« first day (1412 days earlier)      last day (3433 days later) » 

1:22 AM
@xnor Found the bug, will post as answer in a bit
Thanks for the ping, as I'd already given up
 
2:19 AM
@Sp3000 how is the blurring going?
 
It looks better, but it's not giving me the same results as when I used Gaussian
Then compare to the Gaussian one in the post
 
hm, the Gaussian one does look a lot better
 
Yeah, I feel like I might have implemented something slightly differently, and I'm trying to figure out what
Either that or the Gaussian the box converges to is slightly different from the one I was using
 
does a larger kernel help?
 
I think I tried that last night and it didn't seem to. I'll keep playing around with it until I figure it out though
I think I'll output the blurred versions to see where the difference is
Hmm that's weird, I think I'm doing more horizontal blur than vertical
 
2:39 AM
it looked like this in your earlier picture
 
I'm surprised Gaussian actually worked though, after blurring all it does is take the closest colour out of itself and its neighbours
I wonder if I can do something about really thin parts of cells though, that could be something for later
 
hsl
3:16 AM
I think there is a decent challenge at codegolf.stackexchange.com/q/42289/29750, but it's currently on hold with no winning criterion
Am I allowed to edit the post to clarify what I think the OP meant (make it more clear what to do, make it code golf, etc.), or should I leave a comment/hope the OP returns to fix it?
I know we're against editing answers, but I can't remember coming across this for a question
 
if you think you can improve it, do so, and leave a nice comment, saying that you tried to adjust it to our standards, and that the OP should roll back/edit if he doesn't agree with the new version
 
hsl
will do, thanks
 
4:00 AM
@Rainbolt look at the sidebar
 
The awkward moment when nobody realised I got my counter and my splitter the wrong way around for the not sorted alternative
 
@Sp3000 lol
 
-1
 
@Optimizer he already said, it's gonna be his christmas present
 
;)
@MartinBüttner that was with the other lie that we are graduating
 
4:02 AM
no, that was Geobits' message
Rainbolt just said he's gonna change his name and destroy all tips questions.
I'd take him seriously on that :D
 
he said he is giving two things, name change and tips removal
and tips removal was because we are graduating
 
no, that was Geobits with the graduation
 
so tips removal is not because we are graduating
and will happen aanyways ?
 
Damnit Geobits :/ I was confused all through what was it... 3am?
 
@Optimizer actually Rainbolt was indeed the first to say we're graduating, but I considered that to be a joke, while being serious about the tips :D
 
4:04 AM
:P
double standards
:/
 
I took everything seriously :(
 
my nos and yeses too ?
 
... except that :D
 
:(
I was srs all along
 
:(
 
4:12 AM
Hey @NewMetaPosts, are you there? impatiently taps foot
 
Oh, weeklies!
 
@Doorknob冰 I even already got my email notification
 
I think I like the idea of putting the actual ideas/questions up on meta for people to see, but discussing in a chatroom (where we normally leave comments for the sandbox)
Or something like that
 
2
Q: A proposed schedule for weekly challenges

Doorknob 冰The response to the weekly challenge question has been a resounding "yes," so the next step is to decide how we'll manage the schedule. There'll be two simple parts to this, each taking one week: Collect ideas for a theme or concept to base the challenge on, and choose the highest voted after a...

 
@NewMetaPosts Finally. :P
 
4:21 AM
@Sp3000 that's pretty much what I wrote ^^
 
I was going to ping you saying the same thing :P
 
haha :D
 
and I was saying the same thing too
 
@xnor sorry
 
Great minds think alike ;)
 
4:22 AM
it's ok!
if you disagreed, that would be another matter :-P
 
For collab tools to further go into xnor's points, I think docs is unnecessary but Github might be good for more complex questions like KOTHs
 
@xnor then you could have kept your answer :P
 
Also @xnor Hi! Did you see my Pseudoku answer?
 
@Sp3000 yes, that was great
how long did you need to run it?
I'm still surprised that was tractable
 
@Sp3000 actually, let me add that, in the interest of a comprehensive answer ^^
 
4:24 AM
maybe i'm not just in the spirit, but why all the excitement on collaborating over questions?
 
