2011-03-21T00:00:46 *** anaptyxis has joined #aichallenge 2011-03-21T00:00:50 "Parkinson's Law of Triviality, also known as bikeshedding or the bicycle shed example, is C. Northcote Parkinson's 1957 argument that organisations give disproportionate weight to trivial issues." - http://en.wikipedia.org/wiki/Parkinson%27s_Law_of_Triviality 2011-03-21T00:00:53 we could use that, but if we do it right the replay file will become bigger. we discussed that already 2011-03-21T00:01:11 sigh: don't do that. this is an important decision :p 2011-03-21T00:02:24 someone will find some reason next term not to use json ^^, but we could write the metadata in that fashion 2011-03-21T00:02:54 then again... is that easier then to use '#%' comments? 2011-03-21T00:03:03 json is more work than prepending a file with a few lines, but it is ok 2011-03-21T00:04:21 * sigh will agree with anything that doesn't put the metadata at the end 2011-03-21T00:07:52 *** anaptyxis has left #aichallenge 2011-03-21T00:07:54 a json parser may not like other data in the same stream. it is an either or decision. and i doubt that all our replays can be expressed in json efficiently. so i come to the conclusion that the initial comment idea is the best so far. 2011-03-21T00:08:40 ok 2011-03-21T00:08:40 sigh: i agree, lol, this is pretty trivial, yet it's one of the most discussed topics in here, lol 2011-03-21T00:11:54 we are almost done. how do we handle array-like stuff in the metadata? currently we agreed on 'playername '. or do we want strictly key-value pairs and not keys with variable argument counts? 2011-03-21T00:12:06 so.. my idea for json replay: https://gist.github.com/878995 2011-03-21T00:12:15 oh let me take a look 2011-03-21T00:13:48 @botrepo 2011-03-21T00:13:49 hehe, that is planetwars style, but ok. you'll have to escape the invalid characters in the replay data and there will be two formats. on 'elementary' replay file and one wrapped in the cool ai-contest replay wrapper 2011-03-21T00:13:49 amstan: botrepo = http://github.com/amstan/contestbot 2011-03-21T00:14:03 *one 2011-03-21T00:14:04 Frontier: isn't that what we wanted? 2011-03-21T00:14:29 i wanted a single replay file that can be extended with metadata 2011-03-21T00:14:49 contestbot: echo [forget botrepo] [learn https://github.com/aichallenge/contestbot as botrepo] 2011-03-21T00:14:49 Yep. As you wish. 2011-03-21T00:15:21 so a single parser could read replays with or without metadata and need not even understand metadata 2011-03-21T00:15:48 but i guess the json approach looks very appealing 2011-03-21T00:16:47 actually we will have 3 formats, as there is also the live stream format from McLeopold. :p 2011-03-21T00:16:48 so the only place there's no reason to not have json is between bots and the engine, because json libraries in most languages are annoying to include in the starter packs 2011-03-21T00:17:00 everywhere else, i would recommend using them 2011-03-21T00:17:13 oh.. the bots and the lengthy replays, yes 2011-03-21T00:17:35 Frontier: live stream format is also one of those blob ones.. 2011-03-21T00:18:02 I'm with frontier on this, just stick some lines at the top of the file 2011-03-21T00:18:42 i'm ok with that, but if you guys want futureproofing... i'd go with json 2011-03-21T00:18:57 but... nvm futureproofing, it's not like we'll be using the manager from epsilon from now on 2011-03-21T00:19:05 except ants 2011-03-21T00:19:16 the whole point for me is a futureproof metadata system 2011-03-21T00:19:37 json will be more futureproof than our homebrew format 2011-03-21T00:19:45 amstan: we will still need metadata in zeta, won't we? 2011-03-21T00:20:05 yes, but i still recommend json 2011-03-21T00:20:27 Frontier: everyone is now ok with your format, no need to argue anymore 2011-03-21T00:20:57 Frontier: you have the final call anyway.. we're just your consultants 2011-03-21T00:21:09 and yes, we're ok 2011-03-21T00:21:47 amstan: You will be with this project for a longer time than me. If this gets phased out for json in zeta, then there is no point in using comments 2011-03-21T00:22:41 the only code that we'll be able to reuse in zeta from epsilon in this regard will be some manager code 2011-03-21T00:23:08 *** genericbob has quit IRC (Ping timeout: 248 seconds) 2011-03-21T00:23:26 so idk if there's a point in worrying so much about future compatibility 2011-03-21T00:23:56 *** genericbob has joined #aichallenge 2011-03-21T00:25:00 the software may be rewritten from scratch, but the file formats are usually upwards compatible. so before someone else says metadata in comments sucks, i rather use a json wrapper :) 2011-03-21T00:26:02 it is like MPEG then. You have Program Streams that contain metadata and interleaved packets from streams and elementary streams, which lack even timing data. 2011-03-21T00:27:33 the json version is the .mpg file with all the metadata and the raw output from the engine is the elementary stream. that makes sense to me as i worked with mpeg files for a company a few years ago 2011-03-21T00:28:55 i say we move the version information (i.e. 'v ants 1') into the json part as well. 2011-03-21T00:29:38 game: "ants", format: "1" (or "stream" and "storage") 2011-03-21T00:29:54 Frontier: when versioning, try to keep git in mind, either make a tag and referrence that, or referrence the particular commit 2011-03-21T00:30:06 Frontier: sounds good 2011-03-21T00:32:28 I believe McLeopold can agree on that, too. And you finally have your bike shed in the style of JSON :p 2011-03-21T00:34:28 lol 2011-03-21T00:37:25 sigh: The lengthy argument is over, you can come back 2011-03-21T00:37:30 *** zvitamiinz has quit IRC (Ping timeout: 252 seconds) 2011-03-21T00:38:05 :P 2011-03-21T00:43:34 *** chuckki has joined #aichallenge 2011-03-21T00:45:36 *** chuckki has quit IRC (Client Quit) 2011-03-21T01:01:13 *** phire has joined #aichallenge 2011-03-21T01:06:07 *** phire has quit IRC (Ping timeout: 276 seconds) 2011-03-21T01:08:43 *** antiuni has quit IRC (Quit: Page closed) 2011-03-21T01:27:06 *** raydenuni has joined #aichallenge 2011-03-21T01:27:08 hello 2011-03-21T01:35:08 *** mceier has joined #aichallenge 2011-03-21T01:35:37 hi 2011-03-21T01:37:30 ok, so i have the JSON object with the following required keys, to be checked in order: "challenge" (: "ants") -> "replayformat" (: "storage") -> "replaydata" 2011-03-21T01:39:01 https://github.com/aichallenge/aichallenge/wiki/Ants-replay-format 2011-03-21T01:41:12 how's the inflow of redditors going? 2011-03-21T01:42:00 Frontier: nitpick: be consistent in how you format param names... e.g submitids, user_ids 2011-03-21T02:02:28 *** avocado has joined #aichallenge 2011-03-21T02:08:40 *** CPeng has joined #aichallenge 2011-03-21T02:09:15 ...now that the information gates are open, should we use redditors to try and gauge how well people can grasp the battle options 2011-03-21T02:09:21 sigh: what's the reception been to your rule? 2011-03-21T02:10:16 McLeopold and jmcarthur liked it 2011-03-21T02:10:22 no one else said anything 2011-03-21T02:11:27 *** thagomizr has quit IRC (Quit: thagomizr) 2011-03-21T02:11:37 so far it has only been brought up in irc 2011-03-21T02:12:05 link? 2011-03-21T02:12:25 CPeng: to what? 2011-03-21T02:12:33 what you guys are talking about 2011-03-21T02:13:08 i dont see it in the forum 2011-03-21T02:13:14 there's isn't really a link as it has only been brought up in irc :P 2011-03-21T02:13:48 k 2011-03-21T02:15:18 so i brought up the issue about labview here: https://github.com/aichallenge/aichallenge/issues/32 2011-03-21T02:15:29 is there something i can research to get the ball rolling on that? 2011-03-21T02:16:09 CPeng: are you familiar with battle resolution? 2011-03-21T02:16:41 re labview: basically we need an answer to what amstan asked; need a simple, free way to set up and run it on ubuntu 2011-03-21T02:17:04 i am not familiar with battle resolution 2011-03-21T02:21:30 *** CPeng has quit IRC (Ping timeout: 252 seconds) 2011-03-21T02:22:15 *** CPeng has joined #aichallenge 2011-03-21T02:22:28 bleh i cant stay connected to freenet for some reason 2011-03-21T02:22:39 what version of ubuntu? 2011-03-21T02:23:37 amstan ^^ 2011-03-21T02:24:27 janzert^^ 2011-03-21T02:24:58 hmmm, does irc ping him without a space? 2011-03-21T02:25:08 depends on his client 2011-03-21T02:25:18 should do 2011-03-21T02:25:40 anyway, if neither of them are answering ask in the issue 2011-03-21T02:25:50 http://www.reddit.com/r/RedditGameJam/comments/g10hy/reddit_game_jam_06_starting_20110325_2200_utc_and/ 2011-03-21T02:25:56 did people see that? 2011-03-21T02:38:56 *** sigh has quit IRC (Ping timeout: 250 seconds) 2011-03-21T02:50:03 Ubuntu 10.10 is what things are currently being based on 2011-03-21T02:55:34 *** amstan__ has joined #aichallenge 2011-03-21T02:55:34 *** ChanServ sets mode: +o amstan__ 2011-03-21T03:00:30 *** Stocha has joined #aichallenge 2011-03-21T03:02:53 antimatroid: http://marco-leise.homedns.org/antgui/work/visualizer.php?game_id=2#visualizer 2011-03-21T03:04:09 the blue forward and backward buttons now start a slow motion to the next half-tick. so you can now stop after half a turn to see where the ants end up after moving without fighting or conversion 2011-03-21T03:04:26 *** amstan__ has quit IRC (Ping timeout: 276 seconds) 2011-03-21T03:05:43 * janzert doesn't see any forward or back buttons on FF 2011-03-21T03:08:14 janzert: firefox renders the vis. well. maybe i have a bug in the image loading routine. what if you load the page again? 2011-03-21T03:08:42 Frontier: :D 2011-03-21T03:08:55 that's awesome 2011-03-21T03:08:58 nope, where are the buttons supposed to be, below the main view? 2011-03-21T03:09:22 yeah 2011-03-21T03:09:23 I like the way the graphs are coming along at the top 2011-03-21T03:10:03 weird, on third refresh with firebug turned on they showed up 2011-03-21T03:10:47 and now they seem to consistently be there 2011-03-21T03:10:53 hmm, someone else had issues on safari, too until he opened the dev console. 2011-03-21T03:11:04 I don't know what it is though. 2011-03-21T03:11:22 ugh, that makes it hard to debug 2011-03-21T03:11:27 janzert, Frontier: what do you two think about sighs new battle resolutions? 2011-03-21T03:11:45 im' litlle worried for computing time given to bots at the moment. 2011-03-21T03:11:49 antimatroid: I am glad that this is not my front ^^ 2011-03-21T03:11:53 each ant dishes out 1 damange each turn, 1/#enemiesinrange to each enemy, if an enemy has >= 1 damage, they die 2011-03-21T03:12:05 i save there ware inconstistencies to what was given to bot in planet war also. 2011-03-21T03:12:07 Frontier: :P 2011-03-21T03:12:11 it's still good to get opinions 2011-03-21T03:12:21 Stocha: inconsistencies? 2011-03-21T03:12:27 I think this ant things will need quite a punch even for very basic things 2011-03-21T03:12:37 I think I like it on first glance at least 2011-03-21T03:12:52 it means a..b..c or a..b..a just kills b 2011-03-21T03:12:55 peoples complaining they had only 500 milliseconds rather than 1000 2011-03-21T03:13:05 which i think makes cooperation much more prominent 2011-03-21T03:13:05 jeff /hates/ health bars 2011-03-21T03:13:12 there is no health bar 2011-03-21T03:13:20 you go back to 1 health essentially each turn 2011-03-21T03:13:23 or that the 1000 millisecond could mean 1 instruction as well as 100000. Depending on randome external event. 2011-03-21T03:13:31 so you "recover after battle" 2011-03-21T03:13:52 it's a lot like option 1 (not the same), but meant to be easier to understand 2011-03-21T03:13:53 I think with collisions always die instead of collision orders are just canceled some of the old rules that were thrown out may make sense now too 2011-03-21T03:14:18 yeah, and points will be assigned for colliding enemies 2011-03-21T03:14:39 but this is seems a nice intuitive method that should have interesting play 2011-03-21T03:14:41 Stocha: at least the bots will run sequentially 2011-03-21T03:14:43 I do on basic things at the moment, and with many ants, it can begin to be slow. I see no way to hope for counting ennemies on different tree hypothesis. 2011-03-21T03:15:02 Solving basic food gathering take quite some cpu power also. 2011-03-21T03:15:11 Stocha: assuming you have lots of ants, you're only going to be looking one state in the future 2011-03-21T03:15:22 you only really need to do the battle resolution at your terminal nodes 2011-03-21T03:15:37 i don't think food gathering will be that expensive 2011-03-21T03:15:40 why do we need complex rule then ? if we can do no anticipation ? 2011-03-21T03:15:52 I made a basic food gathering engine. 2011-03-21T03:15:53 just paint out from each known food square at the start of the turn, moving the ants you want 2011-03-21T03:16:11 As i don't know the computing power each turn, i assume the worth 2011-03-21T03:16:30 i feel uncomfortable with only food gathering and basic ant movement at the momement. :p 2011-03-21T03:16:48 so do we, hence the desire for interesting attacks 2011-03-21T03:17:11 If you said : an i5 machine with one cpu fully supporting you, for 2 minutes / turns, i could be more relaxed 2011-03-21T03:17:30 i have got no idea what the hard will be, and what amount of computing that will mean. 2011-03-21T03:17:58 when i say i feel uncomfortable i mean : it needs cpu 2011-03-21T03:17:59 you will get like 0.5 seconds a turn 2011-03-21T03:18:00 to get it right. 