« first day (1562 days earlier)      last day (3288 days later) » 
00:00 - 17:0017:00 - 00:00

12:00 AM
Exactly what?
I undeleted it and then spam-nuked it - that's what it's supposed to look like. (Check it again)
 
once its deleted, instead of being marked as spam. it doesnt really matter if I edit the content
 
36 secs ago, by Doorknob
Well, it doesn't really matter at this point.
 
@Optimizer implying you knew that was the only reason editing was fine :P
 
wut ?
 
Yes, it doesn't matter once it's deleted because it's hidden anyway. But please don't edit spam that's still undeleted.
 
12:01 AM
@Doorknob well, I didn't
 
Oh, you edited it after it was deleted?
 
-_________-
 
Sorry, I'm on mobile and trying to juggle several tabs.
Still, it should be spam flagged, and then left alone, not downvoted or delete voted.
2
 
@Sp3000 yes, exactly
 
Hey, today's my mod anniversary! (Just got the Constable badge.) :P
12
 
12:14 AM
@PhiNotPi I have some church music you might like.
 
what is it?
 
Mostly violins, organs, and pianos.
I think there are a few with flutes and trumpets.
 
Do you have a link to it?
 
12:33 AM
@Doorknob Congrats! :)
 
12:43 AM
@PhiNotPi Are you still downloading it?
 
When you said "some church music" I was thinking one or two pieces.
Not 341.
 
1:08 AM
@PhiNotPi Haha
That's one hell of a lot of church music.
 
I'm never going to listen to any of it anyways...
 
Haha
 
It's like:
You - "I found an interesting science news article the other day and thought you might like it."
Me - "Sure, what is it?"
You - <takes out encyclopedia>
 
Haha
 
1:28 AM
@PhiNotPi: Have you considered majoring or minoring in music if you choose to go to college?
 
not really, no
 
Oh yeah? I guess that's sort of surprising.
 
 
6 hours later…
7:33 AM
GL to CodeJam 1C'ers
 
It's starting at 4 AM for me. I'm not sure this is going to go very well :P
 
 
1 hour later…
8:50 AM
10 minutes to go... Anyone else competing?
 
Martin looks like he woke up early just for this, so I assume yes :P
 
Nope
Also, it's 10am
 
Wrong on two counts, nevermind :)
 
I'm starting to think I should have slept before the competition.
 
10 AM on a sunday is early
 
8:56 AM
And mother's day, no less. Actually, what parts of the world celebrate today?
 
9:10 AM
Me competing
11am here
 
9:28 AM
the first one seemed simple enough... but then, it will probably have some strange edge case and I will fail it miserably
 
Aw damn, if you're not competing you can't see the questions :(
 
Sp3000 thats wierd ... the idea u had in this challenge codegolf.stackexchange.com/questions/49876/… matches exactly mine :p (before i take alook at ur answer)
 
I'd be very surprised if you ended up with the exact same thing D:
 
not very , logical thinking leads to same conclusion
 
I give up on sub 33
 
9:38 AM
but the way u dealt with modular was for me a kind of deja-vu
 
The first code jam problem sounds fun.
 
@Optimizer "This converts 0,1,2 to 0 and 7,8,9 to 1." 7,8,9 to 2?
 
err
 
@ASCIIThenANSI "we don't use \n or special codes, but you do not necessarily have to ignore these if we enter them." ... "special codes" seems rather vague. it might be more straightforward just to say "The input will only consist of printable ASCII characters (code points 0x20 to 0x7E)."
 
Also there's no actual scoring criteria yet
 
9:53 AM
I was just gonna say
that's still missing, along with a nice set of test cases
and probably a time limit
 
should a shortest answer be the accepted one if it relies on a bugfix which happened after the question was asked?
 
hm, bugfixes are tricky
 
they are the fine lines I was talking about the other day
 
is there a language specification that says "this is how it should work"? (which is older than the challenge)
 
in Pyth many bugfixes are kind of like new features as I see
 
9:58 AM
I'd say if there's evidence that it was indeed unintentional behaviour, I'd accept it. otherwise I'd probably call it a new feature.
 
like you now how one thing should work with lists too, but not yet implemented
 
@MartinBüttner still, its a fine line. For example, in CJam : was not working with so many operaters, but can we go ahead and modify the answers now that it works ?
that was just a bugfix.
 
