« first day (1793 days earlier)      last day (3050 days later) » 

8:01 PM
@El'endiaStarman Uhhh the bots aren't loading in the spacewar simulation
 
Just tried a random pair. Looks like they do.
 
o_O it's working now
blame caching
 
Hello=)
 
8:07 PM
I think Starman isn't your real name, is it?=)
 
Oh no, it's not. :P
 
Though, on the Internet, it basically is.
 
user image
3
 
0
A: Sandbox for Proposed Challenges

wizzwizz4Parse a tree for pruning code-golf ascii-art Windows supplies users with tree: a neat little tool that converts a directory tree to ASCII-CP437 art. It's a very human-readable format. However, it's a bit useless if you want the directory tree of an entire drive, or even just the Windows folder,...

 
8:08 PM
Does anyone else have the same font problem?
 
Why not Galaxyman??
=P
 
With what? The strikethrough? The badge positioning? The random space in "Do ng"'s name?
 
@El'endiaStarman I think he prefers Downgoat, so that.
 
@NewSandboxedPosts What do you think of Parse a tree for pruning?
No, I wasn't talking to a bot...
 
Yes, you are ;)
 
8:11 PM
@Dennis Oh, I didn't make the connection. You're not the only one with that problem.
58 mins ago, by Sherlock9
Interestingly, when I type your name on mobile, it drops some characters. So I get CNRO'BRIN
 
sᴏ ʏᴏᴜ ᴍᴏʙɪʟᴇ ʜɪᴘsᴛᴇʀs ᴄᴀɴɴᴏᴛ sᴇᴇ ᴛʜɪs ᴀᴡᴇsᴏᴍᴇ ᴛᴇxᴛ?
 
@CᴏɴᴏʀO'Bʀɪᴇɴ How do you output something in Jolf?
I got the code vH*99H
 
@Adnan It's implicit, at the end, but a also does that.
@Adnan repeating H 99 times? what are you doing? XD
 
Wat hahaha
 
8:14 PM
0
A: Sandbox for Proposed Challenges

wizzwizz4Parse a tree for pruning code-golf ascii-art Windows supplies users with tree: a neat little tool that converts a directory tree to ASCII-CP437 art. It's a very human-readable format. However, it's a bit useless if you want the directory tree of an entire drive, or even just the Windows folder,...

 
Huh
How do you assign 9 * 9 to H?
 
oh yeah that's it :P
I forgot I changed that
 
@wizzwizz4 If nobody has any comments about my challenge, I'll just post it as a question.
 
oH*99 or vH*99
@wizzwizz4 ಠ_ಠ
how long have you waited?
 
@wizzwizz4 Best to leave it in the sandbox for a while
 
8:16 PM
I know. Feigning impatience seems to get people's attention though! ;-p
 
ಠ_ಠ
3
 
@CᴏɴᴏʀO'Bʀɪᴇɴ How about this
 
@Adnan Ah, I forgot. v can only be used after defined. Use o.
this should work
 
Ahhh, it works :)
Thanks!
 
8:30 PM
@wizzwizz4 You better appreciate my new profile description.
 
@ಠ_ಠ Illuminati confirmed??
 
I did the math
 
I was referring to this.
 
I know :P
 
hmm...
 
code bowling measured by bytes after some compression method
does that remove the majority of problems code bowling has?
 
Good idea.
 
@NathanMerrill I think we should test it in a simple challenge=)
I really like this idea too.
 
what's a good task that hasn't been overused?
I'll post a sandbox
oooh, a bowling simulator
is .7z accessible to linux users?
by accessible I mean typical/ease of use
 
@Dennis I'm a dong :O
3
 
8:59 PM
@NathanMerrill yes
On Debian it's just apt-get install p7zip
 
Binaries for everyone!
 
