The Nineteenth Byte

6:44 AM

@Doorknob a) Thanks, b) Ah that's unfortunate. So we just repost our challenges in the new sandbox? :-/

@Sparr thanks :D ... and good luck beating it, I never intended to be on top of that leaderboard ^^... it was just a really quick a stab at a half-decent strategy

8:21 AM

@Doorknob I just went through all the drawing questions. There was only one exclusively for ASCII art and one where ASCII art was a substantial component. I retagged the former and only added the tag to the latter. There are several that allow graphical output or ASCII art. I think that's fine though and we could just change the graphical-output excerpt as I suggested yesterday, to only redirect to ascii-art if the challenge is purely ASCII art.

Alternatively, we could write that if ASCII art is allowed, too, both tags should be used.

@Quincunx Do you see that pen next to this post? chat.stackexchange.com/transcript/message/16808464#16808464 ... Guess what I corrected ;)

Btw, while looking through drawing, I found this tag: source-art. It's only been used once, but I'm pretty sure there should be another few questions where it's applicable. Should I remove it or add look for other questions and use it where applicable?

Trimsty

10:01 AM

y'know how flawr mentioned rpscontest on the rpslv tourney?

now I'm hooked xD

rpscontest.com/entry/3830005

woah! another match and it won that one too

THREE rounds, still 100% win rate

4!

10:19 AM

and there it goes

I should try my markovbot there :D

Trimsty

@MartinBüttner yes you should :)

Oh well, win rate isn't what matters

what matters is score

ugh, I'd have to port it to python though

can't be bothered

@MartinBüttner your dominoes have eaten the last 3 days of my life, I've written a program which lets you draw domino layouts, why have you done this to me

Trimsty

my sister got a deviantart. yay ._.

april-black.deviantart.com/art/…

10:37 AM

@VisualMelon damn it, I forgot to add the small print again :P

that's great though... that could be really useful for the follow-up challenge. are you going to add this to your answer?

I wasn't intending to, because it's WinForms based

(I'll post the draft for Domino Circuits in the sandbox later today)

joy of joys!

I wrote the program so I could try out gates without having to align stuff in notepad

@VisualMelon you could put it up on the webs somewhere (like gist.github.com) and just link to it, so it doesn't clutter the post, but people who are interested can still get it

in fact, providing something like this could greatly increase the popularity of Domino Circuits, if people can just try solving these things by hand with a few clicks.

I spent last night trying to get free-hand drawing working

and it's working surprisingly well

(if you only need the dominoes to fall over in one direction)

I might as well put the whole thing on git-hub if you think it might be useful, I usually only do that for code which I'm expecting to break

10:46 AM

that would be great!

I should probably improve the interface first, and add Redo

if I'm going that way for Domino Circuits, I might have to reimplement it in a browser-based way, though... but it would be a great starting point

I'd offer my code up as a starting point, but it's horrendous

ah don't worry, we're not on CR.SE here ;)

11:03 AM

well, that's happened if you're interested github.com/VisualMelon/Dominoes - now I just have to remember to commit to it

11:30 AM

@VisualMelon awesome, thanks

wow, your VS is old :D

I use #Develop for C#, actually

but yeah, it's pre .NET4

are you a student?

indeed

second year CS degree starts in September

but I've been coding for years

dreamspark.com if you want an upgrade ;)

(VBA originally, then VB.NET, then settled on C#, though I do a lot of C++ when C# isn't suitable)

I have #Develop 4.something which does C# 4

but I only use that for my DX11 work

and VS 2008 works fine for my C++ work, except that which I do in VS2010

(express)

11:33 AM

kk

I have lots of IDEs open all the time, so having 5 (VWD 2010 also) that I use makes it easier to organise them on the taskbar

if I had Win8 I'd probably grab VS2013 from dreamspark and find out what Roslyn is

11:49 AM

Okay, I'll add the new tag as per meta.codegolf.stackexchange.com/a/1890/8478 soon

I just need to decide on a name for it

best-complexity perhaps?

For what it's worth, I'm in favor of [fastest-algorithm].

That way it shows up as suggested when people start typing fastest.

@overactor so would [fastest-asymptotic], but it sounds a bit awkward to me (I'm not a native English speaker though)

Neither am I but it seems a bit awkward to me as well.

The meaning might also be less obvious to some people.

