« first day (1163 days earlier)      last day (3683 days later) » 

3:01 AM
Anyone here use android? Perhaps you could help me (I'm trying to get started):
0
Q: AVD Manager, why can't I create an AVD?

QuincunxFor some reason, I cannot create an AVD. I went through AVD Manager - Cannot Create Android Virtual Device and attempted the solutions, and I also fiddled with the different settings in the manager, including changing the name to Device. I made sure that the system images are installed, and I fol...

 
@Quincunx Try this command in terminal android-sdk-location/tools/android avd
 
@Rusher Wow, it's working, but do you know why I can't check Use Host GPU?
 
HAHAHAH I know absolutely nothing about Android. I just Google better than you
It was literally the first result "create avd ok button"
OH hey @ProgrammerDan. Good timing.
I got R installed, added R to my PATH, downloaded the Gist, got the wrapper into my project with the right name, and don't really know what to do with the .R stuff
 
3:18 AM
@Rusher, what's up?
 
I have to admit that I am not very good with terminals
 
:O
what're you struggling with?
 
Well... I have this R code... but where does it <i>go</i>?
really chat? really?
 
use _
`_go_
go
It goes ... where-ever
what's your operating system?
 
> _go_
Error: unexpected input in "_"
> go
Error: object 'go' not found
> _go
Error: unexpected input in "_"
>
 
3:21 AM
lol
 
I am talking about the terminal in R
I used Windows 7
 
right, you don't need to use the terminal directly with R
oh, simple
did you add R to your %PATH% var?
alternatively, replace the invocation line with the absolute path to the Rscript executable
 
@Rusher Thanks.
 
Here is what my path looks like PATH=C:\WINDOWS;C:\WINDOWS\SYSTEM;%R_BIN%
 
and what is the current value of %R_BIN%?
also, if you open CMD, can you type R and get an R prompt?
 
3:25 AM
C:\Program Files\R\R-3.0.3\bin
 
ok cool
 
Ok I deleted R_BIN and my path just has the full path now
 
so, you should just be able to use the invocation I already have in the wrapper, with one change. Replace "WolfCollectiveMemory.R" with the full path to the R file. If there are spaces in the path to the R file, put it somewhere without spaces.
Any luck?
 
lmao nice error message. Starting WolfCollectiveMemory remote process
WolfCollectiveMemory ended catastrophically.
On debug, this is the path for the .R file. I moved it inside of the netbeans project so I could use a relative path - C:\Users\MYNAME\Documents\NetBeansProjects\Wild\src\animals\WolfCollectiveMemory‌​.R
 
:D :D
You can't use a relative path like that on a process invocation, iirc
 
3:38 AM
why is the username on your computer "MYNAME"?
 
String path = System.getProperty("user.dir")+"\\src\\animals\\WolfCollectiveMemory.R";
It comes out to the absolute path
 
That'll get you C:\Users\MYNAME\documents\src\animals\WolfCollectiveMemory.R
 
When I said relative, I meant like it works if you move it
 
kk
Why not just put all your remote scripts in a known, higher-level path?
like C:\RPI-Wolves\
and just replace the WolfCollectiveMemory.R in this line with C:\RPI-Wolves\WolfCollectiveMemory.R?
 
I don't understand what moving to a higher level path solves, but ok.
 
3:42 AM
meh I'm not as familiar with Netbeans, so I figure simplifying is best
 
Sorry I missed your question earlier. When I type R in cmd, I don't get an R Prompt
 
Oh.
You probably installed the 64bit version?
 
yup
 
ok, add x64 to your path
for R
b/c that's the problem :)
 
How do I do that?
 
3:49 AM
Shit, I gotta run
but in your System Advanced properties
you can configure your PATH
just add to the R path the x64 string
so it looks at the right executable
 
Control Panel\System and Security\System\Advanced -> Environment Variables... -> Path -> Edit
 
 
2 hours later…
5:59 AM
@Rusher any luck?
 
6:21 AM
I had a pleasant conversation on The Tavern (@ meta). I thought I'd share it here (starting point): chat.meta.stackoverflow.com/transcript/message/1952280#1952280
 
@Quincunx like this
 
7:14 AM
After all this futzing with apl, I'm thinking of making an array-calculator challenge. '1 1 1+2 2 2' => '3 3 3'. Just 1-D arrays and basic mathops +-*/^. Right-to-left evaluation of ops.
I still haven't gotten inca to do a jot-dot, but I think I'm almost there. It'll accept the syntax, but it just gives you a row- or column- vector, not a full matrix.
The jot-dot part itself actually works, but I need make the mathops smarter.
The mathops currently treat all arrays as 1-dimensional.
 
