The Nineteenth Byte

Sp3000

14:01

I contributed... 2 letters.

Really cool idea though I have to say

Geobits

The U and K I bet :P

Optimizer

the space and &?

undergroundmonorail

the space and the & in "chuck"

Optimizer

yes

Lembik

hi @MitchSchwartz

and all

undergroundmonorail

14:14

i fall into one of those categories

hello

Optimizer

this is much more easier to read than imagined (and any other conventional method) (NSFW language)

undergroundmonorail

yeah i've seen sites that are supposed to train you to read faster

Sp3000

^^ Was about to comment the same, but apps not sites

undergroundmonorail

the idea is that you learn to read without doing the thing most people do where a voice in your head reads the text "aloud" to you

this specific instance is probably cheating for me because i have the text memorized and say it relatively often

Optimizer

why would you say that text often?

are you actually a navy seal?

undergroundmonorail

14:19

whenever any of my friends makes a joke at my expense

i don't usually do the whole thing

NSFW language:

What the fuck did you just fucking say about me... You little bitch.

►

Martin Büttner

Here is Hello, World! in Brian & Chuck btw... (incidentally, the 300th answer to the HW challenge)

0

A: "Hello, World!"

Brian & Chuck, 42 bytes ?Hello, World! !>.>.>.>.>.>.>.>.>.>.>. Introducing my latest esolang, originally submitted for Create a programming language that only appears to be unusable. Each line defines a Brainfuck-like program which operates on the other program's source code. That makes "Hell...

undergroundmonorail

i'm glad english doesn't use "z" very often because that key on my keyboard is kind of sticky

i spilled something on it a long time ago and every key is fine except z

actually 90% of the time z is fine but every once in a while there will be two days in a row when it's kind of sticky

Optimizer

trrr trrrr

Sp3000

Solution: Remap "q" to type "z"

undergroundmonorail

i use q all the time

40% of my communication is complaining about qwerty

(that's not actually true)

Sp3000

14:24

Solution 2: Remap caps lock to type "z". Who even uses caps lock.

Optimizer

remap windows key to z

undergroundmonorail

i actually do use caps lock and windows more often than z haha

Optimizer

power button?

undergroundmonorail

don't have one on my keyboard

i could use the context menu button, i never use that

TimmyD

@Sp3000 Back in High School, there was a person in the keyboarding class (yeah, we had those) that didn't realize SHIFT worked to capitalize letters. He would hit CAPSLOCK, hit the letter he wanted capitalized, then hit CAPSLOCK again to turn it off.

undergroundmonorail

14:25

i did that when i was a kid

I Also Capitalized Every Word Because I Didn't Really Understand Grammar Yet

Sp3000

*punctuation

Optimizer

*programming

undergroundmonorail

Punctuation Programming and Code Golf

Optimizer

Capitalization

Martin Büttner

*and Capitalisation Grammar

Optimizer

14:30

^

undergroundmonorail

perfuct

Optimizer

Programming Punctuation and Capitalization Grammar

undergroundmonorail

that sounds like a stack exchange site for questions about natural language processing

TimmyD

18 hours ago, by Digital Trauma

Protramming Puzzles & Code Golf. Pro-tramming. Seems like a big departure from our previous purpose. Are we now a professional streetcar operators organization?

Optimizer

@undergroundmonorail haven't you realized it yet that we are that site

undergroundmonorail

14:32

...considering how hard it seems to be to create well-specified challenges i'd believe it

Geobits

I no understand you mean.

Optimizer

nope

undergroundmonorail

the last time i golfed in a "real" language was february 20. since then i've only posted answers in pb, brainfuck and the tis-100 architecture

Mitch Schwartz

@Lembik hi

TimmyD

@undergroundmonorail Well, there are answers written in INFORM 7 around here ...

New Main Posts

14:51

0

Q: Create git pack file from a large git repository without using a C library directly

Ok, your task is to create a valid git packfile (sha1.pack) containing the 4 loose objects of that git repository (it takes, less than 40Mb on disk, but don’t misunderstand yourself, it’s several Gb large). Expected result : The packfile along it’s git sha1, or the code to create it. Rules : ...

code-golf

Optimizer

this question only appears to be unclear

Martin Büttner

Quick maths challenge (lifted from an old GCJ): Given two vectors of equal length, how can you permute their elements (independently) such that the scalar product is minimised?

