The Nineteenth Byte

Brun's constant is the value to which the sum of the reciprocals of twin prime pairs (1/p and 1/(p+2) where p and p+2 are both prime) converges. It is approximately 1.902160583104. Given a positive integer N, approximate Brun's constant by summing the reciprocals of the twin prime pairs where bo...

LegionMammal978

The true beauty of code written in Wolfram Mathematica:

SortBy[Map[
  FromDigits, {#[[1]], #[[2, 3, 1]]} & /@
   Select[Import[
       "http://stdkmd.com/nrr/wanted.htm", {"HTML", "XMLObject"}][[2,
      3, 2, 3, 4, 3, 8, 3, 1, 3,
      2 ;;]], #[[3, 6, 3, 1, 3, 1]] == "Free to factor" &][[All,
    3, {1, 2}, 3, 1]], {2}], #[[2]] &]

Sp3000

1:00 AM

Try not to post random code snippets without context :P

Anonymous

@ais523 What do you mean by improvable wrt your Brun's Constant answer?

LegionMammal978

@Sp3000 The tag "The true beauty..." was supposed to be sarcastic

Downgoat

@quartata may I ask how in Pytek you did implementation of a > b > c? Did you handle this at compile-time or something else?

Sp3000

@LegionMammal978 Yeah, but you haven't actually said anything about what it's meant to do or why it's sarcastic, so I'm sure it went over most peoples' heads

ais523

@Mego fewer bytes

LegionMammal978

1:05 AM

@Sp3000 And that is an excellent demonstration of why I can't social

feersum

I know Mathematica and it also means nothing to me...

quartata

@Downgoat You mean chained comparisons?

I'd have to check. It might be a sugar node

but yes it is a parse time thing

Downgoat

@quartata Yes

Random Lex/Flex question: What is difference between %{ ... %} and { ... }

quartata

1:21 AM

Not sure why you're asking how we implement it since our parser is totally different from yours :P

Downgoat

I was just wondering if at parse-time a > b > c should be converted to a > b && b > c

ais523

if I solve the problem via a different algorithm in the same language and it's shorter, do I edit the existing answer or write a new one?

quartata

A new one

Downgoat

@ais523 Depends, if it is your answer then probably edit it unless you really feel like you'll loose something if you remove existing algorithm, otherwise new one

ais523

turns out @Dennis posted it one byte shorter while I was working on it

so that rather solves the problem ;-)

Sp3000

1:29 AM

If it's a significantly different approach it should be fine to post a new answer, for future reference

LegionMammal978

Question: Is there any good set of rules for converting, e.g., non-ASCII German characters to ASCII characters?

Dennis

@ais523 I've edited my answer, so it might not be that similar anymore. No need to delete imho.

LegionMammal978

Wait, nvm

Qwerp-Derp

1:49 AM

Hello

Destructible Watermelon

hi

Sp3000

(Not-so) quick question for Brun's: can k*(k-2) ever be a cube for integral k > 4?

Destructible Watermelon

2:01 AM

Regex crossword, space must be [0-3] but also in [^2013]?

wrong word lol

Mistah Figgins

I'm stuck on double cross #2

Destructible Watermelon

this makes no sense

Mistah Figgins

Which level?

New Main Posts

0

Q: Permanently self-modifying code

Now, we all know most languages have very simple ways to "self-modify" code. However, what if you were to actually modify the code and edit parts of it...on disk? Your goal is to make code that prints a number, then edits its own file to replace the number with the next one in the Fibonacci seq...

code-golf parsing fibonacci

0

Q: Detect if your program has been mutated

Write a program that outputs NOT CORRUPTED If any single byte is changed the program should output CORRUPTED Do not read your source code from a file

code-golf

Destructible Watermelon

tel on volapuk or whatever it is

[^2O13]\.\\*(A|P)?

aw fucking hell

the 0 is an O

can someone post the relevant XKCD please

Mistah Figgins

2:05 AM

[not 0, 1, 2, or 3], then a period, then a backslash any amt of times, then [A or P]

ais523

@Sp3000 I can't quickly do the maths in my head, but according to a quick Jelly program (Rµ2+×µÆE%3SµÐḟ), the answer is no for k < 4000: Try it online!

