The Nineteenth Byte - 2021-05-28 (page 1 of 2)

« first day (3772 days earlier) ← previous day next day → last day (1075 days later) »

12:08 AM

Any feedback on the changed conditions?

Also this?

0

Q: Turn-based RPS Poker

This is a two-player card game which resembles a vastly simplified game of Texas Hold'em, combined with RPS (rock-paper-scissors) and the most basic mechanic of commercial turn-based card games. The deck A single full deck consists of 30 unique cards, where each card has one symbol out of RPS and...

king-of-the-hill javascript card-games

Meanwhile, the ijk challenge became my top #2 challenge in just two days

12:37 AM

TIL you can have animated slide backgrounds in powerpoint. very cool

Redwolf Programs

12:48 AM

> This survey is anonymous. Please enter your name to affirm you have read and understand this.

10/10

That's very sus

just enter "your name" I guess

@RedwolfPrograms link to survey?

Redwolf Programs

It's school related so no :p

aww.

Redwolf Programs

12:52 AM

Unless you want to provide feedback on the quality of my precalculus class

yes, yes I do

I do too!

Me too.

we have observed its ~~horrible~~ effects on you, so we can judge it

Write your name as Redwolf�Programs and watch them freak out

12:53 AM

Same here!

Redwolf Programs

I was going to say it wouldn't matter anyway since you have to enter my full name, but uh...

Y'all know it already lol

You have 5 people wanting to take your survey for you

don't keep the crowd waiting

*bribes*

Redwolf Programs

And apparently "my grade has been recorded"

Looks like someone forgot to change the default message lol

:D school is so much fun

OldSandboxPosts

1:00 AM

Sandbox posts last active a week ago: Introduction, What is the link to your answer?, Print a BigNum, Solve a Quartic Equation, Calculate \$ \lfloor n \log_2(n) \rfloor \$, exactly

Sandbox posts last active a week ago: Memory KoTH

@lyxal Try turning off your dark mode.

It fixes itself when I start to write a message

caird coinheringaahing

user image

Any additions y'all want to add? :P

@OldSandboxPosts "Introduction" can be edited down and deleted, that's been posted.

the fact it's not in alphabetical order is deeply upsetting

@cairdcoinheringaahing Jelly/05AB1E/…: What if everything was about byte count?

@cairdcoinheringaahing sh: What if everything was a file?

@cairdcoinheringaahing Piet: What if everything was a painting?

@cairdcoinheringaahing Scratch: What if everything was a block?

@cairdcoinheringaahing Whitespace: What if everything was ?

@cairdcoinheringaahing Why is the "L" in COBOL not bold?

@cairdcoinheringaahing Node.js: What if Node.js was a programming language?

1:25 AM

Among us: what if the impostor was sus? 😳

CMC: I found a Jelly 14-byter for the dihedral group challenge. 200 rep bounty if anyone can ninja me while I'm writing the explanation for it

heh, i doubt it'll find it even without the time pressure, but i'll give it a shot

Developer survery: what languages have you used in the last year? Me: Vyxal

well I got 24, fml :/

haha, it's the exact same as caird's, but with a different ordering of the steps

Wait, it doesn't work on empty input :/

1:42 AM

mfw java is better equipped for this challenge than jelly

1:57 AM

And it didn't work for other reasons, now all fixed at 17

Revised CMC: I found a Jelly 17-byter for the dihedral group challenge. 200 rep bounty if anyone can ninja me while I'm writing the explanation for it

Unrelated String

lol

2:12 AM

well, i figured out a potential trick, but i can't get it using a shorter approach than just reordering caird's solution to be less efficient :/

Well, the CMC is over :P

alright, well, i still want to try to get to this independently, but i feel like this is some bool-conversion/bitwise memes that i am not gonna figure out :p

one thing i have is that you can just repeatedly replace sr with rrrs first and then handle separately but i don't actually see a good way of doing the latter step without just doing the split-on-1,1,1,1, flatten, split-on-0,0, flatten thing lol

where there is a bubbler there is always a clever mixed base conversion

true

although maybe not in this question

2:24 AM

inb4 it involves something like using 3 and 7 as inputs and converting to base 19 or smth

2:44 AM

Honestly base conversion builtins should come in two versions, one for highest digit first and the other for lowest digit first

3:35 AM

@Bubbler "highest digit first" should probably be called something else, not base conversion

Funny thing is that all of APL/J/Jelly use highest digit first for base conversion

J has "polynomial" for the reverse

@Bubbler huh. what do you mean? i thought "highest digit first" was something like n -> floor(n/b), floor(n/(b^2)), floor(n/(b^3)), .. (and maybe differences between them)

I mean 13 in base 2 is [1, 1, 0, 1] in them, not [1, 0, 1, 1]. 2⊥1 1 0 1 → 13, 2⊥⍣¯1⊢13 → 1 1 0 1

@Bubbler ah.. little-endian. that would make more sense because powers of the base would correspond to the indices in the result.

isn't that big-endian?

wherein the order of digits is how one writes it out, so 13 is 1,3?

3:42 AM

@ngn but it's just the reverse of big-endian base conversion, not fundamentally different

The current languages only support big-endian, and I'm suggesting to have little-endian too

@hyper-neutrino right

it's not too different but (at least in jelly) with base conversion with a right argument your chaining rules can get messed up if you're forced to manually reverse lol

though yes, it's just a matter of reverse

actually wouldn't <base convert> <reverse> <2-link dyad combine> work? it's +2 though, so even a 2b built-in saves in all cases

using 2,1 chaining being F(left ~ dyad ~ right)

3:51 AM

0

A: Sandbox for Proposed Challenges

Compute the generalised XOR code-golf base-conversion The goal of this challenge is to calculate a function which behaves similar to the bitwise xor function, but in any base. Given two non-negative integers \$ x \$ and \$ y \$, and another integer \$ b \$ such that \$ b \geq 2 \$, compute the g...

4:33 AM

this is way too cursed for my liking

4:49 AM

0

A: Sandbox for Proposed Challenges

Drawing the Super User logo The goal is to create the Mathematics Stack Exchange logo using ASCII characters. It looks like this. The exact output should look like this: _______ ____ | \ | | | | | _ \ | |.| ___ | \-/ / | ...

> Mathematics Stack Exchange

5:03 AM

@hyper-neutrino can you delete chat.stackexchange.com/transcript/message/58150271#58150271 ? I had to remove the video from YT because I realised I left some sensitive info in the video

@lyxal done :+1:

l'thank

although, the video is gone (AFAICT) so i don't think the message offers any info anymore? but i've redacted and purged it anyway

I know that, but I thought having the link gone too would be good

true. never hurts to play it safe :p

in any case all that's lost is a tiny bit of context anyway lol

5:08 AM

Meanwhile, inconsistent I/O can make Jelly 13 bytes in two ways

What's another word for "quick" in the Jelly sense? Asking because I'm thinking of making a Vyxal fork that has postfix operators like Jelly's quicks

Transformer? Macro?

adverb? (I think another language uses this; J?)

Either transformer or adverb

@hyper-neutrino and k

5:13 AM

I guess macro would work, since it is the name for the thing that takes part of AST and transforms it

Except it isn't really a macro in the traditional sense

Transformer sounds the best to me

Rust macros exactly do that

@Bubbler i really dislike it when new languages use established terms to mean something else

and there must be some name used by Lisp family too

@ngn Then what is the established meaning of macro? (Just asking because I wasn't aware of that)

@Bubbler something that rewrites source code or transforms the ast

but an adverb/transformer is not doing that

5:17 AM

@ngn that's what I thought the definition of a macro was too

it does transform the AST though, doesn't it? a Jelly quick can change the arity which therefore changes its chaining

and also it just modifies in-place an atom, it doesn't become part of the execution, it just changes the meaning of a function

For my purposes, transforming seems more fitting

@hyper-neutrino can a "quick" be used to make arbitrary ast transformations?

Well... now I'm not sure I know what that means, lol

A quick does a limited form of AST transformation, in that it pops last n "objects" from the source "stack" and pushes a new source object

5:24 AM

@hyper-neutrino i was asking because i'm not very familiar with jelly. actually, it's hard to come up with a realistic example for a useful macro, as jelly has no variables..

A macro also takes some arguments in the form of source object and creates a new source object out of it, I think?

@ngn Bubbler's answer is correct from everything I know. i'm familiar with jelly for the most part, but not so much with the rest of what we're talking about :p

Hi

@Bubbler does that happen before runtime?

Yes, at parsing time

5:28 AM

ok, maybe you're right to call it that then

so, it's not like operators in apl or adverbs in j&k

if i understand correctly, the "quick" transforms the ast and disappears once jelly starts evaluating

whereas adverbs remain as part of the ast and are interpreted in a special way

Most quicks are like a hybrid of a (possibly variadic) adverb and a macro, because they act on source code objects but they also add a special functionality to them (like / in +/ takes a dyad + and creates a monad that reduces with +; +2/ takes a nilad 2 and a dyad + to create a monad that reduces each 2-chunk with +)

A true macro-like quick is ⁺, which duplicates the last source object

oh yeah does A⁺ become AA$ or AA in the source?

@Bubbler that's not my understanding of macro

@hyper-neutrino AA I think

a macro is supposed to perform a source-to-source or ast-to-ast transformation and disappear completely

but the / in +/ remains there until the actual evaluation

5:40 AM

Hmm

oh, $ doesn't error if it doesn't have enough things to combine. hm.

wanted to test by doing AA⁺$ lol

even tie doesn't error

Then you can say most quicks are not actual macros, though a quick can act as one

yeah, it seems to be AA. F⁺€ applied to [[1, 2], [3, 4]] gives [1, 2, 3, 4] which is consistent with FF€ and not FF$€

So F⁺€ parses like this: push F; pop F and push F twice; apply € to the last item

so the source "stack" becomes [F, F€]

which is definitely macro-like and can't just be a modification of an atom

5:45 AM

ok. we can't say that about +/ though

Unrelated String

does the / not pop the + and push a combined +/

I suppose. though i'd argue it's "pop + and push reduce by add"

@hyper-neutrino what if + is an arbitrary user-defined function ("link"? or whatever it's called), what would / push on the source stack?

Unrelated String

it would push reduce-by-that

well it just pops whatever is on the source stack and pushes "reduce by that thing"

doesn't matter if it's a single atom or a custom link; i don't think there's any distinction

Unrelated String

5:48 AM

it parses quicks using stack operations

yeha i'm pretty sure there is no distinction

There's definitely a wrapper that acts as an adverb, but it's kinda hard to say whether the / in the source code is the wrapper

so these "reduce by .." things can't be primitive operations because jelly can't possibly know what function i'm going to reduce. so, +/ must be some structure that preserves information about the /

That's correct

well, I would argue that that's wrong on a technicality - the structure as far as Jelly is concerned isn't a / plus a dyad, it's just some monad

but the monad's functionality technically preserves information that it's being /'d over a dyad in the Python callable that sits on the source stack

so nvm maybe that is right. but one could see it as being either

ok, so it's a philosophical question of whether that / in the structure is "the same thing" as the / in the source code (or ast)

5:52 AM

i guess :P

The object formed from +/ would look like, say, an object of class Reduce containing the function +. The parser sees the token /, pops +, and pushes a Reduce(+). This action is not quite distinguishable from, say, what APL does, as it sees /, consumes the thing on its left, and creates a Reduce AST node

at least from the outside

I mean the difference just comes down to whether you put the / on the AST or hide it within a node as a nested object lol. So I guess there isn't particularly a difference

(and it doesn't help that there's no other way to create a Reduce object in Jelly)

@Bubbler apl does everything at runtime, as until the very last moment it can't even be sure / means "reduce". it could be "replicate".

At least it does parsing for each line before it can run that line

5:58 AM

@Bubbler not really

only tokenization

I think Vyxal wasn't working for me because caching.

@Bubbler classic example: a b c. how would that parse? (a b)c or a(b c)? you can't tell until you know the nameclasses of a, b, and c, and those will be revealed only at evaluation time.

that's one of the big problems with apl that have been solved in k.

I'd argue parsing happens at evaluation time then

@Bubbler exactly

since jelly links are variadic then they'd have to be parsed at evaluation time

actually, maybe i need to run a test on that

is it possible for a link to be valid in dyad but not monad? or vice versa?

6:03 AM

@Bubbler btw, i'm looking forward to your fastest-code ulc evaluation challenge :) (this is not a promise that i'll be able to solve it)

might be easier to dig through the source code to figure out, but I am assuming they are parsed at runtime, rather than Jelly parsing them in all three arities first and just calling on the appropriate one

A Jelly source line is evaluated in two phases: constructing the source stack (quicks are resolved here), and running the source stack with given arity

Okay. so the parsing is at evaluation then

I guess parsing is involved in both stages

i mean, it is possible to parse it at "compile-time" into all three arities and then just select the appropriate one, at least in theory, right?

6:06 AM

Yes, it's certainly possible

what is the arity of a tie, BTW

is it just the max of all of them?

That's a mystery :P

time to dig through source code for like 20 seconds :p

yeah it's the max of 1 and each of the links' arities

ret = [attrdict(arity= max(1, *arities(links)))]

good

this is IMO the only sensible way to do it

6:09 AM

kinda true

maybe max with 0 could work too?

god imagine if the arity changed each time. i don't even think that'd work but that'd be really fun (in a masochistic sense of the word) to try to understand in a more complex structure

Unrelated String

i'm not sure you could even necessarily do that if you apply another quick to the tie but it would be funny

i want to find a sane application of putting a tie into a tie

AB tie C tie would do [A C B C] ... which... i currently can't think of too sane of a use for

Well, +- tie - tie reduce does "add one and subtract next three"

@hyper-neutrino qué

6:22 AM

or --+ tie tie to be precise, as in [a,b,c,d,e,f,g,...] -> a-b-c-d+e-f-g-...

it is shorter than ---+4 tie

huh

@StackMeter the return value is a singleton list containing an attrdict (custom type for Jelly) whose arity is the maximum of 1 and all tied links' arities

Unrelated String

ties are niche enough to begin with

true

it seems they're either part of some super intelligent solution or just not really applicable

except for challenges like "add, sub, mul, div, cycle"

true

Unrelated String

6:30 AM

i feel like mapping ties can come up as sort of an apply different functions to different elements of an array of known length kind of thing

Jonathan Allan's tie in a reduce trick was really clever (for the non-abelian order 8 group question)

6:50 AM

5k rep!

As a celebration, I'm going to finally do my math homework.

congrats :p and good idea

ok so i slept early to try to get up for BMG which i failed to do and now it's 3 AM and i'm not tired and i still need to get up for work tmr pls help

Lol

ah the advantages of having a completely demolished sleep schedule

I can't, I spent too much time on CGCC for a whole week and have done practically nothing :P

About to join "three repcaps on a single week" gang

ooh, nice :p

most i've had is 5 in a row (6 within a week)

6:55 AM

@hyper-neutrino try to count sheep with transfinite ordinals? :)

5 caps then a gap then 3 caps

ah, back when i had so much time and motivation compared to now

@ngn i will give that a shot :p

ok i'm sleep deprived enough that i could probably sleep right now if i wanted. but a) i don't really want to even though it's a shitty idea to not and b) i don't feel tired at least

Bubbler, kick-mute hyper. That should get him to sleep.

Can't kick mute a mod I don't think

i assume you are joking, but one of our mod powers is total immunity to kick-mutes :p

:(

6:58 AM

ninja'd, and now i'm about to get "you're not joking"'d

That's lyxal's job.

@hyper-neutrino you assumed wrong

No I was joking

An understandable error, but still an error

@Ausername you really weren't. You were and still are A Username

also FWIW, for the record kick-muting is announced to all other ROs (IIRC)

7:00 AM

sighs

What does FWIW stand for?

not sure about mods presently in the room - I am not an RO of this room by definition of the room owner list

for what it's worth

:58152161 ninjad

Guys what does idk mean? Everyone keeps telling me that they don't know and it's kinda getting frustrating.

i don't know

sorry, gonna have to ask someone else

Dang it not you too!

7:02 AM

I'm pretty sure no one knows

We're meant to disapprove golfing edits right?

yes

Good cos I just did.

11

Q: When it is and is not acceptable to edit someone else's post

TL;DR: In what circumstances is it ok (by CGCC and politeness standards) to edit directly a post (be it challenge, answer, sandbox question, etc), assuming the edit is done in good faith? I know people can rollback edits and further edit a post, but in the meantime that goes from the unwanted edi...

discussion editing

There's a meta post for that

Also it was by a suspended user (don't ask)

@Ausername At this rate, it'll take me another two months to get to 10k, and another year to finish my homework.

7:21 AM

xigoi just beat Dennis' jelly answer here 5 years after the fact:

1

A: Map a list of indefinite size to a number!

Jelly, 13 12 bytes Encode, 7 6 bytes 2Bṁx‘Ḅ Try it online! Decode, 6 bytes ȯIBŒɠ’ Try it online! Encodes positive integers in the run lengths of the binary representation. Since we may have zeros, we increment all numbers. For example, the list [0,1,4] is encoded as 10011111 in binary. Numbers ...

ooh, very nice. and no new features used, just a new approach!

That's got to be a record... Also, respect Dennis's two ns.

@WheatWizard TWO NS IN DENNIS!

@hyper-neutrino Yeah, I wasn't certain of this since I know very little Jelly but I thought it was the case.

well at least it looks like no new features to me

not too sure about Œɠ actually

but the rest i think have been around for a long time

it's group run lengths, which seems not new. idk tho. in any case it's a clever approach even if it does use a new feature or two so nice

And it is a significant improvement, not just 1 or two bytes. My guess would be 2016 Jelly could still express group, short enough to beat the old one.

7:30 AM

true as well, yeah

7:45 AM

17 hours ago, by pxeger

@ROs, is it worth removing the now finished BMG announcements from the starboard since we no longer need it?

If we have a starboard, where's our port?

that is a question with no answer

o/ @JoKing

heyo

I've been impersonating you, sorry

you've been joking?

is it just me or has it been a while since the three of us have been in chat at the same time? :p

Unrelated String

7:49 AM

seems like it

require feedback and help with testcases pls

speaking of counting sheep, i seriously need to sleep now lol. o/

\o

8:16 AM

what happened here?

@StackMeter sheep.

I don't get it

there's nothing to get

why did @Razetime say "sheep".

well the most recent 2 messages before you entered the room were about counting sheep

sheep

32 mins ago, 1 minute total – 2 messages, 2 users, 0 stars

Bookmarked just now by pxeger

8:23 AM

ok

what happened before that

I dunno

@Bubbler :(

I haven't repcapped once

hi

hi @Wasif

@StackMeter hello

8:33 AM

also nice to see Wezl's put the e back in their name

@Wasif I do not know what challenge to do next

@StackMeter you didn't ever post a question on CGCC until now, are you thinking to make one

yes

(If you are talking about that)

I am talking about that

you have no obligation to post challenges

8:44 AM

I know

but this is the first one that has real potential

the new edit for my question is now out

@RedwolfPrograms I've updated your SE Chat Monospace Fix to include the transcript pages and the 2 other chat domains: gist.github.com/pxeger/935cba86c281003b9f788e7f7f79c44b

it here

9:02 AM

@StackMeter I have an idea for you, determine the quarter of an angle. given an angle in degrees (degree is an integer, positive or negative). Return which quarter of a circle the angle is in.

that's just a%360/90

err no....

how so

but anyway thanks for the formula

what do you mean by "which quarter of a circle" then?

9:05 AM

I used to draw the angles manually for determining them in my math class

@pxeger my English mathematical vocabulary is bad, and it is hard to explain in chat. assume a circle divide It in four equal parts, divider lines are it's diameters. Start to draw an angle clockwise from the center of the circle if it is positive otherwise anticlockwise. determine which quarter it ends. the circle thing is not necessary at all

I just used for ease of imagination

I think we're getting at the same thing, but I don't understand why you said "err no"

10:03 AM

CMQ: PS5 scalping; yes or no?

imo it's bad

Jesus, Mother Meta is a dark place

caird coinheringaahing

10:25 AM

What's wrong with MM?

11:04 AM

@cairdcoinheringaahing there's a lot of overheated discussions that my brain is not ready for

I hereby forbid people to use my username in code golf questions and or answers...searching for Lyxal gives 69 search results (this is a joke of course you can still use it)

brb, going to edit "lyxal" into one of my answers for no reason

lol

welcome to the land of post-69

Caching really do be helping me out rn

why a cache

caches have issues with them

such as syncronization

for example

if you were to get 80% of the way on a big project

then I came

and it loaded your cache

what is to stop me from replacing that with some rude message

or this

11:16 AM

nope

Idk why you thought I'd be that susceptible

bcos you always have been

now, if you excuse me

I have an interpreter to finish

caird coinheringaahing

@lyxal I'll just refer to you as "The Lyx-ler" then :P

@pxeger <whisper>do you think it worked</whisper>

do you know, I think most golfing langs are made from the creator furiously typing "how factorial" in chrome, and things like "how map list python"

the struggle is real

11:50 AM

@StackMeter I didn't actually do it

@StackMeter still nope

There exists a list of 22 tips for rickrolling people I once wrote for @user

Ask them if they'll share it with you

caird coinheringaahing

12:46 PM

@lyxal That list was for me, not user. Here you go @StackMeter

1:13 PM

@cairdcoinheringaahing nice try

requesting more feedback on this meta discussion, specifically in regards to this answer since the other answer is a good suggestion but was tried and didn't work in the long run

caird coinheringaahing

I regret not posting the "as a ticker feed" answer back when the discussion was more volatile in votes

I see pretty much no downsides to it and it's easily reversible like you said if it does become disruptive somehow, so as long as it doesn't get any specific negative feedback I think it's fine to just add it, maybe once the other ROs have confirmed it.

I think a ticker feed for a few more catalogue posts (quine, fizz-buzz?) would be a good idea too

showcase of languages?

let's just do (ticker) feeds for all questions, and remove the ones that get annoying :)

caird coinheringaahing

1:28 PM

I think that if we want and like a Hello World ticker, then we can consider expanding it to a couple more challenges that are used for new languages

@Wezl It'd probably be easier to just set up a chatbot to post every new post :P

but more disruptive

well the chatbot doesn't have to onebox them

how do you disable oneboxing?

caird coinheringaahing

prefix the message with a zero width space

ah ok.

1:31 PM

or just manually do [title](url)

caird coinheringaahing

codegolf.stackexchange.com/questions/226361/…

Observe :P

What about a ticker feed for all new posts? No different than having New Posts open in the background in another tab, and you don't have to switch pages

right, but now we don't know what that post is without clicking :p

caird coinheringaahing

@pxeger Ticker feed dropdowns increase in size with new posts, they don't have a limit

You could leave TNB open overnight on an active evening and get a screen full of random answer links in the morning

oh I thought the ticker would be more like the starboard, with old items rolling off the bottom

1:34 PM

which is fine IMO

caird coinheringaahing

@pxeger I don't think so (they're kinda hard to test tbh)

lemme try it out. I'll use SO's new posts feed to test since it's fast updating

caird coinheringaahing

Try here

I've got a ticker feed on SO's new questions

ok

reading the Vyxal interpreter

@lyxal, surely you could write a cleaner code than to just use the first 100 factorials

nope, it's about a million times quicker than calculating them

1:44 PM

^

it's not like it doesn't support the 101th factorial

it just caches a reasonable amount so it doesn't take forever

there's a built-in in the math library

caird coinheringaahing

That still takes ages for large inputs

it'll still be slower than just caching it

every factorial takes ages for large inputs

caird coinheringaahing

And caching is a way to reduce that by as much as possible

1:46 PM

yeah so why not cache a reasonable amount?

ok

@Wezl now I used it to get a rot13 tool

@Wezl CyberChef is my goto for anything like that

(unless it's quick to write code to do it in my shell or in Python)

right, I'm trying to use it in the shell, I don't want to wait for a webpage to load

you can do a one liner with tr

1:58 PM

that's what I was looking for

tr a-z n-za-m <<< rot13

2:55 PM

@lyxal Here it is

well lyxal already failed #7

@hyper-neutrino Did it work btw?

When you opened the document, did anything happen?

nothing unusual; was I supposed to get rickrolled or smth?

Yeah, a popup should've appeared. Lemme try it again

« first day (3772 days earlier) ← previous day next day → last day (1075 days later) »

Transcript for

May '2128

The Nineteenth Byte

The Nineteenth Byte: General discussion for codegolf.stackexc...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile