The Nineteenth Byte

Quincunx

05:02

Doesn't the current xkcd have a mistake? Shouldn't there be a </blink> instead of a <blink>?

Instead of bobcat, package contained chair.

Doorknob

I think that's part of the joke :)

Sp3000

3

Quincunx

Okay

Sp3000

(thanks for the onebox - not sure what I was missing)

Doorknob

@Sp3000 https doesn't onebox

Sp3000

05:04

Ah

Dennis

05:40

@Sp3000: I'd love to see an explanation of how your Spintax answer works.

Sp3000

k, will do - just trying to iron out bugs

Vioz didn't really specify which chars we should expect in the input

Dennis

People keep complaining about CJam's readability, but I've stared at your code for five minutes and I still don't where to begin.

Must be some sort of magic. :P

@Sp3000 Yeah, I briefly thought about doing some replacements like } -> "][", but that's going to be trouble if the input contains double quotes or backslashes...

Sp3000

Hence my last few updates :P

Dennis

So all that backslash voodoo is just to escape quotes?

Sp3000

Yeah :/

It's still shorter than the other Python one though, surprisingly

Dennis

05:47

@Sp3000 I looked at your first revision and I think I got it now. Every {a|b|c} is replaced by choice("a","b","c"), then evaluated.

If I understood correctly.

Sp3000

Something like that :)

Hm... escaping the quotes and backslashes packs on 18 bytes :(

Yeah, this is definitely not a method that would save anything in CJam/Pyth

Dennis

@Sp3000 Even without dealing with the quotes, it's still 39 bytes.

Sp3000

Ahaha far too long :P

Dennis

goo.gl/cUHzcu

Sp3000

I posted because I thought it'd be an interesting approach, until I realised about quotes and backslashes :/

Dennis

06:02

Still the shortest approach so far in Python.

Sp3000

I'll take a look at splitting if I get bothered :P

Trying to do some %r voodoo but this is really messy :/

Sp3000

06:24

Nope, regex splitting still shorter :P

Dennis

@Sp3000 I don't think I've ever seen %r. What does that do?

Sp3000

String representation

It gets a little funky when you have quotes though, I think the priority is 1) If the string contains no ', wrap string with '' 2) If the string contains ' and no ", wrap string in "" 3) If the string contains ' AND ", escape quotes and wrap in ''

(all the while escaping backslashes)

Dennis

@Sp3000 Sounds interesting (albeit complicated). Is that a Python thing? C printf doesn't seems to have it.

According to the man pages, at least.

Sp3000

Yeah it is - it seems to have been added in Python 2.0

It basically does what ` (backtick) does in CJam

(posted the regex splitting answer :/)

Dennis

@Sp3000 Well, CJam has only one type of quotes, so there a lot less to worry about. Inspect in GolfScript is a pain though. Escapes all non-ASCII and ASCII control characters...

Sp3000

06:34

Eh.

Dennis

Yeah, in many aspects, CJam is GolfScript done right. For example, popping an element of a string pushes its code point, so if you add it back to the string, you obtain "aa65" instead of "aaa"...

And you really can't have a function named print in a golfing language...

Sp3000

base is the one I'm more annoyed about :P

There goes all of Kolmogorov

Dennis

The worst one in converting a string into the array of its code points. [{}/]

5 bytes for such a trivial task...

Sp3000

:) yeah I think CJam is much better in terms of design, so I like it. Still a few quirks, but much better.

Dennis

Hasn't been a match for Pyth lately, but I really like the stack-based approach. I've even been using it in some code challenges.

I can't imagine writing 300 byte Pyth code. All those arities get confusing after a while.

Sp3000

06:43

I was okay with Pyth at first, but I'm starting to not like it as much, mainly because things are a lot harder to remember

Dennis

The documentation is a little sparse as well. I'd probably understand it better if I knew more Python...

Sp3000

Yeah, knowing Python definitely helps. You can check here for docs though, isaac did that a while back

Dennis

My hat goes off to Jakube for this answer. That must have been a pain. Or fun, depending on how you're wired.

Ooh, that's nice. I've only been reading the doc.txt.

Sp3000

:P maybe if Jakube posted the equivalent Python it'd be easier to understand

Dennis

He always posts an explanation. I'm looking forward to it.

Well, I really should go get some sleep. Have to work tomorrow.

Sp3000

06:48

Ahaha k, have fun :)

feersum

07:11

Wow, how did he manage to take so many bytes to implement gcd with a language designed for golf?

Sp3000

Ahaha 122, nice :P

btw if you have parens for 0**(c%g), why not do (c%g<1)?

feersum

it can be negative

Sp3000

... oh k

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Rotate a Chinese checkerboard (or diamond tiling) Your task is to rotate a Chinese checkerboard or a diamond tiling by a multiple of 60 degrees. The input is an integer and a board / tiling, Your program should rotate it by the integer * 60 degrees. You decide whether to rotate clockwise or coun...

Sp3000

@feersum Does a ZeroDivisionError occur for every case where there is no solution?

feersum

07:21

g = gcd(a,b) so yes

Sp3000

Hmm I don't seem to be getting that...

(I'm trying to input 3x+4y=17, which I believe should be input as 4,3,17?)

feersum

x =-17, y = 17

what's wrong with that?

Sp3000

3*3 + 4*2 = 17

You need a solution in nonnegatives

feersum

why nonnegatives?

Sp3000

"Write a program or function that accepts the coefficients a, b and c as input and returns an arbitrary pair of natural numbers (i.e., non-negative integers) x and y that verify the equation ax + by = c, if such a pair exists." (emphasis mine)

feersum

07:27

Example 2 has Output: (2, -1)

Sp3000

That's because example 2 has no solution, and "Your code may behave however you want if no solution can be found, as long as it complies with the time limit and its output cannot be confused with a valid solution."

feersum

Why would someone post a trick question like that?

Sp3000

You didn't think it was weird that [] and -1 were also valid outputs for example 2?

Might drop a comment though

feersum

08:22

How could I divide in Python, rounding away from 0 regardless of sign?

Sp3000

What's rounding away from 0? 1/2 -> 1 and -1/2 -> -1?

feersum

right

Sp3000

08:40

I can't think of anything better than x/y+(x*y<0) :/

feersum

that doesn't work

trichoplax

Integer results need to stay as they are

In python 2, / is integer division, right? Same as // in python 3?

Sp3000

Oh, right... (also I got it the wrong way around anyway, oops)

trichoplax

Oh - it's not as simple as I thought in python 2: "/ is floor division when both args are int, but is true division when either or both of the args are float"

Cool Guy

Why does opening the 9th revision here in [inline] or [side-byside] not show the "show code snippet" button?

Peter Taylor

08:53

@feersum Are both numerator and denominator potentially either positive or negative, or do you have a sign constraint on one of them?

Cool Guy

It only appears when swithcing to markdown and then, switching to [inline] or [side-by-side]

trichoplax

@CoolGuy Revision 8: "Expanded on minimum reqs, minimised stack snippet"

feersum

@PeterTaylor No constraint

trichoplax

Oh hang on - that means reduced the code, not hid the buttons

Cool Guy

I'll create a gif now so you get a better understanding

trichoplax

08:55

@CoolGuy Just had another look - the snippet is hidden (it's a tick box option when making a snippet). This means you'll need to click on "Show snippet" before you'll see the button. It's the same in the question, not just the edit, and it's how it's meant to be.

It just saves a bit of vertical space

Cool Guy

No I wasn't talking about that buttons

trichoplax

Oh sorry - start again... :)

Cool Guy

I was telling that the "Show snippet" disappears

But now, the bug seems to have disappeared

:/

Never mind. Cannot reproduce it anymore :/

Sp3000

@feersum (a+-a%b*(a*b>0))/b maybe?

trichoplax

@CoolGuy I get the opposite. If I open the page with it set to show inline it shows no "show snippet" link. If I then switch to a different view and then back to inline, the "show snippet" link appears

Cool Guy

08:59

I've reproduced that

^

I wonder why....

feersum

@Sp3000 nice, that appears to work, although I can't understand what it does lol

Peter Taylor

@feersum (n+(n^d>0)*(2*(n>0)-1)*(abs(d)-1))/d

trichoplax

I guess it's not too important in this case as the snippet wasn't one of the things edited, so there's no reason for it to show. If it affects edits that did involve the snippet then it would be worth raising it as a bug

Peter Taylor

Ah, Sp3000's is shorter.

Sp3000

Peter, is the ^ in your one xor or exponentiation?

feersum

09:05

n^d>0 would test if they are both positive

Peter Taylor

xor

Sp3000

@feersum If it's negative, Python's round down is correct. If it's positive, add enough to reach the next multiple of b (i.e. (-a)%b)

Hmm if it's xor then I think it works except when n is 0 (writing test cases over here)

Peter Taylor

Ooh, I don't have n=0 in my test cases

You're right: it breaks for n=0, d>0

Because of the (2*(n>0)-1), where I wanted signum but that didn't work

Sp3000

Python 2 has cmp for that

feersum

cmp(x,0) to be precise

Peter Taylor

09:09

And (n+(n^d>0)*cmp(n,0)*(abs(d)-1))/d works

aditsu

what are you guys trying to do?

Sp3000

55 mins ago, by feersum

How could I divide in Python, rounding away from 0 regardless of sign?

Too bad round() rounds towards zero, although floats would probably be a problem anyway

aditsu

can either number be negative?

feersum

My earlier 'problem statement' said yes, but now in my program I'm thinking the denominator is guaranteed to be negative

I've got x/abs(a)*cmp(a,0)

Sp3000

You can drop my previous one to (a+-a%b*(b<0))/b, but with the restriction there might be shorter

feersum

09:23

Excuse me, I meant the numerator is negative

Sp3000

Can the numerator be zero? Or strictly negative?

feersum

strictly

I can use a/b+(b<0>a%b) in this case

Sp3000

You can even do (0>a%b) instead, negative mod only happens when b is negative

feersum

Yes...

aditsu

I got x/y+(x*(x%y)>0)

feersum

09:33

is this in the general case?

aditsu

yeah

feersum

you can remove the parens on x%y

x%y*x

aditsu

ok :)

Sp3000

Nice :)

Peter Taylor

09:44

@aditsu, I've started a side project of making an Android wrapper for the CJam interpreter, taking advantage of it being in Java. It's currently in a barebones alpha form. Would you object to me putting it on GitHub when I think it's ready for other people to try using?

Zgarb

Hello all

aditsu

@PeterTaylor no problem, although I'm not sure how useful it is to have an android version (since you can run stuff in the browser)

Zgarb

Could some Windows user test my Hex-challenge controller? The link is in the sandbox post: meta.codegolf.stackexchange.com/a/5445/32014

Peter Taylor

The full-featured end goal is that it allows you to set up test cases, run them all, load and save to a zip file which wraps the test cases and the script...

aditsu

o_O

trichoplax

10:21

@feersum I'm curious about your sandbox comment on the invisible target

When you say "moving to the center of the arena" do you mean Nathan Merrill's strategy of minimising the maximum path length? The arena wraps so doesn't have a conventional centre. And is your hint at a better strategy something you'd share, or something to save until it is posted to main...?

feersum

@trichoplax if there is a bottleneck leading to an enclosed region with no other players, then you shourld block the etnrance to it

even if the empty area is just 1 square, it will give you better winning chances than being in the middle of a pack

trichoplax

@feersum :) interesting. I guess how many squares you need to make the decision depends on how many other players are in the arena.

feersum

of course if there aren't a lot of others in the middle of a connected region, things could be different

trichoplax

(and how many squares are in the area you aren't guarding - which will change over time)

I'm hoping that with bottlenecks appearing and then being closed off the players should all be constantly changing their minds and scrambling around each other for the best places.

feersum

10:38

I doubt the region will remain toroidal

trichoplax

@feersum Of course, but the centre could end up being anywhere, including on one of the edges, so it only has meaning part way into the game

As well as Nathan Merrill's centre, I guess you could also think of the mean position, or halfway between the extreme horizontal and vertical cells. There are probably lots of possible centres.

feersum

how many players are there again?

trichoplax

I'm leaving it open to as many players as enter - I'm hoping 1024 cells will be enough to accommodate them with plenty of room left for an interesting game.

I'd be surprised if I get enough players to make that a problem, but I suppose ideally I should specify a maximum number of players and run separate games if there are too many for one

I could say 16 or 32 is the maximum number of players in any given game

I'd be shocked if I got 32 answers though

feersum

11:25

@trichoplax Does your controller keep all the programs running for the duration of the game?

New Main Posts

1

Q: Format the given number of bytes to a human readable format

Challenge and origin On Stack Overflow a popular question is: How to convert byte size into human readable format in java? The most up voted answer has a quite nice method for doing this, but this is codegolf and we can do better, can't we? Your challenge is to write a method or program that...

code-golf

trichoplax

@feersum It isn't written yet, but if it uses STDIN/STDOUT then yes it will (for speed)

This is one thing that makes me lean towards a language specific contest - then we could have the speed of calling a function without the overhead of running potentially a dozen programs

feersum

It would be more work but if you could run all the entries from the same language in one instance, then there wouldn't be so much memory overhead

Not sure that there would be enough to even be a problem though

trichoplax

@feersum I was thinking more if all the answers try to store huge arrays to help them plan. I'm not really expecting that but I'm considering functions with no memory or limited memory.

feersum

A bunch of restrictions don't seem necessary..1024 isn't very big.

trichoplax