The Nineteenth Byte - 2014-08-08 (page 1 of 2)

« first day (1287 days earlier) ← previous day next day → last day (3548 days later) »

12:07 AM

@githubphagocyte my attempt at explaining it, which is a bit to much or little verbose in some areas: github.com/marbelous-lang/marbelous.py/blob/master/doc/…

githubphagocyte

@Sparr I'll get round to reading through at some point - I'm jumping between different things at the moment...

I can't help but notice the Python answer to codegolf.stackexchange.com/questions/35471/dont-even-blink claims to take the doctor some way off the board for test case 1...

githubphagocyte

@VisualMelon you should probably add a comment - a lot of people don't see chat

yeah, I'm just trying to get python to work on my machine

so I can check the code actually produces that and he hasn't just pasted it incorrectly or something

(it's the same thing twice)

githubphagocyte

They've used python 2, in case you need to know

12:18 AM

yeah... I have a friend who would disown me if I had python 2 installed

and I don't actually know python, so I might just comment it appears he's pasted the answer twice

githubphagocyte

lol I've only ever used python 3 but I think about the only thing you'd need to change in this code is to add parentheses to the print statement as in python 3 it's a print function

@githubphagocyte I don't see your self-answer on the meta post; did you post it?

> I'm posting my own view as an answer so that the votes will show where it lies in relation to other answers.

githubphagocyte

Sorry I got distracted by a ping - it's being drafted so should be posted shortly

ok

1

Q: The line between art and programming

I really like this question but I know there are complaints of it being more about art than programming. I'd like to hear both sides of this and see what is in the best interest of the community and the site's public beta. I'm posting my own view as an answer so that the votes will show where it...

githubphagocyte

12:21 AM

@Doorknob now that you mention it I could have drafted my answer elsewhere before posting the question and then just pasted it in - don't know why I didn't think of that...

@githubphagocyte There's also a checkbox that lets you post question and answer at the same time, if you want :P

user image

githubphagocyte

@Doorknob thank you that may well come in handy in future!

1:16 AM

When is the Sandbox going to be retired, and a new one created?

didn't that happen like a week ago?

there's more 30 answers on the current one

I think I've got a marbelous main board for the tweetable mathematical art competition. Right now it just outputs R=G=B=i but now I can actually start implementing the functions

140 characters is 7x20 or 20x7... I can accomplish something in that

oh, drat, half that

7x10 or 10x7

still doable!

sadly even R=G=B=i takes four minutes to run for a 256x256 image

I don't think you're going to get many high-frequency traders using marbelous

1:33 AM

:p

1 hour later…

2:37 AM

Time to make more tessellation simulations

(say that 5 times fast)

githubphagocyte

So about an hour for a full 1024x1024 image then...

2:52 AM

I think I'm in the middle of accomplishing a new entry that is going to be technically difficult and also ugly

:(

@EricTressler for which?

still the tweetable art

you should map your tesselations to the poincare disc (like escher did)

Hey guys, I have a question in the sandbox I want to post to the site eventually. It hasn't gotten a ton of feedback, but I'm willing to make changes to it if that's what it takes to make a good question. I think it might end up just getting lost in the sandbox. Should I post it to the site anyway? Or wait in hopes of having more people stumble across it? I didn't want to make a pointless edit to bump it to the top of active, since I don't have any changes I want to make to it.

Or should I ask this question on meta?

no, here is probably the right place to ask

ok thanks! any tips?

2:55 AM

I'm sorry, I'm basically new here. But I'll take a look at it.

I'm new here too

I don't think most languages can accomplish your second task (or your first one, but the second is more problematic)

Also, letters are symbols, but that's a minor quibble

Yeah I wasn't sure what to call them

You're basically going to get a bunch of people using brainfuck and whitespace who can do #1, and I don't know what can do #2

I think even Shakespeare uses square brackets and periods

There are multiple ways to do number 2 in Perl, and if you use Brainfuck or something for #1 then you could use whitespace for #2

but I was hoping people would come up with more ways to do it

since they are far more creative than I am

I guess I know it is possible, but I don't know how good you can get them or how many options there are for both answers

3:00 AM

Well, maybe they are, but I personally could not accomplish #2 in any language

Without learning whitespace

Right so maybe is it too hard or restrictive?

