[00:26] name sounds familiar [00:27] Join: MAD joined #corewars [00:27] Hi Mark :-) [00:27] Hello! I finally made it. [00:27] It's a bit quiet tonight I'm afraid, holiday season! [00:28] I just updated http://corewar.co.uk/icwt1986 :-) [00:29] No problem. Just wanted to say "Hello" somewhat faster than email. [00:29] :-) [00:29] I'm replying to your most recent email right now. I'd do it here, but then the conversation would be broken across two separate paths. [00:30] There's been some interesting activity recently. [00:30] Evolved warriors are competitive with hand-written warriors. [00:31] Just in small cores? [00:31] In fact, there's an evolved warrior at the top of the nano hill right now. [00:32] Also in coresize 8000 [00:32] I'm hoping there'll be one at the top of tiny soon. [00:33] Hi Bvowk [00:34] I might have to take a look at one of those evolved programs to figure out what it is doing. [00:34] There's a few articles listed here http://corewar.co.uk/biblio.htm#evolving [00:37] Did you ever get chance to visit the computer museum before it closed? [00:41] No, I never did. I was only in Boston one time, and even then only in the airport, standing in line for hours, trying to make a connecting flight to New York City. [00:42] * Metcalf lives in Boston, but unfortunately it's Boston U.K. :-( [00:43] I have to go for now. Nice chatting. I'll do it again sometime, for certain. [00:44] Yes, nice to see you here :-) [00:44] If you drop in after the holidays they'll be more people around [00:49] My barometer thinks it's going to snow! [00:58] It's time to get some sleep. [01:03] MSG: Quit: flyduck [02:30] MSG: Ping timeout: 240 seconds [02:51] MSG: Ping timeout: 240 seconds [04:03] Part: MAD left #corewars [06:01] Join: Mannerisk joined #corewars [08:01] MSG: Read error: Connection reset by peer [08:43] * Metcalf wonders [08:43] mornin [08:43] Hi Mizcu [09:14] MSG: Ping timeout: 240 seconds [11:41] Join: NorthStar joined #corewars [12:49] Join: _JoE_ joined #corewars [12:49] <_JoE_> Hi [12:51] hi [12:55] Join: yoR joined #corewars [12:55] Happy New Year! [12:55] * yoR is early today [13:05] MSG: [13:12] Hmmm... [13:12] Hi _Joe_ [13:13] <_JoE_> Yo :) [13:13] Are you new to corewar? [13:14] <_JoE_> yes :D [13:14] <_JoE_> just begin yesterday, i try to understand all this stuff [13:15] What have you programmed before? [13:15] How did you discover Corewar? [13:16] <_JoE_> in the book "the art of computer virus" [13:16] <_JoE_> by peter szor [13:16] <_JoE_> he talks about the game in the introduction [13:17] * Metcalf looks it up [13:18] <_JoE_> http://www.amazon.com/Computer-Virus-Research-Defense-Symantec/dp/0321304543 [13:18] <_JoE_> this one ;) [13:21] 44, out of print [13:21] (thats the Euro-price.. typical) [13:25] <_JoE_> in nano-warrior 3, i read "18 of the top 50 warriors are written [13:25] <_JoE_> by hand. [13:25] <_JoE_> what are the others ? [13:25] <_JoE_> written by foot ? [13:25] <_JoE_> :P [13:26] by an evolver [13:26] <_JoE_> ok, there are some random values and this evolver test the best values ? [13:26] values, instructions, modifiers, etc [13:26] <_JoE_> oh ok [13:28] <_JoE_> so the best players have their personnal evolver ? [13:28] not always [13:29] <_JoE_> ok [13:30] <_JoE_> http://rapidshare.com/files/178446154/The.Art.of.Computer.Virus.Research.and.Defense_MAZ.chm.html [13:30] <_JoE_> if u want to check ;) [13:30] <_JoE_> and you can buy it after :) [13:36] Join: fiveop joined #corewars [13:36] Hi Fiveop [14:02] MSG: <_JoE_> Ping timeout: 240 seconds [15:27] Join: sh0ne joined #corewars [15:36] MSG: Remote host closed the connection [16:29] http://en.wikipedia.org/wiki/How_does_one_patch_KDE2_under_FreeBSD%3F [16:37] Join: fish joined #corewars [16:41] MSG: Client Quit [16:54] Join: Mannerisk joined #corewars [17:31] * Metcalf has been forced to wear a suit :-( [17:31] I'm off now, I hope everyone has a good new year's eve! [18:44] Join: flyduck joined #corewars [18:44] hi guys [18:46] hi [18:57] whats up, Mizcu ? [18:57] no quite much and im fine with that [21:23] Heya flyduck [21:23] hey AndrewBC :P [21:23] How goes it? [21:23] whats up? [21:23] Tiredness :P [21:24] pretty good, stopped hand-writing again, once again working on my evolver [21:24] I had to either sleep, or work on the SICP textbook last night [21:24] I chose to forego sleep in favor of learning [21:24] whats SICP ? [21:24] MIT course textbook called Structure and Interpretation of Computer Programs [21:25] cool : ) [21:25] lambda calculus ? [21:26] well it's going to use Lisp heavily, so I assume it will have things of that sort, yes [21:26] (damn, now it hit me, feel like having head stuffed with cotton) [21:26] I need to break out of my shell and learn some functional programming, and it seems to be a great course for general "computer science" theory, as