The Nineteenth Byte - 2014-08-06 (page 1 of 2)

« first day (1285 days earlier) ← previous day next day → last day (3541 days later) »

12:08 AM

Oooh, I just noticed that we now have only 2 less 10k users than Code Review :D

Is that a good sign?

Are we actually on our way to getting out of Beta?

IIRC, the number of high-rep members (or lack thereof) was one of the main things holding us back.

OK, it appears that for a simplified LOS system I finally have the blink program working

now I just have to implement the rest of the LOS test properly...

... and make the code a bit shorter (3.2kB ATM)

12:57 AM

I vaguely expected 2-4k for my idea on that one, which is why I didn't try

golfing something that big doesn't appeal to me. I hope someone gets it in 200 bytes of golfscript :)

1:29 AM

heh, I'm on 1777bytes of C# right now

I'll probably submit now as it is, and then golf it properly when I'm not asleep

1:48 AM

ok, posted answer to Blink

now I am allowed to sleep, g'night all

4 hours later…

5:57 AM

.. I0 .. ..
.. /\ -- ..
.. .. =0 ++
.. .. -- X0
\\ .. =0 ++
.. .. \/ X1
-- /\ -- ..
S0 .. -- ..
\\ .. S0 ..
.. O0 // ..

githubphagocyte

@MartinBüttner thanks for the messages and for adding me to the marbelous repository. I've been catching up on sleep. Looks like lots of progress made...

@Sparr I've not kept up with the advancements. What are X and S?

# ./pymbl.py fib.mbl 11
89

@githubphagocyte I kinda went off on my own after everyone went to sleep tonight. we have a chat room just for marbelous now. chat.stackexchange.com/rooms/16230/marbelous-esolang-design

githubphagocyte

@Sparr you made a fibonacci generator in marbelous??

X is an output that short circuits the function, causing it to return regardless of whether the other outputs are full

a single function recursive fibonacci generator

S is a subroutine

similar to F in the old spec

@githubphagocyte come over to the other chat room?

2 hours later…

7:45 AM

So TIL that the Bowyer-Watson algorithm was developed independently by 2 people at the same time and both published in the same issue of the same journal

woah

8:29 AM

oh no

missed a rather important clause in Blink

going to add at least 200 bytes I should think

9:01 AM

300bytes!??!?!

Martin Büttner

9:45 AM

interesting, so starred posts with enough stars can float back into the sidebar even if they've already been pushed out

@NathanMerrill would like your input on marbelous functions having a maximum number of outputs, each in a different direction, rather than a stream of an arbitrary number of outputs exiting in the same direction.

Martin Büttner

(I do :))

9:59 AM

and another bug

10:15 AM

I hate this code

I can't bear to make it any worse by subbing character literals for integer literals

@VisualMelon I think you can save a character by replacing struct with class. I know, it's not very much, but still...

no, the struct is there for a very good reason ;)

I make good use of struct comparisons

ah, ok

If I wasn't afraid of the LOS code I'd probably try and compact the 4 for-loops into 2

11:15 AM

2011bytes, this code is horrible

11:46 AM

@Sparr, what do I do about the dogfighting koth?

Martin Büttner

12:12 PM

@Doorknob if I CW something, can I undo that?

New Sandboxed Posts

0

A: Proposed Question Sandbox - Mark XIV

Inspired by lifecompetes.com Multiplayer Game of Life (GOL) There are n players that play the Game of Life (standard rules) on an 50 x 50 grid. (Size, border conditions? Toroidal, Absorbing, Mirroring?) Before first GOL-Step When the game starts, each player has 12 cells that he can place any...

12:34 PM

Flag for a mod if it's important or an accident @Martin

Martin Büttner

@Doorknob well I noticed that one of my own tweetable answers isn't valid as it stands. so it would be only fair to make that CW, too. but I think it's fixable now, in which case I'd like to go back. :D

Ah, ok. Linky?

Martin Büttner

@Doorknob codegolf.stackexchange.com/a/35611/8478

Done

Martin Büttner

thanks

Martin Büttner

1:06 PM

@Doorknob there's voting going on in the Marble spec draft

@Doorknob It's actually a fun interactive process

@MartinBüttner I have an idea

Each square is representable with 3 characters.

the 3rd is the marble

Martin Büttner

unprintable ascii? :/

however, there are 256 machines 00 to AF

...which will produce a marble on the square below

then remove themselves

githubphagocyte

Wouldn't 256 mean 00 to FF? What am I missing?

