The Nineteenth Byte

LegionMammal978

11:00

It's GET / HTTP/1.1<CRLF><CRLF>

wizzwizz4

Golfed: GET/HTTP/1<LF><LF>

zyabin101

@wizzwizz4 O_O

wizzwizz4

I have a dilemma.

LegionMammal978

See above, servers SHOULD NOT accept that

wizzwizz4

I have written a piece of code for a challenge, and it seems that nobody else will.

However, my code doesn't work properly and I can't fix it.

The only other answer has been written by the question asker.

Jester Tran

11:02

How many groups are there of order 2?

wizzwizz4

And they have put a bounty on the question.

15

Q: N-queen-and-equine quine

There is a variant of the well-known N-queens problem which involves queens and knights and is said to be "considerably more difficult" 1. The problem statement is as follows: You must place an equal number of knights ♞ and queens ♛ on a chessboard such that no piece attacks any other piece...

@QPaysTaxes I haven't posted my answer.

LegionMammal978

@JesterTran groupprops.subwiki.org/wiki/Groups_of_order_2

wizzwizz4

I think I should post it as CW; if it's simple to fix, I'll get the bounty.

If it's not, the person who fixes it will.

I can't fix it! That's the problem.

Jester Tran

@LegionMammal978 Thanks, so the answer is 1 as I thought

wizzwizz4

There is a third option:

@QPaysTaxes There's an algorithmic error that's tiny, but important, and I can't quite fix it.

Jsfiddle collab link: jsfiddle.net/wizzwizz4/0z0s6yf0/4/#&togetherjs=pB4TNlat9H

LegionMammal978

11:05

Did you try unplugging it then plugging it back in again?

wizzwizz4

@LegionMammal978 Yes. It doesn't work when the mini-function is unplugged from the wrapper either.

Erm... I might have just fixed it! YAY!!!

zyabin101

@wizzwizz4 For records, I'm Brilliant Beaver here.

trichoplax

@wizzwizz4 I hope you can fix it, but if not, this is not what Community Wiki is for:

14

A: Definitive policy about answers not meeting the challenge specification

Marking an answer Community Wiki does not excuse it from this policy Introducing an answer with "I know this doesn't follow the rules but I worked hard on it so I want to post it here" is not an acceptable excuse. Saying "I've marked it Community Wiki so I won't get any rep from it" does not ch...

@QPaysTaxes Not everyone shares this policy - lots of us are happy to be pinged by anyone. I'd recommend turning off the volume for sleeping if you don't want to be disturbed by pings. If you need to be alertable for other things you can leave the volume on and just turn off pings (top right of chat has a speaker icon).

wizzwizz4

@QPaysTaxes Actually, I saw you came on chat, so I pinged you.

Katenkyo

@QPaysTaxes But that is your special case, and people in this community are generally willing to help fellow codegolfers ^^. Don't take it for you personnaly, and just don't go on a long phrasing, just say that you prefer to be left alone in these cases because you don't feel like helping^^.

trichoplax

11:19

@QPaysTaxes Even if you write in your profile "Please don't ping unless you know me", not everyone will see it and you will still risk being woken from sleep. Because lots of people welcome pings the only way to defend against this is to mute either pings or your phone/computer while sleeping. This isn't meant to be telling you what to do, just hopefully giving you a way to get undisturbed sleep... :)

Katenkyo

@QPaysTaxes (I think it qualifies very well, but didn't remember the word ^^')

trichoplax

@QPaysTaxes Rants are always better received when self described as such :)

@QPaysTaxes I can relate to that. Similar to getting a letter/email and then finding it's only spam/bills

Katenkyo

@trichoplax Or being called to fix a computer... just because it is un-plugged :D

trichoplax

Actually in this case bills, since it was a request for work from you...

@Katenkyo I've had that...

Katenkyo

@trichoplax I know... at least, it's funny afterwards

wizzwizz4

11:24

Sorry for pinging you. :-(

trichoplax

I don't know what etiquette is for pinging. I get pinged because a turtle appeared in a randomly generated xkcd garden. I guess the reason it doesn't annoy me is because I never expect a chat ping to be important news, so it feels inclusive rather than disruptive.