also, "is a cube" is surprisingly verbose in Jelly, it took me a whole five characters

Riker

@Downgoat for the love of all things holy + goats pls update the tf2 server ;-;

Sp3000

Yeah intuitively I feel like the answer is no, but I can't seem to convince myself of it. Basically I'm asking because I'm sure some answer's going to count the number of divisors in k*(k-2).

ais523

let's see: a cube has all its prime factors occur a multiple of 3 times; each of those has to occur either in k or in k-2, with no overlaps; it's clearly impossible for k and k-2 to both be cubes, so at least one factor has to be split between the two sides; that means that k and k-2 are both divisible by the same prime, which must therefore be 2

that's not a proof by itself though

Sp3000

Hmm so I guess the sticking point is showing that k and k-2 can't both be cubes

ais523

2:15 AM

no, that bit's easy; the sticking point's showing that k and k-2 can't be a cube and twice a cube, either way round

Sp3000

... oh right. Hm.

ais523

obviously you can't split a factor other than 2 between the two

IIRC there's a chat room for this sort of maths problem

Primes and Squares

For discussion about programming, math, linguistics, music, sc...

that one, perhaps?

Sp3000

Didn't realise we had such a room

ais523

no, one of the problems with TNB is that everyone goes here, so the other rooms are mostly unused

feersum

Well, having to click through a bunch of tabs to see if there's any chat is pretty horrible

ais523

2:22 AM

if you join a bunch of channels at once

then the most recent comment in each is shown in the sidebar (above the starboard), in blue if there's recent conversation

and if you favourite a bunch of channels, you can join them all at once without needing them open in tabs

feersum

I know, but that's still crappy

For example, if one channel has a conversation you're not interseted in, you won't know about the other ones from that

Mathematica claims to know that there are no solutions to 2b^3 + 1 = a^3 or b^3 + 1 = 2a^3.

Other than |a|,|b| < 2.

wat

PRELIMINARY UI FOR StackChat

feersum

@ais523 I'm solving your Incident question right now. Would setting all the characters in the string that aren't in the output tokens to null characters be an acceptable return format?

ais523

@feersum that doesn't show where the gaps between the tokens are, i.e. it doesn't distinguish between a, a and aa, so I don't think you're constructing the list of tokens at any point; thus I'd probably say no there

feersum

@ais523 But you said it was ok to output them with no delimiter.

This gives equivalent (or more) information.

ais523

2:38 AM

yep, the idea is that you should construct the list internally but I don't really care about how it's output

conceptually this question wants to insist on being a function but I wrote the spec as it is to not exclude languages without functions

feersum

What do you mean "construct the list internally"?

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Dual Kimian (Error) and Regular Quine Polyglot code-golf polyglot quine Create a program that is a non-erroring proper quine in one language... ...and a proper quine that produces its own source code through an error in another language What is an error quine? The compiler/interpreter/r...

ais523

that said, it's a fairly difficult challenge, so if you don't comply with the I/O spec in a minor way like that, I'm not going to be arguing for the post's deletion myself

feersum

I don't care what the requirement is as long as it's clear, but now I'm very confused.

Mistah Figgins

@wat Looks good!

wat

2:39 AM

@MistahFiggins Thanks

Glade just crashed and so I lost it all

Trying to remake it

Mistah Figgins

BTW, is there a way to type "\n" chars in chat? or do you have to copy paste for multi-line?

wat

@MistahFiggins shift+enter

like

this

Mistah Figgins

Got it.
Thanks!

wat

You're

welcome

ais523

I should probably have written it like that originally?

Mistah Figgins

2:41 AM

I'm working on typing out my Del|m|t documentation rn, it's pretty fun

wat

@MistahFiggins what are you using? Markdown?

feersum

@ais523 Huh?

ais523

err, wow that was confusing, I suspect a chat bug

Mistah Figgins

@wat To format it? yea

wat

@MistahFiggins k cool

ais523

2:43 AM

repeating my message from memory in case it disappeared: "basically what you're doing is writing a snippet to output the list of tokens, then adding code to make it into a full program or function"

Mistah Figgins

Cat program:
Argument: #
Code: 7#/##1#3#>#?#9

feersum

If I write the tokens to stdout, then it's OK to have no delimiter?

