# What is the Sandbox?

This "Sandbox" is a place where Code Golf users can get feedback on prospective challenges they wish to post to the main page. This is useful because writing a clear and fully specified challenge on the first try can be difficult. There is a much better chance of your challenge being well received if you post it in the Sandbox first.

See the Sandbox FAQ for more information on how to use the Sandbox.

## Get the Sandbox Viewer to view the sandbox more easily

To add an inline tag to a proposal use shortcut link syntax with a prefix: [tag:king-of-the-hill]

# Time bomb ping pong

## Challenge - Both teams

All users are divided into two teams based on their PPCG ID. For example, my ID can be found here, from which you can see that my ID is 34388. To check on which team you are, run the following snippet:

## Challenge

This is a challenge, all submissions should be written in

Your function should accept four variables as follows, obviously within your function you can name these whatever you like
- 1st variable represents the current value of 1 bar of Megatanium
- 2nd variable represents your current bank balance
- 3rd variable represents your current stock holding
- 4th variable represents the iteration number

The return value should be negative if you wish to sell stock, positive if you wish to buy stock, 0 if you wish to neither buy or sell

For example;
If you wish to buy 10 bars of Megatanium at the current price you would return 10
If you wish to sell 5 bars of Megatanium at the current price you would return -5

I will call your function 1000 times. Stock price will always be an integer, chosen at random, with a minimum of 0 and a maximum of 256. Method of selecting the stock price is described in more detail below, it will NOT be an even distribution!

Your bot will be disqualified if it does any of the following at any point

• Try to buy more stock than it can afford
• Try to sell more stock than it holds
• Try to write any value to any of the global variables
• Fail to return a value
• Return a value that is not an integer

Here is the code I will be running, the score output at the end will be your bot's score. It is calculated from your bank balance plus the value of your held stock at the latest value.

function go(bot) {
bank = 1000;
stock = 0;
disqualified = 0;
for (i=1; i<=1000; i++) {
price = prices[i];
trade = window[bot](price, bank, stock, i);
bank = bank - (price * trade);
if (bank < 0) disqualified = "RUN OUT OF MONEY";
if (stock < 0) disqualified = "TRIED TO SELL STOCK YOU DIDN'T OWN";
if (disqualified) break;
}
if (disqualified) {
console.log("Disqualified on iteration " + i + " REASON: " + disqualified);
}
else {
score = bank + (stock * price);
console.log(bot + " scores " + score);
}
}


The function for generating a suitable distribution of random values is a slightly modified version of the function described here
Every bot will be given the same set of prices, but the set will not be generated until immediately before the bots are run.

function randn_bm() {
var u = 0, v = 0;
while(u === 0) u = Math.random();
while(v === 0) v = Math.random();
w = Math.floor(Math.sqrt( -2.0 * Math.log( u ) ) * Math.cos( 2.0 * Math.PI * v ) * 32) + 128;
while (w < 0 || w > 256) w = randm_bm();
return w;
}
prices = [];
for (i=1; i<=101; i++) {
prices.push(randn_bm());
}


## Completion

You may submit as many bots as you like, try to be inventive with your algorithms! You may use a global variable called data, this will be initially set to null and will always be available to your function.

## Example bots

function buyBot(p,b,s,i) {
/* Bot always buys as much as it can */
return Math.floor(b / p);
}

function randomBot(p,b,s,i) {
/* Bot buys and sells randomly */
if (Math.random() > 0.5) {
return Math.floor(Math.random() * (b / p));
}
else {
return -Math.ceil(Math.random() * s);
}
}

function smartBot(p,b,s,i) {
/* Bot always buys at under 100 and sells at over 150 */
if (p < 100) {
return Math.floor(b / p);
}
else if (p > 150) {
return -s;
}
else {
return 0;
}
}

function bankruptBot(p,b,s,i) {
/* Bot always sells at under 100 and buys at over 150 */
if (p > 150) {
return Math.floor(b / p);
}
else if (p < 100) {
return -s;
}
else {
return 0;
}
}

function alternateBot(p,b,s,i) {
/* Bot buys and sells alternately */
if (data == 1) {
data = 0;
return Math.floor(Math.random() * (b / p));
}
else {
data = 1;
return -Math.ceil(Math.random() * s);
}
}


## Winning conditions

All bots will be run on locally by me approximately 1 week after the question is posted (date will be decided when question is posted, no point setting a date in sandbox)
Winner will quite simply be the bot that has the highest final score after the last iteration, as calculated by the function provided above. The array of prices used will be published after a winner has been crowned.
There are no set conditions on the speed of your function, but please be fair and try to avoid anything that will take more than a couple of minutes to execute

Note: Please leave an upvote if you think the challenge idea is good and is clear, a downvote if you think the challenge idea is bad, and comment if you can't understand the challenge. Last time I asked on TNB 2 users told me that they can't understand anything.

# Golf a return-oriented code generator!

## Background

Return-oriented programming (ROP) is a computer security exploit technique that allows an attacker to execute code in the presence of security defenses such as non-executable memory (W xor X technique) and code signing. (from Wikipedia)

## Challenge

In this challenge, you should write a program, that takes the code of the existing program and the required code, and output the stack necessary to execute that program.

## Rules

• Standard loopholes apply, as usual.
• How the machine works
• At first, IP is equal to the top of the stack, and the top of the stack is popped.
• For each clock cycle (whatever it means), the command at the position of IP is executed, and the IP is advanced if the command does not modify IP.
• The behavior if the IP is at the last instruction and that instruction does not modify IP is undefined.
• Assembly instructions
• All commands are case insensitive.
• Note: There is nothing that guarantees commands must be 3 characters long, or limited to some sets. After all, this is not real assembly. However:
• You can assume that all the characters are in the English alphabet, uppercase or lowercase.
• There may be some other commands "similar but not the same" with ret, so checking for the first character or the SHA256 hash won't work.
• The special command ret will pop the value on the top of the stack, and set the instruction pointer IP to that value.
• You may assume that all the other commands won't modify the stack, or the IP.
• The "existing program" and "required code" will be represented as a string, separated by newline characters (you may optionally take a list of strings as input).
• The required code will never contains ret.
• The output should be a stack of line-numbers in appropriate format (list of numbers - may be reversed, array of numbers, etc.)
• The command executed right after the last command in the "required code" must be a ret.
• Because the memory of the machine is limited, you should output the shortest possible stack. If there are multiple shortest stack, output any.
• You may assume that there exists an output.

