The Nineteenth Byte

00:01

See, I kinda want to nominate the blog for Best Of, but I don't think it really qualifies for "most significant meta impact", and it definitely doesn't qualify for any others :/

Adám

@DLosc Yes, BQN has numbers that are not arrays, although it tends to transparently upgrade them to arrays when needed.

DLosc

@cairdcoinheringaahing Alternate 6-byte answer which might also lead to a different approach

Adám

@cairdcoinheringaahing New category: best blog post

Did mini-golf die?

DLosc

Well, we're doing a CMC right now, so...

caird coinheringaahing

Welcome to the 19th Biweekly Mini Golf. We don't really have any CMCs prepared, but feel free to post any for the next hour or so :)

That works :P

Neil

00:04

@cairdcoinheringaahing the product is 4?

caird coinheringaahing

Honestly, I'm very happy that we've reached 19 - that's almost 10 months of BMG, more or less without fail

Adám

I recently posted this CMC which didn't get a lot of answers:

2 days ago, by Adám

CMC: Given a 5-element list L and a number N, replicate the middle element of L so there are N copies of it. E.g.: L=[3,1,4,1,5];N=3 → [3,1,4,4,4,1,5] and L=[2,7,1,8,1];N=0 → [2,7,8,1]

caird coinheringaahing

@Neil Unfortunately, that takes 5 bytes to express: chat.stackexchange.com/transcript/message/60255866#60255866

DLosc

> Unfortunately

Currently shorter than every other solution :P

caird coinheringaahing

But you can always get shorter :P

even when it's one byte, what if there exists a zero byte solution??? :P

Aaroneous Miller

00:07

@cairdcoinheringaahing wild card?

Adám

Given a matrix, return the cumulative maxima across and down. E.g.

3 1 4
1 5 9
2 6 5

gives

3 3 4
3 5 9
3 6 9

DLosc

@DLosc Yet another 6-byte Jelly solution

caird coinheringaahing

And another

@AaroneousMiller That could work, yeah

DLosc

@cairdcoinheringaahing Literally was just about to post that

caird coinheringaahing

@Adám Assuming starting from the top/left?

Adám

00:10

Yes, as per example.

caird coinheringaahing

Also, nice pi reference :P

Adám

I tend to use various famous constants as examples.

Wheat Wizard

What is done for?

2 1
0 1

Adám

What?

Wheat Wizard

What should the output be?

caird coinheringaahing

00:11

Is the bottom left corner 1 or 2?

Adám

2

caird coinheringaahing

@cairdcoinheringaahing Wait, no, I got that the wrong way around. Should the output look like this?

2 2
2 1

Adám

Every number spreads right and down

There can be nothing in the SE quadrant from a number, which is smaller than that number.

Wheat Wizard

I actually think that the version with 1 in the corner is more interesting.

caird coinheringaahing

So [[2, 1], [0, 1]] -> [[2, 2], [2, 2]]?

Adám

00:15

Yes, that's what I had in mind.

Wheat Wizard

The 2 version is just colmap a scan then rowmap the same scan.

But with 1 in the corner you have to be careful.

caird coinheringaahing

I have 5 bytes for 2 in the corner and 6 bytes for 1 in the corner

@cairdcoinheringaahing Fun fact: for the 6 byter, 50% of the code is the same byte

Adám

I don't get the spec for 1 in the corner.

user

@Adám ⌈\⌈⍀ in APL if I'm interpreting it correctly

Adám

Yes, that's what I had in mind.

"mini" golf.

Wheat Wizard

00:17

@Adám Every element is the max of anything earlier in the same row or column. You don't have bleed between differnt rows and columns.

DLosc

@Adám Lol

(Probably not the shortest possible, but still)

Bgil Midol

nice

caird coinheringaahing

Try it online! - top line is 1 in corner, bottom line is 2 in corner

@DLosc xP

Adám

Wait, so APL beats Jelly on both specs‽ 1 in the corner should be ⌈\⌈⌈⍀

caird coinheringaahing

Jelly doesn't have a "scan by rows" builtin

Adám

00:22

Pathetic ;-)

caird coinheringaahing

My fork ties APL: I add ȥ to be scan by rows

Adám

Do you have reduce by rows too?

caird coinheringaahing

@cairdcoinheringaahing Wait no, I can't read my own code: ȥ is reduce by rows, Ðȥ is scan by rows, and doesn't save any bytes

(it can save bytes, as the behaviour differs from \€ for 3 or more dimensional arrays, but not here)

Wheat Wizard

@cairdcoinheringaahing How does Jelly represent 2d arrays? Isn't that just a map?

DLosc

@DLosc Argh--an approach that seemed more clever is slightly longer.

caird coinheringaahing

00:25

2d arrays are simply an array that contains at least one 1d array

Typically, that's a list of lists

Adám

Jelly doesn't have proper multi-dimensional arrays.

caird coinheringaahing

Nope. Constrained by Python's lack of "being an array language"

Wheat Wizard

In that case rowmap should just be a regular map right?

caird coinheringaahing

@WheatWizard Yeah, scan by rows is just "over each row, scan"

Adám

It could have been built atop NumPy

caird coinheringaahing

00:27

I meant that, while APL has a single char scan by rows, Jelly has to do "each: scan" (\€) instead

Wheat Wizard

Ah yeah I guess I was considing a combination of two things to just be built in.

Adám

CMC: Given a multi-dimensional non-ragged array and a number, replace each element by a list consisting the given number copies of that element. E.g. 3 and

[[[3,1],
  [4,1]],
 [[5,9],
  [2,6]]]

gives

[[[[3,3,3],
   [1,1,1]],
  [[4,4,4],
   [1,1,1]]],
 [[[5,5,5],
   [9,9,9]],
  [[2,2,2],
   [6,6,6]]]]

Wheat Wizard

Haskell fmap.replicate

You have to import libraries to get weird array types but fmap will still work since they are always functors.

lyxal

Vyxal λ-[vx|⁰ẋ (I think)

caird coinheringaahing

Although, Jelly's reduce/scan only scans by columns "on accident": reduce [a, b, c, d] by f technically just does ((a f b) f c) f d, so non-vectorising commands have different behaviour. ,Ṁ¥/ means "reduce by: pair arguments, take max" and »/ means "reduce by maximum", but for multidimensional arrays, they are different: Try it online!

Adám

00:34

@WheatWizard So pretty much the same as in APL: /¨ (/ is replicate and ¨ is each/map)

caird coinheringaahing

The first does max(max([3, 1, 4], [1, 5, 9]), [2, 6, 5]) whereas the second does [max(max(3, 1), 2), max(max(1, 5), 6), max(max(4, 9), 5)]

DLosc

@Adám Oof. Pip doesn't have a good way of mapping arbitrary operations at arbitrary depth. It would be super easy if it were string repetition instead of array, though...

Adám

@DLosc Can you not recurse?

DLosc

@Adám Probably, but that's going to be lots of bytes

lyxal

Imagine not being able to recurse in 1 byte

DLosc

00:37

Probably the shortest way would be to do a string replacement and eval it

Adám

@DLosc Yeah, {0=≡⍵:⍺/⍵⋄⍺∇¨⍵} in APL

Wheat Wizard

In Haskell the obstacle to a lot of these questions is typing. And you often have to get into the weeds of lightly dependently typed stuff in order to get some of these types. But in this one we don't actually care about the shape of the container at all so we abstract all that difficulty out.

Adám

@lyxal How can you do that? Do you not need an action and a stop condition?

caird coinheringaahing

@Adám Jelly, 6 bytes: ßxŒḊ?€

lyxal

@Adám I more meant that the syntax to call the current function isn't 1 byte in pip if there even is a built-in for recursion

Adám

00:40

Oh.

caird coinheringaahing

Aka, f(x, n) {for(e in x) { if(depth(e)) {f(e, n)} else {repeat(e, n)}}}

DLosc

@Adám Pip -p, 10 bytes. Dunno if it's cheating to take and return strings for the arrays, tho...

Adám

Probably allowed by default rules, and it is CMC, so…

DLosc

@lyxal There is a recursion operator in modern Pip, RE, but only for one-argument functions so far.

Of course it could be one byte if I wanted to sell out with a 256-character code page, but I don't. =P

caird coinheringaahing

@DLosc if/when you allow for variadic recursion, avoid Jelly's mistakes :P

DLosc

00:42

@cairdcoinheringaahing What are they, so I can avoid them?

