[07:11] Join: fiveop joined #corewars [13:06] Join: Core_old joined #corewars [13:42] MSG: Quit: Trillian (http://www.ceruleanstudios.com [13:45] Join: Core_old joined #corewars [14:05] MSG: Quit: Trillian (http://www.ceruleanstudios.com [16:51] MSG: Read error: Operation timed out [16:55] Join: fiveop joined #corewars [16:55] MSG: Client Quit [16:55] Join: fiveop joined #corewars [17:01] MSG: Quit: humhum [17:02] Join: fiveop joined #corewars [17:17] Join: bose joined #corewars [17:20] MSG: Read error: Connection reset by peer [17:21] Join: bose joined #corewars [17:21] hi [17:22] MSG: Ping timeout: 255 seconds [17:25] MSG: Read error: Connection reset by peer [17:32] hrm. [17:56] Join: fiveop joined #corewars [18:10] MSG: Read error: Connection reset by peer [18:12] Join: johnkw joined #corewars [18:30] Join: Fluffy joined #corewars [18:31] :) [18:51] bvowk: Have those 3 warriors been your last K-toppers? [18:56] Join: Retro joined #corewars [18:56] evening [18:57] :) [18:57] hey fluffy [18:58] cooked up anything to battle the borgs yet? :) [18:59] Why should I? [18:59] White Moon and most of my other warriors do just fine. [18:59] dunno :) any competion is always nice [18:59] Yes, being koth is nice, but having a really old warrior is, too :) [18:59] true enough [18:59] mine still have some aging left to go [18:59] lol [19:00] yes, but they seems to have a good time atm [19:00] *seem [19:00] yup :) [19:01] I also managed to get a silk onto nano, but it struggled to even keep place 50 [19:01] a nano-paper? [19:04] yup [19:04] a nanosilk [19:04] :) [19:04] using 4 threads :) [19:05] and a payload of 3 instructions and 1 dat being copied [19:05] I might adapt it for tiny, as it'd be hell of a lot harder to destroy there [19:05] (and ofcourse I can add more payload that way) [19:06] The difference between nano and tiny is quite large. So a working nano-warrior usually doesn't do very well on tiny [19:11] yeah naturaly I don't mean to port the nano warrior directly [19:12] but the way it works is interesting as it can be extended to other core sizes and the size of the code that stands to be damaged is rather small [19:16] * Retro submits a quick test [19:21] hrm [19:21] * Retro tweaks and fiddles :P [19:32] still tweaking and fiddling? [19:33] yeah I just tried the paper itself on tiny, impressive amount of draws :P now I have to figure out a nice payload [19:33] I've got some thoughts on it :) [19:33] qscan? [19:34] thinking more in the line of a local bomber [19:34] I could add a qscan too ofcourse, maybe even as a pre-paper stage [19:35] scan -> no luck -> paper + bomb, could be funny [19:37] yes, but so far doing two (longlasting) things after another hasn't been a good idea [19:39] Another thing that you might want to try is to optimize your warriors. Or are you already doing that. [19:39] bvowk: Are you there? [19:41] kinda. [19:41] what do you want? [19:41] hence why the scan would have to be really quick :) [19:41] yeah fluffy I tend to optimize them, but tiny gives me odd results sometimes [19:41] Do you still remember The Human Machine? [19:42] How did you create it? Did you seed your evolver with papers? [19:42] it was evolved w/ ccai [19:42] which was seeded with K [19:42] er.. [19:42] the *CHAINS* were seeded with K, the code ware random.. [19:43] basicly it was a human tweaked version of MWR [19:43] So you didn't push your evolver to create papers? [19:44] well, I wanted it to evolve papers.. [19:44] thats what the chains were for.. [19:44] ok, thanks :) [19:44] to see if I could get papers/scanners by spitting out the right ratios of instructions [19:47] I'm having trouble evloving for tiny using yace, for some reason all the scores are rediculously high, but the programs are crap... [19:47] I tried several testsets but no avail [19:48] Unfortunately I've never *used* yace to evolve sth. So I can help :( [19:48] retro: the biggest problem with evolving is getting a benchmark that reflects the hilll [19:48] try my jnsl, it kinda sidesteps that ;) [19:49] jnsl? [19:49] it makes its own hill [19:49] hehe [19:49] http://www.math.ualberta.ca/~bvowk/files/jnsl-1.1.tgz [19:49] its in python tho.. [19:49] yace makes it's own hill too if you ask it to, I also tried using the top 10 koeningstuhl as a testset [19:50] bvowk: Don't apologize for using Python :) [19:50] it's not bad, great for scripting [19:51] Yes, you can do the benchmarking with a C-Module and do the rest with a few lines of Python :) [19:52] I mean I do tons of stuff in C(++) but I can definately appreciate any language that is simple fast and robust to write in for those things that don't "need" it [19:52] heck I use a quirky language called blitzbasic 3d for a lot of my "fast" work [19:53] I wasn't appologizing for using python.. [19:53] it's not exaclty what you'd call forced correctness, but it works like a charm :P [19:53] As long as a language does what you need I'm satisfied, execpt maybe if it is LITHP ;) [19:54] I was just pointing