(That isn't too hard, although proving it isn't too trivial... for me anyway.)

So now the interesting question is: does a similarly simple result exist if we try to minimise the absolute value of the scalar product?

undergroundmonorail

my constant challenge in life is "figuring out what all the words mean in whatever martin just said"

2

Lembik

@MitchSchwartz how are things?

@MartinBüttner nice question!

Mitch Schwartz

it was the best of things, it was the worst of things

Lembik

14:58

@MartinBüttner rotating the vectors to do the same is fun too

@MitchSchwartz well done on the best part :)

I am contemplating bounties for codegolf.stackexchange.com/questions/62922/…

wondering what the most effective way is :)

Optimizer

most effective way is to be Calvin!

Lembik

what does he do?

Optimizer

ask him

Lembik

@Calvin'sHobbies what do you do re the above?

Sp3000

@MartinBüttner Do you know the answer to this?

Martin Büttner

15:01

@Sp3000 Nope, only to the first one.

The second one seems much harder.

Sp3000

It sounds so similar to many NP-complete problems it's kinda hard to say :P

Martin Büttner

yeah that's exactly what I was thinking

at the same time it seems like a very minor variation of the first problem ^^

Optimizer

^^ should not be the emoji used here

Sp3000

Well it's about as minor as going through all the ~~edges~~ vertices :P

Martin Büttner

heh, I suppose

Optimizer

15:05

Sp3000

Try (1, 2, 3, 4) and (-8, -7, 2, 5) :P

Lembik

@MitchSchwartz are you interested in my challenge by any chance?

he says hoping the answer is yes :)

Rainbolt

I just shouted at a coworker because they tried to tell me for the hundredth time that 12:00 UTC -06:00 and 12:00 UTC 00:00 were the same time, and now everyone is just leaving me alone

Martin Büttner

@Sp3000 [-7 -8 5 2] gives 0

Sp3000

Yeah, the only zero solution

Martin Büttner

15:10

yes

Sp3000

Not sure how to get that without brute force

Martin Büttner

well it's negative decreasing followed by positive decreasing, but I have a feeling that's coincidence ;)

Sp3000

If we do (1, 2, 3, 4, 5, 6) and (-8, -7, -1, 2, 5, 10) then we get (-8, 10, -1, 2, 5, -7)

So... bad luck to you :P

Martin Büttner

heh

thought so

TimmyD

@MartinBüttner Reliably works for one hand-picked example. QED, let's go drink.

Martin Büttner

15:13

^

@Zgarb thanks for the checkmark, finally a nice round reputation again :)

Zgarb

@MartinBüttner It was a bit overdue. I do my accepting in batches once a month.

Calvin's Hobbies

@Lembik What's the question? I rarely award or deal with bounties.

Lembik

@Calvin'sHobbies I don't know.. I was told "Optimizer
most effective way is to be Calvin!"

VoteToClose

Anyone else find it mildly amusing that Vitsy can error quine with any error?

Mitch Schwartz

15:28

@Lembik "I think you can do it in 3^{n/2} time fairly easily" -- how sure are you about that?

Lembik

@MitchSchwartz hmm.. I thought so

is that wrong?

this is just to check a solution

Mitch Schwartz

so, something better than Peter Taylor's canDistinguishInNTurns() ?

Lembik

I think so

I think you notice that the problem is when you can't distinguish between two labellings of coins

consider a labelling of coins as a vector of length n

the difference is a vector with entries that are -2, 0 or 2

there are 3^n of these

but you can check the first half exhaustively, setting the second half to 0

and the second half exhaustively, setting the first half to zero

and then combine the results

you need to code it well to avoid using too much RAM I suppose

@MitchSchwartz does that make sense?

Mitch Schwartz

it might, but i'm not sure how i'm supposed to be looking at it. what do you mean by a labelling? they already have fixed labels, from 0 to n-1

Lembik

@MitchSchwartz sorry.. I mean we want to make a sure a solution can distinguish between all the possible coin weight combinations you might get

is that clearer?

