The Nineteenth Byte

Kritixi Lithos

12:00 PM

I get 5 bytes for Jelly, İS÷@L

totallyhuman

for the harmonic mean one: is lambda l:3/sum(i**-1 for i in l) correct?

idk much about it

Kritixi Lithos

1/i is shorter

Leaky Nun

only for lists with 3 elements is it correct @totallyhuman

New Main Posts

-2

Q: Unwrap an variable number of wrapper types in a single operation

I have a generic wrapper type: public class Wrap<T> { public T Contents { get; set; } } I can wrap any object in a wrap or unwrap a wrapped object: public static class WrapHelpers { public static Wrap<T> Wrap<T>(this T contents) { return new Wrap<T> { Contents = contents }; ...

c#

Kritixi Lithos

@LeakyNun is it possible in just 4 bytes?

Leaky Nun

12:02 PM

replace 3 with len(l) and you're good @totallyhuman

@KritixiLithos it = ?

totallyhuman

ohhhh

sorry i just looked it up on wikipedia :P

Kritixi Lithos

@LeakyNun the Jelly program to your first CMC

@LeakyNun Dyalog APL, `↑,\`

Leaky Nun

@KritixiLithos I don't think it is

Kritixi Lithos

formatting borked, it is 3 bytes in APL

Leaky Nun

@KritixiLithos nice

should I post my second one on main?

Kritixi Lithos

12:05 PM

yes

ASCII-only

@LeakyNun Charcoal, 8 bytes: ＳαＧ↘←Ｌαα

Kritixi Lithos

the first one, if it isn't already a duplicate, can wait if you want to post it. Let's not have too many alphabet challenges together again

I get 7 bytes in Retina: Try it online!

with two trailing newlines

s/8/7/

Erik the Outgolfer

@LeakyNun but decrement that number and it is ;)

Leaky Nun

CMC: given positive integers a and n, apply the collatz function on a, n times.

Kritixi Lithos

btw I was talking about his earlier CMC

user165474

12:11 PM

@EriktheOutgolfer lol that's exactly what I did

user165474

like -2 mins

user165474

you can FGITW i'm going for a bounty right now

Erik the Outgolfer

what bounty?

Leaky Nun

nobody is getting my 3-byte answer

@EriktheOutgolfer try to use my style in the abcab CMC

ASCII-only

@LeakyNun JavaScript, 33 bytes: f=(a,n)=>n--?f(a%2?a*3+1:a/2,n):a

Kritixi Lithos

12:14 PM

19 mins ago, by Leaky Nun

CMC: harmonic mean of a list of positive floats

Erik the Outgolfer

@LeakyNun oh, you mean something like this?

Kritixi Lithos

@ASCII-only use curry to save bytes

totallyhuman

Python 2, 45 bytes: def f(s,i=1):print s[:i];i-len(s)and f(s,i+1)

Erik the Outgolfer

@KritixiLithos huh?

Leaky Nun

@EriktheOutgolfer precisely

Anonymous

12:15 PM

@LeakyNun Python 2, 46 bytes: f=lambda a,n:n and f([a*3+1,a/2][a&1],n-1)or a

ASCII-only

@KritixiLithos Doesn't save bytes

Kritixi Lithos

@EriktheOutgolfer Leaky said "I don't think it is" to my message about doing that CMC I posted and not the newer CMC

Erik the Outgolfer

oh I was confused...

Kritixi Lithos

@ASCII-only oh it's a recursive

Erik the Outgolfer

@KritixiLithos seems impossible to bring under 5

here is my version

user165474

12:19 PM

I got İSİ×L but I don't think I can get it any lower.

ASCII-only

Time ti try it I guess

Erik the Outgolfer

@HyperNeutrino you must at least (afaict) have 1 arithmetic operator, 2 reciprocals, 1 sum and 1 length, those sum up to 5, so I don't think you can get it any lower

user165474

True. Yeah I think it's not possible then.

