The Nineteenth Byte - 2016-08-10 (page 1 of 7)

« first day (2020 days earlier) ← previous day next day → last day (2814 days later) »

New Sandboxed Posts

12:00 AM

0

A: Sandbox for Proposed Challenges

Automate the OEIS We see a lot of challenges here asking for a function to create a sequence from the OEIS. While these challenges are fun, as a programmer I see an opportunity for automation. Your challenge is to make a program that takes the index of a sequence (e.g. A172141) and some intege...

CMC: Observe; this details a sequence [(0, 0), (0, 1), (1, 0), (0, 2), (1, 1), (2, 0) ...] Given N, output the Nth coordinate in this sequence.

12:17 AM

tfw you're browsing image processing challenges and you spot this

user image

2

XD

@Dennis Brilliant, but not perfect

@ConorO'Brien What language do you prefer?

@LeakyNun it's language agnostic

and I actually stated it wrong :/

@ConorO'Brien ?

should be [(0, 0), (1, 0), (0, 1)...

12:23 AM

@ConorO'Brien Jelly, 10 bytes: 0rṗ2SÞị@‘U

cool!

@ConorO'Brien Test suite

@ConorO'Brien Pyth, 9 bytes: _@sD^UhQ2

How come Pyth is shorter than Jelly

Eᴀsᴛᴇʀʟʏ Iʀᴋ

cause the original ftw

Jelly isn't based on Pyth

Original? Jelly isn't even remotely related to Pyth.

12:26 AM

ninja'd

he probably means original golfing lang :P

That would be GolfScript.

I know that. it's more sound tho

Pyth and Jelly are probably the two most dissimilar succesful golfing languages

Exactly

12:32 AM

0

A: I've been alphabet hunting for a while

Cheddar, 146 bytes (non-competing) (s,b=65@"90)->(|>27).any(->s has(b=b.slice(1)+b[0])||s has b.lower||(1|>3).any(j->s.lines.turn(j).vfuse has b||s.lines.turn(j).vfuse has b.lower)) Rather slow but it works ¯\_(ツ)_/¯ . added any function after challenge release date. The input does not need t...

@Downgoat Cheddar has diagonals??

@LeakyNun spec didn't say diagonals. lemme check

> Alphabets may not always read forwards. They can also read backwards, up, and down.

@Downgoat oh, what

brb golfing my Jelly answer

diagonal ∉ { forwards, backward, up, down }

^

12:34 AM

Apparently I'm out of my mind

@LeakyNun Ḷṗ2SÞ⁸ịU 8 bytes w/1-based indexing.

@Dennis I don't know if 1-based is allowed

@Downgoat It's even shorter now

o____o

@LeakyNun it is

my hobby: making challenges where Jelly is over 20 bytes :p

@LeakyNun I'm guessing each item in the array is a line?

12:38 AM

@ConorO'Brien Alright, nice

@Downgoat yep

@ConorO'Brien ... @Dennis shall return with less than 20 bytes

yeah, probably

@Downgoat I like that c? syntax. 10/10 very golfy

thx :D

12:55 AM

@Downgoat did you change the function syntax?

@Dennis Is TIO borked?: D:

Stop executing pid=11065: Should wait until other syscalls to be done

@ConorO'Brien no, why?

@Downgoat maybe because of the c?

I thought it was (args) -> body?

oh, wait, that's an empty-arg func. nvm

:D down to only 4.8x longer than Jelly

@Downgoat Seems to be working fine for me.

12:59 AM

hm :/ weird

huh, ok now it's working for me

I had that error earlier

maybe it was too many proecsses?

@ConorO'Brien With which interpreter?

cheddar

@Downgoat When was beta36 released?

1:07 AM

@LeakyNun uh lemme check

@LeakyNun yesterday‌. I can get more accurate timestamp if needed

@Downgoat ok nothing thanks

The process limit is rather strict for sandboxed interpreters, since I have to prevent fork bombs for the "real" languages. I'll set it a bit higher to prevent this in the near future.

