2011-04-16T00:00:46 *** javagamer has joined #aichallenge 2011-04-16T00:31:48 k thanks 2011-04-16T00:39:33 *** amstan_ has joined #aichallenge 2011-04-16T00:39:33 *** ChanServ sets mode: +o amstan_ 2011-04-16T00:51:21 *** amstan_ has quit IRC (Ping timeout: 276 seconds) 2011-04-16T00:58:11 amtsan done sent the pull request 2011-04-16T00:58:43 feel free to reject it if you feel something isnt right , its just a basic bot that i hacked up using my old code from the last challenge for logging etc 2011-04-16T00:58:57 i think its a little better than the hunterbot 2011-04-16T01:19:21 Ants can see, battle and spawn over water. If a bot issues a move for an ant onto water, it is considered an invalid move and therefore ignored. 2011-04-16T01:19:25 i am guessing this is wrong 2011-04-16T01:19:40 in the Water Blocking section of Ants-Game-Specification 2011-04-16T01:19:45 the first part of that line i.e 2011-04-16T01:21:39 what part do you think is wrong? pretty sure that is indeed the way it works 2011-04-16T01:22:53 battle and spawn over water 2011-04-16T01:23:00 i thought ants can't spawn over water? 2011-04-16T01:23:07 or does that mean they can spawn across from water? 2011-04-16T01:23:20 over as in across, yes 2011-04-16T01:23:29 ah maybe rephrasing that would be better 2011-04-16T01:23:55 there won't ever be food on top of water so an ant won't spawn in water no 2011-04-16T01:24:18 i guess so but it was still confusing when i read it the first time 2011-04-16T01:24:31 hit antimatroid up with it if you want he's doing the spec writing :) 2011-04-16T01:24:50 will do time to go think about how i want to structure my bot for the contest 2011-04-16T01:24:57 this one should be interesting 2011-04-16T01:25:12 noted 2011-04-16T01:25:16 btw, the spec should be taken as a descriptive spec and in general the engine as authoritative 2011-04-16T01:25:39 at least barring bugs in the engine ;) 2011-04-16T01:25:47 :P 2011-04-16T01:25:48 :) 2011-04-16T01:25:50 i was about to say 2011-04-16T01:25:54 antimatroid check out my basic bot that i added 2011-04-16T01:25:59 seems to be beat hunterbot 2011-04-16T01:26:00 for now 2011-04-16T01:26:03 bhasker: I will a bit later 2011-04-16T01:26:26 btw who is writing the engine? 2011-04-16T01:26:41 mcleopold (and sigh) 2011-04-16T01:26:50 sigh is basically taking care of it at this point 2011-04-16T01:27:05 except he is moving state today 2011-04-16T01:27:10 ah k 2011-04-16T01:27:13 so i suspect from here it'll be whoever fixes anything 2011-04-16T01:27:30 i opened a bug ,basically the engine doesn't seem to read the bot's stderr 2011-04-16T01:27:34 so if the bot logs anything ever 2011-04-16T01:27:37 to stderr 2011-04-16T01:27:40 eventually it blocks 2011-04-16T01:27:42 and timesout 2011-04-16T01:27:57 ah yeah okay 2011-04-16T01:27:58 the engine in past contests never read stderr either 2011-04-16T01:28:03 it did 2011-04-16T01:28:13 well at least the local player did 2011-04-16T01:28:16 i used cerr for output in tron? 2011-04-16T01:28:20 not the server one 2011-04-16T01:28:25 maybe 2011-04-16T01:28:26 but it didn't work for pw so i started writing to files, which i prefer anyway 2011-04-16T01:28:33 yeah, server one didn't handle it at all 2011-04-16T01:28:40 we should really just redirect to /dev/null 2011-04-16T01:28:54 the sandbox could probably handle it actually 2011-04-16T01:28:56 k i guess i will just change my logging to write to a file 2011-04-16T01:29:42 so 16 days to beta , and how long is the beta going to be? 2011-04-16T01:30:37 I would guess around 2 weeks but no real plan yet 2011-04-16T01:31:07 k 2011-04-16T01:31:09 also beta really starts as soon as all the parts fit together whenever that may end up being :) 2011-04-16T01:31:34 is any one running a basic tcp server ? 2011-04-16T01:32:08 I don't think anyone has written one yet 2011-04-16T01:32:12 *** Stocha has joined #aichallenge 2011-04-16T01:32:12 ah hmm. 2011-04-16T01:32:20 hi. What's a good map ? 2011-04-16T01:34:50 bhasker: you can use my bot for basic testing if you wish 2011-04-16T01:35:09 although antimatroid had it timeout somehow. others had nor problems 2011-04-16T01:35:27 thanks for the offer will try it out, though for now i am going to spend some time w/ paper and pencil to think my strategy through 2011-04-16T01:35:42 last contest i iterated and edned up with a pile of messy code 2011-04-16T01:35:46 which took forever to clean 2011-04-16T01:35:59 How can you people plan straegy without known what maps will be. that really puzzle me hard. 2011-04-16T01:36:15 maps shouldn't affect strategy 2011-04-16T01:36:26 i don't understand that statement. 2011-04-16T01:36:30 at least the core AI should be adaptable enough to play on any map 2011-04-16T01:36:57 well like the bot i just committed it doesn't even have any knowledge about the map 2011-04-16T01:37:03 I think maps features are more twsited than bots features. 2011-04-16T01:37:05 but it already beats the basic hunter bot 2011-04-16T01:37:23 hunter bot is not even basic. 2011-04-16T01:37:27 it's a starter bot. 2011-04-16T01:37:45 well yea but mostly i don't see how map will change your strategy much 2011-04-16T01:37:47 Simple food gathering with no further moves beat it on the default map. 2011-04-16T01:37:56 default being? 2011-04-16T01:38:05 _10/ 2011-04-16T01:38:08 _10 map? 2011-04-16T01:38:11 *** dlila has joined #aichallenge 2011-04-16T01:38:13 the map that is connected to the play_game script. or whatever 2011-04-16T01:38:19 maybe all that changed :p 2011-04-16T01:38:37 i think its the symmetric_10 2011-04-16T01:38:44 Well, i never heard of hunter bot owning anything. 2011-04-16T01:38:47 on any maps. 2011-04-16T01:38:53 mostly though i believe map analytics can be done at startup 2011-04-16T01:39:05 there is fog. 2011-04-16T01:39:18 yes so you build a map iteratively every turn 2011-04-16T01:39:22 updating your knowledge as you go 2011-04-16T01:39:28 sure. 2011-04-16T01:39:37 but there are stochastics features on maps. 2011-04-16T01:39:42 like food spawning. 2011-04-16T01:39:53 or opponents positions. 2011-04-16T01:39:54 if you consider food spawning to be random 2011-04-16T01:40:06 and assume that in most cases your bot will not know where opponents spawn 2011-04-16T01:40:09 and even then by the time you have an idea what the map looks like the game is ended twice. 