The Nineteenth Byte

TheNumberOne

12:04 AM

@PhiNotPi The best program I've evolved so far always returns the same number :(

PhiNotPi

12:22 AM

@TheNumberOne How many elections are you running?

TheNumberOne

Only 10000

PhiNotPi

Well, there is a lot of randomness involved in each election.

TheNumberOne

The problem is with the program tree generation code.

PhiNotPi

Okay.

How are you generating programs?

TheNumberOne

I'm using JGAP.

Martin Büttner

12:25 AM

@Sparr I generally agree but a) it's meta and b) it's explicitly supposed to list less ground-breaking contributions that aren't worth cluttering the question with.

PhiNotPi

I don't know anything about JGAP. :/

In the meantime, I've made little to no progress towards an OO golfing language.

TheNumberOne

12:43 AM

Is 10000 elections / 17.56519374 seconds(against the 2 default bots) fast enough for a bot?

PhiNotPi

I'm not sure...

Well, given that the final leaderboard needs about 10m or so, which is 1000X more than 10000, probably not.

TheNumberOne

:(

PhiNotPi

17.565 seconds * 1000 = almost 5 hours

TheNumberOne

I'll lower it by a factor of 10. Tell me if it's too slow.

PhiNotPi

Lowering it by a factor of 10 is probably tolerable if I am going to start doing overnight runs anyways.

Well, I was about 2 hours into what I expect to be a 4 hour run.

I'll probably have to cut that off and start a new one to run overnight to include yours in the leaderboard.

Martin Büttner

12:59 AM

this has an interesting question/answer vote ratio (apparently there are also at least 14 deleted answers)

TheNumberOne

@PhiNotPi I could have it time itself and stop after so many ns

The average bot takes 40.81 microseconds per turn.

New Loophole Proposal

1:15 AM

0

A: Loopholes that are forbidden by default

Zero-length answers Considering an empty program a quine was original in the 1994 IOCCC. Eleven years later, if you can solve a task with an empty program, you should just notify the OP.

Sp3000

@MartinBüttner I'd thought about an exec quine, but I'd need to sit down and fiddle to see how it should work

Alex A.

@Sp3000 You can't do it while standing?

@MartinBüttner Looks like a job for Peter.

Sp3000

More like I can't do it on the phone :P

Alex A.

Fair enough. :P

TheNumberOne

@PhiNotPi It tries to take under 200 micro-seconds now.

PhiNotPi

1:27 AM

Okay.

Alex A.

Seeing "o.k." as an abbreviation for "okay" always reminds me of when my friend tried to convince me that O.K. is actually an acronym for Oll Korrect, apparently an old-timey spelling of All Correct.

PhiNotPi

I'm starting a run now.

I expect for it to take 4.5 hours.

Which means I'll have to update it in the morning.

Although.... I might be able to make it a 20m run.

Alex A.

Parallelize your code.

I don't know what you're talking about, but micro-optimization is always the answer.

PhiNotPi

Hmm. the fact that I tell them the partial election results might prevent that.

Alex A.

Nah. They don't need to make informed decisions.

TheNumberOne

1:39 AM

Is that how you vote?

PhiNotPi

I guess that would make it more realistic.

Alex A.

In parallel?

No, I fill out my ballots serially.

PhiNotPi

Dual-wielding writing utensils.

Alex A.

I might need to try that.

Maybe in a small local election...

Also, what the hell:

TheNumberOne

Die bird!!!

Oh, @AlexA. you changed your profile image.

Sparr

1:47 AM

@PhiNotPi you could parallelize your elections by running more than one whole election at once

then you don't have to worry about not having partial results available

Alex A.

@TheNumberOne Is that a bird? It looks like a platypus or something.

PhiNotPi

@Sparr That's a good idea.

Alex A.

@TheNumberOne I did. It's no longer a goose.

TheNumberOne

@AlexA. You silly goose :)

PhiNotPi

I don't know how to do that, however.

Alex A.

1:48 AM

@PhiNotPi Forks?

TheNumberOne

@PhiNotPi docs.oracle.com/javase/tutorial/essential/concurrency/…

Alex A.

