2011-10-09T00:26:19 *** delt0r_ has quit IRC (Ping timeout: 260 seconds) 2011-10-09T00:38:17 *** delt0r_ has joined #aichallenge 2011-10-09T00:40:21 *** dvladim has joined #aichallenge 2011-10-09T00:45:38 *** UncleVasya has joined #aichallenge 2011-10-09T01:02:36 *** dvladim has quit IRC (Ping timeout: 240 seconds) 2011-10-09T01:13:49 *** dvladim has joined #aichallenge 2011-10-09T01:17:47 *** okayzed is now known as okay 2011-10-09T01:21:38 *** antimatroid has quit IRC (Ping timeout: 245 seconds) 2011-10-09T01:28:21 *** antimatroid has joined #aichallenge 2011-10-09T01:45:27 *** antimatroid has quit IRC (Quit: Leaving.) 2011-10-09T01:46:23 *** antimatroid has joined #aichallenge 2011-10-09T01:48:21 *** eashoka has joined #aichallenge 2011-10-09T01:50:39 *** antimatroid has quit IRC (Ping timeout: 248 seconds) 2011-10-09T01:58:32 good morning/evening to all! 2011-10-09T01:59:36 *** dvladim has quit IRC (Ping timeout: 260 seconds) 2011-10-09T02:01:38 *** antimatroid has joined #aichallenge 2011-10-09T02:02:09 can someone pls explain how the game engine passes the hill info.. (in case a hill change hands during the time it was not visible to a bot) ? 2011-10-09T02:05:14 *** prathab has joined #aichallenge 2011-10-09T02:05:31 hi when is the contest starting and ending ? 2011-10-09T02:11:53 prathab: nobody knows. amstan said something like 'next week' on the forum but it can be (and most likely will be) changed. 2011-10-09T02:15:50 prathab: Last one has been active for 3 months but this time admins want to make it shorter (2-2.5 months I think). 2011-10-09T02:16:53 UncleVasya : Oh...Thank you :) 2011-10-09T02:20:44 eashoka: hill information is given like food 2011-10-09T02:20:53 eashoka: I'm not sure I understood your question right so excuse me if this is not the answer: info about hills comes like an info about food; 2011-10-09T02:21:19 once the hill is destroyed the bot won't receive the information about it anymore (since it's gone) 2011-10-09T02:21:31 it means that if it is exists no more than info about it just not comes to you. 2011-10-09T02:22:19 also if you saw hill and see it no more info about it doesn't comes to you 2011-10-09T02:23:30 therefore if hill you saw but do not see anymore was destroyed, when you'll see this square again it will be just land. 2011-10-09T02:25:12 janzert, UncleVasya : thanks. 2011-10-09T02:25:51 eashoka: if you are thinking about how to keep info about hills between turns, I suggest this way: 2011-10-09T02:26:59 - do not clear a list of hills between turns; 2011-10-09T02:27:22 UncleVasya : At the moment, I'm doing it like the same way I'm doing for unseen food. it behaved weird that's why I wanted to confirm whether my understanding of the spec is correct. looks like a bug in the code 2011-10-09T02:28:20 UncleVasya : yes, that's how I do it now. have to find the bug now. :-) BTW, congrats, have seen you up few steps in rankings recently 2011-10-09T02:28:26 - make an 'update list' function which returns [hills you see] + [hills that is not currently visible] 2011-10-09T02:28:42 thanks 2011-10-09T02:29:29 UncleVasya : have few of other qsns for you about your country if you don't mind :-) 2011-10-09T02:29:32 yep, I remember I have fought some stupid bugs with hills keeping. 2011-10-09T02:29:59 qsns - humor? 2011-10-09T02:30:04 I'm interested:) 2011-10-09T02:30:25 oh, question. 2011-10-09T02:30:27 is Mikhail Sholohov (a soviet era write if I have typed the name properly) a ukranian ? 2011-10-09T02:31:14 I don't know this name. Didn't like literature in school. 2011-10-09T02:32:30 UncleVasya : Once I had a kind of passion for Russian and soviet literature. Now that Ukrain is separate I just like to know who were the Ukranian writers I may have already red... 2011-10-09T02:34:52 UncleVasya : ah no worries then. :-) 2011-10-09T02:36:07 *** okay is now known as okayzed 2011-10-09T02:36:57 Russian wiki says he is Russian. 2011-10-09T02:40:00 UncleVasya : thanks. Think my misunderstanding came from getting the geography wrong, I thought river Don belongs to Ukrain 2011-10-09T02:40:29 I'll lurk about 'Sholohov is Ukranian' but my country now tries to take the parnt-rights to almost everything good that was in USSR. There are so many screams like 'this USSRs invention was invented by ukranians!' or 'This famous person was a native ukranian!' 2011-10-09T02:41:33 UncleVasya : hmm interesting :-) You follow chess? Vassily Ivanchuk is favorite player. 