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@Downgoat "non-copmeting"

@EᴀsᴛᴇʀʟʏIʀᴋ I fixed that :/

@Dennis oh, so that's why the forkbomb didn't work ;)

1:11 AM

ಠ_ಠ

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@Downgoat kk

@Downgoat Is || short-circuiting?

no :(

workingo n it tho

@Downgoat readline.js line 953

can you stop throwing external error?

what code are you running?

that is error with REPL not me

1:16 AM

@Downgoat I thought REPL is you

I was doing turn(0)

oh yeah, you're not uspposed to do that

You're not supposed to throw external errors

@Dennis How many syscalls are there?

What does pid mean?

Eᴀsᴛᴇʀʟʏ Iʀᴋ

this is trippy

gfycat.com/SphericalHandsomeIndochinesetiger

Apparently, this is a bug in the sandbox. github.com/tsgates/mbox/issues/18

Eᴀsᴛᴇʀʟʏ Iʀᴋ

woah

i.imgur.com/6ws9sZ5.gifv

1:24 AM

@EᴀsᴛᴇʀʟʏIʀᴋ woah wat

@Downgoat floor? floor division?

My BF answer for codegolf.stackexchange.com/questions/89241/… has hit 361 bytes and I'm not even done yet >_>

@ConorO'Brien n->(m->[m[0]-n+m[1],n-m[1]])((a->[a,a*(a+1)/2])(((8*n+1)**.5-1)/2|0))

(Cheddar)

nice

but wow

@StevenH. how you would output in decimal is beyond me

1:32 AM

Outputting in decimal is something I haven't even touched yet; just figuring out which stall is difficult enough.

@LeakyNun floor n

@Downgoat ok thanks but |0 is still shorter

i should start on docs now >_> holidays are ending next week >_>

@Downgoat Did you see my comment?

no not yet

1:38 AM

These alphabet challenges

The worst part about them is that CG doesn't have a built in constant for the alphabet yet

@quartata Problem?

@quartata oh lol

@LeakyNun, once I have the stall I'll just continue looping x/10 on successive digits until it's less than 10 then add 48 to each digit and then print. That's the easy part.

@StevenH. Alright

@StevenH. With brainfuck, you can give the output in unary, unless the challenge says otherwise.

If I'm outputting in unary does the ASCII code of what I'm outputting matter? (is there a meta link for this?)

@Dennis Does it also work that way for BF derivatives (particularly Sesos?)

1:51 AM

Cureent consensus is that you can only use unary if your languahe cannot handle decomal natively.

@StevenH. With Sesos, just use set numout.

@LeakyNun I didn't even notice it, I think that answer should win since its the only one written with love <3

All right, thanks

@miles lol

does MATL use matlab on tryitonline or octave

Octave

I don't even want to know how much a license that allows me to use MATLAB on a web server would cost.

1:57 AM

Guys, what does "discount" mean in "Discount CGI Cloud City"?

I hear that structure quite often in CinemaSins

Eᴀsᴛᴇʀʟʏ Iʀᴋ

cheap and crappy

@EᴀsᴛᴇʀʟʏIʀᴋ I see, thanks

Meaning they couldn't pay for the real thing so they went for a knockoff version that's similar but not the same

thanks I was just wondering

@StevenH. oh, thanks

1:58 AM

Or even if they could pay for the real thing they chose not to

2:08 AM

@LeakyNun in CinemaSins, it's saying that it's a ripoff, or not the original/optimal example of

@ConorO'Brien I see, thanks

El'endia Starman

Games like Rock Paper Scissors are (mildly) interesting because they are non-transitive. If some subset of options were the best, everyone who had a whit of sense would pick from those. In RPS, though, no option is optimal, so it becomes a mind game. CMC: come up with an interesting way to extend this non-transitivity and uniformity to the set of all integers.

That's the longest chat mini challenge I ever seen.

>>> sin(S('1000.1+i*1000.1', rational = True)).evalf()
sin(10001*i/10 + 10001/10)

Um wat?

El'endia Starman

@LeakyNun ¯\_(ツ)_/¯ I wanted to provide a little background... :P

2:19 AM

lambda x:["rock","paper","scissors"][x%3]? Wait, you said interesting...