9:12 PM
I see no reason compressed code bowling would be an improvement. "HAY I NEED THE VALUE OF PI HERE SO LET ME JUST WRITE IT TO 10000 DECIMAL PLACES BTW PI IS INCOMPRESSIBLE"
 
@quintopia fair point
if there was a language restriction, then limiting the number of statements might work
except not
hmmm
 
booo language restrictions
 
not planning on doing the challenge any more
quintopia made a good point
 
Using the letter O in place of a 0 was extremely convincing.
 
still doing a bowling themed challenge though :P
 
9:22 PM
This is an interesting algorithm
Not superfast, but definitely easy to implement in heavily recursive languages
 
On the topic of compression, this is probably going live soon, unless any major complaints arise in the next two days or so.
 
Related algorithm: bit.ly/1e1EYJv
 
ಠ_ಠ
 
Why should pi be incompressible?
 
@flawr There's no known repetition? Idk
 
9:28 PM
^
 
oooh, what if a cops and robbers style challenge where the cops write a program, and robbers write another program to print the cops' program
 
@AlexA. So? Still doesn't make it incompressible. Or more random.
 
Hence "idk"
 
the cops try to maximize the minimal generating program
 
Because idk
 
9:29 PM
and the robbers try to minimize it
 
Of course pi is compressible
see: pi-calculation algorithms
 
pi = 22/7
 
@Doorknob冰 its not compressible by generic compression algorithms currently available. One has to know one is trying to output pi in order to compress it.
 
pi = 1 in base pi?=)
 
@quintopia [citation-needed]
 
9:38 PM
@mınxomaτ Well, my intuition tells me "you can't take a number and unambiguously turn it into a smaller number reversibly"
 
@mınxomaτ blahblah assuming pi is normal which everyone believes but no one can prove blahblah
 
@Doorknob冰 Yeah, no compression in unambiguous. But that doesn't make things incompressible.
 
@mınxomaτ By "unambiguous" I mean "only one possible output per input"
 
@Doorknob冰 I know.
 
and only one possible input per output
 
9:40 PM
How about lossy-text compression?
Lossy pi-compression?=P
 
Well, for instance, the real number given in binary as 0.1101001000100001000001000000100000001000000001... never repeats but is definitely compressible. No one believes pi is like this....
 
> No one believes pi is like this
Yeah, of course it is. If it is infinite, it can be represented by an infinite number of bytes (i.e.) bits. There's no indication in the (few) first digits that pi has a serial correlation coefficient of 0, actually it's the opposite. Truly incompressible files have a SCC of 0. It's fairly easy to create them, one is included as a sanity check in my compression challenge (link above).
The only reference to your statement is your own discussion at xkcd forums and your friend.
 
Pi =
11.00100100001111110110101010001000100001011010001100001000110100110001001100011‌​001100010100010111000000011011100000111001101000100101001000000100100111000001000‌​100010100110011111001100011101000000001000001011101111101010011000111011000100111‌​00110110010001001010001010010100...
 
9:56 PM
Hello
 
@quartata hi
 
@NathanMerrill Could do something like
I actually think measuring tokens might be a good fit for code bowling
That way having 10,000 digits of pi doesn't matter because it is still one token
 
I'm thinking of doing something similar
 
Part of the problem is that we haven't actually ever had a good either...
 
er, not really
I"m almost done with my sandbox post
then you guys can vote it down
 
10:02 PM
You know, the votes wrap around at -2^31
 
hehe
 
lol
0
A: Sandbox for Proposed Challenges

Nathan MerrillCompressed Code Bowling Cops You need to write some code that takes an array of integers and prints out the sum of the integers. Robbers are going to produce a program that generates your code. You win by making the robber's code as big as possible. Robbers Your program needs to output the ...

posted
 
Mkay.
I might try a hand at an here.
It seems like a decent idea.
 
That sounds....destructive.
 
10:05 PM
The only other way I can think of making code bowling work is a :/
 
Bowling is not challenging without some sort of restriction. There might still be some kind of hybrid bowling that could work.
@NathanMerrill This does seem like a good way to counter the tendency to introduce useless code. I'm interested to see how it turns out (+1)
 
@trichoplax thanks :P
I'm looking through past code bowling efforts, and most of them won't work on my sandboxed post
bah @MartinBüttner
 
@trichoplax I don't see how it avoids the problems of the last .
 