C is the language, btw.

wat

@MistahFiggins four spaces before lines

Mistah Figgins

Thanks (lol)

wat

@MistahFiggins Tom Hanks?

ais523

2:46 AM

@feersum in C it's probably simplest to just output each token with a separate call to a print statement; not only does that comply with the spec, it actually is observable to an outside process whether two characters were output at the same time or at different times, if it goes to enough effort

feersum

@ais523 That's actually quite painful compared to using putchar() in a loop.

ais523

ah, I see

feersum

The only way to output N characters of a string is to use fwrite, which requires a#include.

ais523

is there some point at which you can look at the program's memory and see the string being broken into tokens? (also printf also works but it's rather more verbose; it doesn't need an include on gcc, even though it does on portable C)

feersum

printf won't work unless I temporarily overwrite part of teh input with null characters to mark the end of what I want to output.

What you see currently is an array that has a[i] = N if there is an N-character token starting at position i, a[i] = 0 otherwise.

ais523

2:50 AM

printf("%.*s", length, pointer_to_first_character)

I might have the arguments backwards there

feersum

That's not how it works.

It pads the argument to length if it's shorter than length.

There's no truncation if it's longer.

ais523

no, that's * you're thinking of, this is .*

anyway, I'd say that given the representation in memory you have, just printing the character or NUL is enough

<man 3 printf> "This gives … the maximum number of characters to be printed from a string for s and S conversions."

feersum

OK it works, thanks

Mistah Figgins

3:23 AM

Prints "987654321" by changing the regex at the end of the first iteration:
Argument: [ -?]
Code: %A&B'C(D)[)Z>^8Q?\

Downgoat

@Riker ok

Conor O'Brien

> Stack Exchange App Test

Downgoat

@Riker done, no update was there

Mistah Figgins

Hows This for a wiki?

Now i just need an interpreter...

wat

Hey people

I'm about to post an answer with my new account "Stack Exchange App Test"

Please upvote it enough that it gets to chat

aka 20 rep

Dennis

3:35 AM

Don't beg for upvotes. Make quality contributions.

13

wat

...

I have never gotten more than about 4 upvotes on an answer

even when I was the first one

Mistah Figgins

Quality is not the same as speedy

Maybe even being speedy is detrimental

wat

I know, but the FGITW effect is real

DJMcMayhem

Yeah, but not the only way to get votes

Dennis

Be that as it may, the only thing worse than a crappy contribution from a chat sock account is an upvoted crappy contribution from a chat sock account.

ais523

3:37 AM

being fast increases the chance you get upvotes, but unless the question is so easy that a lot of people can write a good answer quickly, it doesn't help you get more than a few upvotes early

that said, we had a rash of questions where posting a good answer quickly was all that was needed, around the time I joined

wat

this isn't even a bot or anything, it's just for testing the login mechanism without accidentally leaking login data for my real account.

ais523

I was unemployed at the time, it lead to me repcapping a lot

wat

All I need is 2 upvotes

Downgoat

Is anyone here familiar with Lex/Yacc or Flex/Bison?

ais523

I am

Conor O'Brien

3:39 AM

I am somewhat familiar with bison

tho I'd tust ais523 more >_>

Downgoat

@ais523 Could you tell me what the difference between %{ ... %} and { ... } is?

