The Nineteenth Byte

Martin Ender

09:00

oh right, descending sortedness

xnor

did I get it backwards?

Martin Ender

well if you just have lexicographic order and filter by sortedness and then take first, the result will also be sorted

xnor

oh yes

Martin Ender

either you need the lexicographic order of the reversed lists, filter by sortedness and take first; or you need the lexicographic order, filter by descending sortedness and take last

xnor

Here's an even crazier characterization: Write each number 0 to 2**(2*n-1) in binary, keep only those with exactly n ones, and for each one, count the number of zeroes at the end plus one

Martin Ender

09:08

:D

I'm sure eventually you'll find one that is golfy ;)

xnor

combinatorics is it own reward!

2

Sp3000

I'll stick with paths from (0,0) to (n-1,n+1) on a lattice grid :P

xnor

Is there a nice way to index into the n-size subsets of [1..N]? In other words, to get the kth one in lexicigraphic order without counting k steps?

Martin Ender

can you find the kth number with popcount n?

xnor

ah, popcount was the term I was looking for

yes, that's the question

Leaky Nun

09:15

factorial base?

xnor

i know there's that nice way to iterate through numbers of a given popcount, but that's gonna be long

Martin Ender

you don't actually want to iterate through them though, right?

xnor

right, i was hoping to extract it directly

flawr

interesting stuff you're talking about=)

I hope I do not interrupt, but it is a related question: Can you efficiently enumerate/generate all subsets of size k of some given finite set?

xnor

yes, that's the popcount enumerating algorithm

graphics.stanford.edu/~seander/bithacks.html#NextBitPermutation

is this what you were looking for?

flawr

09:22

@xnor Crazy, it seems to work, but I have no idea why

mınxomaτ

I don't know, but the english Google translate voice always sounds like she has an attitude.

flawr

Is pop count short for population count?

xnor

@flawr roughly, given a subset, we can find the lexicographically next subset by including the smallest not-yet-included element to included, removing the next smallest currently include element, then making all elements below it as small as possible

by "lexicographically", here i mean if the elements were listed in decreasing order

flawr

@xnor Thanks, I think I first need to do some further reading=)

trichoplax

I can imagine an animation might help see what is happening

flawr

09:39

Found some sites that seem to explain it and added them to my TBR pile.

Leaky Nun

09:50

CMC: n minus each digit of n. For example, 123 becomes 123-1-2-3=117. code-golf

Martin Ender

CJam, 8: ri_Ab+:-

zʏᴀʙiɴ101

Challenge idea: implement the Tiny Encryption Algorithm.

flawr

Matlab, 24: @(n)n-sum(num2str(n)-48)

Leaky Nun

@zʏᴀʙiɴ101 Please, no sandboxed post.

xnor

>>> f=lambda i:[i for i in range(1,i+1) for i in range(1,i+1) for i in range(1,i+1)];f(3)
[1, 1, 1, 2, 1, 1, 2, 1, 2, 3]

Python, abusing scopes like a champ

Leaky Nun

09:58

@xnor only works for 3?

Sp3000

I just wish we could do [*range(x) for x in range(i)] :P silly PEP clause not getting passed

xnor

@LeakyNun i'm planning on using eval to do the general pattern

zʏᴀʙiɴ101

@LeakyNun Oh look, this sandboxed post was posted on Jun 15 at 19:43 and it is Jun 21 at 13:58. I have to wait to claim this challenge.

Yeah, it will be hard.

Leaky Nun

@xnor Well, that would be repeating many answers already present.

Sp3000

o_O now that I look at it properly, xnor that's quite some scope abuse

Leaky Nun

10:01

@MartinEnder Why doesn't "map" or "each" work on strings?

Martin Ender

?

Leaky Nun

i/ should convert each digit to integer?

Martin Ender

:i

actually no

that converts characters to their character codes

:~

Leaky Nun

That's clever.

Why doesn't r_:~:+- work?

