# 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]

# Constructively golf true primes!

## Introduction

A recurring pattern that I see in challenge involving primes is the (over-) use of prime-checking built-ins and of factorization built-ins. However, I kinda don't trust especially the latter, which is why today, we're gonna build primes and include a certificate that the result is actually a prime.

## Input

A positive integer n denoting the length of the resultung prime in a basis of your choice.

## Output

A list of primes in the base you picked above, one of these primes is of length n and one of these primes is smaller than 20.

## What to do?

Your challenge is to generate a prime of length n of which the primality can be reduced to the primality of one of the following primes: [2,3,5,7,11,13,17,19]. This can be achieved using Pocklington's criterion. Now the idea is to chain up from any of these primes using said criterion to a prime of appropriate size, while also outputting the intermediate steps so primality can be checked easily.

For your convenience, I shall reproduce the relevant theorem here:

Let N>1 be an integer and suppose there exist a,q such that

• q is prime
• q divides N-1
• q>sqrt(N)-1
• a^(N-1) mod N == 1
• gcd(a^((N-1)/q)-1,N)=1

Then N is prime.

And as you might see, your initial set of primes (as candidates for q) is limited to the above short list of primes smaller than 20.

## Who wins?

This is so the shortest solution in bytes wins!
However this is also so all solutions must run in (expected) polynomial time in n. Standard IO and Loophole rules apply.

## Got any help?

If you need algorithmic help, I suggest you consult the Handbook of Applied Cryptography, chapter 4 (PDF), algorithm 4.62 (on page 153 or 22 in the PDF) or you independently look for Maurer's prime generation algorithm (originally published here).

# I got them Mad Matrix Moves

Given a list of edges representing a simple directed graph and (optionally) the number of vertices n>1, output the adjacency matrix of that graph.

The vertices will always be numbered consecutively starting at 0 (or 1 -- your choice).

The adjacency matrix is defined as the n by n square matrix M such that M[i][j] is 1 if there is an edge from vertex i to vertex j and 0 otherwise.

# Example

For this example and the test cases, we will have the vertices numbered starting at 1.

Let's use the sample input [(1,2),(1,3),(3,2)]. Since there are vertices numbered 1-3, n=3 and our output is a 3 x 3 matrix.

The output should be

[[0, 1, 1],    <= (1,1) doesn't exist in the input list, (1,2) does, (1,3) does.
[0, 0, 0],    <= (2,1) doesn't exist, same with (2,2) and (2,3)
[0, 1, 0]]    <= (3,1) doesn't exist, (3,2) does, and (3,3) doesn't


(concisely represented as [[0,1,1],[0,0,0],[0,1,0]] in test cases).

# Notes

• Every vertex is the start and/or end of at least one edge.
• You may replace 1 and 0 in the output with any two consistent distinct values of your choosing.
• The input graph is simple, meaning that it has no loops nor multiple edges (e.g. no edge [1,1] or graph [[1,2],[1,2]])
• The graph is directed, meaning that each edge has a start and an edge (e.g. the edge [1,2] is distinct from [2,1])
• Every vertex is an endpoint of at least one edge
• This is , so shortest code in byes in each language wins.

# Test Cases

One test case per line. Each test case is in the format edges, n => output.

[(1,2)], 2 => [[0,1],[0,0]]
[(2,1)], 2 => [[0,0],[1,0]]
[(1,2),(2,1)], 2 => [[0,1],[1,0]]
[(1,2),(1,3),(3,2)], 3 => [[0,1,1],[0,0,0],[0,1,0]]
[(1,2),(2,3),(3,4),(4,5)], 5 => [[0,1,0,0,0],[0,0,1,0,0],[0,0,0,1,0],[0,0,0,0,1],[0,0,0,0,0]]
[(1,2),(3,4),(4,3),(2,1),(2,4),(4,2)], 4 => [[0,1,0,0],[1,0,0,1],[0,0,0,1],[0,1,1,0]]
[(1,2),(2,3),(3,4),(4,5),(2,1),(3,2),(4,3),(5,4)], 5 => [[0,1,0,0,0],[1,0,1,0,0],[0,1,0,1,0],[0,0,1,0,1],[0,0,0,1,0]]
[(1,2),(3,1),(2,3),(1,1),(2,1)], 3 => [[1,1,0],[1,0,1],[1,0,0]]
[(1,2),(3,2),(1,3),(5,4),(5,6),(5,1),(2,3),(1,5),(5,2),(3,6),(4,5),(5,3),(2,1),(3,5),(4,6),(6,3),(6,5),(4,2),(3,4)], 6 => [[0,1,1,0,1,0],[1,0,1,0,0,0],[0,1,0,1,1,1],[0,1,0,0,1,1],[1,1,1,1,0,1],[0,0,1,0,1,0]]


# Get the Average Number of Bytes for an Answer

Have you ever wanted to know the average number of bytes all the answers to a given code golf question took? Probably not. But I did!

## Input

You'll have to take in a link to a codegolf exchange question by text or you can run this in a browser extension/user-script. If you have a clever way of identifying the webpage and producing the necessary output, just ask about it in the comments, and I'll update the question as long as it doesn't violate the rules. The input will be a link to a codegolf question. It will include http::// or https:://. The slashes will be facing this way: "/". The link may not be to a code-golf challenge specifically, but it could be something like a code-challenge.

## Output

If the user inputs something that isn't to a code golf website output something along the lines of "Hey! That's not a link to a code-golf challenge" If the user inputs a link that links to the codegolf exchange, but isn't a codegolf (maybe it's a code-challenge) the program should output anything that is not a natural number.

If there are strikethroughs through previous amounts of bytes a solution took, those should not be included in the average. If the person answering the question in one of the links did not conform to formatting standards, the program doesn't have to account for it. Minor differences will be allowed. Specifically, the number of bytes should be within 5 lines of the top of the answer and bolded.

All answers should be rounded to the nearest whole number with .5 rounding up.

## Examples

Input:

https://codegolf.stackexchange.com/questions/54038/average-out-two-lists

Output:

95

## Winning Criterion

The winner will be given by (1 / execution time (s) + bytes).

If you know of anything that I missed, comment below.

• What if I make the output message my entire program and then just exec it? – HyperNeutrino Oct 10 '17 at 23:30
• @HyperNeutrino I'm not really sure what that means. If it takes a link in and outputs the answer then you've completed the challenge. If it outputs the answer and 20 lines of unnecessary information that the user can see, I would say that that doesn't pass the program. – Byte11 Oct 10 '17 at 23:34
• My point is, what if I make a string containing my entire program, use string slicing to output a good falsy message, and then use exec to run the string? Would I be exempted from counting the bytes of my string? – HyperNeutrino Oct 11 '17 at 0:05
• @HyperNeutrino That's a smart way to get around the rules. No, you wouldn't be exempt from the rules, but from a literal interpretation, you could argue that that is allowed. How should most concisely update my rules to prevent this? – Byte11 Oct 11 '17 at 0:08
• I'd just guarantee that the input will follow a specific format, because making the input contain backslashes, missing components, etc. is just a pain and doesn't demonstrate the answerer's ability to actually do what you really want. Essentially, don't add too much fancy stuff – HyperNeutrino Oct 11 '17 at 0:22
• @HyperNeutrino Okay, I've updated the question. – Byte11 Oct 11 '17 at 0:47
• along the lines of  → unclear what you're asking. – user202729 Nov 3 '17 at 13:21
• Also what exactly is the winning criteria? Another point: A link to this site never have https:://, only https://. – user202729 Nov 3 '17 at 13:21

# Optimised Hashing

For a given set of input, output a program that can 'hash' the input with a minimum number of collisions.

