The Nineteenth Byte

Stefnotch

9:00 AM

What does the range thing do?

Anonymous

Same as for(int j = m; j > 0; j += -1)

Stefnotch

Oh! :D

Anonymous

Basically, range creates a list of the integers between the first arg and the second arg, with the third as the step

Anonymous

And for j in... iterates over that list

Maltysen

Python has no traditional for loop. It only has a for-each which it calls a for loop.

Stefnotch

9:04 AM

:D (Javascript doesn't have that...aww..)

Anonymous

^^

Maltysen

@Stefnotch coffescript ftw ^.^

Anonymous

^

orlp

reference implementation

def B(n):
    n += 1
    A = [0] * n

    for m in range(n):
        A[m] = 1 / (m + 1)
        for j in range(m, 0, -1):
            A[j-1] = j * (A[j - 1] - A[j])

    return A[0]

Anonymous

@Stefnotch Also note that Python 3's range (as well as almost everything else) acts subtly differently

Stefnotch

9:06 AM

^^^ XD

Sp3000

A bit more than just subtly I believe, as anyone who's done range(99999999999) can confirm :P

feersum

what's way worse is if you do range(10**9) or something like that.

orlp

xrange has just been renamed to range in Python3

Python 2's range has been deleted

Stefnotch

10**9 reminds me of something...JS will get the ** operator!

Anonymous

9:26 AM

@Stefnotch Only like 7 years late

Stefnotch

:D

orlp

I WANT KITTIES

I WUV KITTIES

kitties are the best.

Sherlock9

@xsot I like explaining my solutions because I'm a show-off

@Stefnotch Yeah, Bernoulli numbers are pain in the butt because they're hard to explain and are in a surprising number of places

@Stefnotch Which is why it's impressive that the "first computer program" by Ada Lovelace calculated the Bernoulli numbers

Factorial:
from math import*
lambda n:exp(lgamma(n+1))

@Maltysen ? Then what is a traditional for loop?

Anonymous

@Sherlock9 lazy

Sp3000

9:44 AM

You can do lazy with generators though, right? (Confused)

Sherlock9

@Mego Also, inaccurate as it turns out. Darn

Anonymous

@Sherlock9 You could just do import math; factorial = lambda n:int(math.gamma(n+1))

Anonymous

Though I dunno if that'll also be inaccurate

Anonymous

Also not golfed because reference implementation

Sp3000

If you bring in floats, it's bound to be inaccurate :P

Is there no factorial in the math module?

Anonymous

9:51 AM

@Sp3000 durr

Anonymous

reference implementation

Anonymous

Which means it's allowed to be bad

Sherlock9

@Mego Just checked. Pretty inaccurate

feersum

No, reference implementation means it's supposed to be correct.

Sherlock9

Two functions not golfed at all

def comb(n,k):
 a=range(max(k,n-k)+1,n+1);b=range(1,min(k,n-k)+1);c=1;d=1
 for i in a:c*=i
 for i in b:d*=i
 return c//d

orlp

10:05 AM

today I got to use the word 'parenthesizations'

flawr

^ is this a word?

Sherlock9

def b(m, l=None):
 r = l if l else [1.0]
 if m < len(r):
  return r[m]
 else:
  s = 0.0
  for k in range(m):
   if k >= len(r):r=b(k,r)
   s += comb(m,k)*r[k]/(m-k+1)
  r.append(1-s)
 return r

Alien G

anesthetization is.

parenthesizes also.

Sp3000

(2+3)*4 and 2+(3*4) are two different parenthesizations. <-- I'm guessing this is the meaning?

Sherlock9

>>> b(10)
[1.0, 0.5, 0.16666666666666674, 0.0, -0.03333333333333344, 1.1102230246251565e-16, 0.023809523809523836, 0.0, -0.0333333333333341, 2.220446049250313e-15, 0.07575757575757303]

Need to work on that rounding

flawr

10:07 AM

Yeah, start sanding those edges!

orlp

@Sp3000 yes

Sp3000

What were you using the word for? :P

Sherlock9

@flawr What?

flawr

rounding = sanding edges

too late, i'm sorry for the bad pun=)

Sherlock9

@flawr oic lol

flawr

10:10 AM

I need your opinion on my edge counting challenge in the sandbox. The goal is counting the number of edges in a pixel image.