Sp3000

@xnor Bonus points if you f-string it for -1 byte :P

Leaky Nun

10:04

r_:~:+ gives 123 6, but r_:~:+- gives 123

Martin Ender

@LeakyNun because there's no implicit string to integer conversion

so - is actually set difference there

Leaky Nun

@MartinEnder but it gives 6 already

Martin Ender

you're trying to remove all 6s from the array ['1 '2 '3]

Leaky Nun

The 6 is already an integer

Oh, I forgot to make 123 an integer

So it would be r_:~:+\i\-

xnor

@Sp3000 3.6 is too cutting edge for me :)

Sp3000

10:06

:( but alpha's officially out now

Martin Ender

@LeakyNun r_i\:~:+-

Sp3000

There's... an installer and everything :o

Leaky Nun

@MartinEnder Oh, nice.

@Dennis Not only is there a closed form for your "innovative" idea, but it also turned out to be a duplicate.

Martin Ender

I think he noticed that by now...

Leaky Nun

Alright.

LegionMammal978

10:19

Question: Is x===~~x a valid test for integral numbers in JS?

Leaky Nun

@LegionMammal978 It passes my two testcases.

Why not just write x==~~x?

LegionMammal978

@LeakyNun Safety

Leaky Nun

@LegionMammal978 from?

LegionMammal978

(and consistency, it's used everywhere else)

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Exercise your kids I want you to output the nth verse of this kid's exercise song: Head, and shoulders, knees and toes, knees and toes, Head, and shoulders, knees and toes, knees and toes And eyes and ears and mouth and nose, Head, and shoulders, knees and toes, knees and toes. Each ...

Martin Ender

10:25

@LeakyNun from JavaScript

Leaky Nun

@MartinEnder why?

Martin Ender

because JavaScript's == is cancer

Leaky Nun

@MartinEnder do my challenge in your new language

trichoplax

Javascript ransomware was in the news recently

LegionMammal978

^^^

Leaky Nun

10:27

(if you have an hour of free time)

Martin Ender

797

A: Strangest language feature

JavaScript truth table: '' == '0' // false 0 == '' // true 0 == '0' // true false == 'false' // false false == '0' // true false == undefined // false false == null // false null ...

@LeakyNun which challenge?

Leaky Nun

@MartinEnder 123-1-2-3

Martin Ender

oh

yeah I'm not looking forward to base conversion in that language :D

LegionMammal978

""==~~""
true

Leaky Nun

>> 'abc'==0
<< false
>> +'abc'
<< NaN
>> ~'abc'
<< -1
>> 'abc'==NaN
<< false

LegionMammal978

10:29

""===~~""
false

That's why

Leaky Nun

Can anyone tell me

why the **** do I get these results?

LegionMammal978

@LeakyNun You could call Ecma International at +41 22 849 6000

Leaky Nun

CMC: Check if everything in an array is equal to 1

Just wondering if the Jelly code ;1E could be golfed.

code-golf

LegionMammal978

@LeakyNun Mathematica, 14 bytes: AllTrue[#==1&]

Leaky Nun

@LegionMammal978 Misplaced ampersand?

LegionMammal978

10:40

Hey, @Upgoat, been busy working on rationals

@LeakyNun No, ampersand is postfix operator for anonymous functions/lambdas in Mathematica

Leaky Nun

Isn't it always at the last?

LegionMammal978

@LeakyNun AllTrue can take an anonymous function and return a new anonymous function.

Whereas most code-golf Mma answers are just straight-up anonymous functions

(also, they're technically called "pure functions", but I come from a background in C#)

Leaky Nun

Oh, alright

Martin Ender

@LegionMammal978 actually, I think "pure functions" are a bit misleading, because that usually refers to functions without side effects

LegionMammal978

@MartinEnder Yeah, but I didn't name them

In[1]:= AllTrue[#==1&][{1,1,1}]

Out[1]= True

In[2]:= AllTrue[#==1]&[{1,1,1}]

Out[2]= AllTrue[{1, 1, 1} == 1]

Martin Ender

10:45

1<=##<=1&@@#& works, doesn't it?

LegionMammal978

Yes, it would

Leaky Nun

@MartinEnder why not just 1==##?

LegionMammal978

@LeakyNun Does 1.0 count as 1?