I think so. I think that the only programming tricks you will see is language choice

I thought that allowing people to post half a solution will let more people answer even if they can't do both

and really, both of those are probably routine problems in whitespace

so I don't think it's a particularly good problem, since there is a language that trivializes both of them

ah okay that makes sense

any way to salvage it? I though it was cool to ask for 2 programs that can be different languages

but yeah it does seem pretty restrictive

3:04 AM

on the other hand, it doesn't trivialize getting a good score, so I'm not saying anything definitive

ok cool

I definitely appreciate your thoughts on it!

no problem

so yeah I'm not sure what I'd change to make it better

I'm of the opinion that it isn't salvageable. I've never seen someone try a Piet entry on here. I wonder how you would score it

I've seen a piet entry on here actually

3:06 AM

how did they score it

this was a recent one

20

A: Program your favorite phrase

Piet (honorary answer) Snakes its way downwards, running out the right through blue and re-entering through green. Prints out "Fruit flies like a banana." I used PietCreator to make this, but you can try it out here. Remaking Piet programs is a pain, so I started out at 35x35 and continued th...

they didn't score it

oh, nm, I did see that

odd that I didn't remember it

yeah!

I figured this question might still produce some good half answers

even if you can't get both

because you can do a ton without letters

I would ask again in about 12 hours when other people are around, and get some more feedback before you post it

okay sounds good I'll do that!

thanks Eric :)

3:10 AM

np

1 hour later…

4:13 AM

hhng debugging tessellations

what is that for

16

Q: Implement the Game of Life on anything but a regular grid

Conway's Game of Life is (almost) always played on a regular square grid, but it doesn't need to be. Write a program that implements the standard cell neighboring rules from Conway's Game of Life on a two-dimensional tiling of the Euclidean plane that is not a regular tiling of squares, triangle...

code-challenge game-of-life

interesting. i didn't realize that was alive. also looks like a ton of work

eh, i can do one in ~350 lines of formatted code

granted, i'm using a tiling thats not that difficult to store, but..

4:31 AM

wow i had a silly modulo error zz

5:01 AM

i guess i can run the life model, but i'm not comfortable with all of the fancy graphical output i see in that thread

i should probably go ahead and participate, just to get over it

better hurry before all the relatively easy tessellations are gone :p

5:15 AM

Nah, I go big or go home. Usually home

2

5:28 AM

I might just try a regular pentagonal tiling

Oh, it has to be a tiling of the Euclidean plane

5:44 AM

user image

6:24 AM

@EricTressler there've been quite a few actually. Typically the score is the number of codels.

That makes sense

1 hour later…

Martin Büttner

7:52 AM

@NathanMerrill the last one was up for months and got some 60 questions I think. I'm personally not a fan of retiring the sandbox at all

2

5 hours later…

New Sandboxed Posts

12:26 PM

0

A: Proposed Question Sandbox - Mark XIV

Using all keywords exactly once Overview The task is to write a program in which every keyword of the programming language is used exactly once and to output "Hello World!" The program may not fail, i.e. outputting "Hello world" and then making the program die silently is not permitted. Keyw...

12:41 PM

so I just found this

1 hour later…

1:50 PM

@cjfaure If you follow the links from there to Learn Python the Hard Way, you'll find an incredibly annoyingly laid out site.

Martin Büttner

just found this

geometricolor.wordpress.com/2012/08/19/…

could be interesting for the GoL challenge

I'm still trying to upload all of your finds :p

@Geobits wow that's bad

@MartinBüttner wow that's awesome

@MartinBüttner That's a nice tiling. Time to hit up the store and redo my guest bathroom.

the relation of the sizes is nifty

Martin Büttner

1:55 PM

@KevinL no rush ;)

I'm not sure how else to word this: meta.codegolf.stackexchange.com/a/1949/14215 Anyone else?

