« first day (1541 days earlier)      last day (3296 days later) » 

12:37 AM
hi
 
12:52 AM
o/
 
\o/
 
~o~
 
This place has been pretty vacant recently.
 
i'm here! :D
 
1:11 AM
@PhiNotPi it's called "the weekend"
 
This is my weekend too.
And I'm on here.
I'm thinking of posting my explanation formatter AS-IS.
I know there are different formats that people like explanations to be in, but I'm having trouble with coming up with a formatting description that actually captures them.
I could add a 2nd formatting string, so the 1st shows where to start and the 2nd shows where to stop.
 
1:35 AM
@sirpercival About your meta-posting problem - It turns out I was viewing your main profile (StackOverflow). You might be able to find a challenge such as that has very few or no answers, as people may upvote your question. is another one (this is designed to specifically win by most upvoted answer.) Just find one without any answers (or answers you could make a good improvement on) and you should be able to get enough rep.
 
Yeah, just find some super-generic code-golf or popcon contest somewhere, post some lame answer, and I'll give you an upvote.
^ @sirpercival
Never mind, that particular question was marked as "protected"
 
 
1 hour later…
2:46 AM
That antiquine question's just as impossible to verify as this :/
Considering the new restrictions
 
3:37 AM
Would this be a duplicate?
Input: integer K, sorted integer list L
Output: L with at most K (integer) insertions in a way that the maximal forward difference is minimal
 
At most K? When would it be sub-optimal to just use K?
 
just might be convenient to only use what you need
and answers would have to e.g. pad with the last element and that seems unnecessary hassle
 
given a number N, is it possible to produce a sequence from 1 to N in a random order in O(1) space (for each step)?
I'm looking for something that can take its current position and know its next position given a max N
knowing that the path taken will hit every node once before hitting one twice
 
3:53 AM
@NathanMerrill you could use primitive roots; they're not exactly random but can appear to be
 
@aditsu that would require that N is prime, no?
 
not necessarily
 
say my number is 14. If I start with an even number, every additional number will be even
and vice versa
 
but some numbers may not have primitive roots; in that case, you can use a larger number and skip the values that are bigger than N
 
I actually found a stack overflow question on it
31
Q: Create Random Number Sequence with No Repeats

UnknownDuplicate: Unique random numbers in O(1)? I want an pseudo random number generator that can generate numbers with no repeats in a random order. For example: random(10) might return 5, 9, 1, 4, 2, 8, 3, 7, 6, 10 Is there a better way to do it other than making the range of numbers and sh...

 
4:20 AM
for the primitive roots method, it makes sense to use prime numbers indeed
otherwise you will have gaps
 
 
2 hours later…
6:01 AM
@BrainSteel Sorry, looks like I misunderstood the question. I've deleted my comment.
 
@Sp3000 No problem at all! It does seem the description is rather vague, though. Either his definition of produce or true antiquine should probably change... Technically, my code does output all valid 392-byte C programs. On the other hand, restricting it such that every 392-byte subset of the output was a valid C program would be extremely difficult, if not downright impossible.
 
I got pretty confused with the whole "substring" thing as well, e.g. if abc and def are valid 3-char programs, then given the output abcdef you've also printed bcd and cde?
 
It looks like that is the case. It would be very, very difficult to find a language in which such a thing is feasible, right? The "modified antiquine" is doable, though...
 
this antiquine is easiest in Pyth
 
If I picked a random 20-byte set of ASCII characters, what are the chances that it is valid Pyth?
 
6:08 AM
Not very
 
but the highest!
 
Especially considering that the definition allows for 0-31
 
why would you go to 20 bytes though ? You can do that task in 4 or 5 bytes, where the chances are higher
 
"The program must be at least 20 characters long."
 
oh
 
6:10 AM
You'd be better off picking a language where there are no syntax errors, tbh
 
whitespace?
 
The first one that comes to mind is BF... And that seems preposterously difficult.
 