2011-10-09T02:41:40 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 276 seconds) 2011-10-09T02:41:44 I meant my favourite player 2011-10-09T02:46:56 eashoka: sry, I'm not an interesting person. Mybiggest ability is just to waste a time (tv series like House, Bones; computer game Europa Universalis 3, music with female vocal). Google AI Challenge is the greetest good from all this because it at least makes me do something useful, do some learning and programming practic. 2011-10-09T02:50:17 I have found no info about 'Sholohov is ukranian'. 2011-10-09T02:52:27 eashoka: Did you heard about Rybka's disqualification? 2011-10-09T02:52:42 UncleVasya : yep 2011-10-09T02:52:57 they copied code from some other program or something like that ? 2011-10-09T02:54:26 yes. All results for previous contest (Rybka was the #1 for several years) was annulated. 2011-10-09T02:54:35 *contests 2011-10-09T02:59:07 hmm chess programs are now getting boring. most of them can beat most of human players.. sounds like AI has scored a complete victory there. little room for improvement I think. they have totally cracked checkers for instance. 2011-10-09T03:00:06 As about cracking chess: 2011-10-09T03:00:47 UncleVasya : However a game like Go could have lot of challenge for AI programmers. Sometime back I heard there is a contest for that going on.. if you are interested. :-) 2011-10-09T03:03:49 I've heard a supposition that if we take an athom-sized thing and will use it like 0-1 memory cell than we need a memory card with a size of a Moon to store all info for cracked chess. 2011-10-09T03:06:21 I haven't even played Go but I've heard that it has much more tactic and strategy than chess. 2011-10-09T03:06:31 UncleVasya : I didn't meant chess will be totally cracked that soon but even without totally cracking it, they have pretty much solved the game now. (meaning top programs now will beat all top players hands down) in a programming angle, looks like it is a well trodden path now. openings worked out for 20 moves or so, end game table spaces have all build etc... 2011-10-09T03:07:07 Do you play it? Go I mean. 2011-10-09T03:08:00 UncleVasya : nope. just heard about it. (and red some) it has lot of possibilities so difficult to crack and more strategic than tactical so harder for computer programs. 2011-10-09T03:10:09 It's funny to remember words of Garry Kasparov when journalists asked him 'can you prognose the result of your game with Deep Blue?'. He said: 'I think it's a stupid supposition that computer can beat human in chess'. 2011-10-09T03:13:55 btw, don't you want to try play Go online with me? 2011-10-09T03:15:01 UncleVasya : sure 2011-10-09T03:16:19 So I go to wiki to read about rules :) 2011-10-09T03:16:57 *** ibdknox has quit IRC (Remote host closed the connection) 2011-10-09T03:16:58 sure. let me know once u r making progress. (I need to do the same) :-) 2011-10-09T03:17:29 *** ibdknox has joined #aichallenge 2011-10-09T03:19:25 *** chrisp_ has joined #aichallenge 2011-10-09T03:22:00 *** ibdknox has quit IRC (Ping timeout: 255 seconds) 2011-10-09T03:37:32 *** chrisp_ has quit IRC (Quit: Page closed) 2011-10-09T03:45:39 eashoka: I've registered on http://www.online-go.com . 2011-10-09T03:46:08 or 'I did registered'? 2011-10-09T03:47:07 first is correct, although "I've registered at..." might be a bit more common usage 2011-10-09T03:47:48 also if you're looking for a place to play go the KGS server is one of the more popular servers for realtime play 2011-10-09T03:48:02 http://www.gokgs.com/ 2011-10-09T03:50:19 thanks 2011-10-09T03:52:20 *** prathab has quit IRC (Quit: Page closed) 2011-10-09T03:53:19 *** iljuss__ has joined #aichallenge 2011-10-09T04:02:12 janzert : thanks for the link. 2011-10-09T04:09:18 eashoka: My limit of megabytes per day is almost exhausted so I can't play today:( 2011-10-09T04:10:32 UncleVasya : no prob. lets play on some other day 2011-10-09T04:10:50 on janzert's link? 2011-10-09T04:12:33 UncleVasya : yea let's give that one a try first since the reccomandation has come from a reliable source :-) 2011-10-09T04:12:53 aha 2011-10-09T04:33:28 janzert: As I understand general contest stuff now is pretty suitable for zeta's requirements so [a work on Ants ---> Zeta] must be less labour-consuming than [a work on Planet Wars ---> Ants] ? 