« first day (1298 days earlier)      last day (3538 days later) » 

5:00 PM
How "tailored"? What features are you looking for?
 
@githubphagocyte there aren't, but you just jynxed it.
 
jynxed? You mean now there will never be? Or now there are bound to be?
 
@githubphagocyte I'm already making it.
 
@cjfaure yay! Which languages...? ;)
 
Make sure you golf the source to the IDE also. Preferably in a dedicated golfing language.
 
5:01 PM
/me hits @Rainbolt with a diamond shovel
 
@githubphagocyte i was joking; but i might xD Python probably.
 
@Geobits golf the source?!? Is to ensure it can never have an update...? :)
 
@Geobits The sacrifice between features and code length...
 
@TheDoctor /me Jumps inside of TheDoctor's backback. I suppose it's better than being picked up by an Enderman.
 
I've been Minecraft-sober for 3 months now.
 
5:03 PM
@githubphagocyte How else to prove you know how to golf? Why should I trust golf-related source that isn't even golfed? :P
 
@Geobits Technically, a blank program is a golfing IDE. You open it in your code editor.
i am the king
 
I'd call that a golfing template instead.
It starts with the minimum boilerplate needed.
 
@Geobits "ide" is shorter than "template".
 
/me places @Rainbolt around a beacon to protect it
 
now i wanna play mc with you guys xD
 
5:25 PM
Speaking of golfing tools, I think something that folds constant expressions would be useful
Should be easy to do, too
. o O ( Perhaps easy enough that it could be a suitable golfing task itself )
 
@FireFly Fold in the sense of code hiding?
@MartinBüttner, you do realise it's not a popcon? ;)
 
I assumed @FireFly meant combining expressions involving several constants into one literal
 
@PeterTaylor as in evaluating expressions consisting only of constants
 
I think that's already been asked for integers.
 
5:32 PM
@PeterTaylor yeeaaaah, I do... that's why this gif is all the way at the bottom of the pos... but the specified version of the problem was just a bit too boring :D
 
That seems to be a blueprint for a Trigonometer
They're used to measure trigons
 
6:07 PM
@VisualMelon it looks like the new version of your solver is throwing an index out of bounds exception
did you test it after golfing it down to fit in the character limit?
on another note, depicting Yggdrasil as a Christmas tree is making me incredibly sad :(
 
Where
 
there's an yggdrasil related kolmogorov challenge today
have a look at the front page
 
@MartinBüttner I didn't golf it, I just only used one space for indents... I'll investigate - do you have a line number for the exception?
 
no but I've got
Unhandled Exception: System.IndexOutOfRangeException: Index was outside the bounds of the array.
   at dominoPrinter.Program.<>c__DisplayClass5.<printDominoes>b__4(Int32 x, Int32 y, String str)
   at dominoPrinter.Program.printDominoes(TextReader reader, TextWriter writer, Boolean moreverbosemaybe)
   at dominoPrinter.Program.Main(String[] args)
you've got an avatar!
 
yes, what's the verdict? :P
 
6:21 PM
Too low-contrast
 
it looks like lemony infinity
 
Should be on another color background
 
I can't reconcile your avatar with your name
it's like using an off-centre turtle, my head just doesn't wanna process it
 
@MartinBüttner is it better since I got rid of the off centre dancing turtles and replaced them with just one reasonably central one?
 
much better! ;)
 
6:23 PM
@MartinBüttner you realise you are off centre in your avatar?
 
Maybe his head's just off-center from the rest of him, and you're being rude
 
lol I knew this was coming
 
sorry @MartinBüttner I didn't mean to draw undue attention to your off centre head
2
 
which made me think why photographs should never be centred on what their showing, while we really like symmetry in all sorts of abstract graphics
 
the rule of thirds is pretty accurate for a lot of things
 
6:26 PM
OH DEAR
I didn't do that, really...
no apparently I did
 
you're really fun to "watch" debugging
 
@MartinBüttner it seems I replaced " "*4 with " " instead of tab, so a certain set of important strings containing spaces MAY have broken
 
(about as much as Rainbolt is fun to watch codereviewing)
@VisualMelon oh
 
debugging is always more fun when you're not involved in it
 
@VisualMelon you did keep around the previous version though, didn't you?
because you never sent me the fixed but indented version, so that I'd have it in the git already
 
6:28 PM
@MartinBüttner aye, just submitted
I'm so used to indenting stuff last just before posting that I assumed it wouldn't be an issue (because replacing tabs won't be an issue)
but because it was an afterthought after replacing tabs, clearly it went wrong
 
well you can still have a look at the version in the git to fix it
 
I've uploaded a "fixed" copy
 
@PhiNotPi are you sure you don't program Propeller chips?
 
