The Nineteenth Byte

 1st  528 Sp3000           100 10:59:01
 2nd  599 MarsUltor        100 11:54:49
 3rd 1045 DennisMitchell   100 18:05:39
 4th 1148 Jakube           100 19:16:05
 5th 1182 Sherlock9        100 19:41:32
 6th 1275 aditsu           100 20:47:42
 7th 1287 TheNumberOne     100 21:00:44
 8th 1722 mbue              92 20:15:25
 9th 1932 BrainSteel        80 18:49:58
10th 2481 xsot              75 10:20:40
11th 4320 Maltysen          75 26:57:47
12th 4551 NathanMerrill     70 26:05:47
13th 6635 TheOnlyMego       55  8:05:12

Lembik

Sp3000 ! :)

aditsu

8:09 AM

huh, it had no stars

Lembik

can you remind me how it works... if you answer A-D do you get to the next round?

zyabin101

@Lembik You have to earn 30 points by solving problems.

aditsu

each problem has 2 cases and each case has a number of points; total 100

Lembik

oh ok.. so 30 points is not so hard to get

aditsu

you just need to solve enough to get 30 points in this round

Anonymous

8:10 AM

@Lembik You need 30 points in the qualification round. Each problem has a small version and a large version. Different problems/versions have different points.

Lembik

where was that lovely table that listed number of answers per language?

aditsu

very ninja'd :p

Lembik

@Mego thanks

@aditsu thanks

aditsu

@Lembik go-hero.net/jam/16/languages

Agawa001

C++ rocks...lol

Lembik

8:11 AM

there was one that broke it down by problem

aditsu

if you click a language, you get more details

Anonymous

7/8/10/10/10/20/10/25 were the point values

Lembik

ok thanks

aditsu

@Lembik ah there's also this: go-hero.net/jam/16/round/0

Lembik

so am I right that getting 30 points is not so hard

Anonymous

8:12 AM

See this for details

Anonymous

Yeah solving all the smalls is enough for 30 points, always

aditsu

sure

Anonymous

And the smalls can almost always be brute-forced

Lembik

I see that python people are cleverer than C++ ones :) 609/5533 got perfect

Anonymous

Nah the Python people didn't have to spend as much time chasing down segfaults

Lembik

8:14 AM

@Mego right!

which is a good advert for python imho

Anonymous

I could've gotten 90/100 had I solved D, but I didn't attempt it

Solver

Will it open again at the same time next year?

Lembik

bear in mind I am the person who posed codegolf.stackexchange.com/questions/26323/… :)

Martin Büttner

@Solver roughly

Anonymous

@Solver Roughly the same time

Lembik

8:15 AM

oh is the qualification round closed now?

Anonymous

In 2012 I think it was in February/March

Anonymous

As of 6 hours ago yeah

Solver

@MartinBüttner Good. I'll put a reminder on my calendar.

Lembik

@Mego that's quick!

what happens if you had a dentist appointment that day? :)

Martin Büttner

tough luck

aditsu

8:16 AM

it was open for 27 hours

Lembik

it's google or the root canal

Martin Büttner

the other rounds are only open for 2.5 or so

although there are 3 dates (with different times) for round 1

Lembik

ok what happens next?

and when does it get hard :)

Martin Büttner

rounds 1A, 1B, 1C

aditsu

yeah you have 3 chances for round 1; for the other ones you have to plan in advance

Anonymous

8:17 AM

Round 1 is usually a pretty steep step-up in difficulty from qualifications

aditsu

round 2 gets hard

Martin Büttner

@Mego yeah, at least it was last year

Anonymous

And you're also competing against the clock to be one of the top 1k

Lembik

thanks

actually the haskell people did very well I see

Agawa001

dufff 27170 contestants !!

Lembik

8:18 AM

probably because they are just cleverer :)

Anonymous

Qualifications are easy, Round 1 is medium, Round 2 is hard, and I've never made it past 2

Lembik

is there a prize at some point?

Anonymous

Yep

Lembik

or is it just glory?