Yeah me too - there's about 960-something polyominoes and you might think (960 choose 9) is huge, but then again Sudoku's NP-complete too and the only reason that's tractable is because the problem size isn't very big
 
i can understand for things like KOTH's, but I've found challenges to be mostly a "someone has an idea and someone writes it" type of thing
@Sp3000 it's all cool that you could find all the answers. it means you could add one or two walls and get a unique solution
 
@xnor I don't know why the excitement... it's just something I suggested, and people liked it :D
 
Probably because that way we can combine peoples' expertise?
And you can get questions where one person alone might have trouble writing it
 
I've never participated in a KOTH because stack exchange format is so horrendously bad for it...wish it could have a better way of organization
 
4:26 AM
@xnor I sure hope that's all the solutions - should be though, unless there's any more bugs...
 
@feersum one day we'll have a framework that automatically runs KotHs on some server, and a stack snippet that polls the results...
 
i guess i can see it if it's a "we make a big thing that's too much for any one person"
 
@feersum The BF Joust one went pretty well from memory, but that was because the OP was very dedicated
 
maybe it's also the curiosity of "how good a challenge can we make, if we're really all invested in it."
 
submissions should not be visible to other players IMO
 
4:28 AM
@MartinBüttner I'm more excited about it from the angle of "how good of a golf we can make if we're really all invested in it"
 
@xnor you mean as an answer?
 
yes, exactly
 
@feersum I like the meta game :)
but I feel like we've had this discussion before
 
@MartinBüttner either from collaboration, or having a single focal point for everyone to invest effort into
 
and I probably pointed you to the iterated prisoner's dilemma
 
4:29 AM
Maybe someone else should manage Chat Minigolf :P
 
me
 
@MartinBüttner well, I haven't had this discussion, and I want to say that I feel like the game-theory aspects in challenges i've seen generally take away from it being a programming competition
@feersum
 
it's a continual process of tailioring your solution to beat the other posted answers, which ends whenever OP gets bored of running simulations
 
@MartinBüttner actually, i'm still not convinced about this collaboration thing
 
@xnor doesn't mean we can't try it though, right?
 
4:34 AM
running the thing, sure
@MartinBüttner I guess
 
I mean, what's the worst that could happen?
it dies off? it doesn't work out and we scrap the challenge?
no harm done in either case
 
well, the bad thing would be having it collapse in a paint shed nightmare
when the idea was fundamentally good
 
but maybe we do come up with cool new challenge types or just KotHs that take some work... or finally develop a KotH framework in the process
 
ok, this discussion is inspiring me to make a new meta post
 
too much meta talk going on
 
4:37 AM
@xnor couldn't you just have answered doorknob's first meta question? :P
 
my meta .. bolism is acting weird.
 
@MartinBüttner i actually totally agreed with yours
 
Oh I didn't notice David updated the Mathematic PbN. Spheres looks nice
 
here's a question for everyone: what are things you don't think we have/do enough of in PPCG?
it can be question types or anything else
 
4:46 AM
Data structures, off the top of my head
 
is that a type of question?
 
Yeah
 
we need more fibonacci heaps!!!
 
It'd disadvantage a lot of languages though, which is why we probably don't have many
 
(not that I know what those are)
 
4:48 AM
so do you mean a question to implement a specific data type?
 
(I just know them as the wonky-heaps)
 
@xnor why do you ask? I wanted to keep these to myself until the call for themes on friday :P
 
Well as Lembik suggested yesterday, it could be something to do with the space needed by the structure. Or maybe it could be fastest-code on certain operations.
 
@MartinBüttner mostly curiosity, but also for the weekly challenges
 
"yesterday"
 
4:50 AM
Shh it was 3am at the time
That's yesterday
 
@MartinBüttner wait, you want to do a theme?
 
