2012-04-11T00:24:38 *** epicmonkey has quit IRC (Ping timeout: 260 seconds) 2012-04-11T00:25:08 *** AndrewBC has quit IRC (Quit: leaving) 2012-04-11T01:05:00 *** Fandekasp has quit IRC (Ping timeout: 264 seconds) 2012-04-11T01:06:22 *** Fandekasp has joined #aichallenge 2012-04-11T01:11:07 *** SimonT has joined #aichallenge 2012-04-11T01:28:48 I don't suppose anyone has any idea when the next challenge will be and what it'll be? 2012-04-11T01:29:37 I have an idea 2012-04-11T01:30:42 You can get an RSS feed from the announcements forum on aichallenge.org I think 2012-04-11T01:31:24 other than that ... maybe ... I'd say ... end of the year and ... something around a 2d space shooter with projectiles/missiles 2012-04-11T01:31:35 oh and asteroids flying around 2012-04-11T01:32:51 Look at xpilot, that was inspiration to one of the game ideas for the next contest. While the next game is not set in stone, it is likely to be some clone of that 2012-04-11T01:33:23 Currently I think, the organizers try to make AI Challenge a proper non-profit organization in Canada 2012-04-11T01:35:24 *** pairofdice has joined #aichallenge 2012-04-11T01:36:01 *** kilae has joined #aichallenge 2012-04-11T01:39:56 *** Areks has joined #aichallenge 2012-04-11T01:41:32 mleise: thanks for the info 2012-04-11T01:52:08 *** Garf has joined #aichallenge 2012-04-11T02:12:26 *** SimonT has quit IRC (Quit: rebooting) 2012-04-11T02:16:55 *** archdori has joined #aichallenge 2012-04-11T02:20:45 *** Fandekasp has quit IRC (Ping timeout: 276 seconds) 2012-04-11T02:46:34 *** ivan`` has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2012-04-11T02:48:53 *** ivan`` has joined #aichallenge 2012-04-11T02:58:27 *** amstan has quit IRC (Ping timeout: 276 seconds) 2012-04-11T03:50:14 *** smiley1983 has quit IRC (Ping timeout: 245 seconds) 2012-04-11T03:51:00 *** smiley1983 has joined #aichallenge 2012-04-11T03:59:42 *** mceier has joined #aichallenge 2012-04-11T04:17:14 *** ztfw has joined #aichallenge 2012-04-11T04:18:24 *** ztfw has left #aichallenge 2012-04-11T05:00:52 *** thestinger has quit IRC (Quit: sleep) 2012-04-11T05:15:07 *** Fandekasp has joined #aichallenge 2012-04-11T05:18:50 *** archdori has quit IRC (Ping timeout: 248 seconds) 2012-04-11T05:21:37 *** mviel has quit IRC (Ping timeout: 246 seconds) 2012-04-11T05:23:18 *** replore_ has joined #aichallenge 2012-04-11T05:31:34 *** mcstar has joined #aichallenge 2012-04-11T05:34:24 *** replore_ has quit IRC (Remote host closed the connection) 2012-04-11T05:38:19 *** foRei has joined #aichallenge 2012-04-11T05:41:25 *** foRei has quit IRC (Read error: Connection reset by peer) 2012-04-11T05:41:44 *** mviel has joined #aichallenge 2012-04-11T05:47:36 *** foRei has joined #aichallenge 2012-04-11T05:51:48 *** smiley1983 has quit IRC (Ping timeout: 264 seconds) 2012-04-11T05:52:18 *** smiley1983 has joined #aichallenge 2012-04-11T05:56:23 *** Areks has quit IRC (Ping timeout: 272 seconds) 2012-04-11T06:10:53 *** smiley1983 has quit IRC (Ping timeout: 252 seconds) 2012-04-11T06:11:27 *** Areks has joined #aichallenge 2012-04-11T06:11:44 *** smiley1983 has joined #aichallenge 2012-04-11T06:17:33 *** sigh has joined #aichallenge 2012-04-11T06:18:56 *** Fandekasp has quit IRC (Ping timeout: 244 seconds) 2012-04-11T06:20:37 *** Fandekasp has joined #aichallenge 2012-04-11T06:24:30 *** smiley1983 has quit IRC (Ping timeout: 276 seconds) 2012-04-11T06:25:06 *** smiley1983 has joined #aichallenge 2012-04-11T06:45:50 *** smiley1983 has quit IRC (Ping timeout: 252 seconds) 2012-04-11T06:46:37 *** smiley1983 has joined #aichallenge 