« first day (1651 days earlier)      last day (3209 days later) » 

3:00 PM
As soon as I figure out how to get 31 for n=5 I will post :P Right now I can easily calculate n=9 in ~6 seconds but it's not going to be "right"
 
A quick ballpark: >= 46 for n = 6
 
@Sp3000 nice!
@Vioz- if you read the rules you don't need to be right :)
just better than everyone else :)
 
@orlp Yeah, only problem is the last few thousand are the hardest :/ Or at least that was my experience when I tried a greedy algorithm
 
@Sp3000 I have found 3 81s so far btw
 
Same, mainly due to swaps
 
3:04 PM
how many problems are yoyu guys working on at the same time! :)
 
@Lembik sometimes you just get a bit too invested in one particular problem :)
21
A: Shortest universal maze exit string

orlpC++, 97 95 93 91 86 83 82 81 characters My strategy is fairly simple - an evolution algorithm that can grow, shrink, swap elements of and mutate valid sequences. My evolution logic is now nearly the same as @Sp3000's, as his was an improvement over mine. However, my implementation of the maze...

@Sp3000 ohhh now I know why it got stuck
 
@orlp :)
 
How come?
 
@Sp3000 still had some old code leftover in a condition that kept track of the 'best' so far (which was pure size of the string), and that disallows growing beyond 10 more than the best
 
... right
 
3:07 PM
because otherwise the road from 81 to 100 is 2 seconds
:D
 
From 81 to 100 or from 100 to 81?
 
@Sp3000 if you don't limit growing
 
@Lembik If you want more ballpark estimates, >= 75 for n = 7, >= 106 for n = 8
 
you're going to hit some equilibrium point
 
@Sp3000 that is very interesting. you may win with those .. who knows
@Sp3000 what language are you using?
 
3:09 PM
@Lembik python
from what I know about Sp3000
 
I don't even need to answer :P
 
:)
 
I usually prototype in Python too
 
I look forward to your answer :)
@orlp and then execute in?
 
unless the problem has some really elegant bitwise implementation
 
3:09 PM
Probably not going to give an answer unless nobody does better for a long time, because this is a pretty inefficient method over here
 
@Lembik Python, if it's fast enough, otherwise I port to c++ for multi-day simulations
 
Only using it to give rough estimates for small n
 
@orlp in principle there is cython and numba but I don't see people using them here
 
Correction: >= 107 for n = 8
 
nice!
 
3:10 PM
@Lembik PyPy is the easiest way to speed up your Python number crunching
@Lembik I have used Cython in the past, but usually only to interface with C code
 
@orlp that's only if you exclude all non pure python from your code
 
Not really that nice, it just means that it's not as "easy" to hit any optimal solution as I initially thought
 
I normally have at least one module it can't cope with
 
@Lembik and if you know what you're doing, PyPy is probably the best of both worlds for a lot of problems
(fast prototyping at C speed)
 
@orlp I love pypy but I have to disagree
 
3:11 PM
At not-quite-C-speed :P
 
err
woops
 
@orlp to get good speedups in pypy you have to avoid all things like itertools and also all modules like matplotlib, igraph etc.
 
and if you know what you're doing, >>> numpy <<< is probably the best of both worlds for a lot of problems
that's what I meant
 
ah ok :)
 
I need to get learning with numpy... need to reinstall it
 
3:13 PM
For n = 9 I can confirm the max is >= 158,
 
I find numpy a little mysterious
take this random example
b = np.ascontiguousarray(a).view(np.dtype((np.void, a.dtype.itemsize * a.shape[1])))
_, idx = np.unique(b, return_index=True)
??
 
btw @orlp did you do the maze sorting thing where easier to fail mazes are at the front?
 
46
A: Find unique rows in numpy.array

JaimeAnother option to the use of structured arrays is using a view of a void type that joins the whole row into a single item: a = np.array([[1, 1, 1, 0, 0, 0], [0, 1, 1, 1, 0, 0], [0, 1, 1, 1, 0, 0], [1, 1, 1, 0, 0, 0], [1, 1, 1, 1, 1, 0]]) b...

