« first day (1377 days earlier)      last day (3762 days later) » 

06:05
Not very good yet:
0
A: Sandbox for Proposed Challenges

hosch250Algorithmic Sudoku Solver The challenge here is to create an algorithmic Sudoku solver. There are several possible methods of solving Sudoku, including brute-force, backtracking, exact cover, stochastic search, and constraint algorithms. None of these are banned except the brute-force method. ...

You might want to be particular about what counts as "brute force"
0
A: Sandbox for Proposed Challenges

hosch250Algorithmic Sudoku Solver The challenge here is to create an algorithmic Sudoku solver. There are several possible methods of solving Sudoku, including brute-force, backtracking, exact cover, stochastic search, and constraint algorithms. None of these are banned except the brute-force method. ...

I mean, even with the ones you mentioned there's a big difference between backtracking and heuristics
Yeah.
Out of curiosity, does brute force work in 10 seconds? Backtracking alone, I mean
06:08
No.
Well, yes.
I once helped someone debug a stack overflow in their sudoku solver.
It did if enough values were entered.
Hmm btw have you looked at existing sudoku posts yet?
To see what's around
Yes. Just code-golf.
Peter Taylor said we did not have any algorithmic solvers yet.
Ah right, so brute force wouldn't be useful anyway
yesterday, by Peter Taylor
@hosch250 I don't think so, but you need to be careful how you word it. And expect disputes about Nishio.
yesterday, by hosch250
Have we ever had an algorithmic solver?
Why not the fastest-code tag then? And an idea could be to include time complexity somehow, in addition to clock time
06:12
Yeah, good idea.
Oh wait time complexity wouldn't go well would it...
Bit hard to prove how much a heuristic helps :/
Hmm.
Theoretically, the program should be able to solve it in under 1 second because there will only be one solution.
I think I will leave it at 10 for now though.
:P
I'm trying to find that Wiki page that hard the computationally difficult ones
Can't find it :/
What was it about?
It had test cases which typically took longer thank most other sudokus to solve
So I thought they were good test cases
Ah found it, the Wiki page's changed a lot since years ago
06:15
OK, will stop looking then :)
:P trying to trace it back to a source
What is the page, maybe I can help.
Check the bottom of the page
Nice :)
06:21
I could change the 0 to . and just copy/paste from the links: magictour.free.fr/topn87
Thanks :P
That could work :)
But yeah I just thought harder test cases might make the question more interesting :)
This should be a good read: ams.org/notices/200904/rtx090400460p.pdf
I would probably create dynamic arrays for the possible values, then just choose one. For programs with multiple solutions, I would use look-ahead.
I mean, fill the square(s) with only one possible value, then update the arrays.
I'm not thinking straight anymore - past 12.
Python would be good at that if value in List:.
I did a Sudoku Solver for that project euler question ages ago - it was when I started Python, so it looks horrendous to me now
06:27
:)
(I did implement some nice heuristics though, based off stuff like two cells with the same possible pair of digits left)
I should write a contest to have people's algorithm's play cribbage against each other.
KOTH style.
I will provide the shuffling.
Just a perfect split with alternating swap executed a random amount of times within a limit.
See you.
:) night
 
1 hour later…
07:48
@MartinBüttner I do know JS, but I haven't worked out how that answer is discarding the {}.
@MartinBüttner I won't say I wasn't tempted, but they'd be rather long.
@Sp3000 I wrote a non-brute-force Sudoku solver in an attempt to estimate difficulty. It only knows 4 basic rules, though, none of this advanced X-wing etc. stuff.
Ahaha yeah, I tried working out some of those X-wing stuff and I gave up
Also that Golfscript.
08:09
0
A: Sandbox for Proposed Challenges

KennyTMDecompose a number with bit-xor without the digits 0, 3, 7 code-golf Challenge Write a function or program which takes a positive decimal number, call it A, and output two positive numbers, B and C, such that: A == B bitxor C B and C must not contain any of the digits 0, 3 or 7 in its decimal...