Martin Büttner

I think you might be able to find something if you set up a random initial condition that's mirrored on the vertical centre line

I've seen a few interesting vertically symmetric patterns, but they don't occur often enough and isolated enough to be sure that they will always vanish

(that is for your second tiling)

(btw that's how I was able to reproduce my glider for the other tiling... it happens way to rarely on a completely random grid)

of course increasing the grid size and decreasing the time step also helped

@MartinBüttner K, i'll see if I can do that + a stop/play button + clicking on a cell toggles its state during my lunch break

(in ~2 hours)

maybe also an input method that less verbose (like 21ab 22bc ...)

Martin Büttner

that would be great

2:19 PM

anybody here on dat sci? the chatroom is super inactive xD

2:29 PM

Hello @ChrisJester-Young

Chris Jester-Young

@Rainbolt o/

So now you're Rainbolt instead of Rusher? :-P

@ChrisJester-Young Yea. I wanted to start moving towards my real name

0

Q: Identifying Duplicate Answers

I have a problem with objectively identifying duplicate entries in my challenges, so usually I just don't even try. Some of them are really obvious, but some of them are not. Some solutions are too obfuscated for me to even understand. The moment I start to disqualify some of them, I have to be f...

yeah, yesterday my sister was working on an mlp animation (long story) and one of the characters was called "Rainbolt".

Whaaaat?

2:31 PM

^ my reaction

You all have permission to use my last name in everything you do that would make me look good

My name is already my full name. :D

Chris Jester-Young

@Rainbolt I applaud your direction towards using your real name. I remember the very, very early days of SO. Everyone used their real names.

I come from mathoverflow. There, almost everyone uses their real name

Chris Jester-Young

@cjfaure Really? Really spelt as cjfaure, or more like C. J. Faure? ;-)

2:33 PM

@Rainbolt VTC permission as too subjective.

Chris Jester-Young

@EricTressler Yes, I think it's awesome.

@ChrisJester-Young I added in the "J" for no reason, it's actually just "C Faure"

i have a people-with-second-names complex.

Chris Jester-Young

@cjfaure Ah, like Harrison J. Ford (who has no middle name).

@ChrisJester-Young and J. K. Rowling :D

@Geobits What is VTC?

Chris Jester-Young

2:34 PM

@cjfaure Yes, how could I forget. :-P

vote to close, I don't know what would make you look good.

Or FAO Schwarz

*may not be a real example

On ok. I'm being repressed! Doorknob edits my mismatched curly braces and CJY is voting to close my chat messages. So abused.)

Chris Jester-Young

@EricTressler Let's not forget P. D. Q. Bach. ;-)

@ChrisJester-Young Off-topic but I was doing some of my usual shady link-following and saw that you go to Mensa things - my piano teacher suggested I join and I know nothing about it xD

2:36 PM

@EricTressler Frederick August Otto Schwarz

No, I googled it, nevermind

@Rainbolt my eyes are bleeding

Chris Jester-Young

@cjfaure You're welcome to link-follow, I generally am careful only to post things that are more-or-less public knowledge. ;-)

@cjfaure Pretend like the C in CJY is an open parentheses. It helps calm the mind. ))

@Rainbolt Ah, and thanks for matching them up with close parens - but why isn't there a lowercase ")", then?

Chris Jester-Young

2:38 PM

@cjfaure So, I see Mensa as a social group for people with high IQ, where IQ can be measured by a variety of different tests, and passing any single one is sufficient to get in.

there is, it's \ni

er, no, that's the epsilon

Chris Jester-Young

@cjfaure You sound like a perfect Lisper. :-D

(Disclosure: Scheme is my favourite language.)

there's a symbol for reverse set inclusion in LaTeX, I just forget what it is

@ChrisJester-Young That pun was epic.

c ↄ

2:39 PM

Disclosure ba dum tss

Chris Jester-Young

:-D

@cjfaure Ah, you don't list a location in your profile. Otherwise I'd see if there's a Mensa group in your area. :-P

eheh, I probably won't get in anyway, the only IQ score I've ever gotten was 193 on some weird online test

@ChrisJester-Young indeed! But I can tell you that it's Pretoria, South Africa.

Chris Jester-Young

@cjfaure While online IQ tests are not considered valid, you only need to be 2SD above mean to get in. So for tests that are SD=16, you can get in on 132 IQ.

