The Nineteenth Byte

Doorknob

10:00 PM

oh, it's gotta be something like that, because 0s are powers of two plus one

Adnan

@Doorknob Abacaba sequence?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

I want to do a challenge for golfing PE, but I can't do it on this site because not everyone that wants to do PE hasn't done it yet.

Calvin's Hobbies

Full image for posterity:

10

Doorknob

IT IS DONE.

orlp

LLLETS GET PROMOTED

inb4 "congrats, beta has ended, we conclude codegolf is not valuable and will be closed"

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:02 PM

Mini-challenge: golf the 1000 digit number found here

Doorknob

 1 213 3251634 439265
 1 121 1213121 121312

runs of ones on top, runs of zeroes on bottom

replaced 0/0 with spaces for visual separation

quartata

Graduation shines through the barrier. You feel as though beta is almost at an end. You're filled with DETERMINATION.

Optimizer

so there was no .0 to begin with

i was correct all this time

Doorknob

haha

quartata

er yes sorry

Optimizer

10:04 PM

fuck you all

:P

RikerW

lol

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Electrons bouncing in a wire Imagine a "wire" that has n spaces. Imagine further that there are "electrons" in that wire. These electrons only live for one unit of time. Any spaces in the wire that are adjacent to exactly one electron become an electron. In Game of Life terminology, this is B1/S...

RikerW

Sorry for your troubles.

Doorknob

interesting

RikerW

I like that.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:06 PM

Quick question: given a string with only digits, what is the best way to compress it?

Doorknob

I wonder what causes the pattern of <some number of> 1s and then a number of 0s following the abacaba pattern

RikerW

@CᴏɴᴏʀO'Bʀɪᴇɴ itself

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@RikerW ...

RikerW

@CᴏɴᴏʀO'Bʀɪᴇɴ relavent

Geobits

@CᴏɴᴏʀO'Bʀɪᴇɴ Pump the base.

RikerW

10:07 PM

^

Doorknob

@CᴏɴᴏʀO'Bʀɪᴇɴ if you're 100% sure that there are no patterns, then encoding it in binary and taking bytes of 8 bits is objectively "the best" way to compress it

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@Geobits huhwat

@Doorknob It's an arbitrary number of a huge length.

I've found a pattern

b64 makes it longer.

^

10:08 PM

@orlp yeah?

Doorknob

@orlp for number of ones?

orlp

look at the number of ones from here

El'endia Starman

23 16

orlp

2130325

El'endia Starman

Haven't figured out 22 yet, gimme a moment.

Doorknob

10:09 PM

 1 213 3251634 4392651
 1 121 1213121 1213124

wait...

orlp

bin(2+3) == 0b101 -> 1*0 + 0*1 + 1*2 = 2

Doorknob

the 4 shouldn't come this early

oh, didn't do 22 yet :P

El'endia Starman

Like I said, I haven't got 22 yet.

Doorknob

I can't read

orlp

bin(1+3) == 0b010 -> 1*0 + 1*1 + 0*2 = 1

Doorknob

10:10 PM

 1 213 3251634 439265?1???????
 1 121 1213121 121312141213121

(filled in with sane predictions)

@orlp where are you getting 2+3, 1+3 from?

I think I get the rest

orlp

it breaks down a bit later

>>> def wtf(n):
...  b = bin(n)[2:]
...  return sum(i*int(bit) for i, bit in enumerate(b))
...
>>> wtf(5)
2
>>> wtf(6)
1
>>> wtf(7)
3
>>> wtf(8)
0
>>> wtf(9)
3
>>> wtf(10)
2
>>> wtf(11)
5
>>> wtf(12)
1
>>> wtf(13)
4

but it's spooky

Doorknob

> def wtf

orlp

>>> [wtf(n) for n in range(5, 20)]
[2, 1, 3, 0, 3, 2, 5, 1, 4, 3, 6, 0, 4, 3, 7]

RikerW

function wtf()
print("ಠ_ಠ")
end

Cᴏɴᴏʀ O'Bʀɪᴇɴ

wtf is this

orlp

10:14 PM

it seems that later on it still generates the right numbers, but in a bit different order

Doorknob

[2, 1, 3, 0, 3, 2, 5, 1, 4, 3, 6, 0, 4, 3, 7]
                         6     4           9

with corrections

orlp

>>> [wtf(n) for n in range(5, 30)]
[2, 1, 3, 0, 3, 2, 5, 1, 4, 3, 6, 0, 4, 3, 7, 2, 6, 5, 9, 1, 5, 4, 8, 3, 7]

El'endia Starman

I think I'm gonna have to write a program for n=22.

Doorknob

[2, 1, 3, 0, 3, 2, 5, 1, 4, 3, 6, 0, 4, 3, 7, 2, 6, 5, 9, 1, 5, 4, 8, 3, 7]
                         6     4           9           ?     ?????????????

neat

RikerW

why?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:15 PM

@El'endiaStarman you've been doing it by hand???

RikerW

^

Doorknob

oh oh oh

I think n=22 is 0b11111110

7 and 9 swapped, like orlp said...

El'endia Starman

@CᴏɴᴏʀO'Bʀɪᴇɴ Not exactly. I've been using Variations of Life, but it doesn't have a built-in away to get the period.

@Doorknob And that is...2046?

Doorknob

@El'endiaStarman how'd ya get 1022? O_o

orlp

@Doorknob it seems too good to be unrelated

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:16 PM

@El'endiaStarman There should be.

Doorknob

@orlp yeah

RikerW

Doorknob

definitely somehow related

El'endia Starman

@Doorknob Watched it for a while. :P

RikerW

I used to get FCK a lot here... :P

Doorknob

10:17 PM

@El'endiaStarman O_O

^

@Doorknob holy shit

Dude, mechanize.

@El'endiaStarman You, sir, are brilliantly insane.

orlp

@Doorknob notice the gap between swapped pairs

@Doorknob 1, 3, ...

Doorknob

10:17 PM

yeah I saw that

El'endia Starman

Alright, testing the 2046 hypothesis.

I do have a "fast" mode... :P

Cᴏɴᴏʀ O'Bʀɪᴇɴ

*2044

orlp

I'll write a program for the wire

El'endia Starman

Oh, it's not 2046. It's 4094.

RikerW

marky is smart and doesn't like code-trolling

:P

Doorknob

10:19 PM

1s:    1 213 3251634 43926571???????
0s:    1 121 1213121 121312141213121
orlp:            * *   *   *

Calvin's Hobbies

In case anyone missed it, Geobit's Marky bot is live over here: chat.stackexchange.com/rooms/30332/beep-boop-maggot

Doorknob

too early to see a discernable pattern yet, but...

both pairs are isolated on one side of the "mirror" number in the abacaba thing

and they're all on places where there is 1 0

RikerW

marky has turned his back on us!!!!!

HE LIKES CR BETTER

Calvin's Hobbies

Bad Upbringing

1 min ago, 13 seconds total – 4 messages, 2 users, 0 stars

Bookmarked 50 secs ago by Calvin's Hobbies

6

RikerW

@Quill DON"T LET HIM GO TO YOUR SIDE PLS

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:22 PM

@QUILL HE MEANS SO MUCH TO US :(

Doorknob

orlp:  1 213 3251*3* 43*265*154837??
1s:    1 213 3251634 43926571???????
0s:    1 121 1213121 121312141213121

note that the 7 in the 1s section is a hypothesis

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Is anyone here a JavaScript guru? Requirement: must know more than me.

RikerW

@Doorknob you don't even care. :P

@CᴏɴᴏʀO'Bʀɪᴇɴ ish

shoot

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Is that a no? :|

RikerW

I know a good bit, but not as much as I know python.

Shoot.

Calvin's Hobbies

10:24 PM

What's the question?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Oh, I see.

RikerW

^^

Cᴏɴᴏʀ O'Bʀɪᴇɴ

WHAT DOES THIS MEAN?!?!?!?: uncaught exception: out of memory <unknown>

IT'S DRIVING ME ABSOLUTELY CRAZY

Doorknob

it means you're out of memory

RikerW

^

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:24 PM

HOW TO BLOODY FIX IT!!?!?!>!??!??!?!?!?

RikerW

Even a doorknob knows that.

Doorknob

@CᴏɴᴏʀO'Bʀɪᴇɴ stop using so much memory

RikerW

@CᴏɴᴏʀO'Bʀɪᴇɴ stop infinite loop/recursion

^^

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@RikerW Infinite loop/recursion kills the browser if it uses too much memory.

Doorknob

@El'endiaStarman updates on the 254-hypothesis?

El'endia Starman

10:25 PM

@Doorknob For n=24?

RikerW

@CᴏɴᴏʀO'Bʀɪᴇɴ okay

Doorknob

@El'endiaStarman yeah

RikerW

plz-send-teh-codez @CᴏɴᴏʀO'Bʀɪᴇɴ.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@RikerW :|

@RikerW uhhh I don't think you want me to do that

RikerW

yes i do

Doorknob

10:26 PM

@El'endiaStarman wait, no, n=22

El'endia Starman

@Doorknob n=22 was 4094.

RikerW

what is it supposed to do?

Doorknob

@El'endiaStarman wait WHAT

Cᴏɴᴏʀ O'Bʀɪᴇɴ

It's the code for Jolf.

RikerW

Okay.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:26 PM

All of it.

RikerW

o

Doorknob

so much for @orlp's pattern

RikerW

@CᴏɴᴏʀO'Bʀɪᴇɴ section it

pinpoint the error.

comment bits out until you find it.

Doorknob

                 # #   #   #
orlp:  1 213 3251436 4372659154837??
1s:    1 213 3251634 439265C1???????
0s:    1 121 1213121 121312141213121

RikerW

or give up and play nethack

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:27 PM

Jolf is over 10KB in size.

RikerW

10kb? o_O

I have had single files that are over 1MB...

gfi

Cᴏɴᴏʀ O'Bʀɪᴇɴ

huhwat

Oh, I found the problem.