!$omp parallel do

TheNumberOne

@AlexA. The code link suggests a magpie. However, actual magpies look completely different.

Alex A.

Port all of the code to Fortran

TheNumberOne

No. Java > Fortran

Sparr

1:50 AM

anthropomorphizing birds while keeping different types of birds distinct is hard

Alex A.

@TheNumberOne Indeed they do. No creature on this Earth looks like that thing.

@Sparr False

For example, geese are jerks.

TheNumberOne

@Sparr Haven't you heard of Big Bird? He is very distinct.

Alex A.

He's certainly a large presence in the room.

Sparr

sure. and what kind of bird is he supposed to be?

Alex A.

Emu

TheNumberOne

1:52 AM

@AlexA. Your profile image looks like a Magpie.

Alex A.

It is!

TheNumberOne

:D

Alex A.

Nice work!

Everybody else thought it was either still a goose or an archaeoptryx.

TheNumberOne

Aren't they endangered?

Alex A.

Magpies? Not at all.

Archaeoptryx on the other hand... yes. Extinct, actually.

And geese are too mean to die.

I end up talking about birds every single time I log in here.

PhiNotPi

1:55 AM

You could probably kill half the world population of magpies and they'd still be considered "least concern."

TheNumberOne

They're trying to make Sage Grouse endangered right now :(

Alex A.

According to Wikipedia, the sage grouse is considered "near threatened." So you mean they're going to reclassify it into the "threatened" range?

TheNumberOne

foxnews.com/politics/2015/05/29/…

@AlexA. Either or.

Alex A.

Ugh. Fox News.

TheNumberOne

Is awesome.

Alex A.

1:59 AM

Mm, I think we may have differing views.

To each his own.

TheNumberOne

The funny thing is that, in Idaho, its mainly hunters who are paying for the preservation of the Sage Grouse. (I'm not sure about other states.)

gtg

Alex A.

That's odd. Do they hunt sage grouse?

Bye!

New Main Posts

2:15 AM

0

Q: Chinese checkerboard

The game of Chinese checkers is played on a board with spaces in the shape of a six-pointed star: Image from Wikipedia We can create an ASCII-art representation of this board, using . for empty spots and the letters GYORPB for the six colored starting locations: G G G ...

Alex A.

Is the board game tag relevant for the Chinese checkerboard challenge?

Chinese checkers is a board game, but I guess I don't understand what that tag is meant to signify.

The wiki for that tag should really be clarified...

randomra

3:39 AM

my checkerboard python code turned out way too ugly and complex... probably over 200 bytes properly golfed

at least I have an extra reason to try it in Retina :)

Doorknob

@Calvin'sHobbies uh oh... I've been discovered, abort abort

New question close stats for users with 2k+ rep have just been rolled out.

Katya

yet another motivator for me to hit 2k

New Main Posts

0

Q: Map of Islands (and a river)

Introduction For many centuries, there has been a certain river that has never been mapped. The Guild of Cartographers want to produce a map of the river, however, they have never managed to succeed -- for some reason, all the cartographers they have sent to map the river have been eaten by wild...

code-golf graphs

1

Q: Make a slow error quine maker!

As part of his answer to Make an Error Quine!, @Falko proposed the following algorithm: How to create your own solution in 2 minutes? Open a new file in an IDE of your choice. Bang your head onto the keyboard in front of you. Compile. Replace the code with the compiler error mes...

code-challenge quine busy-beaver

randomra

5:06 AM

"Self Shortening Shelf Sorting" challenge is in the making, although it is more about the title than the actual task :)

Sp3000

Still no solves for checkerboard? I'm surprised

randomra

my solution is very cumbersome, maybe I will search for a better method

or just wait to see it posted :)

it would be fun to use a cellular automaton extension to solve problems like these

idk if there is an adequate one for python

Sp3000

5:22 AM

Hmm I think I can do sub 200

randomra

definitely possible, mine was a mess :)

Sp3000

149 :D

randomra

5:38 AM

nice, somewhy I expected much lower scores initially

New Main Posts

0

Q: XOR encrypt your programs with the source of other programs

Write two programs such that when the bytes of their source code are XOR-d together it produces a third, valid program, all in the same language. The length of each of the three programs in bytes must be identical. The first program should output A. The second program should output B. The third ...

