[02:41] MSG: [05:46] Join: fiveop joined #corewars [06:52] Join: Mizcu joined #corewars [08:35] Join: fiveop_ joined #corewars [08:37] MSG: Ping timeout: 252 seconds [10:00] Nick Change: Cctoide[s changed nick to Cctoide [10:06] Join: Cctoide_ joined #corewars [10:07] MSG: Ping timeout: 252 seconds [10:08] Join: Core29 joined #corewars [10:17] MSG: Quit: Leaving [10:34] Join: Cctoide joined #corewars [12:35] Does anyone here play other programming games? [12:35] I dont really think so [12:46] i know some..but played only a very few [12:47] robosports [12:47] A.I. - The insect mind [12:47] Mindrover [12:47] Join: Roy joined #corewars [12:47] or does anyone remember "Omega"? I had it on my amiga... [12:48] Hi [12:48] one of my first games I actually payed for... [12:48] hi [12:48] Heh, the name sounds very familiar.. [12:49] you had to program a tank. If you won against other preprogrammed tanks you were promoted [12:49] * Roy played some Robot Football game and C++Robots or something [12:49] ya, I know those. Never played it though [12:50] RoboSoccer was pretty cool to do [12:50] Ohw, and RoShamBo! [12:50] roshambo? [12:50] (Rock-Paper-Scissors) [12:50] http://www.cs.ualberta.ca/~darse/rsbpc.html [12:50] Pretty advanced nice stuff there [12:51] btw, what do you guys think of my new website design? [12:51] (http://www.redcode.nl/) [12:51] IIRC there is a list of "programmable" games [12:52] http://www.gameai.com/exaigames.html [12:52] afk for a short trip with Dad's new car :-D [12:52] Hehe Bolo, played that too [12:52] know that too! [12:52] (build a Borg once) [12:54] Join: hwm joined #corewars [12:56] * Cctoide confuses self [12:56] Happens to me all the time :P [12:56] About what? [12:57] hi hwm [12:57] hello [12:58] a lot of people here tonight [12:59] MindRover looks interesting [13:02] tonight hehe [13:03] It's 14:10 here :p [13:03] * Roy wonders what kind of car Core29's dad bought.. [13:03] 15:04 here :P [13:04] No opinions on my webpage design..? [13:05] Not tested in Mozilla/Firefox yet [13:05] Goal for Argentinie [13:07] MSG: Read error: Connection reset by peer [13:07] Join: fiveop joined #corewars [13:08] its 01:07 here [13:08] other end of the earth I guess [13:08] Probably, America? [13:09] I'm a kiwi [13:10] Cool, I'd love to visit that once.. [13:10] there's some lovely terrain here [13:11] it's really the weird animal capital of the world i guess [13:11] in any case I better fly [13:11] i thoughtI'd drop by [13:12] Part: hwm left #corewars [13:15] if only one could write redcode-like warriors to defrag drives :P [13:16] Heh, that will probably defrag more then you might want [13:17] I was remembered of the defrag visual representations by the corewin memory view [13:18] well yeah, there is some likeness [13:40] MSG: [14:14] MSG: Quit: leaving [14:16] Join: pak21 joined #corewars [14:54] Join: Fluffy joined #corewars [14:54] :) [14:54] . [14:55] Any news in the last couple of days? [14:56] not really [14:59] hmm [14:59] oh ... bvowk is here :) [15:00] * Fluffy kicks bvowk [15:00] Let's see how long it takes him to react [15:05] Even a dinosaur should have reacted by now ;-) [15:05] why should a dinosaur have a slow reaction? [15:06] where exactly was Nenad from? [15:06] Don't you know the common "funny" answer to the question, why dinosaurs have become extinct? [15:06] no [15:06] guess not [15:07] would you mind to enlighten me? :P [15:07] Nenad lives somewhere in the former Yugoslavia [15:08] Novi Sad, Serbia [15:09] but what about the funny awnser to the question? [15:09] * bvowk kicks fluffy [15:09] * bvowk then hangs fluffy outside the channel by his entrails as a warning to others. [15:09] Answer: Dinosaurs have become extinct because they took too long to react. By the time they've noticed, that something is wrong, they've long been eaten/... [15:10] Hi bvowk [15:10] bvowk: Please make sure to hang me someplace sunny [15:11] dinosaurs had tail brains tho [15:11] wth are tail brains? [15:12] basicly, many of the dinosaurs were too big to have nerve signals propogate fast enough to keep everything moving right.. [15:13] so they had a "tail brain" or multiple sub brains that would keep motor functions in the way the hell out and gone bits working [15:13] like octopi do, each of their arms is independant [15:13] hehe ... that would make all patents on distributed computing useless because of "prior art" :) [15:13] lol [15:13] yea, I know people too, where one hand does not know what the other does... [15:14] Hi Core30 :) [15:14] Nick Change: Core29 changed nick to Core_old [15:14] :-) [15:14] Usually I don't know what I'm doing [15:14] core: I