The Nineteenth Byte

@Sp3000 when I get to the point where the output format of match mode, grep mode and split mode is configurable, I might look into allowing all modes in multistage usage

I think I might be able to interpret brainfuck with Retina now :D

Sp3000

2:30 AM

@MartinBüttner Wait how would IO work?

Martin Büttner

ah, IO... well I could make the input both the program and STDIN and the output the tape and STDOUT

btw, I just solved the addition challenge myself

2 repetitions, 5 single replacements

(so score 25)

I'm pretty sure it's possible with a single loop, maybe even without loops

Sp3000

Hmm I'm pretty sure about single loops too. Without loops might be a bit harder

Martin Büttner

yeah, but I thought the same about incrementing numbers at some point :D

anyway, I think I'll leave that to the participants

I'm sure user23013 will come up with something crazy

Sp3000

Well at least with incrementing you don't have the problem of the ends of the numbers being in two different positions

Martin Büttner

true

Sp3000

2:34 AM

But yeah user23013 probably :P

Martin Büttner

I'll probably wait with posting until monday

Martin Büttner

3:09 AM

this post is such a treasure chest

1478

Q: Hidden Features of C#?

This came to my mind after I learned the following from this question: where T : struct We, C# developers, all know the basics of C#. I mean declarations, conditionals, loops, operators, etc. Some of us even mastered the stuff like Generics, anonymous types, lambdas, LINQ, ... But what are t...

I keep forgetting it and remembering it every few months and always discover something new

Sp3000

:P there's a pretty good one of Python too

Martin Büttner

C# is such a lovely language... I wish had use for it more often

user23013

4:27 AM

Did you mean the 2D matching challenge? I'm not currently working on it.

Sp3000

@user23013 No, Martin's addition regex problem in the sandbox

user23013

It should be possible with 2 replacements, no repetitions in .net... but a bit too complicated

feersum

6:10 AM

I want to see if I can use the ability to leave slime trails (booleans which cannot be unset) can be used for arbitrary computations

any suggestions on what to compute?

Sp3000

6:24 AM

XOR gate?

feersum

I mean something complicated

requiring iteration etc

Lembik

8:22 AM

@PeterTaylor I don't know how to make a suffix-with-up-to-one-error tree that isn't huge to be honest. Any ideas?

Sp3000

@Lembik If it's like the idea I had previously, you give each search node a counter (here starting at 1). If the next node's char is a match you continue, if it's a mismatch you decrement the counter and either 1) counter < 0 and you stop searching that branch or 2) counter >= 0 and you continue searching

(I'm assuming that's what Peter meant, might be wrong though)

Lembik

@Sp3000 Is a search node a node in the suffix tree?

Sp3000

Yeah

Lembik

@Sp3000 here are some issues.. a) we would need a node after each character. A normal suffix tree only has a linear number of nodes. This isn't a huge problem but it is worth saying b) When you find a mismatch you need to branch somehow to carry on . How do you do that?

@Sp3000 actually I don't understand the idea at all. The standard LCS algorithm makes a suffix tree of the two strings and finds the deepest node which has leaves from both strings under it

@Sp3000 what is the idea here?

Sp3000

The same thing but you branch at most once per path? Hence the relevance of the size of the alphabet

Lembik

8:30 AM

@Sp3000 even if the alphabet is size 2 I still don't get it. When you say branch, it's like you have a pattern you are searching for. But in our case we don't

@Sp3000 are you still making the standard generalized suffix tree of the two strings?

it also doesn't work for pattern matching :) But that's another topic

@Sp3000 or am I just being dumb?

Sp3000

Just making sure I'm thinking the right thing. It's not a strict traditional generalised suffix tree, no

Lembik

ok... the issue is that you would seem to need to branch after every character when making the suffix tree

which would be bad

taking n^2 space and time I think

Peter Taylor

I think it would be O(kn) space where k is the size of the alphabet, but I haven't got time now to discuss it in any detail.

Lembik

@PeterTaylor I don't see that. If you make a suffix tree which adds an extra branch after every character then the added tree that hangs off each branch seems to make it O(n^2)

unless you have some clever way of compressing thta

Martin Büttner

12:13 PM

@user23013 whoa, I'm looking forward to see that

user23013

12:35 PM

I think I'll answer it as a demonstration of those .net features, when I have time. But once you know all these things are possible, it's just something like translating every program to Brainfuck... which isn't too interesting for me.

(isn't too interesting to do that myself. But had a +1.)