code-golf

randomra

but the hexagonal layout never really helps

Sp3000

Hmmm well I'm just doing it line by line...

randomra

I thought you would try to partition the space

my ~200 solution is line-by-line too

but I'm hoping some other approach can beat that

is there a short way to replace list of chars to list of chars? like 'abcdefabc','ac','AC' -> 'AbCdefAbC'

Sp3000

str.translate

(aka what I'm doing as well :P)

randomra

6:02 AM

and by cell automaton I meant something like this:

(type, pos_x, pos_y)*
init='G',3*n,0,'B',0,n,...,'.',3*n,3*n)

(type, move_x, move_y)*:
rules='GxyGxyBxy ... .xy.xy.xy.xy.xy.xy')

print(run(rules,init,owerwrite=False))

Sp3000

6:14 AM

Oh goody, is @xnor doing checkers as well?

xnor

yup!

i have a secret plan!

Sp3000

xnor gets sub-100. Calling it.

Sp3000

6:39 AM

Man, this would be so much shorter if we only cared about visible output :P

xnor

yeah

my method is coming out too long...

probably better just to print each line

Sp3000

Ahaha, trying something fancy like ASCII Sun equations?

xnor

yup

Optimizer

too long - no scrollbars

Sp3000

Ahaha I think we might be bothering DLosc with all these clarifications :P

xnor

7:11 AM

if I make a dict from a list of pairs with duplicate keys, is it specified which one will be used?

Sp3000

The last one

xnor

ok, that might be useful

Sp3000

"If a key occurs more than once, the last value for that key becomes the corresponding value in the new dictionary." (just double checking :P)

xnor

thanks

Sp3000

Hmm I'm having trouble getting as low as I want. Leading and trailing space rules add about 10-20 bytes

randomra

7:34 AM

@MartinBüttner are you planning adding line and/or block comment modifiers to Retina? e.g. #`this is a comment

xnor

I think ~150 is the best I can do with my approach of inequalities in triangular coordinates

Sp3000

Hm... that doesn't sound too good. Line-by-line I guess then :/

xnor

yeah :-/

too bad, it's nice mathematically

randomra

I thought superimposing the 2 big triangles would be good

Sp3000

I thought about that, but it looked like too much work :P

randomra

7:36 AM

as you dont have to be exact with the letters, as any position with any 2 letter is a dot

so in the middle you can choose any letter

xnor

Here was my approach, with the hope that you can make something better of it. Let x and y be the coordinates, centered at (0,0). Let (a,b,c)` be the corresponding triangular coordinates (x+y,x-y,-2*x), which add to 0. Then, the region is determined by which of [a,b,c,-a,-b,-c] are greater than 2*n. If none are, we're in the center. If exactly one is, we're in one of the six triangles. If 2+ are, we're not on the board. Also, if a is odd, put a space instead.

Sp3000

Ooh that sounds nice. Maybe post it as a note in your submission anyway? :P

xnor

sure, i guess i'll post it, even if it's not optimal

randomra

That's nice, cleaner than any of my methods

xnor

unfortunately, python is very clumsy at printing graphs like this because strings are immutable

looping over pairs of coordinates is ugly too

Sp3000

7:50 AM

And using lists would be fine if it wasn't for "".join() being 9 bytes :P

(or ``[2::5] being 8, but still)

xnor

yup

though even generating a list of lists is ugly because [[' ']*a]*b makes them equal by reference

Sp3000

Ahaha yes that too

What's the best way around that situation anyway? map(list,[('',)*a]*b)?

randomra

@MartinBüttner an 'ignore whitespace in replacement string' modifier would also improve readability a lot (if you could write whitespace with escape chars as you planned it wouldn't hinder functionality)

xnor

@Sp3000 hmm, good question

Sp3000

btw your checkboard colours are backwards

xnor

8:02 AM

oh, darn, thought i fixed that

randomra

@Sp3000 my longer try at that: eval(('['+'"",'*a+'],')*b)

Sp3000

@xnor 6*n+1 -> y-~y

xnor

ah, right, that evalulates first

it's also probably shorter to do a double-layer exec-loop

multiplying the code string, adding stuff, and multiplying again

Sp3000

8:26 AM

So how are you going with line-by-line? :P

xnor

i spent enough time coming up with what I had, so i'm giving up on it

i'm waiting for you to post :-P

Sp3000

Hey :P

Mine's not that good though :( ... yet anyway

xnor

it beats mine, right?

