The Nineteenth Byte

ATaco

02:00

I really threw myself off when I started learning German, and I started capitalizing every noun.

Mistah Figgins

@WheatWizard Hey, Same!

> Internet Protoco (IP)l

From a test online that I'm taking.... :/

My CS teacher does this kind of thing way too often

Downgoat

@MistahFiggins once my CS teacher gave us a test and forgot to put the question and the answers >_>

Destructible Lemon

his teacher?

Downgoat

Mistah Figgins

My CS teacher chronically moves assignments back

He moved one back 3 weeks because no one had bothered to do it

If he had just moved it back a day, everyone would temporarily freak out and get it done, but all he did was give everyone time to procrastinate

He also moved a unit test back 2 weeks

ATaco

02:08

So: What languages have people that are currently here written? Out of curiosity.

Mistah Figgins

@ATaco Del|m|t

Riker

@ATaco I wrote that fog one

the one that doesn't work 76% of the time for some of the funcs

Downgoat

@ATaco tastiest cheese ever

Wheat Wizard

Wise!

LegionMammal978

lol

I found a ham radio CW signal

But I don't know any Morse code :/

Riker

02:20

lol

LegionMammal978

Wait, no, it's NFM

Voice

Some redneck-sounding guy talking about "nets"

Riker

lol

LegionMammal978

idk what goes through ham radio operators' heads

Destructible Lemon

@ATaco Turtlèd

also Woefully

Mistah Figgins

@WheatWizard Cool! is input implicitly on the stack, and the stack implicitly outputs? And does it error on a dyad with only 1 thing on the stack?

Destructible Lemon

02:23

but that one is less good

Wheat Wizard

@MistahFiggins Yep, yep and no. There are implicit zeros on the stack

Mistah Figgins

alright, but I assume they aren't affected by ? or !

Dennis

@ATaco Help, Wardoq!, Bubblegum, ShapeScript/Changeling, Jelly/M, and Sesos.

Wheat Wizard

@MistahFiggins yeah

ATaco

Speaking of Changeling, We need another good "Write a language" challenge, perhaps one that doesn't require it to be unsolvably Turing complete.

Destructible Lemon

02:28

I think we have an out of site one

Wheat Wizard

@ATaco Hey

ATaco

Hey

Destructible Lemon

can't remember where it is

ah

docs.google.com/spreadsheets/d/…

Mistah Figgins

It's in my bookmarks, unsurprisingly

ATaco

Oh right, that.

I already wrote Threead for that.

Destructible Lemon

02:33

I haven't worked on mine yet

@ATaco do we really need a challenge?

just make one yourself

like: I will make a bunch of langs

ATaco

Not particularly.

But I would like one.

PhiNotPi

@DestructibleLemon lol @ phigs

Mendeleev

Speaking of langs

I need some help with Tomato

ATaco

What'dya need?

Mendeleev

BTW @Downgoat Your app crashes when you click on the pull requests tab of a repo

@ATaco I need to figure out how to implement the special operators that do something with the stack, and flow control

It's written in Ruby

Downgoat

02:37

hm, weird thought I fixed

Mendeleev

@Downgoat Have you tested with empty repo?

Dennis

> 1/1/2017 20:39:15 Downsheep Swisssheep Evil version of cheddargoat

>_>

Mistah Figgins

@WheatWizard Aw, It's not on TIO...

Wheat Wizard

@MistahFiggins I've used it in one answer once. There wouldn't be much of a point

Mistah Figgins

Hey, it wouldn't be too hard to use it in this challenge

Wheat Wizard

02:42

@MistahFiggins 6 bytes ::^~-&

Mistah Figgins

Would :><^ work?

Wheat Wizard

@MistahFiggins Yep I was just about to say that

PhiNotPi

Sometime I should (re-)start on Phigs.

Dennis

@MistahFiggins I like this language already.

Destructible Lemon

language design challenge: language where smileys aren't syntax but they appear often

Mistah Figgins

02:46

@WheatWizard can I post?

Wheat Wizard

@MistahFiggins Be my guest. I'm happy for someone else to use my language.

obarakon

https://i.imgur.com/QlGpd.gif

ATaco

With Credit, I rarely see answers in any of the languages I've written.