That would depend whether you count unmatched brackets as invalid BF
 
Solid point. Hadn't thought about that.
 
@BrainSteel Left a comment on the question summarising our problem. Would you like to take a look?
Just to make sure I didn't misunderstand anything else :P
 
 
1 hour later…
7:36 AM
@NathanMerrill Formally, no, because you have O(lg N) output per step.
 
hi all
 
8:19 AM
Waddup?
 
 
3 hours later…
11:12 AM
still thinking about the assembler
 
11:34 AM
Hah, awesome!
f x = x
x = 2*6+3
y = x*2
main x = f y
-> http://i.imgur.com/NAXAVlt.png
 
you know, reflect.parse can do that for you .
 
Shhh.
I'm learning here
And isn't parse, for like, parsing?
I mean, I have now written a parser, semantic analyzer and a bare-bones compiler.
 
 
1 hour later…
12:56 PM
1
A: Sandbox for Proposed Challenges

ScimonsterOperator, i can't remember my phone number... code-challenge quine "Hello? Operator? How do i get my phone to tell me its own number?" The challenge is to write a quine -- from your cellphone. Not a smartphone with a virtual keyboard, mind you. One of the old phones, where to get a 'c' you have...

I while ago i posted a challenge to the sandbox that was very similar to an existing challenge, but with a different scoring method. @MartinBüttner left a comment saying that while the scoring is cool, the challenge itself is too similar (essentially identical). I changed the challenge, but honestly prefer my original idea. I'm posting this now to get some other people's opinions, whether it actually is too similar.
Here's a link to the original revision: meta.codegolf.stackexchange.com/revisions/4695/1
 
whee in answered my first question! lol. pretty boring answer...
*i answered
 
(you can edit messages in chat for a short while after posting them, mouse over the left part, and a button for a drop-down should appear)
 
@VisualMelon You can also just hit the up arrow.
 
indeed, that does work
 
oh hey that's cool
i'm used to IRC where once it's out there, it's out there
 
1:09 PM
you can even remove evidence of your own stupidity ;)
 
@PhiNotPi i answered a popularity contest: codegolf.stackexchange.com/a/49014/31863
 
1:27 PM
@sirpercival Have an upvote. Now you can use the sandbox. (I would also like to inform you that your answer probably isn't a valid solution.)
 
@Sp3000 I fell asleep, haha. You haven't misunderstood me, but it's possible we've both misunderstood something else :P
 
I sure hope we haven't or I'd feel bad for putting on hold
 
1:44 PM
@randomra I'm not aware of a challenge like that. Sounds quite nice :)
@Scimonster So is the challenge just "write a quine, but score it in keypresses instead of bytes"?
@Scimonster Also, I didn't see your comment on the challenge, because I only get notified when you use @MartinBüttner in your comment (unless it's on my own post). (There's tab auto-completion for that, so no need to write the umlaut yourself.)
 
@Sieg I believe it can generate AST from JS and vice versa too
 
@Optimizer I have a parser.
 
@Sieg I didn't say that you haven't
 
@PhiNotPi haha you're totally right. whoops... anyway, thanks!
 
2:06 PM
timer <- ticker 0
main <- timer*2
->
var timer=(ticker)(0);
var main=FRP$.lift([timer],function(timer){return timer*2;};

mut
timer <- ticker 0
main = timer*2
->
var timer=(ticker)(0);
var main=timer.get()*2;
Awesome.
Apparently the compiler missed a brace.
 
@MartinBüttner No, i want to know about the original challenge -- the calculate keypresses one - scored in keypresses.
@MartinBüttner I thought you also get pinged if it's my post and you're the only other commenter.
 
@Scimonster Well, I don't think my opinion has changed. If the scoring is the only novelty of the challenge, it's more of a gimmick than anything else.
@Scimonster Yeah, that seems to work sometimes, but not always, I don't know...
@PhiNotPi When are you planning to post your explanation formatting challenge?
 