(Pretty much all IQ measurements are adjusted to have mean = 100.)

@ChrisJester-Young oooh :D Isn't there some sort of fancy-techno future science neuro-screening to find your IQ accurately though?

goes to duckduckgo

@cjfaure Scientology? :D

2:46 PM

> fancy-techno future science neuro-screening

I'll probably never sound that informed again :D however I did find this particularly smart sounding [wikipedia article](https://en.wikipedia.org/wiki/Neuroscience_and_intelligence)

IQ tests in general have been widely discredited, at least insofar as they imply anything much beyond the test. On the other hand, 2 standard deviations on a logic test is probably at least meaningful.

But anyone who's sitting around doing code golf puzzles is bound to do better on a logic test than the general public

Chris Jester-Young

@EricTressler Oh, I don't ascribe any meaning to it at all, other than that people usually have to be some kind of nerd to want to assign a number to something like that. :-)

So if Mensa is just a social gathering of nerds, I'm okay with that.

I'm about to post a marbelous entry for tweetable mathematical art

This is why they used craniometry :D

I think beyond early childhood, it probably has more to do with your interest in math and logic than it does general intelligence

Phrenology? heh

2:47 PM

because the size of your head is obviously proportionate to your intelligence!

it's only going to take an hour or three to run...

I can't remember for sure, but I think the guy who was the main proponent of that had an unusually smooth skull, or else had bumps where they were supposed to indicate intelligence

stupid multiplication function

i.e. his theory predicted that he was a genius. what a coincidence!

@Sparr !!!

2:49 PM

@cjfaure I had the framework done last night. 4 minutes to generate a 256x256 ppm image where every red=green=blue=x+y

@EricTressler Well, really everybody's definition of intelligence is "whoever thinks the most like me" :P

also, an hour was probably an exaggeration. maybe more like 15 minutes

Martin Büttner

@Sparr how are you checking the tweetability?

that is what are you counting and what aren't you counting?

Yeah, that's true. Ugh, I've been running this tweetable art program over and over since 12 hours ago. I fell asleep last night waiting for it to finish

@Sparr kewl :D though the images are supposed to be 1024x1024, perhaps you should generate 16 images?

2:50 PM

I'm counting the contents of the red, green, and blue functions

@cjfaure other people are doing 512x512, or 8-bit instead of 10-bit color, so I don't feel bad constraining the output. if you extend the main() of my code to 1024x1024 you'll just get the same output tiled 16 times

Chris Jester-Young

@cjfaure I'm okay with that. If Mensa gives people a place to belong to, who otherwise wouldn't belong anywhere else, I think that's a good thing. :-)

@ChrisJester-Young yeah :D

I could easily alter my main to run 16x as long and produce the appropriate output, but that would just be a huge waste of processing time

@MartinBüttner it's just bytes inside the function, right?

That's basically the motto of PPCG: "A huge waste of processing time"

3

@Geobits sure, but usually the results are distinct :p

Martin Büttner

2:53 PM

@Sparr yeah sure

I was just wondering how much boilerplate you were adding

err, as much as all the normal entries?

Martin Büttner

@KevinL I love the spaceship on your second tiling so much :D

@Sparr well in terms of available maths functions

a main that spits out the ppm header, two nested loops of 0-255, calls red/green/blue functions

oh, hmm

I totally forgot about maths functions. I didn't add any

Martin Büttner

e.g. I think it should be legal to implement an integer square root and use that

I guess to be fair I should add square and cube and cube root

however, the runtime is already ridiculous even having just one multiply instruction in my whole code

Martin Büttner

2:55 PM

well you don't need to if you don't use them

I don't think I'll be doing 65536 square root operations in a marbelous program until we have a much more efficient interpreter.

(and probably not even then)

@Sparr I wonder if they wrote Earth in Marbelous

unsigned short red_fn(int i,int j){return ((i/2)+(~j))*4;}
unsigned short green_fn(int i,int j){return i*j;}
unsigned short blue_fn(int i,int j){return ((i)+(j/2))*4;}

that's the C equivalent of what I've implemented in marbelous

using about 1/4 of the available bytes in total

oh jeez, that 6 period oscillator shows my task manager for 2 frames

red is 27, green is 53, blue is 22

if I had a function for squaring, then squaring would be cheaper than multiplying, and my choice of functions would change

2:59 PM

Hmm. It turns out that random vines look pretty good as a background if you green up the color and extend it to 3840x1080 :) Just too bad imgur mangles large pictures.