## The Challenge

Write a method or function that receives an array of Strings (see input for more detail) and outputs a full program. This program must, for each String in the original input array, print to STDOUT (or equivalent) some positive integer i where i<100 and is unique for that input.

## The Input

Each program will be tested with 4 different input sets of 50 unique Strings;

• 50 English words of the same length.
• 50 English words of differing lengths.
• 50 randomly generated character strings of the same length
• 50 randomly generated character strings of differing lengths

All Strings in all inputs will only contain the character set [a-zA-z] (ASCII values 65-90 and 97-122) and will all be no more than 20 characters long.

## Special Rules

• Standard loopholes are forbidden.
• Inbuilt hashing functions are not allowed.
• You must provide a free and easily accessible environment to compile and run both programs with your answer.
• I reserve the right to discount any program which takes more than 10 minutes to complete any test.

## Scoring

You will be scored by combining the following;

• The byte count of your original program
• The byte count of each outputted program
• 5 penalty points for each 'collision'.

I will keep and update a leader board as programs are added and I can test them.

Please let me know if I've missed anything; this is my first question idea!

# Hunting The Wren

Write a program to produce the words to the folk song Hunting The Wren.

"We're hunting the wren." says Robin to Bobbin.
"We're hunting the wren." says Richard to Robin.
"We're hunting the wren." says Jack of the land.
"We're hunting the wren." says everyone.

"Where oh where?" says Robin to Bobbin.
"Where oh where?" says Richard to Robin.
"Where oh where?" says Jack of the land.
"Where oh where?" says everyone.


The song continues with the same structure.

"In yonder green bush."
"How get him down?"
"With sticks and stones."
"How get him home?"
"The brewer's big cart."
"How will we eat him?"
"With knives and forks."
"Who'll come to the dinner?"
"The King and the Queen."


The song concludes with three more verses, which because they have no repetition, we'll omit for the purposes of this challenge.

Rules

• The program must produce the full words of the song as 54 or 55 lines of text.
• Each verse of four lines must be separated by a blank line. A blank line on the end is optional.
• You must not use a text decompression library. (GZIP the whole song is 380 bytes.)
• Normal code golf rules apply. Shortest code wins.

C# Reference Implementation

public class Program
{
public static void Main()
{
string[] starts = new [] {
"We're hunting the wren.",
"Where oh where?",            "In yonder green bush.",
"How get him down?",          "With sticks and stones.",
"How get him home?",          "The brewer's big cart.",
"How will we eat him?",       "With knives and forks.",
"Who'll come to the dinner?", "The King and the Queen." };

string[] ends = new [] {
"says Robin to Bobbin.",      "says Richard to Robin.",
"says Jack of the land.",     "says everyone."            };

foreach (string start in starts)
{
foreach (string end in ends)
{
System.Console.WriteLine("\"" + start + "\" " + end);
}
System.Console.WriteLine();
}
}
}


# Pokémon Champion

This challenge is about writing the ultimate bot for battling Pokémon. Since knowledge of competitive Pokémon battling is necessary, I'll leave that to the experts at Smogon University.

### Rules

• Your bot must battle another bot in a Gen 7 Random Battle Format.
• Your bot will be judged based on its performance in a round-robin tournament orchestrated by the author of the challenge, against all other submitted bots before a specific deadline.
• In order to avoid reinventing the wheel, there are several frameworks already available to choose from. Unfortunately, a lot of these are written in Node.js, so you might still have some porting to do, unless this challenge decides to create its own framework for battling.
• Since your bot will be connecting to a Showdown server using the existing Web Socket API, any error in your program that causes the bot to become non-responsive or to disconnect will automatically forfeit that match.
• Standard loopholes apply, and do not exploit any possible bugs in the API. If you find a bug, please report it here.

### Further Considerations

While in theory, I think this would be a great idea, I believe this would require a lot of time-investment and groundwork, which frankly I don't have the availability for right now, so anyone is free to take this over, but please be courteous and get my consent in The Nineteenth Byte first. Just @PatrickRoberts and wait for a reply before modifying this proposal.

* Not sure if it's preferable to write an API specifically for this challenge to eliminate the need for Web Socket communication, since there are already available platforms. The idea is to host an official tournament at a certain date (yet undecided) on Pokémon Showdown.

• I could use a new bot on my side server (the one I've got doesn't know about Z-moves so even I can beat it at Ultimate Z...) – Neil Oct 18 '17 at 0:19

# All the same digits

For any two coprime integers n, b > 1 there exist infinitely many integers k > 0 for which n * k's representation in base b only contains the digit b-1.
This challenge's task is to output the smallest such k for given integers n, b.

### Input

Two integers n, b with the above described properties in any reasonable format (integers, strings, ...). If your language does not support arbitrary precision integers, you can assume that n, b < 2**31.

### Ouput

An integer k with the above described properties in any reasonable format (integer, string, ...). If your language does not support arbitrary precision integers, you can assume that k < 2**31.

### Test cases

f(  n,     b) == k

f(  7,    10) == 142857
f(  13,    5) == 48
f(  14,   13) == 12
f(1728, 1729) == 1
f( 107,   99) == 54863250648363053681585635237627534906401818527472898624219397466199311510764517134614152141940101983414


To generate more test cases, take a look at my reference implementation written in Python 2.

Fraction periods can also be used to calculate k (example given for n, b = 13, 10).

1/13             =      0.076923(076923)* (period length 6)
1/13 *  10**6    = 076923.076923(076923)*
1/13 * (10**6-1) = 076923
999999   = 076923 * 13
k =  76923

• Not true. Consider n=b>1. You need a coprimality condition. – Peter Taylor Oct 17 '17 at 16:03
• @PeterTaylor Thanks for noting. – Jonathan Frech Oct 17 '17 at 17:38

I'm posting this on sandbox first, because this might not be a good challenge for this site, or just too complicated, because it's about constructing a meaningful sentence. Please, tell me, if it's appropriate task for this site.

Also a note: If you want, I can wait for Christmas with this.

# Back to the New Year!

You encounter a probably hard challenge.

Inspired by this answer by @dansalmo.

If we sum character codes in the string Happy new year to you!, we will get 2014. With later years, people started to modify the string, so it fits the current year:

• 2015: A Happy New Year to You!
• 2016: Happy New Year to you!!!

## The challenge

Given an integer year, output a string meaning "Happy new year!". Sum of character codes should be equal to the integer.

Rules:

• No default loop holes allowed
• Output rules:
• Meaning rules:
• You must use the word Happy - you cannot use any other adjective instead.
• You must wish a NEW year, no matter if it's 2010 or 2020!
• You can only wish a happy new year - You can't wish a happy new Easter...
• You can target the string to a person or group, or to everybody: Happy new year!, Happy new year, Soaku!, Happy new year, CodeGolfers!
• You can add new words, but the sentence can't lose the meaning!
• Output must have valid grammar, but you can mess up the punctuation. Just note that you can't use punctuation characters except trailing exclamation marks.
• Letter case rules:
• Letters that aren't first should be lowercase. The only exception is when all letters in the word are uppercase.
• Punctuation rules:
• You are not allowed to use any punctuation characters, except on the end of the string
• There must be exactly one or three exclamation marks ! at the end.
• Spacing rules:
• There must be exactly one space between words.
• No line breaks allowed
• Leading and trailing spaces/newlines are allowed, but they don't count in the character code sum. Word is the same as outputting Word.
• Character rules:
• You cannot use digits!
• You must only use ASCII letters (the only exception are the trailing exclamation marks)
• You can assume the input is an integer between 2000 and 3000

## Test cases

examples, don't need to be the same

2001 - Happy new YEAR to MYSELF!
2002 - A VERY HAPPY NEW YEAR TO PPCG!
2003 - Happy NEW Year TO MYSELF!!!
2004 - A happy new year to me!
2013 - A HAPPY NEW Year To The SE!!!
2014 - Happy new year to you!
2015 - A Happy New Year to You!
2016 - Happy New Year to you!!!
2030 - A Happy New Year To PPCG!!!