2012-04-11T06:53:03 *** smiley1983 has quit IRC (Ping timeout: 245 seconds) 2012-04-11T06:53:52 *** smiley1983 has joined #aichallenge 2012-04-11T07:21:31 *** Fandekasp has quit IRC (Ping timeout: 250 seconds) 2012-04-11T07:29:09 *** smiley1983 has quit IRC (Read error: Operation timed out) 2012-04-11T07:30:40 *** smiley1983 has joined #aichallenge 2012-04-11T07:46:25 *** smiley1983 has quit IRC (Ping timeout: 265 seconds) 2012-04-11T07:47:04 *** smiley1983 has joined #aichallenge 2012-04-11T08:05:13 *** mikewintermute has joined #aichallenge 2012-04-11T08:44:25 *** antimatroid has quit IRC (Read error: No route to host) 2012-04-11T08:44:30 *** antimatroid has joined #aichallenge 2012-04-11T08:50:47 *** mviel_ has joined #aichallenge 2012-04-11T08:54:39 *** mviel has quit IRC (Ping timeout: 276 seconds) 2012-04-11T09:24:52 *** kurnevsky has joined #aichallenge 2012-04-11T09:43:38 *** thestinger has joined #aichallenge 2012-04-11T09:46:29 *** pairofdice has quit IRC (Quit: It is time for you to leave) 2012-04-11T10:11:15 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2012-04-11T10:11:36 *** smiley1983 has joined #aichallenge 2012-04-11T10:14:41 thestinger: hai 2012-04-11T10:14:47 hi 2012-04-11T10:14:58 still doing crypto? 2012-04-11T10:14:58 mcstar: loving the algo+crypto classes btw :) 2012-04-11T10:14:59 yeah 2012-04-11T10:15:07 i dropped algo 2012-04-11T10:15:18 i was late last week 2012-04-11T10:15:32 and couldnt afford time to it this time 2012-04-11T10:15:51 yeah, lots of videos to watch 2012-04-11T10:16:14 I really liked the strongly connected component thing 2012-04-11T10:16:19 week3 prog. assignment was fine 2012-04-11T10:16:41 strongly connected what? 2012-04-11T10:16:44 algo? 2012-04-11T10:17:02 yeah 2012-04-11T10:17:10 i dont think i saw that one 2012-04-11T10:17:17 parts of a directed graph where you can get to/from everywhere 2012-04-11T10:17:26 a -> b -> a 2012-04-11T10:17:38 i did the crypto optional prog. assignemnt in c++,haskell,f#,python 2012-04-11T10:17:53 the brute force hash one? 2012-04-11T10:17:57 yeah 2012-04-11T10:18:17 I haven't done problem set 4 yet for crypto 2012-04-11T10:18:21 I have for the algo class 2012-04-11T10:18:23 that was the SCC one 2012-04-11T10:18:24 neither did i 2012-04-11T10:18:54 f# was a bit faster than c++ 2012-04-11T10:19:01 python was fast but ate a lot of memory 2012-04-11T10:19:08 haskell had very slow sha256 2012-04-11T10:19:11 which lib did you use? 2012-04-11T10:19:13 crypto++? 2012-04-11T10:19:16 yeah 2012-04-11T10:19:27 I think crypto++ is fairly slow at most things 2012-04-11T10:19:35 it also spewed out a ton of warnings when I compiled :) 2012-04-11T10:20:09 really? 2012-04-11T10:20:26 wait, did you compile the lib for yourself? 2012-04-11T10:20:35 no, when I compiled my little program 2012-04-11T10:20:46 using gcc 4.7 (I think they just haven't caught up) 2012-04-11T10:20:48 i havent noticed any warning 2012-04-11T10:20:51 ah 2012-04-11T10:21:01 i have 4.7 too 2012-04-11T10:21:45 no warnings here 2012-04-11T10:21:52 did you use their hex dump thing? 