@NathanMerrill I like the idea of being able to reduce whitespace and variables, but it might still be hard to define what changes are acceptable. At the extreme, I know you don't want robbers that just write code that outputs valid but unrelated source code, but I don't know how a line could be drawn clearly.
 
10:12 PM
1
A: Sandbox for Proposed Challenges

Nathan MerrillCompressed Code Bowling Cops You need to write some code that takes an array of integers and prints out the sum of the integers. Robbers are going to produce a program that generates your code. You win if the smallest generating code is bigger than any other cops' submission. Robbers Your p...

 
@NewSandboxedPosts is the weather nice?
 
@MartinBüttner I'm really not up to date - I'll go and have a look...
 
@trichoplax yeah, that's definitely the hardest part of writing this challenge
 
@trichoplax it's been a while. essentially you encode your code in a unary number and get that number from a string's length.
 
@MartinBüttner would disallowing "eval" be reasonable?
it seems that much of the problems of code bowling has to do with the eval statement
 
10:15 PM
@NathanMerrill that's cool, I'll just write a Unary interpreter.
(and write a list-summing program in Brainfuck/Unary to embed in my code)
 
Hmm... how should I define a token exactly
Language dependent :/
 
@MartinBüttner Under the sandbox post's rules, it seems you could output another string of the same length, and it would count as a valid robber. Which means complicating the string won't help the cop. Does that improve anything?
 
@quartata the problem of every atomic code golf ever.
 
@NathanMerrill Then again, what exactly defines "eval"?
 
@trichoplax I can probably include some checksum which breaks the program if the actual characters change.
 
10:19 PM
@flawr take this string and execute it
 
@MartinBüttner Oh yeah...
 
@MartinBüttner but unary is easy for a program to produce
and short
wait, can unary be any character?
 
@NathanMerrill well, I meant Lenguage, not Unary. the point was not the specific language, but that instead of using eval I can just interpret the code myself, by interpreting a different language.
 
@NathanMerrill Sure, you can use any symbol you like to represent a digit
 
I'm defining "eval" as taking a string and executing it (don't care about the language the string is in)
 
10:21 PM
ummm
if I write that executing code myself, I really don't see how you're going to find an objective definition for what qualifies as "eval"
 
As an alternative to code golf, there are challenges with an upper limit on code length, and you have to perform as much as possible without going over that limit. So an alternative to code bowling would be a challenge with a lower limit on code length, and you have to perform as little as possible without going under that limit. This seems to confirm that code bowling is almost impossible to make an interesting question for...
 
@quartata not in the language Unary
 
@MartinBüttner Oh, I didn't know he meant uppercase Unary :P
@NathanMerrill Technically blocks are strings in GolfScript, meaning that evaluating a block (say for a loop) falls under your category for eval
It's not quite a bullet-proof definition sadly
 
I'm ok with a broad definition of "eval" that would of-necessity exclude languages (such as Golfscript)
 
@trichoplax I like that idea too.
 
10:26 PM
@trichoplax define "as little as possible"
 
@NathanMerrill I'm not sure I understand your definition of eval, but it seems to contradict Rice's theorem.
 
Maybe there could be a challenge to produce code of a fixed (question-defined) length, and your score is the minimum number of character edits required to make it run without error.
Has that been done?
 
@trichoplax are you trying to maximize the minimum?
 
@trichoplax seems like I can choose Unary (the language), and just use N non-0 characters
 
@NathanMerrill Yes - make the code as difficult to fix as possible
 
10:28 PM
(or any other language, which errors on unknown characters and use N of those unknown characters)
 
@MartinBüttner Hmmm. Yes.
 
@trichoplax I'd just post a series of random characters and say the language is Java
 
@NathanMerrill that is not guaranteed to get a score of N ;)
 
@NathanMerrill Requires changing each character at least N times before errors cease. Score: N^2
 
0
A: Sandbox for Proposed Challenges

I Can Haz HatsIdentifying spies in "Resistance" This is still very rough, but I think it has the makings of an interesting challenge. Please help me improve it. What do I need to add? What do I need clarify? Resistance is a party game that pits "resistance members" against "imperial spies" on a series of mis...