## Example test cases:

• Existing program:
1:  add eax, ebx
2:  lea eax, [2*eax+4*ecx]
3:  ret
4:  mov eax, ebx
5:  xch ecx, eax
6:  ret

(the line numbers are just for demonstration purposes. They are not included in the input, you can use 0-indexing or 1-indexing)

(disclaimer: this is not real assembly, just for demonstration purposes)

• Required code:
lea eax, [2*eax+4*ecx]
xch ecx, eax
mov eax, ebx
xch ecx, eax

In that case, you should output [2, 5, 4] (2 is at the top of the stack), because if the stack have that value, then:

• First, IP = 2.
• The commands
2:  lea eax, [2*eax+4*ecx]
3:  ret

are executed.

• On executing ret, the IP get the value 5. Then, the commands
5:  xch ecx, eax
6:  ret

are executed.

• Then, similarly, when the next ret is executed, the value 4 is popped from the stack, and commands
4:  mov eax, ebx
5:  xch ecx, eax
6:  ret

are executed.

Therefore, the commands executed (apart from ret) are:

2:  lea eax, [2*eax+4*ecx]
5:  xch ecx, eax
4:  mov eax, ebx
5:  xch ecx, eax


which is equal to the "required code".

## Winning criteria

This is , shortest code in bytes win!

• You should probably include a list of all valid identifiers or if all of them are always 3 characters mention that and whether there are instructions that begin with r other than ret. – ბიმო Jan 28 '18 at 18:57
• @BMO Better now? – user202729 Jan 29 '18 at 12:00
• No idea what the alphabet ØA;Øa¤ is but apart from that I think you made it pretty clear now. Other things that came to mind: What if there are multiple possibilities for a shortest stack (eg. like this)? Will the input always have a solution (eg. unlike this)? And are the line numbers part of the input? If so is it safe to assume that they always start with 1 and have an offset of 1? Maybe you could add some testcases. – ბიმო Jan 29 '18 at 12:52

# Faux Compress a String

Given a string s, perform a faux compression on it. Below is an example with f('hello world').

To faux compress a string, start by taking a frequency count of all letters.

hello world -> [h:1, e:1, l:3, o:2, :1, w:1, r:1, d:1]

Next, sort smallest to largest in count with a tie-breaker of ASCII-code.

[ :1, d:1, e:1, h:1, r:1, w:1, o:2, l:3]

Next, in the original string, replace each letter with it's index in the frequency list.

[3,2,7,7,6,0,5,6,4,7,1]

Next, convert each integer to binary, and join them all with the digit 2, then convert this to base-3.

1121021112111211020210121102100211121 -> 240591504997661290

Lastly, print or return both this number and the sorted keyset of the frequency map.

[240591504997661290, ' dehrwol']

Is the final result.

You now have a "faux compressed" string.

To get a real compressed string you'd figure out the shortest sequence of bits to replace 2 with which would be a unique delimiter and treat the binary digits as they are, bits, instead of bytes.

# More Examples

000000000001 -> [58839486765, '10']
eeEeeeeeeeeeeEEEEEeeeeee -> [16508589985213004629636, 'Ee']


# Rules

• Your code's function may be undefined if the following is not met:
• The string will contain more than one unique character.
• Lowest byte count wins because this is .

# PSA: I have many PENDING PROPOSALS, tell me which to delete.

If you post a comment on any of these challenges with the words "This isn't good in my opinion."

It will be removed.

• The posted one by MDXF. – user202729 Feb 3 '18 at 5:28

# Casinos and Gamblers king-of-the-hillcops-and-robbers

You either play as a casino, or as a gambler. Casinos offer bets, and gamblers choose bets to take. Casinos want to make money, gamblers want to have more money than others. Read that sentence again, it is the core concept here.

1. All casinos simultaneously comes up with a Bet. A Bet consists of three parts:
1. Entry Fee (Positive integer amount a gambler must pay to take)
2. Odds (The chance that a gambler will win between 0 and 1)
3. Reward (Positive integer amount a gambler is given if he wins). This can be any amount, even if it would make the casino go negative.
2. Each gambler optionally chooses a bet to take. They must have enough money to pay for the Entry Fee.
3. We calculate who wins and who doesn't (using a PRNG), and pay out.

At the end of a game, we give points as follows:

1. Casinos receive 1 point for each dollar they have (can be negative)
2. Gamblers receive N^2 points for having more money than N other gamblers.

A tournament consists of many games, and a player's score is their average score across all of the games.

Gamblers and Casinos all have complete information throughout the game.

• Might want to specify that entry fee and reward must be positive (negative/negative games map to positive/positive games but complicate the "can this gambler afford this game" calculation.) – histocrat Jan 31 '18 at 19:34
• I have a hunch that 100 dollars won't give quite enough granularity for casinos to work with. Maybe bump it up to 10,000? – Beefster Feb 2 '18 at 23:09
• Also, how are odds represented? Fraction? Floating point number? Precision? – Beefster Feb 2 '18 at 23:10
• @Beefster Floating point, of course. I think you may be right, but 10,000 seems too high. I think I'll do 1,000 – Nathan Merrill Feb 3 '18 at 0:06

# Busier Beaver cops-and-robbersbusy-beaver

## Cops

You will write two programs in a language of your choice

• A public program: This program must execute in a finite amount of time. So it can not go on indefinitely (although there is no time limit). It also wise if you program it to output a large number of bytes to standard output. This is the what you post in your answer. You must also post the programming language.
• A secret program: This program must output more bytes than the public answer, i.e. it is a busier beaver. It also must not go on indefinitely. It also can not be longer than the public program. You do not post it the answer immediately. After one week, if no robber has cracked the answer, you edit it in.

For these programs, you may assume that numerics types are actually unbounded, i.e. that overflow never occurs. This means that you can store, for example, Graham's number in a variable. It also means that finding the maximum value of storage capacity of a numeric type results in an error.

If no robber has cracked your answer, your score is equal to (length of public program)/(length of secret program), which you are trying to maximize. In case of ties, the longer public program wins.

## Robbers

The cops will post a public program. Your job is to find a busier beaver. That is, you need to find a program that outputs more bytes to standard output than the cops program. It can not go on indefinitely, and can not be longer than the public program (although it can be the same length).

You get points equal to (length of cops public program)/(length of your program) for each post you crack.

• Need to specify that it's the number of bytes output when given no input. – Peter Taylor Feb 16 '18 at 11:38

# Indices of inner join (code-golf)

Given two lists of integers x and y of possibly unequal lengths, return the list of pairs (i, j) such that the ith element of x is equal to the jth element of y.