2012-04-11T10:22:01 CryptoPP::HexEncoder 2012-04-11T10:22:09 CryptoPP::SHA256 sha; 2012-04-11T10:22:15 sha.CalculateDigest(output, input, 32); 2012-04-11T10:22:17 thats it 2012-04-11T10:22:52 http://sprunge.us/ZTZh 2012-04-11T10:23:14 ah, I used their code for printing it 2012-04-11T10:23:23 that's where the warnings come from now that I actually read them :P 2012-04-11T10:24:03 i found the crypto++ documentation awful, so i didnt read it 2012-04-11T10:24:22 yeah, if I did it again I would just use openssl 2012-04-11T10:26:33 http://paste.pocoo.org/show/xfXqwi5fXncxMKSfugnm/ that was my code for the crypto one 2012-04-11T10:26:46 it's just an application for DFS (the topics were BFS and DFS basically) 2012-04-11T10:27:23 you go once through the graph and order the nodes (1 for the first with no unexplored nodes) 2012-04-11T10:27:25 ?that was my code for the crypto one 2012-04-11T10:27:37 mcstar: I know :P 2012-04-11T10:27:47 I really liked the algo one though 2012-04-11T10:27:52 I wasted some time making it fast 2012-04-11T10:28:10 thestinger: you said the pasted code is for the crypto one 2012-04-11T10:28:16 but i saw some graph stuff 2012-04-11T10:28:20 i mean i see 2012-04-11T10:28:24 oh, my brain messed up 2012-04-11T10:29:09 it uses up a huge amount of memory 2012-04-11T10:29:20 at first reading in the input data took 8s 2012-04-11T10:29:37 haha 2012-04-11T10:29:40 using Boost.Spirit gets that down to 0.6s 2012-04-11T10:30:18 well, i dont want to waste time on algo, since i decided not to do it 2012-04-11T10:30:31 i hope you understand 2012-04-11T10:30:35 yeah :) 2012-04-11T10:30:57 anyway, this problem was my first monadic haskell stuff 2012-04-11T10:31:14 oh, and System.Random is slow too 2012-04-11T10:31:16 beware 2012-04-11T10:31:26 heh 2012-04-11T10:31:30 i changed it to mersenne-random package 2012-04-11T10:32:07 thestinger: do you know about monadic lifeters? 2012-04-11T10:32:12 lifters 2012-04-11T10:32:24 sort of, I've read articles but haven't really coded much haskell 2012-04-11T10:32:47 i didnt have to use them so far 2012-04-11T10:32:56 but seems to me its pita 2012-04-11T10:33:00 that's the monad transformer stuff iirc 2012-04-11T10:33:03 yeah 2012-04-11T10:33:06 I looked over it but never used it 2012-04-11T10:33:52 im in a bit of a struggle, not sure what should i pursue first, haskell or f# 2012-04-11T10:34:11 cause f# looks a bit messed up, but quite fast, and has the whole .net 2012-04-11T10:39:09 mcstar: you're alive! 2012-04-11T10:39:21 I ended up discovering a weird, scary corner of C++ 2012-04-11T10:39:27 incomplete types :) 2012-04-11T10:40:01 antimatroid: hi! 2012-04-11T10:40:03 http://sprunge.us/UNiU this tiny example compiles with gcc and clang and gives no warning (no matter how many switches you turn on) 2012-04-11T10:40:09 and it's undefined behavior... 2012-04-11T10:40:34 if you switch map -> unordered_map it won't compile 2012-04-11T10:40:44 but it's undefined behavior with map too 2012-04-11T10:41:58 mcstar: how's life? 2012-04-11T10:42:17 scuks 2012-04-11T10:42:20 sucks 2012-04-11T10:42:49 thestinger: whats with template<>? 2012-04-11T10:44:56 mcstar: it's a silly example, pretend 'int' is std::vector 2012-04-11T10:45:13 it's undefined behavior to use incomplete types with the standard lib containers 2012-04-11T10:45:19 * thestinger got bitten by it 2012-04-11T10:45:29 I could _not_ figure out why that works with std::map and not std::unordered_map 2012-04-11T10:45:41 with more complex code 2012-04-11T10:45:46 SomeType > 2012-04-11T10:46:40 http://sprunge.us/NUWb that 2012-04-11T10:46:48 that's still undefined behavior * 2012-04-11T10:48:11 anything like a tree or linked list that refers to itself is an incomplete type during the definition, so you can't use any standard lib containers in it 2012-04-11T10:48:24 but you can use the boost implementations of map, etc. 2012-04-11T10:48:31 just not the standard lib :\ 2012-04-11T10:49:00 so how are linked lists or trees done? 2012-04-11T10:49:06 in stl 2012-04-11T10:49:37 mcstar: you can do it with templated types, just not the standard library ones 2012-04-11T10:50:22 if you replaced map with boost::map it wouldn't be undefined behavior 2012-04-11T10:51:48 http://stackoverflow.com/questions/9860503/templated-recursive-data-types I actually had to ask a SO question because I couldn't figure it out 2012-04-11T10:53:34 maybe you can do it with a mutually recursive solution 2012-04-11T10:53:58 mcstar: no, because it still needs itself defined to be complete 2012-04-11T10:54:14 the answers suggested using a smart pointer, which works (but adds overhead) 2012-04-11T10:54:36 the real solution was just using the boost container library, which doesn't have that rule 2012-04-11T10:55:24 I think implementing containers that work in that case is hard, so they said it was undefined behavior to make them easier to implement 2012-04-11T10:57:08 did you look at the boost implementation? 2012-04-11T10:58:16 *** kilae has quit IRC (Quit: ChatZilla 0.9.88.2 [Firefox 11.0/20120312181643]) 2012-04-11T10:58:37 anyway, all that template crap and OOP design patterns and c++ corner cases makes me puke 2012-04-11T10:58:49 yeah, this really put me off the language 2012-04-11T10:59:05 the fact that neither gcc or clang gives a warning for something that's apparently clearly undefined 2012-04-11T11:00:42 how is this done in python? 2012-04-11T11:00:56 i mean how do you write generic code? 2012-04-11T11:01:05 it's not a problem since it's dynamically typed 2012-04-11T11:01:16 ofc there's a performance and safety cost for that 2012-04-11T11:01:21 still, you need to knwo the types sometimes 2012-04-11T11:01:49 it's usually regarded as bad design in python to use type() 2012-04-11T11:02:23 lets say you want to make a hashtable container in python 2012-04-11T11:02:38 and obviously you need different hashing wrt different types 2012-04-11T11:02:48 how do you solve that? 2012-04-11T11:03:07 you define the __hash__ method 2012-04-11T11:03:26 and duck typing takes care of it 2012-04-11T11:03:42 *** dici has joined #aichallenge 2012-04-11T11:03:44 can you define that method on existing types? 2012-04-11T11:03:48 or classes 2012-04-11T11:03:55 you can inherit from them and define it 2012-04-11T11:04:15 i dont get it 2012-04-11T11:04:24 lets say you want to use integers as a key 2012-04-11T11:04:34 so, the dict calls __hash__ and finds the bucket 2012-04-11T11:04:37 how do you make sure int will have __hash__ 2012-04-11T11:04:47 mcstar: it just calls it 2012-04-11T11:04:52 if it doesn't have it, an exception is thrown 2012-04-11T11:05:00 i know 2012-04-11T11:05:08 but how do you make an int have __hash__ 2012-04-11T11:05:24 you mean for a user-defined type? 2012-04-11T11:05:29 no, for int 2012-04-11T11:05:35 but int has it :P 2012-04-11T11:05:36 but you dont use a built in dict 2012-04-11T11:05:41 you do it yourself 2012-04-11T11:06:05 you have to inherit from the built-ins to add stuff to them 2012-04-11T11:06:24 if pypy was the reference implementation that might not be the way things are 2012-04-11T11:06:53 so if i want a mydict, that calls __myhash__ on its objects, basically couldnt use built-in types with mydict 2012-04-11T11:07:42 or, i should use if(typeof(obj)=='int') return obj.