(to the question)
 
ah okay
 
6:30 PM
@VisualMelon if you start with the unbroken version, can you replace "\n " rather than " "?
 
did you click "submit" yet? :D
@githubphagocyte or just "^ "
editors are usually in multiline mode
 
@MartinBüttner I did, but I didn't notice this message "Body is limited to 30000 characters; you entered 30028."
give me a minute
 
user image
3
I fixed it
 
6:31 PM
I'm reasonably sure.
 
Well, I'm not.
 
oh
well back to code
 
AH, I reimplemented strness, I'll have to remove a comment
@MartinBüttner reposted
 
6:33 PM
@MartinBüttner I fixed @VisualMelon's avatar
 
oh lol
I didn't even notice that those eyes were lemons
wait
lemon <-> melon O.O
mind blown
 
It's an anagram!
 
I know, I'm working on putting googly eyes on a watermelon for him
 
@MartinBüttner you missed @VisualMelon's explanation earlier...
 
oh
link anyone? :)
 
6:35 PM
Speaking of avatars and anagrams and things, how obvious is it that my avatar is an ambigram?
 
@FireFly very
 
so, you guys reckon the contrast is too low?
 
(in your profile, not the scaled down version in here)
 
Hehe
 
@MartinBüttner I found it but I don't know how to link to a specific chat post
 
6:37 PM
found it myself, don't worry
 
@VisualMelon I think I'm the only one who complained
 
@EricTressler not everyone can have such a high contrast avatar as you... :)
 
but if you hover over it, there's a link button on the left (if you're in the frozen transcript) or just the usual menu on the left, where there's a permalink option
 
@MartinBüttner yes, that's exactly where I hovered and still managed to miss the permalink option...
 
addressing everyone other Eric whom has already expressed his a sour opinion - how can I improve my rather yellow/green profile picture?
 
6:40 PM
@VisualMelon how about purple and orange stripes in the background?
 
@VisualMelon cell closeup! yeeaaah
 
For the sake of the tiny avatar in chat, you could zoom in slightly so the lemons touch the edges of the box
 
@githubphagocyte well, that sounds like a very good suggestion, but I've managed to not save the iteration where I cut the cut lemon out, so it would require alot of effort
 
@VisualMelon okay, I'm quickly testing an improvement of the controller, and then I'll get you some results for your solver
 
6:42 PM
@MartinBüttner roger
 
for what it's worth, I like it as it is. Won't be mistaken for anyone else.
 
I think it's fine as is, @VisualMelon
 
lol, my avatar comes from the same place as @EricTressler's
 
ok, if I'm not feeling lazy some time I made add a mad purple background, but if it's not too objectionable as it is I'll defer that thusly
 
6:44 PM
now Eric is it look easy!
/me samples the purple
 
where's full res? I wanna try ;D
 
my apologies to anyone who doesn't want the chat filled with pictures of melons lemons
 
3 hyphens
 
three dashes :P
yay
 
Huh, I thought ~~foo~~ was the semi-standard markdown extension
 
6:49 PM
okay, I'm done.
 
Yay. :P
 
Hm
 
i could do better, lol
 
I like the background though
 
Hmmm.. in-chat popularity contest... :)
 
6:52 PM
I wonder how I managed to get transparency working in my avatar.. I uploaded a new one now and it didn't like the transparency anymore
perhaps it doesn't like semitransparent pixels
 
I think transparency works via .gif and only .gif :P
 
:(
 
Single frame gif?
 