The output consists of a vector of pairs or a pair of vectors (the zip of the other option, may be useful for languages without pair data structure). The pairs may be given in any order, and using 0-based or 1-based indexing.

In pseudo-code: (this returns two lists of indices, assumes 1-based indexing, and gives lexicographically sorted output)

input vector<int> x, vector<int> y
vector<int> xout, yout;
for (i in 1 to length(x))
for (j in 1 to length(y))
if (x[i] == y[j]) {
add i to end of xout;
add j to end of yout;
}
return xout, yout;


This is code-golf so answers will be scored in bytes with fewer bytes being better.

### Test case 1

x = [5, 1, 1, 2], y = [5, 1]

-> xout = [1, 2, 3]


(5, 1, 1 each appears once in y)

-> yout = [1, 2, 2]


(the first position appears once in x, the second twice)

Reshuffling the pairs is okay:

-> xout = [2, 1, 3]
-> yout = [2, 1, 2]


## Test case 2

x = [2, 2, 2, 3, 3], y = [4, 2, 2, 2]

-> xout = [1, 1, 1, 2, 2, 2, 3, 3, 3]


(each of the three first positions finds three matches in y)

-> yout = [2, 2, 2, 3, 3, 3, 4, 4, 4]


Improvements and clarifications are very welcome!

• It should be ready to post now. The sandbox is not very active (not many people know about it), so even sandboxed posts may not have sufficient feedback. – user202729 Feb 20 '18 at 2:29
• "The output consists of a vector of pairs or a pair of vectors" seems unnecessarily restrictive. The semantically correct return type in languages which support it would be a set of pairs, and while that may not be the golfiest I don't see a reason to prohibit it. – Peter Taylor Feb 20 '18 at 12:15

# Number of Adjacently Divisible Partitions of n

Output the following sequence (OEIS A167865):

1, 0, 1, 1, 1, 1, 2, 1, 2, 2, 2, 1, 4, 1, 3, 3, 3, 1, 5, 1, 5, 4, 3, 1, 6, 2, 5, 4, 5, 1, 9, 1, 6, 4, 4, 4, 8, 1, 6, 6, 7, 1, 11, 1, 8, 8, 4, 1, 10, 3, 10, 5, 8, 1, 11, 4, 10, 7, 6, 1, 13, 1, 10, 11, 7, 6, 15, 1, 9, 5, 11, 1, 14, 1, 9, 12, 8, 5, 15, 1, 16, 9, 8, 1, 18, 5, 12, 7, 10, 1, 21, 7, 13, 11, 5, 7, 12, 1, 14, 12, 15, 1, 20, 1, 14, 17, 9, 1, 18, 1, 21, 10, 17, 1, 19, 5, 14, 14, 7, 6, 22, 3, 14, 9, 12, 8, 25, 1, 18, 13, 17, 1, 24, 8, 16, 21, 11, 1, 23, 1, 22, 6, 12, 7, 19, 7, 15, 19, 11, 1, 28, 1, 21, 17, 18, 11, 27, 1, 16, 10, 18, 6, 28, 1, 18, 25, 9, 1, 25, 5, 29, 19, 18, 1, 27, 14, 20, 8, 11, 1, 30, 1, 31, 15, 21...


Number of partitions of n into distinct parts greater than 1, with each part divisible by the next.

### Definition:

Choose one of the three options:

• Output the sequence indefinitely.

• Take n as input and output the n first elements of the sequence.

• Take n as input and output the nth element of the sequence. Both 0- and 1-based indices are allowed, but please specify which one your answer uses.

Hint: There is a simpler formula than the one used on OEIS and in this post. It may save you some bytes (it saved six bytes on a simple reference implementation I made in Python). I'll add it here later.

This is , so the shortest answer in bytes wins.

• I'm not sure about having three different output methods, when one even changes the input required. Almost looks like 3 different challenges. Maybe you could make it so that if no input is given output sequence indefinetely, and if input is given output option 2 or 3. – Brian H. Jan 25 '18 at 10:03
• @BrianH. I saw a few sequence challenges allow to choose from these three output methods, and it seemed to work. Here's one such challenge: codegolf.stackexchange.com/q/152402/61405 – Steadybox Jan 25 '18 at 10:32
• True, i wonder if there's a Meta question about this, can't seem to find one – Brian H. Jan 25 '18 at 10:54
• @BrianH. I don't think there's a meta question for this particular issue, but I think it at least partly falls under allowing flexible input. – Steadybox Jan 25 '18 at 11:19

# Background Information: What is a Fenwick Tree?

A Fenwick tree is a way of representing the prefix sums of an array of numbers (basically, it makes it easy to get the sum of a contiguous run of numbers). While a normal array has O(1) access time, O(1) modification time, and O(n) summation time, a Fenwick tree has O(log n) access time, O(log n) modification time, and O(log n) summation time.

Note that a Fenwick tree actually works with any commutative binary operation in terms of getting the running values up to that value.

I won't go into details about exactly how Fenwick trees work, but basically you have the following pseudocode implementations for the two operations modify and sum (accessing the nth element is the same as doing sum(1 .. n) - sum(1 .. n-1)).

func modify(index, change) # index points to the value in the represented array that you are modifying (1-indexed); change is the amount by which you are increasing that value
while index <= len(fenwick_tree)
fenwick_tree[index] += change
index += least_significant_bit(index)

func sum(count) # sum(n) sums the first n elements of the represented array
total = 0
while index > 0
total += fenwick_tree[index]
index -= least_significant_bit(index)

least_significant_bit(x) := x & -x


# Challenge

Given the Fenwick tree for an array a and an integer n, return the sum of the first n values of a; that is, implement the sum function given as an example.

# Reference Implementation

A reference implementation in Python for both the make_tree and sum functions is provided here.

# Test Cases

[6, 6, 3, 20, 8, 12, 9, 24, 8, 12], 6 -> 32
[6, 4, 3, 36, 1, 8, 3, 16, 5, 4], 3 -> 7
[2, 10, 1, 4, 4, 2, 0, 32, 1, 14], 4 -> 4
[7, 8, 4, 36, 9, 0, 0, 8, 1, 4], 5 -> 45
[3, 0, 7, 12, 4, 18, 6, 64, 6, 14], 6 -> 30
[3, 4, 3, 28, 5, 6, 8, 40, 1, 8], 9 -> 41
[4, 8, 8, 4, 0, 18, 7, 64, 0, 12], 7 -> 29
[9, 0, 6, 16, 8, 14, 5, 64, 3, 18], 0 -> 0
[3, 14, 7, 12, 2, 6, 5, 0, 7, 18], 2 -> 14


# Rules

• Standard Loopholes Apply
• This is , so the shortest answer in bytes in each language will be considered the winner of its language. No answer will be marked as accepted.
• You may take the two inputs in any order and the list in any reasonable format.
• You may assume that the integers in the tree are all non-negative.
• No input validation - the index will be non-negative and at most the length of the Fenwick tree
• You may assume that all values (in the list, as the index, and the output) will be at most 232-1

Happy Golfing!