1:08 PM

sorry, yes FF

my bad

so, the format is still the same

Martin Büttner

but how do you represent a marble with a single byte?

but we can still take an intermediate layout and save it

...an ascii character

the 3rd one

Martin Büttner

so what about unprintable ascii?

we've been trying to avoid that like the plague

githubphagocyte

That's a relief - I thought there was some weird marbelous voodoo going on that I didn't understand...

0

Q: Convert Not an Answer to Community Wiki

I noticed a couple of answers to the tweetable art question have been converted to community wiki. These are questions which, although interesting, didn't keep to the question requirements (the code was longer than 140 characters and additional functions were added outside the specified red, gree...

githubphagocyte

1:10 PM

(well, I guess there is, a lot, but at least I get that one bit...)

@NewMetaPosts I was going to ask that very same question

@MartinBüttner escape codes?

githubphagocyte

@Rainbolt I'm torn because I find both the answers fascinating to look at but it seems unfair to leave the images there as it makes it look easier than it is to produce interesting images

Martin Büttner

@NathanMerrill that's 4 bytes, then you can use hex again

Ninja'd by the culprit lol

1:12 PM

internal memory storage would only be 3 bytes

input output may have a few extra characters

@MartinBüttner Doesn't fit anywhere, but make it CW because we want to let it stick around anyway. This sounds like something else I know...

Martin Büttner

@NathanMerrill but 4x1 would look quite bloated... and we're already considering to expand by one character to allow 256 synchroniser and portal groups

@NathanMerrill I also think Sparr has already solved this

he wants to add an optional parsing mode where each cell is 2x2

Btw when was the last time someone asked a non challenge question? They've been awfully quiet

hmmm

that's not a bad idea

Martin Büttner

@Rainbolt I think it must have been the JS question I answered a while ago

1:15 PM

I like it

the marble goes on top?

githubphagocyte

@Rainbolt what is a non challenge question?

Martin Büttner

yes

@Rainbolt this one codegolf.stackexchange.com/questions/30285/…

@githubphagocyte tips/advice questions

@githubphagocyte meta.codegolf.stackexchange.com/questions/1723/…

what's this about allowing another character?

@NathanMerrill It's to allow more than 36 synchonizer/portal groups

@githubphagocyte Some time ago people used to create these giant list questions where ALL tips for golfing in a language would go. Now we expect questions just like every other site: specific, narrow enough to answer with a single answer, general enough to be useful to future visitors, and all that jazz. But we never actually get any of these questions.

2

Martin Büttner

1:18 PM

@NathanMerrill wanna come to the marble room?

yes

link?

Martin Büttner

Separate chatroom for Marbelous esolang design

4

I'm promoting marbelous :D

githubphagocyte

I get it now :)

1:29 PM

@Martin Nice; I'm on mobile and have to go soon but I'll try to vote tomorrow morning! /cc @overactor

@Doorknob See you

Oh, and what @Rainbolt said is spot on, and I would star it if mobile chat didn't make it such a pain -_-

@Doorknob Did it for you

2:09 PM

wow, only just noticed that @Geobits does specify what to print if there is no safe path for Blink - on the bright side, "0" is shorter than "No way out"

@VisualMelon I caught that in your answer, but I assumed you did it for kicks, or to look better. Since it cost you extra, I figured there was a reason :)

no, I'm just incapable of reading specs ;)

I'll fix it if I managed to get the byte count down a sensible amount, doesn't seem worth editing when it's blatantly fixable

2:42 PM

Meh.... I can't run the Dining Philosophers until I post at least one more scoreboard for Good v Evil. It would just feel like a cheap cop out where I rack up tons of rep and then ignore the challenge.

2:56 PM

@MartinBüttner I came up with a name for the Github repo for the King of the Hill framework/library/whatever it is we are making. What do you think of "Hill of the King"?

Martin Büttner

I like that :)

HotK

well

I don't like the acronym too much :D

but Hill of the King is great

Awesome :)

HK is short enough

And holy crap

It's my Dad's cow brand

Martin Büttner

lol

@MartinBüttner How do I invite you to a repository?

Martin Büttner

settings, collaborators?

3:03 PM

@MartinBüttner username?

Martin Büttner

mbuettner

@MartinBüttner Do you need to be notified or did you just get the invitation?

Martin Büttner

I did

get notified

with an invitation

does that answer your question? I can't tell.

Sentences are hard. Yes.

3:40 PM

@overactor I guess it's time for a bounty on dogfight? not getting much attention otherwise :(

4:25 PM

so they released the richter magnitude of yesterday's earthquake -- 5.5

Martin Büttner

how far away was it?

@MartinBüttner It was in Orkney, which is about a 3 hour drive

definitely felt it though

everybody's freaking out because we're not used to earthquakes xD

indonesia had a 6.1 just 5 hours ago, no news about that

found it's earthquake db entry: earthquake.usgs.gov/earthquakes/eventpage/usb000s0b3

(obligatory one-upper comment) Eh, I was an hour from Fukushima.

Seriously, though, where are you that people are freaking out about a 5.5?

@Geobits we're not near any tectonic plate edges...

Martin Büttner

@Geobits he's a three hour drive away from Orkney

4:39 PM

the 2 main theories atm are 1) a mine exploded 2) the continent is ripping apart

Ah, yea. I can see that, not much happening in the UK quakewise.

Martin Büttner

UK?

too many orkneys

Unless you're more specific, Orkney shows up in the UK when I google it... that's why I asked.

Martin Büttner

@Geobits I believe he meant this: en.wikipedia.org/wiki/Orkney,_North_West

@MartinBüttner yup

Martin Büttner

4:42 PM

(and my comment was intentionally unspecific :P)

I used to live right next to it actually :I

Hmm, you are pretty far from a plate boundary. Interesting.

@MartinBüttner I saw your comment about using the Manufactoria esolang for my challenge.

Martin Büttner

yep?

I guess I could add two options for submissions: either using the actual game or the esolang

I was able to complete the challenge in the 13x13 grid.

New Sandboxed Posts

4:53 PM

0

A: Proposed Question Sandbox - Mark XIV

Paired Programming: Down Periscope! king-of-the-hill This is very basic for now, and lacks specific numbers as I flesh out the details and play with the speeds involved. I/O will be pretty basic, but it's not fully outlined yet. Four submarines are placed in an arena, two on each team. The...

2

Martin Büttner

@PhiNotPi hm interesting. Still I think it's discouraging for users if they have to be "good enough" to even participate.

Is the esolang the exact same as the game?

Martin Büttner

I never tried it, but I'd hope so

I think it handles conveyor crossings differently

Martin Büttner

esolangs.org/wiki/Manufactoria

ask @Quincunx ;)

4:57 PM

I was just looking at that.

Martin Büttner

@Geobits sounds like a good challenge! :)

although communication with arbitrary bots sounds hard

especially since there is no common information given to the team mates

Communications with subs is hard in general :p

Martin Büttner

well I meant that 16 byte message

The communication is there to use or not use as you see fit, really.

Of course, if you submit two entries...

Martin Büttner

there's no way to make the chance of a team mate decoding it higher than that of an opponent

5:00 PM

I was thinking of sending a team identifier with pings, but I'm unsure as of yet.

Like 0 for friendly ping, 1 for enemy.

Martin Büttner

I don't see how that helps

if I can't put some form of salt into the message which only my team mate knows

Hmm.

Martin Büttner

you could probably make that quite simple... a simple cypher with a 1-byte key known two both team mates

probably rather 2 bytes

anyway, the key should be large enough to not be decipherable with a single message, but possibly if you collect multiple messages

That's an interesting idea.

Even a simple ROT would probably work, since messages are naturally going to not conform to standard language letter frequency.

Or, use a checksum of your teammate's sub's name to encrypt it. Then the simplest way to decode it is to figure out who you're fighting :)

Martin Büttner

@Geobits I like that :)

in that case you might want to ask players to finalise their submissions

it would be really annoying if someone goes to the lengths of trying to fingerprint every single submission, and then they all make minor tweaks to thwart those attempts

5:12 PM

True, but I think fingerprinting might only work to a point.

why not let the programs allow their own encryption?

Martin Büttner

@NathanMerrill who would decrypt it?

the people listening

Martin Büttner

sorry, typo

How?

Martin Büttner

5:14 PM

but then how are the team mates better at it than the enemies?