2011-10-09T04:36:27 *** Akranis has joined #aichallenge 2011-10-09T04:45:38 *** UncleVasya has quit IRC (Ping timeout: 244 seconds) 2011-10-09T04:53:12 *** dvladim has joined #aichallenge 2011-10-09T05:00:46 *** iljuss__ has quit IRC (Ping timeout: 276 seconds) 2011-10-09T05:06:59 *** iljuss__ has joined #aichallenge 2011-10-09T05:37:01 *** iljuss__ has quit IRC (Ping timeout: 255 seconds) 2011-10-09T05:38:01 *** olexs has joined #aichallenge 2011-10-09T06:09:07 *** foRei has joined #aichallenge 2011-10-09T06:17:46 *** g0llum has joined #aichallenge 2011-10-09T06:23:12 *** eashoka has quit IRC (Quit: Page closed) 2011-10-09T06:31:06 *** mcstar has joined #aichallenge 2011-10-09T07:14:01 *** dvladim has quit IRC (Ping timeout: 276 seconds) 2011-10-09T07:27:06 *** mleise has joined #aichallenge 2011-10-09T07:46:37 whoa 2011-10-09T07:46:42 java is FAST 2011-10-09T07:50:12 lol 2011-10-09T07:50:25 roflmao 2011-10-09T07:50:35 Seems like you last used Java 1.3 2011-10-09T07:50:47 me? 2011-10-09T07:50:54 no mcstar 2011-10-09T07:51:06 oh, i was going to say, i realise java is nowhere near as bad as i make out 2011-10-09T07:52:40 And it is so easy to parse that it was possible to write amazing IDEs around it 2011-10-09T07:53:28 Imagine you had to write refactoring for a language that supports macros, templates and operator overloading of anything. 2011-10-09T07:56:07 what 2011-10-09T07:56:44 mleise: you told me you never talk to me again 2011-10-09T07:56:58 pls, liase through antimatroid 2011-10-09T07:57:21 well, you could say you didnt,hm 2011-10-09T07:58:05 anyway, im just trying out some tight loops in several languages, and java beats sbcl, too, but its because java is typed 2011-10-09T07:59:53 actually i never used java before 2011-10-09T07:59:57 *** onensora has joined #aichallenge 2011-10-09T08:04:21 *** smiley1983 has joined #aichallenge 2011-10-09T08:05:35 antimatroid: whats the word im looking for? i cant spell it: lieasion, its not lease, but sounds like liease would, its like a mitigator? a third party that communicates between some other two 2011-10-09T08:05:57 liason? 2011-10-09T08:06:05 i dunno 2011-10-09T08:06:17 *** ben_ has joined #aichallenge 2011-10-09T08:06:25 liason isnt in the dictionary 2011-10-09T08:06:26 anyone about? 2011-10-09T08:06:30 but yeah, sonds like it 2011-10-09T08:08:04 liaison 2011-10-09T08:08:10 missing an 'i' 2011-10-09T08:08:10 liaison 2011-10-09T08:08:20 damn, keystrokes too late :P 2011-10-09T08:08:27 : 2011-10-09T08:08:29 ) 2011-10-09T08:08:42 argh, thx, i dont know how i missed that 2011-10-09T08:08:47 ^ broken attempt at a smiley 2011-10-09T08:08:50 i usually can spell english 2011-10-09T08:08:59 smiley can't smiley 2011-10-09T08:09:07 indeed :) 2011-10-09T08:09:08 ):)) 2011-10-09T08:11:00 i watched tron legacy yesterday 2011-10-09T08:11:13 i wonder what they did to jeff bridges to look that young 2011-10-09T08:13:48 they did a story on it at some point 2011-10-09T08:13:54 it was pretty cool 2011-10-09T08:14:00 cgi, it seems: http://3danimationcgi.com/young-cgi-jeff-bridges-as-clu-in-tron-legacy-3d/ 2011-10-09T08:14:31 http://www.guardian.co.uk/film/filmblog/2010/dec/06/jeff-bridges-tron-legacy 2011-10-09T08:20:30 huh 2011-10-09T08:20:46 sometimes it looked a bit artificial 2011-10-09T08:22:15 i didnt like 13 there though, it was odd 2011-10-09T08:22:47 and lol to the isomorphic algorithms 2011-10-09T08:23:04 iso is a standards, damn it 2011-10-09T08:26:31 *** delt0r_ has quit IRC (Ping timeout: 252 seconds) 2011-10-09T08:27:05 *** olexs1 has joined #aichallenge 2011-10-09T08:29:13 *** olexs has quit IRC (Ping timeout: 260 seconds) 2011-10-09T08:34:19 now let me share this with you: 2011-10-09T08:34:27 c: 5010ms 2011-10-09T08:34:35 sbcl: 5240ms 2011-10-09T08:34:40 java: 19ms 2011-10-09T08:35:06 what 2011-10-09T08:37:25 3x nested loop 2011-10-09T08:37:33 incrementing 1 variable 2011-10-09T08:37:39 basically its a combinatorical counting 2011-10-09T08:37:53 c -01: 2032ms 2011-10-09T08:37:59 c -02: 10ms 2011-10-09T08:38:04 c -03: 12ms 2011-10-09T08:38:36 now i just have to figure out, how to speed up sbcl 2011-10-09T08:40:29 *** delt0r_ has joined #aichallenge 2011-10-09T08:41:14 i wonder what -O2 does in this case though, will it unroll a 1000 iterations? 2011-10-09T08:41:42 or somehow it un-nest the nested loops 2011-10-09T08:41:48 un-nests* 2011-10-09T08:42:32 yes, guys, your opinions are perfectly welcomed! 2011-10-09T08:45:35 i cant get below 1300ms with sbcl 2011-10-09T08:52:32 *** iljuss__ has joined #aichallenge 2011-10-09T08:53:41 *** iljuss has joined #aichallenge 2011-10-09T08:57:30 *** iljuss__ has quit IRC (Ping timeout: 258 seconds) 2011-10-09T08:59:39 *** ben_ has quit IRC (Quit: Page closed) 2011-10-09T09:18:26 *** olexs1 has quit IRC (Quit: Leaving.) 