The Nineteenth Byte

2:59 AM

same inquiry as last night... anyone avoiding Dogfight because of the java requirement? I'd love to help someone use my wrapper to make a non-java entrant

Martin Büttner

7:26 AM

@VisualMelon 2 so far

Martin Büttner

10:38 AM

where is everyone?

ProgramFOX

@MartinBüttner It's always quiet here in the morning, not?

Martin Büttner

It wasn't yesterday. Generally, yes it doesn't get really busy until noon/early afternoon (UTC), but usually at least there's some posts before ^^

ProgramFOX

Probably most discussion about the domino and dogfight challenge is done now.

Martin Büttner

hehe

I hope the domino discussion starts again when I actually post the challenge ^^

VisualMelon

I'm alive now

ProgramFOX

10:46 AM

Hi @VisualMelon!

VisualMelon

howdy

Martin Büttner

morning!

@ProgramFOX I think he sensed someone saying "Domino"

ProgramFOX

:D

VisualMelon

12:40 PM

everybody loves worst-case high-overhead optimizations

Martin Büttner

1:28 PM

@VisualMelon congrats on the bounty!

VisualMelon

oh, fabulous

I've never won a bounty before

thank you kindly

Martin Büttner

getting bounties on PPCG is pretty tough, I think

VisualMelon

easier than on Stack Overflow, where the mighty Skeet answers anything interesting

Martin Büttner

I only got one so far, and I think that was mostly due to being the first to answer it (and the criterion being upvotes)

@VisualMelon learn regular expressions :D

I even got the regex gold badge before Jon Skeet (by a couple of months or so)

VisualMelon

I've never had much use for regexs

whenever I need to do a ridiculous find and replace I just fumble around until it just about works and then ask someone to fix it

Martin Büttner

1:31 PM

:D

my main use for regex was 25k rep on SO :D

VisualMelon

that sounds like a very good use for regexs

I don't even have a "nice answer" badge yet, but I'm slowly gaining rep for an answer which didn't actually work over on SO

Martin Büttner

haha

Sparr

2:04 PM

@ProgramFOX I hope discussion about dogfight begins anew when we officially open it up to other languages

ProgramFOX

@Sparr Yes, but doing dogfight in other languages seems complicated.

You have a lot of helper classes/functions. If you would want to use them all from your code, you'd have to create an input format for that, which will be quite complicated.

Or re-implement every class in your code, and only send the necessary things as output. But that looks much work.

Martin Büttner

I usually don't like taking votes to the chat, but would anyone like to help reopen this question (and subsequently close its duplicate) before the existing reopen votes expire? codegolf.stackexchange.com/questions/34926/…

as per the votes on this meta post and the lack of votes on Peter's answer, I'd gather there are some people who agree on this: meta.codegolf.stackexchange.com/questions/1875/…

Sparr

@ProgramFOX no more complicated than doing it without helper classes in java, which is about where you stand for any other challenge.

ProgramFOX

2:19 PM

@Sparr Yes, without the Java helper classes it is not that complicated; it just takes more time to write your program.

Geobits

2:37 PM

@MartinBüttner While cleaning up my picture folder I found this from last year:

Martin Büttner

@Geobits nice :D

I'm hoping to present the next screen shot in 5 days :D

Geobits

9999, right?

Or are you planning to go back down to 7777? :D

Martin Büttner

@Geobits I'm gonna downvote A LOT!

yeah 9,999 hopefully ;)

Geobits

Getting downvoted would be quicker, but it would be a lot harder to do that without tripping the serial-vote script. Maybe you just need a ton of sockpuppets.

All with the necessary rep to downvote, of course.

Martin Büttner

hmmm

I think I'll just go for 9,999... seems easier

Geobits

2:49 PM

At the rate you're going, probably.

Martin Büttner

this week was mental so far :D (see user leaderboard for the week^^)... but that's mostly due to the bounty I think

Geobits

So you're saying I shouldn't wait until you're at around 9,980 and upvote a handful of your posts?

Martin Büttner

:D

VisualMelon

I think I'll take a break from trying to think and just have some fun with forest simulation instead

Martin Büttner

@VisualMelon I'm sure you'll be back when the challenge goes live :P

Geobits

3:46 PM

New Meta Post! Can I post a cooperative challenge where there isn't a single winner?

