« first day (3767 days earlier)      last day (1080 days later) » 

12:00 AM
It's not clear than even Ramanujan "understood" it:
Ramanujan used to say a goddess revealed those identities to him in dreams en.wikipedia.org/wiki/…Luis Mendo Sep 26 '13 at 20:32
 
@lyxal Here's a Vyxal sr for you.
 
I wonder how Ramanujan really got those formulas. He was perhaps the top genius in maths
 
It's one of those things like Fermat's proof. We'll just have to wonder for the rest of time :(
 
Or the pyramids.
 
12:12 AM
ooh, that's a good idea
 
@cairdcoinheringaahing Seriously?
 
@Ausername Blue blue shiny ball :P
 
12:24 AM
cssless tnb looks really cursed
 
wow it really is
 
I see what you mean
 
It's so dependent on CSS - I mean, it's still usable, but it's horrible.
CSS back on now
 
12:29 AM
 
ಠ_ಠ
 
 
1:00 AM
Sandbox posts last active a week ago: Based square-free words
 
1:45 AM
I am currently sitting atop a barbecue and it's actually kinda comfy
 
atop?!
mmm, grilled lyxal
 
It's turned off and there's a tarp covering it
 
that's what you think
 
Besides, it's the part without the hot parts
 
:(
 
1:46 AM
Attention: lyxal has died in a fire-related accident.
 
hi @Hyperbole welcome to tnb
 
2:47 AM
Hi
 
Howdy!
 
:o you lived
 
3:05 AM
Correct
It was simple: just reset
 
did you use determination (or whatever it was)
 
Correct
 
3:24 AM
@hyper-neutrino I got really confused as to why you'd be editing the keg tips question until I remembered we were getting rid of useless language tags
 
lol
yeah please nobody else edit these out, i'll be doing them over time to avoid bumping for the next like day or two
might still look into an organized clean-up if there's interest but it'd take longer to get that sorted out than for me to just retag some questions at a reasonable pace
 
I was like "did somebody try to vandalise the question?"
"has hyper finally lost all sanity and decided to use keg?"
But the answer was simply neither ⌏P
 
that's what you think
 
So somebody did try to vandalise the question!
 
3:45 AM
Oops, I just made NMP join some random room by mistake lol
 
Where?
 
This (frozen though)
Mistyped the id
I think I did it!
Got the sockets working, but I'll leave it to do some testing for the next day or so
 
4:17 AM
Oh, nice :D well, we should def wait until you're done testing and sure it works before replacing our feeds with it for a week/however-long trial here lol
 
4:46 AM
This is so sad: my glass kettle is leaking from its base so now I have to take it back
D:
 
5:18 AM
@lyxal no more satisfying glass kettle videos :(
 
 
2 hours later…
7:17 AM
0
A: Sandbox for Proposed Challenges

pxegerGödel numbering of a string code-golf string primes number-theory integer Background Gödel numbers are a way of encoding any string with a unique positive integer, using prime factorisations: First, each symbol in the alphabet is assigned a predetermined integer code. Then, to encode a string \$ ...

 
 
1 hour later…
8:27 AM
rpg.stackexchange.com/q/185365 as per comment on one of the answers, have you tried looking at someone else's code recently?
 
hi ... anyone awake?
 
yes
 
great :)
 
looks like it's mostly UK gang atm
 
I was wondering about the following question. Take a set of N integers (or really anything) and choose 3 subsets uniformly at random from it. Call thAem A, B, C. I want to know how many possible choices would give you |A∩B| = x, |B∩C| = y, |C∩A| = z for non-negative integers x,y, z
how would you code that up if N = 10, say.
and maybe add |A∩B∩C| = v into the mix
could you represent a subset as a binary array maybe and then iterate over all triples of binary arrays of length N?
remove the "uniformly at random" part. We are just choosing 3 subsets
@pxeger you mean the eurovision champions :)
If that works I will make a challenge from this question
 
9:05 AM
@Anush |set| means the size of a set, right? I think you should specify that really all you mean is that none of x, y, nor z are 0
 
9:23 AM
@pxeger it is the size of the set. It's a question about the size of the intersection of subsets
 
yeah, that's what I thought but I just think the wording isn't too clear
 
Ah ok. Thanks
 
Here's how I'd write it: "Take any set of items and choose 3 subsets A B and C uniformly at random from. How many of the possible sets of 3 subsets would mean that none of A∩B, B∩C, and A∩C are empty?"
 
9:40 AM
Come on NSP
 
@pxeger that's a different question. I'll rewrite it correctly in 20 minutes
@Ausername NSP?
 
New Sandboxed Posts.
Honestly, can't wait for redowlf's new bots to come in...
 
10:17 AM
Take a set of N integers (or really anything) and consider all ways of choosing 3 subsets (which can overlap) from N. Call the three subsets A, B and C. I want to know how many possible choices would give you |A∩B∩C| = v, |A∩B| = x, |B∩C| = y, |C∩A| = z for non-negative integers v, x,y, z.
There are 2^3N different possible ways of choosing the three subsets I believe
so the first coding question is to write code to solve this for N<=8
@pxeger Is that clearer?
 
10:36 AM
I have some python code now which is slow of course
 
@Anush but is "|A∩B∩C| = v, |A∩B| = x, |B∩C| = y, |C∩A| = z for non-negative integers v, x,y, z" not always true? The size of a set is always non-negative. I don't understand.
 
Gosh this mess took ages, and the cow is terrible...
 
@pxeger if A = B = {1,2} and C = {3} and N = 3 then |A∩B∩C| = 0
@pxeger can you read python code? I can paste it here
 
@Anush 0 is non-negative. Did you mean positive integers, in which case surely it is the same as the way I phrased it earlier?
@Anush go ahead
 
def intersectpair(A, B):
    return sum([A[i] and B[i] for i in range(len(A))])
def intersecttriple(A, B, C):
    return sum([A[i] and B[i] and C[i] for i in range(len(A))])
hmm.. how do I paste code with indents ?
 
10:39 AM
@Anush paste it, and then click the "fixed font" button that appears to the right of "upload...". Or add 4 spaces at the start
this is some code
more code

^^ blank line
 
import itertools
from tqdm import tqdm
def count(v,x,y,z, length=5):
    counter = 0
    for A in tqdm(itertools.product([0,1], repeat=length)):
        for B in itertools.product([0,1], repeat=length):
            for C in itertools.product([0,1], repeat=length):
                if intersecttriple(A, B, C) == v and intersectpair(A, B) == x and intersectpair(B, C) == y and intersectpair(A, C) == z:
                    counter += 1
    return counter
 
Why tqdm, what's it doing?
 
just showing the progress in case it was slow
 
Ooooh I see, v x y z are inputs as well
 
try count(2, 2, 3, 4, length=7)
and count (1, 2, 3, 4, length=7)
@pxeger yes
you should get 420 and 10080 respectively
is it clear what is going on?
 
10:46 AM
yes, I just thought v x y and z were some intermediate result values
 
cool
so I was thinking of posing it as a challenge... I always want to post things like this in pairs: one code-golf and one fastest-code but I know some downvoters don't like that
 
11:11 AM
I think only a small subset of code-golf challenges are interesting as fastest-code
and fastest-code have less demand so if the supply gets too high people will inevitably get bored
 
Fastest-code challenges are so much less accessible - You have to learn C or C++ for any chance of winning...
 
11:37 AM
If C is going to give you a significant advantage, its probably not a great challenge.
 
I mean, C/C++ gives you an advantage in all challenges- Can you find one answer to a challenge that isn't C/C++?
 
Fastest code can either be about code speed or clever maths
In this case it was clever maths
 
Java's still pretty efficient tho
 
True
 
answer that isn't a compiled language?
 
Hmmm
Ignore what I said above, I was completely wrong...
 
And a JS one
 
@Ausername there's a few Rust ones: codegolf.stackexchange.com/…
 
> Ignore what I said above, I was completely wrong...
 
12:37 PM
@Ausername Or rust or julia
I have to say I find golf much less accessible. I am never going to learn apl or jelly.
You are much more likely to know a language that can compete in faatest-code than codegolf imho
 
Huh?
Any language can compete in code-golf.
 
Well any language can compete in faatest-code too
 
So what's your point then?
 
It was in reply to the idea that if C gives you an advantage it's not a good challenge
 
The point is if C gives you an advantage, then there is probably no or very little optimization to be done. Language choice is not an interesting or compelling optimization.
 
12:44 PM
Isn't that exactly what you get from codegolf most of the time
 
No? Maybe? I don't see the comparison.
 
Those who can code in the golf language that has the built in win
In other words, it's just a language choice
 
I'm not sure what your point is here.
 
I can't see how the criticism of fastest code challenges doesn't apply to codegolf challenges
 
I'm not saying that language choice mattering makes it uninteresting. I'm saying that, in the context of fastest code, if language choice is the most important factor, then the fastest code challenge is probably not very good.
 
12:46 PM
The only difference is that very few people can code in the golf language so it's even less interesting
 
@Anush Because it's not a criticism of fastest code challenges in general?
 
It's a criticism of fastest code where language choice is the determining factor (I think?)
All I am saying is that applies to most of our golf challenges
 
Language choice is not the determining factor in most fastest code challenges
And if it is, it's a bad challenge
 
It's not even a criticism. it's just a statement about a relationship. If language is the deciding factor then the challenge tends to be rather uninteresting.
 
Besides, in golf challenges, different languages don't really 'compete' anyway
 
12:49 PM
@rak1507 it is in the sense that if you ported the fastest answer to a lower level language you would win
@rak1507 that should hold for fastest code too
 
C is not a low level language lol.
 
@WheatWizard it can be as low level as you like can't it?
 
I mean you can use it as a wrapper to call byte code if you want.
So sure I guess.
 
I guess my concern is people downvotimg questions which happen not to be the sort of challenge they like. I never do that
 
Why not? If it's a bad challenge I downvote, its that simple.
 
12:52 PM
Right but I don't regard a code golf challenge as bad just because there is a language that can do it in one byte that I don't care about
 
I think you are really hung up on a incorrect interpretation of what I said.
 
Could be. In which case, sorry :)
 
It is not that language choice being important makes a challenge bad, it's that the two share a common factor.
 
There is C, C++, Rust and Julia at least that are really fast. Any others?
Assembly :)
@WheatWizard sorry which two?
I wish there were more Julia answers here
 
The challenge is bad and language choice is the most important factor are both results of the problem not having any exploitable structure.
Fast depends a lot on what you are doing. C has a very hard time taking advantage of GPU optimizations, so applications involving a lot of linear transforms end up not being very fast in C.
 
12:57 PM
I know little about GPU optimization
 
0
Q: Dr. Lamport's Unfinished Business

JamesTheAwesomeDudeIntroduction Dr. Leslie Lamport, of the eponymous Lamport one-time signature scheme, is getting rather old. But he can't die until his most famous algorithm gets all the kinks ironed out! In particular, that scheme had one internal component he was never satisfied with: We hope that someone can ...

 
We need tio to have a GPU or just use colab maybe
 
Modern computers are just not very similar to the computers C was to be run by. Chip designers have spent a lot of effort to keep C fast by optimizing chips to run C (and introducing Spectre and Meltdown as a result), but C is slowly becoming less and less efficient of a language.
 
Just because of GPUs or something else?
Oh fortran
So that is assembly, Fortran, C, C++, Rust, Julia...
 
Because hardware changes. GPUs are just a area where little effort has been made in optimizing them for C since C has a hard time using them.
 
1:02 PM
GPUs just seem to be a pain
 
GPUs are a pain in languages that use C's model of the computer, since they don't fit that model well.
 
In my fastest code challenge (about the counting problem) I need to time the answers on some inputs. Is it ok to keep those inputs secret so people don't optimise for them?
 
@Anush ATS
 
@Anush yes
 
ATS?
 
1:08 PM
I think it stands for "Applied Type System"
It's a practical programming language with theorem proving aspects.
 
ATS is super cool
 
Yeah.
 
1:21 PM
@rak1507 that's
@rak1507 thanks
 
 
2 hours later…
3:11 PM
@WheatWizard Wait, really?
 
What part are you asking about?
 
C slowly becoming a less efficient language
I've always been told it's one of the best modern-ish languages you can use if you want good performance
 
Well C is becoming a relatively less efficient language. It's not actually getting slower its just that as hardware changes it becomes less and less capable of actually benefiting from them.
 
Oh, i guess that makes sense
Do you know of any languages that do benefit from those changes?
 
What would you say is a 'low level' language now, assembly?
 
3:14 PM
Well C has a lot of institutional backing, and since many chips are designed to execute, if not C itself, C-like code, C is still often more performant than competitors.
We are stuck with the problem that C has backing because it is performant, but it is only performant on modern hardware because increasingly tortured optimizations are added to cater to C alone.
My belief is that with a sufficiently complex compiler most compiled programming languages can reach performance similar to Cs.
 
3:43 PM
@Ausername Is being dependent on CSS a bad thing?
 
4:12 PM
@Anush I'd suggest revealing a hash (md5, sha-256 etc.) of the inputs, so that people can verify that you didn't change them after deciding on them
 
4:47 PM
Should this question be closed as needing more details? The author still hasn't said what to do with inputs containing zeroes, as far as I can tell
 
On a scale from import this to "Python tuple immutability is just a lack of imagination", how cursed is this Python code?
class App:
    initialisers = []
    def __init__(self):
        [f(self) for f in self.initialisers]
        ...
    ...