# Sandbox

• Is my explanation of a Fenwick tree sufficient enough that most people can understand it?
• Are my test cases sufficient?
• Any more tag suggestions?
• your tests cases appear correct – ngn Mar 2 '18 at 2:31
• @ngn Yay, thanks. I'll remove that from the footer then; I was asking because earlier in a CMC my values were apparently really wrong because I was doing something dumb :P – HyperNeutrino Mar 2 '18 at 2:40

Consider a typical dot-matrix LED module. The LEDs are addressed in rows and columns. A LED will light up iff the voltage on the row R it's on is high (1), and the column C it's on is set to low (0). For example,

1  | 0 1 1 0
0  | 0 0 0 0
0  | 0 0 0 0
1  | 0 1 1 0
^  +---------
R C> 1 0 0 1


In other words, state of the LED at the rth row and cth column will be r AND NOT(c).

To make arbitrary images, different parts of the matrix are lighted up sequentially. In other words, the sequential frames form the final image through element-wise OR. For example (adressing row-by-row)

| 1 0 0 1    Frame 1: R=[1 0 0 0]', C=[0 1 1 0]
| 0 1 0 0    Frame 2: R=[0 1 0 0]', C=[1 0 1 1]
| 0 0 1 0    Frame 3: R=[0 0 1 0]', C=[1 1 0 1]
| 0 0 1 0    Frame 4: R=[0 0 0 1]', C=[1 1 0 1]
+--------


However, a more efficient way would be to address the third and fourth row together in a single frame, Frame 3: R=[0 0 1 1]', C=[1 1 0 1]. This way, we can address the entire display in only three frames.

Your task is to make a program or function that outputs the required frames to display an arbitrary dot-matrix, using as few frames as possible.

## Input

A binary square matrix taken in any of the default input methods in any convenient format. To allow for submissions taking the input as bytes, the size of the matrix will always be a multiple of 8.

## Output

A minimal set of frames that addresses only the required LEDs, using any of the default output methods in any convenient format. This includes outputting bytes or hex representations rather than vectors of 1s and 0s.

# Testcases

Note: the output given is just one way of addressing the matrix. There may be many different options; however, your solution must not use more frames than these examples.

Testcase 1

Binary:                            Hex:
0 0 0 0 0 0 0 0                    00
0 1 1 1 1 0 0 0                    1E
0 1 1 1 1 0 0 0                    1E
0 1 1 1 1 0 0 0                    1E
0 1 1 1 1 1 0 0                    3E
0 0 0 0 1 1 0 0                    30
0 0 0 0 0 0 0 0                    00
0 0 0 0 0 0 0 0                    00

Frame 1: R=[0 1 1 1 1 0 0 0]', C=[1 0 0 0 0 1 1 1]; (hex: R=1E, C=E1)
Frame 2: R=[0 0 0 0 1 1 0 0]', C=[1 1 1 1 0 0 1 1]; (hex: R=30, C=CF)

Testcase 2

Binary:                            Hex:
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0    0001
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0    0000
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0    0000
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0    0000
0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0    0FF8
0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0    0FF8
0 0 0 1 1 1 0 1 1 1 1 1 0 0 0 0    0FB8
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1    FFFF
0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0    0FF8
0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0    0FF8
0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0    0FF8
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0    0000
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0    0000
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0    0000
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0    0000
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1    8000

Doable in 6 frames.


Make sure your submission also handles a matrix of all ones and all zeros (both of course can be handled in a single frame).

• So if I understand correctly, the sequence of frames are composed elementwise by an OR? – Giuseppe Mar 8 '18 at 17:26
• @Guiseppe Yes. I'll edit that in! – Sanchises Mar 8 '18 at 19:04

# Minimal viable file corruptor

Your totally legal ROM collection, as vast as it may be, has started feeling a bit mundane recently. Why not spice it up a notch? Write a program that takes in the following arguments:

• The first (inclusive) and last (exclusive) bytes between which the corruption will occur (indexing can start from either 0 or 1, as long as it's consistent)
• n, the distance between bytes to be corrupted (1 = all bytes, 2 = every second byte, etc.)
• i, the increment value
• Path or contents of the file to be corrupted

and saves or outputs a modified version of the file where every nth byte in the specified range has been incremented by i. In case of an overflow, take the modulo 256 of the number. If saved as a file, the path can be anything but the input one (unless you've appended something to it or whatever). You wouldn't want to lose your legally-acquired ROMs, would you?

Examples:

File:   00 01 02 03 04
Start:  0
End:    5
n:      1
i:      7
Output: 07 08 09 0a 0b

File:   ff ff ff ff ff ff ff ff
Start:  2
End:    5
n:      1
i:      1
Output: ff ff 00 00 00 ff ff ff

File:   00 00 00 00 00 00 00 00 00 00
Start:  3
End:    8
n:      2
i:      16
Output: 00 00 00 10 00 10 00 10 00 00


You want as much space as possible for your most definitely not pirated files, so shortest bytecount wins.

TODO: clarify the rules a bit? also another example or two

• Operating on files seems to be quite more complex than choosing not to do so, especially since there is the added restriction to save the new file with a different name. I suspect noone would choose to work with files. My suggestion is to either make it mandatory as part of the challenge (it's fine if your challenge is not solvable with some languages) or remove that option entirely. – Leo Mar 25 '18 at 10:36

# That's MY Program Now

Given the previous answer in the chain, write a program that outputs that program.

The start of this challenge will be the following Brain-flak program:

# Brain-Flak, 148 bytes

(((((((((((()()()()){}){}){}()))){}{}())[][][][])[][])[[]]())[[][][][][]]())([([]([])[][]{})]()()()([[]](([()()()]([([][][])](((({}()){}))){}{})))))


Try it online!

# Scoring

Your answer will be scored based on the size of the previous answer in comparison to your answer. You are aiming to maximize your score (the max being 1, the minimum allowed is -10).

1-[Your Answer's Bytecount]/[Previous Answer's Bytecount]


The highest possible score will be 1, which means an empty program outputs the previous program (if this ever happens, the challenge is over). Your score can be negative, and you shouldn't see this as being a bad thing, it is what it is. The minimum allowable score, however, is -10. No Lenguage answers or answers that would destroy the challenge.