Erik the Outgolfer

coincidentally, those 5 are all 1-byte atoms

user165474

well actually

Leaky Nun

12:21 PM

CMC: define a bijective function from the natural numbers to itself, such that the image of the even numbers under that function is a proper subset of the even numbers

user165474

Kritixi did it using reciprocal -> sum -> division [reversed] -> length

Erik the Outgolfer

[reversed] corresponds to reciprocal

user165474

@LeakyNun ??? So that means that some even numbers have to be mapped to odd numbers? Or am I being dumb

user165474

@EriktheOutgolfer not quite but I guess

user165474

well it does the same thing ¯\_(ツ)_/¯

Erik the Outgolfer

12:22 PM

the reciprocal of a / b is b / a that's why

Kritixi Lithos

heh

Leaky Nun

@HyperNeutrino Some odd numbers need to map to even numbers

user165474

What do you mean by image? I think I get what it means but I'm not sure based on your last message.

user165474

@EriktheOutgolfer whoops >_< I'm just being dumb ~~today~~ always don't mind me

Leaky Nun

f[B] subset B, where B is the set of even numbers

Kritixi Lithos

12:24 PM

Would f=B=>B+1 work?

Leaky Nun

f[B] means {f(b) : b in B}

user165474

Oh okay that makes sense thanks

Kritixi Lithos

what does f[B] mean?

Leaky Nun

@KritixiLithos the image of the even numbers is the odd numbers, which is not a subset of the even numbers

user165474

32 secs ago, by Leaky Nun

f[B] means {f(b) : b in B}

Leaky Nun

12:25 PM

@KritixiLithos 2 messages above you

Erik the Outgolfer

@LeakyNun f(x ∈ B) = x?

where B is the set of even numbers

Leaky Nun

@EriktheOutgolfer proper subset

user165474

Wait so how is it possible to make it bijective then... nvm there are infinite elements

user165474

f(x) = x^2? random guess

Kritixi Lithos

what does a proper subset mean?

Leaky Nun

12:27 PM

:37966206 that's left to you

@HyperNeutrino not bijective

Erik the Outgolfer

@HyperNeutrino not bijective

Leaky Nun

@KritixiLithos subset but not equal

Erik the Outgolfer

@LeakyNun that's what happens when you reply to a removed message I guess :P

user165474

hm

Mayube

@Mego thanks for taking the time to prove that js is pretty dumb in some cases

user165474

12:28 PM

Wait how so not

Mayube

@Mego i.imgur.com/suE1nQK.png

Anonymous

@Mayube It's less JS and more IEEE-754 floats. But JS is dumb for only using doubles.

user165474

Leaky Nun

@HyperNeutrino bijective is injective and surjective

your function is only injective

user165474

12:29 PM

@LeakyNun which one is which again? I can't remember >_<

ASCII-only

@HyperNeutrino It has to be a 1:1 mapping

user165474

Oh wait I ... nvm just being dumb

Mayube

@Mego I suspect that the issue may still be due to IEEE-754 floats storing the exponent in only 52 bytes, but js doesn't shorten the number at all as it tends to do for larger numbers, so I'm not sure

Leaky Nun

@HyperNeutrino what is mapped to 2?

user165474

Ok I see.

Erik the Outgolfer

12:30 PM

@LeakyNun f(x ∈ B) = x + x mod 2 * 2 - 2

does this count as proper subset?

user165474

CMC: Given a group G and an element a in G and a positive integer k, determine if the cyclic subgroups generated by a and a^k are equal. You may assume that |<a>| is finite.

ASCII-only

@EriktheOutgolfer Wait this isn't bijective nothing maps to the odd numbers

user165474

And an operator f as a function (though theoretically you don't need this btw)

Leaky Nun

@HyperNeutrino basically "is k coprime with the order of a?"

user165474

Does such a mapping exist?

user165474

12:31 PM

@LeakyNun darn you're too fast

user165474

:P

user165474

yeah that's what the problem reduces down to

Leaky Nun

@HyperNeutrino yes of course

Anonymous

@Mayube No, JS follows the IEEE-754 standard. The standard is the issue - accuracy of computations decreases with the size of the input(s). After 2**53-1, not every integer is representable, so that fact can be (ab)used to find values n where (n**.5)**2 != n. That is true in any language that uses IEEE-754 floats.

user165474

ok just wondering :p

Erik the Outgolfer

12:32 PM

@HyperNeutrino yeah his speed can be too fast sometimes

Leaky Nun

@EriktheOutgolfer what does f(x in B) mean?

user165474

@EriktheOutgolfer Wait so if his speed is too fast, then his speed has a high speed, which means he has a good acceleration rate? :p

Erik the Outgolfer

possibly...