Then again, It's poorly documented.

Mistah Figgins

That is the main problem

@ATaco I would be using the hell out of TacO if I could get a command list

Wheat Wizard

I have the opposite problem; I love making specifications but I hate writing computer code

ATaco

02:50

Oh, that's not hard to write.

It's all in Functions.lua anyway.

Helka Homba

What's a good python ide?

Dennis

→ 8 messages moved to talk.tryitonline.net

Destructible Lemon

@HelkaHomba I use pycharm but I never did I big search for the best one

so that'd be a start

Downgoat

@Mendeleev I think it should work, I have very few implicitly unwrapped optionals

ATaco

github.com/TehFlaminTaco/TacO

Readme now contains a function list.

Good luck writing in a language that takes 29 bytes to do %2.

Helka Homba

03:03

@MistahFiggins Way to be Wise :p

Mendeleev

wait, so, for the golflang time capsule

Will people have to create a polyglot with all of these?

Seems impossible with e.g. WSF, Tomato, and SpaceChem

ATaco

Not all of them, but as many as possible.

Mendeleev

Oh, that's how scoring?

Mistah Figgins

@HelkaHomba yeh

Wheat Wizard

@MistahFiggins The TIO for wise is live now thanks to Dennis.

Mendeleev

03:09

Apparently my glue gun gets hot enough to extrude PLA O_o

Mistah Figgins

@WheatWizard Yay! Thanks @Dennis

Mendeleev

If only I had some 3mm PLA...

I could basically use it as one of those 3D printer pens. Sadly all I have is 1.75 mm

Mistah Figgins

@WheatWizard uh oh...

it prints 0 every time for me, no matter what

even the code :

Wheat Wizard

Odd

Mistah Figgins

Ah, command line arguments

Wheat Wizard

03:13

It does not do that for my local copy

Mistah Figgins

weird

Dennis

Yes, Input is STDIN.

Mistah Figgins

There should be inplicit input at the beginning of the program, correct @WheatWizard?

Wheat Wizard

@MistahFiggins yes

Mistah Figgins

AFAIK there is no "take input" command

Wheat Wizard

03:15

@Dennis would you like me to make the option to take input from a file?

Dennis

I'm piping input. I have no idea why that wouldn't work.

I don't think I've had this problem before.

Nathan Merrill

@Mego you here?

Anonymous

Yep

Anonymous

What's up?

Nathan Merrill

if Windows simply pipes commands to the processor, then wouldn't windows executables be trivially different to linux executables?

ATaco

03:24

I have implemented BigInteger scale Randomization for RProgN2.

Which was annoying, because BigIntegers don't like being random.

I'm rather proud of my system.

Anonymous

@NathanMerrill That's not what I said at all. Windows loads executables and libraries in memory for the processor to execute. The format of the executables differs between Windows and Linux because Windows and Linux load executables and libraries differently.

Nathan Merrill

I'm still not sure how those differences would make that big of a difference

but ok :)

like, you could have different libraries?

Anonymous

Different libraries, yes. Also different library formats, different linking methods, different memory layouts, etc

Nathan Merrill

so, if you compiled a program, one for linux, the other for windows, you'd essentially find large chunks of the program that are identical

chunks where you are doing a bunch of cpu calculations

Anonymous

Possibly, depends on the compiler(s) used

Nathan Merrill

03:31

right. Assuming the same compiler with the same flags :)

but even that depends on implementation

Anonymous

If you use gcc-mingw for Windows and gcc for Linux on the same machine, you'd probably end up with executables that are very similar, aside from the format

Mendeleev

@NathanMerrill Wine is actually mostly a shim that interprets Windows executable formats and translates library calls.

Nathan Merrill

doesn't that mean you could write some sort of program that converts exe files into linux executables?

Mendeleev

All the CPU instructions stay the same.

@NathanMerrill Wine.

Nathan Merrill

I don't mean Wine, because that's at runtime

Anonymous

03:32

For example, here's what a HW executable looks like on Linux: tio.run/nexus/…

Nathan Merrill

I mean you could "recompile" a program into linux

Anonymous

@NathanMerrill Theoretically, yes. Realistically, no.

Nathan Merrill

ok. good to know. Thanks!