Mitch Schwartz

15:37

sorry if dumb question, still trying to understand

Lembik

I imagined there were n coins that were labelled -1 or 1 but that is a bit confusing you are right

New Main Posts

4

Q: An infinite FTW

The infinite Fibonacci word is a specific, infinite sequence of binary digits, which are calculated by repeated concatenation of finite binary words. Let us define that a Fibonacci-type word sequence (or FTW sequence) is any sequence ⟨Wn⟩ that is formed as follows. Commence with two arbitrary...

code-golf math sequence fibonacci

Dennis

@NewMainPosts Slowest bot in the west.

Mitch Schwartz

@Lembik are you considering {n, k, {x_i}} where k is the number of steps as a solution, or something else?

i mean, we are testing whether a solution is correct, and {n, k, {x_i}} is one way to define a solution

BrainSteel

@MartinBüttner A bit late, but about your vector problem: if x = (x1, x2, ..., xn) and y = (y1, y2, ..., yn) with xi >= x(i-1) and yi >= y(i-1), we know that the minimum dot product of the two vectors x d y = y1 * xn + y2 * x(n-1) + ... y(n-1) * x2 + yn * x1, right? So, if we want the minimum value of abs(x d y), can't we just sort the original vectors into lists (x1*, x2*,..., xn*) and (y1*, y2*,..., yn*) such that abs((x or y)i*) >= abs((x or y)(i-1)*)and minimize the first way?

I'm trying to see if I can prove that.

Mitch Schwartz

15:48

note that canDistinguishInNTurns() can check a solution of this sort -- {x_i} gives mask, and n gives coinsets

Cᴏɴᴏʀ O'Bʀɪᴇɴ

I wrote my first Batch program :D

Mitch Schwartz

(and "k" as i wrote it is the "n" in the function (method))

Dennis

@CᴏɴᴏʀO'Bʀɪᴇɴ Are you OK?

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@Dennis Probably not. It's a HQ9+ interpreter.

BrainSteel

@MartinBüttner Never mind, that definitely doesn't work.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

15:51

AFAIK it works fine, just has a memory limit

Dennis

Hahaha.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Ikr. 9 prints 99 to 31 bottles.

Ze program:

@ECHO OFF
SETLOCAL EnableDelayedExpansion
SET LF=^


IF [%1] EQU [] (
  ECHO Error: no arguments passed.
  EXIT /B
)
IF /I %1 EQU -r (
  SET "prog=%~2"
) ELSE (
  FOR /F "Tokens=* Delims=" %%x in (%1) do set "prog=!prog!%%x"
)
SET "newprog="
FOR /F DELIMS^=^ EOL^= %%A IN ('cmd /u /v:on /c ECHO(^^!prog^^!

^|more') DO SET "newprog=!newprog!,^%%A"
SET "newprog=!newprog:~1!"
SET accumulator=0
SET "output="
FOR %%i IN (!newprog!) DO (
  IF %%i EQU H SET output=!output!Hello, World^^!
  IF %%i EQU Q SET output=!output!!prog!

Mitch Schwartz

@Lembik also note that canDistinguishInNTurns() places no restrictions on how/when to choose the 0 coin -- it tries all possibilities

Lembik

@MitchSchwartz In {n, k, {x_i}} I assume k is the number of weighings. You need to specifiy the initial set of coins you weigh and whether you add coin 0 at each new weighing

Mitch Schwartz

but whether you add 0 depends on the result of the weighing

Lembik

15:57

@MitchSchwartz it might or might not

it depends if you think being adaptive helps

Mitch Schwartz

to be general

Lembik

in that case yes

Mitch Schwartz

and yeah by steps i meant weighings

Lembik

@MitchSchwartz so my 3^{n/2} method works once you have fixed when you weigh coin 0 as well

VoteToClose

Hey, for anybody who does Java - How can I test if the first object of the first object of each object in a 3D array list are not something?

