[06:26] Join: Mizcu joined #corewars [07:08] Join: pak21 joined #corewars [08:15] MSG: [08:29] Join: Fluffy joined #corewars [08:30] MSG: Client Quit [09:31] Join: willvarfa joined #corewars [09:32] MSG: Remote host closed the connection [10:51] Join: bigk joined #corewars [10:51] hi [10:51] anyone? [10:53] Part: bigk left #corewars [12:13] Join: bigk joined #corewars [12:13] Join: Mizcu joined #corewars [12:15] hi [12:15] mizcu [12:15] Nick Change: bigk changed nick to caesar [12:23] MSG: Quit: Bye... [12:50] Join: multimode joined #corewars [12:52] multimode? thats a new nickname here [12:52] Yep. [12:52] It was supposed to be "multimodel" but it seems the "l" got cut off. [12:52] Name length restriction I guess. [12:53] the server has been down for couple days, so thats the reason if you've had problems [12:53] Naa, this is my first time here. [12:53] the server borked and the host didnt have instant replacement [12:54] so, you are here to learn or to otherwise speak about corewars, right? [12:54] I've been looking for things to practice on programming wise that land in that rare part of the spectrum that is neither boring, nor takes so long to complete you'll give up before you're done. Games like Core War look like a good match. [12:55] some people would consider corewar boring as hell, but if you are already a programmer then that wont be program [12:56] http://www.corewar.info/beginner.htm [12:56] I'm not sure why it'd be considered particularly boring. [12:56] Yeah, I've been going through the pages on beginner stuff. I don't consider myself to know much yet, but I know where to learn. [12:57] I've gone through the whole Aiforge list. The games I'm interested in are Core War, RoboCom, and Hoverbots. [12:57] not particulary boring, but some people just dont understand programming for fun [12:57] I think programming for fun is fairly easy to explain. [12:57] everything spoken at this channel is logged at www.koth.org/irc-logs/ , just to note [12:58] e-mail addys are censored so thats not a problem [12:58] There are two things that will allow you to create anything you can imagine. They are your mind, and a computer. The computer is special, in that you can show it to someone else, and know that they perceive exactly what you meant. Using words to transmit thoughts doesn't work as well. [12:59] if you get any questions, just ask me, or wait until something answers in the logs [12:59] Programming is the closest we'll probably ever get to be, to being a high-fantasy wizard, or a god. [13:00] Okay. I'm sure I'll have questions later. [13:04] Hmm. I just had an insight explaining that. You could look at programming like "backwards physics". Math is the descriptive language. Physics figures out the "program" that produces the object. Programming produces the object by writing the program. [14:18] Join: Neo joined #corewars [14:18] wooooh [14:19] back to business... :) [14:45] Hi. :) [15:02] MSG: Ping timeout: 255 seconds [15:30] Join: Neo joined #corewars [15:34] MSG: Ping timeout: 255 seconds [16:50] Join: Owner joined #corewars [16:50] Hi :-) [16:50] Nick Change: Owner changed nick to John [16:52] hi John [17:20] Can any suggest a good name for a new pixel ad site? [17:22] "ripoff-site" [17:23] Well, I was thinking free pixel ads [17:47] MSG: Quit: Chatzilla 0.9.74 [Firefox 1.5.0.4/2006050817] [17:50] Join: John joined #corewars [18:02] What makes a "pixel ad" different from a "ad". I haven't heard that term before. [18:07] A pixel ad page is a website with an image on which pixels are sold, or given away [18:07] Some guy started the craze last year by having a page with 1 million pixels, and he sold them all for $1 [18:08] http://milliondollarhomepage.com/ [18:18] Why would anybody want to buy a individual pixel? [18:19] There are much cheaper advertising methods. [18:20] Well people are buying pixels, usually in 10 x 10 blocks [18:20] I though I'd make a free pixel ad site, with google ads [18:21] I still don't understand why anybody would do it. Banner ads can be put up for cheap, and that's far larger. [18:21] Some of the pixel ad sites get > 1 million hits per month [18:21] No idea why [18:21] I've