Anonymous

The amount of time and work it would take to convert a Windows executable more complicated than HW to an ELF would greatly dwarf the time needed to go buy a second computer, download and install Linux, install gcc, copy the program over, and compile it.

Anonymous

Once Visual Studio finally finishes loading, I'll show you what a Windows HW exe looks like

Mendeleev

03:37

@Mego What's with the piles of zeros?

Anonymous

@Mendeleev I dunno. Probably padding for some reason.

Mendeleev

ehh

Anonymous

The best answer I can come up with is "because the compiler said so"

Anonymous

I'll use this for reference for Linux instead, because VC++

Anonymous

This is what it looks like when compiled with VC++

Anonymous

03:46

And for the sake of completeness, this is the output from gcc on cygwin

Mendeleev

@Mego Why is there XML embedded in your hello world executable

Anonymous

Because Windows

Anonymous

I swear I can hear a sigh of relief from my computer when I close Visual Studio

Mistah Figgins

@WheatWizard Time to answer a bunch of questions with Wise that were never meant to happen

Wheat Wizard

@MistahFiggins Good luck! Wise is pretty bad even at things that involve bitwise manipulations.

Mistah Figgins

03:50

So be it

Mendeleev

@Mego Why not use mingw and gcc?

ATaco

Also Mistah, the TacO readMe now has a command list.

There aren't a lot of commands.

Mistah Figgins

@ATaco Yay!

ATaco

HMU if there's anything that doesn't make sense.

Mistah Figgins

Alright, will do

Wheat Wizard

03:53

@MistahFiggins By the way here is addition in Wise. Its about the most complicated program I've ever written in Wise.

Mistah Figgins

Thanks!

ATaco

This is the most complicated thing I've written in TacO.

I'm just amazed it actually works.

Anonymous

@Mendeleev I don't have mingw installed and I don't care to install it for only this

Mendeleev

@Mego Oh

Destructible Lemon

04:16

hey I found my bones file with the amulet of life saving!

died

goddamn hobbits

Anonymous

@DestructibleLemon You should really consider whether or not your messages will be relevant or interesting to anyone else before sending them. TNB is not a place to dump your stream of consciousness.

2

Mistah Figgins

@WheatWizard What's the easiest way to subtract 1?

Wheat Wizard

-~

Mistah Figgins

oh yea

Destructible Lemon

@Mego people seemed kinda interested when I posted the part when I died with the bones file character

Mistah Figgins

04:31

Still, a little context would help. Nethack?

@WheatWizard Is there a quine yet?

Wheat Wizard

Nope, I started but I never got around to finishing it

Riker

04:49

@Mego fwiw @DestructibleLemon I play nethack daily, and I still get tired of the chatter. TNB just isn't the right place, sorry man. No offense intended.

Mistah Figgins

Nethack

Mistah Figgins

05:08

@WheatWizard Multiplication: [?:[?~-!-~]|!-~]|

Wait no

Is there a way to switch the top 2 elements?

Wheat Wizard

05:26

@MistahFiggins well this works but its really long

Mistah Figgins

nice

Wheat Wizard

OH I'm dumb, this can be done real easily

@MistahFiggins Here's one that isn't utter trash

Mistah Figgins

That's a lot better

Wheat Wizard

Another one that is the exact same length

In fact they are anagrams

Mistah Figgins

"?!" -> ""

Wheat Wizard

05:38

Good point

Mistah Figgins

Maybe we/you should compile some of these on the github page

Wheat Wizard

Good idea. I'll start the wiki

@MistahFiggins Here is the wiki. Feel free to contribute.

mniip

05:54

I should make a golflang someday. Perl no longer does its thing in modern codegolf questions

Mistah Figgins

@WheatWizard How do I contribute to the wiki?

mniip

"Negativize" that's called negate

DJMcMayhem

06:13

@MistahFiggins you need to be a collaborator I think

Qwerp-Derp

@WheatWizard How does - work on a stack consisting of 1010?