 
10:33 PM
@MartinBüttner What if the original code has to give no compile time errors, only run time errors?
 
I think that the only way I could get to work is if the answers had to be in a certain language
 
(or rough equivalent for interpreted languages)
 
@trichoplax I'm not sure that's a well-defined distinction for interpreted languages
 
I just can't think of a good general definition of token here.
 
ah yes, "rough equivalent" :P
 
10:34 PM
@MartinBüttner Yes that would be the difficulty :)
But would it even work for compiled languages?
 
@quartata the only way ever works is by defining simple toy-languages for the challenge.
 
I don't mind excluding a subset of languages
 
What are tokens in, say, Befunge?
 
@El'endiaStarman Individual non-whitespace chars.
 
10:36 PM
why non-whitespace? clearly, the whitespace is vital for the program to work?
 
They are just empty spaces on the grid.
 
I can't remove them without breaking the program though
 
Some of it is necessary, though.
 
especially the linefeeds
but neither the spaces
 
Trailing spaces can usually be removed.
 
10:37 PM
Right, but removing the whitespace results in the non-whitespace chars being in a different location on the grid which is why it breaks
 
yeah, so the whitespace is definitely significant for the program to work
 
I don't consider the whitespace to be tokens, they just make up more the shape of the grid
 
so I don't see why it shouldn't be counted
 
The interpreter itself doesn't do anything with the whitespace specifically
 
anyway, atomic code golf is also susceptible to encoding the actual code in a single string and eval-ing it
 
10:38 PM
This is code bowling in this case.
So that would be somewhat detrimental
 
ah right, what's stopping me from blowing up the code here?
 
With a good definition of tokens, I don't see a way to arbitrarily pad code. Whitespace doesn't count, comments don't count...
 
there is an infinite number of programs (ignoring whitespace and comments) which solve any given problem. clearly, these must be arbitrarily large.
 
Right, but it actually takes effort to make them.
The issue with code bowling was that usually it was too easy to pad programs
 
@quartata There are obfuscators. They blow up the code massively while having no dead code.
 
10:40 PM
but then it's just an exercise in patience of adding another pointless indirection to the algorithm
it is still way too easy to pad programs
if I need a numeric literal I'll just replace it with an arithmetic expression with 5000 operands.
 
Hm
I suppose that depends on the task.
I was planning on making the task a primality test
 
Maybe someone needs to design a bowling language that doesn't have these problems.
 
I'm not sure what the task has to do with it.
 
No literals are involved, so that eliminates a place for padding
 
@trichoplax by not being Turing-complete? ;)
 
10:43 PM
But I guess you could just calculate a literal and not use it or something
 
@MartinBüttner At least :)
 
@quartata how are there no literals involved? I'm sure I'll need a for-loop somewhere...
 
0
Q: Not Truely False

NO_BOOT_DEVICEThe goal of the program is to break your language so that the check if false: passes. You may not fake us out by setting something that looks like the variable false then checking that. You must actually override a False check. Bonus points: Make if false and not true pass. Bonus bonus points: M...

 
@MartinBüttner Oh, you mean something like for(int i = (2-2)+(2-2)+(2-2)+...
 
10:45 PM
Is not banned?
 
(I haven't solved it yet, so no spoilers!)
 
@trichoplax unfortunately not
@quartata for instance...
 
@trichoplax I know, I still think it is too much like
luckily we don't see too much of it
 
@Doorknob冰: Look a few messages up for a chess puzzle.
 
@quartata Can we just make it a synonym and skip the meta discussion?
 
10:46 PM
@El'endiaStarman I think I've seen that exact one before, heh
 
@trichoplax pffft
 
@Doorknob冰 ah, lol
 
Sadly I don't think we can since the tag was already made
@MartinBüttner All right, forget it. Too much effort for something that won't even be that fun in the end...
 