Martin Ender

@LeakyNun umm...

yeah...

Leaky Nun

@LegionMammal978 whatever is more convenient

LegionMammal978

10:47

Can it take both?

In[3]:= 1.==1

Out[3]= True

Leaky Nun

I said, whichever is more convenient.

LegionMammal978

Because if we had to choose, then it would have to be 1===##

Kinda like JS in that respect

Leaky Nun

Alright.

LegionMammal978

So far: 1==##&@@#&

flawr

bbc.com/news/technology-36575687

LegionMammal978

10:51

Why does Windows still have the option to run any executable in an email?

At least force it to be an attachment

(Then it would be the user's fault)

Leaky Nun

@LegionMammal978 Could you explain it to me?

LegionMammal978

@LeakyNun <func>@@#& is kind of like JS .apply, it uses a list as a function's arguments

1==##& just inserts == between its arguments, so 1==##&[{3,2,1}] would be 1==3==2==1

It's more complicated, but you get the point

mınxomaτ

@flawr It's not JavaScript. It's JScript, a server-side legacy .NET language with JS syntax.

Sp3000

@flawr Did you double up on @(a) in your answer or is it meant to be like that?

Leaky Nun

@LegionMammal978 alright, nice.

flawr

10:55

@Sp3000 That should indeed be there only once, thanks!

@mınxomaτ But it's close enough =D

Sp3000

Hmm does [[0,1,0];[1,-1,1];[0,1,0]] work? (should return truthy)

LegionMammal978

The thing with Mathematica is that everything is a function applied to arguments (or a number, or a string, or a symbol)

mınxomaτ

@flawr Well, actually no. The source explains that pretty well, but the BCC article implies ordinary browser-engine JS.

LegionMammal978

For example, 1==##&@@#& is actually Function[Apply[Function[Equal[1, SlotSequence[1]]], Slot[1]]]

mınxomaτ

@flawr Write a Cryptolocker in browser-JS and I'm impressed. Otherwise, this is not news. JScript cryptolockers have been around for more than a decade.

LegionMammal978

10:57

So pretty much a Lisp syntax replacement with pattern matching?

Leaky Nun

@LegionMammal978 Well, SlotSequence[1] is ## and Slot[1] is #, right?

flawr

@Sp3000 Let me check

Leaky Nun

What is @@ and &?

LegionMammal978

@LeakyNun Yes..

@LeakyNun @@ is Apply and & is Function

(and == is Equal)

Leaky Nun

Alright.

LegionMammal978

11:01

This has some interesting consequences:

In[4]:= 1==##&@@#&[[1,1,1,0]]

Out[4]= Equal

(i.e., the zeroth element of the first element of the first element of the first element of 1==##&@@#& is the symbol Equal)

New Main Posts

0

Q: Downward Number Race

Your task is to create a program that adds random numbers to previous sums in the ultimate number racing showdown. Each racer (column) starts at 0 and adds either 1 or 0 to the previous sum at each step of the race until all of the racers reach the score required to win. The 1 or 0 should be cho...

code-golf

LegionMammal978

11:19

['^', (LHS, RHS) => {
    if (RHS.constructor.Name === "Number") {
        if (RHS.value === ~~RHS.value) {
            if (RHS.value < 0)
                return HelperInit(LHS.constructor, Math.pow(LHS.den, -RHS.value), Math.pow(LHS.num, -RHS.value));
            else
                return HelperInit(LHS.constructor, Math.pow(LHS.num, RHS.value));
        }
        else
            return HelperInit(CheddarNumber, 10, 0, Math.pow(LHS.num / LHS.den, RHS.value));
    }
    else if (RHS instanceof LHS.constructor) {

Rational exponentiation is irrational

Leaky Nun

@LegionMammal978 rational exponentation? Like a rational to the power of another rational?

LegionMammal978

In both ways

Leaky Nun

But a rational to the power of another rational isn't necessarily rational... why would you support that?

LegionMammal978

> irrational
adjective
3. not in accordance with reason; utterly illogical.
noun
8. (Mathematics.) irrational number.

Leaky Nun

So you have irrational numbers?