 
2 hours later…
10:27
@PeterTaylor I figured he probably just wrapped the entire code in them... but I'm not sure what to do with the rest.
@MartinBüttner Welcome! Going to post another CJam quote? :)
hehe, I'll see
@Sp3000 Dennis has already done that ^^
but apart from that there aren't really any new ones that I'm interested in cracking, and I doubt I'm able to crack Dennis's
Ahaha k
Well I was trying the Golfscript one - there's an interesting pattern to the numbers
But unfortunately I don't think I know enough GolfScript :(
10:37
what's that pattern?
1234(blah)...
4582 4583 4584 4585 4586
(blah)...
142924 142925 142926 142927 142928
302928
497409 497409
grc
grc
@Sp3000 I like your last python one :)
ie there's two sections which are 5 consecutive numbers
@grc I just had to do it ;)
I also like my 37 one - if it manages to get through the week I'll be surprised, because I don't expect short Python programs being able to manage unless people start getting bored
the first four might also be 1 2 3 4
Yeah, probably is
I got really confused about where to put the two commas though
10:43
you could make another list with it and discard it with ?
Oh? Hmmm
I think I've been using ? as pow this whole time because I thought the ascending numbers were relevant, not sure
or wait, maybe GS doesn't have ? as the ternary operator
oh yeah it doesn't, so nevermind... that's CJam only
"? raise to power, index, find" is what I've got in my notes here
yeah that's correct
ugh, / has some weird "unfold" semantics... I thought it's just "foreach"
I think the closest I got was 22,{.4&.?+}/ which is missing a comma
10:50
, could be "map" instead
"Oct 31, 2011: Happy Halloween! '?' operator now works correctly for finding substrings in strings, or strings in arrays. Thank you Peter Taylor!" <- it'd be funny if that was a hint
I think map needs a condition, which makes it a bit hard
Also if you map you can't loop, right?
what?
map is a loop
and why does it need a condition?
oh wait
it's different map semantics
the map I was thinking of is % (just as in CJam) and , is actually "filter" (as in CJam)
Sorry not condition, another block or something
(looking at the site's example 10,{3%}, -> [1 2 4 5 7 8] )
... I'm just confusing myself, don't mind me
10:54
@MartinBüttner I didn't feel anything wrong reading this, because yes you can discard it if the previous item is a smaller nonnegative integer
@user23013 ummm... example? :)
7 9,? returns 7 in GolfScript.
oh, right
so I think we're looking at this sequence of numbers:
1
2
3
4
260
885
2181
4582
4583
4584
4585
4586
25322
53883
92299
142924
142925
142926
142927
142928
302928
497409
497409
that's 23, which is a good argument for starting with 22,
the step from 142928 to 302928 is also pretty mysterious
142928 - 302928 = 160000 = 20^4
So pow 4?
let me check the others
yes
the step from 92299 is 15^4
10:59
Huzzah we have a pattern
before that it's 14^4
So it goes up by powers of 4, but not consistently
so that step isn't always taken, but if it is, it's the fourth power of the current index
Hang on... going up by 1 is just 1^4
Maybe that's where bitwise & comes in?
let me see which ones are there
(1), 4, 5, 6, 7, 12, 13, 14, 15, 20, 21
11:01
That's the third bit from the right, isn't it
n & 4
grc
grc
@Sp3000 hmm... StringIO or os.devnull?
or are they both red herrings?
@grc Ahaha no red herrings in this one, you can relax on that :P
oh, and we also know that the block leaves a copy of the current element on the top, so we probably have a . at the end
@Sp3000 yes, you're right. which also means that 1 is not part of those numbers.
okay, I think this is where the second bloody comma comes in :D
do you have it?
11:08
Nope, you?
no, I thought that :O was some realisation ^^
That was at your realisation :P
lol okay
actually I think the end of the block might be doing some more preparation than that
ugh, this would be so easy if we had a / or @ ^^
I need a debugger so I can see what's going on with the arrays :(
actually, I'll move to CJam now
I think the relevant operators should be the same
and I can see the stack there
11:13
Hm?
or maybe not... it seems that GS has unary +, which we might need
which is actually a really good clue
Surely we need the + to add though, right?
yes, but the first iteration is a special case, because we don't have the stuff from the last iteration on the stack yet
I don't think GS has unary +. You are probably adding something with the empty input string.
ah, right, I forget about that
oh, wait
TADAAAAA
@user23013 thank you so much
the , is used to turn that empty string into a 0 to kick things off
,22,{.4&?+.}/
11:20
:o
Damn, would never have guessed
@grc Nice! Exactly the same, barring the interchange of u and _
actually, the next thing I was going to try is what happens if I put , at the beginning, but I didn't think that the stack wouldn't be empty
thanks @Peter, that was a really fun puzzle :)
@Sp3000 do you wanna post it? you figured out the sequence.
:P nah would never have gotten it
I was stuck on 22 being first
(just thought it'd help you solve it though, I'd like to see the scrambled backlog cleared because it's long (I'm not helping there though))
heh, okay
Almost wish SE had collaborative answers, but then again codegolf's just a special place
we should probably look into Dennis's next ^^... but I need to shower and eat something
11:29
:P k
Hmm now I'm out of Python tricks
grc
grc
11:52
Ahaha nice
grc
grc
anyway I'm off
thanks for the cool challenges @Sp3000 and others
:) cya grc
@Sp3000 added an explanation and stuff: codegolf.stackexchange.com/a/41143/8478
For some reason my head didn't think that 4& pushes either 0 or 4 back on the stack, and then you pow that
And it was trying to do 4& and 4? at the same time
XD
12:02
okay, let's see what we can find in Dennis's code
B and J are variables initialised to 11 and 19.
hi @PeterTaylor
That's a lot of different letters there :/
I think it should be just something like "Beware the aberwock, my son!"Jb
does your code have an early stopping rule or does it check all possible sums and the count?
@user23013 There's probably a bit more to it, don't you think? (I was actually looking for a quote with b myself, but then Stop, Hammer time! was just too good :D)
Oh yeah... I know what more there is to it: the order of the letters in the string -.-
and it's probably "..."JbBb or "..."BbJb
although just Jb does give the right length
12:08
So how are characters converted?
they are implicitly converted to code points
and then to base 19 or 11. and if you have another [BJ]b after it, then that number is converted back to an array of digits (which are printed back-to-back)
Jb has the result with the right length, I think I should try that first
yeah, that's likely
the trouble is that the char codes are all bigger than 19, so it's not trivial to reverse the operation
but if it's any help, here are the "normal" digits in base 19
{6, 7, 5, 0, 7, 5, 10, 2, 0, 8, 11, 0, 12, 12, 17, 18, 8, 10, 15, 15, 11, 17, 14, 11, 12, 3, 1, 7, 6}
12:15
Hmm I'm confused why "a"Jb gives 97 - does 97 not get converted to base 19 or something?
It's converted back from base 19.
Oh.
"a"Jb is the same as [97]Jb, which means that it's trying to convert a base-19 number with a single digit (being 97) to decimal
Single digit being 97 XD
In fact I posted something like this in the tetris question: codegolf.stackexchange.com/a/40171/25180 which should look like 95bKG#b:c , and has the right length, but in fact it was 5bPi~C#b:c
12:20
okay, I think I've found a way that should find the correct solution fairly quickly if automated
but since I'm having breakfast now, user23013 will probably beat me to it :P
@MartinBüttner I reckon the entire code might be wrapped in () to avoid the REPL giving an error, but it would really help to know what REPL he used to test. Or for OP to intervene and declare that a REPL doesn't meet spec.
@user23013 the number also has 29 digits in base 19, which is also correct (the string length is 28)
I found myself too lazy to do high precision base conversions in a sane language after learning CJam. Better start to learn Python... Or just golf a search program.
if you can't be bothered I'll do it in Mathematica later
I'll take my time to learn Python or have a look at Mathematica then
12:40
Would anyone mind me promoting a non-ppcg question of mine here?
13:00
But I did that anyway.
ah you got it
did it by hand?
By program.
k
okay, cleaned up the backlog a little
Woo
Still long :P
yay, silver badge for Avoiding Rivers :)
13:07
Nice :)
it's time for a new CJam quote
Ahaha maybe I should join the CJam fun
Did you come up with something that is possible doing by hand? I searched for that.
@user23013 Starting from the least significant digit, pick letters whose code point mod 19 gives the the right remainder. That branches out a bit though. Not a lot, but it was slightly annoying to do by hand.
My program just did that. But it should still have some branches
13:26
grrr. I wanted to do Programming Puzzles & Code Golf (without quotes), but there is not a single token that includes an n in CJam :/
guess I'll have to add quotes
Oh yes, but 'n then
Wish we could have joint answers and just demolish the backlog :P
@user23013 Yeah, I think I'll use single quotes
13:43
@Sp3000 What do you mean?
Answers tied to more than one user or something
right
yeah I don't see that happening
we can still demolish the backlog though ^^
Ahaha k
Most of the remaining ones are in languages I'd probably have to actually do something to obtain :/
there's still that 10 byte ES one
Is it hard? Or just poorly specified?
13:47
hm, I'm not really getting anywhere with Programming Puzzles & Code Golf (either single or double quotes)... all the cool things I could do with those letters are missing some other letters to set up the stack
@Sp3000 a bit of both I think
judging by his other answers, it's also just something you paste in the browser console
but even so, I don't know what to do with the letters that doesn't yield an undefined reference error
Maybe 1e<something>?
interesting
no that's not valid
Is it easier in Pyth?
13:53
I don't know Pyth, but possibly
feel free to post it if you find something interesting
I don't know Pyth either, although I think I was the second person who posted an answer in Pyth in this site.
Oh hell
btw can't the ECMAScript one be brute forced in some way?
Or does console only make that not possible
it's probably possible, but I can't be bothered to write code that iterates through all permutations in JavaScript
I took some code for permutations and it crashed my browser tab due to memory :(
yeah, you'd have to generate them one by one
from itertools import ...

:(
14:23
lol yeah
@Sp3000 I've just tried that. Every single permutation threw an exception when I tried to eval it.
(Chrome)
Ah :/
Thanks
@PeterTaylor You should leave a comment with that on the answer.
Either my code's wrong, or FF doesn't work either
14:36
Anyone would like to remove the last PHP one. I seem not so great at brute-forcing big number output… (codegolf.stackexchange.com/a/40990/9409)
@Sp3000 you were looking for Python submissions? codegolf.stackexchange.com/a/41155/8478
Oh my
Bets $10 that we have from future import braces
@bwoebi I can see a printf in there but that's as much time as I'm willing to spend on it right now ^^
I guess also a for which means that the printf must be inside the for
@MartinBüttner For your Programming Puzzles and Code Golf idea, Pyth could do something with it. You have G which contains "abcdefghi....", the char->int function, map, the prime factor function, sorting and sums, etc.
@FryAmTheEggman Go ahead then ;)
14:42
Then leading junk with a space character will suppress output
@MartinBüttner Yeah but I don't want to brute force operations order … On a small number, I'd be on board…
For me that's not a challenge, but just mind fucking brute-force…
So, I've got for(;;)printf("%d",); ... the remaining characters are $$$$$$$$+++,3557999<====dddooooo ... the format specifier can't be %o, because there are 8s and 9s in the output.
You can probably assume safely that the characters are paired individually with the dollar signs
14:58
(:/ I don't have enough x's)
@MartinBüttner exactly. But I don't like to brute-force out what number fits where
yeah, neither do I
that's what some people seem to have severely misunderstood about this challenge
15:13
well not really... from the cops' perspective the answers aren't meant to be cracked ;)
I almost want to withdraw my first submission because it's like that - pretty boring
But yeah that does make it hard to crack I guess (although I do like posting puzzley type ones now)
@MartinBüttner Eih, I mean, provide something crackable, but not only by brute-force...
15:39
Does anyone know of a simple way to test/run SAS online?
0
A: Unscramble the Source Code

