« first day (4448 days earlier)      last day (397 days later) » 
00:00 - 18:0018:00 - 00:00

12:26 AM
i got fanatic badge :)
 
@NewPosts Possible extension to this: Shortest distinguishable subsequence
 
what do you mean
 
Instead of restricting it to a single contiguous subsequnece, allow selecting any group of characters/indices.
 
hm, i don't think it would make it easier or harder but it is an interesting idea
 
would kind of turn it into a multi-enklact
definitely a different challenge yeah
 
12:33 AM
they should make mini-code-golf for gamepigeon
 
1:00 AM
Sandbox posts last active a week ago: Vertically Aligned JSON formatting
 
1:13 AM
@NewPosts I genuinely think I just don't get this challenge :/
 
Ok let me explain it to you
 
think about it like codegolf.stackexchange.com/questions/164208/… except no modular indexing and it's a slice instead
 
the slice (1, 4) means the substring from index 1 through index 3 in a given string
 
alternatively, read my awful jelly answer
 
@Jacob Okay, first of all, I hate this
Just from a challenge writing perspective, this is a bad way to do indices
 
1:16 AM
why, it's how slices work in most programming languages
 
Not neccesarily.
 
it really isn't
The most obvious way to do slices is to index from x to y, inclusive, and let the langauge decide 0 or 1 indexing
 
I'd've assumed (1, 4) refers to the span of 4 characters starting from and inclusive of character at index 1 (0 indexed)
 
But, either way, that doesn't impact my understanding of the challenge
 
If I were you (jacob) I'd allow 0/1 indexing and inclusive/exclusive slices. (since it's just another way of writing the same thing)
 
1:18 AM
> Your program may either use zero-based indexing or one-based indexing, but please specify which is being used in your submission.
good point on inclusive/exclusive though
 
can you explain what you would like changed in terms of inclusive/exclusive slices in the spec
i'm not sure we're speaking quite the same language here
 
Okay, is [2, 2] a valid output for the first test case?
As you can identify the string based on the second character
 
that would be a zero-length slice
 
So it would need to be [2, 3]?
 
yes
which is a valid output as seen in the test cases
 
1:22 AM
Cause, [2, 2] to me, as a slice, is "start and end and index '2'", i.e. "only index 2"
 
what language does slices like that
 
Jelly
Multiple languages that do slices through range indexing
 
do any practlangs do that
 
Praclangs? In my CGCC?
 
:(
 
1:23 AM
@Jacob It's a combination of 1 or 0 indexed, and semi-exclusive ranges
 
i mean in my experience i have only seen languages do slices the way i explained in the challenge (e.g. in JS and python)
 
Are inclusive range allowed?
 
@cairdcoinheringaahing zero or one indexed are both allowed, that's fine
by inclusive range do you mean that 1-indexed slice (2, 3) in string "abcdefg" would be "bc"?
 
For the first, is [1,2] allowed for 1-indexed, inclusive?
I.e. the first two letters?
 
for the first test case?
 
1:26 AM
Yes
 
@Jacob not that I know of
Jelly seems to be the edge case
 
@cairdcoinheringaahing sure, because "ha", "an", and "hu" are all different
i will modify the spec to allow inclusive range now if it makes you happy -.-
 
It's nothing about making me "happy"
It's about clarity
ranges and indexes can be fuck-y with inclusive/exclusive and 0/1-indexed
 
well before it was first index inclusive, last char exclusive, but now i changed it so it can also be both inclusive
i have to leave now so goodbye
 
@Jacob I think what you had is pretty standard for 0 indexing
 
1:30 AM
I'll finish up my answer then clean up the comments, thanks for clarifying :)
Also, jelly why are you like this????
 
skill issue tbh
/s
 
don't you fucking start
 
heh heh
 
I don't fucking believe this
After all that
I couldn't even outgolf Unrelated
Imma go cry face down in bed
 
i’m glad you’re crying
 
1:36 AM
I will nuke your fridge
 
@cairdcoinheringaahing I requested it be done this way while it was in the sandbox because C++ STL does [start, end)
 
That's the opposite of what feedback you should provide in the Sandbox
 
well, I think I asked that it be allowed to be [start, end), not that [start, end] should be forbidden
 
Challenges should be as agnostic as possible
@Bbrk24 Oh, fair
 
because at the time it was only [start, end] and I said "well what about [start, end + 1) like in the STL"
Swift lets you do either, spelled arr[i...j] and arr[i..<j] respectively
 
 
2 hours later…
3:51 AM
Y'all know how everyone is all about memory safe languages lately? Well I think there should be a memory sabotaging language. Instead of preventing memory leaks with things like borrowing, the compiler inserts memory leaks into your code for you. And it also never reports compiler errors, instead classifying good code as compiler errors.
 
4:13 AM
That.. would certainly make a new kind of esolang
 
Intercal vibes
 
it also wouldn't tell you where it placed the memory leaks
 
What are integer solutions for (log_2 x)^8 < x
 
0
Q: Single assignment for +/- INF and +/- NAN

RARE Kpop ManifestoUsing one single statement, numerically assign the IEEE 754 : +INF to variable a — inf also acceptable -INF to variable b +NaN to variable c — a sign-less nan being displayed is acceptable as long as the code demonstrates how it arrives at a positive nan -NaN to variable d — a sig...

 
I know x=1,2 are solutions
But what else?
 
4:24 AM
nothing
those are the only 2 integer solutions
> berdfy
one letter away from greatness!
 
@lyxal what about x=3?
 
39 is not less than 3
 
I put a lot of work into constructing and golfing this, and find the result to be quite satisfying: .NET regex for matching primes in decimal up to a specified maximum
 
@lyxal berdly more like nerdly
 
@lyxal ah yes. What about really big x?
 
4:28 AM
Wolfram Alpha says 0.527363<x<2.14354 or x>1.2961163241337846319×10^13
integer solutions would be 1, 2, and all x from 12961163241338 inclusive
 
Thanks!
 
Why do you want integer solutions though?
 
@Bbrk24 What if the index of the element to be accessed is inputted?
But the optimization of [>>...>>] is definitely what one should use to do that.
 
@NewPosts Does someone want to comment on this and help the OP out?
It seems to be in some need of modification
 
@Bubbler I was wondering when arxiv.org/abs/2203.03456 was the fastest algorithm
 
4:33 AM
If [>>...>>] is optimized to O(1), one can do something like binary search
And at the very least, even if there are modifications on the tape
the interpreter can build a bunch of balanced trees and do [>>...>>] in O(log n) anyways
 
@Bubbler the answer is, never realistically :)
 
@Simd It's new, which means there are spaces to improve.
 
@TwilightSparkle that is true. Getting 8 down to 6 would make it almost useful
 
5:06 AM
> near-linear algorithm
> O(m log^8 n log W)
I guess near-linear has become more of a buzzword in the area of algorithms
 
5:38 AM
@Bubbler Relatively closer to linear, anyways.
On a completely unrelated topic: is there a brainfuck visualizer with more memory?
 
6:11 AM
I haven't seen any, other than the copy.sh one (visualizer shows 30000 cells iirc)
 
 
1 hour later…
7:20 AM
@Bubbler yes, sadly
 
so what does near linear actually look like?
 
technically it means anything in the form of n (linear) times any power of log n, log log n, etc.
it's called "near linear" because it grows slower than any n^(1+x) for any x>0, as n goes to infinity
 
so why it become buzzword
and why is so bad...
 
because everyone is hyped with fast-looking algorithms
but an n (log n)^8 algorithm is actually slower than n^2 one until n > 10^13, which means the former is pretty much useless in reality
 
near-linear mean faster??
 
7:36 AM
an n log n algorithm is usually faster than n^2 for like thousands or millions of items (e.g. quicksort vs. bubble sort)
so yes at the basic level
 
@Bubbler So how do you use copy.sh to visualize?
 
but the common sense breaks down when you get stuff like n (log n)^8
@TwilightSparkle set "dump memory at char", insert that char somewhere in the program, and "view memory"
@Bubbler it also happens when the log power is pretty low (2 or 3) but the constant overhead is too large
 
ah so near linear is actually not fast at all??
 
sometimes yes
 
wat, so is near linear fast or not fast
wait log is base 10 here right
 
7:45 AM
base 2
 
wat now
 
do you know what "time complexity" is?
 
isnt log by itself base 10, u need to write log_2 for base 2 right
@Bubbler ya its that O thing with n inside
 
log with base 2 is common in computer science
 
huh i never see it use like that before
so computer science they write log_10 instead?
 
7:47 AM
there is simply no use for log with base 10 when talking about algorithms
 
huh why not
log base 10 is pretty useful
like finding how many digit inside a number
 
because computers use bits
 
so what about base 10 number
 
well, they might use log_10 in that case
 
that is confuse me, they gotta stick with one convention lol
i learn it like log is base 10
 
7:51 AM
i mean literally log_10, not log for base 10
in computer science
in math and other parts of science log is base 10, yes
 
wat so they use log for both base 2 and base 10 in computer science??
 
nonono
just base 2
 
that is strange
huh
so log isnt base 10?
 
just in computer science
 
have i been lied to my entire life...
this is hurting my brain lol
 
7:53 AM
it's just that different fields use different notation
it'll be common when you get to college, or read papers
 
oh
this kinda thing is normal? its like more confusing than helpful ngl
 
yeah it's confusing, but it's life (and math and science)
 
i think eng tends to use 10, math tends to use e, and cs tends to use 2, but a) I could be wrong b) different areas of them also probably use the other ones
 
eng means english??
 
engineering
 
7:56 AM
wait isnt ln base e, wat now
i see ln in math all the time...
 
it is
 
okok so log in math is base 10 right
 
yes until high school
 
sometimes
 
i can't guarantee for college or higher
 
7:57 AM
basically it's just inconsistent and you just need to be sure you know which one's being used lmao
 
yes until high school????
 
some college texts do use log for base e, iirc
 
tf they be doing in college???
 
^^^^
 
yeah. I had at least one math class in uni write log for base e
 
7:58 AM
@Bubbler ok just no, that is gonna fuck with my brain too much
that seriously cant be a thing, like just stick with one notation, it really cant be that hard right
 
idk lol
 
unfortunately it really is apparently too hard to stick to one notation
math uses i for the complex unit but engineering uses j (typically, IIRC)
python uses the latter as well
 
j??? i thought that was only a weird python quirk...
 
wait until you see "natural number" has two definitions
one including 0, and the other not
 
o.0
 
8:01 AM
who da hell see imaginary number and think ah yes, we will use j to denote that
like just no
@Bubbler bruh what natural number dont include 0, aint no way ppl say it does
 
why i though
 
cuz imaginary start with i lol
 
"i"maginary I suppose? but that's also not always what they're called either
 
ya they call complex number or smth but who da hell gonna start using c for sqrt(-1)...
 
g for gauss would've been cool
 
8:03 AM
and j makes even less sense than i
whats the rationale behind j anyways
 
not sure if this is actually why
but this seems like a compelling rationale
 
Real gamers use ° or ı
 
@hyper-neutrino huh i never hear about i j k vector before
 
real gamers use quaternions
 
da hell are quaternions, i keep seeing ppl say that here but like tf is it
@lyxal gamers using imaginary numbers lol, as if
and also what kinda symbols are those, theres no way u gonna use a degree symbol for imaginary number
 
8:07 AM
(vyxal reference)
 
Try it online! for the second symbol
 
@AidenChow excellent question, I looked at the wikipedia page and didn't understand a thing
@lyxal oh wait did we change it for v3? I thought you were referring to jelly there lol
 
@lyxal ok bruh literally that is why i dont touch golfing langs, the notations are going to fuck with my brain
 
Actually changed
@hyper-neutrino ^
 
8:10 AM
make notation actually make sense, da hell is degree symbol for complex number...
 
@AidenChow I'm halfway done with this fricking 60 page explanation
 
@AidenChow what you choose when complex numbers weren't part of the original design
 
@mathscat 60 page, what now
 
@AidenChow you might like literate mode which uses i
 
that literally looks like a wiki page with that table of contents u have there...
@lyxal ye using i make more sense
 
8:12 AM
yeah, I forgot most of it
pretty sure 3b1b has made a video about quaternions though
 
ok i literally scroll to a random section of quaternion and i see slerp, bruh some interesting acronyms there
 
@hyper-neutrino ah right
 
@AidenChow only when you stop caring about things being 1 byte :p
 
 
2 hours later…
10:10 AM
@AidenChow hello
 
@Bubbler Thanks
 
11:03 AM
0
Q: Is byte count on TIO valid, specifically for non-ASCII C code?

anatolygWhen I have non-ASCII characters in my C code, TIO counts them as 1 byte. For example, the following C program prints 56 music-related characters on the terminal. Link TIO reports its length as "96 chars, 96 bytes (SBCS)". I understand "SBCS" as follows: C never standardized its source code form...

 
why the hell is c listed as sbcs on tio?
 
sbcs??
 
single byte character set
 
ah ok
 
11:34 AM
0
A: "Hello, World!"

PlaceReporter99HQ9+ 1 byte H HQ9+ is a joke programming language made by Cliff L. Biffle.

 
I'm gonna make an anti-golfing language
based on python
 
Java already exists
8
 
based on two words:
sesquipedalian
pneumonoultramicroscopicsilicovolcanoconiosis
 
how quickly can you compute the Earth mover's distance between two arrays of n numbers?
 
@Fatalize oof
 
11:53 AM
0
Q: A003056 in FRACTRAN

user1100017I have a question: How could we write a FRACTRAN to calculate the n-th term of sequence a(n+1) = a(n-a(n)) + 1, a(0) = 0 (A003056)? (What is the most common trick here for Recursion, which of these examples much more complicated than sequence A003056 that I have just mentioned?). I need your help...

 
don't can it, post it :p
 
@mathscat I see you mention an image to add and I agree that an image of examples would be helpful
 
What makes a polyomino "free"? — Bbrk24 Mar 17 at 18:47
 
ah, I should probably paraphrase the second sentence
 
12:05 PM
CMM: , given the number of questions we get about it?
 
@lyxal ASCII is a SBCS I guess
 
Huh, I see
 
12:36 PM
0
A: Sandbox for Proposed Challenges

emirpsCosine Similarity The cosine similarity of two vectors \$A\$ and \$B\$ is defined using their dot product and magnitude as: \$\frac{A\cdot B}{\|A\|\|B\|}\$ Or in other terms \$\frac{\sum_{i=1}^nA_iB_i}{\sqrt{\sum_{i=1}^nA_i^2}\sqrt{\sum_{i=1}^nB_i^2}}\$ Challenge Given two vectors (can be taken a...

 
I was about to write an answer in TI-BASIC and then I remembered that's still in the sandbox :facepalm:
 
protip: do not upgrade to chat improvements 4.0
it messes up the top bar and indents the starboard
 
Can't upgrade something you don't have :rollsafe:
 
that's chat without chat improvements
I'd show a screenshot of with
 
12:52 PM
since when does chat have a dark mode? How do I activate it?
 
but for some reason
it doesn't want to upload
@Bbrk24 universal dark mode extension :p
 
ah
 
that's with chat improvements
it timestamps messages, adds a helpful top bar, makes the user list look nicer, makes the "rooms you're in" list look nicer, and a bit more
that, combined with the SE chat modifications userscript and SE scroll the starboard userscript makes chat much nicer
 
1:10 PM
on mobile, all messages already have timestamps
@lyxal do you have a link to that last one?
 
@Bbrk24 not all of them
 
all of them do on mobile, at least for me. Not all of them do on desktop
 
Oh, I see what you mean. I don't mind those being missing, and it's still more than what desktop has by default
 
not that userscripts work on mobile anyway :p
it is worth having the nicer "users in room" list though
 
1:31 PM
I know you can use Promise.all and friends to combine parallel promises into one, but is there a built-in/primitive to combine serial promises? Right now I'm just using an IIFE, like
(async () => {
  await this.bot.guilds.fetch();
  await this.serverCountChanged();
})()
 
@cairdcoinheringaahing Tbf as other people have said exclusive ranges are definitely the default for programming, at least in every praclang I've enocuntered, and for IMO very good reasons.
@lyxal Problem is, people would just write build tools that search for and fix the memory issues. The compiler needs to obfuscate the code, add checksums and self-modifying code, and search the system for programs which aim to interfere with it and insert memory errors into them
 
Of course
All steps needed to prevent gaming the system
 
I just had an awful idea
What if operating systems were garbage collected, and instead of programs having their own garbage collection and getting big chunks of RAM from the OS, they'd ask the OS for individual variables
Even worse: What if there could be naming conflicts
"Sorry, you can't have length, that's already being used by some other program on the system"
 
That could make for interesting malware
A virus that closes every other program and uses up every possible variable name
 
And don't worry, this wouldn't allow for any data privacy concerns, as long as you remember to scope all of your variables correctly
 
1:43 PM
You could have like a variable registry like how you register urls
 
@Bbrk24 No, but you could just use a for loop if they're in an array
for (prom of proms) { await prom; }
 
Also, I was scrolling through old TNB transcripts, and look at how awful my prediction ability is:
Jun 3, 2022 at 3:20, by caird coinheringaahing
Mod team is still running smoothly, so no election soon
3 months later:
Sep 1, 2022 at 13:51, by caird coinheringaahing
👀 mod election?
 
Wait you only became a mod that recently?
Ohhh right
Feb 28, 2021 at 1:17, by caird coinheringaahing
idk the physics behind it, but time has been on some hardcore drugs for the past 18 months
 
@cairdcoinheringaahing sounds more like an alabi to me. Where'd you bury doorknob caird? :p
 
(also I think we could all probably communicate just as effectively purely through caird transcript quotes :p)
8
 
1:48 PM
May 11, 2021 at 1:23, by caird coinheringaahing
@lyxal frick, only 10/10 not 69/420? I failed as a crewmate, hella sus. Let me just go commit frick and die
Huh, guess you're right :p
 
The first time my computer shut down due to overheating, I thought it was never going to work again, so I said that my computer "went commit die"
 
Why is MathJax not working for me on the ask question page
Can y'all repro?
 
That's fair. It do be like that sometimes
@Bbrk24 ^
 
Mar 20 at 17:56, by lesobrod
Inline Latex not work!
 
@RydwolfPrograms Inline doesn't, block does
 
1:53 PM
Weird, neither works for me
 
Block takes a few seconds of thinking time though
Anyhow, it's bed time for this lyxal
o/
 
What about the other lyxals?
Will one take over in your absence?
 
Okay okay phew I'm here
7
 
@RydwolfPrograms I'll take that as a compliment :P
 
2:40 PM
CMQ: Should I start on my language ranking script again? One thing I've been planning to do with it is to use the language rankings to create "golfer rankings", sort of like ELO for how good you are at golfing. Like, if Jelly is expected to be 6× golfier than Python, and your Jelly answer is 8× shorter than the shortest Python answer to the challenge, your golfer ranking would go up
(of course, it'd be inaccurate when a user is so good at using a language and so prolific in doing so that they single-handedly raise the language's expected golfiness, harming their own score, but I don't think that'd bias it too much)
 
You can do it for fun but the sampling bias will always be way to great to draw any kind of conclusions from this
 
Sampling bias?
 
sampling bias as in which languages are used more frequently i think
 
No, that doesn't change the ranking in any way
 
No what languages people pick for certain problems
 
2:49 PM
I'm guessing they mean samplig bias as in people won't use languages as often for challenges they're bad at, amplifying the scores of languages that focus more on specific domains
ninja'd
Yeah, that's a factor, but it wasn't as big of one as you'd think
 
People don't use golfing languages for complex challenges so they seem much better than they actually are
 
That's not just the case for golfing languages though
Praclangs have the don't-use-it-if-it-doesn't-work-well effect too, so the bias partially cancels out
 
There's also code-golf vs fastest-code
 
I filter out everything that isn't vanilla code golf ofc
The filter I used for the old Node.js ranking script was:
questions.filter(q => q[1].includes("code-golf") && !["tips", "polyglot", "restricted-source", "source-layout", "cops-and-robbers", "code-challenge", "popularity-contest", "fastest-code", "fastest-algorithm", "answer-chaining", "programming-puzzle", "rosetta-stone"].some(n => q[1].includes(n)))
Plus a check for bonuses
 
@RydwolfPrograms Practlangs tend to be viable on a much wider range of use cases
that's their purpose
 
