« first day (1323 days earlier)      last day (2126 days later) » 

9:59 AM
@EricTressler yes, that sounds reasonable
10:12 AM
Hello, everyone!
hi there
(you won't find "everyone" in here on a weekend)
@MartinBüttner I'm working on a proof-of-concept cheat bot that recognises whether the opponent is your RandomBot or the Center bot and (basically) has fun. ;-)
^ Unsportsmanlike indeed.
Well, no, if you write a bot that actually tries to perform well by fingerprinting the opponent and using a dedicated strategy against each of them, I think that's fine.
What I was referring to is a bot whose only purpose is to defeat one other bot. We had that recently, where a user had several submissions which were all beaten by some other guy. So he added another bot whose only purpose was to incapacitate the winning bot.
Have you had a look at the arbiter?
The code, I mean.
No I haven't followed that KotH at all.
10:19 AM
Q: Should we split per-language tips for different versions?

Martin BüttnerThe recently added list for golfing tips in ECMAScript 6 sparked a bit of a discussion in the comments (and the close/reopen votes) whether tips for different versions of one language should go in the same list or whether we should create a new list. There is a precedent for creating separate qu...

@Rainbolt feel free to add a third answer along the lines of "We wouldn't have this problem if we used individual questions instead of lists!" ^
@PeterTaylor thanks, fixed the link
Is anyone interested in a Spellbinder/Waving Hands KotH (there's one in the Sandbox)//
That was Peter's idea. I'd love to see that realised (or realise it myself), but I think Peter dropped it. I left a comment, but he never replied.
@SohamChowdhury, there seemed to be a lot of interest from other people.
Oh, sorry, will have a look.
That could be awesome if done properly. I just had a look at the rules today, seems to provide a lot of options in terms of AI-making.
10:27 AM
But my position is that I think KotH was better in the early days when the games were simpler. When I saw that the trend is to games where just understanding it takes an hour or two and writing a decent competitor takes all your spare time for two weeks, I reconsidered its suitability.
I think the rules are also not entirely unambiguous, so coming up with a solid spec and a solid controller will be tough.
@PeterTaylor I see. In that case, maybe Rainbolt should get his dining philosophers on track ;)
Have you heard of this?
Chopsticks (also called Numbers, Swords, Sticks, Fingers, Magic Fingers, Split, Cherries, and Bananas) is a hand game for two players, in which players extend a number of fingers from each hand and transfer those scores by taking turns to tap one hand against another. The basic version is an example of a solved game in that so long as both players play optimally, the game's victor can be predicted at any point. == Rules == Each player uses both hands to play the game, the number of digits extended on a hand showing the number of points that the hand has. Both players start with each hand having...
@MartinBüttner, on the different versions meta-question: you might want to make it a bit clearer what kind of "different versions" are in scope. I think that the heart of the issue is about new versions which make significant syntactical changes, but I might be wrong.
Yes, that's right. I'll clarify.
Chopsticks reminds me of one of the questions in the sandbox. Possibly one of Eric's ideas?
10:31 AM
Regarding Waving Hands, I still think it would work quite nicely because you can start off simple and build on it. It seems more promising than my two KotHs in the sandbox.
The tank battle proposal is also cool.
the one with the customisable tanks?
the ones I'm really looking forward to are Mafia and Tetromino Game
I need to go buy some RAM before the shops shut. TTYL
alright, see you
Yes, the Tetromino thing is cool.
10:39 AM
@cjfaure you still with us? ;)
(hope that ping still works)
@MartinBüttner you German, right?
1 hour later…
11:52 AM
okay, stop it right there
how is it going?
12:02 PM
my computer rebooted overnight
i was running things, and now I can't even remember what, exactly
oh, yeah. I was making a ridiculous version of that bifurcation diagram, because I want to print it out
I really need to change that code so that it's not run-once
that bifurcation diagram?