out that it was in python and its not really.. uh.. standalone ;) [19:54] hehe [19:54] that's ok, I'm well versed enough to get it to run :) [19:55] I really find lisp a pleasure to work with fluffy.. [19:55] you should try it :) [19:55] I hacked jnsl in lisp first.. [19:55] I have, but taunted you is fun, too :) [19:55] *taunting [19:56] and then redid it in python because it's easier to get running on the wildly ecclectic assortment of machines I've got access to [19:57] "eclectic assortment" ... hmm [19:57] they are just computers [19:58] running random assortments of unix, on random processors :) [19:58] And how many of those random computers are running your evolver now? [19:58] uh.. [19:58] none ;) [19:58] but thats not the point! ;) [19:59] none? [19:59] about what about the non-random computers? [20:01] I've got it running on only a few [20:01] new=20? [20:01] 10? [20:02] err ... [20:02] well.. [20:02] *few [20:02] 10 or so, and mitzi [20:04] I still can't understand why you don't start to hit the tiny/lp/94nop hills [20:04] its ok, you don't understand what mathematicians do either. [20:04] lol [20:05] one day, they may be able to help you with your shoe drooling problem too. [20:05] lol [20:05] (not the mathematicians.. they don't care..) [20:05] so bvowk those new entries of yours evolved using that network? [20:05] all of them are, Retro! [20:05] MSG: Read error: Connection reset by peer [20:05] all! [20:06] Join: Roy joined #corewars [20:06] retro: I have access to hundreds of processors [20:06] likely more computrons than you can possibly imagine. [20:06] Hoi Retro (and the rest!) [20:06] Nice work on Nano, impressive! [20:07] bvowk: I still remember one of my first courses at university. There was this guy, who insisted on having a towel on his left should. ALL THE TIME! That probably has shaped my view of mathematicians ;) [20:07] *shoulder [20:07] He probably also thought the answer was always 42? [20:08] Join: johnkw joined #corewars [20:08] and no, he didn't know the Hitchhiker's guid to the galaxy then [20:08] fluffy: you don't have a towel!? [20:08] *guid [20:08] tnx :) [20:08] * bvowk snaps fluffy with a wet towel [20:08] No, all the dirt tends to fall off after a while without me doing anything [20:08] bvowk: I don't know I can imagine a lot... :) [20:09] Hi Roy! Are The Netherlands already flooded? It is raining here! [20:09] nah it ain't flooded here [20:09] retro: there's a room here with 900 cores in it! and thats just one of many! [20:09] Pretty dry here, I went to Winterberg this saturday, some serious storm-damage there! [20:10] working at a hosting company or render cluster setup or something? [20:10] no, I work at a university [20:10] Retro: You forget the obvious! [20:11] ah yes :P [20:11] naturaly [20:11] with mathematicians! [20:11] bvowk: Did I mention that I know a little bit about math, too? [20:11] ;) [20:11] and chemists next door, physicists to the south, and engineers to the west! [20:11] terrorists! [20:12] * Roy sucks at math [20:12] they use chemicals! [20:12] they have to be terrorists [20:12] hehe [20:12] fluffy: this whole place is red circled, you know that knowledge is dangerous.. [20:12] Somebody told me a little riddle yesterday, and I didn't believe it at first (now I do) [20:12] even the sad little arts students could be deadly if placed near an american.. [20:12] bvowk: knowleedge? what? [20:12] so bvowk why haven't they beaten the borgs up yet? :) with all that computing power... [20:13] (americans shoot to kill if you make them feel stupid) [20:13] retro: because I've been told not to abuse the hill that badly.. and I've already got like 15 warriors on the hill [20:13] bvowk: Don't they shoot themselves first? [20:13] haha ok ;) [20:14] This is how it goes: There is game show with three doors, behind one door is a price. You are allowed to pick one door. After that the gameshow-host opens a empty door and asks: Do you really want to go with the door you picked or do you want to switch? What do you do....? [20:14] switch [20:14] switch [20:14] And why :) [20:14] roy: quit gambling and get a real job! [20:14] just calculate the probabilities [20:14] more chance of winning :D lmao [20:14] it's simple maths really [20:15] as fluffy says [20:15] Yeah, when you switch you indeed have 2/3 change to win, but its pretty unnatural feeling, it feels like a 50/50 change when I heard it at first [20:16] Roy: That's the first rule of doing math, that you have to learn. Don't think! [20:16] But when you apply some simple math you get 1/3 - 2/3 :) [20:18] bvowk: I actually have a pretty decent real job, I program Java.. and if you just throw around self-made-up-design-pattern they think you are very smart and they pay you more [20:19] I'd feel dirty using java all day ;) [20:19] I feel dirty supporting java for people.. [20:20] Wouldn't that depend on the amount of money that is given to your? [20:20] *you [20:20] not really, I feel dirty and stupid when I have to work on