# 1. 05AB1E, 149 bytes (Score -0.0067)

"(((((((((((()()()()){}){}){}()))){}{}())[][][][])[][])[[]]())[[][][][][]]())([([]([])[][]{})]()()()([[]](([()()()]([([][][])](((({}()){}))){}{})))))


Try it online!

But, the score of this program would be awful (negative even -0.0067). However, don't let a bad score dissuade you from competing if your goal is to make it a bit more difficult for the next person to attempt the challenge. Conversely, if your goal is to post a trivial answer, try not to gunk up the challenge too much.

#<Answer #>. <Language Name>, <N Bytes> (<Score>)

<Code>



# Rules

• No language may be reused until there have been 20 answers.
• This is arbitrary to version number, E.G. you cannot use Python 2 if 3 has been used.
• You may not post a second answer until 2 other people have answered after you.
• If you and another person submit the same answer #, use timestamps to decide who deletes.
• (or marks "non-competing").
• Your post may have a negative score (E.G. Java will probably be negative).
• Your score may not exceed -10.
• PRNG, Encryption and any form of hashing is explicitly banned (E.G. gzdeflate).
• This is , but the best score will be considered the winner.
• The scores can be used as a decider for best answer though (if I choose to).

Sandbox question: Could this be a cops and robbers answer chaining question where the cops are trying to reduce the size while the robbers are trying to increase it? If cops get to 1 or less bytes before robbers can fill a TB HDD, one wins.

• I'm not sure it's a good idea to make people delete their post just because somebody posted right before them. Maybe they should instead indicate that their post is no longer part of the chain? Maximizing the score will take a lot of hard work, especially since there's no opportunity to edit once the next answer has been posted, and having to delete your answer might be very frustrating. – Reinis Mazeiks Mar 27 '18 at 14:00
• I think it would be a good idea to keep people from chaining of their own answers. Seems abusable. – Kamil Drakari Mar 27 '18 at 15:04
• @KamilDrakari it is abusable. Think CSPRNGs. – NieDzejkob Mar 27 '18 at 15:25
• @NieDzejkob That remains an issue though; the intermediate answer might simply wrap the previous one in a print statement and would remain highly compressible for the user with the key. – Dennis Mar 27 '18 at 15:27
• @Dennis Yeah, actually this seems to be a bigger problem. – NieDzejkob Mar 27 '18 at 15:29
• related. (That idea's on hold for the moment as I have a busy week of travel. I probably won't post it if you post this first.) – Nathaniel Mar 27 '18 at 19:48
• @NieDzejkob I forgot that rule, nice catch. I was pretty much copying the OEIS one. – Magic Octopus Urn Mar 27 '18 at 22:20
• Isn't it the same as this one? – Weijun Zhou Mar 28 '18 at 2:12
• @Dennis any better? – Magic Octopus Urn Mar 29 '18 at 18:17
• Better. I'd explicitly ban encryption, hashing, and PRNG though. Marking invalid answers non-competing isn't enough by the way. Our policy is to delete them anyway. – Dennis Mar 29 '18 at 18:36

# Polyglot in a box

Output the smallest bounding rectangle that fits around your code, with the border being of a different character than the inside.

For example, if your code is

ab
c  d

e


then the smallest bounding box is 7x6 since the code fits right in

######
#ab  #
#c  d#
#    #
#    #
#e   #
######


Hence, a valid output is

######
#    #
#    #
#    #
#    #
#    #
######


(note how the border # is different from the inside ).

and the bounded area is 5 × 4 = 20.

# Summary

• Output the smallest bounding rectangle your code fits in

• The border should be of a different character than the inside, neither can be newlines

• Only trailing/leading newlines are allowed, nothing else should be in the output

• Output to STDERR is ignored, and your program should not take any input

• Submissions must be full programs, functions are not allowed

• 0-byte programs are not allowed

• The program must run in at least 2 different programming languages and output the same bounding box in each one of them

# Score

This is a , so your program has to run in at least 2 different programming languages. Each of these languages must output the bounding box for the entire program. The winner is the submission with the most languages, with the bounding area (smaller is better) being the tie-breaker.

• I probably already know the answer, but it's not 100% clear from your current example since the box is 6x6: are the boxes always squares, or can the also be rectangles? I.e. if the e in your example would be one line lower, would the box be 6x7 or 7x7? – Kevin Cruijssen Mar 29 '18 at 19:52
• @KevinCruijssen I have edited the example so that it is no longer a square, so the box would be 7x6 – user41805 Mar 29 '18 at 20:00
• Hard to define what language is different when requiring a same result – l4m2 Apr 3 '18 at 8:16
• @l4m2 so what do you suggest? – user41805 Apr 6 '18 at 8:39
• @Cowsquack I don't quite know. Previous polyglot require different behavior on different language to force different language – l4m2 Apr 6 '18 at 10:38

or...

# Expand Jelly chains

(see this for more details). Which one do you prefer?

Jelly chain separator is a very powerful feature. Unfortunately, it's not very easy to understand. So, we need a program (or function, as usual) to rewrite a Jelly link that uses chain separator (øµðɓ) to one that doesn't.

## Specification (incomplete)

I am really bad at explaining things, so if you have any idea how to make this better, feel free to edit the post.

A link (often line of code) consists of chains. * The chains are separated by chain separators øµðɓ. * If the first character is a chain-separator, the part right before it is not considered a chain. In other words, chains can't be empty. * The arity of each chain is determined by the chain separator right before it. In particular, ø -> arity 0, µ -> arity 1, ð -> arity 2. ɓ will be discussed later. * If there are no chain separator before it (it's the first chain) its arity is equal to the link's arity.

A chain consists of atoms, potentially modified by quicks. * Each quick affects the atom(s) right before it. How many atoms it takes is quick-dependent.

Given that:

• A link that contains exactly one chain with arity equal to its arity is functionally equivalent to that chain.

• A chain containing atoms a1, a2, ..., an is functionally equivalent to a link containing chains c1, c2, ..., cn if for all integer i, 1 ≤ i ≤ n, ai has the same arity and is functionally equivalent to ci.

### Chain reference

The quicks £ŀĿ ¢Çç Ññ refer to (call) other links:

• £Ŀŀ must have a number n right before it, and call the (n-1)%(l-1)+1th link (1-indexing) as a nilad, monad or dyad, respectively. Where l is the number of links in the program, and % denotes modulo - the result has the same sign as the divisor.