@Sparr um, you can square by bit-shifting, right?

wait, no, that's multiplication

We should host a competition for generating a cool looking screensaver

Nothing has ever been as good as pipes

@Rainbolt +10, and I'm working in base infinity :D

@cjfaure thanks for the reminder, I need to rewrite the multiply function now that we have shifts. it can probably be faster now

@Rainbolt it's +inf ;_;

3:02 PM

So people like to remove/edit things, huh? Fine

I'll just do that, too.

oh yeah oops ^^;

ninjedited

Yes!

main board in my marbelous program is running about 100 ticks per second right now, and it has 100k ticks to go, and it's slowing down. :(

I should write a compiled implementation of Marbelous.

it's plausible. it's not even a very complex language, in terms of components to implement.

3:06 PM

@Sparr the trouble is paradigm conversion :P

there are a lot more options between here and there, though

sparr, can you commit your latest code?

even my python interpreter could be MUCH faster with a bit more work on the data structures used

i.e. i don't want the generated code to actually store the board.

@NathanMerrill already done, why?

3:08 PM

I'm going to write a GUI for it

@cjfaure I think what you'd end up with is a queue of actions to take, and each device would be augmented with the places its outputs end up and how far away those places are, to add that next action to the queue

now that I've written that, it occurs to me that that might be a goal for the interpreter as well :)

right now I track two falling marbles out of a splitter, iterating the fall every tick until they reach something else interesting.

instead, at the moment they leave the splitter I could look ahead to that next interesting thing, and add "marble X does interesting thing Y on tick Z" to position Z in the future queue

@NathanMerrill that will be awesome. I won't be working on it much more today, or likely tomorrow, so you've got some merge-conflict-free time

Ok

I want to pass some stuff passed you though

ok, switch to the other room?

sure

There's a new problem to output graham's number. Ron Graham was my thesis advisor

And one of my coauthors.

Interestingly, the number there is not his original definition of the number. I'm not sure where it got changed, but I think it may have been simplified in correspondence with Martin Gardner

Martin Büttner

3:14 PM

guys, can we write down some design goals at some point... we've all been designing around our individual visions, and so far we were able to reconcile them fairly well... but I think if we don't have some overarching design written down, someone may at some point question some previous important design decision and replace it with something that just seems like a good thing at the time

1. Make Rainbolt a sandwich.

2. Sudo make Rainbolt a sandwich.

nice xkcd ref there

What kind of shell uses numbers as the prompt that way?

My shell that increments the prompt between every command

Duh

Every 100 commands it tells a funny joke

Guest comic by Bill Amend of FoxTrot, an inspiration to all us nerdy-physics-majors-turned-cartoonists, of which there are an oddly large number.

3:19 PM

Hmm. Since there's no output from the first command, it's hard to say, but I'm guessing you don't have both 'Make' and 'make' commands. Try using lowercase without the sudo.

alias make='Make'

Sudo make Rainbolt a sandwich.

(three minutes worth of Googling right there)

"However, I think you're right; this question seems too difficult for this site's community." - ljacqu :o

@Rainbolt Ok, you're a sandwich. badum tss

2

I can't decide if I'm insulted or illiterate right now

3:48 PM

I liked that problem @Geobits

Which? The sandboxed one?

0

A: Tweetable Mathematical Art

Marbelous The actual output image is rather mundane compared to other entries here, the equivalent of about 25 bytes of C. I had more fun writing the main loop than the R/G/B functions. Note that spaces between cells are optional, included here for clarity. The Gren function is actually 53 by...

@cjfaure ^

@Sparr :D

this is all your fault

@Sparr i feel so guilty :D

Small print: this schedule will eventually drive one stark raving mad.

my new schedule.

Um, the 28 hour thing xD

4:00 PM

I used that schedule in college. It's neat.

i don't think my parents will approve though

if you only talk to them on weekdays, they might not know :)

i wish :P

4:23 PM

Idea for a golf: **Light some fireworks**
Given a list of fireworks that need lighting and the times which they should light, arrange the fuse assembly so that they will all light at the right time. Each unit of fuse takes one unit of time to burn. Example input and output:

a 3 b 6 c 6 d 8 e 9 f 9

---a
  |
  |-c
--|--d
| b |
f   e

Now, how do I prevent people from taking too much fuse and just using a single branch for each firework? I'm not sure how I can prove optimality in all test cases.

easy avoidance of the single branch solution is to feed them test cases where 3+ fireworks have the same timing

are you sure this is a "solvable"/golfable problem? maybe it needs to be a best-answer challenge?

Yea, that's what I'm trying to figure out.

Originally the plan was golf, but the more I think about it the less sure I am.

Sparr, the timing condition you mentioned doesn't stop single-branch answers

------a
 ||||
 |||e
 ||d
 |c
 b

All length 6

oh, that

sure, but try three fours, three fives, and three sixes

the naively constructed tree branches will start to overlap

4:28 PM

his diagram generalizes to n_i of length k_i

wait, maybe it doesn't

sorry

@EricTressler either I misunderstand that statement or it's wrong

It couldn't handle seven 2s for sure, but nothing could with a single lit point.

right, everything has to be sufficiently large

At first glance, the problem is too hard

I think it's a good best-answer problem

At least for golf. Maybe for best-answer

4:31 PM

break ties in golf style?

so if someone DOES find a perfect algorithm, then it turns into a golf competition

Sounds like a good plan.

If you have a large enough set of test cases, there should be no ties. But yeah, I think that's fair

You're right, if there's an optimal algorithm, then I guess golf is a good tie-breaker

there will be solution-ties for any two implementations of the same algorithm

Hmm. Can you think of a straightforward way to generate valid test cases?

I could just count invalid ones as zero-score and generate random numbers, but...

@Geobits ever do a topcoder challenge?

4:32 PM

Nope. Glanced over the site, but never participated.

You could just generate test cases and let the players decide if they're valid

they give you a bunch of test cases. you solve the problem. then everyone looks at all the solutions, and you get extra points for coming up with test cases that break others' solutions

no one has done that yet on PCCG but we COULD...

Yeah, but it's too interactive for this site, unless you want to do a lot of work on OP upkeep

IIRC, Martin had people submitting extra test cases for Vector Racing. The idea was that if you had a test case that your algorithm performed well on where others didn't, it gave an advantage overall.

it would be KOTH-type work

Martin Büttner

4:50 PM

Yes, I did that with a code challenge, and I intend to do that with the next code challenge as well. Someone also recently suggested running two challenges in parallel (overactor I think), where you could pit solutions from one challenge against solutions of the other. A suitable challenge for that would be a two-part thing, where one challenge is to design complicated mazes and the other is to solve mazes (there's a sandbox proposal for that).

I see people all the time pop up with "You failed this test case I just made up [link to jsfiddle]"

@MartinBüttner The solver challenge would be invalid until at least one answer exists on the creator challenge

I guess you could seed it with a simple creator

sure, but you could seed the creator challenge with a bunch of test cases

just like a KOTH should be seeded :p

Mine never are, since the sizes of most of my maps/groups/whatever depend entirely on the number of submissions

you can say the seeds will be omitted in the real competition

or you can not, since if you don't then someone will submit seed-like bots anyway

such as my pacmen

I had a submission once that only joined when the number of voters was even (to force an odd number of participants)

4:55 PM

also, writing seeds helps you find bugs in your controller

@Sparr What if the submissions are using random number generators? What if they wrote something that is time sensitive?

For example, if the submission was "output how long it takes to run this statement in CPU cycles;", I don't understand how I would seed that

err

by "seed" I just mean providing a few entrants that do something valid

Seed the competition, not the PRNG gen.

Oh, when I thought of seed I was thinking the goal was to get predictable results

As in, seed everything

Lunch time!

« first day (1287 days earlier) ← previous day next day → last day (3548 days later) »

Transcript for

Aug '148

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