« first day (2892 days earlier)      last day (393 days later) » 

1:07 AM
@AdmBorkBork yeah, I think this has been the worst Winter Bash of all the few that I paid any interested in.
A plea of help to all mathy people here
A: Sandbox for Proposed Challenges

QuintecSurreal Numbers Surreal numbers are one way of describing numbers using sets. In this challenge you will determine the value of a surreal number. Intro A surreal number consists of two sets: a left and right. The value of the surreal number must be greater than all numbers in the left set and ...

(And non-mathy people, actually)
3 hours later…
4:07 AM
@flawr Except the parsing html with regex questions
4:46 AM
CMC: Write a simple Haskell type (using only as, bs, cs etc. and ->) that has an infinite number of pure instantiations.
for example a -> a has 1 instantiation and a -> a -> a has 2.
Integer -> Integer or even a -> [a] have an infinite number of representations, but those are cheating since they use concrete types.
@WîtWisarhd I'm a Haskell dummy - what's an instantiation?
@Mego An object with that type.
realistically a function with that type.
Since there is no pure function of type a.
Perhaps I ought to give an example, for the type a -> a -> a the two instantiations are f a b = a and f a b = b.
Any pure function with that type will be equal to one of them.
So Integer -> Integer has an infinite number of representations because Integer is a bignum and thus can represent a theoretically-infinite set?
And a -> [a] has an infinite number of representations because you can always nest an object deeper in a list
5:03 AM
Wouldn't (a -> a) -> a suffice, since that's the type definition for the Y combinator?
Can you write an instance of (a -> a) -> a? I can't come up with any.
The standard definition of the Y combinator is not a pure function I believe, it would be fix f = f (fix f), which I think is the same as fix = undefined.
Yeah, just give it a function that has no fixed point and it will never halt.
I don't think that type has any instances.
5:19 AM
@WîtWisarhd How is it not pure?
It has 0 side effects, and performs consistently when evaluated with the same arguments repeatedly - it either returns the fixed point if one exists, or runs infinitely if no fixed point exists.
5:53 AM
Apparently Mathematica has a builtin called
2 hours later…
7:36 AM
@WîtWisarhd How about this: f :: (a -> a) -> a -> a
I think it can support f x = x, f x = x.x, f x = x.x.x etc.
3 hours later…
10:32 AM
@lirtosiast if you make all uppercase character lower case (except the first) one might be tempted to think it is german
1 hour later…
11:44 AM
@flawr In Danish it would be stationærvektorlineærmodelsbegyndelsesværdisdistribution.
11:55 AM
@Adám bucketlist.push(learn danish)
@Mego wouldn't you need to somehow choose a "natural" initial value? for that to exist you'd probably need some type restrictions
12:29 PM
@Mego It's not pure because it doesn't always produce output. For example (+1) has no fixed point and thus loops forever.
CMC: Given an integer list L and an integer N, reduce all runs of N in L to a single N. E.g. L=[3,3,1,3,2,2,3,3,3,0], N=3[3,1,3,2,2,3,0]
12:44 PM
@ngn I can do 11 in APL
@Adám Pyth, 16 bytes: r9m,^hdnedQedr8E
I can do 12 in APL
13 in k: {x@&~0*':y=x}
12:54 PM
I think there's a very short solution in Jelly out there but ¯\_(ツ)_/¯
1:31 PM
@Adám Jelly, 8 bytes: Œgḟ€¹ȯ€F
7 bytes: ŒgḢf¡€F
2 hours later…
3:46 PM
will we need this again?
Give us the benefit of the doubt, then get a mod to post that when trusting us inevitably fails
This WB was absolutely disastrous for PPCG imo
4:02 PM
I haven't checked the Main challenges within the last few days and went o_O when I saw the timestamp of the most recent challenge :P
oh wow lol
@FrownyFrog Almost the same as what I had. Just remove
4:32 PM
@Mr.Xcoder Disastrous how? It seems to me like we just didn't really do anything
@cairdcoinheringaahing Exactly, we just didn't do anything. The site was particularly inactive when I would have thought it would be even more active due to WB. Not to mention that we're really low on hats on the site-wide leaderboard.
@Mr.Xcoder Ah. I reckon its just because we lost quite a few members over this year
Including 2 mods, sadly.
IIRC last year, me, you and user202729 were the top hat getters, and we've both scaled back our activity
Q: Find equally-weighted complete graphs

ArcturusGraph theory is used to study the relations between objects. A graph is composed of vertices and edges in a diagram such as this: A-----B | / \ | / \ | / E | / / |/ / C-----D In the above diagram, A is linked to B and C; B is linked to A, C, and E; C is linked to A, B, an...

Can't you file a copyright infringement, or something like that? Seems as though that could significantly impact APL
@cairdcoinheringaahing Nobody has copyright to the name APL. It is an ISO standard. IBM owns APL2, MicroAPL owns APLX, APLNow owns APL+…
I don't think copyright comes into play here because Amazon isn't even using any ideas from APL. Trademark comes closest, I think, since APL is sorta like a "brand", but even then, it hasn't been marketed as a brand or anything like that, so there's no ground to stand on there.
But it will surely cause confusion in both camps. There has already been two SO Alexa questions wrongly tagged .
Imagine if Amazon had called it C…
Yeah. While I don't think there's any legal ground to force Amazon to change their naming, maybe a letter or something co-written by the major developers of APL could make a case.
5:16 PM
@Adám I can't :(
@FrownyFrog Hint: Note that N (your ) is a scalar.
got it!
thank you
A: Sandbox for Proposed Challenges

Khuldraeseth na'BaryaAverage Strictly Increasing Sequence code-golf number random Imagine rolling a six-sided die. Roll, record, and repeat, until one roll is no greater than the previous. You'll end up with a strictly increasing sequence of at least one integer, but no more than six, having rolled seven times at m...

I tried that already but separating with a space hasn't crossed my mind lol
"didn't cross"?
6:17 PM
I think Geometry Dash could qualify as a programming language here
6:42 PM
Guys, what would you suggest a label be in my programming languae
@MilkyWay90 You can edit your messages for a few seconds.
@MilkyWay90 Label as in goto label? You mean what the syntax should look like? Not label:?
So far I have "Crossed out __ is __ :(" for pushing a number to the stack, "Explanation of an unreadable Pyth answer:" for a comment, and "Post a question on the Sandbox" to print the top of the stack in it's Unicode value
It is supposed to be based off the site
And yes a goto label
@MilkyWay90 How about [tag:nnnn]?
Sure, but that sounds too practical programming language-based
I'd say something like "Make another BF derivative called __"
for labels
So like commands like those
that relate to PPCG or esotoric programming language
@MilkyWay90 So it is supposed to read like Chef, but for "doing stuff" on PPCG?
6:47 PM
Sort of, yes
And a goto could be "Make a translator from BF to __ in BF"
Or "Make a hyperdimensional language out of __"
or "Fix the bugs with __"
Yeah, I think I'll go with the bigfixes one, but worded differently
Okay, so the final goto command name will be "Upload a bugfix to __"
Have there been mandelbulb challenges?
@Anush What are those?
To make my programming language Turing-complete, I'll need to decrement, increment, and be able to access any part of the stack
Then I'll add in more features for an update
Oh but I need input too
@MilkyWay90 "Ask in TNB"
@Adám This is TNB, right?
@MilkyWay90 Yes.
6:58 PM
@Adám Then why did you tell me to ask in TNB if I'm already here?
@MilkyWay90 I didn't. I suggested that the command to prompt for input should be "Ask in TNB". Note the quotes!
@Adám Oh, okay. Thanks for the suggestion!
"Add one byte in a code-bowling submission" An incrementor
"Save one byte in a code-golf submission" A decrementor
"Post a submission on the __th prime number question derivative" Access nth stack element
Any suggestions on this?
"Ask a question on The Nineteenth Byte" Take input
@Adám is "Ask a question on The Nineteenth Byte" good enough for taking in input?
@MilkyWay90 I think so, but I'd personally say in not on. Do you need to specify which kind of input?
@Adám 150 :P
@Adám Okay! And no, it turns it into a string
Or I could make it an int if it is a number, and string if not.
Or it's Unicode value
I'll decide when I start implementing it.
7:06 PM
CMC: Given a matrix A and a matrix B, return the number of times A occurs as a submatrix in B. Occurrences may overlap. E.g. A=[[1,4],[3,1]], B=[[3,1,4,5],[6,3,1,4],[5,6,3,1]] → 2. A=[[3,1,4,5],[6,3,1,4],[5,6,3,1]], B=[[1,4],[3,1]] → 0
For the implementation, should I use regex to parse the lines?
Or will I get another problem?
@Adám I don't know a lot about matrixes (I'm to the end of Geometry), so I can't help you with that.
uh... that's a CMC...
i.e. "Chat Mini Challenge"
@MilkyWay90 Also, you don't need to know anything about matrices to solve this
@Adám oh yeah
So... is using regexs okay for making programming languages to parse the lines?
@MilkyWay90 If the phrases are rigid, and cannot be nested, then I guess, yes. But just looking at the first few characters to identify operation would be enough, no?
@Anush I don't think there is already a question for that. Be sure to post it on the sandbox, though.
@Adám Okay
seems like a nice thing to draw for ppcg
Also, I created a chat room for PPCG on my website
In case anybody wants to check it out
www.randomstuffonline.com/chat.html with the chatroom being called "ppcg", all lowercase
7:23 PM
Has anyone else been working on this? I remember a few people were interested when I first posed the question
Nov 25 at 0:59, by Mego
Chat Mega Challenge: There are roughly 43 quintillion valid states for a 3x3x3 Rubik's Cube. What is the largest subset of valid, unsolved states that can be solved without passing through any intermediate state that is also part of that subset, using only the basic twelve moves (F, B, U, D, L, R, F', B', U', D', L', R')?
@Mego Wasn't on this chat on that time
A trivial lower bound that was pointed out quickly is 12 - the set of states that are 1 move away from solved.
132 should also work - all unsolved states that are 2 moves away from being solved (12**2, less 12 for conjugate pairs).
What are the apostrophes
7:31 PM
Moving the face in the opposite direction of the turn without the apostrophe
There are 6 faces (Front, Back, Up, Down, Left, Right), and two ways to rotate each face 90 degrees (clockwise and counterclockwise)
7:48 PM
@JonathanAllan so, I can't really understand the challenge here, but it looks like you can, can you please tell me a bit more? (note: I've frantically tried not to see your answer so far... :P) also, the input you've provided to your answer doesn't look very symmetric to the diagonal to me
@EriktheOutgolfer ah, my input is not symmetric - it works regardless of input symmetry.
what don't you understand?
the challenge in general
maybe I can clarify it instead of VTCing it
also... how can you test with an asymmetric input when the challenge is only defined on symmetric ones? :P
@EriktheOutgolfer I see it as "How many ways can you pick a subset of the nodes, of size n, such that the subgraph between those nodes [is symmetric and] has all non-diagonal weights equal?"
yeah, because the diagonal ones are always 0
so I just pick the symmetric submatrices?
they can be anything (from comments)
7:56 PM
yeah I was wondering too
not just symmetric - all non-diagonals must be equal
so this challenge can do without the graphs... :P
but... mustn't the picked matrices be along the diagonal too?
like, otherwise there are nodes with only a single connection
and not any sub-matrix, only those representing sub-graphs - i.e. the top-right 2-by-2 of a 4-by-4 (or larger) is a sub-matrix but not one which represents a subgraph
yeah that's what I thought above ^^^ too
yeah, hence the graph talk :D
7:59 PM
well... I suspect all of them represent subgraphs, but some don't represent complete subgraphs :P
so... yeah, actually, I don't have to check if they're symmetric ;P
so... how did you pick an asymmetric input to test your answer?
like, that can't represent a graph at all
...and I'm not confident I have the golfiest way
yeah, you might've figured out I'm trying to disprove that... >:D
just being dumb about choosing an input
8:03 PM
> I've frantically tried not to see your answer so far... :P
@JonathanAllan wait... isn't your solution supposed to output 1?
no, why?
because I can only see one valid submatrix of size 3 over this:
(ignore the asymmetry for now)
like, even with wrapping around
call them ABCDE - pick either ACE or BCE
ah, so I have to count those as well... :P heh
yeah, forget about
sub-matrices: [[9,2,2],[2,6,2],[2,2,0]] and [[8,7,7],[7,6,7],[7,7,4]]
8:13 PM
yeah I noticed
makes sense now then?
yep, it's just that I mistakenly tried to use and it managed to convince me otherwise :P
@JonathanAllan alright, the initial draft is at 23 bytes, let me try to knock you out (as a sign of gratitude, of course ;) )
21 20
happy golfing
8:26 PM
@JonathanAllan 15 ;-)
let me see if I can suggest it or post it
meh, post it
oh wait
that was testing for symmetry
@JonathanAllan 17
and a small fix to my real coding style after ensuring you won't ninja the 17 from me :P
16, you say? hm...
and now I wish I didn't need an to stick out there in the middle...
FYI I replaced my input with a symmetric one
8:38 PM
can you post here? I don't want to accidentally look at your code
same sets of 3
because I was putting my hand strategically on my eyes to filter out lines I can see from my screen before
so... if a Dennis suddenly comes across... would it be surprising?
no :p
hm... that gives me confidence... ;P
9:00 PM
@JonathanAllan hah 16 I think :P
dunno if the same as yours
how didn't I think of merging the two s... >_>
unfortunately, it now turned into two lines... :/
both check all; you dequeue a diag'd like I used too
what? you mean it's the same approach?
or you meant "I used to" instead of "I used too"?
oh so you changed the approach
I'm currently having trouble with fitting the stuff into one line... not cool, Mr.
only by filtering out equal [x,y] indices and being flat already
i.e. removing diag indices rather than values
9:10 PM
ah so you used multidimensional indexing?
I still do
maybe I can save with 2 lines too
oh it's in one line?
if you have two s, try merging them into one
because I don't think mine can go into one line without trouble
yeah, looks like an approach with œị is still 16 bytes for me
two lines
but with my bottom
I golfed 1 byte with full program
9:17 PM
lol completely unrelated to my approach
@JonathanAllan two lines don't help?
no & no
it's the same overall approach - get all submatrix indices, then check equality of non-diag elements add 1s and 0s of results
yeah but you used "full program"? that's interesting...
yeah allowed me to do a 4-chain monad for each
4-chain? like, 3 or 4 chain separators? lol
4 link chain so I can use Ʋ€
9:25 PM
ah, I also had Ʋ€ before... you sure you need a full program for that?
not /sure/
phew, and I thought you made four entire chains...
CMC: Given a year as input, output how many leap years have occurred/will occur in the years from 1 to N AD (inclusive). Note that the Julian calendar was used until 1582 when it was replaced with The Gregorian calendar. For this challenge, assume that we will use the Gregorian calendar until the end of time
@JonathanAllan ah, saw your answer
looks like you do need a full program for that, yeah
@DJMcMayhem Each country switched from Julian to Gregorian in its own year. The first was Spain in 1582.
9:40 PM
@JonathanAllan btw, you could've used µ...) as well
@Adám TIL
I was just going off of Wikipedia
@JonathanAllan (double conversation ping) unfortunately, it looks like that won't suit my approach... :P
@DJMcMayhem Anyway, for the purposes of your CMC, we shall assume 1582 as switchover year, and use a proleptic Julian calendar before 46.
@Adám um... that 46 you're referring to is probably 46 BC, which we don't care about
9:44 PM
@EriktheOutgolfer Oh, right. From 1. Sorry @DJMcMayhem
@Adám Since you apparently know way more about this then I do, why did the Julian calendar start in 46 BC?
If I was introducing a new calendar system, I would definitely the introduction of the calendar as 1 or maybe 0
@DJMcMayhem It was only the leap year system that was begun then, not the actual year label. For a long time that continued to be "year of current ruler's reign".
Ah, ok. So it's the anno domini system
yeah, that's what "AD" means
Yeah, I knew that. I just wrongly assumed that BC/AD comes from the Julian calendar
9:58 PM
@DJMcMayhem Test cases?
10:17 PM
@DJMcMayhem is 501 correct for 2018?
@Adám Sorry, I had to go. I'll write some up in just a minute
@EriktheOutgolfer I believe so
@EriktheOutgolfer I get 504 for 2018.
@Adám um... sure? because I believe 501 is correct
@Adám 400 -> 100. 1600 -> 400. 1700 -> 424. 2018 -> 501. Those are off the top of my head though
@Adám 2018 / 4 == 504. Minus 1 for each of 1700, 1800, and 1900 which were not leap years and that gives 501
10:22 PM
1700 -> 424
Yes, sorry
And 2000 -> 497
Something is definitely going wrong here.
Does my explanation make sense? Every multiple of 4 so far has been a leap year other than 1700, 1800, and 1900
not sure why Adám has a problem here
@EriktheOutgolfer Very simple: Sleep deprivation. ○/
10:34 PM
sleep deprivation at 22:33... guess somebody had much work today :P (Sunday?)
does 1582 count as a leap year? ("When the new calendar was put in use, the error accumulated in the 13 centuries since the Council of Nicaea was corrected by a deletion of 10 days")
@JonathanAllan is that divisible by 4 though? tio.run/##y0rNyan8/9/k4Y5eQ1MLo///AQ
yeah, -10 days is more likely to be 10 leap years less
days were leapt, but not divisible by 4
lol... unfortunately, a leap year has 366 days and a normal one 365 days... one with 355 days is -10 leap years :P
@JonathanAllan Ugh, that makes it more complicated. It should but technically doesn't
Let's say it does for the purpose of this challenge
10:46 PM
Oh, yeah they deleted :/ how does that help, were there 366 previously >confused<
>_< yeah needed to delete as they had failed to skip. I surprise myself sometimes!
so not really a leap year, as Erik said
I have 59 in Python if anyone wants to try to beat me
@EriktheOutgolfer Which new rules?
5 mins ago, by DJMcMayhem
Let's say it does for the purpose of this challenge
To keep it simple, we're treating 1582 as a leap year even though it technically wasn't
10:52 PM
*less simple
Oh wait I'm dumb
1582 isn't divisible by 4
catching, sorry
My bad, I thought it was. 1582 is definitely not a leap year
they skipped 10, didnt add any :)

« first day (2892 days earlier)      last day (393 days later) »