Sp3000

141, not by that much

randomra

~250 byte at first run with Retina, more tedious than fun though

but I did it :)

xnor

8:37 AM

btw, that's a nice trick with map to do map(list,[('',)*a]*b)

you can use lists to shave a char map(list,[['']*a]*b)

Optimizer

hate it when I have to add 9 more bytes to a CJam code.

xnor

it's funny mapping list to lists

map(list,[' '*a]*b) works for spaces

Sp3000

Ahaha mapping lists to lists

xnor

two shorter:

eval(`[['']*a]*b`)

Sp3000

XD

xnor

8:45 AM

hahaha

Sp3000

You should make that a tip

xnor

it's the poor man's copy.deepcopy

Sp3000

That actually sounds like a really useful idea in general

xnor

can you think of any other uses?

Sp3000

eval(`L`.replace("1","0")) and such?

That's pretty specific though

xnor

8:48 AM

ooh

I had a thought to deep-reverse a tree this way, but [] don't flip

Sp3000

>>> L = [[1, 2, 3], [4, [5]], 6]
>>> print(eval(str(L)[::-1].translate({91:93,93:91})))
[6, [[5], 4], [3, 2, 1]]

(I don't know what I'm doing)

xnor

unfortunately, i think at that point a recursive lambda is probably shorter

Sp3000

Probably, yeah

>>> L = [1, 2, 3, 4]
>>> eval(`L`.replace(",","*"))
[24]

Poor man's reduce?

xnor

that's hilarious

Sp3000

I think a lambda's still shorter there though

But it might be more useful in Python 3, which doesn't have reduce builtin

xnor

8:54 AM

Python 3 can't do the short repr though

Sp3000

Still better than import functools :P

xnor

for Python 2, i get a tie with reduce(lambda a,b:a*b,L,1)

Sp3000

Yeah :(

eval(`L`.replace(","," or "))

Short circuiting, maybe? You need to be careful with whitespace though

Like any but actually returning the item

... nvm filter wins here

xnor

well, i wrote a tip just for the matrix, feel free to expand with more ideas

Sp3000

I've been annoyed by the matrix thing for a while now, so I'm glad there's a better solution :)

xnor

9:03 AM

it's still 8 chars of overhead if not for the mutability issues

is there any way one could overcome this when you modify?

Sp3000

I guess it depends on what you're doing with the lists

I think for fibclock I did [[" "]*a]*b because I zipped straight after

xnor

in the type of case i'm thinking of, you generate the matrix, but there's only one statement that modifies an entry

like for these ACSII art challenges where you draw in the symbols

Sp3000

Not sure...

New Main Posts

1

Q: Shift characters in a string

In this challenge, you are required to shift characters in an inputted string n number of times and output the shifted string Input Input will first contain a string. In the next line, an integer, which denotes n will be present. Output If n is positive, shift the characters in the string to...

code-golf string

Sp3000

I posted my attempt - I'm sure you could smash it if you sat down and tried though :P

randomra

9:12 AM

J shifts the other direction, no 2 char solution

and stdout is another 5 bytes

Sp3000

10:01 AM

@xnor I think the "Choosing one of two numbers based on a condition" tip could do with some examples, esp the last one

Martin Büttner

@randomra you can already use comments in the regex, if you use the free-spacing modifier x

that doesn't help much in -s mode of course, because you still can't include newlines

I'm not sure I'll add comments to replacement strings. I think in most cases the trickiness will be in the pattern so that any quirks in the substitution can be explained there

but I'll keep it in mind for if I ever rewrite Regex.Replace

N3buchadnezzar

10:23 AM

Heya

hello

moin

:D

hi

@PhiNotPi Really nice question you asked =)

PhiNotPi

10:27 AM

Thanks.