7:49 AM
I've also added procedures but not local variables or functions. procedures can interact with global vars only.
 
8:06 AM
@Quincunx You do realise that some people catch up on chat over breakfast?
 
 
2 hours later…
9:59 AM
but, rainbows.
 
10:25 AM
What about a translator into the stick-figure alphabet from Sherlock Holmes?
I never did track down a complete alphabet, yet.
I don't even know what this says, I just copied the pictures from the tv show.
But pictures are always fun. Translators are always fun. You get to come up with encodings and stuff.
 
10:47 AM
Or royal measurements unit convertor. nb.
1 tun = 2 pipes
1 pipe = 2 hogsheads
1 hogshead = 2 barrels
1 barrel = 2 casks
1 cask = 2 bushels
1 bushel = 2 kennings
1 kenning = 2 pecks
1 peck = 2 gallons
1 gallon = 2 pottles
1 pottle = 2 quarts
1 quart = 2 pints
1 pint = 2 cups
1 cup = 2 gills
1 gill = 2 jacks (<-- jack&jill was a drinking song!)
1 jack = 2 jiggers
1 jigger = 2 mouthfuls
1 mouthful = 4 drams
1 dram = 60 minims
Hm. I think I'll hold that one until I get inca to do base conversions. hehe
Well, them's my pipe dreams. worth about 2 hogheads.
 
 
2 hours later…
1:27 PM
I like to write Ms. Windows, like it's a wee girlie-girl.
 
 
2 hours later…
3:47 PM
@ProgrammerDan Nope. I can run the R program from cmd line but can't run it from java using the exact same string.
 
@Rusher I've been doing a lot of testing
the version of @plannapus' program you have won't work, he chews up file pointers like you wouldn't believe
but I fixed it
about to post a new version. As for getting it working for you, let's dive into that
can you post the ProcessBuilder invocation line here?
 
Path to WolfCollectiveMemory.R is C:\WolfCollectiveMemory.R
ProcessBuilder pb = new ProcessBuilder("Rscript C:\\WolfCollectiveMemory.R".split(" "));
 
and the error message you receive?
 
holy shit it worked
I guess I needed to reboot... because I changed NOTHING today
 
:D
yeah with the version you have, after about 150 commands or so, the process will fail due to too many file pointers
grab the new gist here gist.github.com/ProgrammerDan/… and you'll be golden
 
3:55 PM
Yea one of them failed to move already
It was that damn Wolf number 25
 
yeah that's due to the file issue. Basically plannapus was opening a new connection to stdin every program loop
 
Did you edit both files or just the r
 
and opening a new file connection to his "collectivememory.txt" every command process
just the R
I'm getting to know R a bit -- not the excuse to learn it I was expecting, but okay
 
Huh... his wolves are looking morel lively now
 
they always just sit there for me :D
but I haven't added any of the other players yet
 
3:57 PM
I meant they are moving faster. Before it was slow
 
how do you feel about the submissions that subvert the Wolf character?
ahhhh yeah
not reading from a file all the time will help with that -- now all his collective memory is indeed in memory.
 
Sweet. Now I can post results unless another submission was made.
You have been a huge help. Too bad you wiki'd your wrapper
Oh, that's interestinmg
 
Hehe, not a problem, I still got a badge out of it. And it was fun to write.
 
I modified the controller to loop 5 times for 1000 iterations rather than just "while true"
And now every CollectiveMemory after the first trial fails to initialize (but they are still present and moving on the board lol)
 
That won't work with my current wrapper...
lol
 
4:02 PM
I'll try to fix it
And I might use this as an excuse to learn github
 
Goooood
You'll need to explicitly stop the WolfProcess thread, and set the static version of it to null inbetween each "round"
 
Oh, ok. I'll use this as a chance to learn multithreading then
 
I thought you were going to stop and start the Wild program inbetween each round, I don't currently have any support for "reset", because there are no hooks in the Wolf construct to allow the Wolf to know when the simulation begins or ends
 
I can do that and it would work fine, but programmers are supposed to be lazy
 
Enjoy :). I make use of them all throughout the wrapper, to handle basically all inter process communication.
Maybe you should add to the Wolf base class a "cleanup" or "reset" method
all submissions with static memory or other things like that will have the same problem, they won't ever restart, and just "continue" -- perhaps with new child Wolves, but the same "shared memory"
 
4:08 PM
Am I right in assuming that after I stop the main thread all of the other threads just die?
 
Not always
 
Great. Now I have to check my processes to see what garbage is leftover lol
 
Nothing is there. And even more surprising, it doesn't look like the memory for the program is growing at all (and I have all submissions in there)
 
I think the process will die in this case b/c the STDIN/OUT pipes will get closed when the JVM terminates, and R will exit, gracefully or not
lemme throw a static gracefulReset() into the wrapper for you.
 