THE bifurcation diagram
hah. I mean the logistic one I submitted to tweetable art
the logistic map?
though, to be fair, it really is the most famous bifurcation by far
I'm trying to output it at poster-size
yeah true
12:06 PM
i'm encountering major problems with 1) fine-tuning and 2) time
I generated one of my random paintings on 3k x 4k or something
I don't know much about printing, but you want like 300 ppi, right?
so for poster size you need a LOT more
why do we even use inches as the standard there? anyway
I forget what the dimensions of a poster are, but let's say 600x1000 cm
I doubt it :D
12:09 PM
that would be a sick poster :D
60*100 cm
okay so that's what... 24x40 inches?
is that not a reasonable size? i was eyeballing it
that's roughly 7200x10000 pixels
maybe 18x24 inches is more normal
we could settle this right now with google, but arguing is fun :)
actually I need to get some lunch
12:12 PM
12:30 PM
I hope someone will give it a go!
I do have an idea myself if anyone is interested
^ Write one, noh? It'd be cool.
ah.. you don't know I am the OP :)
my nick is just mucked up
I don't want to ruin the challenge
@user2179021 If you change your Stack Overflow nick, it also will be reflected in chat. Or you can change your parent site to PPCG, that also works.
1:14 PM
is anyone else annoyed by the fact that the text area for SE posts isn't wide enough to fit in lines, which do fit in a monospaced code block in the answer?
I'd not noticed, but I'm sure I will be if/when I do notice it
it's really annoying when you're trying to use the full length of the code block without creating a scrollbar (happens when you want to inline comment a regex or other code)
because then every bloody line wraps in the textarea
@MartinBüttner that does sound like something worth fixing. Has anyone ever raised a meta post on MSE or MSO?
no clue, I can have a look around
In your particular case would it be practical to use a separate text editor to get it just right and then paste it in?
1:27 PM
I always write my answers in notepad first so I'm unlikely to suffer this, and when I'm golfing for code-golf I comment the code as I go along because I know I'll be too lazy to do it when I'm ready to post (and don't worry about aligning too much, because it's C#)
yeah, but then I'd have to first check how long each line can be etc. it would be nice if that was immediately obvious in the browser
@MartinBüttner if you find/post a meta feature request let us know so we can upvote...
@githubphagocyte will do
@VisualMelon well I noticed this long before I came to PPCG, when posting more complicated regexes on SO
@MartinBüttner ah, sorry, I forgot that SE had practical uses
ah, yes, I've seen that answer before ;)
1:40 PM
Okay: 1. Search for/write feature request. 2. Write sandbox proposal. 3. Clean up desk. 4. Write Tetris controller. 5. ??? 6. Profit!
@Sparr I added the offset after talking to ccarton (the owner of the bot that was the biggest offender)
1:54 PM
@githubphagocyte hm I just noticed, that that's actually only true on some (including beta) sites. SO uses a larger font for the code, so this doesn't actually happen there
yeah, so uses 14px, beta sites use 12px
@MartinBüttner a larger font? Do you mean the presented answer code is in a larger font, so the maximum line length before scrolling is lower, so the shorter lines fit into the text entry box without line breaking?
So if we had that change made, it would be a restriction of what we can currently fit into an answer. Would there be other circumstances where this would bother you?
I could ask for the beta sites layout to use 14px, but I'm sure CR would cry out that they need the smaller font to fit more code on the screen
Although in Code Review the number of characters in a code line is already more than most languages would recommend restricting to.
I think it's over 90, and reviewers generally recommend restricting to 80
1:59 PM
hm k
I know how strict that is depends on the language, but I think asking for a review of code with very long lines is a little inconsiderate anyway
still... I'm not gonna request a feature which will affect tons of other sides which I don't know
well the font size also affects how many lines you get before scrolling
Perhaps we need a meta question for things we are going to request one day for our full site after beta
...one day ^^
@MartinBüttner ah yes I hadn't considered that - in that case I'd prefer to have more lines
2:01 PM
point 2 on the list: write sandbox proposal
what's the proposal?