2011-10-09T09:24:36 *** smiley1983 has quit IRC (Ping timeout: 240 seconds) 2011-10-09T09:25:06 *** smiley1983 has joined #aichallenge 2011-10-09T09:30:22 i figured out, the -O2 optimization skips the inner loop 2011-10-09T09:30:47 just increments the variable with a different number (the number of the iterations of the inner loop) instead of 1 2011-10-09T09:31:03 *** antimatroid has quit IRC (Ping timeout: 258 seconds) 2011-10-09T09:31:30 with this, written manually, sbcl runs in 7ms 2011-10-09T09:33:34 *** antimatroid has joined #aichallenge 2011-10-09T09:47:01 *** iljuss has quit IRC (Read error: Connection reset by peer) 2011-10-09T09:54:56 it seems, pypy can be as fast as unoptimized C 2011-10-09T10:19:21 *** iljuss has joined #aichallenge 2011-10-09T10:21:34 *** dvladim has joined #aichallenge 2011-10-09T10:28:00 *** mleise1 has joined #aichallenge 2011-10-09T10:28:10 *** mleise has quit IRC (Ping timeout: 255 seconds) 2011-10-09T10:31:00 *** iljuss has quit IRC (Ping timeout: 260 seconds) 2011-10-09T10:31:02 *** iljuss_ has joined #aichallenge 2011-10-09T10:40:53 *** iljuss_ has quit IRC (Read error: Connection reset by peer) 2011-10-09T10:44:58 *** sigh has quit IRC (Remote host closed the connection) 2011-10-09T10:47:02 *** amstan has joined #aichallenge 2011-10-09T10:47:02 *** ChanServ sets mode: +o amstan 2011-10-09T10:55:44 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-10-09T11:15:15 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-10-09T12:16:23 *** onensora has quit IRC () 2011-10-09T12:19:27 *** Palmik has joined #aichallenge 2011-10-09T12:40:37 *** dvladim has quit IRC (Ping timeout: 252 seconds) 2011-10-09T12:53:02 *** olexs has joined #aichallenge 2011-10-09T13:05:19 *** smiley1983 has quit IRC (Quit: leaving) 2011-10-09T13:18:16 *** temp has joined #aichallenge 2011-10-09T13:24:54 *** temp has quit IRC (Ping timeout: 252 seconds) 2011-10-09T13:43:26 *** onensora has joined #aichallenge 2011-10-09T14:07:33 gcj is really fast too 2011-10-09T14:07:59 on average it beats the C codeű 2011-10-09T14:17:36 *** dlila has joined #aichallenge 2011-10-09T14:46:45 *** mceier has joined #aichallenge 2011-10-09T14:52:47 *** onensora has quit IRC (Ping timeout: 255 seconds) 2011-10-09T14:54:06 *** kaemo has quit IRC (Remote host closed the connection) 2011-10-09T14:54:37 *** onensora has joined #aichallenge 2011-10-09T14:58:02 *** kaemo has joined #aichallenge 2011-10-09T15:07:11 *** mceier has quit IRC (Ping timeout: 244 seconds) 2011-10-09T15:07:30 *** mceier has joined #aichallenge 2011-10-09T15:07:57 *** onensora has quit IRC (Read error: Connection reset by peer) 2011-10-09T15:10:09 *** george___ has joined #aichallenge 2011-10-09T15:17:17 *** ibdknox has joined #aichallenge 2011-10-09T15:17:34 *** justasu has joined #aichallenge 2011-10-09T15:22:32 *** onensora has joined #aichallenge 2011-10-09T15:25:33 *** ibdknox has quit IRC (Remote host closed the connection) 2011-10-09T15:37:57 *** ibdknox has joined #aichallenge 2011-10-09T15:41:10 *** ibdknox has quit IRC (Remote host closed the connection) 2011-10-09T15:47:29 *** Akranis_ has joined #aichallenge 2011-10-09T15:47:54 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-09T15:51:25 *** Akranis has quit IRC (Ping timeout: 276 seconds) 2011-10-09T15:58:04 *** rogue780 has joined #aichallenge 2011-10-09T15:58:17 hey, how's it going? 2011-10-09T15:58:39 any word regarding the 2011 ai challenge? 2011-10-09T15:59:55 *** Accoun has quit IRC () 2011-10-09T16:03:57 :p 2011-10-09T16:06:44 ~p 2011-10-09T16:06:58 steaming bowl of soup 2011-10-09T16:10:53 *** justasu has quit IRC (Quit: Leaving) 2011-10-09T16:11:20 *** justasu has joined #aichallenge 2011-10-09T16:13:36 *** Accoun has joined #aichallenge 2011-10-09T16:26:40 *** delt0r_ has quit IRC (Ping timeout: 260 seconds) 2011-10-09T16:40:23 *** delt0r_ has joined #aichallenge 2011-10-09T16:40:59 *** Palmik has quit IRC (Remote host closed the connection) 2011-10-09T16:48:52 :\ 2011-10-09T16:50:23 aichallenge: Pavel Vasin * r5846111 / (3 files): add hills to java starter bot - http://git.io/TnCQ2g 2011-10-09T16:50:24 aichallenge: Janzert * r148e416 / (3 files): 2011-10-09T16:50:24 aichallenge: Merge pull request #281 from rat4/epsilon 2011-10-09T16:50:24 aichallenge: update java starter bot - http://git.io/c8f8nA 2011-10-09T16:53:37 so…it's just in the planning stages then? Is there a target date to get it kicked off? 2011-10-09T16:54:13 we're basically just waiting on some approvals from Google at this stage 2011-10-09T16:54:34 pending that, hoping to launch within the the next week or so 2011-10-09T16:54:38 cool 2011-10-09T16:54:51 I'm quite looking forward to it this year 2011-10-09T16:54:54 stress on the "or so" 2011-10-09T16:55:28 well, I really think once the approvals are finished we'll be launching within a few days 