2011-04-16T01:40:22 i don't see why the overall map matters 2011-04-16T01:40:35 nobody seems to. Maybe i'm just crazy. 2011-04-16T01:40:46 all that matters is the area you can really influence at any turn 2011-04-16T01:40:52 since ants can't fly 2011-04-16T01:40:56 and they move only 1 unit 2011-04-16T01:41:15 the only thing knowing the map might help is with pathingg 2011-04-16T01:41:19 pathing* 2011-04-16T01:41:26 that's a strong statement. 2011-04-16T01:41:42 pathing will allow certain strategies which simple bots won't be able to do 2011-04-16T01:41:49 like going around obstacles 2011-04-16T01:42:05 pathing is the single feature that nobody ever tested a bot without so far. 2011-04-16T01:42:16 thats probably the first thing i am going to implement 2011-04-16T01:42:18 a* 2011-04-16T01:42:27 and see how fast i can move the ants 2011-04-16T01:42:37 with an a* returning next steps to target 2011-04-16T01:42:59 this game is perfect for a* 2011-04-16T01:43:05 there is still no upper bound to the number of ants. 2011-04-16T01:43:16 well maybe 6000 if we keep to what's been said. 2011-04-16T01:43:28 yea well i guess at some point you will reuse paths 2011-04-16T01:43:33 and not compute for every ant 2011-04-16T01:43:40 256x256 with at most 5% initial food and 4 food/turn. 2011-04-16T01:43:54 1000 turn long 2011-04-16T01:44:04 that should be interesting 2011-04-16T01:44:04 i think that's the worst upper case so far. 2011-04-16T01:44:11 though i doubt that will work 2011-04-16T01:44:15 i doubt you would find it in though. 2011-04-16T01:44:17 games will slow down to a crawl 2011-04-16T01:44:26 i have tested 256x256 games with 10 000 ants. 2011-04-16T01:44:30 and how does it 2011-04-16T01:44:33 perform 2011-04-16T01:44:35 100 ms/ turn with 4 bots. 2011-04-16T01:44:41 i have applets 2011-04-16T01:44:45 you can look at it. 2011-04-16T01:44:52 oh k 2011-04-16T01:44:59 although it's only 1 frame very 4 turn. 2011-04-16T01:45:13 ouch 2011-04-16T01:45:26 mostly though i am wondering if i should continue in python 2011-04-16T01:45:31 or just go back to c++ or something 2011-04-16T01:45:33 Big map size http://ns368394.ovh.net/stocha/Unreason_128x128_AntDieOnSameSquare/ Huge map size (upper bound that McLeopold gave for map size) http://ns368394.ovh.net/stocha/Unreason_256x256_AntDieOnSameSquare/ 2011-04-16T01:45:45 if you have 64 bits system that should be the same sort of timing i have. 2011-04-16T01:46:24 i'm with bhasker 2011-04-16T01:46:31 i ignore map types when writing my bot 2011-04-16T01:46:39 your applet doesn't seem to load very well on my mac laptop 2011-04-16T01:46:50 bhasker: I wrote the c++ starter bot, use that :P 2011-04-16T01:46:52 i have a mac laptop too 2011-04-16T01:47:05 Mac book pro 2011-04-16T01:47:10 hmm maybe its chrome then let me try with someother browser 2011-04-16T01:47:20 do you have java ? 2011-04-16T01:47:26 yea i do 2011-04-16T01:47:30 anyway will check it out later 2011-04-16T01:47:39 antimatroid are you writing in c++ ? 2011-04-16T01:47:44 of course 2011-04-16T01:47:46 i miss the flexibility of python 2011-04-16T01:47:51 but the speed blows 2011-04-16T01:47:55 yeah, i'm slowly learning haskell 2011-04-16T01:48:00 i don't like python 2011-04-16T01:48:03 i have no problems with chrome also. 2011-04-16T01:48:07 hmm weird 2011-04-16T01:48:12 probably something messed up on laptop 2011-04-16T01:48:20 the bot should have no effect on what happens in the browser 2011-04-16T01:48:23 antimatroid had problem with my bot too. 2011-04-16T01:48:29 given the browser is working off replay data 2011-04-16T01:48:36 maybe java just run when it pleases. 2011-04-16T01:48:57 My applet are real time 2011-04-16T01:49:01 bhasker: give a1k0n a prod for us too would you? :P 2011-04-16T01:49:03 not replays. 2011-04-16T01:49:35 i would like to see how he does with ants if he has the time/wanted to 2011-04-16T01:49:48 There is the server calculating the result of each turn, and all 4 bots making up their plan and sending them to the server. 2011-04-16T01:49:50 will do, will talk to him tomorrow 2011-04-16T01:51:10 oops i mean monday 2011-04-16T01:51:23 yeah, i was wondering about that :P 2011-04-16T01:52:32 *** Stocha has quit IRC (Quit: Page closed) 2011-04-16T01:55:50 btw any idea what version of python will be used this time? 2011-04-16T01:55:51 2.6? 2011-04-16T01:55:53 2.5? 2011-04-16T01:56:02 or 3.0? 2011-04-16T01:58:18 whatever is in ubuntu, although I'm not sure now whether we'll be using 10.10 (maverick) or 11.04 (natty) now 2011-04-16T02:00:52 btw the python starter bot has a bug 2011-04-16T02:01:00 it uses time.clock() 2011-04-16T02:01:04 which actually measures cpu time 2011-04-16T02:01:06 and not wall time 2011-04-16T02:01:09 def time_remaining(self): 2011-04-16T02:01:09 return self.turntime - int(1000 * (time.clock() - self.turn_start_time)) 2011-04-16T02:02:16 My guess is McLeopold wrote that while developing under windows 2011-04-16T02:02:21 :} 2011-04-16T02:02:56 certainly should be changed though 2011-04-16T02:05:00 *** jmpespxoreax has quit IRC (Ping timeout: 252 seconds) 2011-04-16T02:05:44 *** jmpespxoreax has joined #aichallenge 2011-04-16T02:07:06 yea on windows it measures the wall time 2011-04-16T02:07:12 but on unix it measures cpu time 2011-04-16T02:07:23 * janzert nods 2011-04-16T02:07:30 I've been caught with that before 2011-04-16T02:16:23 *** amstan_ has joined #aichallenge 2011-04-16T02:16:23 *** ChanServ sets mode: +o amstan_ 2011-04-16T02:19:13 *** amstan__ has joined #aichallenge 2011-04-16T02:19:13 *** ChanServ sets mode: +o amstan__ 2011-04-16T02:21:30 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-04-16T02:21:48 *** amstan_ has quit IRC (Read error: Operation timed out) 2011-04-16T02:30:24 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-04-16T02:33:24 *** boegel has joined #aichallenge 2011-04-16T02:39:31 it would be awesome to use clock time, but it's always sounded too messy 2011-04-16T02:39:49 pesky windows :P 2011-04-16T02:41:19 *** dlila has quit IRC (Quit: Leaving) 2011-04-16T02:43:27 *** sigh has joined #aichallenge 2011-04-16T02:44:22 * sigh is now a resident of sydney 2011-04-16T02:45:27 