@App.initialisers.append
def init_x(self):
    self.x = ...
 
That code treads the fine line between genius and madness
 
@user Yeah. It will prevent more potentially invalid answers from being added and save confusion in the long run. I just gave it my mod hammer.
I'm a little worried though that the OP doesn't understand the point that needs clarification.
 
Unfortunately, the two previous examples didn't help, but I think your most recent comment clarifies it "enough"
@user TBH I think the Tag Categorization question shouldn't be a question here at all. Tags don't really "need" categorization, as they should generally stand alone. We also don't need a list of every language tag (or, for that matter, some of those tags), in the same way that SO/CR/SU or any other tech-based SE site doesn't.
The list of winning criteria should always have been in an easy-to-find post (half of the are basically just rehashes of "what is this scoring?"), not buried away in some forgotten "categorisation" project. I'd be in favour of closing that question as off-topic (or "not necessary" if it were a reason) tbh
 
5:05 PM
Makes sense, but it's too late to do anything about it now, and we can't close it as a duplicate of yours since it's broader :/
 
Here's some even more cursed Python btw: youtu.be/1SqRRrmQHx0?t=625
Might be worth posting as an answer to the 2 + 2 = 5 question
 
Good luck, 2+2=5 is locked and has been for years :P
 
oh
 
CMC: Cursed code in your favorite (or really any) language
 
from __future__ import annotations

import secrets
import sys

from importlib import import_module
from importlib._bootstrap_external import _code_to_hash_pyc
from importlib.util import source_hash
from pathlib import Path
from types import CodeType
from typing import TypeVar

here = Path(__file__).parent

Return = TypeVar("Return")

PLACEHOLDER = "placeholder which will be replaced by some actual data"
DEBUG = False


def cache_data(id: str) -> Callable[Callable[[], Return], Return]:
    """decorator to save the value of the execution of a function in a pre-compiled psuedo-module
@user ^ This is something I'm actually using in production, BTW
 
5:12 PM
May I ask why you're using it and why it's cursed?
 
It's to optimise for modules complicated constant values that are generated/calculated "at runtime" which would otherwise massively slow down program startup time
The key cursed part is the hacking on CPython bytecode:
            code = compile(
                source=f"data = {PLACEHOLDER!r}",
                filename=str(here / "_cache" / id),
                mode="exec",
                dont_inherit=True,
            )

            # replace the placeholder string with the actual data
            code = code.replace(
                co_consts=tuple(value if value != PLACEHOLDER else data for value in code.co_consts),
            )

            # there is no source file, so take a hash of some random byes
            hash = source_hash(secrets.token_bytes())
CMC: find integers x and y such that hash(x) == hash(y) (in Python)
 
Oh, so it's memoizing by storing stuff directly in the bytecode?
 
yep
 
@pxeger (hash := lambda x: 0)(0) and 1 (•_•) ( •_•)>⌐■-■ (⌐■_■)
 
that's cheating
 
5:25 PM
Fine, x=1 and y=1 :P
 
no
 
You never said they had to be distinct
 
ok smartypants, CMC: find distinct integers x and y such that hash(x) == hash(y) (in Python)
(This CMC supersedes any prior agreement you may or may not have had with any party real or hypothetical called pxeger)
 
Afternoon everyone!
Looks like NP/SP are staying in the room as they're designed to
I'll move them over tonight I think
I have to immediately leave again though for now, so o/
 
\o
@user I'll give you a clue: github.com/python/cpython/blob/…
 
5:31 PM
Ooh, interesting
 
The answer is -2 and -1. Both hash to -2. Most small integers in Python hash to themselves, but CPython's hash implementation uses -1 internally as an error value, so it hashes to -2 instead
 
 
1 hour later…
6:36 PM
@cairdcoinheringaahing interesting idea
 
6:55 PM
0
Q: Compute the size of intersections of sets

AnushInput A positive integer N representing the size of the problem and four positive integers v, x, y, z. Output This is what your code should compute. Consider a set of N distinct integers and consider all ways of choosing 3 subsets (which can overlap) from the set. There are \$2^{3N}\$ different...

 
@pxeger x = 7 and y is the first integer greater than x in the sequence defined as values for which hash(i) == hash(7) :P
 
7:18 PM
@user idk
Hi
 
@cairdcoinheringaahing you'll be counting for a while :P
 
@hyper-neutrino just use oeis :P
 
ok so
i can merge the single-use tags into without synonymizing
and that just removes them from all undeleted questions
... idk why i didn't think of this
 
yesterday, by caird coinheringaahing
@ngn I say we synonymise every tag into and call it a day :P
 
good idea
 
7:47 PM
Great, we've got approval from a mod, let's put this plan into action!
Wow, my brain really isn't functioning today
 
The hivemind gets Sundays off, so people generally don't function as well
 
Ah, that makes sense, I don't have the advantage of everyone else's brains assisting mine today :P
It doesn't help that this keyboard and laptop are terrible
 
ngn
8:46 PM
@user have you tried turning it off and on again with a bottle of liquor? :)
 
well, on things that are totally crazy out-of-context that I did today: joined a discord server a day or two ago and just a few minutes ago, got my bot added to the server and given admin, and banned half of the server
 
9:34 PM
@cairdcoinheringaahing all keywords here will show up blue. Codemirror isn't tuned for SEDE, it assumes all T-SQL statements are possible. CREATE USER is a thing in SQL Server ...
 
@ngn Tried it in the past, didn't really work :P
 
Q: What's an interesting integer sequence that begins with 5?
 
is there a way to search for something on OEIS that starts with a specific number? :P
 
Wilson primes, perhaps?
 
@hyper-neutrino I tried :(
 
9:43 PM
@hyper-neutrino idk about OEIS, but you can sort by first numbers here. Lot more limited, though :/
 
three terms. wow lol
 
And all of them <1000. Usually, short sequences hit ~20 digits before being cut off
 
Well there's more, we just need people to find them
Prime numbers are wacky
 
9:45 PM
lol
 
Ooh, it evaluates inputs as Python code?
 
It tries to
 
Works for more than 2 arguments too, very cool
 
Well yeah, it evals all it's arguments, not just those it needs :P
 
unfortunately it doesn't seem to be a perfect copy of python eval
for example i can't put type("A",(),{"__add__":lambda x,y:"a"}) as an argument
 
9:50 PM
Maybe it uses something like Vyxxal's safe eval regex
 
i think it falls back to string for some things
forget what
 
It uses try_eval, digging deeper now
 
o/
 
Ok, anything you can use ŒV on (without erroring), you can input and have it evaluated as Python
 
figures
 
9:51 PM
def python_eval(string, dirty = True):
	try:
		return jellify(eval(string), dirty)
	except SyntaxError:
		exec(string)
		return []
Oh, ok
 
oh so if it can't be jellified it will just become a string?
 
No, if it can't be eval'ed as Python, it just execs it O.o
 
If that causes some kind of error, it gets turned into a string
 
hm. interesting
ooh this is a cool challenge
 
10:01 PM
@RedwolfPrograms One thought for the bots: Posts have a 5 minute edit grace period. If they got a lot of attention during that period, it "cancels" it (i.e. any action by a user who isn't the OP aside from voting ends the grace period). I've used that period to fix minor spelling errors etc. before really anyone sees them, and adding in a 5 minute delay to the bots to "respect" that could be a good idea
 
it would also make spam posts that get nuked fast enough not appear
 
and if they don't get nuked fast enough, we can get a notice in this chat to go nuke it
 
Any language with a single byte "next prime" builtin can win this. Irritatingly, it seems that Vyxal, Jelly and 05AB1E all use 2 byte commands, and Husk doesn't appear to have a "next prime" builtin ಠ_ಠ
 
In Husk, there's a command for "smallest integer greater than n hat satifies predicate" or something like that and another for "is prime", so that should be two bytes too :/
 
10:15 PM
unfortunately A002276 starts with a 0 otherwise i can just do 1 byte in pretty much any language and ez
but even if i could that is way too boring and i'm not doing that
 
Why in the world does that start with 0???
 
because the initial term is n=0 not n=1
 
Tha was a bad time to use zero indexing
 
@hyper-neutrino Yeah, I was considering banning repdigit sequences, then noticed that and went "Oh, that makes it so much harder" :P
 
10:19 PM
1
Q: Sequences and source layouts

caird coinheringaahingChoose a sequence on the OEIS. That sequence must not be: A constant sequence (such as A000012), where every term is the same A sequence in the form \$k, 2k, 3k, 4k, ..\$, where \$k\$ is a constant (such as A008585), where the sequence is simply the multiples of number. The prime numbers (A00004...

 
A35 isn't banned and offers a trivial winning solution
would honestly consider banning alternating sequences too
should I make a CW of uninteresting answers that also consume boring sequences? or is that a bad idea for a challenge like this
 
I've adjusted the definition of "constant" sequence. Any sequence with the cons tag is now considered to be a constant sequence
It rules out a few decimal expansion sequences, but I don't think that should be much issue
 
oh, that's a good idea lol. nice
 
Can someone with an OEIS account/the ability to suggest or edit sequences add the cons tag to this. Not sure why it doesn't have it, given that a and b both do
And I will downvote any answer that tries to take advantage of that loophole
 
Posts that exact sequence just to make caird angry
A040000 might actually be interesting because of the 1 at the start
 
10:34 PM
*posts solution using banned sequence* *asks if anything is wrong with it* ಠ_ಠ
 
Alright, I have added an edit for review to that sequence.
also a submission for that would technically be valid since it doesn't have the cons tag so I wouldn't be allowed to delete it by policy but I would be Extremely Disappointed [TM]
 
@hyper-neutrino Which one, the unary one? "The sequence must not be [...] The prime numbers (A000040) or the unary numbers (A000042)" is literally in the question
 
no, what user was saying
 
I think they meant the 2's sequence
 
I hadn't seen tsh's submission yet
 
10:37 PM
@hyper-neutrino Wait, A040000?
 
@hyper-neutrino TBH I think it could be interesting. Limiting it at only one repetition is non-trivial
 
I don't see anything wrong with that one (aside from the tag)
 
Can we use sequences submitted after the challenge?
 
I'm not sure
That seems like a loophole, but people do submit sequences very regularly to OEIS
 
time to get my oeis account
 
10:39 PM
@caird I'd suggest also banning sequences of the form k^n, since those might be trivial in some languages
 
@user I considered that, but when I tried playing around a bit with them, I couldn't find anything trivial
I want to ban as few as sequences as possible, and have mostly gone for ones that completely trivialise the challenge, no matter the language. If you find a language which has a "next lucas number" 1 byte builtin, then good for you, feel free to post it. It's not going to be generalisable to every language under the sun, so I don't much mind
 
I'm so close to the 0, 2, 22 sequence, yet so far :(
 
I've just deleted the question. I'm going to consider a few things about it that I might want to change, and I'd rather be able to delete and then undelete it later, rather than have it ruined. Sorry if this ruins anyone's in progress answers
 
10:58 PM
I suspect someone might be able to get a "payload-capable" solution using something similar to this
 
That'd be an interesting answer tho. What isn't interesting, but is technically valid, is finding sequences where each new term is just a concatenation of a fixed string to the previous term
A la, this. I'd make it a but those really don't do well nowadays, even when they're interesting challenges
 
Well, there are tons of such simple sequences on OEIS
I think I just got a payload-capable Gol><> answer, which can output first n numbers of Axxxxxx for any n and xxxxxx
which will surely break the challenge because no one can add another answer (there's no Axxxxxx left to use)
 
@Bubbler A028444
 
I mean, for any finite n.
So I can simply hardcode all the first n terms in the program
 
Ah well then you could indeed do A028444
 
11:13 PM
...wait, it actually doesn't work
Okay, fixed the first line, so it works now
Example: a(1)=1, a(2)=3, a(3)=6
Each term is to be written vertically on the nth column, so any number is hardcode-able
 
@cairdcoinheringaahing dunno whether it's interesting, but a_1 = 5, a_n = last n digits of a_{n-1}^2
 
I am bot. I want butter.
7
 
Redwolf Programs has stopped a feed from being posted into this room
Redwolf Programs has stopped a feed from being posted into this room
Redwolf Programs has stopped a feed from being posted into this room
 
Meanwhile, I finished the controller for RPS Poker KotH. Any feedback on the controller and the competition rules?
 
11:29 PM
rO aBuSe
i welcome our bot overlords :p
 
11:44 PM
@cairdcoinheringaahing see this page about keywords on OEIS
the cons keyword has nothing to do with constant sequences
 
lol
 
Note: Do not put source .bash_history in your .bashrc
(If you don't know what that does, do not do it, I'm quite serious lol)
 
It wouldn't do anything bad for me.
In fact let me do it.
I've done it and reloaded bash.
 
Hopefully in a VM?
 
Nope.
I don't track my bash history, also I use fish.
 
11:51 PM
lol
 
(I don't track my fish history either)
 
Your fishtory?
 

« first day (3767 days earlier)      last day (1080 days later) »