2011-10-09T16:55:35 I've always wondered how the contest has google's name, but as far as I can tell no real google affiliation 2011-10-09T16:55:54 the question is just how long it will take for the go ahead to happen 2011-10-09T16:56:38 janzert: have you convened with google yet, or did marco said next week? 2011-10-09T16:56:45 say* 2011-10-09T16:57:02 rogue780: Jeff Cameron, who was a former Google intern, rebooted the contest and asked someone at Google for permission to use the name. 2011-10-09T16:57:13 Jeff is our link to Google and I believe he is meeting with the appropriate person(s) early this week 2011-10-09T16:57:35 ? its already next week where you live? 2011-10-09T16:57:53 sunday is considered the start of the week here yes 2011-10-09T16:57:58 oh 2011-10-09T16:58:23 indeed. It's good to be 'mercan 2011-10-09T16:58:29 i thought its was a priting type that english calendars start from sunday 2011-10-09T16:58:36 printing typo* 2011-10-09T16:58:40 hehe 2011-10-09T16:58:41 * mcstar stfu 2011-10-09T16:58:59 Canada, here. Weeks do indeed start on Sunday. 2011-10-09T16:59:03 mcstar, where are you from? 2011-10-09T16:59:15 from a much saner world, it seems 2011-10-09T16:59:33 so, not Japan, then 2011-10-09T16:59:51 iie 2011-10-09T17:00:10 I'm guessing Austria or Germany. 2011-10-09T17:00:37 Hostname: adsl-89-135-200-126.monradsl.monornet.hu 2011-10-09T17:00:37 Or I could just right click on your name and figure it out, i suppose 2011-10-09T17:00:39 Hungary 2011-10-09T17:00:45 pff 2011-10-09T17:00:51 i hate these hackers 2011-10-09T17:01:10 they can get their dirty hands on any info they want 2011-10-09T17:02:26 jbroman: you dont see it? im waveing for your satellite 2011-10-09T17:02:43 it's what we do. Watch out, next I'll be buying blowup sex dolls using your credit card 2011-10-09T17:03:30 mcstar: Oh, I'm watching you via the car across the street. Better angle. 2011-10-09T17:04:10 can your parabolic microphone handle trance music? im gonna turn it up i swear 2011-10-09T17:04:16 aichallenge: Janzert * r233ded7 / ants/dist/starter_bots/coffeescript/ants.coffee : 2011-10-09T17:04:16 aichallenge: Merge pull request #283 from aeter/epsilon 2011-10-09T17:04:16 aichallenge: coffeescript bot - adding hills; cleaning white space - http://git.io/RQCR_A 2011-10-09T17:08:16 @later tell _flag If you get a chance it'd be great if you could update the C starter bot to handle the new hill information 2011-10-09T17:08:16 janzert: I think that worked... 2011-10-09T17:10:17 *** AlliedEnvy_ has quit IRC (Remote host closed the connection) 2011-10-09T17:11:50 *** onensora has quit IRC () 2011-10-09T17:13:50 *** AlliedEnvy has joined #aichallenge 2011-10-09T17:13:52 *** justasu has quit IRC (Quit: Leaving) 2011-10-09T17:14:26 rogue780: why is your pptp port open? 2011-10-09T17:22:09 because I use pptp when I'm not home to access my network. Why are you portscanning me? 2011-10-09T17:22:30 it was just a lucky guess 2011-10-09T17:22:41 ? 2011-10-09T17:23:48 come on, you wanted to buy me dolls that blow up 2011-10-09T17:23:59 thats a bit weird 2011-10-09T17:25:43 Gah! this lab is killing me. For the life of me, I can't wrap my mind around solving a maze by using a queue. I did it with a stack, no sweat, but the queue is killing me! 2011-10-09T17:26:45 is it a high-dimensional maze? 2011-10-09T17:26:53 rogue780: BFS vs DFS? 2011-10-09T17:28:40 jbroman: is he talking about "solving" a maze by graph-searching? 2011-10-09T17:29:06 I'm speculating. 2011-10-09T17:29:12 *** ltriant has joined #aichallenge 2011-10-09T17:29:18 it's not specified in the assignment. Here, let me put the requirements in a paste bin. Any hints would be appreciated, but please don't just write code for me (though, I doubt you were going to) 2011-10-09T17:29:19 http://pastebin.ca/2088477 2011-10-09T17:29:43 is that homework? cause you know the rules... 2011-10-09T17:30:46 it's late homework :\ that's why I don't want someone to just give me code, just a few hints. here's my stack solution http://pastebin.ca/2088478 2011-10-09T17:31:09 rogue780: Advice: do not share your assignment solutions. 2011-10-09T17:31:14 There could be someone else in your class in here. 2011-10-09T17:31:54 good call. I'll keep that in mind, though the assignment is past due at this point 2011-10-09T17:32:24 thats code is too long 2011-10-09T17:32:26 -s 2011-10-09T17:33:11 rogue780: do you know rosetta-code? 2011-10-09T17:33:15 no 2011-10-09T17:33:28 http://rosettacode.org/ actually 2011-10-09T17:33:44 there are all kinds of stupid problems coded up in several languages 2011-10-09T17:33:55 i can imagine your solution lies there 2011-10-09T17:34:09 By the way, sorting algorithms are fun. 2011-10-09T17:34:12 I love sleep sort. 2011-10-09T17:34:18 i love random sort 2011-10-09T17:34:22 (Not relevant, just reminded me.) 2011-10-09T17:34:27 mcstar: random sort == bogosort? 