Katenkyo

@trichoplax I think there's no etiquette

the community pings when they don't want to discuss something in comments, when they need advice from a person, when there's something that might be interesting

trichoplax

I wonder if we should make one. Chat doesn't have a meta but I guess it could be on topic on our meta if the etiquette is PPCG specific, or on mother meta if it's general

Katenkyo

but no clear "social rule" about that, or at least, I didn't feel them

wizzwizz4

It was serious discussion!

Katenkyo

11:29

@QPaysTaxes this chat is "childish" compared to others on SE if I can allow myself to say so

wizzwizz4

About a seriously fun code-golf challenge!

Katenkyo

We're a pretty active chat, and often speaks about everything

trichoplax

@QPaysTaxes This room definitely has less chance of serious discussion (not suggesting yours wasn't @wizzwizz4...)

Katenkyo

@trichoplax I think it would be sad to make one, because people would want to watch closely how they ping from that point

trichoplax

In some rooms a ping seems like a summoning but hear it is often just a mention "for your information" style

Katenkyo

11:31

it would result in questions like "should I ping @x for y subject"

trichoplax

I would hope that the community would reach consensus on a pretty permissive etiquette, but yes it would be awful if it became a set of strict rules

Katenkyo

community like objective things, a permissive ain't as we can't define which level of permissivity to apply

@QPaysTaxes are the hours in GMT+0?:)

@QPaysTaxes it would mean pinging while I sleep, easy :D

but yeah, as I stated, I don't think it would work

wizzwizz4

@zyabin101 jsfiddle collaborate has crashed :-(

Here's the link again: jsfiddle.net/wizzwizz4/0z0s6yf0/#&togetherjs=pB4TNlat9H

Katenkyo

@QPaysTaxes Have a nice day pal!

Lembik

hi @TonHospel

trichoplax

11:42

@QPaysTaxes You too :)

Ton Hospel

@Lembik: Hi

Lembik

@TonHospel you must have a very recent g++ . g++: error: unrecognized command line option ‘-std=c++14’

g++ (Ubuntu 4.8.4-2ubuntu1~14.04.1) 4.8.4

is c++11 no good?

Ton Hospel

@lembik: Ubuntu is sooo behind the times

Lembik

@TonHospel is it hard to make it c++11 compliant?

Ton Hospel

@lembik: It compiles and runs for me with C++11

I don't seem to use any 14 features

Lembik

11:46

ok let me try that!

@TonHospel hmm.. it will take a few minutes as my computer is updating itself and then I need to install armadillo

@TonHospel argh..it's updating sagemath. That's 10 minutes then :)

@TonHospel in the meantime I can ask.. what score did you get?

Ton Hospel

@lembik: It depends on how you determine m when you generate the initial mxn matrix M when constructing your positive definite matrix

For m close to n it's fast, for m=1 it is slow

Lembik

@TonHospel oh that's surprising. Can you tell me more? I have just arbitrarily chosen m around n/2

for m = 1 there is probably a closed form solution, no?

Ton Hospel

It's very sensitive to the shape of the elipsoid xT P x == 1

Lembik

@TonHospel say m is fixed, some P will need large boxes I assume?

Ton Hospel

It constructs a box around that and then grows that box. For low m it has to grow the box a lot, for high m it needs almost no extra points beyond (0,0,0,0...)

Lembik

12:01

@TonHospel can that intuition be formalized mathematically?

I assume you are good at math :)

Ton Hospel

I think m very much influences the the difference in eigenvalues.

Lembik

the eigenvalues are typically all around m I believe for a randomly chosen M

well.. unless they are 1

Ton Hospel

@Lembik: I'm a physicist by training, but with a strong math background. Though it's all quite some time ago :-)

Lembik

there are n-m eigenvalues which are 1 exactly and m which are close to m with high probability

@TonHospel sounds good!

does that make sense ^^^

@QPaysTaxes err?

ah

welcome back

Ton Hospel

@Lembik: Anyways, for m=n/2 it should be around 12, maybe 13

Lembik

12:09

@TonHospel great! One day sage will finish installing :)

@TonHospel bpaste.net/show/47d367e4d90b

any ideas?

