[01:57] MSG: Quit: humhum [02:21] MSG: Read error: Operation timed out [02:28] Join: Mizcu joined #corewars [03:19] Join: brx_ joined #corewars [03:23] MSG: Ping timeout: 252 seconds [04:17] Join: johnkw joined #corewars [05:48] MSG: Quit: using sirc version 2.211+ssfe [06:42] MSG: [06:46] Join: Mizcu joined #corewars [08:21] morning [08:21] Nick Change: brx_ changed nick to brx [08:27] 'ning [08:27] * brx has a new webcam [08:31] aaa pilot of babylon 5 [10:31] Join: fiveop joined #corewars [11:49] Join: fiveop_ joined #corewars [11:52] MSG: Read error: Operation timed out [14:10] Join: Tenoch joined #corewars [14:11] ooh, a new nick [14:12] hi Tenoch [14:12] hi Mizcu [14:12] and everyone [14:12] I just discovered CoreWars [14:12] that's an awesome game [14:13] yup [14:13] I'm just having troubles running the pMars GUI on my mac [14:13] the SDL frontend crashes [14:13] (segfault) [14:14] And I'm too lazy to check the code myself... [14:14] This irc-channel tends to be a bit silent at times, but you can usually just leave questiong here and someone will answer later; everything is logged and watcheable at koth [14:14] ok, great [14:15] I cant help myself with pmars-problems, but there was some talk about getting pmars to compile under MacOs at rec.games.corewar [14:16] I'll check that, then :) [14:16] I can run jMARS though [14:19] Mac? Voluntarily? *faints* [14:19] =) [14:20] Better faint than say stupid things :) [14:20] Did you get the pmars from sourceforge or koth? [14:20] Koth has ollld versions [14:20] he said SDL [14:20] Ah [14:22] Mizcu, When is the next tourney [14:22] Tourney = motivates me to actually play : / [14:22] Planter is doing as expected on 94b though [14:23] He fears no change. [14:23] Bring on the stones! [14:24] data, no idea [14:24] I'll host on in spring or something [14:24] as there is no volunteer for Frenzy's newest round [14:25] Ah [14:25] I'll do it [14:25] you got a good idea for round? [14:25] As long as I don't have to be creative or anything and pick the theme [14:25] Damnit [14:25] Thought so [14:26] * brx remembers providing gentoo linux with its pmars-sdl ebuild [14:57] silence -_- [15:07] * datagram tumbleweeds [15:09] * Tenoch already hates CoreWars [15:09] ;°) [15:11] It's 7 AM here [15:11] Don't ask too much of me [15:11] * datagram hasn't slept yet [15:11] * datagram probably want sleep until 5 PM [15:13] 4 PM here [15:13] Where ya at [15:15] france [15:16] Far out [15:16] :) [15:22] Gotta run [15:22] See you all later [15:22] ok [15:22] bye ! [16:49] MSG: Quit: Tenoch [18:40] Join: Tenoch joined #corewars [18:40] Hey folks [19:28] . [19:33] hi [19:39] hello, [19:46] Join: Fluffy joined #corewars [19:46] :) [19:47] hmm ... bvowk isn't there :-( [19:48] hello fluffy. [19:48] :) [19:49] I am always willing to help if I can, you know that :) [19:49] Join: Neogryzor joined #corewars [19:49] hi Neogryzor [19:49] hey neo. [19:50] hi all [19:50] Hi Neo! [19:50] brx: I wanted to ask bvowk a little bit about the circle contest ... but that can wait [19:51] ah, well I am not rokicki [19:51] brx: In any case ... the Redcode parser is now working :-) [19:52] Join: beretta joined #corewars [19:52] Fluffy: congratulations :) so how long do you estimate until we have a working version of your speedy mars with beautiful python gui? :) [19:52] Hi beretta! [19:52] hello all [19:52] hello beretta [19:53] hi. Some new names here [19:53] first timer here. [19:53] brx: Don't know. I hope to release PyCorewar 0.1.0 (with new parser an ICWS '94 support) within the next three weeks [19:53] Neogryzor: :°) [19:53] RedCode is not really simple, when you're used to code with C [19:54] Oh, welcome the freaky CW comunity so :) [19:54] brx: Let's simply say, that I hope to have a GUI by summer [19:55] Fluffy: do you have anything special planned? I imagine gui work not to