@Optimizer hm, true...
 
I can think of 10+ answers which had to come up with longer alternatives
 
I think, the simplest solution is just not to accept any changes.
 
9:59 AM
yeah
 
ok, I think I will go with any modification disqualifying the answer from being accapted
@MartinBüttner yep :) ^
 
10:26 AM
@Sp3000 @NinjaBearMonkey Retina update: intermediate stages are now silent by default (use : to get their output and ; to make the final stage silent, although I'm not sure what that's good for).
 
For debugging? I guess
 
or something...
now let's see if I can get those multi-stage loops
@Sp3000 do you have a better idea for syntax than introducing ( and ) options which start and end a loop?
I'll probably make leading ( and trailing ) optional as well.
 
<>? Not sure
 
I was gonna use <|> or [|] for when I allow specifying custom delimiters.
but my question more like if there's a better approach conceptually, not about the particular characters used :P
 
okay, I can't really do b, I'm really bad at probabilities
 
10:34 AM
@rcrmn the overlapping makes it really tough
 
it makes it impossibly tough to do it in the naive way, at least
 
@MartinBüttner Well are you making the language usable or for golfing? :P
 
there is probably some way to simplify it though
 
@Sp3000 for golfing, obviously :P
 
duh
 
10:38 AM
Then I'm not sure what you can do other than 1 char delimiters :/
 
hmmm... () in a single configuration string makes sense (would probably be the same as +)... but what about )(? That doesn't really work, does it?
 
I think I give up
I did A, and I think I got it right, but B and C are too hard for me to do in so little time
 
what was C again?
ah, the coins
 
yes, the coins...
 
weirdly, according to the point distribution, they thought all problems were equally difficult this time.
 
10:49 AM
@Sp3000 that bounty . -__-
 
@Optimizer what's wrong with it?
 
just that python only
like python answers do not get enough upvotes
 
the jealousy...
that challenge could use a leaderboard snippet
@Calvin'sHobbies ^ (ascii sun)
@Sp3000 yay for generics: var StageTree = new Stack<List<Stage>>();
I think I'll disregard the order of ( and ) for now, and always treat them as if all ( are before all ).
 
I think I'm out too. Doing B and C in 30 minutes seems really unlikely.
 
11:06 AM
@Sp3000 I think I'm almost done with the multi-stage loops... would you mind trying to come up with some test cases and help me bullet-proof it?
(loops can be nested)
 
Well, at least we tried
 
@Sp3000 okay, I've pushed the changes. printing is still awkward if the last stage is part of the loop. the best thing to do currently is to add an empty stage at the end (which sort of defeats the purpose of having ) at the end implicit, but I haven't sorted out nice semantics for printing within these loops yet)
 
You can encode Collatz in unary :P
 
do I need nested loops for that?
 
That was just for multi-stage
 
11:13 AM
ah right, yeah I've already tested simple multi-stage.
The nesting is what's tricky ;)
 
Hmm when would you need a nested loop... hm...
 
when do you need nested loops in normal code? :P
I guess you could try moving a marker through an input grid, one character at a time.
 
Well it doesn't quite translate like that :P
My first thought was bubble sort, but you can do that with one loop and global
 
transposing a grid?
 
this is not image processing at all, right ?
 
11:21 AM
depends on how you define "image"
 
neither ascii-art
@MartinBüttner there is no image here. input is text, output is truthy/falsy
 
why is it not ascii art? o.O
@Optimizer "depends on how you define image"...
 
because there is no art
 
ascii-art on questions which have only ascii input and no ascii output is already a fine line
 
11:23 AM
we should nuke the ascii art tag... it's obviously lacking an objective application criterion.
 
@MartinBüttner no, we should clear its meaning and limit it to output
 
@Optimizer I'd say we should clarify its meaning and explicitly say it also applies to input :P
 
:/
being a mod, you should be more serious
 
at least more than me :P
 
11:24 AM
I don't see why we wouldn't want the tag for input
do you have a better suggestion for a name for an ascii-art input tag?
 
no, but relate it with other tags
 
code-golf, for ex. does not mean that the input has to be the least bytes
math/string/array-manipulation
 
that comparison doesn't make any sense
 
etc are all related to the answer and output
 
11:25 AM
are they?
I don't think so
 
only a few are there who relate to just input. like image-processing
but in any case, that particular question is not IP
 
I don't see why... if the input was an actual image file with white, grey and black pixels, that would only change how I read the input, but not how I process it.
ascii art is just another (low-res) representation of images.
 
then that holds to all possible questions who take input
 
ehm...
 