Also BTW Wise looks really cool :)

Mendeleev

@ГригорийПерельман This animation is amazing youtu.be/ddJiP3jRVJ8

Music is cool too

ATaco

06:46

10 imaginary internet points if someone can prove TacO TC or not

Jan Dvorak

Humans are bounded storage machines, and certainly not Turing complete

ATaco

:/

Qwerp-Derp

Yey I made a Wise multiplication function

tio.run/nexus/wise#@2@vW2dvpcgVzaWgYG9lDySj7et0FXXrYmtAIrp1XLH//…

More clean version: tio.run/nexus/…

Григорий Перельман

07:08

@ATaco can you make an infinite loop yet?

ATaco

Nope, but that's not required by the definition of Turing complete

A program must be able to run for an arbitrarily long time that is not limited

New Main Posts

0

Q: Decide what option to choose

This is not code golf! Simple and clear answers win :) I have a program that needs to decide what Check to do. I want to make it as efficient as possible, but readability is more important. Each check has a countdown to 0. When the countdown hits 0, the check is due. Rules: A-Checks should...

Jan Dvorak

@ATaco isn't that the same as infinite?

Григорий Перельман

Yes it is required. A TC language must not always halt.

ATaco

A TC program must be able to solve any Computable Algorithm

Then again, the official definition is, "able to simulate a single tape Turing machine"

TidB

07:14

@NewMainPosts Is this some kind of elaborate "program this for me"?

Jan Dvorak

If the halting problem is solvable in your language, it's not Turing complete. The halting problem is trivial if your language can't even loop forever.

Григорий Перельман

@ATaco ...which doesn't always halt.

Jan Dvorak

@TidB seems to me. Voted too broad. I admire the audacity, though.

Besides, "clean and readable code" is not an objective winning criterion.

Григорий Перельман

I voted off topic: No objective winning criteria.

Riker

08:13

@ГригорийПерельман midnight dark souls was a really bad idea

Now I really can't sleep :p

Next area tho

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Efficiently find the median Background Computer scientists have spent a long time looking into ways of sorting data faster. One of the known discoveries is that if you can use the actual values of the data, sorting can be faster than if you can only compare them. Finding the median of a list i...

Riker

Paused rn at bonfire

Destructible Lemon

something something idle chatter

something something stream of conscioussness

Lembik

@PhiNotPi Could you let me know some numbers? I would be interested in even unspectacular figures :)

betseg

08:43

@Riker did my message on the APL room ping you?

Qwerp-Derp

Does anyone here know VBA?

New Main Posts

09:46

0

Q: Which libraries for testing pseudo-random numbers generator?

For research purposes we are coding a generator of pseudo-random numbers in a given range. Do you know of any recent tool/library for testing such sequences of numbers?

number random statistics

Luis Mendo

10:15

@DJMcMayhem This one is digital, FM-based. FM synthesis has a name for being difficult, but you can achieve great results

@Riker In Spain I have sometimes heard that, but it's very uncommon. And it's always pronounced Spanish-like (softer "g" than in English)

New Meta Posts

0

Q: Default I/O for audio?

I am about to write a challenge about Text To Speech and 'Hello, World!'. But it would be crucial to know what I/O types works. Could you construct the consensus for it?

discussion defaults

Lynn

10:40

@xnor Sometimes I give up and try to use every single thing in Prelude. ^^;

shows a b == show a++b is probably worth being in Tips. I’ve seen it before in old solutions of henkma’s.

Qwerp-Derp

11:09

I'm really proud of my Wise thing :)

New Main Posts

12:10

0

Q: sum of 2<sup>Pi</sup> mod 1000000007 for all i where P<sub>i</sub> is sum of numbers in ith subset of a set X

I am stuck on a problem in which I have to print sum of 2Pi mod 1000000007 for all i where Pi is sum of numbers in ith subset of a set X. Length of set can be upto 100000. Value of element in the range [0,1012]. Here's the link of the Problem. Problem Statement I could not find any approach ...

math combinatorics set-theory

PhiNotPi

12:31

@Lembik for example, I conjecture that

  1  0  1  0  0
  1  1  0  1  0
  0  1  1  0  1
  1  0  1  1  0

with a ratio of 1.25 is the largest ratio achievable for a matrix in which each column differs from the previous column by the same amount (in this case, 3 bits are flipped and 1 bit stays the same).

I've also looked at matrices where the last several bits of the first column, and the first several bits of the last column, are 0s.