Mitch Schwartz

16:05

@Lembik when i asked "so, something better than Peter Taylor's canDistinguishInNTurns() ?" you replied "I think so" -- wouldn't it have made more sense to say that what you have in mind does something fundamentally different from canDistinguishInNTurns() ?

Lembik

@MitchSchwartz If I had known that, yes!

Martin Büttner

@Dennis do you have less than 13 for FTW in CJam?

*12

Mitch Schwartz

@Lembik i see no reason to expect non-adaptive strategies to be as good as adaptive strategies in all cases

Martin Büttner

@Dennis can we take inputs in the opposite order?

I think then I can do 11

Optimizer

16:33

i can do 10

Dennis

> You may accept the words in any order

Optimizer

we are talking about push ups, right?

Martin Büttner

@Dennis that doesn't answer my second question ;)

(has anyone noticed that Optimizer basically stopped golfing altogether?)

Mitch Schwartz

@PeterTaylor i think you can use a more efficient structure than Set / HashSet, because there are never any duplicate elements

Dennis

My stupid internet connection sent the ping, but not the message...

Martin Büttner

16:35

@Dennis must have overlooked that

and you're saying you can do 10?

Dennis

No, I was pointing to the message I never posted. I had 15.

Martin Büttner

oh

I am disappoint

Optimizer

ed

Martin Büttner

-.-

Optimizer

ur welcome

Geobits

16:41

yo'e

Dennis

@MartinBüttner I seem to be doing nothing else today. >.<

Martin Büttner

other than disappointing people?

Anonymous

@Dennis Have you mortarboarded yet today?

Optimizer

@MartinBüttner and outgolfing

Martin Büttner

@Mego when has he not?

@Optimizer no, he's being outgolfed apparently

Optimizer

16:43

WAT?

Dennis

@MegaTom I'm actually fairly disappointed by the given solution. I was expecting something significantly more clever than 92.9% useless code. — primo 2 hours ago

Martin Büttner

@Optimizer well he said he had 15 for his own challenge. I'm at 11 now.

Optimizer

WAT?

Its okay, we still love you Dennis:

Dennis

@Mego Not even close.

Anonymous

@Dennis disappointment

Dennis

16:45

>.<

Geobits

I can't decide if he's throwing the plant down because it's in a talking flowerpot, or the pun quality. Either works I guess.

Anonymous

We're more than halfway through the UTC day, what do you mean you haven't mortarboarded yet?

Optimizer

@Geobits No, he's already committed

user2284570

Hello, what isn’t clear in this question body ?

1

Q: Create git pack file from a large git repository without using a C library directly

Ok, your task is to create a valid git packfile (named as sha1.pack) containing all the 4 loose objects of that git repository (it takes less than 40Mb on disk, but don’t misunderstand yourself, it’s several Gb large). Expected result : The packfile, corresponding to my repository, along it’s g...

code-golf

Dennis

@Mego I spent too much time with ShapeScript...

Geobits

16:48

@user2284570 I haven't read it too deeply, but is it just outsourcing an actual need? The language restriction and netbook comment make it appear that way.

Anonymous

@user2284570 The restrictions are very loosely and vaguely defined, it's in broken English, and doesn't explain the format at all (a link doesn't count).

user2284570

The point is near all libraries for handling git repositories are written in C, or use the official git project (which is itself written in C).

Optimizer

then you need it for your own purpose, right?

Anonymous

@user2284570 So?

Geobits

If you're going to be using libraries anyway, why would it matter what language the library was written in?

Anonymous

16:51

^

Geobits

I don't even know off-hand exactly which libraries use C++ code in Android, for instance. It's all transparent to the API.

user2284570

@Geobits : no it isn’t. The repository I linked in my question is just an example to highlight the specifics rules of my code-golf. Anyway, you’re free to implement it in any language.

Geobits

As long as you don't use C libraries, even indirectly?

user2284570

YEs, because of manual memory management. (but python use a C library, ruby also does, so that’s why I wrote indirectly)

@Mego : Sorry, if my english isn’t perfect, please edit the question in that case. Concerning, the format I’m afraid the documentation wouldn’t fit in a question.