El'endia Starman

@StevenH. ~~Where did I say anything about mod 3?~~ Oh, yeah, that's a kinda trivial extension.

@El'endiaStarman I don't quite follow. You want to extend RPS to the integers? :/

El'endia Starman

@ConorO'Brien Pretty much, yeah.

@StevenH. That produces ties.

Fixed to make clearer

So does Rock and Rock, Dennis.

2:21 AM

I'm generally confused

:P

I know, but I assume 0 and 3 shouldn't tie here.

El'endia Starman

@ConorO'Brien Well, do you know about Rock Paper Scissors Lizard Spock?

Well, as to how you'd make them not tie therein lies the interesting-ness, right?

@El'endiaStarman vaguely.

@El'endiaStarman How do you measure if 0 wins and loses against "the same" number of options?

2:23 AM

user image

El'endia Starman

@ConorO'Brien It's nothing more than RPS with 5 choices. For any given choice, half of the others beat it and it beats the other half.

And that's RPS 15 ^^.

What could RPS Infinity look like?

< and >?

-1 no Spock

sponge beats devil. seems legit.

But once again > and < don't quite meet the interesting criterion

2:24 AM

@StevenH. I'm not sure if they meet the wrap around criterion

@StevenH. That works as long as the only requirement is that both sets are infinite.

@LeakyNun The CMC didn't specify a wrap-around criterion. The unfortunate fact is that we don't find as interesting a game in > and <, since the best way to play is to determine a larger (or smaller, depending on winning criterion) represent-able number than you think your opponent has come up with.

@StevenH. what does RPS infinity even mean?

I'm more interested in RPS N*i :P

El'endia Starman

@LeakyNun That's basically the challenge. :P

2:31 AM

@El'endiaStarman I don't even know what we are supposed to come up with

What is a generalization of RPS?

^^

RPS infinity would mean that in an infinite set N there's a mapping R between every pair in N^2 such that if R(x,y) is 1 (or win), R(y,x) is -1 (or loss) [cont.]...

El'endia Starman

@LeakyNun The image that Steven oneboxed above is RPS 15.

@El'endiaStarman Do you agree with the definition above?

@StevenH. So basically a well-ordering

El'endia Starman

@LeakyNun I was just thinking that that's actually a pretty good way to summarize the problem.

2:33 AM

@El'endiaStarman So basically a well-ordering

El'endia Starman

It should be non-transitive, though. A well-ordering is transitive.

@El'endiaStarman So if R(x,y) is 1 and R(y,z) is 1, R(x,z) may be 1 or -1

El'endia Starman

@LeakyNun Yes.

@StevenH. Oops, your ordering is transitive

Not sure if anyone is interested, but this Python Q just came through and OP is seeking a code review, with the caveat of their code being a one-liner if at all possible...

-1

Q: need to have 3 expression in one line

I have the code below (I INSIST TO KEEP IT IN THE ONE LINE FORMAT IF POSSIBLE): master_indices = dict(((r[0]), i) for i, r in enumerate(csv.reader(master))) I need to add the statement that if the line contains '/' than enumarate and make dic I have tried: master_indices = dict(((r[0]), i) f...

python beginner python-3.x csv

2:35 AM

> I INSIST TO KEEP IT IN THE ONE LINE FORMAT IF POSSIBLE

I INSIST

Right...

DON'T IGNORE MY WILL

maybe if he asks here, he'd get a better response :P

El'endia Starman

@Phrancis This reads more like an SO question to me.

Not sure why, that seems like a silly requirement

@El'endiaStarman Does the code look like it's actually not working the way they want it to?

2:36 AM

Where R(x,y) == 0 (tie) iff x == y, ∑R(x,y) = 0 ∀y∈N, and R is not transitive. I think that covers everything. Also, yes, unfortunately > and < are transitive.

El'endia Starman

@Phrancis Err, yes? The "I need to add the statement..." bit tipped me off, as do his attempts.