@LeakyNun am I required to allow odd numbers too?

Mayube

@Mego right but surely in that case it can then be said that those n values are not accurately representable by the language, and thus not valid inputs?

Leaky Nun

@EriktheOutgolfer of course

user165474

12:33 PM

@EriktheOutgolfer "Bijective function from the natural numbers to itself"

Kritixi Lithos

@HyperNeutrino does F=x=>x+2 work?

user165474

What maps to 1?

Kritixi Lithos

-1

user165474

Not a natural number

Anonymous

@Mayube In the case of 2**54-2, it is representable within the language, but ((2**54-2)**.5)**2 is not, so the inaccuracy causes incorrect results.

Erik the Outgolfer

12:34 PM

@LeakyNun hmm, then it's more complicated than that...

user165474

Also, proper subset. :P

Mayube

@Mego hmm alright, I'll leave my answer as deleted then

user165474

?

user165474

Or is 0 in N

ASCII-only

12:36 PM

@LeakyNun Does this count: n=>n%2?n===1?0:n-2:n+2 :P

Leaky Nun

@HyperNeutrino up to you

user165474

yay

user165474

@ASCII-only What does 1 map to?

Erik the Outgolfer

@LeakyNun you say that the image of the set of even numbers under the function must be a proper subset of them, while you also say that some odd numbers must map to even numbers, while the function is bijective. then, though, since the function is bijective, the even numbers resulting from odd numbers must return the odd numbers, however this violates the image rule...impossible?

Leaky Nun

@HyperNeutrino 0 of course

Anonymous

12:37 PM

@Mayube Clearly the solution is to use a language that uses symbolic computation :P

user165474

oh wait I forgot about the extra ternary

ASCII-only

Wait 0 isn't a natural number whoops

Anonymous

Bonus points if that language has a builtin that is short... Say, 9 bytes or so?

Erik the Outgolfer

@ASCII-only it is actually

user165474

@ASCII-only It's up to you whether or not it is

Leaky Nun

12:38 PM

and @ASCII-only is the winner

user165474

congrats! :D :P

user165474

ok anyway I'm going back to work cya o/

Leaky Nun

@EriktheOutgolfer how does that violate the image rule?

the rule is that even number -> even number

some odd numbers also -> even number

just that no even number -> odd number

Erik the Outgolfer

how would it be bijective then?

ASCII-only

n=>n%2?n%4===1?n+1:(n+1)/4:n*2 should work as well?

Leaky Nun

12:41 PM

it is surjective and injective

ASCII-only

f=n=>n%2?n%4===1?n+1:(n+1)/4:n*2;
Array(25).fill().map((_, n)=>f(n))

Erik the Outgolfer

@LeakyNun oh you mean that every result is only mapped to by one natural number, and that one natural number doesn't map to any other results?

but not necessarily f(f(x)) = x?

ASCII-only

@EriktheOutgolfer Yeah, 1:1 mapping

Leaky Nun

@EriktheOutgolfer f f x = x is involution...

Erik the Outgolfer

yeah...

*really giant facepalm*

Leaky Nun

12:44 PM

why would bijection imply involution...

orlp

what are you guys talking about?

Erik the Outgolfer

@LeakyNun btw those terms always confused me

Leaky Nun

@orlp define bijective f:N->N such that f[2N] is a proper subset of 2N

orlp

2N is the set of even numbers?

Leaky Nun

yes

orlp

12:45 PM

and N = N0 or N+ here?

Erik the Outgolfer

@LeakyNun wait would f(x) = 2x work then?

Leaky Nun

@EriktheOutgolfer not surjective

@orlp up to you

Erik the Outgolfer

@LeakyNun huh? any example?

Leaky Nun

@EriktheOutgolfer that is only injective

@EriktheOutgolfer what is mapped to 1?

Erik the Outgolfer

2

and what is mapped to 2? 4

oh wait that isn't surjective?