I can do it right now, I guess.
 
no rush, I was just wondering
 
@PeterTaylor I don't think so. Each step I output a single number, which is O(1)
 
2:21 PM
@PhiNotPi If you want to simplify it slightly, change the order of upper and lower case prioritiees.
@NathanMerrill His point is that printing the number N takes log N digits. Hence "formally".
 
@MartinBüttner That makes sense, ASCII order and all.
 
3:01 PM
I'm about to post Formatting
Any title ideas?
"Code Explanation Formatter"
"Create a Code Explanation Formatter"
etc.
 
"Do create a code explanation formatter"
Please ? Dare you to ?
 
3:19 PM
Can I haz code explanation formater?
 
Bonus points for solving in LOLCODE.
 
score/10 if written in an esolang
 
^ minimum score wins!
 
In fact, the only acceptable languages are those without Wikipedia articles.
 
@PhiNotPi I feel like that could lead to some serious vandalism on Wikipedia :D
3
 
3:28 PM
git checkout wikipedia
rm -rf wikipedia/*
language neutrality!
 
"Only languages created after this challenge may be used."
 
"Only languages created after this challenge was posted to sandbox and before to main site may be used"
 
That would be the cue to create INTERCAL--
 
Now where is aditsu when we have a nice challenge that could have used that variable I requested!
 
What's the line between feature requests and harassment?
 
3:38 PM
A ray..
of hope ?
 
@PhiNotPi Line(x=self.amiability, y=self.persistence)
 
you cannot make a line using a single point
Whats your other point ?
 
@Optimizer those are arrays
in the language of my brain, which is a cross between python and IDL apparently
i'mma have to start golfing IDL, since i don't recall ever seeing it on here... that'll be "fun"
 
@MartinBüttner how long is yours ? (solution to code formatter)
 
4:14 PM
does anyone else use IDL?
 
how does it give me a notification even though the comment did not contain my nick
 
magic
 
fail
 
eic fial
 
9 kid
 
4:19 PM
Did I answer your question with the formatting?
 
yes
 
It's kind of a weird way to do it.
 
no one would do it, but a spec is a spec is a spec
 
4:32 PM
I wrote a brute-forcer for the cube-numbering problem. It's been running for a few days now. No results yet. :/
I honestly don't expect it to find anything.
 
@PhiNotPi What is it looking for?
 
A shorter way to write a portion of my Element solution.
 
Lots of choices there to bruteforce.
 
I actually pretty certain that I've hit "maximum golf" with my Element solution.
Three characters must be along the lines of _4: to take input and make copies of it. The top-right corner must be the face number (1 character), otherwise I must append a newline to the calculated number (+ 3 chars). The top-left number must then be calculated independently of the bottom row (or stack manipulation costs characters).
The shortest possible way to derive 1 of the other 3 corners is -5+` (4 chars) to subtract it from five, because simple adding 2+` will not work (I tried). Then, that leaves 10 characters for the second line in order to get it to be equal to my current 18 char solution.
Of those 10 chars, the last one must be ` (1 char to print), so 9 chars are left.
 
4:50 PM
trrr... trrr...
 
Since I'm only interested in solutions that are actually shorter than my current version, I'm looking for 8 characters to fill in the blank.
@Optimizer ?
It just gave a result!
Now to verify, by hand, that the result makes physical sense. I only programmed checks to make sure that obviously incorrect things are excluded, like solutions with two corners being the same number and such.
 
Couple of days ago I asked your favorite programming language. Now, I ask what language has struck you as the most interesting or innovative.
 
Did someone reply back then ?
 
Some did.
 
@PhiNotPi so... why is there no aa pair in the last line ?
since they are there in the second line, it appears that the pair will show up if any thing inside them is still showing up
so bb pair show up because ee exists, thus, aa pair should also show up
 
5:04 PM
The brackets continue to show up until all characters within it have appeared. On the last line, all of bbccdd have already appeared on previous lines, so the aa pair is not printed.
 