Martin Büttner

@aditsu would you consider changing e[ and e] such that if you give them an arraylist to pad with, they pad with the contents of the arraylist, not the arraylist itself? e.g. "abc"8"Xx"e] --> "abcXxXxX"

wow, 91 bytes for chinese checkers in cjam o.O

Optimizer

wow

you make @Sp3000 happy so much

Sp3000

??

Optimizer

(in 2x range of cjam code)

Sp3000

10:42 AM

Ahaha, I think that just means the 91 is long here :P

Martin Büttner

I guess with xnor's approach it would be closer to 70

Optimizer

@Sp3000 in good words. yes. :P

Martin Büttner

I just tried the straight forward way of building one quadrant with a couple of loops and then mirroring twice

Sp3000

I think I'd skip the horizontal mirror and just go with vertical, personally

The odd/even offsets look a bit fiddly

Martin Büttner

then I need to pad spaces on both sides though to get the rectangle

Sp3000

10:48 AM

... oh right. CJam doesn't have .center()

Strip spaces on the right instead?

Martin Büttner

I need the spaces to simplify the vertical mirror

87 now

Optimizer

he's still happy

Sp3000

Er... by vertical mirror do you mean left half/right half? (because I meant top/bottom oops)

Martin Büttner

@Optimizer have you tried/are you trying?

Optimizer

nope

Martin Büttner

10:50 AM

@Sp3000 top/bottom

Sp3000

Trailing spaces simplify reflecting top/bottom? Interesting...

PhiNotPi

Current OO language concept art for a simple "Fibonacci object" with an initial conditions and an "update" method:

F:[x:y:1;u:[~x,~y=y,x+y]]

Martin Büttner

@Sp3000 yeah, just make it a single string and reverse it

(like we did for that random rug some time ago)

Sp3000

Ahaha right, I see

N3buchadnezzar

Anyone here seen Ex Machina?

Martin Büttner

10:57 AM

May 21 at 14:32, by Rainbolt

Has anyone seen Ex Machina?

I don't know if he asked because he wanted to talk about it or because he wanted to know if it's worth seeing

N3buchadnezzar

I just found a particular scene funny. They are "hacking" into a security system, and the code they are writing is in python... The explicit code the protagonist uses is the "Sieve of Eratosthenes ".

I giggled.

Martin Büttner

not as bad as Tron: Legacy

"let's try the backdoor... su backdoor"

Sp3000

@MartinBüttner A naive port got me 77 :P

N3buchadnezzar

@MartinBüttner Kung fury has some glorius moments "I can hack you back in time".

Optimizer

I have seen ExM

randomra

11:07 AM

@MartinBüttner haven't noticed the x modifier, that seems useful and I guess you can use something like |$ this is like a comment if you want to

what annoyed me was that I can't put my work-in-progress text somewhere in the file

but now realized I can do (double) comment lines with $

N3buchadnezzar

@Optimizer enjoyed it?

Optimizer

not that much, but yeah

PhiNotPi

I found a video titled "Hack All The Things" where a guy briefly explains how to hack 20 devices in 45 minutes.

Martin Büttner

11:28 AM

@randomra you can still use free-spacing mode for that

x`# comment

randomra

@MartinBüttner wow, I can't even read the second line of a two line text, my bad

Martin Büttner

I think (?# ) is essentially a block comment, but still not usable with newlines in -s mode

randomra

yeah, I find -s very practical, not planning to leave it

or the other mode impractical :P

Martin Büttner

yeah the other mode is a pain to use... the only reason it's the default mode is for golfing when you need newlines in pattern or replacement

randomra

loophole to use a 257 char alphabet :)

Martin Büttner

11:36 AM

except you're not ever gonna use most of the control characters anyway, so I could have used one of those instead, but that would be even less usable

I might ditch multifile mode once newline escapes work in replacement strings. I could also maybe make -s more sophisticated by not splitting patterns on newlines if that would invalidate the regex... so that if you have a literal \n in a character class or a group, it won't be considered as a split because the pattern isn't complete yet

flawr

I think this is the last time I bother you with my 'Catch the cat' assymmetrical koth challenge: Do you agree that I should make two threads (like in cops and robbers) one for the cats, one for the catchers submissions? (As cats only compete within cats, and catchers only compete within catchers)

Martin Büttner

has anyone used programs to replace flash cards when learning stuff? I'm just looking for something really simplistic, but all the tools I've found so far seem a bit overkill while not being as flexible as I'd like them to be.

