The Nineteenth Byte

PhiNotPi

4:00 AM

I don't think it counts as a "video" if it's just a still image with music.

Alex A.

@quartata What does Toxic Garbage Island have to do with anything?

quartata

@PhiNotPi I didn't even click on it

Anonymous

quartata

@Mego gaben pls

Anonymous

Oh wow the quality on that came out terrible

Helka Homba

4:01 AM

Is it a rotating clown leg?

Anonymous

It's the Steam logo

quartata

No, that's the steam logo

Anonymous

Hence steamrolled

Alex A.

Gojira - Toxic Garbage Island (Live at Vieilles Charrues Festival 2010)

►

Helka Homba

@quartata You say potato

Alex A.

4:01 AM

Is that better?

@El'endiaStarman Trust me, the advice I'd give has nothing to do with programming. :P

El'endia Starman

@AlexA. Hahaha, well, would it be good advice for someone dating a programmer?

Anonymous

Another good piece of advice is "don't complain unless you want it fixed"

Helka Homba

@El'endiaStarman Programmers are normal people too.

Alex A.

^ I was going to say

Anonymous

10 mins ago, by Mego

The advice I'd give is "expect me to take everything too literally, but it's not always because I'm joking around"

Anonymous

4:07 AM

@HelkaHomba citation-needed

Alex A.

I can't think of any advice that wouldn't apply to folks of other professions

Though I guess "don't be surprised if your significant other spends a lot of time on the computer"

Sherlock9

Welp, Google Code Jam has me stuck already

El'endia Starman

@HelkaHomba Of course, but there's plenty of generic relationship advice out there. My fiancée was more interested in the programmer-specific side of things.

Sherlock9

Might not help that I'm programming in class when I'm supposed to be learning about databases

Anonymous

The thing is, it's not specifically programmers. Rather, it's people who tend to think in similar ways - taking things literally, having a problem-solving mindset, etc. It just so happens that programming is a field where those traits are very beneficial.

El'endia Starman

4:09 AM

@Mego That's true.

Maltysen

@QPaysTaxes a wall you want broken?

Helka Homba

@El'endiaStarman My advice: Take them outside once in a while.

Anonymous

You can see the same traits in many mathematicians, analysts, and engineers, too

El'endia Starman

@HelkaHomba Ha ha! Yep... >_>

Maltysen

@HelkaHomba why would you ever tell people to do that to programmers, you inhumane monster?!

Anonymous

4:10 AM

@QPaysTaxes Banging your head on the wall to get rid of a bug is not an optimal solution. Have you tried a shoe, or rolled-up newspaper?

Anonymous

@Maltysen Because the best way to solve a tricky problem is to step away from it for a while.

Helka Homba

@Mego Speaking of which, do you know how to disable the "update news" popup when you open Steam?

Anonymous

@HelkaHomba Don't open Steam?

Maltysen

@Mego yeah, but you can just take a shower

Anonymous

@QPaysTaxes No that's alcohol.

Nathan Merrill

4:12 AM

oooh, I found an improvement on google code jam

Anonymous

For example, I'm stuck on problem C in GCJ. So I'm playing Learn To Fly 3 while it marinates in my subconscious.

Nathan Merrill

say you find two numbers when n = 5

Anonymous

Which problem? C?

Nathan Merrill

and you find a similar number on n = 6

yes

Maltysen

codeJam is making me need a lot of sugar. I've had like a mountain dew, a pack of skittles, and now I have a bowl of ice cream. I should stop

Nathan Merrill

4:13 AM

you can use the difference to find another number on n = 6

Sherlock9

@Mego I have no idea why my submission is wrong

Anonymous

Sugar is not good for concentration and thinking

El'endia Starman

.............I'm seriously considering getting into Code Jam now, but...man, that'd take quite some time.

Anonymous

My C-small submission was right, but I did it with brute force

Sherlock9

Well, it's only today, so

Anonymous

4:13 AM

I'm still trying to come up with a more efficient way to do the large

quartata

@El'endiaStarman I might do next year's.

Sherlock9

My A-small submission is annoying me

quartata

I honestly never heard of it until recently...

El'endia Starman

@quartata I was thinking that too. I want to focus on Pytek for now.