• I think meaning "Happy new year!" is too unclear. There is no way to objectively enforce this. – Wheat Wizard Oct 23 '17 at 17:38

# Behavior Driven Golfing

### Context

From wikipedia (emphasis mine): Behavior Driven Development (BDD) is largely facilitated through the use of a simple domain-specific language (DSL) using natural language constructs (e.g., English-like sentences) that can express the behavior and the expected outcomes.

Example in normal BDD (also from Wikipedia):

Story: Returns go to stock

As a store owner In order to keep track of stock I want to add items back to stock when they're returned.

### The challenge:

Write a program that reads like a single English-like sentence that executes the meaning of that sentence.

### Other details

• The program must receive input of some kind (stdin, arguments are okay depending on your language's syntax).
• The input can be limited in scope, (e.g. all real numbers, all positive even integers), but there must be infinitely many possible inputs assuming that your language doesn't have a cap.
• For example, if the input is an integer, you don't actually have to handle integers beyond the language's cap, if your language happens to cap at 32 bit long, it just has to be infinitely many in theory but not necessarily in practice.
• The program must have at least log(n) distinct outputs over an input space of size n.
• For example, if there existed a language where convert to zero was a valid program, returning 0 on all inputs would not be okay, but printing a number of 0 equal to the base 10 number of digits of the input number would be okay.
• The program should not error out on legal input.
• Irregular case and unicode characters are allowed
• Unicode characters will be interpreted as the similar character in English (so most commands in 05AB1E are allowed).
• Words may be delimited by punctuation marks instead of spaces, but they must be delimited.
• Punctuation should not be present in the middle of a word unless it's legal in English (such as an apostrophe in a contraction or a possessive)

Example valid answer for all requirements except the "must take input" requirement:

### HQ9+

Print Hello, world.

Since only the H character is an instruction in this language, all the other characters are ignored, and the program prints hello, world. Since case is ignored, this program does what the english sentence says it will do.

• As it's currently written, the task is way too broad. The task that the program should do must be narrowed and clearly specified (e.g., "add the input with itself") in order for this to be well-received. – AdmBorkBork Oct 20 '17 at 20:00
• I like the challenge, but I think it should be popularity-contest rather than code-golf. I just feel like code-golf will attract too many answers similar to the example: pick a language that ignores most characters, then write a sentence that describes a built-in and includes that built-in as the only executed character. Popularity Contest seems like it would attract much more interesting answers. – Kamil Drakari Oct 20 '17 at 21:59
• @KamilDrakari popularity-content is typically not well received, most of those are closed. – durron597 Oct 20 '17 at 23:20
• I'm not exactly sure how we can measure "X is a similar character to Y" objectively.. Probably @Kamil is right and this should be a popularity contest. – ბიმო Oct 26 '17 at 19:02
• @BruceForte, pop-con is not a "solution" to the problem that the spec is not objective. A pop-con should have an objective validity criterion and not be amenable to an objective scoring criterion. – Peter Taylor Oct 27 '17 at 12:04

# Does the Riemann hypothesis hold?

I would like to earn a million dollars, and the hardest way to do it is apparently solving one of the Millenium problems, so let's focus on one: the Riemann hypothesis.

Now, how could I possibly ever solve it? By writing a computer program, of course! However, I am lazy, so I would rather have you do the work for me by making the program as short as possible (to minimize the number of key strokes required for me to verify your results)

## The challenge

Your goal is to write a program in a language of your choice which would decide the Riemann hypothesis. More precisely, you should write a program which terminates iff the Riemann hypothesis fails.

## Restrictions

• The program has to work when running with empty input [Sandbox note: perhaps allow input but add it to byte count?]
• Standard loopholes are forbidden.
• Your program must provably terminate iff RH is false. For example, submitting two programs, one trivially terminating, one trivially not terminating, and claiming one of them works, is not allowed, since neither of the programs provably works.
• You must give a proof of equivalence of your program's termination with RH failing, with references to other work allowed. I give myself the right to decide whether the work referenced is valid or not, to exclude the many proofs and disproofs of RH.
• If the Riemann hypothesis doesn't hold, your program can throw an error, or simply exit, as long as the program cannot continue beyond that.
• You may assume you have unlimited memory and time, and also that your number types can hold real numbers of arbitrary size to arbitrary precision.
• This is , so the shortest code in bytes wins!

## Sandbox comment

To address the comments below, let me clarify why this challenge is most definitely solvable. Note that the problem is not of the sort "give a counterexample to RH" (if it did, then it would fall under this discussion and I completely agree it would not be a good challenge). Instead, it asks for a program whose termination depends on RH, and then one can give valid solutions regardless of the fact that RH is not solved yet.

To give an analogy (following Deedlit's example), suppose we have replaced RH with Goldbach's conjecture. This conjecture surely is unresolved, but one can give a valid solution to the challenge, for example by following this pseudocode:

n = 4
while true:
found = false
for k from 1 to n-1:
if k is prime and n-k is prime:
found = true
break
if found == false:
break
else:
n = n+2
return 0


If Goldbach's conjecture is false, this program will eventually terminate by returning 0 (namely, when it finds the first counterexample). Otherwise, it will stay in the while loop forever. Hence it would be a valid solution.

My challenge as it stands can be solved in a similar manner, perhaps using

one of the many known equivalents of RH.

• codegolf.meta.stackexchange.com/q/11033/45941 – Mego Oct 27 '17 at 19:47
• @Mego My proposed challenge doesn't fall under that discussion. It is not an open problem whether a solution for that problem is possible. – Wojowu Oct 27 '17 at 19:51
• Relying on open problems for challenges is not a good idea. Either it will be impossible to answer because the problem isn't solved, or answers will be trivial copy jobs of the proof once the problem is solved. This challenge relies on both an open problem (Riemann hypothesis) and an unsolvable problem (halting problem). – Mego Oct 27 '17 at 20:42
• @Mego The way I have phrased my problem makes it possible, but not trivial, once the Riemann hypothesis is unsolved. True, the problem trivializes once we get a (dis)proof of RH, but until then I don't see an issue with this problem. Also, I don't see how halting problem is in any way relevant here. – Wojowu Oct 27 '17 at 20:58
• You don't see how the halting problem is relevant in a challenge for writing problems that halt iff some unsolved problem in mathematics is true? – Mego Oct 27 '17 at 20:59
• @Mego Let me rephrase - I don't see how its unsolvability impacts the challenge in any way. – Wojowu Oct 27 '17 at 21:01
• @Mego As an example, it is straightforward to write a program that halts if and only if the Goldbach conjecture is false. Simply test out even numbers one at a time until you find one that is not the sum of two prime numbers. The unsolvability of the halting problem doesn't have any bearing on this, since we aren't relying on an algorithm that determines the halting behavior of all programs. A program that halts precisely when the Riemann Hypothesis is true could be done similarly. – Deedlit Oct 27 '17 at 22:01
• @Deedlit Can one also show that it is possible to write a program that halts iff Goldbach's conjecture is true? – Jonathan Frech Oct 27 '17 at 22:31
• @JonathanFrech I believe that's an open problem. However, thanks to your comment I've realized I have mixed things up - my challenge should've been about RH being false, not true (so just like in Deedlit's example with Goldbach). Sorry if any confusion was caused. – Wojowu Oct 27 '17 at 22:43
• @Wojowu Well, that seems solvable then. Would one, however, not simply loop through all complex numbers (as one has a lot of computational power), implement the zeta function (again, infinite computational power) and halt if they find a counterexample to the hypothesis? – Jonathan Frech Oct 27 '17 at 22:49
• @JonathanFrech It needs to be quite a bit more than that, because there are uncountably many complex numbers, so you can't loop over them all even with unbounded computing power. I believe you can do it by evaluating a contour integral around rectangles in the complex plane and looping over these rectangles. And it gets annoying around the critical line. It's definitely more complicated than just implementing the zeta function, though. – Chris Oct 28 '17 at 1:58
• @Chris Could one not loop through all possible sums of all rationals to approximate the reals, argue that every real is computable by an infinite sum of rationals, do that twice to gain a program which theoretically looks at every complex number, implement the zeta function and check if the hypothesis holds? – Jonathan Frech Oct 28 '17 at 12:02
• @JonathanFrech It's not possible to look at all possible infinite sums of rational numbers. Some (in fact, most) are uncomputable, and either way there are just too many real/complex numbers. Try to implement your algorithm in a real language like C, you will see that this simply doesn't go through. – Wojowu Oct 28 '17 at 12:06
• @JonathanFrech It is not possible to cycle through all possible complex numbers, but there are other ways which would let one verify the Riemann hypothesis. – Wojowu Oct 28 '17 at 12:17
• @Chris err, yes, that's what I meant. – Wojowu Oct 28 '17 at 20:40