2011-10-09T17:34:30 yes 2011-10-09T17:34:36 *** mceier has quit IRC (Quit: leaving) 2011-10-09T17:34:40 its clever isnt it? 2011-10-09T17:34:49 mcstar: Not guaranteed to finish, ever. :) 2011-10-09T17:34:58 thats the beauty 2011-10-09T17:35:25 Sorting by genetic algorithms. :) 2011-10-09T17:35:33 Permute the list, and use a heuristic to measure sortedness. 2011-10-09T17:35:52 ive seen a "hello world" example of genetic algorithms 2011-10-09T17:36:10 the fitness was the closeness to the above string 2011-10-09T17:36:32 Yeah, I saw that. 2011-10-09T17:38:19 Also, bruteforce sorting. Compute all permutations of the list then find the sorted one. 2011-10-09T17:53:44 rogue780: i studied the problem for a bit, and i dont see whats the problem with the queue 2011-10-09T17:54:04 (i wont read your code though) 2011-10-09T17:54:39 and its quite funny, that the solution is in the assignment 2011-10-09T17:54:52 what was the point of it? learning c++? 2011-10-09T17:56:29 *** antimatroid has quit IRC (Quit: Leaving.) 2011-10-09T17:56:59 *** antimatroid has joined #aichallenge 2011-10-09T17:58:56 *** Akranis_ has quit IRC (Ping timeout: 252 seconds) 2011-10-09T18:03:13 *** antimatroid has quit IRC (Read error: No route to host) 2011-10-09T18:03:22 *** antimatroid has joined #aichallenge 2011-10-09T18:04:28 mcstar, that's the point. The class is entitled , "Datastructures in C++" 2011-10-09T18:04:46 so 2011-10-09T18:04:57 if i can help you let me know 2011-10-09T18:05:07 mcstar, as far as the solution being in the assignment, I'm assuming you're referring to the part where it says, "do the same as with stack, but replace pop and push with dequeue and enqueue" right? 2011-10-09T18:05:44 it says "substitute the queue" 2011-10-09T18:05:59 i cant help with the code 2011-10-09T18:06:27 but if you dont understand how to use the queue for the pathfinding, there i might be able to help 2011-10-09T18:06:48 i mean, i dont see what could be the problem 2011-10-09T18:07:08 right…but that doesn't work. the queue is fifo, so if I reach a dead end and I start to dequeue, it pulls off the first element (the starting point) and when the queue empties, my last element is at the dead end 2011-10-09T18:07:50 your queue wont be empty 2011-10-09T18:08:01 unless theres no way to reac the exit, and you tried all tiles 2011-10-09T18:08:34 as far as i see, this is only a technical question, using queue instead of a stack 2011-10-09T18:08:54 they are different algorithms, but the assignment isnt about this 2011-10-09T18:09:34 ok, let's work through this. So I start out with a newly instantiated queue with no elements. I enqueue the starting square. Then I check around it to see what square I can go to next and then I enqueue it, then I enqueue the next one, etc. 2011-10-09T18:09:59 When I reach a fork in the maze, I choose one path and take it. If that path is wrong…then what? 2011-10-09T18:10:01 lets just say "new queue", instantiagiated isnt necessary 2011-10-09T18:11:10 with the stack, I would just pop the top element off until I get to a square that has a direction that I haven't tried yet and then continue on from there until I reach the end. with the queue, I can't pop the end element off. I can dequeue the first element, which is the starting point 2011-10-09T18:11:10 if that proves to be a dead end, you access the queue from the other way, that is the earliest enqueues element 2011-10-09T18:13:49 I don't understand 2011-10-09T18:15:16 I wish this was an in-person class vice and online class 2011-10-09T18:18:36 mcstar: Not having read his assignment, not all queues are double-ended. 2011-10-09T18:19:47 jbroman, mcstar indeed. the queue class we are provided doesn't allow us to take off from the back 2011-10-09T18:20:54 I mean, it sounds like straightforward bread-first search to me. 2011-10-09T18:21:04 *** foRei has quit IRC (Ping timeout: 248 seconds) 2011-10-09T18:21:37 you queue all the neighbots you encountered 2011-10-09T18:21:42 yes, bfs 2011-10-09T18:21:50 neigbors 2011-10-09T18:22:52 rogue780: the stack based is the depth first, meaning you scan a whole path 2011-10-09T18:23:03 that would mean using a tree, right? for BFS? 2011-10-09T18:23:09 this is breadth, you check every possible tile for up to a given length 2011-10-09T18:23:32 you can reconstruct it as a tree 2011-10-09T18:23:39 but you dont store you tiles in a tree 2011-10-09T18:23:50 (whats with my typos) 2011-10-09T18:24:14 s/bread/breadth/ 2011-10-09T18:24:19 (unless you are in a bakery) 2011-10-09T18:24:36 actually, you could, if you were to implement a priority queue based on a fast heap 2011-10-09T18:24:51 The queue implementation is not up to me 2011-10-09T18:24:59 doesnt matter 2011-10-09T18:25:07 do you know what we are talking about? 2011-10-09T18:25:39 So, I would only have one queue, then? Or are you talking about multiple queues here? 2011-10-09T18:25:47 no 2011-10-09T18:25:55 The One Queue 2011-10-09T18:26:06 ok. so I have one queue. I am at the starting point at 0,0 2011-10-09T18:26:29 I look in every direction and enqueue every square that I can move on 2011-10-09T18:27:24 then I do the same thing with the most recently enqueued square, and so on…? 2011-10-09T18:29:02 yes 2011-10-09T18:29:18 you know where you are from the bottom of the queue 2011-10-09T18:29:28 (the earlies enqueued element)ű 2011-10-09T18:29:34 and you add the neigbors 2011-10-09T18:31:07 of course this requires a lot of testing 2011-10-09T18:31:36 so, let's say I'm at 0,0 2011-10-09T18:31:53 i mean, i would remove the position from the end of the queue 2011-10-09T18:32:12 but then i would need to store paths on the board 2011-10-09T18:32:25 are you allowed to use anything other than the queue? 2011-10-09T18:32:27 i guess not 2011-10-09T18:32:43 it can be done, but i dont think its efficient 2011-10-09T18:33:01 0,1 and 1,0 are good paths. I do queue.enqueue(Square01) and queue.enqueue(Square10) so now Square10 is the last thing on my queue…and I can't get rid of it if it turns out bad unless I clear the whole queue 2011-10-09T18:34:59 I'm just having a hard time thinking of how to keep track of the paths that I haven't taken yet so I can make sure I go back and check them out too 2011-10-09T18:35:06 *** george___ has quit IRC (Ping timeout: 252 seconds) 2011-10-09T18:35:22 i guess its better if you could use multiple queues 2011-10-09T18:35:49 you must have more thoroughly read the assignment than me, can you or not? 2011-10-09T18:36:51 no, i think im thinking in terms of stacks again 2011-10-09T18:37:05 it doesn't explicitly say, though at the end of the criteria it says "When the program ends, all stacks and queues must be empty." so it does pluralize queues... 2011-10-09T18:37:48 so, I suppose I could have several queues 2011-10-09T18:37:59 I did solve it using a single stack, thoguh 2011-10-09T18:38:03 *though 2011-10-09T18:38:09 yes 2011-10-09T18:41:23 ok 2011-10-09T18:41:29 typing 2011-10-09T18:42:12 so lets say youcant remove elements from the queue, you start adding tiles to the queue, until you hit a dead end 2011-10-09T18:42:20 than you go back to the start-pointer 2011-10-09T18:42:36 lets call this thing the start -pointer, or it could be last-good-choice-pointer 2011-10-09T18:42:38 whatever 2011-10-09T18:42:51 at first it points to the earliest element of the queue 2011-10-09T18:43:01 after you hit a dead end, you go back to it 2011-10-09T18:43:19 you try another path, that you havent tried before 2011-10-09T18:43:28 sou you take another turn, from the start 2011-10-09T18:43:56 if the start position doesnt provide you any other free directions, you advance the start-pointer, until you fiond a tile that does 2011-10-09T18:44:21 you walk throught that new path too, maybe its the exit, maybe another dead end 2011-10-09T18:45:07 if it is, you go back to the start-pointer, and if there are unexplored paths, go for them, otherwise advance the pointer for a new tile that provides free direcitons 2011-10-09T18:45:14 this is not BFS 2011-10-09T18:45:24 yuo will be able to recover the path 2011-10-09T18:45:40 with bfs, and one queue, i dont thinkthat'd be possible 2011-10-09T18:45:54 rogue780: what do you think? 2011-10-09T18:46:13 <_flag> So, wasn't launch supposed to be today? 