Anonymous

@user2284570 If the specification won't fit in a question, it's probably not a good question

Geobits

16:59

And this has nothing at all to do with this question?

user2284570

@Mego : But those who works with git perfectly understand the format (including me)

Anonymous

@user2284570 You understanding the format does not mean everyone does. And people who don't use git, or who don't use it for more than pull-add-commit-push won't understand it immediately.

user2284570

because of manual memory management, libgit2, the official git project…

They also suffer from bugs allowing remote code execution, and in the case of my example, they just crash, because of bugs, whereas my repository is valid. An implementation which don’t use C won’t suffer from those memory corruption bugs.

Geobits

I think the correct course of action would be to file a bug report or wait for an answer to your question on SO, personally. Shifting it to a challenge here (and then claiming it's not for a personal task) seems quite shady to me.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

^ So does talking to perfect strangers ;)

Geobits

17:06

Good thing I'm typing, not talking.

TheDoctor

6

Q: Display a MIDI Track

Background MIDI files are quite different from WAV or MP3 audio files. MP3 and WAV files contain bytes representing a "recording" of the audio, while MIDI files have a series of MIDI messages stored in MIDI events informing a MIDI synthesizer which virtual instrument to play or a MIDI sequencer ...

code-golf ascii-art music

user2284570

@Geobits : More precisely fill hundreds of bugs on each projects.

Anonymous

@user2284570 Regardless, it's not our problem. I think you drastically misunderstand the purpose of this site.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@Geobits Well played.

Geobits

@Mego You mean it's not "create a sock to outsource work to other programmers"?

TimmyD

17:12

Holy cow, I must need more caffeine or something ... First I concatenated $e=$b+$a then decided to print out the substring of $e starting at $b's length, rather than just printing $a ...

Lembik

@MitchSchwartz It would be very interesting to know if there is a difference

Cᴏɴᴏʀ O'Bʀɪᴇɴ

^^ you need some alchohol

Martin Büttner

@TimmyD You should have converted $a to a regex, searched for a match inside $b, and sliced $b from the start to the end of the match to print that.

TimmyD

@MartinBüttner Brilliant.

Geobits

Why not just reverse loop through $e, popping a$.length characters off into another string? Then just reverse it and you're done.

Mitch Schwartz

17:15

@Lembik it should not be hard to check -- you could try all 2^(k-1) possibilities and then you don't need to loop on "d" anymore, you can just use the selected value

Martin Büttner

Markets have been rocked by a second day of uncertainty after someone set up a giant Ouija board on the NYSE wall controlled collectively by the movement of the stock tickers.

3

@Calvin'sHobbies look, Randall made a comic about your binary run length encoding sequence!

Cᴏɴᴏʀ O'Bʀɪᴇɴ

XD

user2284570

No, as you can see, my example is purely an example, it doesn’t contains any real work *(I won’t have to use the result)*. I wanted to ask for codegolf first, but I thought, it would get a better attention on so.

But generally, so I think peoples should afraid in terms of security. I also inadvertently caused a large downtime on several storage GitHub servers recently *(due to a bug in the official git project)*.

Geobits

Then you should tell them about it. They should know if there's a bug that can take down servers, no?

Mitch Schwartz

@Lembik (the quick-reject code would also change)

user2284570

17:18

Anyway, I think the question

-1

Q: Create git pack file from a large git repository without using a C library directly

Ok, your task is to create a valid git packfile (named as sha1.pack) containing all the 4 loose objects of that git repository (it takes less than 40Mb on disk, but don’t misunderstand yourself, it’s several Gb large). Expected result : The packfile, corresponding to my repository, along it’s g...

code-golf

doesn’t have the right close reason.

Geobits

The asker very rarely thinks his own question is unclear, whether it is or not.

Anonymous

@MartinBüttner -1 not enough jquery

Martin Büttner

If you have regex you don't need jQuery. Everyone knows that regex can do anything jQuery can (and more).

user2284570

@Geobits : and the closer seems to rarely comment (in order to explain what peoples don’t understand) or try to edit before closing.

aditsu

@MartinBüttner I think we have a relevant challenge: codegolf.stackexchange.com/questions/33059/draw-with-your-cpu

Geobits

17:22

@user2284570 sigh Yes, because closing is supposed to be fast, to prevent unwanted answers before the question is "good to go". It's a feature, not a bug.