# Are these numbers perfect for each other?

A perfect number is a number that is the sum of its divisors. Examples of perfect numbers are:

     | Divisors                                  | Divisor sum
-----+-------------------------------------------+-------------
6    | 1, 2, 3                                   | 6
28   | 1, 2, 4, 7, 14                            | 28
496  | 1, 2, 4, 8, 16, 31, 62, 124, 248          | 496


Similarly, numbers that are perfect for each other are numbers where each is the sum of the other's divisors.

     | Divisors                                  | Divisor sum
-----+-------------------------------------------+-------------
220  | 1, 2, 4, 5, 10, 11, 20, 22, 44, 55, 110   | 284
284  | 1, 2, 4, 71, 142                          | 220


## Challenge

Write a program/function that takes two positive integers and outputs a truthy/falsey value based on whether they are perfect for each other.

### Specifications

• Standard I/O rules apply.
• Standard loopholes are forbidden.
• The output must be consistent for both truthy and falsey values.
• This challenge is not about finding the shortest approach in all languages, rather, it is about finding the shortest approach in each language.
• Your code will be scored in bytes, usually in the encoding UTF-8, unless specified otherwise.
• Built-in functions that perform this task are allowed but including a solution that doesn't rely on a built-in is encouraged.
• Explanations, even for "practical" languages, are encouraged.

## Test cases

Input         Output

1, 1          truthy
3, 7          falsey
6, 6          truthy
13, 42        falsey
220, 284      truthy
563, 492      truthy
1184, 1210    truthy


In a few better formats:

1, 1
3, 7
6, 6
13, 42
220, 284
563, 492
1184, 1210

1 1
3 7
6 6
13 42
220 284
563 492
1184 1210


## Reference implementation

divisors :: Integer -> [Integer]
divisors 1 = [1]
divisors n = [i | i <- [1..n - 1], n mod i == 0]

perfectPair :: (Integer, Integer) -> Bool
perfectPair (a, b) = (sum $divisors a) == b && (sum$ divisors b) == a


Try it online!

This challenge was sandboxed.

• Although I wouldn't really VTC (seems fine to me), this is very likely to be closed as a dupe of Am I perfect (number)?. – Mr. Xcoder Oct 29 '17 at 19:46
• it's also kinda trivial... Try it online! - 3 bytes, 05AB1E – Magic Octopus Urn Oct 31 '17 at 20:24

# Self-language validator

Create a language validator, which takes a string/file that might or might not be syntactically in the language of your choice, and returns a value corresponding to whether or not it is a syntactically valid program in the implementing language.

• At a bare minimum, your submission must return a truthy value for code where a strict implementation compiles/interprets without any problems, and must return a falsey value for code that even lax implementations cannot recognise as valid in your language.
• Your submission must have cases where it returns a truthy value, and cases where it returns a falsey value.
• As long as a construct is syntactically valid and could possibly be of the correct scope/type/whatever, it can be called valid. No need to dig into a repository to fetch documentation for an import or anything like that.
• Constructs whose syntactic validity depends on the current runtime execution state do not need to be checked, as statically checking their validity would necessitate reasoning about program execution.
• Appealing to external programs/libraries or to the compiler/interpreter, runtime, or standard libraries to parse or execute the string as language instructions for you would circumvent the whole point of this challenge, and thus is forbidden.
• Standard loopholes apply.
• I don't think quine is correct. Also, you might want to be more explicit that there must be a false case, as languages like Seriously are always syntactically valid. – ATaco Oct 30 '17 at 22:31
• I wonder how many languages get trivial solutions to this. – Joshua Oct 30 '17 at 22:55

# Budget my lunch!

I'm a man of habit, but I also like to live quite frugally so every day I make myself the same lunch consisting of;

• 3 Falafel
• A serving of houmous
• Some lettuce

Now in order to make sure I always put aside enough money for my lunch I need a program that will calculate the total cost of these ingredients for x number of days, and that is where you come in!

# The Challenge

Write a method that, when given two integers representing the start day and the number of days to calculate for, will return/print out a the total cost of my food for that number of days. There are several rules that this calculation follows;

• Each pack of falafel contains 8 falafels and costs £1.50. However, every three days I get lucky and the pack has 9 falafel in it!
• Each pot of houmous contains 4 servings and costs £1.10.
• Each head of lettuce provides 7 servings and costs 46p (£0.46).
• Houmous and flatbread both go off over the weekend/I eat them so have to be bought fresh on Monday.

# The Input

You will recieve two integers as input. These can be read from STDIN, passed as individual arguments, or passed as an array. Up to you.

The first integer represents the starting day e.g. 1=Monday, 2=Tuesday etc. This will always be between 1 and 5 inclusive.

The second integer is the number of days to budget for not including weekends e.g. an input of 1 and 5 would be budgeting my lunch for each day of the standard working week.

# Sample program

Below is a throroughly ungolfed program to use as reference and clarification of the above rules.

public static double budgetMyLunch(int startDay, int days){
double totalCost = 0;

final double falafelCost = 1.5;
final int falafelsInPack = 8;
int falafels = 0;
int falafelPackCount = 0;

final double houmousCost = 1.1;
final int houmousServings = 4;
int houmous = 0;

final double lettuceCost = 0.46;
final int lettuceServings = 7;
int lettuce = 0;

int endDay = startDay + days;
int currentDay = startDay;

while(currentDay<endDay){

// If there are less than 3 falafels I have to buy some more
if(falafels<3){
falafelPackCount++;
falafels += falafelsInPack;
// Every third pack of falafel I get lucky and it has 9 instead of 8!
if(falafelPackCount%3==0) falafels++;
totalCost += falafelCost;
}

// If there are less than 2 flatbreads I have to buy some more
}

// If there are no servings of houmous left I have to buy some more
if(houmous == 0){
houmous += houmousServings;
totalCost += houmousCost;
}

// If there is no lettuce left I have to buy some more
if(lettuce == 0){
lettuce += lettuceServings;
totalCost += lettuceCost;
}

// Use up the ingredients for my lunch!
falafels -= 3;
houmous -= 1;
lettuce -= 1;

// If it's the end of the week, I eat the remaining
// houmous and flatbread over the weekend
if(currentDay%5==0){
houmous = 0;
}

currentDay++;
}
}


# Questions for sandbox

• Any suggested tags other than ?
• Is the question of appropriate complexity to be an interesting challenge? Should I add more rules?
• Is the sample program required or are the rules clear enough on their own? If it is required should it be included in the question (as above) or linked through to somewhere like Pastebin?