take that long [19:55] Tenoch: That will soon pass and Redcode becomes easier :) [19:55] Manipulating data AND code in the same time is very disturbing [19:55] for the begginer :) [19:57] brx: So far I've planned to support all Redcode standards and then implement some helper functions, which let's you easily create a hill [19:57] brx: after that I want work on the debug-versions of the MARS. then a GUI (with PyGTK) follows [19:57] brx: nothing more yet ;-) [19:58] Tenoch: I've learned it by trying to understand other people's warriors. [19:58] brx: As a byproduct I will have a simple evolver and an optimizer [19:58] newbie question: is submitting a evolved warrior a no-no? [19:59] brx: Well ... I already have an evolver, but it is not that userfriendly [19:59] no-no? [19:59] beretta: You can submit whatever you want :) [19:59] beretta: And creating a good evolved warriors is still a hard task [19:59] Question : you can never see other's players code on KOTH ? [20:00] Tenoch: nope [20:00] ok [20:00] Tenoch: That is the fun part ... to guess, what you opponents do [20:00] But a lot of warriors are published at rec.games.corewar and Koenigstuhl [20:00] yes, I saw that [20:01] I spent the day wandering the web [20:01] I've learned to play with my population, tweaking mutation rates and introducing new species is a must. [20:02] beretta: I'm still trying to evolve something decent, but apart from some simple clears I didn't get anything good :-( [20:02] Fluffy: so you plan on releasing a fullblown redcode ide? :) [20:02] Fluffy: good! :) [20:03] brx: No IDE, only an debugger [20:03] Fluffy: don't be modest [20:03] *a [20:03] :) [20:03] brx: Wait until I've written it [20:03] brx: But at least I already have the second fastest MARS for '88 [20:04] brx: the '94-MARS will be a little bit slower but I'm aiming for about twice the speed of pMARS [20:05] I probably try to rewrite my scripts for the score surfaces to use PyCorewar ... [20:05] It is about time to calculate some new [20:06] Fluffy: I started from total randum warriors, and soon started seeing imp bombers and replicators develope. Sometimes when the population became stagnant, i'd up the mutation rates, and introduce "premade" warriors. [20:06] Anyone one of you use Ares ? [20:07] Fluffy: beating fmars speedwise is probably not a goal [20:09] beretta: What program are you using? [20:10] brx: Beating fmars might be possible but it probably have to break portability for that, which I don't want [20:11] Fluffy: my own. I call it "Chaos", basically a fancy script, frontend for pmars [20:11] brx: And under normal circumstances you won't notice the difference between 9.3 MIPS (PyCorewar) and 10.4 MIPS (fmars) [20:11] Fluffy: remember worse is better. [20:11] aaah [20:12] beretta: Nice ... probably I still lack the competence to write a decent evolver :) [20:12] brx: That was good for the parser, but for speed ... I don't know [20:13] Having written the fastest MARS should be still fun. [20:13] I probably have to invest some time into writing a new warrior for '94nop ... HFB is dead long enough [20:14] "HFB" ? [20:14] Fluffy: I had plenty of fun writing it. I'm looking through an old log file now. it took 5175 generations to double the average score of a population of 1000 played against a "clasic" hill of human made hill. [20:15] "Harmless Fluffy Bunny", a paper [20:15] ok [20:16] * Tenoch reads the code [20:16] wow [20:16] :D [20:16] Fluffy: For what hill do evolve? At the moment I'm trying to get a good '88 warrior [20:16] Tenoch: Don't think of writing individual instructions, but of writing building blocks [20:17] * Fluffy is quite stupid to address himself ... [20:17] Fluffy: :d [20:17] * beretta grins [20:18] Tenoch: note that for a quick corewar snack you can always go to nano hill :) [20:18] Tenoch: A big part of HFB consists of