2:57 PM
Well aside from questions requiring interfacing with files/the internet/dates, that doesn't really matter
And there's another filter removing those questions
For every huge challenge with a 1000-byte Python answer there's a 100-byte Pyth/Japt/Jelly/05AB1E/Vyxal answer
 
You also have to account for the fact that some answers specify e.g. Python 2 vs Python 3, while others use one or the other and just say "Python"
or that e.g. Perl 6 and Raku are the same thing
 
Yep, the way I handle that is hierarchical languages
So the "Python" score includes every Python answer, including "Python 2" and "Python 3.9.1"
 
There's also the wonderful world of shells -- POSIX sh, macOS zsh, pure bash, macOS bash, bash + coreutils, dash, ...
 
@RydwolfPrograms That's not true at all
 
The process of running the script is pretty involved. First it generates hundreds of pairs of similar-looking language names, and has you dedupe them. Then it generates tons of likely parent-child pairs for the hierarchy, and has you start to build that tree structure. Then you manually go through the list fixing any weirdness.
@Bbrk24 Yep. I put all of those under an sh hierarchy, but keep them separate
 
3:02 PM
Many complex challenges have just one answer and it's usually python or even Java or JS
 
And questions with one answer have nothing to score against
So they don't matter
 
My point is any potential scoring mechanism can never be better than terrible
 
Well you're not doing a very good job of defending it :p
Flawed? Sure. Terrible? That is definitely not what I've experienced
 
Do you still have the whole script?
 
There is a meta question that explains it, let me find it
 
3:04 PM
@Bbrk24 Yeah, but it's kind of in a messy state right now since I stopped in the middle of some improvements
 
3:25 PM
@Bbrk24 Otherwise there were some rounding errors. For example, LOG(81)/LOG(3) should be 4, and displays as 4, but I think under the hood it's actually 3.99999999999 or something because when you take the INT you get 3.
 
ah, makes sense. My Swift answer might do that too then
 
In Rust, is there a way to "flatten" a Result where Ok and Err are the same type?
E.g., OsString::into_string returns Result<String, OsString>. If I map_err the OsString into a String, which gives me a Result<String, String>, is there a way to turn that into a String without a match?
 
I would probably just use match rather than map_err in the first place, but I also don't use Rust
 
Ooh wait it looks like unwrap_or_else does what I want
Yep, worked
 
@RydwolfPrograms yes
 
3:33 PM
Jan 7 at 3:02, by caird coinheringaahing
Don't think so
 
Dec 23, 2022 at 0:26, by caird coinheringaahing
@Niko Pretty sure this has been done
May 18, 2021 at 20:10, by caird coinheringaahing
It's definitely possible, but might be a little trickier
 
Jan 11, 2018 at 22:43, by caird coinheringaahing
@Mego Yeah, a surprising amount of people have that attitude on this site.
 
I should make a script to automate finding the closest cairdpost to a string
(I need to make sure to use good coding practices tho, otherwise this could happen:)
May 6, 2021 at 1:49, by caird coinheringaahing
Does that script actually work? Everytime I've tried it, it hasn't actually done anything
 
Nov 28, 2022 at 4:15, by caird coinheringaahing
That's fair. Mind if I ask why tho? The community's opinion is more valuable than ours :P
 
Oct 2, 2021 at 16:05, by caird coinheringaahing
Probably because I don't know an easier way :P
caird would be incredibly useful to have as an ally in this game due to their power to generate arbitrary transcript messages by themselves at will :p
 
3:40 PM
Apr 11, 2022 at 19:43, by caird coinheringaahing
True gamer right here
 
Jun 6, 2021 at 23:40, by caird coinheringaahing
That's like so much easier and open to exploits
Exploits including bribery :p
 
4:19 PM
me making an anti-golfing programming language 🙄
Please dont ban me
 
4:47 PM
@RydwolfPrograms Behold: a quote
3
 
@PlaceReporter99 that’s called a taript
@Bbrk24 git commit -m die
3
 
5:15 PM
@Sʨɠɠan someone botta go make a tarpit named "taript" lol
2
 
5:38 PM
@Sʨɠɠan I'm calling mine 'hippopotomonstrosesquipidelian'
 
00:00 - 18:0018:00 - 00:00

« first day (4448 days earlier)      last day (397 days later) »