4:13 PM
I think I can do it
 
Cool. Going to submit me a pull request once fixed ? :D
 
Something like that XD
 
Hehe, awesome.
 
4:28 PM
Hmmm I'm at a weird obstacle. I could just stop your thread in main, but then the main class is basically babysitting the wrapper. Is there a method that gets called when you are garbage collected?
 
there is, but it's bad to depend on, b/c the GC is tempermental
also, my wrapper isn't leveraging finalize to cleanup, for better or worse
 
Ohhhhh I can't actually run the trial three times in a row. I think that some others may be persisting data as well, and if I don't sanitize the entire thing, they will be misled
Just going to run it five times using 5 fresh copies of the project
 
Yeah, that's what I was saying above :P
@Rusher any thoughts on the submissions that subvert the 'W' character? legal/not?
 
4:44 PM
Wolf
Only wolves that are submitted as an answer will be included. Represented by 'W'. Moves with any Move. Attacks with any Attack
Not legal
If not because of the spec, then because I said so
Was there another one besides doorknob that I missed? I'm pretty sure he was just joking
 
@Rusher Yes, this one:
This one:
0
A: Survival Game - Create Your Wolf

Ilya_GazmanChameleonWolf V 1.2 Especially designed to hunt lazy wolfs ;). Working pretty well on sheep too. The idea is moving as collective. When one wolf decide to move all the wolfs moving in the same direction. So If I decide to hunt my self, I will flee. I implemented this using AtomicInteger to suppo...

 
^^^ that one
 
That randomly chooses which animal.
 
Oh dang. That's a new one. I knew I should have said something, but I thought Doorknob was trolling and I didn't want to spoil his joke
If someone did manage to build a Wolf that could gain an awareness of it's relative location on the map, then they could attack LazyWolf in a sort of pincer movement. LazyWolf is winning btw.
 
Way to go lazyWolf :D
the new Node.js submission is broken
now I get to learn Node.js -- it's not clear to me that people using my wrapper are actually trying to invoke their programs
b/c if they did, they would find most of their problems just at the command line, using STDIN attached to console
 
4:56 PM
I don't understand how gamblerwolf works
 
Yes. There are a lot of wolves alive, but it is really hard to understand.
 
GamblerWolf tries to pre-read the next random value from the random number generator
in an attempt to "look ahead" at what the other wolf will choose to pick, in RPS
Not sure the approach will work, but it's damn clever
 
Ah, ok. He's not manipulating anyone else's wolf so that seems fair
Might want to roll a few times to survive that guy lol
 
Now I can't decide between GamblerWolf or GatheringWolf as a favorite
The latter is just so CUTE. He named a variable "loneliness"
 
5:04 PM
lol
 
5:18 PM
Time to go watch MTG. Got a huge tournament this weekend in Dallas and I need to be prepared
 
nice, have fun
 
@Rusher Are you officially banning wolves that aren't a 'W'? If so, I'll stop working on a counter to ChameleonWolf. LazyWolf was actually worried about that one.
 
@Rusher Is my Wolf class working okay for you?
When I run it against the opposition it tends to be in the lead for the first 500 moves (while it's in avoid mode) but then it goes downhill relatively quickly from there.
 
@Geobits Try beating Camo now! :D
 
5:35 PM
@ProgrammerDan My understanding was that he was replacing the PRNG with one that always returns the same value? I'm not a Java coder though, so I may be wrong about that.
 
@Gareth That's what it looked like to me, too, but I didn't test it out much to see if it actually did.
 
@Geobits I tested it, and there were still 99 wolves alive.
And now there were 100 wolves alive.
 
Well, I threw it in my tests, but it doesn't actually win. With all the straight-Java wolves included(beside the Chameleon), my LazyWolf was still winning.
BTW, I modified my version of tester with a flag to show/skip the GUI and just report a sorted list of results to the console. It runs 1000 rounds in a second or two. If anyone's interested, I can post it. It makes testing a hell of a lot faster.
@Doorknob At least in my tests, I had to change exactly nothing to still beat Camo :p
 
@Geobits Gah! What were the statistics? :P
 
Not an average, but pretty representative results:
> LazyWolf : 58
Sheep : 44
GatheringWolf : 18
CamoWolf : 17
EmoWolf : 12
Wion : 6
GamblerWolf : 5
WolfWithoutFear : 3
Bear : 1
Lion : 1
Stone : 0
 
5:50 PM
Lol, some did worse than EmoWolf? O_o
Wait, how does that even work? Shouldn't EmoWolf have 0?
 