Nathan Merrill

Geobits, I like it

Geobits

Me too, but I may be biased.

Nathan Merrill

Can anybody start a team, or is there a fixed number of teams?

Geobits

For my particular challenge, it would be fixed. However, if a new challenge genre is accepted, I don't see why either couldn't work.

Nathan Merrill

Hence why I asked it here :P

Geobits

3:52 PM

If you look back over chat from yesterday, there's a lot of talk about it :D

Nathan Merrill

bah, working on my own challenge

Geobits

In essence, it's capture-the-flag.

Nathan Merrill

I finally got the map and everything generating properly

New Meta Posts

2

Q: Can I post a cooperative challenge where there isn't a single winner?

I've got an idea for a team based challenge, where entries will choose the team to participate on, and teams will compete against each other. I don't want to get into the details much here, so let's keep it somewhat vague and high-level. If I can't determine a good way to pick a single winning ...

Geobits

Ah, I was wondering if that was going to kick in. Impatience FTW.

Sparr

3:54 PM

I'm developing a habit for answering early on questions that won't get much traffic, and answering late on questions that have had a lot of traffic :(

Nathan Merrill

new image :) i.stack.imgur.com/WUTX1.png

VisualMelon

cool

I once wrote a maze-evolver that ranked mazes on how many cycles it took a BFS to solve it in

Nathan Merrill

...a labyrinth (1-path maze) would get the highest score

VisualMelon

no

the result was always disapointing

Nathan Merrill

then how do you define a cycle?

VisualMelon

4:05 PM

I'm not sure actually, was a while ago, I'll dig it up and check

it appears to be the sum of the number of active nodes each iteration

Sparr

@NathanMerrill can't wait to submit a really crappy pacman :)

Nathan Merrill

oh, that reminds me

I was going to change the rules

so that if you try to walk through a wall, it just stops you

instead of killing you

VisualMelon

a 1-path maze therefore would be the lowest score (it rejects unsolvable mazes)

Nathan Merrill

but an open maze would get the most, I think

VisualMelon

that sounds about right... but it doesn't seem to select for more-open mazes, maybe I've given it a minimum number of walls or something

Sparr

4:13 PM

@NathanMerrill what happens when a square contains one ghost, one invincible pacman, one vulnerable pacman?

@NathanMerrill do dead pacmen ever respawn? the line of sight rules make it possible to be completely surprised by a ghost, with no chance to avoid it.

VisualMelon

it's also a heuristic algorithm rather than an optimal one, so will select for mazes with "less obvious" routes, I think that was what I was trying to achieve, but I failed

Nathan Merrill

@Sparr Invincible lives, vulnerable dies, ghost teleports.

@Sparr I guess I should include a 45 degree ray as well

Sparr

that ray would need to be able to see through walls/corners if there's a path around the wall/corner

naively, it would just hit the corner

Nathan Merrill

each square in the code is only 1 square big

I'm not doing physics ray casting

Sparr

sorry, what I mean is...

let's say I'm facing west down a hallway

and the square ahead of me has a south exit

and there's a ghost one square southwest of me

if I move west and the ghost moves north, it gets me

but I can't see it

Martin Büttner

4:25 PM

@Geobits 10 minute refresh interval ;)

Sparr

@Geobits I deleted my off topic comment

Martin Büttner

ah the disappointment when I see a green +10 appear in the status bar, but then it's just some random old SO answer

2

Nathan Merrill

I understand Sparr

if you are facing west, you are going to get W, SW, S, NW, and N squares

VisualMelon

@MartinBüttner I live for that moment

Nathan Merrill

So, if you are going to head south, and the ghost is south east of you, and he moves east, then he gets you

Sparr

4:34 PM

@NathanMerrill one square W SW S NW N, and line of sight W S N?

hmm, you're right, that scheme still lets ghosts catch you when you turn around a corner that is a U-turn

or even one that's a zig-zag

pacmen would be incentivized to prioritize T intersections of hallways

Geobits

@Sparr Cool. I was really just trying to stay on topic there. Hopefully I didn't come off as an ass or anything :)