windows.. that won't go away no matter how much I get paid [20:21] I have to disagree with that, Java is a pretty clean language but it takes long to make something nice with it... obese but readable/understandable, and the longer I take, the longer I get payed ;-) [20:21] if it didn't hurt to use inferior tools, nobody would bother making better ones! [20:21] I feel only frustrated, if I have to work with Windows. [20:21] bvowk: Who says I have to work on Windows, Java is 100% designed to be platform independand [20:21] roy: that is *SO* much bullshit tho! [20:21] Roy: No, it isn't [20:22] java doesn't work everywhere, it hardly works anywhere.. so badly so that companies ship their own JVM to go with their programs! [20:22] GO JAVA :P [20:22] so you'll end up with a couple dozen shit ports of JVMs on your box because they all need the magic version w/ the magic patch for whatever. [20:22] Well, not all ideas work very well, but it was DESIGNED to be cross-platform :) [20:22] they are switching to teaching java in college these days [20:23] its even so bad you'll get the same version of the program using different JVM ports for different operating systems.. [20:23] and let's not even get into java on mobile phones... :P [20:23] no, don't, experience with that :( [20:23] or the fact that the JVM is a wildly flailing target [20:24] so code that works today, might not work tommorrow, but it could work the day after.. [20:24] haha poor Roy [20:24] I agree on all those things, but the code we make runs on multible JVM's on Unix (server) and on Windows (during development) [20:24] yeah Roy, I can see the good sides of java obviously [20:24] despite it's failings... [20:24] getting paid much? [20:25] nah I'd get paid more coding in C(++) I'm sure [20:25] but doing a large multiplatform project in that isn't exactly ideal either [20:26] Its not always bad, just most of the implementations aren't very good.. If you want to counter that you should look at .NET, one implementation (Microsofts) and no hassle with different implementations etc, just one way... [20:27] All the tools are made by the same company, very good right? But then you lack a lot of open-source tooling and its also a BIG drawback to be Microsoft dependend [20:27] (costs a lot..) [20:27] oh yeah roy, I was just saying that crossplatform C(++) isn't ideal either [20:28] <-- unixy [20:28] So you can't just bring Java down by saying it doesn't have 100% compatible JVM's all the time without giving us another (better?) option :) [20:29] I spew one liners of perl/python/awk like some people make TV quotes :) [20:29] C is a better option.. [20:29] C usually just works ;) [20:29] * bvowk ducks [20:29] Like I said, our new piece of server code (5000+ lines of code) works perfectly on Unix and on Windows on the same JVM with the same settings [20:30] Its not always bad, and it pays good enough for me ;-) [20:30] * Roy is even getting some more Unix action lately messing in the cron etc [20:32] yeah bvowk, there is the other point that in general Java might be on the rise because there isn't enough "great" programmers anymore... [20:33] we don't teach programmers to be great anymore, we hobble them with java.. [20:33] like your mother always said, it stunts your growth. [20:33] point being that C(++) is faster and offers more options, yet is simply more complex and requires more dicipline to be even remotely re-useable [20:33] you have to understand the metal under you to really be able to program effectively.. people keep trying to convince themselves otherwise, but they are fools. [20:34] Things like this:http://markbyers.com/moinmoin/moin.cgi/ShortestSudokuSolver proves Java is bulky :-P [20:34] teach someone how to use a stack pointer, and you've taught them how to program. [20:34] nice philos [20:34] philosophy bvowk* [20:34] bvowk: They are going to add some more pointer-like things in Java 7 btw, not working around it anymore [20:35] roy: that doesn't mean students are going to go anywhere near that corner of java.. [20:35] you really can't program a machine intelligently without understanding the metal.. [20:36] Sure they are, where I studied they started with the most basic stuff like making a linkedlist in C and using the pointers etc, of course you have to know these things when you are going to program Java [20:36] now, I'm likely a little biased, as I've got a computer engineering background, but every great programmer I've ever met knows exactly whats happening under the hood, why it has to happen that way, and has a long list of ways it could be improved. [20:37] bvowk, that's true, but those programmers often are too expensive for non performance, non cutting edge stuff such as business apps... that's often why these new java only guys are popping up [20:38] I also think as a programmer you should know the the O and when you should use what kinds of lists/maps/arrays [20:38] Roy, I think that would be one of the most basic requirements, better would