I assume you mean a new challenge, not a proposed change to the sandbox?
yes new challenge
implement floating point square root using integer and bit arithmetic
Is this a result of discovering how poorly implemented the square root is in most languages?
it's the result of wondering how I could use my newly created tag :D
my initial idea was just implementing a simpler operation like addition
I thought it was bordering on real world useful for a moment ;)
2:04 PM
Q: Implement an IEEE 754 64-bit binary floating point number through integer manipulation

Joe Z.(I've tagged the question "C" for the time being, but if you're aware of another language that supports unions, you can also use that.) Your task is to build the four standard mathematical operators + - * / for the following struct: union intfloat{ double f; uint8_t h[8]; uint16_t i...

@MartinBüttner that looks like it was posted by you but with a very pixelated avatar...
You added to that since?
peter did
I'm trying to think of what else the tag would apply to. Maybe something impractical
Like manipulating unsigned integers as if they were floats
2:09 PM
I might scrap my proposal for now... getting the standard down to a presentable form that makes the challenge still sound interesting is beyond me at the moment
@githubphagocyte that's what I meant though ^^
I wouldn't use the struct approach Joe Z. had
type-unsaftey ?
OK - so you'd take an integer and output an integer, but the binary representations would be the same as a float and it's float square root?
just take an integer, return an integer, and the bit-representations should be interpreted as floats (by the programmer, not the program)
As long as you can refer to a standard for both int and float so it's not language specific, you should be able to get the description and rules fairly short (I think, without considering it properly...:) )
well IEEE 754 isn't available free of charge. and all the references I find are a bit more convoluted than necessary. so I'd want to include a full specification of the format in the question.
and there are a lot of gritty details like implied leading 1, exponent bias, infinity, NaN, etc...
2:14 PM
I wonder if it would be interesting with strings too? Find words/phrases that are made up only of upper and lower case letters and spaces, and whose square root also fits the same pattern. It would have to be restricted to sizes that make sense to interpret as floats somehow
I can't think of any clever ways around it though
hm, I can't quite visualise how that would work
the floating point format might be too rigid for that
I guess I won't get around cleaning my desk any longer.
I need some fast music :D
2:40 PM
@NathanMerrill offset like we had temporarily in CB1?
@MartinBüttner I will assist in your desk tidying by stopping talking. Now you just have to shut up the rest of the internet... :)
I'm actually already done (well, not "done done"... but done enough :D)
@NathanMerrill that Borg bot in CB2 looks pretty awesome
1 hour later…
3:55 PM
hi @Sparr
4:14 PM
@Sparr yes to both
4:29 PM
4:47 PM
@githubphagocyte did you ever calculate the fractal dimension of those images?
@EricTressler I didn't, no.
@githubphagocyte >:-O
hi @githubphagocyte
what eric said
@user2179021 hi
4:51 PM
there are programs out there that already do it, but they're a mix between 1995-2005 programs, paid software, or both
it's insane
It would be an interesting thing to add to the answer. I haven't ruled it out.
the algorithm isn't that complicated
It would be straightforward using a grid approximation, yes
i guess there was a spike in interest in fractals 15 years ago
would anyone be interested in codegolf.stackexchange.com/questions/37620/… ?