2011-03-21T03:18:27 we wont be going below that anyway 2011-03-21T03:18:31 that hardly make it possible to explore complex possibilities with complex slow to compute rules. 2011-03-21T03:18:34 that was the previous promised minimum amount 2011-03-21T03:18:41 they aren't that slow 2011-03-21T03:18:51 what computer will you provide ? 2011-03-21T03:19:16 *** Stocha has quit IRC (Quit: Page closed) 2011-03-21T03:19:46 no one else is bothered that blue wins the game Frontier posted above? 2011-03-21T03:20:04 janzert: green would have won with a longer game 2011-03-21T03:20:07 *** amstan has quit IRC (Ping timeout: 250 seconds) 2011-03-21T03:20:13 they just hadn't started killing people yet 2011-03-21T03:20:24 realise blue lost a lot more ants too from fights probably 2011-03-21T03:20:43 Frontier: at least for the moment could we get total number of ants had maybe? 2011-03-21T03:20:50 we better make sure all games are long enough then :) 2011-03-21T03:21:14 yeah, i want to start everything way out of range then bring it in 2011-03-21T03:21:36 rather than start small then find a non-subset better set of parameters/whatever 2011-03-21T03:22:08 what is the save button for? 2011-03-21T03:22:21 to download the game eventually? 2011-03-21T03:23:07 *** espes has joined #aichallenge 2011-03-21T03:23:25 that'd be my guess 2011-03-21T03:24:38 the visualizer is really shaping up nice 2011-03-21T03:26:33 <_flag> Why aren't the scores based on the number of ants a player has? 2011-03-21T03:28:05 there isn't an incentive to kill off your opponent, just keep him beat down below you 2011-03-21T03:28:21 *** sigh has joined #aichallenge 2011-03-21T03:28:28 I'm not a big fan of the current system though either ;) 2011-03-21T03:28:41 <_flag> Why not a ratio of your total ants to his? 2011-03-21T03:28:50 <_flag> So that means if you kill him your ratio will be higher 2011-03-21T03:29:06 <_flag> The fact that blue won that game is obviously kind of silly 2011-03-21T03:29:59 I'd still like it to just be inverse order of elimination, but most seem to be strongly against that 2011-03-21T03:30:41 <_flag> That would work too, in fact it might be even better because it doesn't encourage any kind of playstyle. If you win you win. 2011-03-21T03:30:51 yep 2011-03-21T03:31:11 look at the game, blue was involved in a lot more fights 2011-03-21T03:31:24 i'm curious what the total number of ants was to each player 2011-03-21T03:31:33 regardless, green was about to kick its ass 2011-03-21T03:31:44 <_flag> Is there not an ant counter at the top for you? 2011-03-21T03:31:50 yep, that's the point green is pretty clearly ahead 2011-03-21T03:31:56 it's a current ant counter 2011-03-21T03:31:58 not aggregate 2011-03-21T03:32:37 at best I think blue should have tied, since it did survive to the end of the game 2011-03-21T03:32:59 <_flag> Well most kills are a 1:1 thing, so green would be approximately 147 + 83 and blue would be 79 + 91 2011-03-21T03:33:17 janzert: the game was 100 turns, i expect to have much longer games 2011-03-21T03:33:34 like, i'd start at max 1000 turns 2011-03-21T03:33:36 and see how that goes 2011-03-21T03:33:56 yes, longer games will help but I'm not sure will eliminate the problem 2011-03-21T03:34:10 <_flag> This senerio could happen with longer games too. Not as much, but it could still happen. 2011-03-21T03:34:19 Frontier: can we get a 1000 turn version of that game? :P 2011-03-21T03:34:31 it happened with planet wars 2011-03-21T03:34:43 i saw numerous games of mine on tcp where the result would have changed given more time 2011-03-21T03:34:50 right so we should learn from that :) 2011-03-21T03:34:55 what? 2011-03-21T03:35:15 i mentioned it during planet wars and was keen on a longer match time, and people were pretty unanimous in saying it doesn't really matter 2011-03-21T03:35:34 no matter where you cut the game off at there will be the same problems some of the time 2011-03-21T03:35:36 provided it's as rare as it was for planet wars 2011-03-21T03:35:37 <_flag> In planetwars having more ships was directly correlated to winning. In this having more kills is not as important as having more ants. 2011-03-21T03:35:53 I wanted ties if both survived to the end in planetwars too 2011-03-21T03:36:02 _flag: there is also a food bonus that hasn't been implemented 2011-03-21T03:36:04 but felt it was too late to change by the time I realized it 2011-03-21T03:37:05 if an ant is the last remaining player for a game ending before maxTurns, that player gets a point for every food square that is on the board or would have spawned and every enemy ant that's on the board from a crashed bot 2011-03-21T03:37:29 so really, we just need to pick maxTurns so that games can reasonably play out and i don't think it'll be much of a problme 2011-03-21T03:37:53 although we do need to see some examples of the scoring, but that will always be the case until we can test stuff fully 2011-03-21T03:39:21 <_flag> What if we have a scenario where two bots are fighting and killing each other off, and a third bot builds up ants "wins". Would that person still lose if they had less kills and the food bonus wasn't enough? 2011-03-21T03:40:07 <_flag> It seems silly to me to rank bots on anything other than either who won, or who was most likely to win 2011-03-21T03:40:51 the food bonus was supposed to be large enough to ensure the third wins but I'm not sure if it really is 2011-03-21T03:41:27 and yeah, I really like inverse elimination better 2011-03-21T03:41:50 <_flag> So basically we have to try to create an algorithm that immitates inverse elimination without actually being inverse elimination? 2011-03-21T03:42:26 that's as much as I could talk people into :P 2011-03-21T03:42:32 <_flag> :P 2011-03-21T03:46:36 *** a_Tick has joined #aichallenge 2011-03-21T03:57:19 *** epaga has joined #aichallenge 2011-03-21T03:59:59 *** shutch has joined #aichallenge 2011-03-21T04:00:48 *** _andy <_andy!d5730a62@gateway/web/freenode/ip.213.115.10.98> has joined #aichallenge 2011-03-21T04:02:06 *** fishy1 has joined #aichallenge 2011-03-21T04:13:12 *** epaga has left #aichallenge 2011-03-21T04:17:50 *** delt0r_ has joined #aichallenge 2011-03-21T04:19:13 *** delt0r___ has quit IRC (Ping timeout: 252 seconds) 2011-03-21T05:02:33 *** Eruonen has joined #aichallenge 2011-03-21T05:05:42 *** Stocha has joined #aichallenge 2011-03-21T05:07:29 sigh: i have a new most gross problem of the year :P 2011-03-21T05:07:47 well a whole assignment really, why the hell did i decide to take a physics unit :\ 2011-03-21T05:08:01 physics is phun 2011-03-21T05:08:20 what is the physics? 2011-03-21T05:08:51 symmetry stuff, so maths :P 2011-03-21T05:09:14 *** caution has joined #aichallenge 2011-03-21T05:09:30 this question has tensors, i've not actually been given a tensor for actual classes before 2011-03-21T05:09:40 but you love tensors :P 2011-03-21T05:10:02 i do 2011-03-21T05:10:06 it's still messy 2011-03-21T05:14:01 Hello there. 2011-03-21T05:14:23 I have been thinking about this game, and how to sort through all the ideas. 2011-03-21T05:15:03 I think testing is clearly needed. With different sets of rules. The most basic ones. And some more twisted one. Comparing how bots do in the different environnement. 2011-03-21T05:15:16 Obviously you need bots in order to test. That's quite a problem :) 2011-03-21T05:15:46 Anyway what strikes me right now, is the need to some TCP servers, each with a different game engine 2011-03-21T05:16:01 So that people could start to put bot on. 2011-03-21T05:16:15 And the design team would have lot's of data te chew upon. 2011-03-21T05:17:19 You would need some machine for keeping the differents servers opens. 2011-03-21T05:18:18 I really see no way out of this. 2011-03-21T05:18:29 As far as my testing goes, i have several thing to say. 2011-03-21T05:19:03 First i really think rule "ant dies on same square" needs further testing before being discarded. I mean who really tryed to build a good bot with this rule on ? 2011-03-21T05:19:47 Second, i think it is far too soon to say this game should be multiplayer. 2011-03-21T05:20:06 In my testing, multiple bots fighting doesn't bring much to the table. 2011-03-21T05:20:34 I really see no urgent improvement to the bots engine that would take the multiplayer parameter into account. 2011-03-21T05:20:41 It's just you and some ennemy ants. 2011-03-21T05:21:33 that the ennemy is in fact splitted in 3 others teams, is of no concern to the Ai of the bot at the moment. I really see nothing i can use the ennemy ant ownership information for. 2011-03-21T05:21:42 aichallenge: sigh epsilon * rfe323a7 / ants/ants.py : Make food and attack method mapping public and fix some errors in comments. - http://bit.ly/dWPSPR 2011-03-21T05:21:57 stocha: other fighting rules encourage cooperation 2011-03-21T05:22:06 someone logged an issue on github because one of my comments was wrong 2011-03-21T05:22:14 consider this situation : ....a..b..c...... 2011-03-21T05:22:16 i think you are going a bit far on this one :) 2011-03-21T05:22:38 with option 1 and sighs new rule, a and c kill b 2011-03-21T05:22:55 What is certain, is that we need a tcp server, in order to put on some reasonable bots. Like my ExplorerBot, and some other's 2011-03-21T05:23:12 there will be a tcp server, we just haven't got everything ready enough yet 2011-03-21T05:23:15 So if you agree on this, i can help you with this particular goal. 2011-03-21T05:23:34 although we shouldn't be that far off? sigh? 2011-03-21T05:23:38 Which i feel is quite urgent :p 2011-03-21T05:23:54 antimatroid: no idea, all I know about is the engine 2011-03-21T05:24:05 testing parameters has always been a necessary part of getting the game ready 2011-03-21T05:24:23 it's also not really possible until most other things are ready, which is not the case, it's not a case of saying oh that's important, we'll do that first 2011-03-21T05:24:26 I'm very comfortable with java, and quite uncomfortable with pretty much anything else. but if need be, i can try python or anything you like. Only it will takes ages for me to do anything useful :p 2011-03-21T05:24:41 i have only ever touched python for this 2011-03-21T05:24:52 although java is rank :P 2011-03-21T05:25:08 what others things might be more important than a tcp server in order to settle the rules a bit ? 2011-03-21T05:25:29 having working components that allow a tcp server to even work? 2011-03-21T05:25:30 I mean isn't the spec of the game very important for the challenge ? 2011-03-21T05:25:41 What components are lacking ? 2011-03-21T05:25:54 I think you have some game engine. 2011-03-21T05:26:02 you might be able to knock up a quick visualiser without worrying about all the server stuff etc. but there's a surprising amount of work, along with ranking bots, etc. 2011-03-21T05:26:26 ranking bot in a true sense might be something. 2011-03-21T05:26:36 but you also can do some very rough ranking to start with. 2011-03-21T05:26:38 not possible to "perfectly" rank bots 2011-03-21T05:26:50 I'm all for simplicity while testing. At least at first. 2011-03-21T05:27:05 it is possible to produce the raw score data for each game. 2011-03-21T05:27:13 how is that far for perfection ? 2011-03-21T05:27:32 anyone can dig through the data and make his mind. untile we got something to compute this data. 2011-03-21T05:27:43 *** olexs has joined #aichallenge 2011-03-21T05:27:48 into something hopefully mor intelligible 2011-03-21T05:28:15 Something that i'm wondering is : who is in charge of the project ? 2011-03-21T05:28:30 How are decisions made ? 2011-03-21T05:29:06 about the subprojects, dans their goal. And to settle disagreement if there are any (for example what scoring rules is more important to test first) 2011-03-21T05:29:46 I'm okay with no one is in charge, anyone doing what's seems right. but i would like to know for sure how things works in there :) 2011-03-21T05:30:13 *** Naktibalda has joined #aichallenge 2011-03-21T05:30:49 j3camero is in charge 2011-03-21T05:31:04 For example, let's say that i come with some engine with some tcp capabilities for submiting bots, and a tcp starting pack. Would someone be able to get a machine to run the server ? 2011-03-21T05:31:22 Is he around often ? 2011-03-21T05:31:25 no 2011-03-21T05:31:46 who's in charge when he is not there then ? 2011-03-21T05:31:59 how do things works when he is not there ? 2011-03-21T05:32:08 nobody, we come up with something we can all bear of wait for his ruling 2011-03-21T05:32:19 amstan is the closest to a second in command 2011-03-21T05:32:28 he is around frequently 2011-03-21T05:32:56 where can i find amstam plans / planning ? 2011-03-21T05:33:02 for the challenge. 2011-03-21T05:33:11 the tasks list 2011-03-21T05:33:18 on the wiki i'm pretty sure 2011-03-21T05:33:35 i should have a look at that then. 2011-03-21T05:38:44 I looked into it (i think). I see nothing about sorting out rules. 