Nathan Merrill

Ah, 1 square would be fine I think

Sparr

any scheme that doesn't give pacs line-of-sight around corners leaves some cases where a ghost can surprise a pac

making an in-place U-turn will always be dangerous, so making an around-the-corner U-turn being dangerous seems appropriate

but turning into a zig-zag hallway that's strictly ahead of you shouldn't be dangerous

Nathan Merrill

if you want, however, it to be safe

you can just stay there for a turn

but then you miss out on a pellet

and you have to assuume there's no ghost right behind yo

assume*

Sparr

staying for a turn doesn't make you any more safe, though

whichever turn you turn on, a ghost could be there

Nathan Merrill

4:40 PM

no, but you can then see all directions

its there in the description somewhere

Sparr

if Stay lets you see behind you, that's cool

but that doesn't solve the problem of turning towards a surprise ghost

| |G|
|   |
|P|G|

I move north, the ghosts sit still (which they wouldn't, but for the sake of compact illustration...), and I can't see them

if I then move east, I'll probably get eaten, without ever having seen a ghost that is ostensibly in front of me

the one "behind" me when I turn, the south one, might be a reasonable risk

but the one "ahead" of me when I turn, the north one, seems unfair and unreasonable

Nathan Merrill

if you move north, then you will see the one NW of you

you will not see the one SW of you

I updated the description to say so

if you move north, then stay

you will see both the one NW and SW of you

Sparr

ahh, nifty

(east, not west)

but only because there's a gap in the wall between us?

Nathan Merrill

yes

Sparr

ok, so the only risky behaviors are: standing still with a ghost right behind you, stepping backwards with a ghost two squares behind you, turning in a direction where there could be a ghost one square in the direction you turned from.
?

Nathan Merrill

4:49 PM

and stepping backwards with a ghost 1 square behind you

no swapping ghosts

Sparr

oh, yeah?

good to know

Martin Büttner

@Geobits I think it's time we thought about decent tag names for your new genre ;)

Sparr

so, in this situation, showing the moves we just made, if I move east then I'm going to get eaten with no warning, right?

err

|P  |
|^|G|
| |^|

but if I stand still, the ghost will move north and I'll see him. and I'd see him even if he was one space further back than illustrated?

Nathan Merrill

If you stayed, and he was where is arrow is, you wouldn't see him

if you stayed and he was where is drawn, you would see him

check the edit: meta.codegolf.stackexchange.com/questions/1847/…

Sparr

5:41 PM

edit looks great

in your example below that, about 23,70, why dont you see the square at 21,70 with the power pellet?

Nathan Merrill

You're right

Sparr

also, for consistency, shouldn't your current square say P?

Nathan Merrill

Oh, I never thought about that

actually...no

it'll say X

that way, you know if there is a pacman under you

Sparr

but there's a pacman in your current square...

ahh, ok

that's fair

Nathan Merrill

I'll put an X there

Sparr

5:48 PM

can I distinguish between the ghosts other than by watching their behavior?

Nathan Merrill

no

I've debated about that

I want the IO to be simple

hence, only 1 character for the ghosts

Sparr

ANSI escape codes to color the character! :)

take THAT, "plain text"

Geobits

6:07 PM

@MartinBüttner Any suggestions? I guess it needs to be broad enough to permit more than just two-team contests.

Martin Büttner

6:30 PM

@Geobits Yes I already thought of that... no good idea yet, but I'll let you know... something "metaphorical" like rosetta-stone would be neat.

I guess "symbolic" is the better word

I never understood the difference between metaphor, simile, allegory and symbol in literature classes...

Geobits

To be honest, unless you're a writer or trying to be pedantic online, that's not necessary knowledge.

Sparr

a metaphor is like a simile. except this one, which is a simile, but that's metaphorical.

Martin Büttner

@Sparr I see now

Sparr

:)

Geobits

Were you blind as a bat before?

Martin Büttner

6:33 PM

probably :P

Sparr

so, I have a strategy for a challenge, but I don't have the time to implement it right now, and I don't expect to be able to choose the best language to implement it.

is there any appropriate place/method for describing the strategy?

Geobits

Probably chat. I can't think of anywhere it would fit on main or meta.

