time" exactly means

Dude coinheringaahing

12:59 AM

"output infinite numbers in finite time" sounds like something WW would say, then spent an hour mind-controlling us here to think was a totally normal statement

hyper-neutrino

accurate

emanresu A

^

OldSandboxPosts

Sandbox posts last active a week ago: Display a number in Toki Pona, Is this Flow Free puzzle trivial? [code-golf] [grid] [decision-problem] [puzzle-solver] [array-manipulation]

Bubbler

Yeah, having WW here would have definitely helped :/

exedraj

@OldSandboxPosts ah yes, tags in the header

wonderful

Dude coinheringaahing

1:01 AM

Stand back, I'll summon him: Brain-flak, Haskell, Lost

Damn :/

COMBINATOR PARSING

Let me try.

LAMBDA CALCULUS

There ain't enough punks in this palace

wow I really thought that would have worked ;p

emanresu A

Let me try

@WheatWizard

Dude coinheringaahing

1:02 AM

> And we make a lot of noise, because we love each other.

It's no longer punks

exedraj

well if there ain't any code golf to do, I'mma go back to wikiracing

this is where I'm playing if y'all want to join

Dude coinheringaahing

Group Code: 859364

Link

@Dudecoinheringaahing Why are you like this?

exedraj

I'm in

Bubbler

Edited in a correct algorithm

Dude coinheringaahing

I have no way of knowing whos in the the group, this is awful design :P

Bubbler

1:07 AM

Well, if this doesn't get reopened, you know why l4m2 does not have a good question history

Their challenges are hard and don't give any explicit ways to solve it. They're not even CS theory-based, so it is very hard to persuade the general CGCC public that it is solvable. (Bad English makes it worse but we can fix it. We can't fix the core of the challenge though)

Dude coinheringaahing

You can make difficult to solve, highly specific challenges that are high quality and understandable. They all too often make the assumption that everyone knows that they're talking about, and anyone who is confused just doesn't understand the topic

Some basic background/explanation should be the bare minimum behind a challenge. As much as we like short code, challenges shouldn't aim to be short. They should aim to be clear and concise. l4m2's are very often concise, but not clear

exedraj

@Dudecoinheringaahing can you see how many people are in the wiki race group?

Dude coinheringaahing

@exedraj Nope

exedraj

oh

wonderful.

Dude coinheringaahing

Yeah, it's not very well designed

Bubbler

1:19 AM

@Dudecoinheringaahing Honestly I have no idea what kind of background information is needed to explain this one. Should I add a full intro to probability theory?

Probably that's a sign that it is simply not a good idea to bring to CGCC...

Niko

@NewPosts Can't understand Almost always finite

emanresu A

@exedraj How do you start?

Niko

is Almost always finite means it's finite, or it just most likely finite

exedraj

@emanresuA idk, caird has to start the round

Bubbler

@Niko Almost always

emanresu A

1:21 AM

@exedraj O

Dude coinheringaahing

The two key parts of the challenge are "For each positive integer i, output it with 1/i^2 probability" and "However, your program should halt almost always in finite time". The first part is, I hope, understandable to most. The second part needs some explanation. Infinities are weird, and especially so when you combine them with finite numbers/spaces

emanresu A

Code golf > Lyxal Watson? Was that intentional?

exedraj

?

Dude coinheringaahing

@exedraj Give me a start and end :p

Bubbler

It means with 100% chance, but there can be an infinitesimal chance that the program does not finish

Like bogosort

exedraj

1:22 AM

@Dudecoinheringaahing Code Golf to Shrek

Niko

the can't I just 1+3+5+7+9...?

emanresu A

@Dudecoinheringaahing Shrek > Kansas City Chiefs

Dude coinheringaahing

@Bubbler So why is the pseudocode not valid?

That'll halt with a 100% chance, as i tends to infinity

Bubbler

First one is not valid because you never break out of the while loop

so there will be no further output with 100% chance but it still runs for nothing

Dude coinheringaahing

So would Try it online! be valid?

So who is aprt of the wiki race game?

exedraj

1:24 AM

me

Bubbler

@Dudecoinheringaahing It doesn't make each number printing independent

emanresu A

mde

Dude coinheringaahing

@Bubbler Why not? The calculation is made anew each time, with the updated i

Bubbler

Printing 3 depends on 2 being printed, for example

Dude coinheringaahing

Started the race

user

1:26 AM

@exedraj Can’t say I do (want to be a big plate tgat is)

emanresu A

@Dudecoinheringaahing How did you beat me

Dude coinheringaahing

@Bubbler So for each integer, it should have a non-zero independent chance of being printed, but the program should terminate in finite time?

emanresu A

Oh I see

less random clicking

More thinking

Dude coinheringaahing

I have no idea how to consolidate those two

exedraj

I screwed up with a few misclicks at the start

emanresu A

1:28 AM

@exedraj French Third Republic>?

exedraj

@emanresuA that's me trying to get to United States

Bubbler

@Dudecoinheringaahing Exactly, except that each integer should have exactly 1/i^2 chance of being printed independently

not just nonzero

exedraj

@emanresuA because I didn't go down the path I initially wanted, I had to make a few corrections

Dude coinheringaahing

@Bubbler Yeah, that bit is fine with me

emanresu A

@exedraj Code golf → Data compression → Data file → Computer file → File format → List of computer standards → Software standard → Technical standard → Standard