2 numbers as input, sure , they can be fed as an image
with black and white pixels
a string as an input, why not ?
 
11:28 AM
yeah, but you're not actually doing something with it that is like finding structures/patterns in an image.
 
you are finding patterns of alphabets
to convert it to string/number
can you give examples of any 3 existing questions which do not have image as an input and have image-processing tag ?
 
no, but that doesn't seem like a relevant argument
 
and again, forget what counts as an image, forget that input could have been an image. What matters is that its not
 
just because we had no question use that tag like this yet doesn't mean it's the wrong tag
 
the question, as of now, does not involve any image processing . no answer will do any image processing, neither the input will ever have an image as an input
 
11:31 AM
@Optimizer if I search long enough, I'm sure I'll find an image format that is just a text file whose byte values are the intensities of the pixels
ugh, this isn't going anywhere. take it to meta if it actually bothers you. I don't think I can help you here.
 
@MartinBüttner and ?
what you are talking about is the opposite case.
@MartinBüttner since the input here is text, your weird image format won't matter because in this question, the input format is clearly defined.
 
@Optimizer no. if that exists, the input to that challenge is actually an image file.
 
Having said that I still think tags need reworking though, but anyway...
 
@Sp3000 yeah, that in general.
 
11:35 AM
Well, I got A-large wrong, so 11 points overall... so sad!
 
@MartinBüttner but the OP clearly says that its not
@Sp3000 then these need to be worked too. image generation is different from image processing
 
@Optimizer "In modern sciences and technologies, images also gain much broader scopes due to the ever growing importance of scientific visualization (of often large-scale complex scientific/experimental data). Examples include microarray data in genetic research, or real-time multi-asset portfolio trading in finance."
 
sure, then I am going to add image processing to every challenge that takes any for of ascii art as input. I think that is alright , okay ?
 
@Optimizer sounds reasonable to me
@Sp3000 I think the loops are actually working :)
 
@MartinBüttner wow
 
11:43 AM
Huzzah
 
I need a version number
0.5.0? (CJam style)
okay, done
anyone wanna do the Pauli matrices challenge today?
 
only if you add image-processing tag to it
 
so bitter
 
I am talking your language only now
 
you know, we could just agree to disagree and leave it at that.
 
11:56 AM
@Optimizer Don't worry, at least image-processing isn't the worst tag
 
@MartinBüttner I agree
 
@Sp3000 Are you up for Pauli matrices? (even without an image processing tag? ;))
 
how tag has 16 questions?
 
Pauli sounds good but I can't guarantee a submission from me
If you do get one though it means that, once again, I'm not doing what I'm supposed to
 
awww... I was hoping someone would take the new retina loops for a walk ;)
anyway, posted it:
0
Q: Multiply Pauli Matrices

Martin BüttnerThe Pauli matrices are a set of 2x2 matrices which appear very commonly in quantum physics (no, you don't need to know any quantum physics for this challenge). If we include the identity in the set, the four matrices are: σ0 = σ1 = σ2 = σ3 = [1 0] [0 1] [0 -i] [1 0] ...

 
12:04 PM
can you put in all possible multiplication of 2 pauli matrices in there ?
like 13 gives -i2
 
@Optimizer that sounds boring
 
@MartinBüttner but helping
 
@MartinBüttner I think it's obvious but maybe not that you have to multiply matrices from left to right
 
@randomra matrix multiplication is associative
 
my bad, that's embarrassing
 
12:05 PM
@Optimizer it was meant to be part of the challenge (although you can just look it up on the linked wikipedia article)
 
Woo Klein-4
Oh wait
 
@MartinBüttner bringing in identity matrix is kind of useless
 
@Optimizer you need it for the output, so I thought it wouldn't hurt in the input.
unninja'd o.O
 
^^ Word of god mod
 
sp3000 was first though
 
12:26 PM
fyi, I got 77 in retina (I don't intend to post it though)
 
232 is same as 223 is same as 03 is same as 3, right ? Am I doing something wrong ?
 