@El'endiaStarman OK we'll just close it then, wasn't sure since my Python is very unimpressive

An example would be lambda x,y:cmp(x%3,y%3) or cmp(x,y)

or, in English: compare them mod 3, if equal, compare them instead

wait, I could build larger cycles indefinitely

so 0<1<2<0, 3<4<5<3, 6<7<8<6

El'endia Starman

Isn't that transitive? Suppose that x = 6, y = 9, and z = 12...

@El'endiaStarman 0<1<2<0 so it isn't completely transitive

El'endia Starman

2:39 AM

Hmm, you're right.

@El'endiaStarman Or are you saying that if f(x,y) = 1 and f(y,z) = 1 then f(x,z) must equal -1?

And then I could make 0<3<6<0, ...

El'endia Starman

@LeakyNun It'd be interesting to see if that's possible.

and then build cycles indefinitely

We could try to make it indefinitely non-transitive by adding a hashing function stronger than %?

I'm about to compare their base 3 representations from right to left

2:43 AM

or maybe f=lambda x,y,z:cmp(x%z,y%z) or f(x,y,z**2) with initial call f(x,y,3)? Not sure how sound of a solution that is

El'endia Starman

A vague idea I have is that it shouldn't be possible to go into an escalating race. For instance, if we say that x beats y if x > y, then you have a race to infinity, essentially. There is no such race in RPS 3 - you have to try and guess what your opponent will pick. So it should, in a sense, always be possible to beat a larger number with a smaller number. In that regard, cmp(x%3, y%3) or cmp(x,y) works.

And by right I mean the least significant digit

El'endia Starman

I feel like modding by n isn't the best solution, though, since it effectively turns it into an RPS n game.

In other words, lambda x,y:0 if x==y else cmp(x%3,y%3) or f(x//3,y//3)

I proudly present reticular, 99% documented!

2:44 AM

That's why I had a variable n rather than a static n @El'endiaStarman

@ConorO'Brien For reasons I can't explain I automatically associated the name to rectum

@LeakyNun ಠ_ಠ

@El'endiaStarman I thought what you want is an RPS n game?

@El'endiaStarman @StevenH. How's my solution?

El'endia Starman

@LeakyNun Perhaps that's the only kind of solution, but maybe there's another, more interesting one.

That seems like it could work, but if we can come up with something else

2:46 AM

Oh, I don't know if you all are interested, but I've been playing around with a database of network-wide SE stats (not SEDE)

user image

@ConorO'Brien -10/10 can you at least provide one example program

@LeakyNun fiiiiine

El'endia Starman

@Phrancis Did you build the database yourself?

@LeakyNun I agree, seems like it could work.

@El'endiaStarman No, it was built by someone else on CR, using the SE API. They gave me access to it

El'endia Starman

@Phrancis Ahhh, cool.

2:47 AM

Do you know of any good bijections over the integers? We could apply one of those before the first recursion of f to make it a little more interesting

@StevenH. Define a good bijection

@El'endiaStarman If you know some SQL and are interested, you can have (read) access too

but I just came up with an interesting one

group them into groups of size 1, 2, 3, 4, 5, ...

and then rotate them one to the left

i.e. (1), (2,3), (4,5,6), (7,8,9,10), ...

@LeakyNun I actually do

and then (1), (3,2), (6,4,5), (10,7,8,9,)...

1,3,2,6,4,5,10,7,8,9,...

@ConorO'Brien where?

2:49 AM

@LeakyNun here

I love how where is just w + here

@ConorO'Brien ugh, in the docs?

3 mins ago, by Conor O'Brien

@LeakyNun fiiiiine

Good bijection, as I had imagined it, would have been defined as meeting the criterion that any two adjacent numbers in the set are no longer adjacent after the bijection is applied. Your suggestion could work, though.

@StevenH. (2,4,1,3), ... would fit your criterion

My code is here @El'endiaStarman in case you are curious how I came up with these numbers

WARNING: Not Golfed^

2:52 AM

Sure. I see nothing wrong with that... should I?

basically, lambda x:x//4+[2,4,1,3][x%4] (roughly)

I know it doesn't work due to 0-indexing

but you get the idea

El'endia Starman

@LeakyNun lambda x:4*(x//4)+[1,3,0,2][x%4] would suffice.

@El'endiaStarman just making the idea here

That could work. Are there any homogeneous bijections? Homogeneous meaning that, if the bijection is called B, for any natural number n there is a number m such that B(m)-B(m+1)>n

@StevenH. example?

2:56 AM

I don't know if there is one, that's why I'm asking

CMC: (real numbers) come up with the most interesting bijection between a finite interval to an infinite interval; you may specify the finite interval and the infinite interval; the infinite interval can be one-sided or double-sided

El'endia Starman

@LeakyNun It would also work to just reverse the groups. 1, 3, 2, 6, 5, 4, 10, 9, 8, 7, ... It seems like it's still subject to the race-for-infinity problem.

@LeakyNun that's... impossible?

El'endia Starman

@ConorO'Brien Not if they're continuous.

@ConorO'Brien tan(x)

@El'endiaStarman what is that problem?

2:57 AM

@El'endiaStarman explain?

@ConorO'Brien tan(x)

El'endia Starman

@LeakyNun The best way to win is to try and guess a larger number than your opponent.

@LeakyNun thanks for the 4 letters and 2 symbols, helps a lot.

El'endia Starman

@ConorO'Brien There are just as many real numbers between 0 and 1 as there are real numbers between 0 and infinity.

@ConorO'Brien tan(x) maps from (-pi/2,pi/2) to (-infinity,infinity)

El'endia Starman

2:58 AM

That's the magic of uncountable infinities...

@El'endiaStarman d'oh I've been working with integer intervals recently. I assumed integer interval >_<

@El'endiaStarman so back to my comparison of base-3 representation from right to left?

@ConorO'Brien "real" is literally the first word in my challenge

@LeakyNun it wasn't before.

also, parens

@ConorO'Brien fair point

With your bijection lambda x:4*(x//4)+[1,3,0,2][x%4], any two adjacent numbers are guaranteed to have a difference of at most 3 (1,3,0,2,5,7,4,6...)

El'endia Starman

2:59 AM

@LeakyNun You edited that in on the second edit. :P

ninja'd

@StevenH. but you wanted no continuous

so it cannot be homogenous

by your definition

If we're sampling over the integers and not real numbers, you should be able to have noncontinuity and homegeousness (spelling? Wording?)

@StevenH. eh, you said you must have integers m such that f(m)-f(m+1)=1?

f(m)-f(m+1) > 1. Big difference, there.

3:03 AM

oh...

ind@@\
  ;p<<
p%~1<;

maps [0, 1] to [0, infinity)

And in English?

>_>

actually, codomain is [0, infinity) \ (0, 1)

roughly, f(x) x = 0 ? x : 1 / x

@ConorO'Brien Alright

@ConorO'Brien In other words: {0} U [1,infty)

yes

3:13 AM

@ConorO'Brien Challenge: make [0,1] to another continuous infinite range

continuous? hmm

1/x?

@ConorO'Brien undefined for x=0

I define it to be infinity

#programmer

infinity doesn't exist

Wouldn't it have been slightly simpler to just map (0,1) to (1, infinity) using 1/x and not tackle the x == 0 ? case?

3:14 AM

min(floor(x), 1) * (1 / x)

By not including it in the domain

@StevenH. I was talking about [0,1]. Of course it would be easy to use (0,1)

@StevenH. this sub-challenge is [0, 1] mapped to an infinite continuous domain

@ConorO'Brien In other words: x or 1

let's try that

3:15 AM

@ConorO'Brien I was saying that your mapping maps every number save 0 to 1

which is not bijective

this one does?

@ConorO'Brien ??

min(floor(x), 1) * (1 / x)

I changed it

El'endia Starman

What about x -> 1/(1-x)?

@ConorO'Brien so it's basically every number save 1 to 0

@El'endiaStarman undefined for x=1

3:17 AM

hm, you're right

El'endia Starman

@LeakyNun x -> limit as y->x of 1/(1-y)

@El'endiaStarman undefined for x=1

El'endia Starman

@LeakyNun No, it's infinity.