@xnor k, in that case, you'll have to wait a couple more days ;) ...
@xnor well I've got two rough ideas, so I thought I might post them as suggestions
 
i thought your meta post was saying that themes tend to wear thin
 
@Sp3000 the situation is not improved by me pulling an all-nighter right now
 
@MartinBüttner i guess i'll have to see what you mean
 
4:51 AM
@xnor that's why my suggestion was to do one challenge per theme
the original "weekly challenge" concept is to post a bunch of question surrounding the theme
 
@MartinBüttner My sleep times for the last three days have been 10, 10, 4. Now I get why people say SE is addictive
 
and I said that's not a good idea for PPCG, because we don't want 6 graph theory questions in the same week.
 
i don't know how to imagine a theme for a single challenge, but that's ok, i'll just wait and see
 
instead we should try to do one really good challenge around the theme
 
it sounds like your have some great ideas, looking forward to seeing them
 
4:52 AM
@xnor why is that difficult? a theme can be as simple as a tag.
 
but that's just coming up with the challenge, right?
 
@Sp3000 well I'm actually doing homework :D
 
:P
 
@xnor well... kinda... except a theme could be more general... I'd probably rather call it a vague idea (and I think I used the term "idea" in my meta post)
 
actually, i think i see the appeal of a theme now
restrictions breed creativity
so specifying a theme can encourage people to write good questions even if only one will be chosen
 
4:59 AM
Hmm now that you put it that way...
 
 
3 hours later…
7:49 AM
@MartinBüttner Found my bug. I mixed up i and j...
 
ah, the usual.... ^^
 
Oh, you're still awake?
 
all nighter and stuff...
 
Ah... well here's much-more-time-efficient-blur
I think it looks similar enough now (was struggling because I didn't realise I had to bump the radius waaaaay down)
 
that looks really amazing :)
any idea how to get rid of the very small cells?
 
7:55 AM
Rerun the first few steps of the algorithm, modified :P
If I do this right I think I get to be a bit less wasteful with how many cells are actually used
 
 
4 hours later…
11:45 AM
does zlib use a random algorithm? I'm getting different lengths when I compress the same thing
 
11:57 AM
how do you add latex to a question?
I think there was some way to render it as an image online and then insert it
but I can't remember how
 
I don't know about zlib specifically, but there are certainly cases in LZ compression where ties must be broken, and there are zip tools which run themselves lots of times picking random choices to try to find the best ones.
 
thanks
@PeterTaylor I put up my exact probability challenge
which I hope people find interesting
 
You seem to be obsessed with inner products of vectors over {0, 1}
 
@PeterTaylor :) You could say so
@PeterTaylor I think them because they are simple to state and the math appears beyond our understasnding
as far as I know, no one knows how to analyze these things exactly
although I do have another question in the sandbox currently
which has no inner products
also, I think it's an interesting programming challenge to compute probabilities exactly.. at least I hope people find it interesting
 
12:28 PM
what if I still do it via PnC and maths ?
Its not that hard.
 
12:49 PM
uhoh someone else submitted on regex comparison!
now i'm gonna have to compute optimal variable names for huffman coding!
maybe i should name them all a, aa, aaa, and aaaa
 
Ahaha :P I'm pretty amazed you guys even got so far with the question
 
dammit how to I put a \n without a \r
'bytearray' was suggested for binary data put it still puts one!
 
Hm what's wrong?
 
Just in case nobody's noticed yet: winterbash2014.stackexchange.com
 
@Optimizer What is PnC?
 
1:04 PM
Aww it's a countdown :(
 
@PeterTaylor Permutation and combination. a typical maths topic shortform :)
 
1:25 PM
I just tried to make a program that finds the shortest string literal to express a string in Python and now I'm totally confused
this would make a tough codegolf
there are 8 kinds of strings (single quote, double quote) X (single, triple delimiters) X (not raw, raw) plus the possibility of putting multiple string literals after one another
 
boo.py
 
2:09 PM
hi
 
Hi
 
hi
 
2:31 PM
I am impressed by the number of people who have already seen codegolf.stackexchange.com/questions/42299/…
do we know if these are different IPs ?
or could they all be the same person?
 
nah it was just me I clicked refresh 100 times
 
ah.. I see this room is still in crazy mode :)
 
Views are counted per unique IP, per 15 minutes, IIRC.
 
it really didn't use to be crazy
@Geobits thanks
 
So if I visit it once an hour, it counts each, but doesn't if I visit it twice in ten minutes.
 
2:33 PM
@Geobits got you
I am wondering what language the first answer will be in.. I feel I should open a book on it :)
 
you are worrying too much dude .
 