Sparr

well, it's for yours :p

Geobits

But it's not exactly permanent then.

My initial thought was to create a graph with 4*spots, one for each facing. a function for safety-check, and traverse from start to finish. What's yours?

Sparr

for blink, what I want to do is build a graph with four nodes for each space on the map, one for each direction you might face. each node is linked to the neighbors with the same facing direction, and to the same space with other facing directions

heh, yes, that :)

Geobits

6:36 PM

Nice.

Sparr

my idea for safety is not to check per-space, but to loop per-angel and delete nodes in the graph for any unsafe space+face combos

Geobits

Ah, good point. That would save time, but for byte count I'm not sure. Depends on how it's done I guess.

Sparr

I can certainly do this with perl Graph::Undirected but it would not be very golfy at all

two days for the caveman duels bounty... if I can figure out my multithreaded file access problem I am certain I can move up from second to first place

Martin Büttner

just read the challenge, sounds interesting

but I think just getting the complete graph is the simplest

Nathan Merrill

7:00 PM

I feel like there should be two loophole posts

one for KOTH, the other for the rest

as nearly none of the answers on the standard loop holes apply to KOTH

Geobits

I think the tag wiki is a good place for that, honestly.

Nathan Merrill

mmm, I agree

Geobits

I think Martin added some common ones to it recently.

Nathan Merrill

Geobits, that is a fantastic new code-golf

Geobits

I thought the overlap in programmers and Whovians might work to my advantage :D

Nathan Merrill

7:08 PM

there's an overlap between any field and Whovians

Martin Büttner

7:27 PM

@Geobits I don't think I added loopholes... more like common practices

turns out I did

Sparr

I'm trying to get my caveman entry to learn from previous matches. I can only run about half the contestants locally, unfortunately

Martin Büttner

but more as a caution to challenge hosts to include these in their restrictions... because some of them might actually be desirable in challenges tailored to those concepts... so they aren't really standard loopholes

Geobits

Yea, I knew I'd seen something like that. It's not exactly an explicit list of banned methods, though.

Assuming the host wanted to include them all, a simple link to the wiki would be fine, in my opinion.

Martin Büttner

yeah exactly, and it's not supposed to be

Geobits

Typing them out gets tedious ;)

Nathan Merrill

7:36 PM

bah

Martin Büttner

another sheep :)

I rewrote that loophole post: meta.codegolf.stackexchange.com/a/1916/8478

as always, improvements always welcome

now to fix my french counting

Nathan Merrill

so, apparently this site's license just expired: code.jquery.com

it's breaking the internet

Martin Büttner

:D

that is amazing

Sparr

because people are hotlinking the jquery.min.js?

and they are linking the https version, for some reason

Nathan Merrill

yeah, the reference it in their html

Geobits

7:38 PM

Most likely. Hilarious.

Sparr

reddit is broken by this??

Geobits

Nah, looks like they link to the ajax.googleapis.com version.

Only 19 question upvotes to go until 4k.

Maybe I should answer something instead.

Nathan Merrill

oh, I just had a great idea...

for another KoTH

give two (maybe more) players a square

er....a grid of squares

and then then can turn each square on or off

and then I play the game of life

after 500 turns, the player with the most on squares wins

Sparr

I've played that game

interactively

it was a windows 95 game, in a package of other games

Nathan Merrill

...yeah, you modify it each turn, right?

this one, you only set the initial state

Martin Büttner

7:49 PM

it's been suggested twice

once in the sandbox and once in chat I think

let me look for it

Sparr

in the one I played, it was modified every turn, toggling a single square, with two colors of squares

Nathan Merrill

yeah, I don't like that one as much

Sparr

you're suggesting my program just output a grid that will have the most on-squares after 500 turns?

that seems like a solvable problem

Martin Büttner

here is the first one: meta.codegolf.stackexchange.com/a/1332/8478

Geobits

How do you split up the initial grid? I assume both players can't start with the same square.