Destructible Watermelon

@Downgoat one has % signs

ais523

they both insert literal code; however %{ … %} is used by yacc and bison (EDIT: and lex and flex) as a sort of compiler directive to insert code that it doesn't really look at, typically near the start of the program

Downgoat

what do you mean by "doesn't really look at"?

ais523

whereas { … } is used by rules in yacc/bison (and lex/flex if you want your code to stretch over multiple lines) to specify the code that implements the rule

by "doesn't really look at" I mean it just copies it byte-by-byte to the output

and only reads it to know where it ends

Downgoat

3:41 AM

<expect>[0-9]+"."[0-9]+      {
            printf( "found a float, = %s", yytext );
            }
<expect>\n           %{
            BEGIN(INITIAL);
            %}

ais523

oh, that's ridiculous, I'd use {} in both cases

Downgoat

@ais523 ah ok, I see

ais523

%{ … %} blocks are idiomatically only used at the start of the file (and the end, although often there's a more lightweight syntax to put them at the end)

Dennis

@wat And there's nothing wrong with that. I have a sock as well. Asking for upvotes is where the problems commence. All upvotes should be cast because of the post, not because of the poster.

Conor O'Brien

@Dennis wait is ಠ_ಠ your sock

ais523

3:43 AM

the bison documentation outright says you're only meant to use %{ … %} once in a program

Qwerp-Derp

Hey does anyone want to play Black Hole?

Maybe like a three player game thing

IDK

I can actually make a decent KoTH now because I know Java yey

ais523

flex documentation says that lines between %{ and %} with leading whitespace are copied verbatim to the output; unlike {} it doesn't seem to make guarantees about where it ends up

so I'm guessing the %{ … %} is exploiting undefined behaviour to get the BEGIN(INITIAL) into the right place

not sure what the leading whitespace is for, it's probably to avoid violating a standard that specifies what to do with unindented lines

Qwerp-Derp

Uhh how to make KoTH in Java?

Dennis

@ConorO'Brien Sure, I learnt Octave to create a sock account and give it enough rep to chat, then posted a Julia answer to make everyone think it was Alex's, and finally an ES6 answer just for the heck of it. That sounds exactly like something I would do.

2

ais523

making a KoTH is much more complex than I can fit into PPCG chat, really

Destructible Watermelon

3:47 AM

damn, I was gonna go to arqade to ask a game identification question but I don't have a screenshot or whatever

Qwerp-Derp

Is there a challenge where you have to do something like this?

ais523

@Dennis it's the sort of thing I would do if I had some reason to get a sock reputation without people knowing it was me, but I don't see why I'd do that

Conor O'Brien

@Dennis ... maybe that's how you feed the 11?

Dennis

No, this is how I feed the 11.

a contracted slim noun

21 4

bitwise code-golf integer

Qwerp-Derp

There's an equilateral triangular grid, numbered 1, 2, 3... all the way until the last dot.

  1
 2 3
4 5 6

Etcetera etcetera

Conor O'Brien

3:50 AM

@Dennis yeah I have no idea what that means lol

or why its post has a downvote

Qwerp-Derp

And the task is, given a number and the size of the grid, return the neighbours it has

ais523

wait, if its post is at +2 / -1, why does it have 21 rather than 19 rep? accepted edit?

also, in retrospect, that post is using a very Dennis algorithm; different golfers have their own style

Dennis

@ConorO'Brien The downvote was cast automatically by the Community user when I edited the answer. I consider this a bug. And I posted the answer precisely to corroborate Doorknob's theory regarding that bug.

ais523

oh, cleared VLQ flag downvote?

I think that's a bug too, if Community is autoflagging things as VLQ and the post gets edited, that's not a reason to convert the Community VLQ flag into a downvote, on the basis that Community should have downvoted rather than flagging

Dennis

I would have done it with my regular account, but it's very hard to get a VLQ flag from Community with a high-rep account.

Conor O'Brien

3:53 AM

ahh

Dennis

Nov 30 '16 at 16:22, by Dennis

@Geobits Close enough. Dennis' mortal account

ais523

there's a partial fix, in that the downvote is cleared if the post is ever upvoted, IIRC

Nathan Merrill

since the real Dennis is immortal

ais523

but only if the upvote happens after the post is edited, for some reason

(you'd think an upvote beforehand might be a clue that the answer isn't actually that terrible…)

Dennis

@ais523 That's true for regular VLQ flags, but not those cast by the Community user.

ais523

3:55 AM

oh, that's surely got to be a bug, or at least fairly stupid

come to think of it, maybe Community might get some rep on PPCG if the bug were fixed

(Community's rep is nearly always 1, but that's not due to a special case or anything like that; it's because it downvotes so many things)

Doorknob

@ais523 Quite stupid, yes.