2011-10-09T18:47:01 (when you are at the start-pointer, and looking for a new tile, that isnt after the start-pointer, you should search the queue for it, and the search must fail, for you to continue on that path) 2011-10-09T18:47:30 btw, by pointer i dont mean c pointer, just my finger pointing to it, but sure, you can implement it as a pointer 2011-10-09T18:47:52 but im not sure c++ lets you do that if that class or whatever is abstracted out 2011-10-09T18:48:06 can you index the queue that is provided for you? 2011-10-09T18:48:23 or do you must access it elemnt-wise, like a list? 2011-10-09T18:49:02 _flag: eat as often as you can, famine is approaching 2011-10-09T18:49:27 rogue780: id welcome a response 2011-10-09T18:49:50 <_flag> mcstar: Are they lowering the food spawning rate, or does that cryptic message mean something else? 2011-10-09T18:49:55 btw tell your lecturer, that his map sucks 2011-10-09T18:49:57 sorry, I was trying some code. let me catch up on what you were saying mcstar 2011-10-09T18:50:30 _flag: its not cryptic, soo there will be wars fought over water and wheat 2011-10-09T18:50:36 eat as much as you still can 2011-10-09T18:50:43 <_flag> Okay, will do 2011-10-09T18:51:10 *** ibdknox has joined #aichallenge 2011-10-09T18:51:24 _flag: bosses say talks are w google are in progress 2011-10-09T18:51:37 - first are 2011-10-09T18:52:21 <_flag> mcstar: I see, do you know if there are going to be new maps for the launch? I heard we were getting rid of the ridiculously long maze ones 2011-10-09T18:52:24 mcstar, that sounds like it might work. Let me try to implement it. As far as accessing elements of the queue, I can only access the first element 2011-10-09T18:52:36 _flag: no idea sry 2011-10-09T18:53:12 rogue780: thats bad 2011-10-09T18:53:23 maybe im missing something 2011-10-09T18:58:31 *** olexs has quit IRC (Quit: Leaving.) 2011-10-09T19:23:25 *** ChrisH__ has joined #aichallenge 2011-10-09T19:25:30 @later tell McLeopold here is an example of a come from behind victory. My bot was down by 2 points with 4 ants and no hills and it captured two more hills to win with only one ant left. http://aichallengebeta.hypertriangle.com/visualizer.php?game=20919&user=5 2011-10-09T19:25:30 ChrisH__: Ready to serve, my lord. 2011-10-09T19:36:26 I think…I think I just figured it out 2011-10-09T19:42:40 *** ChrisH__ has quit IRC (Quit: Page closed) 2011-10-09T20:09:06 *** janzert has quit IRC (Ping timeout: 258 seconds) 2011-10-09T20:16:08 hmm…this new ants game looks promising 2011-10-09T20:16:24 no objective-c, though 2011-10-09T20:17:21 there is 2011-10-09T20:17:26 oh 2011-10-09T20:17:29 thats c# 2011-10-09T20:17:38 did you get yours going? 2011-10-09T20:17:44 mine found the exit 2011-10-09T20:31:24 2:30, ive got to sleep 2011-10-09T20:31:27 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-10-09T20:41:29 *** senorpanqueques has joined #aichallenge 2011-10-09T21:15:33 @later tell mcstart yes, I got it to work. I managed it using only one queue. I felt like a retard when I figured it out. 2011-10-09T21:15:33 rogue780: Yep. 2011-10-09T21:16:43 *** okayzed is now known as okay 2011-10-09T21:16:53 rogue780: I think you mean mcstar. 2011-10-09T21:16:59 I know who you mean, but contestbot likely does not. 2011-10-09T21:23:57 @later tell mcstar rogue780 later told you: "yes, I got it to work. I managed it using only one queue. I felt like a retard when I figured it out" but he mistyped your name 2011-10-09T21:23:57 amstan_: Yep. 2011-10-09T21:24:01 *** amstan_ is now known as amstan 2011-10-09T21:28:06 *** ibdknox has quit IRC (Remote host closed the connection) 2011-10-09T22:11:04 <_flag> amstan: any word on a launch date? 2011-10-09T22:11:30 _flag: we're waiting for jeff to get name rights and funding 2011-10-09T22:11:34 they have a meeting tomorrow 2011-10-09T22:11:49 *in 2 days* 2011-10-09T22:12:31 *** ibdknox has joined #aichallenge 2011-10-09T22:17:06 *** janzert has joined #aichallenge 2011-10-09T22:38:15 *** okay is now known as okayzed 2011-10-09T22:49:09 <_flag> janzert: I read awhile back that we were going to get rid of the super long maze maps, is that still happening? 2011-10-09T22:53:45 *** ChrisH_ has joined #aichallenge 2011-10-09T22:54:22 *** ChrisH_ has quit IRC (Client Quit) 2011-10-09T23:08:40 _flag: map development is planned to continue on throughout at least most of the contest 2011-10-09T23:09:10 but yes some of the current maze maps have a problem with overly long routes, e.g. maze_2 and maze_3 2011-10-09T23:13:21 <_flag> janzert: Also, how does the game determine which hill an ant will spawn out of? If I block all my own hills with my own ants will all the ants funnel out of the remaining hill? Or do I have to wait for that hill to come up in the cycle? 2011-10-09T23:16:21 *** okayzed is now known as okay 2011-10-09T23:20:03 *** dlila has quit IRC (Quit: Leaving) 2011-10-09T23:39:16 jbroman, amstan good call. now I feel like even more of a retard 2011-10-09T23:39:30 rogue780: it happens to the best of us 2011-10-09T23:39:48 well, I'm off to bed. adios 2011-10-09T23:40:33 *** okay is now known as okayzed 2011-10-09T23:43:17 _flag: it spawns at the least recently occupied hill, so yes if only one hill is unoccupied it will keep spawning there every turn