Sherlock9

Changing r.append(1-s) to r.append(round(1-s,6)) doesn't work very well >>> b(10) [1.0, 0.5, 0.166667, -1e-06, -0.033332, -1e-06, 0.02381, -0.0, -0.033334, 1e-06, 0.075758]

flawr

I want to allow (but penaltize instead) answers that do not perfectly solve the task, in order to allowe creative solutions that may not work perfectly.

feersum

@flawr Your polygons' edges appear to need some sanding.

flawr

@feersum I know, right=)

What is your opinion on this, would you allow non-optimal answers?

feersum

How many test cases are there going to be?

flawr

10:14 AM

I think about 10-12.

Depending on how many interesting ones I'll find=)

feersum

How many do you think it needs to solve to be interesting?

flawr

I do not know, I just thought of different cases like with convex and concave vertices etc.

@feersum So would you disallow non optimal answers completely?

feersum

Do you mean optimal as in solving all the test cases or should it theoretically work with some other inputs as well?

flawr

Finding the correct number for each test cases.

orlp

I posted a new challenge \o/

good luck

feersum

10:26 AM

If there are limited test cases then you could already take some shortcuts, probably

flawr

(The ansers should work for other inputs as well anyway, as optimizing for testcases is not allowed by default, I think?)

New Main Posts

0

Q: Fully parenthesize expressions

Today your challenge is to produce all possible full parenthesizations of an expression. Your input is a single line of printable ASCII containing one or more variables separated by operators. The input might also contains spaces - you must ignore these. A variable is [a-zA-Z], an operator is [^...

code-golf

orlp

10:46 AM

@Optimizer you here?

it doesn't matter what the terms/operators are anyway

it's always TERM OP TERM OP TERM OP TERM

so you can just strip spaces and split

Sherlock9

@orlp Right. Now the pain in the ass is getting all the parentheses

orlp

12:09 PM

still no answers after an hour

damn

Optimizer

sunday..

orlp

@Optimizer isn't that the best day?

Optimizer

not on holidays

thanks giving

blah blah

overactor

12:36 PM

It's been a while since I've been here last but I just had an interesting code golf related thought

are there challenges (likely koth) where one could compete by employing branch prediction?

orlp

@overactor you mean implementing?

overactor

No I mean employin it, as in:

12759

Q: Why is processing a sorted array faster than an unsorted array?

Here is a piece of C++ code that seems very peculiar. For some strange reason, sorting the data miraculously makes the code almost six times faster. #include <algorithm> #include <ctime> #include <iostream> int main() { // Generate data const unsigned arraySize = 32768; int data[arr...

java c++ performance optimization branch-prediction

orlp

ah

perhaps

I was thinking of a code challenge where you must implement a branch predictor in limited space

overactor

basically, suppose there's a challenge which requires you to recognise patterns in a series of ones and zeroes

@LegionMammal978 well, one thing it could do is predict what comes next in a series of 1's and 0's

LegionMammal978

How would you define a "pattern"?

01010010001000001000000001 has a pattern

overactor

12:49 PM

Though I'm not sure how clever branch prediction is

i know it recognises 11111111111111111 as a pattern

probably 1010101010 too

LegionMammal978

It can't use too many resources

overactor

as well as 110011001100

@LegionMammal978 good point

LegionMammal978

Or it would defeat the purpose

New Main Posts

2

Q: Help the architects visualize the skyline

As part of a city planning project, you've gotten the assignment of creating a program of function that will display the city skyline, given some input from the architects. The project is only in the startup phase, so a very rough sketch is sufficient. The easiest approach is of course to simply ...

code-golf ascii-art

LegionMammal978

Mathematica is too verbose

See my most recent solution as an example :P

overactor

12:52 PM

It would be more of a fun answer than an actually competitive one obviously

LegionMammal978

Especially string manipulation

Also, how do you put posts into chat?

Calvin's Hobbies

1:05 PM

@LegionMammal978 Just paste in the link

Nathan Merrill

If I have an object that represents a Vote, would it contain a "Voter" and "Votee", or a "Voter" and "Candidate"?

Dennis

Depends on what the votes are cast on. For votes on PPCG, votee is the better choice.

Nathan Merrill

well, anybody can vote for anybody

so, it sounds like Votee then?

(its for the controller for my next KoTH)

Oh, I could also do "Voted"

Dennis

1:22 PM

Without more details, I don't know if candidate would fit. But votee/voted fits always.

Not sure if votee is an actual word, but whatever.