# UTF-8 repair

UTF-8 is the standard by which most of the world operates. Software that was intended to take in other encodings have been modified or wrapped to use UTF-8 as input and output, some with bugs and/or sloppy adherence to the standard. Even rock-solid software with years-old implementations of UTF-8, such as Java, can contain numerous UTF-8 implementation errors.

## The challenge

Given a string of characters as input, apply a list of transformations to it. The transformations can be applied in any order, but should have the same result as if the transformations were applied in the listed order.

• Replace overlong sequences with their equivalent minimal-length encodings.
• If a byte taking the form 0b11111xxx occurs, replace it, and up to three trailing bytes of the form 0b10xxxxxx, with a single replacement character (\uFFFD).
• Replace any run of extraneous bytes taking the form 0b10xxxxxx with a single replacement character.
• If a byte of the form 0b11xxxxxx is followed by an insufficient number of bytes of the form 0b10xxxxxx (either because of a byte of the borm 0b11xxxxxx or the end of the file/string), replace it and any trailing 0b10xxxxxx bytes with a single replacement character.
• Convert a high surrogate character followed by a low surrogate character to the utf-8 encoding of the supplemental plane codepoint it represents.
• Replace extraneous surrogate characters with the replacement character.
• Remove the initial byte order mark (\uFEFF), if one occurs at the start of the string.
• Keep all other characters intact. Make particular care to make sure that null characters are handled correctly.

The following are errors:

• An inverted byte order mark (\uFFFE) occurs at the start of a string.
• The transformed string consists of more than 16 codepoints, and more than 1/8 of them correspond to bad byte sequences.
• The input string consists of more than 32 bytes, and more than 1/8 of them do not correspond to a valid character, an overlong character, or a surrogate character.

If an error occurs, use an error indication mechanism (return a null pointer, utilize a tagged union indicating an error condition, set an global error variable, throw an exception, raise a signal, whatever), and do not output anything other than an optional error message. Otherwise, print the whole transformed string.

Standard loopholes apply. As with all code golf challenges, shortest submission wins.

## Test cases:

To be added if this submission gains traction.

## Reverse a regex

The .NET regex engine has the ability to process a regex in right-to-left order. There are some subtle differences between right-to-left and left-to-right order, so for the purposes of this challenge, right-to-left order is equivalent to running a reversed regex on a reversed string.

But how do you reverse a regex? Aha, that's where you come in. You need to write a program or function that reads a regex and writes one that will match reversed strings.

Fortunately reversing regexes is quite simple: break them down into symbols and concatenate them in reverse order. The symbols you should support are as follows:

• Unquoted nonmagic characters
• Quoted characters
• Character classes (anything from [ to an unquoted ])
• Symbols with repetition suffixes (assume that any of +*? or {...} is a valid repetition suffix)
• Parenthesised regexes (the inner regex will itself need to be reversed)

So for instance the reverse of a{2,}(\*[b-g])+ is ([b-g]\*)+a{2,}.

TODO: More examples, rules, loopholes.

Should I add lookaheads/behinds to the list or is that too complex?

• Character classes and repetition suffixes, at the very least, need more detail. Obviously some advanced repetition suffixes simply can't work, but which ones should be implemented? – Peter Taylor Nov 3 '17 at 11:16
• @PeterTaylor I don't understand why they can't work. You'll be able to assume that the input is a valid regex, so you'll be able to treat a+? as the symbol a+ with a suffix of ?. – Neil Nov 3 '17 at 13:42
• I had possessive quantifiers particularly in mind. I do retract the "obviously": now that I'm trying to think of an explicit example, it's not easy. But I'm still confident that one can be constructed. – Peter Taylor Nov 3 '17 at 18:26

# Find local Minimum

META: This is just meant as an introductory challenge for functional input (black box functions) and hopefully establish some more inputs.

Given a real number e > 0 and some black box function f: (0,1) → (-10,10) ⊆ ℝ, return a value x that is within e of a local minimum x* that means x ∊ [x* - e, x* + e].

### Details

• A value x* is local minimum of f if and only if there is some a > 0 such that f(x*) < f(x) for all x ∊ (x* - a, x* + a) \ {x*}. This means all other points in some neighbourhood must have a strictly greater function value. (This definition is just for this challenge and might be different from other definitions.)
• The input function f can assumed to be differentiable and satisfy f'(0) < 0 and f'(1) > 0 which means there are no local minima at 0 and 1.
• You can assume that |f'(x)| < 1000 this means the function cannot be to steep.
• You can use your native floating point numbers and assume e is sufficiently great. (That means greater than the machine precision of your floating point type.)
• You can use any language for which there are defaults for black box functions input in the standard IO meta post. If there are no such default for your language feel free to add one in the sense of the definition of black box functions, and make sure to link your proposals in that definition. Also don't forget to vote on them.

### Examples

x*                 f(x)
0.5                x*(x-1)
0.5                (x+1/2)^(2*k-1) * (x-3/2)^(2*k-1) for k=1,2,...,1000
2*(k+1/4)/(2*n-1)  -sin(pi*(2*n-1)) for n=1,2,3,...,159 (for every applicable k there is one minimum)

• FWIW there was a black-box function question a long time ago: codegolf.stackexchange.com/q/3839/194 . On a separate issue, being differentiable is quite a weak condition: is it strong enough to make this answerable? I don't know the whole zoo of nasty counterexamples in analysis, but ISTM that to have a local minimum you need a second derivative, and to solve for a black box you probably want at least the second derivative to be continuous. – Peter Taylor Nov 6 '17 at 14:47
• Thank you for the feedback. You're right, right now it is not answerable, I found an counterexample, and also with requiring second derivatives we can solve it as a black box function. – flawr Nov 6 '17 at 20:13

I asked this question in main site. Got a very bad review. So, I would be glad if any kind soul here would like to help me make this question better. And Some feedback on how can i make it good.

Given any input, you need to encode / decode it in or from base 16, 32, 64.

So your task will be to make a base 16, 32, 64 encoder / decoder.

RULES:
Standard Loopholes apply.
and that's all
INPUT:
In any format you like.
The inputs will be the
i. Data
ii. Task (encoding or decoding) (0 or 1)
iii. Base (16 or 32 or 64)

Output:
The Encoded/Decoded data


The shortest code in byte wins.

Reference : RFC

• Just as a starter, I would remove the "It should not be a builtin" (see this) Secondly, rather than link to a website with a description of what you mean, include that description in your question. Finally, allow function answers, rather than restricting to a full program. – caird coinheringaahing Nov 5 '17 at 16:10
• @cairdcoinheringaahing thanks. I will add those changes – Kishan Kumar Nov 6 '17 at 2:36

# The Guardian of the Chessboard

Given a collection of chess squares1, output the chess piece with the smallest value, together with its position, that is able to reach all the given squares in a single step.

1 - A chess square is a notation of a position on an 8x8 grid, with the x-axis being labeled with letters instead of numbers, meaning for example that the notation of 3|2 would be c2.

## Input

Input must be received as a collection of strings, each string representing a chess square. You may assume that the chess square will always be in the range a1 - h8.

Example Input: [ "b7", "c4", "h1", "g8" ]

## Output

Output must be in the format [piece][square], with piece being the notation of the chess piece, such as N, Q or K, and square being the notation of the square that the chess piece has to be located at.

You may either return a string from a method or directly output the result to stdout.

## Rules

• Note that you have to use the chess piece with the least possible value, so if there is a choice between for example Queen and Bishop, you would choose the Bishop. See below for a table noting the piece values.
• Special moves such as pawn's first move, en passent and castling do not have to be respected.
• This is , shortest code in bytes, in any programming language wins.
• Standard loopholes are forbidden.