Ton Hospel

@lembik: Looks like a link error. Did it install libarmadillo itself ? Try apt-get install libarmadillo6

Lembik

The following NEW packages will be installed
libarmadillo-dev libarpack2-dev

that's what it installed

I could install libarmadillo4

Ton Hospel

@lembik: Yeah, whatever is the latest on your ubuntu. dpkg -l "libarmadillo*[0-9]" should say something is installed

Lembik

12:26

@TonHospel bpaste.net/show/3eb3b39af9ba

so I am a bit stuck now

what to do?

Ton Hospel

@Lembik: What is the output of dpkg -l "libarmadillo*-dev" ? Is it the same library version ?

Lembik

I am just purging and reinstalling

bpaste.net/show/cba5f3493994

Ton Hospel

@lembik: Ah, try to add -llapack -lblas to the link options

Lembik

g++ -std=c++11 -O3 -fomit-frame-pointer -march=native -llapack -lblas -larmadillo -s infinity.cpp -o infinity

like that or does it have to be in a particular order?

Ton Hospel

@Lembik: The order shouldn't matter

Lembik

12:31

bpaste.net/show/4980a5a1cd8b :(

Ton Hospel

@Lembik: Paste the ouptut of dpkg -l "libarmadillo*"`please

Lembik

bpaste.net/show/2e7712939a43

Geobits

In lieu of the usual "good morning", here's this:

4 Non Blondes - What's Up

►

Ton Hospel

@Lembik: That looks right. Don't know why it can't find the missing symbols

Lembik

@TonHospel The FAQ is at arma.sourceforge.net/faq.html#support but I can't say I fully follow it

Ton Hospel

12:51

@Lembik: is eleanora your alter ego ?

Lembik

@TonHospel no.. it's just confusing. I share the computer and now stackoverflow has decided that's who I am on SO. I can't find how to fix it

it hasn't decided that for any other .se site

presumably someone using this computer was logged in as eleanora once

Ton Hospel

@lembik:Weird. Anyways, indeed try adding #define ARMA_USE_LAPACK to the top of the source code

@lembik: And/Or add #undef ARMA_USE_WRAPPER amd drop the -larmadillo from the link set

@lembik: mm, no that undef makes no sense. That would only work by editing the defaults file

Lembik

paste.ubuntu.com/15649559 and g++ -std=c++11 -O3 -fomit-frame-pointer -march=native -s infinity.cpp -o infinity gives the same problem :(

am I doing it wrong?

Ton Hospel

Ah, you can undef the wrapper by defining ARMA_DONT_USE_WRAPPER

@Lembik: Yes :-) The define must come beforte the armadillo include

Lembik

paste.ubuntu.com/15649621

now?

this is really quite annoying :(

Ton Hospel

13:03

@Lembik: Right, and drop the -larmadillo on the compile line but keep lapack and blas. Still failing ?

Lembik

like this? bpaste.net/show/26241ea1eb45

Ton Hospel

@Lembik: Probably also define ARMA_USE_BLAS

zyabin101

@Mego you can now safely publish the caucus results.

Lembik

@TonHospel same failure

zyabin101

If you do so, please include the ballot file with which you've done it.

Lembik

13:05

do we really need the -s ?

Ton Hospel

@Lembik: No. you can drop it. Shouldn't make a difference though

Lembik

it doesn't

it's just all a total failure :(

Ton Hospel

@Lembik: Yeah, and I can only randomly guess since it works perfectly on my system :-)

Lembik

@TonHospel what is your system?

Ton Hospel

@Lembik: Debian sid

Lembik

13:07

hmm.. not so different

Ton Hospel

Usually close enough to ubuntu, but less old

Lembik

hmm.... maybe someone on SO will reply helpfully

@TonHospel ahem... g++ -std=c++11 -O3 -fomit-frame-pointer -march=native infinity.cpp -o infinity -larmadillo

that works :)

it matters where the -larmadillo is apparently

Ton Hospel

@Lembik: Surprising!

Lembik

yep

@TonHospel testing your code at long last

@TonHospel how do you work out when to stop?

Ton Hospel

13:34

@Lembik: If the growth in a dimension is less than (a bit below) 1e-4/dimensions. Since the growth goes down steadily (with sometimes small increases due to more points being added) this is reasonably robust. However it has a possibility to stop too early so i also added a test that all dimensions have to be tried at least once (though that makes it run too long when m == n)

Lembik

@TonHospel your method sounds generally awesome. Could you add some explanation to your answer? I am also going to add a bounty soon I should say

Ton Hospel

Since each growth step gets slower than the previous and the test is "was the last step useless" it actually wastes too much time in that last step.

Lembik

that makes sense

my current problem is that my testing code is slower than your code!

Ton Hospel

@Lembik: hehe

Lembik

:)

or do you prefer not to add explanation for fear someone will copy you :)?

Ton Hospel

13:37

@Lembik: So what n does your code reach ?

Lembik

@TonHospel about 9 or 10

it's running time varies depending on P

Ton Hospel

@Lembik: Hey, I have 2 comment lines. That's a lot for me :-)

Lembik

:)

the explanation you have given here would be great in the answer

zyabin101

@StackCodeGolf/moderators after a second election, will the current mods retain their diamond?

Ton Hospel

@lembik: It also wastes a lot of time evaluating the corners of the box that contribute nothing significant, but I couldn't think of a good way to stay close to the elipsoid without wasting a lot of memory and search time

Lembik

13:41

@TonHospel If you document the method other people can work on it too :)

(I'll stop now...)

Mars Ultor

@zyabin101 Yes

Ton Hospel

@lembik: the message is clear :-). I'll add some context

Lembik

@TonHospel thanks! I still don't fully understand whether your method is guaranteed to be correct or not (within 0.0001)

is it?

Ton Hospel

@lembik: It probablly isn't absolutely guaranteed, but it should be very close

Lembik

@TonHospel ok that makes sense

did you try it on my example with output 8.1443647932 ?

Ton Hospel

13:45

You can see the "change" per step in the output. Only if that would suddenly increase would the estimator be bad. You can see it increase very slowly for directions that are equivalent and then suddenly drop when it goes to the next worse dimension

Lembik

trying n = 12 now

Ton Hospel

@lembik: yes, it's correct to 7 digits after the period for that one

Lembik

@TonHospel your code is seriously awesome! Just finished testing n = 12

it also varies hugely in running time depending on P

Ton Hospel

@Lembik: Yes. It probably won't make 13

zyabin101

@Downgoat @EasterlyIrk and other GitHub people: Verification of commits signed with GPG signatures is now automatic.

Read more about GPG signatures at GitHub's helpful docs.

Lembik

13:53

@TonHospel bounty added

@TonHospel I will be impressed if anyone beats yours I have to say

zyabin101

@Lembik For what? FA@EOBP? Or for Ton's answer?

Lembik

@zyabin101 that's too confusing

wow.. thanks

zyabin101

@Lembik FA@EOBP ==> Fastest answer at end of bounty period.

Lembik

@zyabin101 I meant Ton's answer is great so if anyone does better that's even more impressive

hi @quartata

Rainbolt

Hi

Oops wrong window

2

zyabin101

13:57

Hi @quartata

Cᴏɴᴏʀ O'Bʀɪᴇɴ

HIYA @QUART

zyabin101

@zyabin101 Speaking of GitHub, I don't know what to do for zyabinVI :/

quartata

14:16

I leave for eight hours and come back and everyone's pinging me

Lembik

@quartata too popular!

Cᴏɴᴏʀ O'Bʀɪᴇɴ

You cannot leave @quartata

You are here to stay @qua

New Main Posts

0

Q: KOTH: Everybody loves tokens

In this game two players compete to eat the most points worth of tokens, but there's a twist! Eating multiple tokens in a row of the same color gives an ever-growing bonus, but watch out, or your opponent will thwart your plans by eating the tokens you want before you can! Rules: 1 versus 1 n ...

grid king-of-the-hill

New Meta Posts

14:37

0

Q: Add a tag for brackets?

