« first day (3203 days earlier)      last day (631 days later) » 
01:00 - 14:0014:00 - 00:00

1:06 AM
@Jono2906 "code-gulp"
@Bubbler Btw, that has had a whole lot of good knock-on effects. Thank you again!
3 hours later…
A _
3:57 AM
@Jono2906 Well, I didn't eat my lunch in order to chat here. (In other words, I guess I won the CMC.)
1 hour later…
5:18 AM
Q: And how you solve this challange?

KingWrite a Java Program to reverse the contents of the array (Java)

5:38 AM
Esolang idea: Code is made of a numerator and a denominator, which form a fraction. The decimal expansion of that fraction is the program, where the individual digits are instructions. There are no loops, instead to loop you have to form an infinitely repeating decimal with the correct instructions.
A _
That makes me think of FRACTRAN.
So there are 10 instructions in total...
5:59 AM
Or rather, use continued fraction and take its denominators as instructions.
So the golden ratio solves the "infinite loop without I/O" challenge
How do you specify the number?
Then the code can be an arbitrary math expression that evaluates into a real number.
that may be hard to interpret
You could get golden ratio by having it be root of a polynomial.
I wonder if that would be enough for all programs.
@Bubbler also, if there's no limit to the denominator, so how many instructions would there be?
6:05 AM
It could be solved in multiple ways, e.g. take a modulo or let one denominator encode multiple instructions by base conversion
And at least one of those instructions should be a jump of some kind, otherwise it'd be hard to encode two indefinite loops
No jump. It's like, barely even esoteric if you have jump.
@feersum Surely there has to be some sort of conditional though
@JoKing Eh? I know you're not new here...
Anyway you could have some kind of conditional stack opreation or soemthing if it's really needed.
I guess the question here should be "will the language be Turing complete?"
^^ I mean it could still be turing complete without an explicit jump, but that would probably push it beyond the point of usability
2 hours later…
A _
8:31 AM
@JoKing Draft idea. The language operates on the input (which is similar to a fraction). Instructions: 0: Denominator + 1 1: Numerator + 1 2: If the denominator is 0, halt the program. 3. D - 1 4. N - 1 5. If the fraction is 1, stop the program. 6. If the fraction is 0, stop the program. 7. If the fraction is not finite, stop the program. 8: TODO 9: TODO
So the decimal 0.132132132... adds the values of the numerator and the denominator.
The fraction form is 44/333.
Another sample program. In order to create a Truth-machine(output 0 & halt if input = 0, output 1 indefinitely if input = 1), here is my code:
Yeah, I was thinking something similar:
A _
I'd be glad to hear your idea. (P. S. 0.4444... solves this problem)
Skip to next zero if 0, Add one, Subtract one, Recipricol, Input, Output, Terminate, ??? (probably structure related). I'm mostly non-decided on what data struture to use as storage... maybe stack based?
A _
Maybe use a fraction?
Well yes, but how should I represent multiple fractions? I don't just want a single fraction since that's rather hard to work around
A _
8:42 AM
My idea is fractions nesting fractions... and then access them using some kind of multiplication operation.
Huh, that's interesting. That might work
Also, we could replace halt with whenever 0 is the denominator
hi all
A _
I was also wondering if we should do an eval instruction, haha
8:46 AM
@A_ I guess traversing through a nested fraction would look like a binary tree
wondering if this could be made a challenge here...bpaste.net/show/QGIL4
and +1 and -1 on a fraction would naturally be an addition and subtraction
A _
@JoKing This might be copying the fraction to the source code...
"inverse" and "simplify a nested fractional node" should serve value swapping and multiply/divide
I am guessing a subset of the codegolf population will hate it :)
A _
8:49 AM
(Naturally we are going to need improper fractions to represent the result.)
and the inverse could be used to move around many values at once
A _
Looks really interesting.
@Anush Does A-1 mean the inverse of A?
But halting when any fraction has zero denom can be a trouble, so halt only when "simplify" on the root node results in zero denom
A _
If it halts on any fraction with 0 denom then you can't store 0 anymore.
8:53 AM
@JoKing yes.. there was a typo in the first version I pasted
Instead, zero denom at current node can be a breaking condition of a while loop (or whatever looping construct it will have)
A _
Remember that the only looping construct is repeating decimals.
@JoKing you can compute the answer exactly, which is fun
@JoKing but I don't fully understand what can be a here
can you for example just give a win to the first answer that computes something ?
Questions have to be competitive, you can't just have a winner with no chance of beating them
8:57 AM
right but being the first is not a win criterion?
aha.. thanks that's very helpful
Well, technically there are zero byte answers to questions too, but they're rare
so I could have "compute this to as many decimal places as you can" for example?
clearly not here as there is an exact answer :)
or where there is an exact answer there is
@Anush Erm, maybe, though you shouldn't let infinite precision be possible
gets increasingly irrelevant as the problem gets harder
@JoKing right.. so it's no good here as there is an exact solution