LegionMammal978

11:24

In both ways, yes

Leaky Nun

:o

LegionMammal978

And thus, rational exponentiation is irrational.

Q.E.D.

Also, under what conditions does (a%b)/n = (a/n)%(b/n) hold?

Okay, I'll take that as a 2

orlp

11:47

@LegionMammal978 only by coincidence

unless you define / to be the inverse of multiplication in a finite field

LegionMammal978

I'm not talking about modular arithmetic here, I'm talking about % in the sense that pi%(1/2)=pi-3

orlp

that's still modular arithmetic, albeit an extension

LegionMammal978

Okay, then arithmetic in the interval [0,b) for a%b

orlp

@LegionMammal978 so if I understand your a%b correctly (assuming positive reals)

is the smallest possible number x left over after repeatedly subtracting b from a

LegionMammal978

Yes

orlp

11:56

ok

then let's say a%b is a - k*b for some integer k

(a-k*b)/n is what we have on the left side then

but if we divide both operands by the same number, we know that the process of repeated subtraction will stop with the same number k of subtractions

so we can rewrite the right hand side using the same k

LegionMammal978

makes sense so far

orlp

(a-k*b)/n = ((a/n)-k*(b/n))

LegionMammal978

So they would be equal.

orlp

if you now multiply both sides by n you'll see that they are always equivalent

yes

assuming both positive operands

LegionMammal978

Thanks, know some terminology but have never been much of a mathematician

orlp

12:00

so your "%" (often called fmod) can be distributed over by division

LegionMammal978

Yeah, was implementing rational mod but didn't want to go through finding multiplicative inverses

Leaky Nun

@LegionMammal978 You know, a%b = a-floor(a/b)

beware floor(-4.5) = -5

LegionMammal978

Eventually came up with (n1/d1)%(n2/d2) = ((gcd(d1,d2)*n1/d2)%(gcd(d1,d2)*n2/d1))/(d1*d2/gcd(d1,d2))

Which only uses integer %

Leaky Nun

so (a/b)%(c/d) = ((ad)%(bc))/(bd)

LegionMammal978

More like (a/b)%(c/d) = g((ad)%(bc))/(bd/g)

Leaky Nun

12:09

You can simplify the fraction later

orlp

@LeakyNun well

LegionMammal978

Have JS integer width to consider here, so I can't multiply huge numbers

orlp

that's the funny thing

you can choose different ways of truncating integer division

and those in turn determine the sign of a%b for when the signs differ

LegionMammal978

The outermost division of the RHS is actually my arbitrary-precision Rational type

So no integer division is necessary

orlp

relevant post: python-history.blogspot.nl/2010/08/…

LegionMammal978

12:17

ad < bc implies that a/b < c/d, correct?

Leaky Nun

Only if the signs of b and d are the same

LegionMammal978

(which they are)

Leaky Nun

Then yes.

LegionMammal978

And that applies to >, <=, etc., correct?

Leaky Nun

well just check that a and c are not zero and those nasty things

orlp

12:20

@LegionMammal978 yes, you can manipulate inequalities just like equalities

LegionMammal978

@LeakyNun You mean b and d?

orlp

except division/multiplication you need to flip the sign if you multiply/divide by a negative amount

Leaky Nun

It is equivalent to dividing both sides by bd

LegionMammal978

@orlp Yeah, was asking because that keeps catching me :p

Leaky Nun

@LegionMammal978 Well, all four of them.

LegionMammal978

12:30

@LeakyNun Why can't a or c be negative?

Leaky Nun

@LegionMammal978 Oh, I meant zero.

LegionMammal978

Why not?

Leaky Nun

Well, I am not sure what 0%0 is

LegionMammal978

Oh, I'm talking about ad < bc → a/b < c/d

Leaky Nun

again, as long as the signs of b and d are the same.

mınxomaτ

12:32

Because I needed this today: Very fast float to int array conversion in C/++ with x86 inline ASM. It's usually min. 6 times faster because it skips the FPU pipeline flushing mandated by the C std.

Leaky Nun

@LegionMammal978 and neither b nor d is zero

LegionMammal978

12:44

@Upgoat I figured out a way to do what you were trying to do with Git: git rebase -i <commits> can actually squash commits together!

(Just read git-scm.com/docs/git-rebase for more awesome things it can do.)

Question: How does one get all accurate digits of a number in JavaScript?

flawr

12:59

Goin to the dentist

Hippo Brushing Teeth

►

TimmyD

Chat is a big ole' heap of grey. Legion's avatar is just a white box for me, Leaky's is a broken image, and Minxomat's is greyscale. And then comes flawr busting up the combo.

Quill

13:16

cccccccombo breaker

Leaky Nun

@TimmyD Which part of my image is broken?

TimmyD

@LeakyNun Facebook is blocked here, so that's why.

Leaky Nun

@TimmyD where is "here"?

TimmyD

work

Leaky Nun

I see

Sp3000

13:24

@LeakyNun Actually I don't know why, but your avatar does seem a bit weird on some browsers (Chrome here)

Leaky Nun

@Sp3000 Seriously, maybe because it is imported from Facebook

Sp3000

Woop, yes that does seem to be the problem

Just turned off Disconnect (browser extension) and it showed

LegionMammal978

@Sp3000 Wait... It's disappeared even more than mine has!

(Works fine for me using FF)

Optimizer

I blame cache

I cannot see mine on the left of the text box

LegionMammal978

13:39

@Optimizer shouldiblamecaching.com

TimmyD

Weird ... www.shouldiblamecaching.com doesn't exit (DNS lookup error). -1 for lack of standards support

LegionMammal978

@TimmyD Not on their end

TimmyD

I don't understand what you mean by that.

LegionMammal978

It's your problem, not theirs

zʏᴀʙiɴ101

@TimmyD For me, it works fine.

TimmyD

13:42

It's my problem that they're not publishing a WWW record?

zʏᴀʙiɴ101

@LegionMammal978 Yes, according to caching.

El'endia Starman

@TimmyD There's no www. at the front.

TimmyD

Right. That's my point.

El'endia Starman

I too get a DNS lookup error if I try www.shouldiblamecaching.com, but not if I try shouldiblamecaching.com.

LegionMammal978

Uh... toolbox.googleapps.com/apps/dig/#A/…

TimmyD

13:44

@LegionMammal978 Right, it returns the SOA record, equivalent to "Does Not Exit"

LegionMammal978

Oh, I see

So it doesn't support www.

TimmyD

Right

LegionMammal978

Blame FF for being misleading then

(And blame caching, that's what it tells me)

TimmyD

@MorganThrapp Ooo, 400 bytes is agonizingly close

Morgan Thrapp

@TimmyD You're telling me.

Only 3 bytes to go!

If I didn't have to pad the time, I'd have it already. :P

zʏᴀʙiɴ101

14:02

I had an emergency outage at my home a few hours ago because the AC had nowhere for the water to flow.

Anyway, it's now safe for me to use the computer. :)

Poke

@zʏᴀʙiɴ101 mon

Morgan Thrapp

@TimmyD Check it! 392!

zʏᴀʙiɴ101

@Poke mon

Morgan Thrapp

Wait, oops, hang on, broke it.

zʏᴀʙiɴ101

Pro tip: a blank FM radio frequency is a decent random number generator.

I took 20 seconds from a blank FM radio frequency with my trusty-dusty Lenovo smart phone.

And now parsing the result.

LegionMammal978

14:11

Pro tip: a GPS signal is a decent time function.

zʏᴀʙiɴ101

Yes, because some phones come equipped with a "Sync time via GPS" function.

LegionMammal978

Well, GPS has perfect accuracy, running on an atomic clock to measure the exact SI second

Morgan Thrapp

Actually below 400 now! 397!

LegionMammal978

@MorganThrapp What are you talking about?

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@Upgoat htf did i come online

i was probably either puking my guts out or sleeping

Morgan Thrapp

14:20

@LegionMammal978 The FAT32 thing, I was hovering right around 400 bytes last night and I finally got below it.

Poke

@LegionMammal978 @TimmyD's challenge

LegionMammal978

Thx

TimmyD

@MorganThrapp Nice.