ASCII-only

12:47 PM

@EriktheOutgolfer 1 is mapped to 2 in your function, not the other way round

orlp

@LeakyNun I don't believe it's possible to have it be a proper subset

every even number must still be even after f has been applied

Erik the Outgolfer

@ASCII-only Leaky said not necessarily involution though

Leaky Nun

@orlp continue

orlp

actually never mind

Erik the Outgolfer

@LeakyNun oh I see what you mean

Leaky Nun

12:48 PM

@orlp scroll above for a solution

orlp

no

Leaky Nun

@ASCII-only here @orlp

orlp

@LeakyNun I said I didn't want to see it

Leaky Nun

oh sorry

orlp

f(n) = 2^n if n is even, otherwise smallest positive value not taken by any f(k) with k < n

ASCII-only

12:50 PM

@LeakyNun btw idk if this is also valid

Leaky Nun

@orlp good but not golfy

Erik the Outgolfer

@LeakyNun so...surjective means that if f(N) returns N then all inputs must correspond to all numbers in N?

Leaky Nun

@EriktheOutgolfer surjective only means that f(N)=N, dunno wt the conditonal means

Erik the Outgolfer

that's what I understand at least

@LeakyNun oh so f(x) = 2x is f(N)=2N not f(N)=N which is not surjective right?

Leaky Nun

@EriktheOutgolfer correct

Erik the Outgolfer

12:52 PM

and I can't use bitwise xor either, since f(2N) must return a number in 2N

flawr

@LuisMendo matl ftw=)

orlp

f(n) = 2*n if n is even, otherwise n+1 - ceil((n+1)/3) + 1

Kritixi Lithos

would f=n=>n%2?n-1:n+1 work?

ASCII-only

@KritixiLithos No, even numbers don't map to a proper subset of even numbers

orlp

which can be simplified

f(n) = 2*n if n is even, otherwise n - floor(n/3) + 3

Kritixi Lithos

12:56 PM

so the set of even numbers have to map to the set of even numbers?

ASCII-only

@LeakyNun Wait it has to be golfy? 18 bytes, ES6

@KritixiLithos Yes but not every even number so e.g. an odd number must map to 0, or maybe every number divisible by 4

Erik the Outgolfer

of course it has, we're golfers here

orlp

eh actually my 'answer' is broken

@LeakyNun ok looking at that solution it's not that interesting =/

that's just removing 1 element to get the subset

ASCII-only

17 mins ago, by ASCII-only

f=n=>n%2?n%4===1?n+1:(n+1)/4:n*2;
Array(25).fill().map((_, n)=>f(n))

This is slightly less boring but still super simple

Leaky Nun

@orlp nobody said it had to be interesting

it just has to be golfy

orlp

12:59 PM

@LeakyNun then make it a popcon :D

Leaky Nun

every popcon is closed

orlp

how can a CMC be closed?

Leaky Nun

alright, it's a popcon then

Anonymous

@orlp I can show you if you really want :P

Leaky Nun

@EriktheOutgolfer f[2N] is a set, so "f(2N) must return a number" is ~~nonsense~~ false

Anonymous

1:01 PM

-> 1 CMC moved to Trash

Mayube

wow

Erik the Outgolfer

@Mego please don't, this is actually a good cmc

Leaky Nun

@EriktheOutgolfer he's just joking

orlp

I believe he's joking people

Kritixi Lithos

@Mego looking for this symbol →?

Erik the Outgolfer

1:02 PM

@LeakyNun ikr what do you think I'm serious?

ASCII-only

JavaScript (ES6), 29 bytes: n=>n%2?n%4==1?n+1:(n+1)/4:n*2

Erik the Outgolfer

@LeakyNun Python 2, 27 bytes: lambda x:x%2and x-x/2or x*2 (x/2 is the same as floor(x/2))

Adám

@KritixiLithos You've got your APL keyboard always on, right?

8

Leaky Nun

@EriktheOutgolfer what does 2 and 7 return, respectively?

ASCII-only

@EriktheOutgolfer Not valid

Erik the Outgolfer

1:06 PM

oh injection fail

ASCII-only

@Adám Well I use WinCompose for that

Leaky Nun

I don't know why people starred it, but it got 3 stars, so I gotta star it.

9

flawr

@LeakyNun that is how starring here works in general.

Kritixi Lithos

@Adám most of the times the APL keyboard is on, occasionally I switch to a keyboard more suitable for V/Jelly while golfing (this time I had Jelly).

flawr

@KritixiLithos you do actually have an APL keyboard?