Martin Büttner

7:51 PM

here is the other one: chat.stackexchange.com/transcript/message/16164270#16164270

Nathan Merrill

they start with a 10 by 10 grid of square

Martin Büttner

I'm becoming the Peter Taylor of challenges that have never been posted :D

Nathan Merrill

I also wouldn't distinguish between owners of squares

I don't like the network protocol

Sparr

@NathanMerrill I don't think we fully understand what you're proposing. How do the two players interact with each other?

Martin Büttner

@NathanMerrill nope, it just doesn't work out unless you have thousands of players

Nathan Merrill

7:53 PM

why?

you place the two squares next to each other

...the interact

they*

I think 1 on 1 would be the best

Geobits

But if you don't distinguish, is it just the count of live squares in their initial home area?

Nathan Merrill

yes

Geobits

Oh. Interesting.

Sparr

so the actual game grid is 20x10?

Nathan Merrill

2n*n

but yeah

Sparr

7:55 PM

do we play more than once?

Nathan Merrill

I think 1 on 1 against each player

Sparr

what decision making does the program do, if only playing 1 match against each enemy?

Nathan Merrill

it returns a square

that's it

hmmm...

it's not really a program then

Sparr

why would I submit a program with any logic, rather than a hard-coded "best chance" square?

Nathan Merrill

no, I was thinking you just submit your best chance

Sparr

7:56 PM

oh

Geobits

If you had the chance to flip on/off a single square each turn....

Nathan Merrill

but that isn't really a KoTH

Sparr

each new entry could just brute force the best answer to every previous entry

Nathan Merrill

actually, I like that idea Geobits

Sparr

lol

Nathan Merrill

7:58 PM

anyways, its an idea, but I need to finish my current KoTH

Sparr

idea: each player has a side of the grid. you can turn on any cells that you want in the 1/4 closest to your side, but your score is the live cells in the 1/2 closest to the other side

Martin Büttner

btw, while we're talking about KotH ideas (totally disregarding the fact that I already have two sandboxed), I've got the idea of designing a KotH around genetic algorithms, but it's gonna take a while until it happens... dibs.

Geobits

Wait, so you're calling dibs on the entire field of GA? ;)

Martin Büttner

only in the context of Koth :D

Nathan Merrill

dibs on all KoTH around battle

Sparr

8:00 PM

speaking of genetic algorithms... Google bought a company a while back that had some fun non-koth programming challenges as part of their application process. I really want to get them to release one of them so I can post it to a website like SPOJ or codegolf.

Martin Büttner

@NathanMerrill ummmm

Nathan Merrill

hey, if we're playing dibs...I want as much as possible

Geobits

Well hell, dibs on any teamwork based challenge.

Martin Büttner

nah, obviously if someone goes ahead with an original idea and makes a good GA KotH from it that's my fault for being too slow ;) ... it would only be annoying if that didn't turn out the way I would have liked it but too close for mine not to be a duplicate :D

@Geobits ever?

Geobits

If I'm feeling generous I may open it to the public domain one day :D

Martin Büttner

8:03 PM

that's nice, I guess

Geobits

Once I squeeze all the sweet sweet rep from it.

Sparr

maybe I should try save the last bullet

Sparr

8:17 PM

:/ at Coward hard coding the names of other bots that it knows won't shoot it

Geobits

@Sparr My Rifter for the Battlebots koth actually called other programs and gathered their output to see what they'd do on their next turn. I believe that would be frowned upon today, but it wasn't in the rules...

Sparr

it's explicitly forbidden in most KOTH lately, I think

someone tried to do it in another recent koth, but encountered a sandbox in the way

Geobits

I also had certain bots hardcoded to mimic, and could update it to whatever was leading.

Ultimately, though, it didn't win. Out of three bots I entered, it performed the worst in final standings.

Sparr

that's... weird

if you CAN predict all of your opponents' moves, how would you lose?

Geobits

Yea, it was performing better in my tests, but there were a few bots I wasn't running locally.

Also, some of them were random (or semi), and it was easily possible to get into a stalemate position.

It did get the most upvotes and a +100 bounty though :D

I could probably "fix" it now to do better, but I don't think there will be another run.

Sparr

8:25 PM

ahh

I've been part of some battle-bot-style communities in the past. It seems weird to me to see a single short competition for that kind of thing.

I'm used to seeing strategy and meta-game and engine-quirk finding over the course of years

Geobits

Most of the "old generation" koth's were somewhat time limited. It's only recently I think that's changed here.

Even without an explicit limit, people just lose interest in time due to * shiny! *

Sparr

oh, sure

very different environment

hundreds of people playing one game for years

vs people playing a new contest every day or week

like chess players or go players compared to board gamers :)

Nathan Merrill

<3 board games

Martin Büttner

<3 go

(other board games are nice, too, though)

Nathan Merrill

I'm reading the KoTH prisoner's dilemma

I find it fascinating that tit-for-tat did so well

Martin Büttner

9:04 PM

2048 is driving me nuts :D

it was easier when I couldn't control the spawning and chose every move by gutfeel :D

Geobits

@MartinBüttner I sincerely hope you mean the common definition of gut feel, and not this one.

Martin Büttner

:D

I do

(the common one)

I'd just give up on this, but I reaally want that revival (and maybe necromancer) badge :D

Geobits

I don't have those here either! :o

Martin Büttner

I've got two eligible answers, but no one cared to upvote them :D

well two for revival, one of them for necromancer

Geobits

I have 7 Revival on SO, but that's a bit different. I don't think I have a Necromancer anywhere.

Martin Büttner

9:18 PM

I've got two necromancers on meta.PCCG :D

Sparr

I don't keep track of badges, and I definitely don't go seeking them.

I wonder how much actual benefit the site gets from people seeking badges

Geobits

I don't either, really. Hats on the other hand, that's a different story :D

Sparr

hats?

Martin Büttner

on SO I've got one revival badge (it was one of my first answers ever in september 2013, if not the first one)... but it only received upvote number 2 this april

@Sparr this winter game... basically the same thing, but it only happens in december

Sparr

same thing as what? winter game?

Martin Büttner

9:21 PM

regarding badges... I used not to care about them, but then I set out to take over my SO account in badges (with my PPCG account)... so I started looking through the badges list... and before you know it I'm hunting badges for some reason. :D

@Sparr hats <- same thing as -> badges

Geobits

Blog - Stack Exchange: Winter Bash 2013 is here!

Abby T. Miller on December 15, 2013

The gatherings of family and friends, the giving and receiving of gifts, the making and/or breaking of New Year’s resolutions – however you and yours celebrate, the end of a calendar year heralds many traditions.

Here at Stack Exchange, we wanted to get each of you an awesome, personal gift, and mail it to you as our way of saying “thanks.” But our accountant pointed out that there are 4.5 million of you, which promptly reminded us that the holidays aren’t about gifts. The real spirit of the holidays can only be captured with… …

Martin Büttner

you get them for random (partly secret) tasks and can adorn your profile picture with them

Geobits

Completely temporary and meaningless. Which of course makes it more fun.

Sparr

oh, I remember those

I think I might have gotten a couple

Geobits

Basically all my rep on MSE is from that time span. Go figure.

I also won an SO T-shirt and mug in the haiku contest ^^

Sparr

9:24 PM

pardon me while I go try out the bracket numbers challenge

that one appeals to me greatly

Martin Büttner

yeah it looks pretty nice

@Geobits what's on the shirt?

Geobits

Just the SO logo. Like this (the black one): meta.stackexchange.com/q/38018/212780

And the mug: meta.stackexchange.com/q/208393/212780

Martin Büttner

ah I've the white one with the stackexchange logo

I don't have that mug though unfortunately ^^

Geobits

I have two :)

I got the other one for this nonsense: meta.stackexchange.com/a/212477/212780

Martin Büttner

I barely ever go on MSE

Geobits

9:31 PM

I did for a short period, not so much now.

Unless they're doing some sort of contest/event, there's really not much besides literally the same posts over and over again. "Why was my post closed/deleted/downvoted", "Allow low-rep users to comment", etc.

Nathan Merrill

No, its all "What's up with MSE and MSO" right now

Geobits

10:38 PM

I can imagine. I haven't been there much at all since a bit before the split.

Sparr

11:30 PM

ok, I think I've got a partial solution for bracket numbers

not perfect, but close, for 1-256

77MB of text for a solution for the bigger numbers... that's probably not optimal

11kB seems a lot more palatable :)

Sparr