:)
 
@Sp3000 pip3 install --user numpy scipy matplotlib scikit-image requests pillow beautifulsoup4 scikit-learn sympy
@Sp3000 yes
@Sp3000 two lines of code :)
 
for n=10, >= 218
 
3:14 PM
My problem is more the fact I'm getting errors on installation which I need to sort out
 
@Sp3000 in solves_all, all you have to do is add std::swap(mazes[i], mazes[i/2])
 
@Vioz- I am hoping you will post an answer with all this in
 
after failing a maze
 
Ah k - I took a bit of a different approach, but it should be similar (chuck a thousand random strings at the mazes before beginning and sort by #failures)
 
@Sp3000 that's more complex and worse
 
3:17 PM
Complex, yes, but it does roughly as well I'd imagine
 
@Sp3000 no
@Sp3000 random strings are of bad quality and thus will fail on easy mazes
as you get in the low 90s/80s, different mazes start becoming the failures
 
Well from testing most strings fail within the first 100 mazes or so...
 
@Lembik So I can post code that doesn't hit 31 for n=5, but does do up to n=11 in a reasonable time?
 
@Sp3000 mine are around 20-50
def check(string):
    for i, maze in enumerate(mazes):
        start, finish, walls = maze
        cell = start
        for c in string:
            cell = move(c, cell, walls)
            if cell == finish:
                t = mazes[i//2]; mazes[i//2] = mazes[i]; mazes[i] = t
                return True

    return False
you don't even need that global there
 
It's more a reminder to myself
 
3:22 PM
but it's literally two lines of code to do it in check :)
and is guaranteed to be optimal
rather than testing random strings and praying that those random strings failed on the same things your small strings will
 
I haven't actually tested swap but I'll see how it goes
Okay, overhead's negligible and it does slightly better, swap it is
You know, something else I've been considering is somehow excluding patterns which cannot be part of a solution
e.g. NNN and such
 
@Vioz- ah.. you need to get 31 to get a higher score :)
@Vioz- do you understand why it doesn't get 31?
 
@Sp3000 they get picked up so quickly anyway, I don't think it's worth spending time ruling those out manually
 
I'm not sure how well that'd fit in with the current algorithm though
But if there was an algorithm that built up solutions then it might help
 
I have no idea. As far as I know it's just like Martin's approach
 
3:35 PM
vv incoming another useless bonus except for golfing langs
 
0
Q: Magic Square Challenge

SirParselotA magic square is a square where the rows, columns, and main diagonals all add up to some number n. Your Task Create a solver that given an unsolved square as a list and some number n on stdin will print the solved square as a list. You should be able to put any solvable square in it and get a...

 
Lembik, I think I've figured out a way to get 31. I'm now testing my original function, but instead of starting off at a bunch of zeroes, I'm seeding with a random starting point.
 
@Vioz- >= 225
 
For which n?
 
10
But it's almost definitely >, based on previous results
 
3:49 PM
Yeah, still trying to figure out how I'm 2 short with the same bruteforce Martin is doing
Also, can say n = 11 >= 319
 
You'll have to give me a while for numbers that big :P
 
Mine with pypy takes ~3 minutes for n=11 :P
 
Mine will probably take 10. Slow brute force is slow.
 
What I'm getting: pastebin.com/MHPDMCCK
I mean, mine isn't particularly fast
all(any((U>T*2)+(T>U*2) for U,T in zip(M,B)) for B in R) is basically it
 
(On a side note, is anyone working on a Bogo-magic-square-solver? :P)
 
3:54 PM
Is that where you buy one and get one free? :P
I just assume there's something builtin to Mathematica for this, so I'm waiting for the IsMagicSquare[Permutations[n]]-like answer ;)
 
Got my bogo solver up and running :P
Not sure how 3 people think this has been posted before.. codegolf.stackexchange.com/questions/54239/the-untouchables
 
4:34 PM
@Vioz- >= 321 I believe
 
Woah, what's happened to the questions per day? 😨
1652
Code Golf & Programming Puzzlescodegolf.stackexchange.com

Beta Q&A site for code golfers and for those who interested in code golfing (from beginners to experts), and programming puzzles.

Currently in public beta.

 
We have our highs and lows.
 
Oh I thought it stayed constant above 5
 
I'm pretty sure it was below 5 for a long time
 
Any time it's much above 5, it normally means we've been having a lot of poor challenges posted, then closed.
Having >5 good ones per day would be a new thing I think.
 