@Optimizer :)
 
and 113 is not really a huge number :)
 
that's all :)
managed to install fsharp for linux...
 
maths SE has everything super lower than us
users, views, etc.
(except for questions per day)
(which is okay as this is a different kind of SE)
 
2:38 PM
Probably something to do with the fact that wait an hour and your question's off the front page
 
true
 
makes sense
 
3:13 PM
HEYY
 
hi @TheDoctor
 
sup?
 
3:30 PM
0
A: Significant Whitespace

TheDoctorPython 2.7, 16+9=25 bytes Counter: len(x.split())-1 Splitter: x.split() Not sure if this is valid

 
... er... pretty sure that's not valid
On multiple counts
 
i have a lot of problems with reading too :(
 
Main one being they're not equivalent up to whitespace
 
oh well
 
The other being you're not reading from STDIN/function argument
 
3:40 PM
wait...
the top answer doesnt read from stdin either
 
Which one?
 
27
A: Significant Whitespace

Sp3000Python, 54 + 56 = 110 bytes Counter: m=lambda x:sum(y.isspace()for y in x) +1 0<9or x.split() Splitter: m=lambda x:sum(y.isspace()for y in x)+10<9or x.split() For the counter, we use the fact that Python is okay with having just an expression on a line. It's necessary to split up +1 and 0<...

 
That's function argument
 
ah
 
(You can run it with m("some string"))
Why the -1 btw?
Out of curiosity
I'm not sure how that counts whitespace
 
3:46 PM
is my answer valid now?
 
Well I don't think your counter actually counts whitespace for starters...
(also output is via returning from a function or STDOUT)
 
if you run it in the interpreter
 
Even if REPL was allowed (which I don't think it is, but you might want to ask Martin about that), I think your splitter would output the len and -1 wouldn't it?
But you still haven't answered my question about how the counter counts whitespace...
 
split() with no args splits on whitespace
 
It splits on runs of whitespace, so what happens if the input has more than one whitespace in a row?
e.g. input is 3 spaces
 
3:53 PM
>>> "hello world           a bunch of spaces        lalala".split()
['hello', 'world', 'a', 'bunch', 'of', 'spaces', 'lalala']
 
I'm not questioning the splitter, I'm questioning the counter...
 
the question was unclear...
therefore my downvote
 
I think the question was pretty clear. It says the first program needs to "Count the whitespace in a given string."
Wonder where @MartinBüttner is though...
 
but does that mean that 8 spaces is counted as one segment of whitespace? that is my understanding
 
8 spaces
is 8 spaces
 
3:58 PM
(If you want clarification from the OP you can leave a comment. But yes it would have been good if Martin's example had consecutive whitespace...)
 
4:43 PM
@Sp3000 I had to catch up on some sleep...
 
Good thinking :P I should really follow your example
 
@TheDoctor I thought the definition was pretty clear. "whitespace" refers to individual characters. so unless stated otherwise, 3 consecutive spaces should be counted as three. I can use the second example for both parts though.
I hope it's clearer now.
 
Hmm I'm a little conflicted about the modified k-means. It seems to do better in some cases, worse in others (still writing up my new answer)
 
examples?
 
The pro with the spheres is more sphere colours... but it loses the whole background
But the wave is much less blue
(this is for N=20, P=5)
 
4:51 PM
hm I see
how does it perform on the spheres for larger N/P?
 
Better - I think overall it does better on large N/P, but may be worse for smaller
It does lose a few details sometimes though, since they're small
e.g. centre of this nebula i.stack.imgur.com/RXY7X.png
It's kind of amusing seeing what changes, e.g. American Gothic losing the top window but the lady gets the red in her clothes :P
 
@Sp3000 I'd probably sacrifice the very small ones for better results on the larger ones then
 
:P yeah, well you'll see soon
Just the updated GIF to go
Hm... looks like I'll need to golf a bit of my Markdown
 
(the code could do with some tidying up too though, but I'm feeling lazy atm)
(so it's got copy-pasted code parts [/I'm the only one doing the maintenance so it's fine, right?])
 