I see. thats a rather difficult thing to figure out.
 
Otherwise, a pair of interweaving brackets abab is not resolvable : a appears if b appears, b appear if a appears.
 
(as per my algorithm)
 
@Sieg Most innovative? Lambda calculus. I think that speaks for itself.
 
5:20 PM
I calculate yo lambdas.
 
6:10 PM
UGH my controller is still not working correctly!!
also it's slow... i should probably cythonize it but i suck at cython
 
@Optimizer why do you think I have one?
 
you don't have one ?
 
I dunno if this is a really great fit — it's definitely a puzzle to me, but isn't something that started as wanting to construct a puzzle
 
6:26 PM
@Optimizer It also doesn't sound like a very enjoyable task in CJam.
 
@MartinBüttner its challenging, thus, enoyable
 
@Optimizer Are you a fax machine?
 
@AlexA. what year are you living in?
 
@Optimizer I don't think there is necessarily a correlation between the two (either way)
 
there is for me
one way.
 
6:27 PM
@Optimizer Not sure. Let me check my sun dial and star chart and get back to you.
 
they can't tell the year
 
Hmm.
Well... then let me consult my inscribed Mayan calendar.
 
oh, if mayan can tell, then you definitely are like 3 years behind :P
 
Seriously. The Mayans even had pagers.
They all went off at the end of the world in 2012.
"Excuse me, I'm getting a page." Boom. Mass extinction.
@PhiNotPi I believe you mean "I can haz." 1.bp.blogspot.com/-CzqzzBV2tMk/TxBM3ar18MI/AAAAAAAAPm0/…
 
 
1 hour later…
7:41 PM
@Optimizer I'm getting an error from your CJam program: "Syntax error: java.lang.RuntimeException: ee not handled"
 
ctrl refresh
 
Are you going to use your new code explanation formatting tool to format the explanation of your code?
 
okay, it works
 
@BrainSteel maybe ;)
 
You can show output and explain things in one step!
 
7:44 PM
@Optimizer get rekt
 
58 byte
 
And here I am, halfway through my ~200 byte C solution...
 
in pyth ?
 
I'm writing explanation at the moment
Yes
 
7:45 PM
similar algorithm ?
 
I don't think so.
 
well, given the amount of loops in the code, Pyth out to win it anyways
 
@MartinBüttner So if i also include keypresses for stuff outside of the normal range, would that be different enough? I mention that scoring at the end of the current version. That wasn't in the original.
 
s/out/ought
 
So, posted.
Even with a nice example.
 
7:58 PM
So, this particular explanation format seems to look a lot better with CJam than Pyth.
Due to its stack-based nature, I guess.
 
take that Pyth!
 
^^
Also it may be my personal style of doing an explanation.
 
8:14 PM
Some day, I'll write a code golf solution in Pyth just to mess with everybody.
4
 
8:51 PM
I ran Code Formatter on the QWERTY keyboard:
          a
                       b
                     c
            d
  e
             f
              g
               h
       i
                j
                 k
                  l
                         m
                        n
        o
         p
q
   r
           s
    t
      u
                      v
 w
                    x
     y
                   z
3
 
@PhiNotPi: Oh, okay. Now I understand the keyboard.
 
I think I understand the keyboard less now...
 
9:45 PM
@Sieg Not technically a language, but I love Minsky register machines as a model of computation.
 
how do you measure how many bytes your code is?
 
If your code is just made up of "regular" letters/symbols, then byte count = character count.
 
@sirpercival This is a good resource!
 
@PhiNotPi, @BrainSteel thanks!
 
There's also this byte counter.
 
9:49 PM
@PhiNotPi Boooo, off-site resources :P
 
We are both off-site resources.
 
At least Meta is still on SE!
 
At least the off-site one works in my browser :P
 
@Sp3000: Are you on Safari? Almost none of the stack snippets work for me in Safari.
 