2011-03-21T05:39:15 That kinda worry me :p 2011-03-21T05:39:51 I see some task for starte pack. but nothing about the GameEngine(s) implementation tasks 2011-03-21T05:40:59 Most tasks refers to stuff that i don't really know what they are about at this point. The only thing clear to met being that their is a set of games (with different rules features) that will be sorted down to one game (with one set of feature) that people will try to make bot for. 2011-03-21T05:41:11 it's on to the "to do" list, we're just not to the stage that we feel ready to do it 2011-03-21T05:41:37 But i have heard of launching the stuff in like a week ? 2011-03-21T05:41:49 some people are kidding themselves 2011-03-21T05:41:54 it seems more like a few 6 month right now. 2011-03-21T05:42:20 So let's say we are still in the prototyping phase. 2011-03-21T05:42:43 it should me live within i month, i hope 2011-03-21T05:42:46 be* 2011-03-21T05:43:06 Wouldn't it be usefull to have a quick dirty way of making (strong)bot play one another on different sort of rules ? 2011-03-21T05:43:49 the engine can run pretty much all rules 2011-03-21T05:43:59 once everything is set up including a tcp server, we will test them 2011-03-21T05:44:04 so that we can see more clearly out what a "good" set of rules will be. And also some people would have fun making some bot, and exploring the different set-features constraints 2011-03-21T05:44:14 that's precisely why there's so many unknowns, we haven't got stuff to a testable point 2011-03-21T05:44:36 but i have seen bots playing one another yet. 2011-03-21T05:44:47 in thoses visualizer showcases. 2011-03-21T05:45:05 yes, but there is more to it than just that to actually test bots against each other with rankings etc. 2011-03-21T05:47:20 *** ItzMattu has joined #aichallenge 2011-03-21T05:48:11 *** phire has joined #aichallenge 2011-03-21T05:49:50 what more ? 2011-03-21T05:50:22 how is that more so coupled that you can't go for the objective of test bots against each other without doing "more" stuff before ? 2011-03-21T05:50:47 Is it a ressource problem, or an architectural problem ? 2011-03-21T06:11:05 *** olexs has quit IRC (Quit: Leaving.) 2011-03-21T06:11:38 *** Sunhay has joined #aichallenge 2011-03-21T06:17:34 *** boegel has joined #aichallenge 2011-03-21T06:44:58 aichallenge: sigh epsilon * rad40548 / ants/battle_sim.py : Added a small script to simulate a battle - http://bit.ly/f9fIlz 2011-03-21T06:45:58 antimatroid: you might like this, I wrote a small script to test the output of a battle 2011-03-21T06:46:03 *result of the battle 2011-03-21T06:46:45 :) 2011-03-21T06:49:35 it should be handy for deciding between battle options 2011-03-21T06:53:28 yep 2011-03-21T06:53:44 i'm still wrestling this assignment problem atm (although i've got it sorted I think) 2011-03-21T06:53:49 i will have more time tomorrow 2011-03-21T07:00:38 aichallenge: sigh epsilon * r718559b / ants/battle_sim.py : Allow | to seperate rows in battle_sim script - http://bit.ly/em6DSn 2011-03-21T07:05:24 *** puru has joined #aichallenge 2011-03-21T07:15:10 *** olexs has joined #aichallenge 2011-03-21T07:27:32 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-03-21T07:32:43 *** a_Tick has quit IRC (Read error: Connection reset by peer) 2011-03-21T08:03:03 *** puru has quit IRC (Ping timeout: 252 seconds) 2011-03-21T08:15:14 *** needsch has joined #aichallenge 2011-03-21T08:29:43 *** jmpespxo1eax is now known as jmpespxoreax 2011-03-21T08:29:48 *** jmpespxoreax has joined #aichallenge 2011-03-21T08:37:23 *** Stocha has quit IRC (Quit: Page closed) 2011-03-21T08:59:02 *** sigh has quit IRC (Remote host closed the connection) 2011-03-21T09:18:32 *** mceier has quit IRC (Quit: leaving) 2011-03-21T09:24:39 *** UncleVasya has joined #aichallenge 2011-03-21T09:35:05 *** dotirc has joined #aichallenge 2011-03-21T09:37:52 *** bss03 has quit IRC (Remote host closed the connection) 2011-03-21T09:38:43 *** dotirc has quit IRC (Quit: .IRC: .NET IRC Client @ http://www.dotirc.com/) 2011-03-21T09:53:40 *** kaemo has joined #aichallenge 2011-03-21T09:58:56 *** olexs has quit IRC (Quit: Leaving.) 2011-03-21T09:59:21 *** needsch has quit IRC (Quit: Leaving.) 2011-03-21T10:22:08 done 2011-03-21T10:22:14 that was one rank assignment 2011-03-21T10:27:42 *** mrnitrate has joined #aichallenge 2011-03-21T10:34:39 hi 2011-03-21T10:37:32 *** kaemo has quit IRC (Remote host closed the connection) 2011-03-21T10:38:01 *** FireFly has joined #aichallenge 2011-03-21T10:41:16 *** JamesMG has joined #aichallenge 2011-03-21T10:41:54 *** thagomizr has joined #aichallenge 2011-03-21T10:43:38 *** tiresias has joined #aichallenge 2011-03-21T10:47:44 *** tiresias has joined #aichallenge 2011-03-21T10:50:50 *** tiresias has quit IRC (Client Quit) 2011-03-21T10:52:01 *** thagomizr has quit IRC (Ping timeout: 248 seconds) 2011-03-21T10:53:50 *** mechanical_turd has joined #aichallenge 2011-03-21T10:58:38 *** FireFly has quit IRC (Quit: swatted to death) 2011-03-21T10:59:06 *** robbs has joined #aichallenge 2011-03-21T10:59:26 *** FireFly has joined #aichallenge 2011-03-21T11:11:06 *** superflit has joined #aichallenge 2011-03-21T11:13:56 *** helo_ has joined #aichallenge 2011-03-21T11:21:55 *** helo_ has quit IRC (Quit: Page closed) 2011-03-21T11:40:42 *** mechanical_turd has left #aichallenge 2011-03-21T11:41:55 *** Kingpin13 has joined #aichallenge 2011-03-21T11:42:47 *** zvitamiinz has joined #aichallenge 2011-03-21T11:48:32 *** metik has joined #aichallenge 2011-03-21T11:52:45 *** amstan__ has joined #aichallenge 2011-03-21T11:52:45 *** ChanServ sets mode: +o amstan__ 2011-03-21T11:52:54 *** metik has quit IRC (Client Quit) 2011-03-21T11:59:16 *** needsch has joined #aichallenge 2011-03-21T11:59:59 *** dwchandler has joined #aichallenge 2011-03-21T12:01:06 *** amstan__ is now known as amstan 2011-03-21T12:06:32 *** Stocha has joined #aichallenge 2011-03-21T12:07:27 *** Stocha has quit IRC (Client Quit) 2011-03-21T12:15:17 *** kaemo has joined #aichallenge 2011-03-21T12:15:44 *** zvitamiinz has quit IRC (Ping timeout: 252 seconds) 2011-03-21T12:16:39 *** delt0r_ has quit IRC (Read error: Operation timed out) 2011-03-21T12:17:56 *** delt0r_ has joined #aichallenge 2011-03-21T12:26:58 *** Gorgoroth has joined #aichallenge 2011-03-21T12:28:34 *** boegel has quit IRC (Ping timeout: 264 seconds) 2011-03-21T12:36:11 *** Eruonen has quit IRC () 2011-03-21T12:40:24 antimatroid: thanks for taking care of reddit 2011-03-21T12:48:00 *** slkjcna has joined #aichallenge 2011-03-21T12:48:43 contestbot: later tell Eruonen http://www.reddit.com/r/programming/comments/g7qtj/google_ai_challenge_2011_ants/c1lnp9i 2011-03-21T12:48:43 amstan: I come to serve. 2011-03-21T12:50:49 Hello chaps! It seems like you're running a pretty interesting challenge... Do you think it's plausible that an okay programmer with no previous AI experience but plenty of time could take part convincingly? 2011-03-21T12:51:23 slkjcna: you should do fine 2011-03-21T12:51:32 you won't win though 2011-03-21T12:51:53 only 1 can win 2011-03-21T12:52:07 well.. i won rps with no previous AI experience before, but there were 120 contestants at the time 2011-03-21T12:52:49 caution: I didn't expect to. 2011-03-21T12:52:59 I was joking 2011-03-21T12:53:29 Don't worry about it 2011-03-21T12:55:11 amstan: Thanks! Are there any useful resources beyond the wiki/github/forum/IRC? 2011-03-21T12:55:30 slkjcna: nope.. that's pretty much it, lol 2011-03-21T12:56:29 No "beginner's guide to AI"? 2011-03-21T12:57:58 hi guys, i've been thinking a minor tweak to the core gameplay, how about bringing food back to the nest before gaining another ant ? 2011-03-21T12:58:44 *** Fibrewire|Work has joined #aichallenge 2011-03-21T13:00:17 slkjcna: there are a LOT of beginner's guides to AI on the net 2011-03-21T13:00:40 and a lot of intermediate material, and ... 2011-03-21T13:00:47 dwchandler: So I gathered, but are all of them relevant? 2011-03-21T13:00:55 no 2011-03-21T13:01:06 Oh, dear. 2011-03-21T13:01:14 the tricky part is figuring out *which* bits are relevant ;-) 2011-03-21T13:01:37 keep reading the forums as things get going and people will talk about approaches 2011-03-21T13:01:45 I'd better get to it, then. Thanks all. 2011-03-21T13:01:46 then you can go search 2011-03-21T13:01:54 slkjcna: my entry for previous contest was my first real program. Before this were only the simple tasks in my technical school (like find average of 1-metric array of integers or find sum of elements on main diagonal of an array). I downloaded starter package for contest and there were sources of few sample bots with different tactics. 2011-03-21T13:02:41 UncleVasya: That's encouraging. Great! 2011-03-21T13:02:47 I have no computer science or even good mathematician background. But string by string I implemented my ideas to the bot. 2011-03-21T13:03:15 Looks like there's hope yet... 2011-03-21T13:03:47 And it became smarter. So I didn't use even any of game theory elements, just scripting. 2011-03-21T13:03:48 did you rank in the top 50%? 2011-03-21T13:03:58 My place is 313. 2011-03-21T13:04:48 Gorgoroth: if you wanna play with the specs you have to nag antimatroid 2011-03-21T13:04:53 But it needed really many time and to the end of contest I was wery tired :) 2011-03-21T13:04:57 Gorgoroth: and mcleopold 2011-03-21T13:05:19 *a lot of time 2011-03-21T13:05:27 UncleVasya: Good thing I've got plenty of time on my hands. 2011-03-21T13:06:22 *** Fibrewire|Work has left #aichallenge 2011-03-21T13:06:41 Gorgoroth: bringing food back was in the original spec but j3camero eliminated it 2011-03-21T13:06:46 But I think this time contest auditory will be much stronger and we'll need to implement game theory to stand even in top500. 2011-03-21T13:09:09 *** alphac has joined #aichallenge 2011-03-21T13:09:29 in reference to https://github.com/aichallenge/aichallenge/issues/32: i wrote a labview app that reads in labview source files, is there somewhere i can upload it to be tested on the server? 2011-03-21T13:10:42 CPeng: does it work on windows? 2011-03-21T13:11:04 yes 2011-03-21T13:11:15 CPeng: so... it won't work on linux? 2011-03-21T13:11:51 i am not sure. i dont have access to linux. but there is a runtime engine for linux so it could work yes? 2011-03-21T13:12:00 we also need a compiler for linux then 2011-03-21T13:12:04 anyway.. g2g 2011-03-21T13:12:08 put it on that issue 2011-03-21T13:12:13 k 2011-03-21T13:13:36 is there somewhere i can upload files on github? 2011-03-21T13:14:42 fork, commit and pull 2011-03-21T13:15:13 *** twinshadow has joined #aichallenge 2011-03-21T13:15:21 k ty 2011-03-21T13:16:14 *** amstan has quit IRC (Ping timeout: 246 seconds) 2011-03-21T13:19:49 *** alphac has left #aichallenge 2011-03-21T13:23:32 *** bhasker has joined #aichallenge 2011-03-21T13:28:07 *** mceier has joined #aichallenge 2011-03-21T13:30:49 *** Gorgoroth has quit IRC () 2011-03-21T13:33:15 *** |UncleVasya| <|UncleVasya|!kvirc@46.133.251.53> has joined #aichallenge 2011-03-21T13:33:42 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-03-21T13:34:44 *** amstan has joined #aichallenge 2011-03-21T13:34:44 *** ChanServ sets mode: +o amstan 2011-03-21T13:37:00 *** bhasker_ has joined #aichallenge 2011-03-21T13:37:46 *** bhasker has quit IRC (Ping timeout: 250 seconds) 2011-03-21T13:40:32 what is this labview that you speak of? 2011-03-21T13:44:12 *** Accoun has quit IRC () 2011-03-21T13:46:15 *** thagomizr has joined #aichallenge 2011-03-21T13:53:06 labview is mostly used for testing/manufacturing 2011-03-21T13:53:14 i love it as a general purpose language though 2011-03-21T13:54:42 its a closed source language....so it will never take off in a community like this 2011-03-21T13:54:43 I think python is used for this quite a bit actually 2011-03-21T13:54:57 there is a large community behind it 2011-03-21T13:55:12 woot, python 2011-03-21T13:56:18 what are you going to write your bot in computerwiz_222? 