Yeah. Perhaps input could be size of the array?
9:00 AM
@JoKing true!
but what is the challenge? Solve for as big a matrix as you can?
Even if the problem gets harder, it's usually fine if it can fit in a few hundred bytes in Python
@Anush Probably , lol
@Bubbler true. I just meant that if the main challenge is solving the problem at all, golfing it seems irrelevant
@JoKing it could be right?
people may complain that you just need a bigger and more expensive computer of course but I have seen recent challenges where that would hold
in fact almost all 's seem to have that property
ok so let's start off with that question for n = 2 :)
Well, we mainly assume submissions have infinite time and memory
If you limit that and have the goal be the biggest solved array in a limited time, then it would be
9:06 AM
And you can't really have the winning criteria be based off the answer's computer since that would be unfair
both are similar to what I had in mind but have not been shot down interestingly
despite clearly suffering from the "but how fast is your computer" problem
I'm one of the people who upvoted the comment: I think you should pick a time limit (say, 60s) so the contest isn't about how long someone spent running their code. Generally, all the code should be run on the same computer so they are more comparable
right.. I know there are problem with TIO but it would be great to be able to use that for this
Also, does TIO not support C#?
oh it does! Hmm.. I wonder why the answer says it doesn't
codegolf.stackexchange.com/a/195360/9207 asks how you do the byte count. Is there a good link for this I could give them?
@Anush I don't see anything about TIO in that answer. It just uses a different website.
You'd should wait until the questions about it's validity are addressed, since it's currently 7692 bytes by the dotnet link
@JoKing ah right. I could swear there was a TIO comment. But maybe it has been deleted
@JoKing I feel people should be helping the person rather than criticising them
the SE message even says that ! :)
(new user message)
9:24 AM
Is telling them how their submission is invalid not helping?
@JoKing But having an arbitrary time limit may raise a question about "on which computer?".
um, did you read the second half of my message?
Oh, sorry
my question which looked hard is quite simple now codegolf.stackexchange.com/q/195216/9207
in that I think just about any numerical integration technique/library will work quite well
... maybe :)
@JoKing "Also, you didn't even make the effort to count the number of bytes in your question." seems unfriendly
@Anush Okay yeah, that might be a bit unfriendly, but there are two other comments asking more politely about it (yours and mine).
9:36 AM
@JoKing true
I just fear they might be driven away completely given its their first answer
Well, you expect a modicum of sense with new users, like looking at other answers to pick up on obvious standards like headers containing scores
Obviously that's no excuse for being rude, but even fair points about deleting invalid answers or what the scoring method is may come off as elitist without any way to mitigate that
@JoKing they asked me how to do the byte score but I didn't know the answer
is there a link I could give them
Well, just pasting it in TIO should give a byte count
@JoKing Thanks.. I'll add that
as long as the actual submission is valid
9:46 AM
I suspect they may know nothing of TIO
As it is, the byte count for the code block in the question would be invalid
by the way.. how on Earth does TIO work? I mean which OS is it running?
@Anush Here's the github for setting TIO up. It says it runs on Fedora 26, though that may be outdated
so how does it manage C# ??
It says .NET Core with a link to Microsoft
@Anush I don't understand. What's wrong with running C# on Linux?
hmm, it does say supported for Fedora 29+
ah yes, the readme is outdated since recent commits reference Fedora 29
9:56 AM
@Anush Bash 17 bytes: cat /proc/version
Interesting, there are leaked plans to increase rep from questions to +10 instead of +5
for some sites, including this one, it makes a lot of sense
Which is something we've asked for for years
"the reasoning behind this is to benefit women because they believe the change would impact women and help them on the site."
Yeah, that kinda stuck out to me as weird too
We know that people who contribute by asking have a harder time earning privileges than people who focus on answering. Independent research suggests this disproportionately affects women:

We also see that women contribute differently to building the community’s knowledge base: they are asking more questions. Stack Overflow’s current system strongly incentivize answering by rewarding upvotes on answers twice as much upvotes on questions.
10:14 AM
Waiting for the days you’ll have to scan your ID card to have a SE account, and your rep calculation will depend on the gender stated on your card
this sort of motivation seems almost obsessive to me
It wouldn’t even surprise me
Well, we know what to do next time we have a feature request.
@feersum Go to Twitter?
A: Sandbox for Proposed Challenges

KrishnaList Max: Return the maximum value in the final Array In this challenge, start with an array initialized to zeros with indices starting at 1 and a series of operations to perform on segments of the list. Each operation will consist of a starting and ending index within the array, and a number t...

@A_ Um, yes?
1 hour later…
11:54 AM
@Adám good point!
@JoKing I just didn't know it was possible!
I remember the good old days of mono
CMC Let A be the 20 × 20 matrix whose entries are zero everywhere except for the primes 2,3,5,7,… along the main diagonal and the number 1 in all the positions a_{i,j} with |i-j| = 1, 2, 4, 8,.. What is the (1,1) entry of A^{-1} ?
@Anush 0.5952301124
@Adám is that an exact answer?
@Anush You want more decimals?
there should be an exact fractional answer
we can start with something smaller, like 4 x 4
@Anush You mean this, right?
2 0 1 0
0 3 0 1
1 0 5 0
0 1 0 7
@Anush 5÷9
12:11 PM
so the challenge is to the same in code for the 20 x 20 case
@Anush Well, I did it numerically and recognised the fraction ;-)
shall I write it as a challenge?
@Anush No.
@Adám go on....
Because it is a KC challenge for a short uninteresting output.
12:16 PM
it has an input parameter n
Let A be the n x n matrix....
OK, but will you require exact results?
Or is decimal approximation good enough?
exact only
I don't like easy questions :)
Then no. It'll basically be a problem only for languages that have exact stuff built-in or as a library.
Just generating the matrix is already involved for many languages.
@Anush Do you know the answer?
12:31 PM
@Adám I know a general method, yes
I think I see the question of a challenge being more difficult in some languages differently
first, all challenges are per language
and inevitably for any given challenge, it will be easier to code in some languages than others
@Adám very nice!
is that in APL?
Now if I can only spot the system here…
@Anush Yes.
has anyone ever paid you to write APL code?
@Anush Uh, yes. I have a full-time APL job ;-)
12:34 PM
can you say more about it?
who wants APL code?
A _
People want APL code.
@Adám I do think that's great
@A_ Mostly people, some animals…
@A_ but which paying customers?
12:37 PM
@Anush I work for Dyalog Ltd. which sells a commercial implementation. I work mainly in the development and maintenance of APL tools, customer support (when it is the APL code they struggle with), teaching, and "spreading the work". (So you have been warned: I will try to let me teach you APL!)
Thanks.. I am still interested who the paying customers are
Not sure this is public info
@Anush Yes ^^ but the biggest are financial institutions like investment managers and financial service providers, but also e.g. a (super) large oil/chemical company (can't mention names, obviously).
tl;dr APL causes global warming
12:42 PM
7 mins ago, by Adám
@A_ Mostly people, some animals…
@Adám fascinating... thank you
I never would have guessed
APL to me equals 100% unintelligible
@Anush They tend to not mention it. Some are afraid their competition will find what gives them the advantage…
@Anush Let's fix that! Up for a quick intro?
1:19 PM
Just pick random unicode characters and you'll eventually make Doom
(apparently they have had this going on for a while, but I had never heard about it)
Seems fun
@Adám I would love to but not today! :)
(my 2 cents: APL is tied for the easiest language to get started in along with Japt) and both are equally fun
@Anush Do you know when it will fit you, or will you ping me?
1:31 PM
APL's ok to start with
A _
@Anush My previous quick intro was really boring. I already know everything taught in that intro.
@Quintec Cool to know. APL has the benefit over Japt of being an actual real-world production language.
I didn't find it exceptionally easy, but it was mostly intuitive
Compared to the amount of work to start in most serious languages (so like not BF) it was pretty nice for an intro
@A_ I'm sorry for that. It can be hard to know exactly what the student knows if they don't communicate it.
@Adám I will!
I do want to learn Julia
before the language dies :)
in fact I would love it if it became at least slightly popular
1:33 PM
Can a programming language truly die?
@Adám no not fully... but lots are very close to dead
I guess, if they are specific to very specialised hardware…
A _
Lisp was dead when it was released.
cobol is not very alive
@Adám Objective-C?
A _
1:34 PM
(Because it (Lisp) is too impractical to implement)
@Anush You'd think, but:
COBOL (; an acronym for "common business-oriented language") is a compiled English-like computer programming language designed for business use. It is imperative, procedural and, since 2002, object-oriented. COBOL is primarily used in business, finance, and administrative systems for companies and governments. COBOL is still widely used in legacy applications deployed on mainframe computers, such as large-scale batch and transaction processing jobs. But due to its declining popularity and the retirement of experienced COBOL programmers, programs are being migrated to new platforms, rewritten in...
A _
COBOL is a common language.
I think the NHS in the UK is still running windows XP.. or something like that :)
Okay the thing is some really old languages are still used pretty commonly and I want to say the main reason is just that people either just refuse to adopt to new tech and adapt, or just don't want to change old infrastructure
CJam is an extinct language on PPCG. It didn’t adapt to its predators, mostly 05AB1E
1:37 PM
Not sure if maybe that many people just enjoy COBOL or VBA though
A _
@Fatalize GolfScript is an extinct language on PPCG. It didn’t adapt to its predators, mostly CJam
The university I'm applying to teaches Racket lmao, which makes sense since it wants to teach a functional language, but I'd much prefer Haskell
@HyperNeutrino did they develop Racket?
^ probably this
sometimes universities stick with the language they invented/developed
I had a language called Orwell which is basically Miranda but a version they made
when I was at university
Miranda is a lazy, purely functional programming language designed by David Turner as a successor to his earlier programming languages SASL and KRC, using some concepts from ML and Hope. It was produced by Research Software Ltd. of England (which holds a trademark on the name Miranda) and was the first purely functional language to be commercially supported.Miranda was first released in 1985, as a fast interpreter in C for Unix-flavour operating systems, with subsequent releases in 1987 and 1989. Miranda had a strong influence on the later Haskell programming language. == Overview == Miranda is...
hmm... we need some miranda answers here
A _
1:40 PM
If I can post in the language without the effort of finding a duplicate, I would love to use the language.
"Commentary is introduced into regular scripts by the characters ||"
No thank you
and cobol answers!
@Anush That would make sense, but nope.
oh racket is still developed!
codegolf.stackexchange.com/questions/195268/… has really had a lot of up and down votes!
is there some way of seeing the record ?
that is question with a huge number of both
Probably can make a dataSE query
1:44 PM
Heh, APL still winning that one with a large margin ⍥
+19/-5 isn't even that much
@HyperNeutrino how do you get that information?
A _
Click the votecount. You will eventually be able to do it once you have enough reputation.
@Anush Once you have enough rep, you can ^
aha.. I have +17/-8 for my other question
1:50 PM
@Anush data.stackexchange.com
can anyone beat +17/-8?
A _
Beat the downvote votecounts?
what are you trying to maximize
I think it is the ratio.
I guess I am looking for a ratio near 1 and a high absolute number
+100/-100 would be good :)
A _
1:52 PM
Looking for challenges with votecounts of 0... (Pretty bad that my browser is slow)
@Anush Can you give us a score formula?
@AdmBorkBork that's awesome!
I think that's been the most controversial question since I've been around.
A _
@AdmBorkBork Downvote the challenge!
1:54 PM
@Adám I think it's just the largest total number of votes for each score. So that's one formula for a vote of 0, one for a vote of 1 etc.
The key to get high vote count with lots of both + and - is a boring challenge, simple enough for lots of languages to solve.
@Adám do you still think the matrix question from earlier is not suitable?
even after your awesome APL code for it
@Anush Then a +1000/-0 would be more interesting than a +100/-100‽
@Anush But my code doesn't work for high N.
@Adám they are in different categories. The first is in the 1000 vote category and the second in the 200 vote category
the optimum is a curve
@Anush OK, so how do we compare them, since few will have same number of votes.
1:57 PM
@Adám right.. hence the need for a challenge. How high can you go with your code?
@Adám we don't. We can bin the vote counts though if few have the same votes
so 0-10 votes, 10-20 votes etc.
@Anush I think 22:331288986896183÷556542348309597. I suspect 23:70957683559613÷119203561594927 is incorrect.
@Anush Or, is 20:4777819749183÷8026844828572?
01:00 - 14:0014:00 - 00:00

« first day (3203 days earlier)      last day (631 days later) »