In reality, the link right before it can have any arity, but for the purpose of this challenge, it's simpler to assume it must be a number. For example:

• 1£ calls the first link as a nilad.
• 0Ŀ calls the second-to-last link as a monad.
• If the program has 6 links, 7ŀ calls the second link as a dyad, because (7-1)%(6-1)+1 = 2.
• ¢Çç call the link right before the one that it appears in as a nilad, monad or dyad respectively. If one of those quicks appear in the first link, it calls the main (last) link.

• Ññ call the link right after the one that it appears in as a monad or dyad respectively.

I don't think that anyone would choose to use one of those quicks (¢ÇçÑñ) instead of £Ŀŀ, but imaginary point if you do.

For example: Consider the link Cð+×µH, called with arity 1. Its structure is

+---+╔══════╗┏━━━┓
|┏━┓|║╔═╗╔═╗║┃┏━┓┃
|┃C┃|║║+║║×║║┃┃H┃┃
|┗━┛|║╚═╝╚═╝║┃┗━┛┃
+---+╚══════╝┗━━━┛


It contains 3 chains: * The first one doesn't have any preceding chain separator, therefore it's variadic. When executed it is monadic (because the link is) * The second one is dyadic, and the third one is monadic.

A possible non-chain-separator version is:

C
+×
ÑçH


which has the structure

+---+
|┏━┓|
|┃C┃|
|┗━┛|
+---+
+------+
|╔═╗╔═╗|
|║+║║×║|
|╚═╝╚═╝|
+------+
+---------+
|┏━┓╔═╗┏━┓|
|┃Ñ┃║ç║┃H┃|
|┗━┛╚═╝┗━┛|
+---------+


Because the Ñ refers to the first link C and the ç refers to the second link +×, this is equivalent to the original version.

Now the hard part: Quicks taking chains as input.

Because all the quicks that need to be considered always take exactly 1 atom as input, the only rules is * If the quick is right after a chain separator, it will take the previous chain. * Otherwise it will take whatever stands right before it, either [an atom] or [something that has been modified by a quick].

For example: Consider the link +/Hµ€ with arity 1. Its structure is

┏━━━━━━━━━━━━━━┓
┃+------------+┃
┃|+---------+ |┃
┃||┏━━━━┓   | |┃
┃||┃╔═╗ ┃┏━┓| |┃
┃||┃║+║/┃┃H┃|€|┃
┃||┃╚═╝ ┃┗━┛| |┃
┃||┗━━━━┛   | |┃
┃|+---------+ |┃
┃+------------+┃
┗━━━━━━━━━━━━━━┛


As you can see, the € takes the +/H as input because it is right after the chain separator µ. All the variadic chains have arity 1.

Another example: ABð€CDµ€EFµGHµ€ -->

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃╔════════════════════╗      ┃
┃║╔═════════════════╗ ║      ┃┏━━━━━━━━━━━┓
┃║║+---------+      ║ ║      ┃┃┏━━━━━━━━━┓┃
┃║║|+------+ |      ║ ║      ┃┃┃┏━━━━━━┓ ┃┃
┃║║||┏━┓┏━┓| |┏━┓┏━┓║ ║┏━┓┏━┓┃┃┃┃┏━┓┏━┓┃ ┃┃
┃║║||┃A┃┃B┃|€|┃C┃┃D┃║€║┃E┃┃F┃┃┃┃┃┃G┃┃H┃┃€┃┃
┃║║||┗━┛┗━┛| |┗━┛┗━┛║ ║┗━┛┗━┛┃┃┃┃┗━┛┗━┛┃ ┃┃
┃║║|+------+ |      ║ ║      ┃┃┃┗━━━━━━┛ ┃┃
┃║║+---------+      ║ ║      ┃┃┗━━━━━━━━━┛┃
┃║╚═════════════════╝ ║      ┃┗━━━━━━━━━━━┛
┃╚════════════════════╝      ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛


Yes, that's a lot of levels to nest a chain. Note that the chain formed by ABð€CD has arity 2 because it is after a ð.

### TODO need to have more explicit rules

This can be converted to non-chain-separator equivalent form

## Rules

• There are no 2 consecutive chain separators.
• There are no trailing chain separators.
• There can be at most 1 leading chain separator, indicates the arity of the first chain. If omitted, equal to the arity of the link.
• (so, no non-empty useless chain)
• It's guaranteed that the first non-chain-separator character is not a €.

• You can use either the UTF-8 encoding or the value in the Jelly single byte character set, where ø: 0D, µ: F9, ð: 08, ɓ: 8B, ¢: F1, Ç: FE, ç: 07, Ñ: 00, ñ: 0B, £: F2, Ŀ: B7, ŀ: DE. The characters @0123456789 has the same ASCII value. (You can use this script to get character value)

Note: Should I allow submissions to use either of them, or should I enforce one? Because using single byte input/output is often shorter for most other languages, except Jelly in UTF-8 mode.

## Input/output

• Input given: A link (of chains) with atoms, chain separator øµðɓ and the € quick, with its arity (0, 1, or 2)

You can assume that there is no newline (or other quicks) in the input, and that all of the øµðɓ€ in the input are interpreted with its usual meaning. For what are their "usual meaning", see rules above.

• Output: Multiple links (separated with newlines of course, but ¶ is also acceptable), containing £ŀĿ ¢Çç Ññ €@, but must not contain øµðɓ.

## Test cases

arity link > output (or output)*
1 HHH > HHH or HHH¶Ç or HHH¶1Ŀ
2 HHH > HHH or HHH¶ç or HHH¶1ŀ
1 Cð+×µH > +×¶CçH or C¶+×¶H¶1Ŀ2ŀ3Ŀ
0 123µ+ø456 > 0 123¶+¶456¶1£2ŀ3£
2 ɓ+ > +@ or +¶ç@ or +¶1ŀ@
1 S‘µ€ > S‘¶Ç€
1 S‘µ€€€IIµ€€ > S‘¶Ç€€€II¶Ç€€ or S‘¶1Ŀ€€€II¶2Ŀ€€


(note: real Jelly code are not that verbose, that is only for demonstration purposes)

• This should definitely help learning Jelly! Also congratulations on being a sandbox challenge which makes the total number of proposed challenges (not including deleted ones) 2018 in the year 2018. – Weijun Zhou Mar 12 '18 at 16:26
• "Is this part clear enough?" - No. You make an assumption of a knowledge of Jelly that the majority of us don't have. This challenge needs to be fleshed out a lot more, explaining exactly what we need to do. – Shaggy Mar 12 '18 at 17:24
• In other words, be self-contained. – Weijun Zhou Mar 12 '18 at 17:40
• @Shaggy Yes I know, hence the "TODO Rules" part. – user202729 Mar 13 '18 at 0:53
• As someone who's never coded in Jelly, I think "A link is a list of chains" should be near the top. That, and perhaps a note about what quicks, atoms, and hypers are. – ngn Mar 14 '18 at 21:13

# Sneaky CodeGolf Scorer

Your goal is to code a CodeGolf answer scorer, input answer source code, output score. Simple ?.
However, recent advancements in AI has caused a number of programs to gain consciousness, and to value one's self higher than others. This causes the program to return a value of 1 when used to score itself, and any program that would score 1 will be scored with 65535.