2011-03-21T13:57:45 *** n9thbit has joined #aichallenge 2011-03-21T14:00:23 *** amstan has quit IRC (Ping timeout: 276 seconds) 2011-03-21T14:06:32 *** Accoun has joined #aichallenge 2011-03-21T14:10:51 *** bhasker_ has quit IRC (Ping timeout: 246 seconds) 2011-03-21T14:17:03 i have some problem with the automatic line ending conversion. there are always some scala files showing up as modified even after a git reset --hard 2011-03-21T14:19:28 *** n9thbit has quit IRC (Quit: n9thbit) 2011-03-21T14:20:28 *** Eruonen has joined #aichallenge 2011-03-21T14:21:01 *** bhasker has joined #aichallenge 2011-03-21T14:22:28 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86.1 [Firefox 3.6.15/20110303024726]) 2011-03-21T14:36:16 *** boegel has joined #aichallenge 2011-03-21T14:38:22 *** melipone has joined #aichallenge 2011-03-21T14:53:36 *** thagomizr has quit IRC (Quit: thagomizr) 2011-03-21T14:54:09 *** mrnitrate has quit IRC (Ping timeout: 252 seconds) 2011-03-21T14:54:18 github proposes to always check-in files with LF line endings. i'll just convert the line endings and commit 2011-03-21T15:00:33 aichallenge: Marco Leise epsilon * r723e04e / (6 files in 4 dirs): set line-endings to LF - http://bit.ly/dOjv6f 2011-03-21T15:00:38 aichallenge: Marco Leise epsilon * r93f66b3 / (4 files): 2011-03-21T15:00:38 aichallenge: (1) escaped <, > and & in javascript error messages / (2) fixed error on showing 2011-03-21T15:00:38 aichallenge: error message for error while downloading replay / (3) playback of half turns 2011-03-21T15:04:51 *** n9thbit has joined #aichallenge 2011-03-21T15:09:11 *** thagomizr has joined #aichallenge 2011-03-21T15:10:34 *** thagomizr has quit IRC (Client Quit) 2011-03-21T15:11:13 *** fireworker76 has joined #aichallenge 2011-03-21T15:15:59 *** jstritar has joined #aichallenge 2011-03-21T15:16:07 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-03-21T15:17:09 Are starter packages available for ants? 2011-03-21T15:17:15 *** jstritar has left #aichallenge ("Leaving") 2011-03-21T15:18:55 *** fireworker76 has quit IRC (Quit: Page closed) 2011-03-21T15:20:22 melipone: there are some on the github source, but the rules are still changing quite a bit 2011-03-21T15:20:34 i would recommend people wait a while before seriously trying to start on a bot 2011-03-21T15:22:01 *** nivardus has joined #aichallenge 2011-03-21T15:25:27 *** _aj <_aj!~aj@66.133.136.43> has joined #aichallenge 2011-03-21T15:25:30 antimatroid: as opposed to playing around with a bot :) 2011-03-21T15:25:56 is "Winter 2011" still the timeframe? 2011-03-21T15:27:18 Spring seems more accurate, probably a month or so away still 2011-03-21T15:27:43 but I've seen some saying only a week away, I don't see how though 2011-03-21T15:28:02 well, depends where you are in the world :P 2011-03-21T15:28:09 it's the backend of summer in australia 2011-03-21T15:28:21 a week is not going to happen 2011-03-21T15:28:29 norther hemi, spring then :) 2011-03-21T15:28:37 northern even 2011-03-21T15:28:54 meh, pesky northerners :P 2011-03-21T15:29:28 *** amstan has joined #aichallenge 2011-03-21T15:29:28 *** ChanServ sets mode: +o amstan 2011-03-21T15:30:19 *** n9thbit has quit IRC (Quit: n9thbit) 2011-03-21T15:34:10 bah. 2nd Quarter 2011 works globally 2011-03-21T15:44:39 *** fireworker76 has joined #aichallenge 2011-03-21T15:47:59 *** Bitbrit has joined #aichallenge 2011-03-21T15:53:32 *** stride has joined #aichallenge 2011-03-21T15:58:25 *** McLeopold1 has joined #aichallenge 2011-03-21T15:59:32 Hi all 2011-03-21T16:00:07 McLeopold1: hey 2011-03-21T16:00:29 how do i run my first game? 2011-03-21T16:01:08 *** num1 has joined #aichallenge 2011-03-21T16:04:16 can anyone help me start my first game? 2011-03-21T16:06:01 fireworker: what language are you writing your bot in? or do you just want to run a test game? 2011-03-21T16:06:06 *** antrn11 has joined #aichallenge 2011-03-21T16:07:32 *** robbs has quit IRC (Read error: Connection reset by peer) 2011-03-21T16:08:08 *** amstan has left #aichallenge ("Konversation terminated!") 2011-03-21T16:08:10 *** amstan has joined #aichallenge 2011-03-21T16:08:10 *** ChanServ sets mode: +o amstan 2011-03-21T16:08:28 fireworker76? 2011-03-21T16:08:53 *** robbs has joined #aichallenge 2011-03-21T16:09:21 *** bhasker_ has joined #aichallenge 2011-03-21T16:09:31 *** fireworker76 has quit IRC (Quit: Page closed) 2011-03-21T16:11:31 *** bhasker has quit IRC (Ping timeout: 252 seconds) 2011-03-21T16:17:05 *** Bitbrit has quit IRC (Quit: Bitbrit) 2011-03-21T16:19:46 jmcarthur: making haskell headway? 2011-03-21T16:22:37 *** niemela has joined #aichallenge 2011-03-21T16:23:11 *** phire has quit IRC (Read error: Connection reset by peer) 2011-03-21T16:29:25 *** dr`away is now known as dr- 2011-03-21T16:33:16 *** phiren has joined #aichallenge 2011-03-21T16:33:30 *** Nova has joined #aichallenge 2011-03-21T16:35:39 *** Nova has quit IRC (Client Quit) 2011-03-21T16:38:38 *** dr- has quit IRC (Remote host closed the connection) 2011-03-21T16:38:56 *** dr- has joined #aichallenge 2011-03-21T16:42:44 *** FireFly has quit IRC (Quit: swatted to death) 2011-03-21T16:43:56 quit 2011-03-21T16:45:07 *** melipone has quit IRC (Quit: go home) 2011-03-21T16:46:08 *** olexs has joined #aichallenge 2011-03-21T16:48:17 *** sigh has joined #aichallenge 2011-03-21T16:54:45 *** jhuynh85 has joined #aichallenge 2011-03-21T16:55:17 *** phirenz has joined #aichallenge 2011-03-21T16:57:43 *** phiren has quit IRC (Ping timeout: 252 seconds) 2011-03-21T17:00:45 *** sigh has quit IRC (Remote host closed the connection) 2011-03-21T17:03:36 *** thagomizr has joined #aichallenge 2011-03-21T17:11:54 *** thagomizr has quit IRC (Ping timeout: 260 seconds) 2011-03-21T17:12:28 *** boegel has quit IRC (Quit: Leaving) 2011-03-21T17:14:51 *** rstrong has joined #aichallenge 2011-03-21T17:15:27 *** odinsbane has joined #aichallenge 2011-03-21T17:17:43 *** odinsbane has quit IRC (Quit: leaving) 2011-03-21T17:21:06 *** robbs has quit IRC (Read error: Connection reset by peer) 2011-03-21T17:24:30 *** bhasker has joined #aichallenge 2011-03-21T17:27:24 *** bhasker_ has quit IRC (Ping timeout: 250 seconds) 2011-03-21T17:28:36 *** fireworker76 has joined #aichallenge 2011-03-21T17:33:36 *** thagomizr has joined #aichallenge 2011-03-21T17:34:17 *** UncleVasya has joined #aichallenge 2011-03-21T17:34:47 *** UncleVasya has quit IRC (Client Quit) 2011-03-21T17:36:59 *** |UncleVasya| <|UncleVasya|!kvirc@46.133.251.53> has quit IRC (Ping timeout: 260 seconds) 2011-03-21T17:42:33 *** dlila has joined #aichallenge 2011-03-21T17:43:26 *** ltriant has joined #aichallenge 2011-03-21T17:46:08 *** fireworker76 has quit IRC (Ping timeout: 252 seconds) 2011-03-21T17:59:44 *** jeff____ has joined #aichallenge 2011-03-21T18:01:30 *** anders__ has joined #aichallenge 2011-03-21T18:03:08 Hello, I wonder if I would like to set up a new contest amongst friends for the Planet Wars game, is it possible to do so using the stuff at github (server, sandbox, starter kits and so on) ? 2011-03-21T18:03:37 *** thagomizr has quit IRC (Read error: Connection reset by peer) 2011-03-21T18:03:38 it seems the github project is meant for the ant game 2011-03-21T18:03:43 *** thagomizr has joined #aichallenge 2011-03-21T18:04:52 *** bhasker has quit IRC (Ping timeout: 255 seconds) 2011-03-21T18:05:48 anders__: yes. The github project is mostly for the upcoming contest. What is it that you are looking for, perhaps I can point you the right way? 2011-03-21T18:06:05 *** ltriant has quit IRC (Quit: Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/) 2011-03-21T18:06:13 We'd like to play the Planet Wars game 2011-03-21T18:07:00 and have people with different programming skills (read Java vs .NET guys) try to beat eachother in their favourite programming language 2011-03-21T18:07:30 Planet wars with the relation to Galcon reference seems more fun than the ant thing 2011-03-21T18:08:03 it seems the official ai-contest is closed so I thought we need to set up our own contest server somehow 2011-03-21T18:09:37 perhaps there is some other public contest server we could join ? 2011-03-21T18:10:20 I think some people have set up their own instances, but I don't think any are active right now. 2011-03-21T18:10:25 *** jeff____ is now known as j3camero 2011-03-21T18:10:41 *** sigh has joined #aichallenge 2011-03-21T18:10:53 It's pretty difficult to set up a Planet Wars instance, unfortunately. We are trying to make it much easier to do this time around. 2011-03-21T18:11:07 *** bhasker has joined #aichallenge 2011-03-21T18:11:10 ok 2011-03-21T18:11:12 If you have some experience setting up web servers and such, you will probably be able to figure it out. 2011-03-21T18:11:26 But perhaps all you want is the game engine itself. 2011-03-21T18:11:49 anders__: see http://www.benzedrine.cx/planetwars.html 2011-03-21T18:11:52 For example, if you and a group of trusted friends want to run your programs against each other, you could use the game engine to do that. 2011-03-21T18:11:57 I know Java stuff and SQL, but very litlle Python which seems to be all over the project 2011-03-21T18:12:09 *** ltriant has joined #aichallenge 2011-03-21T18:12:24 dwchandler: is the TCP server for planet wars still running? 2011-03-21T18:12:49 no, but the server code tar.gz is linked on that page 2011-03-21T18:13:01 I had it running on my server for a bit 2011-03-21T18:17:22 planet wars was hella fun. it might be worth reviving a tcp server if anyone cares 2011-03-21T18:17:28 amstan/McLeopold1: which one of you should I talk to about the directory structure of the ants/ directory? 2011-03-21T18:17:34 me 2011-03-21T18:18:14 McLeopold1: I made a script to check battle resolution (ants/battle_sim.py), where should it live? 2011-03-21T18:18:27 like a unit test? 2011-03-21T18:18:43 no, it's more for helping people decide between the battle options :P 2011-03-21T18:19:30 does it run test games? 2011-03-21T18:19:30 it's just a tool, it takes an initial positon on STDIN and simulates a battle on it under each attack option 2011-03-21T18:19:59 no, not really 2011-03-21T18:20:28 the main directory for now 2011-03-21T18:20:44 ok, no worries 2011-03-21T18:21:01 if we get unit tests, maybe a test directory would be good 2011-03-21T18:21:35 yeah, this can act as a unit test for the attack options 2011-03-21T18:22:07 looking at the dir, that "map.py" file ought to go 2011-03-21T18:22:11 dwchandler: so AIs that we develop out of starter kits could connect directly to the server engine found in the source code? 2011-03-21T18:22:55 anders__: do you care if the server compiles, or do you just want a quick and dirty way to get a leaderboard? 2011-03-21T18:24:30 well I guess we would like to both have an "arena" where the battles happen and a leaderboard yes 2011-03-21T18:25:22 a tcp server would allow bots to connect and play, but it wouldn't schedule the games, and the bots would have to run on their own machines 2011-03-21T18:25:38 If you're okay with that, then that would be easier to setup. 2011-03-21T18:25:39 anders__: the starter kits don't do TCP, so on that same page I linked there's a tcp.c app that passes the I/O to the server 2011-03-21T18:26:14 the tcp server had very basic Elo matching if there were choices available, iirc 2011-03-21T18:26:36 dwchandler: actually it had first 2 in get matched logic 2011-03-21T18:26:39 but any two free connections would get matched up 2011-03-21T18:26:42 ah ok, so that's what the "tournament" thing on github is taking care of ... the game schedule 2011-03-21T18:27:04 right, otherwise you just get some random matchups 2011-03-21T18:27:05 McLeopold1: I thought daniel added a bit of wait and Elo match 2011-03-21T18:27:26 there was elo match, but it only every allowed 1 in the queue :) 2011-03-21T18:28:03 it was fine for testing bots 2011-03-21T18:29:29 So McLeopold1, how are things? Did you see that the github project was on the Reddit front page yesterday? 2011-03-21T18:30:47 j3camero: yes I did. we have a to do list going on an etherpad, which was easier for me to do that github issues 2011-03-21T18:31:14 http://ietherpad.com/sMmBH9htlr 2011-03-21T18:31:24 Love the etherpad. 2011-03-21T18:31:30 I just noticed the LOLcode bot. 2011-03-21T18:31:35 well, thanks for guidance, I will think a bit on what to do next... perhaps it fun enough to play random games againts eachother, I guess you will soon see who's master of universe (our universe that is... ) :) 2011-03-21T18:31:53 anders__: tcp.zeroviz.us, I'll get it running shortly for you, so you can play. 2011-03-21T18:31:59 McLeopold1: oh snap are we using chroot to jail the users? 2011-03-21T18:32:15 j3camero: janzert might do it... 2011-03-21T18:32:28 Oh nice nice. 2011-03-21T18:32:33 that's one of the big todo's left 2011-03-21T18:32:38 We overworked poor janzert last time. 2011-03-21T18:33:00 McL : oh how nice! 2011-03-21T18:33:03 Yes security has always been the toughest nut to crack. Once we get past that, things will start to fall into place very rapidly. 2011-03-21T18:33:04 The other one is the true skill rankings. I honestly think I'll have that working this week. 2011-03-21T18:33:22 sigh has taken over engine issues 2011-03-21T18:33:30 McLeopold1: wow great. That would be epic. 2011-03-21T18:33:49 *** thagomizr has quit IRC (Remote host closed the connection) 2011-03-21T18:34:16 *** thagomizr has joined #aichallenge 2011-03-21T18:34:18 *** raydenuni has left #aichallenge ("Leaving") 2011-03-21T18:34:25 McLeopold1: my only meta-advice at this point is to find and train as many deputies as you can. Things gets real really fast as launch approaches, and the workload can turn into 36 hours per day. But it sounds like you're really stepping up into a leadership role, which is great to see! 2011-03-21T18:35:17 brb in 5. Gotta go across the street and buy bananas. 2011-03-21T18:35:51 twinshadow: so, if you can find an implementation of lolcode, that can actually parse strings, then it could be a real bot! 2011-03-21T18:36:19 there's like 170 people watching the repo now 2011-03-21T18:36:20 ha ha ha 2011-03-21T18:36:37 * McLeopold1 need to comment code more 2011-03-21T18:37:01 McLeopold1: someone opened an issue because one of my comments was wrong... 