Adám

Related CMC: Given two identically structured non-ragged arrays, return an array identical to one of them, except each element is copied to N copies, with N being the corresponding number from the other array. E.g.

[[2,7,1],
 [8,2,8],
 [1,8,2]]

and

[[3,1,4],
 [1,5,9],
 [2,6,5]]

gives

[[[3,3],
  [1,1,1,1,1,1,1],
  [4]],
 [[1,1,1,1,1,1,1,1],
  [5,5],
  [9,9,9,9,9,9,9,9]],
 [[2],
  [6,6,6,6,6,6,6,6],
  [5,5]]]

caird coinheringaahing

@DLosc Basically, if you have a dyadic link f(x, y), and in that dyadic link, you have a monadic chain g(k) that calls e.g. f(k, k²), then ß (the recursion operator) is forced to be a monad, which changes the parsing. So, you'd expect ß² to call f(k, k²), but it actually tries to do f(k)² :/

Adám

(SE, stop rate-limiting high-rep ROs!)

Wheat Wizard

Haskell: Data.Zip.zipWith replicate

DLosc

@cairdcoinheringaahing Huh. I'm not 100% sure I understand, but that definitely doesn't apply to Pip, so I think we're good :)

Adám

00:45

Btw, this is still /¨ in APL because why not map over two arrays?

Wheat Wizard

I can think of a few reasons.

DLosc

@DLosc Though, parsing issues are why RE is only unary at the moment. Operators in Pip can have multiple arities, but I'm finding that keeping some of them unary-only can help with parsing sometimes.

caird coinheringaahing

@DLosc So if you have the Jelly version of f(x, y) { if(x > 0) then g(x) else g(y) }; g(k) { if(k prime) then f(k, k-1) else f(k, k-2) } (except that g is defined inside f), then it doesn't like it

Adám

@WheatWizard tbf, it does get a bit uglier if you need to map over three or more.

caird coinheringaahing

ß's arity should be defined based on the link (aka f) it's in, but, due to a bug, it's actually defined based on the chain (aka g) its in

@Adám Ooh, e making an apperance :P

@Adám Jelly, 6 bytes: ßxŒḊ?"

Takes the arguments in the opposite order than how you specified

Adám

00:50

I didn't really specify.

@cairdcoinheringaahing You didn't see my latest webinar, did you? (Not sure if you've seen any…)

DLosc

@cairdcoinheringaahing Redwolf's least favorite mathematical constant :P

caird coinheringaahing

@Adám To be honest, I didn't know you've been releasing webinars. Do you have a link?

Adám

dyalog.tv

DLosc

@Adám This isn't so bad, actually. 13 bytes, assuming all lists are nonempty.

caird coinheringaahing

Tuvalu must be making bank of all these websites using .tv :P

Adám

00:53

Did you know that .io is for British Indian Ocean Territory?

caird coinheringaahing

@Adám Hmm, TIL that apl.chat redirects to the Orchard

Adám

Yeah, I bought that up, together with apl.wiki

DLosc

@Adám This one is 15 bytes for a recursive solution, using some global variable shenanigans.

Adám

Anyway, if you watch my latest webinar, keep your mathematical, physics, and chemistry constants' bingo card available…

@DLosc What do you mean by the lists being non-empty? You mean the input arrays?

caird coinheringaahing

@Adám I'm going to see if I can recognise them without googling :P

Adám

00:58

Good luck.

DLosc

@Adám I guess this only applies to the ragged arrays one, but: there won't be anything like [[1,2], [], [3, 4]].

Oh, wait, you said non-ragged both times.

Adám

I explicitly stated that the input arrays are non-ragged.

DLosc

I can't read :P

OldSandboxPosts

Sandbox posts last active a week ago: Move the next greater number, with fewest fails to take everything

DLosc

Well, also assuming there won't be anything like [[], []], which is technically not ragged but is rather pointless for these challenges

Adám

01:00

Oh, right.

