2011-07-08T00:14:59 *** computerwiz_222 has quit IRC (Remote host closed the connection) 2011-07-08T00:39:34 *** ArchMonkey has quit IRC (Quit: ArchMonkey) 2011-07-08T01:07:51 *** pcavs has quit IRC (Ping timeout: 255 seconds) 2011-07-08T01:08:05 *** filter has joined #aichallenge 2011-07-08T01:09:09 *** filter_ has quit IRC (Ping timeout: 252 seconds) 2011-07-08T01:36:36 <_flag> My newest bot has been compiling for several hours now, and the server is still playing games with the old one 2011-07-08T01:36:47 <_flag> Anyone want to look into this? 2011-07-08T01:55:33 *** amstan has quit IRC (Ping timeout: 255 seconds) 2011-07-08T02:02:44 *** smiley1983 has joined #aichallenge 2011-07-08T02:03:35 *** Chris_0076 has quit IRC (Remote host closed the connection) 2011-07-08T02:05:32 *** Chris_0076 has joined #aichallenge 2011-07-08T02:19:42 *** VilleH has joined #aichallenge 2011-07-08T03:19:10 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-07-08T03:19:42 *** gobang has quit IRC (Ping timeout: 255 seconds) 2011-07-08T03:24:45 *** ltriant has quit IRC (Quit: Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/) 2011-07-08T03:27:31 *** gobang has joined #aichallenge 2011-07-08T03:28:53 *** gobang has quit IRC (Read error: Connection reset by peer) 2011-07-08T03:38:09 *** Palmik has joined #aichallenge 2011-07-08T04:14:22 *** mleise has joined #aichallenge 2011-07-08T04:16:53 *** Naktibalda has joined #aichallenge 2011-07-08T04:25:15 *** mcstar has joined #aichallenge 2011-07-08T04:43:19 *** sigh has joined #aichallenge 2011-07-08T05:01:12 *** zaphus has joined #aichallenge 2011-07-08T05:26:24 *** _0x47 <_0x47!~slash@p5B12F708.dip.t-dialin.net> has joined #aichallenge 2011-07-08T05:49:35 @rankings 2011-07-08T05:49:46 mcstar: Error: There was a problem accessing the interface to ai-contest.com 2011-07-08T05:52:18 *** onensora has joined #aichallenge 2011-07-08T06:02:58 *** FireFly has joined #aichallenge 2011-07-08T06:25:53 *** mcstar has quit IRC (Ping timeout: 246 seconds) 2011-07-08T06:35:31 *** ibdknox has quit IRC (Remote host closed the connection) 2011-07-08T07:02:42 *** sigh has quit IRC (Remote host closed the connection) 2011-07-08T07:09:56 *** VilleH has quit IRC (Ping timeout: 252 seconds) 2011-07-08T07:15:44 *** GeorgeSebastian has joined #aichallenge 2011-07-08T07:15:44 *** GeorgeSebastian has joined #aichallenge 2011-07-08T07:39:52 *** olexs has joined #aichallenge 2011-07-08T07:41:45 *** olexs has left #aichallenge 2011-07-08T07:48:59 http://www.reddit.com/r/askscience/comments/ijc5y/what_is_onedimensional/c247w4i 2011-07-08T07:49:07 not sure if that'll help in the end aha 2011-07-08T07:49:12 it's hard to explain what dimensions are 2011-07-08T07:52:19 *** Palmik has quit IRC (Remote host closed the connection) 2011-07-08T07:59:11 *** Palmik has joined #aichallenge 2011-07-08T08:01:52 *** foRei has joined #aichallenge 2011-07-08T08:02:20 *** mcstar has joined #aichallenge 2011-07-08T08:03:57 *** sir_macelon has joined #aichallenge 2011-07-08T08:04:05 hi 2011-07-08T08:04:20 hey 2011-07-08T08:04:27 seems again workers are down and beta is not playing any games for 5 hours now... 2011-07-08T08:05:38 :( 2011-07-08T08:06:51 and I don't see any progress over any of the 5 issues for the launch milestone... 2011-07-08T08:06:51 im gonna kill myself... 2011-07-08T08:07:34 i have more pressing problems 2011-07-08T08:07:53 my linux just hang for the second time in 2 weeks 2011-07-08T08:09:19 wait a minute, hang hung hung? 2011-07-08T08:09:26 yes 2011-07-08T08:09:34 just hung, right antimatroid M 2011-07-08T08:09:36 ? 2011-07-08T08:09:42 what did it do? 2011-07-08T08:09:46 just freeze up? 2011-07-08T08:09:50 yep 2011-07-08T08:09:54 mine does that if i am using too much memory 2011-07-08T08:10:06 i have to be careful to close my browser windows every few days, it sucks 2011-07-08T08:10:14 that and paint are my only complaints about linux 2011-07-08T08:10:27 no, if that was the case, it would start swapping, which is noisy, and its not unresponsive when it does that 2011-07-08T08:10:37 what browser do you use 2011-07-08T08:10:43 it just freeze 2011-07-08T08:10:48 when i was compiling 2011-07-08T08:11:18 rabidus: it cannot be the browsers fault, and i usually have 3 gb free ram 2011-07-08T08:11:31 mcstar: does it run xscreensaver? and does it pass memtest? 2011-07-08T08:11:48 xscreensaver is the most reliable way to crash linux I've found 2011-07-08T08:12:02 the other one being the use of dodgy hardware 2011-07-08T08:12:39 smiley1983: my first guess i faulty memory, ill check with memtest 2011-07-08T08:12:48 but it can be a kernel/cpu problem too 2011-07-08T08:12:56 xscreensaver never crashed my box 2011-07-08T08:12:59 or then: something else 2011-07-08T08:13:46 hell, im not sure i have memtest as an option 2011-07-08T08:13:55 i use grub2 with gpt 2011-07-08T08:14:04 and i didnt configure it to have memtest 2011-07-08T08:14:10 ubuntu install discs have the option built in 2011-07-08T08:14:17 if you have one of those lying around 2011-07-08T08:14:24 i dont use optical drives :) 2011-07-08T08:14:32 they can be transferred to USB 2011-07-08T08:14:44 but i might have a pendrive lying around 2011-07-08T08:15:05 though because Ubuntu is silly, it needs special software to do the transfer (unlike the superior Arch isos) 2011-07-08T08:15:15 oh, my 4gb pendrive actually has win7 installer on it 2011-07-08T08:15:35 smiley1983: youre an archer too? 2011-07-08T08:15:42 yes :) 2011-07-08T08:15:56 I was a Debian user until I discovered Arch 2011-07-08T08:16:03 ok, there are 3 confirmed arch cases on this channel 2011-07-08T08:16:12 its spreading 2011-07-08T08:16:33 who is the other one? 2011-07-08T08:16:38 debian/suse/kubuntu/arch 2011-07-08T08:16:48 jm-carthur 2011-07-08T08:16:57 (i dont want to wake him) 2011-07-08T08:17:21 right. It makes sense that people will migrate to Arch, though, because it's so great :) 2011-07-08T08:18:16 there's no TCP server or working server code at the moment, is there? 2011-07-08T08:18:28 my highlights are: good packaging(modularity), easy build system(abs,aur) and bsd style init scripts 2011-07-08T08:18:48 oh, and rolling release style 2011-07-08T08:19:02 there was one, from berak 2011-07-08T08:19:10 but i didnt see him in quite a while 2011-07-08T08:19:34 berak pulled mcleo's tcp server 2011-07-08T08:19:59 but there was talk to add couple of functions to the engine's sandbox, so that it could function as a tcp server 2011-07-08T08:20:31 smiley1983: and arch has the nicest wiki 2011-07-08T08:20:42 When the test server was running, though, it wasn't telling me about water, so I think it's still incomplete 2011-07-08T08:23:10 yes, all the things you've mentioned about Arch are great. I was surprised at first by how stable everything is considering the rolling release model. 2011-07-08T08:23:41 memtest added to grub 2011-07-08T08:23:45 im gonna reboot 2011-07-08T08:23:49 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-07-08T08:51:54 *** sir_macelon has quit IRC (Quit: Page closed) 2011-07-08T08:53:16 smiley1983: that water thingie was a bit strange 2011-07-08T08:53:30 beacause it didn't go that way on some bots 2011-07-08T08:55:44 rabidus: do you know if there's a TCP server now? 2011-07-08T08:55:52 no, i don't know 2011-07-08T08:56:47 oh well, I'm guessing it will probably become a higher priority once the contest starts 2011-07-08T08:57:11 me too 2011-07-08T08:57:35 i'm not sure anyone's even working on a tcp server 2011-07-08T08:57:41 i suspect mcleopold is busy with other stuff? 2011-07-08T08:57:54 he is slacking 2011-07-08T08:57:55 :D 2011-07-08T08:58:09 he's done more than you :P 2011-07-08T08:58:18 (And i) 2011-07-08T08:58:23 yeah, well i'm a huge slacker 2011-07-08T08:58:41 isn't that common thing with programmers? 2011-07-08T08:58:59 if you are not coding, then you are thinking (slacking really) 2011-07-08T08:59:00 with people* 2011-07-08T08:59:14 i wouldn't have a clue how to set up tcp stuff 2011-07-08T08:59:22 me neither 2011-07-08T08:59:33 but great idea overall 2011-07-08T08:59:46 awesome actually 2011-07-08T09:00:17 it would be fun if every game would leave a tcp server behind, and contest goes on over there after it 'offically' ends 2011-07-08T09:00:24 like there is now pw tcp server 2011-07-08T09:01:41 tcp server doesn't need that much resources to run 2011-07-08T09:01:50 just talks to clients and check that data is valid 2011-07-08T09:01:59 very straightforward 2011-07-08T09:06:44 *** pcavs has joined #aichallenge 2011-07-08T09:21:05 *** Palmik has quit IRC (Remote host closed the connection) 2011-07-08T09:23:25 rabidus: hopefully with zeta we could revive servers (possibly just with tcp, but at least with tcp) for all old contest games 2011-07-08T09:23:28 possibly with some new ones 2011-07-08T09:23:44 the vision is like an ai training ground i think, at least that's what I envision 2011-07-08T09:24:13 *** jmreardon has joined #aichallenge 2011-07-08T09:24:34 i think everyone is looking forward to looking at tron again 2011-07-08T09:24:44 antimatroid you know what 2011-07-08T09:24:50 bhasker: what? 2011-07-08T09:24:54 i felt planetwars was more of an interesting challenge 2011-07-08T09:25:06 than tron? 2011-07-08T09:25:10 than ants 2011-07-08T09:25:14 i feel* 2011-07-08T09:25:15 rather 2011-07-08T09:25:38 i've stated my opinion on this a few times, i think planet wars was a more interesting problem, but ants offers more in the way of ai 2011-07-08T09:25:46 in that there's more obvious things to work on and see clear results from 2011-07-08T09:25:57 *** mleise has quit IRC (Quit: Leaving.) 2011-07-08T09:26:21 although that can also be a bad thing, as people are all doing very similar things 2011-07-08T09:26:37 at the start of planet wars everyone was like shit, how the hell am i going to do something with this? 2011-07-08T09:28:11 with zeta? what do you mean? 2011-07-08T09:28:19 *** mleise has joined #aichallenge 2011-07-08T09:28:53 rabidus: the current "framework" is being developed under the name epsilon, there is plans for a future one completely rewritten called zeta, which would provide a framework for multiple contests etc. 2011-07-08T09:29:10 mmh.. very nice 2011-07-08T09:29:25 as to how long that will take if it ever happens, we will see, i hope it does happen 2011-07-08T09:29:32 we intent to do it at least :) 2011-07-08T09:29:38 then there can be like .. poker games or .. chess or something similar with zeta framework ? 2011-07-08T09:29:38 yea thats what i felt, all the bots look the same right now 2011-07-08T09:29:43 very, very nice idea if that's so 2011-07-08T09:29:43 there doesn't seem to be much of a difference 2011-07-08T09:29:57 and more importantly the games look boring to watch 2011-07-08T09:30:02 planetwars was a lot more fun 2011-07-08T09:30:05 to watch 2011-07-08T09:30:12 yeah, i don't like the flooded maps 2011-07-08T09:30:37 i would prefer less ants on the map, but so far people seem to think it's been necessary for games not to take forever? 2011-07-08T09:30:45 i wouldn't have thought it would make that much difference 2011-07-08T09:31:03 so it was more strategic and you could actually see the game/battle unfold 2011-07-08T09:31:24 but i don't know how much will vary strategy wise 2011-07-08T09:31:40 battle code can vary, but how noticable will it be? 2011-07-08T09:33:18 yea battle code is rather simplistic 2011-07-08T09:33:30 and there is no incentive to control territory 2011-07-08T09:33:42 all ants are spreading out 2011-07-08T09:34:32 i think we should allow stacking 2011-07-08T09:34:38 that would change the game a lot 2011-07-08T09:35:06 you could have super squads that can't be blocked. 2011-07-08T09:35:29 though visualizing it might be harder 2011-07-08T09:37:15 *** pcavs has quit IRC (Ping timeout: 255 seconds) 2011-07-08T09:42:40 stacked ants could also do wider range damage, so they can be larger on map 2011-07-08T09:43:23 that actually would change the entire game. 2011-07-08T09:43:32 i liked antimatroids version more 2011-07-08T09:44:28 the dmg distribution method? 2011-07-08T09:46:50 stacked ants and food into the nest 2011-07-08T09:47:29 we should have done my original ants game :P :( 2011-07-08T09:47:33 but alas, way too late 2011-07-08T09:48:30 heh 2011-07-08T09:48:41 oh well, 2011-07-08T09:48:44 *** GeorgeSebastian has quit IRC (Ping timeout: 240 seconds) 2011-07-08T09:48:48 mine was like now, but you had to use a move to "attack" and adjacent ant, you had hp for ants, you could fuse your ants together and their hp would sum, you had home bases that you had to harvest and return food to to generate new ants, and you had to defend your base because if an enemy invaded it you died (like future cop lapd preccint assault) 2011-07-08T09:49:03 and you had to use a move to harvest a single food item 2011-07-08T09:49:27 and why that idea was crushed? 2011-07-08T09:49:32 "too complicated" 2011-07-08T09:49:56 well it is much more complicated than the ants game now, but... i think we need a bit more complexity 2011-07-08T09:50:28 frontier made an awesome mock up visualiser for it too 2011-07-08T09:50:33 i liked a lot of that 'move harvested food into the nest if you want your son' 2011-07-08T09:50:37 the ants were actual walking (animated) ants 2011-07-08T09:50:37 yeah, i saw that 2011-07-08T09:51:42 with multiplayer, we could have given like 10+ points for taking out a queen/base and then the bot who lost the base couldn't generate more ants 2011-07-08T09:52:07 or the bot who lost the base loses all their ants and the bot who took the base out gets that many points 2011-07-08T09:52:16 yea the current game is just too simple 2011-07-08T09:52:30 not too simple, but it is simple 2011-07-08T09:54:06 i don't really see how battles are simple though? 2011-07-08T09:54:31 i still hope that the best bot will be one that is somewhat offensive 2011-07-08T09:58:37 *** Kingpin13 has joined #aichallenge 2011-07-08T09:58:42 lets see 2011-07-08T09:58:49 most bots seem to be following a strat of 2011-07-08T09:58:58 don't enter battle as long as its an equal tradeoff 2011-07-08T09:59:07 they just seem to go back and forth till they have a majority 2011-07-08T09:59:08 and then fight 2011-07-08T10:06:21 i think ant spawning should be tuned too, maybe if it was made so that the bigger your ant cluster the more chances of attracting an ant to spawn would be nicer 2011-07-08T10:11:31 any ideas on how to do an efficient getNearbyAnts query 2011-07-08T10:11:57 sql? 2011-07-08T10:12:09 no i meant like in the bot 2011-07-08T10:12:19 what data structure are you using atm? 2011-07-08T10:13:43 its crappy right now since ants are stored in a vector 2011-07-08T10:13:56 i am assumin gi can just search the grid 2011-07-08T10:14:01 in a square around the current position 2011-07-08T10:14:10 rather than do a real distance calculation 2011-07-08T10:14:14 between all ants 2011-07-08T10:14:28 i am wondering if there was something better than can be done 2011-07-08T10:14:47 like maybe precompute it in one pass across the map for all ants 2011-07-08T10:14:54 rather for all squares 2011-07-08T10:14:55 ya, right now I put all my ants in a 2d array of bools, keep a precalculated list of grid offsets, then add them to the point and query those array entries 2011-07-08T10:15:28 ah yea i was thinking of doing something similar 2011-07-08T10:15:38 i think those queries are killing my bots performance right now 2011-07-08T10:15:51 my entire ai is actually being done by precalculating various values on the entire map 2011-07-08T10:15:57 nice 2011-07-08T10:15:59 > 100 skill 2011-07-08T10:16:06 which bot is it? 2011-07-08T10:16:12 works fairly well, but it's hard to do things like make only one ant go after a food 2011-07-08T10:16:15 jmreardon 2011-07-08T10:16:21 sir_macelon 2011-07-08T10:16:39 oh 2011-07-08T10:17:07 btw why does the worker keep dying? 2011-07-08T10:17:23 no games in last 10hrs or so? 2011-07-08T10:17:47 not sure 2011-07-08T10:17:50 not my area 2011-07-08T10:19:30 need to go to drink some beer 2011-07-08T10:19:32 almost +30 2011-07-08T10:25:03 *** mcstar has joined #aichallenge 2011-07-08T10:25:32 smiley1983: memtest showed no errors, but i gave up, it takes too long 2011-07-08T10:31:30 so what now? the strategy-conversation just stopped? 2011-07-08T10:33:02 lets see, what i have to offer 2011-07-08T10:33:37 i think food spawning could be modified, so that you would have to collect 2 squares within lets say 10 turns, to get 1 ant 2011-07-08T10:34:29 *** VilleH has joined #aichallenge 2011-07-08T10:35:15 *** dude_ has joined #aichallenge 2011-07-08T10:35:24 hey, mcstar this is really a good idea 2011-07-08T10:35:37 dude_: you agree? 2011-07-08T10:35:43 yeah, totally, man 2011-07-08T10:35:47 rabidus: i miss summer :( 2011-07-08T10:35:58 hey, dude_ i already like you 2011-07-08T10:36:05 can you be my best friend? 2011-07-08T10:36:14 mcstar: i was going to ask you the same thing 2011-07-08T10:36:28 Hello 2011-07-08T10:36:29 gtg now, sorry, bye mcstar 2011-07-08T10:36:37 *** dude_ has quit IRC (Client Quit) 2011-07-08T10:36:41 Just started 2 weeks vacation :) 2011-07-08T10:36:42 ah, dude left 2011-07-08T10:37:46 antimatroid: http://www.ouka.fi/index_kamera.asp this city camera 2011-07-08T10:38:31 rabidus: i rode to uni today with track pants under jeans, 2 jumpers, gloves and a scarf 2011-07-08T10:39:20 VilleH: 2 weeks of programming* 2011-07-08T10:39:48 :) 2011-07-08T10:40:01 *** mleise has quit IRC (Read error: Connection reset by peer) 2011-07-08T10:40:36 1 week of fishing & drinking somewhere where isn't any electricity or mobile signals :) 2011-07-08T10:41:01 *** jmreardon has quit IRC (Quit: jmreardon) 2011-07-08T10:41:08 VilleH: where are you from? yeah finland, but where? 2011-07-08T10:41:23 And maybe 1 week of programming :) 2011-07-08T10:41:36 VilleH: i went camping during the summer, like half the people had smart phones with internet 2011-07-08T10:41:40 Rovaniemi, Lapland. Home of lapdancing. 2011-07-08T10:41:49 i don't even own a mobile, completely different experience in the one campsite :) 2011-07-08T10:41:53 we were gone about a week 2011-07-08T10:42:09 VilleH: oulu here :) 2011-07-08T10:42:26 rabidus: Thought so :) 2011-07-08T10:42:39 yeah, but now that beer, cu -> 2011-07-08T10:46:39 *** mleise has joined #aichallenge 2011-07-08T10:49:45 *** jmreardon has joined #aichallenge 2011-07-08T10:53:24 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-07-08T10:56:38 *** jmreardon has quit IRC (Quit: jmreardon) 2011-07-08T11:03:55 *** ArchMonkey has joined #aichallenge 2011-07-08T11:13:34 my recursive command logic just for a couple of times in the first turn, then never again 2011-07-08T11:13:37 i suspect a bug 2011-07-08T11:13:50 couple of thousand* 2011-07-08T11:16:18 i wonder if i program the same silly way as i type: run*, run for...... 2011-07-08T11:17:19 VilleH: you there? 2011-07-08T11:17:28 Yes 2011-07-08T11:17:34 http://aichallengebeta.hypertriangle.com/visualizer.php?game=67742&user=17 2011-07-08T11:17:38 take a look at this game 2011-07-08T11:18:11 my ants must have bribed the engine, otherwise i cant understand why did my bot win 2011-07-08T11:19:24 Haha, seems so :D 2011-07-08T11:20:14 http://aichallengebeta.hypertriangle.com/visualizer.php?game=67602&user=17 2011-07-08T11:20:18 another example 2011-07-08T11:20:40 when it wins, it wins the same way, with much fewer ants than other bots 2011-07-08T11:21:11 i left a few bugs in it, and it seems to be suicidal 2011-07-08T11:21:19 and still wins, i dont understand 2011-07-08T11:21:48 by suicidal i mean, frequently 2 or more of my ants step on each other.. 2011-07-08T11:22:24 Yeah, my first bot did that also 2011-07-08T11:22:54 the problem is that its not my first :D 2011-07-08T11:24:23 Hehe, nice comeback on that last game from 4 vs 59 ants to victory :D 2011-07-08T11:24:52 *** eashoka has joined #aichallenge 2011-07-08T11:25:05 @rankings 2011-07-08T11:25:16 mcstar: Error: There was a problem accessing the interface to ai-contest.com 2011-07-08T11:25:18 I think you are winning with small amount of ants because the battle system is rewarding based on how much your ants are contributing to the killing of other ants 2011-07-08T11:25:45 Though your ants are getting beaten to death at the same time :) 2011-07-08T11:25:48 yeah, i wonder how much better it would be without the continuous suicides :( 2011-07-08T11:27:15 I wonder how my bot would do against any of the "real" bots on the server as it is currently beating those local bots all the time.. 2011-07-08T11:27:43 as the c# or mono doesn't work on server I cant test them there 2011-07-08T11:28:58 too bad 2011-07-08T11:29:29 what can i say? experiment, find it out 2011-07-08T11:30:34 Yeah, I'm running my own bots against each other but hard to say if they are even close to what others have done.. 2011-07-08T11:31:08 hi, anybody knows when would the server be up again ? 2011-07-08T11:31:49 mcstar: remember you are likely collecting points when battling enemies 2011-07-08T11:32:24 antimatroid: ok, but the only thing that bot does well, is collecting food quite optimally 2011-07-08T11:32:28 so even if you don't have the largest army, if you're churning through ants a lot faster it's possible for your score to be high 2011-07-08T11:32:38 and as i said, it dies often from its own fault 2011-07-08T11:32:49 my bot is doing NOTHING in the way of battles 2011-07-08T11:32:54 Actually it seems that c# has started working again, now if the worker would be doing its work I could try it out 2011-07-08T11:32:54 it's just churning through lots of ants 2011-07-08T11:33:01 antimatroid: mine neither 2011-07-08T11:33:11 absolutely has no idea about enemies :) 2011-07-08T11:33:20 mine sents ants towards enemies 2011-07-08T11:33:35 but it doesn't do it smartly, and will only do it if "it thinks it has enough ants" 2011-07-08T11:33:38 but yeah 2011-07-08T11:34:03 antimatroid: cant you help VilleH ? youre cpp, youre closer to it then me 2011-07-08T11:34:14 (to c#) 2011-07-08T11:34:32 what's the problem? 2011-07-08T11:34:38 i've never done anything with c# 2011-07-08T11:34:59 i forgot, but he cant use it on the server 2011-07-08T11:35:10 VilleH: maybe you could tell again.. 2011-07-08T11:35:29 https://github.com/aichallenge/aichallenge/issues/207 2011-07-08T11:36:10 amstan opened issue on it before and now it seems that C# should be working also, if only the worker was running 2011-07-08T11:36:33 i don't think i'd be much help sorry 2011-07-08T11:36:57 No worries, I think the problem has been sorted out allready 2011-07-08T11:37:33 Testbots are kicking again allready: http://aichallengebeta.hypertriangle.com/profile.php?user=184 http://aichallengebeta.hypertriangle.com/profile.php?user=87 2011-07-08T11:39:34 we need those ec2 server asap 2011-07-08T11:39:42 s 2011-07-08T11:45:14 *** nux67 has joined #aichallenge 2011-07-08T12:07:16 *** carlop has joined #aichallenge 2011-07-08T12:09:22 *** VilleH has quit IRC (Quit: Page closed) 2011-07-08T12:18:46 aichallenge: McLeopold epsilon * r907c1ab / (4 files in 2 dirs): changed points for killing an ant to 2 to reward good fighting bots - http://bit.ly/nSVtCH 2011-07-08T12:18:47 aichallenge: McLeopold epsilon * r22b5149 / (10 files in 5 dirs): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://bit.ly/onneRW 2011-07-08T12:28:51 aichallenge: McLeopold epsilon * rc3d23f5 / ants/ants.py : create food dead zone along symmetric mirrored lines - http://bit.ly/pOMjGG 2011-07-08T12:44:35 *** amstan has joined #aichallenge 2011-07-08T12:44:39 *** ChanServ sets mode: +o amstan 2011-07-08T12:45:44 _flag: still have that problem? 2011-07-08T12:46:04 _flag: there's no workers running though 2011-07-08T12:56:08 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 5.0/20110615151330]) 2011-07-08T12:58:36 *** smiley1983 has quit IRC (Quit: leaving) 2011-07-08T13:00:26 <_flag> amstan: I fixed it by reuploading, but it was stuck for about an hour and a half before the workers went down 2011-07-08T13:00:39 <_flag> See submission 20: http://aichallengebeta.hypertriangle.com/profile.php?user=7 2011-07-08T13:00:40 *** sir_macelon has joined #aichallenge 2011-07-08T13:00:58 yeah 2011-07-08T13:01:04 i think the worker crashed on your submission 2011-07-08T13:01:16 and the reason it crashed was because of http connection problems 2011-07-08T13:01:25 and that's because of the crappy host 2011-07-08T13:01:27 <_flag> Well I got 3 games in with the new upload 2011-07-08T13:01:36 <_flag> So it was still working after I resubmitted 2011-07-08T13:01:51 amstan: and it seems rankings command is not working again 2011-07-08T13:02:11 maybe one of the workers crashed and then the second after some time 2011-07-08T13:02:11 contestbot: later tell janzert I'm going to be starting the worker on the main host again, there's no other worker 2011-07-08T13:02:11 amstan: Job's done. 2011-07-08T13:02:28 sir_macelon: not sure why contestbot doesn't like it 2011-07-08T13:03:11 amstan: is it related to the worker crashes? 2011-07-08T13:03:53 sir_macelon: probably to the http errors 2011-07-08T13:05:21 amstan: what do you mean "on the main host"? 2011-07-08T13:05:30 other than the beta server? 2011-07-08T13:05:34 mcstar: same as the website 2011-07-08T13:05:36 no.. 2011-07-08T13:05:43 sir_macelon: if this fails: http://aichallengebeta.hypertriangle.com/ranking_json.php?page=1 2011-07-08T13:05:46 sir_macelon: it'll error 2011-07-08T13:05:58 sir_macelon: i'm using urllib2, idk what the timeout on it is 2011-07-08T13:06:07 probably 10 seconds or something 2011-07-08T13:13:18 amstan: so maybe timeout is too short? 2011-07-08T13:13:36 you really want to wait 10 seconds to get a reply for rankings? 2011-07-08T13:51:15 *** carlop has quit IRC (Quit: Page closed) 2011-07-08T14:00:08 amstan: well, I would like to eventually get an answer, not an error every time; maybe there's a better solution, I don't know 2011-07-08T14:00:47 @rankings 2011-07-08T14:00:48 amstan: Top 10 players: sir_macelon(101.4), Sto03RepliCrush(85.3), antimatroid(82.0), TRavary(81.7), davidjliu(81.3), Zaphus(80.7), Zaphtest02(80.3), OlexandrT(80.3), paratrechina(77.6), pachalmars(77.5) 2011-07-08T14:00:59 sir_macelon: just wait till we get a proper server 2011-07-08T14:01:00 amstan: workers are still down, I thought you were going to start them again? 2011-07-08T14:01:31 sir_macelon your bot is quite interesting to watch 2011-07-08T14:02:07 amstan: It's nothing important, just if there is a simple solution we could use it, but if it is to complicated then we have to live with it during beta 2011-07-08T14:02:22 2011-07-08 22:02:04,038 - 18107 - WARNING - Bad json from server during post result: 2011-07-08T14:02:24 {"hash":"67e6ab18c29fc4ef77a61d8700ccc199"}{"hash":"67e6ab18c29fc4ef77a61d8700ccc199"} 2011-07-08T14:02:25 whatever that means... 2011-07-08T14:02:59 contestbot: seen mcleopold 2011-07-08T14:02:59 amstan: mcleopold was last seen in #aichallenge 4 days, 1 hour, 19 minutes, and 1 second ago: in games per minute, that is 2011-07-08T14:03:10 bhasker: thanks, but it's still rather simple I think 2011-07-08T14:03:18 not sure.. something seems broken 2011-07-08T14:03:41 well i think most bots are still simple, seem to be chase food, if nothing to do chase an ant, if still nothing explore 2011-07-08T14:04:05 or reverse the last two orders maybe 2011-07-08T14:04:43 the difference between bots seem to be, a) how fast they can explore b) not to die in suicide moves 2011-07-08T14:04:48 excuse me, my bot is rather complicated, it makes me tired to improve it 2011-07-08T14:05:13 people are still reluctant to implement any battling, because the final rules are not yet settled 2011-07-08T14:05:22 yeah when does the actual contest start? 2011-07-08T14:05:37 mcstar: you'd better start refactoring then :) 2011-07-08T14:05:48 yeah, constantly :D 2011-07-08T14:06:05 a1k0n_: I am asking the same question 2011-07-08T14:06:33 how long it was for the planet wars to launch? 2011-07-08T14:06:47 does anybody know, how long was the beta period there? 2011-07-08T14:08:55 though i wonder if the fog of war actually hampers smart bots or really smart strategies from evolving 2011-07-08T14:09:00 very short in comparison 2011-07-08T14:12:15 bhasker: it puts more pressure to exploration strategy and therefore some luck 2011-07-08T14:12:28 yea but its making every bot the same 2011-07-08T14:12:34 every bot is just spreading out 2011-07-08T14:12:41 also there doesn't seem to be any incentive in sticking together 2011-07-08T14:12:44 beyond combat 2011-07-08T14:13:02 so we don't see like any marching ants or any grand battles really 2011-07-08T14:13:13 my take is the current structure of the game makes for a very boring watch 2011-07-08T14:13:21 compared to planetwars which was a lot more fun 2011-07-08T14:15:00 it's an early stage still, so I wouldn't expect some great battles yet 2011-07-08T14:15:24 the beta has been long enough for some sophisticated bots to emerge 2011-07-08T14:15:32 the beta has been as long as the contest for planetwars 2011-07-08T14:15:37 aichallenge: McLeopold epsilon * r6b00236 / sql/2_generate_matchup.sql : small improvement to map and opponent selections - http://bit.ly/qZkXvY 2011-07-08T14:16:07 also maybe the wraparound is not a great idea , makes for a larger territory to protect 2011-07-08T14:16:26 non wrapping maps may lead to greater territory control wars 2011-07-08T14:16:34 which would be more fun to watch 2011-07-08T14:17:02 i hope there will be maze-maps 2011-07-08T14:17:13 antimat. is working on generating them 2011-07-08T14:17:40 bhasker: but I agree planet wars contest was very interesting, I regret I found it so late and couldn't compete really 2011-07-08T14:17:56 i think fow is important and realistic, and the wrapping makes the game more interesting 2011-07-08T14:18:16 dunno wrapping and ant spawn rates don't match up making territory defense very hard 2011-07-08T14:18:25 so you can't like easily mark out territory 2011-07-08T14:18:26 bhasker: you still can generate maps surrounded by water, island-like 2011-07-08T14:18:38 i guess one could 2011-07-08T14:19:00 having a wrap around map doubles the perimeter you need to defend 2011-07-08T14:19:17 actually more than double 2011-07-08T14:19:23 quadruples 2011-07-08T14:19:26 *** McLeopold has joined #aichallenge 2011-07-08T14:19:40 bhasker: if you want you can choose to use water as a defensive guard 2011-07-08T14:19:57 the maps right now don't have enough water 2011-07-08T14:20:03 some had 2011-07-08T14:20:12 and ant spawn rates make it impossible to block anything effectively 2011-07-08T14:20:16 mcleo was experimenting with high symmetry maps 2011-07-08T14:20:22 since the perimeter is much much larger than available ants 2011-07-08T14:21:09 as far as i can see, if anything kills this contest, it might be the too simplistic battle resolution 2011-07-08T14:21:18 but its more like a guess 2011-07-08T14:23:38 amstan: what's up? 2011-07-08T14:23:49 the worker is crashing 2011-07-08T14:23:58 check the screen on the main server 2011-07-08T14:24:07 ok 2011-07-08T14:24:55 I see some kill commands... 2011-07-08T14:25:44 yes, run it and see what happens 2011-07-08T14:25:55 something about bad json 2011-07-08T14:25:58 :( 2011-07-08T14:26:07 something bad about my python code, actually 2011-07-08T14:26:42 arg, I'm an idiot 2011-07-08T14:28:30 aichallenge: McLeopold epsilon * r1fc9650 / ants/ants.py : fix to locations set code - http://bit.ly/n4Kt40 2011-07-08T14:30:29 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-07-08T14:32:06 amstan: do we really want a worker on the main server? I think we are going to get a lot of timeouts 2011-07-08T14:32:39 McLeopold: i'm not really sure, i don't think i can access any worker if i don't have their ip 2011-07-08T14:32:55 I can't touch workers either 2011-07-08T14:35:55 *** Naktibalda has joined #aichallenge 2011-07-08T14:39:50 aichallenge: McLeopold epsilon * r0a93dd0 / ants/ants.py : update to food dead zone - http://bit.ly/pZv0LY 2011-07-08T14:41:26 so now we will have 2pts for killed ant is a final rule or just a test? 2011-07-08T15:03:03 *** Accoun has quit IRC () 2011-07-08T15:05:17 a quick poll: do you, experienced programmers use pen and paper? or do you think everything through in your pretty heads? 2011-07-08T15:07:27 mcstar: i find myself opening a new file for ideas pretty often 2011-07-08T15:07:50 a scratch file for code? 2011-07-08T15:07:55 yes 2011-07-08T15:08:04 yeah, i use that too 2011-07-08T15:08:16 and pull debugged code from it 2011-07-08T15:16:30 *** Accoun has joined #aichallenge 2011-07-08T15:18:44 *** McLeopold has left #aichallenge 2011-07-08T15:19:23 mcstar: depends on actual stage of the project; at the beginning I usually start with pen and paper before writing any code; then if I have some backbone already, I start the implementation and unit tests 2011-07-08T15:20:00 ok, im glad im not the only one who has a "real" notebook 2011-07-08T15:20:03 at later stages new ideas are just some TODOs in the code 2011-07-08T15:20:21 where they would actualy fit 2011-07-08T15:20:50 sir_macelon: do you comment for yourself? 2011-07-08T15:21:19 of course and write java-docs everywhere 2011-07-08T15:21:37 it's better then to analyse back the "old" code 2011-07-08T15:21:48 what I had in mind when writing this and why 2011-07-08T15:21:53 i guess, but i cant really do that infortunately 2011-07-08T15:22:00 u 2011-07-08T15:22:31 why? 2011-07-08T15:22:41 im just not used to it 2011-07-08T15:23:11 i wasn't also, but then after trying that I've found a very good practice 2011-07-08T15:23:14 and i rewrite things quite often(at least in this stage) so it wouldnt be of much use 2011-07-08T15:24:07 well, if I am unsure of something I don't document it yet 2011-07-08T15:24:26 but at least leave a short comment what's the general idea behind it 2011-07-08T15:24:46 maybe later I'll change, but idea would stay the same, if it was good 2011-07-08T15:25:21 so it's rather about giving only some rough background, then explaining deeply what does the code do 2011-07-08T15:27:08 ill try to more "commentive" than i am now 2011-07-08T15:28:33 you will surely benefit from that 2011-07-08T15:29:01 http://aichallengebeta.hypertriangle.com/visualizer.php?game=67833 java crashes are back again? :( 2011-07-08T15:36:43 *** jmreardon has joined #aichallenge 2011-07-08T15:47:54 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 5.0/20110622232440]) 2011-07-08T15:51:12 *** ibdknox has joined #aichallenge 2011-07-08T16:01:28 I'm getting some fun errors when I upload 2011-07-08T16:01:38 it appears Java is no longer on the server? 2011-07-08T16:01:43 "Could not create the Java virtual machine." 2011-07-08T16:07:24 hm and worked after a third try 2011-07-08T16:07:55 amstan: can you check those jvm crashes? why they occur now? 2011-07-08T16:39:59 i want to cry 2011-07-08T16:40:34 i nil-led the vector that held the accumulating water information 2011-07-08T16:40:37 every turn 2011-07-08T16:40:41 ahrghargfh 2011-07-08T16:51:07 sir_macelon: the host has that problem, the normal workers are fine 2011-07-08T16:53:45 amstan: but it didn't occur for a long time before, how this was possible? 2011-07-08T16:54:03 yes, till now we had the ec2 workers, but for today idk how to access them 2011-07-08T16:54:11 mcstar: that's what the unit tests are for! :) 2011-07-08T16:54:53 amstan: only janzert knows that I suppose...? 2011-07-08T16:55:12 sir_macelon: unit tests? 2011-07-08T16:55:22 as in measurement? m,sec,kg? 2011-07-08T16:55:58 http://en.wikipedia.org/wiki/Unit_testing ^^ 2011-07-08T16:56:05 i was kidding 2011-07-08T16:56:16 but honestly how would you unit test this? 2011-07-08T16:56:18 heh :) 2011-07-08T16:56:31 write a small map, where you know the exact conditions? 2011-07-08T16:56:41 exactly! 2011-07-08T16:56:48 seems too much work 2011-07-08T16:56:51 :D 2011-07-08T16:56:52 unit tests should as small as possible 2011-07-08T16:57:37 i just nil-led a lot of variables all at once, and silly me left in the water vector too :S 2011-07-08T16:57:51 i guess i was low on coffeine at that moment 2011-07-08T16:58:22 sir_macelon: i promise you, if i become a professional, ill unit tests 2011-07-08T16:58:29 ill do* 2011-07-08T16:58:37 "shit happens", but unit tests are always there to help :) 2011-07-08T16:59:04 it's more a matter of a good habit and practice than being a professional 2011-07-08T16:59:29 sir_macelon: do you insert assert's in to the code? (idk whats that called in java, but you get the point) 2011-07-08T17:00:01 mcstar: only in the unit tests 2011-07-08T17:00:10 ok 2011-07-08T17:00:16 http://aichallengebeta.hypertriangle.com/visualizer.php?game=67911 0-turn game :( 2011-07-08T17:00:21 it not uncommon to have asserts in production code 2011-07-08T17:00:28 it's* 2011-07-08T17:00:36 not just the unit tests 2011-07-08T17:00:46 I don't like asserts outside unit tests 2011-07-08T17:00:56 I consider it as a bad practice 2011-07-08T17:01:07 unit tests should cover such cases 2011-07-08T17:01:08 another "problem" with lisp, is that i dont have to use types 2011-07-08T17:01:24 it makes debugging a little harder 2011-07-08T17:01:53 but if i can pull a feature out, maybe put it in a separate package, then i use type declarations 2011-07-08T17:01:55 I don't know anything about lisp 2011-07-08T17:02:05 but its mainly to speed up the code 2011-07-08T17:02:29 in my a* algorithm, all but a few functions have types declared in them 2011-07-08T17:02:45 sir_macelon: sorry to hear that 2011-07-08T17:03:15 the bigger problem is that i can debug my running program 2011-07-08T17:03:27 cause it has the 1sec limit, and the engine kills it 2011-07-08T17:04:21 known issue, you would need to have your own engine without timing considerations 2011-07-08T17:04:33 i have 2011-07-08T17:04:35 that's why I use unit tests for debugging 2011-07-08T17:04:52 but im sort of lazy to put my logic class into it, cause it would need a small rewrite 2011-07-08T17:05:05 but if i finish my snake now, then i think ill put it in my engine 2011-07-08T17:05:07 mcstar: you have? but how can you be sure it works just the same as the original one? 2011-07-08T17:05:39 its not exactly the same, but could run my logic before a rewrite 2011-07-08T17:05:42 you're writing a snake bot? it is already there 2011-07-08T17:05:49 it has different battle resolution 2011-07-08T17:05:59 sir_macelon: i havent written one yet 2011-07-08T17:06:07 so im sure mine is not already there :) 2011-07-08T17:06:55 I don't see a point in that, but as you wish :) 2011-07-08T17:07:07 and i want to test my recursive-resolve framework 2011-07-08T17:07:09 as a learning exercise? 2011-07-08T17:07:14 sure 2011-07-08T17:07:14 I consider it a as a waste of time 2011-07-08T17:07:16 what else? 2011-07-08T17:07:30 how is that a waste of time? 2011-07-08T17:07:56 you would learn more from trying to improve your own bot rather than trying to do the snake one 2011-07-08T17:08:04 :S 2011-07-08T17:08:05 lol 2011-07-08T17:08:05 that's my opinion 2011-07-08T17:08:21 i sort of dont see the point, but sure 2011-07-08T17:08:37 that's faulty logic 2011-07-08T17:08:51 since a snake bot could be the improvement of your own 2011-07-08T17:09:02 *** FireFly has quit IRC (Read error: Connection reset by peer) 2011-07-08T17:09:05 *** McLeopold has joined #aichallenge 2011-07-08T17:09:30 sometimes it's just fun to mess around too. 2011-07-08T17:09:31 how come it can be an improvement? 2011-07-08T17:09:57 if the snake bot is doing better than your implementation, what have you done? 2011-07-08T17:10:04 improved your bot 2011-07-08T17:10:12 this is a ridiculous semantic argument 2011-07-08T17:10:25 and with a snake bot you can improve on some specifics 2011-07-08T17:10:31 yes, you can also try to write a bot which forms his ants to mimic letters and words, trying to say something :) 2011-07-08T17:10:38 than what works, pull it into your other code.. 2011-07-08T17:10:52 sir_macelon: what is wrong with that? 2011-07-08T17:10:57 and there's nothing wrong with that 2011-07-08T17:11:05 ibdknox: snake bot is a dead end strategy 2011-07-08T17:11:06 obviously that would require a nice robust control framework 2011-07-08T17:11:17 *** FireFly has joined #aichallenge 2011-07-08T17:11:19 sir_macelon: snake bot is a different way of solving the problem 2011-07-08T17:11:24 like you would base your strategy on a bad idea 2011-07-08T17:11:27 sir_macelon: i didnt say i consider it good strategy-wise 2011-07-08T17:11:28 lol 2011-07-08T17:11:32 obviously it sucks 2011-07-08T17:11:42 you completely fail to understand the point of this entire thing 2011-07-08T17:11:53 the point isn't just to win 2011-07-08T17:11:57 if it is, you're doing it wrong 2011-07-08T17:12:01 the point is it doesn't solve the problem efficiently and has upper bound limit 2011-07-08T17:12:03 you will learn a very narrow swath of AI 2011-07-08T17:12:29 sir_macelon: i understand that a snake has huge limitations but the control of it can be smarter than another bot's that is using less logic 2011-07-08T17:12:37 i would recommend to try to write a tron bot rather then a snake bot in ants 2011-07-08T17:12:41 a much better outcome, beyond simply winning, is to learn 2011-07-08T17:12:53 sir_macelon: he's not trying to win 2011-07-08T17:12:58 your advice is good 2011-07-08T17:13:00 if he was 2011-07-08T17:13:11 moreover, you often find great insight in the things that don't win 2011-07-08T17:13:13 ibdknox: ill try, but not with the snake :D 2011-07-08T17:13:15 well, we have different goals then 2011-07-08T17:13:39 to look only at winning solutions is folly 2011-07-08T17:13:44 i would do that maybe after the competition is over 2011-07-08T17:13:56 it hasn't even begun yet ;) 2011-07-08T17:13:59 sir_macelon: you see, it hasnt even started, whats the hurry? 2011-07-08T17:14:29 but really there's so many things you could try implement to learn that you really won't have enough time for 2011-07-08T17:14:40 hell, I'm just trying to go the Clojure stuff to work at this point 2011-07-08T17:14:52 get* 2011-07-08T17:15:19 ibdknox: how are you gooing with it? 2011-07-08T17:15:43 mcstar: there seems to be something weird going on with Java on the server? 2011-07-08T17:15:47 no hurry, but I prefer to use the time to improve my bot rather than trying ideas which doesn't help me in achievieng my goal :) 2011-07-08T17:15:48 im interested in clojure, i find the idea nice: lisp on java's framework 2011-07-08T17:16:07 ah, in terms of Clojure itself, it's wonderful :) 2011-07-08T17:16:14 ah, ok 2011-07-08T17:16:27 I highly recommend it 2011-07-08T17:17:10 do you know racket? 2011-07-08T17:17:31 some, though racket isn't nearly as pragmatic as Clojure and, as such, less interesting to me 2011-07-08T17:17:57 pragmatic? 2011-07-08T17:18:11 Clojure was designed to build production software in 2011-07-08T17:18:16 more functional? 2011-07-08T17:18:30 Racket is trying to move there from Scheme, which is extremely academic 2011-07-08T17:19:16 i saw you can interface directly with java's library(i dont know whats that called) 2011-07-08T17:19:30 you have access to any and all Java code 2011-07-08T17:19:41 which is very useful when you're trying to build something real 2011-07-08T17:20:24 in a lot of ways, Clojure is better at writing Java than Java is ;) 2011-07-08T17:20:37 well, i hope so 2011-07-08T17:20:42 hehe 2011-07-08T17:20:49 yeah, I'm not a fan 2011-07-08T17:21:08 the JVM, however, is amazing 2011-07-08T17:21:47 is it compiling to machine code? 2011-07-08T17:21:51 like sbcl? 2011-07-08T17:22:04 Clojure compiles to Java Byte Code 2011-07-08T17:22:12 and runs on the JVM 2011-07-08T17:22:26 so jvm cant execute machine code directly 2011-07-08T17:22:38 not to my knowledge 2011-07-08T17:23:16 i really like sbcl, its obviously a high level language, but you can very much control the code it outputs 2011-07-08T17:23:31 and always compiles to macine code 2011-07-08T17:23:58 mm 2011-07-08T17:24:09 the need for machine code in most applications is fairly low 2011-07-08T17:24:11 im glad that i considered lisp for the contest, im not sure if i could do this in c 2011-07-08T17:24:19 and the JVM can actually beat it in certain scenarios 2011-07-08T17:24:24 due to runtime optimizations 2011-07-08T17:24:25 well, you can make high performace code 2011-07-08T17:24:36 turning tight loops into jump tables and such 2011-07-08T17:25:51 ibdknox: can you tell me when [] is used? 2011-07-08T17:26:01 it's a vector 2011-07-08T17:26:23 (defn square [x] 2011-07-08T17:26:25 (* x x)) 2011-07-08T17:26:27 here? 2011-07-08T17:26:27 so an indexed array where addition occurs at the end 2011-07-08T17:26:30 right 2011-07-08T17:26:36 so to make things easier to parse 2011-07-08T17:26:46 i dont get this 2011-07-08T17:26:49 Clojure uses vectors for all of its "binding forms" 2011-07-08T17:26:53 aham 2011-07-08T17:27:01 hm 2011-07-08T17:27:03 this is done because order matters 2011-07-08T17:27:08 aham 2011-07-08T17:27:37 you mean like if i push to a list, the new element appears on the front? 2011-07-08T17:27:45 correct 2011-07-08T17:27:48 why is this a problem for bindings? 2011-07-08T17:28:06 when you travel on the cdr's you get the same order 2011-07-08T17:28:12 i mean from left to right 2011-07-08T17:28:28 it's not specifically a problem in the binding case versus a list, it is a specific argument against maps 2011-07-08T17:28:58 in the defn case it's purely to make it more obvious what the params are (as far as I know) 2011-07-08T17:29:05 but take the (let) form 2011-07-08T17:29:09 which looks like this: 2011-07-08T17:29:20 (let [x 1 y 2] (println x y)) 2011-07-08T17:29:50 most people would assume a map makes more sense, as this represents a key and a value 2011-07-08T17:30:09 but you need those to be evaluated in the ordered they're set in 2011-07-08T17:30:18 cl has let and let* 2011-07-08T17:30:22 yep 2011-07-08T17:30:48 sorry, i dont get it 2011-07-08T17:30:59 x bind first, the y 2011-07-08T17:31:02 s 2011-07-08T17:31:10 x binds first, then y 2011-07-08T17:31:12 yes 2011-07-08T17:31:27 so that I could use x in y's definition 2011-07-08T17:31:30 couldn't do that with a map 2011-07-08T17:31:37 but I went off on a tangent 2011-07-08T17:31:39 yes, but why do you need []? 2011-07-08T17:32:10 the reason for not using a list is not a technical argument, I don't believe 2011-07-08T17:32:18 -don't 2011-07-08T17:32:44 i think its nice that i have to use * 2011-07-08T17:32:48 Rich believed it was easier to make lists primarily mean one thing: execute this 2011-07-08T17:32:56 it explicitely shows me that theres something fishy there 2011-07-08T17:33:15 aham 2011-07-08T17:33:22 makes more sense this way 2011-07-08T17:33:46 which way? 2011-07-08T17:34:03 as you lastly put it 2011-07-08T17:34:16 yeah, sorry got off on the wrong foot there 2011-07-08T17:34:35 so anywhere you would use a list as an actual list, you use vector instead 2011-07-08T17:34:38 can a symbol has function value and value at the same time? 2011-07-08T17:34:52 can you give me an example? 2011-07-08T17:34:57 I can do this: 2011-07-08T17:35:06 (def my-func (fn [x] x)) 2011-07-08T17:35:13 (my-func 1) 2011-07-08T17:35:47 so no 2011-07-08T17:35:56 its like scheme then 2011-07-08T17:36:27 clojure is a Lisp-1 if that's what you're asking 2011-07-08T17:36:27 i can say (list list) => (4) if list had the 4 as value, and the list functions as its function value 2011-07-08T17:36:31 yes 2011-07-08T17:37:08 here's a great video targeted at CL guys: http://blip.tv/clojure/clojure-for-lisp-programmers-part-1-1319721 2011-07-08T17:37:23 it's worth watching 2011-07-08T17:37:45 thx 2011-07-08T17:39:25 one more thing, can you splice lists directly? 2011-07-08T17:39:38 destructure, you mean? 2011-07-08T17:39:58 like this (fun @(list 1 2 3 4)) => (fun 1 2 3 4) 2011-07-08T17:40:15 with macros, yes 2011-07-08T17:40:15 its not a correct syntax, i used it for the idea 2011-07-08T17:40:18 without 2011-07-08T17:40:23 no 2011-07-08T17:40:30 but that is easily solved with apply 2011-07-08T17:40:33 yes 2011-07-08T17:40:40 (apply fun [args1 args2]) 2011-07-08T17:40:56 i tell you what i had in mind: 2011-07-08T17:41:44 let say fun takes 4 arguments, but i have 2 lists which i would need to destrucutre(use first and second) to call fun 2011-07-08T17:42:05 (fun (first list1) (second list1) (first list2) (second list2)) 2011-07-08T17:42:13 and id like to do this: 2011-07-08T17:42:22 (fun @list1 @list2) 2011-07-08T17:42:47 (apply fun (concat list1 list2)) 2011-07-08T17:42:55 but actually i think i can do this with a reader macro 2011-07-08T17:42:56 you might not even need the concat 2011-07-08T17:43:08 you might just be able to do 2011-07-08T17:43:13 (apply fun list1 list2) 2011-07-08T17:43:15 I'm not sure 2011-07-08T17:43:20 apply conses up its arguments 2011-07-08T17:43:24 so it wouldnt work 2011-07-08T17:43:40 neither with funcall 2011-07-08T17:43:47 I meant in Clojure :) 2011-07-08T17:43:54 i meant in cl 2011-07-08T17:44:13 i dont think clojure would differ 2011-07-08T17:44:17 but you could try :) 2011-07-08T17:45:00 (apply + 1 2 '(3 4)) ; equal to (apply + '(1 2 3 4)) 2011-07-08T17:45:07 it appears that does work 2011-07-08T17:46:24 (cons 1 (cons 2 (list 3 4))) 2011-07-08T17:46:46 yep 2011-07-08T17:46:49 you're right 2011-07-08T17:47:23 the simple solution with concat works fine though 2011-07-08T17:47:31 yes, i guess 2011-07-08T17:47:36 but its not that clear 2011-07-08T17:47:48 and there can be other things than lists in the arguemnt list 2011-07-08T17:47:56 and you would need to first listify those 2011-07-08T17:48:08 In Clojure, concat works with all sequences 2011-07-08T17:48:33 but yeah 2011-07-08T17:48:37 if you had single calues 2011-07-08T17:48:38 can you concatenate a number with a list? i doubt it 2011-07-08T17:48:38 values* 2011-07-08T17:48:54 mhm 2011-07-08T17:49:21 Just depends on what you expect the inputs to be 2011-07-08T17:49:45 it's not hard to deal with single values, and there's no reason you can't do this completely generally 2011-07-08T17:49:55 tuck it away in a function somewhere and never worry about it again :) 2011-07-08T17:50:12 i forgot the name of the macro im loooking for 2011-07-08T17:50:28 that would tranform the code the way id like it 2011-07-08T17:52:02 set-macro-character 2011-07-08T17:54:22 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-07-08T17:55:05 ibdknox: i added a new lambda, whith which you dont have to specifiy the arguments 2011-07-08T17:55:12 with* 2011-07-08T17:55:40 as in it has variable arguments? 2011-07-08T17:55:46 that too 2011-07-08T17:55:59 like ($ (+ $1 $2 $3)) 2011-07-08T17:56:06 ah 2011-07-08T17:56:14 Clojure has that too: 2011-07-08T17:56:23 #(+ %1 %2) 2011-07-08T17:56:31 nice 2011-07-08T17:56:37 implicit argument declaration 2011-07-08T17:56:45 cl lacks this by default 2011-07-08T17:56:53 nice 2011-07-08T17:56:59 can you nest them? 2011-07-08T17:57:01 how much code did it take to implement that? 2011-07-08T17:57:09 ill take a look 2011-07-08T17:57:16 nope, they're meant solely for simple one off functions 2011-07-08T17:57:36 I would think the nested %1's would get dreadfully confusing anyways 2011-07-08T17:57:55 http://pastebin.com/y0cihEsB 2011-07-08T17:58:18 i can nest mine, as long as obviously you use $N correctly 2011-07-08T17:58:47 if one $N appears after another $, then obviously it will belong to that 2011-07-08T17:58:57 yeah 2011-07-08T17:59:17 but i can say: ($ (funcall ($ (+ $1 $2)) $1 $2)$1 $1) 2011-07-08T17:59:29 haha 2011-07-08T18:00:59 ah sorry i f'ed it uo 2011-07-08T18:01:02 p 2011-07-08T18:01:07 but i can say: ($ (funcall ($ (+ $1 $2)) $1 $2)) 2011-07-08T18:01:13 how about this to make the game more interesting... in order to spawn an ant, you have to "control" the food for n turns, where "control" means it's uncontested and uninterrupted 2011-07-08T18:01:17 *** eashoka has quit IRC (Quit: Page closed) 2011-07-08T18:01:27 that would encourage fights more often 2011-07-08T18:01:51 jmcarthur: you mean to guard it? 2011-07-08T18:01:52 another effect would be to lower the overall number of ants in the game 2011-07-08T18:01:54 that would definitely change how you play the game 2011-07-08T18:01:55 yeah, basically 2011-07-08T18:02:03 thats good too 2011-07-08T18:02:10 as in the strategies would be very different 2011-07-08T18:02:15 this is re: the earlier conversation about making the game less simplistic 2011-07-08T18:02:15 but i still prefer mine 2011-07-08T18:02:33 i mean , i dont see the implications of any of them clearly :S 2011-07-08T18:02:33 what was yours? 2011-07-08T18:02:44 i skimmed very quickly, so i don't know what yours is either 2011-07-08T18:03:02 well, that to gain 1 ant, you need to eat 2 foods, within a couple of turns 2011-07-08T18:03:15 ah 2011-07-08T18:03:31 the same ant needs to eat 2 foods? 2011-07-08T18:03:33 or any ant 2011-07-08T18:03:35 ? 2011-07-08T18:03:47 id say any 2011-07-08T18:04:14 how does this change things? i'm not seeing a big strategic difference 2011-07-08T18:04:16 the same any would require ants to have "roles" 2011-07-08T18:04:18 that doesnt make much sense, do restrict it to the same ant 2011-07-08T18:04:24 ah 2011-07-08T18:04:33 making it the same is a bit more interesting maybe 2011-07-08T18:04:51 yeah 2011-07-08T18:04:57 that means that once that ant has eaten one food it is now "special" and might need to be guarded 2011-07-08T18:05:04 mhm 2011-07-08T18:05:07 would this specialness be visible to your opponents? 2011-07-08T18:05:14 no 2011-07-08T18:05:22 i think it might be nicer if it is 2011-07-08T18:05:31 would encourage opponents to be more offensive 2011-07-08T18:05:35 jmcarthur: if the same would need to eat again, the time limit would need to be much longer 2011-07-08T18:05:37 then it turns into a hunting game 2011-07-08T18:05:47 mcstar: i agree, but don't think that's a bad thing 2011-07-08T18:05:58 ibdknox: and a guarding game 2011-07-08T18:06:08 indeed 2011-07-08T18:06:09 jmcarthur: actually i like your idea more 2011-07-08T18:06:14 use your lesser ants to guard your greater ants 2011-07-08T18:06:21 i think i do too 2011-07-08T18:06:27 but yours isn't bad either 2011-07-08T18:06:28 a combo of the two might be very interesting 2011-07-08T18:07:41 but again, its too late for this kind of changes i believe 2011-07-08T18:07:49 we have a beta for a reason 2011-07-08T18:08:14 but we proposed nice battle startegies, and all that went to waste 2011-07-08T18:08:19 because of time constrints 2011-07-08T18:08:22 a 2011-07-08T18:08:46 we have already switched battle resolution methods since the beta started, haven't we? 2011-07-08T18:09:04 yeah, but its still too simple 2011-07-08T18:09:15 agreed 2011-07-08T18:09:26 and people are pressing to open up the contest 2011-07-08T18:09:27 my point is just that we have already set a precedent for changing things in the beta 2011-07-08T18:09:51 people have been pressing since before the game was even implemented :P 2011-07-08T18:09:54 jmcarthur: its not like im some kinf of authority, id welcome those changes 2011-07-08T18:10:24 i think i'd like to ping some of bigger names in the contest about it 2011-07-08T18:10:29 i think i know what j3camero would say 2011-07-08T18:10:33 (no) 2011-07-08T18:10:35 out of curiosity, who *is* the authority? 2011-07-08T18:10:39 j3camero: 2011-07-08T18:10:48 without the colon of course :\ 2011-07-08T18:11:03 i believe there are others too 2011-07-08T18:11:11 lets mention mcleo too 2011-07-08T18:11:12 yeah, but jeff has the final word, really 2011-07-08T18:11:19 why is that? 2011-07-08T18:11:32 i never saw him say anything here 2011-07-08T18:11:34 <_flag> Something's not right with the server. Previously stable bots are crashing and timing out 2011-07-08T18:11:36 these are rebellious words, but that's a good question :) 2011-07-08T18:11:43 yeah, he's not around enough 2011-07-08T18:11:59 _flag: I've noticed this too 2011-07-08T18:12:25 <_flag> eg. http://aichallengebeta.hypertriangle.com/visualizer.php?game=67911&user=16 2011-07-08T18:12:43 there seems to be a problem with Java 2011-07-08T18:12:50 <_flag> It's not just java 2011-07-08T18:13:01 what're you seeing it with? 2011-07-08T18:13:07 what else? 2011-07-08T18:13:10 <_flag> My bot is doing it to, and so is Sto03RepliCrush, which I've never seen crash before 2011-07-08T18:13:23 yours is? 2011-07-08T18:13:25 <_flag> Wait, he's java to 2011-07-08T18:13:32 <_flag> Unless I just have a bug 2011-07-08T18:13:38 are you flagcrapper? 2011-07-08T18:13:44 <_flag> Without the r, yeah 2011-07-08T18:13:52 thats c right? 2011-07-08T18:13:57 and you crash too? 2011-07-08T18:13:57 <_flag> yup 2011-07-08T18:14:08 <_flag> I timeout though, might be my fault 2011-07-08T18:14:22 "flagcrapper", hah 2011-07-08T18:14:25 <_flag> Although it doesn't happen on my end 2011-07-08T18:14:39 jmcarthur: i couldnt pass this joke :) 2011-07-08T18:14:47 i hope he isnt offended 2011-07-08T18:15:08 well for certain, the JVM isn't booting 2011-07-08T18:15:17 about half the time 2011-07-08T18:25:36 *** _0x47 <_0x47!~slash@p5B12F708.dip.t-dialin.net> has quit IRC (Quit: Leaving.) 2011-07-08T18:32:29 what's happening? 2011-07-08T18:32:47 lots of bots are crashing on start or timing out 2011-07-08T18:33:17 specifically I see this for JVM ones: Could not create the Java virtual machine. 2011-07-08T18:33:22 yeah, we are running a worker on the beta server, which doesn't have good memory 2011-07-08T18:33:35 example: http://aichallengebeta.hypertriangle.com/visualizer.php?game=68205&user=346 2011-07-08T18:33:40 only people those who control the ec2 instances aren't around to restart them 2011-07-08T18:33:48 ah 2011-07-08T18:34:05 We could shut down the beta worker, but then there would be no games 2011-07-08T18:43:23 *** Kingpin13 has quit IRC (Quit: this used to be the life, but I don't need another one) 2011-07-08T18:58:39 *** sir_macelon has quit IRC (Quit: Page closed) 2011-07-08T19:02:19 *** mleise has quit IRC (Quit: Leaving.) 2011-07-08T19:04:39 *** mleise has joined #aichallenge 2011-07-08T19:16:24 *** mleise has quit IRC (Quit: Leaving.) 2011-07-08T19:20:30 Good night, Zion. Sweet dreams. 2011-07-08T19:20:33 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-07-08T19:48:53 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2011-07-08T20:01:46 *** pcavs has joined #aichallenge 2011-07-08T20:53:32 aichallenge: Jake McArthur epsilon * rf63be9e / (4 files in 2 dirs): add support for .txz and .tbz extensions (alternative extensions for .tar.xz and .tar.bz2), and hopefully also fix previous issues with .tar.xz and .tar.bz2 - http://bit.ly/pwUzNh 2011-07-08T20:53:33 aichallenge: Jake McArthur epsilon * ra96df89 / (5 files in 3 dirs): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://bit.ly/qIuwmX 2011-07-08T21:01:00 *** FireFly has quit IRC (Quit: swatted to death) 2011-07-08T21:05:26 *** nux67 has quit IRC (Quit: Page closed) 2011-07-08T21:13:50 *** ibdknox has quit IRC (Remote host closed the connection) 2011-07-08T21:17:05 aichallenge: Jake McArthur epsilon * r88275fb / worker/compiler.py : add a heap size limit as a GHC runtime flag for haskell bots - http://bit.ly/oOgbWT 2011-07-08T21:19:15 lovely: "Page rendered in 30641 milliseconds" 2011-07-08T21:23:11 *** sigh has joined #aichallenge 2011-07-08T21:31:09 *** jmreardon has quit IRC (Quit: jmreardon) 2011-07-08T21:32:42 *** ArchMonkey has quit IRC (Quit: ArchMonkey) 2011-07-08T21:37:38 *** jrdnllrd has joined #aichallenge 2011-07-08T21:43:07 *** ibdknox has joined #aichallenge 2011-07-08T21:46:47 *** AlliedEnvy has quit IRC (Remote host closed the connection) 2011-07-08T21:49:13 *** AlliedEnvy has joined #aichallenge 2011-07-08T21:49:21 *** jrdnllrd has quit IRC (Quit: Page closed) 2011-07-08T21:51:53 *** Accoun has quit IRC (Read error: Connection reset by peer) 2011-07-08T21:53:09 *** Accoun has joined #aichallenge 2011-07-08T22:00:25 *** Accoun has quit IRC (Ping timeout: 260 seconds) 2011-07-08T22:43:06 *** onensora has quit IRC () 2011-07-08T22:47:10 *** ibdknox has quit IRC (Remote host closed the connection) 2011-07-08T23:02:27 *** mleise has joined #aichallenge 2011-07-08T23:08:38 *** foRei has quit IRC (Read error: Connection reset by peer) 2011-07-08T23:08:59 *** mleise has joined #aichallenge 2011-07-08T23:11:02 ah this is nice. my linux now boots 60% faster with ureadahead 2011-07-08T23:11:20 *** sigh has quit IRC (Remote host closed the connection) 2011-07-08T23:12:40 it involved a bit of manual kernel patching, because the last kernel patch i could find was outdated and also some fiddling with init scripts, but all in all it was worth it 2011-07-08T23:13:18 but I guess you ubuntu and arch users already have that in the standard distribution 2011-07-08T23:15:19 *** pcavs has quit IRC (Ping timeout: 250 seconds) 2011-07-08T23:27:07 *** pcavs has joined #aichallenge 2011-07-08T23:27:41 mleise: what's ureadahead? 2011-07-08T23:33:23 aichallenge: Alexandru Stan epsilon * r5794e2e / (3 files in 2 dirs): 2011-07-08T23:33:23 aichallenge: Merge pull request #209 from ibdknox/epsilon 2011-07-08T23:33:23 aichallenge: add clojure starter bot and fix the classpath for clojure support 2011-07-08T23:33:23 aichallenge: Thanks - http://bit.ly/oNyzxL 2011-07-08T23:33:31 *** pcavs has quit IRC (Ping timeout: 250 seconds) 2011-07-08T23:33:36 aichallenge: Alexandru Stan epsilon * r3062ddc / setup/worker_setup.py : 2011-07-08T23:33:37 aichallenge: Merge pull request #205 from jbrechtel/epsilon 2011-07-08T23:33:37 aichallenge: Update scala version to latest stable (2.9.0.1) - http://bit.ly/rqcmPq 2011-07-08T23:33:40 amstan: it is a service that interrupts the boot process 2011-07-08T23:34:05 then it loads some files into the disk cache and the boot process continues 2011-07-08T23:35:35 *** Chris_0076 has quit IRC (Quit: Leaving) 2011-07-08T23:35:37 You first start ureadahead in 'tracing' mode for x seconds, where it will find all files loaded by the system during that time. It then looks at their physical location on the HDD and writes the names down in optimal order. 2011-07-08T23:35:46 *** Chris_0076 has joined #aichallenge 2011-07-08T23:36:25 and then every time you boot it reads them in optimal order based on that list? 2011-07-08T23:36:42 The files in this list are loaded during the next boot. This takes ... I think 20 seconds or so for me. But after that the boot process finishes as if it used a solid state drive. 2011-07-08T23:36:51 epic... why isn't this default? 2011-07-08T23:37:03 *** ibdknox has joined #aichallenge 2011-07-08T23:37:54 It still requires that you manually run a trace session when you changed a lot of files 2011-07-08T23:38:40 even reordering the files on the same disk would be a bad idea. so it is more something for geeks i guess 2011-07-08T23:39:23 anyway I simply added a boot line to grub so i can switch between tracing and non-tracing boot easily 2011-07-08T23:39:54 it was just a matter of adding "gentoo=force-tracing" to the line 2011-07-08T23:40:26 well, can't it record what files that are not on the list? or that were never used? 2011-07-08T23:40:33 that way it's dynamic 2011-07-08T23:49:36 <_flag> melise: you say arch has this by default? 2011-07-08T23:49:51 <_flag> mleise: you say arch has this by default? 2011-07-08T23:52:38 _flag: I think it is an arch tool, yes. but i'm not 100% sure on this 2011-07-08T23:53:10 mleise: ubuntu already has this? i never knew 2011-07-08T23:53:50 amstan: I don't know why they made it so static. It looks as if in 'normal' boot mode it doesn't do any special kernel magic and just loads the files in the most robust way 2011-07-08T23:54:14 just check if you have a /var/lib/ureadahead directory 2011-07-08T23:54:28 yep 2011-07-08T23:54:49 <_flag> I ask because I'm an arch user who would probably find it worthwhile to go through a major hassle for a small speed upgrade :P 2011-07-08T23:55:17 _flag: do you have a /var/lib/ureadahead directory? 2011-07-08T23:55:27 <_flag> nope 2011-07-08T23:55:33 =-O 2011-07-08T23:55:40 then go through the hassle 2011-07-08T23:55:48 <_flag> I can just look it up in the wiki then: https://wiki.archlinux.org/index.php/Ureadahead 2011-07-08T23:55:48 http://forums.gentoo.org/viewtopic-p-6748142.html#6748142 2011-07-08T23:56:18 oh yes, use the arch-linux link of course 2011-07-08T23:56:45 <_flag> Is this something that's going to make it into the main-line kernel? Do you know? 2011-07-08T23:57:26 I have no clue, but I read something about rebasing the patch everytime the syscall changes... it didn't sound promising 2011-07-08T23:57:44 then again, look at ubuntu :) 2011-07-08T23:57:50 <_flag> Yeah, I wouldn't want to have to repatch the kernel everytime I do a kernel update 2011-07-08T23:58:24 <_flag> Ubuntu has the advantage of only doing major updates once every 6 months though 2011-07-08T23:59:07 "[...] and if none are found or if the packfiles are older than a month, it starts tracing the boot process" Ahhhh! Cool stuff! 2011-07-08T23:59:11 mleise: apparently it already does reprofile: http://ubuntuforums.org/showthread.php?t=1434502 2011-07-08T23:59:18 mleise: i have to delete the pack file, then reboot 2011-07-08T23:59:45 amstan: why would you delete the files? 2011-07-08T23:59:58 mleise: i can force a reprofile that way