Wednesday, 2010-03-31

*** tpb has joined #tp00:00
*** ChanServ sets mode: +o tpb00:00
*** mithro has joined #tp00:43
*** Erroneous has quit IRC00:54
*** jnengland77 has joined #tp01:19
*** tansell-laptop has quit IRC02:26
*** jdlkem has joined #tp02:42
*** nash has joined #tp02:54
* llnz wanders off03:42
llnzlater all03:42
*** llnz has quit IRC03:43
*** tuna-fish has joined #tp04:18
*** sara1089 has joined #tp04:37
*** AlphonZeus has joined #tp04:49
*** StupidIncarnate has joined #tp06:03
*** Reginleifu has quit IRC06:03
StupidIncarnatebit off topic, but why do every technology seem to have an animal representation?06:07
StupidIncarnateor at least some representation that's completely unrelated most of the time06:07
StupidIncarnatesaw an xp book with a bull from06:08
StupidIncarnatefrog*06:08
StupidIncarnatethink gwt has a russian troll dude06:08
StupidIncarnatephp has one I can't remember as does java06:09
*** Reginleifu has joined #tp06:16
kerbethey add certain 'cute' factor for them06:16
kerbeit's most of the time matter of imago if something gets used or not, and especially if you're buying book, it is cover what sells. :)06:17
StupidIncarnateya I guess06:18
StupidIncarnatebtw what version of the server are you using on that gran....fr domain?06:19
kerbe.fi domain. close, but not so. :P06:20
StupidIncarnatefine, you made me scroll up gladiolus.fi06:20
kerbe0.6.2 tpserver-cpp06:21
StupidIncarnateah it must be tp04 then :/06:21
kerbeyou wished it was newer?06:21
StupidIncarnatebit older I think06:22
kerbemhh, there are few releases which I could emerge, though not sure how stable / functional they are06:23
kerbeor are their dependencies how good condition. :)06:23
StupidIncarnateare you on linux?06:23
kerbeyep.06:23
kerbeI don't touch on windows servers, unless I really have to06:24
kerbeusually I can get someone else touch them. :P06:24
StupidIncarnatewould it be too much trouble to do 0.4.1 for a bit?06:25
StupidIncarnatewanna test something06:26
StupidIncarnatethough I don't know how long it takes to compile that stuff06:26
kerbethat's not rèadily available, but I suppose I can find source tar somewhere, or getting it directly from repo06:26
StupidIncarnateit's on source forge06:27
StupidIncarnatehttp://sourceforge.net/projects/thousandparsec/files/C%2B%2B%20Server/0.4.1/tpserver-cpp-0.4.1.tar.gz/download06:27
kerbeI'll take a look in a while, do few urgent things for work first. :)06:27
tpb<http://ln-s.net/5VJO> (at sourceforge.net)06:27
kerbethanks, that helps :)06:27
StupidIncarnateno thank you06:27
StupidIncarnateI gotta go to bed soon, but i'll be back before you go to bed probably06:28
kerbeas you gave me direct link to download, it's compiling now06:30
StupidIncarnateoh ;306:30
kerbeand it fails to compile06:31
StupidIncarnatethings like that can fail?06:32
kerbehttp://pastebin.com/gCWwz2A106:32
tpbTitle: /bin/sh ../../../libtool --mod - Kerbe - gCWwz2A1 - Pastebin.com (at pastebin.com)06:32
kerbeeasily. :)06:32
StupidIncarnatechange in linux's version of dlls?06:32
kerbemost likely06:32
StupidIncarnatelinux's dll sort of files related to c++ compiling*06:32
StupidIncarnateah06:33
kerbesimilar issue was preventing latest version to compile06:33
kerbeuntil they released new version06:33
kerbeso it'll take a while to get older version up, while tracking what has to be downgraded. :)06:33
StupidIncarnatenaw, you don't have to do all that. I'll either see if tote has an older version lying around, otherwise I'll bang my head on my comp till I get the python mod working with some sort of server06:35
kerbelinux is fairly easy to get along, even with older versions. just need a bit work, if wanting to get older stuff work. :)06:39
StupidIncarnatewell, don't get too fustrated over it. I'll just check back with you when I wake up then06:40