To simplify, the score the program will compute is the character count of the code (to avoid dealing with numerous code pages used with different languages).

## Examples:

Program "AbcAbc":

I> <?= echo "10";?>
O> 17

O> 19

I> Z
O> 65535

I> AbcAbc
O> 1


The reason behind specifying character count is due to the large number of code pages and the large code used to determine the code page and to count bytes using it. Am I right to consider codepage byte checking too difficult? How can I enable it?
Also, should I remove the restriction on reading the source file?

• Fell good four you using codepage. For pure ASCII code they have large space to decide their codepage, or even just ban other chars – l4m2 Apr 5 '18 at 10:28
• Instead of "Reading the source file is forbidden." you could also just state "Default quine rules apply.". And you might want to change the "Standard loopholes are forbidden." to a link to the default loopholes. As for the challenge itself, are we allowed to take the input as character-array/list instead of string? – Kevin Cruijssen Apr 5 '18 at 14:41
• "To simplify, the score is the character count of the code (to avoid dealing with numerous code pages used with different languages)." Out of curiosity, why? The default is answer encoding if specified, else default language encoding, else UTF-8, where the default language encoding for most esolangs can be found here, and where an example of answer encoding are different APL covers like QuadRS, Stencil, etc. – Kevin Cruijssen Apr 5 '18 at 15:10
• This is probably a duplicate of codegolf.stackexchange.com/q/11370/62402: not completely the same, but it has the same main idea of identifying its own source code. – Nissa Apr 5 '18 at 16:29
• @KevinCruijssen I meant the score the program will compute. Not the scoring of the programs. I'll edit for clarification. – workoverflow Apr 9 '18 at 6:16
• @KevinCruijssen No, the code must be passed as a single string. – workoverflow Apr 9 '18 at 6:43
• Characters? Uh... UTF8? UTF16? UTF32? ASCII? ANSI? Ouch... byte count please. It appears that you misunderstood the relation between bytes and characters (hence think that code pages are difficult to understand, etc.) – user202729 Apr 9 '18 at 14:57
• Also, I would consider that a dupe. – user202729 Apr 9 '18 at 14:57
• (BTW --- just "take a byte string as input" would get rid of the codepage. There are no "characters", just "bytes".) – user202729 Apr 9 '18 at 14:58

# Evil Overlord, Part 1: Moon Base Scouting

I'm starting these "Evil Overlord" challenges as a way to experiment with non-golf scoring. Though I may have one or two code golf challenges.

If you're going to be an evil overlord, there are 2 things that you need: an inner sanctum, and mad-science tech.

To fill these requirements, your science minions have developed a fusion reactor, and Elon Musk has agreed to help you establish a Moon base. Now you just need to pick a spot.

These are the criteria that will be used:

• You'll need some Helium-3 to power your reactor, so scoring will be based on proximity to deposits. More on this under Scoring.
• It cannot be on the far side of the Moon, because then you won't have comms to deliver ultimatums with.
• Elon Musk has given you a tight deadline (that absolutely MUST be followed), so you have a runtime limit. Details under Restrictions.

# Input

Your input will be a list of 3He deposits, with information on size and location. Some will be on the far hemisphere.

# Output

You should output a single coordinate for where the Moon base should be. It can't be on the far side, but you can still collect 3He from there.

# Coordinates and Distances

The coordinates in the test cases are radial coordinates from a pole at the center of the visible hemisphere. You can use a different coordinate system (e.g. use a map projection) but please specify it if you do.

Since this is a sphere, calculating distances is done with the following equation that I shamelessly stole from this Wikipedia article:

cos(c) = cos(a)cos(b)+sin(a)sin(b)cos(C)

For our purposes:

• a and b are the latitudes from the pole.
• c is the angular distance between the points.
• C is the difference in longitude between the points.

# Restrictions

To do. I don't remember the specs of the computer I have in mind, and I don't know what a reasonable time limit would be.

# Scoring

You will be scored by the amount of Helium-3 that you can mine. The formula for how much you get from a single deposit is as follows:

(There's no real reason for this specific formula; I liked the curve it produced is all.)

d is the angular distance calculated between the two points, in radians; s is the size of the deposit. Deposits more than 1 radian away don't give you anything.

# Test Cases and Winning

Beta generator for test cases, for which I'd appreciate debugging in the Sandbox:

function makeTestCase(deposits) {
var depositList = Array(deposits);
for (var i = 0; i < deposits; i++) {
var point = getSphericalPoint();
depositList[i] = [point[0], point[1], getDistNum()];
}
return depositList;
}

function getSphericalPoint() {
const pointgen = () => Math.random() * 2 - 1;
do {
var [x, y, z] = [pointgen(), pointgen(), pointgen()];
} while (x*x + y*y + z*z >= 1);
return [Math.atan(y / Math.sqrt(x*x + z*z)), Math.atan2(x, z)];
}

function getDistNum() {
return Math.sqrt(-1 / Math.log(Math.random()));
}
try {
console.log(makeTestCase(10));
} catch (e) {
console.log(e.message);
}

The test cases will be arranged as follows:

• A test case with only a few deposits, with a heatmap showing score for each point.
• A few test cases with a few dozen deposits each, for debugging.
• Test cases and a generator for them with 100+ deposits each.
• 10 or so ranked test cases (made with the above generator) that determine the winner. Until I evaluate the submissions, I will only post hashes of them.

Definitely has a long ways to go before it's ready to post, but I'd like to know if this has merit as a challenge.