the quickscan, which is quite easy to use, because it mostly the same for every warrior [20:23] Fluffy: I forget... lemme check... I think it's '94 ruleset... *I think* [20:24] beretta: with pspace? [20:26] Fluffy: maybe. are LDP and STP p-space instructions? [20:27] beretta: i think so [20:27] load and save to p-space [20:27] ok... sorry, I haven't played with redcode in 4-5 years. Then yes, my evolver is p-spaced [20:28] beretta: hehe ... it would be nice to have a good and evolved pspacer [20:29] I alway thought it would be cool to create a SETI-like multiple computer evolver to allow for larger populations, and quicker results [20:30] beretta: You should wait until bvowk is back [20:31] beretta: He's quite into nano-evolving have has sometimes a LOT of computers to use for his evolver [20:32] Fluffy: I believe "SETI-like" is a tad greater and might be a fun thing to try [20:32] ahhh... sounds great! [20:32] Fluffy: "come help freaks have fun with your processing power" [20:32] brx: sounds like a good tag-line to me! [20:33] better yet, commit fraud. "help develope a cure for cancer" [20:34] How about "Help to create new score surfaces."? [20:34] Fluffy: sorry. [20:34] :) [20:35] Creating ONE score surface for 94nop should still take about 100 years at the moment [20:35] yes, and think about how useful that score surface will be when we have it! :) [20:35] What do you call score surface? [20:35] !scoresurfaces [20:35] http://corewars.jgutzeit.de/score_surfaces/index.en.html [20:35] oh, bvowk our channel bot is not here today [20:36] whats a score surface? [20:36] hey, that looks great [20:36] :D [20:37] Tenoch: No, it is not about fun ;-) It is a VERRY SERIOUS TOP SCIENCE ;-) [20:37] Oh, I never doubted it :) [20:37] It is a color dot picture that shows a warrior's score depending values of two constants against a certain opponent [20:37] but, it has the adventage of being appealing to the eye [20:37] Fluffy: lol [20:37] ohhh... very nice! [20:38] damn ... there goes the science part ;-) [20:38] Fluffy: Btw, now i can obtain 3-D surfaces and spheres [20:38] Neo: From me? [20:38] err ... forget it [20:38] from the data files... :) [20:39] Neogryzor: Are they online? [20:39] the images [20:39] nope :( [20:39] i think i can crete upload upload with ease [20:41] Hmm ... it is really a pity, that I can't multitask very well :-( [20:41] It "have to" work on PyCorewar, want to update the score surface pages, write the next part of the quickscanner tutorial [20:41] I can't do that all at once :-( [20:42] buy the multithread module [20:42] at your grocer store [20:42] :) [20:42] implant a secondary head :) [20:43] yeah... be just like Zaphod Beeblebrox [20:44] Nobody wants to have a second head like mine ;-) [20:46] How about you all write the texts and code and I only write my name on the top of it [20:46] Sounds quite fair to me ;-) [20:47] i can write the odd words and you the evens :) [20:48] Mmm, anyone uses Ares here ? [20:48] strange behavior... [20:49] Ares: nope. [20:49] Neogryzor: OK, I will wait for your first page then ... [20:50] Tenoch: sorry, but no :( [20:50] ok [20:50] too bad [20:51] Fluffy A ·-D test picture, see http://www.estadium.ya.com/neogryzor/sub1t2.jpg [20:52] A 3-D i mean [20:52] Neogryzor: mmm, I would *eat* that [20:53] it is just a small piece of the score already [20:53] Fluffy: Do you think it could be useful? [20:54] which part is it [20:55] I mean: what part of the surface did you use? [20:55] 0-100,0-100, seen from 0,0 [20:55] stupid me ... I should have looked at the image [20:55] :) [20:55] yap vs muskrat i think [20:56] I definitely like it, but an interactive 3d-model would be nice [20:57] Fluffy: yes. [20:57] If I could rotate the image in realtime, I could probably see more details than I could on the "normal" score surface [20:57] * brx chuckles and