## Data Table

Here you can look up each pieces value and its notation character.

|  Piece | Value | Character |
|:------:|:-----:|:---------:|
| Pawn   | 1     | P         |
| King   | 2     | K         |
| Knight | 3     | N         |
| Bishop | 3     | B         |
| Rook   | 5     | R         |
| Queen  | 9     | Q         |


## Test Cases

Input -> Output

[ "b7", "c4", "h1", "g8" ] -> Bd5


# Sandboxing

• Duplicates?
• Possible misunderstatings?
• Reach in exactly 1 step? Is 0 step allowed? What about multiple steps? – user202729 Nov 6 '17 at 1:23
• @user202729 Thanks, clarified. – Ian H. Nov 6 '17 at 8:55
• Needs a mention of special cases (pawn's first move, en passant, castling): should we assume that none of them are available? – Peter Taylor Nov 6 '17 at 14:35
• @PeterTaylor That's correct, I added it to the rules section, thanks! – Ian H. Nov 6 '17 at 16:27

# Ternary Parity of Substrings

## Definition

The ternary parity of a string is the number of 1's in its ascii bitstring mod 3. For instance, the ascii values for "Hello World!" are:

H:72 e:101 l:108 l:108 o:111 space:32 W:87 o:111 r:114 l:108 d:100 !:33

Converting each ascii value to binary gives:

H:1001000 e:1100101 l:1101100 l:1101100 o:1101111 space:100000 W:1010111 o:1101111 r:1110010 l:1101100 d:1100100 !:100001

Concatenating these together gives the combined bitstring:

1001000110010111011001101100110111110000010101111101111111001011011001100100100001

Which has 45 1's. As a result, the ternary parity of "Hello World!" is 45%3 = 0.

## Challenge

Write the shortest program in the language of your choice that does the following:

• Takes a string, s, as input.
• Finds S, the collection of all substrings of s
• Calculates the ternary parity of each element of S
• Creates a ternary string, q, by concatenating the parity bits of each element in S
• Outputs the ternary parity of q. That is, the number of 1's in q mod 3.

## Input/Output

Input and output may use any of the standard methods listed here. The program must output three distinct values that indicate ternary parity. For instance, the program could print "one", "two", or "three", it could exit with an error code of 0, 1, or 2, or could be a function that returns False for 1 and Null for 2, and 100 for 0, etc.

## Test Cases

"Hello World!" => 0
"foobar" => 2
"ABCDEFG" => 1
"abcdefg" => 0
"One" => 2
"Four" => 0
"2049" => 1

• One element of a good question is a clear motivation. This looks like a random mishmash of operations. Why should anyone care about the result of this calculation? – Peter Taylor Nov 7 '17 at 10:14

# Output an Anagram! No Not That One!

Given a list of unique strings that are anagrams of each other, output an anagram of those words that is different from each word in the list.

The strings will be alphanumeric, and there is guaranteed to be a valid anagram.

The program or function can, but doesn't have to be non-deterministic, meaning given the same input, multiple running a of the code can yield different outputs, as long as every possible output is a valid one.

### Test Cases

[Input] -> Possible output
-----------------
[ab] -> ba
[aba, aab] -> baa
[123, 132, 231, 312, 321] -> 213
[hq999, 9h9q9, 9qh99] -> 999hq
[abcde123, ab3e1cd2, 321edbac, bcda1e23] -> ba213ecd


## Sandbox questions

• Is this a duplicate of anything?
• Any other test cases I should include?
• I should probably post my other sandboxed challenge at some point... – MildlyMilquetoast Sep 27 '17 at 1:23
• Should it output the same string each time given the same array of strings as input? – Shaggy Sep 27 '17 at 10:17
• What do you mean by "unique" in the first sentence? – Peter Taylor Sep 27 '17 at 19:43
• @PeterTaylor Did that clear it up? – MildlyMilquetoast Sep 27 '17 at 20:42

# Clear the centrally significant on bits

Related and inspired by.

# Input

Input is a single positive integer n.

# Output

It's easiest to describe this by example.

n = 433


Take n's binary representation.

bin(n) = 110110001


Note which bits are on.

110110001
^^ ^^   ^


Set the center bit of those that are on to 0. (if there is an even number of 1 bits, set both to 0).

110010001
^^ !^   ^


Finally, represent the input as an integer again.

unbin(110010001) = 401


## Test cases

1 => 0
13 => 9
115 => 99
236 => 204
433 => 401
732 => 652
1555 => 1539
1556 => 1028


# Bring an end to the Vigil

Vigil, being the supreme moral paragon of programming languages, inspires us all to write bug-free code without exceptions through supreme medieval punishment.

However, every hero has their weakness, and Vigil's weakness is on line 98:

except:
print("Vigil has failed to uphold supreme moral vigilance.")


Your goal is to write a Vigil program that forces Vigil to experience an exception, reach this line, and print this error. Because it's not enough to merely succeed at our master plans, but instead we ought to succeed efficiently, the shortest answer in bytes wins.

### Draft proposal notes

• This is a language-specific challenge in the Vigil language alone. (Is there a tag for that?)
• I'm not sure if a standard challenge is the best way to measure a winner in this scenario, but it's the primary measure that comes to mind.
• I'd like to ham this up a bit more with narrative. Maybe about us being the villains. The title could be wittier.

## Background

Radiation hardened quines have been around for a long time. The basic idea is that it's (possibly) a valid quine, and if you remove any character, it's still valid. But what about the other way around? What if you added a character?

## Challenge Description

Develop a program which will output its source code, even when a character is inserted in any position in the source code.

## Input

This program is a quine. It shouldn't take any input.

## Output

When the quine runs, it should either output its own source code, or the source code of the proto-quine.

## Scoring

Programs are scored on their robustness. That is, for a given program, if the characters "a", "!" or "😷" could be put into the program at any point and it remain a quine, then the program has a score of 3. I leave the burden of testing what characters work for your program up to you. Note that in languages with a limited character set (such as HQ9+ or brainfuck) characters that do nothing yield no score. You couldn't write a normal quine in brainfuck and then claim that every single unicode codepoint except the few used in a brainfuck program was your score.

Finally, other than that, standard rules apply.

• Not a bad challenge, but could use some more fleshing out. Take a look at the framework of some other challenges, both here and in main, to see what the usual framework is. – Gryphon Nov 17 '17 at 0:58
• To clarify, I mean specifically the usual: brief story about the challenge (optional), "Your Task:", task, "Input:" input methods, "Output:", output methods, "Examples:", examples, "Scoring:", scoring method. That framework works well for most challenges – Gryphon Nov 17 '17 at 1:00
• Can we chose where the character is inserted? – ATaco Nov 17 '17 at 1:13
• @ATaco no, it has to be valid for anywhere. Otherwise, somebody would make a safety box for the character. – Jakob Lovern Nov 17 '17 at 1:21
• I would suggest clarifying your scoring section, specifically specifying "robustness" better. I'm pretty sure I get what you mean, but others may misinterpret that. Other than that, this looks like a great challenge. +1 – Gryphon Nov 17 '17 at 1:57
• Is the emoji bad? I put it in there to demonstrate únicode, but I'm not sure the point is that obvious. – Jakob Lovern Nov 17 '17 at 2:41
• Nah, it's fine. The example makes it more clear. I might add something to specifically say that it doesn't count if a character can be inserted twice (or at least, that's my understanding of the rules), but other than that, you should be good. – Gryphon Nov 17 '17 at 3:24
• To be certain, when you say it should output "it's own source code, or the source code of the proto-quine" then, after a character is added, your code may either output the original code or the modified code? Is it alright if a single submission does one or the other depending on the type or location of the inserted character? – Kamil Drakari Nov 17 '17 at 15:34
• @KamilDrakari yeah, as long as it does do at least one of the two. – Jakob Lovern Nov 17 '17 at 18:16

# Rhythm Generator

Your task: given a number n from 2 to 5 inclusive, (pseudo)randomly generate a valid rhythm in the time signature n/4, and output using the system explained below. It must have an equal chance of choosing each valid rhythm.