Using gif worked
 
what the... the controller just exploded to 5 gig RAM when validating the 3-bit adder
what have I done
 
6:56 PM
@MartinBüttner has it become self aware?
 
ah no, Visual Melon's solution must have been ridiculously large
the solution got ditched before being evaluated, because it had more than 8M cells
(51M in fact)
 
Hm, I guess transparency only works for gravatars for some reason
 
I think I need to improve my controller to do that check first before doing some array-based manipulation on the circuit :D
 
using a gif with transparent pixels caused them to become white when uploading to SE
 
interestingly, the multiplier (which has more outputs) had only 43M cells)
@VisualMelon you might be interested in these numbers ^
 
6:59 PM
for >4inputs, the size is proportional to the number of 1s outputable
 
if anyone wants to look at more numbers:
15
Q: Domino Circuits

Martin BüttnerScoreboard Here are the raw scores (i.e. domino counts) for VisualMelon's submission. I'll turn these into the normalised scores described below, when more answers come in. The existing solution times out on three circuits and produces too large circuits on 2 others, but solves the rest successf...

 
if you read that and thought "I don't want to look at more numbers" consider yourself in the wrong
 
7:14 PM
I wonder how well a brute-force solution would do
I think it might be able to solve the 2-bit circuits optimally
but would obviously fail for the large ones
so it might even be somewhat balanced with your submission
but I wonder how small you can actually get those 2-bit circuits
huh?
 
wrong brute force
 
? did you brute force the keyboard?
 
I brute the expression tree, then it's all pretty tame after that
 
ah yeah
no I meant brute forcing it cell by cell
 
aye, took me a bit of time to work that out
 
7:34 PM
So apparently Twitter seems to like us, especially the Tweetable Mathematical Art challenge (not surprising). Yay for more users and traffic!
 
Nice :)
 
Nice
 
o.O
two tweets by Jeff Atwood
both for constests
this is sad
actually no, the first one is a code golf, just the title seemed like it was underhanded
 
8:00 PM
We should take this and turn it into a challenge to write a program to generate a Golfscript program to generate an input and still be shorter than the input, and then compress the script it outputs to create a text compression format.
 
I had to read that a few times too often to get it
 
@MartinBüttner lol, wait, let's also make one to compress Golfscript code and then use the best submissions in conjunction
LZMA2 is toast ;D
 
I wonder if GS programs have enough entropy to effectively zip
I guess since they're ascii, yeah
 
You can compress them further if you know they'll follow specific syntax rules.
the only actual syntax construct in golfscript is a string, right? :P
 
@EricTressler How do you even determine the actual entropy of a program? There's so much information "hidden" in the implicit meaning of the syntax.
 
8:05 PM
This is going to something like TarGolfscript, where the interpreter first unzips the source, then runs a GS interpreter on it
 
Well
 
@MartinBüttner You can't directly measure entropy (in practical examples), but compression algorithms are a good approximation
 
entropy is actually really simple to measure if you have a list of every possible thought that could ever exist.
 
Entropy is simple in a mathematical context
But you people, with your "real world", ruin it
3
 
@cjfaure I've seen that either in a question or in the sandbox.
 
8:08 PM
@PeterTaylor ow.
 
I do actually have a Java program which generates GolfScript programs for pure questions.
 
i can haz?
 
I actually had an idea for a code challenge
 
(i won't use it on questions, I swear! xD)
i wanna make that compression format
 
@PeterTaylor Maybe you know something about this?:
 
8:10 PM
I'll have to see whether it looks presentable enough to make it public.
 
@PeterTaylor Presentable code is overrated
 
I wanted the challenge to be taking two measures of randomness, and to produce 2 strings, one which scores as high as possible on metric A and low on metric B, and vice versa for the other string
but then I realized that I didn't know any simple metrics for randomness, and the best I could do is rip things out of the Dieharder tests
Since you can't directly measure Kolmogorov complexity
One metric could simply be compressibility under LZMA
do you have any ideas for an easy second choice?
 
An interesting proxy is to lower-bound with two different compression algorithms. It might be quite an interesting challenge to find strings which have very different compressability with LZW and bzip.
 