matrix multiplication is not commutative
232 ≠ 223
 
ok
 
(in the case of the Pauli matrices, 232 = -223 though)
 
but 203 is still 23
 
12:28 PM
yes
 
I think the hardest part of this challenge... will be the output formatting :/
 
i know someone would beat my stumpy dumpy C code
 
@Sp3000 depends on how you solve it I guess
 
The actual calculation's not too bad... in Python at least
 
aww , im not list-tailing anyway :D thank u MickyT
 
12:36 PM
oh, had a bug... 86 in Retina now.
 
the states of a complete state-machine are 32 (or 16) bytes. it makes me want to skip the whole thinking/trickery part :)
 
every byte is an agony :D
 
Hmm I got everything right except 1320130100032. Uh oh
 
hm, let me double check that
 
grc
@MartinBüttner can we use numpy?
 
12:48 PM
wow this is long
 
@grc sure
@Sp3000 hm no, I think i2 is correct
 
>>> 0-1j
-1j
>>> -1j
(-0-1j)
>>> -0
0

that (-0-1j) happens because complex numbers are always float internally right? (Python3)
 
@MartinBüttner I screwed up on output formatting :/
 
grc
@randomra yes
 
56 :(
 
12:55 PM
@Optimizer hey, at least you're not beaten by Retina :P
 
like retina can beat anyone (Except once)
 
I've won challenges with regex before retina was a thing (so retina would have won there as well)
 
but didnt
:P
 
I'm not gonna start searching for all the challenges where Retina won, but it's definitely more than 1
 
mind that whether the identity matrix is multiplied by itself or any other one it doesnt change anything
i mean σ0
 
1:10 PM
no, the identity matrix doesn't change anything
 
and thats an interestin clue
 
These conditionals are turning out crazier than integer grades :(
 
ok 50
worst part is calculating the prefix
as cjam doesnt have complex
 
all the complication resides in third complex matrix
 
1:31 PM
@MartinBüttner Wow, this is actually really nice. One part of it anyway.
 
23 = -32 ?
 
@Optimizer yes
 
@Optimizer in base Sqrt[-305/2], yes.
 
@Sp3000 looking forward to it :)
 
but not for 03 :/
 
1:38 PM
nope
@PhiNotPi in base -1 actually
 
ughh, 10 bytes for just that
60 total
 
@Optimizer was I missing a test case?
 
no
 
@PhiNotPi unless it's a -3 digit instead of -(32), then it's base -1/5.
 
Woo I'm within 2xCJam, goal looking good
 
1:45 PM
@Sp3000 so many ternary conditions that there is almost nothing cjam specific
 
@MartinBüttner I totally messed up the calculation. :/
base -1 is right
 
so is base 0 :)
 
depending on your definition of 0^0
 
^
0^0 = 1
so no
 
Only in computing!
 
1:55 PM
I'd rather have a discontinuity in 0^x than in x^0 ;)
 
@Sp3000 how much is urs ?
 
also, 0^0 = 1 is the right limit for x^x.
 
112, of which 30-something is output
 
^ python?
 
Yeah
 
1:57 PM
did you find something mathy ?
 
For one of the parts, yes
The other part's too long for my liking atm though :/
 
optimiser , u bear ur name well , dude
 
Optimizer*
 
is it american english ?
or just a proper noun
 
both :)
@Sp3000 you are at borderline now :P
 
2:04 PM
tch
 
+1 , despite im not habitual with that kind of talisman-coding
 
@MartinBüttner 000 => 0 might be a useful testcase if you are planning on removing all 0's in preprocessing
 
hm, true
that destroys my nice Fibonacci sequence though :D ... but I guess I can live with that.
 
replace 123 with 000
 
nah, I like that one
 
2:10 PM
make that 01023 :P
waits for J or Pyth answer
is ti just me or now the cjam links do not work almost 95% of the time in Firefox ?
(earlier they used to work 50+% )
 
Okay, I give up. This is hard :/
 
2:28 PM
@Sp3000 what on earth :D
 
:D
 
@Sp3000 hah, I was actually able to save two bytes on my retina solution by using nested loops :D
 
XD nice
 
got 80 now
hm, found an equivalent solution that doesn't nest () loops, but uses + inside an () loop.
 
headdesk boy I feel like an idiot - why am I using complex again?
 
grc
2:44 PM
@Sp3000 haha yeah I was about to suggest this:
x=y=0
for m in map(int,raw_input()):x+=(m-y)%3*(m*y>0)*3/2;y^=m
print'-i'[x%4<2:1+x%2]+`y`
 
Ahaha nice :P
Yeah I'm not sure why it took me 2 hours to realise that
 
grc
how do you format code in chat again?
 