Calvin's Hobbies

2:53 PM

9

Q: Help the architects visualize the skyline

As part of a city planning project, you've gotten the assignment of creating a program or function that will display the city skyline, given some input from the architects. The project is only in the startup phase, so a very rough sketch is sufficient. The easiest approach is of course to simply ...

code-golf ascii-art

@NewMainPosts It's interesting that the vote count in chat linked messages doesn't update. That must mean they keep track of the exact time of each vote.

Unless the message history is hiding something.

orlp

we should make a leader board

"most challenges won"

@Calvin'sHobbies are you going to answer codegolf.stackexchange.com/questions/65210/…

we need more answers from you :(

Sherlock9

@orlp I'm currently confused about parentheses and bogged down by assignments

I will try to answer though

Incidentally, around what time does @aditsu come on?

aditsu

I never leave :p

orlp

SLEEP IS FOR MORTALS

Calvin's Hobbies

@orlp data.stackexchange.com/codegolf/query/284338/…

Dennis, Howard, Peter, isaac, Martin...

Sherlock9

3:10 PM

@aditsu Your CJam implementing the prime divisors thing for sums of squares

It's driving me NUTS that I can't get it to work in Pyth

ARG!

aditsu

didn't somebody already do it in pyth?

Sherlock9

... Did they? I thought it was all array of pairs of squares then filter?

aditsu

ah, you did

are you saying it doesn't work?

Sherlock9

Yeah, I was the one who posted it

It works

It's 41 bytes!

That's far too bloated!

aditsu

anyway, I can't help with pyth

Sherlock9

3:15 PM

Ah fair enough

Pyth is kicking my ass

At this rate, I'm going to switch to CJam anyway

aditsu

yay :)

orlp

3:32 PM

@Calvin'sHobbies not a very good indicator

most challenges do not accept answers =/

anOKsquirrel

@aditsu how did you go about implementing the [ and ] operators? How do they deal with stuff like 7[;] ?

aditsu

I have a stack of marks (positions in the main stack); when you pop something, it moves back any mark that was at that position

anOKsquirrel

huh?

ohhh, I see

I think

GamrCorps

@aditsu so [;] is an empty array?

Calvin's Hobbies

@orlp But if they all did the order would presumably be about the same. I doubt Dennis and co. only try to answer questions made by users who accept answers.

aditsu

3:36 PM

yeah, assuming there was something before the [

anOKsquirrel

so if I did [, the stack would contain 0?

the mark stack that is

aditsu

depends what's before the [

anOKsquirrel

well, if the stack was only 0

orlp

@Calvin'sHobbies maybe the relative ordering has some form of accuracy, but the absolute numbers are worthless

anOKsquirrel

0 length*

aditsu

3:38 PM

if the whole program is [, then you get an error

orlp

@Calvin'sHobbies if you go through my answers you'll find that I've won a good chunk of them, yet barely any are accepted

anOKsquirrel

hm? why would that be?

aditsu

oh wait, is the comma part of the program or not?

anOKsquirrel

no, it's not

just the [

aditsu

I see.. ok, then it just adds a 0 mark (internally)

anOKsquirrel

3:39 PM

cool!

thanks! I'm probably adding arrays to my language next so that will help

Calvin's Hobbies

@orlp Well you go ahead and score every question without an accepted answer and tell me the proper results.

orlp

@Calvin'sHobbies It should be possible, using the same logic the per-question leader board uses

was more of a potential fun side project idea :P

@Calvin'sHobbies I also had an idea for a code-challenge

anOKsquirrel

@orlp tell me

orlp

writing a branch predictor

anOKsquirrel

what?

so I added carrots to scg

@aditsu have you seen usandfriends.github.io/CJam101/ ?

I use it all the time :P

aditsu

3:44 PM

yeah I linked it from the wiki

anOKsquirrel

really?

Cool.

Hello, everyone!

Hi Sock

Hi @anOKsquirrel!

So I got the Pyth solution I've been bashing my head on to 35 bytes

J4VrPQ8=J*J?nhtN2.&tt%htN4hhN1;?QJ1

orlp

3:50 PM

@Sherlock9 what challenge

Sherlock9

@orlp Sums of squares

@orlp Remember last night when I posted that ugly 45 byte solution?

anOKsquirrel

My power went out

Power outages suck

orlp

26

Q: Count sums of two squares

Given a non-negative number n, output the number of ways to express n as the sum of two squares of integers n == a^2 + b^2 (OEIS A004018). Note that a and b can be positive, negative, or zero, and their order matters. Fewest bytes wins. For example, n=25 gives 12 because 25 can be expressed as ...