Kritixi Lithos

1:09 PM

yes, I installed an APL keyboard layout

It's very useful

orlp

@LeakyNun Python 2: lambda n:(2*n+n%2)/(1+n%2*2)

Erik the Outgolfer

@KritixiLithos so that's why you can post those carrots so easily

user165474

CMC: Make a program with at least 2 bytes that when given dimension x will output that code new-line delimited to fit inside a rectangle of width x with each line as wide as possible. Tokens and identifiers may be split over a line.

totallyhuman

inb4 charcoal answer

ASCII-only

:| Charcoal has no newline

totallyhuman

1:13 PM

o0

user165474

(you may assume that if l is the length of your program that 0 < x <= l

Leaky Nun

@HyperNeutrino I don't understand the CMC

user165474

Let's say you have program abcdef

user165474

Then if given input 1, it gives:

user165474

a
b
c
d
e
f

user165474

1:13 PM

If given input 2:

user165474

ab
cd
ef

user165474

If given input 3:

user165474

abc
def

ASCII-only

So the result doesn't have to be a valid program right?

user165474

no

user165474

1:13 PM

If given input 4:

totallyhuman

O_O

orlp

@LeakyNun remember those modular power towers?

user165474

abcd
ef   < You may optionally have spaces here but not necessarily

New Main Posts

-1

Q: Goodbye Cruel World!

Background Hello golfers! I would like to learn all the programming languages! But I kinda have a short attention span... and copying all the Hello World examples gets boring... but I like fire! ^w^ Challenge So here is the plan! I want you all to write the smallest code that will compile, pr...

code-golf hello-world

totallyhuman

@NewMainPosts i'm not sure how to react to that challenge

ASCII-only

1:14 PM

Oh no Charcoal can't do this very golfily

Leaky Nun

@orlp yes?

orlp

2

Q: Modular power tower mapping, is it injective?

Given an infinite sequence $a_1, a_2, \dots$ where all $a_i > 1$ we study $a_1^{\,a_2^{\,\cdots}} \bmod m$. While this is an infinite power tower that grows without bound, I argue that it can be assigned a value. If $f(n)$ is the power tower with the first $n$ terms of $a$, there is a constant $c...

modular-arithmetic power-towers

maybe you can think of a counterexample :)

Business Cat

@HyperNeutrino CJam, 15 bytes: {"_~"]sri/N*}_~

orlp

basically by writing an infinite sequence as a power tower

then taking that power tower mod 1, mod 2, mod 3, etc...

we get another infinite sequence

is this operation injective?

user165474

:o that is cool

ASCII-only

1:17 PM

$a_1$ maps to $a_1^{{a_2}^{...}} mod 1$ right?

orlp

no this mapping applies to the sequence as a whole

if b = MPT(a) then you are right that b_1 = a_1^{a_2}^{...} mod 1

ASCII-only

Unique for all inputs = unique for all valid input sequences?

Leaky Nun

@orlp just let $a_1=6$ and see that $\operatorname{MPT}(a)(2) = \operatorname{MPT}(a)(3) = 0$.

orlp

@LeakyNun I don't understand, can you elaborate?

Mayube

CMC: Port jQuery to your language

js: 0 bytes ``

Shaggy

1:26 PM

Holy hell, I made it to the top of the HNQ with my first challenge! :o

Leaky Nun

@orlp I mean, if $a_1=6$, then $b_2 = b_3 = 0$, right?

@Shaggy congrats

Mayube

@Shaggy 24 answers in 2 hours is pretty good

Leaky Nun

@HyperNeutrino Jelly, 13 bytes: “v0¡µṾ;ḢsƓY”v

Erik the Outgolfer

@LeakyNun Python, 38 bytes: def f(x):y=x-x%2;return y+y%4*-2+2+x%2

orlp

@LeakyNun yes

Leaky Nun

1:27 PM

@orlp so it isn't injective

orlp

@LeakyNun no, the question is whether MPT is injective

with input domain the set of all infinite sequences of integers > 1

Leaky Nun

@orlp wow, meta.

Stephen S

@Shaggy it's not that hard to get a ~trivial, decent PPCG challenge a ton a hotness points

Mayube

^ true

orlp

and output domain a set of infinite sequences of nonnegative integers

Leaky Nun

1:28 PM

@orlp so it is in some sense N^N -> N^N...