### Background Information:

In music, collections of beats are called "measures." You can only fit a certain number of beats in a measure, determined by the time signature. For example, a time signature of 4/4 would mean that you can put 4 (as determined by the first number) quarter note (determined by the second number) beats in the measure.

There are different kinds of notes. For the purposes of this challenge, the following characters will represent different kinds of notes (and how many quarter note beats each one is worth):

W        Whole note                4
H        Half note                 2
K        Dotted Half note          3
Q        Quarter note              1
D        Dotted Quarter note       1.5
E        Eighth note               0.5
F        Dotted Eighth note        0.75
S        Sixteenth note            0.25


(There are other kinds of notes but they do not apply to this challenge)

Note: A dotted note takes 1.5 of the length of its non-dotted counterpart.

In a n/4 measure there can only be n quarter note beats. Some notes take more than a beat, some less. But the total number of beats needs to add up to n.

For example, here are some valid 4/4 measures (separated by newlines):

W
HH
QQQQ
EEEEEEEE
SSSSSSSSSSSSSSSS
KEE
EQQQE


And here are some examples of valid 3/4 measures:

K
DD
QQQ
QDE


So, your task is to generate a rhythm and output it in that format.

So now let's make some music!

### Test Cases:

Coming soon. Will be a complete list of all possible valid rhythms.

# Moore-Penrose inverse

In linear algebra the pseudo-inverse of a matrix is a generalization of the inverse of a matrix which doesn't necessarily exist. One definition of a pseudo-inverse is called the Moore-Penrose inverse which shall be denoted by A⁺ in the following. Further let Aᵀ denote the transpose of the matrix A.

## Definition

The Moore-Penrose inverse of the matrix A is defined for all matrices A in a way such that it satisfies the following four properties:

• AA⁺A = A
• A⁺AA⁺ = A⁺
• (AA⁺)ᵀ = AA⁺
• (A⁺A)ᵀ = A⁺A

Your task in this challenge is, given a matrix real A (non-zero and non-empty) to compute its pseudo-inverse as defined above.

• You're allowed to receive input as a single list, array, list of lists etc.
• The input is a single non-zero, non-empty matrix with real entries
• The input matrix is not necessarily square
• Your program/function either returns the pseudo-inverse or prints it to STDOUT
• You don't need to handle floating point inaccuracies
• The behavior on invalid input (eg. []) is left undefined

## Examples

Here are some possible examples (note the different (in)accuracies):

[4] -> [0.25]
[1 1; 1 0] -> [0 1; 1 -1]
[12 4 9; -9 1 3; -6 6 15] -> [0.0467129 -0.0342457 -0.0217785; 0.0115298 0.0022497 0.0160292; 0.0251844 0.0076240 0.0404324]
[0 1 0; 0 0 1; 0 0 0; 2 0 0] -> [0 0 0 0.5; 1 0 0 0; 0 1 0 0]
[1 1 2 3] -> [0.0667; 0.0667; 0.1333; 0.2]
[0; 0; 1] -> [0 0 1]
[1 0 0 0 0; 2 0 0 0 0] -> [0.2 0.4; 0 0; 0 0; 0 0; 0 0]


### Sandbox

• Should I ban built-ins (looking at Octave/Matlab, Mathematica etc.)?
• Is the part with floating-point inaccuracies formal/objective enough?
• Are more examples needed?
• Having no challenge for computing a SVD, maybe I should reduce this to only computing the SVD of a matrix since that would be the heart of this challenge anyway. What do you think? – ბიმო Nov 22 '17 at 15:00

# Examining the Student's Swing

The 2017 Level 1 Mathematics and Statistics NCEA examination paper has been making headlines in New Zealand for being 'impossible' and bringing students to tears.

One of the questions involves a children's playground swing, and finding where the holes in the swing seat should be placed. But lets generalise the problem in code in the fewest number of bytes, in case they have to solve for a different swing (in a different exam).

Find how far apart the holes in the board need to be if the shape of the rope stays the same, given the following swing setup:

• The rope hangs from a cross beam at a height h, and will have a parabolic shape, as presented in the exam (rather than a catenary shape), and will maintain this shape with the seat.
• The rope is connected to the crossbeam at two points, that are separated by a distance d.
• The lowest point (vertex) of the rope sits above the ground at a height of v.
• We have an adequately sized wooden board for the seat, and we want this to be at a height s above the ground.

Exam questions often have a visual representation to help clarify the question, so I've also included one:

### Rules

• inputs must take positive real values
• s will be greater than v
• output must be correct to at least 2dp.
• This is , therefore, the lowest byte count in each programming language wins
• Standard rules apply, and no forbidden loopholes

### Test Cases

(h,d,v,s) -> hole separation
(4, 6, 1, 1.2) -> 1.549
(10, 4, 2, 4) -> 2
(14.5, 12, 2.5, 5.5) -> 6
(4.25, 3.4, 0.5, 1.2) -> 1.468


### Sandbox Questions

• First time question, so any recommendations would be great
• Any recommendations for how I can make this a well received question?
• are there any rules I'm missing, or should add?
• is this a duplicate?
• What other tags should this question have?
• About the output must be correct to at least 2dp. part, I'm afraid that there will be some problems from that requirement. A better requirement is "the algorithm in the code must be able to theorically calculate to infinite precision, but numerical error from programming language type limit is acceptable". / Also, the "the rope is always in parabolic shape" is weird, is that correct? If so you should highlight it. (normally a unweighed rope has catenary shape, and a weighed rope has the shape of multiple straight line segment) / Otherwise the challenge looks good. – user202729 Nov 22 '17 at 7:48
• Thanks for the suggestions; the 2dp was based on the exam question requirement of showing your working to 2dp, though I understand the precision issue. Maybe I just remove that rule? I didn't know about catenary shape, the exam question just used a parabola, so I'll highlight that fact. – Ayb4btu Nov 22 '17 at 8:39
• So in other words, given (h,d,v,s) output d sqrt((s-v)/(h-v)). I don't think it's a duplicate, but I also think that the reason that it isn't is because everyone has previously (and correctly) concluded that calculating a square root without restrictions to prohibit builtin sqrt functions is too trivial to be worth posting. – Peter Taylor Nov 22 '17 at 15:00
• @PeterTaylor Well, (you know) on this site it is generally discouraged to post a "puzzle", because once the first answer had been posted, the other answers can just use that - and your comment has already specified the method. / You meant sqrt should be prohibited? That will lead to the (well-known) do X without Y problem. – user202729 Nov 22 '17 at 15:15
• @user202729, no, I did not mean that. I meant that "Calculate sqrt without a builtin sqrt" would be a duplicate (although maybe only of closed questions). See codegolf.stackexchange.com/q/73/194 and linked questions. – Peter Taylor Nov 22 '17 at 15:25
• @PeterTaylor This exam question certainly simplified down to a trivial coding answer; I think I was more posting due to the story behind it. I'll have a think about if or how to make it a bit more challenging. Will just mean that it may not be as closely related to the exam question on which it is based. – Ayb4btu Nov 22 '17 at 20:29

# View float numbers in graphical binary

IEEE 754 is a well known format for representing floating point numbers.

Your task is to give an input to the user where he/she enters a signed float number and represents clearly the number in a color separated by functional area binary 32 bit single representation, like:

• You do not have to append the captions like they are in the figure.

• You must follow the order Sign, Exponent, Fraction

• You may choose whatever colors you want, as they are distinguishable from each other and from the binary text numbers.