StupidIncarnatesee you later06:40
kerbegood night. :)06:40
StupidIncarnatenight06:40
*** StupidIncarnate has quit IRC06:40
*** Reginleifu has quit IRC07:00
*** jnengland77 has quit IRC07:01
*** jdlkem has quit IRC07:15
*** AlphonZeus has quit IRC07:16
*** tansell-laptop has joined #tp08:04
*** mithro has quit IRC09:01
*** yubing has joined #tp09:46
*** mithro has joined #tp09:47
*** yubing has quit IRC10:00
*** nash has quit IRC10:30
*** bzawisto has quit IRC10:49
*** verhoevenv has quit IRC11:06
*** null_000 has joined #tp11:21
null_000hi all11:25
mithrohey null_00011:44
null_000my proposal is almost finished11:46
null_000I'll just add a little more text11:46
CIA-70jiaming tpclient-pywx-tp04 * r7819809d63e9 /windows/ (winAccount.py winServerBrowser.py):11:48
CIA-70Fix a bug in the Server Browser window.11:48
CIA-70(When this sever browser window shows up, it hides the ConnectTo window.11:48
CIA-70However, when this window is closed by clicking the cross, it does't make the11:48
CIA-70ConnectTo window visible again.)11:48
mithroI wanted to chat with you about that11:48
null_000sure11:49
mithronull_000: so what is the "core" of your proposal11:49
mithroIE what is the major pass/fail point11:50
null_000AI for MTSec and TAE  and at least basic support for coding AI in python11:52
mithrobut what does "at least basic support for coding AI in Python" actually mean?11:53
mithroand MTSec is a pretty big ruleset11:54
mithrowhat is AI support for MTSec mean?11:54
mithroare you allow to hard-code the ship designs?11:54
null_000I want to make daneel-ai in to a framework where AI coder doesn't have to study the client and protocol libraries but can focus on AI logic11:54
mithroso why is daneel-ai not already that?11:55
mithroit seems that way to me11:55
null_000It is if you use CHR syntax but not if you want to do it in pure python11:56
mithroplus the client library hides most of the protocol level11:56
mithrowhat about the Risk ruleset?11:57
mithrowhy TAE?11:57
mithrohave you asked about which rulesets are the most playable?11:57
mithrohave you gotten anywhere in your minisec coding process?11:57
null_000I choose TAE because risk already has an AI11:57
mithronull_000: how well does that AI work?11:57
mithrowhat about bugs in the server?11:58
mithroI'm sure you'll come across plenty11:58
null_000I've come across some already11:58
mithroso you need to plan that you'll come across blocking bugs11:58
mithrohow are you going to test how much better you AI is?11:59
mithroplaying against it manually is going to take a long time11:59
mithroand not actually be reliable indicator11:59
mithronull_000: sorry to fire so many questions at you, just trying to brain dump before I head to bed11:59
mithroyour schedule seems way to aggressive too12:00
null_000I'm not sure if I have any other option than to manualy test AI12:00
mithronull_000: sure you do12:00
mithronull_000: for minisec you can play against tpsai-py if you can get it working12:00
mithrothere is also genconai12:00
mithroand older versions12:01
mithroIE version 2 now always beats version 112:01
null_000yes for minisec theres tpsai-py but for MTSec there isn't anything (as far as I know)12:01
mithroI think you also need to plan how you are going to build your AI better12:02
mithrowhat type of approach are you going to take12:02
mithrosimple task list with dependencies like tpsai-py?12:02
null_000I was thinking that  AI for rulesets that dont have AI yet would be most apritiated12:02
null_000I will probably start  with a simple tasklist12:03
null_000and build from there12:03
mithronull_000: but your second version of the ai should be able to easily beat your second version12:03
mithros/second/first12:03
null_000ofcourse12:03
mithroso in 100 rounds, v2 should always beat v112:03
null_000for advanced tactics i'll have to spend more time playing12:03