@PhiNotPi i'm working on a formatter
it's not going to be anywhere near short enough to win
but hey
 
9:55 PM
Chrome actually. For some reason snippets work when they feel like working (which, at least, is more often than not)
 
@Sp3000: Unrelated: Did I ever tell you that I looked at that gallery you sent a while back when I asked about your avatar? It's freaking incredible. So cool.
 
:P one day I'll update the thing so it doesn't take 10 minutes to run. One day...
 
sandbox autopost is lazy:
0
A: Sandbox for Proposed Challenges

randomraSmoothing out a list code-golf number sequence You should write a program or function which takes a non-negative integer k and a sorted integer list Las input and outputs or returns a smoothed list M. M is created from the ascending list L by inserting at most k integer elements while keeping ...

 
Looks familiar, I wonder where I've seen that before... (but for the one I remember I think k was the max diff)
Oh I remember where. Here it is.
 
10:15 PM
Just FYI, the off-topic reason for general programming questions has been updated.
6
@orlp you might want to accept that answer ^
@Sp3000 Isn't it essentially House of Pancakes?
 
Is it?
 
Well sort of. I think the approach is the same
You take the initial gaps, and then you keep adding numbers (equidistantly) into the gap that currently still produces the largest differences.
 
(btw I think ((n=#)+#^2+(f=#&@@#@Range@Max[1,2n]~Nearest~n&)@Prime+f@Fibonacci)/4& is 69)
 
hah, crazy, thanks
 
Oh, yeah if you put it that way it does feel the same :P
 
10:29 PM
@randomra doesn't the algorithm I just described also solve the bonus?
 
@MartinBüttner it seems to be, I'm not sure yet it's polynomial in output size (obv. can't be in input)
if you can compare gap A with k inserts and gap B with j inserts, it's probably good
 
I'm pretty sure it's something like O(N^2 log N)
N^3 log N at worst
 
which is probably just A*j>B*k
 
I think if you're doing it like pancakes it can be better than O(N^2 log N), although I think the complexity should have k in there somewhere
 
you might want to require polynomial time or something. golfing-wise it might be shortest to just generate all n-choose-k distributions (where n is the number of available slots) and then check which has the smallest forward difference.
@Sp3000 N was meant to be |M| (i.e. k + |L|)
 
10:41 PM
Oh, right
 
also, I think this might be a bit trickier, because the division isn't as simple as with pancakes, and if you want lexicographic order you might also have to compare lists instead of just numbers.
hence the conservative guess of N^3 log N
 
@MartinBüttner Right, I will add the "Your solution has to solve any example test case under a minute on my computer (I will only test close cases. I have a below-average PC.)." part.
 
yeah, that's good
 
@randomra Can you also put the maximum forward distances in the test cases? Since that's what people would need to check
 
@MartinBüttner In my idea you need a sorted modifiable structure to store the (gap,insertion number) pairs and always take the topmost and increase insertion number by one.
 
10:49 PM
@MartinBüttner I'm unreasonably excited for your telescopic parentheses challenge. I don't really know why.
 
hm, right you don't need to resort each time
you only need to push down the top thing to the correct position.
@BrainSteel I'm glad to hear that. :D I'll post it in about 12-16 hours, I think.
 
I may have gotten a bit of a start on it :3
 
I don't mind that :P
 
In before everyone uses zip :P
 
11:12 PM
@Sp3000 added max difs
 
11:50 PM
@ChrisJester-Young Poor Rainbolt will wonder what pinged him there ;)
 
@MartinBüttner You'd think the pings would disappear too. ;-)
 
You'd think...
 
(In practice, they don't. The count stays as before, and when you click them, they do highlight the deleted messages. So you're right. :-P)
 
@Chris I think it works for comments though.
 
@MartinBüttner Yep. (I think.)
 

« first day (1541 days earlier)      last day (3296 days later) »