__hash__() 2012-04-11T11:09:50 *** X-Scale has joined #aichallenge 2012-04-11T11:11:01 mcstar: python isn't as dynamic as it could be due to the way CPython was coded 2012-04-11T11:11:23 it sucks that the solution for performance problems is "write stuff in C" 2012-04-11T11:11:27 and not "make the language faster" 2012-04-11T11:11:31 *** sigh has quit IRC (Remote host closed the connection) 2012-04-11T11:12:13 yeah 2012-04-11T11:12:20 well, i dont intend to use python for big things 2012-04-11T11:12:31 but i definitely gonna use it where i would have used bash 2012-04-11T11:12:47 (for >2 lines of scripts) 2012-04-11T11:13:04 bash sucks 2012-04-11T11:13:16 i mean last time i run into it i was pulling out my hair 2012-04-11T11:13:26 yeah, bash is pretty awful :P 2012-04-11T11:21:32 i should write something for windows and sell it 2012-04-11T11:22:22 or an iphone app :P 2012-04-11T11:22:33 yeah 2012-04-11T11:22:42 im not sure what though 2012-04-11T11:23:32 maybe a good tiling wm helper for developers 2012-04-11T11:23:49 i dont think windows progrmmers know the benefit 2012-04-11T11:24:18 so theres a possible market for that, if the softw. would be real high quality 2012-04-11T11:25:39 thestinger: did you know theres a shell mode for ipython? 2012-04-11T11:25:45 yeah 2012-04-11T11:26:06 I never really used it though 2012-04-11T11:26:42 do you have some secret money making idea? 2012-04-11T11:26:48 nope :( 2012-04-11T11:27:13 yeah, dont let anyone know it 2012-04-11T11:28:18 the issue with iphone apps is that you need OS X 2012-04-11T11:28:37 and ofc an ipad/iphone... 2012-04-11T11:29:09 the issue is that i have absolutely no idea for a mobile app 2012-04-11T11:29:15 other than some farting apps 2012-04-11T11:29:21 but that has already been done 2012-04-11T11:29:41 or 'instant nude pictures' 2012-04-11T11:29:51 they censor nudity :P 2012-04-11T11:30:24 hm, so the idea is an alternate AppStore? 2012-04-11T11:30:50 mcstar: it exists (cydia) but you have to jailbreak 2012-04-11T11:32:33 I should make a tablet game 2012-04-11T11:34:46 lets just write another opensource theorem prover in haskell in the meantime 2012-04-11T11:35:04 lol 2012-04-11T11:35:10 I'm actually thinking about a game now 2012-04-11T11:35:18 there are no good strategy games for the ipad 2012-04-11T11:35:24 just arcade/casual stuff like risk 2012-04-11T11:35:46 and it's a perfect platform for an overhead view + clicky interface 2012-04-11T11:35:55 thestinger: like, sitting down with buddies an playing a collborative strategy game? 2012-04-11T11:35:58 over wifi? 2012-04-11T11:36:07 and* 2012-04-11T11:36:22 that would be kinda cool 2012-04-11T11:36:26 or against the AI 2012-04-11T11:36:30 ah 2012-04-11T11:36:33 that would suck 2012-04-11T11:36:42 lol 2012-04-11T11:36:48 you can do that with anything 2012-04-11T11:37:13 you have to incorporate the uniqe nature of the ipad 2012-04-11T11:37:19 but the only fun part would be implementing the AI 2012-04-11T11:38:07 lets write a multiplayer hex game 2012-04-11T11:39:49 hmm 2012-04-11T11:40:19 or, lets port red alert to ipad 2012-04-11T11:40:55 a google search reveals a lot of strategy games for the ipad though 2012-04-11T11:41:03 so i will just forget about this 2012-04-11T11:41:56 i have a better idea: solve the automatic parallelization problem 2012-04-11T11:42:03 and build a proof of concept compiler 2012-04-11T11:44:05 *** Areks has quit IRC (Ping timeout: 272 seconds) 2012-04-11T11:50:02 thestinger: whats with asteroids? 2012-04-11T11:50:14 is it going to be the next game? 2012-04-11T11:50:37 no clue :P 2012-04-11T11:52:41 lets change asteroids to balls, and the ais would have to kick balls, and if the ball hits another ai, it takes some HP off of that ai, and this would be inversely proportional to the travel time of the ball 2012-04-11T11:53:27 *** kilae has joined #aichallenge 2012-04-11T11:53:52 but the real-time aspect of all this is probably would be too hard to do correctly 2012-04-11T12:05:16 we could start coding stuff for asteroids right now and be way in front of others 2012-04-11T12:06:05 think of solving collision equations for different numbers and types of objects 2012-04-11T12:07:16 i dont have time for that 2012-04-11T12:07:22 neither do you 2012-04-11T12:07:46 lol 2012-04-11T12:08:03 I'm too lazy to write code for a game that might not be chosen 2012-04-11T12:08:10 won't have anyone to play with :P 2012-04-11T12:08:18 it is like solving a project euler problem 2012-04-11T12:08:52 mleise: we dont have time for PE either 2012-04-11T12:08:57 XD 2012-04-11T12:09:32 may i remind you that you don't have time to chat here? 2012-04-11T12:09:44 so right 2012-04-11T12:10:05 mcstar: just don't have kids :P 2012-04-11T12:10:18 what do you mean? 2012-04-11T12:10:28 then you'll _never_ have time again 2012-04-11T12:10:31 they steal your prrecious time 2012-04-11T12:10:32 its not like im an hermaphrodite 2012-04-11T12:19:55 for future study: http://www.haskell.org/haskellwiki/Typeclassopedia 2012-04-11T12:23:03 *** iglo has joined #aichallenge 2012-04-11T12:24:08 *** amstan has joined #aichallenge 2012-04-11T12:24:08 *** ChanServ sets mode: +o amstan 2012-04-11T12:25:18 *** kurnevsky1 has joined #aichallenge 2012-04-11T12:28:17 *** kurnevsky has quit IRC (Ping timeout: 246 seconds) 2012-04-11T12:29:59 *** kurnevsky has joined #aichallenge 2012-04-11T12:32:32 *** thestinger is now known as thestinger_ 2012-04-11T12:32:58 *** Palmik has joined #aichallenge 2012-04-11T12:45:25 *** mikewintermute has quit IRC (Quit: mikewintermute) 2012-04-11T12:58:20 *** thestinger_ is now known as thestinger 2012-04-11T12:58:30 mcstar: How much does Microsoft Research Labs pay you for advertising Haskell every day :p 2012-04-11T13:01:05 wow there are a lot of things i have never heard of in this typeclassopedia. monoidal classes, ArrowPlus, do rec notation. It must take years to understand all that 2012-04-11T13:01:21 mleise: or a class in category theory 2012-04-11T13:02:33 mleise: I don't think it's really that complicated, just alien to us :P 2012-04-11T13:03:20 yeah, like Klingons coming to earth and presenting us their programming language 2012-04-11T13:05:01 I learnt QBasic on my own, VB, Pascal, even C++. But these FP languages are just so different, that I don't feel any urge to learn them. 2012-04-11T13:05:37 mleise: wrong, i would be advertising F# and F* then 2012-04-11T13:05:58 hey, guys have you tried F# and F* ? they are both great languages, and FAST 2012-04-11T13:06:01 and interoperable 2012-04-11T13:06:04 try it today 2012-04-11T13:06:07 for free* 2012-04-11T13:06:10 the mono implementation? 2012-04-11T13:06:19 shit, no. M$ is involved in Haskell, I tell ya! 2012-04-11T13:06:39 no really, well PSJ works for them thats true 2012-04-11T13:07:02 thestinger: the mono implementation what? 2012-04-11T13:07:05 just search for "haskell microsoft-research" on google 2012-04-11T13:07:59 mcstar: of F# 2012-04-11T13:08:05 I read about that on the D forums, where someone posted an article about the "Downfall of imperative languages" 2012-04-11T13:08:21 thestinger: it is still an incomplete question 2012-04-11T13:08:26 just like your SomeType 2012-04-11T13:08:30 lol 2012-04-11T13:08:32 what are you asking? 2012-04-11T13:08:36 is that what you're using? 2012-04-11T13:08:41 yes 2012-04-11T13:09:38 i dont think they will ever fall down 2012-04-11T13:09:48 the computer is imperative 2012-04-11T13:10:16 yay 2012-04-11T13:10:18 haskell 2012-04-11T13:10:34 well, the argument was that multi-core systems can easily exploited with pure functiona languages 2012-04-11T13:11:00 it is never easy 2012-04-11T13:11:04 where imperative languages run into race-conditions that will kill patients in hospitals 2012-04-11T13:11:05 afaik it is hot research 2012-04-11T13:11:06 mleise: but GPU programming is imperative, with manual memory management 2012-04-11T13:11:16 some languages support more automatic parallelization than others 2012-04-11T13:12:15 GPU programming ... well you just write a filter kernel and have no write access to anything other than your current pixel 2012-04-11T13:12:18 mleise: one solution to that is STM, like in clojure 2012-04-11T13:12:23 but it has a performance hit 2012-04-11T13:12:30 STM means? 2012-04-11T13:12:40 software transactional memory 2012-04-11T13:12:53 and you never have to lock anything again 2012-04-11T13:12:56 mcstar: haskell has it too 2012-04-11T13:13:03 yeah, but not part of the language 2012-04-11T13:13:20 (and im not sure about the details,like doest have agents too?) 2012-04-11T13:13:47 another one is message passing, which is the advertised way to do it in D 2012-04-11T13:13:49 *** dici has quit IRC (Read error: Connection reset by peer) 2012-04-11T13:13:54 and in erlang 2012-04-11T13:14:02 but erlang is crazy slow 2012-04-11T13:14:08 you have to write c code for it 2012-04-11T13:14:19 lol, like in Python now and then 2012-04-11T13:14:27 erlang is slower than python most of the time 2012-04-11T13:14:33 yeah, could be 2012-04-11T13:14:39 but python has GIL 2012-04-11T13:14:55 mcstar: stackless python has lightweight threads and message passing though 2012-04-11T13:14:56 It's sad, when a language is uber cool, but you have to resort to hacks anyway to get it fast enough or consume less memory 2012-04-11T13:15:05 mcstar: and pypy has an implementation of that now 2012-04-11T13:15:08 gil doesnt appect green threads 2012-04-11T13:15:13 affect* 2012-04-11T13:15:33 there is something like that in standard python, no? 2012-04-11T13:15:43 pypy uses a lot more RAM than python if I read that correctly 2012-04-11T13:15:49 mleise: due to the JIT 2012-04-11T13:15:53 but objects consume less memory 2012-04-11T13:15:59 so it scales better 2012-04-11T13:16:05 it's just the initial overhead that's a lot higher 2012-04-11T13:16:09 ah, ok 2012-04-11T13:16:10 the startup time + initial memory usage 2012-04-11T13:16:12 mleise: by language you mean D? 2012-04-11T13:16:33 *** foRei has quit IRC (Ping timeout: 272 seconds) 2012-04-11T13:16:35 mleise: similar to how browsers use a ton of memory now :) 2012-04-11T13:16:49 mcstar: Python, Erlang, D. Every language with nice productivity enhancments comes with a caveat somewhere 2012-04-11T13:17:24 i dont think any of that deserves 'uber cool' 2012-04-11T13:17:38 yeah, I wouldn't call python cool :P 2012-04-11T13:17:48 it's clean/consistent and pragmatic 2012-04-11T13:17:56 at least compared to most other languages 2012-04-11T13:18:06 it is easy too 2012-04-11T13:18:15 Sure, and that's uber cool alltogether 2012-04-11T13:18:22 Java is uber cool too. 2012-04-11T13:18:24 mcstar: yeah, but I think that's because of how consistent it is 2012-04-11T13:18:31 yeah ok, sry for asking mleise 2012-04-11T13:18:35 since the same syntax is used everywhere 2012-04-11T13:18:40 thestinger: lisp 2012-04-11T13:18:58 you brought up syntax... 2012-04-11T13:20:23 mcstar: python reads like pseudo-code/english though 2012-04-11T13:20:31 sorta, true dat 2012-04-11T13:20:40 i just said that the other day on #python 2012-04-11T13:20:48 haskell uses too many weird characters and operators to be that readable 2012-04-11T13:20:48 we had a great discussion 2012-04-11T13:20:51