Conor O'Brien

and it still hasn't been fixed... :/ would offering a bounty help remotely?

Dennis

It would at least burn some meta rep.

Destructible Watermelon

my contribution on elendia starman leaving and stuff

ais523

not normally, bounties tend to attract regular users who can answer the question, but the developers don't need reputation much

Destructible Watermelon

3:57 AM

ais523

(also, I don't really have SE-meta rep to burn, my meta rep's still low enough that I can make a noticeable gain from suggesting an edit there)

Conor O'Brien

mine's at 100

Dennis

@DestructibleWatermelon What on Earth is that supposed to be?

Destructible Watermelon

@Dennis the new starman

Dennis

Suddenly I want to leave too...

ais523

3:59 AM

association bonus + 1 downvote?

Conor O'Brien

@ais523 some convoulated combination of upvotes and downvotes and a bounty

@Dennis NO PLEASE DON'T :(

Qwerp-Derp

I don't know how to do this :(

I'm making Black Hole in KoTH probably for multiple players

Destructible Watermelon

@Qwerp-Derp what is this koth?

@Dennis ._.

Nathan Merrill

@Qwerp-Derp is this black hole from tom scott?

Qwerp-Derp

@NathanMerrill yup

Nathan Merrill

4:03 AM

I actually already made a controller for that, but never posted it

and I think it is much better as a 2 player game

Destructible Watermelon

what is this black hole you speak of?

Nathan Merrill

Black Hole - Emma Blackery vs Katie Steckles - Game On 1x03

►

its a pretty interesting game

Dennis

Yay, first restricted-source silver. codegolf.stackexchange.com/help/badges/148/restricted-source

Mistah Figgins

@NathanMerrill So do you have to play in order?

Nathan Merrill

yes

Qwerp-Derp

4:09 AM

Yup

Uhhh I kinda want to try out a game of Black Hole in chat just to see if there's any strategy

Maybe I should make a room?

@NathanMerrill Do you have the repo for the controller?

Nathan Merrill

nope. I don't have the code on me either, as that computer is getting repaired

Qwerp-Derp

Ah

Nathan Merrill

but it should be really easy to write with KoTHComm

Qwerp-Derp

You need to write specs for KoTHComm though

Nathan Merrill

I'd guess around 3-4 hours

I know

Qwerp-Derp

4:12 AM

@anyone Does anyone want to play BlackHole in a separate chat room?

Nathan Merrill

ok, I'll write up a bunch of specs for KoTHComm

Qwerp-Derp

So what does KothComm do exactly?

Nathan Merrill

everything but write the game

it can communicate with players over piping, download players from a question, run multiple games, aggregate the scores into a single ranking

Qwerp-Derp

Oh

Whoa that's cool

@DestructibleWatermelon Do you want to play a game of Black Hole?

Destructible Watermelon

@Qwerp-Derp maybe

Nathan Merrill

4:15 AM

@Qwerp-Derp I'd highly recommend making the board bigger for the real tournament. I feel like a 18-grid is too easily brute forced to find the optimal solution

Destructible Watermelon

make the room

Qwerp-Derp

Black Hole

Playing games of Black Hole here.

I've made a challenge spec on Meta, hang on I'll send you a link @NathanMerrill

Destructible Watermelon

maybe you should alter the game to work with an infinite grid

Qwerp-Derp

meta.codegolf.stackexchange.com/a/8944/49561 here it is

Destructible Watermelon

or, wait, I have an idea

Qwerp-Derp

4:16 AM

@DestructibleWatermelon Go on the room

Nathan Merrill

@DestructibleWatermelon that doesn't work. You need a point to finish and score

Destructible Watermelon

multiple bots, big grid, memory between games

@NathanMerrill I know

Qwerp-Derp

I was thinking multiple players, but with multiple black holes

So there's not just one black hole, and it makes the game much harder to bruteforce

Destructible Watermelon

it would be a large modification to the game if that happened

also why not hyperbolic play field so that each space has adjacent 10 spaces

Qwerp-Derp

@NathanMerrill Do we start off with 10 or 1?