Optimizer

1:28 PM

You forgot Java :D :D

XD

Martin Büttner

oh right

I actually thought about it when writing .NET

Martin Büttner

1:41 PM

@Optimizer fixed (and added notes about character class operations to all flavours)

wtf did I just find... the character class documentation on MSDN has a | inside a character class: msdn.microsoft.com/en-us/library/… o.O

Martin Büttner

2:11 PM

Doorknob

I remember when that was at 1337. :D

PhiNotPi

"beta"

Martin Büttner

Not much longer! 14th of May!

PhiNotPi

Really? (I can't remember whether or not that was a running gag at one point)

Martin Büttner

no, not really

I'm pretty sure if they have any intention of letting us graduate at all, they'll probably want to complete the graduation of CR first.

PhiNotPi

2:20 PM

After a certain point, beta = final release

To be honest, their whole scheme of "graduation" seems rather broken.

Martin Büttner

@Doorknob oh, while you're here, the sandbox got unfeatured

Doorknob

What? I see the featured tag right there.

(and there's no yingluck tag this time, for that matter)

2

PhiNotPi

yingluck tag?

Doorknob

Oct 12 '14 at 23:19, by Geobits

@Doorknob Before I edit it out, is there any reason you tagged the sandbox with [yingluck]? It seems a bit odd...

Sp3000

2:50 PM

Python 3.5: "Number of capturing groups in regular expression is no longer limited by 100."

I like the idea that, somewhere out there, there was probably someone going "Oh no my regex needs 101 groups - stupid Python"

PhiNotPi

Sounds like it's time for some epic regexes.

Martin Büttner

also what sort of limit is 100?

that sounds so arbitrary

(unless it was actually 99 and their parser only checked for 2 digits)

Sp3000

"It is 100 to avoid a syntactic ambiguity between numbered groups and octal numbers, if I remember correctly."

Optimizer

100 ?

Martin Büttner

haha I see

Sp3000

2:55 PM

>>> re.search("\100","@")
<_sre.SRE_Match object; span=(0, 1), match='@'>

Well that's confusing.

feersum

did you mean to try r"\100" ?

i.e. "\\100"

what you did there is search('@', '@')

Sp3000

re.search(r"\100","@")
<_sre.SRE_Match object; span=(0, 1), match='@'>

Yes, I did mean that - same result though

Optimizer

100 ?

Martin Büttner

he's stuck in a loop again

Optimizer

100 ?

Martin Büttner

3:01 PM

wait, how did you do that with duplicate message detection?

Optimizer

100 ?

ProgramFOX

@MartinBüttner That detection only stops messages if they're posted in a certain amount of time.

Optimizer

so I think now you would have guessed that there is a time check

ninja'd

Martin Büttner

huh, I would have thought that time limit is longer

Optimizer

100 ?

ProgramFOX

3:04 PM

I think it also has something to do with the amount of messages posted by other people between the time of your duplicate messages, but not sure about that.

goes to Sandbox

Optimizer

you mean meta ?

ProgramFOX

Sandbox

Where you can play with chat features (except flagging) and ch...

2

4

Martin Büttner

still can't repost that deleted message

ah now

wow my repeated message "ban" is up to 10 seconds now...

test

interesting... I can repost if I post something in a different chatroom in between

test

very odd

PhiNotPi

I was trying to interject some Jackson lyrics.

ProgramFOX

Hmm... it looks like ban time increases when the system detects that you post a duplicate message -- during my first test, it took 2 mins, and during my second test, I still couldn't do anything after 2,5 mins.

or perhaps it doesn't increase, but just start over

and now I could after 22 seconds, so apparently it does start over even if your message couldn't be posted

PhiNotPi

3:16 PM

Check this out: ` - You can perform this action again in 1 seconds - retry / edit / cancel
` - You can perform this action again in 3 seconds - retry / edit / cancel
` - You can perform this action again in 1 seconds - retry / edit / cancel
` - You can perform this action again in 5 seconds - retry / edit / cancel
` - You can perform this action again in 5 seconds - retry / edit / cancel
` - You can perform this action again in 4 seconds - retry / edit / cancel
` - You can perform this action again in 4 seconds - retry / edit / cancel

Optimizer

100 ?

Martin Büttner

were those 22 seconds after you'd posted in this room?

ProgramFOX

@MartinBüttner 22 seconds between those messages, without any or message posted in between

PhiNotPi

It looks like there's a 3- 5- 7- 9- etc. second pattern.

Martin Büttner

weird

on my very first attempt I couldn't repost for 4 or 5 minutes

ProgramFOX

3:17 PM

So figuring out the exact time will be hard.

Optimizer

I have special dupe repost prevs

ProgramFOX

@MartinBüttner Has the system caught some of your duplicate messages during testing? It looks like the ban time starts over if it does.