(ideally I'd just like to read in a bunch of markdown card definitions and flick through them in my browser... might just put something together myself if I can't find anything)

I guess this comes close

randomra

ahh, that kind of flash card, that makes more sense

Martin Büttner

well not images... just text really

randomra

just saw it, today I read just half of the sentences...

Martin Büttner

11:51 AM

I can work with that. I can work with that. Is this better? Is this better?

randomra

nope, this is just "I can work with that. I can work with that." I need "Relevant text. Gibberish."

PhiNotPi

@flawr It's up to you. The main motivation for having two separate thread for Cops and Robbers is so that the Robbers can submit answers as opposed to comments. Since everyone in CTC is going to submit an answer either way, it doesn't matter.

Martin Büttner

@flawr I haven't read the sandbox post, but "asymmetrical KotH" screams "separate threads" at me

PhiNotPi

Might as well reap the rep.

randomra

@flawr as the tasks are different I would vote for separate questions too

flawr

11:56 AM

ok thank you very much for the feedback, if everything goes as planned it will be released today or tomorrow=)

Martin Büttner

it allows you to accept a cat and a catcher separately

flawr

yes thats the idea

Martin Büttner

maybe I should add a transliteration mode to retina

N3buchadnezzar

@MartinBüttner I was in need of that for the rubiks cube, but the number of flash cards seemed a bit daunting to me.

Ideally I would like to have subcases of images or texts. Practice these 30, then these 30 and finally the 60 mixed togheter.

Martin Büttner

@N3buchadnezzar that brainthud one seems good... I don't know if you mix decks though

@trichoplax btw my upvotes from tweetable maths didn't get reverted the other day. how about yours?

TheNumberOne

12:31 PM

@AlexA. No, we don't :)

N3buchadnezzar

12:55 PM

@MartinBüttner well it is a bit more advanced than that though

New Main Posts

1:12 PM

0

Q: Assymetrical KOTH: Catch the Cat (Cat Thread)

Assymetrical KOTH: Catch the Cat This challenge consists of two threads, this is the cat thread, the catcher thread can be found here. The controller can be downloaded here. This is an assymetrical KOTH: Each submission is either a cat or a catcher. There are games between each pair of each a...

king-of-the-hill java

0

Q: Assymetrical KOTH: Catch the Cat (Catcher Thread)

Assymetrical KOTH: Catch the Cat This challenge consists of two threads, this is the catcher thread, the cat thread can be found here. The controller can be downloaded here. This is an assymetrical KOTH: Each submission is either a cat or a catcher. There are games between each pair of each a...

king-of-the-hill java

Optimizer

no catchness thread ?

@aditsu working on next version ?

PhiNotPi

@flawr I would recommend adding RandomCat/RandomCatcher answers to help kick-start the competition.

trichoplax

1:29 PM

@MartinBüttner No mine didn't get reverted either. I wonder whether there was more than one person looking on the same day. I got another the next day, for the same answer as one of the previous day's, so at least that one was from a different person.

flawr

1:44 PM

@PhiNotPi Thanks for the suggestion

randomra

in a Java koth if I don't want players (classes) to keep state/info between calls is forbidding static variables enough? apart from directed hacks like reflection

flawr

@randomra As here on codegolf.SE the competitions are not automated (like rps-contest.com) I do not think it is necessary to make restrictions? The competitors have to post and explain their code so any attempt to cheat will be discovered I think..

Sp3000

(Friendly reminder: CodeJam in 10 minutes, if anyone's participating)

flawr

CodeJam?