sigh: :) 2011-04-16T02:46:45 and I have internet, which is all I really need 2011-04-16T02:47:00 haha, spoken like a true nerd 2011-04-16T03:14:35 *** computer1iz_222 has quit IRC (Ping timeout: 240 seconds) 2011-04-16T03:15:43 *** computerwiz_222 has joined #aichallenge 2011-04-16T03:32:59 *** antimatroid1 has joined #aichallenge 2011-04-16T03:33:00 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-04-16T03:34:52 *** kaemo has joined #aichallenge 2011-04-16T03:52:01 *** sigh_ has joined #aichallenge 2011-04-16T03:52:16 *** sigh has quit IRC (Read error: Connection reset by peer) 2011-04-16T03:56:53 *** kaemo has quit IRC (Ping timeout: 248 seconds) 2011-04-16T04:06:04 *** sigh_ has quit IRC (Remote host closed the connection) 2011-04-16T04:23:51 *** amstan__ has quit IRC (Ping timeout: 246 seconds) 2011-04-16T05:19:46 wtf chrome 2011-04-16T05:19:50 apparently anymore isn't a word 2011-04-16T05:19:56 and pidgin 2011-04-16T05:20:09 is it? 2011-04-16T05:25:23 *** Cyndre has quit IRC (Quit: Leaving) 2011-04-16T05:35:12 *** antimatroid1 has quit IRC (Ping timeout: 240 seconds) 2011-04-16T05:36:55 *** antimatroid has joined #aichallenge 2011-04-16T05:38:38 *** kaemo has joined #aichallenge 2011-04-16T05:47:06 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2011-04-16T05:49:00 *** sigh has joined #aichallenge 2011-04-16T05:57:04 *** berak has joined #aichallenge 2011-04-16T05:58:35 *** kaemo has quit IRC (Ping timeout: 258 seconds) 2011-04-16T06:07:59 *** antimatroid has joined #aichallenge 2011-04-16T06:26:14 *** berak has quit IRC (Read error: Connection timed out) 2011-04-16T06:27:22 *** berak has joined #aichallenge 2011-04-16T06:29:14 *** antimatroid has quit IRC (Ping timeout: 246 seconds) 2011-04-16T06:33:44 *** Ice_Harley has joined #aichallenge 2011-04-16T06:33:54 *** antimatroid has joined #aichallenge 2011-04-16T06:37:33 i need a relation to say a node in a tree preceeds another one 2011-04-16T06:37:38 like v1 preceeds v2 2011-04-16T06:41:50 *** sigh has quit IRC (Remote host closed the connection) 2011-04-16T06:48:52 *** delt0r_ has joined #aichallenge 2011-04-16T06:49:56 *** berak has quit IRC (Read error: Connection timed out) 2011-04-16T06:50:21 *** delt0r___ has quit IRC (Ping timeout: 258 seconds) 2011-04-16T06:52:19 *** berak has joined #aichallenge 2011-04-16T07:04:14 *** Apophis has joined #aichallenge 2011-04-16T07:09:21 *** skynetz has joined #aichallenge 2011-04-16T07:13:17 *** berak has quit IRC (Read error: Connection timed out) 2011-04-16T07:14:21 *** berak has joined #aichallenge 2011-04-16T07:22:26 *** needsch has joined #aichallenge 2011-04-16T07:37:38 *** Stocha has joined #aichallenge 2011-04-16T07:37:45 antimatroid what do you mean 2011-04-16T07:37:46 ? 2011-04-16T07:40:11 you want a the isAncestor relation with your tree node. But where is the question ? 2011-04-16T07:44:24 what should the relation be :P 2011-04-16T07:44:31 *** Stocha has quit IRC (Ping timeout: 252 seconds) 2011-04-16T08:47:58 *** Ice_Harley has quit IRC (Ping timeout: 252 seconds) 2011-04-16T08:59:19 *** sigh has joined #aichallenge 2011-04-16T09:00:16 *** berak has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2011-04-16T09:07:52 *** FireFly has joined #aichallenge 2011-04-16T09:15:55 *** sigh has quit IRC (Ping timeout: 252 seconds) 2011-04-16T09:20:29 *** sigh has joined #aichallenge 2011-04-16T09:20:42 *** FireFly has quit IRC (Read error: Operation timed out) 2011-04-16T09:25:41 *** FireFly has joined #aichallenge 2011-04-16T09:30:19 *** andy_ has joined #aichallenge 2011-04-16T09:31:14 *** mceier has joined #aichallenge 2011-04-16T09:32:01 anyone around? 2011-04-16T09:32:44 antimatroid? 2011-04-16T09:32:58 i'm about 2011-04-16T09:34:29 any reason why server sends us a list of dead ants? 2011-04-16T09:34:37 at a particular square? 2011-04-16T09:38:10 ah nm saw the spec 2011-04-16T09:43:18 *** jmpespxoreax has quit IRC (Changing host) 2011-04-16T09:43:18 *** jmpespxoreax has joined #aichallenge 2011-04-16T09:44:27 what's wrong with a list of dead ants? 2011-04-16T09:47:56 no i was just wondering if there was any reason why the server was sending it other than informational 2011-04-16T09:48:39 sorry, i was being productive :P 2011-04-16T09:49:03 need to ride that train whenever it's there 2011-04-16T09:50:52 btw antimatroid you should rename your bug class to something like logger? 2011-04-16T09:51:13 bug seems like some game object rather than a logging class 2011-04-16T09:51:20 :( 2011-04-16T09:51:22 in the starer bot i.e 2011-04-16T09:51:23 i like bug 2011-04-16T09:51:26 i used it for pw 2011-04-16T09:51:32 not all logs are bugs :) 2011-04-16T09:51:41 bugshit 2011-04-16T09:51:42 :p 2011-04-16T09:51:57 otherwise its a nicely written starter bot 2011-04-16T09:52:02 i like it better than the python one 2011-04-16T09:52:16 how'd you like my communist approach? 2011-04-16T09:52:19 everything is public 2011-04-16T09:52:29 state should be public anyway 2011-04-16T09:52:33 since its a struct 2011-04-16T09:52:46 yeah, i always just use structs 2011-04-16T09:53:11 although i've recently broken free of using (abusing) vectors for everything 2011-04-16T09:53:22 not always, but not as much 2011-04-16T09:53:51 well vectors are meantt o be abused:) 2011-04-16T09:54:01 vector/map/stack/queue etc 2011-04-16T09:54:21 yep 2011-04-16T09:54:35 maps are so slow though 2011-04-16T09:54:42 as in? 2011-04-16T09:54:50 maps or stacks 2011-04-16T09:54:56 like storing information about visited squares 2011-04-16T09:55:18 well you can always use hash_maps 2011-04-16T09:55:24 and you can't have a matrix of bools that can be indexed matrix[location] :( 2011-04-16T09:55:40 sure you can why cant you? 2011-04-16T09:55:55 oh as in for maps? 2011-04-16T09:55:57 because you can't return a reference to an element of a vector of bools in c++ apparently 2011-04-16T09:56:08 you could always just use straight up arrays for that 2011-04-16T09:56:09 latest stupid thing i've come across (this week even) 2011-04-16T09:56:19 for a matrix? 2011-04-16T09:56:21 yes 2011-04-16T09:56:26 antimatroid: don't you just use iterators? 2011-04-16T09:56:31 can they be variable sized? 