be to have implemented each of them before. [20:39] they're only expensive if you're running the current IT industry standard of not solving problems.. [20:39] yeah I know bvowk, it's not like I agree with them :) [20:39] there's no money in making things work.. but there's plenty of money in never ending series of fixes. [20:40] Very true, most programmers make the best software in their spare time [20:40] also true [20:40] Not while making some dull business application.. but why should they? [20:40] but roy, take for example radix sort, are you familiar with that? [20:40] Sure I am [20:40] (yes I'm sure you are bvowk :) ) [20:41] Sure he is :) [20:41] But the customer is paying me much more money is I quickly do Collection.sort() in java then when I implement my own Radix sort [20:41] as little as 2-3 years ago I read a paper by someone in a university here which was about the O'ness of sorting algo's [20:41] is-if [20:42] and it actualy stated that radix sort was an old algorithm that could only sort char arrays and integer values [20:42] and since it didn't deal with floats/doubles it wasn't usefull for a lot of numerical sorting these days [20:43] this comming from a university is pretty shocking :P [20:43] You could partially do a radix and finish it with another algorithm.. [20:43] radix sort runs through floats and doubles just fine with a minor adaptation [20:43] it just requires a quick swap pass afterwards [20:46] ofcourse the reason for this is that these guys simply don't know the inner workings of a radix sort and have never programmed one themselves, they write a paper based of other people's mistakes... [20:46] Like this? http://www.stereopsis.com/radix.html [20:46] yup [20:47] Thats another mistake most programmers make, they don't look stuff up... even IF he didn't know that radix could be very well used on floats/doubles, before you put it in a document CHECK IT! took me about 20 seconds with google.. [20:48] Heard it once and take it for granted [20:48] what's google? [20:48] ;) [20:50] hehe ... just found a really nice RFC [20:50] But this is a big difference between Java programmers (that just need to get something to work pretty well) and C programmers who want something to work very good very fast. [20:50] 3514 :) [20:51] Sigh... [20:55] The thing I want to do when programming at work: 1. Make it readable, 2. Make it reusable 3. Make it work well enough... Something a C/C++ programmer would say I think you be 1. Think of the best algorithms, 2. Code the algorithms, 3. Review and do 1 again [20:57] But thats just the market... a customer will give me money if I say I will program some Webservice in 3 days, but he will shout at me it I say I made my own double-radix-sort and the sorting now takes 2 seconds instead of 5 [20:58] [3jane:~/work-tmp/jnsl$] find clean_save -name \*.rc | wc -l [20:58] 570743 [20:58] yow. [20:58] I've got 570,000 nano warriors! [20:58] and thats filtered w/ a 110 point cutoff vs K! [20:58] Nice.. but how do you know which one is best suited for the hill ;-) [20:59] if I knew that roy, I'd have taken the world already. [21:00] no, you wouldn't [21:01] (poor fluffy, he's a little slow) [21:01] No, I believe Fluffy, he probably did the math first [21:02] yeah, but he doesn't believe in solving differential equations.. so he really only did half the job. [21:02] Knowing that a solution exists, *is* enough! [21:12] http://www.sudokutron.com/daily-Insane <-- ok, that one really is insane [21:13] why? [21:15] because usually I have no trouble solving them [21:17] Only because you can't solve it, it doesn't really mean that it is "insane" [21:19] I didn't say I couldn't solve it.. [21:19] it was just actually a difficult puzzle. [21:21] * Fluffy waves [21:21] MSG: Quit: fluffy.i < 1, # 42 [21:39] MSG: [21:45] MSG: Quit: Lost terminal [22:39] Join: jon_ joined #corewars [22:42] good morning/afternoon/evening [22:44] lo jon [22:44] nalgas is still on the beginners hill [22:47] hehe I won't submit anything there :) [22:47] so you're safe from me atleast [22:50] oh good [22:50] it's my goal to submit something better this week [22:50] its basicly a stone thrower with an imp gate.. [22:50] well I have 4 bots on top of nano atm so it'd be a bit mean for me to start submitting beginner bots now, lol [22:51] it'd be pretty funny though [22:52] meh, spoils the fun :) [22:52] enough Borgs out on the lists already [22:53] C&C 3 demo is out :) [22:53] command and conquer? [22:55] yup [22:55] oh man [22:55] that looks cool [22:55] if i only i had the hardware [22:55] they are bringing back the FMV sequences [22:55] and kane has returned :) [22:55] my box can barely run pmars :( [22:56] ic [23:04] have you ever been to the HAL/HOPE/etc ? [23:11] doesn't sound familiar [23:20] ah [23:20] the hacker camp things in .nl [23:27] oh yeah, not been out there [23:28] not to say I wouldn't be tempted, I mostly have a love for gamedevelopment though [23:30] ah [23:30] i went to the one in 2001