@ProgramFOX that would include space complexity, and after giving it some thought I don't think these should be thrown into the same tag, simply because that would still mean that the tag itself does not imply a specific winning criterion without the OP stating what he actually wants to go by

11:58 AM

Ah, ok.

also overactor has a really good point regarding auto-completion

yeah, that's true

(yay, nice question badge on meta)

New Sandboxed Posts

0

Domino Circuits code-challenge It is possible to build simple logic gates from dominoes. Hence, by combining these or otherwise, arbitrary binary functions can be computed with dominoes. But of course, everyone who has played with dominoes (except Robin Paul Weijers) has experienced the disapp...

damn it, I had hoped to throw both proposals into the same sandbox feed update

(there's another one coming in 8 minutes then I guess)

12:14 PM

ah, scored by number of dominoes, I assumed it would be by size of grid, I like this

New Sandboxed Posts

0

Underwater Survival Game king-of-the-hill Lions and bears are extinct, the wolves dominate the fauna on land. But in the depths of the sea the battle for survival rages on. This is the spiritual sequel to Rainbolt's Survival Game, with a few additions. The arena is the sea and hence three-dime...

@MartinBüttner the ///////// in your examples are a bit confusing

I don't know...

nah, probably fine

Well they need to be part of the output

yeah

NAND, NOR, and XNOR could be interesting

I already have 30 examples in my benchmark set (which I'll later today, I think)

it includes all 16 binary functions

as well as de/multiplexers, a half adder and a full adder... should be interesting ^^

with this one I'll be happy if I get a single working submission :D

12:26 PM

fabulous

I'll try not to disappoint ;) though I won't be going for compactness

that would be awesome. I think if there is already an answer that shows that it is possible, that might be quite encouraging... because then other people could focus on compressing solutions if they can see that the baseline case is doable.

is it unsportsmanlike to start work on the problem before you've posted it?

no, I don't think so. there is a chance that the spec will change in some details though. ;)

understood

but I sometimes already do golf challenges which are still in the sandbox

12:39 PM

how hard it can it be to get two bloody upvotes for a Revival badge? :D

12:52 PM

I don't think I'll be using your power line if I do this

You need to for some functions

you can't implement negation or constant 1 without the power line

@MartinBüttner I might not be able to read, but I don't think it's quite clear how the thing is tested - do you initially replace all the left-most | with D ?

I'd provide my own power with forwarders

well only those left-most | will be replaced with D which are 1 in the input... your grid has to work for any arbitrary combination

too much effort thinking about how to get the power-line to where it needs to be ;)

ah, I see, so you can't assume that the | will be replaced in the 0th frame

so you /can't/ rely on your own forwarders timing

OK

very well, I /will/ be using your power line then ;)

basically, in any case where output for (0,0,0,0...) isn't all zeroes you need the power line

12:56 PM

@MartinBüttner I was intending to run a NOT line along all of the inputs, with their own forwarder by the input

but as you can't time that, I'll have to do as you say

what do you mean by forwarder?

WEDCXZAQ

ah okay... yeah basically you can't get any additional input anyway

in the left most column you get only exactly M dominoes

and no other dominoes apart from some of those and the power line will be pushed over in the first frame

yeah, I understand now

k

12:58 PM

there goes my genius master-plan

;)

really don't want to have to add zooming to this program...

@MartinBüttner do you expect all the outputs to arrive at the same time?

1:18 PM

no

I'll simulate the grid until no dominoes are falling any more, then I'll read off the output

ok

New Sandboxed Posts

1:50 PM

0

Use the whole alphabet in your code Your code has to use the whole alphabet (abcdefghijklmnopqrstuvwxyz) in your code (uppercase or lowercase does not matter) and you have to use it in the right order. You can use as many characters between each of those (which can be characters of the alphabet...

2:13 PM

ok, I think I've worked out the dominoes bit... now I need to work out the maths bit

sweet :D

2:58 PM

Hi guys, I added a bunch of code for my KOTH challenge and threw it around quite a bit, can you check it out? meta.codegolf.stackexchange.com/questions/1847/…

I'm sure it's entirely flawed though;

Any help would of course be greatly appreciated :^D))]

Or feedback, feedback is cool too.

4:48 PM

@Rainbolt you're supposed to be on vacation!

(it's nice to see you here though :P)

@MartinBüttner i.imgur.com/592WBNQ.jpg

Took a picture just for you :)

:D not bad

K I have orders to carry umbrellas down to the beach and dig holes and set up the elders with some shade.

Later guys!

hf!

@Doorknob Do we retag closed questions? codegolf.stackexchange.com/q/11530/8478 This one says something about complexity... but the scoring is so weird, if I retag it at all, I might just remove the fastest-code and be done with it.

@Rainbolt Where exactly are you at?

@MartinBüttner why is your name on here sometimes Martin and sometimes Martin Büttner?

5:02 PM

lol if you google-image-search for it, you get mostly pictures of snowy mountains

@overactor I hope it's always Martin Büttner? Where did you see just Martin?

@MartinBüttner i.imgur.com/yJcFJWv.jpg?1

Rather peculiar

Probably it's overflowing.

But it is strange that it doesn't always happen.

And there are also no profile pics on your screenshot...

You have one and so does rainbolt

I don't see any for me or Martin.

Perhaps it's blocking of a specific image provider.

Are those images not hosted on SE servers?

5:09 PM

Yes, both your pic and Martin's pic are linked to Gravatar, while mine is not.

@overactor Most images are stored on i.stack.imgur.com.

But if you create an account, your image is linked to Gravatar.

aha

Only if you manually upload an image, it will be uploaded on i.stack.imgur.com.

Oh, I figured it out

Ghostery is blocking Gravatar

Did anyone happen to check my code for the KOTH challenge?

nope sorry... my Java isn't really up to speed anyway

(that'll hopefully change soon, because I intend to write the Underwater KotH in Java)

@overactor I think this can be shortened:

if (player1Moves.length <= 0) {
moves[0] = new Move(player1Planes[0].getDirection(), false, false);
} else {
moves[0] = player1Moves[0];
}

if (player1Moves.length <= 1) {
moves[1] = new Move(player1Planes[1].getDirection(), false, false);
} else {
moves[1] = player1Moves[1];
}

You can replace it by:

5:22 PM

Good luck with that Martin, java isn't tooo hard

^ Pressed Enter too soon...

happens

Errmm... wait, I think I'm wrong. I didn't take in account that there was also an else statement.

The controller code is far from elegant

	if (Math.abs(NS) + Math.abs(NS) + Math.abs(NS) < 3) {

Can be shortened:

if (3 * Math.abs(NS) < 3)

5:26 PM

oh

that should be: if (Math.abs(NS) + Math.abs(WE) + Math.abs(DU) < 3) {

Ah, ok.

thanks for pointing that out

Not sure whether this will compile:

	NS = (NSDir>0)?1:(NSDir==0)?0:-1;

If it does not, wrap the second ternary operation inside parentheses.

I'll try it

It might compile, not sure about Java's rules on this.

5:28 PM

I have a few of those

collided[i] = true;
collided[j] = true;

The following also works, and is shorter:

collided[i] = collided[j] = true;

oh, that's neat

In the Plane constructor:

 this.position = new Point3D(position.x, position.y, position.z);

If you copy this because it is a reference type, then the above line is good. If it is not, then it is not necessary to create a new Point3D.

if I do this.position = position;

it's a reference type I think

position is an instance of the class Point3D

are all classes reference types?

In C#, they are. In Java, not sure. Probably they are.

Ah, found it:

> Every object is either a reference or primitive type. Reference types all inherit from java.lang.Object

A class inherits from Object by default, so then a class is a reference type.

Then nevermind what I said about the constructor :)

5:41 PM

yay, I did it right

You have many similar code in your constructors. You can put the equal code in another constructor and then call this one from the two constructors you have now: leepoint.net/notes-java/oop/constructors/constructor.html

    //============ Constructor
    public Point(int x, int y) {
        m_x = x;
        m_y = y;
    }

    //============ Parameterless default constructor
    public Point() {
        this(0, 0);  // Calls other constructor.
    }

You can do something like that.

That does make sense.

Though I don't want people calling parameterless constructors

@overactor You can make it private, I think (that's what I would do in C# at least).

ooh

Just found that it's also allowed in Java: stackoverflow.com/a/2816139/2619912

And I think you cannot do very much with a parameterless constructor. You will have to make the private one with arenaSize and direction as parameters, and the public ones have arenaSize, direction and the point indicators as arguments.

In both public constructors, you can call the private constructor on the first line.

5:51 PM

I now did this

public Plane(int arenaSize, Direction direction, int x, int y, int z) {
this.arenaSize = arenaSize;
this.position = new Point3D(x, y, z);
alive = position.isInArena(arenaSize);
this.direction = direction;
}

public Plane(int arenaSize, Direction direction, Point3D position) {
Plane(arenaSize, direction, position.x, position.y, position.z);
}

should work too right?

Oh, yes, that's much better than what I had in mind!

Doesn't always work I suppose, but it did here

Thanks for your help

does the code read okay?

Especially the way participants can use the helper functions.

@overactor Does it compile? I'm asking because the tutorial uses this(arguments here) and not ClassName(arguments here).

@overactor I think it does, but it is a good idea to document in your question how to use the controller.

Ah, for testing purposes.