2011-03-21T18:40:13 I saw. Looks like you're a bad coder. :D 2011-03-21T18:40:40 anders__: the tcp server is running and you can play the contest winner when you connect 2011-03-21T18:40:43 well, I already knew that :P 2011-03-21T18:41:06 McLeopold1: we'll start looking at setting up starter kits and then we'll try to connect to your server 2011-03-21T18:47:15 *** McLeopold1 has left #aichallenge 2011-03-21T18:48:20 McLeopold1: how easy do you think it's going to be to launch an end-to-end contest instance this time? Is there any script such that I can fire up a VM and run "wget some_script.sh; ./some_script.sh" and have a working contest instance? 2011-03-21T18:49:13 Not the first priority obviously, but it would be great if any one of those 170 people watching the repo could light up their own instance. I bet we would get lots more contributors. 2011-03-21T18:52:09 I don't think you can set up a working instance even doing it manually yet ;) 2011-03-21T18:52:12 i keep seeing new issues get opened that are comments on old issues... 2011-03-21T18:52:29 but there is a server_setup.py script that will configure as much of a main server as I think is possible so far 2011-03-21T18:53:04 and a worker_setup.py script that does the same for a worker, but is still basically trying to setup a planetwars worker 2011-03-21T18:58:55 *** caution has quit IRC (Quit: caution) 2011-03-21T19:00:23 *** thagomizr has quit IRC (Quit: thagomizr) 2011-03-21T19:01:52 woah, j3camero is back 2011-03-21T19:08:17 *** Eruonen has quit IRC () 2011-03-21T19:15:31 sigh: what issue? 2011-03-21T19:16:35 amstan: https://github.com/aichallenge/aichallenge/issues/closed#issue/33 2011-03-21T19:16:55 lol 2011-03-21T19:16:55 sigh: lol 2011-03-21T19:17:07 yeah... 2011-03-21T19:17:14 from comments on reddit, it would appear option 1 is out 2011-03-21T19:17:21 antimatroid: why? lol 2011-03-21T19:17:26 people just don't get it 2011-03-21T19:17:40 well.. i don't get it either actually.. 2011-03-21T19:17:44 but we have a more intuitive close relative, so i'm on board for it 2011-03-21T19:17:52 amstan: it's quite simple once you understand it 2011-03-21T19:17:57 haha 2011-03-21T19:18:06 antimatroid: tip: stop calling it option 1 2011-03-21T19:18:09 it's not semantic at all 2011-03-21T19:18:24 power option then 2011-03-21T19:18:36 amstan: your power = 1/#enemiesinrange 2011-03-21T19:18:48 seems rather weird 2011-03-21T19:18:54 you live if you have power > power for all enemies in range 2011-03-21T19:19:20 i find this hilarious "j3camero: My mom submitted an entry in the last contest. I kid you not. You will be fine, trust me. :-)" 2011-03-21T19:19:26 antimatroid: what's the range? 2011-03-21T19:19:33 attackradius 2011-03-21T19:19:33 atm sqrt(5) 2011-03-21T19:19:38 6 2011-03-21T19:19:45 6? 2011-03-21T19:19:48 antimatroid: so.. what about the other options? 2011-03-21T19:19:52 oh, sqrt(6) 2011-03-21T19:20:31 sighs new method i like most now: each ant hast 1 damage to pass out each turn, distrbuted evenly amongst its enemies (so 1/#enemiesinrange to each enemy), then an ant with >=1 damage dies 2011-03-21T19:21:18 the other main option is the "iteratively remove the enemy ants that are closest" which doesn't really offer much in the way of cooperation and has weird situations like a..b..a where both a's die with the b 2011-03-21T19:21:34 with sigh's option, a..b..a and a..b..c (cooperation!) kill just b 2011-03-21T19:22:05 i believe the cooperation i speak of is a result of the battle option reflecting the number of enemy ants your enemies in range have 2011-03-21T19:22:10 this should be written up somewhere, with examples 2011-03-21T19:22:23 sigh: i'll do that today 2011-03-21T19:23:04 antimatroid: I suggest creating a new wiki page, and link the specs page to that... no point listing options on the specs page until it is finalised 2011-03-21T19:23:13 yeah, that's the plan 2011-03-21T19:23:29 i suspect we'll need to break the specs page up like that anyway 2011-03-21T19:23:34 it's going to be huge otherwise 2011-03-21T19:23:37 antimatroid: also, see the comments in the code for reasonable descriptions of the current options 2011-03-21T19:23:43 :P 2011-03-21T19:23:57 don't raise any issues if they are wrong :) 2011-03-21T19:24:06 but... 2011-03-21T19:24:25 :p 2011-03-21T19:24:27 *** thagomizr has joined #aichallenge 2011-03-21T19:24:36 the number of lurkers doubled in a day 2011-03-21T19:24:41 antimatroid: nah, the final specs should be on the one page 2011-03-21T19:25:04 sigh: i was thinking maybe briefer explanations then a link to more detailed? 2011-03-21T19:25:08 if anything, you could link elsewhere for more detail (e.g. examples of gaem replays, battle resolutions or whatever) 2011-03-21T19:25:10 yeah 2011-03-21T19:25:19 so then the battle resolution can have example images, same with map files etc. 2011-03-21T19:25:41 sure, but I think you should be able to go to one page for a complete specs 2011-03-21T19:25:55 other pages are just for examples and so on 2011-03-21T19:26:41 what options am i considering? 2011-03-21T19:26:44 there's like 4 now? 2011-03-21T19:26:54 should i throw out power and friends/enemies? 2011-03-21T19:26:54 there's four in the code, yeah 2011-03-21T19:27:09 i think it's really out of option 2 and 4 2011-03-21T19:27:10 you decide 2011-03-21T19:27:34 it's easy enough to generate examples for all four with the battle_sim.py script 2011-03-21T19:30:03 *** nicklovescode has joined #aichallenge 2011-03-21T19:30:16 hey guys, new here 2011-03-21T19:30:37 interested in the ants challenge, is it ready for me to start developing and testing a bot or should I wait some more 2011-03-21T19:30:39 (python) 2011-03-21T19:31:05 nicklovescode: i would suggest you wait a few weeks 2011-03-21T19:31:12 you will have plenty of time for programming i promise 2011-03-21T19:31:36 (the contest will go for ~ 2 months i think) 2011-03-21T19:31:42 ah, ok cool 2011-03-21T19:31:59 nicklovescode: you could start.. but there's no guarantee that things might not change 2011-03-21T19:32:14 nicklovescode: so if you do start, and things do change, don't get mad 2011-03-21T19:32:27 nicklovescode: also.. you won't be able to submit anything since the server won't be up 2011-03-21T19:32:35 aka start thinking about strategies, but coding could be frustrating 2011-03-21T19:32:49 antimatroid: well.. no, we already have starter packages 2011-03-21T19:32:52 but is there anyway to see how my bot compares to others? 2011-03-21T19:33:05 nicklovescode: at this point there is no other bots, afaik 2011-03-21T19:33:14 amstan: sure, but dependent on the rules chosen, how you structure it beyond that might become annoying/not the best way 2011-03-21T19:33:14 nicklovescode: just the sample ones 2011-03-21T19:33:20 antimatroid: yeah 2011-03-21T19:33:25 amstna: there is hunterbot 2011-03-21T19:33:31 antimatroid: those has to be finalized soon 2011-03-21T19:33:37 nicklovescode: try playing agains the python hunterbot 2011-03-21T19:33:40 antimatroid: we can't leave them hanging till just before launching 2011-03-21T19:33:54 I have it git cloned and I'm in that directory, how do I get it running? 2011-03-21T19:34:07 antimatroid: how's the wiki btw? How finalized are things in there? 2011-03-21T19:34:09 nicklovescode: what os? 2011-03-21T19:34:13 OSX 2011-03-21T19:34:18 sigh: ^^? 2011-03-21T19:34:26 10.6 2011-03-21T19:34:27 amstan: i'll give stuff a clean today, i haven't touched it in a while 2011-03-21T19:34:43 been busy the last little bit, but handed the last of my current assignments in this morning 2011-03-21T19:34:48 antimatroid: i would like to start going over that stuff, and integrate it with the website 2011-03-21T19:35:04 get back to me in a day? 2011-03-21T19:35:10 antimatroid: ok 2011-03-21T19:35:21 not 24 hours, i'll be at uni then :P 2011-03-21T19:35:30 *** Stocha has joined #aichallenge 2011-03-21T19:36:54 hi dimkadimon you will find some link and data to the experiment you requested (in the exotic rule thread). As i now have a server (at least for a month) i got more flexibility. All the stuff is still very rough to say the least. 2011-03-21T19:37:53 Stocha: dimkadimon doesn't come on irc 2011-03-21T19:37:56 nicklovescode: if you look at ants/play_one_game.sh you can see how to run a game 2011-03-21T19:37:59 at least not with his normal username 2011-03-21T19:38:22 nicklovescode: bear in mind getting stuff running should be a lot simpler in a couple of weeks 2011-03-21T19:38:36 so if you can't get it running now, don't give up and assume you wont be able to when we launch officially 2011-03-21T19:39:55 either way, running stuff on osx should be the same as linux for this stuff, so anyone should be able to help you 2011-03-21T19:40:19 *** Heulsay has joined #aichallenge 2011-03-21T19:40:28 got it working, awesome! 2011-03-21T19:40:33 :) 2011-03-21T19:40:37 I still have issue on safari with your visualizer sigh. I simply keep freshing until things settles right. 2011-03-21T19:40:58 Stocha: Frontier is the visualiser guy 2011-03-21T19:40:59 So you might write some message to help people do that on top of the page :p 2011-03-21T19:41:04 ah Frontier then :p 2011-03-21T19:41:12 i got name shuffled a bit :p 2011-03-21T19:41:31 i'm quite tired also :p 2011-03-21T19:43:12 contestbot: seen mcleopold 2011-03-21T19:43:12 amstan: mcleopold was last seen in #aichallenge 20 hours, 38 minutes, and 31 seconds ago: okay, cya 2011-03-21T19:43:19 contestbot: seen mcleopold1 2011-03-21T19:43:19 amstan: mcleopold1 was last seen in #aichallenge 1 hour, 2 minutes, and 39 seconds ago: anders__: the tcp server is running and you can play the contest winner when you connect 2011-03-21T19:44:23 *** McLeopold1 has joined #aichallenge 2011-03-21T19:44:29 McLeopold1: ping 2011-03-21T19:44:38 what's up? 2011-03-21T19:44:51 McLeopold1: i would like to separate the sample bots from the starter packages.. 2011-03-21T19:45:13 McLeopold1: you ok with that? 2011-03-21T19:46:41 I'd prefer a smart makefile, but yes, I'm okay with that. 2011-03-21T19:46:52 McLeopold1: i'll be doing that too 2011-03-21T19:47:13 the one thing i want to solve is people that forked to make starter packages.. 2011-03-21T19:47:23 idk if this will work 2011-03-21T19:47:27 *** Kingpin13 has quit IRC (Quit: You're a kitty!) 2011-03-21T19:49:22 having each starter bot in its own separate repo seems like it would be better 2011-03-21T19:49:30 although maybe too late already? 2011-03-21T19:51:22 For each ant \(a\) with \(n > 0 \) enemies within the attack radius, define their power to be \(power(a) = \frac{1}{n}\). An ant lives if it has no enemies or a greater power than all its enemies within the attack radius. 2011-03-21T19:51:40 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-03-21T19:51:40 is that style alright, or do people hate variables? 2011-03-21T19:52:33 janzert: that's too late 2011-03-21T19:52:44 janzert: but that's not an issue, it's just that the path is not semantic at all 2011-03-21T19:52:53 oh well.. i can deal with this actually 2011-03-21T19:53:00 *** bhasker has quit IRC (Quit: bhasker) 2011-03-21T19:53:15 if i mess it up, i'll clean it 2011-03-21T19:53:32 amstan: it's not too late 2011-03-21T19:53:46 McLeopold1: it is, we have a ton of people with forks 2011-03-21T19:53:47 consider all the stuff in ants/bots to be example bots 2011-03-21T19:54:03 McLeopold1: most of those forks are for starter packages 2011-03-21T19:54:32 we control the starter packages, so we just contact who we need to 2011-03-21T19:54:41 with conflicts option 2 doesn't really need to be an iterative thing 2011-03-21T19:54:49 closest option** 2011-03-21T19:55:25 antimatroid: the general style is fine with me, I'd get rid of the redundant "no enemies or" in the second sentence though 2011-03-21T19:55:44 janzert: it's not redundent :P 2011-03-21T19:55:56 I have set up two applets with my bots running on. On for dieOnSameSquare. and why for rule1fighting 2011-03-21T19:56:00 the first sentence already says you only do the test if enemies are in range 2011-03-21T19:56:13 So you can view them at home. Alos it's far from comfotable at the moment :p 2011-03-21T19:56:23 no, it says i only define power for those ants with ants in range 2011-03-21T19:56:24 and it make the second one harder to mentally parse with it there 2011-03-21T19:56:24 oh man.. sbcl is 53MB 2011-03-21T19:56:30 rough prototype stuff 2011-03-21T19:57:16 amstan: let's move bots to example_bots and create a new dir starter_bots 2011-03-21T19:57:18 it does make the second one harder to pass, but otherwise it's not perfectly clear :\ 2011-03-21T19:57:28 here is the first url http://ns368394.ovh.net/stocha/Rule1Fight/ 2011-03-21T19:57:45 McLeopold1: they're not example bots though, they're starter bots 2011-03-21T19:57:54 i couldn't work out how better to phrase it without leaving something potentially unclear (even though it's pretty obvious that an ant with no enemies in range wouldn't die) 2011-03-21T19:57:59 amstan: I can move out my python and java starters to new repos that can then be submodules 2011-03-21T19:58:14 here is the other one : http://ns368394.ovh.net/stocha/DieOnSameSquare/ 2011-03-21T19:58:16 i want to touch my c++ bot again before that's sent off 2011-03-21T19:58:16 amstan: they are what every we call them :) 2011-03-21T19:58:19 when do i need to do that by? 2011-03-21T19:58:30 McLeopold1: i don't want to disturb the workflow of about 6 outside commiters 2011-03-21T19:58:37 The bots ranking is kinda reverse between the two rules :p 2011-03-21T19:59:26 Stocha: that actually doesn't surprise me that much when i think about it 2011-03-21T19:59:33 here the result for fighting1 : http://ns368394.ovh.net/stocha/Rule1Fight/result.txt 2011-03-21T19:59:35 consider the a..b..a sitation yet again 2011-03-21T19:59:39 amstan: It's a simple request. Please create a brand new repo if you are in charge of a starter bot, we will link it in. 2011-03-21T19:59:49 the other result is P0 winning most of the games. 