Indent 4 spaces (or hit the button that pops up)
 
@grc put it in its own message. hit Ctrl+K
as there's no markdown in multiline messages, you can mix code with normal text
 
grc
print('ah thanks')
 
2:46 PM
@Sp3000 it's nice that you could reuse the character selection trick for -i :)
 
:P thanks for that
 
maybe I should add that to the Python tips
 
:P
Must... golf... down... x+= ...
 
@Sp3000 your algorithm is around 45 bytes in cjam
 
Sounds about right :P
 
2:56 PM
:D
but i am messing up some operator precedence ..
 
@Sp3000 done... could probably use some polishing from someone who actually uses Python :P
 
can you properly add brackets in this : x+=(m-y)%3*(m*y>0)*3/2 ?
 
Looks good to me
%*/ all have the same precedence
 
the same thing in cjam is not working.
so my understanding is
 
So it's x+=((((m-y)%3)*(m*y>0))*3)/2
Oh right, Java
-5 3% gives -2 instead of 1 as it would in Python
 
3:01 PM
s/Python/a good language/
 
grc
x+=m*y and(m-y)%3*3/2
^ does that work @Sp3000
 
-5 3% gives 1 in python ?
 
as it should
 
@grc I think it does, thanks :P Trying out a few other ideas over here though, will see how this goes :)
 
@Optimizer oh sick, I just found a way to get "proper" modulo in 2 bytes in CJam
just use ,= instead of %
to the tips!
 
3:04 PM
That's pretty nice :P
 
(I guess the naive way is :X%X+X%)
 
3:17 PM
This has to be the closest I've ever been to beating CJam in Python :D
 
@Sp3000 wow that new indexing is clever
 
I could almost switch x+= to x-= but then the index becomes ~-x%4 :/
 
and lik iv said before , some one already did it
 
oh nice, I'm halfway to my first Steward badge
 
0
Q: Should [tips] also be tagged [code-golf]?

Martin BüttnerWe've got 93 tips questions. 61 of those also have the code-golf tag. I suspect that actually all 93 of them (or at least 90) are actually tips for code golfing. So this seems like fairly inconsistent tagging practice, and I can see why there's confusion about it. It would be nice to decide how t...

 
3:30 PM
and more C-jumble
i dont think C-jam needs any more golfing thu
 
The new challenge looks familiar.. goes dupe hunting
 
there are like 10 incarnations of it, and they are always marginally different and incredibly hard to search for
 
it explains why r arithmetic-puzzles done so fast sometimes
 
Okay I give up :/
 
4:26 PM
Okay, Jakube's got the same matrix method :P
 
down to 78 in Retina
two bytes at a time...
 
Too bad Retina doesn't have XOR :P
 
I love how Pyth answers have no shame at all in terms of copying algorithms, just to be the smallest :D
and then they claim that its better than CJam :D
@Sp3000 pretty sure if you convert, it would be smaller.
 
It will be, but I don't feel like posting :/
It's not really that different either
 
its not like you are copying or anything
 
4:42 PM
@Optimizer this :/
 
@MartinBüttner I know right. Even the author promotes it :/
 
tbh I wouldn't be surprised if Jakube got the xor thing on their own, because they update the coefficient differently (and a bit longer than mine I think)
 
its not that straight forward too though
 
btw in CJam it'd be shorter to do the -1^ that Jakube's doing, just a heads up
Unfortunately it was too long for Python though :/
 
I do not intend to use this algorithm..
 
4:47 PM
ah, crap, I forgot to remove a byte from my Mathematica answer earlier, now I'm late in the tie :D
 
lol
see !
 
:P just saying, in the interest of CJam-research's sake
@MartinBüttner "tie" - giiiiive me a sec :D
 
... :( it's the same as OP's - except they're using 4 space indents and str instead of backticks
 
perl has 50 byte solution
 
4:52 PM
yep
 
wut : "--+" [1 2 3].+ed
ahhh
 
00:00 - 17:0017:00 - 00:00

« first day (1562 days earlier)      last day (3288 days later) »