unignores fiveop [20:58] i can take smaller or bigger pieces, change color and shadding, even rotate and create a movie, but i cannot change the dimensions of the picture :( [20:58] Neogryzor: Did you use all 100x100 points for the image? [20:58] Fluffy: yes, and if you could walk on it! [20:58] we could build houses on them .. [20:58] plant some trees [20:58] use snow cannons [20:58] and play dark ages [20:58] Fluffy: No, i took a bit bigger matrix, but reduced axis to 100x100 [20:59] hehehe [20:59] Neogryzor: It looks a little bit too smooth to my eyes, but I could be wrong [20:59] Yes, colors are interpolated [21:00] brx: Too late ... while the first score surface has been calculated I took the liberty to name all features on the score surface. NOW IT IS ALL MINE! [21:00] brx: You must pay to live on the score surface ;-) [21:00] Neogryzor: Do you have time to create an image for say 200x200 with as less interpolation as possible? [21:01] yes [21:01] i'm on it [21:01] cool :) [21:04] Fluffy: ;) [21:04] uploading... [21:06] .... [21:06] ..... [21:06] ...... [21:06] hehe ... I know that [21:06] When I uploaded the score surfaces it took ages, too :) [21:07] http://www.estadium.ya.com/neogryzor/yvmtest01.jpg [21:08] you'll see a bigges area than 200x200 because i used the complete score matrix [21:08] bigges equ bigger [21:08] Neogryzor: it might be nice to retain the coloring [21:09] yeaaaah [21:09] i made a self replicator [21:09] which suppositedly protexts himself from imps [21:09] :d [21:09] * Tenoch happy [21:09] Tenoch: Have you sent it to the hill? [21:09] Fluffy: don't, he is happy now. [21:09] :) [21:09] oh good idea [21:10] it'll be the third one to be pushed off before even having got on it [21:10] :d [21:10] gimmpe a second [21:10] Tenoch: Which hill do you use? [21:10] mm [21:10] 94nop [21:10] Neogryzor: Now it looks more like I've imagined it [21:11] though, the only thing i can see there are thos low-score valleys [21:11] Neogryzor: I've hoped to see new structures on it, but there aren't any new :( [21:12] Wanna see houses there? [21:12] Tenoch: Maybe you should try the beginner's hill at SAL? [21:12] Neogryzor: YES! [21:12] Fluffy: isn't it down ? [21:12] Neogryzor: And trees, too, please [21:13] :) [21:13] Fluffy: i can sell you a 50x50 square [21:13] Tenoch: http://sal.math.ualberta.ca/hill.php?key=94b ... no, I don't think so [21:13] Neogryzor: You can't. It is already MINE! [21:13] ok [21:14] Fluffy: it's in euclidean space.. [21:14] Fluffy: Heh, you can only sell psichodelic flat carpets, not 3D :) [21:14] brx: No, it isn't! [21:15] It is a torus! [21:15] Neogryzor: I'm still waiting for the house and the trees ... [21:15] :) [21:15] And as far as I know the geometry on a torus isn't euclidean [21:16] But I've never been that good at topology [21:17] Tenoch: Did it work? [21:17] Fluffy: it might only be a matter of perspective after all! [21:17] * Neogryzor tries to build a whole planet [21:17] Fluffy: i'm going to test [21:17] i sent it on KOTH [21:17] brx: Let's simply assume, that I'm right and you are wrong ;-) [21:18] brx: It is easier for me that way ;-) [21:19] Tenoch: http://www.koth.org/lcgi-bin/list.pl :-) [21:19] * brx shrugs [21:19] Fluffy: yeah, my brave little warrior [21:19] Neogryzor: Trees? Houses? [21:19] empty minds are susceptible to exterior mindstuff [21:19] :) [21:20] * Fluffy 's mind is always emptier than empty [21:20] Especially, if I'm in stupid mode [21:21] Fluffy: you always assume a defensive posture! [21:22] Yes, that way you will be surprised, when I start to attack. I can overrun your defences very easy then :) [21:23] I think my PC is suffering fatigue.. [21:23] And if that doesn't help, I can resort to simple violences :) [21:23] Fluffy: where can I find the email to send the warrior ? I counl't find it on SAL [21:23] Neogryzor: why is that [21:23] Tenoch: [21:23] thanks [21:23] Tenoch: But I think, that you have to register first, but I'm not sure [21:23] It is hard to handle so much data and calculations [21:24] yes [21:24] See http://www.estadium.ya.com/neogryzor/yvmtest03.jpg [21:24] Yeah, I scored 11 on the koth 94nop hill :) [21:24] Unfortunately it is not detailed enough [21:24] 11, wow! [21:26] Fluffy: :) [21:26] Tenoch: hmm ... very hard to achieve such a score, even an imp get around 55 :) [21:26] :D [21:26] yeaaah [21:27] Neogryzor: It is Fluffy's Planet :) [21:27] Tenoch: Simply try again ... it took me ages to get my first warrior onto '94nop [21:28] Tenoch: ... and then it was pushed off just with the next challenge [21:28] i should try to make it better first :) [21:28] Tenoch: That would probably help ;-) [21:29] Tenoch: You can always publish you warrior here or on r.g.c. and ask for help/hints :) [21:29] brx: Still here? [21:29] see also yvmnpole01.jpg and yvmnpole02.jpg [21:31] uh, no, wait [21:31] Neo: I can't access theses images :( [21:31] just a moment [21:31] yes Fluffy [21:32] brx: I still have problems with recursive EQUs [21:32] what is it this time? [21:32] brx: ... and I can't think clearly now ... maybe you can help [21:32] brx: I want to recognize then as soon as possible (without any additional work) [21:32] my head is stuffed with jms notes on babylon 5. [21:33] brx: I have to make three passes to get the final warrior [21:33] brx: Shouldn't it be enought to simply replace all EQUs once, evaluate them and then reevaluate them again [21:34] brx: Everything not resolved by then is recursive, right? [21:35] brx: B5 is always the best :) [21:36] Tenoch: Just found sth., that is way harder than Redcode: http://esoteric.voxelperfect.net/wiki/Malbolge [21:36] http://sal.math.ualberta.ca/hill.php?login=203&key=94b [21:37] i'm on the hill [21:37] yeah [21:37] 16 points :) [21:37] Tenoch: :) [21:37] congrats [21:37] ahah :) [21:37] But 94b usually has no free slots [21:38] Tenoch: In any case: Congrats :) [21:38] there's one more [21:38] i'd have been on the hill anyway :) [21:38] YES and that's all, what counts! [21:38] Give them hell! [21:38] :) [21:38] Yeah [21:38] TenTen rocks [21:39] now i'll improve it [21:39] ok, try the images now [21:40] oh i know [21:40] i'll spread processes everywhere [21:40] I was having only one two the same time [21:41] Neo: works now [21:41] Neogryzor: Maybe I should take the time and use the raw data as input for a raytracer [21:42] Neogryzor: Then I could add trees and houses, too ;-) [21:42] You can accomodat brx on a small area of the north pole :) [21:43] cold but sunny [21:44] And where is the north pole? [21:44] No magnetic field there ... [21:45] at the north... [21:46] Maybe where the lowest score is to be found? [21:46] Unfortunately then there will be several north poles [21:47] Same for sourth pole (= hightest score) on some surfaces [21:48] Fluffy: sorry, no. [21:48] brx: no = not right? no = don't know? [21:49] Fluffy: not right [21:49] Do you have a counterexample? [21:49] Would be nice to have some kind of good test [21:49] Fluffy: just replace recursively, and raise a circular if you encounter the symbol you are trying to resolve [21:49] north pole= 0,0 location [21:50] brx: thanks ... will try that [21:51] argh ... yet another thing, I have to do :-( [21:52] I still haven't published the docs for PyCorewar :( [21:52] * Fluffy will code now a little bit and then get some sleep [21:52] * Fluffy waves [21:53] Fluffy: how did you reach the 3 times conclusion? [21:53] brx: By not thinking ... it is probably too late to think about such things ... will try tomorrow again [21:53] Fluffy: I have a trivial counterexample for parallel evaluation [21:54] which is? [21:54] just a long enough chain of references [21:55] you mean " a EQU b ... b EQU c ... c EQU d ... d EQU e ...... z EQU a"? [21:55] * brx nods [21:56] no, that would be work with 3 passes [21:56] i.e. recognized as resursive [21:56] ... yes it would. [21:56] the parser (PyCorewar's and pMARS') tries to expand every EQU as much as possible the first time it encounters it [21:57] think about a one element longer chain whcih is valid [21:57] ah so it does [21:57] well that is nearly what I suggested anyway. [21:58] recurse into unresolved labels [21:58] why labels? They are replaced with their values in the second pass [21:59] In the first pass all labels and equs are collected (and FOR/ROFs expanded) [21:59] sorry, I refer to equs as labels [21:59] ah ... ok [22:00] during equ expansion, start with one and expand until it is completely expanded. expand other equs as needed. [22:02] that should eliminate the problem of finding circulars. [22:03] How does that eliminate the problem? [22:05] I am A EQU B + 1, I expand, it is needed to expand B EQU C for that, but C equ A + 3, just keep track of what you are expanding at any given stage during the dive. [22:05] you note that you got to expanding C from trying to expand A [22:05] circular [22:06] BUT: [22:07] A EQU C + D ... C EQU C [22:07] yes? [22:07] and D EQU whatever [22:07] I start with A [22:07] and come to C [22:07] what about it? is that valid` [22:07] but there is a circle [22:07] if not I see no problem [22:07] yes, that circle is catched [22:07] I never come back to A to find a circle for A [22:07] yes. [22:08] I should be more clear I suppose [22:08] * Fluffy thinks again ... [22:08] "at any given stage" [22:09] *caught [22:09] I think, I get it now. [22:09] you expand C for A, but also for C. if during the expansion of C C or A is referenced [22:10] so the above example is not an issue [22:11] you have the call stack [22:11] Yes and that's what I wanted to avoid, but it seems, that there is no way around [22:11] mh [22:12] I don't see why you should want to avoid it. you have to expand each equ anyway. the problem is the checking overhead. [22:12] building a graph is of course overkill [22:13] * brx tries to think this time [22:14] actually I don't think there is an easier way to check for circularity [22:14] but I am listening to hiphop currently and it is distracting. [22:15] At the moment I'm back at 3 passes thingy ... [22:15] I find an EQU ... all following occureces of that EQU will immediately replaced by it [22:16] hrm [22:16] Apart from a check of selfreferences (a EQU a + 1) I've now eliminated that EQU [22:16] right? [22:17] go on [22:17] Let me start again [22:17] We have an EQU: a EQU ... in line 10 [22:18] After line 10 there will be no reference to 'a', if I replace 'a' immediately [22:18] yes [22:18] Before line 10 there still might be references [22:18] In the next pass, I will replace all possible 'a' in lines < 10 [22:19] So after the second pass there will be no more reference to 'a' [22:19] right? [22:19] if there is no selfreference, no. [22:19] I mean yes [22:21] If I do that with all labels after the second pass there will be only labels left, that are recursive, right? [22:21] So in the third pass (when I evaluate all expressions) and find a label, I immediately know, that it is recursively defined, right? [22:23] ah ... hell ... I will simply look at the sources of pMARS and see how it is done [22:25] * Fluffy waves [22:25] (time to get some sleep) [22:25] Part: Fluffy left #corewars [22:25] * Neogryzor waves [22:43] fluffy won't like seeing the recursive version in pmars. [22:48] time to go for me too [22:48] * Neogryzor waves [22:48] * brx waves [22:48] MSG: Quit: Physical Wreck IRC webscript: http://www.pswclan.com/cgi-bin/irc/irc.cgi [22:57] MSG: Quit: Tenoch [22:57] Join: Tenoch joined #corewars [22:57] MSG: Client Quit [23:01] Join: Tenoch joined #corewars [23:10] baaack