Anonymous

I've done Code Jam every year since 2012

Anonymous

4:15 AM

A tip for problem A: 0 is the only input where all digits will never occur

Mars Ultor

did everything with brute force

Anonymous

A is brute-forceable by virtue of how simple it is

Maltysen

@Mego i did brute force it to verify that 0 was the only one

pypy is magic

Doorknob

I used a HashSet and everything for A thinking that the large test case would be at least somewhat performance-critical, but nope

Maltysen

took like 3 secs for 0-1e6

Anonymous

4:16 AM

I used sets too, but that was because I was too lazy to do my own duplicate checking

Maltysen

@Doorknob yo same!

Mars Ultor

wait

B didn't need brute force

but using brute force for C

Doorknob

@AlexA. Remember when you asked about languages that you really like but there's one thing about them that annoy you?

#![feature(step_by)]

Maltysen

@MarsUltor B-small does

Doorknob

^ this is that for Rust

Maltysen

4:17 AM

dijkastra

Mars Ultor

@Maltysen ?

PhiNotPi

@KennyLau Are no-ops allowed in cracking attempts (even though they are banned for cops?)

Alex A.

@Doorknob Oh yeah, I liked that question; that got some interesting responses. Yeah, Rust's feature importing thing bugs me and I've only used Rust a couple times.

Maltysen

@MarsUltor if you write it as an unweighted directed graph it only has 1024 nodes, easy with dijkastra

now I need to write dijkastra…

Sherlock9

Ah, Dijkstra. I should learn that some time

Mars Ultor

4:21 AM

@Maltysen No, you can only lift from top of stack; you can't flip the middle. B doesn't need brute force

Maltysen

@MarsUltor right, you write each possible flip operation as an edge

Kenny Lau

@PhiNotPi added

Maltysen

yes you don't need brute-force, its not even good enough for large, but i'm lazy

Anonymous

You could make it into a tree by just not allowing the same flip you just did

Mars Ultor

@Maltysen It's way easier without djikstra

Anonymous

4:23 AM

Wait actually no

Kenny Lau

@Maltysen replace with capture group in pyth?

Anonymous

You can still have cycles that way

Mars Ultor

just did some conditional counting for B

wait a second

damnit

Kenny Lau

How to replace with capture groups in Pyth?

Maltysen

@KennyLau its in the python docs

Pyth uses python's re

Kenny Lau

4:26 AM

@Maltysen How do you do so in python?

Alex A.

2 mins ago, by Maltysen

@KennyLau its in the python docs

El'endia Starman

@quartata At least I can still look at the problems... :D

Kenny Lau

@AlexA. finding

docs.python.org/2/howto/regex.html

Sherlock9

I didn't realize that the inputs changed with each attempt

Whoops

El'endia Starman

Hey wait haven't we had something like B as a challenge on PPCG?

Sherlock9

4:27 AM

Well, A-small is done

Anonymous

@El'endiaStarman Something similar, yeah

Anonymous

20

Q: Flipping pancakes

In pancake sorting the only allowed operation is to reverse the elements of some prefix of the sequence. Or, think of a stack of pancakes: We insert a spatula somewhere in the stack and flip all the pancakes above the spatula. For example, the sequence 6 5 4 1 2 3 can be sorted by first flippin...

El'endia Starman

@Mego I remember attempting it with Pancake Stack, the programming language.

quartata

(define chunkNum (car (if (zero? (modulo room 2)) ((λ () (let ([y (list-tail side1 (- (length side1) 1))])
   (set! side1 (take side1 (- (length side1) 1)))
    y
)))((λ () (let ([y (list-tail side2 (- (length side2) 1))])
  (set! side2 (take side2 (- (length side2) 1)))
   y
))))))

Oh my god why does this work what have I done

El'endia Starman

You have transcended into parentheses hell. :P

Anonymous

4:30 AM

@El'endiaStarman I remember working on that with you :)

Maltysen

@Mego not optimal though, otherwise I would have submitted the golfscript answer :/

Anonymous

I think our final conclusion was that Pancake Stack can't sort

Anonymous

@Maltysen It's a completely different puzzle

El'endia Starman

Yeah, and Pancake Stack is Turing-complete only as a technicality.

Anonymous

Pancake sorting != pancake flipping happy making

Maltysen

4:31 AM

oh sorting

Chris Jester-Young

@quartata In most cases, excessive use of set! means you're Doing It Wrong.

Anonymous

Pancake Stack is not TC

quartata

@ChrisJester-Young I only used set! once trying to avoid it

Chris Jester-Young

Also, your code looks like it has way too many lambdas. :-)

quartata

I have two lists side1 and side2 and all I want to do is pop (remove) an element from side1 if room % 2 == 0 and from side2 otherwise

The empty lambda is so I can have multiple lines, the let is so I can save the list-tail

Chris Jester-Young

4:32 AM

Right, normally you'd use a named let to do that. Lemme see if I can demonstrate.

quartata

Wait a second

Why do I need the lambda

Anonymous

Pancake Stack is TC except for a certain subset of problems that 2CMs can't solve

Anonymous

Because Pancake Stack is essentially equivalent to a 2CM

El'endia Starman

Yeah. Hence my "technicality" remark.