What's for testing purposes?

5:55 PM

I don't have eclipse set up on this computer

Ah, right.

You can always use ideone.com to test code online.

well, the controller is only needed for me and for people who want to test their entry against others

@overactor You can use Rainbolt's approach for his Wolf KotH contest. He stored all participating classes in an array, so if someone wants to test, they just have to adjust one file.

yeah, I was thinking of doing that.

On Ideone, I found out that you have to use this(arguments), because ClassName(arguments) does not compile.

ideone.com/cAQCN5

Main.java:12: error: cannot find symbol
		Test(0,0);

With this, it compiles fine.

6:02 PM

ok, thanks

You're welcome!

I was trying it too, but got other weird errors.

Did you add the Test class, and kept the other class? That's something I don't like about Java: every class has to be in another file, otherwise you get an error.

I copied relevant code and rewrote it a bit.

I saw what my mistake was though

Thanks a lot for your help

I'll get going now

If you find think of anything about the challenge, leave a comment

If you don't feel like thinking about it, That's fine too.

Cya

See you!

6:50 PM

@MartinBüttner re Underwater: How do you win / When does the game end?

when humanity boils away the oceans

So..... not long now?

yeah, you better hurry with getting your entrant together

Crap. I'm pretty sure I want to be a shark if plain survival is the goal. They move faster and can apparently outrun anything.

Probably squid for attack. Long range vision allows better formations.

Wait, do members of a species/tribe have knowledge of others of their own kind, or is it just like Rusher's?

@Geobits Ha, see I knew I forgot something. Same way as in wolf: most surviving specimen.

@Geobits The idea was to be able to distinguish species but not bots (basically like Rainbolt did). I'm open to suggestions though

6:58 PM

Survival after how many rounds?

@Geobits 1000ish? Will depend on how fast the controller turns out to be.

Hmm. I'm not sure if 1000 turns is enough. Your board looks like it will be pretty sparse as far as contestants go (~1000 cells per animal), so it might take some time for them to find each other on a regular basis to fight/die.

I think I'll tweak the numbers to make it more dense

Wait... how many members per entry? I thought it was 10 for some reason, but I don't see that now that I look again.

should be 100

to be honest, I copied most of the numbers from Wolf for a start

7:06 PM

Oh, then 100 cells per animal. Not nearly as bad.

I'll definitely have to add up and down-going nets well. Otherwise, dodging them is trivial.

They are supposed to be well deadly to counter the whales' boon

@overactor Orange Beach

@MartinBüttner Lol your name is Martin now

Except when I ping you it populates your entire name

@Rainbolt It's been like that for a while ^^ ... I think I changed it a day or two after you did

oh you mean just Martin

wait

see here chat.stackexchange.com/transcript/message/16825051#16825051

aaaaaaaaaaaaaand rep cap at 415 rep, yay! :D

codegolf.stackexchange.com/questions/35257/… ... is he gonna leave with a KotH? :D

7:28 PM

Oh you're back to Martin

It appears that when you say a lot of things in a row, your full name and also your rep show up

@Rainbolt cool

Say, you're like the king of KOTH challenges, could you check out what I did so far?

1

Dogfight (Work in progress, looking for help) Sandbox note: Since no one has offered hosting this challenge I've started writing some code for it, I don't think I could host it the way I envisioned it so if I do this, it'll be java only. The advantage is that there will be a lot of boilerplate c...

2

Unless you're too busy being on holidays of course

Also, do you surf?

@overactor I have crowned kings, but never been one. You may call me the Chancellor if you wish.

I'll check it out real quick :)

7:45 PM

Is that a tag wiki or is that a tag wiki? codegolf.stackexchange.com/tags/fastest-code/info (of course feel free to improve it further!)

@overactor I can write a wrapper for you in Java that will call other entries, but you would have to modify your requirements slightly.

Basically, the contestants need to include a command that can be used to run their entry. Like python MySubmission.py or ruby MySubmission.rb

If you wanted your life to be easy, you could just download compilers for Java. Python, Ruby, and C++ and limit it to those four languages. I feel like that would satisfy most people

@Rainbolt While you're floating around the sandbox... would you mind having a quick glance at my two proposals? ;)

Oh I have another idea

Alternatively, go and enjoy the sea. :D

What if we zipped up like 15 compilers and placed it on the Google drive for people to download

And instead of adding it to your PATH or whatever Unix equivalent is, you could just unzip them in the same directory as the submissions

7:57 PM

Assuming any/all licenses were abided by....

@Rainbolt Sounds like a good thing to include with the framework

As long as my reputation is worthless, I think this counts as noncommercial use

Right. I'm just saying you'd have to check each individually. Some things are free, but not distributable.