can do one better, I used to know a guy who didn't know what the hell he was doing! [15:14] bvowk: Now you know two :) [15:14] hehe [15:15] mouth speaking without brain thinking... [15:15] I've just read about Rubberhose (encrypted filesystem) [15:15] arg won 6:0... [15:15] * bvowk rubber hoses fluffy [15:15] In one of the doc file a guy is described, who use different names [15:16] He had one for even days [15:16] and one for odd days :) [15:16] bvowk: I'm immune to torture [15:16] everyone says that until they're tortured [15:16] nobody resists torture.. its one of the facts of life [15:16] bvowk: I'd probably wet my pants before being tortured [15:16] well, maybe not chuck norris.. [15:17] but certainly everyone else :) [15:17] problem is that people will say anything to make it stop, so its utterly useless for intelligence [15:17] That's not the problem. [15:18] well, aside from it not being polite? [15:18] The problem is, that you will be tortured nonetheless [15:18] there is that too.. [15:18] cause tortureing is fun! [15:18] -!+? [15:19] :P [15:19] not actually [15:19] bvowk: By the way ... you've forgotten Rambo. He's immune to torture, too [15:19] so whats everyone doing today.. speak up.. don't make me torture you.. [15:19] its a crap job [15:20] yeah, but stallone is so unintelligable it wouldn't matter if he did crack, so he doesn't count! [15:20] lol [15:21] bvowk: I'm doing nothing ... just reading (again) lots of texts about steganography [15:22] bvowk: And if you torture me, I'll tell you things, that you never wanted to know :) [15:23] why don't you work on pycorewar? :) [15:23] mizcu: you haven't been running one of those CIA rendition houses have you? [15:23] yes, but at the moment I only change an average of 100 bytes/day [15:24] bvowk: i dont work with 'mericans [15:24] do you work on them then? [15:24] Mizcu: They aren't called "'mericans". In most countries they are just aliens ;-) [15:26] I work with the Finnish police's ninja-squad [15:27] heh [15:27] * bvowk hires mizcu for the international corewars security squad [15:27] bvowk: Any what should be its task? [15:27] * bvowk hires fluffy too.. might need bait. [15:27] To eliminate all evolved Redcode? [15:28] no! to eliminate the enemies of redcode! [15:28] like those damn robot battlers! [15:28] Mizcu: You should take the job and demand "proper" payment from bvowk [15:30] (and beware not to be payed with those useless Canadian $) [15:30] won't be long before our dollars are worth more than USD [15:31] http://finance.yahoo.com/currency/convert?amt=1&from=CAD&to=USD&submit=Convert [15:31] yes, probably everybody will pay collector prices as soon as Canada is invaded [15:31] heh [15:32] if the US has trouble invading and holding iraq.. wth are they going to invade canada with?! [15:32] Microsoft? [15:33] I don't think microsoft have enough employees to occupy canada [15:33] Maybe they start to clone BG [15:36] I have a little problem: I have a huge bitmap, which shows which sector of a harddisk is used and which is not. How do I choose a random free sector without holding the entire bitmap in memory? [15:36] Did I mention, that it should be fast algo? ;-) [15:37] MSG: Quit: Trillian (http://www.ceruleanstudios.com [15:45] hmm [15:46] hm [15:48] I guess, that means, that I have to hold the entire bitmap in memory AND have a very bad performance, especially, if the harddisk is almost full :( [15:49] MSG: Read error: Connection reset by peer [15:49] Join: fiveop joined #corewars [15:49] hrmpf [15:49] dont use a bitmap in the first place? [15:50] But how do I choose a truly random free sector, if I don't have an allocation bitmap? [15:52] make a list of the free sectors [15:52] Try a truly random sector. See if it's free. Try again if it's not. Obviously slow if the disk is close to full. [15:54] why do it randomly? [15:54] pak21: That would mean, that I have to hold the entire allocation bitmap in memory. If not I'd have to read from disk to check for free sector, which would make it slower [15:54] bvowk: steganography [15:54] fill the entire disk with high entropy data, and then pick a random offset, and mod that with the actual disk size in sectors.. that'll be your start point, allocate from that point [15:54] Fluffy: wth is that? [15:55] fiveop: wikipedia [15:55] deletions will leave openings.. take the first opening.. [15:55] (that's faster than explaining it here) [15:55] bvowk: that's not random [15:55] (cache the last free sector) [15:55] fluffy: if its all a mass of random high entropy data.. [15:55] there's no way to tell where you're beginning and ending assuming there isn't a flaw