  1  0  1  1  1  0  0  0  0
  1  1  0  1  1  1  0  0  0
  0  1  1  0  1  1  1  0  0
  0  0  1  1  0  1  1  1  0
  0  0  0  1  1  0  1  1  1
  0  0  0  0  1  1  0  1  1

(ratio 1.5, best I've found although I didn't search too hard)

Whenever I set additional restrictions, the ratios decrease dramatically, in some cases, they appear to peak early on and then decrease for larger column sizes.

I've also looked at symmetrical matrices that are near-URAs, best ratio I've found is:

13:
  0  0  0  1  1  0  1  0  1  1  0  0  0
  5  2  2  2  1  2  1  1  2  1  2  2  2
 1.3333333333333333  1  7
  0  1  0  1  1  0  0  0
  1  0  1  0  1  1  0  0
  1  1  0  1  0  1  1  0
  0  1  1  0  1  0  1  1
  0  0  1  1  0  1  0  1
  0  0  0  1  1  0  1  0

Followed in second place by this much larger matrix:

29:
  0  1  0  0  0  1  0  1  1  1  1  0  0  1  1  1  0  0  1  1  1  1  0  1  0  0  0  1  0
 15  8  7  7  8  8  8  8  7  8  7  7  7  8  7  7  8  7  7  7  8  7  8  8  8  8  7  7  8
 1.3076923076923077  1  27
  0  1  1  1  0  0  1  1  1  1  0  1  0  0  0  1  0
  0  0  1  1  1  0  0  1  1  1  1  0  1  0  0  0  1
  1  0  0  1  1  1  0  0  1  1  1  1  0  1  0  0  0
  1  1  0  0  1  1  1  0  0  1  1  1  1  0  1  0  0
  1  1  1  0  0  1  1  1  0  0  1  1  1  1  0  1  0
  1  1  1  1  0  0  1  1  1  0  0  1  1  1  1  0  1

Anonymous

12:47

@DestructibleLemon There's a significant difference between directly addressing one user and referencing a previous conversation, and simply outputting whatever happens to be at the forefront of your mind into chat. Making specific complaints would be far more useful than petulant passive-aggressive messages.

Poke

12:59

Fatalize

@NathanMerrill (discussing what you mean by encoding input is probably easier here than in the comments)

I assume you're asking about how I construct the feature vector given to the MLP? In that case yes I should probably explain that (though it's a pretty direct way)

Nathan Merrill

sure? Its been long enough since I've used ML that my terminology is gone :)

you have a bunch of input nodes

I'm asking how you convert the test case into the input nodes

Fatalize

Yeah yeah that's it

I'm writing it up but basically I construct a 76 elems vector

where the first 64 elems are the cells and the last 12 are each type of piece

Nathan Merrill

also, can say how odd it is that I've gotten 3 answers in the past hour, despite getting no activity for basically the first 24 hours :P

(I'm exaggerating a bit, but I went to bed with the question unanswered, and woke up to a bunch of activity, so it feels that way)

Fatalize

Well it takes more times than other challenges to tackle it

It's the same as the trivial challenges (they get all initial answers at the same time), but with a half a day overhead :p

Nathan Merrill

13:09

lol

@Fatalize oh, also, as a tip: did you see in my question that I'm using 10% of all board states?

I'd highly recommend training your NN on the first 80% of the data (of each color)

as opposed to a random 80%

because some test cases use the same game

Fatalize

I'm not sure I understand why that would change much

Nathan Merrill

It may be identifying the games

Fatalize

Boards from the same game are consecutive in the data?

Nathan Merrill

yes

I intentionally made it that way for this very reason

Fatalize

I'll see about that

Jan Dvorak

13:20

CMC: When your input is 2, output "xy". When your input is 3, output "xyz". When your input is 4, output "wxyz". You may use all-uppercase instead.

Nathan Merrill

I also would not be opposed to giving out my generation program so you can generate your own

orlp

13:34

@isaacg hey, did you see my efficient computation algorithm? did you have a similar one when you made the challenge?

Anonymous

@JanDvorak Python, 36 bytes: lambda x:"wxyz"[x*(x<4)-x+(x<4):x+1] (there's got to be a better way)

Fatalize

@NathanMerrill It does seem to make a significant difference (for the worse), but since the dataset is already fairly small I'm not sure whether this is due to the fact that the first games are less representative than others or if it really mostly learns the games sequence

How many different games are there in total in the data?

