[00:11] MSG: Quit: Lost terminal [11:03] Join: fiveop joined #corewars [11:16] i r back [11:26] no, you isn't ;) [11:27] FU [11:27] ok [11:28] uh, i need sleep or can of - if not tear gas - caffeine [11:28] on the other half, i have again succeeded i wiriting other page of pseudocode for it [11:29] pseudocode for what? [11:29] the evolver [11:29] and real code? [11:29] about one page [11:30] Then it will probably be finished tomorrow, won't it? ;) [11:31] but the big problem in writing it will be the goddamn instruction -generation variance -table [11:32] which does what exactly? [11:33] it is used as a weight when generating instructions [11:34] weighted random, y'know [11:34] why don't you set all weights to 1 and then see how it works. Maybe you don't even need them [11:37] there is more planned that just throwing instructions in the warrior-generation/mutation [11:40] Do you plan to throw in anything else? [11:40] not exactly [11:42] great, table of 18 instructions times 20 values [11:42] plus couple more for the "else" part [11:44] why don't you start with a reduced instruction set [11:44] sth like spl/mov/djn/jm*/add/sub? [11:45] i most likely will disable them in the table by default [11:45] but the parse of the table will still be a pain in the butt [11:45] then write it directly into the source code [11:46] programming -teach wont like that [11:46] and if i am going to release it, it would be nice for the setting to be customizable without recompile [11:47] now be serious, who won't want to recompile it ;) [11:47] John [11:48] then don't tell John about it and you'll be fine [11:48] already did [11:49] then you have only one choice. Invent a time machine and undo it [11:51] or just make the goddamn table [11:54] c'mon! You can't seriously want to do it. [11:54] Play some nice game. [11:55] do you play nice against Bvowk? [11:55] Or do you part for accelerating global warming [11:55] *your [11:55] why should I play against bvowk? [11:55] I've already won! [11:56] you know, even if i save cycles from the table-reading, they will be used on running exhaust anyway [11:56] yes? [11:57] so i contribute to global warming anyway [11:57] :) [11:58] then comes the question of how the code the decision tree [11:59] as a tree? [11:59] maybe ill just nest 6-7 switch/cases [11:59] (waiting for the "urgh") [12:00] what's wrong with 6-7? I have more than 170 cases in the big switch for '88 [12:02] sound like a submission to (former)thedailywtf(today known just as wtf) [12:03] no, because I didn't want to create "good", but really fast code [12:03] so I knew what I did :) [12:03] at least I pretend to know ;) [12:05] My Z80 code has something over 1000 cases in its main switch. Stresses the GCSE pass of gcc a bit :-) [12:05] s/code/core/ [12:06] See, Mizcu! Use many cases. Your teacher will love it! [12:07] riiight [12:08] See it as a challenge. All you need is to properly reason with him. [12:08] Then you can do whatever you want [12:09] will return:ing from a Case cause a Fubar, or work just fine? [12:10] huh? [12:10] In C(++), it's completely fine. [12:10] Look up Duff's device if you want something almost completely insane with switch(...). [12:11] :) [12:12] Mizcu: Does anything depend on the code? Do you get a grade? [12:13] most of the grade comes from the project, and the rest from online coding-"challenges" [12:13] and the evolver is *the* project? [12:14] eg. we input c -code into nifty java-applet, which then compiles it in secure enviroment, feeds it values and seems what it pukes out [12:14] the evolver is my planned project, yes [12:15] Try to convince him the next time to write some Redcode [12:16] would work better on the assembly -programming course [15:50] Join: John joined #corewars [15:51] Hi :-) [15:53] h [15:55] Hi Mizcu [16:05] I've had 36000 visitors to my website in the last 24 hours [16:06] sound suspicious [16:07] Different IP addresses [16:27] it was me [16:29] Thanks :-) [16:34] * John waves [16:34] MSG: Quit: mov.i #1,1 [17:49] hmm [17:57] hmmmmmmmm? [17:57] hm [18:06] Join: Core_old joined #corewars [18:08] Hi old Core! [18:15] hi! [18:39] MSG: Read error: Operation timed out [18:40] MSG: Ping timeout: 255 seconds [18:41] oops [18:41] MSG: Ping timeout: 255 seconds [18:41] MSG: Quit: fluffy.i < 1, # 42 [18:41] Join: fluffy joined #corewars [18:41] hrm [18:42] Join: fiveop joined #corewars [18:42] Join: Core_old joined #corewars [18:45] Nick Change: fluffy changed nick to Flffy [18:45] grmpf [18:45] Nick Change: Flffy changed nick to Fluffy [18:45] that's better [18:45] sascha: 146.7 [18:53] Join: Mizcu joined #corewars [19:02] http://pluto.jhuapl.edu/soc/data/jupiter/level2/lor/jpeg/003486/lor_0034866662_0x630_sci_2.jpg [20:42] Fluffy: what's 146.7? [20:43] A number ... just a little bit below 150. [20:43] sascha: 146.8 [20:45] ... and now it is a little bit more closer to 150 :) [20:46] jon_: btw I'm still waiting for you to take over nano. [20:46] which warrior? :p [20:46] huh? which warrior? [20:47] which warrior has a score of 146.8 [20:47] on which hill [20:48] That warrior isn't on any hill [20:48] ... yet [20:51] But I guess, that you already know, that there are several benchmarks out there, which you can use, jon_ [20:52] really? [20:52] yes [20:53] you can take you warrior and let it fight against the benchmark [20:53] got 110 on the nano [20:53] not good enough to get on the hill though [20:53] oh wait [20:53] are those the wilmoo style things? [20:53] yes, but wilmoo is very old [20:55] http://www.corewar.info/bench.htm [20:55] for '94nop the fsh94nop 0.3 is the most recent [20:56] http://www.corewar.info/optimax/dload/ [20:57] nice [20:57] yes :) [20:58] Almost all warriors for '94nop are optimized agaist the fsh94nop [21:00] hmmz [21:00] got problems? [21:01] yes [21:01] need me to pimp up your warrior? [21:01] its ok :) [21:02] what is ok? [21:07] MSG: Quit: Trillian (http://www.ceruleanstudios.com [21:07] Join: Core_old joined #corewars [21:08] MSG: Client Quit [21:33] jon_: Do I write United Kingdom or Great Britain on a letter to the UK to specify the country? [21:33] The other USA? [21:36] UK, i would say [21:36] but not in short? [21:37] not in short [21:37] I think, that "United Kingdom" is good enough [21:38] thanks [21:38] And you really don't want to write "United Kingdom of Great Britain and Northern Ireland", do you? [21:39] Tomorrow my application for cambridge will be on its journey :) [21:39] Fluffy: no :) [21:39] Good luck, fivop! [21:40] thanks [21:41] And why Cambridge and not Oxford? [21:42] Fluffy: both works [21:43] hehe .. that mistake is now recorded and be automatically send to Cambridge *AND* Oxford! [21:43] *will be [21:44] Fluffy: because I study mathematics and not somethign else :) [21:44] Nobody wants to study math! [21:44] Brx does [21:45] Maybe he doesn't know yet [21:45] brx's is doing it [21:45] and Fluffy's done it :P [21:45] I'd never admit that [21:45] you have [21:45] You can't proove it [21:45] *prove? [21:46] anyone got the url to the logs for me? :P [21:46] You mean http://koth.org/irc-logs/ ? [21:47] yes [21:47] ok, then don't use that link ;) [21:47] :) [21:48] Now a period of stress has ended. Was kind of a task to gather all needed documents (and translations, and certified copies and what not) :) [21:49] Then it is the perfect time to write a warrior :) [21:49] it's th perfect time to get some rest :P [21:49] and drink some water [21:49] good idea [21:49] :P [21:50] * Fluffy calls Fluffy.nighty_night() [22:28] fiveop: i've seen you on freenode i think? [22:47] i just had something really rare happen [22:47] Yace's evolver scores went up to ~115, then took a dive, and current best is mov mov dat (without spl) [22:53] Part: jon_ left #corewars [22:54] and now its going all voodoo [23:00] hm [23:00] 132 with, certainly interesting warrior [23:01] MSG: Quit: Lost terminal