only put up free ones myself [18:22] Oh well, this just confirms my theory about humanity. :P [18:23] :-) [18:26] Does anyone know the current status of Datagram's tournament? [18:30] Join: willvarfa joined #corewars [18:30] evening all [18:30] Hi Will :-) [18:41] * John waves [18:41] MSG: Quit: mov.i #1,1 [18:42] MSG: Remote host closed the connection [19:39] Join: tax0n joined #corewars [19:41] hi tax0n [19:41] hi Mizcu [19:42] please allow me a moment - need to visit kitchen [19:43] no problems - I'm just watching TV here [19:48] so, how may i be of assistance? [19:48] I'm not sure - sorry. I was hoping to just listen in to people talking about cw, so I could get learn more about the game. [19:49] this is quite silent channel [19:49] but if you feel like reading, www.koth.org/irc-logs/ [19:49] I tried to play two warriors against each other the other day but got an error and gave up [19:49] ok, thanks [19:50] what program you used? [19:50] I think it was pmarsv [19:50] www.geocities.com/corewin2/ [19:50] the one for win98/NT/Me [19:51] corewin is more userfriendly [19:51] thanks Mizcu [19:52] I'm okay with cmd line - I'm more used to linux than windows, but can't use my linux partition atm for lack of a proper mouse [19:52] are there many people who regularly play cw? [19:53] about a dozen active and dozen semiactive [19:54] and the newsgroup will hold a lot of stuff to read, http://groups.google.com/group/rec.games.corewar is you dont want to install newsreader [19:54] newsreader? an RSS feed? [19:55] usenet [19:55] Join: cywhal1 joined #corewars [19:55] mm, i am not exactly sure about the naming, but whatever [19:55] hi cywhale [19:55] Hi [19:55] yeah, I'll probably just use the group; I'm not keen on using usenet [19:56] hi cywhale [19:56] three newbies already today, i guess it has been alerady a half a year since previous spike [19:56] ;) [19:57] though i remember cywhale from somewhere [19:57] I found out about cw in Peter Szor's book "The Art of Computer Virus Research and Defence" [19:57] Wrote something about a cw blog i'm working on here at IRC a few weeks ago [19:58] he recommends playing cw over learning to write virii to become a better virus researcher [19:58] i dont think thats exactly a good recommendation, because of the different mechanisms [19:59] Mizcu: Ssh. I don't want people writing more viruses I have to cope with at work. :P [19:59] but maybe if viruses become more aggressive and complex in future, then yes [19:59] I wouldn't know - I'm a beginner in both areas and I know only the very basics of assembler [20:00] corewars is sort of really simplified assembly [20:01] yeah, I was reading: redcode - 17 | 18 instruction code set [20:01] well, redcode is. Corewars is kinda generic term. [20:02] It sounds a pretty cool game [20:02] cool in a geekish way [20:02] I'm not about to start playing publically soon because I know I'll get trounced. That said, which hill do people seem to play on most of the time? A lot of them seem to have empty queues with no recent records. [20:02] koth 94'nop is the most active [20:03] Guess I'll start playing there once I think I know what I'm doing. [20:03] it takes time to write and optimize the warrior, and as i said earlier, there is only a dozen active players [20:03] sal.math.ualberta.ca has a beginner -hill where you can send stuff without shame [20:04] Stride there is based upon a really really really tough warrior, and there are no warriors really there capable of going against [20:04] i could make something to smack it, but that wouldnt really matter [20:05] as warriors at the hill are dropped once they reach high enough age [20:05] Well, I'm planning on downloading a crapload of warriors to play against, and then not play publicly until mine manages to hold its own fairly well. I have read that it seems that most warriors either do very well against some opponents and very poorly against others, or mediocre against all of them. I'm wondering if it's possible to make a decent overall warrior. [20:05] I didn't find code for any of the warriors - this is probably a stupid question, but do people no publish code of winning