To be clear, you can send any 16 bytes you want. If you encrypt it, so be it, although it won't be helpful to your teammate.

but if its encrypted, and your teammate is better than your enemies at decrypting

then he gets the message

Martin Büttner

@Geobits well if you're introducing the checksum of the team mates name as the key, then I think that makes fingerprinting an important component of gameplay.

@NathanMerrill But in another game your team mate will be your opponent. And your opponent your team mate.

yep.

Martin Büttner

Then it's pointless on average, right?

5:16 PM

In another game, even you might be your opponent.

Martin Büttner

@Geobits wat?

Well, if all pairwise teams are in a bracket, how could you avoid playing yourself if two teams you're on got near the top?

That's why teammates matter. You+A != You+B

Martin Büttner

ohhhh

I didn't read the tournament system careful enough

@NathanMerrill I think if the rules themselves don't give an advantage to intra-team communcation then there will be no advantage on average, regardless how clever a communication system you devise.

On that point, should communication even be there?

Geobits, why would you ever compete against yourself?

just declare yourself the winner

5:19 PM

It's not "you vs you", it's "you + alice vs you + bob".

You can't just declare one the winner.

I have another idea for a KotH challenge

oh, that's right

sorry

Martin Büttner

@Rainbolt go on, we're currently really lacking ideas. ;)

...you guys are?

You are in a tournament bracket. Every time you lose, your strength increases by 5 and you are out of the tournament. When you win, your strength decreases by 1. You can concede a match that you would otherwise have won. You have full knowledge of all standings and the strengths of other players. Repeat this process some number of times. You are scored based on wins.

5:22 PM

I'm not

I'm just trying to write code for my KoTH fast enough

Martin Büttner

@NathanMerrill no we aren't

ok good :P

Anyway time for lunch. Had to get that out of my system.

@Rainbolt I assume you win/lose based on strength, but how? Is it random if two players have the same strength (I'm guessing this is true at least for the start)?

Can both players concede? If so.... ?

5:46 PM

i yearn to play chess

eurgh i'm the most boring person ever ;_;

6:18 PM

i yearn for the game of life on non-regular grid to be reopened cause i finished my answer

i'm such an "overzealous user" ;_;

Martin Büttner

help them fix it then ;)

2

i dunno how :/

Martin Büttner

what's your score? :)

i dunno, i have like 10 instances constantly generating random states to try and find some interesting starting states

so "0"

The scoring there looks pretty arbitrary, imo. A spaceship gives 15, but then finding another distinct one is only another 15, for example, while I'm sure finding the second is more challenging. There's also no tiebreaker, unless I missed it.

6:27 PM

oh actually its like 15

@geobits up/downvotes could probably be the tiebreaker

Could, but that would probably just mean that it's down to a popularity contest, since most "serious" entries will try to max out the score.

@Geobits Highest strength wins. Ties broken randomly unless a concede is made.

Double concessions?

I don't know lol

One of them has to win in order to ascend the bracket

Unless you give the next opponent a bye, but that could cause a bit of confusion :)

6:30 PM

well the issue with the scoring is that i'm not entirely sure that for every tessellation pattern, you can find things that give points

But that's all technical details that I can work out later. Does the challenge seem interesting at all?

@Rainbolt It sounds interesting, yes. I'm just wondering if there's a provably-optimal strategy. It seems like there should be.

so maximizing points is probably going to be simulating a bunch of tessellations until you find a turing machine, which is boring :(

@Geobits I think that it's basically a max distance algorithm on a graph, which I have asked before but got only one answer.

@KevinL Agreed, boring, and crazy since the bonus is so high for it.

6:32 PM

And part of the reason I got only one answer is that it had to be optimal per the restrictions of the challenge

This way you could limit your search depth to 10-13 levels and do alright

@Rainbolt For a single bracket it probably is something like that. I'm not sure since you're repeating it starting at the bottom each time.

Ok, so maybe that is the search depth

Then again, the branching factor is crazy. Each match has several outcomes.

If I do this I am going make it a tennis themed challenge and add a rule that nobody except the author can use Rafael Nadal as the name of their submission.

So maybe there is something optimal, but it would take heat death to prove it.

Sounds like a good case for Monte Carlo.

7:03 PM

anybody here know haskell

@NathanMerrill i know about haskell

I'm trying to run the new entry

and its throwing an error

not sure who is at fault

What's the error?