4:43 PM
The closed challenges count towards that?
Seems flawed :P
 
btw need to head off soon, but if the OP could clarify the example I gave in the comments, this question might be good to go
 
It's just that the questions per day metric simply doesn't apply to codegolf
a question on codegolf requires way more work than an answer
 
(usually)
 
and while on other sites answers are the once providing value to most users, on codegolf it's the questions that provide value
 
I agree. After spending so long on PPCG, it feels wrong to give answers which require so little thought and effort on other sites.
 
4:48 PM
Depends on the question :P
@BetaDecay I thought we had 1652 upvotes, but then I realised it's days in beta...
 
Aim for 3000. We can do it!
 
Except for that whole "graduating on Aug 27th" thing.
 
Maybe we'll stay in beta forever. Then we'll all stay with the same privileges and not have to worry about a silly theme ;)
 
Oh hey, Sexuality started private beta
Also... law has one now?
 
4:51 PM
Yea, I've peeked over at law a few times.
@BetaDecay We should move into gamma phase instead. Nobody gamma tests long enough IMO.
4
 
And gamma sounds cool
 
Way better than anything with 'beta' in the name anyway :P
 
:O
I'll just change to Gamma Decay then
@Optimizer
 
Sounds like Donkey Kong's Grandma.
 
Can we work our way all the way up to omega??
 
4:55 PM
That reminds me of Planet of the Apes for some reason
 
@trichoplax I have a feeling the degree of interest will hit its maximum in delta.
 
Everything changes with delta
3
 
@trichoplax Are you a sloganeer for Delta Airlines now? :P
 
Hopefully we can get past delta, though. Epsilon is where you find most errors.
 
All this talk about deltas and epsilons is really testing the limit of my patience
3
 
5:01 PM
Calculus pun? Oh God, it's this again
:D
 
Just let it happen.
2
 
Well that sucks. I don't know calculus, so I have to decide whether to power through this series of messages or just diverge altogether.
3
But, I guess that's normal.
 
@BetaDecay Calculus puns are a pretty integral part of mathematical humour
 
They're pretty easy to differentiate from other mathematical puns.
 
These puns are becoming derivative...
 
5:06 PM
Great minds are bound to converge at some point though
 
@MartinBüttner At least partially.
 
@PeterTaylor thank you for the title improvement!
 
It's probably approaching a local maximum, though.
 
although you have made it look a little scary now :)
 
I'm looking forward to Peter Taylor's contribution to this series of comments
 
5:07 PM
Quick! Umm... Err... Squeeze Theorem... Umm... Damn it.
3
 
@Vioz- :)
 
@BrainSteel That was a pretty limit inferior attempt
 
We all need to get real and make these puns less complex
 
@Sp3000 It's really a saddle point in the conversation.
 
don't get all hyperbolic
 
5:10 PM
These puns... sine...
 
Things are getting diracstically out of hand.
 
there's no hope Riemanning
 
I Laurent the loss of humor in these puns.
 
Are Euclidding me?
3
 
Why Conway Look and Say what happens?
 
5:13 PM
Kill me now
 
My secret to getting stars is being terrible at puns
 
@BrainSteel I believe you mean incredible
 
I like to think of myself as good at terrible puns instead of terrible at good puns.
 
are y-Escher?
 
^ nope
 
5:14 PM
haha
shot down
 
are we having pun again?
 
No pun fortunately
 
4 mins ago, by Geobits
Kill me now
 
parse error
 
Apparently not
 
5:17 PM
It's always me ;(
 
I guess it's time to spin the wheel again
 
catch(ChatTimeError e){ print("Ugh."); }
 
Spirit Bomb!!!
 
@Sp3000 Did I win??
 
@Geobits You win a whole mess of blame. Congratulations!
 
5:19 PM
@Geobits You win Alex's swag. You have to break into his house first, but you still won it ;)
 
@BetaDecay I'll make sure to bookmark this so I can bring it up during my deposition.
 
@Geobits You win a new question vv
 
He'll probably just sneak in like my neighbor's cat and I'll find him in my living room.
 
@Sp3000 Plan ruined :D
 
Have people from PPCG ever actually met in person before?
 
5:20 PM
I've met myself
 
@BrainSteel I go with @Doorknob on all of his family vacations.
2
 
I've met my dog. Sometimes he likes to chat in here.
3
 
I've met one ex-PPCGer, if that counts
 
I met 1 or 2 people... on irc
 
@Geobits On the Internet, nobody knows you're a dog. We certainly didn't know he was a dog.
 