I ask simply because these brace/bracket/parenthesis questions are more numerous than the sands of the sea.... Should we create a tag for brackets? I don't think they're all that interesting (so I don't want to encourage it), but maybe having a tag for it will make people less likely to rewri...

discussion tags

Dennis

@wizzwizz4 Current consensus is that invalid answers should not be posted. The bounty doesn't change that. Sorry...

Btw last seen keeps updating as long as you have a chat tab open somewhere. It doesn't mean the user is actually active, or even awake. ;)

Mars Ultor

@zyabin101 Make your own esolang like most other people

Solver

@Dennis BTW you still need to update your Jelly language competition post XD

Hi @quartata!

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@zyabin101 I can help! I've probably made the most esolangs here :P

HEY @QUARTA

Mars Ultor

@CᴏɴᴏʀO'Bʀɪᴇɴ Wait, what? How many?

zyabin101

14:46

"Stop abusing my mention counter!"---quartata

Solver

@CᴏɴᴏʀO'Bʀɪᴇɴ So far I've been working on a working lang, but an esolang: CHALLENGE ACCEPTED :P

zyabin101

@CᴏɴᴏʀO'Bʀɪᴇɴ Then contribute something to the org. :)

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@MarsUltor At least 10, probably around 20

Mars Ultor

Hello @qüärtätä

@CᴏɴᴏʀO'Bʀɪᴇɴ That's too many

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@MarsUltor Most of my semi-documented languages

muddyfish

14:50

Can someone help me debug some code?

zyabin101

@CᴏɴᴏʀO'Bʀɪᴇɴ That's a lot of langs :O

muddyfish

pyke.catbus.co.uk/code?hFD+4<;*)&input=5

Mars Ultor

Um

Cᴏɴᴏʀ O'Bʀɪᴇɴ

@muddyfish Dammit I'm in class!

muddyfish

lol

sorry

Cᴏɴᴏʀ O'Bʀɪᴇɴ

14:52

it's good

I muted it before it played too much :P

muddyfish

Well

I got 3 people

ven

I'm goddamn tired of SO kicking me out of rooms after a day.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Set up a bot that types a . every N hours

zyabin101

@CᴏɴᴏʀO'Bʀɪᴇɴ Uhh...

Solver

Engineers do that by using an Arduino

zyabin101

14:55

I don't think bots are allowed in rooms other than the Sandbox.

Solver

I do it with my hands

zyabin101

And/or site-specific bot sandboxes.

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Set up a bot that stars a random message every N hours :P

zyabin101

@CᴏɴᴏʀO'Bʀɪᴇɴ Uhh...

^^^^^

^^^^

Cᴏɴᴏʀ O'Bʀɪᴇɴ

Set up a bot that display random "Set up a bot" message

zyabin101

14:56

@CᴏɴᴏʀO'Bʀɪᴇɴ Uhh...

^^^^

Cᴏɴᴏʀ O'Bʀɪᴇɴ

This conversation could've very well been between two bots :P

Solver

Bye bye @zyabin101

Cᴏɴᴏʀ O'Bʀɪᴇɴ

bai

Solver

Good luck with @CᴏɴᴏʀO'Bʀɪᴇɴ bots

quartata

@Dennis Is there a better way to create base256 compressed strings aside from doing it by hand?

zyabin101

14:57

@Solver Bye!

quartata

I'm about to add something super spicy to Cinnamon Gum

Dennis

@quartata For what? Bubblegum?

quartata

@Dennis Yes

Dennis

import fileinput, sys

def bb96decode(bbytes, a = 0, s = []):
    bbytes = list(bbytes)
    for byte in bbytes:
      if byte == 10:
        byte = 127
      a = 96 * a + byte - 31
    while a:
      r = a % 256
      s = [r] + s
      a //= 256
    return bytes(s)

input = ""

for line in fileinput.input():
  input += line

sys.stdout.buffer.write(bb96decode(input.encode('utf-8')))

quartata

Oh, thanks

Solver

14:59

@Dennis You wrote that right now, didn't you?

Transcript for