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@Upgoat Tor stopped working, and the proxy settings screwed stuff up.

Chrome alternated between france, germany, and japan. wtf

TimmyD

@MorganThrapp To be fair, it's not really FAT32. There are a lot of additional specs that I pared out to make it a decent challenge (as opposed to writing a whole file I/O subsystem as an answer).

Morgan Thrapp

14:30

@TimmyD Shhh, let me have this moment. ;)

It was actually really interesting, because I've always been interested in how file systems work, but not interested enough to read about them.

TimmyD

Well, we've come a long way in file systems since this challenge's spec, which is a trimmed down version of FAT16

I mean, FAT16 came out over 30 years ago

Bálint

Anyone knows this fractal?

Because I don't

Morgan Thrapp

Oh yeah, that's Frank. Good guy, 2 kids.

Bálint

@MorganThrapp Much funny. Very laugh

TimmyD

Hmm ... neither TinEye nor GRIS come back with a result.

Eᴀsᴛᴇʀʟʏ Iʀᴋ

14:36

Any advice for this?

2

A: Sandbox for Proposed Challenges

Do not repeat yourself metagolf Given a number m, output a program that prints out the string Do not repeat yourself. Now for the challenging part: Each program the original code prints out has to have the byte at place m*n be the same as m*n - 1 and the same as m*n - 2 .... all the way down t...

Doorknob

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@Bálint looks kinda a weird koch snowflake

Bálint

@EᴀsᴛᴇʀʟʏIʀᴋ Yeah, the genration is pretty similar, but instead of putting triangles on triangles, I put circles on circles

@Doorknob I got that far

It's kind of hard to explain

Eᴀsᴛᴇʀʟʏ Iʀᴋ

(removed)

how do grey text chat

Bálint

Here, with better sample size

Is there a list of fractals online?

Eᴀsᴛᴇʀʟʏ Iʀᴋ

14:49

idk

TimmyD

@Bálint ics.uci.edu/~eppstein/junkyard/fractal.html

Bald Bantha

github.com/epicTCK/Cheddar/commit/…

Eᴀsᴛᴇʀʟʏ Iʀᴋ

> Best guess for this image: monochrome

@BaldBantha 0/10 brb remvoing

Bald Bantha

It's on my fork

Eᴀsᴛᴇʀʟʏ Iʀᴋ

yes

Bald Bantha

14:52

so you can't remove it

right?

Eᴀsᴛᴇʀʟʏ Iʀᴋ

brb warning @Upgoat DON"T MERGE ANY PR's FROM EPICTCK

Sp3000

@Bálint Also: en.wikipedia.org/wiki/List_of_fractals_by_Hausdorff_dimension

Bald Bantha

I don't want him to merge it, I would have made a pull req in that case. I just wanted to fork it and mod it to my liking, as the license lets me

zʏᴀʙiɴ101

@BaldBantha Li cense.

Without the s.

Bálint

Nothing

I may have just found an unnamed fractal

Bald Bantha

14:54

@Bálint name it after yourself

Eᴀsᴛᴇʀʟʏ Iʀᴋ

reddit.com/r/talesfromtechsupport/comments/4p3dha/…

Bálint

@BaldBantha I don't think you can without publishing it

Bald Bantha

...? so you have to publish a paper about it?

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@BaldBantha license says either delete that repo or put a NOTICE.(md|txt) file that says you don't own it.

Bálint

@BaldBantha Yeah

Bald Bantha

14:57

@EᴀsᴛᴇʀʟʏIʀᴋ DANGIT y no MIT licence

Bálint

I don't think you can just register it

Eᴀsᴛᴇʀʟʏ Iʀᴋ

because Apache

Bálint

I may ask the mathematics SE

Bald Bantha

brb deleting repo

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@BaldBantha ^^^

@BaldBantha it's fine, just add 1 simple txt file

Bald Bantha

14:58

and say "i don't own this"?

Eᴀsᴛᴇʀʟʏ Iʀᴋ

yeah

"I don't own this, all credit goes to XXX YYY the Goat"

Transcript for