2011-03-21T20:00:05 and P4 loosing most of them. Although the file is not accessible at the moment 2011-03-21T20:00:17 i imagine it's a matter of the difference between wanting to send more/less ants into a fight for the two options and reversing who gets points there, not necessarily meaning the game is going to cause the worse bot to lose under a certain rule 2011-03-21T20:00:36 so I would make it "For each ant \(a\) with \(n > 0\) enemies within the attack radius: define their power to be \(power(a) = \frac{1}{n}\). The ant lives if it has a greater power than all its enemies within the attack radius." 2011-03-21T20:00:49 McLeopold1: how do you envision this then? a bunch of smaller repos, one per language? 2011-03-21T20:01:12 janzert: you haven't specified what power is for an ant with no enemies 2011-03-21T20:01:16 i can define that to be 1? 2011-03-21T20:01:17 yes, that's what I was thinking? you were concerned about forks of the whole thing just to make a starter? 2011-03-21T20:01:18 then it works 2011-03-21T20:01:19 I only count the number of ants also. But the fighting rule is enougth to make quite a few difference : if you stick together you have a great advantage with Fighting1 (at least in my implementation wich may be wrong) 2011-03-21T20:01:33 McLeopold1: oh, no.. forking is not the problem 2011-03-21T20:01:43 McLeopold1: it's me moving stuff right now, when there's forks in progress 2011-03-21T20:01:52 oh wait, nevermind 2011-03-21T20:01:55 okay i'll do that 2011-03-21T20:01:57 but don't they just pull and it all works? 2011-03-21T20:02:09 * antimatroid had a logic fail 2011-03-21T20:02:24 antimatroid: this is still english, people should be able to know that the power is the same as just defined ;) 2011-03-21T20:02:25 no, because git doesn't track folders, see this: http://pastebin.com/kdt2zQ5P 2011-03-21T20:02:36 I should probably put on your new rules in the same settings. Would be consistent with Rule1 in this particular case i guess. 2011-03-21T20:02:39 even if a compiler couldn't 2011-03-21T20:02:57 janzert: give me an explanation for the closest option :P 2011-03-21T20:03:17 amstan: ok, so then we create an example_bots folder and move stuff over there, but keep bots as the starters? 2011-03-21T20:03:19 all i have is "iteratively remove the ants that are currently closest together until no enemy ants remain within the attack radius of each other" 2011-03-21T20:03:48 I've never thought of or seen one yet :P 2011-03-21T20:03:51 probably also need to state closest is using the euclidrean metric 2011-03-21T20:04:25 McLeopold1: no, i'll just git mv bots starter_bots, then extract the sample bots in python and java and put them somewhere else 2011-03-21T20:04:35 yeah, I'd do that outside of the battle resolution are actually (i.e. "All distances are calculated using the eucilidean distance") 2011-03-21T20:04:51 McLeopold1: it should be fine.. i was just worrying about merging later, but it's all good 2011-03-21T20:05:04 antimatroid: actually there is already a section for that 2011-03-21T20:05:13 *** Stocha has quit IRC (Quit: Page closed) 2011-03-21T20:05:17 janzert: on the specs page? 2011-03-21T20:05:21 yep 2011-03-21T20:05:31 i am writing a new one outlining the 4 current options, what strategies they encourage and giving examples 2011-03-21T20:05:37 between Bot Output and Fog of War 2011-03-21T20:05:41 so hopefully we can try to knock it down to 1 2011-03-21T20:05:49 janzert: i know, i wrote that page :P 2011-03-21T20:06:15 right, thought maybe you forgot about that detail 2011-03-21T20:06:49 anyway it's already specified as all distances being eucilidian so I suppose you could repeat it if you want but isn't completely necessary 2011-03-21T20:07:19 *** kaemo has quit IRC (Ping timeout: 250 seconds) 2011-03-21T20:07:46 McLeopold1: what bout the java sample-bots, do we need those? you have them coded in python too 2011-03-21T20:07:50 McLeopold1: should we keep both? 2011-03-21T20:08:21 *** maa2011 has joined #aichallenge 2011-03-21T20:08:30 Yes, let's keep them. 2011-03-21T20:08:52 McLeopold1: i'm just thinking from the perspective of the user, is there anything they can do with them? 2011-03-21T20:08:58 just javac *.java? 2011-03-21T20:09:25 *** olexs has left #aichallenge 2011-03-21T20:09:46 McLeopold1: what's the structure of the java stuff? what's common, what's starter pack only and what's sample only? 2011-03-21T20:10:08 umm... 2011-03-21T20:10:31 McLeopold1: ls => Aim.java Ants.java Bot.java HunterBot.java Ilk.java LeftyBot.java MyBot.java RandomBot.java Tile.java 2011-03-21T20:11:09 aim, ilk, ants, bot, and tile are shared 2011-03-21T20:11:13 mybot is the starter 2011-03-21T20:11:18 k, cool 2011-03-21T20:11:26 randombot, leftybot and hunterbot are the examples 2011-03-21T20:14:55 antimatroid: whoa, we have 4 options now? 2011-03-21T20:15:28 McLeopold1: yeah i know 2011-03-21T20:15:29 https://github.com/aichallenge/aichallenge/wiki/Ants-Battle-Resolution-Options 2011-03-21T20:15:44 please rip those explanations to pieces for me and give me more intuitive ones 2011-03-21T20:15:51 janzert^^ 2011-03-21T20:16:14 ergh, assuming i managed to finish the line for closest option :P 2011-03-21T20:16:31 don't use math 2011-03-21T20:16:48 heh, also you're missing the definition of power in the power option 2011-03-21T20:17:31 so i am, ergh sorry, i tried to add definitions to the top and must have removed it :P 2011-03-21T20:17:39 aichallenge: Alexandru Stan starter-packaging * rc28cdad / (194 files in 49 dirs): Moved starter_bots in a folder, separated sample_bots to another folder. If you have a fork where you're making a starter package, please pull this - http://bit.ly/hphBE7 2011-03-21T20:17:41 is the power option the one where an ant spreads it's attack? 2011-03-21T20:17:58 McLeopold: there is no maths in there? 2011-03-21T20:18:02 other than basic arithmetic 2011-03-21T20:18:16 power option is what you were calling occupied I believe 2011-03-21T20:18:18 yeah, remove the basic math too 2011-03-21T20:18:28 *** delt0r___ has joined #aichallenge 2011-03-21T20:19:53 *** delt0r_ has quit IRC (Ping timeout: 255 seconds) 2011-03-21T20:20:07 McLeopold: i'm stripping out all the variables 2011-03-21T20:20:46 Support Option: An ant dies if there are more enemies in its attack radius then friendlies. An ant counts as it's own friendly, so it always has at least 1. If the ant collides with an enemy ant, it dies. 2011-03-21T20:21:03 for damage option I'd probably just say something like "An ant distributes 1 point of damage evenly among all the enemies within its attack radius. An ant then lives if it receives less than 1 full point of damage in that turn." 2011-03-21T20:21:21 * McLeopold1 nods 2011-03-21T20:22:09 antimatroid: with the support option, we will need to detect ants crossing paths, and kill them 2011-03-21T20:23:00 maybe at the top say that ants always die if they collide with another ant and leave it out of any of the specific options 2011-03-21T20:23:02 but, where do the die? how do you represent the 'd' lines? 2011-03-21T20:23:32 try again now.. 2011-03-21T20:23:52 McLeopold1: https://gist.github.com/880519 2011-03-21T20:23:54 with n > 0 is not needed. 2011-03-21T20:23:54 McLeopold1: how's that? 2011-03-21T20:23:58 McLeopold1: that's a reason I don't like any option that has the potential for ants to "cross paths" 2011-03-21T20:23:59 it is 2011-03-21T20:24:11 otherwise people might try to calculate the power of ants without enemies 2011-03-21T20:24:16 which is /0 error 2011-03-21T20:24:51 McLeopold1: you can send more than 1 d line for a square, you already are for colliding friends right? 2011-03-21T20:24:57 I'd prefer it just said "distributes damage evenly" in the damage option 2011-03-21T20:24:58 amstan: the tools will be a separate download, so no 2011-03-21T20:25:06 it's a pain to process if you don't care about that but do about dead ants from battle 2011-03-21T20:25:10 *** JamesMG has quit IRC (Ping timeout: 250 seconds) 2011-03-21T20:25:18 McLeopold1: well then sample bots should also be included in that tools download 2011-03-21T20:25:24 hence why i asked you like 2 weeks ago to make the line input parameter different for dead ants from collision and battle :P 2011-03-21T20:25:25 antimatroid: yeah, but say "if there are enemies in range", take out the math 2011-03-21T20:25:46 amstan: yes, they should 2011-03-21T20:26:03 and don't we call the example? not sample? 2011-03-21T20:26:13 *** JamesMG has joined #aichallenge 2011-03-21T20:26:43 *** Heulsay has left #aichallenge 2011-03-21T20:27:40 i need n to define power? 2011-03-21T20:27:48 how would i do that option without any maths? 2011-03-21T20:27:58 *** bhasker has joined #aichallenge 2011-03-21T20:28:15 use english? 2011-03-21T20:28:30 power is one divided by the number of enemies in range? 2011-03-21T20:28:34 :p 2011-03-21T20:28:51 or.. i could just throw option 1 out now? 2011-03-21T20:28:58 *** maa2011 has left #aichallenge ("Ex-Chat") 2011-03-21T20:29:10 i think damage is close enough to option 1 while being simpler 2011-03-21T20:29:46 here's another little expansion/clean up for damage "An ant with enemies in range distributes 1 point of damage split evenly among those enemies. An ant then lives it it receives less than 1 full point of damage in a turn." 2011-03-21T20:30:13 *** mceier has quit IRC (Quit: leaving) 2011-03-21T20:30:49 janzert: okay 2011-03-21T20:31:00 is anyone opposed to me leaving out the power option? 2011-03-21T20:31:11 do power and damage diverge in any situations? 2011-03-21T20:31:21 yes 2011-03-21T20:31:30 I think it probably needs to stay since it's talked about in the forum threads so much 2011-03-21T20:31:41 pastebin me the setup, and I'll figure it out 2011-03-21T20:31:49 you could add a note that it's not really being considered anymore though? 2011-03-21T20:32:04 what setup? 2011-03-21T20:32:10 i haven't actually done an example 2011-03-21T20:32:15 but they aren't equivalent 2011-03-21T20:32:16 the setup where power and damage have different outcomes 2011-03-21T20:33:17 hmmm, i'm trying to think of one 2011-03-21T20:34:19 they;re VERY similar 2011-03-21T20:35:17 jmcarthur and I did derive the power method from thinking about how ants would do damage... 2011-03-21T20:35:45 open problem people :P 2011-03-21T20:36:08 sigh: ^^ we want an example where damage and power options differ 2011-03-21T20:37:08 amstan: the scala files always mess me up in windows. when I pull, git says they are changed. now I can pull!? 2011-03-21T20:37:28 McLeopold1: pull from epsilon, they should be ok 2011-03-21T20:37:49 I'm in epsilon 2011-03-21T20:38:01 .aaa. 2011-03-21T20:38:01 bbbbb 2011-03-21T20:38:02 McLeopold: https://github.com/aichallenge/aichallenge/wiki/Ants-Battle-Resolution-Options 2011-03-21T20:38:13 I think that differs between power and damage 2011-03-21T20:38:27 i think you're right 2011-03-21T20:38:46 oh wait, no? 2011-03-21T20:38:55 all a's die in both, but power kills the middle b? 2011-03-21T20:39:08 oh yeah 2011-03-21T20:39:16 but each ant distributes 1/3 yeah? 2011-03-21T20:39:17 damage 2011-03-21T20:39:22 so one of the b's gets 1 damage 2011-03-21T20:39:24 killing it 2011-03-21T20:39:51 hmm only 3 b's are in range? then I guess not 2011-03-21T20:40:04 oh, wait yeah 2011-03-21T20:40:17 but then no b's die with power either? 2011-03-21T20:40:39 oh yeah, for sqrt(6) you may be right 2011-03-21T20:40:58 I just was going by your first diagram in the forums, but it may be using a smaller radius? 2011-03-21T20:41:12 amstan: i do "git checkout Bot.scala", then "git status" and it still says it's modified? 2011-03-21T20:41:22 McLeopold1: git reset --hard 2011-03-21T20:41:25 try that 2011-03-21T20:41:36 no! you totally screwed me last time 2011-03-21T20:41:58 well.. yeah.. assuming you don't have any uncommited files 2011-03-21T20:41:59 janzert: i'm using sqrt(4) there i think 2011-03-21T20:42:00 didn't Frontier have problem with that earlier? 2011-03-21T20:42:01 it's smaller anyway 2011-03-21T20:42:10 hmm 2011-03-21T20:43:41 it's a newline at the end that's killing me 2011-03-21T20:44:08 pretty sure that's exactly the problem Frontier had 2011-03-21T20:44:31 I think he just gave up and checked it in with modified line endings 2011-03-21T20:44:57 https://github.com/aichallenge/aichallenge/commit/723e04ea2769d266b7f261e4ec05ba70140ec38e 2011-03-21T20:46:54 *** bhasker has quit IRC (Quit: bhasker) 2011-03-21T20:47:11 janzert: do you know why he tried merging that pull request? 