RikerW

what was it?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

I had too many tabs open. >_<

RikerW

wtf?

Adnan

10:29 PM

Hahaha

RikerW

@CᴏɴᴏʀO'Bʀɪᴇɴ ಠ_ಠ

Doorknob

@El'endiaStarman got anything >=24 yet?

El'endia Starman

Still working on n=24.

Doorknob

ok

El'endia Starman

Heh, I would've expected orlp to have that program written by now. :P

Cᴏɴᴏʀ O'Bʀɪᴇɴ

10:31 PM

@RikerW I had three chats open, four githubs, two instances of MDN, two instances of Jolf, fourteen codegolf.se tabs, and project euler.

El'endia Starman

24 2046

RikerW

@CᴏɴᴏʀO'Bʀɪᴇɴ ಠ_ಠ

Doorknob

10 zeroes, 1 one

orlp's pattern is definitely breaking down...

El'endia Starman

muahaha

Cᴏɴᴏʀ O'Bʀɪᴇɴ

*mwuhaha

El'endia Starman

10:33 PM

25 252

Doorknob

     0         1         2         3
     0123456789012345678901234567890
                 # #   #   # ###
orlp:  1 213 3251436 4372659154837??
1s:    1 213 3251634 439265C1A69????
0s:    1 121 1213121 121312141213121

next one's a power of two minus two, I already know :P

El'endia Starman

26 1022

:P

Doorknob

orlp's thing has completely stopped making sense

orlp

btw, how does one have a period of 0?

El'endia Starman

The electrons die out.

27 56

orlp

10:37 PM

That's still a 1-cycle period

El'endia Starman

If it helps, you can think of it that way.

orlp

if it alternates a, b, a, b

it's a 2 cycle period

El'endia Starman

Lemme see if that's on OEIS.

Doorknob

   0         1         2         3         4         5
   012345678901234567890123456789012345678901234567890
1s:  1 213 3251634 439265C1A693
0s:  1 121 1213121 121312141213121 1213121412131215121

El'endia Starman

nope

RikerW

10:40 PM

Anybody here use something other than chrome?

Doorknob

I use Chromium .-.

and used to use Firefox

orlp

46 takes forever

Doorknob

oh awesome

El'endia Starman

Oh gosh. I was working my way up through 2^n-2 for 28.

Glad I didn't have to go all the way. :P

orlp

@El'endiaStarman try 36

:)

El'endia Starman

10:44 PM

hahaha, yeah...

Mind if I take your list for the test cases in my sandbox post?

orlp

that list is my copyright!

eternally mine!

:(

(go ahead)

:)

(for the small fee of one virtual cookie)

El'endia Starman

10:46 PM

Should I include the case n=0?

Doorknob

woah woah woah

RikerW

what was it?

Doorknob

n=36 blows the pattern out of the water

101010101010101010

El'endia Starman

O_O

RikerW

cool

El'endia Starman

10:47 PM

WAT.

orlp

Oh, my n = 0 is faulty

or well, computational method

Doorknob

check your n=36...

RikerW

Anybody know how to get sympy in julia?

orlp

for simplicity I would just ignore n = 0

El'endia Starman

yeah, that's what I was thinking too

Doorknob

10:51 PM

             1         2         3         4         5
    123456789 123456789 123456789 123456789 123456789
1s:  1 213 3251634 439265C1A693E45 54C3?9C2A675CB
0s:  1 121 1213121 121312141213121 1213?21412131215121

RikerW

@Downgoat

El'endia Starman

Hey wait, I don't get 48 for n=39.

Wait, yes I do.

It's one of those lengths that doesn't repeat its starting configuration.

Doorknob

n=36 is really a mystery

RikerW

^

Doorknob

             1
             10
             1
            110
             100
           1110
             1
           1110
            1100
         111110
             1000
        1111110
           11100
          11110
             1
          11110
           11100
     1111111110
            11000
        1111110
         1111100
   111111111110
             10000
    11111111110
        11111100
     1111111110
           111000
111111111111110
          111100
         111110
             1

orlp

10:56 PM

# Brent's cycle detection, modified from wikipedia.
def brent(f, x0):
    power = lam = 1
    tortoise = x0
    hare = f(x0)
    while tortoise != hare:
        if power == lam:
            tortoise = hare
            power *= 2
            lam = 0
        hare = f(hare)
        lam += 1
    return lam

def electron_period(n):
    get = lambda w, i: w[i] if 0 <= i < len(w) else 0
    next_wire = lambda w: [get(w, i-1) ^ get(w, i+1) for i in range(len(w))]
    return brent(next_wire, [1] + [0]*(n-1))

El'endia Starman

@Doorknob Fascinating.

orlp

err

@El'endiaStarman can you delete that version above?

I forgot to rename a variable

El'endia Starman

I can edit it.

orlp

gist.githubusercontent.com/orlp/6cf78554b4386f45ca24/raw/…

Bad Upbringing

Transcript for