2011-04-16T09:56:35 yes 2011-04-16T09:56:36 delt0r_: not for vectors 2011-04-16T09:57:09 bhasker: i'll do that thanks :) 2011-04-16T09:57:17 np 2011-04-16T09:57:20 beats using chars or matrix[location.row][location.col] 2011-04-16T09:57:24 been a while since i did STL... but i really thought they all support iterators 2011-04-16T09:57:30 they do 2011-04-16T09:57:30 i am going to do some c++ bot and see, i am tired of the amount of cpu the stupid python one takes 2011-04-16T09:57:37 but it's just easier to not use them when possible :P 2011-04-16T09:58:40 i decided i didn't need two templates for my games in my game theory library 2011-04-16T09:58:43 i think that was a mistake 2011-04-16T09:59:06 it would be reasonably easy for people to straight up use it in bots with 2 2011-04-16T09:59:35 what does your game theory library provide? 2011-04-16T10:00:46 it's more for general game theory stuff, but it does 2 player zero sum games, n player normal form and extensive form (with imperfect information) games 2011-04-16T10:00:58 ah k 2011-04-16T10:01:01 https://github.com/antimatroid/GTL 2011-04-16T10:01:05 it's slowly being moved to github 2011-04-16T10:01:13 there is some wiki stuff for normal form games 2011-04-16T10:04:24 ah damnit, github messed with their wiki settings 2011-04-16T10:04:37 it screwed me over with the ants spec and that 2011-04-16T10:04:45 and that's linked on my gsoc application 2011-04-16T10:04:51 damn github !! 2011-04-16T10:04:55 heh somehow i still find git too complicated :-\ 2011-04-16T10:05:06 so did i, but then it's not too bad 2011-04-16T10:05:07 git sucks 2011-04-16T10:05:16 i prefer svn + google code 2011-04-16T10:05:33 git is the ordinary linux/unix-style crap 2011-04-16T10:05:45 in theory everything is fine but in practice it sucks 2011-04-16T10:05:59 i switched to ubuntu like a month ago 2011-04-16T10:06:09 so far it's frozen 3 times, i can't remember the last time xp froze on me 2011-04-16T10:06:12 otherwise i love it though 2011-04-16T10:07:06 git is nice to fork and track stuff, but really its too complicated to use for most people 2011-04-16T10:07:20 after reading the damn manual multiple times i still can't keep the stuff straight in my head 2011-04-16T10:07:27 i love mercurial. especially for its brilliant documentation. 2011-04-16T10:07:57 you get a short yet complete description of what any command does with "hg help ". 2011-04-16T10:08:23 git does *not* offer you that. thats why the wannabe-elite coders love it 2011-04-16T10:08:48 but as a developer i dont have the time to fight through such a badass documentation like git's 2011-04-16T10:08:50 every time again 2011-04-16T10:09:10 lol, github wont let me use > in part of a LaTeX section 2011-04-16T10:09:13 that's ridiculous 2011-04-16T10:09:25 i guess so, i should learn mercurial or something but really i hate learning new source cod control systems 2011-04-16T10:09:32 cvs->svn was a simple enough switch 2011-04-16T10:09:41 but mercurial and git are both very different 2011-04-16T10:09:42 yes, I hate that, too. :( 2011-04-16T10:11:19 \) > \( what a silly fix 2011-04-16T10:12:04 (/\)0o 2011-04-16T10:12:32 it's a sideways pacman 2011-04-16T10:13:22 antimatroid: What? I have latex in my git 2011-04-16T10:13:30 it accepts any and all text 2011-04-16T10:13:39 delt0r_: it wont take < from me anymore 2011-04-16T10:13:42 it would before 2011-04-16T10:13:49 i recently had to update the ants spec page too 2011-04-16T10:13:49 its not git 2011-04-16T10:14:12 he's talking about github, not git 2011-04-16T10:14:12 i don't see how things can stop working that did before and not be git 2011-04-16T10:14:13 git doesn't care about the content --other than the diff 2011-04-16T10:14:19 I think? 2011-04-16T10:14:24 yes, github 2011-04-16T10:14:36 it still git... and i use git hub 2011-04-16T10:14:51 its just git... oh wait are you using some lame web interface? 2011-04-16T10:15:17 he's updating the wiki using the web interface 2011-04-16T10:15:19 i'm making the wiki on their site yeah 2011-04-16T10:15:48 It will be the html sanitation code... or lack there of 2011-04-16T10:16:05 whatever they did, they annoyed me 2011-04-16T10:16:11 I have never used html portal anything with only gmail as an expction 2011-04-16T10:16:26 html does not like < or > 2011-04-16T10:16:26 i still have to go and add my email to the end of my pass key every time i want to pull/commit 2011-04-16T10:16:32 i should really contact them about that 2011-04-16T10:16:47 it worked for me a couple of weeks ago 2011-04-16T10:17:00 antimatroid: you can try cloning the GLT.wiki repo and editing locally 2011-04-16T10:17:13 *GTL.wiki 2011-04-16T10:17:29 sigh: surely the problem is with them displaying it and not saving the file? 2011-04-16T10:17:44 *shrug* i dunno 2011-04-16T10:18:03 antimatroid: I would guess so 2011-04-16T10:18:05 if it's screwing up pages that were working before, then you are right 2011-04-16T10:18:08 that page was broken then before i even touched it, and i doubt the file automatically changed 2011-04-16T10:18:16 that page being https://github.com/antimatroid/GTL/wiki/Normal-Form-Algorithms 2011-04-16T10:18:16 ok then, yeah 2011-04-16T10:18:18 which i fixed 2011-04-16T10:21:27 antimatroid another change i would recommend for the c++ starter bot 2011-04-16T10:21:34 make the logger a singleton 2011-04-16T10:21:42 and don't tie it to the state class as an instance variable 2011-04-16T10:21:56 i shouldn't need to have a state object to log something 2011-04-16T10:22:13 you lost me at singleton :P 2011-04-16T10:22:34 well singleton is a pattern, mostly you create at most one instance of it at any time 2011-04-16T10:22:47 it's like a global variable :P 2011-04-16T10:22:47 and then it's global? 