Anonymous

Geobits

You've gotten feedback in chat, but you don't seem responsive to any of it, merely defensive.

Martin Büttner

@Mego I'm reading "pork".

user2284570

@Geobits : a feature which have some many meta post to prevent that kind of use.

Anonymous

@MartinBüttner I can't help you there

Geobits

17:24

To prevent what use? The very reason it's called "on hold" is so that it can be fixed. It's the whole point. You close it up while it's being fixed, not fix it before closing. Having an open vague/broad/whatever question attracts answers which may be disqualified later by edits.

3

If you know a (well-received) post on our meta site that contradicts that, please link it.

user2284570

Anyway, I edited the question, the edit is in the review queue.

Geobits

Your edit doesn't really address any clarity issues IMO:

> If you really don’t know how git works, you can try to start by reading the long documentation about the pack file format.

Mitch Schwartz

@Lembik @PeterTaylor sorry if bugging you with an idea that won't pan out, but could it work to consider, instead of a coinset, a list of ternary digits with 0=0, 1=1, 2=[undecided, could be 0 or 1] ? so each of these objects would describe potentially many coinsets at once. i'll try to think about it

Thomas Kwa

@Geobits Is it advised to approve edits that improve the question (e.g. fixing grammar errors), but don't address a major issue?

user2284570

@Geobits Ok, updated the edit again.

@ThomasKwa I posted this

-1

Q: Create git pack file from a large git repository without using a C library directly

Ok, your task is to create a valid git packfile (named as sha1.pack) containing all the 4 loose objects of that git repository (it takes less than 40Mb on disk, but don’t misunderstand yourself, it’s several Gb large). Expected result : The packfile, corresponding to my repository, along it’s g...

code-golf

as a guest, so I can’t edit it myself directly.

Anonymous

17:39

All your edit did was make the post even more unreadable

Anonymous

This is fundamentally not a good challenge

user2284570

@Geobits : It’s like downvotes, they even ended up by putting a comment telling you should write a comment.

Geobits

@ThomasKwa Sometimes. I mostly do that case by case I guess.

user2284570

@Mego : still about the first rule, or more generally ?

Geobits

@user2284570 That's a whole other discussion. While I'm normally up for a debate about commentless downvotes, I just don't have it in me right now.

Anonymous

17:44

@user2284570 "The should be able to on so it would be a bonus"

Cᴏɴᴏʀ O'Bʀɪᴇɴ

^ that sounds like a request rather than a challenge imo

Geobits

That's because at its root, this is an SO question that hasn't received an answer.

Anonymous

@CᴏɴᴏʀO'Bʀɪᴇɴ That's what I've been saying. It's basically the same as his SO question, poorly reworded into a vague code golf challenge. It's not a good challenge, and no amount of editing will make it a good challenge.

user2284570

@Mego Updated with : The git repository is just an example for showing on what the code should be able to work on. You’re free to not do it, but using it would be a bonus.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@Mego I know, just supporting your point ;)

Anonymous

17:53

@user2284570 See my previous message

Geobits

My advice would be to just let this one go. If you want to join the community, take a look around at other challenges, maybe solve a few before jumping in with your own. If you're just looking for a bug workaround, you're in the wrong place, as has been said.

2

user2284570

@Geobits : Ok, but it’s not a single bug (hundreds bugs). and in that case this just make the wrong close reason.

@Optimizer : Please not it is not really better suited on so.

Optimizer

umm, what?

Geobits

One bug or many doesn't change anything. And even if I agreed that the close reason was wrong (which I don't), reopening it just to close it as off-topic or something else doesn't really accomplish anything.

Lembik

@MitchSchwartz sorry I missed your comment ..let me see if I can find it

Mitch Schwartz

18:01

@Lembik @PeterTaylor i think my ternary idea isn't very helpful -- it can reduce the size of "coinsets" at higher levels of the recursion, but resolving the 2's to 0's and 1's takes a lot of operations, and limits the ability to use bitwise operators

Lembik

ah ok

@MitchSchwartz I am thinking of offering a double bounty :) 50 rep for a week and then 100

would that be acceptable?