Peter Taylor

I'm refraining from casting a dupe vote because of my superpowers; do other people think that codegolf.stackexchange.com/q/47510/194 is close enough to codegolf.stackexchange.com/q/25416/194 to be a dupe?

Optimizer

I have special dupe repost prevs

PhiNotPi

How many other people have dupe-nuke superpowers?

Peter Taylor

3:19 PM

Four

Doorknob

@PhiNotPi 5 in total

Peter Taylor

Plus mods

Martin Büttner

@ProgramFOX oh I see, yeah I did try in between

Optimizer

oh noes, my perfect rate of 5 upvotes per answer towards the race to dupe-nuke power is down by 8 upvotes :(

Martin Büttner

@PeterTaylor I thought about that but I didn't think so

the core of the old challenge is parsing the dice format, the core of the new challenge is the interactivity

Optimizer

3:34 PM

@MartinBüttner David's answer looks like O(m*n) no ?

Martin Büttner

yes

Optimizer

ok

feersum

a_~f~x_:=Max[k #+#2&@@@a/.k->#]&/@x lmao, is that actually code?

Optimizer

Have you seen Pyth answers ?

Martin Büttner

@feersum looks very readable to me ;)

Optimizer

3:48 PM

so does Pyth

Martin Büttner

to you maybe :P

no, but I mean, I find that much more readable than CJam for instance... which I know pretty well, too

feersum

what is @@@ ?

PhiNotPi

It's readable only in the sense that the light from the characters is hitting my retina.

Martin Büttner

@feersum f@@@list is syntactic sugar for Apply[f, list, {1}], which applies f to each list contained in list... it's like a hybrid of Map and Apply

feersum

@@@ is a single token?

Martin Büttner

3:54 PM

@PhiNotPi actually the characters are the bits that don't emit light (or very little, at least)

@feersum yes

hm, might as well post the regex challenge now

Optimizer

4:16 PM

now ?

or now ?

PhiNotPi

I posted some ideas in the Lego chatroom.

Optimizer

what about now ?

PhiNotPi

Martin's not even in the room right now.

Optimizer

I wonder what he meant by "now"

PhiNotPi

"sometime in the future"

Optimizer

4:38 PM

NowNotNow ?

PhiNotPi

4:52 PM

I'm (finally?) pretty close to being ready to post Core War.

Optimizer

KOTH ?

PhiNotPi

yes

Since it is a purely-deterministic KOTH where the competitors are just text files and the controller is in Java, it's about as system-independent as possible.

Optimizer

"controller is in Java" - talk about system-independent

PhiNotPi

I assume Java is not nearly as system-independent as it claims to be, then?

Optimizer

oh can the answers be in any language ?

PhiNotPi

4:58 PM

Core War is a game in which "programs" fight for control of a "virtual machine." Since the game's concept dates from a time where "Core" was a type of memory, it's not that complicated.

Competitors are a form of assembly language. Each instruction can add/subtract/move/etc. data. The goal is to corrupt the opponent's memory and cause it to crash.

Core War is a real thing. The version I am posting is about as simple as Core War variants come.

By "the competitors are just text files" that's literally what I mean: the competitors aren't written in a "real" programming language.

Optimizer

oh

PhiNotPi

The most similar KOTH held so far was CodeBots: codegolf.stackexchange.com/questions/36978/…

@Optimizer You sound unenthused.

Optimizer

koth not my thing

but lets see

Optimizer

5:36 PM

what about now ?

I changed my mind

now ?

you changed it now ?

6:04 PM

has it ever occurred to you have putting spaces before punctuation isn't very optimised?

Optimizer

is it ?

maybe its Optimized ?

Martin Büttner

I doubt it

PhiNotPi

M a y b e   y o u   s h o u l d   p u t   s p a c e s   i n   b e t w e e n   e a c h   l e t t e r   a s   w e l l .

Does anyone else want to cast their vote for Lego challenge ideas?

and/or post their own idea?

Optimizer

L I K E T H I S ?

PhiNotPi

N O , L I K E T H I S !

ProgramFOX

6:19 PM

L
i
k
e

t
h
i
s
!

Optimizer

L k       h  s
i  e   T i   !

PhiNotPi

Martin Büttner

why, guys?

ProgramFOX

Martin saw us! HIDE!

:P

Optimizer

O         o  s   !
h     N e     !

PhiNotPi

6:25 PM

(╯°□°）╯︵ ┻━┻

^ unmatched closing bracket

PhiNotPi

7:22 PM

I'm uploading my Core War controller to GitHub now.