Oh I see

@Rainbolt So basically, people who don't want to use java will have it more difficult?

but at least they could participate

?

8:00 PM

They literally just have to understand how to call their program from the command line. It's something you learn in your first ever programming class

@overactor It's not even necessarily that more difficult.

Of course, just having to implement a type-safe interface is a lot more idiot proof.

Coming up with python MySubmission.py really isn't that hard.

But for you it might be, because you would have to learn every language

But in general people would probably rather write something in a language they're comfortable with if they can.

Wait, would they have access to the helper functions I wrote?

gotta go

Nope

(but I could wrap those too lol)

8:02 PM

enjoy your holidays

see ya

cya

Java is freely distributable

@Rainbolt have you even seen the RSPLV KotH? I think that happened after you left

@MartinBüttner No not yet

Python is freely distributable

Somehow I've been leading the last two leaderboards with a fairly simple submission ^^

Ruby is freely distributable

gtg play dota

8:07 PM

hf!

(gcc and gfortran are obviously distributable too)

I've got an idea for a challenge. It's team-based, but not koth-style with a winner. Competitors would choose a team and add their bot to it, and one team wins (say red/blue). My hope is that people would choose the weaker team to help even it out, but that seems naive. How could I make it cooperative in this way without being unbalanced?

Think capture-the-flag.

Well, for starters by coming up with gameplay that puts the minority at an advantage

which brings limited resources to mind

Well... I don't mean there would be an actual minority on the field. Each team would have 180 competitors, divided up by how many entries on that team (remainder some default-bot).

My concern is that once one side starts winning, every entry will want to be on that team so they "win", even if their bot doesn't contribute/improve the result.

that could be odd as well though... my bot could be really good in the beginning because it relies on having a lot of instances, but then there are suddenly 89 other entries and now I only get two bots

Right.

8:11 PM

but maybe that's what you intend

simplest solution: you assign a team randomly

Yea.... I wanted to avoid that, allowing people to pick (say if they saw a weakness in a certain bot, they might choose the opposite team). If I can't find a way, I might make it alternating rather than random.

It is the simplest, but also not as fun.

so you do want to allow people to not side with the losers but you don't want them to side with the winners? ;)

Also, my intent was to allow multipl submission from a person, but only for the same team. Nobody can play both sides.

Yea, pretty much. Sounds like a losing battle.

There has to be some way to incentivize people to choose the currently losing team, but I don't know what it is.

I mean... you could make the winner the best submission from the best team... so if a player joins the losing team (and manages to turn the tides) he likely has fewer contestants in that team than in the already winning one

what would make for even more interesting dynamics is to let the trophy go the best player in the losing team :D

Right, but I don't know how to define "best". Maybe the one who actually captures the flag? Or carries it the longest distance toward home.

That would discourage defensive strategies, though.

8:20 PM

Well... depending on how long one trial takes, you could redo the trial once for each bot (without that bot) and see which bot had the greatest positive impact

that might become infeasible for a large number of bots or time intensive trials though

My rough estimate shows it might be time-intensive.

The number of bots doesn't matter as much, since the field is set to 360 total agents, but there would probably be many turns and calls to each of those agents.

Not sure exactly how long it would take, but redoing it n times every time sounds like a bit of work :D

parallelisation, yeah

hm, I'll think about it

btw, have you seen the Domino Circuits proposal as well? any comments on that?

Oh, haven't seen that one yet. Will look later. Time for me to head out ^^

alright

see you

Doorknob

8:55 PM

@MartinBüttner The goal of closing is to improve the question later, so yes, do everything you can to improve closed questions

@MartinBüttner Alright, tags are officially synonyms!

@MartinBüttner Uh, I guess. I'm not exactly the local sandbox expert :P

@Sparr Alright, I still have to go through all the new comments and submissions on the post, so I may not have seen your most recent comments. But if you want to talk about it here in chat just ping me.

9:51 PM

@Doorknob All right. Then I've got 6 more to retag tomorrow. I'll do three in the morning and three in the afternoon.

Tag wikis have been populated.

(for fastest-code and fastest-algorithm)

Sparr

10:06 PM

@Doorknob the comments in question are regarding clearing the data directory between tournaments. the RPSLV contest specifies that the data directory will get cleared before each tournament. Caveman Duels does not. This is a good thing for my markov bot when I start making it learn from previous matches, but a bad thing for people who want to track behavior just within one tournament.

also, running multiple threads is probably hurting a lot of bots that rely on storing mid-match data, but that's a different issue