Nathan Merrill

4:23 AM

1

and you make it 2 player, and 1 black hole

anything else makes things too undeterministic IMO

Qwerp-Derp

Mmhmm

I think the first few moves are random and then it just gets strategic at the end

@NathanMerrill You want to join the room?

Nathan Merrill

nope. Working on the spec for KoTHComm

@Qwerp-Derp that's not true.\

Riker

@Downgoat ... well it updated so

I tested just when I sent the frist message, just did agin and it work

Nathan Merrill

especially not for computer players

Qwerp-Derp

@NathanMerrill Kindofbutnotreally

Nathan Merrill

4:27 AM

you want your low numbers to touch as many spaces as possible

Qwerp-Derp

so in the middle?

Nathan Merrill

yeah, but not necessarily in a cluster

Nathan Merrill

4:38 AM

@Qwerp-Derp it turns out I did push it to github

github.com/nathanmerrill/BlackHole

that said, it is not using the latest version of KoTHComm

Qwerp-Derp

Oh

Wait when did you make KothComm?

Nathan Merrill

its been a WIP for about a year now, I think

Qwerp-Derp

Oh nice

What version is it now?

Or does it have no version yet

Nathan Merrill

I really should start to version it\

I've been wanting to put it on maven

so people could easily grab a jar

but as of now, I work off of the repo

Qwerp-Derp

Oh

Nathan Merrill

4:43 AM

typically, I add KoTHComm as a git submodule within the project

Qwerp-Derp

How much specs do you have to do?

If you're finished do you want a game of BH?

Qwerp-Derp

4:59 AM

I think Black Hole is kindof counterintuitive

Anonymous

@Dennis Wait that's not what you did?

Anonymous

(disregarding the fact that we've all-but-confirmed that disapproval-face is VTC's sock)

Dennis

@Mego Sure, VoteToClose learnt Octave to create a sock account and give it enough rep to chat, then posted a Julia answer to make everyone think it was Alex's, and finally an ES6 answer just for the heck of it. That sounds exactly like something VTC would do.

(It's fun to see y'all guessing when I can check. :P)

Nathan Merrill

@Qwerp-Derp added a readme

I'm going to bed

Mistah Figgins

cya, i should too at some point

Anonymous

5:07 AM

@Dennis The thing is, VTC absolutely would do that :P

Mistah Figgins

too busy working on an interpreter for del|m|t

Dennis

@Mego Would, but hasn't.

feersum

If I posted an ES6 answer, that'd be the ultimate deep cover.

Anonymous

So if I keep guessing, you'll keep sarcastically responding, letting me know that the guess was wrong, thereby allowing me to work out the true owner via process of elimination

Anonymous

Here's an interesting idea for a statistic: which PPCG user has used the most diverse set of languages?

Dennis

5:12 AM

Guven that we have 46k users, that might take a while.

Anonymous

Oh

Anonymous

It's minxomat

Anonymous

I think

Anonymous

Maybe not... Luis is a decent candidate

Anonymous

I mean, this comment kinda gives it away:

Anonymous

5:17 AM

I'm only missing some eval here to make it really EVIL :-) — Luis Mendo Oct 5 '15 at 0:11

Anonymous

But at the same time, I hope I'm not right, because trying to figure out who owns disapproval face has been one of the few things everyone in TNB can rally around. It'd be sad if the mystery was gone.

Dennis

Guesses are a dime a dozen, diamonds are priceless. :P

Anonymous

One day we will figure it out. And then we'll be sad because the great mystery is over, and we'll have to find something else with which to occupy our time.

feersum

5:33 AM

Huh? I thought it was commonly known to be flawr.

Anonymous

@feersum This is news to me

Destructible Watermelon

maybe feersum is trying to confuse people

and is actually feersum

Doorknob

"Maybe feersum is actually feersum." —Destructible Watermelon, 2017

5

Mistah Figgins