[00:28] MSG: [00:44] MSG: Ping timeout: 240 seconds [00:44] Join: Metcalf joined #corewars [03:05] MSG: [03:32] Join: myndzi joined #corewars [03:32] MSG: Client Quit [03:53] Join: myndzi joined #corewars [07:38] Join: fiveop joined #corewars [09:22] Join: OoS joined #corewars [09:24] MSG: Ping timeout: 240 seconds [11:55] Join: CwarsV1 joined #corewars [14:15] Hi! [14:27] Hi CwarsV1 :-) [14:30] Hi Oos [14:30] Do you know a beginner hill? [14:31] (using the oldest redcode) [14:32] Do you mean a benchmark you can test against at home? [14:34] hmm [14:34] yes. [14:34] i have not yet tested my warrior. [14:35] See http://corewar.co.uk/bench.htm [14:35] The first one, Wilkies, is considered a beginner's benchmark [14:36] ahh [14:36] You'll need a script to run the benchmark. I normally use MTS (mars tournament scheduler) [14:37] hmm [14:37] i got linux [14:37] and only pmars [14:51] MTS comes with C code. [14:52] Or there's a Perl scheduler [14:52] yes [14:52] i just try the warriors manually. [14:53] It does not look good! :-( [14:55] i loose against all of them! [15:25] You need to run 100 or more rounds again each warrior to get a more accurate result [15:26] Try to improve your warrior until you have at least 30% wins [15:38] 30% against one or all? [15:40] On average [15:42] okay [16:59] MSG: Remote host closed the connection [18:56] Join: Fizmo joined #corewars [18:57] Join: yoR joined #corewars [18:57] JM: Got one a single instrucion smaller ;-) [19:01] Hi yoR [19:01] hi Roy [19:01] In that case, so have I :-P [19:02] You do? :P [19:02] 97 instructions - +++++++++++[>>+>++>+++>+++<[++++++<]<-]>>-----.>>++.+++++++..+++.>-.<<-.>.+++.------.--------.>+. [19:02] Hi Fizmo [19:02] Why does everyone always turn up at meal times! [19:02] don't know ;-) [19:03] +++++[>+++<-]>[>>++>++++>++++>+++>-[+++<]<-]>>---.>----.>+++..+++.>>++.<---.<.+++.------.<-.>>>+. [19:03] I posted the 98 instruction version to twitter because it's got 3 nested loops! [19:03] Because in Europe its just after dinner ;) [19:04] Hmmm... actually it didn't have 3 nested loops! [19:04] (hour difference) [19:06] I'm pretty sure < 97 is possible... [19:06] :) [19:11] Preparing food! [19:40] Back! [19:41] No luck yet ;) [19:41] Roy: have you seen any other programming languages worth implementing in Redcode? [19:41] No not really... [19:41] I was thinking of some kind of Forth [19:41] I enjoy playing with bf now [19:41] you have a BF roy? [19:41] * bvowk ducks. [19:42] Are you still trying to reduce that hello world? Can't you write a program to do it? [19:42] Hi Bvowk [19:43] Heh... [19:44] Yeah, I have a beautiful (girl)friend :) [19:59] MSG: Quit: ChatZilla 0.9.84 [Firefox 3.0.5/2008120122] [20:13] Roy: ++++[>++++<-]>+[>>->++>+>+>++<[+++++<]<-]>>++++.>>>-.<++++++..+++.>>--.<<<.>.<-----.>---.>-.>+. [20:15] why don't you dudes evolve a brainfuck optimizer ;P [20:15] We did :) [20:16] We'll I did/tried [20:16] Hi nescience [20:16] That's cheating JM, you changed the case again ;-) [20:16] Someone wrote a bf optimizer and the best Hello World it finds is 118 (which includes the comma) [20:17] ha, oh really [20:17] Oh, didn't realise! [20:17] hi [20:23] too bad bf doesn't have like xor and stuff amirite? [20:23] or, really, anything useful :P [20:24] an idea re: hellow world [20:24] well, i dunno maybe it already works sorta like that [20:24] i see a lot of >s and it seems like it might be worth having multiple registered initialized to various ascii ranges [20:25] registers* [20:25] No xor in bf :-( [20:25] and it'd be shortest to do that in a loop [20:26] what do you think of a loop that'd initialize like 5 registers to a convenient number before subsequently skipping between whichever ones are most convenient? [20:26] actually, m aybe that's what it's doing, i see a nested loop too [20:26] lol [20:27] Yeah, we generate numbers close to the onces we need ;) [20:27] ++++++[>+++<-]>[>>+++>+++>->++++[+++<]<-]>>----.---.>..+++.>----.>-------.<<.+++.------.<-.>>+. [20:27] Even another case variation :) [20:30] i mean, not only that, but it'd be wasteful to go back and forth on a single register, maybe could be done shorter if you use more registers [20:31] ++++[>++++<-]>+[>>->++>+>+>++<[+++++<]<-]>>++++.>>>-.-<++++++..+++.>>--.+<<<.----->.---<[.>] [20:33] i need a three letter nick then i can twiddle bytes and compare immediately too [20:33] :> [20:33] "Hello world!" in 92 :-P [20:34] We do that too ;-) [20:34] Generate a couple of registers with characters close to eachother [20:36] Your case variantion in 89 yoR: ++++[>++++<-]>+[>>++>+>+>++<[+++++<]<-]>>>>++.---.-<++++++..+++.>>--.+<<<.----->.---<[.>] [20:36] yoR: that's what it was beginning to look like to me [20:36] but as i said, i've no talent for interpreting brainfuck in my head [20:36] and no time to sit down and study them yet [20:37] i'd rig up an interpreter in mirc but i bet none of you use it :> [20:37] nes: yoR and I are always trying to outsmart each other. [20:37] (one of us has more success than the other!) ;-P [20:38] (one of them has Koen #1..) [20:38] There's an interpreter on freenode in #esoteric [20:38] Just type ^bf before the code [20:39] lol [20:40] what is koen? [20:41] koeningstuhl, biggest hill of them all [20:42] ah, that koen [20:51] We both like to challenge eachother and invent new stuff in the process ;) [20:55] that's the fun way to do it :) [20:56] It depends, some people can't stand it [20:57] But we both know its in good spirit, its more challenging eachother, not outsmarting [20:58] I'm still waiting for yoR to beat my redcode bf interpreter ;-) [20:59] Quite a few pairs of bf instructions compile to one redcode instruction! [20:59] Yes, that was a nice find, I didn't think of that! [20:59] I was pretty smug with my implementation :) [21:00] i dunno, i'm in favor [21:00] i'm the kind that can get awfully determined ;) [21:01] I'm quite please with my implementation! [21:01] The next step would be to read bf in from a file. [21:02] I need to find a way to compile exmars first [21:10] now i want to play around with brainfuck hehe [21:11] gonna take lunch and bring my laptop [21:11] Laptop not needed :) bf is mostly done on paper ;-) [21:11] Although a debugger (wrote one) is helpfull [21:12] This oneline interpreter works fine http://www.iwriteiam.nl/Ha_bf_online.html [21:13] I found another which doesn't work properly. I thought my bf code was broken :-/ [21:13] Mine just outputs the memory after each instruction, works good enough [21:14] Next would be an interactive interpreter with step forward/back... but nah :) too much work [21:15] There's some challenges here [21:15] http://golf.shinh.org/ [21:15] MSG: Ping timeout: 240 seconds [21:15] I've done number 1, "Hello, world!" [21:16] in 24 instructions? :P [21:17] I think that's some kind of cheat [21:17] 94 is more realistic [21:29] Join: nescience joined #corewars [21:33] df [21:33] er [21:39] hello world! [21:39] yay [21:39] //say $bf(++++[>++++<-]>+[>>++>+>+>++<[+++++<]<-]>>>>++.---.-<++++++..+++.>>--.+<<<.----->.---<[.>]) [21:39] hehe [21:41] Nice ending :) [21:41] nah, that's copied from above [21:41] just grabbed one to make sure it worked [21:41] (but i did think of that on the way here lol) [21:41] which is why i noticed it and was pleasantly surprised :> [21:56] This is the smallest I can manage, 84: ++++[>++++<-]>+[>>++>+>+>++<[+++++<]<-]>>>++.>-.<++++..+++.>>--.++<<<.---->.--<[-.>] [21:58] how about the smallest "HeLlO, wOrLd!" ? ;) [22:05] nescience: if you can get it smaller than 84 :-P [22:05] Thats easy: ++++[>++++<-]>[>+++++<-]>--.+. [22:06] MSG: Ping timeout: 240 seconds [22:14] Shorter: +++++++[>+++++++++++<-]>+.