@OldSandboxPosts I guess that means BMG is over.

caird coinheringaahing

CMC: Given 9 digits, output the check digits, as described in @Adám's webinar. You may output either the entire 11 digits, or just the last two check digits

I have 18 bytes in Jelly

Adám

For those not wanting/able to see the video, this is Brazil's CPF numbers' two check digits.

(Beware, some online descriptions are wrong.)

What does | do in Python?

caird coinheringaahing

Bitwise or

DialFrost

01:26

guys i've been thinking of doign this puzzle

Conway's 99-graph problem

en.wikipedia.org/w/index.php?title=Conway%27s_99-graph_problem

caird coinheringaahing

@Adám Ooh, I managed to golf 3 bytes from my Jelly solution thanks to your video :P

Fmbalbuena

CMQ: How to gain 60 rep without posting a challenge?

caird coinheringaahing

Tho I think you can golf the (t≠10)×t← with 10|, no?

@Fmbalbuena Post a good challenge

Fmbalbuena

Any feedback?

caird coinheringaahing

@Adám I see pi and 42, then in the memory there is: pi, e, phi, avogadro's constant, 4 I don't know, then root 2, and I don't know the rest :P

Ah, googling gives them as: spoiler. Had heard of all except for 13 and 14 (and the unknown ones) before now, just didn't recognise their decimal expansions

DialFrost

01:42

guys is this a good problem en.wikipedia.org/w/index.php?title=Conway%27s_99-graph_problem

caird coinheringaahing

Not knowing the last one is annoying, I should've looked closer instead of just dismissing it

DialFrost

i dont know how to make this into a qns tho

ta Radvylf srik su shilani

@DialFrost If an unsolved problem in mathematics has its own wikipedia article you will not be able to solve it yourself

DialFrost

im not going to thats my point

i want to change it so that u can solve it

but idk how to

caird coinheringaahing

@taRadvylfsriksushilani Unless you think it was set as homework

DialFrost

01:44

haha

ta Radvylf srik su shilani

I doubt there are many remaining problems nowadays that you could do that with though :p

DialFrost

lol

caird coinheringaahing

idk, if you get lucky, you could stumble across the integer that disproves the Collatz conjecture. I'll be nice tho, and let you find it yourselves instead of telling you and spoiling it :P

DialFrost

haha

caird coinheringaahing

@cairdcoinheringaahing OEIS tells me that 137035999 is spoiler but still no idea about 119814024

@cairdcoinheringaahing Speaking of, I rewatched Good Will Hunting the other day - still a great movie, but I now understand the math they were talking about that was "unsolved" until Will comes along, and... it's like, very basic graph theory stuff

One of the questions is literally "what is the adjacency matrix of this graph?"

Unrelated String

02:04

lmao

user

@cairdcoinheringaahing You’re calling that basic? Damn, you must be smarter than even Will Hunting! :p

caird coinheringaahing

I mean: en.wikipedia.org/wiki/Good_Will_Hunting#Mathematics :P

> When asked for a problem that Will could solve, Kleitman and Bohman suggested the unsolved computer science P versus NP problem, but the movie used other problems.

user

Lmao they hited a guy for realistic dialogue and still got ez problems

caird coinheringaahing

Damn, Matt Damon almost proved P = NP :P

user

Or perhaps not

Reminds me of a book i read where a guy has a friend who finds a way to factor primes in not exponential time or something…except the friend is actually a con man (spoiler alert)

caird coinheringaahing

02:14

@user I have discovered a truly marvelous proof that P = NP, which this chat message is too small to contain

5

Fmbalbuena

Any feedback?

ta Radvylf srik su shilani

Too small to be intersting

Kolmos are boring by default

Fmbalbuena

@taRadvylfsriksushilani why?

caird coinheringaahing

It's at -2, I'd suggest abandoning it as a challenge idea

ta Radvylf srik su shilani

Unless you put effort in to find something unique, the same techniques will just be recycled again and again

DialFrost

02:19

i have an idea of a bracket chlannge its quite ez tbh

Input:
( ()) ((()()())) (()) ()
Output:
['(())', '((()()()))', '(())', '()']
Input:
() (( ( )() ( )) ) ( ())
Output:
['()', '((()()()))', '(())']

smh like this

Fmbalbuena

@DialFrost Dupe (probably)