Mitch Schwartz

bounty would not help my motivation, but maybe would for others

user2284570

@Geobits : So basically, you’re telling your’re understanding the question and you’re not understanding at all. Pretty incoherent.

VoteToClose

So, I've discovered that attempting to turn a stack-based language into a stack-based-object type language is a major error in judgement.

2

I've spent two hours on it and I still can't get the non-static methods to work right.

Anonymous

@user2284570 We have explained to you over and over why your question is not suitable. It's time to stop arguing.

Mitch Schwartz

18:07

@Lembik my general ideas (maybe not helpful) are: (1) use peter taylor's approach beyond brute force range using random masks, or maybe with heuristic and/or local search, (2) general heuristic is to minimise the size of largest partition at each step, (3) trying to join coinsets into classes of coinsets, like with the ternary idea

Lembik

18:20

@MitchSchwartz I think some heuristics to look for plausible solution is a great idea

exhaustive search doesn't :)

VoteToClose

@user2284570 I left you a comment on your post explaining why I, at least, voted to close it. I agree with @Mego about the basis for the question not being suited for code golf, and the fact that you asked a similar question on SO confirms this.

That moment you realize you can't expand votes on other sites. o-o

Optimizer

you are not alone

VoteToClose

Peter Taylor

@MitchSchwartz To be perfectly honest, I wrote that code while the question was in the sandbox just to check that there wasn't a trivial pattern. The question went up while I was away on a business trip, and I was sure that a better answer would be posted before I got back, but when that turned out to be wrong I just posted it without any further work on optimisation.

mbomb007

@MartinBüttner Contrats on hitting exactly 63,000 rep!

Optimizer

18:30

LOL

we saw it, you can't hide it ..

mbomb007

is was a mistype

I was literally looking at the number while I typed that

Optimizer

still funny.

Calvin's Hobbies

Contrats?

3

mbomb007

63000 doesn't seem quite as round as the number I typed incorrectly, lol

oh well

If it were car mileage, I'd celebrate it.

Mitch Schwartz

@PeterTaylor oh, cool, thanks for the info

(above, i meant minimise the size of the largest set in the partition, not the largest partition)

Peter Taylor

18:36

I understood. It's probably a good heuristic.

Alternatively, it might be interesting to gather some statistics on all optimal trees for n up to 14 and see whether any useful conclusions can be drawn about how useful it is to add in x_0, and what kinds of initial mask tend to work best.

Nathan Merrill

if we define a composite prime to be a prime number where all substrings of the number starting at position 0 are prime

Peter Taylor

Without actually looking at it at all I would guess that you want fairly evenly spaced bits in the initial mask, but that is a thoroughly unscientific guess.

Nathan Merrill

then are there a finite number of composite primes in base 10?

(so 2 is prime, 23 is prime, 233 is prime, ect)?

Peter Taylor

@NathanMerrill That's a hideous name, but see A085823

Nathan Merrill

I know

couldn't come up with a better name

that has the restriction of all substrings. Mine has a looser restriction, namely that all substrings starting at 0

Peter Taylor

18:44

oeis.org/A024770

Nathan Merrill

how do you find these things

oO

Peter Taylor

oeis.org/…

Nathan Merrill

I didn't know they had a word search

I've always just used their number search

Peter Taylor

I've probably spent more time than is healthy on that site.

Nathan Merrill

lol

do they have some notation as to whether the sequence is finite?

Dennis

18:48

> The sequence ends at a(83) = 73939133

Nathan Merrill

I don't think so

I've found a bigger one

nevermind

Dennis

That would make a nice challenge IMHO.

Print all 83 right-truncatable primes.

Nathan Merrill

I was thinking about it

but I wanted to know if it was finite

that changes the format of the challenge

oooh, the list for suffix primes is much longer

but still finite

Martin Büttner

@mbomb007 thanks ;)

Nathan Merrill

357686312646216567629137

Dennis

18:53

4260 terms. Do you have a specific format in mind?

