The Nineteenth Byte

caird coinheringaahing

00:04

Ooh nice, 47777 rep :P

Just need 30k more :P

Bubbler

incidentally I have 65777

New Posts

1

Q: Persistence of a number

The persistence of a number \$x = d_1d_2d_3...d_n\$, with \$d_1 \ne 0\$, under some function \$f : \mathbb N_0 \times \mathbb N_0 \to \mathbb N_0\$ is defined as the number of applications of \$f\$ to the digits of \$x\$ before it reaches a single digit integer. That is, if we have the map $$I_f:...

lyxal

00:49

@NewPosts just in relation to this: vyxal has an element that lets you override the contents of STDIN (that is, whatever was given as input originally is discarded). Is it valid to use it to take a function (and all the other required inputs) as input on the provision it happens in the header and there's no regular STDIN given?

Steffan

anyone wanna try brute force 81, 82, 75, 66, 78, 80 => 9, 5, 4, 3, 3, 1? I've tried and failed. bitwise ops are fine

caird coinheringaahing

@lyxal I'd say no

If it doesn't work when STDIN is given normally, then no

lyxal

Except functions can't be given through STDIN

The function object is already on the stack in the header

This is just setting STDIN manually instead of having it given on invocation of the interpreter

Bubbler

If there is a way to take function input, I suppose you need to use that

caird coinheringaahing

I'd say you can take the function object on the stack, as that's equivalent to assuming it's already defined in the program

Bubbler

00:52

and imo Vyxal has a proper way to do it: write a function that takes a function from the stack

caird coinheringaahing

^

Bubbler

emphasis on "write a function"

caird coinheringaahing

@lyxal The problem isn't the function, it's the integer input: that should be provided in a default I/O method

lyxal

So what if there was an append to STDIN built in?

Could a function object be added to STDIN from the stack?

caird coinheringaahing

I think that should be discussed on meta

@lyxal But, my personal opinion is that you'd have to count the code that adds it to STDIN

Bubbler

00:56

I mean, you're hardcoding the function input either way, which is not normally allowed

OldSandboxPosts

Sandbox posts last active a week ago: Does there exist a synchronizing word? code-golf decision-problem string, Contrasting Colors code-golf

thejonymyster

@Steffan so far i have 3-(x%6)+4*(x%2) which gets all but the first two

Bubbler

80 81 82 -> 1 9 5 sounds hard

thejonymyster

yeah haha

im sure theres a better way involving like

writing a program to find it

which is probably what was meant by brute force

Steffan

yeah

I tried lots of different forms but couldn't find any

thejonymyster

01:03

lol i got p close on my own though :P

but yea fair

Steffan

let me try of form a-x%b+c*x%d like yours

thejonymyster

yea mods let you get away with tons of noisy stuff since its like

you might want to also add something to the x's themselves or something to offset the modulos

Bubbler

@lyxal I guess you can introduce a notation for a function on stdin, e.g. Vyxal code surrounded by some delimiter

thejonymyster

so youre not relying on just the way numbers are

i have no idea what terms to use for this

Steffan

nope :/

eliminating the first two gives tons of possibilities

thejonymyster

01:10

i have something if you can use bools as 1 or 0

a little long though

Steffan

sure

this is python

thejonymyster

3-(x%6)+4*(x%2)+(x>80)*(x-76) lol

Steffan

ah nice lol

thejonymyster

since its just those two above 80 haha

u can always go back and golf it later when you wanna throw away time and/or computing power

Steffan

unfortunately that's longer than hardcoding it :/

thejonymyster

01:12

yea that was the fear LOL hm

Steffan

the bigger numbers are char codes so it's shorter than using num lists

thejonymyster

well how long is hardcoding it

just to know

Steffan

so it's just b" "['QRKBNP'.find(x)]

emanresu A

Have you tried using hash?

thejonymyster

nice

Steffan

01:14

byte-strings help a ton

in python 3, hashes differ every time

for strings

only works in python2

emanresu A

:/

thejonymyster

unrelated but check out my regex to match even integers in english (ur|lv|[dothxy])(te)?e?n?$ :P

Steffan

lol

lyxal

@Steffan gosh dang it I know exactly what this is for now :p

caird coinheringaahing

@thejonymyster My regex to match odd integers in English: e :P

5

thejonymyster

01:22

thirteen

but also lol

Bubbler

seven

thejonymyster

(etc)

lyxal

@cairdcoinheringaahing 8 confirmed to be an odd integer

caird coinheringaahing

My bad, I golfed 1 byte off it :P

thejonymyster

well tbf that isnt "while also not matching even integers" so

caird coinheringaahing

01:23

@lyxal I never said it only matches odd integers :P

thejonymyster

yeah haha

Bubbler

actually sounds like an interesting challenge

thejonymyster

which way around

Bubbler

match all odd numbers and no even numbers; support up to (excluding) a billion (for example)

or even just a regular decision problem

thejonymyster

well my regex works if you allow inverse output :P

i wonder if its shorter the other way around though

emanresu A

01:29

@thejonymyster What about billion?

thejonymyster

ends with n

emanresu A

Isn't matched though?

Oh wait nvm the (te) isn't

thejonymyster

my regex correctly matches all even integers and correctly matches no odd integers

emanresu A

Nice :p

thejonymyster

im gonna try going at it the other way though

might be shorter

(might not be, but we never get places by not trying them)

Feeds

01:34

1

Q: Can Function Objects be Appened to STDIN From the Stack?

This is mostly specific to Vyxal, but I've worded the language so that any hypothetical stack language could utilise this Say you have a stack-based language which supports function objects on the stack. And say that there's no way to take function objects on STDIN - they have to be placed in the...

discussion policy code-golf defaults

lyxal

Hmm yes 24 minutes

Man I miss the bots

thejonymyster

odds but not evens is way harder

Jo King

just run it through an answer from this, ez

thejonymyster

...yknow, im curious, i will do that

Jo King

though i suppose that is a modified regex flavour

thejonymyster

01:55

oh yeah doh it only takes 1s and 0s haha

emanresu A

You mean this rigt?

thejonymyster

sadly doesnt seem to work either :P

for my porpoises

Jo King

02:21

@emanresuA oh yeah, that's what i was looking for

Bubbler

03:12

:61933025 Why deleted?

Steffan

Misread a message

Bubbler

For anyone interested, someone on the internet solved this problem in Prolog, with nice explanation for each step involved: john-h-kastner.github.io/regex-complement-prolog — Bubbler 7 mins ago

Steffan

it's linked in the comments of one of the answers

I never got around to golfing it, but here's regex complement implemented in Prolog — ankh-morpork Feb 20, 2020 at 1:44

Bubbler

Oh there is, somehow I didn't notice it even though I did answer the challenge

maybe it's because it was posted after my answer

it deserves a little bit better visibility anyway

Steffan

03:29

also the link is broken

I had to trace it back to the GH repo, then it redirected to the new username

Bubbler

oh

thejonymyster

03:42

i got the 'odds no evens' regex to two bytes longer than the 'evens no odds'

a bit stuck atm

(?<![^efinr]r?te+|lv|o)[en]$

oh wait it breaks nvm

i have no idea when it last actually worked because it was breaking for something i hadnt accounted for

:P rip that i guess

Do you think anyone would be upset if i submitted the working regex to the even or odd decision problem :P

actually in general;

Steffan

well it doesn't say you can take it in english

thejonymyster

right yeah lol

but i am curious on peoples opinions in general;

Bubbler

what's with the semicolons

are you talking in C

thejonymyster

im about to say a thing is why

namely:

CMM: if a problem takes numerical input, would it be OK to take it in english?

or would that be a violation of policy, either by not being a serious contender (i personally disagree), or being "cheaty" input format / not "reasonable"

or anything else

Bubbler

I'd say no because it'll trivialize "convert number to english" challenge, and the english reading of large numbers are not well defined iirc

thejonymyster

03:54

fair

Unrelated String

trivializing a specific challenge is a bad argument because challenges can always retroactively override defaults (and in cases like this with existing defaults often do)
but i would agree that it's too poorly defined and doesn't have any real rationale

graffe

06:25

If you want to write code that is really fast and you want to get it to work in 2 days or less, what language would you learn?

Clearly not C++ :)

Say you can already code in python

mousetail

@graffe Rust

My catch all answer to all language decision questions

2 days ago, by Radvylf Programs

My parser worked first try after I got all the compiler errors cleaned up

Of course getting the compiler errors cleaned up is the hard part

graffe

@mousetail surely that can't be done in 2 days???

mousetail

Depends how fast you learn

What's your goal?

How advanced knowledge do you need?

graffe

@mousetail let's say I want to be able to re-implement codegolf.stackexchange.com/a/251540/108721

mousetail

I think it's possible

graffe

06:30

I can imagine doing that in Julia in 2 days from what I know of it

mousetail

Then why not use Julia?

graffe

But rust looks like a huge amount to learn for a python coder

mousetail

It's really fast though

You asked for a fast language and fast languages always need to cut some corners in terms of usability. IMHO rust cuts the least corners possible so is the easiest

But if you know some Julia already and think that's fast enough I'd recomend that

Bubbler

Looking at the list of imports, I might not be able to port it to Rust in 2 days

the main blocker being matplotlib

mousetail

True, the graphs would be more difficult. I assume he just wants to re-implement the algorithm not specifically the visualization

I did find this: docs.rs/plotters/latest/plotters

seems a capable alternative to matplotlib

pxeger

06:47

Yeah you could just use rust to calculate the raw data and then feed it into a python script to make the graphs

thejonymyster

CMC: given a positive integer N, output a positive integer M with exactly N digits, where N is a substring of the decimal form of M

fixed

graffe

@mousetail I asked for a fast language a newbie could implement something fast in in 2 days :)

mousetail

It's a trade off

graffe

@mousetail yes just the algorithm

mousetail

I think a beginner could do it in rust in 2 days, if they have sufficient prior knowledge of linear algebra and statistics

graffe

06:56

@mousetail what would you point them at to do that?

@mousetail that looks pretty cool

mousetail

Do the hello world and the binary search guessing game first

graffe

I see rust has einsum

docs.rs/numpy/0.13.1/numpy/macro.einsum.html

And in fact numpy!

mousetail

I wouldn't use it. It includes a python interpreter so you wouldn't get the performance benefit

graffe

Ah....

So the einsum part might be a problem too

mousetail

There are probably ohher libraries that offer that

att

07:53

@thejonymyster apl, string output ⊢⍴⍕

Fatalize

08:16

@thejonymyster Brachylog, 6 bytes: ~l.s?∧

pxeger

08:51

CMC: determine if a given integer is exactly two less than a power of two

emanresu A

@thejonymyster Vyxal, 2 bytes: Try it Online!

pxeger

You can assume it's positive

emanresu A

Vyxal, 2 bytes: Try it Online!

I think all[bin] works in Whython

Oh wait nvm that's not how bin works

lyxal

@emanresuA two less, not one

emanresu A

Oh. Increment then

New Posts

10:17

1

Q: Number of ways to make an amount with coins

This is not a duplicate of Sum of combinations with repetition. This question considers 1+2 to be the same as 2+1. The other question doesn't. OEIS link - A000008 Background If you have 4 types of coins (1, 2, 5 and 10 cents), in how many ways can you make an amount? There are 4 ways of making ...

code-golf number sequence

pxeger

10:33

spam: codegolf.stackexchange.com/a/251678

emanresu A

They're getting smarter lol

pxeger

another one codegolf.stackexchange.com/a/251679

there's more

emanresu A

frick

Soon they'll be correctly formatting them

pxeger

keep watching FA queue

emanresu A

:D smokey picked them up

@pxeger Where?

pxeger

10:39

a third one at least, codegolf.stackexchange.com/a/251681

emanresu A

Saw that, smokey did not pick it up

Hm I wonder which mod that was

pxeger

all deleted by Community, so no mod, just getting enough spam flags from known users

Jo King

i mean, technically i cast the last spam flag...

emanresu A

Except six spam flags should cause six automatic downvotes right?

(and mod spam flags are binding)

pxeger

@emanresuA I didn't think spam flags caused automatic downvotes

emanresu A

10:47

IIRC they cause system downvotes or something?

pxeger

Maybe it's because I always reflexively downvote before spam flagging

Can mods see deleted posts on sites they don't moderate without having 10k rep there?

forest

No.

emanresu A

What I don't get is what they were trying to advertise

pxeger

just looks like some spammy websites rehosting games

emanresu A

I guess

graffe

12:10

I was reading some lecture notes where it has ibb.co/tMb5ZzF . What does the dot over the union sign mean?

I guess it is the symbol for disjoint union

pxeger

12:28

> V = A Ù B (V is the disjoint union of A and B)

no way

lyxal

12:44

Funny story: I was doing some vyxal development and went to try out the changes I made on the online interpreter thinking I was running it locally. Turns out I was actually testing on the actual online interpreter, which explains why nothing was changing when I changed things lol

graffe

13:12

@pxeger have you ever seen that symbol before?

pxeger

Probably

graffe

:)

what probability?

pxeger

Unknown unknown

graffe

:)

Ginger

@lyxal hell yes

New Posts

14:02

0

Q: The "Fly straight, dammit" sequence

Background "Fly straight, dammit" (OEIS A133058) is a sequence of integers, which has these rules: \$a_0 = a_1 = 1\$ \$a_n = a_{n-1}+n+1\$ if \$gcd(a_{n-1}, n) = 1\$ Otherwise, \$a_n = \frac{a_{n-1}}{gcd(a_{n-1}, n)}\$ In pseudocode: a[0] = a[1] = 1 if gcd(a[n-1], n) == 1 then a[n] = a[n...

code-golf sequence number-theory

Sandbox Posts

0

A: Sandbox for Proposed Challenges

Render a triangle in vulkan vulcan The vulkan API is famous for requiring graphic engines to be very verbose. I am curious to see how far we can escape this trend. Task Using vulkan and no other rendering API, have a triangle rendered to screen with a color pattern as seen in this image: For ren...

mousetail

@NewPosts Is it ok to answer your own question immediatly?

lyxal

it's generally discouraged, but not forbidden

Adám

@mousetail Yes, but usually better to give others a week, especially if you use a more popular language.

mousetail

I guess in this case it helps illustrate the problem

Adám

14:05

You could add a not-golfed reference solution to the problem description.

Razetime

14:21

@mousetail usually you put your own answer as a reference impl in the question

graffe

14:38

What is the hardest math challenge we have ever had?

You know, one that a professional mathematician wouldn't be ashamed to have answered

user

Determine if the Collatz conjecture is true or not :P

graffe