It's not TC in a useful way.

quartata

Oh nevermind the lambda is completely unnecessary

Anonymous

4:33 AM

2CMs are TC iff you can initialize the counters to any value

Kenny Lau

@Maltysen Still can't get it to work

Tried "\1" and "\\1"

quartata

(define chunkNum (car (if (zero? (modulo room 2)) (let ([y (list-tail side1 (- (length side1) 1))])
  (set! side1 (take side1 (- (length side1) 1)))
  y
 )(let ([y (list-tail side2 (- (length side2) 1))])
  (set! side2 (take side2 (- (length side2) 1)))
  y
 ))))

Only marginally better of course...

Anonymous

But because Pancake Stack is equivalent to a 2CM with counters initialized to 0, the caveat makes it not TC

Anonymous

@KennyLau Use r'\1'

Kenny Lau

@Mego Can't find r in Pyth

Chris Jester-Young

4:34 AM

@quartata Can you describe in pseudocode what your code is trying to do?

I'll try to write a version in idiomatic Scheme. :-)

if for no other reason than so you have a pattern to work with.

quartata

@ChrisJester-Young int chunkNum = (room % 2 == 0 ? pop(side1) : pop(side2));

Anonymous

@KennyLau You need a raw string, I'm not sure how to do that in Pyth

quartata

literally the line I'm trying to translate

Kenny Lau

@isaacg Feature request lol

Chris Jester-Young

@quartata Can you show me the rest of the function, with the context of side1, side2, room, etc.?

Maltysen

4:35 AM

try "\\\\1"

that might work

Anonymous

@QPaysTaxes After new submissions are no longer competitive (when a winner is chosen)

Anonymous

So like for Kenny's CnR, after April

Chris Jester-Young

@Mego That's a thing? I make a point to update my accepted answer if a shorter answer is posted, no matter how far into the future.

quartata

(let ([side1 (shuffle (filter (λ(x) (= (bitwise-and x 2) 0)) (range 0 16)))]
      [side2 (shuffle (filter (λ(x) (= (bitwise-and x 2) 2)) (range 0 16)))])

Anonymous

It's customary for cops to reveal their solutions after the competition, usually with spoiler tags so that people can still try to solve it

quartata

4:38 AM

side1 and side2 are both lists.

(for ([room (range 0 (random 6 11))])

Anonymous

@ChrisJester-Young A lot of people use cutoff dates for CnRs and KoTHs

quartata

^^ that's the for loop my code is in

Chris Jester-Young

@Mego *nods*

quartata

I'd paste the whole code but I don't want to wall of text the 19th Byte

Anonymous

Use pastebin

Chris Jester-Young

4:38 AM

@quartata Put it in a Gist. :-)

Anonymous

Or a gist

Anonymous

Ninjo'd

quartata

ok

pastebin.com/tQ1LYPJf

Anonymous

My cat is being obnoxious and bathing in my lap while I'm trying to program :(

Anonymous

But she's cute so I can't be upset

quartata

4:40 AM

Currently all I have actually lol

Anonymous

Less cute when she's mopping Australia with her tongue, but still cute

Chris Jester-Young

@quartata What's the code supposed to do at a high level? I'm trying to avoid an XY problem. :-)

quartata

@ChrisJester-Young It's a random room generator for a little ASCII art dungeon maker.

So basically

Sp3000

@Mego "Please do not discuss the problems during the round."

quartata

You have the 80x24 screen split up into 19x4 chunks with 1x4 and 19x1 pockets of dead space between them

side1 represents chunks on the left side of the screen

side2 on the right

Maltysen

4:42 AM

@Sp3000 you're allowed to for qual

Sp3000

Where does it say that?

Maltysen

FAQ IIRC

or maybe terms and conditions

quartata

The loop alternates picking chunks from the left and the right side, picks a random point in the chunk and draws a rectangle between it and the farthest corner of the chunk

Sp3000

"Collaborating with anyone else during any round of the contest, with the exception of the qualification round, is strictly prohibited and will result in your disqualification."