7:23 PM

File name does match module name:

Saw: `Main'

Expected: `Pacman'

but in Pacman.hs, there is no Main

7:34 PM

here's marbelous calculating fibonacci(13) recursively:
https://gist.github.com/sparr/3a2933b8dc5b84a2d7ae

@Sparr THAT is a big gist.

it's a lot of recursion :p

1.6^n lots of recursion? :o

7:49 PM

This one is pretty -- but I can't reproduce it at all! Closest I can get is when the PPM is generated improperly (LSB instead of MSB) and even then it just looks like a variety of alpha-blended circles of different color. — DreamWarrior 6 mins ago

???

what is it with platform inconsistencies? ;_;

8:22 PM

I saw someone commit a file to github with windows newlines in it recently

:/

@Sparr Are you using a tool that can't handle Windows newlines?

I was

it might have been Python

You created a tool that can't handle Windows newlines?

no

Python was the tool

Python is a language...

8:26 PM

/usr/bin/python

(or wherever it lives right now)

You mean Python can't compile programs written with Windows newlines?

I'm really confused

python on osx, compiled with homebrew with the default options, does not recognize windows newlines

Probably because it just grabs the system newline and looks for that.

11

A: Handling \r\n vs \n newlines in python on Mac vs Windows

I guess it may depend on what you're reading from, but the built-in open() function takes a 'mode' parameter, and if you pass 'U' for the mode, Python will take care of the newlines in a cross-platform way transparently. It requires that Python be built with universal newline support, but test it...

(I admit to not having read it entirely, but it looked useful)

it's not just a python thing

I have to tell git to ignore the newline differences if I'm going to fork/contribute to the project

8:30 PM

Oh GitHub has an entire article on the subject: help.github.com/articles/dealing-with-line-endings

yep

in other news, here's a new version of my marbelous interpreter: gist.github.com/sparr/fb1105eecfbd5f8a6439

I'm learning GitHub so this is good to know. Especially since I work on Windows

@NathanMerrill I can help debug if you want

@Moop If you want

I understand the issue

aka, I understand what you are describing

8:46 PM

ok

letters = ['P' if len(self.square.players) > 1 else 'X']

seems not necessary

You actually shouldn't be getting a P at all

oh

that's legacy code

i am looking here github.com/thenameipicked/Pacman/blob/master/Pacman/…

I see where

that's the right function

i get the 'P' on line 82

yeah

you were right

that was the code until becoming a ghost was possible

8:55 PM

i see

it seems like the code for connection squares together is too complicated

@NathanMerrill oh no, I've been bested at pacman! I need to make a pathfinding bot.

connecting*

I figured out the bug

at least 1

that's good

why do keep track of walls and not paths?

why do you*

why would I do it the other way?

9:01 PM

because then you can represent the playing field as a graph

that's what i do with ghostbuster

I've never considered it that way

walls are just edges that don't exist

i think doing it that way makes a lot of things easier

(not asking you to rewrite)

as long as it works

@PhiNotPi @MartinBüttner Yes, the language handles conveyor crossings differently. Instead of 4 different types of crossings, there is only one (#). It simply sends the robot in the direction it was moving when it reached the bridge. In the game, the bridge could send the robot back instead.

@sparr thanks for the some of the logic

githubphagocyte

9:16 PM

@Moop - random interruption - did you know you can edit a post for a few minutes after typing it? I was delighted to find this out...

@githubphagocyte Yeah, but only for 5 minutes. I missed the cutoff

githubphagocyte

Ah I see. Sorry for telling you old news then...

no worries!

@NathanMerrill what was the first bug?

back

self.direction doesn't update

also, line 71

while cur_square.is_connected_to(next_square, wraps=True) and cur_square!=first_square:

won't cur_square == first_square all the time?

since you set both of them to the self.square

i think you want

and next_square != first_square

Martin Büttner

9:54 PM

@githubphagocyte fixed the Hofstadter sequence answer!

I didn't even need the forward declarations for the second case, but without the new function names it still would have been too long

githubphagocyte

@MartinBüttner that's great - I'm off to look...

Martin Büttner

in fact it's amazing how many characters I've got left now

has anyone really made use of all three functions calling each other to produce some sick grayscale image?

« first day (1285 days earlier) ← previous day next day → last day (3541 days later) »

Transcript for

Aug '146

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