DialFrost

awwwwwwwwwww

ta Radvylf srik su shilani

@taRadvylfsriksushilani And quite often, small patterns can only be efficiently solved with string literals and, at best, concatenation

DialFrost

okey

proably bruh

caird coinheringaahing

@Fmbalbuena Most kolmos have an approach of "print this", but compressing bits that can be compressed. In order to be worth posting, there should be something unique and special about the output that means that are multiple approaches. Unfortunately, I don't see it for this one

DialFrost

02:19

tbh i dont think so

idk

how bout find the indexes of numbers of a given list below a given threshold

caird coinheringaahing

18

Q: Parentheses Clusters

Write a program that groups a string into parentheses cluster. Each cluster should be balanced. Examples : split("((())d)") ➞ ["((()))"] split("(h(e(l)l)o)(w(o)r)l(d)(w)h(a(t)(s)u)p") ➞ ["((()))", "(())", "()", "()", "(()())"] split("((())())(()(()()))") ➞ ["((())())", "(()(()()))"] Input may ...

DialFrost

like [100,[1,59,14,10,40]]

thx caird

any suggestions or issit duped

for the find the indexes of numbers of a given list below a given threshold

caird coinheringaahing

@DialFrost What would this output?

DialFrost

actually this is way too ez nvrm

i've been thinking like [given range,[numbers here]]

then given a number in the "numbers here" list find the index of it in the given range

but its abit too easy right?

caird coinheringaahing

@DialFrost something I've noticed about your challenges that I think you should work on: loosen your input formats. instead of saying [given range, [numbers here]], let people take given range and the list of numbers here however is most natural for their language

DialFrost

02:25

it was an exmaple

not a fiexd input

caird coinheringaahing

This is good for two reasons: one, people don't like haing to deal with unnecessarily complicated input formats, and two: if you ignore the input format, the challenge difficulty has to come from the main task, instead of the input format making an easy challenge seem diffiult

@DialFrost I know, but it reminded me of a bit of feedback I've been meaning to give you

DialFrost

hmmm ic

Sandbox Posts

0

A: Sandbox for Proposed Challenges

Sum Every two elements Task Your task is to Sum every two elements. If there is a list and a number, evaluate [a1 + b, a2 + b ... aN + b] But there is both lists, evaluate (the length of a is 2 and b is 5)[a1 + b1, a2 + b2, a1 + b3, a2 + b4, a1 + b5] You must Flatten to array with depth 2. Every ...

Fmbalbuena

^ Any feedback?

ok, i have to gtg now.

caird coinheringaahing

@SandboxPosts @Fmbalbuena this would be a better challenge if it was just "given two inputs, do the sum as follows" instead of making it about doing the sum for pairs of elements in the input

Unrelated String

02:29

^

caird coinheringaahing

Can we reopen this so it can be hammered as a dupe of this? As noted in the comments, the first is just \$a(n) = 98a(n-1) + a(n-2)\$, so they're essentially identical, and closing it as a duplicate makes it more useful than leaving it closed as unclear

DialFrost

for my new qns that im goign to post

do i limit the output format to [[index,index][...]]

thejonymyster

"do i limit the output format" probably not

DialFrost

okey

thejonymyster

just lkie in general

Sandbox Posts

02:41

0

A: Sandbox for Proposed Challenges

Ragged Matrix A ragged matrix, is a matrix that has a different number of elements in each row. Your challenge is to write a program in any favorable language to find the indices of all occurrences of target in the ragged matrix. Input: A list of ragged lists (can be empty) of positive integers a...

DialFrost

prob gna change the format ltr

for the output/niput

SuperByte

02:52

@taRadvylfsriksushilani can you now unclose my post

ta Radvylf srik su shilani

Have you made improvements to it?

SuperByte

yes

ta Radvylf srik su shilani

It still needs some things clarified:

1. What's the output format?

2. Can the matches overlap?

3. What do we do if there's a tie (or will there never be one)?

4. Will there always be exactly three suspects and three crime scene strands?

DialFrost

i've added this

You can output your answers in any way as long as it is distinguishable what the indexes of the number and matrix are

lol idk

ta Radvylf srik su shilani

5. Will the crime scene strands always be three bytes long? If not, can they be different lengths? Can one contain another?