(I get about 91.73% on all data if I use 80% for training and 63.74% on the remaining 20% of test data when I select data in order instead of randomly)

Nathan Merrill

@Fatalize there are 2600 games

Fatalize

(compared to 94.5% and 77.08% with random selection)

Nathan Merrill

oh, wait

my 2600 is wrong

because that includes ties

1354 white wins vs 760 black wins

Fatalize

13:41

So roughly 3 boards per game

orlp

this goes too far IMO

half my screen is filled with this obnoxious ad

Fatalize

(Also is that intentional that there are more white wins?)

Nathan Merrill

no, that's just the data

Fatalize

A naïve algorithm would achieve better than average by always saying white wins

orlp

@Fatalize white has an advantage is chess

Nathan Merrill

13:43

I was honestly surprised when I found that

orlp

that's well-known

that's why the stronger player plays black

Nathan Merrill

but 1354 vs 760?

that's almost double the wins

orlp

oh that's an anomaly for sure

Nathan Merrill

ok, so in my second set of data, I have 3946 white wins vs 2696 black wins

so, not as drastic, but still pretty drastic

Fatalize

@NathanMerrill You can extend your edit to "If you're using a machine learning model" rather than a NN, because that "problem" would stay true for other algorithms like SVMs

It's not that much of a problem that there are significantly more white wins than black as long as the test cases you use have a similar distribution to this

If you test answers on a set that has significantly more black wins than white wins then that's a bit of a problem because that means the original set you provide is not representative of the test data

PhiNotPi

13:50

white vs. black matters much more for high-level players

Fatalize

60% is a pretty ridiculous advantage

PhiNotPi

(note y-axis scale)

Fatalize

Has there never been attempts to balance that back like in Go?

e.g. allow two pawn moves for black, or forbid white from moving his first pawn 2 cells as first move?

I guess it might balance it too much the other way around

orlp

I mean

consider that in high level most games end in ties

Fatalize

But what you suggest is that if it doesn't, then most games end with white winning

Wheat Wizard

14:04

@MistahFiggins You can clone it on the side and make a pull request.

Nathan Merrill

@Fatalize I'm going to add a bunch more test cases

ok, I found a huge database. It apparently contains 2.2 million games

Fatalize

That might too extreme in that case, I don't want to have my PC working for 2 weeks :p

Nathan Merrill

oh, I won't use it for my final test case

but if you don't want that much data, then don't use it :P

oh yeah, this file is huge

1.47G

Downgoat

14:28

Oh no ;_; 2 year PPCG consecutive streak broke :'(

Fatalize

You sanity is saved

New Meta Posts

1

Q: Can a language's True/False be used in place of their values in a sequence?

I'm asking this to find a consensus on whether it's okay for an answer to use True or False in place of the language's corresponding integer values -- as output in a sequence, for example. So if the challenge is to output the Fibonacci sequence forever, and your output is False, True, True, 2, ...

discussion policy

Nathan Merrill

I've processed 16% of all games :)

New Main Posts

14:45

0

Q: I need help to divide my program into threads and assign threads to processors i don't know codes and also other stuff in discription

created a large array of random integers and calculated sum and average of odd integers between 25 and 75 (inclusive). So far the program is running but the result its giving is too long in other words its showing every addition and average step individually how can i displayed the result after a...

array-manipulation java multi-threading

Downgoat

15:02

@PhiNotPi what's the white line in the middle?

Wheat Wizard

Can one actually earn epic on the meta?

PhiNotPi

@Downgoat that's the measured value, grey is a confidence interval of some kind.

DJMcMayhem

@WheatWizard I've gotten mortarboard, so I assume it's possible. Just not easy

Riker

15:33

@LuisMendo huh, cool

Riker

15:57

is there an easy way to get the favicon for a site?

Anonymous

It's usually http(s)://root.url/favicon.ico, but not always

Riker

hm, thanks

that works

Anonymous

So for PPCG it's https://codegolf.stackexchange.com/favicon.ico

Anonymous

Which actually redirects elsewhere, but close enough :P

Nethack

Transcript for