warriors? [20:06] www.koth.org/planar/ has some older - [20:06] http://www.ociw.edu/~birk/COREWAR/koenigstuhl.html - almost everything, including new killers [20:06] http://para.inria.fr/~doligez/corewar/ [20:07] tax0n: The KOTH site has a bunch on its FTP server, and doligez's site has ~1000. [20:07] and you can try finding them at the newsgroup [20:07] uups, sorry, same as koth->planar [20:07] Join: Neo joined #corewars [20:07] there really is no overall warrior [20:08] certain type of stone/papers can score well against wide variety of warriors, but even then they have weaknesses against some warriors [20:10] I'm wondering if that's due to resource limits. Given enough resources it should be possible to analyze enemy behavior and adapt as appropriate. You might be taken off guard by a new strategy but it would keep you from being weak against existing ones. [20:11] yes, but it is really hard, because once you have enough info to choose your strategy, opponent is already ripping you apart [20:11] there is p-space which stores data between rounds, but p-warriors have their own problems too [20:12] and the most active hill is nop, meaning no p-space [20:13] how much room do you get in p-space? [20:13] Part: cywhal1 left #corewars [20:13] Could be. I don't know that much about Core War yet. Some strategy games reward simple tactics like rushing so much that complex tactics are pointless (rush is the optimal strategy). Usually adjusting some parameters (most games won't let you, but it should be possible with a open source game) fixes that (e.g. make defense easier at first, then offense easier as time goes on). [20:14] couple hundred numbers [20:14] i dont remember exactly how large the setting was at koth [20:14] at sal, you can see it at the same page as the hill [20:15] * Neo wakes up [20:15] hi all [20:15] hi Neo [20:15] hi neo [20:16] hi Mizcu, tax0n, multimode [20:16] Hi Neo. [20:17] p-space is usually 1/16th of core [20:17] * multimode starts wondering if something dataflow-ish wouldn't work well generally... Just use some strategy at start, don't bother monitoring the enemy, if what you're doing starts to fail, pick a strategy that isn't vulnerable to the same things... Should minimize the overhead. [20:18] Reactive rather than proactive. [20:18] there is a little problem at your thinking [20:18] Most likely. [20:18] there are warriors that attack by making you create useless processes, causing you to slow down [20:18] p-switchers already do that [20:19] and once the opponent is asleep, they go for the killing blow [20:19] yes, but p-space and non-pspace are quite different enviroments in that meaning [20:19] Hmm. [20:20] Is there some good reason p-space is avoided? If you didn't like it, you wouldn't have to use it, but could still play on that hill, right? [20:20] in p-space you rather think how the rounds go and act accordingly, when there is no pspace you have to give your all during rounds with no chance to think [20:21] I have to say i dont know myself why [20:21] I havent used p-space myself because i am not exactly pro yet [20:21] Well, i have used switchers, but not created anything competitive [20:22] i think people are more interested in polishing single warriors, instead of improving multiple warriors and the p-brain at same time [20:22] what defines a switcher? [20:23] well, at p-space position 0 is the result of last round [20:23] Sometimes p-switchers contain a bit too specialized components, to beat other p-switchers. Those components often have problems to beat heavy defensive opponents, like paper/imp. [20:23] and if you store what component you used last round, you can use another component and hope for better [20:24] Neo: well, there is the anti-imp clear that can beat most paper/imps.. [20:24] you mean the d-clear? [20:25] nonono, Nenad used it on oneshot some time ago [20:25] an smd-clear [20:25] (ssd, smd, you get it?) [20:25] mmm, i have to see those clears... [20:25] spl/mov/dat i guess [20:28] * tax0n checks out CoreWin [20:30] there have been couple of times with contest about creating best possible p-brain [20:31] I remember Neo winning one with switcher that tries to predict