I know a lot of these seem really nitpicky, but if it's not clarified, it's hard to know if your answer is valid

I think if you clarify all of these things and make it clearer that the input given is a test case and not the only possible input, it'll be reopen-able

Honestly, I'd recommend deleting it, reposting it in the sandbox while you make your changes, and then posting a new challenge when it's cleaned up.

That way you get a fresh start, without the downvotes and votes to close, and your in-between edits don't keep bumping it and sending it to the review queues

DialFrost

02:59

hmm?

ta Radvylf srik su shilani

(wasn't sure if you were still in the room, sorry about that)

Oops you're not even who I was talking to lol

Ignore that

Was thinking about feedback for two different users' challenges and got them mixed up :p

DialFrost

lmaooooo

lyxal

got'em

ta Radvylf srik su shilani

CMQ: Opinions on my new (well, not-that-new) pfp, the Go programming language, and nuclear energy?

lyxal

yes, no, maybe (in no particular order)

ta Radvylf srik su shilani

03:12

I'm actually really starting to like this pfp...it'll be a shame to have to give it so someone else for a little while when we do the Greater Profile Swap of 2022

DLosc

@taRadvylfsriksushilani Looks fine, don't like the syntax, probably a good idea if done right

emanresu A

^^^

lyxal

When is the profile swap?

ta Radvylf srik su shilani

It'll depend on when we're available

Time wise, that is. Date wise, maybe a week from now?

Maybe I should make an anonymous google form where we type in our username and available times of day, and what day we should do it

lyxal

Sure

DLosc

03:21

Out-of-context tweet of the day:

> (if you don't watch im gonna name someone in my park after you and drop them in a lake and then save them right before they die cause i'd feel bad otherwise)

ta Radvylf srik su shilani

I'm thinking a friday during the quieter part of the day/night, to maximize confusion when everyone logs in on saturday

DialFrost

lmao

SuperByte

@taRadvylfsriksushilani I asked 20 people if my post made sense and 20/20 people said it made sense so i mean can you unclose it

ta Radvylf srik su shilani

Doesn't matter if 20 people think it makes sense, it doesn't fit our standards for clarity

I wrote a list of suggestions for what to clarify above

Like I said then, I'd recommend deleting it for now and moving it to the sandbox

That way you can post it again as a fresh start

SuperByte

btw, @taRadvylfsriksushilani I mean I was very depressed when I took LOTS of time to think of this idea and it just got closed, so maybe like unclose it later, I'm gonna fix it now

DialFrost

03:25

no comments for my post yay

...

ta Radvylf srik su shilani

Sign up for the Profile Swap of 2022: forms.gle/rYfmxdtJc4fN4M6fA

DLosc

@SuperByte Yeah, writing a good challenge is hard. I know it's easy to say, but try not to take it personally. Listen to the feedback, rewrite accordingly, and when it gets reopened it will be a much better challenge.

emanresu A

@taRadvylfsriksushilani Pin?

DialFrost

bruh for wheatwizard he's js breezing

SuperByte

thanks @DLosc

lyxal

03:30

Just in case you've ever wanted to see what my pfp looks like full size, here it is

PyGamer0

@taRadvylfsriksushilani what, why, when, where, how

lyxal

PyGamer0

i just dont like my current name

should probably change it to APLgamer0 :p

ta Radvylf srik su shilani

What: We swap usernames and pfps with other users, to confuse people
Why: For fun
When: Friday/Saturday
Where: CGCC, or just chat
How: Changing your username and pfp to an assigned one

PyGamer0

ok sk duration is 2 days?

and after that our profiles will get 11d back to normal?

ta Radvylf srik su shilani

03:31

Well, however long it lasts before a mod gets mad at us and reverts it all :p

DialFrost

sheeeeeeeeeeeeeeee

lxyal luv it!

:)

SuperByte

@taRadvylfsriksushilani You said the crime strands were 3 bytes long: Untrue. I did it in python, and only 4 kb total, which is impossible since 3x3=9, so unclose it please

ta Radvylf srik su shilani

I'm not sure what you mean

I'm asking if the crime strands will always be three base pairs

Or if you could get something like GGGG

DialFrost

anyone have suggestions

to improve my qns

Transcript for