As you. can see, I got a bit lost

Dude coinheringaahing

1:29 AM

The bit I still don't get is the fact that every integer has to be considered, yet it still terminates in finite time

exedraj

@emanresuA If I'd seen Andrey Kolmogorov, I'd have gotten to US quicker

emanresu A

True

Dude coinheringaahing

Additionally, by that interpretation, your pseudocode isn't correct, as values greater than floor(1/r) are never considered

Bubbler

@Dudecoinheringaahing That's where my added formula kicks in. This allows to handle the infinitely long tail of integers with a single random variable

Dude coinheringaahing

@exedraj Amateur :P

exedraj

1:30 AM

@Dudecoinheringaahing Sushant City, Jaipur to Ryan Mishler

target wikipedia page

Dude coinheringaahing

Someone give me another wiki page, I don't trust lyxal :P

exedraj

I literally got those using the random page link

Dude coinheringaahing

Oh yeah, that's a thing :P

Bubbler

@Dudecoinheringaahing The values greater than floor(1/r) are excluded from consideration because every integer greater than that is known to be not printed by the single random variable

Dude coinheringaahing

@Bubbler How is it "known"? It's a random variable?

Aaron Miller

1:32 AM

@Dudecoinheringaahing The wiki race sounds fun, I've gotta do a thing, but I'd be down for it in a few minutes

Dude coinheringaahing

@exedraj @emanresuA Y'all ready for rounds 2, 3?

exedraj

yep

emanresu A

Yep

Dude coinheringaahing

Ok, I'll do round 2, then wait for Aaron

Aaron Miller

You don't have to wait; I might be a little while, so I'll just let you know when I get back.

exedraj

1:35 AM

dang it

I got to Jaipur

Dude coinheringaahing

ಠ_ಠ I was on the page for India

emanresu A

Have we started?

exedraj

oh we got nowhere near it

Bubbler

@Dudecoinheringaahing I define r to represent exactly that. The chance of 1 being the maximum in the entire output is 1/2, the chance of 2 being so is 1/6, 3 -> 1/12, etc. If 1/2<r<1, 1 is the maximum (and therefore I know 2 and onwards won't be printed, so I just terminate the program after printing 1), if 1/3<r<1/2, 2 is the maximum, etc.

exedraj

@Dudecoinheringaahing looking at the "what links here", you would have had to go to Bollywood > Hindi Medium > Anand Lok, Jaipur > Target

emanresu A

1:37 AM

have we actually started?

exedraj

we already finished

Dude coinheringaahing

We've finished :P

emanresu A

O

Dude coinheringaahing

I've set up the next one to be 3 minutes, with 10 seconds in lobby. Ready?

exedraj

No

emanresu A

1:38 AM

Yep

exedraj

Not ready

Give me a minute

Dude coinheringaahing

It's 2:30am here, and I have to be up in 8 hours. I Think I can get away with that

exedraj

ight ready

emanresu A

ready

Dude coinheringaahing

started

:P

emanresu A

1:40 AM

I see what you did there

Ryan is. a senator

Dude coinheringaahing

Applaud my stupidly close following of US politics :P

emanresu A

Oh well I won

Can you give one of us admin, since you need to sllep?

exedraj

just make a new group

Dude coinheringaahing

Yeah, it's really easy :p

Aslo, sleep? What sleep?

emanresu A

Ok, 342445

exedraj

1:44 AM

I'm in

I have best username

Dude coinheringaahing

I'm out :P

exedraj

"Lyxal Code Golf"

emanresu A

We can wait for Aaron, and Dude can goto sleep;.

exedraj

actually i gotta go now too

emanresu A

Oh ok

Bye

Browncat Programs

1:49 AM

@Dudecoinheringaahing You're wrong about why new posts broke :p

It has nothing to do with header changes since the headers are never interacted with

Dude coinheringaahing

Why did it break?

Browncat Programs

They switched a lot of stuff from grid to flex layouts

It was around the same time, and the class names changed for a lot of stuff

emanresu A

O

Dude coinheringaahing

cough cough

emanresu A

Btw, you need to match BBRBRRRRROOOWWWWNNNNNCAAAATTTTTTTT!!!!!!11!1!!111!!! as well now

exedraj

1:52 AM

@emanresuA I back now

emanresu A

OK

Start? I'll think of end

exedraj

Central Park, Cowdenbeath

emanresu A

Ok

End is Q150

Should I start?

exedraj

has aaron joined

emanresu A

Aaron's AFK I think

Browncat Programs

1:55 AM

Aw man, the change to flexbox totally broke the layout for New Posts :/

exedraj

@BrowncatPrograms come wikirace with us

thewikigame.com/group

code = 342445

Aaron Miller

I have returned

Browncat Programs

> No games yet!
Waiting for the Group Admin to create some Challenges...

Bubbler

@BrowncatPrograms It's somewhat like SE dev got annoyed with the HTML changelog request and went "Hey all userscript writers, take this!" :P

emanresu A

@BrowncatPrograms ready?

Browncat Programs

1:57 AM

@emanresuA I guess so

@exedraj Ready?

yep

Ok, go!

Uh...where am I going

exedraj

Q150

Q150 was the sesquicentenary (150th anniversary) of the Separation of Queensland from New South Wales in 1859. Separation established the Colony of Queensland which became the State of Queensland in 1901 as part of the Federation of Australia. Q150 was celebrated in 2009. == Celebrations == The Queensland government and other Queensland organisations celebrated the occasion with many events and publications, including: announcement of the 150 icons of Queensland by the Queensland Premier Anna Bligh placement of a time capsule in the grounds of Old Government House the creation of monuments at...

Browncat Programs

1:58 AM

Oh wait I had to reload

10/10 game design

emanresu A

I didn't even realise JoKing was playing

Transcript for