Is CamoWolf going to be included since it's bending the rules somewhat? I just want to know whether I should include it in my tests. :-\
 
Sheep is the only one I've seen beat me lately, about 5-10% of the time. I need to work on getting the non-Java entries in.
 
Also, is thinking of more strategy to survive
 
@Doorknob Emo only kills itself if it gets to attack. The survivors must have just not been stepped on.
 
@Geobits Ha. So perhaps not moving unless absolutely necessary is a good idea :D
 
5:51 PM
My LazyWolf agrees.
 
@Gareth No idea, we'll need @Rusher to clarify that. But Lazy, Sheep, and Gathering are all beating me anyway. :P
 
Well, Lazy and Sheep are pretty consistently in #1-2, but the others fluctuate a lot.
 
@Doorknob My wolf relies on there being a lot of opponents killing each other and the rest of the opposition during the first 500 turns. I do better in my tests when Camo and Chameleon are both in the game.
 
@Gareth No. And neither will the other one that modified his letter.
 
@Rusher Are we allowed more than one entry?
 
5:57 PM
@Rusher Why not? It's not against the rules. (The other one is, since it doesn't fit the required code format, but mine does.)
 
@Doorknob Read the spec again.
 
@Rusher Ok, I did. It says that all 'W's will be user submissions, but it never says all user submissions will be 'W's.
 
It says all wolves will be 'W'. Did you submit something besides a wolf?
 
Can I enter another wolf who only attacks wolves that move(shamelessly boosting Lazy in the process) ? Feels kinda cheap, but...
 
@Rusher No, I submitted a wolf with camouflage. :D
 
6:03 PM
Despite this being sandboxed for nearly three weeks, I didn't really have time to think of all the corner cases you guys could create, because I am not as smart as 8 of you. I am invoking the don't be a jerk rule and saying that 1) You can't change your letter and 2) If you submit a Wolf that intentionally boosts another Wolf, you'll just be excluded.
Also, @Gareth, I barely got the R submission working and haven't had time to learn how to include yours. I promise I'll work on it this week.
 
@Rusher Neither of those are in the rules, however. And what about wolves that specifically hunt down other wolves? For example, if someone created a LazyWolfHunter that attacks every wolf it sees.
 
@Rusher Cool. Just trying to clarify before I work on another one is all. I don't want to get it done and then find out it's banned.
 
@Rusher No problem. If you have any difficulties with it let me know.
 
@Doorknob If you have a suggestion, I'm all ears. If you only have problems, I don't really care.
 
@Rusher I'm replying to your message. None of those things were disallowed in the rules. And bending the rules is part of the fun of the site.
 
6:58 PM
@Doorknob Proof: Assume Wolves are represented by 'W'. Your Wolf is not represented by 'W'. Therefore, you are not a Wolf.
 
 
1 hour later…
8:12 PM
@Gareth @Geobits You guys are correct about the Gambler Wolf. I didn't look close enough at what was going on, but you both nailed it :)
 
8:33 PM
@PeterTaylor I'm going over that link you posted (stats.stackexchange.com/questions/15979/…) trying to improve that query (data.stackexchange.com/codegolf/query/180833/tag-love). My interpretation of the issue with Evan Millers' approach is that the lower bound is unrealistically low for items with small rating counts; and that's the primary issue -- correct?
 
 
1 hour later…
vzn
9:50 PM
hello again golfdudes
 
vzn
whats new
 
not a whole hell of a lot
 
@ProgrammerDan The real Java solution to this would be to use a different classloader for each trial...
 
vzn
so, anyway, prob spent 30m last time chatting with doorknob re his campaign & whether he could lead/motivate anyone to vote for it....
think there was about +1 after, what, some days? so "results" are in, ok!
its that magic of se collaboration/teamwork live in action! :p
 
10:03 PM
@JasonC My interpretation of the issue is that it's like using a hammer to drive screws or a screwdriver to drive hammers: just the wrong tool for the job.
2
 
vzn
> if you have a hammer everything begins to look like a nail —anonymous :)
se ofc the ultimate hammer :\
 
@PeterTaylor Thanks!
@vzn Or everything begins to look like ham.
Mmmmm...
 
vzn
mmm, bacon, delicious :)
←recently tried to buy unicoins with bacon, didnt work out :(
ps the se dataexplorer/datamining challenges are way cool, keep up good work :)
theres so much unexplored potential there for analyzing/improving sites quality...
lol the answer by the Gaz Man
 
10:26 PM
Did I type that? I mean to say using a screwdriver to drive nails. I'm obviously too tired to be left in front of a keyboard.
 

« first day (1163 days earlier)      last day (3683 days later) »