@El'endiaStarman infinity is not a number

Good luck mapping from infinity back to 1

log(x+1)/x

3:19 AM

Γ(1/(1-(x-floor(x))))?

works for [-1, 0]

El'endia Starman

@LeakyNun For our purposes it is. It's impossible to answer your challenge otherwise, since some number in [0,1] must map to infinity.

^

@El'endiaStarman Oh, what a fool am I.

let's go back to my main challenge

24 mins ago, by Leaky Nun

CMC: (real numbers) come up with the most interesting bijection between a finite interval to an infinite interval; you may specify the finite interval and the infinite interval; the infinite interval can be one-sided or double-sided

Γ(1/(1-(x-floor(x))))

[0,1]

3:21 AM

@StevenH. domain and codomain?

Codomain is [1, infinity)

@StevenH. I thought 0-floor(0) = 1-floor(1)

That's true. Do we really need to have a double-closed finite interval?

@StevenH. eh... since El'endia said above that it's impossible, it isn't necessary

In that case, either (0,1] or [0,1) can act as the domain.

3:26 AM

x! over [-5, -4]

El'endia Starman

:31588189 You're better than Downgoat and his darned hooves.

@ConorO'Brien interesting

Conor, x! is not defined for x<0. The extension into negative numbers is Γ, which also is not defined for negative integers.

@El'endiaStarman, I don't get the reference, sorry.

El'endia Starman

@StevenH. Downgoat often blames his terrible typing on his hooves.

The extension of the factorial to real numbers would be Pi, not Gamma.

3:29 AM

To read about gamma distribution, see Gamma distribution In mathematics, the gamma function (represented by the capital Greek letter Γ) is an extension of the factorial function, with its argument shifted down by 1, to real and complex numbers. That is, if n is a positive integer: Γ ( n ) = ( n − 1 ) ! {\displaystyle \Gamma (n)=(n-1)!} . The gamma function is defined for all complex numbers except the non-positive integers. For complex numbers with a positive real part, it...

Real and complex, sorry

@StevenH. I think he meant (-5,-4)

@StevenH. not bijective

@ConorO'Brien not bijective

El'endia Starman

Hoo, I'm tired now. Gonna head to bed. G'night all!

Whereas Γ over (0,1) is bijective

Night, @El'endiaStarman

@StevenH. that is true

@LeakyNun Tangent for (-pi/2,pi/2) to real numbers, x->x/(1-|x|) for (-1,1) to real numbers.

3:36 AM

@Dennis nice

Alternatively, B(x,x) where B is the Euler Integral of the first kind over (0,1] to [1, infinity)

@LeakyNun it isn't?

@ConorO'Brien nope

why not?

I thought there was an inverse gamma function and therefore an inverse factorial function?

@StevenH. In other words, gamma(x^2)/gamma(2x)

@ConorO'Brien just look at the graph

3:40 AM

oh

derp

[-5, -4.635] is revised domain

maps to [0.245, Infinity] (basically)

-10/10 very precise

Why not do [-4.6353, -4)?

[-8, -7.5] is less precise

sorry, that may be wrong

bah, why not [-2, 0]

or even [-1, 0]

@ConorO'Brien online interpreter for rectum?

desmos.com/calculator

oh, wait, for reticular?

@LeakyNun for reticular? ಠ_ಠ

3:49 AM

@ConorO'Brien i'm gonna call it rectum from now on

@LeakyNun you're worse than VTC >_<

@ConorO'Brien what did VTC do

@VTC

what did VTC do

a while back I had a language called simplex. he called it limp sex.

3:49 AM

omfg that's funny af

this always seems to happen to me

@ConorO'Brien Online interpreter?

@Dennis could you try adding reticular to TIO? The only instance of eval is evaluating a char in "\#{char}"

and it's in ruby! you don't have to mess with node!

is there any ruby-based language on TIO?

uhhhhh

probably?

3:52 AM

I'm just asking

ik

but idk

« first day (2020 days earlier) ← previous day next day → last day (2814 days later) »

Transcript for

Aug '1610

The Nineteenth Byte

The Nineteenth Byte: General discussion for codegolf.stackexc...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile