« first day (3666 days earlier)      last day (128 days later) » 

1:17 PM
so.. how can we write a challenge which maximises the amount of Tamil in the answer?
maybe a quine where you score is the length of the code / the number of Tamil characters?
that's probably not a good idea
need better ideas please!
What, how can that one get 2 stars.
A possibility is to simply require at least X (Tamil) characters for some tasks...
and people will have to figure out how best to include it. (variable names? in comment is not a good idea)
3? :)
Although it still seems arbitrary.
Or... minimize the number of other characters.
people could encode the information in tamil characters maybe?
Weird things to do, and might spawn some lenguage characters.
1:42 PM
@Neil like this? Try it online!
or just make a tamil ascii art challenge
lot of work
tamil ascii art sounds interesting!
2:05 PM
@Razetime New MetaGolfLang: pick the character encoding where your solution is 1 byte.
that's not new, actually
i think the name is a knowing reference to metagolfscript-n
2:22 PM
any ideas how to make a tamil ascii art challenge?
CMC print all the letters in the tamil alphabet
I might pose that as a challenge on main if it looks interesting enough
any thoughts?
Easier for languages with unicodedata.
Might want to look for similar challenges with other alphabets first.
@user202729 yes and yes
Doesn't look like a contiguous block, unfortunately
interesting to see how different langauges do it
it should be 12 vowels, 18 consonants and the special symbol ஃ
so only 31 things need to be printed
2:29 PM
What does the special symbol visarga do?
@Anush brb, gonna make a golflang with the tamil alphabet
@user visarga?
@Razetime cool!
paste the character here
It says ஃ is Tamil Sign Visarga
"The special letter ஃ, represented by three dots and called āytha eḻuttu or akh. It traditionally served a purely grammatical function, but in modern times it has come to be used as a diacritic to represent foreign sounds. For example, ஃப is used for the English sound f, not found in Tamil."
2:31 PM
prefix for some consonants for alternate pronunciation
Sounds like APL's ⎕ :P
anyone against my posting this challenge?
put it in the sandbox for a few days
@user It's like Japanese dakuten
@Anush Doesn't seem like a bad idea
hmm.. sandbox for a trivial challenge
2:33 PM
@Anush Even for a trivial challenge, it's always a good idea
Although Sandboxed posts don't get as much attention as they should :(
it will be ignored I bet
feels sorta like the greek alphabet one
can you do it in Jelly?
It's mostly generating a number range and converting to charcode
@Anush it'll very likely be easy
I look forward to seeing that!
2:36 PM
another idea to add is all the diacritics and prefixes
@Razetime yes I did wonder about that
vowel modifier variations
I haven't seen anyone post any code to output tamil here yet!
A: Sandbox for Proposed Challenges

AnushChallenge Print all the Tamil characters in any order you like. They should be 12 vowels, 18 consonants and the special symbol ஃ . Shortest code wins

please free to edit and comment
2:39 PM
@Anush staxlang.xyz/#c=%22%26F%25*%5C%22%21E%7Cr&i=
packed representation is -1
I need to see the letter! :)
run it
I got &F%
copy the full link
and open it
@Razetime you forgot to remove the Decremented, but yes
2:41 PM
so, sort of :)
it has the entire tamil alphabet in it
I really want a bigger font
@Razetime :)
you won't get it here lol
edited the question to ban your solution :)
2:44 PM
which language was it in ? Stax?
yep stax
@Razetime your fault for suggesting the sandbox :)
3:41 PM
I guess one risk is that the shortest code will be to put all the characters in a string and print them
@Anush With compression, it would be pretty easy (like the Stax solution @Razetime provided)
could still be cool though
and what would it be in C??
2 hours later…
5:19 PM
Q: .NET System.Collections.Generic.IEnumerable

DennisI hope this is the correct place to post this. I am trying to call an F# DLL function (in Nuget Package Excel.FinancialFunctions) this is using the following data types as parameters: System.Collections.Generic.IEnumerable1[System.Double], System.Collections.Generic.IEnumerable1[System.DateTime] ...