at the moment it is very easy to win :)
4:52 PM
I wonder whether the occasional spikes in interest in fractals exhibits fractal behaviour
you have to use a grid approximation, because your images are discrete
but the history of Mandelbrot is interesting
he was discounted by his colleagues for a long time
I recently read about that
because he doesn't really do any serious math
apparently he had no formal training and his geometric/visual approach to everything wasn't taken very seriously for ages
yeah, but it's par for the course, and perhaps for good reason
he sounded like a crackpot
and he was treated as one for a long time; there are many more crackpots than visionaries in math
people came around
4:55 PM
thanks for the upvote someone :)
now I just need an answer too :)
what is your name
why not make it your username across SE
4:56 PM
I have discussed this a number of times. It was once
then something went wrong
and now you can't be bothered to fix it :P
I think it's caused by multiple people using this computer and stackexchange's desperate need to hold on to cookies
my apologies, i wasn't party to that
because it will just break again in 2 minutes
same happens when I fix my gravatar identicon on Area 51
4:57 PM
if stackexchange were less aggressive about holding on to cookies or gave you an easy "switch user on all stackexchange sites" button that would help
It just reverts to the incorrect one next time I sign in
@user2179021 I'm sorry, I just want people on SE to use their real names
it's stupid, but it subverts a lot of things
I don't even understand how it holds on to them so aggressively. If I delete all cookies it still seems to keep track of me unless I quit the browser and restart
@EricTressler can you suggest a "switch user" option?
nah, nobody listens to me
you may have some things in cache, btw
so if you delete cookies, it might only be in stored memory and not cache?
@user2179021 you can suggest a switch user option
4:59 PM
it must be
that's what meta.SE is for
@MartinBüttner who to?
just post a question tagged with feature-request on meta.SE
If so, i will do it now
5:00 PM
But the primary use of that would be to post something under a pseudonym and then treat it as coming from outside
so if I were a mod, I'd naysay it
It would make it easier for bots to get SE points
sock puppeting is already a thing
@EricTressler don;t you have to have a valid email address to be able to get an account?
what does that mean to you?
my main problem is trying to work out the right tag :)
5:04 PM
I use gmail
having additional pseudonym (sock puppet) accounts for posting answers
if someone compromises that, then I am slain
why would it be any easier for a bot?
that's allowed unless you do anything that you couldn't do with a single account
I mean a bot can delete cookies more easily that a person can
Q: Suggestion: a switch user id option

LembikStackexchange seems to try very hard to hold on to your identity. In particular, even if you delete all cookies and log out, it still somehow often remembers who you are. If you have multiple people using the same computer and different stackexchange sites it can all get very confused and confus...

5:06 PM
do you understand how?
do you know how it tracks users or not?
only through cookies
so I think the answer is 60%
"somehow" implies that you do not
I certainly dno't have a full understanding
I in particular don't understand memory caching of cookies, when this happens, on what browsers etc.
5:07 PM
well, 1) It's not inherently bad, but 2)
2) you should know what's going on before you post a question
@user2179021 you might want to clarify that your main use case is multiple people using the same browser
If multiple users use the same browser, you should create different profiles.
what extra-cookie steps can a browser take?
That's not just for Stack Exchange. That's for everything.
take me as an adversary. what can I do to you to thwart the basic guards against malicious cookies?
5:12 PM
@Dennis I added some more clarification
gmail helps you do this
it's not such a strange request :)
5:39 PM
Not saying I'm against it, but one profile per user is still a good idea imho.
Logging your out of all SE sites at once could be a hassle. As it is, you're logged in on each site separately, using once of the two gazillion different authentication methods.
6:22 PM
@Dennis right.. I don't know how hard it is technically to do that
@Dennis I wonder if the Eban numbers code could be even shorter if CJam supported mixed-base conversion :D
@EricTressler Depends on what you're printing. How far away will people stand to look at it?
@user2179021 Are you using OpenID? It might be cookies from a different site which you need to clear.
6:40 PM
@PeterTaylor Thanks for recommending CJam the other day. I needed a language with concise base conversion for large bases yesterday, so I looked into it, and it was actually good fun (although I've obviously been missing lots of tricks and Dennis golfed off about a third)
Glad you enjoy it, although it does mean one more CJammer to beat my GolfScript entries.
I'm not sure yet if I'll start golfing regularly in it other than to learn the language.
Golfing in Ruby and Mathematica is still fun, of course.
But Ruby doesn't handle bases beyond 36 and Mathematica... well... the code was more than three times as long ^^
damn I need to finish the marbelous interpreter
moving to a new city really slowed down my silly projects temporarily
So CJam was an obvious choice there... I guess I'll keep it like that in general... if the languages I want to golf in lack the features to implement the algorithm I have in mind, I might switch to CJam
@Sparr I really need to get back on track regarding all things marbelous :P (and then write a JS-based interpreter ^^)
@EricTressler, these printers recommend 100dpi for scientific posters.
6:47 PM
but I don't think it's gonna happen any time soon
I really want to post some marbelous challenges
I got all the code contributors to agree to a license change, so we can use stuff from SE
@PeterTaylor Also, you could.... you know... switch to CJam if being beaten by it really bothers you ;) (which I actually doubt)
btw, how come there is not a single FlogScript submission on PPCG? I only heard about it when you mentioned Anarchy Golf recently (for the domino circuits scoring), because it seems to be used quite a lot there.
(I hate it when I try to google something for C and only C++ and C# answers pop up)
I've invested enough time in learning GolfScript and writing tools for it that I don't want to switch. Actually, I recently came across Ruboto and now I'm wondering about trying to fork its REPL project and put GolfScript on my PDA.
That would give much faster feedback than my current approach when AFK of writing GolfScript with pencil and paper.
So I'm writing the Tetris Tournament in C... does anyone know if C's rand being as bad as it is would make the tetromino sequence too easily predictable?
@PeterTaylor No I just click on "log in to stackexchange"
and type in my name and password
I don't mind really.. It just all looks like something fixable
@MartinBüttner you can literally predict all the numbers if you are given two in a sequence I believe
I mean two consecutuve
7:03 PM
Well, you are given the numbers modulo 7
No that's not even true
My random generator will be a bit more involved
it's not a bad math.stackexchange question :)
A lot more information will be lost on the way to the player, so it probably doesn't matter
I have this random number generator, is it easily crackable
7:04 PM
I'll be generating 7 numbers in a go, and will use their order to determine the order of all 7 tetrominoes. In addition you don't (initially) know where a bag of 7 tetrominoes starts and ends.
do you think you will try my puzzle?
I may at long last found a challenge too hard for PPCG :)
Sorry, currently not. I'll spend the time I have over the weekend on the tetris controller, testing submissions for False Positives, and maybe some simple recreational golf if a good question comes up. But I don't feel like spending time on a code challenge this weekend.
@MartinBüttner ok I understand. There are 7 days left of the bounty so hopefully you will get a chance before then
@Sparr what is a marbelous challenge?
I have some more, possibly hard challenges up my sleeve
I don't quite understand what is acceptable for a pop-con yet
can anyone give me any hints?
7:10 PM
cjfaure came up with it, I made a push to come up with a spec and turn it into an actual language and Sparr jumped on it to implement the first interpreter :)
@user2179021 Hm, not really... it's really a case-by-case thing.
If you ask Peter, the answer is "nothing" ;)
right :)
ok so here is a mathsy challenge
I think graphical contests are often best done as pop-cons, because you can't "validate" the output. But at the same time you have to be careful that it's still actually a programming challenge and doesn't turn into an art contest.
@user2179021 no
true although I did really like the tweet art ones
which were really art contests ;)
@MartinBüttner then my question won't be very interesting to you :)
basically I would like a challenge where your code has to automatically produce math questions with their solutions
@user2179021 I believe that one in particular was very much a programming contest. Have you seen the golfing magic some of the answers needed to use to squeeze their code into 3x140 bytes?
one type would be recurrence relations
@MartinBüttner you are right but I think there was a lot of artistic skill in some too, Did you see the compression one?
7:14 PM
@MartinBüttner Not quite. The answer is "Posting in the chatroom rather than on the main site"
Of course, there is, that's part of why it was so fun, but this one definitely made sure that you needed to golf out every last byte if you wanted some crazy results, like phagocyte's action painting.