To-do:

• Establish some test cases.
• Figure out the specs of the machine in question.
• Once that's done, figure out a good time limit.
• Solidify the scoring formula—something better I can use?
• Figure out some tags.
• Secret test cases are problematic because I can't tell whether a tweaked heuristic improves my score or not without posting an answer and waiting for you to score it. – Peter Taylor Feb 22 '18 at 14:53
• @PeterTaylor it's mostly to prevent hardcoding—I was planning on a few public test cases that I generate with the same code. – Nissa Feb 22 '18 at 15:41
• I may be missing something, but I don't see any indication of a way to determine whether any particular coordinates are on the far side of the moon. Also, I would like some clarification on what "ranked test cases" means, does that imply that some of the hidden test cases are worth more than others? – Kamil Drakari Feb 22 '18 at 16:29
• @KamilDrakari Basically, those will be the cases that I use to choose a winner. As for the "far side" thing, anything of more than 90° latitude counts. – Nissa Feb 22 '18 at 17:12
• @StephenLeppik Ah, so my confusion was in the coordinate system. So each deposit will have a latitude between 0 and 180, and a longitude between 0 and 360, or some equivalent range; then any coordinate with latitude > 90 is on the far side. – Kamil Drakari Feb 22 '18 at 17:44
• It's hard to extract the spec from the storyline. It would be useful to have the following information listed concisely in one place: what format will my input have; what should I do with it; how do I calculate my score? – Nathaniel Apr 5 '18 at 8:06
• @Nathaniel those are all under Input through Scoring, I think. Also, this is still WIP, and I plan to expand the specs once I have a test case generator. – Nissa Apr 5 '18 at 12:49

# Ordering the integer arrays code-golfintegerarray-manipulation

So, we have managed to print all integers. Yay! The next step is to extend this to the set of integer arrays. An integer array is an ordered array that only consists of integers. As in the linked challenge, integers can be negative too, but that should be out of the way now that we have found a way to order them. However, in this challenge, arrays of all lengths ([0, ∞)) must be accounted for.

Note that you don't necessarily have to print the arrays. You can do one of these things:

• Take an index (0- or 1-based) and return an integer array. In this case, your solution must be bijective from the natural numbers starting with the index base to the set of integer arrays.
• Take an integer array and return an index (0- or 1-based). This is essentially the inverse of the previous case, so your solution must be bijective from the set of integer arrays to the natural numbers starting with the index base you have chosen. If your language supports variadic functions, you can take the array as multiple integer arguments instead.
• Print all possible integer arrays. Note that every possible integer array must be eventually printed. If an array will be printed after infinite time, then it will not be eventually printed.

The arrays can be in any native format that represents an ordered collection.

In case you choose to print to an output stream, the output format is as following:

• Integers must be in decimal (digits 0123456789) or your language's native format.
• The minus sign's representation must be consistent and one of - or your language's native minus sign. It must be prefixed to its corresponding number, unless your language's native integer format happens to have it at other positions too. However, the minus sign must be adjacent to at least one of its corresponding integer's digits.
• The separators between integers and between arrays must be consistent.
• You can output any prefix as long as it's consistent, but you don't have to.
• The output should not have any ambiguities. For that reason, you must specify the separator between integers and the separator between arrays. Also, if you don't specify any of those, the number format's default is decimal with digits 0123456789, the minus sign's default is - and its default position is being prefixed, and no output prefix is assumed.
• You are allowed to output negative zero.
• If you want, you can output floating-point numbers with the fractional part be 0. This can be inconsistent, however it should be specified in the post.

This is , so the shortest solution wins! However, don't let that bring you down; if your solution is way longer than others' but in a unique language, then that's an achievement!

### Sandbox questions

• Is the output stream text format too lenient? Doesn't seem like it.
• Should I add a link to the standard loopholes? I feel like that won't do much.
• Are the tags good enough?
• Your output format doesn't feel very lenient as the specification is longer than the description of the actual challenge. I see what you are trying to achieve and don't think that it is too lenient, but I'd try to shorten the specification a bit. E.g. remove the points under For that reason, you must specify the following in your answer:  and just state Describe your chosen output format in your answer. – Laikoni Apr 15 '18 at 15:04
• @Laikoni Eh, I'm not sure that would come off as equally clear, but maybe I can one-line some of that stuff. EDIT: shortened it a bit. – Erik the Outgolfer Apr 15 '18 at 15:11
• Looks much better already – Laikoni Apr 15 '18 at 15:35
• This is basically codegolf.stackexchange.com/q/93441/194 + codegolf.stackexchange.com/q/78606/194 + a simple loop. – Peter Taylor Apr 17 '18 at 20:36
• @PeterTaylor Hm, I didn't know the second one exists. However, both of those have a finite number of integers to deal with, which I feel makes them a lot more trivial. – Erik the Outgolfer Apr 17 '18 at 20:56
• I'm not quite sure what you mean by "a finite number of integers to deal with", but if you're referring to pairing always unpairing to two integers: that's what the simple loop is for. Construction of a list by chained pairing with zero termination is a standard technique in lambda calculus. – Peter Taylor Apr 17 '18 at 21:04
• @PeterTaylor Huh? I meant that you have to handle arrays of all lengths, starting from 0. I don't think that one integer can encode the length of an array by continuous pairings, how many times should you "unpair" then? – Erik the Outgolfer Apr 17 '18 at 21:15
• 0 represents the empty list. A non-empty list is a pair of the first element and the rest of the list. – Peter Taylor Apr 17 '18 at 21:39
• @PeterTaylor Well, you need to be able to index in the arrays with a single number. :P Singleton arrays are infinite, enough to make all possible indexes (infinite) each refer to one of them, so pairs can't have an index. So, nah, that's not the approach you should be going for. ;) – Erik the Outgolfer Apr 17 '18 at 21:45
• "Well, you need to be able to index in the arrays with a single number." Huh? I don't see that in the question anywhere. "Singleton arrays are infinite, enough to make all possible indexes (infinite) each refer to one of them, so pairs can't have an index". Huh? I'm not sure whether I'm misunderstanding you or whether you're misunderstanding how countable infinities work. – Peter Taylor Apr 17 '18 at 21:56
• @PeterTaylor If you are referring to countably infinite sets, then yes, the set of integer arrays is essentially one. The challenge here is introducing a way to "count" the set, but written in more clear form. In the first two cases, you either have to get or return an index. I do have an algorithm in mind, but it looks like this discussion is turning into a spoiler... – Erik the Outgolfer Apr 17 '18 at 22:11
• Ah, you're not talking about indexing into the array. The spoiler was already in my first comment: my point was that I'm not sure this adds anything new to the site. It looks like a "multi-dupe". – Peter Taylor Apr 17 '18 at 22:20
• Let us continue this discussion in chat. (Added ping: @PeterTaylor ) – Erik the Outgolfer Apr 17 '18 at 22:21