in your crypto system [15:56] now, it gets tougher if you're trying to hide it in something that exists as lower entropy data [15:56] Which is exactly how CSS was broken :-) [15:57] they tried to hide things in low entropy data? [15:57] The keys [15:58] k [15:58] In the rest of the program; they just looked for a section of the code with particularly high entropy. [15:58] I'm just playing with some ideas: [15:58] hm [15:58] take steganographic file system. it works as follows [15:59] fill a partition with random data [15:59] I'd like to have a mathematical defined measure for entropy :P [15:59] take a key and encrypt some data and scatter it all over the partition [15:59] from the key you know where to look for the encrypted data [16:00] take another key and encrypt some other data and scatter it all over the partition again [16:00] In oder to avoid overwriting data, you have choose the sectors to be written to, from the list of yet free sectors [16:00] fiveop: there are a couple [16:00] bvowk: any urls? [16:01] If you only know one key, you have no way to proof the existance of another set of encrypted data [16:01] (of course you have to know all keys, if you want to write securely to the fs) [16:01] If you're going to be writing data to the disk anyway, does it matter if you take a small amount of time to find out if a sector is free or not? [16:03] It isn't a problem, if you have almost no data of the fs, but if you have, it takes a lot of time to find a free sector [16:03] (free and random sector) [16:03] http://www.fourmilab.ch/random/ [16:04] thats the one off the top of my head fiveop :) [16:04] and Fluffy if the totalsize/sectorsize ratio isn't too big, you can easily place a bitmap in ram [16:05] fiveop: No, I can't and that's the problem [16:05] Suppose I have a 1 TiB partition (it isn't that big) [16:05] size of sector: 1 KiB [16:05] they'll be 512 bytes on pretty much everything [16:06] That means: 2**40/(2*10 * 2*3) = 2**27 bytes of memory for the bitmp [16:06] err [16:06] Fine. That's less than a GiB. [16:06] 2*40(2**9 * 2**3) [16:06] (you know, what I want to write) [16:06] [16:06] pak21: Yes, but I sill want to use it on my laptop) [16:07] Put more RAM in. [16:07] If you want this kind of security, you're going to have to pay for it, either in speed or in space ( = money ). [16:07] and unfortunately I have to use that much memory for each key, that I use [16:07] but your laptop hasn't got 1TiB :P [16:07] fluffy, increase your sector size [16:07] a sectorsize of 4kB shoudn't be that bad [16:08] and you're down to 2**25 ;) [16:08] still 33 gigs :P [16:08] no [16:08] Even better, variable sector size. Have a mix of big and small sectors on the disk. [16:08] fluffy: screw that.. [16:08] This reminds me of the problems they have with TLB caches for processors. [16:08] its a silly idea. [16:09] no bitmaps.. [16:09] Which is essentially the same problem. [16:09] pak21: And how do I know, which sector size to use? [16:09] If you're writing lots of data, use a big sector. If you're writing less, use a small sector. [16:09] define an interleave size for your disk, so you only touch every n sectors, where n is a random number of potiential keys [16:10] then don't keep track of the blocks that are allocated, just treat it as an encrypted block device and run a real filesystem on top of it.. [16:10] pak21: And what do I do, if I delete data? [16:10] Possibly waste some space. Run a compactor over the data occasionally. That may be slow, but you don't need to do it too often. [16:11] bvowk: That doesn't sound very random [16:11] I could deduce the number of keys from the interleave pattern [16:12] use the key to pick the start sector using a random modulo of the sectors you operate on, encrypt each sector with a different key, created from the sector number (after the mod, you'll need to know the key to know where the numbering starts) [16:12] pak21: Sounds like defrag :) [16:12] fluffy: a) if you fill the disk with random information first there's no way to tell whats what, and if you set N to like 30 or 40, there's really no way anyone could expect you to have memorized that many good solid passwords.. [16:13] (should be a b in there somewhere) [16:13] and then you're using a real filesystem at least, and you don't have to reinvent the wheel here :) [16:13] bvowk: the typical use of a steganographic fs is to reveal some keys (with mildly compromising material) [16:13] I think this depends on whether you're worried about active or passive attacks. If someone can view your disk access patterns, this whole thing is pointless. [16:14] bvowk: and to keep the key