5:21 PM
I've tried to get people I know to come here, but I don't think that counts.
Yeah, I find hitting my head on a keyboard comes with a 50% chance of coming up with a real message from Geobits.
 
New Q is surely a dupe of something?
 
Wait for it....
 
@BrainSteel The other 50% are Pyth programs
5
 
New Q is in Star Trek Voyager methinks
 
5:24 PM
Who said anything about Star Trek?
 
1 min ago, by Sp3000
New Q is surely a dupe of something?
 
I thought...But you...and Q...and....
 
why does he look like a French commander from the 19th century?
 
Q's fun like that.
 
Picard surely would disagree.
 
5:25 PM
Or twisted, sadistic, or just bored. Take your pick.
 
@aditsu You are not worthy to be in this chat room if you don't know the answer to that ;)
 
@BetaDecay That's no way to recruit people ;)
 
@BetaDecay He invented CJam. He can do whatever the hell he wants.
 
it's been maybe 20 years since I've seen Q on tv, I forgot some details :p
 
5:26 PM
You're forgiven.
At least you knew at some point
 
I have seen Q in this chat room
 
5 mins ago, by Alex A.
^ Q
Found it
 
@Sp3000 Related
 
0
Q: Arithmetic... tock... tick... tock

nobillygreenThis question brought to you by a game I like to play when stuck in long phone meetings. Given any two times from a 24 hour clock (from 00:00 to 23:59), how many valid mathematical equations can be generated with all of the times in between using only basic arithmetic operations? Input: two fou...

 
It's about.... time.
 
5:32 PM
.. I get back to my arithmetic questions ..
 
@BetaDecay Did you make that?
 
@BetaDecay What on earth is 2^-1 mod 7 o'clock?
 
@AlexA. Nope, that's someone else
@Sp3000 I have no idea. None of the equations are correct IIRC, but it looks nice :)
 
Well the other 4 make sense, that one just bugs me
 
Ahh maybe it's just some
 
5:34 PM
... oh it's field inverse, right. Not 1/2
 
I'm guessing it's about half o'clock.
 
So 4 o'clock
 
I'm thinking it's probably the left half
 
yep, 4 should be correct
 
... I'd get that if there were about 1000 of these expressions so you'd have to think every time
 
5:37 PM
@AlexA. These are the three that I've put up:
 
by the time you figure it out, it may be obsolete
look what I found
 
2^-1 mod 7 again :D
 
Same thing for 10 also.
 
I've yet to seen one of those clocks I'm impressed by :( they're all too simple
 
Looks like somebody got "inspiration" somewhere....
 
5:40 PM
but I like this one:
 
haha
 
A friend of mine had a backwards clock. It screws with your head in a serious way.
 
like a mirror image?
 
Kinda. The numbers go in reverse order and he hands move anticlockwise
 
5:45 PM
This is the watch I want:
The Citizen Astrodea
 
@BetaDecay a clock that goes anticlockwise.. that's messed up :)
must be an anticlock
2
 
It's the perfect loophole for any challenge that includes "clockwise" in it
 
@BetaDecay a watch that can show you the universe but not the time?
 
@aditsu Well you can guess from the way the hands are oriented
 
You mean the hands aren't saying it's Jan 12? :P
 
5:49 PM
There's a lot of maths clocks, but trying to look up "programming clock" is pretty underwhelming
 
more likely, from the position of the stars :p if you're hardcore enough
 
Apparently I program in 2 o'clock
 
@Geobits Actually I think it's mid September
 
@Sp3000 there are binary clocks
 
@Sp3000 There's a gap in the market for you then :D
 
5:52 PM
My first attempt was "Group theory clock". Not very interesting.
 
@BetaDecay What if he doesn't shop at Gap?
 
Steve Wozniak wears a pretty weird watch
 
Haha what is it?
 
That's cool
 
5:54 PM
Nixie, nice
 
> If I wanted to buy a watch that guaranteed I would never get laid, I certainly wouldn't have to spend that much on it.
 
I'd like my microwave timer to look like that
 
@Geobits you know.. I hate to say it .. but....
@Geobits some women are nerds too :)
2
 
Not my opinion... just a funny quote from the page linked.
 
5:56 PM
I can't imagine what @Geobits wife/girlfriend looks like
Or maybe he just split in half like an amoeba
 
@BetaDecay Are we really going to go personal...?
 
Looks like we need a timeout
 

« first day (1651 days earlier)      last day (3209 days later) »