I see, well this is just confusing

quartata

But all I have is the "picking chunks from the left and the right side" currently

Sp3000

4:43 AM

Cos their twitter says "The #CodeJam2016 Qual Round has officially started! Conversation about the problems during the round is prohibited."

o__O

quartata

That's what that monster if is supposed to be

Anonymous

Probably auto-generated

Anonymous

Their FAQ has more legal basis than a tweet

Sp3000

The first quote with the ping was from the main page though

But oh well

Chris Jester-Young

@quartata I see. I'll just translate your pastebin contents into how I'd write it (without changing the algorithm).

PhiNotPi

4:47 AM

@QPaysTaxes It uses Perl's "auto-typecasting." Meaning that it's converted to a number if used as a number.

quartata

Numbers are (usually) actually stored as strings in Perl so

We usually refer to "numbers" and "strings that look like numbers" as numerals.

Why?

I'd think that would help your solution...

This is a code golf, right?

PhiNotPi

Cops-and-Robbers

quartata

Oh this is that new CnR

PhiNotPi

@QPaysTaxes What happens when you draw from an empty stack depends, IIRC, on the operator.

Chris Jester-Young

@quartata What's the [room (range 0 (random 6 11))] about?

quartata

4:52 AM

@ChrisJester-Young Oh, that's how many rooms will be generated.

PhiNotPi

I think, popping from an empty stack gives an empty scalar (empty string).

quartata

6-10.

So I'm iterating through 0 - random number

Chris Jester-Young

@quartata Okay.

PhiNotPi

@QPaysTaxes {} uses the control stack, and in that case an empty control stack is a falsey value

so, {} will be a no-op

@QPaysTaxes It's not a no-op in my original solution, but if you find a solution that doesn't use it, then that is also acceptable.

"No-ops can be used in robber's solution." codegolf.stackexchange.com/q/77420/2867

Maltysen

@Mego have answer for smallB crossing my fingers, trying it out

Chris Jester-Young

4:58 AM

(let-values ([(side1 side2) (partition (lambda (x)
                                         (zero? (bitwise-and x 2)))
                                       (range 16))]
             [(rooms) (random 6 11)])
  (let loop ([result '()]
             [side1 (shuffle side1)]
             [side2 (shuffle side2)]
             [room 0])
    (cond [(>= room rooms) result]
          [(even? room) (loop (cons (car side1) result) (cdr side1) side2 (add1 room))]
          [else (loop (cons (car side2) result) side1 (cdr side2) (add1 room))])))

Anonymous

I almost have an answer for smallB

Chris Jester-Young

@quartata ^ That's probably one of the more idiomatic ways to write your code in Racket. :-)

Sp3000

runs away from TNB in case of Code Jam spoilers

Mars Ultor

@Mego smallB is easy, but only if you do it the naïve way - counting the number of contiguous sections

Or maybe I just got insanely lucky that it worked

PhiNotPi

@QPaysTaxes : always takes two arguments (the top 2 things on the main stack). The topmost (most recent) argument is the number of copies, the other is the thing to copy. Performing : when there's not 2 things on the main stack is a no-op.

Maltysen

5:00 AM

@Mego \o/ 35 points (well assuming I got A-large correct)

Sherlock9

Is sympy allowed in GCJ submissions?

PhiNotPi

@QPaysTaxes Yes.

Maltysen

@Sherlock9 prolly

Sherlock9

Because brute-forcing C is driving me nuts

Thank God

Maltysen

I used networkX

so kinda cheating

but I would have copy pasted from python cookbook anyway so ¯\_(ツ)_/¯

Anonymous

5:02 AM

How would sympy help with C?

Mars Ultor

@Mego faster prime factors?

maybe faster base conversion

Anonymous

Base conversion is a builtin in Python

Mars Ultor

wait, what?

as in number to number in binary?

Dennis

That's bin.

Mars Ultor

well, C needs to convert to bases 2-10

Maltysen

5:05 AM

@MarsUltor for c you need the opposite

that's int(n, base)

Mars Ultor

@Maltysen No, you don't need it in binary at all

just set it to minimum and keep on incrementing

Chris Jester-Young

(let-values ([(side1 side2) (partition (lambda (x)
                                         (zero? (bitwise-and x 2)))
                                       (shuffle (range 16)))]
             [(rooms) (random 6 11)])
  (let loop ([result '()]
             [side1 side1]
             [side2 side2]
             [room 0])
    (cond [(>= room rooms) result]
          [(even? room) (loop (cons (car side1) result) (cdr side1) side2 (add1 room))]
          [else (loop (cons (car side2) result) side1 (cdr side2) (add1 room))])))

@quartata ^ Slight improvement, with only one shuffle.