mithro(with random universe)12:04
mithrowhat about improving the risk ai?12:04
mithrowhat about producing ai demo battles - like the ones we have produced previously between tpsai-py and galaxie12:05
null_000I could do that but I thought that AI for new rulesets would be more appreciated12:05
mithrohttp://www.youtube.com/watch?v=vG2YwaBN0Kc12:05
tpbTitle: YouTube - Thousand Parsec - Battle of AIs (at www.youtube.com)12:05
mithronull_000: sure - I do think you need to target a new ruleset, but I think if you are going to change daneel-ai you also need to provide it's worth on the current rulesets12:06
mithroyour proposal also needs to have stronger goals12:06
mithroI don't think you'll be able to write a *good* AI for mtsec in 1 month, let alone 1 week12:07
null_000I could drop TAE from the proposal and focus more on MTSec12:08
null_000and there is time planed for advanced tactics discovery during the community bonding period12:09
mithronull_000: how much of that time will be spent fixing bugs however?12:10
mithroand other work?12:10
mithronull_000: you need to understand *why* mtsec is going to take a lot of work12:10
mithroremeber the AI can't cheat12:10
null_000I talked to daneel-ai creator about the time needed to create an AI and his estimate was a few weeks tops12:11
mithrothat's to create an AI12:12
null_000If hard results are needed I could focus more on rulesets with existing AI and making sure my AI beats the old one12:12
mithronull_000: I think you should try two things12:12
mithro1) concentrate on MTSec12:12
mithro2) any improvements to daneel-ai need to be proved in all rulesets12:13
mithro3) you need a better plan for testing the MTSec AI12:13
mithroI want to see you have thought about these type of things12:14
null_0002.) the improvements I was planning don't change the way existing AI works but provides a different way of coding new AI12:14
mithrowhat about a "zerg rush"12:14
null_000zerg rush?12:14
mithrowhere you rush the opponnet with hundreds of cheap units early on12:14
mithroI don't even know if MTSec allows that tatic12:15
mithrohow important is ship design? It's suppose to be a core part of the game play12:15
mithrois it?12:15
mithrothese are questions you need to answer when creating an AI12:15
mithroam I making sense?12:16
null_000the main problem for me now is that I didn't have an oprotunity to play against other players12:16
mithroezod and alanp (balki) are the only people who have really played MTSec12:16
mithronull_000: sure - and uncertainty means things always take longer then you expect12:17
mithronull_000: btw - getting a minisec AI up and running is probably trivial - less then a day or twos worth of work12:18
mithrobut it would be a good thing to show off12:18
null_000I'll drop TAE as I can always do it if there is more time12:18
mithroI also think you should do your simplication in parellel on an *as needed* basis12:18
null_000I'll have minisec AI done this week, but I can't dedicate too much time because I have an exam on the 8th12:18
mithronull_000: sure, but you need to understand the uncertainty and plan for it12:19
mithroand plan for when things go wrong12:19
mithrowhat if MTSec is totally unplayable?12:19
*** bddebian has joined #tp12:20
null_000since everyone said its the most stable ruleset it would be my best bet12:20
mithrostable is a relative word :)12:20
null_000^_^12:21
mithrowe don't really have enough players to say "you can finish a MTSec game without the server crashing in most cases"12:21
mithroI guarantee that you'll come across crasher bugs12:22
mithronull_000: maybe proposing an AI grind test for the server would be cool proposal12:22
null_000I was just going to propose that12:22
mithroobviously an AI is a poor subsitute for proper tests and real human players12:23
mithrobut it's better then what we have now :P12:23
null_000I could turn the buggy state of the server to my advantage in my proposal ^_^12:23
mithroan AI's who focus is to test the server is quite different to an AI who wants to win against a player12:24
mithroalso different from an AI who wants to emulate a human player12:24
null_000any AI good or bad could have at least some use for testing the server12:25
null_000I also have an idea about AI evaluation12:25
mithrocool12:26
mithrowell I should head to bed12:26
null_000I could write a stupid AI and try to evaluate my progress with how many stupid AI can my smart AI beat (and how quick)12:26
null_000AI zerg rush12:26
mithronull_000: good idea12:27
null_000well than you for your time12:27
null_000I do need to turn my proposal around but at least I know what I can do to have better chances of being accepted12:28
mithronull_000: if you see daneel-ai's creator, tell him to sign up to be a mentor!12:28
mithronull_000: much better for it to happen now then when there is nothing you can do about it :)_12:29
null_000i talked to him about that yesterday12:29
mithronull_000: what did he say?12:29
null_000he said he'll try to apply as a student if i understood corectly12:30
mithroahh okay12:30
null_000but he said he would be happy to answer all my questions even if he's not my mentor12:30
mithroanyway, I'm really heading to bed now12:30
null_000good night then ^_^12:31
mithrogood luck!12:31
null_000thank you12:31
mithroif you see students, tell them to send email to the list and not directly to me!12:31
null_000ok12:31
null_000if I recognize them12:32
null_000btw how much competition do I have (how many proposals)? ^_^12:33
mithroat the moment 112:34
mithrobut looks like there are a bunch going onto the wiki12:34
null_000aren't there 5 proposals on the wiki (not sure if there any good)12:34
null_000thanks for the info12:35
mithrowe normally get about 50-80 proposals12:41
mithroa lot are not very good however12:42
mithrognight for real :)12:42
null_000good night12:44
*** sara1089 has joined #tp12:44
*** bddebian has quit IRC12:45
*** mithro has quit IRC12:47
*** verhoevenv has joined #tp12:48
*** Reginleifu has joined #tp12:57
kerbenull_000: if you want to have game at some point, I have server & time during easter weekend. then you have some experience about game. :)13:06
null_000that would be great13:11
kerbegreat.13:12
null_000maybe there should be a TP-players channel on irc for people who want to play13:12
kerbeain't this for those? ;913:12
null_000^___^13:12
null_000kind of... but most people here are away13:13
null_000(or just really quiet)13:13
kerbeso you want two channels where people are away? :P13:13
null_000^___^13:13
kerbethat's nature of irc ;)13:13
kerbeanyway, I'll head home, back in hour or two. :)13:14
kerbe->13:14
*** jmtan has joined #tp13:16
*** null_000 has quit IRC13:21
sara1089hello13:37
sara1089anyone interested in gsoc?13:37
ezodin what way?13:39
*** null_000 has joined #tp13:46
balkii'm interested in it13:54
balki;)13:54
*** Greywhind has quit IRC14:19
*** null_000 has quit IRC14:23
*** null_000 has joined #tp14:29
*** Reginleifu has quit IRC14:34
*** Greywhind has joined #tp14:40
*** sara1089 has quit IRC14:45
*** mhilmi has joined #tp14:54
*** null_000 has quit IRC15:55
*** Noughmad has joined #tp16:04
*** null_000 has joined #tp16:06
*** Erroneous has joined #tp16:07
*** StupidIncarnate has joined #tp17:59
StupidIncarnatehey kerbe, how'd it go?18:01
kerbehaven't had time yet to investigate it18:14
StupidIncarnateoh ok18:15
kerbework took my day, and relaxation rest. :)18:15
StupidIncarnatewhat'd you do?18:15
kerbebad company 2 (or were you interested in my work? :P)18:16
StupidIncarnatewhat you do, ya18:16
kerbein fact I'm currently eating this, having short break... http://en.wikipedia.org/wiki/M%C3%A4mmi18:19
tpbTitle: Mämmi - Wikipedia, the free encyclopedia (at en.wikipedia.org)18:19
StupidIncarnate>.> ewww?18:20
StupidIncarnateoh easter's coming up isn't it18:20
kerbenext weekend, yep18:20
StupidIncarnatelol not that I do anything with it18:20
kerbe2 free days from work, that's what it is good for :P18:20
StupidIncarnateoh, I don't work on the weekend. School don't let students18:21
kerbeI mean friday and monday are free, weekend of course is free, like most work places have :)18:22
StupidIncarnatereally18:32
StupidIncarnateI'll have to check with my work18:32
StupidIncarnatethough they are school jobs so probably don't have them off18:33
*** verhoevenv has quit IRC18:48
*** jiaming has joined #tp18:48
*** jmtan has quit IRC18:52
*** Demitar has quit IRC18:55
*** Demitar has joined #tp19:02
*** ezod has quit IRC19:06
*** ezod has joined #tp19:08
*** Noughmad has quit IRC19:25
*** Noughmad has joined #tp19:27
*** StupidIncarnate has quit IRC19:38
*** Demitar has quit IRC19:49
*** llnz has joined #tp19:52
*** Demitar has joined #tp19:58
jiaminghi llnz20:22
llnzhi jiaming20:22
jiamingi write a new proposal at http://www.thousandparsec.net/wiki/Google_Summer_of_Code/Proposals2010/Jiaming-Scenario20:23
tpb<http://ln-s.net/5WFb> (at www.thousandparsec.net)20:23
jiamingif possible please take a look at it20:23
jiamingand give me some feedback :) thanx20:23
llnzlooking now20:29
*** jiaming has quit IRC20:35
*** jdlkem has joined #tp20:42
*** jiaming has joined #tp21:27
*** llnz has quit IRC21:39
*** llnz has joined #tp21:40
llnzi hate windows21:40
*** bisc has joined #tp21:53
bischi all. anybody to discuss gsoc ideas for tpclient-pywx?21:54
null_000llnz: I can relate to that ^_^21:57
llnzi have to use it here at work21:57
null_000what kind of work do you do?21:58
llnzSoftware engineering/development and computer network preformance consulting21:59
null_000I worked 3 years for a small company that did web apps (mostly for the ministry of education) and not only did I have to use windows I had to learn and program in C# and worse Visual Basic.net22:01
llnzhehe22:02
null_000but now im mostly MS free ^_^22:02
llnzmost of my development work has been in MFC C++ and java, but currently working on python to be run on a linux server22:02
llnzbisc: i can help in general terms, but you will need mithro or Greywhind for specifics22:04
biscllnz: well... I'm in need of certain details, thanks for info22:07
biscGreywhind: can you tell me how to get your gsoc 2009 work running? I've changed for tp04 branch of tpclient-pywx, but the GUI is rather different from what's on screens of your blog. Am I doing smth wrong?22:08
Greywhindbisc: just using the tp04 branch of tpclient-pywx with the tp04 branches of libtpclient-py and libtpproto-py should be all that's required22:09
biscGreywhind: oh, thanks. I forgot to change branch for those two libs22:10
null_000llnz: this might be a stupid question... but the man page for tpserver-cpp says that you can enter console commands... how do I do that?22:29
null_000I've tried running it in debug mode...22:29
llnzah, well...22:29
llnzthe console has been removed, and replaced by an admin client22:29
null_000oh...22:30
llnztpadmin-cpp22:30
llnzit has basically the same functionality22:30
llnzpatch for the man page is welcome22:30
null_000it does say 2006-01-02 in the man page... so should see that one coming ^_^22:30
llnz:-)22:31
llnzthe man page doesn't need updating often22:31
llnzjust now22:31
null_000^_^22:31
null_000if nobody has complained about this I can see how often it is used22:32
*** Noughmad has quit IRC22:33
*** bisc has quit IRC22:37
*** quaternary has joined #tp23:21
*** null_000 has quit IRC23:38
tanselljiaming, ping?23:54
tansellGreywhind, ping?23:56

Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!