@user2179021 a challenge using only the marbelous language
I agree
I have to say stackoverflow.com/a/919989/2179021 is really beautiful
@user2179021 There have been a few of those IIRC
although not on this site :)
7:15 PM
@PeterTaylor is it a pop-con if you can't add a pop-con tag to it? ;)
@Sparr hi!
@user2179021 that challenge is crazy!
@MartinBüttner yep :)
@MartinBüttner It was a contest which required programming skill to enter, but I suspect that the voting was 90% on the pictures, so as a contest it was really an art contest.
@PeterTaylor Hm, fair point.
7:18 PM
@Sparr as someone who did such an amazing job with my last challenge. Can I tempt you with codegolf.stackexchange.com/questions/37620/… ?
not right now, perhaps later
@Sparr ok cool :)
I do have some ideas of my own
but they are unlikely to be as good as those the community comes up with
7:49 PM
@MartinBüttner Probably. I've never used a language with such a feature. Looks like it would have helped with my electrons answer.
I don't know if such a language exists
I think not even Mathematica can do it, except for the special case of dates
might be worth suggesting to aditsu
I currently have 11 open tickets on his SF page. I think I'll take a little break from suggesting things. :P
I made a version of that Twitter Image Encoding challenge a while back: codegolf.stackexchange.com/questions/11141/…
Oh I wanted to check whether irrational bases work in Mathematica
thanks for reminding me
they should though, even symbolic bases work
sick, they do
7:56 PM
<opens mathematica>
@Dennis I created a ticket
8:22 PM
Q: Posting closely related challenges: Bonuses or separate questions?

DoorknobConsider this: I have an idea for a challenge, such as "Find the optimal foos such that the bar is greatest." I also realize that I could alter it slightly and add a more interesting twist to it: "Create an animated GIF that shows the bar approaching the greatest value as the foos change." I don...

Are there any particle physics-based challenges?
@PhiNotPi hmmmm standard model and stuff?
I'd be up for that :)
(writing and participating)
Soon, code golf will have all standard programming tasks, and all you need to do is patch together different questions to make any application, pre-minified!
@NathanMerrill search for "duct tape coding"
(on PPCG)
I wouldn't know enough about the standard model to write a challenge.
8:38 PM
@MartinBüttner there should be more of those
I second that.
What would happen if we combine duct tape and Twitter?
9:33 PM
I've got 3 duct-taping ideas on my list
but I never bothered to write them up
a) duct tape + code golf instead of pop-con
b) polyglot duct tape
c) prove that duct taping is turing complete
9:48 PM
Turing-complete duct tape seems the most interesting out of the three.
i just woke up. but thank you, Peter, if you get online later
I guess a problem is that there have been several challenges in the past like "Interpret [Turing-complete language]" which could make it too easy.
oh, he's online now
@PeterTaylor thanks for the information about science posters
10:04 PM
Does anybody know what this thing is?
*----------*          *----------@
|          |           \
|          |            \
|  *----------*          *----------*
*  |       *  |       *----------*  |
|  |        \ |        \         |  |
|  |  *----------*      \   *----------*
|  *  |       *  |       *---\------*  |
*   \ |    *-----|----*       \  *     |
 \   \|  *----------*  \       *----------@
  \   *  |   \   *  |   \   *----------*
   *     |    *   \ |    *   \      *
*--|-----|-*  |    \| *--|----\--*  |
 \ |     *  \ |     *  \ |     *----------*
It's an ASCII art fractal of some kind.
10:17 PM
Miscellaneous challenge idea: protein folding
I've taken up a game called Foldit.
10:39 PM
@PhiNotPi Looks like a 3D Hilbert curve
@PhiNotPi you can't take code from PPCG
just SO
@PhiNotPi a simplified version would have been one of my Gödel, Escher, Bach challenges
1 hour later…
11:57 PM
@MartinBüttner in what context can you take code from SO but not PPCG?

« first day (1323 days earlier)      last day (2126 days later) »