Maltysen

@MarsUltor no, converting from bases 2-36 is int

Mars Ultor

@Maltysen ?

yeah, but don't we need to convert to, not from?

Maltysen

@MarsUltor no you interpret it as bases 2-10

Mars Ultor

5:08 AM

oh wait

realy needs to read the question better

Maltysen

you don't even need to bruteforce C-small in advance

its really fast

Anonymous

int(str, base) interprets str as a base-base number

Anonymous

That's all you need for C, other than factoring

Maltysen

and itertools.product

PhiNotPi

@QPaysTaxes I don't think there's a case where it does nothing. It gives 0 on an empty stack.

@QPaysTaxes It does both a number->character and a character->number conversion, with the char->num result on top. One of those two results is typically useless/meaningless.

Maltysen

5:15 AM

@Mego what's your nickname?

PhiNotPi

@QPaysTaxes It's kinda weird, I hacked together two functions into a single operator (ran out of ASCII).

112, leaves the stack with 49 on top (the character code of 1), and p underneath (with p being 112).

Yes

Mars Ultor

hmm

are we allowed to share outputs for the qualifying rounds?

PhiNotPi

@QPaysTaxes That's actually a pretty smart solution.

Double check my original?

Anonymous

@Maltysen On what?

PhiNotPi

Yeah, it works (that's what I used to write it, too)

Maltysen

5:24 AM

@Mego on gcj

so i can add u as a friend

PhiNotPi

You're not going crazy though... there's definitely some weird behavior with the language that is important to my solution.

There's actually not that much to the source code: github.com/PhiNotPi/Element/blob/master/InterpreterTIO.plx

Nope, not a bug in Perl.

But, if you look over my list of operators, I'm sure you'll find something useful.

Anonymous

@Maltysen TheOnlyMego

Maltysen

added

Anonymous

Coolio

PhiNotPi

It means something along the lines of "last index of", but whatever you're looking at is completely irrelevant

Chris Jester-Young

5:29 AM

(require srfi/26)
(take (call-with-values (cut partition (compose zero? (cut bitwise-and <> 2))
                                       (shuffle (range 16)))
                        (cut append-map list <...>))
      (random 6 11))

^ a much golfier version of @quartata's code, if we compress all the spaces out and make identifiers single-character.

PhiNotPi

@QPaysTaxes I can tell you which entry to look at, if you're desperate.

(I'm trying to balance not giving out hints to make it difficult, with giving enough so that you don't give up)

Maltysen

> you may exclude it as long as you put a comment in your source code explaining where the library is available.

shit

do you think they check everyone's source code for the qualifying round?

Anonymous

Probably not

Mars Ultor

um

xsot

Sigh I discovered a mistake in my reasoning for D large

Mars Ultor

5:37 AM

D: i have no idea why i'm getting C small wrong

Anonymous

If you forget to list a third-party library, they can usually figure it out

Maltysen

@Mego ok then, as long as they aren't too strict, networkx is pretty well known

Anonymous

It's literally the top result on Google for "networkx"

Anonymous

If they can't use their own search engine, there are bigger issues here

PhiNotPi

@QPaysTaxes ¯\(°_o)/¯

Anonymous

5:38 AM

@QPaysTaxes That's not very nice

Anonymous

If you failed to expect something, maybe the problem lies within

Anonymous

Most unexpected behavior can be explained with the presence of alcohol

Anonymous

I never said anything about drinking it

Mars Ultor

oh whoops

Anonymous

Just the presence of a bottle of Jack can make interesting things happen

Anonymous

5:41 AM

Close enough

Anonymous

But seriously, calling someone a dick is not Being Nice

Mars Ultor

\o/ finished A through C

Dennis

@QPaysTaxes That's not acceptable here. Name calling will not be tolerated.

Anonymous

With the exception of calling Alex wrong, of course :P

Anonymous

I should have B done. I don't know why it isn't working.

Mars Ultor

5:45 AM

@Mego ?

how are you doing it?

Anonymous

Graphs

Mars Ultor

right now I have a 50% change my B large is right (i.e. assuming flipping any number of pancakes only reduces the number of contiguous sections by at most 1)

Anonymous

But apparently I messed it up somewhere

Maltysen

@Mego I seem to have brute-forced C-large

Mars Ultor

@Maltysen what

why

just skip if the prime factor exceeds 10k

Anonymous

5:47 AM

I'm going to use better factoring to try brute-forcing C-large while I'm asleep tonight

Maltysen