+. [22:16] Join: nescience joined #corewars [22:19] We're also talking about bf in #esoteric on irc.freenode.net [22:21] Join: ehird joined #corewars [22:23] Hi ehird [22:23] hi :) [22:24] Everyone's hard at work! [22:25] MSG: Ping timeout: 240 seconds [22:26] idling is the hardest job on all IRC [22:27] it's pretty touch. [22:27] er.. [22:27] tough [22:28] i know. idling is so hard it takes away your ability to spell. :D [22:28] heh [22:29] I wonder if there's a treatment program for my addiction to redheads. [22:31] what wrong with that? [22:31] I suggested to the current one that perhaps I just need to hit redhead bottom, so I needed at least a half dozen more before I could really admit it wasn't a problem. [22:31] she wasn't especially impressed with the idea. [22:32] er.. that it *WAS* a problem. [22:33] Hi ehird :) [22:34] What are you up to met? [22:38] bvowk: I'm trying to write smaller bf programs than yoR! [22:39] I got that much :) [22:39] I was hoping to get hello world in < 88 instructions, so the interpreter and demo program fit in 100 instructions in exmars streams [22:40] yoR needs to aim for < 75! [22:43] And I'm outputting the nr E in bf... [22:43] And I suck at it :) [22:43] (last place by a long marging) [22:56] I think my computer is in the process of crashing :-/ [23:00] MSG: Quit: humhum [23:17] Okay, I did echo to discover yoR got there first! [23:20] Do E :-) [23:21] echo? [23:21] Golf bf echo [23:21] ,[.,] [23:21] eof confused me for a few minutes, I thought it was 0 [23:21] it is. [23:21] on decent implementations [23:22] It's -1 :-/ [23:22] Only on bad implementations :-) [23:22] but for -1: [23:22] ,+[-.,+] [23:23] (You can't really get shorter than that, BTW.) [23:23] Or am I misunderstanding "echo"? [23:25] That's the same as my solution [23:26] ok, get yourself a decent implementation and use ,[.,] :-P [23:29] No, we are playing anarcgy golf :) [23:29] http://golf.shinh.org/l.rb?bf [23:29] ahh [23:29] yeah, anagolf's implementation sucks [23:30] BTW: the cheating ones are sneaky. [23:30] They exploit a bug in anagolf's implementation, which lets you output arbitrary strings. [23:30] (by accessing the codespace) [23:33] yeah, I was wondering how they did that [23:34] Hm. [23:34] Apparently it's switched to bfi.c, so I don't think that would work again [23:34] http://esoteric.sange.fi/brainfuck/impl/interp/BFI.c A rather, uh, naive implementation. [23:35] Heh I looked through that code, looking for the bug :) [23:35] Still no boundary checking.. [23:35] and it checks for brackets every loop x_x [23:35] it used to be in Ruby, like the rest of anagolf [23:36] < will result in x[-1] though.. (?) [23:36] yep [23:36] * ehird writes BF program to slaughter memory, submits to anagolf [23:36] So some hacking could still be possible :P [23:37] Not really; it's run under a strict chroot. [23:37] (That's why you can give it arbitrary C programs...) [23:37] Well, internally, by accessing another array while running out? [23:38] Maybe. [23:38] http://golf.shinh.org/p.rb?Brainfuck+FIXED [23:39] Those are some large interps... [23:39] [I submitted that] [23:41] Nice.. is there a way in bf to go forwards/backwards the amount of registers with a value? [23:41] Like: do 20 > (instead of writing >>>>>>>>>>>>>>>>>>>>>>>>) [23:41] Maybe in a loop..? [23:42] Nope. [23:42] You have to make a tape. [23:42] 1 val 1 val 1 val 0 [23:42] looping is trivial from there [23:42] Yeah.. [23:42] I'm puzzled at the results for the number e [23:43] I can't understand how they can make it any smaller them +/- 800 lines [23:43] cheating. [23:43] (outputting a constant string) [23:43] well [23:43] wait [23:44] yoR: 800 lines?! [23:44] Not all I think [23:44] calculating 800 is so not 800 lines... [23:44] Fox seems to be legit, he has a score of 295 :) [23:45] Curious.. [23:52] http://golf.shinh.org/p.rb?Print+out+a+lot+_56K+BEWARE_ [23:52] Try this (my submission) [23:52] To generate the 100 digits (which are irrational) I start out with this snippet: >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>+[+++++++++++++++++++++++++++++++++++++++++++++++<+] [23:53] Which makes 100x '48', and the I add the stuff to generate the sequence.. [23:53] +++++++++++++++++++++++++++++++++++++++++++++++ [23:53] That can be shortened. [23:53] And you don't need to go right so many times [23:53] Well yeah, but thats pretty minimal :) [23:54] Why not right..? [23:54] It's not very minimal, and because it takes a lot of codespace :P [23:55] No, but how else would you print the number e..? Thats my big question! :D [23:55] Beats me :) [23:57] Is anagolf down...? Seems so. [23:57] works here.. [23:58] Me too, now. [23:59] On freenode, #anagol houses the main anarchy golf people, BTW. Might be helpful.