PhiNotPi

7:36 PM

github.com/PhiNotPi/NanoCoreWar

Jakube

8:02 PM

What do you think of a golfing language, where the commands are huffman coded. I mean commands that appear quite often use only 3 or 4 bits and commands that are used rarely take 10 bit.

For instance: I took about 500 chars of Pyth code and Huffman coded the letters. Afterwards I could translate the short Pyth code '%eQr1hhQ' (8 byte) to 00100 111101 1001 110101000 110110 10000 10000 1001 (5.5 byte).

PhiNotPi

I wonder how short you can get stuff to be.

Although I feel that some people will consider compressing the source code to be cheating (because you can't write code like that).

Jakube

My quick analysis (500 chars code) tells, that on average you can shorten code by about 25%.

PhiNotPi

Some of that is a result of the fact that there's only about 97 usable characters anyways (6.6 bits per character).

Jakube

And you don't have to compress the code. I mean there can be a doc that says range = 1011, read numbers from input = 11101, ... You just set each bit individual.

Optimizer

"setting each bit individually" is hard

PhiNotPi

8:16 PM

I think there is some precedent for that. I'm pretty sure I've seen hex humps on here before.

Like this golfscript answer: codegolf.stackexchange.com/a/6325/2867

TheBestOne

Mostly for assembly programs

Optimizer

@PhiNotPi this is self converted via scripts

nobody sets the bits manually

PhiNotPi

okay

There would be some limit on setting the bits manually, so as to fit the .txt file format.

TheBestOne

Don't we usually use UTF-8 to measure the bytes in code?

PhiNotPi

I honestly don't know.

Optimizer

8:20 PM

@TheBestOne no

PhiNotPi

8:36 PM

Can someone please look over my Core War proposal? meta.codegolf.stackexchange.com/a/2479/2867

I'm ready to post it, but I want someone to make sure that the instructions make sense.

Martin Büttner

9:31 PM

@TheBestOne depends

PhiNotPi

@MartinBüttner Are you willing to look over my KOTH?

Martin Büttner

sorry, I'm a bit knackered and I still need to get some work for uni done

maybe tomorrow if you don't mind waiting that long

PhiNotPi

okay

Optimizer

CH is cooking up something

I can smell it

PhiNotPi

"seen 7 mins ago"

maybe you're right

Calvin's Hobbies

9:49 PM

@Optimizer You say that like every three days.

I actually not cooking. Just thinking of ingredients.

Optimizer

all coz you post so much

PhiNotPi

There is an interesting article about my high school in the newspaper, talking about disparities in education funding.

Off topic, I know, but it's interesting to read.

aditsu

10:34 PM

hey guys, check this out: cjam.aditsu.net/#code=1e6000%20998001%2F

PhiNotPi

I'm not quite sure what that's supposed to be.

A really fast division?

aditsu

look at the digits in the result

PhiNotPi

I already noticed, they look weird.

aditsu

1..2..3..4..5..

...999 :p

but one of them is missing

PhiNotPi

which one?

aditsu

10:42 PM

this makes it clearer: cjam.aditsu.net/#code=1e3000%20998001%2F%0As%2200%22%5C%2B3%2FN*

PhiNotPi

998?

aditsu

yep

998,001 and its Mysterious Recurring Decimals - Numberphile

►

PhiNotPi

I'm trying to "golf" my AP CompSci homework.

I say "golf" because real golfing would be a lot more extreme.

It's also difficult because I want a decent grade.

FUZxxl

11:07 PM

@PhiNotPi If you want to annoy your TA, rewrite the homework in J or APL.

Martin Büttner

11:22 PM

@aditsu the comments on that make me cringe so bad

aditsu

that's normal for every youtube video..

Martin Büttner

lol I knew you'd say that

yeah I agree for most videos (there are rare exceptions)... but this one was particularly bad

(like most things popsci, I guess)

aditsu

which comment in particular? :p

Martin Büttner

I just started reading the reply thread to the top comment and couldn't take it any more after a few

PhiNotPi

Yeah, some of the comments are pretty bad.

aditsu

11:31 PM

hmm, looking back at chat.stackexchange.com/transcript/message/20439284#20439284 , I may need to encode "*" too

New Sandboxed Posts

11:52 PM

0

A: Sandbox for Proposed Challenges

Digitsum of powers in a base code-golf math number base-conversion You should write a program or function which outputs or returns the sum of the digits in base b representation of the first n natural power of a in ascending order. Example for a = 2, b = 3 and n = 4 The base10 numbers are 2^...

Calvin's Hobbies

How does the New Sandboxed Posts user know when the Sandbox updates? Is it some fancy moderator tool I know nothing of?

Sandbox

Transcript for