@NewMainPosts viewed by 4 users: the OP, me, Wezl and WW :P
'I hope this is the correct place to post this' facepalm
@RedwolfPrograms Do you have a link to your quick comments userscript?
:57031186 No, I think it's Jelly :P
@cairdcoinheringaahing You're right, although it errors as it is :P
That's probably the ('
5:30 PM
@cairdcoinheringaahing I thought it ran only the last line?
The ( creates a new chain, discarding everything before it, and ' is a quick which needs something before it
@user Yeah, but it parses the whole program
Ah, ok
5:46 PM
Q: Minimum of a Polynomial in Python

ibanez221What is the shortest amount of code that can find the minimum of an inputted polynomial? I realize that you can import packages like Numpy and others, but using only user defined functions, what is the shortest way to do this? For example, if 6x6 + 4x3-12 is entered (ignore parsing issues), what ...

^ How about turning that into a proper challenge?
6:14 PM
@cairdcoinheringaahing gist.github.com/RedwolfPrograms/…
Internet's been really spotty, there's a lot of lightning even though it's below freezing, and the school's wifi goes down for an hour if you lightly breathe on it
Are there any compiled languages with eval?
@RedwolfPrograms Java allows scripting for JavaScript and other languages, if that's what you mean
some like lisp are usually regarded as interpreted, but are compiled behind
And Scala's Toolbox lets you do that, albeit a bit verbosely
also many interpreted langs, like js, have compilers
so js could count as a compiled language with eval
@RedwolfPrograms Toolbox.eval
6:27 PM
@RedwolfPrograms what is this for? just curiosity?
Yeah, just curious :p
You'd probably have to include the whole compiler in the resulting binary, which seems sort of suboptimal
or package a slower interpreted version
@Wezl also prolog
If you're compiling and running code at runtime, it's going to be kinda suboptimal anyway, though
ig eval probably works best with interpreted languages
var f = n => n + " > 1 ? " + n + " * eval(f(" + n + " - 1)) : 1"
This is such a hilariously stupid way to find the factorial of a number
@Wezl Guess Python kinda counts too, then
6:33 PM
Wait I change my mind, it circumvents the recursion limit
also it's kind of lazy, bc it returns a string
tho thunks are better than strings
Wow, a "hilariously stupid way" is safer than the normal way to do factorials :)
Wait is it really lazy?
because recursion in an imperative language is a sticky mess
@user no, but it returns a string instead of immediately evaluating
@Wezl Oh ok. Just one layer of laziness then :(
I guess that means it's delayed, while true laziness would automatically evaluate when needed and remember the result afterwards
6:37 PM
var f = n => n > 1 ? n * eval("f(" + (n - 1n) + "n)") : 1n; is a more reasonable approach, with bigints too
does it really circumvent recursion? it's just changing the stack from
f > f > f
f > eval > f > eval > f
Yeah, JS is weird
Is that what they call corecursion?
But f(4000n) works, even though it takes two seconds :p
Wait no
It works without the eval too
I guess Chrome is optimizing it
So the recursion limit didn't exist all this time (at least not in Chrome)?
6:41 PM
That code would error a few versions ago, I swear
21837 - that's a lot higher than 256.
Actually, 9643 now.
I've been lied to!
my testing says "Infinity"
oh yeah, I didn't use bigints
Dang it, the eval one doesn't circumvent the max stack size :(
6:47 PM
7:13 PM
> 11,990 questions
It was 12000 yesterday :P
So you're saying we'll have another 12kth question?
Yep, assuming we ask faster than we delete
Or maybe stack exchanges follow the same patterns as populations:
1. Asking and closing/deleting ratios are both high
2. Sudden increase in quality, dropping deletion ratio and rapidly increasing question count
3. Dropping asking rates, causing it to even out
4. Low asking and deletion rates; high average quality
never used SEDE before, can someone help me make a query that searches for all APL answers that use ⍥?
@rak1507 @Bubbler is your man, I think.
7:20 PM
Pretty sure no other language uses , so you could probably just look for answers that contain
right, that would work too
@RedwolfPrograms why would increased Q count lead to dropping asking rates?
It probably doesn't. That's basically the stages of the Demographic Transition Model applies to SE questions :p
oh, looks like SEDE doesn't match diacritics
so LIKE '%⍥%' matched %O% which you can imagine matched... quite a few things
Well I wasted a lot of time then
7:30 PM
@rak1507 Ah, I was wondering why that didn't work :(
@rak1507 How in the world did it get that was similar to O‽ Does it look at visual similarity?
oh maybe it matched something else
I think it's O, not 0
but either way, definitely not ⍥
@user I tried something similar
Honestly I wouldn't be surprised if SQL matches by visual closeness
7:31 PM
@RedwolfPrograms scribbles notes for Hatred
but yeah... anyone know how to match ⍥ not O/0/whatever
So 25 matches ZS
@rak1507 Character code maybe?
how would I do that lol
7:32 PM
@RedwolfPrograms Hence LIKE and not MATCHES
Or you could use some kind of CNN to rate similarity
Wait if you change the font in the text box maybe they won't look similar
I'll try it
A: Sandbox for Proposed Challenges

Beefster(WIP) Settling the Lands of Codegolfia king-of-the-hill grid The challenge controller will randomly generate a 200x200 map representing the terrain of the land Your task is to write an AI whose goal is to have the largest population after 500 turns. Start Each player begins with 1 cell claimed a...

surely I can't be the first person who's ever wanted to search for non ascii text
Given that SO.ru and similar exist, I doubt it
7:39 PM
Oddly enough, this answer's body matches %APL%⍥%.
Unicode support is so annoying, you either don't need it and spend forever on it, or do need it and are working with something that doesn't have it
actually maybe if I search just for APL answers, the fact ⍥ will be O is not too bad as APL answers are less likely to use it? worth a shot
I believe there's a rust crate for grapheme manipulation... and if you don't need unicode awareness, you can juggle bytes instead.
@rak1507 It doesn't seem to be working for me.
me neither
7:43 PM
This is frustrating
why can't it just work!!!
As far as I can tell, the first one doesn't even have Os in it
who knows what it's doing
Okay, just wrote the code for one of Ash's most important features...time to test it
@RedwolfPrograms That's the one byte builtin for 69 right? ಠ_ಠ
Looking at the users sorted by edits (link), it's insane that Martin's totalled up 1600 edits
7:47 PM
Don't tell Lyxal I removed it in favor of 36
1632 is almost double the next highest (mbomb on 877) and mbomb is a solid 200 ahead of Riker, Xcoder and me
It's working surprisingly well but seems to think [1, 2].map(n => -n) (in pseudocode) is [-2]
What is this builtin?
, map
It's the first one that uses non-static arguments
So you have higher-order functions too? That's pretty cool
7:54 PM
Sort of, yeah
Hang on wait, it's actually working the output format's just wrong
First try!
It works!
As an example, Ṃ₀o, in addition to being the sound a dying cow makes, will map over an inputted array and increment each item
@RedwolfPrograms Is Ash designed to be a golfing lang?
3 bytes to increment each element in an array doesn't sound competitive for a golfing language
That's 1 byte shorter than Golfscript
8:09 PM
APL: 1∘+
really just 1+ if you're doing other things with the array
There's a shorter way to do it, but I'm just using map so that I have a simple example of how map works
@RedwolfPrograms Ah, good :P
Note that that example does include input, though
is there implicit input/output?
Yes, but it doesn't work well with operators that take functions as input
It costs more for really simple operations, but it's surprisingly compact for more complicated stuff
8:13 PM
will it be able to outgolf jelly?
Not sure, it'll probably be situational
Ash will probably do better for some stuff, but no idea how much
it's got some stuff like multidimensional array manipulation that it'll probably do really well at
Personally I think its kinda impressive that despite having no dev for the past 2 years, Jelly is still competitive
I think the current ones are pretty close to optimal. Until something revolutionary comes along.
I'm kinda scared that the "revolutionary" thing will be an improved form of compression
and that all golfing languages will basically just be Stax
After Ash, I'll attempt a language with sub-byte commands
8:18 PM
I tried that, it never ends up shorter
The limited command set combined with the fact that we require answers to be full bytes means that it's very good at a few specific things and awful at the rest
If you had maybe 48 commands that were 6 bits and a few hundred that were ten or twelve, that'd probably do well
sledgehammer does well in challenges I've seen
Perhaps a system where you take six bits, then if it starts with 11 you take six more, that gives you 48 short commands for havily overloaded common operations, and 1024 operators for other things.
Sure three bytes in a typical language gives you 2^24 possible sets of operations, but most are useless. If you divided that into two operators instead of three, you'd get many more useful combinations
Not good for trivial challenges, but potentially OP for complicated stuff if you chose a operators that made it competitive in a large range of challenge types. It'd probably worm well with multiple people, with experience in different things (quines, matrices, string stuff, math, etc.)
the best way would be to utilise every possible combination of bytes
at the moment, there are certain combinations in jelly that will just never get used
so they should do their own thing
that + some form of compression would easily win
At some point compression will be useless because of the entropy
I doubt compressing every Jelly answer on the site would be much better than like 10% compression
8:30 PM
potentially less than one byte compression though
Q: Generate a "pi" number

fasterthanlightGenerate the smallest integer of the form \$x^x\$ with \$n\pi\$ digits, given \$n\$. If there is no such number, take the nearest number that has as many digits of \$\pi\$ in its digit count. Test cases: Rules: Shortest amount of bytes of code wins Your program must complete the operation for \...

9:15 PM
A: Sandbox for Proposed Challenges

fasterthanlightGenerate the smallest integer of the form \$x^x\$ with \$nπ\$ digits, given \$n\$. If there is no such number, take the nearest number that has as many digits of \$π\$ in its digit count. Test cases: Rules: Shortest amount of bytes of code wins Your program must complete the operation for \$n≤5...

Wheat Wizard's new profile picture keeps throwing me off :p
Hi all
Which languages make it tricky to print Unicode characters?
How do you do it in C?
Not necessarily printing, but working with unicode in JS is awful because it's like the only remaining thing in the world that still uses UTF-16
9:30 PM
WinAPI does, too
@RedwolfPrograms well, Java and C# too
Then there's Java, which uses utf-16 so hard that when they tried to implement utf-8, they ended up with ... CESU
Not sure why so many programming languages use UTF-16...seems like the worst of both worlds
Uses double the memory for ASCII, but you still have to deal with variable sizes characters so you don't get the speed of UTF-32
@dzaima JS at least has [...string] to get a list of characters, which is reasonably similarly usable to just the string
Of course it's not useful for actual web development because of the browser support :/
Code golf is nice because I can pretend I've never heard of internet explorer and then I become 50% happier
9:36 PM
@RedwolfPrograms fwiw java stores ASCII-only strings with 1 byte per char (and i'm pretty sure handling UTF-16 is at least a bit faster than UTF-8)
pretend you've never heard of what?
This: what an awful way to ruin a fun challenge
@RedwolfPrograms tip: make sites only targeting sane people who have normal browsers
UTF is the best of both worlds if a lot of your text is non-ASCII and none of it is emoji or any of the fancy stuff like that. Which is, almost never.
@dzaima I think I read somewhere that UTF-32 is fastest, followed by UTF-16, then UTF-8, although on modern computers I'm sure the speed differences are minimal
9:38 PM
@cairdcoinheringaahing what ruins it?
@RedwolfPrograms right, UTF-32 is of course very fast as it's very trivially vectorizable and stuff
"For invalid inputs (non-integers, integers out of range), throw an exception or output/return nothing/null."
"An integer with leading/trailing whitespace as input is considered invalid."
"An integer with a + as sign character as input is considered invalid."
"If your language allows you to pass an already-parsed integer as input, the above parsing rules (except the range one) don't apply, because the int is already parsed."
"Use of built-in prime generators and primality testers (this includes prime factorization functions) is not allowed."
oh right
We should just make the default input and output methods so horribly awful we get used to it, and when we get something like that it's actually nice because it's easier than usual
> @ProgramFOX But we're still required to treat inputs that start with a + as invalid, even if our language's string->int function parses them fine? – Runer112