of the really important info. [16:14] pak: there's nothing to say you can't have the filesystem layer manipulate the unallocated N slices to create traffic to defeat the maid attack [16:14] So it is very important, tht there is no way to get to know the real number of keys [16:15] (and now we are back to torture ;-) [16:15] your bitmap system wouldn't save you tho.. [16:15] if you revealed anything that would give you the bitmap, they'd know [16:16] Yes and no [16:16] Fluffy: just make wrong keys return rubbish. Then you can give always give them n keys and just say "I never got around to using those ones yet." [16:16] on a modern disk you could set N pretty insanely high [16:17] since most people dont use much of their 60GB disks [16:17] setting N to 60 gives you 10GB [16:17] ? [16:17] hell, you could set it to 600 and play with a couple 1GB slices :) [16:17] if someone wants to beat 600 keys out of you, that would be entertaining :) [16:18] (ok, so it slices by sectors and not by bytes.. but you see what I mean) [16:18] They wouldn't have to; they'd just beat out of you the location of the piece of paper with 600 keys on. Almost nobody can remember 600 strong keys. [16:18] if you're planning on storing large sections of data, you're going to have to be upfront about it [16:18] pak21: How do I know, what are wrong keys? [16:18] pak21: I mean from the program's point of view? [16:18] fluffy: you don't! [16:19] and they won't expect you to know 600 keys.. [16:19] which is an easy task :) [16:19] they'll beat you for the first couple, and then once they find the embarrassing collection of granny porn, they'll assume thats it [16:19] only you'll know how many real keys exist [16:20] It is probably useless to implement steganographic fs. Nobody would use them [16:21] and if you're not going to see at least a couple others with strange, upto dateish random embarrassing stuff that they might expect someone to keep quiet, its not worth considering the exercise. [16:21] fluffy: well, not with that attitude! [16:21] hehe [16:21] I keep all my boxes encrypted [16:22] at the block level.. [16:22] using GBDE or GELI or CGD [16:22] bvowk: How many time have you tried to convert other to use crypto? [16:22] GBDE is pretty cool.. [16:22] bvowk: How many times have you succeeded? [16:22] why would I care.. if everyone starts using crypto, they'll start developing legal techniques to counter it [16:22] Fluffy: have you looked at stegfs? The original paper behind it is by Ross Anderson, who generally knows what he's talking about [16:22] Anyway, I should go home... [16:23] * pak21 waves [16:23] pak21: yes, I hve [16:23] * Fluffy waves [16:23] don't forget fluffy, you'll also need to encrypt your swap and temporary file space [16:24] bvowk: Yes :) [16:24] goal [16:24] fiveop: huh? [16:24] and scrub your deleted filespace on the unallocated unencrypted disk regularily [16:24] roy's guys have scored a goal [16:24] * Fluffy kicks fiveop for mentioning that stupid sport [16:24] dev/ad0s1e.eli 989M 16K 910M 0% /tmp [16:24] dev/ad2s1a.eli 43G 429M 39G 1% /home [16:24] dev/ad0s1g.eli 34G 7.0G 24G 23% /disk [16:24] like so. [16:25] Fluffy: why is it stupid? [16:25] well I like it, as long as I don't have to watch it :P [16:25] and in the regular /home (under the mount point) I have some randomly touched updated computer geek hacking stuff [16:25] why randomly touched? [16:25] so if anyone just went for a quick peek, they'd likely assume I don't use the box much [16:25] so its up to date [16:25] fiveop: 22 people in search for a ball. Once someone has it, he kicks it away. Sounds stupid to me :) [16:25] they'll check the file dates :) [16:26] I've also got an encrypted swap partition [16:26] but why would anyone assume, that you don't use the box much? [16:26] because there's not that much stuff there.. just some random corewars project stuff.. [16:26] etc [16:26] if it's randomly touched, there are always files lately touched [16:26] :) [16:27] if you're going to encrypt disk, make sure you scrub it with a good random source first... [16:27] /dev/brain [16:27] that's pretty much random data here [16:27] write to every sector at least a couple times from /dev/random, somewhere that you'll get good entropy stirring.. [16:28] bvowk: I hope, that you have good backups, too [16:28] ie: connect it to a busy ethernet segment and capture the nic interrupt inter-packet timing data.. [16:28] fluffy: why is that? [16:29] MSG: Remote host closed the connection [16:29] I've got a firefox extension that mentions every goal that's achieved :P [16:29] scored [16:29] what ever [16:29] poor fiveop :) [16:29] Join: bvowk joined #corewars [16:29] wtf was that. [16:29] CIA? FBI? [16:29] Fluffy: well i don't care [16:29] but it's funny :P [16:29] heh [16:30] no, I hit ctrl-S in the wrong window and my irc client died [16:30] odd. [16:30] anyways.. [16:30] why would I need good backups fluffy? I've got a trustworthy filesystem (its just standard UFS/UFS2) on an encrypted block device [16:31] bvowk: backup ... have you ever tried to recover from a bad encrypted disk without a good backup [16:31] same as recovering from anything else. [16:32] hmm [16:32] UFS UFS? [16:32] same as UFO, but with a S instead of the O [16:33] Join: sf joined #corewars [16:33] UFS is unix file system.. the basic filesystem of the BSD family, SCO unix, solaris.. [16:33] Hello [16:33] Hi Bvowky :) [16:33] Hi sf :) [16:33] Hi Fluff [16:33] dev/ad4s1d.bde 849G 690G 91G 88% /home [16:33] thats my personal favorite filesystem :) [16:33] encrypted with GBDE [16:34] where's the key stored? [16:35] each sector is encrypted with a different key, derived from the master key material and some "other" information, you can have up to 4 user keys.. any user can destroy the master key material.. and the master key material is either stored off disk, or stored in a sector that is derived from the key by a hash [16:36] geli will let you xor two different bits of key material together to generate your key.. so you'll need to have a blob of data stored on your USB keychain, and a key you keep in your head [16:36] last one is nice ;) [16:37] uses pkrs for the passphrase->key generation [16:37] makes it (unless pkrs becomes vulnerable in the future) much more difficult to abuse the passphrase [16:38] PKRS lets you set a computational requirement for every passphrase->key mapping the attacker wants to try.. [16:39] I set the one on my notebook really high, so it takes like 10 seconds to make the key from my passphrase [16:39] (on a 1.8Ghz P4) [16:39] its really annoying to get the passphrase wrong now.. [16:40] anyways.. [16:40] thats my crypto rant.. [16:40] HI SF! [16:40] how are you? [16:40] You finally noticed I'm here! [16:41] sf: You could kick him now :) [16:41] sorry, I got lost in crypto land.. took me a while to walk back [16:41] whatcha doing today? [16:43] Working on some new stuff [16:43] sf: hehe ... I've just noticed, that you are my next opponent :) [16:43] Am I [16:43] I thought I was fighting Neo [16:44] nope [16:44] Oh [16:44] Now I don't need to make as much effort :) [16:44] :P [16:44] I'll submit a scanner then [16:44] hehe [16:45] Maybe I'll send a paper, which is only vulnerable to stones :) [16:47] Maybe I'll send a stone [16:48] Or an imp again ;) [16:48] Did you send an imp last round? [16:48] Yes :) [16:48] bvowk: Do you use anywhere IP6 seriously? [16:48] damn keybord [17:02] fluffy: no.. [17:02] ipv6 is generally just a headache [17:03] Maybe I should start to patent ipv8 :) [17:10] sf: I hope, that you won't send an imp this time. [17:10] sf: It is more fun, if I'd have a good opponent. Even, if I loose :) [17:11] bvowk: 'in accordance to the rules' [17:11] what's a better term for accordance (if there's any) [17:12] "according to the rules" [17:12] hey, wait [17:12] my name isn't bvowk [17:12] in line with the rules? staying within the limits defined by the rules? according to the rules? [17:12] staying within the boundaries defined by the limits, you can find in the rules? [17:13] "do not cheat"? [17:13] "whether the placement of a [17:13] stone is in accordance to the rules." [17:13] in line sounds good [17:13] whether the placement of the stone is a valid legal move? [17:13] according to the rules [17:14] thanks! :) [17:14] s/valid legal/valid, legal/ [17:14] not invalid [17:15] What's this about placing stones? [17:15] MSG: Quit: Leaving [17:15] * bvowk places a stone squarely against fluffys forehead from across the room [17:15] sounds like he's playing go :) [17:16] I want to write a go webapp [17:16] :) [17:16] * Fluffy sells each stone for 1 Canadian Dollar, because they were touch by the holy hands of bvowk [17:16] *touched [17:16] holly? [17:16] hum :P [17:16] -l [17:17] fiveop: Yes, bvowk is the God of all Redcode evolvers ;-) [17:17] Every time you kick him, you might be granted a wish [17:18] * Fluffy kicks bvowk [17:18] fluffy: thats not actually true, the dave hillis, microgp, and loren are all better at evolving than me. [17:18] fluffy: your wish is: evisceration! it is done! [17:18] It is not about being good at evolving. It is all about being God! [17:18] he is humble too [17:18] * Fluffy kicks bvowk [17:19] * fiveop kicks fluffy [17:19] * Fluffy kicks fiveop [17:20]