Martin BüttnerCJam, size 19 Code "Can't touch this!" Output hunt 12tacos (Probably not as difficult as the last one, but I had to post a sequel.)

@Geobits no, but I haven't looked into finding one
I think Sp3000 had concluded he'd need to install a VM to test it, so he probably looked for online testers
Yea, after finally giving up (for now) on the APL one, I wanted to give the other short one a try, but I haven't had any luck. I can find a command line version for linux, but not on the right PC for that at the moment.
Yeah I found the site for SAS and gave up cos I don't have much disk space on my laptop
I feel like the APL one really can't be that difficult for someone knowing APL
whoa
I don't know it, but the basic operations seem simple enough. I can get all the right numbers in the wrong order, or 27 3 27 4.34..., etc. I just can't figure out how he got it that way.
15:47
Wow fast
Then again it was obvious what went in the string :P
well I had hoped more people weren't aware of t being cyclic
maybe I should have hidden a different character than !
I'll try to come up with a longer one tonight.
I should take a break from this for now...
seriously, I haven't been able to get anything done in the last 48 hours ^^
I haven't gotten anything done in the past week and I have an assignment due tomorrow
:D
16:05
0
A: Unscramble the Source Code

durron597Java, size 108 Code forprintmymostfaultSent66{{{}}}.......;""(((())))%,*++//\\--''''''''''''main static class void public String Output 3.141593

due for a decent java one imo
I realized the number in the APL answer isn't what I thought it should be after posting it. So it may be somewhat unexpected.
@user23013 Wait, what? Is the output you have listed correct? I've wasted hours on that one :P
It is correct, at least in the online interpreter
Ok, that's what I've been using.
But the number isn't calculated from the formula I had in my mind.
16:17
Ah. Well I can get the stupid number, it's just getting the others in the proper order that I can't seem to figure out.
Well it is supposed to be not easy to get enough 3
At one point I had 3 3 27 4.434..., but that was about it. I'm still baffled by the +. I can't seem to find anything it's useful for, so I've been assuming it's a red herring. Probably not, though ><
Ah, yes, I added the + in the last minute.
Ah, I shouldn't spoil that
That's not much of a spoiler. I'm not sure if you meant it is or isn't a diversion. If you'd like to clarify, that's fine by me ;)
16:38
hi
anyone here interested in the rice challenge? I would like to ask a variant where you have to play a dot on each grain of rice
but I don't want to restrict it to just these 10 images as it seems too easy to overfit
so I was wondering how to make a large number of variants of the images sensibly
It is a red herring. In fact the number I thought isn't supposed to be so easy to figure out what it is. But since it is obvious how you can get all these numbers without a +, I didn't think it would work...
any ideas?
How about actually taking photographs?
16:53
Let me know when my profile picture updates
Dick :P
Everyone will finally be able to see another 1/4 of your face
I'm pretty sure I didn't crop it dead center though, so it's probably going to look a bit strange.
I'm sure you look at bit strange anyway. We shall soon see.
Meh, better than a block of dirt
16:57
Tomorrow I think I'll change my name to VisualBanana.
Part of me wants to not tell you that you can only change it back after 30 days...
But the other part says it warns you anyway when doing it.
@Geobits Unless you change it only on one site, then you can revert it by going to another site, enter the display name you want and press "Save and copy to all sites".
@ProgramFOX Nice loophole. I wish I'd known that when I changed mine to "GenericHolidayName" last year :D
17:39
@Geobits What you don't know is that he didn't mirror the image... he actually managed to obtain the part of the picture you cropped away.
That would be pretty impressive, since the original didn't go that far over anyway (it's not cropped much).
Fortunately, since his needs to be to the right of mine, it isn't up to him to line it up. All I have to do to prevent it showing up right is not talk directly after him :D
(assuming others aren't helping him by removing themselves from between us)
who would do that?
I'm thinking you would :P
Anyone know why my Golfscript answer got downvoted?
@Beta Not mine, but it could be because it looks like "chucked chars at an interpreter and this is my output". Your comment of "I have no idea..." probably doesn't help that image.
17:53
Update the backlog... I think it's time those Ruby submissions were cracked.
@Geobits Say something
He must
Wait... it's still showing up as a block of dirt in the top right, but not in the bottom left
@Geobits Ahh okay. Any tips on how to phrase it?
no it's fine for me
17:57
Yea buddy. Any tips on how to phrase it?
dude... let people post after you...
haven't you been following the conversation? :P
Huh?
You talking to me or Rainbolt?
They are conspiring with Geobits to keep the images from lining up.
17:59
Ahh, actually my question was for Rainbolt ;)
Really? Because I was talking to Martin.
BTW, @Rainbolt, your image is much sharper than Geobits.

« first day (1377 days earlier)      last day (3762 days later) »