5:03 PM
lol
 
You see, normally with PPCG questions you don't go "Hmmm I think I'll need this in three days. Better plan ahead and use a class."
 
what's with the recent problem of people not understanding the whitespace spec... the first 11 answers were all correct and none of the last four were.
 
HNQ? Maybe
 
they were all active PPCG users I think
yep
 
Time spent reading post = 60s * post upvotes ^ -1 ?
:P
 
5:15 PM
I think it's the (number of answers)^-2 factor
 
This warrants further study ;)
Anyone know where to find undergrad psych students?
 
lol, the favicon challenge is now my third challenge that gets to exactly 57 upvotes when it leaves HNQ
@FryAmTheEggman I'd try a university
 
maybe you could offer people 3 rep points to install tracking software
 
Sorry, motivated undergrad psych students
 
@feersum 3? upvote a question, downvote another? ^^
 
5:18 PM
sure that works
 
5:31 PM
Eh I give up, posted
 
haha, that allrgb image is fun :D ... have you tried with one of those awesome coral-like things from the winning post?
or some of the stuff from tweetable maths?
 
I was going to, but I probably won't put it in the post (it's pretty long)
Any you want me to try? (Specify N and P as well)
 
I'll have a look later, gotta head out now.
 
Ahaha k :P
 
You could still put them them in an external gallery, like I did for tweetable maths itself.
 
5:34 PM
does the challenge actually specify to make it look like paint by numbers or is it to look most like oroiginal?
 
@feersum it should approximate the original, but the voting criteria suggest to look for smoothness of cells
@Sp3000 what's next? outputting an actual PbN template? :P
 
That's one thing I could do
I'm actually trying to think of whether there's a good way of finding "thin" areas
Since that's about the only thing non-PbNlike left
 
TFS code reviews sure are clunky. Almost every time I have to get up and walk all the way to my team leaders desk just to discuss something.
 
finding them is easy, but not making them in the first place...
 
Well finding and fixing. Fixing's the hard part :P
(Do you make it wider or do you annihilate it?)
 
5:39 PM
well, I'm off... later, guys...
 
:) see you
 
6:32 PM
//9*
-623.
(sorry)
 
sorry the spec forbids comments
 
@MartinBüttner Heading off to sleep now, here's two pics (one of which is yours :P)
 
bye!
very nice pics!
 
 
2 hours later…
8:36 PM
"Say you're in an elevator. I could do two things to you and you wouldn't know the difference. I could pull the elevator up with a rope and you'd begin to feel heavy. Or, I could replace the planet beneath you with a bigger planet and you'd feel heavy. Now most likely I'll do the first one. But you can't tell the difference!" ~ Dr. Ramamurti Shankar
That one was listed as a joke. I don't think it is very funny, but I do think it is interesting.
Ugh. There is a guy at my workplace whom I can't seem to communicate with. If I think the job needs to be done in SQL, he wants to do it in C#. If I think the job is suited to C#, he wants to do it in SQL. I can spend all day explaining, and he just responds with a series of illegible grumbles.
Wow. SuperUser has ads on every question. Why are they special?
I wonder when it will be time to refresh our community promotion ads. It looks like they always do that in December.
I'm getting the silent treatment :(
 
8:55 PM
yes
 
Wait... have you been fixed @Optimizer?
 
yes
 
He has achieved golf nirvana: everything he does is as short as possible
 
no
 
@FryAmTheEggman Even sex?
 
8:58 PM
no
 
He obviously hasn't reached the pinnacle. Otherwise, he'd just be using 'y' and 'n' instead of 'yes' and 'no'.
 
yes
 
Better yet, cache answers until you have 8, then respond with a char representing them all.
 
 
2 hours later…
11:23 PM
0
A: Return the clock emoji closest to the current time when run

feersumC, 56 main(){while(time(0)%43200);printf("\xF0\x9F\x95\x9B");} Algorithm Wait for the clock to strike twelve. Print the UTF-8 bytes of CLOCK FACE TWELVE OCLOCK.

this answer works around the "print the current time" requirement by simply delaying until the current time is the only time it can print a correct answer for.
should this be disqualified somewhere as a [new?] standard loophole?
 
11:54 PM
@Sparr Do you not think it's already covered by this one?
 

« first day (1412 days earlier)      last day (3433 days later) »