Shaggy

@StephenS Let me have my moment! :p :D

orlp

@LeakyNun can you find two distinct infinite sequences such that after application of MPT they have the same output?

that's the question

Leaky Nun

I can't.

Erik the Outgolfer

@LeakyNun is my function valid?

it seems to be bijective and following your image rule

Stephen S

@Shaggy well... I might, but my challenge was on top of HNQ before you came along xD

Erik the Outgolfer

1:30 PM

oh wait...is it a proper subset though?

Leaky Nun

@EriktheOutgolfer it isn't.

Erik the Outgolfer

yeah just realized

Shaggy

@StephenS Oops, sorry!

Stephen S

@Shaggy it's no problem, just being persnickety :)

Erik the Outgolfer

@LeakyNun damn this cmc is hard

Leaky Nun

1:32 PM

@EriktheOutgolfer you just need to have the right intuition

ASCII-only

Hmm I wonder whether not getting a design has something to do with all the work hours PPCG wastes :P or maybe the number of questions it has on HNQ

Erik the Outgolfer

I think I've come closer to the solution though

if that's the intuition

Mayube

Shout out to @Arnauld for that time he implemented the entire MD2 hashing algorithm into his JS answer to my first ever challenge :P

@ASCII-only or maybe we're just not active enough

New Meta Posts

0

Q: Temporarily close my own challenge

I have a challenge that I posted which I'd like to put on hold while I refine it in sandbox. Is this possible? I don't want to delete it because I'd like to recycle bits of it, and I'm guessing if it's deleted then I'll lose the work I've done/ there will be a time limit on how long I can still ...

support

Erik the Outgolfer

@LeakyNun does this count as a proper subset?

lambda x:x%2and max(x-2,0)or x+2

generally, if x is even, it returns x+2, if x is 1 it returns 0 otherwise it returns x-2

might be able to shorten though...

Leaky Nun

1:37 PM

does not return 0 for x=1 because of the or

in general, you only need and ... or if you need short-circuiting.

Erik the Outgolfer

@LeakyNun as I said above, not ready yet, but you get the idea

of course I can use a ternary or even some math

Leaky Nun

So lambda x:[x+2,max(x-2,0)][x%2] fixes the problem and is shorter

ASCII-only

@EriktheOutgolfer Just port this

Erik the Outgolfer

hmm...

ASCII-only

lambda n:[n+2-4*n%8,0][n==1]

Erik the Outgolfer

1:40 PM

@LeakyNun how about this?

lambda x:max(x+x%2*-4+2,0)

PyRulez

Would a programming language based on Wang tiles work?

Since they're turing complete?

Leaky Nun

@EriktheOutgolfer wonderful

ASCII-only

@PyRulez Yes

@EriktheOutgolfer lambda n:max(n+2-4*n%8,0) or lambda x:max(x-x%2*4+2,0) should work

Erik the Outgolfer

I already figured out second one...

...it's just that 2.5 minutes already passed

Wheat Wizard

How are meta key strokes counted? is M-^ one, two or three bytes?

Erik the Outgolfer

1:47 PM

@LeakyNun of course this is 9 bytes in Jelly hehe

so I guess I'm the new winner?

Leaky Nun

lol

ASCII-only

@EriktheOutgolfer But it's a popcon

Erik the Outgolfer

is there a popularity-contest tag anywhere?

ASCII-only

48 mins ago, by Leaky Nun

alright, it's a popcon then

Erik the Outgolfer

oh sneaky...yet it seems that my function is coincidentally same as yours, only a bit more math involved, and more math involved means more interesting, so I'm still the new winner...

ASCII-only

1:50 PM

@EriktheOutgolfer A bit more math?

Erik the Outgolfer

yeah, it uses max instead of ternary for n=1

basically it doesn't check for n=1 at all

Adám

@flawr Well, there is an APL keyboard (software layout) and an APL keyboard:

ASCII-only

:37968592 Oh no it's 1 byte longer

Erik the Outgolfer

@ASCII-only this is the fixed version, still 1 byte longer

even though you can compute x <= 1 in 1 byte, here you need x == 1 since for input 0 behavior is different

Transcript for