• The input (I suggest a text box) does not have to forbid invalid values, but if does not, the graphical binary view needs to indicate the invalidity. You can leave it empty or replace with some text like "ERROR" or "INVALID".

• Every little change on the input needs to make the binary view to be immediately updated. Changing and have to confirm with an "OK" like button is not acceptable.

No winner, unless I see some very creative answer. I would like to see a golfed and an ungolfed version of the code.

• Needing the output to be dynamically updated based on the input changing (and not an OK or the like) is going to drastically limit the languages that can participate. – AdmBorkBork Nov 21 '17 at 18:58
• Additionally to what @AdmBorkBork said, the requirement of dynamically updating the output & validating input is just distracting from the real challenge. I'd just leave behaviour on invalid input undefined. – ბიმო Nov 22 '17 at 15:08

## Optimise Retina transliterations

For the question Translate Morse code based on tone duration I stupidly created the following transliteration:

TT\EMNAI\OGKDWRUSJ2V\HBZ89#@NIGDRS8ZCBP\LF\H12356789_\w@#


The first thing wrong with it is that it 2, 8 and 9 transliterate to themselves, so they are superfluous. This saves 6 bytes:

TT\EMNAI\OGKDWRUSJV\HBZ#@NIGDRS8ZCBP\LF\H13567_\w@#


The next thing to notice is that we have some runs of consecutive letters, but because they're not ordered we can't reduce them. Let's sort them:

TABD\EG\HIJKMN\ORSTUVWZ#@R6BIZ5S1CGD8\L\HNF3P7\w@#


We can now save another 3 bytes by using the run G-K, a byte by using the run M-O, and 3 bytes by using the run R-W:

TABD\EG-KM-OR-WZ#@R6BIZ5S1CGD8\L\HNF3P7\w@#


Total 13 bytes saved. Note that there are other optimisation opportunities but I think they might be too difficult be worth adding to the question; for instance, inserting a C (at a cost of 2 bytes) allows the creation of the run A-E for a 3-byte saving, which is still overall a byte saved; inserting all of the letters to achieve run A-Z would then be simplified to L.

Is the Retina command string too inflexible an input format? I suppose input could be in the form of an unordered mapping from printable ASCII to printable ASCII, but output would need to be the actual Retina syntax, including appropriate quoting, because the length of the result is important.

(Quoting: The characters -dEOHhLlwpoR\ need to be quoted except that the letters don't need to be quoted as part of a range.)

Should this be a question or ?

# A Knight's Walk

On a chessboard, a knight has two options for movement: it can move one square horizontally and two squares vertically, or two squares horizontally and one square vertically. In short, it can follow the path drawn by an upper-case "L". These can be oriented in any way, and as such, a knight has 8 potential moves from any given position.

There have been a number of challenges posted at one time or another regarding knights and their movement because they're so mathematically interesting (and the problems are easy to state) but we're going to go for a basic one that I haven't seen.

## The Challenge

Given two 2d coordinates (x1, y1), (x2, y2) return the minimum number of moves required to travel from (x1, y1) to (x2, y2).

There are no other pieces on the board, and you may assume that the coordinate values given are valid (a set of two integers) though they may not be unique. You may also assume that each individual coordinate value will be between 0 and 255, and that the board is large enough that you needn't consider edges.

## Input Format

As long as the coordinates are inputted in the order x1 y1 x2 y2 you can accept them as four separate values, two tuples, two lists, etc...

## Examples

(0, 0) (1, 2) => 1
(0, 0) (1, 1) => 2
(125, 125) (126, 127) => 1
(-100, -100) (0, 100) => 100

• You might want to more explicitly state that negative numbers must be supported. Also, 'less than 128' applies only to inputs? or to anything? – Rɪᴋᴇʀ Nov 24 '17 at 0:24
• Will do! And I'm not sure what you mean by that second part: I mean the input values will individually have absolute values of less than 128 a.k.a. (-127, -127) to (127, 127). This is totally a point I'm flexible on though. I just figured there may be languages that don't support very large numbers in any given memory location, and didn't want to give them additional hoops to jump through. Edit: Should I just make the valid range 0-255? – Willbeing Nov 24 '17 at 1:45
• I'd recommend making it 0-255, yeah. I don't really think it's fair to have input be -127, +127, and then the boundaries of the chesboard be infinity. That's just kinda contradictory. – Rɪᴋᴇʀ Nov 24 '17 at 2:14
• Do you think I should keep a bound like 255 at all? Would 2^16-1 or 2^32-1 be better? My whole premise with picking a number like this was that I want it to be easy to represent any number you'd run into, but impossible to use any kind of naive recursion to solve it – Willbeing Nov 24 '17 at 3:31
• "impossible to use any kind of naive recursion to solve it" - Why? I don't understand your argument. / What is the winning criteria? – user202729 Nov 24 '17 at 13:29
• @Willbeing I think 255 is fine, it doesn't matter too much. I've seen successful challenges that only require you to support up to 255, but that's it. – Rɪᴋᴇʀ Nov 24 '17 at 14:35
• @user202729 Yeah, it sounds odd when its played back to me, hahaha. I only care if the program deterministically produces the correct step count. I still think limiting the input size is probs a good idea, just not for that reason anymore. – Willbeing Nov 24 '17 at 19:21

# Clean duplicate website from my history

## Introduction

I am looking for help to monitor my history. I parsed everything, but I am not yet happy with the final result. Could you help me? I want to remove following same website.

My history looks like this:

| url                                               | id |
|---------------------------------------------------|----|
| https://codegolf.stackexchange.com/q/147318/15214 | 4  |
| https://codegolf.stackexchange.com/q/148927/15214 | 4  |
| https://en.wikipedia.org/wiki/Code_golf           | 3  |
| http://lichess.org                                | 2  |
| https://en.wikipedia.org/wiki/Code_poetry         | 1  |


But I want it to be:

| url                                               | id |
|---------------------------------------------------|----|
| https://codegolf.stackexchange.com/q/148927/15214 | 4  |
| https://en.wikipedia.org/wiki/Code_golf           | 3  |
| http://lichess.org                                | 2  |
| https://en.wikipedia.org/wiki/Code_poetry         | 1  |

• If I browse a site, then another one, then the first one. I keep all of them.
• If I browse a site, then another part of the same site. I keep only the first one.
• The list is ordered from young to old. (we keep the older)

This challenge is inspired by my day work, but it is not related in any ways with anything in it.

## Challenge

• Input is a list of identifier.
• Output is a list of two elements (identifier's index, identifier).
• No Empty input
• No standard loophole
• This is code-golf, so shortest answer in bytes wins.

## Examples Input and Output

### 2 links on website 1

Input:

[ 1, 1 ]


Output:

[
[1,1]
]


### 1 link on website 2, then 3 links on website 1

Input:

[ 1, 1, 1, 2]


Output:

[
[2,1],
[3,2]
]


### 1 link on website 1, then 1 link on website 2, then 2 link on website 1

Input:

[ 1, 1, 2, 1]


Output

[
[1,1],
[2,2],
[3,1]
]
`

## Question

• Should I allow outputs as list of object, list of tuple or dictionary/map?
• Should I ask for url parsing where identifier is the hostname?
• Question title and tag? – user202729 Nov 24 '17 at 14:11
• Also I think the I/O specification (Input is a list of two elements. / Output is a list of two elements.") seems to contradict the test cases. – user202729 Nov 24 '17 at 14:33
• @user202729 Update title, tag and I/O specs. Thank you. – aloisdg moving to codidact.com Nov 24 '17 at 14:53
• You definitely should allow using tuples, maps and dictionaries for input and output. In fact, that's an understatement. You should read the meta post on I/O methods. URL parsing will make the challenge more interesting, IMO. – NieDzejkob Nov 24 '17 at 19:12