2011-04-16T10:22:54 well its different from globals 2011-04-16T10:22:55 ah, i always wanted stuff like that 2011-04-16T10:23:01 globals are initialized at startup 2011-04-16T10:23:12 singletons lifetime is managed from first invokation 2011-04-16T10:23:17 invocation* 2011-04-16T10:23:39 and more importantly constructors are hidden away 2011-04-16T10:23:47 one can never directly construct a singleton class 2011-04-16T10:24:05 normally singletons will expose a getInstance() or some method named like that 2011-04-16T10:24:12 which will intenrally construct the instance and keep a reference 2011-04-16T10:24:24 either via a private static member variable 2011-04-16T10:24:28 or a function static 2011-04-16T10:24:47 sounds complicated for a "starter bot" 2011-04-16T10:24:57 let me give you a simple logger class 2011-04-16T10:25:03 that works 2011-04-16T10:25:07 okay 2011-04-16T10:25:14 normally i just have Bug *bug in state 2011-04-16T10:25:27 so i can have my states in the game tree nodes 2011-04-16T10:25:57 but i didn't want to give people a starter bot with pointers 2011-04-16T10:26:15 *bug << isn't as nice 2011-04-16T10:28:10 i guess thats fine 2011-04-16T10:28:43 one thing your class lacks is debug levels which are kind of useful 2011-04-16T10:31:41 yeah they could be 2011-04-16T10:32:04 would sort of be hard to do it with the << format 2011-04-16T10:32:09 yes it would be 2011-04-16T10:32:19 and i love the << format :P 2011-04-16T10:32:30 i guess so 2011-04-16T10:32:57 i don't understand why they don't overload them the other direction 2011-04-16T10:33:05 as in? 2011-04-16T10:33:17 so you could be like if((n << cin) < 10) 2011-04-16T10:33:26 you would do 2011-04-16T10:33:30 cin >> n ? 2011-04-16T10:33:35 that returns cin 2011-04-16T10:33:38 ah 2011-04-16T10:34:16 i guess you could overload it that way 2011-04-16T10:34:36 or can't you hmm so long since i wrote overloaded operators etc in c++ 2011-04-16T10:34:41 i have at times for things, but i don't understand why that isn't standard :\ 2011-04-16T10:35:04 cause you can't chain? 2011-04-16T10:35:04 you can do it for everything with templates even 2011-04-16T10:35:19 the point of it is to use it inline 2011-04-16T10:35:34 yes i guess if its important enough you could write one up but stl doesn't cover every case 2011-04-16T10:35:38 *** sigh has quit IRC (Remote host closed the connection) 2011-04-16T10:35:44 stl tries to :P 2011-04-16T10:35:45 you know what i am realizing why i don't like coding in c++ as much lol 2011-04-16T10:35:52 its too goddamn verbose to do anything 2011-04-16T10:35:56 yeah, i am trying to get away from it 2011-04-16T10:36:04 but i don't like python 2011-04-16T10:36:11 why is that? 2011-04-16T10:36:13 and i don't understand haskell enough to jump yet 2011-04-16T10:36:29 it's slow and not very pretty 2011-04-16T10:36:49 i'm not overly convinced on all of haskell either, some of it i really like, other bits i'm like hmmm 2011-04-16T10:37:00 python is fine as long as you understand what parts not to use 2011-04-16T10:37:06 as long as use tuples and lists you are fine 2011-04-16T10:37:09 just don't use dictionaries 2011-04-16T10:37:19 hehe, tell that to sigh and McLeopold 2011-04-16T10:37:20 especially don't use python classes 2011-04-16T10:37:28 they insisted i use a dictionary when writing a map analyser 2011-04-16T10:37:50 well dictionary is fine but its slow especially if you have to copy them around 2011-04-16T10:38:04 i had to rewrite a lot of crap for my planetwars bot 2011-04-16T10:38:07 to use tuples 2011-04-16T10:38:19 i had a state class and deep copying it was SLOW as hell 2011-04-16T10:38:32 so i rewrote the state as a bunch of tuples 2011-04-16T10:38:36 yeah i remember you guys complaining about that :P 2011-04-16T10:38:45 and copying tuples was way way faster 2011-04-16T10:38:53 but then you lose the pretty syntax 2011-04-16T10:38:58 and everything becomes indexed 2011-04-16T10:39:02 which is kinda annoying 2011-04-16T10:39:08 i wish c++ had tuples 2011-04-16T10:39:17 hmm structs? 2011-04-16T10:39:18 i have Triple and Quad defined already for my ants bot 2011-04-16T10:39:22 are the same thing? 2011-04-16T10:39:45 a plain struct with no data members is pretty much the same thing 2011-04-16T10:39:58 oops not data members 2011-04-16T10:40:04 yeah, they're structs 2011-04-16T10:40:05 i meant member fns etc 2011-04-16T10:40:09 but it's silly having to define them 2011-04-16T10:40:14 as in? 2011-04-16T10:40:26 c++ gives you nice syntatic sugar to access individual parts 2011-04-16T10:40:34 like template struct Triple... 2011-04-16T10:40:42 ah 2011-04-16T10:41:15 then one needs to define < for any ordering etc. 2011-04-16T10:43:00 ah true 2011-04-16T10:43:12 k its done i am going back to python for ants no matter how slow it is 2011-04-16T10:43:21 i can't deal with c++'s syntactic issues 2011-04-16T10:43:33 :p 2011-04-16T10:43:45 you'll be back 2011-04-16T10:43:59 people are going to need the speed i think 2011-04-16T10:44:08 i think i will not make the mistakes i did with planetwars 2011-04-16T10:44:31 only thing that concerns me is pathfinding might be a bit too slow on python 2011-04-16T10:44:37 especially solving for a large number of ants 2011-04-16T10:44:58 you can search from all ants simultaneously, but still need to move them all 2011-04-16T10:45:23 A* isn't very helpful when you don't really have an obvious target when you start searching 2011-04-16T10:45:50 what are you doing for food collection? 2011-04-16T10:46:19 my base example is as follows: 2011-04-16T10:46:19 *..a...* 2011-04-16T10:46:19 *..a...* 2011-04-16T10:46:30 do you move them optimally there? 2011-04-16T10:47:20 right now i haven't even thought of my strategy 2011-04-16T10:47:40 well there will be targets 2011-04-16T10:47:45 like move to closest enemy 2011-04-16T10:47:48 or closest food 2011-04-16T10:48:12 and a* will be useful in going around obstacles 2011-04-16T10:48:18 otherwise your ant may run into walls 2011-04-16T10:48:35 but you don't know what the closest food/enemy square is until you search 2011-04-16T10:48:46 it's easy to not search onto water 2011-04-16T10:49:08 well i would just simply look for food by distance around the ant 2011-04-16T10:49:33 and maybe try a couple of the closer ones create a path 2011-04-16T10:49:34 if you add all ants locations to a queue, you can do a bfs from all simultaneously to find the ant closest to a food square 2011-04-16T10:49:37 and measure path length 2011-04-16T10:49:50 hmm probably 2011-04-16T10:49:59 if you update when a square will be "available" on each iteration, you can move them optimally in the above situation 2011-04-16T10:50:03 that's what i'm currently doing 2011-04-16T10:51:34 lets see what i come up with right now i need to think a bit more on this, i am also going to see if flocking would be useful 2011-04-16T10:51:47 like breaking up the ants into squads and making them flock around 2011-04-16T10:51:52 yep 2011-04-16T10:51:57 i don't think it will be 2011-04-16T10:52:15 why do you think it wont be 2011-04-16T10:52:21 given the damage rules 2011-04-16T10:52:25 i'm planning to do food collection somewhat like above, a game tree for any ant that may land in a battle and then heuristic everything else 2011-04-16T10:52:26 i think it will be useful 2011-04-16T10:52:45 then the game tree should handle flocking in a way 2011-04-16T10:52:48 but only when it's needed 2011-04-16T10:53:04 i doubt i will use any sort of game tree but lets see 2011-04-16T10:53:19 but yes, i was very much in favour of the fighting rule specifically because it encourages bunching up for fights 2011-04-16T10:53:30 everything else already encourages spreading, it's good to have balance 2011-04-16T10:53:35 this rule is fine 2011-04-16T10:53:40 since it spreads damage 2011-04-16T10:53:46 having more of your ants closeby 2011-04-16T10:53:46 this is the second best rule 2011-04-16T10:53:48 will help 2011-04-16T10:53:52 but people didn't understand the best rule 2011-04-16T10:54:45 my heuristics for moving non battle and non food collecting ants is going to be a combination of moving towards squares not recently seen and pairing ants up with visible enemy ants, so that should handle sending them closer to the enemy 2011-04-16T10:54:50 i just have to play around with all that 2011-04-16T10:55:08 yup my greedy bot does something like that at a very basic level 2011-04-16T10:55:17 it moves to food if its closeby 2011-04-16T10:55:24 if no food then moves to nearby enemy 2011-04-16T10:55:32 i've discovered it's very hard to move ants to "not recently seen squares" in a smart manner 2011-04-16T10:55:36 if neither then does a random move to a non explored square if possible 2011-04-16T10:55:39 as you end up sending too many to the one patch 2011-04-16T10:56:00 well i just add the square to a site not be sent to 2011-04-16T10:56:02 the best option is to probably bomb visibility where you plan to send the ant on each iteration, but that's costly 2011-04-16T10:56:20 hmm 2011-04-16T10:57:00 otherwise you end up with ants all over the place moving back and forth, often more than is necessary in each section 2011-04-16T10:58:02 true 2011-04-16T10:58:16 anyway gotta run now, will spend some time thinking on the game before i write any code this time 2011-04-16T10:59:18 yep 2011-04-16T11:06:33 *** amriedle has joined #aichallenge 2011-04-16T11:12:04 *** Ice_Harley has joined #aichallenge 2011-04-16T11:16:11 *** Mathnerd314 has quit IRC (Ping timeout: 252 seconds) 2011-04-16T11:16:51 *** berak has joined #aichallenge 2011-04-16T11:17:46 *** Eruonen has joined #aichallenge 2011-04-16T11:35:57 *** skynetz has quit IRC (Quit: .) 2011-04-16T11:36:08 *** Ice_Harley has quit IRC (Quit: Page closed) 2011-04-16T11:53:20 *** Kingpin13 has joined #aichallenge 2011-04-16T11:54:29 *** FireFly has quit IRC (Read error: Operation timed out) 2011-04-16T12:09:26 *** FireFly has joined #aichallenge 2011-04-16T12:15:31 *** jmpespxoreax has left #aichallenge 2011-04-16T12:21:59 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-04-16T12:24:59 *** Chris_0076 has joined #aichallenge 2011-04-16T13:24:26 *** Sunhay has joined #aichallenge 2011-04-16T13:31:08 *** bhasker has quit IRC (Ping timeout: 252 seconds) 2011-04-16T13:32:25 *** Ice_Harley has joined #aichallenge 2011-04-16T13:37:58 *** bhasker has joined #aichallenge 2011-04-16T14:11:50 *** Naktibalda has joined #aichallenge 2011-04-16T14:12:48 *** amstan has joined #aichallenge 2011-04-16T14:12:48 *** ChanServ sets mode: +o amstan 2011-04-16T14:38:56 *** FireFly has quit IRC (Remote host closed the connection) 2011-04-16T14:47:06 *** Ice_Harley has quit IRC (Quit: Page closed) 2011-04-16T14:48:53 *** delt0r___ has joined #aichallenge 2011-04-16T14:49:05 *** kaemo has joined #aichallenge 2011-04-16T14:50:19 *** delt0r_ has quit IRC (Ping timeout: 248 seconds) 2011-04-16T14:52:30 *** andy_ has quit IRC (Disconnected by services) 2011-04-16T14:58:38 *** FireFly has joined #aichallenge 2011-04-16T15:04:38 *** Ice_Harley has joined #aichallenge 2011-04-16T15:46:56 *** Accoun has quit IRC () 2011-04-16T15:58:23 *** Accoun has joined #aichallenge 2011-04-16T16:08:01 *** smellyhippy has quit IRC (Remote host closed the connection) 2011-04-16T16:08:35 *** smellyhippy has joined #aichallenge 2011-04-16T16:38:11 *** kakwa__ has joined #aichallenge 2011-04-16T16:52:03 amstan do we have a final python version for the contest? 2011-04-16T16:52:17 cause i want to make use of certain syntactic sugar which is only available in 3.x 2011-04-16T16:52:25 3.x, no way 2011-04-16T16:52:46 unless we implement it as a separate language 2011-04-16T16:52:49 ah k 2011-04-16T16:52:52 so 2.5 or 2.6? 2011-04-16T16:52:55 go with 2.6 2011-04-16T16:52:56 k 2011-04-16T16:52:58 thanks 2011-04-16T16:53:16 if 2.7 is supported in the new ubuntu, and we switch to it, we might use it 2011-04-16T16:53:18 at least 2.6, possibly 2.7 2011-04-16T16:53:33 k cool thanks 2011-04-16T16:53:40 and for sure not 3.x :) 2011-04-16T16:54:06 kk 2011-04-16T16:54:20 *** Naktibalda has left #aichallenge 2011-04-16T16:54:23 well, I was thinking of trying to make it available for contestants but don't rely on that in any way shape or form 2011-04-16T16:55:34 i was just mostly looking for the syntactic sugar to access tuple indexes by name 2011-04-16T16:55:48 will make it a lot more readable 2011-04-16T16:56:22 really I'm probably more likely to put effort in getting pypy available for contestants before py3 ;) 2011-04-16T16:57:24 use a namedtuple 2011-04-16T16:57:36 for your access tuple indexes by name that is 2011-04-16T16:57:57 thats what i was looking at 2011-04-16T16:58:01 is that available in 2.6? 2011-04-16T16:58:10 since 2.6, yes 2011-04-16T16:58:15 i forget i was trying to use it for last contest and couldn't since 2.5 doesn't support it 2011-04-16T16:58:29 dictionaries in python are super slow:-\ 2011-04-16T16:58:34 and objects are horrible 2011-04-16T16:58:51 well a namedtuple will almost certainly be slower 2011-04-16T16:59:05 does it also use a dictionary? 2011-04-16T16:59:06 since it's just a class 2011-04-16T16:59:14 sigh 2011-04-16T16:59:28 I'm not sure of the internal implementation but it's going to be doing the object method lookups at least 2011-04-16T17:00:07 you really want pypy with its jit :) 2011-04-16T17:00:20 i guess so :) or just go down to c++ 2011-04-16T17:00:43 i had to rewrite a large part of my codebase last time since stupid object copy was ultra slow 2011-04-16T17:00:51 had to move all state information into tuples 2011-04-16T17:00:55 yeah, I have a feeling this contest will be more like tron than planetwars in regards to language speed being important 2011-04-16T17:01:20 i.e. it will be fairly important 2011-04-16T17:01:27 hmm 2011-04-16T17:01:54 so if you found python too slow for planetwars I expect it will only be worse for ants 2011-04-16T17:02:18 *** boegel has quit IRC (Quit: This computer has gone to sleep) 2011-04-16T17:03:15 i guess pythong was slow also because of my lack of knowledge of costs involved with certain operations 2011-04-16T17:03:27 only learnt those things towards the end of the competition 2011-04-16T17:03:37 but i found a few things annoying overall 2011-04-16T17:03:45 object copies are extremely expensive 2011-04-16T17:04:14 so i had to break encapsulation and access member variables directly and store stuff in tuple of tuples 2011-04-16T17:04:16 *** Ice_Harley has quit IRC (Quit: Page closed) 2011-04-16T17:08:10 do you mean method calls are expensive? (which is certainly true) I'm not sure how breaking encapsulation, etc. follows from expensive object copies 2011-04-16T17:08:23 I suppose I'd have to have seen the code :) 2011-04-16T17:18:38 *** berak has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2011-04-16T17:31:30 oh sorry unrelated things 2011-04-16T17:31:54 i was using getter/setters and it was ultra slow vs accessing member variables directly 2011-04-16T17:32:10 and then i was trying to build a min max tree which required me to copy state 2011-04-16T17:32:24 ahh, makes more sense :) 2011-04-16T17:32:28 and when i was storing the state in a nice class it was like an order of magnitude slower 2011-04-16T17:32:40 than storing in a tuple an copying tuples around 2011-04-16T17:32:55 the problem with latter was that everything has to be subscripted by indexes 2011-04-16T17:33:01 which makes for really ugly looking code 2011-04-16T17:33:48 then logging can really slow down python even if nothing is output 2011-04-16T17:34:04 just constructing the log strings alone added quite a bit of slowdown 2011-04-16T17:34:34 how much of a speed up does pypy give? 2011-04-16T17:34:49 i tried it for the planetwars and didn't see that much of an increase frankly 2011-04-16T17:34:58 on my local machine i.e 2011-04-16T17:35:32 depends on the code being run of course http://speed.pypy.org/ 2011-04-16T17:36:58 hmm looks like it only supports 2.5 features 2011-04-16T17:37:10 so if we want to support 2.6 then that rules out 2.5 kind of 2011-04-16T17:37:19 oops rules out pypy 2011-04-16T17:37:53 the just merged in their 2.7 compatibility branch recently 2011-04-16T17:37:55 *** amriedle2 has joined #aichallenge 2011-04-16T17:38:04 but it hasn't been included in a release yet, no 2011-04-16T17:38:19 ah k 2011-04-16T17:38:37 and pypy would only be an alternative if supported at all 2011-04-16T17:38:58 i.e. would have to mark your submission in some way to have it run with pypy instead of cpython 2011-04-16T17:39:33 the default python for submissions will stay with a cpython version 2011-04-16T17:40:06 ah k 2011-04-16T18:01:36 bhasker: I think we can do both python 2.7 and python 3.2 2011-04-16T18:01:57 We can use the extension "py3" for python 3.2 2011-04-16T18:02:01 oh that might be interesting 2011-04-16T18:03:15 hmm... ubuntu seems to be stuck on 2.6 and 3.1 :( 2011-04-16T18:03:46 heh its fine, i think 99% i am going to code in c++ this time 2011-04-16T18:04:00 McLeopold: Natty has 2.7 and 3.2. 2011-04-16T18:04:10 i am tinkering around with antimatroid's starter bot 2011-04-16T18:04:21 and changing things around a bit 2011-04-16T18:05:24 jbroman: is version 10.10 map to a named version? 2011-04-16T18:05:39 McLeopold: 10.10 is Maverick, 11.04 is Natty. 2011-04-16T18:05:41 10.10 is maverick 2011-04-16T18:06:07 so, natty is still unstable right now? 2011-04-16T18:06:09 Natty is scheduled for release in 12 days. 2011-04-16T18:06:17 Currently beta. 2011-04-16T18:06:26 close enough 2011-04-16T18:06:34 that's when we release :) 2011-04-16T18:09:41 bhasker: what should I use instead of time.clock? 2011-04-16T18:09:48 time.time() 2011-04-16T18:09:54 that uses gettimeofday internally 2011-04-16T18:10:01 which always gives wall clock time 2011-04-16T18:10:35 the docs say you can't be ensured a better resolution than 1 second 2011-04-16T18:11:07 you will get at least millisecond resolution on linux and windows though 2011-04-16T18:11:21 I'm not sure what platform gives less actually 2011-04-16T18:11:46 I guess it just means to check the platform for the resolution 2011-04-16T18:41:06 *** sigh has joined #aichallenge 2011-04-16T18:46:41 *** jaycobcoleman has joined #aichallenge 2011-04-16T18:49:55 Even though it isn't explicitly given, couldn't you pretty easily determine the number of players based on the dimensions of the map? 2011-04-16T18:51:08 we are trying to prevent that 2011-04-16T18:51:19 How? 2011-04-16T18:51:40 You'd have to find a second ant and then assume the rows/cols are a multiple of where it probably started 2011-04-16T18:52:19 antimatroid uses a system of shearing the map so that no player starts in the same row/col, so you can't make that assumption 2011-04-16T18:52:45 So I saw. That's smart 2011-04-16T18:53:29 Even so, by the time you get the information on the maps size, you probably have seen all the players, and vice versa 2011-04-16T18:54:28 indeed. So probably not worth the effort trying to exploit in the first place 2011-04-16T18:54:48 Even if you were told the number of players, what would you do with it? 2011-04-16T18:55:07 gather data specific to each bot 2011-04-16T18:55:12 You still need to multiple, explore the map, kill enemy ants and such 2011-04-16T18:55:51 It'll help me determine the best course of action when dealing with a specific bot 2011-04-16T18:56:01 I think the only player information that would be valuable would be where there are probably going to attack next, and a simple number of players won't really help with that. 2011-04-16T18:56:30 Yes I agree. I was simply curious 2011-04-16T18:56:52 You won't be told who you are fighting against, so you can't code for specific bots. All you can do is observe behaviour 2011-04-16T18:57:15 That is the plan as of now. 2011-04-16T19:15:00 mcleopold are you planning to implement some interesting bots lke last time? 2011-04-16T19:15:05 it was fun seeing your bots in action 2011-04-16T19:15:10 *** Frontier has joined #aichallenge 2011-04-16T19:17:15 *** mceier has quit IRC (Quit: leaving) 2011-04-16T19:23:16 *** jmpespxoreax has joined #aichallenge 2011-04-16T19:40:13 bhasker: I told you you'd be back :P 2011-04-16T19:40:18 that didn't even take a day 2011-04-16T19:42:32 also, determining the number of players gives you (or at least really helps you to get) row and col translate, which gives you information about n water squares every time you see one 2011-04-16T19:45:26 seems a little like cheating to me 2011-04-16T19:47:01 yes but i am rewriting parts of your code to make it a little easier to play around with 2011-04-16T19:47:51 i just noticed your starter bot actually dies in 2 moves:-\ 2011-04-16T19:48:05 it doesn't check if the spot its moving into is already occupied 2011-04-16T19:48:16 and i think sometimes it moves its 2nd ant on top of the first one 2011-04-16T19:48:17 gg 2011-04-16T19:48:56 also another thing that will be useful will be to store per square if there is ant there and the owner of the ant 2011-04-16T19:49:01 right now it marks it if there is an ant there 2011-04-16T19:49:04 but not the owner of the ant 2011-04-16T19:58:50 *** FireFly has quit IRC (Read error: Operation timed out) 2011-04-16T20:04:48 jacobcoleman: what is cheating? 2011-04-16T20:05:09 bhasker: i was told i have to do that in the starter bot 2011-04-16T20:05:37 giving your ants an awareness that they're part of a cookie-cutter world 2011-04-16T20:05:40 the starter bot strategy only blocks movement into water, otherwise it's a place for people to start from 2011-04-16T20:07:23 and no it isn't really cheating, I know. But cheating in spirit 2011-04-16T20:07:51 it's not cheating in spirit 2011-04-16T20:08:01 if something is possible, it's easier to just declare it fair play 2011-04-16T20:08:09 it's like saying screen watching on goldeneye is cheating 2011-04-16T20:08:20 it's not if you just declare that fair play, which is the only way to stop people cheating 2011-04-16T20:08:22 Haha, some people do 2011-04-16T20:08:32 alls well and good until you end up on a map that was generated in a way you aren't expecting :) 2011-04-16T20:08:39 I'm totally fine with that 2011-04-16T20:08:51 yep, hopefully we can have lots of different types of maps 2011-04-16T20:09:08 i have put a bit of thought in trying to make it VERY hard for people to get information like mentioned 2011-04-16T20:09:31 Oh it's on 2011-04-16T20:09:46 with some testing, i think all maps should be 120x120 2011-04-16T20:09:55 Why is that? 2011-04-16T20:10:12 it's the lcm of like (2,3,4,5,6,8,10,12) 2011-04-16T20:10:21 Very very good point 2011-04-16T20:10:29 so you can use number of players from that set and not given them information about the number of players 2011-04-16T20:10:56 I was definitely thinking about that yesterday 2011-04-16T20:11:49 i might make another map gen later 2011-04-16T20:12:01 to do a completely different symmetry 2011-04-16T20:12:21 we need to update the map file format so the engine can know what type of map it has, and use the appropriate spawning method 2011-04-16T20:14:13 *** FireFly has joined #aichallenge 2011-04-16T20:18:40 *** Mathnerd314 has joined #aichallenge 2011-04-16T20:23:08 *** spacebat has quit IRC (Remote host closed the connection) 2011-04-16T20:29:11 *** amriedle2 has quit IRC (Quit: Lost terminal) 2011-04-16T20:30:42 animatroid: Can't we just keep different map types in different directories based on type and choose spawning method based on that? 2011-04-16T20:44:36 jacobcoleman: it would be better if the engine could just be passed any map file 2011-04-16T20:54:01 *** suicidolt has joined #aichallenge 2011-04-16T21:04:40 *** sigh has quit IRC (Remote host closed the connection) 2011-04-16T21:08:33 *** antimatroid has quit IRC (Quit: Leaving.) 2011-04-16T21:09:53 *** antimatroid has joined #aichallenge 2011-04-16T21:14:49 *** Frontier has quit IRC (Quit: Leaving.) 2011-04-16T21:21:58 *** kaemo has quit IRC (Ping timeout: 276 seconds) 2011-04-16T21:28:38 *** suicidolt has quit IRC (Remote host closed the connection) 2011-04-16T21:40:14 *** suicidolt has joined #aichallenge 2011-04-16T21:43:56 *** jaycobcoleman has quit IRC (Quit: Page closed) 2011-04-16T21:46:24 *** antimatroid1 has joined #aichallenge 2011-04-16T21:46:32 *** antimatroid has quit IRC (Quit: Leaving.) 2011-04-16T21:48:15 *** antimatroid1 has quit IRC (Client Quit) 2011-04-16T21:48:39 *** antimatroid has joined #aichallenge 2011-04-16T21:51:46 *** suicidolt has quit IRC (Remote host closed the connection) 2011-04-16T22:06:29 antimatroid well the python one ensures that it doesn't do that 2011-04-16T22:06:47 doesn't do what? 2011-04-16T22:08:04 *** needsch has quit IRC (Quit: Leaving.) 2011-04-16T22:10:00 that it doesn't walk onto other ants 2011-04-16T22:16:11 McLeopold: are you avoiding collisions in your starter bot? 2011-04-16T22:16:34 also, where does unix store my application stuff like for latex? 2011-04-16T22:25:03 *** Eruonen has quit IRC (Ping timeout: 258 seconds) 2011-04-16T22:25:56 *** Eruonen has joined #aichallenge 2011-04-16T22:26:53 *** jmpespxoreax has left #aichallenge 2011-04-16T22:33:59 *** Mathnerd314 has quit IRC (Ping timeout: 246 seconds) 2011-04-16T22:38:39 *** Eruonen has quit IRC (Ping timeout: 248 seconds) 2011-04-16T22:39:56 *** amriedle has quit IRC (Ping timeout: 246 seconds) 2011-04-16T22:44:18 *** Mathnerd314 has joined #aichallenge 2011-04-16T22:49:14 *** delt0r_ has joined #aichallenge 2011-04-16T22:50:41 *** delt0r___ has quit IRC (Ping timeout: 246 seconds) 2011-04-16T23:34:28 k finally got the damn bot working