Nathan Merrill

would it be too tough to make a challenge where if you run the program forwards, it prints the nth right-truncatable prime, and backwards it prints the left ones?

Martin Büttner

that's a restriction that can be worked around quite easily in most languages

Nathan Merrill

let me sandbox

brb

Martin Büttner

that's the spirit

Peter Taylor

@Dennis I had a look because I thought I remembered a challenge along those lines, but there isn't one on main. I'm not sure whether one has previously been sandboxed.

Martin Büttner

18:55

@PeterTaylor I seem to recall something like that as well

ah, Alex posted something related: codegolf.stackexchange.com/q/55406/8478

that was probably what I remembered

Anonymous

I'd say that's different enough to not be a dupe

Anonymous

Considering that sequence is infinite :P

Martin Büttner

yeah definitely

Dennis

I answered that question and didn't remember it. :P

Martin Büttner

just saying it explains why I thought it sounded familiar (and might have been the reason why Peter thought it sounded familiar)

Dennis

18:58

I have a feeling you're going to win feersum's challenge.

Optimizer

@Dennis you are disappointing more and more

Dennis

@Optimizer This isn't news. Me memory has always sucked.

anOKsquirrel

yay my programming language is kind of working!

Dennis

\o/ (kinda)

TimmyD

@MartinBüttner Also vaguely related -- codegolf.stackexchange.com/q/59736/42963

Peter Taylor

19:06

On the subject of the sandbox, any comments?

Anonymous

@PeterTaylor I see no issues with it

Anonymous

thumbs-up

Anonymous

If this Seed brute-forcer finds a Truth Machine Seed program within my lifetime, I'll be ecstatic

Anonymous

All the rep will be mine :P

anOKsquirrel

wat

Optimizer

19:14

wat

Κριτικσι Λίθος

wat

Optimizer

I guess, you should not trust google's first result that much

TimmyD

Watt?

anOKsquirrel

Wheat?

Κριτικσι Λίθος

Cwarrot?

Optimizer

19:16

Nathan Merrill

0

A: Sandbox for Proposed Challenges

Right and tfeL truncatable primes A right-truncatable prime is a prime where every prefix is a prime (in base 10). A left-truncatable prime is exactly the opposite, where every postfix is a prime. Both of these sequences are finite (There are only 83 Right-truncatables, while there are 4260 Le...

Peter Taylor

@Mego I tried to find a Seed program for an earlier challenge by doing better than brute force.

Anonymous

@PeterTaylor How were you doing it?

Peter Taylor

The idea was to first write the Befunge, then generate a description to pass to a "satisfiability modulo theories" solver.

Anonymous

I understood a few of those words

Peter Taylor

19:22

Unfortunately, it seems that the Mersenne twister implementation used by Seed is not the one documented on Wikipedia.

New Sandboxed Posts

1

A: Sandbox for Proposed Challenges

Right and tfeL truncatable primes A right-truncatable prime is a prime where every prefix is a prime (in base 10). A left-truncatable prime is exactly the opposite, where every postfix is a prime. Both of these sequences are finite (There are only 83 Right-truncatables, while there are 4260 Le...

Nathan Merrill

oooh, I like the avatar for the sandbox

2

Anonymous

Python and C++11 both have built-in Mersenne twister algos

Martin Büttner

@Dennis I don't know. It can certainly be cracked.

Lembik

bounty awarded on codegolf.stackexchange.com/questions/62922/…

Dennis

19:26

We'll see. I was fairly certain mine wasn't going to make it trough 8 days...

Martin Büttner

@Dennis I did think of the line-wrapping string trick, but couldn't be bothered to try and crack

(plus, I was busy cracking B&C myself ^^)

I think if mine makes it through the week, it'll be because people have lost interest in the challenge.

Dennis

ShapeScript turned out far more usable than I anticipated. I managed to do a primality test in it.

Looks like the ShapeScript -> Changeling part deterred most robbers.

Rainbolt

19:54

I missed the drama :(

anOKsquirrel

yay codegolf.stackexchange.com/a/63151/46787

Wait, I already did JS darnit

Transcript for