Martin Büttner

@trichoplax I got three upvotes twice... I think maybe the script is clever enough to notice if someone just upvotes a bunch of answers on one question

or generally if someone just upvotes a lot or specifically targets one user

randomra

2:12 PM

@flawr it's more about avoiding misunderstanding about what "do not store data between calls" means than catching cheats

and what should I pay attention to in submissions to spot data storing

Optimizer

2:56 PM

@MartinBüttner not posting ?

Martin Büttner

?

Optimizer

chinese checkerboard

Martin Büttner

oh, I closed the tab when Sp3000 said he had 77

Optimizer

in CJam ?

Martin Büttner

yes

Optimizer

2:57 PM

wow, only 2 hrs left before Dennis wakes ups .. gotta hurry

Sp3000

You have 73 to beat

Also this is hard

Optimizer

I probably have ~60 to beat

New Meta Posts

0

Q: Can numeric input/output be in unary?

Some string-based languages (including but not limited to sed, Retina, ///) can't really handle integer arithmetic in decimal (or other normal bases). Therefore, these almost always have to start by converting the input from decimal to unary, and the output back from unary to decimal, which can e...

discussion policy code-golf answers

randomra

@MartinBüttner do you think other languages could accept unary too whose standard number representation (whatever that means) isn't unary?

Martin Büttner

well yes

I just mentioned the languages as examples where it's sort of necessary, but the default is supposed to be valid for all languages equally

randomra

3:14 PM

and if someone wants you to use decimal, he/she should give the input as string, because otherwise it is just the inner representation

(this sound a bit unclear, was clearer in my mind)

Martin Büttner

well this of course only applies to reading input from STDIN or something

just found a bug while posting that meta poll:

0

Q: Making an immediate self-answer community wiki also CWs the question

Normally, users can't make questions community wiki (only moderators can, after the question is posted), but they can manually turn answers to CW. Now I just noticed that when you self-answer immediately (i.e. by checking the "Answer your own question" checkbox while posting the question), and s...

bug community-wiki self-answer

randomra

I'm having trouble with this (unary). How is it different from accepting binary input (which is forbidden.)
I feel like it should be more about "accepting input number as a standard representation in your language".

Martin Büttner

is binary input forbidden?

I find "standard representation" way too vague

randomra

well, i suppose, we often convert the input to binary right at the start

@MartinBüttner that's true unfortunately

Martin Büttner

8

A: Can numeric input/output be in the form of byte values?

Yes, numerical input and output may be given as a character code As long as this doesn't give an unfair advantage (e.g. for a challenge that is "given a character code, output the character"), I don't see any problem with this. It's similar to allowing each language to use its native string repr...

randomra

3:25 PM

although we already do that with list inputs and many kind of output

Martin Büttner

for those we allow any convenient/unambiguous representation, not necessarily the standard representation

people can still choose whether they want an array literal or a space separated list for instance

randomra

so can I take a list as 1*2*3 if they ask for the reduced product?

> The languages in question are probably not going to win normally anyway, and I don't like to discourage answers in interesting esoteric languages, just because the author couldn't be bothered to include the atoi implementation.

This is the goal here too I guess, and unary doesn't give much advantage but I would say other bases might do.

adding my code to the end of the input is also a convenient/unambiguous representation in my mind, so I don't like that

Martin Büttner

I think this is where we use common sense :P

randomra

for me that is what standard representation means, maybe standard isn't the best word, as that implies only one

BrainSteel

3:40 PM

I think I'm against unary being used as a default. In some challenges, it makes a legitimate difference and I'm not sure it should be the asker's responsibility to specify numbers must be in decimal.

Martin Büttner

right but how many question askers are gonna explicitly allow it when it doesn't make a difference?

BrainSteel

That's a good point too. I'm very uncertain about this whole idea. Making Retina users ask about what to do every time is certainly not optimal either.

Martin Büttner

even if you ask them about it, if it's not an accepted default I can see a lot of challenge authors being reluctant to allow it

I'll go get some food... later, guys.

BrainSteel

Ah, yeah. Perhaps specifying decimal numbers should be more common practice when it matters. It's another thing the sandbox can easily catch, but will probably disrupt a bunch of main posts until it catches on.

Transcript for