@quartata we could.
 
Really?
 
10:47 PM
there's a distinction between the two though
code trolling was "make something that doesn't do what the question asks"
underhanded is usually "make something that appears to be doing x, but does y instead"
it's a lot more specific, but usually not much better, with very few exceptions
 
I think we need more reeeeeeeally hard fgitw challenges
 
What's a fastest-gun-in-the-west challenge?
 
a really easy one?
 
a popularity contest?
 
I'm not sure, but I think it's like this one
 
10:59 PM
@quintopia is it code-golf-is-boring-let's-resurrect-a-challenge-type-that-was-never-a-good-idea day? :)
5
 
29
Q: Catalan Numbers

quartataThe Catalan numbers (OEIS) are a sequence of natural numbers often appearing in combinatorics. The nth Catalan number is the number of Dyck words (balanced strings of parenthesis or brackets such as [[][]]; formally defined as a string using two characters a and b such that any substring starti...

Since you need to be quick
 
That's totally oxymoronic combined with "reeeeeeeally hard" though.
 
we used to have a meta post that asked for ideas for new challenge types (instead of a meta post proposing a single new challenge type)... where did that go...
 
challenge idea: Take my random string of characters, insert characters into it (any location/order you like) to make a program that does X
 
that doesn't sound too bad
although it will be hard to strike the balance between string length and task difficulty
 
11:07 PM
last panel.
 
lol
 
challenge-idea: Make a program that pokes Randall with a stick if he's procrastinating a(nother) What-If? release.
 
19
Q: Let's create some new types of challenges!

Doorknob 冰Our site needs unique content. The more different and new a question is, the more interesting it will probably be. Therefore, let's propose new genres of challenges! What exactly is a challenge genre? It's one of these. The most commonly used ones are code-golf, code-challenge, and popularity-co...

@MartinBüttner This one?
 
how did I not find this when searching for "new challenge types"
anyway, this is the answer I was looking for and what I thought quintopia was talking about:
-7
A: Let's create some new types of challenges!

NathanielFirst Past The Post To be used for challenges that are genuinely difficult, and/or require solving some kind of puzzle that only has one answer. The winner is simply the first person to post a correct solution.

 
@MartinBüttner My round about search method was: search for "challenge is:question" and then sort by votes and read through until I find it...
(because I vaguely remembered it being highly voted...)
 
11:20 PM
@MartinBüttner exactly. What's so bad about a problem solving challenge so difficult that being able to solve it at all is an accomplishment, but once someone has shown how to do it, golfing it is trivial?
 
@quintopia see my comments on the answer I just linked
 
I like seeing code challenges that are difficult, but I don't see that this should make subsequent answers redundant. Even if the question is a deliberate puzzle with an approach in mind, someone may still find something better (unless like me, you post a question which someone proves an optimal solution to).
 
@MartinBüttner how about multihole challenges with hidden solutions
 
I'm not sure what you mean, but it reminds me that I wanted to write up a meta post about multihole challenges...
 
1
Q: The Abelian Orders

Fax MachineSome background In math, a group is a tuple (G, •) where G is a set and • is an operation on G such that for any two elements x and y in G, x • y is also in G. For some x, y, z in G, basic group axioms are as follows: G is closed under •, i.e. x • y in G The operation • is associative, i.e. x...

 
11:38 PM
Oh it worked
If you edit the feeds post it updates
 
Yes...but I don't see an edit icon?
 
good night guys
 
G'night!
 
Good night!
@El'endiaStarman Downward arrow on the left side of the post when you hover over it, click, "edit" in the bottom left. It's not there for you?
 
@AlexA. I mean, I know that if you edit a feed post, it re-oneboxes. But I don't see the pencil edit icon on the most recent one.
 
11:42 PM
Oh really?
It shows up for me
 
I also don't see a history link.
[refreshes] There it is.
Strange. That's not usually a problem.
 

« first day (1793 days earlier)      last day (3050 days later) »