[00:02] MSG: Read error: Connection reset by peer [10:50] Join: Core_old joined #corewars [12:36] MSG: Ping timeout: 255 seconds [12:56] Join: Core_old joined #corewars [13:11] MSG: Ping timeout: 255 seconds [15:34] Join: Core_old joined #corewars [16:53] Join: sf_ghoul joined #corewars [16:59] Hello [17:08] MSG: Quit: Trillian (http://www.ceruleanstudios.com [17:41] MSG: Quit: Bersirc 2.2: Looks, feels and sounds (?!) different! [ http://www.bersirc.org/ - Open Source IRC ] [19:03] Join: Fluffy joined #corewars [19:03] :) [19:06] hi floff [19:06] Hi Mizcu ... do you still have o- and u-key swaped? :) [19:07] nu [19:08] Mizcu: Did anything interesting happen during the last days? [19:08] hehe [19:10] I guess, that means no [19:11] my paper/imp turned out to be crap [19:11] *sigh* [19:12] is it on the hill? [19:12] didnt get on either [19:13] what kind of paper did you use? [19:13] I've tried to write a p/i, too, but failed .. probably because I liked the idea of an anti-imp-paper + imp [19:16] hmm ... before I forget [19:16] Roy: I've read you emails and will answer soon [19:16] Sascha: dito :) [19:40] * Fluffy waves [19:40] MSG: Quit: fluffy.i < 1, # 42 [20:26] Join: John joined #corewars [20:26] Hi [20:28] hi John [20:28] Hi Mizcu [20:28] I still haven't got a name for my Pixel site! :-( [20:28] pixie [20:30] Hmmm... I think pixie has gone [20:31] Gone :-( [20:31] "dots-freakin-dots" [20:31] ? [20:32] :-) [20:32] www.pix.ie [20:32] Nice name ;-) [20:33] I can see it now. "Dubya-dubya-dubya, dot, dots freakin dots, dot, com." [20:33] But I would prefer .co.uk because it's free with my hosting [20:34] You probably wouldn't want snort.co.uk (I'm probably not the first to think of that one either :P ). [20:34] pixelcrime is free [20:35] tiny-graffiti ? [20:35] cartesian-quarter ? [20:36] pixelgraffiti is free :-) [20:38] pixelparty and patchworkpixels are also free [20:43] Does anyone think pixelcrime is a bad idea? [20:44] yes [20:44] :-/ [20:45] sounds like some alternative computer-graphic group [20:47] patchworkpixels then? [20:47] no comments [20:49] :-) [20:55] * John waves [20:55] splashdot :P [20:55] MSG: Quit: mov.i #1,1 [21:08] *yawn* [21:39] Hmm, I just started wondering about something... Is it possible for a Core War battle to go into an infinite loop? It seems like it would be, ping-pong style. [21:39] if core is overwritten by an imp, yes [21:39] or the whole core is overwritten by anything non-lethal for that matter [21:40] Does that happen often? I guess they have some limit on running time and just call it a draw. [21:40] yes, and yes [21:41] just look how many ties there are going at the koth'nop [21:41] Hmm. [21:41] though not many of them are imp-sourced [21:41] mm [21:42] never mind, half of them are from imps, rest are from paper vs paper -battle [21:43] the time for draw in normal core is 80000 cycles [21:43] which is little less than i think is good [21:44] normal core being 8000 instructions, max 8000 processes [21:45] I suspect things would tend to draw out longer the better the AI was (assuming both sides are pretty well written). [21:45] well, its actually rather spikey [21:46] How's that? [21:46] stones defeat or lose to scanners in under 10k cycles, paper vs paper and stone vs stone is almost always draw, and scanner vs paper tends to be resolved either early or very late [21:47] if someone would figure out a way to make scanner more resistent, then your hypothesis would be true [21:47] but currently multi-process scanners are beyond our ability [21:48] But if someone wrote something complicated like a switcher it would probably take a fairly long time. [21:49] switcher tries to choose best component against the opponents one [21:49] if battle stretches out longer and longer, the bigger is the chance for something to go wrong [21:50] Switcher A starts with tactic 1 of 2. Switcher B (roughly same design) starts with tactic 2 of 2. Switcher A notices and switches tactics to 2. Switcher B notices and switches tactics to 1. Wash, rinse, repeat. [21:50] I suspect that doesn't happen often but with lots of adaptation one would probably not die real fast. [21:51] the uncertainity-rating will throw a spanner to ping-ponging [21:51] as "best" doesnt mean 100% rate [21:51] and nearly all switching warriors have some threshold for switching [21:52] plain switch-on loss is bad with the uncertainity, because if your good component loses once with bad luck, you switch to a warrior which has even bigger chance to lose [21:53] I gathered there was no fixed best strategy. I'm still not too sure if Core War is "solved" or not. It sounds like it sort of borderline is, since there tends to be an equal and opposite response to everything. But it looks like the "fine points" keep it from being something you can approach mindlessly. And as I said, at this point, I'm not sure that impression is even accurate. [21:55] there are still things to discover [21:55] like Inversed did some reseach with paper-defeating papers a little time ago, and it turned out to be easy to turn even a normal paper to defeat other papers [21:56] and p-switching is still quite murky area [21:57] If nothing else, being able to react to your opponent should stay an open problem, since you don't know what you'll face. The risk of a move in chess for the general case is pretty predictable if you look lots of moves ahead, but knowing the likelihood of a particular move is probably important (I gather that's what Deep Blue did to win against Kasperov). [21:59] well, there are circumstances in which you can react to your opponent [21:59] even without p-space [21:59] Fizmo created few years ago "Argon" which is a scanner that switches to paper if it meets many bombs in core [21:59] I'd been wondering about that since I suspected you could make 'computable storage space' somehow, no matter what, given code = data. [22:00] Argon did kinda badly, but the point is that the analyzing can exist [22:12] me goes catch some sheeps --> [22:12] MSG: [23:58] Join: Core_old joined #corewars