its opponent's decisions [20:31] but it was so big and slow that it worked only for such competition [20:32] I think I might like that kind of competition best. [20:32] I have reduced the size :) [20:33] can you interact with your warrior during a round or between rounds - i.e. receive reports or issue commands? [20:33] nope [20:34] I think they're basically dead data outside the round. Just source code. Nothing is running to interact with. [20:34] exactly [20:35] the code has to be run through a simulator so no problems with your warriors running loose :) [20:36] the most common P-Brain is currently P^3 which is able to work very fast, as the decisions are put in a table beforehand [20:36] Yeah, it would be embarrassing if it got loose and took over the Internet or something. "I swear, it was supposed to be a Core War player. I didn't /mean/ for it to be a sentient computer virus!" :P [20:38] the decision table is the heart and can fry your head when you try to think the best switching method [20:38] Something Prolog inspired would probably work well. [20:41] and warrior that used to be #1 at the draft (p-space enabled) hill called Grilled Octopus switched between three switchers(!) [20:44] MSG: Ping timeout: 255 seconds [20:44] Join: Neo joined #corewars [20:45] and then there is combatra which is piece of genious, but i dont think ill explain it today [20:46] which hill does combatra play on? [20:46] currently non [20:47] none? [20:47] you can find it at the ociw.edu address in storage [20:48] thnx [20:48] it couldnt score reliably evenly, because it took time for it to learn about its opponent [20:49] sometimes it would learn what it wanted during the first fights in a battle, sometimes during last ones etc. [20:52] man - I'm going to have trouble learning this game! [20:52] dont worry [20:52] Bah, forget p-space now, that's quite advanced... [20:52] though - I want to learn assembly and low level programming and this is probably as fun a way to do it as any other [20:54] I'm watching TV drinking beer atm - I'll give it a proper shot tomorrow when I've a clearer head. [20:55] how long is it this game 's been going on, '84? [20:56] well, first tournament was in '86 [20:57] anyone still around from then? [20:57] Chip (winner of the tournament) is writer of corewin [20:57] and he is doing well in the most current tournament [20:57] cool [20:57] but doesnt send warriors to the hills [20:58] Yeah, he did a very succesful comeback [20:59] so the hills and the tournament are separate: like the hills host the battles but the tournament hosts the war proper? [21:00] tournaments consists of multiple different rounds, each round using different rules [21:00] hills are generally automated, tournaments are usually run by hand or script [21:02] Wasn't Core War proposed back when undergrads had to keep coal shoveled into the computers constantly. The late 50s or something? :P [21:02] i believe you mean Darwin [21:02] http://en.wikipedia.org/wiki/Corewar [21:02] Could be. [21:03] MSG: Ping timeout: 255 seconds [21:06] Join: Neo joined #corewars [21:06] :S [21:06] right folks, I'm off! Thanks for the chat, hope to see you around later, bye all. [21:07] bye [21:07] MSG: Quit: Leaving [21:13] Morris' warrior is interesting. [21:14] *blink* *blink* [21:14] dont remember any Morris right now [21:14] i blame it being midnight here [21:14] It's described on the Darwin page. [21:15] It basically uses "antibodies" (that's how I'd describe it). [21:17] the starting position in corewar is varied, so there is no such strategy to exist. Or, sort of is [21:18] As pmars has bit of problems with the randomization of starting positions, you could exploit it with some work [21:19] but such exploit had other problems, which means they cant survive [21:19] (not only because of repairing it) [21:21] Well, the starting position aspect isn't what I found interesting, it was the fact it spawned custom copies of itself to take out certain kinds of programs. [21:21] Join: Core_old joined #corewars [21:21] hi core_old [21:21] hi :-) [21:22] John? [21:30] nah, other person [21:31] oh [21:43] time to go for me [21:43] * Neo waves [21:43] MSG: