[01:46] Nick Change: X-Scale2 changed nick to NorthStar [07:11] MSG: Ping timeout: 240 seconds [07:42] Part: NorthStar left #corewars [08:23] Join: M0wLaue joined #corewars [09:14] MSG: Ping timeout: 240 seconds [14:29] Join: fiveop joined #corewars [15:44] MSG: Quit: Lost terminal [15:45] Join: fiveop joined #corewars [19:06] Join: impomatic joined #corewars [19:06] Hi :-) [19:06] hi [19:08] got reaaaally boring during the "basic of entrepreneurship" -lessons so i actually went pen&paper coding again [19:16] :-) [19:17] I need a "basics of entrepreneurship" lesson! [19:17] damn, im rusty [19:18] now im thinking how the hell 2*step + n*(2*step) can become step [19:31] ..maybe it would help not having dat 1,1's as placeholders when using a djn-train [19:31] Could be tricky :-) [19:43] (muttering) [19:44] 21 processes and i could've gotten this to sync.. [19:45] not processes, steps [19:54] 1 step wrong.. [19:59] and now i have a mystery inc in place where it isnt supposed to be [20:03] :-( [20:04] mystery solved [20:17] damn, my head has a frost [20:18] ha, i've written redcode on paper on lunch breaks and stuff before :( [20:18] not like i ever "started" doing it.. [20:18] now im wondering why shifting a pattern will not affect a self-hit in the way i think it would [20:21] (expecting selfmodifying code give same results when skipping a part, ha!) [20:38] gg [20:43] can you think of a bomb that'd do a good amount of damage to an spl carpet? [20:43] i guess an imp of some sort would at least help wipe it out, but i don't know how i'd make it kill itself [20:44] maybe depend on the values in the spl carpet [20:44] but that'd only work on certain kinds [20:44] hm, or... [20:45] since it's in an spl carpet there's plenty of processes, something like mov loc, there really arent other solutions than a carpet of dat's [20:45] MSG: Quit: humhum [20:45] in case of the imp, the created processes will be running the imp and still doing useless thing [20:46] hmm.. mov.i {0, }0 [20:46] hit near the bottom of carpet [20:49] ooh, i like that [20:51] its less boring than mov #0, -1 [20:52] and faster acting [20:52] possibly [20:52] not really [20:52] only thing is it doesn't help the carpet after where the instructions are [20:52] well, mov 0,-1 requires processes to execute in each location [20:52] oh nevermind [20:52] :\ [20:52] i forgot it resets itself [20:55] impomatic: [20:56] is it worth spending one instruction/paper in order to separate a paper from the imp in a binary-launched warrior? [21:00] (dont tell me you havent tried) [21:09] Mizcu: I've experimented with it before, but I don't think it ended up in a finished warrior [21:09] i got Recon 2 in mind, since it uses 2667n based stepsize [21:09] Assuming you want a mov imp, somewhere in the paper. mov imp, theres better instruction for that [21:10] Although wouldn't it need 2 extra instructions? [21:11] mov hop, }pstep+hop , imp #impposition, istep [21:11] replace impposition #bomblocation [21:13] the first process will copy the imp to the supposed location, and the rest will move additional instructions elsewhere [21:15] Wouldn't you also need a jump to the imp? [21:15] of course [21:16] That's quite a few extra cycles just to seperate them [21:17] but i consider mov imp, {<>}target to be inferior, since the imps create a large carpet which is a little bit more vulnerable to scanners instead of 1 and separate carpet [21:17] but it is huge amount of time [21:18] Better for the carpet to hit an imp than a paper [21:18] better to hit unused imp-instructions [21:18] I think there'd be better ways to use the time. Bombing for example [21:20] true, but im now contemplating the time when scanners get really overoptimized [21:21] How about MOV 1, }1 / MOV.I #2660, 2667 [21:21] Then if one of the papers has been spl carpeted, it'll squash the spl carpet and put the imp back [21:21] (hopefully before the next one gets hit by recon) [21:22] sure, it is better to spend time on bombing right now, but isnt it a nice puzzle to try majorly defeating a scanner with a P/I? [21:22] i guess you missed what i meant [21:23] this is probably offtopic but i had a litle fun making a deal that ran like an imp ring until it got hit somewhere [21:23] it kept "pushing" the relevant code ahead of it around the core hehe [21:24] not very useful though, because it'll push damaged code around too :P [21:24] it was fun to watch though [21:24] sliding around like a ring then blossoming into a paper or something [21:24] :-) [21:25] Sounds an interesting piece of code, even if it isn't practical [21:25] http://users.metropolia.fi/~mikaos/pitemp.txt [21:25] i don't know if i saved it to a file ever, but i had to synch the processes so they executed in stacks, in turn [21:26] 5 here, 5 there, 5 on the third poitn [21:26] point* [21:26] or 3, 4, whatever the quantity was [21:28] silly typo fixd [21:28] Mizcu: now I understand :-) [21:30] Also, why not try jmp gap, }0 to make spirals [21:31] -\|/-\|/-\|/- [21:31] :cogs: [21:33] then youd need only one imp included.. [21:33] ..in danger of over-engineering [21:36] MSG: Ping timeout: 240 seconds [21:53] http://www.nomorepasting.com/getpaste.php?pasteid=26045 [21:53] ^ a (poor-ish) example of what i meant earlier [21:53] lol now that he quit [23:48] Join: lllll joined #corewars [23:48] Part: lllll left #corewars