I don't know how bzip works; I'll have to read about it
 
But I'm not sure that it would be a very good question for this site.
 
8:13 PM
but yes, that's the general idea. I wanted to choose two things that are trying to measure the same thing, but to thwart them
 
The key element of bzip is the Burrows-Wheeler transform. (Yes, Wheeler is the W of LZW).
 
No? I was thinking of asking people to write their own function to generate the strings from thin air, no using random libraries, and only 32 bits for magic numbers
 
You'd have to find a suitable wording to make the question about code rather than about the strings per se. It's a similar problem to the one with the md5 of ffff...
 
No math constants, etc.
 
I wrote a compression format once.
I build a sort of matrix around generating a new character mapping to squish the unused bytes out of a string, sometimes going into very many dimensions,
 
8:16 PM
Have you seen the code that compresses any file by 1 byte? (Except an empty file, obviously)
 
@PeterTaylor Do you think, if I could plug all of the obvious holes, that this could work? Or still a bad idea?
 
and printed the sum of all the values and the product of all the values.
 
@githubphagocyte that's impossible to do losslessly. Is it a joke?
 
Ok, that might work. 32 bits for magic numbers needs a bit of bulletproofing ("But 68 is only 7 bits, even though I'm storing it in a 32-bit type!")
 
Decoding it, was, of course, O(n log n), but hey
 
8:17 PM
I think it was somewhere between a joke and an educational example. It was lossless.
 
it worked really well
 
@EricTressler I read the code and found that it worked, and then resented having wasted the time to read and understand it once I got the loophole
 
@PeterTaylor Yeah, but I could just say "you have an array of 32 booleans that you can initialize to whatever you want" or something
@githubphagocyte What's the loophole? Does it delete unimportant file metadata?
 
@EricTressler that wouldn't work iteratively down to zero bytes
@EricTressler hint: doesn't work on Windows
 
Ahhh....... I don't know, does it push it onto dev/random/?
I don't know enough about how linux works. Does it delete the initial '/' from the path?
 
8:20 PM
@EricTressler I'm happy to tell you once you stop guessing - it's really disappointing
 
Ignore that last one, that doesn't work iteratively. Was stupid to say. Okay, tell me
 
@EricTressler it only works because most file systems don't include the size of the filename in the size of the file
 
@EricTressler Why would I need 32 booleans? 2 is enough, it's all in the indexes you use to get at them.
 
@EricTressler but that means it won't work on a system that has a cap on filename length
 
8:22 PM
@PeterTaylor I guess I should also include a byte limit on the string generating function
Don't worry, I won't post it unless I can prove to myself that it's not ill-conceived.
 
@EricTressler I knew it was impossible while reading through the code, but I was expecting a more interesting loophole :(
 
Yeah, but really, any OS that allows some more nefarious way to hide bits of files is doing kind of a bad job
 
Heh. It could corrupt the least significant bits of the loading splash screen, for instance
 
@EricTressler that's a reasonable amount of spare space...
not very portable though
 
8:26 PM
Or I guess even your desktop wallpaper
That would be interesting, to just store all of your passwords in your wallpaper in case you forget
 
@EricTressler not now you've mentioned it publicly ;)
 
I've always been interested in steganography and things like numbers stations. I really wish someone would declassify something that would shed some light on that whole business
It's pure idle curiosity, but I'd really like to know
 
@EricTressler it is indeed fascinating. I don't have much knowledge of it but it's intriguing. I'd be interested to know how much it's in practical use.
 
I don't think there's much to know except that there are radio frequencies that just broadcast sequences of numbers
It could be used to convey any kind of information, or it could be a prank
 
But there's also a whole world of other places to hide information.
 
8:41 PM
Yeah, but coming up with secrets worth that kind of effort is hard
 
@EricTressler Or some half-way house. A way to keep the boss' nephew busy. A service which had cryptographic value back in the 70s but nowadays survives because no-one can work out which office's budget is paying for it.
 
not the boss' nephew's son then?
 
@PeterTaylor New conspiracy theory: Imgur is actually run by the CIA to coordinate its spy network
You heard it here first!
 
hmm
 

« first day (1298 days earlier)      last day (3538 days later) »