Anonymous

$15k is top prize, minimum prize for making it to finals is $100

aditsu

8:18 AM

top 1000 in round 2 get t-shirts

Anonymous

Also t-shirts, the most important prize

Lembik

got you

aditsu

I have a collection ^_^

Lembik

so basically not much for the amount of work

Anonymous

Also finalists get noticed by Google and are often solicited for employment opportunities

Agawa001

8:20 AM

what a glory, to wear a tshirt from google company and boost 'bout it next to ur fellows

Lembik

@Mego that always seem a little funny to me to be honest. It suggest that had you just applied their system would not have been good enough to select you

aditsu

the bigger prize for making it to finals is flights + accommodation :p

Agawa001

really seems nicer then a handful of cash

Lembik

who needs money after all ;)

Anonymous

@Lembik Google gets mind-boggling amounts of applications daily. You have to really stand out in order to even be noticed.

Lembik

8:21 AM

@Mego I can imagine

(numbers of applications.. :) )

Anonymous

Even internships at Google are super-competitive

Anonymous

One of my classmates from university was despised by our entire Software Engineering class because they got an internship

Lembik

on another note.. what is the ppcg policy on machine learning type challenges?

Agawa001

what is the deadline for the competition?

Lembik

sort of mini kaggle challenges

Anonymous

8:22 AM

@Lembik Go for it

Anonymous

If you can make it objective and clear, there's no reason not to

Lembik

@Mego I sort of assumed people would object on the grounds that they are tough to do in lots of languages

thanks

aditsu

17 mins ago, by Sherlock9

Ah yes. So 6 hours ago

Anonymous

58

Q: Build a working game of Tetris in Conway's Game of Life

Here is a theoretical question - one that doesn't afford an easy answer in any case, not even the trivial one. In Conway's Game of Life, there exist constructs such as the metapixel which allow the Game of Life to simulate any other Game-of-Life rule system as well. In addition, it is known that...

Lembik

has there ever been a successful one?

one feature of kaggle type challenges is that some data is held back until testing

Anonymous

8:23 AM

Tough challenges are allowed :)

Lembik

@Mego :)

Martin Büttner

do you consider genetic algorithms machine learning? :P

Anonymous

We've had challenges where the test cases used for scoring were not revealed until after a winner was chosen

2

Lembik

@MartinBüttner if they work better than stochastic hill climbing. Otherwise I call them fraud :)

( I have gone star mad this morning :) )

Anonymous

Nowadays the preferred method is to ban optimizing for the test cases, and reserve the right to change the test cases if necessary

Lembik

8:24 AM

@MartinBüttner I am a genetic algorithms sceptic (denier?)

@Mego that's very interesting.. thanks

Martin Büttner

well we had codegolf.stackexchange.com/q/44707/8478

Lembik

@MartinBüttner Oh I saw that. It was great!

ah and I see the author.. no surprise then :)

Anonymous

But as long as the challenge is clear and objective, you don't have to reveal the scoring cases immediately (just as long as they're available at some point for verifying scores). Even test cases aren't required (though highly recommended).

Martin Büttner

@Lembik well, as the credits show it was a huge collaborative effort

Lembik

@MartinBüttner true

@Mego thank you this is very helpful

I do have one complaint.. or query. codegolf.stackexchange.com/a/77212/9206 is completely awesome. Is it not so upvoted due to the lack of explanation?

Anonymous

8:27 AM

> 84,465 specimens were harmed in the making of this challenge.

Sp3000

@aditsu I'm amused that there was actually a submission in HSP (Hot Soup Processor)

Agawa001

@aditsu oh no! wud is the next set-up

Anonymous

@Lembik That's the main reason why I haven't upvoted it

aditsu

@Agawa001 you mean next chance to compete? next year

Martin Büttner

@Lembik and due to lack of insight into the actual problem.

Lembik

8:28 AM

I guess the lack of insight is the same as the lack of explanation?

Anonymous

The answer is just "here's the code", no attempt at explaining the process that led to that code, or why the code works, etc.

Lembik

I mean Ton Hospel has had a lot of insight to write the code

Martin Büttner

yeah but I don't :P

Mars Ultor

@Mego Unless stated otherwise (e.g. the dictionary hash challenge)

Lembik

I wonder if he can be persuaded to write any explanation. I tried to ask him but he seems reluctant

Martin Büttner

8:28 AM

I mean even if there was an explanation for the answer, if I don't have a good grasp on the actual problem, I'm not sure how much it's going to help

Lembik

which is confusing because he has explained it more here on chat!

Agawa001

not such a big deal, got time for training :)

Lembik

@MartinBüttner I feel that a good explanation of the answer would help explain the problem too

Anonymous

I also have a poor grasp on the problem itself, other than knowing what the formula means

Lembik

@Mego I wonder what I can do to help

@Mego do you know basic linear algebra?

Anonymous

8:30 AM

Yes, I do

Anonymous

My minor is in mathematics

Solver

@Lembik Meh. Calculus!

Anonymous

I would suggest giving an overview of how the formula relates to the Riemann Theta function

aditsu

@Mego actually I have some bad news about that

Lembik

ok.. I think he has basically implemented page.math.tu-berlin.de/~bobenko/papers/2003_Bob_DHvHS.pdf with some optimizations of his own

@Mego ok let me add something now

aditsu

8:31 AM

a bunch of people wrongly submitted the input for B as the code

2

Anonymous

@aditsu Oh lol

Martin Büttner

:D

Anonymous

And it looks like brainfuck, so... That's hilarious

aditsu

yep

Anonymous

Next year I'm going to submit Seriously and Brainfuck answers :P

Sherlock9

8:34 AM

I should really start learning a Lisp dialect soon

Or Haskell

I just want to try functional programming :D

aditsu

I tried Haskell this time (started reading during the qual)

I'll probably check out Scheme later

Lembik

@Mego updated the question

@MartinBüttner ^^

I am not sure how helpful my update is

any suggestions for how I can make it more helpful gratefully received

Sherlock9

@KennyLau I may take a look at Codeforces soon. Thanks for the idea :D

Kenny Lau

@Sherlock9 np

@Mego stackoverflow.com/a/3801917

Anonymous

@Lembik Please reserve stars for things that are actually important/interesting/amusing enough to be worthy of a spot on the starboard.

Kenny Lau

8:42 AM

@Mego "There is no way in Java to get available operators."

Anonymous

@KennyLau I didn't mean calling the operators. I meant changing the mapping between the operators and the opcodes in the compiled byte code.

Kenny Lau

@Mego How would you actually do that?

I mean in Java

Anonymous

I'm not sure

Solver

Anyway, in the end, I'm joining Topcoder.

Kenny Lau

alright

Anonymous

8:43 AM

There's a Java answer on the 2+2=5 challenge that does something similar

Kenny Lau

Anyone here familiar with reflection?

@Mego That one changed 4 to 5

aditsu

I know some reflection

Kenny Lau

@aditsu Do you know how to do that?

Solver

Yeah, I do. Just take a mirror.

aditsu

which "that"? :)

Kenny Lau

8:44 AM

@aditsu change the opcode of +

aditsu

I don't think you can do it via reflection

you use reflection to access/invoke fields and methods

Anonymous

You might be able to get more help in another room:

Anonymous

The Whiteboard

General Discussion for softwareengineering.stackexchange.com P...

Neil

the way I know is to change the integer autoboxing cache so that the cached Integer for 4 is Integer(5).

aditsu

yeah that cache is an array which is a field in a class

Kenny Lau

8:46 AM

@aditsu codegolf.stackexchange.com/a/28818/48934

but is there a way to change +?

@Mego That room has nobody

zyabin101

> Pants optional when telecommuting.

lol

aditsu

@KennyLau well, if you generate bytecode yourself, you can do that

Kenny Lau

@aditsu What do you mean?

Anonymous

You could do it a simple way: get source, string replace + with something else, run modified source

Anonymous

And use a flag that would have different values depending on whether or not + has been changed to decide whether or not to do the modification

zyabin101

8:48 AM

RAPTFTD

aditsu

you could take a class, either the source or already compiled, and generate/fiddle with the bytecode for it at runtime

there are some libraries for that kind of stuff

zyabin101

Dead simple interactive Python version management

Anonymous

@aditsu My original idea was along those lines

Anonymous

I know it can be done, but I don't know the specifics on how to do it

aditsu

same

probably using asm.ow2.org

Lembik

8:50 AM

@Mego ok.. did you get a chance to look at my update?

Kenny Lau

@aditsu So it isn't official?

aditsu

what do you mean?

Kenny Lau

Does ASM use the Java libraries?

is ASM in Java ?

aditsu

it probably uses some java libraries

it's written in java afaik

Lembik

have we had challenges like codegolf.stackexchange.com/questions/47912/… for Euler's constant?

aditsu

8:53 AM

that's my challenge :)

Kenny Lau

@aditsu From what I understand it is separate from Java?

Anonymous

@Lembik I looked at it, but I'm too tired to parse it

aditsu

@KennyLau well, it's a 3rd party library

Lembik

@Mego oh... if I can do anything to make it clearer please do let me know

or maybe you just need a rest :)

Kenny Lau

@Lembik Would it be a duplicate?

Anonymous

8:54 AM

6 hours of sleep and 2 cups of coffee afterward should do the trick :)

Kenny Lau

@aditsu From what I understand, you compile a .java to a .class, and then run the ASM instead of the Java?

aditsu

@KennyLau no, you run java and your program calls some methods from the ASM library to create/modify the bytecode for a class

Anonymous

Why not just tinker with a class's compiled byte code and (re-)load the class using reflection?

Anonymous

Oh that's what you're talking about

Anonymous

Disregard :)

Kenny Lau

8:56 AM

@aditsu Then basically I can do it without the ASM?

Solver

wait, what? reflection = mirrors

aditsu

@KennyLau if you can reimplement the bytecode manipulation stuff, then yeah I think it's possible

Kenny Lau

@aditsu wait..... a class modifying itself?

aditsu

I'm not sure it can modify itself (maybe not), but it could at least modify another class

you can load a class from an array of bytes (and it doesn't involve reflection)

Anonymous

Your class could easily modify itself. Load its own byte code, make changes, change the name and the filename, load it as another class.

Anonymous

9:01 AM

It couldn't be the class with the main method though, because it would have to be out of the call stack in order to be unloaded (I think), unless you're fine with not unloading the old version

Solver

Hello, @Sp3000

Kenny Lau

@aditsu What is reflection?

Lembik

@KennyLau sorry... would what be a duplicate?

Mars Ultor

@KennyLau Looking at source code, or bytecode

Lembik

@Mego :)

Solver

9:05 AM

@KennyLau It's - mirrors

aditsu

@KennyLau an API to access classes as objects, and their fields and methods (by name and type)

Kenny Lau

@Lembik just to change π to e

Lembik

@KennyLau I not sure it would. The math is completely different as are the algorithms

but other people may not agree of course

Kenny Lau

@Lembik Well, related...

Lembik

@KennyLau i am not sure how related it would be

except they are both constants :)

Mars Ultor

9:07 AM

@Mego there's also a Piet submission

Kenny Lau

@Lembik Well, ok...

Lembik

but as I say.. other people may disagree

@MartinBüttner what do you think?

Kenny Lau

@aditsu And how to change final objects?

Mars Ultor

the number of 100 point python contestants are only slightly less than that of C++

Kenny Lau

**Testing** Testing

Lembik

9:08 AM

@MarsUltor that is very impressive

aditsu

@KennyLau you can override them using reflection

Lembik

not so many in R I see

Kenny Lau

@aditsu Sorry I'm new to reflection, could you teach me?

Martin Büttner

@Lembik context?

Lembik

@MartinBüttner challenges to compute e (Euler's constant) to high precision along the lines of previous Pi ones

aditsu

9:09 AM

@KennyLau take a look at this and similar questions and answers

Mars Ultor

@Mego it would be much easier to create a Brainfuck macro system, then use that

Martin Büttner

I don't know enough about the matter to know if answers would be essentially identical or completely different

Ton Hospel

@lembik: Hi

Lembik

@TonHospel Hi! How things?

@MartinBüttner OK

aditsu

@KennyLau in particular, you can change the modifiers of the field (at runtime) so that it's no longer final

Kenny Lau

9:11 AM

@aditsu Could I also change private to public?

aditsu

@KennyLau maybe, but you don't need to, you just force-make it accessible

zyabin101

:O

Kenny Lau

@aditsu So I could basically set up a BigInteger by giving it an integer array...

Ton Hospel

Calculating e is easy. `perl -e 's!\w+!$%=($z=$&.0+$%)/($a=--$?||10);$z%$a!eg,print$%for(-1x65536)x1e3' (wait long enough for the result)

2

Should print 1000 digits

aditsu

@KennyLau yeah, looks like it has some non-public constructors for that, you could call them via reflection; the issue is they may be jvm-implementation-specific

same with the "mag" field

Kenny Lau

9:15 AM

@aditsu Can I hack into SecurityManager?

Ton Hospel

(uses a spigot algorithm for e which is easier than that for pi)

zyabin101

@TonHospel Is the 1e3 changeable to print a specified number of digits?

aditsu

@KennyLau a security manager may prevent you from doing this kind of things; if not, then you also have no need to hack into it

zyabin101

For example, 500 instead of 1000. Or 100.

Kenny Lau

@aditsu If there's a security manager, can I hack into it?

Lembik

9:16 AM

@TonHospel awseome perl!

aditsu

@KennyLau I don't think so, otherwise it would be pointless :p

Lembik

@TonHospel about my challenge.. I am excited to know if you have any more plans

zyabin101

@Lembik github.com/hachiojipm/awesome-perl

Ton Hospel

@zyabin101: Yes, though it uses a dirty trick with $? to limit the number of loops, so that can't be extended indefinitely. Fixing that will make the code slightly longer but a lot faster since now it does way too many loops

Lembik

@TonHospel you mentioned improving the library usage already

en.wikipedia.org/wiki/Feigenbaum_constants might be more fun to compute....

@zyabin101 thanks

Kenny Lau

9:18 AM

@aditsu What kind of security issue would this lead to?

aditsu

@KennyLau don't know a lot about that, but an example would probably be a web application hacking into another one running on the same server

Ton Hospel

Still this e code was a side joke I came up with when doing a series for pi. It could be fun to seriously golf it some more and generalize it

Kenny Lau

@aditsu Could you elaborate?

Anonymous

If a SecurityManager locks you out of doing reflection stuff, there's not really anything you can do to get around it

Ton Hospel

@lembik: I'm looking at the siegel modular transforms but I don't understand 1 of the transforms, the one with diag b. And that one is very important because until now i was able to keep z=0 and I want to keep it like that. But i find very little about the modular group for the riemann theta function on the net (hoping to find a source that derives the formulas so I understand what is going on and if I can fix it)

aditsu

9:23 AM

@KennyLau you could have a java application server / servlet container running webapps developed by different people; they are normally separated via different classloaders, but with unrestricted reflection you might be able to get past that barrier

Kenny Lau

@aditsu Which is why java disabled virtually every application online?

Mars Ultor

@KennyLau ?

aditsu

wat?

Kenny Lau

@MarsUltor most java applications online are disabled...

Optimizer

but it broke my website where i used to run JavaScript in Rhino engine in a Java servlet container...

Kenny Lau

9:26 AM

due to security issues

Mars Ultor

@KennyLau That's just your Java security settings then

Optimizer

How else am I supposed to run JavaScript on a browser?

Kenny Lau

well...

java applet

not javascript

aditsu

@KennyLau if you're talking about most shared hosting providers not supporting java, it's probably because it's difficult to get it right, and now VPS's are cheap and much better

Optimizer

@MarsUltor there are things called exploits common to both flash and java applets

aditsu

9:28 AM

@KennyLau applets have nothing to do with web applications

applets run in the browser, webapps run on the server

Kenny Lau

alright

aditsu

applets are a different can of worms :p

Anonymous

Solution: don't use Java

Solver

^^^

Anonymous

Java was designed to do many things, but securely run client-side code is not one of them

aditsu

9:31 AM

well, java web start is still fine

Mars Ultor

D: go-hero doesn't detect Retina

aditsu

speaking of calculating e, if you find a way to compute (1+9^(-4^(7·6)))^(3^(2^85)), you get 18457734525360901453873570 accurate digits

Kenny Lau

Optimizer

java applets?

Anonymous

PHP?

Kenny Lau

9:36 AM

@aditsu Not even wolfram alpha

New Main Posts

0

Q: Virtual Keyboard Text Input

On modern game consoles and other devices without traditional keyboards, trying to input text is a nightmare. Having to type with a few buttons and a joystick on a virtual keyboard is annoying, and I like to make as few movements/button presses as possible. The keyboard you will be using looks l...

string code-challenge keyboard

aditsu

@KennyLau it's listed at mathworld.wolfram.com/eApproximations.html btw

Kenny Lau

@aditsu wolframalpha.com/input/…

aditsu

@MarsUltor it will probably be fixed within the next several hours / couple of days

Mars Ultor

@aditsu How are you so sure?

Kenny Lau

9:41 AM

@aditsu I am not so amazed. It is basically (1+1/n)^n, n->infty

aditsu

@MarsUltor in previous years foxlit went through the unknown submissions and updated them (sometimes with help from other users)

Mars Ultor

@aditsu How do they have access?

aditsu

@MarsUltor he made that site

Anonymous

@MarsUltor All submissions are publicly available

Mars Ultor

@Mego He has a crawler?

aditsu

9:43 AM

yeah you can download the data from the scoreboard, they might even have a full archive somewhere

Anonymous

Something similar, supposedly

zyabin101

[gist.github.com/schas002/1a8bda6f34f7eca8a47faa103d6f3937

wat

Martin Büttner

Finished GCD in Hexagony and Labyrinth they're surprisingly short. \o/

flawr

Time to show off.

zyabin101

Challenge idea: golf Unipants' Golfing Language numbers.

That's what I really need, as I want to be able to output "Hello, World!" in the fewest bytes.

Also, a very big bullet point: make a prime checker. Else I wasted these 6-8 minutes and failed to make this golfing language.

Martin Büttner

9:48 AM

@flawr indeed.

aditsu

what unipants really needs is a suitable picture

zyabin101

wai upload same image again ;_;

aditsu

zyabin101

@aditsu O_o

Unicornpants is very obvious.

Panticorn is very insert reason here.

flawr

@aditsu What the hell are these here for?

I probably do not even wanna know.

aditsu

9:51 AM

zyabin101

@aditsu >.<

flawr

Unicorns are stupid. At least most of the internet ones.

Martin Büttner

@Mego the other challenge requires the entire computation to be output

plus, it requires the Euclidean algorithm

aditsu

@flawr just helping zyabin with the language "design" :p

Anonymous

@MartinBüttner Silly me, I read the "The Challenge" section and assumed that it would be descriptive and complete

zyabin101

9:53 AM

@flawr v

i - input an integer into the stack
I - input a character into the stack
o - pop the top of stack and output an integer
O - pop the top of stack and output a character

flawr

So what happens if there is a character on the stack and you try to output an integer?

zyabin101

Characters are actually integers, just that they are kept as their ASCII ordinals.

So the letter A is actually kept in stack as 65, the ASCII ordinal for A.

flawr

Oh right, I feel dumb now.

=)

@zyabin101 So where can we find an interpreter for your language?

zyabin101

@flawr Nowhere. I have to get help programming it.