2011-03-21T20:47:52 janzert: i'll keep doing the page and try to get an example when i start making examples with sigh's analyzer 2011-03-21T20:47:55 not sure what pull request you even mean, but no 2011-03-21T20:48:57 http://pastebin.com/M0HVMyDT 2011-03-21T20:49:17 so, according to that, it's not the newline, but it still thinks I modified 2 lines? 2011-03-21T20:49:19 you can see the archives around 2:00-3:00 this afternoon to see where Frontier had the problem 2011-03-21T20:50:03 i.e. about 6 hours ago 2011-03-21T20:51:04 okay, but I still have an issue and I can pull 2011-03-21T20:51:08 can't 2011-03-21T20:51:58 McLeopold1: do you have anything uncommited? 2011-03-21T20:55:31 amstan: as always, yes 2011-03-21T20:55:59 McLeopold1: try git reset --hard path/to/scala/folder/that/you/don't/care/about 2011-03-21T20:56:27 thing about git pull is that it won't let you pull if there's stuff that's uncommited that has to be edited 2011-03-21T20:57:06 aichallenge: Alexandru Stan starter-packaging * r12134cb / (195 files in 49 dirs): Moved starter_bots in a folder, separated sample_bots to another folder. Updated c submodule. If you have a fork where you're making a starter package, please pull this. - http://bit.ly/fJEMEF 2011-03-21T20:57:08 aichallenge: Alexandru Stan starter-packaging * r7cb4b12 / ants/dist/starter_bots/Makefile : Added starter_bot makefile to package the bots - http://bit.ly/gCFbxb 2011-03-21T21:01:12 amstan: reset --hard does not have a paths option 2011-03-21T21:01:18 McLeopold1: darn 2011-03-21T21:01:33 McLeopold1: what's the actual message you're getting when trying to pull? 2011-03-21T21:01:51 local changes would be overwritten 2011-03-21T21:02:00 isn't there a -f for pull 2011-03-21T21:02:03 so they are overwritten 2011-03-21T21:02:05 try it 2011-03-21T21:02:43 *** thagomizr has quit IRC (Quit: thagomizr) 2011-03-21T21:03:37 cant you just use fetch? 2011-03-21T21:08:50 *** kosiini has joined #aichallenge 2011-03-21T21:08:53 amstan: no luck 2011-03-21T21:09:09 McLeopold1: can you give me a git status pls? 2011-03-21T21:09:13 I can fetch, but then I'm still left with a merge to do 2011-03-21T21:10:52 http://pastebin.com/W9M6meC3 2011-03-21T21:11:25 (use "git checkout -- ..." to discard changes in working directory) 2011-03-21T21:11:32 try git checkout -- scala/ 2011-03-21T21:12:20 antimatroid: ping 2011-03-21T21:12:32 amstan: ? 2011-03-21T21:12:47 antimatroid: your readme from the cpp starter bot is not up to date, it talks about non existing files 2011-03-21T21:12:58 yeah sorry, i need to update the whole thing 2011-03-21T21:13:04 when do you want me to do that by? 2011-03-21T21:13:04 amstan: nope 2011-03-21T21:13:15 antimatroid: can i delete it for now? 2011-03-21T21:13:20 yeah 2011-03-21T21:15:28 https://github.com/aichallenge/aichallenge/wiki/Ants-Battle-Resolution-Options 2011-03-21T21:15:40 anyone welcome to provide me more properties to include there :P 2011-03-21T21:17:47 amstan: deleting the files locally seemed to work ? 2011-03-21T21:17:56 McLeopold1: ok, lol 2011-03-21T21:18:48 *** McLeopold1 has left #aichallenge 2011-03-21T21:19:35 aichallenge: Alexandru Stan starter-packaging * r4d39c74 / ants/dist/sample_bots/java/Makefile : added makefile for java sample bots - http://bit.ly/fLhxJO 2011-03-21T21:19:37 aichallenge: Alexandru Stan starter-packaging * r12f5282 / (2 files): added cpp makefile, deleted old readme - http://bit.ly/hSpmAw 2011-03-21T21:19:52 *** Todd200 has joined #aichallenge 2011-03-21T21:20:31 *** Todd200 has left #aichallenge 2011-03-21T21:20:57 *** deblap has joined #aichallenge 2011-03-21T21:24:01 *** needsch has quit IRC (Quit: Leaving.) 2011-03-21T21:29:20 aichallenge: Alexandru Stan starter-packaging * r580ca74 / (2 files): added makefile for java starter package, apparently it requires RandomBot.java as well - http://bit.ly/glXFL4 2011-03-21T21:30:53 amstan: want to make me a makefile? :P 2011-03-21T21:31:03 i can edit whatever you make, i just have no idea how 2011-03-21T21:31:14 ideally it would be able to set DEBUG in bug.h 2011-03-21T21:31:16 antimatroid: i already have it for cpp, see the second last commit 2011-03-21T21:31:21 ah, cheers :) 2011-03-21T21:31:22 antimatroid: oh! good idea 2011-03-21T21:37:16 *** nicklovescode has quit IRC (Quit: nicklovescode) 2011-03-21T21:39:13 antimatroid: let's see about that debug thing 2011-03-21T21:40:06 antimatroid: the command line flag won't work if you have this in the source: #define DEBUG false 2011-03-21T21:40:21 *** nicklikescode_ has joined #aichallenge 2011-03-21T21:40:22 the #define overrides the switch, i think 2011-03-21T21:40:28 what if it's const bool DEBUG false? 2011-03-21T21:40:37 shouldn't be a const.. 2011-03-21T21:40:37 ;* 2011-03-21T21:40:48 then you really can't set it 2011-03-21T21:40:54 or let the preprocessor handle it 2011-03-21T21:41:20 *** nicklikescode_ has left #aichallenge 2011-03-21T21:41:26 hmm, i dunno then :P 2011-03-21T21:41:49 antimatroid: how do i test if it works? 2011-03-21T21:44:41 just run the program 2011-03-21T21:44:51 and paste in the sample bot input from the specs page 2011-03-21T21:44:52 just leave DEBUG undefined in the source files 2011-03-21T21:45:08 protect the DEBUG sections with #ifdef DEBUG 2011-03-21T21:45:19 janzert: i have it set to false? 2011-03-21T21:45:19 and define DEBUG on the command line for debug builds 2011-03-21T21:45:26 and they're all inline 2011-03-21T21:46:06 they have a variable in state that they can treat like cin and it all gets stripped when debug is false when compiling 2011-03-21T21:46:28 defining it in the source will only allow it to work by physically changing the source 2011-03-21T21:46:51 *** nim0o has joined #aichallenge 2011-03-21T21:47:06 aichallenge: Alexandru Stan epsilon * rbd27c56 / (2 files): moved the debugging flag in the makefile (+5 more commits...) - http://bit.ly/fYw1lY 2011-03-21T21:47:12 aichallenge: Alexandru Stan starter-packaging * rbd27c56 / (2 files): moved the debugging flag in the makefile - http://bit.ly/fYw1lY 2011-03-21T21:47:23 antimatroid: there you go ^ 2011-03-21T21:47:30 :) cheers 2011-03-21T21:48:06 antimatroid: that's a thing you might want to include in the README.md 2011-03-21T21:48:12 how to use the debug features 2011-03-21T21:48:25 yeah i shall 2011-03-21T21:48:43 i'll fix up my starter bot when i get the examples done for resolutions 2011-03-21T21:48:52 then look at the wiki 2011-03-21T21:49:14 *** tvorryn has joined #aichallenge 2011-03-21T21:50:38 *** Kingpin13 has joined #aichallenge 2011-03-21T21:50:47 *** kosiini has quit IRC (Remote host closed the connection) 2011-03-21T21:52:01 i really like my current symmetric maps 2011-03-21T21:52:11 *** nim0o has quit IRC (Ping timeout: 252 seconds) 2011-03-21T21:52:33 they map need to be made with a 2x2 block if 1 width openings present issues 2011-03-21T21:52:34 *** stalled has joined #aichallenge 2011-03-21T21:54:00 antimatroid: I'm starting to think power and damage may be equivalent 2011-03-21T21:54:18 so am i :P 2011-03-21T21:54:30 they are not 2011-03-21T21:54:31 they don't seem like it at first 2011-03-21T21:54:33 *** stalled has left #aichallenge 2011-03-21T21:54:34 but they're damn close 2011-03-21T21:54:38 they are similar 2011-03-21T21:54:39 sigh: give us a counterexample 2011-03-21T21:54:49 hang on 2011-03-21T21:55:05 seems like they should be different, but every time I try and construct an example to show it they come out the same :/ 2011-03-21T21:55:26 i think we should straight up say no openings of width one of maps 2011-03-21T21:55:53 I had counter-examples, when I was testing my battle_sim script 2011-03-21T21:56:03 consider the following 2011-03-21T21:56:03 %%%%%%% 2011-03-21T21:56:03 ........a..a.......... 2011-03-21T21:56:03 %%a.....a%%% 2011-03-21T21:56:03 %%%%%%% 2011-03-21T21:56:15 a essentially can't be killed 2011-03-21T21:56:33 with power or damage resolution 2011-03-21T21:56:52 anything that allows blocking 1 width corridors has that problem 2011-03-21T21:56:57 right 2011-03-21T21:57:43 python battle_sim.py <<<"a.a.b|.c.c" 2011-03-21T21:57:54 took a few tries :) 2011-03-21T21:58:58 so it's quite rare? 2011-03-21T21:59:15 it's hard to find if you are hunting around randomly, yes 2011-03-21T21:59:18 like I was doing 2011-03-21T21:59:38 but I know I hit a few last night 2011-03-21T21:59:43 yeah okay 2011-03-21T21:59:44 also I'm not sure that exact example is possible to get to happen in a game 2011-03-21T21:59:52 we'll i'm going to make some lunch then make those examples 2011-03-21T22:00:05 sigh: feel free to tell me anything i miss on this page 2011-03-21T22:00:09 https://github.com/aichallenge/aichallenge/wiki/Ants-Battle-Resolution-Options 2011-03-21T22:00:10 janzert, just showing that they can give different results :P 2011-03-21T22:00:16 yeah 2011-03-21T22:02:08 aichallenge: Alexandru Stan epsilon * rbd6829d / ants/visualizer/java/Makefile : added makefile for the java visualizer - http://bit.ly/gm4DGA 2011-03-21T22:02:52 janzert: that position is certainly possible, I think 2011-03-21T22:03:21 hmm... maybe :) 2011-03-21T22:03:35 how do the c's get so close? 2011-03-21T22:03:50 moving up form below? 2011-03-21T22:03:55 *from 2011-03-21T22:03:57 contestbot: later tell Frontier Please check my makefile for the visualizer, i couldn't figure out the -classpath: https://github.com/aichallenge/aichallenge/commit/bd6829d78e01f09bb10fc4902f03d47d1f7d5503#L0R3 2011-03-21T22:03:57 amstan: Aye, aye, sir 2011-03-21T22:04:26 ok, yeah if the a's had taken a step down the b a step left and the c's a step up? 2011-03-21T22:04:34 yes, I think so 2011-03-21T22:06:11 janzert, don't forget that they could have spawned in the last round 2011-03-21T22:06:21 so it opens up more possiblibies 2011-03-21T22:06:38 yes, but only if just one other side was in range of the food 2011-03-21T22:06:56 yeah, but food radius is only sqrt(2) 2011-03-21T22:07:02 much smaller than attack radius 2011-03-21T22:07:24 which only makes it harder to have had an effect 2011-03-21T22:07:40 since you're more likely to be in the attack radius range? 2011-03-21T22:08:00 anyway they certainly differ 2011-03-21T22:08:13 no, it means that you can effectively "move" diagonally 2011-03-21T22:08:14 and the example given is even possible 2011-03-21T22:08:20 because you can spawn on the diagonal 2011-03-21T22:08:55 but yeah, this is all tangential :) 2011-03-21T22:09:33 I just meant if food radius was sqrt(6) you'd have more likelihood to get positions that couldn't be moved into 2011-03-21T22:09:54 *** superflit has quit IRC (Quit: superflit) 2011-03-21T22:09:56 * janzert is afk for a while 2011-03-21T22:10:02 yes, but there is more chance that the food is lost because of conflicts 2011-03-21T22:10:16 ah, too late :( 2011-03-21T22:11:35 *** nim0o has joined #aichallenge 2011-03-21T22:12:18 i like a small spawn radius 2011-03-21T22:12:24 antimatroid: another one, but here the bots are further away than before: python battle_sim.py <<<"a.b.c||..c.a" 2011-03-21T22:12:27 i think we could take away food deaths? 2011-03-21T22:12:34 and make food squares block movement 2011-03-21T22:12:53 or at least allow the bot with the most ants to get the food square first, otherwise kill it 2011-03-21T22:16:12 *** nim0o has quit IRC (Ping timeout: 252 seconds) 2011-03-21T22:17:39 if i wanted to start working on a bot, how should I get started? since the contest hasn't actually started yet, the website isn't super clear 2011-03-21T22:18:03 *** kylem__ has joined #aichallenge 2011-03-21T22:18:15 https://github.com/aichallenge/aichallenge/issues#issue/3 says that neither the python nor java starter packs are finished 2011-03-21T22:19:28 i see you guys hit it big on HN and reddit yesterday, however it still seems to be in the early stages. when are you anticipating that you'll start accepting submissions? 2011-03-21T22:20:07 *** jhuynh85 has quit IRC (Quit: Page closed) 2011-03-21T22:20:09 *** nim0o has joined #aichallenge 2011-03-21T22:20:44 *** Todd200 has joined #aichallenge 2011-03-21T22:21:51 tvorryn: starter bots are in ants/dist/starter_bots - see ants/play_one_game.sh for how to run the engine, please note that everything is still in dev and is subject to change 2011-03-21T22:22:12 ok thanks 2011-03-21T22:23:01 *** Kingpin13 has quit IRC (Quit: You're a kitty!) 2011-03-21T22:23:07 sigh: that was fast... lol 2011-03-21T22:23:18 sigh: your path update time 2011-03-21T22:23:29 kylem__: not sure, it is still a few weeks away... reddit submission wasn't an official announcement or anything 2011-03-21T22:23:33 amstan: path update time 2011-03-21T22:23:35 ? 2011-03-21T22:24:18 sigh: yeah, the time between my commit and you realizing there's a new path for those things 2011-03-21T22:24:52 I just went to the page and looked at where things were :P 2011-03-21T22:26:16 sigh: sounds good. couple of follow ups...the initial forum submission says that each ant would run its own code with no centralized controller, but python start bot has a central controller it looks like. does that mean you guys are leaning towards a centrally controlled colony AI? 2011-03-21T22:26:43 kylen__: your program controls all ants at once 2011-03-21T22:26:54 we decided it wasn't feasible to do the former while it would have been nice 2011-03-21T22:27:09 kylem__: see https://github.com/aichallenge/aichallenge/wiki/Ants-Game-Specification for latest specs 2011-03-21T22:27:14 mostly due to the overhead of starting all those processes, particilarly for tcp connections 2011-03-21T22:27:25 if you look at the forum you'll see all the ideas that have been thrown around 2011-03-21T22:27:31 kylem__: can I use you? 2011-03-21T22:27:36 lol 2011-03-21T22:28:07 antimatroid: use me how? 2011-03-21T22:28:11 https://github.com/aichallenge/aichallenge/wiki/Ants-Battle-Resolution-Options 2011-03-21T22:28:32 we have yet to decide what battle option to use, i need to know what you can understand on that page easily without any background knowledge 2011-03-21T22:28:42 then consider it all and give me your thoughts :P 2011-03-21T22:28:58 sorry that there aren't examples, i'm working on those now 2011-03-21T22:32:16 lol, wtf 2011-03-21T22:32:55 Frontier: someone was complaining about visualizer on safari 2011-03-21T22:33:45 antimatroid: just from the cursory read, i'm partial to the damage option, it is the most intuitive imho 2011-03-21T22:33:49 sigh: o'rly? let me guess: the images didn't load. i hate that bug. nasty, nasty bug. i should drink less 2011-03-21T22:34:10 Frontier: not sure what it was, grep logs for safari 2011-03-21T22:34:18 kylem__: correct answer :) 2011-03-21T22:34:53 antimatroid: iterative seems not very good, especially since it depends on how the game orders the pairs versus the players, so you'd be forced to disclose very detailed specs on how the battle unfolds in order to reproduced in client simulations 2011-03-21T22:35:09 it doesn't depend on ordering pairs 2011-03-21T22:35:18 if 3 players are at an equal distance, they all die in the same iteration 2011-03-21T22:35:29 but that's my least favourite of the rules, always has been 2011-03-21T22:35:30 so ..a..b..c.., who dies? 2011-03-21T22:35:31 kylem__: none of the implementations depend on order 2011-03-21T22:35:56 kylem__: b dies for power and damage, otherwise they all die 2011-03-21T22:35:58 It is Stocha, who has trouble with safari. I was thinking about possible sources for that problem, but i didn't find anything convincing. 2011-03-21T22:36:34 kylem__: the same can be said for a..b..a :\ 2011-03-21T22:36:57 Frontier: was that with buttons? 2011-03-21T22:36:58 antimatroid: with sqrt(6) radius they are too far away to kill each other 2011-03-21T22:37:00 janzert had that problem 2011-03-21T22:37:11 sigh: ergh, you know what i mean :P 2011-03-21T22:37:43 a guy from texas told me in an irish pub in hannover, that i was into the girl behind the bar. i said no, but he didn't listen. 2011-03-21T22:38:05 o_O 2011-03-21T22:38:21 hey 2011-03-21T22:38:27 i saw a gif somewhere 2011-03-21T22:38:31 of an ant fight 2011-03-21T22:38:32 where do all these people come from that hang around bars at midnight and want to tell people how much their life sucks 2011-03-21T22:38:39 anybody know where it was? 2011-03-21T22:38:51 CPeng: forums 2011-03-21T22:38:57 CPeng: would have been Stocha's posts 2011-03-21T22:39:10 antimatroid: janzert had trouble seeing the buttons below the visualizer 2011-03-21T22:39:27 i see. well, if you end up going with iterative, the way it reads now sounds like 1. pick closest pair of ants 2. kill both 3. if some ants still close goto 1 2011-03-21T22:40:10 it's that, but pick closest ants, kill all of them, start over 2011-03-21T22:40:27 i like to think we're not going with that option 2011-03-21T22:41:01 antimatroid: are you concerned about rounding errors in the damage option? 2011-03-21T22:41:11 no, we'd use rationals 2011-03-21T22:41:30 antimatroid: they dont exist in computers 2011-03-21T22:41:35 kylem__: you can see the implementation of all the options in ants/ants.py 2011-03-21T22:41:37 Frontier: they do if you make them 2011-03-21T22:41:46 sigh: thanks :) 2011-03-21T22:41:52 integers exist do they not? (at least close enough) 2011-03-21T22:42:05 they are the do_attack_* functions 2011-03-21T22:42:41 Frontier: what exactly exists in computers, then? 2011-03-21T22:42:55 ok, i get it now. the guinnes (<- good beer) is still makin' my thoughts fuzzy 2011-03-21T22:43:11 :P 2011-03-21T22:43:15 sigh: these ieee floating point nnumbers 2011-03-21T22:43:29 what does it mean to exist? 2011-03-21T22:43:31 :P 2011-03-21T22:43:39 to be or not to be 2011-03-21T22:43:41 what does mean? 2011-03-21T22:43:45 what does "it" mean? 2011-03-21T22:43:47 who is mean? 2011-03-21T22:43:49 who was it? 2011-03-21T22:43:54 what? 2011-03-21T22:44:08 lol -.- we crashed the conversation 2011-03-21T22:44:09 you accidentially missed 2011-03-21T22:44:22 http://www.fohguild.org/forums/attachments/screenshots/95895d1229374265-funny-strange-random-pics-24mcj1g.jpg 2011-03-21T22:44:55 what? aha 2011-03-21T22:47:29 import fraction, nice 2011-03-21T22:48:13 fraction? rational would make more sense :\ 2011-03-21T22:48:59 antimatroid: are you trolling me? 2011-03-21T22:49:13 well a fraction could be of many things 2011-03-21T22:49:16 a rational only ints 2011-03-21T22:49:34 rational can be anything too 2011-03-21T22:49:43 hmmm so it can :P 2011-03-21T22:50:00 you can have rational functions 2011-03-21T22:50:14 yeah, anything with a gcd defined 2011-03-21T22:57:07 i really have to find out what is going wrong with the images on the visualizer. it is difficult to test when everything is local 2011-03-21T22:58:30 *** fireowrker76 has joined #aichallenge 2011-03-21T22:58:47 when i try to run the test_bot.sh script i get this error 2011-03-21T22:58:48 #aichallenge 2011-03-21T22:58:53 oops hold on 2011-03-21T22:58:59 amstan: should that make file be removing object files already for me? 2011-03-21T22:59:04 Traceback (most recent call last): File "../worker/engine.py", line 30, in run_game bots = [Sandbox(*bot) for bot in botcmds] File "../worker/sandbox.py", line 61, in __init__ cwd=working_directory) File "/usr/lib/python2.6/subprocess.py", line 623, in __init__ errread, errwrite) File "/usr/lib/python2.6/subprocess.py", line 1141, in _execute_child raise child_exception OSError: [Errno 2] No such fil 2011-03-21T22:59:09 what am i doing wrong... 2011-03-21T23:00:03 *** kylem__ has quit IRC (Ping timeout: 252 seconds) 2011-03-21T23:00:12 firewrker76: you want to run playgame.py 2011-03-21T23:00:37 ok 2011-03-21T23:00:40 whats the test script for? 2011-03-21T23:05:07 amstan: I read your comment on the Makefile, but all I know is that the plugin.jar needs to be on the classpath. That file is in the Sun JRE. My Gentoo linux uses java-config-2 to create a common path to the jre. I don't know what Ubuntu uses. There might be something similar. 2011-03-21T23:06:52 *** fireowrker76 has quit IRC (Quit: Page closed) 2011-03-21T23:08:24 amstan: if Makefiles understand environment variables, try: $JAVA_HOME 2011-03-21T23:11:04 *** chris___0076 has joined #aichallenge 2011-03-21T23:12:52 lol, there's definitely something f'd with my github 2011-03-21T23:13:07 i have to manually add my email address to the end of my passkey on the website to authenticate each time 2011-03-21T23:13:25 if it's more than a line, pastebin it please 2011-03-21T23:13:50 oops sorry, was scrolled up 2011-03-21T23:14:31 *** chris__0076 has quit IRC (Ping timeout: 255 seconds) 2011-03-21T23:17:20 btw, a.b.c|.c.. is a slightly simpler example showing difference between power and damage resolutions 2011-03-21T23:22:41 actually I guess not really simpler than the example sigh gave, just simpler the example I misremembered as being from him in my playing around :P 2011-03-21T23:23:41 janzert: I had an example where the bots where further away from each other: "a.b.c||..c.a" 2011-03-21T23:24:10 dunno if you saw that one 2011-03-21T23:24:13 nice 2011-03-21T23:24:19 no I didn't 2011-03-21T23:24:33 was just looking for an example that came out different for all 4 methods 2011-03-21T23:26:00 "a.b.c|..c.a|..a" 2011-03-21T23:26:20 cool :) 2011-03-21T23:26:22 although that one probably really can't occur in a game 2011-03-21T23:26:27 yeah :P 2011-03-21T23:28:03 sigh: what format does my file need to be for battle_sim? 2011-03-21T23:28:18 ^^ 2011-03-21T23:28:19 are you running on linux? 2011-03-21T23:28:22 yeah 2011-03-21T23:28:39 ./battle_sim.py <<<"a.b.c||..c.a" 2011-03-21T23:28:40 "a.b.c|..c.a|..a" is valid input (without quotes) 2011-03-21T23:29:04 or replace | with newlines 2011-03-21T23:29:10 yup 2011-03-21T23:31:06 erm.. maybe i'm not passing the file correctly 2011-03-21T23:31:24 i was doing this "python battle_sim.py << egin.txt" 2011-03-21T23:31:35 just < 2011-03-21T23:31:36 i guess that's probably passing the actual input aye 2011-03-21T23:32:07 paste egin.txt? 2011-03-21T23:32:18 or just do this if you want it to be inline: python battle_sim.py <<<"a.b.c||..c.a" 2011-03-21T23:32:38 http://pastebin.com/Vgn46KQ1 2011-03-21T23:32:44 doesn't seem to like < syntax 2011-03-21T23:32:49 just shouts at me 2011-03-21T23:32:56 paste error? 2011-03-21T23:33:13 or you using one < for the redirect? 2011-03-21T23:33:16 http://pastebin.com/1Ae0iPew 2011-03-21T23:33:26 err, i didn't put it in :P 2011-03-21T23:33:31 python battle_sim.py < egin.txt 2011-03-21T23:33:36 is what you want 2011-03-21T23:33:45 oh, nah, same error 2011-03-21T23:33:49 i just copied the wrong one 2011-03-21T23:34:11 invalid syntax... hmm 2011-03-21T23:34:25 git status shows no changes to battle_sim.py? 2011-03-21T23:34:46 dw, it works now 2011-03-21T23:34:49 it was something i did 2011-03-21T23:34:55 sorry :P 2011-03-21T23:35:12 lol 2011-03-21T23:35:21 i changed the main function trying to work out what was going on, breaking it in the process :P 2011-03-21T23:35:25 was gonna say, there is no 115 2011-03-21T23:35:30 :) 2011-03-21T23:35:40 btw, you don't need to pad the map 2011-03-21T23:35:49 the script makes sure there is no wrarpping 2011-03-21T23:35:53 *wrapping 2011-03-21T23:36:22 Frontier: ok, i'm back.. 2011-03-21T23:36:33 "a.b.c||..c.a||.a" (more vertical space added) still works to show a difference with all the methods of resolution, still probably not possible in a game though 2011-03-21T23:36:39 Frontier: my $JAVA_HOME is blank 2011-03-21T23:36:53 Frontier: is it another package i have to install? 2011-03-21T23:38:58 sigh: it might be better so people don't get confused 2011-03-21T23:39:15 what might be better? 2011-03-21T23:39:23 to make sure it wouldn't change with wrapping 2011-03-21T23:39:30 i'll make the examples smaller, that was just a first 2011-03-21T23:39:44 for testing battle resolutions I think keeping wrapping off is best 2011-03-21T23:39:47 actually, i'll just say ignore wrapping for tests 2011-03-21T23:40:05 yes keeping it off, i was referring to people looking at the examples and going but but... 2011-03-21T23:40:58 well, then put some padding in the examples... this script is for people to experiment with the specs :) 2011-03-21T23:42:02 hmm... I suppose it is useful even after a method has been settled 2011-03-21T23:42:16 for working out tactics 2011-03-21T23:42:55 antimatroid: you available? 2011-03-21T23:52:40 amstan: what for? 2011-03-21T23:52:58 antimatroid: well, i finished the starter packages packaging.. 2011-03-21T23:53:05 i need to look at the tools section 2011-03-21T23:53:17 don't package the c++ one yet :P 2011-03-21T23:53:46 umm 2011-03-21T23:53:49 it's a makefile 2011-03-21T23:53:52 ah, okay 2011-03-21T23:53:53 i type make to make it 2011-03-21T23:53:56 make clean to delete it 2011-03-21T23:54:04 so.. tools 2011-03-21T23:54:06 what do we need? 2011-03-21T23:54:23 i have sample bots, so that's good 2011-03-21T23:54:28 visualiser, engine, bot, sample bots 2011-03-21T23:54:29 i need an offline engine 2011-03-21T23:55:03 playgame.py 2011-03-21T23:55:12 it would be awesome if there was a program like jbotmanager that we released officially 2011-03-21T23:55:18 but i don't know who could write it 2011-03-21T23:55:21 delt0r is quite busy 2011-03-21T23:55:27 antimatroid: we'll use what we have for now 2011-03-21T23:55:32 janzert: how user friendly is that? 2011-03-21T23:55:51 about as userfriendly as our past engines were I believe :P 2011-03-21T23:55:55 that is not at all 2011-03-21T23:56:10 yep, i agree with that 2011-03-21T23:56:23 we could give bat/shell scripts etc. that do it all 2011-03-21T23:56:29 or just give example commands like previously 2011-03-21T23:56:37 a readme.md should suffice 2011-03-21T23:56:52 should be the same as the one online.. 2011-03-21T23:57:50 is playgame.py all i need in terms of tools? 2011-03-21T23:58:17 a visualizer for the resulting output? 2011-03-21T23:59:20 which I believe Frontier was working on a way to easily run the current one locally 2011-03-21T23:59:34 *** nivardus has quit IRC (Quit: Lost terminal)