@Runer112 Yes, you are. I will not drop that requirement, because it makes a nice addition and makes the challenge a bit more difficult. – ProgramFOX
9:41 PM
@Adám Yeah, sure.
Did it work?
@Adám oh cool that guy I saw one of his tweets was about APL
@wizzwizz4 It did. Thanks.
@JohnDvorak oh is that why this method is so stupid and caused me so many headaches
Morning y'all
9:50 PM
Am I the only one who would absolutely love to see an argument between people who didn't know about time zones about whether it was day or night
IIRC there's a place in Norway which is getting rid of time because they're so far north
@RedwolfPrograms wdym? It's clearly morning!!!
@RedwolfPrograms Evening
@cairdcoinheringaahing no!!!!! It's Morning!!!!!!
@Lyxal haha time zones go brrr
10:01 PM
Q: Threshold for votes in the Sandbox before posting in the main site?

fasterthanlightRecently, I posted my first post on the main site for Code Golf, and it was poorly received because I am a newcomer to the site. I was asked to post it in the sandbox, here. As of now, it has one vote, and I doubt that is sufficient enough to have confidence to post it on the main site. So how ma...

@RedwolfPrograms 69 or you don't get the upvote again
I told you the rules
And you reneged on the deal
So now you have a downvote again
Which will be removed once again when I can have the number 69 in a single ash byte
10:21 PM
A: Sandbox for Proposed Challenges

fasterthanlightmathpifastest-codecode-golf Generate the smallest integer of the form \$x^x\$ with \$nπ\$ digits, rounded down, given \$n\$. If there is no such number, take the nearest number that has as many digits of \$π\$ in its digit count. For example, if \$n=5\$, then the number of digits would be 31415. ...

any suggestions here?
personally I just don't think it's that interesting but I don't really have any other feedback
@fasterthanlight I highly recommend against bonuses: codegolf.meta.stackexchange.com/a/8106/66833
@fasterthanlight I'm excited that you've joined this site and are enthusiastic about writing challenges, which I think we could definitely use more of nowadays. So let me pass on some great advice that Adám gave on another Sandboxed challenge, which I think applies in spirit if not in the exact specifics:
"I highly recommend that you answer a few existing challenges before attempting the, frankly monumental, task of writing your own. This will give you an appreciation for well-specified challenges. I've been here for years, and have an order of magnitude more answers than challenges, and yes, my early challenges were problematic too."
@fasterthanlight Also, is difficult to get right, even for experienced challenge authors. If you do it, you'll need to test each solution on your computer to make it fair. I'd just suggest removing the fastest code part and scoring on alone
I fully agree. Writing a challenge is hard, fastest-code makes in 10x hard, and combining with other scoring makes it 10x harder still. Consider simplifying it. If you do think these are integral for your challenge, let it wait! There's no rush, you can come back to it after you've had some practice answering and writing other challenges, and with your experience you can make the best it can be.
10:36 PM
that link doesn't appear to work for me
Hmm, works for me. Does this link to the Sandbox post work? codegolf.meta.stackexchange.com/a/20703/20260
yep, thanks
First time on mobile
Not snowing, but everything's frozen and beautiful
Help I switched to desktop mode and can't get back
@RedwolfPrograms Link in bottom right corner?
10:51 PM
Got back home, now I can use my laptop again :p
@RedwolfPrograms wow you've only just used mobile se chat?
Also my hands are no longer purple
I've being using it for years now
And I'm using mobile right now
@RedwolfPrograms you want to zoom in on the bottom right corner and click the little link that says mobile
Lol I'm using mobile on my laptop right now
you what
10:57 PM
@RedwolfPrograms Heretic
Burn the witch
@RedwolfPrograms erm
I'm mean, I get using desktop on mobile
To see what happens :p
But not mobile on desktop
@Lyxal I dont, that's awful
10:59 PM
Lyxal can you please un-downvote me, I'm trying to make a golfing language and a one byte constant for 69 is not practical
There are at least three ways to do it in two bytes
69 being the most well known

« first day (3666 days earlier)      last day (128 days later) »