well [21:26] * AndrewBC pokes Mizcu's head [21:26] feels like it's stuffed with brain and skull! [21:27] functional programming is neat : ) [21:27] :) [21:27] AndrewBC: ive been having couple stiff drinks for the new year; im too old to throw firecrackers, too geek to bother going out with friends [21:28] Mizcu, peace and quiet sounds good to me. [21:28] I'll be taking care of a baby for half the night, so roommates can go shopping, or out to eat, or whatever they're doing. Then I'll be going to a friend's house for pizza. [21:29] She's dating a preacher, so there won't be any alcohol, though. [21:30] I only drink like twice a year, but this year it looks like it may be once :( [21:30] or none! [21:40] I was thinking about writing an evolver in lisp... [21:41] perhaps it would be easier to create some smart systems [21:41] Mannerisk, why lisp? [21:43] it's used in AI because of the way it works [21:43] well, from what the authors of SICP have to say about Lisp, it's great for AI [21:43] so I could see it being great for an evolver as well, because it's apply-eval loop [21:43] yes [21:44] also, I'd love to rewrite a simulator to use the GPU [21:45] ah but a genetic algorithm is really just some simple bit-manipulations, unless you wanted an evolver that itself 'evolved' with AI i dont see how lisp would be better, was what i was saying :P [21:45] the brute force methods are making slow progress [21:45] flyduck, could write self-modifying corewar warriors in Lisp [21:45] lisp is good for programs that rewrite itself, but a genetic algorithm remains the same, while it produces new generations, you know? [21:46] a self-modifying warrior is actually closer to reality, methinks. It should react to it's environment... [21:46] so I'd like to find a way to do more intelligent evolving [21:46] ah, but to evolve self-modification, do you need a self-modifying algorithm? see what im sayin :P [21:47] it'd add another level of complexity and it certainly could work, definitely [21:47] but i was just saying i think imperative languages are probably easier for straight-forward genetic algorithm implementations [21:48] oh, easier overall? maybe. [21:48] but Lisp would make it easier than other languages, at the same task [21:48] no doubt, if you are just mutating bits of a warrior, C is just fine [21:49] Lisp could help to do it to all the warriors simultaneously perhaps, though that might not be useful [21:49] i think if you use lisp, you start to get into territory where your methods of evolution are themselves evolving [21:49] which is an exciting idea :o [21:50] however, if it is easier to represent the warriors in a meaningful way in Lisp, you can make intelligent, not random, changes [21:51] the meaningful way is theoretical, but it would be a huge leap in AI [21:53] and my other idea is to use CUDA to rewrite pmars [21:53] ideally has a 20x speed gain [22:10] Join: tom joined #corewars [22:13] happy and prosperous 2009 from EET [22:37] what's EET [22:38] eastern european time? :P [22:38] Oh [22:39] i dunno, some timezone that's over 9 hours away [22:39] east europe sounds about right [22:39] I thought it was a club or something [22:39] Also, Mannerisk, flyduck, I would agree, it would be fairly exciting to see that in Lisp [22:46] version 0.1 of my evolver is about to go into effect : o [22:46] i will be trying to evolve an imp-gate, heh [22:47] :D [22:47] good luck, mate [22:52] hmmmm [22:52] hehe :D [23:00] well [23:00] right off the bat, i got JMZ.I #0, #0 [23:00] which ties a basic imp :P [23:00] interesting ! [23:03] it evolved an imp ! :D [23:06] well that was a failure [23:06] kind of [23:06] sort of [23:09] havent touched my evolver since the upgrade of system - until now [23:09] seems like ok speed while running only on single core :] [23:10] (im holding another for other uses) [23:10] um.. [23:10] speaking of lisp evolvers.. [23:10] I've got exhaust bindings for sbcl and chicken scheme if anyone wants them. [23:11] pretty much the same as the python evolver. [23:11] just using exhaust and FFI [23:13] ok, i just got [23:13] DJN.I$0,<-5 [23:13] MUL.I>43,#5 [23:13] SPL.I>13,*-15 [23:13] o_O [23:14] hehe [23:15] i have no clue how that works, but it seems to be beating the imp.. [23:22] its the djn 0, <-5 [23:23] also, hi bvowk [23:24] bvowk, I would like that :D [23:24] dunno if I'd ever use it, but I'd probably peruse it! [23:24] im enthused [23:24] but ive got a long way to go before this evolves competitive warriors [23:25] versions 0.1 is a success though : ) [23:25] progress is progress! well done [23:26] 125 against bench.. (~115 sal-nano) [23:30] MSG: Ping timeout: 240 seconds [23:35] mov spl mov mov mov, bleeeh [23:38] id still prefer to implement "RedTree -approach", but id also rather do something other anytime soon, including schoolwork